u
A,“ ,.._
k5 . £523

h‘

3':-
J-L‘c ‘

c ‘1
‘. , ‘
7‘19-
H‘c
.

L,

‘ :1

'2. .
1::

C 1‘

‘ ‘uJT. A
1 f?

1...‘
Lu .
11‘ (323‘ “

5...“

~:_ {.w
225

. ' W
.1..

H.151 .
1N7}!
, .\
l a

:41.“
_<.,- ..
‘...
i. ‘1.

 

L‘L

‘.L
{‘29. i an
r
.P l -
. .
.

‘ ”33;“?
193%
4 ‘VY. -

$.39

¢
‘
s

Q
3&9.

M .

‘ ‘27 "v.5:
{3-23. V
#:3sz

. 3:?“

‘3‘5‘4’7'?

. y

; ,3:

3:. H-
‘Wfixm .

J. m
"£91122” .J
W2:- 3

fim'ioméi
- .1“; .3

. \ .

«a... #5?» :«fm‘i:
tut

cv

Effiwr
.r‘ if???"
3. ‘3'

 

. UNIVERSITY L'IBRARE
Th: ‘

|LL|LLLL|L|LLLL LLLLLLL LLL LLL

3 1293 01029 879

LLLLL

 

 

 

 

 

 

               

 

 

 

 

 

 

 

This is to certify that the

dissertation entitled

RICINOLEIC ACID BIOSYNTHESIS IN RICINUS COMMUNIS

 

presented by

Frank Joost van de Loo

has been accepted towards fulfillment
ofthe requirements for

 

Ph-D degree in Botany & Elant Pathology

    

Major professor

Date 3 /\‘ /q7

MSUis (m Affirmatiw A 1‘11”!) w'L'qua/ Opportunity lnsmurion 012771

 

 

 

 

 

 

LIBRARY
Mlchlgan State
University

 

 

 

PLACE ll RETURN BOXto monthl- duckomm your record.

1'0 AVOID FINES Mum on or baton dd. duo.

L DATE DUE DATE DUE DATE DUE

 

IAN O 6 201‘}

 

 

 

 

L
L

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU I. An Affirmatlvo Actionled Oppommny Inflation
. WWI

RICLNOLEIC ACID BIOS YNTHESIS IN RICINUS COMMUNIS
By

Frank Joost van de Loo

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Botany and Plant Pathology
and
Department of Energy Plant Research Laboratory

1993

ABSTRACT
memouarc ACID BIOSYNTHESIS IN RICINUS COMMUNIS
By

Frank Joost van de Loo

Ricinoleic acid, the hydroxylated fatty acid synthesized in developing seeds of
castor (Ricinus communis L.) is the industrially-important constituent of castor oil.
Little biochemical and no molecular information was available concerning the nature
of the oleate-lZ-hydroxylase responsible for this unusual fatty acid. Characterisation
of this enzyme and isolation of the encoding gene(s) were the goal of this work,
toward the eventual aim of engineering ricinoleic acid production in an alternative
oilseed.

The induction of hydroxylase activity concomitant with differentiation of the
cellular endosperm during seed ontogeny was characterised. Anti-cytochrome b5
antibodies were used to demonstrate that cytochrome b5 is the electron donor to the
hydroxylase. That molecular oxygen is the source of the hydroxyl oxygen was
demonstrated using "‘02 and mass spectrometry.

Intractability of the hydroxylase to purification prompted genetic and molecular
approaches to cloning the hydroxylase gene. Of 5300 yeast clones expressing
individual cDNAs from developing castor endosperm, none showed altered fatty acid
composition when analysed by gas chromatography. Hypothesizing that the
hydroxylase is homologous to microsomal fatty acid desaturases, a clone of the

Brassica napus Fad3 desaturase was used to isolate hybridising castor cDNAs. A

gene isolated by this approach was identified as a different desaturase (Fad7). Fad7
was also isolated from castor seed mRNA by PCR using a conserved desaturase
sequence motif, but no putative hydroxylase clone was identified by these approaches.

Expressed Sequence Tags were generated from 468 moderately-abundant seed-
specific cDNA clones of developing castor endosperm. Of these, 213 could be
identified (Blastx score 2 80), including the identification (by homology to
desaturases) of two putative clones encoding the hydroxylase. An analysis of the
possible role of the identified clones in castor md metabolism is presented. The
putative hydroxylase gene (pFL2) showed features expected of this gene: strong, seed-
specific expression, and ca. 37% amino acid sequence identity with membrane-bound
desaturase genes. Expression of pFL2 in transgenic yeast and plants did not result in
detectable accumulation of ricinoleic acid, nor other changes in fatty acid
composition.

Based on this work, experiments can be designed to test the identity of pFLZ

as the oleate-lZ-hydroxylase gene.

"To seek it with thimbles, to seek it with care;
To pursue it with forks and hope;

To threaten its life with a railway-share;
To charm it with smiles and soap!

"For the Snark’s a peculiar creature, that won’t
Be caught in a commonplace way.

Do all that you know, and try all that you don’t:
Not a chance must be wasted today!”

iv

ACKNOWLEDGEMENTS

I warmly thank Dr. Chris Somerville for sharing his enthusiasm, knowledge
and wisdom, and for his support and confidence in my abilities. I also thank Drs
John Ohlrogge, Pam Green and Mike Thomash‘ow for their guidance.

I thank all members of the lab for their help and assistance at every turn, as
well as other members of the Plant Research Laboratory.

I thank Dr Brian Fox for helpful discussions on oxygenase chemistry. Svieta
Ndibongo, and Susan Lootens of the PRL Plant Biochemistry Facility, gave patient
and competent technical assistance to the work described in chapter 5.

I am particularly grateful to Vlada Orbovié and Ed Cahoon, Scott Peck and
Antje Heese-Peck, Gary Vercruysse and Catherine Tamareille for their hospitality,
not to mention their friendship.

Finally I thank the people of the United States of America for their
commitment to the research enterprise and their generosity to those such as me who

profit so splendidly from our involvement in it, and from our sojourn on this soil.

TABLE OF CONTENTS

LIST OF TABLES
LIST OF FIGURES
ABBREVIATIONS
CHAPTER 1: INTRODUCTION
UNUSUAL FATTY ACIDS OF PLANTS

TAXONOMIC RELATIONSHIPS AMONG PLANTS
ACCUMULATING UNUSUAL FATTY ACIDS

TISSUE LOCALISATION OF UNUSUAL FATTY ACIDS

POTENTIAL PHYSIOLOGICAL ROLES OF UNUSUAL FATTY
ACIDS IN PLANTS

GENERAL BIOCHEMISTRY OF UNUSUAL FATTY ACID
ACCUMULATION IN PLANTS

CASTOR (RICINUS COMMUNIS) As A SOURCE OF RICINOLEIC
ACID

RESEARCH GOAL
REFERENCES

CHAPTER 2: BIOCHEMICAL STUDIES
ABSTRACT

INTRODUCTION

vi

Page

xiii

xvii

13

16
17
18
26
26

27

MATERIALS AND METHODS

DEVELOPMENTAL ANALYSIS OF RICINOLEIC ACID
BIOSYNTHESIS

ELECTRON TRANSFER TO OLEATE-lZ-HYDROXYLASE

PATIAL PURIFICATION ATTEMPT S

POLYCLONAL ANTIBODIES RAISED AGAINST CRUDE
ENZYME EXTRACTS

DISCUSSION AND CONCLUSIONS

REFERENCES

CHAPTER 3: GENETIC EXPERIMENTS

ABSTRACT

INTRODUCTION

MATERIALS AND METHODS
EXPERIMENTAL DESIGN
PRELIMINARY EXPERIMENTS
RESULTS

DISCUSSION AND CONCLUSIONS

REFERENCES

CHAPTER 4: EXPERIMENTS USING DESATURASE GENES IN
ATTEMPTS TO ISOLATE THE HYDROXYLASE GENE

ABSTRACT

vii

3O

35

36

43

45

50

51

55

55

56

61

67

68

71

76

78

82

82

INTRODUCTION
MATERIALS AND METHODS
RESULTS AND DISCUSSION
CONCLUSIONS
REFERENCES

CHAPTER 5: LARGE SCALE SEQUENCING OF SEED SPECIFIC
CASTOR CLONES

ABSTRACT

INTRODUCTION
MATERIALS AND METHODS
RESULTS AND DISCUSSION
CONCLUSIONS
REFERENCES

CHAPTER 6: ANALYSIS OF PUTATIVE OLEATE-lZ-HYDROXYLASE
CLONES

ABSTRACT

INTRODUCTION
MATERIALS AND METHODS
RESULTS AND DISCUSSION
CONCLUSIONS

REFERENCES

viii

83
89
93
104

105

109
109
109
112
116
167

167

175
175
176
177
186
218

218

CHAPTER 7: CONCLUSIONS AND PERSPECTIVES
REFERENCES

APPENDIX: MASS SPECTROMETRIC STUDIES: POSITION OF
HYDROXYLATION AND SOURCE OF HYDROXYL OXYGEN

INTRODUCTION
EXPERIMENTAL DESIGN
MATERIALS AND METHODS
RESULTS AND DISCUSSION

REFERENCES

ix

221

231

232

232

234

235

239

242

LIST OF TABLES

Table 1. Minimum summary of the kinds of fatty acids found in plants. The
numbers were compiled by counting the numbers of distinct fatty acids
described in compilations. "2

Table 2. Diversity of unusual plant fatty acids'.

Table 3. Oleate-lZ—hydroxylase activity measured in extracts of endosperm
isolated from castor seeds at different stages of development.

Table 4. Oleate—lZ—hydroxylase activity (pmol ricinoleic acid mg“1 protein
min") following incubation of enzyme (325 pg protein) with immunoglobulins
(IgG; 16.6 pg) from rabbits immunised against castor proteins of different
molecular weight (MW) ranges. In experiment 1, IgG were incubated with
enzyme for 45 min on ice, before assay. In experiment 2, IgG were incubated
with enzyme (2 h, on ice) and then precipitated with Stapphylococcus aureus
(S.A.) cells, before assay of the supernatant.

Table 5. B-galactosidase activities of cultures of CGY2557 harbouring a
random clone (pYESr) or pCGSZ 86, supplemented with either glucose or
galactose, at the time of sugar addition (a) and on subsequent days. All values
are means of duplicate assays of 5 cultures (n= 10), except where indicated
(‘n=8).

Table 6. Identification (blastx database search result) of sequences cloned by
PCR amplification using a sequence conserved among desaturases.

Table 7. Redundant sequences obtained in batch 1 sequences, used to make a
probe for screening of batch 2. Frequency of these sequences in batch 2 (after
the screen) is compared with batch 1 (before the screen) to indicate
effectiveness of the screen.

Table 8. Clones for which the most similar match in the databases was from
a prokaryote. The clones are listed by the suffix of their clone number
(pCR8265 etc.). Information for each clone is structured as follows: blastx

Page

37

49

70

102

117

score, (probability of random alignment), database (G: GenBank; S:

SwissProt; P: PIR), accession number, protein description, organism. This is
followed by a discussion of the possible function of each clone. Information

for these discussions is drawn from annotations to the database entries, as well

as standard textbooks, and original references where cited. 130

Table 9. Clones for which the most similar match in the databases was from

a eukaryote other than a higher plant. Ribosomal proteins are not included.

The clones are listed by the suffix of their clone number (pCRSZ66 etc.).

Information for each clone is structured as follows: blastx score, (probability

of random alignment), database (G: GenBank; S: SwissProt; P: PIR),

accession number, protein description, organism. This is followed by a

discussion of the possible function of each clone. Information for these

discussions is drawn from annotations to the database entries, as well as

standard textbooks, and original references where cited. 136

Table 10. Clones homologous to ribosomal proteins. Those having homology

to ribosomal proteins for which higher plant sequences are already available,

and those having homology to ribosomal proteins apparently previously

unsequenced from higher plants (see text), are listed separately. The clones

are listed by the suffix of their clone number (pCRS4l6 etc.). Information for

each clone is structured as follows: blastx score, (probability of random

alignment), database (G: GenBank; S: SwissProt; P: PIR), accession number,

protein description, organism. 143

Table 11. Clones for which the most similar match in the databases was from

a higher plant, other than ribosomal proteins. Clones are grouped according

to the general area of metabolism or cell function with which they may be

involved (see text for discussion). The clones are listed by the suffix of their

clone number (pCRS792 etc.). Information for each clone is structured as

follows: blastx score, (probability of random alignment), database (G:

GenBank; S: SwissProt; P: PIR) , accession number, protein description,

organism. 145

Table 12. Database (dbEST, GenBank) accession numbers of sequences
obtained in this study, corresponding to clone numbers used in the text. 161

Table 13. Sequence similarity at the nucleotide (NT) and amino acid (AA)
levels between pFL2 and membrane-bound desaturase genes. Nucleotide

xi

sequence of the clone ELI72 was not available for comparison, and the
deduced amino acid sequence used here is considered preliminary data only.

Table 14. Relative content (%) of the abundant fatty acids in transgenic carrot
roots. Fatty acid methyl ester abbreviations: 16:0 hexadecenoic acid, 18:2
octadecadienoic acid, 18:3 octadecatrienoic acid, 18:1 octadecenoic acid (two
isomers were resolved but not identified; carrot contains petroselenic acid, A6-
18:1, in addition to oleic acid, A9-18zl), 18:0 stearic acid.

Table 15. Relative content (%) of the abundant fatty acids in transgenic carrot
roots and tobacco shoots. Fatty acid methyl ester abbreviations: 16:0
hexadecenoic acid, 18:0 stearic acid, 18:1 octadecenoic acid, 18:2
octadecadienoic acid, 18:3 octadecatrienoic acid.

Table 16. Relative content (%) of the abundant fatty acids in transgenic
tobacco shoots (S), calli generated on leaf explants (C), and calli obtained by
NT-l cell transformation (NT). Fatty acid methyl ester abbreviations as for
Table 15.

Table 17. Formation of ricinoleic acid from oleoyl-CoA by cell-free extracts
of developing castor endosperm. Extracts were incubated with 18-trideutero-
oleoyl-CoA in the presence of air or 1802. The position at which oleic acid
was hydroxylated as well as the source of the hydroxyl oxygen was determined
by GC-MS analysis of fragmentation ions (m/z 187 to 192) derived from the
recovered trimethylsilyl ether of methyl ricinoleate. Each treatment was
replicated two times and each replicate was analyzed three times by GC-MS.

xii

192

213

215

217

241

LIST OF FIGURES

Page

Figure 1. Oleate—12-hydroxylase activity as a function of protein assayed. 39

Figure 2. Stability of oleate-lZ-hydroxylase activity of enzyme incubated on
ice. 40

Figure 3. Oleate-l2-hydroxylase activity as a function of anti-cytochrome b5
immunoglobulin (IgG) added. Immunoglobulins were from pre-immune mouse

ascites fluid (closed circles) or from a mouse that had been immunized with
cauliflower cytochrome b, (open circles). 41

Figure 4. Western blot strips stained with polyclonal antibodies from rabbits
immunised with different molecular weight ranges (shown at bottom, kD) of
developing castor md proteins. The same protein preparation was separated

on a preparative SDS-PAGE gel, and transferred to give the blot shown. Each

strip was developed separately with either preimmune (p) or immune (i)

serum. The migration of proteins of given molecular weight (kD) is shown at

left. 47

Figure 5. The vector pYE82.0 (Invitrogen) used for expression of cDN A in
yeast. Castor seed cDN A was ligated at the BstXI sites. 62

Figure 6. Gas chromatogram of fatty acid methyl esters from sample 145

(cultures 8H5 through 8H9). An attenuation was used which favoured visual
inspection for small peaks. Vertical axis: units of signal strength (flame

ionization detector); horizontal axis: retention time (min after injection).

Elution times of abundant yeast FAMEs in this chromatogram were: 16:0

4.960 min, 16:1 5.342 min, 18:0 6.414 min, 18:1 6.846 min. 73

Figure 7. Gas chromatogram of fatty acid methyl esters from sample 145

(cultures 8H5 through 8H9), spiked with authentic methyl-ricinoleate. Note

the elution of TMS-methyl ricinoleate at 8.610 min. (Attenuation and axes as

for Figure 6). 74

xiii

Figure 8. Gas chromatogram of fatty acid methyl esters from sample 143
(cultures 9A5 through 9A9). Note the elution of a small peak at 8.540 min, at
the same elution time as authentic TMS-methyl ricinoleate (see Figure 7).

This peak was not seen when cultures 9A5 through 9A9 were re—grown and
analysed individually. (Attenuation and axes as for Figure 6).

Figure 9. A Southern blot of castor genomic DNA digested with restriction
enzymes EcoRI (E) or HindIII (H), was hybridised at moderately low
stringency (52°C, 4 X SET) with the 32P-labelled Brassica napus fad3 cDNA.
Migration of DNA standards (kb) is shown to the left.

Figure 10. Sequence similarity between pFLl (centre) and other membrane-
bound desaturases at the nucleotide (above) and amino acid (below) levels.
The pFLl coding sequence and untranslated regions are shown as solid and
checked bars, respectively. The percent sequence similarity between pFLl
and the other genes is shown in the shaded regions over which the similarity
was averaged.

Figure 11. Comparison of the deduced amino acid sequences of four
membrane bound desaturase genes: RCFAD7 Ricinus communis Fad7 (pFLl ,
above); ATFAD7 Arabidopsis thaliana Fad7;9 BNFAD3 Brassica napus Fad3;4
SDESA Synechococcus DesA.26 The conserved GHDCGH and HXXHH
motifs are shaded.

Figure 12. Autoradiogram of plates 28-36 probed with 32P-labelled cDNA
from developing seeds. The positions of the wells in the original 96—well
plates is indicated (borders), and the position of clones from each 96-well plate
relative to other 96-well plates is indicated in the box.

Figure 13. Autoradiogram of plates 28-36 probed with 32P-labelled cDNA
from developing leaves. The positions of the wells in the original 96—well
plates is indicated (borders), and the position of clones from each 96-well plate
relative to other 96-well plates is indicated in the box.

Figure 14. Autoradiogram of plates 28-36 probed with 32P-labelled DNA
from redundant clones sequenced in batch 1. The positions of the wells in the
original 96-well plates is indicated (borders), and the position of clones from
each 96-well plate relative to other 96-well plates is indicated in the box.

xiv

75

94

97

120

121

122

Figure 15. Fates of clones selected for sequencing by differential screening.

Figure 16. The vector SLI4K1 constructed by J. Jones, and used for
preparation of constructs for expression of pFL2 in plants.

Figure 17. Alignment of the 3’ sequences of various castor cDNA clones
isolated with pCRS677. Note that the linker sequences (CTCTAAAG) have
not been removed. Similarity between 3cvii and the other clones begins at
position 313 of this alignment.

Figure 18. DNA and deduced amino acid sequence of the clone pFL2.
Motifs conserved among membrane-bound desaturases are shaded.

Figure 19. Alignment of the deduced amino acid sequence of pFL2 with
pFLl (Fad7 from castor: chapter 4), and Brassica napus Fad3.12

Figure 20. Northern blot analysis of pFL2 expression in castor. A 32P-
labelled probe corresponding to ~ 700 bp of the 3’ end of clone pFL2 was
hybridised to poly(A)+ RNA from leaves (L) and developing seeds (S) of
castor. Panel A: the blot was exposed to film for 30 min. The migration of
RNA standards (kb) is shown to the right. Panel B: the same blot was
exposed for 16 h. Panel C: the same blot was hybridised to a 32P-labelled
probe made from the Colletotrichum graminicola B-tubulin gene TUBZ."

Figure 21. A Southern blot of genomic DNA from Arabidopsis thaliana and
castor (Ricinus communis) digested with restriction enzymes EcoRI (E),
BamHI (B), or HindIII (H), was hybridised at high stringency (65°C) with the
32P—labelled insert of clone pFL2. Migration of DNA standards (kb) is shown
to the left.

Figure 22. The Southern blot shown in Figure 21 was hybridised again with
the pFL2 probe under low stringency (52°C). (See legend to Figure 21 for
details). '

Figure 23. Typical gas chromatogram (above) of yeast fatty acid methyl
esters (CGY2557 harbouring pFL2, grown on glucose medium), and
chromatogram of methyl-ricinoleate standard (below). Detector signal is
plotted against retention time (min).

XV

127

183

188

191

193

196

198

199

202

figure 24. Typical gas chromatogram (experiment 1) of transformed carrot
root TMS-derivatised fatty acid methyl esters (root 20-1). Detector signal is
plotted against retention time (min). Peaks identified by co-chromatography
with standards, and by comparison of mass spectra to a mass-spectral data
library: 17.38 min, 16:0; 20.88 min, 18:2; 21.04 min and 21.19 min, 18:1;
21.69 min, 18:0.

Figure 25. Gas chromatogram of trimethylsilyloxy-methyl-ricinoleate
(experiment 1). Detector signal is plotted against retention time (min).
Retention time of standard: 25.1 min.

Figure 26. Mass-spectrum of ions derived from material eluting from the
column at 25.165 min, in the chromatogram shown in Figure 24.

Figure 27. Mass-spectrum of trimethylsilyloxy-methyl-ricinoleate, taken from
the chromatogram shown in Figure 25 at 25.120 min.

xvi

205

207

209

211

18:1

IgG
SDS-PAGE
SDS
FAME

GC

TMS

EI

MS

ABBREVIATIONS

Standard fatty acid abbreviations are used. These are of the format
(number of carbon atoms in linear acyl chain):(number of double
bonds). Position of unsaturation may be referred to by numbering from
the carboxyl end (e. g. A9,12—18:2 for linoleic acid), or from the methyl
end (e.g w-9,w—6-18:2).

Immunoglobulin

Sodium dodecyl sulphate-polyacrylamide gel electrophoresis

Sodium dodecyl sulphate

Fatty acid methyl ester

Gas chromatograph

Trimethylsilyloxy

Electron impact

Mass spectrometry

xvii

CHAPTER 1

INTRODUCTION

UNUSUAL FATTY ACIDS OF PLANTS

Extensive surveys of the fatty acid composition of seed oils from different
species of higher plants have resulted in the identification of more than 210 naturally
occuring fatty acids which can be broadly classified into one of eighteen structural
classes (Table 1). The classes are defined by the number and arrangement of double
or triple bonds and various functional groups, such as hydroxyls, ketones, epoxys,
cyclopentenyl or cyclopropyl groups, furans or halogens. This level of structural
diversity is similar to that of some of the least diverse families of plant secondary
metabolites, a class of compounds which have been estimated to contain as many as
100,000 different structures.3 A summary of the range of structures to be found in
plant fatty acids, and a small number of examples of representative fatty acids, is
presented in Table 2. Extensive lists of the amounts and sources of plant fatty acids
are available in earlier reviews."2'7'“

The most commonly occurring fatty acids, which may occur in both membrane

and storage lipids, are a small family of 16- and 18—carbon fatty acids which may

have from zero to three, methylene-interrupted, cis unsaturations.

2

Table 1. Minimum summary of the kinds of fatty acids found in plants. The numbers
were compiled by counting the numbers of distinct fatty acids described in
compilations. 1'2

 

 

Type Number of Structures
saturated l4
monounsaturated 1 1
diunsaturated‘ 9
triunsaturated‘ 9
tetraunsaturated‘I 6
pentaunsaturated‘I 3
hexaunsaturated‘ 1
nonconj ugated ethylenic” 32
conjugated ethylenic 15
acetylenic 26
monohydroxy 33
polyhydroxy 12
keto 5
epoxy 9
cyclopentenyl 15
cyclopropanoid 6
furanoid 2
halogenated 2

 

‘all double bonds are methylene-interrupted cis-isomers
”configurations other than methylene-interrupted cis-double bonds

 

338.83. 5.83%....

5.3.2.... 5.... .38 35:33....— eca 3.38.6.3. 3.3.8.2.»: 33.5....
.5... .38 35:25 ..o..o8....=......m 2:5...”53. Eziew d»... 53...?“
fiesta do... 33...... 893...... .2 32:8. Es... ..c 393....» a. «nu...»

38389.8(. 3535.... 82.3.3355 Es...
.38 3:62.535... new 3.38.5...95 SSS 235:0 EC... 38 3......«0

.88....2853... .93 33.2.3.5... ..:.. 32.3.... 6,352.3: 3.5.5

3... 4:55.... 6982.... 5.3 6383885 9... :. fizz—=3 .3039 a. 38
.a.u.u........o .....:.x..... as... 2.... ..83«.£E.::E.m £355.? 54:555....
a..:.:..~ z. 3.3.3 5.3 8 cm... .3... 6....» 5.5.3 .3 .93 .38 3:29..
68$: e.........u...ew .3... 5.8.5.3... 5.3.2.52 c. .98 6.8 3.3..88
..m.o. 8983.38 .88 898...... o... c. 58: 8.38 32.995. bo>

.88853... .9... 3.2.3... 8:... 3.3.5.2. a... .0388...

....=....§.....u 3.3.3.... z: .25.. 8.5. can .8o88.< 685...... 380. :3
555.5 .5... 32.5.... a... ”.38 3.2.9....33... 5.389556 6387...»:
.3 3.8.3.35 3.3.»... 2.... 885...... :. 2.87.5.5 9...... 3 3......

:82... 8.3.2.... 33.... .2. .83.. .853.

a.
5.882.312... (I\/\/\(Is‘ ( ( «a
9.
382.931.. (Ix/2(4\( (a:
a!
an“... .8. t .. .(Ao 3:32 .N
iii
.55. u .. .3. 581...... o
a. 3593.93.85 ..
.x....a........a§ o 53.2.5833 .m
2
O
A“. _ 3. is a!
O

.3. _ 89.!

9
880918. \I\/\/\I\I\/\/\I\I\Ao. Ear—U NE.- bo> .N

9.
8 2. 9.3::- \/\/\I\/.\I\I\A

32. 2|: \l\l\/\I\I\Ac

a. 5.2.9 25.—55. ..
35...... 2(1)}. 3.9.3 .355 .<

9.22:... E.- oE-z 2.2.2.. .2825

.238 b8. .8... .832... .8 3.9.35 .N 03¢...

.... n...9...........9 ............9 9... 3.9... .35.... _.3 3.... 3 99...... .a.9.9........9
.......,.:...:. 1...... 9... n. :2 .229 £9.99... 5...: 25...... E 599.. .29... 9.9...
.: 9.....m ..9..99..97.£:... ................ ........:........ Sc... 29... 3.23:... .z...
.,.9..99....»9....<. .._.... ......9....... i...» .,...................a. 2...... 2...... 9.9......9.....n.
..9..99..........._..m_. ................. .....,..._..< .2... 2.7.39 :5... 3...... 9.92229...

.92....

.29.. 9.92.9 33.5.9... .3..........9 .93... 3.3.; .93.... 3......35 :99.
2:... 33...... 3.... $9393.92 1.... ..........................i 5.9. 3... 199.. 2:3
.3 a.9>9. 32...... 2. .2. 39.. 5...... ............9 .2832... a 28 3:39.34...
........9 ..9..........9 ..m.9 .9..99.......< 5.... 898...... 9... .3 3.92.32... .39..
3.3.9259. ..9..99.................. .5... ... ...........a...... 3................ :5... 9n .ucm

.9..99..........m 3.... 939938.: 9... E ..9.....
.599: £39.. 3:91:99: 3.9.59 9...... 3.5.5 ..9..99...9.$... i... 3......
5:... 39.. 9...»..9..9.9 E... .9..99....:.........m. ......... .................. 5:... 1.9.. 33......

8.38

.34.. 9:9... 3......9?:9::= .3... ”529.» ..... 5.2.9 .5... 29.. 33.5.3. .3
53.3.1393 b. 1952.... £39.. 2 .5. 9:9... 19.99.72... .x.......£...=. .925...
.28 9....23 ..u.9. 3...... 92...... 2...... «c.5356 .3... 3.29:9» .m9..........9
.93... 32...; ..9..99a.€.:...:m. 38% 3......93. ..... as... :. a.....9.9........U

.9259. N 033.

3

.I.
.929: a 3 .1: 9.1.3.3.
a.
c C
a.
z: a
.6. . .9!— 9.8.4.839.
.4.- .. .‘3 391.3:
.

.9: .2. 8.892.... c
3.
o

a!
a
.83.. 35.33....
a;
a!

niacin-31"! 3.
3.2.3.1.. \(l\l\/\I\/\I\A

.1;

3.3.5. 2.3.3.3...

3—3-8nd33‘3l

.335»: ..
39:39:}. .0

3.3.3.... .9325 .n

3:92:99... ...

39...»...3..U .n

_h.

C
9.
6
"5.3;: 9.
43855535 Etckzui ctcEEeG 5 3.9a 3.3. 3...: .23 73.3: . .2555: c o
5...... 2.! “I...
23. .v

5:: gaseous—“.2355Uv 95.... 2:.Ecezafab 5 v.33 SS— 505: :33 Y”— a
-xv ”3395:”. 63%: c..==.....: z: 8.3:: .3 Bug bi. 5.3:. a 6233.:

O on.

O

at... «2.1.2.553.

9:8" 2.5.2... 9.
/\/\/\Amr\/\/\/\Ao 2:55.”. .m

.3382:=E=w 6.3.5.533 cpitv

_.:8.:w_»ua_= .83. .93.: 3 26:25.3 5?... a 53.9.)..— .4985
-26. use 3535559 aziiuiuizu EESSHM Es... ~“w_->x.x_u-~_.o

9.

.8.....E.~:.~=x.!! e H...
49:83:25.

5.5:. c5356 :5: Nun—$5.5 2.2 we. 29833.339 3.321%“.
3.35:: Eegzizm .38qu35 5.252355. 5.55.; 5::
3.5253... \/\/\/\/wv/\/\/\/\AH... xix—m .N

2.3 o::Eo> guano—".935 QESEE :euiekcfi. Es: aux—$5.9 2.0

.9253 N 2an

.159... 92 :5. .29.. .4 $5....b. ._ ”9:53.99: I a ”9:53.29 9 “9593.29 2:... . "9:52:29 :9 9 3......23212 .
599......» 9.9.: 5:... 922.32. .5. 9.9:... :93» 9.... 39:39.9. 553:. .N 15. _ 39:93.3— .5... 3.9:... 39:55.9 ..
529.. .95.. 12.55. _: 99:95.99: 9... .5 9w:.9>:9 939.533.559 a 93» 3.9. 9... 5 39:9 £59391 .35.
099.53 :9... :5. 529: 5:... 9... 5:: _: 3.55:3 .3 .955... :53 a 5.3 59.5.9.5» 29.. aze :33 .3 5.393: 95 .2. 5:39.35. 5. 92m :. 39.59.... a. 93:. 2.2. ‘

../\/\/\/\/\/\/\/\AH...
33 35:6. .2
a.
..9..99:_.:9._:._9_C. 4:222...
..:..::......:. .25.:955...5 .2. 4.999. 5 1:5: 929.. 95. 59> .. .55... .n.
o A!
29.59. .c 59539.. 9... E 39...: .2352... =: 395:...2—9 E 929.. £35: Q/\/\/\/\/\Ao
_x_.9.:95._9xu 49:32:55.2”: 55:35 553.55%: .2: 395......339 21531.3. \/\/\/\/V/\/\/\/\Aa.
.5... 29.. 2:39:53 :5. 29:99:55.3. 559...... 3.5.1:... 5:... w:
39: 9__:9.9.n:.1x.=s 2939933955. $9....Ex 3:29.95. 5:... 39.. 9.1.9.95 23.3. \/\/\/\/V/\/\/\/\Aa 9:19 .N
5.9.9.... 5 55559 9...... ._9::. 9.... :5. 5...... 5 9...: 9... :.
n39: =_:._9,19._9=.:a 5493995513....:um. .25.... .=..=..\:.==< .3 3:25: o: 3.3. /.\/\/\/\/\/\/\/\Ae
955.2...— ._= 99:3: 93.3.2.9» .95. 933.929.. 5 9.29.. 19:95:: .95.. \L\«...
.23 92.920. .A9a99a_a...i:m. .555. 5.3.0 .2. 2.93 1.3 .59... .5... 3.1.3.3.... M.
39.. 9:»: 15. 3395.5 3,595.53...» 9599.2... .3 .52... 5:... 3.9.. 9:992 3.1.1: 3...... \JXo 5.2.9 59:95:: .—

353 d

.8980. a 2.3

7

All members of the family are descended from the fully saturated species as the result
of a series of sequential desaturations which begin at the A9 carbon and progress in
the direction of the A15 carbon.l2 Fatty acids which cannot be described by this
simple algorithm are generally considered ”unusual” even though several, such as
lauric (12:0), erucic (22:1) and ricinoleic (12-0H, 18: 1) are of significant commercial
importance. The biosynthesis of ricinoleic acid in castor (Ricinus communis) seed is
the focus of this work.

Much of the research to date concerning unusual plant fatty acids has been
focused on the identification of new structures or cataloguing the composition of fatty
acids found in various plant species. Relatively little is known about the mechanisms
responsible for the synthesis and accumulation of unusual fatty acids, or of their
significance to the fitness of the plants which accumulate them. This chapter gives a
general review of unusual fatty acids in plants and a review of the data pertaining to
biosynthesis of ricinoleic acid available when this work was commenced. Finally the

goals of this work are summarised.

TAXONOMIC RELATIONSHIPS AMONG PLANTS ACCUMULATING

UNUSUAL FATTY ACIDS

The taxonomic relationships between plants having similar or identical kinds of
unusual fatty acids have been examined. "7 In some cases, particular fatty acids occur

mostly or solely in related taxa. For example, the cyclopentenyl fatty acids have been

8

found only in the family Flacourtiaceae, although the presence of
cyclopentenylglycine, the biosynthetic precursor of the cyclopentenyl fatty acids, in
the Passifloraceae and Tumeraceae suggests that these acids may also be found in
these other families of the order Violales.”l4 Petroselinic acid is most commonly
found in the related families Apiaceae, Araliaceae and Garryaceae, but has also been
observed in unrelated families.“ In other cases there does not appear to be a direct
link between taxonomic relationships and the occurrence of unusual fatty acids. For
example, lauric acid is prominent in both the unrelated families Lauraceae and
Arecaceae. Similarly, ricinoleic acid has now been identified in 12 genera from 10
families. “”7 However in the large genus Linum, ricinoleic acid was found in the seed
oil of all species tested from the section Syllinum but in no species from other sections
of the genus." If any conclusion can be drawn from the taxonomic distribution of
unusual fatty acids, it would seem to be that the ability to synthesize some unusual
fatty acids appears to have evolved several times independently, while for others it

may have evolved only once.

TISSUE LOCALISATION OF UNUSUAL FATTY ACIDS

A feature of unusual fatty acids is that they are generally confined to md
triacylglycerols.2‘"33 Analyses of vegetative tissues have generated few reports of
unusual fatty acids, other than those occurring in the cuticle. A small number of

exceptions exist in which unusual fatty acids are found in tissues other than the seed.

9

The A‘—desaturated acids 7-linolenic (486332-183) and octadecatetraenoic (A°'9"2"5-
18:4) are found in leaves and seeds of some plants of the Boraginaceae.“'35 The
cyclopropenoid fatty acids of the order Malvales are not restricted to the seeds.Ems
The cyclopentenyl fatty acids of F lacourtiaceae seed oils are also found in smaller
proportions in the phospholipids and glycolipids of various tissues.39 Acetylenic fatty
acids are found in root, stem and leaf in the Santalaceae.”36 The small
triacylglycerol fraction of rape leaves and siliques contains erucic acid."5 Similarly,
petroselinic acid is found in the pericarp (a maternal tissue) as well as within the
seeds of ivy (Hedera helix).36 Latex of the rubber tree, Hevea brasiliensis, contains
triacylglycerol in which a furanoid acid (Table 2) is the major component,40 while
seeds of the same tree contain no unusual fatty acids.“1 An interesting observation is
that branched chain fatty acids accumulate in the yellow-white chloroplast-deficient
parts of leaves of plastome mutants of snapdragon and tobacco, which have reduced
levels of unsaturated fatty acids. The branched chain fatty acids are completely

absent from normal leaves and normal green parts of mosaic leaves.‘s

POTENTIAL PHYSIOLOGICAL ROLES OF UNUSUAL F A'ITY ACIDS IN

PLANTS

Since the ability to synthesize various unusual fatty acids must have evolved
independently, the common feature of confinement to seed triacylglycerol indicates

some selective constraint or functional significance. One possible function of unusual

10
fatty acids is that by being toxic or indigestible they protect the seed against

herbivory. Some unusual fatty acids may be inherently toxic, such as the acetylenic
fatty acids, or some of their metabolites described below, which have antibiotic
properties." Other unusual fatty acids are toxic upon catabolism by the herbivore,
such as the w—fluoro fatty acids of Dichapetalum toxicarium.42 One of these acids
(threo-lS-fluoro-9, lO—dihydroxystearic acid) was lethal to rats when injected
intraperitoneally at 25 mg kg", probably due to catabolism to toxic fluoroacetate.
The cyclopentenyl fatty acids were long used in the treatment of leprosy, and activity
of hydnocarpic acid (Table 2) against many Mycobacterium species has been
demonstr'ated.:"""43 These acids were also deleterious to the patients, causing a range
of side—effects. The cyclopropenoid fatty acids also appear to have biological
activities, possibly due to the accumulation in animal tissues of partial catabolites
containing the cyclopropene ring, which inhibits B-oxidation.39 Three effects of
cyclopropenoid fatty acids have been described but are poorly understood: there is
some alteration of the properties of membranes, there is an inhibition of fatty acid
desaturase activity, and there is a carcinogenic effect, or a co-carcinogenic effect with
aflatoxins.39'“"5 The tumor promoting effect of cyclopropenoid fatty acids may be
dependent upon their incorporation into membranes.“ Interestingly, malvalic and
sterculic acids inhibit the growth of seed—eating lepidopteran larvae and may be part
of the defense of cotton plants against these insects." These fatty acids may also be
effective antifungal agents, inhibiting the growth of some plant pathogenic fungi at

concentrations that appear biologically relevant.”48 The most intensely studied of the

11

unusual fatty acids from a dietary viewpoint is erucic acid, due to fears that the
consumption of rapeseed oil may be detrimental to human health. Chronic feeding of
erucic acid to experimental animals has a range of deleterious effects ,‘9 but whether
these are sufficiently severe to propose a herbivore-defense role for erucic acid in
seeds is questionable.

It is possible that the use of unusual fatty acids as a carbon source may require
adaptations of lipases or B-oxidation enzymes not present in the herbivore so that it
cannot catabolise them and remains unrewarded for eating seeds in which unusual
fatty acids make up a large component of stored carbon. The purgative properties of
castor oil, in which triricinolein is the predominant lipid, are well known. However
the significance of ricinoleic acid in this context is unclear since the castor seed seems
already well protected by the presence of toxic and allergenic proteins.

The question of why so many unusual fatty acids (Table 1) have evolved in
plants may be considered as a subset of the same question concerning the extreme
diversity of plant secondary metabolites.”0 Indeed, in some cases, unusual fatty acids
may be starting points for plant secondary metabolism. For example, crepenynic acid
(Table 2) is believed to be the precursor of most of the large range of polyacetylenes
synthesized by a small group of plant families. These polyacetylenes may be involved
in plant-plant or plant-animal interactions, and some have toxic or antibiotic
properties. However, the presence of acetylenic fatty acids in other plant families
such as the Santalaceae seems not to be accompanied by the accumulation of

polyacetylenes. ‘5

12

It is apparent that the accumulation of unusual fatty acids in lipid bodies of the
seed does not incur any selective disadvantage. By contrast, some unusual fatty acids
might be expected to disrupt membrane structure and function if incorporated into
membrane-forming lipids. There is relatively little direct biological evidence for a
disruptive effect of unusual fatty acids upon membrane structure. The few available
studies have exploited mutants of Escherichia coli and of yeast which are incapable of
fatty acid desaturation and require exogenous unsaturated fatty acids for growth. The
fatty acids supplemented to the growth medium become incorporated in the
membranes, and thus provide a technique for correlating fatty acid structure with
ability to augment saturated fatty acids in the formation of functional membranes.

Data from such experiments are somewhat contradictory, perhaps due to toxic
impurities in some of the fatty acids used. In yeastf”4 it appeared that cis-
polyunsaturates supported growth regardless of double bond positions. Cis-
monounsaturates appeared to be effective only for certain desaturation positions,
though this was due to an incompatibility with long stretches of contiguous saturation
rather than an incompatibility with certain bond positions. Trans-unsaturation was
less effective than cis. Certain hydroxy fatty acids (e. g. ricinoleic) were compatible
with growth, though they were less effective than the non-oxygenated analogues (e.g.
oleic acid). The use of some unusual fatty acids in membranes was associated with
their modification, such as the acetylation of hydroxystearic acids.” Results of
studies with E. coli are broadly similar; branched chain, brominated and trans-

unsaturated fatty acids could support growth, as could cyclopropanoid acids which are

13

in fact "usual” fatty acids for this organism‘“8 (see below). These studies show that
at least some unusual fatty acids can exist in some functional membranes, but it
remains quite possible that their inclusion in normal plant membranes would be
deleterious.

A new technique for modifying plant membrane fatty acid composition in vivo
has been described,” which might be used to address the question of the compatibility
of unusual fatty acids with membrane function. When fatty acids (15:0, 17:0, 17:1,
18: 1) were applied as their Tween esters to leaves or other organs they were
extensively incorporated into membrane lipids. Techniques also exist for the study of
physical properties of multilamellar liposomes containing unusual fatty acids,
generated in viva.”61 One such study‘50 showed that cyclopropanoid fatty acids are in
fact eminently suited to the formation of functional membranes stable over a broad
temperature range. As mentioned above, such fatty acids are normal components of

some bacterial membranes.”63

GENERAL BIOCHENIISTRY OF UNUSUAL FATTY ACID ACCUMULATION

IN PLANTS

In addition to the presence of enzymes involved directly in the synthesis of
unusual fatty acids, plants which accumulate unusual fatty acids may require other
specialized proteins. Germination of the seeds in which they occur requires that the

catabolic enzymes, such as lipases and the enzymes of B-oxidation, must be able to

14

accept the unusual fatty acids and that unusual structures formed during B—oxidation
ean be processed. I am aware of only one pertinent study. When [“C]ricinoleate was
catabolised by homogenates of germinating pea and castor seeds, B-oxidation was
blocked at the C10 level in pea, but went to completion in castor.“ The C10 product
identified in pea, 4-keto—decanoic acid, was presumably not further metabolizable,
eausing the arrest of B—oxidation at this point. A pathway was proposed for the
degradation in castor of 4-hydroxy-decanoic acid via 2-hydroxy-octanoic acid or 4-
keto-decanoic acid, 2-keto-octanoic acid, and heptanoic acid, but the operation of this
pathway has not been verified.

As mentioned above, unusual fatty acids accumulate almost exclusively in the
triacylglycerol fraction, and are in some way excluded from the polar lipids. This is
particularly intriguing since diacylglycerol is a precursor of both triacylglycerol and
polar lipid. With castor microsomes, there was some indication that the pool of
ricinoleoyl-containing polar lipid is minimised by a preference of diacylglycerol
acyltransferase for ricinoleate-containing diacylglycerols.“ A similar result was
obtained with Cuphea lanceolata microsomes66 where the diacylglycerol
acyltransferase is highly active and selective for diacylglycerol containing medium-
chain fatty acids. In addition, the lysophosphatidic acid acyltransferase was selective
for both donor and acceptor acyl groups such that didecanoyl and dioleoyl species of
phosphatidic acid accounted for the majority synthesized. The lysophosphatidic acid
acyltransferase of palm (Syagrus cocoides) microsomes also preferentially acylates

lysophosphatidic acid containing medium-chain (12:0) acyl groups with medium-chain

15
acyl-CoAs, again favouring dilauroylglycerol over mixed (12:0, 18: 1)

diacylglycerol.67 In borage (Borago ofi‘icinalis), y-linolenic acid may be efficiently
acylated to the snl- and 3712- positions of g1ycerol-3-phosphate, but in fact
accumulates preferentially at the sn3-position of triacylglycerol, and is altogether
absent from the sn-1 position. The diacylglycerol acyltransferase preferably uses 7-
18:3-CoA and this may minimise the pool size of 7-18z3-CoA, such that y-l8z3 is
concentrated in triacylglycerol.68 Data accruing from a number of other studies6M
tend to strengthen the view obtained from those described above, that targetting of
unusual fatty acids to triacylglycerol can be at least partly explained by the relative
activities and selectivities of the acyltransferases.

In some cases the discrimination of acyltransferases actually limits the unusual
fatty acid content of triacylglycerol. A well-known example is the ”66% barrier" to
erucate content in triacylglycerol of Brassica seeds. In this case the lysophosphatidic
acid acyltransferase does not accept erucoyl-CoA as an acyl donor, limiting erucate to
the 5711- and sn3-positions of triacylglycerol. This specificity is observed in several
species but the enzyme from meadowfoam (Limnanthes alba) has been recognised as
an exception.72 An understanding of enzyme specificities and the ability to exploit
those enzymes with desirable properties, such as the meadowfoam lysophosphatidic
acid acyltransferase, will be important facets in the future manipulation of oil crops
through molecular techniques. Likewise, it is important to investigate the metabolic
fate of unusual fatty acids introduced (in vitro, and eventually through molecular

techniques) into plants lacking unusual fatty acids. One such study73 showed that the

16

targetting of unusual fatty acids to triacylglycerol is maintained, but the rate of

esterification of the alien fatty acids was slow.

CASTOR (RICINUS COMMUNIS) AS A SOURCE OF RICINOLEIC ACID

Castor (Ricinus communis L.) is a minor oilseed crop, the large seeds being
harvested for their oil content ( ~ 50%) which is rich (85-90% of total fatty acids) in
the hydroxylated fatty acid, ricinoleic acid (l2D-hydroxyoctadec—cis—9-enoic acid).74

Castor is a widely variable plant of warmer climates, and is the only species of
the genus Ricinus, in the Euphorbiaceae (spurge family) (for classifications of this
family, see references 75-77). Castor probably originated in the Ethiopian-East
African region, and has been cultivated for millenia, although it remains semi-wild in
many regions, from which much of the total production of developing countries may
be harvested.78 Castor has 2n=20 chromosomes, with some claims that it is an
allopolyploid (x =5). The plants are generally monoecious, bearing racemes separated
into upper female flowers and lower male flowers. Castor is somewhat unusual
among dicots in that the endosperm, rather than the cotyledons, is the storage tissue.

Oil pressed or extracted from castor seeds has hundreds of industrial uses,
many based upon the properties endowed by the hydroxylated fatty acid. The most
important uses are production of: paints and varnishes, nylon-type synthetic polymers,
resins, lubricants, and cosmetics?8 In addition to oil, the castor seed contains the

extremely toxic protein ricin, allergenic proteins,” and the pyridine alkaloid ricinine.80

17

These constituents preclude (see reference 81) the use of the seed meal (following oil
extraction) as a livestock feed, normally an important economic aspect of oilseed
utilisation. Furthermore, with the variable nature of castor plants and a lack of
investment in breeding, castor has few favourable agronomic characteristics. For a
combination of these reasons, castor is no longer grown in the United States. In
1965, US production of castor oil was approximately 65 million pounds, from around
80 000 acres cropped. Current annual imports of about 92 million pounds cost
around $45 million.“2

The major producers of castor oil are India, Brazil and China. Supplies have
been somewhat unstable and subject to artificial pricing, and regain of a domestic
source of castor oil would be attractive.82 The production of ricinoleic acid, the
important constituent of castor oil, in an established oilseed crop (such as sunflower
or rapeseed) through genetic engineering would be a particularly effective means of

circumventing the problems of castor production.

RESEARCH GOAL

The difference between ricinoleic acid and the usual plant fatty acid, linoleic
acid, is simple, the acyl chain being hydroxylated at the C12 position instead of the
usual desaturation at this position. The goal of this work was to isolate a gene or
genes necessary for this hydroxylation in castor, with a view toward eventual

production of the hydroxylated fatty acid in a different plant.

18

REFERENCES

10.

Hitchcock, C. , and Nichols, B.W. , Plant Lipid Biochemistry, Academic
Press, London, 1971, chap. 1.

Gunstone, F.D., Harwood, LL, and Padley, F.B., The Lipid
Handbook, Chapman and Hall, London, 1986, chaps. 1 & 3.Webster,
G.L., The saga of the spurges: a review of classification and
relationships in the Euphorbiales, Bot. J. Linn. Soc. , 94, 3-46, 1987.

Wink, M. , Plant breeding: importance of plant secondary metabolites

for protection against pathogens and herbivores, Theor. Appl. Genet. ,
75, 225, 1988.

Kleiman, R., and Payne-Wahl, K.L., Fatty acid composition of seed
oils of the Meliaceae, including one genus rich in cis-vaccenic acid, J.
Am. Oil Chem. Soc., 61, 1836, 1984.

Shibahara, A., Yamamoto, K., Takeoka, M., Kinoshita, A., Kajimoto,
G., Nakayama, T., and Noda, M., Application of a GC-MS method
using deuterated fatty acids for tracing cis-vaccenic acid biosynthesis in
kaki pulp, Lipids, 24, 488, 1989.

Radunz, A. , On the function of methyl-branched chain fatty acids in
phospholipids of cell membranes in higher plants, in The Metabolism,
Structure, and Function of Plant Lipids, Stumpf, P.K., Mudd, 1.8.,
and Nes, W.D., Eds., Plenum, New York, 1987, 197.

Hitchcock, C. , Structure and distribution of plant acyl lipids, in Recent
Advances in the Chemistry and Biochemistry of Plant Lipids, Galliard,
T. and Mercer, E.I., Eds., Academic Press, London, 1975, chap. 1.

Smith, C.R., Jr., Occurrence of unusual fatty acids in plants, in
Progress in the Chemistry of Fats and Other Lipids, Vol. 11, pt 1,
Holman, R.T., Ed., 1970.

Markley, K.S. , Fatty Acids; Their Chemistry, Properties, Production,
and Uses, Vol. 1, 2nd ed., Krieger, Malabar, 1983.

Smith C.R., Jr., Unusual seed oils and their fatty acids, in Fatty Acids,
Pryde E.H., Ed., American Oil Chemists’ Society, Champaign, 1979,
chap. 2.

ll.

12.

l3.

14.

15.

l6.

17.

18.

19.

20.

21.

22.

23.

19

Hopkins, C.Y., Fatty acids with conjugated unsaturation, in Topics in
Lipid Chemistry, Vol. 3, Gunstone, F.D., Ed., John Wiley & Sons,
New York, 1972, chap. 2.

Browse, J. , and Somerville, C. , Glycerolipid synthesis: biochemistry
and regulation, Ann. Rev. Plant Physiol. Plant Mol. Biol. , 42, 467,
1991.

Spener F. , and Tober, I. , Cyclopentenylfettsauren im Brennpunkt,
Fette, Seifen, Anstrichm., 83, 401, 1981.

Tober I., and Conn, E.E., Cyclopentenylglycine, a precursor of
deidaclin in Tumera ulmifolia, Phytochemistry, 24, 1215, 1985.

Harborne, J.B. , and Turner, B.L. , Plant Chemosystematics, Academic
Press, London, 1984, chap. 8.

Brieskom, C.H. , and Kabelitz, L. , Hydroxyfettsauren aus dem cutin
des blattes von Rosmarinus officinalis, Phytochemistry, 10, 3195, 1971.

Green, A.G., The occurrence of ricinoleic acid in Linum seed oils, J.
Am. Oil Chem. Soc., 61, 939, 1984.

Garg, S.P., Sherwani, M.R.K., Arora, A., Agarwal, R., and Ahmad,
M., Ricinoleic acid in Vinca rosea seed oil, J. Oil Technol. Assoc.
India, 19, 63, 1987.

Ahmad, M.U., Husain, S.K., and Osman, S.M., Ricinoleic acid in
Phyllanthus niruri seed oil, J. Am. Oil Chem. Soc., 58, 673, 1981.

Daulatabad, C.D. , and Mirajkar, A.M. , Ricinoleic acid in Artocarpus
integrifolia seed oil, J. Am. Oil Chem. Soc., 66, 1631, 1989.

Daulatabad, C.D., Desai, V.A., Hosamani, K.M., and Jamkhandi,
A.M., Novel fatty acids in Azima tetracantha seed oil, J. Am. Oil
Ozem. Soc., 68, 978, 1991.

Daulatabad, C.D. and Hosamani, K.M., Unusual fatty acids in
Brunfelsia americana seed oil: a rich source of oil, J. Am. Oil Chem.
Soc., 68, 608, 1991.

Yunusova, S.G., Gusakova, S.D., Glushenkova, A.I., Lipids from the
seeds of Securinega suffruticosa, Iflzim. Prir. Soedin. (Tashk), 3, 277,
1986.

24.

25 .

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

20

Kleiman, R., Spencer, G.F., Earle, F.R., Nieschlag, H.J., and
Barclay, A.S. , Tetra-acid triglycerides containing a new hydroxy
eicosadienoyl moiety in Lesquerella auriculata seed oil, Lipids, 7, 660,
1972.

Sreenivasan, B., Kamath, N.R., and Kane, J.G., Studies on castor oil.
1. Fatty acid composition of castor oil, J. Am. Oil Chem. Soc., 33, 61,
1956.

Badami, R.C., and Kudari, S.M., Analysis of Hiptage madablota seed
oil, J. Sci. Food Agric., 21, 248, 1970.

Nikolin, A., Nikolin, B., and Jankovic, M., Ipopurpuroside, a new
glycoside from Ipomoea purpurea, Phytochemistry, 17, 451, 1978.

Rao, S.V., Paulose, M.M., and Vijayalakshmi, B., A note on the fatty
acids present in oilseed phospholipids, Lipids, 2, 88, 1967.

Persmark, U., Main constituents of rapeseed lecithin, J. Am. Oil Chem.
Soc., 45, 742, 1968.

Vijayalakshmi, B., and Rao, S.V., Fatty acid composition of
phospholipids in seed oils containing unusual acids, Chem. Phys.
Lipids, 9, 82, 1972.

Kondoh, H. , and Kawabe, S. , Phospholipids in castor seeds, Agr. Biol.
Chem, 39, 745, 1975.

Nasirullah, Werner, G. , and Seher, A. , Fatty acid composition of lipids
from edible parts and seeds of vegetables, Fette, Seifen, Anstrichm. ,
86, 264, 1984.

Prasad, R.B.N., Rao, Y.N., and Rao, S.V., Phoshpholipids of palash
(Butea monosperma), papaya (Carica papaya), jangli badam (Sterculia
foetida), coriander (Coriandrum sativum) and carrot (Daucus carota)
seeds, J. Am. Oil Chem. Soc., 64, 1424, 1987.

Jamieson, GR, and Reid, E.H., Analysis of oils and fats by gas
chromatography. V. Fatty acid composition of the leaf lipids of
Myosotis scorpioides, J. Sci. Food Agric., 19, 628, 1968.

Iamieson, GR, and Reid, E.H., The leaf lipids of some members of
the Boraginaceae family, Phytochemistry, 8, 1489, 1969.

36.

37.

38.

39.

41.

42.

43.

45.

46.

47.

21

Appelqvist, L.-A. , Biochemical and structural aspects of storage and
membrane lipids in developing oilseeds, in Recent Advances in the
Chemistry and Biochemistry of Plant Lipids, Galliard, T. and Mercer,
E.I., Eds., Academic Press, London, 1975, chap. 8.

Shenstone, F.S. , and Vickery, J.R. , Occurrence of cyclo-propene fatty
acids in some plants of the order Malvales, Nature, 190, 168, 1961.

Schmid, K.M. , and Patterson, G.W. , Distribution of cylclopropenoid
fatty acids in malvaceous plant parts, Phytochemistry, 27, 2831, 1988.

Mangold, H.K., and Spener, F., Biosynthesis of cyclic fatty acids, in
The Biochemistry of Plants, Vol. 4, Stumpf, P.K. , Ed. , Academic
Press, London, 1980, chap. 19.

Lie Ken Jie, M.S.F., and Sinha, 8., Fatty acid composition and the
characterization of a novel dioxo Cls-fatty acid in the latex of Hevea
brasiliensis, Phytochemistry, 20, 1863, 1981.

Gandhi, V.M., Cherian, K.M., and Mulky, M.J., Nutritional and
toxicological evaluation of rubber seed oil, J. Am. Oil Chem. Soc. , 67,
883, 1990.

Harper, D.B., and Hamilton, J.T.G., Identification of threo-l8-fluoro—
9,10-dihydroxystearic acid: a novel w-fluorinated fatty acid from
Dichapetalum toxicarium seeds, Tetrahedron Lett. , 31, 7661, 1990.

Abdel-Moety, E.M. , Cyclopentenylfettsauren als ausgangsmaterial zur
gewinnung neuer wirkstoffe, Fette, Seifen, Anstrichm. , 83, 65, 1981.

Greenberg, A. , and Harris, J., Cyclopropenoid fatty acids, J. Chem.
Edu., 59, 539, 1982.

Schuch, R., and Ahmad, F., Structure and biological significance of
triacylglycerols containing cyclopropene acyl moieties, Fat Sci.
Technol., 89, 338, 1987.

Pawlowski, N.E., Hendricks, J.D., Bailey, M.L., Nixon, J.B., and
Bailey G.S. , Structural-bioactivity relationship for tumor promotion by
cyclopropenes, J. Agric. Food Chem, 33, 767, 1985.

Chan, B.G., Waiss, A.C., Jr., Binder, R.G., and Elliger, C.A.,
Inhibition of lepidopterous larval growth by cotton constituents,
Entomol. Exp. Appl, 24, 294, 1978.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

22

Schmid, K.M. , and Patterson, G.W. , Effects of cyclopropenoid fatty
acids on fungal growth and lipid composition, Lipids, 23, 248, 1988.

Kramer, J.K.G., Sauer, F.D., and Pigden, W.J., High and Law Erucic
Acid Rapeseed Oils, Academic Press, Toronto, 1983, chaps. 11-21.

Williams, D.H., Stone, M.J., Hauck, P.R., and Rahman, S.K., Why
are secondary metabolites (natural products) biosynthesized?, J. Nat.
Prod., 52, 1189, 1989.

Proudlock, J.W., Haslam, J.M., and Linnane, A.W., Biogenesis of
mitochondria 19. The effects of unsaturated fatty acid depletion on the
lipid composition and energy metabolism of a fatty acid desaturase
mutant of Saccharomyces cerevisiae, Biaenergetics, 2, 327, 1971.

Wisnieski, BL, and Kiyomoto, R.K., Fatty acid desaturase mutants of
yeast: growth requirements and electron spin resonance spin-label
distribution, J. Bacterial, 109, 186, 1972.

Lands, W.E.M., Sacks, R.W., Sauter, J., and Gunstone, F., Selective
effects of fatty acids upon cell growth and metabolic regulation, Lipids,
13, 878, 1978.

Nes, W.D., Adler, l.H., and Nes, W.R., A structure-function
correlation for fatty acids in Saccharomyces cerevisiae, Exp. Mycol. , 8,
55, 1984.

Light, R.T., Lennarz, W.J., and Bloch, K., The metabolism of
hydroxystearic acids in yeast, J. Biol. Chem, 237, 1793, 1962.

Silbert, D.F., Ruch, R., and Vagelos, P.R., Fatty acid replacements in
a fatty acid auxotroph of Escherichia coli, J. Bacterial. , 95, 1658,
1968.

Silbert, D.F., Ladenson, R.C., and Honegger, J.L., The unsaturated
fatty acid requirement in Escherichia coli. Temperature dependence and
total replacement by branched-chain fatty acids, Biachim. Biophys.
Acta, 311, 349, 1973.

Machtiger, N .A., and Fox, C.F., Biochemistry of bacterial
membranes, Ann. Rev. Biochem, 42, 575, 1973.

Terzaghi, W.B. , Manipulating membrane fatty acid compositions of
whole plants with tween-fatty acid esters, Plant Physiol. , 91, 203,

61.

62.

63.

65.

67.

68.

69.

23
1989.

Dufourc, E.J., Smith, I.C.P., and Jarrell, H.C., Role of cyclopropane
moieties in the lipid properties of biological membranes: a 2H NMR
structural and dynamical approach, Biochemistry, 23, 2300, 1984.

Isaacson, Y., Riehl, T.E., and Stenson, W.F., Nonelectrolyte
permeability of liposomes of hydroxyfatty acid-containing
phosphatidylcholines, Biachim. Biaphys. Acta, 986, 295, 1989.

Grogan, D.W., and Cronan, J.B., Cloning and manipulation of the
Escherichia coli cyclopropane fatty acid synthase gene: physiological
aspects of enzyme overproduction, J. Bacterial. , 158, 286, 1984.

Grogan, D.W., and Cronan, J.B., Characterization of Escherichia coli
mutants completely defective in synthesis of cyclopropane fatty acids,
J. Bacteriol, 166, 872, 1986.

Hutton, D., and Stumpf, P.K., Fat metabolism in higher plants LXII.
The pathway of ricinoleic acid catabolism in the germinating castor
bean (Ricinus communis L.) and pea (Pisum sativum L.), Arch.
Biochem. Biaphys., 142, 48, 1971.

Bafor, M., Smith, M.A., Jonsson, L., Stobart, K., and Stymne, S.,
Ricinoleic acid biosynthesis and triacylglycerol assembly in microsomal
preparations from developing castor-bean (Ricinus communis)
endosperm, Biochem. J., 280, 507, 1991.

Bafor, M., Jonsson, L., Stobart, A.K., and Stymne, 8., Regulation of
triacylglycerol biosynthesis in embryos and microsomal preparations
from the developing seeds of Cuphea lanceolata, Biochem. J. , 272, 31,
1990.

00, K.-C., and Huang, A.H.C., Lysophosphatidate acyltransferase
activities in the microsomes from palm endosperm, maize scutellum,
and rapeseed cotyledon of maturing seeds, Plant Physiol. , 91, 1288,
1989.

Griffiths, G., Stobart, A.K., and Stymne, S., A‘- and A‘z-desaturase
activities and phosphatidic acid formation in microsomal preparations
from the developing cotyledons of common borage (Borago officinalis),
Biochem. J., 252, 641, 1988.

Mukherjee, K.D., and Kiewitt, 1., Changes in fatty acid composition of

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

24

lipid classes in developing mustard seed, Phytochemistry, 23, 349,
1984.

Cao, Y. -Z. ,and Huang, A. H. C., Acyl coenzyme A preference of
diacylglycerol acyltransferase from the maturing seeds of Cuphea,
maize, rapeseed, and canola, Plant Physiol., 84, 762, 1987.

Fehling, E., and Mukherjee, K.D., Biosynthesis of triacylglycerols
containing very long chain mono—unsaturated fatty acids in seeds of
Lunaria annua, Phytochemistry, 29, 1525, 1990.

Cao, Y.-Z., Oo, K.-C., and Huang, A.H.C., Lysophosphatidate
acyltransferase in the microsomes from maturing ms of meadowfoam
(Limnanthes alba), Plant Physiol. , 94, 1199, 1990.

Battey, LE, and Ohlrogge, J.B., A comparison of the metabolic fate of
fatty acids of different chain lengths in developing oilseeds, Plant

Physiol., 90, 835, 1989.

Weiss, E.A., Oilseed Craps, Longman, London, 1983.

Pax, F., and Hoffmann, K., Euphorbiaceae, in Die Naturlichen
Hanenfamilien, Zweite Auflage, 19c, Engler, A., and Prantl, K., Eds., W.
Engelmann, Leipzig, 1931, 11-233.

Webster, G.L. , Conspectus of a new classification of the Euphorbiaceae,
Taxan, 24, 593-601, 1975.

Webster, G.L. , The saga of the spurges: areview of classification and
relationships in the Euphorbiales, Bat. J. Linn. Soc. , 94, 3-46, 1987.

Atsmon, D., Castor, in Oil Craps of the World, Robbelen, 6., Downey, K.R.,
and Ashri, A., Eds., McGraw-Hill, New York, 1989, chap. 24.

Kanerva, L., Estlander, T., and Jolanki, R., Long-lasting contact urticaria
from castor bean, J. Am. Acad. Dermatology, 33, 351-355, 1990.

Rizk, A.-F.M. , The chemical constituents and economic plants of the
Euphorbiaceae, Bat. J. Linn. Soc., 94, 293-326, 1987.

Horton, J. , and Williams, M.A. , A cooker-extruder for deallergenation of
castor bean meal, J. Am. Oil Chem. Soc., 66, 227-231, 1989.

Hawkins, R., Bring back the U.S. castor cropl, J. Am. Oil Chem. Soc., 66,

1562-1563, 1989.

25

CHAPTER 2

BIOCHEMICAL STUDIES
ABSTRACT

Experiments aimed at a better understanding of oleate-12-hydroxylase are
described. An earlier report1 suggested that hydroxylase activity arises in the castor
seed at a stage in development subsequent to the differentiation of the cellular
endosperm and the onset of lipid deposition. Results are presented contradicting this
report: hydroxylase activity was readily detectable at all stages of seed development in
which cellular endosperm was present. 3

Inhibition of oleate-12-hydroxylase activity following incubation with anti-
cytochrome b, immunogobulins demonstrated that cytochrome b, is the electron donor
to the hydroxylase.

Oleate-lZ-hydroxylase activity was lost during various attempts to fractionate
active endosperm extracts, thwarting efforts at partial purification.

In an alternative strategy, rabbits were immunised with denatured endosperm
proteins. Polyclonal antibodies produced, if inhibitory to oleate-lZ-hydroxylase,
could be purified and used to isolate the hydroxylase protein. However, these
antibodies had no effect on hydroxylase activity.

Biochemical features of the hydroxylase are summarised and compared to

features of the microsomal fatty acid desaturases.

26

27
INTRODUCTION

The biosynthesis of ricinoleic (12D-hydroxyoctadec—cis—9-enoic) acid from
oleic acid in the developing endosperm of castor (Ricinus communis) is relatively well
studied. Morris2 established in elegant double-labeling studies that hydroxylation
occurs directly by hydroxyl substitution rather than via an unsaturated-, keto- or
epoxy-intermediate. Developing endosperm slices were incubated with a mixture of
[1-“C]oleic acid and erythra- 12, 13-ditritio-oleic acid. The 3H/“C ratio of ricinoleate
synthesized was found to be 75 % of that of the substrate mixture, and upon chemical
oxidation to 12-keto-oleate the ratio was 55 % , close to a predicted 50%. These
results can only be obtained by a hydroxyl substitution mechanism (’H/“C ratios in
the ricinoleate and 12-keto-oleate would be, respectively, for a keto-intermediate:
50% and 50%, for an unsaturated intermediate: 50% and 25%, and for an epoxy-
intermediate also 50% and 25%).

Hydroxylation using oleoyl-CoA as precursor can be demonstrated in crude
preparations or microsomes, but activity in microsomes is unstable and variable, and
isolation of the microsomes involved a considerable, or sometimes complete loss of
activity."4 Oleic acid can replace oleoyl-CoA as a precursor, but only in the presence
of CoA, Mg2+ and ATP,3 which suggests that activation to the acyl-CoA is necessary.
Furthermore, radioactivity from oleoyl-CoA is rapidly lost during hydroxylation
assays, appearing first in lipid-esterified oleate, and subsequently in lipid-esterified

ricinoleate. This and other experiments demonstrated a better substrate-product

28

relationship between lipid-esterified oleate and lipid-esterified ricinoleate than oleoyl-
CoA and lipid-esterified ricinoleate, and no radioactivity could be detected in
ricinoleoyl—CoA.‘ The hydroxylase is sensitive to cyanide and azide, and dialysis
against metal chelators reduces activity, which could be restored by addition of
FeSO4, suggesting iron involvement in enzyme activity,3 although contradictory
results have also been reported.‘ There is no conversion of oleoyl-CoA to ricinoleic
acid in the absence of molecular oxygen,“ although earlier studies had indicated
otherwise.” Hydroxylation requires NAD(P)H, and NADPH supports lower rates of
hydroxylation than NADH.3'4 Carbon monoxide does not inhibit hydroxylation,
suggesting that a cytochrome P450 is not involved.3'4 Data from a study of the
substrate specificity of the hydroxylase show that all substrate parameters (i.e. chain
length and double bond position with respect to both ends) are important; deviations
in these parameters caused reduced activity relative to oleic acid.6 The position at
which the hydroxyl was introduced however, was determined by the position of the
double bond, always being three carbons distal.

Fatty acid hydroxylation in organisms other than castor is now considered for
comparison. The only other organism in which ricinoleic acid biosynthesis has been
investigated is the ergot fungus, Claviceps purpurea. Ricinoleate accumulates (up to
40% of the fatty acids) in the glycerides produced particularly by sclerotia of
anaerobic cultures.7 As this suggests, oxygen is not necessary for the synthesis of
ricinoleic acid in Claviceps, and the precursor of ricinoleic acid in fact appears to be

linoleic acid.8 However, ricinoleic acid may not be formed simply by hydration of

29

linoleic acid, since there are no free hydroxyl groups in ergot oil. Rather, the
hydroxyl groups are all esterified to other, non-hydroxy fatty acids, leading to a range
of tetra-acyl-, penta-acyl- and hexa-acyl-glycerides. These estolides may be formed
by a direct enzymic addition of non-hydroxy fatty acids across the A12 double bond
of linoleate.’ Ricinoleic acid may, therefore, be merely an artifact of the hydrolysis
employed to study the fatty acid composition of the oil.

Investigations of the biosynthesis of isoricinoleic acid (9-hydroxyoctadec-cis-
12—enoic acid) in Wrightia spp.‘°'11 have provided no conclusive evidence relating to
the pathway of biosynthesis. Interestingly, 9-OH-18:0 is detected in the seed oil, and
can be labelled after incubation of developing seed halves with [“C]acetate, indicating
that the substrate for hydroxylation may be stearate (by direct hydroxylation) or oleate
(by double bond hydration). These workers" favoured the latter possibility because
under their conditions [“C]18:1 and [”C]18:2 could act as precursors for [“C]9-OH-
18: 1, and this conversion was greater under anaerobic conditions. However it seems
that these anaerobic conditions may not have been rigorous, and incubation times
were very long; indeed the conversion of [“C]18:2 to [“C]9-OH-18:0 and [“C]9-OH-
18:1 at equivalent rates must be viewed with some concern.

Little information pertinent to hydroxylation in castor is given by these
comparisons with other organisms. The state of knowledge of the biochemical nature
of the castor hydroxylase is summarised as follows. Oleate is the precursor of
ricinoleic acid, and hydroxylation procedes by a direct hydroxyl susbstitution

mechanism. The immediate enzyme substrate may be lipid-esterified oleate, and

30

some activity may be found in membrane fractions, but evidence that the enzyme is
membrane-bound rather than soluble is weak. NAD(P)H is required, as probably is
molecular oxygen. Preference for NADH over NADPH and insensitivity to carbon
monoxide mitigate against the involvement of a cytochrome P450. No purification or
partial purification of the enzmye has been reported. The present studies of
hydroxylation in castor therefore commenced with the objective of acquiring a better
knowledge of the biochemical nature of the hydroxylase, particularly focused on any
information which may suggest methods of purification which could further the stated

goal of cloning a gene(s) for the hydroxylase enzyme.
MATERIALS AND METHODS

Plant Material

Castor (Ricinus communis L. cv Baker 296) plants were grown throughout the
year in the greenhouse. Developing inflorescences were removed for isolation of
endosperm tissue in the laboratory. Cellular endosperm and embryo were removed
from seeds at development stage HI to stage V,12 identified chiefly by the phase of
rapid expansion of the opaque white endosperm. The tissues were isolated directly
into liquid nitrogen, ground to a powder in a mortar and pestle, and could be stored

at -80°C. This frozen powder is hereafter referred to as "endosperm".

3 l
OleateIZ-Hydroxylase Assays

Developmental Analysis of Ricinoleic Acid Biosynthesis. Frozen endosperm
was ground with two volumes of extraction buffer (50 mM PIPES-KOH pH 7.1, 10%
(w/v) sucrose, 1 mM cysteine, 5 pg ml'1 leupeptin) in a glass-in-glass homogenizer.
The homogenate was micro-centrifuged (~13 000 g) for 15 min, and the supernatant
was decanted from the pellet and fat pad and used for hydroxylase assays. Each
assay contained enzyme (50 p1), assay buffer (50 mM PIPES-KOH pH 7.1, 10%
(w/v) sucrose), NADH (0.4 mM) and 1-“C-oleoyl-CoA (80 000 dpm, 52 Ci mol"), in
a total volume of 0.5 m1. Assay tubes were incubated at 30°C for 60 min, then
stopped by the addition of 15 % methanolic KOH (0.5 ml). Lipids were saponified by
heating to 80°C for 30 min, followed by recovery of the free fatty acids by
neutralising with 2.5 M HCl (0.5 m1) and extraction with 3.5 m1 hexane:isopropanol
(3:2)/ 2.5 ml 0.2 M NaZSO4. The organic phase was dried under nitrogen and
redissolved in 50 p1 chloroformzmethanol (2: 1) for spotting (alongside authentic oleic
acid and ricinoleic acid standards) on a silica TLC plate (Baker Si250). The plate
was developed in a paper-lined tank containing benzene:ethyl etherzethanol
(100:30:2). Fatty acids were detected by staining with iodine vapour. Radioactivity
could be quantitated by autoradiography followed by scraping and scintillation-
counting of the oleic acid and ricinoleic acid spots, but equivalent quantitation was
routinely obtained using a BioScan 2000 scanner.

Several lines of evidence demonstrate that the radioactive compound

32

quantitated in these assays is truly ricinoleic acid. When reaction products were
esterified with methanol, and separated by TLC in a petroleum ether:ethyl ether (1:1)
system they again co-migrated with authentic standards. Furthermore, in the standard
solvent (CJ-IozEtQO:EtOH), a predominant iodine-staining spot (ricinoleic acid is a
predominant fatty acid of the developing castor endosperm) readily overloads the
plate, resulting in a tailing pattern that is precisely matched by tailing of the
radioactivity thought to reside in ricinoleic acid. Finally, unequivocal proof of
ricinoleic acid as the reaction product was obtained by using a tri-deuterated substrate
in parallel to the radioactive substrate and analysis of the products by gas
chromatography-mass spectrometry. This work was a collaboration with Dr. B.W.

Underhill and is presented in the Appendix.

Electron Transfer to Oleate-1 2-Hydraxylase. Endosperrn was extracted and
assayed as described above, with the exception of the following changes. The
extraction buffer contained only 50 mM PIPES-NaOH, pH 7.1, and centrifugation
was at 1000g (SS-34 rotor) for 15 min. The supernatant was desalted by collecting
turbid fractions from a small Sephadex G-25M column equilibrated with the extraction
buffer. The same enzyme preparation was stored without loss of activity by freezing
drop-wise in liquid nitrogen and maintained at 80°C, and used for this series of
experiments. The assay buffer was neutralised with NaOH in place of KOH, and
contained 1 mg ml‘1 fatty acid free bovine serume albumin in addition to the other

components.

33
Polyclonal Antibodies Raised Against Crude Enzyme Extracts. Endosperm was

extracted and assayed as described above, with the exception of the following
changes. The extraction buffer contained only 50 mM PIPES-NaOH, pH 7.1, and
leupeptin (5 pg m1“), and centrifugation was at 1000g (SS-34 rotor) for 10 min.

Assays used only 25 pl enzyme, and contained no sucrose.

Protein Determination

Protein was assayed by the method of Bradford,13 using BioRad reagents, with

bovine serum albumin as standard.

Polyclonal Antibodies raised against crude enzyme extracts

Endosperm was extracted as described above, but the supernatant was further
processed as follows. MgCl2 (1M) was added to a final concentration of 50 mM, and
the preparation was micro-centrifuged for 10 min. The supernatant was passed
through an 0.2 pm filter, and the filtrate desalted by chromatography on Sephadex-
G25M as described above. This preparation was active in hydroxylase assays, and
was loaded (3.7 mg protein per gel) onto two preparative (1.5 mm) polyacrylamide
(10%) gels. “ Following electrophoresis, protein was transferred“ to nitrocellulose
filters. Strips of the filters were baked at 80°C in vacuo for 30 min, chopped finely

with a razor, followed by homogenization in a glass-in-glass homogenizer with

34

deionised water. The nitrocellulose powder was pelleted and resuspended in sterile
phosphate-buffered saline solution. ‘6 These samples were divided into three aliquots,
and used to immunise female New Zealand White rabbits. The rabbits were bled one
day prior to injection, injected subcutaneously,17 then boosted one month later, and
again after a further three weeks. Immune sera used in the experiments described
were obtained one week after the second boost, and had higher titre (as determined by
signal strength on western blots) than sera obtained one week after the first boost.

Serum was cleared at 5000g for 10 min and purified by ion exchange
chromatography with DEAE-Sephacel as described18 and stored at -20°C. Further
purification by affinity chromatography on protein G-Separosel9 was necessary prior
to incubations with enzyme, to remove presumed proteases.

A western blot” was prepared from a preparative (0.75 mm) polyacrylamide
(10%) gel containing protein (0.4 mg total) from the same enzyme preparation as
used to immunise the rabbits. The blot was cut into strips (each representing 40 pg
protein loaded) and blocked with BSA. Each strip was then incubated with diluted
(1/200) immune serum or preimmune serum (from the same rabbit), and developed
using goat-anti-rabbit alkaline phosphatase (Kirkegaard & Perry) with NBT/BCIP.‘°

Staphylococcus aureus cells were the gift of Dr. John Shanklin, prepared

according to Kesslerzo, and were washed extensively with assay buffer prior to use.

35
DEVELOPMENTAL ANALYSIS OF RICINOLEIC ACID BIOSYNTHESIS

A report in the literature‘ suggests that oleate-12—hydroxylase becomes active
in the developing castor endosperm at a stage subsequent to and distinct from the
differentiation of the storage endosperm and the commencement of storage lipid
synthesis. This could be exploited by examining the possibly small number of
proteins which appear in the endosperm concomitant with the onset of oleate
hydroxylation. An initial experiment was therefore done to confirm the previous
observations.‘

A comprehensive descriptive morphology of the developing castor seed has
been developed by Greenwood and Bewley.12 In this timetable the cellular
endosperm, which is the major lipid accumulating tissue, first arises at stage III (heart
shaped embryo). Lipid accumulates very rapidly from stages IV (early cotyledon) to
VII (full cotyledon),21 concomitant with the rapid enlargement of the cellular
endosperm. Lipid accumulates more gradually during stages VII to X (maturation).

Endosperms were dissected from developing inflorescences removed from
greenhouse-grown plants. Inner integument, nucellus and free nuclear endosperm
were isolated from stage II seeds. Cellular endosperm was isolated from stage 111
seeds, with a minimum of adhering nucellus and inner integument. Stage IV to stage
VIII cellular endosperms were isolated together with the enclosed embryo. The
isolated tissues were frozen, homogenised and assayed for hydroxylase activity and

protein content.

36

The conversion of C“—oleoyl-CoA to C“-ricinoleate is shown in Table 3 for
each of the developmental stages. No hydroxylase activity was detectable in tissues
from stage II, which precedes the differentiation of the cellular endosperm in which
ricinoleate—containing lipid accumulates. However, hydroxylase activity was
measured in cellular endosperm from stage III and all subsequent stages,
demonstrating that ricinoleic acid biosynthetic capacity arises in parallel with the
storage tissue.

This result demonstrates that there is no’ period of lipid accumulation prior to
the onset of oleate hydroxylation, contrary to the observations of James et al.1
Rather, all stages of the seed in which the cellular endosperm is present and
triglyceride synthesis occurs are also competent for oleate hydroxylation. Since the
appearance of the oleate hydroxylase is concomitant with the differentiation of the
cellular endoperm and the host of biosynthetic capacities which characterise it, it was
decided not to pursue an examination of protein profiles in the seed before and after

the appearance of the oleate hydroxylase.
ELECTRON TRANSFER TO OLEATE-lZ-HYDROXYLASE

Castor oleate-12-hydroxy1ase is dependent upon NAD(P)H}4 This was also
observed in the current study. When enzyme preparations were desalted by
chromatography on Sephadex G-25M, activity could only be recovered by addition of

NADH, or less effectively, NADPH. Another class of plant fatty acid oxidative

37

Table 3. Oleate-12-hydroxylase activity measured in extracts of endosperm isolated
from castor seeds at different stages of development.

Oleate-lZ-Hydroxylase Activity

 

Stage‘ pmol ricinoleate/mg f w/h pmol ricinoleate/mg protein/h

II 0 0

1H 0.14 18.2
IV 0.59 39.8
V 0.71 38.7
VI 0.62 48.3
VII 0.52 35.8
VIH > 0. 38" 33.3

 

'Stages according to Greenwood and Bewley.12

”This value is a minimum bound to the true value, since a little of the enzyme
preparation was spilled, precluding accurate calculation.

 

38

enzymes, the microsomal desaturases, also require NAD(P)H, and it has been shown19
that donation of electrons originating from NADH to the desaturase is via cytochrome
b,. This is also the case for mammalian desaturases, where NADH—cytochrome b5
reductase and cytochrome b, are involved in transfer of electrons to the desaturase.”23
Cytochromes P450 on the other hand, including plant fatty acid hydroxylases and
epoxidases involved in cutin biosynthesis“3O typically utilize NADPH preferentially
over NADH, and NADPH-cytochrome P450 reductase has been shown to be the
electron donor to these enzymes.

An experiment was done to test whether cytochrome b, is also the electron
donor to oleate-12-hydroxylase. This was by the same approach as Keams et al.,‘9
who used antibodies raised against cytochrome b, to hinder its interactions with either
the NADH-cytochrome b5 reductase, the desaturase, or both. Strong inhibition of
oleate-l2-hydroxylase by anti-cytochrome b5 antibodies would be further evidence that
this enzyme is not a cytochrome P450.

Preliminary experiments showed that maximal hydroxylase activity (20 pmol
ricinoleate formed h") was achieved with 111-125 pg of an enzyme preparation, more
enzyme (to 277 pg) resulting in the same measured activity (Figure 1), indicating that
these enzyme quantities were saturating. Furthermore, the enzyme could be stored
for at least 5 h on ice before initiation of the assay without loss of activity (Figure 2).

Enzyme (83 pg protein) was incubated on ice for 6 h before assay with pre- or
post-immune IgG from a mouse injected with purified cauliflower cytochrome b,.19

The results presented in Figure 3 demonstrate that while incubation with pre-immune

39

 

 

30
:5 25 ‘-
e
E O O
_Q_) 20 ‘” o o O
E O
.9 o
f: 15 ——
O
E
O.
V
>\ 10~—
7‘: 0
.Z
'5
< 5 —— o
0
Q)
0 i 1 1 1 1

 

 

0 BO 100 150 ZOO 250 BOO

Enzyme (micrograms protein)

Figure 1. Oleate-lZ-hydroxylase activity as a function of protein assayed.

40

 

i3 300
\\

C

E:

O

a 250——
O

E

B 200 “o—
8

_ W
.g 150--
6

E

‘v lOO~~
.3

E

6

< 50 r_
2

E

2':

m D i i i i ee i

 

 

 

Time on ice (h)

Figure 2. Stability of oleate-l2-hydroxylase activity of enzyme incubated on ice.

41

 

180

 

Specific activity (pmol ricinoleate/mg protein/h)

 

 

l 1 I;

O i i ‘1 T134! 1 IV
D 51015 20 25 30 35 4O

 

IgG (micrograms)

Figure 3. Oleate-12-hydroxylase activity as a function of anti-cytochrome b5
immunoglobulin (IgG) added. Immunoglobulins were from pre-immune mouse ascites
fluid (closed circles) or from a mouse that had been immunized with cauliflower
cytochrome b, (open circles).

42

IgG had essentially no effect on oleate-12-hydroxylase activity, anti-cytochrome b5
IgG strongly inhibited hydroxylation. Complete inhibition was achieved with 15 pg
IgG, or an IgG/enzyme protein ratio of 0.18. The same antibody was required at a
much higher ratio (4 mg IgG/ mg microsomal protein) in order to acheive 93 %
inhibition of oleate-12-desaturase in microsomal membranes from safflower,19
although a shorter (2 h) incubation of enzyme and IgG was used in that work.

The assay used here to measure oleate-12-hydroxylase activity employs oleoyl—
CoA as substrate, but this may not be the direct substrate of the hydroxylase enzyme.
In fact, it has been reported31 since this work was done that hydroxylation occurs on
oleate esterified at sn-2 of phosphatidylcholine, as is the case for the microsomal
desaturases. Therefore, the assay measures, at least, both acyltransferase activity and
hydroxylase activity. It was shown by Keams et al.19 that the antibodies used in these
experiments had no effect on the incorporation of oleic acid into phospholipids,
including phosphatidylcholine, in safflower microsomes. Furthermore, the antibodies
also block transfer of electrons from cytochrome b, to cytochrome c, and inhibition of
desaturation mediated by the antibodies could be completely quenched by the addition
of purified cytochrome b5.” This shows that inhibition of desaturation is due to the
involvement of cytochrome b, as the electron donor to the safflower microsomal A12
desaturase, and by extension, also to the castor oleate-12-hydroxylase. Since the

work described here was done, similar results have been reported by Smith et 01?"2

43
PARTIAL PURIFICATION ATTEMPTS

Various strategies for partial purification of oleate-12-hydroxy1ase activity
were investigated in over 30 separate experiments. Since none of these experiments
led to any worthwhile purification of the enzyme, they are not presented here in
detail. Since it had previously been reported that oleate-12-hydroxylase could be
recovered in microsome preparations by differential centrifugation, though often with
significant loss of activity};4 attempts were made in these studies to prepare active
microsomes. .

Preparation of microsomes by standard procedures (removal of cell-debris and
fat by low-speed centrifugation followed by centrifugation of the supernatant at
105000 g for 1 h) generally resulted in complete loss of activity, even if the pellet
was resuspended in the supernatant. Control enzyme, incubated at the same
temperature for the same period of time without ultra-centrifugation, retained full
activity. A linear relationship could be established between the amount of activity
remaining and the duration (or field strength) of centrifugation. These experiments
also revealed an intriguing phenomenon. If an enzyme preparation was partially
inactivated by centrifugation, and the pellet resuspended in the supernatant, the
quantity of this enzyme sample (pg protein) which saturated the oleate-lZ-hydroxylase
assay was similar to the quantity of the uncentrifuged enzyme (pg protein) required to
saturate the oleate-12-hydroxy1ase assay. For example, if quantities less than 50 pg

protein were in the approximately-linear range for the hydroxylase assay, and

44

quantities greater than 50 pg gave no more activity than did 50 pg (enzyme
saturation), then the same was also observed for the centrifuged enzyme, even though
the activity of 50 pg was considerably lower.

These experiments suggested that some property of the oleate-12-hydroxylase,
such as a requirement for a multi-component complex in a membrane, was destroyed
by the force exerted during centrifugation. I am aware of only one other enzyme, an
acyl-CoA A5-desaturase of meadowfoam (Limnanthes alba), for which such a
phenomenon has been ob served.33

Alternatives to differential centrifugation were investigated. Fractionation of
membranes in sucrose density-gradients also generally resulted in complete loss of
activity, probably due to the extensive centrifugation (83 000 gm 3 h) required to
achieve separation of membrane fractions. In some cases, activity could be recovered
from a mixed gradient, but not from any individual fractions of an identical gradient
(nor from pair—wise combinations of these fractions). Activity could be retained by
briefer centrifugation (such as in Percoll gradients), but at the expense of
fractionation, since activity was then recovered throughout the gradient.

Though complete inhibition of the hydroxylase activity by low concentrations
of detergents (<16 mM CHAPS, < 1% Triton X-100) suggested that the enzyme was
membrane-bound, it could not be assigned to a particular cell fraction on the basis of
low-speed centrifugations. After low-speed (e.g. 15 000 g x 15 min) centrifugation,
hydroxylase activity was found in both pellet and supernatant fractions.

As an alternative to ultracentrifugation, attempts were made to prepare

45

microsomes by precipitation with magnesium ions.“ In this technique, the
supernatant from a low-speed centrifugation is adjusted to 50 mM Mg“ and then
centrifuged again, microsomes being obtained in the pellet. However, oleate-12-
hydroxylase activity pelleted in the second centrifugation independently of the Mg2+
concentration added (in the range 0 to 100 mM).

These results generally suggested that the cellular fraction containing oleate-
12-hydroxylase activity does not behave independently of other cellular fractions
during attempts at fractionation. Possibly, as membrane fraction containing the
enzyme adheres non—specifically to various cellular components and therefore gives no
simple pattern of fractionation.

Another alternative fractionation method investigated was iso-electric focusing
using a Rotofor aparatus (BioRad). No activity could be recovered in any fraction.
Precipitations of protein with ammonium sulphate did not give any apparent

purification of the activity.

POLYCLONAL ANTIBODIES RAISED AGAINST CRUDE ENZYME

EXTRACTS

Because difficulty was experienced in even partial purification of the oleate-12-
hydroxylase activity from developing castor seeds, an alternative strategy was
investigated for identifying proteins involved in the hydroxylation reaction. It was

reasoned that if polyclonal antibodies raised against a complex mixture of proteins,

46
including the hydroxylase enzyme, could be found to inhibit hydroxylase assays, then

a purification of the effective antibody could be attempted. A problem with this
approach was the presence in castor mds of the extremely toxic protein ricin. Two
strategies were chosen for raising antibodies in rabbits. In the first strategy, a washed
membrane fraction was prepared from seed extracts, and treated with dithiothrietol, to
separate the subunits of any remaining ricin, thereby rendering it harmless.
Unfortunately, this treatment was not sufficient, due to the extreme lethality of ricin,
and this strategy was abandoned and is not described further. The second strategy
was to separate SDS-denatured proteins by electrophoresis, then transfer the proteins
to a nitrocellulose filter, which was powdered for injection. Such an approach has
been used to generate antibodies that inhibit native enzyme activity.”

Proteins (3.7 mg) from an enzymically-active oleate-12-hydroxylase
preparation were separated on a preparative SDS-polyacrylamide gel, and then
transferred to a nitrocellulose membrane. Coomassie-staining of portions of the gel
before and after transfer indicated that > 50% of the protein was transferred. The
membrane was divided into five horizontal strips bearing proteins of the following
molecular weight ranges: >95 kD, 69-95 kD, 46-69 kD, 30—46 kD, and 15-30 kD.
Protein was fixed to the membrane, then powdered to a form which could be mixed
with aqueous buffer to make an injectable suspension. Rabbits, each immunised with
protein from one of the molecular weight ranges, provided serum which was used to
stain proteins blotted from a gel on which the original enzyme preparation were

separated (Figure 4). Antibodies in the sera react principally with proteins of the

47

 

200-

95-

 

 

46..

30-

 

 

 

 

 

 

 

 

 

 

Figure 4. Western blot strips stained with polyclonal antibodies from rabbits
immunised with different molecular weight ranges (shown at bottom, kD) of
developing castor md proteins. The same protein preparation was separated on a
preparative SDS—PAGE gel, and transferred to give the blot shown. Each strip was
developed separately with either preimmune (p) or immune (i) serum. The migration
of proteins of given molecular weight (kD) is shown at left.

48

expected molecular weight range (i.e. that with which the rabbit was immunised),
with some cross-reactivity to proteins outside the expected range, possibly due to
shared epitopes.

Immunoglobulins from the sera were purified by affinity to protein G, and
then incubated (45 min on ice) with castor developing seed extract, before assay of
oleate-12-hydroxylase activity. There was no effect of incubation with any of the
immunoglobulins on hydroxylase activity (Table 4). In a second experiment, enzyme
was incubated with immunoglobulin for a longer period (2 h), and then mixed with a
10% (v/v) slurry of Stapphylacaccus aureus cells. These bacteria carry a surface
protein (protein A) which binds immunoglobulins non-specifically, and can therefore
be used to precipitate immunoglobulins and immunoglobulin-protein complexes. The
cells were removed after 30 min by centrifugation, and the supernatant was assayed.
These incubations also had no effect on hydroxylase activity (Table 4).

Several explanations may account for the lack of effect of the polyclonal
antibodies upon enzyme activity. If the hydroxylase is an integral membrane protein,
the polyclonal antibodies may have been generated against epitopes buried in the
membrane of the non-denatured, membrane-bound enzyme. Such antibodies would
not hinder the reaction of the native enzyme. Alternatively, the hydroxylase may be a
poor antigen, so that the titre of antibodies specific for the hydroxylase is low.
Possibly, other conditions or ratios of antibody to enzyme might be found which

would give inhibition of hydroxylase activity. This was not pursued.

49

Table 4. Oleate-l2-hydroxylase activity (pmol ricinoleic acid mg‘1 protein min“)
following incubation of enzyme (325 pg protein) with immunoglobulins (IgG; 16.6
pg) from rabbits immunised against castor proteins of different molecular weight
(MW) ranges. In experiment 1, IgG were incubated with enzyme for 45 min on ice,
before assay. In experiment 2, IgG were incubated with enzyme (2 h, on ice) and
then precipitated with Stapphylacaccus aureus (S.A.) cells, before assay of the
supernatant.

 

Experiment 1 Experiment 2

 

 

- IgG : 14.6 - IgG, - S.A. : 14.6

- IgG, + S.A. : 15.2

 

 

 

 

 

MW range pre-immune immune pre-immune immune
15-30 kD 14.2 13.8 12.5 11.5
30-46 kD 15.0 14.9 12.2 14.0
46-69 kD 15.3 14.3 12.4 12.5
69-95 kD 13.1 15.0 13.5 12.4

> 95 kD 14.5 15.8 12.2 12.8

 

50
DISCUSSION AND CONCLUSIONS

The synthesis of ricinoleic acid involves the stereospecific hydroxylation of
oleic acid, found in all plants, at the A12 position. Biochemical investigations of this
hydroxylation activity have produced very similar results to those of other fatty acid
modifying enzymes, namely the microsomal desaturases. All plants have a
microsomal oleate desaturase active at the A12 position. The substrate of this
enzyme36 and of the hydroxylase31 appears to be oleate esterified to the sn-2 position
of phosphatidylcholine. The modification occurs at the same position (A12) in the
carbon chain, and requires the same cofactors, namely electrons from NADH via
cytochrome b, (Figure 3, references 19,32) and molecular oxygen (Appendix).
Neither enzyme is inhibited by carbon monoxide}4 the characteristic inhibitor of
cytochrome P450 enzymes.

This last observation is in contrast to another set of plant fatty acid modifying
enzymes studied, namely the hydroxylases and epoxidases involved in cutin
biosynthesis. These enzymes act at the terminal positions (or, w-l) of the carbon
chain and are characteristically P450 enzymes. The possible evolutionary origin of
the castor oleate-12-hydroxylase is discussed further in chapter 4. It is merely
observed at this point that the greater similarities to the microsomal destaturases make
these the more likely progenitors of the castor seed hydroxylase than the hydroxylases
of cutin biosynthesis. The opposite relationship may apply for another fatty acid

oxidase, responsible for linoleate 12,13-epoxidation in Euphorbia Iagascae, which has

5 l
the characteristics of a cytochrome P450.”

The studies reported in this chapter provided new information relating to the
developmental pattern of oleate-12-hydroxylase activity in developing seeds, and
demonstrate the involvement of cytochrome b, and molecular oxygen in
hydroxylation. However, little direct progress was made toward the stated goal of
cloning a gene(s) encoding the hydroxylase. Attempts at partial purification of the
enzyme indicate that this approach would be very difficult. I am not aware of the
purification of any membrane protein from an oilseed. Difficulty in purification has
also been encountered for microsomal desaturases, leading to attempts to clone genes
by genetic approaches, such as chromosome walking”. Therefore, the next step

toward the goal was also to take a genetic approach (chapter 3).

REFERENCES

1. James, A.T. Hadaway, HQ, and Webb, J .lP.W., The biosynthesis of
ricinoleic acid, Biochem. J., 95 , 448-452, 1965.

2. Morris, L.J. , The mechanism of ricinoleic acid biosynthesis in Ricinus
communis seeds, Biochem. Biaphys. Res. Camrnun., 29, 311, 1967.

3. Galliard, T., and Stumpf, P.K., Fat metabolism in higher plants XXX.
Enzymatic synthesis of ricinoleic acid by a microsomal preparation
from developing Ricinus communis seeds, J. Biol. Chem, 241, 5806,
1966.

4. Moreau, R.A., and Stumpf, P.K., Recent studies of the enzymic

synthesis of ricinoleic acid by developing castor beans, Plant Physiol. ,
67, 672, 1981.

5. Canvin, D.T., The biosynthesis. of long-chain fatty acids in the developing
castor bean, Can. J. Bat., 43, 49-62, 1965.

10.

11.

12.

13.

14.

15.

16.

17.

18.

52

Howling, D., Morris, L.J., Gurr, M.I., and James, A.T., The
specificity of fatty acid desaturases and hydroxylases. The
dehydrogenation and hydroxylation of monoenoic acids, Biachim.
Biaphys. Acta, 260, 10, 1972.

Kfen, V. , Rezanka, T. , and Rehacek, Z. , Occurrence of ricinoleic acid in
submerged cultures of various Claviceps sp., Experentia, 41, 1476-1477,
1985.

Morris, L.J., Hall, S.W., and James, A.T., The biosynthesis of ricinoleic acid
by Claviceps purpurea, Biochem. J., 100, 29c-30c, 1966.

Morris, L.J., Mechanisms and sterochemistry in fatty acid metabolism,
Biochem. J., 118, 681-693, 1970.

Ahmad, F., Schiller, H., and Mukherjee, K.D., Lipids containing
isoricinoleoyl (9-hydroxy-cis—12-octadecenoyl) moieties in mds of
Wrightia species, Lipids, 21, 486, 1986.

Ahmad, F., and Mukherjee, K.D., Biosynthesis of lipids containing
isoricinoleic (9-hydroxy-cis-12-octadecenoic) acid in seeds of Wrightia
species, Z. Naturfarsch., 43c, 505, 1988.

Greenwood, J .S. , and Bewley, J.D. , Seed development in Ricinus communis
(castor bean). 1. Descriptive morphology, Can. J. Bat., 60, 1751-1760, 1982.

Bradford, M.M. , A rapid and sensitive method for the quantitation of
microgram quantities of protein utilising the principle of dye-binding, Anal.
Biochem., 72, 248-254, 1976.

Laemmli, U.K., Cleavage of structural proteins during the assembly of the
head of bacteriophage T4, Nature, 227, 680-685, 1970.

Towbin, H. , Staehelin, T. , and Gordon, J ., Electrophoretic transfer of
proteins from acrylamide gels to nitrocellulose sheets: procedure and some
applications, Prac. Natl. Acad. Sci. USA, 76, 4350-4354, 1979.

Sambrook, J ., Fritsch, ER, and Maniatis, T., Molecular Cloning: a
Laboratory Manual, 2nd ed., Cold Spring Harbor Laboratory Press, 1989.

Diano, M., Bivic, A. 1e, and Him, M., A method for the production of highly
specific polyclonal antibodies, Anal. Biochem., 166, 224-229, 1987.

Hum, B.A.L., and Chantler, S.M., Production of reagent antibodies, Meth.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

53
Enzymal., 70, 104-142, 1980.

Keams, E.V., Hugly, S., and Somerville, C.R., The role of cytochrome b, in
A12 desaturation of oleic acid by microsomes of safflower (Carthamus
tinctart'us L.), Arch. Biochem. Biaphys., 284, 431-436, 1991.

Kessler, S.W. , Cell membrane antigen isolation with the staphylococcal
protein A-antibody adsorbent, J. Immunol., 117, 1482-1490, 1976.

Greenwood, 1.8., and Bewley, J.D., Seed development in Ricinus communis
cv. Hale (castor bean). II. Accumulation of phytic acid in the developing
endosperm and embryo in relation to the deposition of lipid, protein, and
phosphorus, Can. J. Bat., 62, 255-261, 1984.

Oshino, N ., Imai, Y., and Sato, R., A function of cytochrome b, in fatty acid
desaturation by rat liver microsomes, J. Biochem. (Tokyo), 69, 155-167, 1971.

Strittmatter, P., Spatz, L., Corcoran, D., Rogers, M.J., Setlow, B., and
Redline, R. , Purification and properties of rat liver microsomal stearoyl
coenzyme A desaturase, Prac. Natl. Acad. Sci. USA, 71, 4565-4569, 1974.

Ferrante, G., and Kates, M., Characteristics of the oleoyl- and
linoleoyl-CoA desaturase and hydroxylase systems in cell fractions

from soybean cell suspension cultures, Biachim. Biaphys. Acta, 876,
429, 1986.

Salaun, J .-P., Weissbart, D., Durst, F., Pflieger, P., and Mioskowski,
C., Epoxidation of cis aand trans A9-unsaturated lauric acids by a

cytochrome P-450-dependent system from higher plant microsomes,
FEBS Lett., 246, 120, 1989.

Blee, E., Schuber, F., Stereochemistry of the epoxidation of fatty acids

catalyzed by soybean peroxygenase, Biochem. Biaphys. Res. Commun. ,
173, 1354, 1990.

Grechkin, A.N., Kukhtina, N .V., Gafarova, T.E., and Kuramshin,
R.A., Oxidation of [l-“C]linoleic acid in isolated microsomes from pea
leaves, Plant Sci., 70, 175, 1990.

Janistyn, B. , Evidence for conversion of arachidonic acid to
hydroxyeicosatetraenoic acids by a cell-free homogenate of maize
seedlings, Phytochemistry, 29, 2453, 1990.

Fahlstadius, P., Absolute configurations of 9,10—epoxydodecanoic acids

30.

31.

32.

33.

34.

35.

36.

37.

38.

54

biosynthesized by microsomes from jerusalem artichoke tubers,
Phytochemistry, 30, 1905, 1991.

von Wettstein-Knowles, P.M. , Waxes, cutin, and suberin, in Lipid Metabolism
in Plants, T.S. Moore Jr., Ed., CRC Press, Boca Raton, 1993, 127-166.

Bafor, M., Smith, M.A., Jonsson, L., Stobart, K., and Stymne, 8.,
Ricinoleic acid biosynthesis and triacylglycerol assembly in microsomal
preparations from develOping castor-bean (Ricinus communis)
endosperm, Biochem. J., 280, 507, 1991.

Smith, M.A., Jonsson, L., Stymne, S., and Stobart, K., Evidence for
cytochrome b, as an electron donor in ricinoleic acid biosynthesis in

microsomal preparations from developing castor bean (Ricinus communis L.),
Biochem. J., 287, 141-144, 1992).

Moreau, R.A., Pollard, M.R., and Stumpf, P.K., Properties of a A5-fatty
acyl-CoA desaturase in the cotyledons of developing Limnanthes alba, Arch.
Biochem. Biaphys., 209, 376-384, 1981.

Diesberger, H., Muller, C.R., and Sanderman, H., Jr., Rapid isolation of a
plant microsomal fraction by Mg2+-precipitation, FEBS Lett., 43, 155-158,
1974.

Christie, P.J., Ward, J .E., Gordon, MP, and Nester, B.W., A gene required
for transfer of T-DNA to plants encodes an ATPase with autophosphorylating
activity, Prac. Natl. Acad. Sci. USA, 86, 9677-9681, 1989.

Heinz, E., Biosynthesis of polyunsaturated fatty acids, in Lipid Metabolism in
Plants, T.S. Moore Jr., Ed., CRC Press, Boca Raton, 1993, 33-89.

Bafor, M., Smith, M.A., Jonsson, L., Stobart, K., and Stymne, S.,
Biosynthesis of vernoleate (cis-12-epoxyoctadec-cis-9-enoate) in microsomal
preparations from developing endosperm of Euphorbia lagascae, Arch.
Biochem. Biaphys., 303, 145-151, 1993.

Arondel, V., Lemieux, B., Hwang, 1., Gibson, 8., Goodman, H.M., and
Somerville, C.R. , Map-based cloning of a gene controlling omega-3 fatty acid
desaturation in Arabidapsis, Science, 258, 1353-1355, 1992.

CHAPTER 3

GENETIC EXPERIMENTS

ABSTRACT

A genetic approach to isolating a gene encoding oleate-12-hydroxylase is
described. A cDNA library was prepared from developing castor endosperm, in a
yeast expression vector. Clones of the library were expressed in yeast cultures,
which were analysed by gas chromatography for fatty acid composition. Two
methods were used to pool and analyse cultures for the appearance of ricinoleic acid
or other changes in fatty acid composition. In experiment 1, five cultures were
pooled per analysis, and of 2300 clones screened, none showed differences in fatty
acid composition. In experiment 2, 2500—3000 clones were analysed as pools of 500,
with enrichment of polar fatty acids prior to gas chromatography. No ricinoleic acid
was detected.

It is suggested that some problem related to heterologous expression or the
nature of the oleate-lZ-hydroxylase precluded detection of its expression in this

experiment.

55

56
INTRODUCTION

Since the biochemical experiments described in chapter 2 were not productive
towards my overall goal of gaining a better molecular understanding of the oleate-12-
hydroxylase, I decided to employ a different strategy. One genetic approach to
cloning a gene for oleate-12-hydroxylase is the subject of this chapter. I sought to
identify a Ricinus communis cDNA which when expressed in growing yeast cells
would be actively transcribed and translated, leading to enzyme activity which could
be detected by the accummulation of ricinoleic acid. The rationale for this

experiment is described below.

Expression of Heteralagaus Enzymes in Yeast

Expression of active cytosolic enzymes from eukaryotes has been possible in
E. coli, but expression, targetting and processing of eukaryotic proteins normally
found in other compartments of the cell requires expression in another eukaryote,
such as yeast. In many cases, proteins are targetted in the yeast cell in a very similar
manner to that in the native cell. Mammalian proteins have been functionally
expressed in yeast cells. For example, cytochrome P450 laurate w—l hydroxylase
from rabbit accumulates in yeast microsomes in an active form when the rabbit cDNA
is expressed in yeast cells.“ Similar results have also been obtained with cytochrome
P450MC from rat liver.‘ Likewise, a mouse liver cytochrome P450 cDN A expressed

in yeast was integrated into the microsomal membrane in a fully functional form,

57

interacting with the endogenous cytochrome P450 reductase.‘s Translation and proper
targetting of heterologous proteins is, however, not uniformly the case. When
cDNAs for the four subunits of the nicotinic acetylcholine receptor from electric ray
were transformed into yeast, all were transcribed, but only two of the proteins
accumulated.7 For the other two subunits it was necessary to replace the original
signal sequences with that of a yeast gene.

Plant cDNAs have also been functionally expressed and targetted in yeast.
Barley a-amylases 1 and 2 were efficiently secreted into the culture medium under the
direction of their own signal peptides.8 Similarly, when the bean vacuolar protein
phytohemagglutinin was expressed in yeast, it was efficiently targetted to the yeast
vacuole.9

In an example particularly relevant to this work, single copy expression of the
rat stearoyl-CoA desaturase gene in a yeast mutant (aleI) deficient in the homologous
enzyme, complements the mutant.10 These rat and yeast stearoyl-CoA desaturases are
microsomal proteins which form part of a three-component system involving NADH-
cytochrome b, reductase, cytochrome b,, and the desaturase. The functional
replacement of the yeast gene by the transformed rat gene therefore demonstrates that
the rat gene product is correctly targetted to the microsomes and can functionally
interact with yeast cytochrome b,. This is in spite of the fact that the rat and yeast
desaturase proteins share only 36 % identity, suggesting that aspects of these proteins
involved in cytochrome b, interaction and cytochrome b, itself are particularly

conserved. Alternatively, no specific sequence is required for the interaction. Since

58

the experiments of this chapter were done, expression in yeast of various plant
proteins has been reported, including functional expression of a plant cytochrome

P450 (see Discussion and Conclusions).

Availability of Substrates and Cafactars

For functional expression, the plant enzyme must have access to all necessary
substrates and cofactors. It is clear that NADH and oxygen are present in yeast, and
it seems at least quite possible that the yeast NADH-cytochrome bs/cytochrome b5
system will be able to provide electrons to castor oleate-lZ-hydroxylase as they do to
the rat stearoyl-CoA desaturase. The true substrate of the oleate-12-hydroxylase was
not known at the time of this experiment, but was thought to be either oleoyl-CoA
(which is the substrate supplied to assays of the enzyme in castor endosperm
extracts), or oleate esterified to phoshatidylcholine. Oleate is a major component of
yeast membranes, and phosphatidylcholine is the major phospholipid of yeast
microsomes,ll so it appears likely that both oleoyl-CoA and phosphatidylcholine-
esterified oleate should be available at the yeast microsome. It was considered that
pH, other ionic and solute differences between castor and yeast microsomal
environments should be relatively subtle and not sufficient to abolish activity of the

plant enzyme in yeast.

Single-Gene Assumption

One fundamental assumption of this experiment was that a single castor cDNA

59

is sufficient to give oleate-12-hydroxylase activity. There is no evidence available for
or against this assumption, so that for this reason in particular, this experiment was

viewed as risky.

Level of Accumulation of Ricinoleate
Ricinoleate produced by the enzyme expressed in yeast must accumulate to

detectable amounts. Since ricinoleate is an unusually polar acyl group, and is found
only in the triglyceride fraction of castor lipids,“13 it was thought that high levels of
accumulation in yeast might be disruptive to membrane function and were, therefore,
not to be expected. In experiment 1, total fatty acids of the yeast cultures were
analysed for the presence of ricinoleic acid. It was estimated that ricinoleic acid
could be detected if it constituted approximately 0.105% of the fatty acids of a

particular culture.

Possibility of Further Metabolism of Ricinoleate

For ricinoleate to accumulate to detectable levels, it must also not be
metabolised further, or the metabolite should also be detectable. There is a report“
that hydroxystearic acids support the growth of yeast in anaerobic condtions when
they are not otherwise able to grow, due to their inability to make unsaturated fatty
acids (similar to experiments reviewed in chapter 1). This appeared to be due to
acetylation of the hydroxyl group and use of the acetoxy acids as substitutes for oleic

acid in growth. The acetylase activity was also active towards ricinoleic acid. Thus

60

synthesis of ricinoleic acid in yeast may be followed by its acetylation or other
modification. Chromatograms were therefore analysed not only for a peak
corresponding to ricinoleate, but also for any other novel peaks.
Number of Clones to be Screened

All the above conditions being met, the success of the experiment would
depend upon being able to screen a sufficient number of clones, based upon the
expected frequency of the oleate-12-hydroxylase cDNA in the library used. The
library was analysed for the freqency of stearoyl-ACP desaturase" clones, to use as
an approximation for the frequency of clones for oleate-lZ-hydroxylase. This
assumption of similar frequencies is made on the basis that both are involved in the
same pathway (i.e. producing ricinoleate for triglyceride synthesis) and that both
enzymes accomplish fatty acid oxidations, therefore being required to overcome the
same high energy of carbon-hydrogen bond cleavage, and so might have similar
turnover numbers (discussed further in chapter 4). Since the library was constructed
by a non-directional approach, on average only half the clones encoding the pututative
hydroxylase will be in the correct orientation for expression. Furthermore, a
percentage of clones will not contain a full-length copy of the coding sequence,

causing some further increase in the number which need be analysed.

61
MATERIALS AND METHODS

Construction of cDN A Library

Total RNA was purified from developing stage III to stage Vl6 cellular
endosperm plus embryo by the technique of Puissant & Houdebine,l7 with the
following modifications: tissue was ground under liquid nitrogen instead of at 4°C,
and both the LiCl and CHCl3 extractions were repeated once. Poly(A)+ RNA was
enriched by two rounds of chromatography on oligo dT cellulose18 and analysed by
electrophoresis through formaldehyde-containing agarose gels. ‘8 The final poly(A)+
RNA still contained obvious ribosomal RNA bands but these were less prominent
compared to the background smear than in the total RNA. Complementary DNA was
prepared using a kit ("Librarian IV", Invitrogen) according to the instructions of the
manufacturer. First strand cDNA (1.65 pg) was synthesised from poly(A)+ RNA (5
pg) by priming with oligo dT and extension by avian myeloma virus reverse
transcriptase. The RNA was nicked by E. coli RNaseH, forming primers for second-
strand cDNA synthesis by E. coli DNA polymerase 1. Any nicks in the dsDNA were
repaired with E. coli DNA ligase. Ends of the dsDNA were made blunt with T4
DNA polymerase for ligation of BstXl non-palinmomic linkers. The cDNA was size-
selected by agarose-gel electrophoresis, and molecules larger than ~ 750 bp were
ligated into the BstXl-digested pYES2.0 vector (Figure 5) and transformed into E.
coli strain INVlaF’, yielding four pools containing a total of 1.42 X 10°

transforrnants.

62

Xbal.Sphl.Xhol.Notl.Xma||l.BstXl.EcoRl.BstX|.Xmal|l.BamHl.Sacl.Kpnl.Hindlll

Termination

   

T7 Promoter

Gal 1 Promoter

F1

Amp

Figure 5. The vector pYE82.0 (Invitrogen) used for expression of cDNA in yeast.
Castor md cDNA was ligated at the BstXI sites.

63
Screening of Library for Stearoyl-ACP Desaturase

Three plates of the pYE82.0 library were grown until small colonies (53 000
total) were visible. A nitrocellulose filter (Schleicher & Schiill BA85) was laid on
each plate, its position marked, and lifted off to a fresh plate, the adhering colonies
now facing upwards. Care was necessary that both plate and filter were not too
moist, to avoid smearing of the colonies. The original plate was incubated for a few
h at 37°C to recover colonies, while the filters'were processed as folllows. Each
filter was sequentially placed, colony side up, on Whatman 3MM paper moist with
10% SDS (3 min), denaturing solution (0.5 M NaOH, 1.5 M NaCl; 5 min),
neutralising solution (0.5 M Tris-Cl pH 7.4, 1.5 M NaCl; 5 min), and 2 X SSC (0.3
M NaCl, 0.03 M Na-Citrate, pH 7.0). The filters were then air-dried for ca. 1 min
before pressing twice between sheets of filter paper to remove cell debris. After air-
drying a further 30 min, DNA was fixed to the filters by baking in vacuo at 80°C for
1-2 h.

The filters were prehybridised in a minimal volume of 0.25 % nonfat dry milk,
6 X SSC, 10% dextran sulphate, at 68°C, before addition of the probe and
hybridisation overnight. Probes were labelled by random priming18 and purified of
unincorporated nucleotides by ethanol precipitation in the presence of ammonium
acetate.“ The filters were washed three times in 2 X SSC, 0.1% SDS at 68°C, then
exposed to X-ray film. Hybridizing colonies were picked and replated for a second

round of hybridization yielding positive clones.

Transformation of Yeast

DNA of the four pools of the library was prepared from 300 ml E. coli
cultures by an alkaline lysis procedure18 and purified by banding in CsCl gradients.
This DNA was transformed into yeast by electroporation according to Becker &

Guarente. ‘9

Selective Yeast Growth Media

Media for selective growth of yeast contained (per litre):
6.7 g yeast nitrogen base without amino acids (Difco)
10 ml amino acid stock solution
amino acid stock solution (per 200 ml):
8 g casamino acids (Difco)
0.4 g adenine sulphate
0.6 g L-leucine
0.4 g L-tryptophan
0.6 g L-lysine-HCl
0.6 g L-histidine—HCl
Media contained 2 % carbon source (glucose, galactose, or sodium lactate) added after

autoclaving. Media for plates were solidified with 2 % Dico Bacto Agar.

65
B—Galactosidase Assays

B-galactosidase activity was assayed by measuring a-nitrophenol hydrolysed
from ONPG (a-nitrophenyl-B(D)-galctopyranoside). Absorbance (A =600 nm) of the
yeast culture was measured and duplicate samples (1 ml) were pelleted in a microfuge
(5 min). Pellets were washed with 1 ml Z-buffer (0.06 M NaQI-IPO4, 0.04 M
NaI-IZPO“ 0.01 M KCl, 0.001 M MgSO4, adjusted to pH 7.0 with NaOH, and 0.03
M B-mercaptoethanol added just prior to use). Z-buffer (950 pl) was added to the
pellet, followed by 2 drops chloroform (from a 9-inch pasteur pipette) and 2 drops
0.1% SDS. The cells were vortexed (10 s) at maximum speed. 200 pl ONPG (4 mg
ml" aqueous solution) at 30°C was added at a recorded time, vortexed, and incubated
at 30°C. The time was recorded when the reaction turned yellow, and 0.5 ml 1 M
N a2CO3 added and vortexed to stop the reaction. The tubes were microcentrifuged 2
min, and absorbance 0:420 nm) of the supernatant measured
spectrophotometrically. Results (nmol a-nitrophenol/Am/ min = A420 x 79.812/(A600 x

time)) were calculated using an extinction coefficient for a-nitrophenol of 21300 M'

l -1

cm.

Gas Chromatography of Yeast Fatty Acid Methyl Esters (FAMFs)

For experiment 1, yeast cells from five 5 ml cultures were pooled, pelleted

(SS-34 rotor, 10 000 rpm, 5 min), resuspended in methanol (2 x 0.5 ml) and

66
transferred to a glass scew-cap tube. 1.5 M methanolic HCl (2 ml) was added, the

tube capped with a teflon-lined cap, and heated to 80°C for 1 h. Upon cooling, 2 ml
hexane:isopmpanol (3:2) and 1 ml 0.2 M NaZSO4 were added and the FAMEs
removed in the hexane phase, which could be stored at —20°C. The samples were
dried under a stream of nitrogen, redissolved in hexane (100 pl) and transferred to a
gas chromatograph (GC) vial. Any hydroxyl groups were derivatized at room
temperature by addition to the vial of 1 p1 trimethylsilylimidazole reagent ('I‘riSil Z,
Pierce). Some samples were halved and one aliquot spiked with authentic methyl-
ricinoleate (4 pg, Sigma).

For experiment 2, the cell pellet was thawed in 5 ml 15 % methanolic KOH
and saponified at 80°C for 1 h. The fatty acids were neutralized with 2.5 ml 6 M
HCl, then extracted by addition of 1 ml 1 M NaCl and 3 ml hexane. The hexane
phase was removed and the aqueous phase further extracted with hexane (2 x 1 ml).
The pooled hexane phases were dried under a stream of nitrogen and redissolved in
chloroform. A slurry (1:1) of silicic acid (BioSilA) in chloroform was added (1 ml),
then the solvent removed under nitrogen. The fatty-acid containing silicic acid
powder was transferred to a small column and eluted with 4 ml chloroform (neutral
fraction) followed by 2 ml methanol (polar fraction). These fractions were dried
under nitrogen and redissolved in methanolic HCl (1 M) and heated at 80°C for 1 h.
The methyl esters were extracted into hexane, dried down, redissolved in a small
volume (100 pl) of anhydrous hexane, and transferred to a GC vial. After GC

analysis of these FAMEs, any hydroxyl groups were derivatized by the addition of

67
trimethylsilylimidazole and analysed by GC again, comparing the chromatograms of

underivatized and derivatized samples.
Gas chromatography was done on a Hewlet-Packard 5890 series 11 instrument
using SP-2330 glass capillary columns (30 m long, 0.75 mm internal diameter, 0.20

pm film thickness) with helium as carrier gas.
EXPERIMENTAL DESIGN

Complementary DNAs expressed in developing castor endosperm during active
oleate hydroxylation were cloned into a plasmid vector suitable for expression of the
cDN As in yeast (Saccharomyces cerevisiae). The plasmid chosen (pYES2.0) had
several important features (Figure 5). Firstly, it has a ColEl origin of replication and
a B-lactamase gene allowing maintenance and selection in E. coli, respectively.
Secondly, it contains the 2p origin of yeast plasmids which confers high copy levels
in yeast cells and thus favours high levels of expression of the insert cDNA.

Selection in yeast depends upon complementation of uracil auxotrophy in the host
strain by the plasmid URA3 gene. Transcription of the insert cDNA is under the
control of the GAL1 promotor in GAL+ host strains, so that transcription can be
repressed by growth on glucose, or activated by growth on galactose in the absence of
glucose. This particular feature was considered desirable in the possibility that the
inserted cDN A would be toxic to yeast cells when expressed.

The yeast host strain CGY2557 (derivative of AB1380 x X15-3A from

68
Douglas Smith, Collaborative Research, Inc., Bedford, MA. MATa, GAL”, ura3-52,

leu2-3, trpI, adeZ-I , lys2-I , his5, can] -100) was chosen for the experiment on the
basis of its GAL“ phenotype and uracil auxotrophy.

Since a novel peak was sought in gas chromatograms, it was possible to pool a
number of yeast cultures for each analysis, thus increasing the overall number of
clones that could be analysed. In experiment 1, five cultures were pooled for each
analysis. The number of cultures pooled was greatly increased in experiment 2, based

on the reconstruction experiment described below.
PRELIMINARY EXPERIMENTS
Testing GAL Phenotype of Yeast Host Strain

A preliminary control experiment was done to test the GAL+ phenotype of the
the yeast strain CGY2557. This is essentially a test of its ability to grow on galactose
and induce the GAL1 promoter. The plasmid pCGS286 was obtained from Dr Susan
Gibson and has a GALl/lacZ fusion under the control of the GAL1 promoter. A
random clone from the library constructed in pYE82.0 and pCGS286 were used to
transform CGY2557 to uracil prototrophy. Transforrnants were inoculated into
selective medium with 2 % lactate as sole (GAL-promoter neutral) carbon source.
These cultures were grown for 3 days at 30°C until visibly turbid (lactate supports

low growth rates), and then supplemented to either 0.2% glucose or 0.2% galactose.

69

Cultures were harvested and assayed for B-galactosidase activity at the time of sugar
addition and on subsequent days. Results (Table 5) demonstrate that B—galactosidase
activity was detectable only in pCGS286—bearing cells, and only when these were
supplemented with galactose. The highest B-galactosidase activities were attained
within one day, declining in subsequent days. Thus induction of the GAL1 promoter

occurs under the conditions used in these experiments.
Reconstruction Study for Experiment 2

In experiment 2, larger pools of cultures were analysed simultaneously to
facilitate the analysis of a greater number of clones. To achieve this, the samples
were first enriched for polar fatty acids before analysis. A preliminary reconstruction
experiment was therefore done to verify the sensitivity of the techniques used, and to
give an estimate of the content of ricinoleic acid which might be detected. Soybean
phoshatidylcholine (50 mg) was dissolved in a minimal volume of chloroform, then
saponified with 2 ml 15 % methanolic KOH at 80°C for 30 min. The fatty acids were
extracted into hexane, and different samples were spiked with 0-100 pg methyl-
ricinoleate, then dried under a stream of nitrogen and redissolved in chloroform. A
slurry (1:1) of silicic acid (BioSilA) in chloroform was added (1 ml), then the solvent
removed under nitrogen. The fatty-acid containing silicic acid powder was
transferred to a small column and eluted with 3 ml chloroform (neutral fraction)

followed by 2 ml methanol (polar fraction). These fractions were dried under

70

Table 5. B-galactosidase activities of cultures of CGY2557 harbouring a random
clone (pYESr) or pCGS286, supplemented with either glucose or galactose, at the
time of sugar addition (a) and on subsequent days. All values are means of duplicate
assays of 5 cultures (n= 10), except where indicated (‘n=8).

pYESr pCGSZS6

 

Time Glucose Galactose Glucose Galactose

 

units of activity

a 0.00983 0.0781
a+ 1d 0.0193 0.0205 0.0229 10.35
a+2d 0.0134 0.0124 0.0364 5.97
a+3d 0.0169 0.0167 0.0828 3.95'

a+4d 0.0256 0.0241 ’ 0.138 3.62

 

71
nitrogen and redissolved in methanolic HCl (1 M) and heated at 80°C for 1h. The

methyl esters were extracted into hexane, dried down, redissolved in a small volume
(100 p1) of anhydrous hexane, and hydroxyl groups were derivatized by the addition
of trimethylsilylimidazole. These samples were analysed by gas chromatography for
TMS-methyl-ricinoleate. TMS-methyl-ricinoleate could only be detected in the polar
fraction from samples originally spiked with 10 pg or 100 pg methyl ricinoleate.

This indicated that if the original sample (36 mg ‘ fatty acids) represented the fatty

acids extracted from 500 yeast cultures, ricinoleate could be detected if it comprised

approximately 10% of the total fatty acids of an individual culture.

RESULTS

Experiment 1

Yeast transformants from the cDN A library constructed from developing castor
endosperm RNA were ordered into 96-well plates. Each clone could then be used to
inoculate a 5 m1 culture, which was pooled with four other cultures and analysed by
gas chromatography for any changes in the fatty acid methyl ester (FAME) profile. If
a gas chromatogram appeared different from normal, fresh cultures were grown of the
five clones comprising that pool, and each culture analysed individually. If there was
a true difference caused by the cDNA expressed in one clone, then the change in the
FAME profile was expected to have a five-fold greater magnitude than that observed

in the analysis of the pool. This could be confirmed by comparing growth in glucose-

72

and galactose-supplemented medium; expression of the cDNA is repressed by glucose,
hence a change in the FAME profile caused by expression of the cDNA should only
be observed when grown with galactose in the absence of glucose.

An example of a normal gas chromatogram from a primary screen is shown in
Figure 6. Periodically, samples were halved, and one aliquot spiked with methyl
ricinoleate to verify that the derivatisation reaction was working satisfactorily, and to
recalibrate the elution time of TMS-methyl ricinoleate. The sample shown in Figure
6 was such a sample; the spiked chromatogram is shown in Figure 7. An example of
a chromatogram in which there is a putative difference in the FAME profile and
which was selected for rescreening is shown in, Figure 8.

A total of 2300 clones (460 pools) were analysed by this method. No
repeatable differences in the FAME profile were found. Thus, of 2300 cDNA clones,
none had a detectable effect on fatty acid metabolism of the yeast cell. This may be
due (for various reasons discussed in Introduction) to problems associated with
heterologous expression of randomly chosen cDNA clones, or due to the absence
from this sample of 2300 clones of any which might be involved in fatty acid
metabolism (e.g. a fatty acid desaturase or hydroxylase). It was desirable to screen a
larger number of clones, and so a procedure was devised which allowed larger pools

to be analysed, in experiment 2.

73

6-846

 

5.9'91

 

 

pr
(0
0
Mimi,.mtuutflui‘a
A.
4721

 

I
7.072

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 6. Gas chromatogram of fatty acid methyl esters from sample 145 (cultures
8H5 through 8H9). An attenuation was used which favoured visual inspection for
small peaks. Vertical axis: units of signal strength (flame ionization detector);
horizontal axis: retention time (min after injection). Elution times of abundant yeast
FAMEs in this chromatogram were: 16:0 4.960 min, 16:1 5.342 min, 18:0 6.414
min, 18:1 6.846 min. '

74

H H H
N U h
0 0 0
...t‘....t....irr
J
:1346
ééfiml—éfiaas
7
5 9963

 

 

 

 

 

 

 

 

 

I\
1 1e4— di
1 0:42 “1
90001

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 51 on
o 0 0
0 o o
0 o 0
11111.11 1111
I __—_

137369
9 145
éQég?
$10428
46 a

 

 

 

 

 

 

 

 

Figure 7. Gas chromatogram of fatty acid methyl esters from sample 145 (cultures
8H5 through 8H9), spiked with authentic methyl-ricinoleate. Note the elution of
TMS-methyl ricinoleate at 8.610 min. (Attenuation and axes as for Figure 6).

75

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l
1.4-4—1 T t:
3 ti
1.:3e4-:
12e43
L1e4€
3 c
LOe4: i .0
- b
I l 03
9000: »‘ 0
3 ti ii' ’ ‘ a” v
8000—: (x. J Orig? (it)? 011:]
hm'mw 0.
.. ‘4 m ' ('30
7000-: PI O'H
-i i fi d H
soooe J“
5000{
40003
o i i i - 5 T i I i {0

Figure 8. Gas chromatogram of fatty acid methyl esters from sample 143 (cultures
9A5 through 9A9). Note the elution of a small peak at 8.540 min, at the same elution
time as authentic TMS-methyl ricinoleate (see Figure 7). This peak was not seen
when cultures 9A5 through 9A9 were re-grown and analysed individually.
(Attenuation and axes as for Figure 6).

76
Experiment 2

Clones in 96-well plates were replicated onto galactose-containing agar plates
using a 96-prong device. The clones were grown to patches and then the cells from
each plate were scraped together in 5 ml water. The cells from five plates
(representing five 96-well plates) were used to inoculate a 400 ml culture of
galactose-containing medium and grown at 30°C for approximately 24 h. The cells
were pelleted, stored briefly at -20°C, and analysed for FAME profile by gas
chromatography. Between 2500 and 3000 clones were grown and analysed by this
method. No difference in the gas chromatographs could be detected before and after
TMS-derivatization, indicating that no hydroxylated fatty acids were present. This
indicates that ricinoleic acid did not constitute as much as 10% of the fatty acids of

any individual culture.

DISCUSSION AND CONCLUSIONS

In experiments 1 and 2, a total of approximately 5000 yeast clones were
analysed for fatty acid content. In experiment 1, 2300 clones showed no differences
in the overall fatty acid profile, and no ricinoleic acid could be detected. The
sensitivity and resolution of the gas chromatograph system was such that ricinoleic
acid could have been detected if it were present at 0.105% of total fatty acids of an
individual culture. In experiment 2, 2500 to 3000 clones were screened by a more

rapid technique, where larger pools of cultures were first enriched for polar fatty

77

acids by chromatography on silicic acid. A reconstruction experiment indicated that
ricinoleic acid would be detectable if it constituted 10% of the fatty acids of an
individual clone. No ricinoleic acid was detected.

As explained in the introduction, this experiment was viewed as risky. It was
intended only to screen a reasonable number of clones, on the assumption that clones
for the hydroxylase should be moderately abundant, and to terminate the experiment if
it appeared likely that some other problem precluded success. By the design of the
experiment, the negative result obtained also does not provide much new information
regarding the nature of the oleate-12-hydroxylase. The critical factors of the
experimental design are discussed briefly to put the results in context.

Heterologous expression of a range of eukaryotic proteins was described in the
introduction. In many cases, proteins are targetted in the yeast cell in a very similar
manner to that in the native cell. Since the experiments reported here were
completed, many more such reports of successful heterologous expression have
appeared. These include plant soluble proteins2023 and correct targetting of plant
mitochondrial proteins?” nuclear proteins,2°’27‘ and plasmamembrane proteins?“0
Furthermore, a plant cytochrome P450 has been functionally expressed in yeast, and
is localised in microsomes where it forms an active complex with yeast NADPH:P450
reductase?n This suggests that the failure to detect functional expression of an oleate-
12-hydroxylase in the experiments reported here is probably not due to a problem of
heterologous expression.

Various other criteria for successful expression of an oleate- 12-hydroxylase in

78

yeast were discussed in the introduction. It must be assumed that for one or mare of

these reasons, no expression could be detected. In addition to the hydroxylase, it was

hoped that differences in fatty acid composition might also be detected due to the

expression of another cDN A, such as a microsomal desaturase. However, this was

also not observed. Difficulties have independently been encountered, however, with

functional expression of a plant microsomal fatty acid desaturase in yeast (V .A.

Arondel, unpublished). A putative oleate-12-hydroxylase clone also gave no

functional expression in yeast (see chapter 6).

REFERENCES

Imai, Y., Characterization of rabbit liver cytochrome P-450 (laurate w—l
hydroxylase) synthesized in transformed yeast cells, J. Biochem. (Japan), 103,
143-148, 1988.

Imai, Y., and Nakamura, M., The importance of threonine-301 from
cytochromes P-450 (laurate (w-1)-hydroxylase and testosterone 16a-
hydroxylase) in substrate binding as demonstrated by site-directed mutagenesis,
FEBS Lett., 234, 313-315, 1988.

Imai, Y., Uno, T., Nakamura, M., and Yokota, H., Structure-function
relationships of cytochrome P-4501aurate (w—l)-hydroxylase, Drug Metabolism
Reviews, 20, 467—478, 1989.

Uno, T., and Imai, Y., Identification of regions functioning in substrate
interaction of rabbit liver cytochrome P-450 (laurate (w-1)-hydroxylase), J.
Biochem. (Japan), 106, 569-574, 1989.

Oeda, K., Sakaki, T., and Ohkawa, H., Expression of rat liver cytochrome P-
450MC cDNA in Saccharomyces cerevisiae, DNA, 4, 203-210, 1985 .

Cullin, C., and Pompon, D., Synthesis of functional mouse cytochromes P-
450 P1 and chimeric P-450 P3-1 in the yeast Saccharomyces cerevisiae, Gene,
65, 203-217, 1988.

10.

11.

12.

13.

14.

15.

l6.

17.

18.

79

Yellen, G. , and Migeon, J .C. , Expression of Torpedo nicotinic acetylcholine
receptor subunits in yeast is enhanced by use of yeast signal sequences, Gene,
86,145-152, 1990.

Sogaard, S. , and Svensson, B. , Expression of cDNAs encoding barley oz-
amylase 1 and 2 in yeast and characterization of the secreted proteins, Gene,
94, 173-179, 1990.

Tague, B.W., Dickinson, C.D., and Chrispeels, M.J., A short domain of the
plant vacuolar protein phytohemagglutinin targets invertase to the yeast
vacuole, Plant Cell, 2, 533-546, 1990.

Stukey, J .E., McDonough, V.M., and Martin, CE, The OLE] gene of
Saccharomyces cerevisiae encodes the A9 fatty acid desaturase and can be
functionally replaced by the rat stearoyl-CoA desaturase gene, J. Biol. Chem. ,
265, 20144-20149, 1990.

Zinser, E., Sperka-Gottlieb, C.D.M., Fasch, E.V., Kohlwein, S.D., Paltauf,
F. , and Daum, G. , Phospholipid synthesis and lipid composition of subcellular

membranes in the unicellular eukaryote Saccharomyces cerevisiae, J.
Bacterial, 173, 2026-2034, 1991.

Rao, S.V., Paulose, M.M., and Vijayalakshmi, B., A note on the fatty
acids present in oilseed phospholipids, Lipids, 2, 88, 1967.

Kondoh, H., and Kawabe, S., Phospholipids in castor mds, Agr. Biol.
Chem, 39, 745, 1975.

Light, R.J., Lennarz, W.J., and Bloch, K., The metabolism of hydroxystearic
acids in yeast, J. Biol. Chem., 237, 1793—1800, 1962.

Shanklin, J. and Somerville, C. , Stearoyl-acyl-carrier-protein desaturase
from higher plants is structurally unrelated to the animal and fungal
homologs, Prac. Natl. Acad. Sci. USA, 88, 2510, 1991.

Greenwood, 1.8., and Bewley, J.D., Seed development in Ricinus communis
(castor bean). 1. Descriptive morphology, Can. J. Bat., 60, 1751-1760, 1982.

Puissant, C. , and Houdebine, L. , An improvement of the single-step method
of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform
extraction, BiaTechniques, 8, 148-149, 1990.

Sambrook, J ., Fritsch, E.F., and Maniatis, T., Molecular Cloning: a
Laboratory Manual, 2nd ed., Cold Spring Harbor Laboratory Press, 1989.

19.

20.

21.

22.

23.

24.

25 .

26.

27.

28.

29.

80

Becker, D.M., and Guarente, L., High efficiency transformation of yeast by
electr0poration, Meth. Enzymal., 194, 182-187, 1991.

Alderson, A., Sabelli, P.A., Dickinson, J.R., Cole, D., Richardson, M.,
Kreis, M., Shewry, P.R., and Halford, N.G., Complementation of snfl, a
mutation affecting global regulation of carbon metabolism in yeast, by a plant
protein kinase cDNA, Prac. Natl. Acad. Sci. USA, 88, 8602-8605, 1991.

Deforce, L., Tomizawa, K.-I., Ito, N., Farrens, D., Song, P.-S., and Furuya,
M. , In vitro assembly of apophytochrome and apophytochrome deletion
mutants expressed in yeast with phycocyanobilin, Prac. Natl. Acad. Sci. USA,
88, 10392-10396, 1991.

Wahleithner, J.A., Li, L., and Lagarias, J .C., Expression and assembly of
spectrally active recombinant holophytochrome, Prac. Natl. Acad. Sci. USA,
88, 10387-10391, 1991.

Blechl, A.E., Thrasher, K.S., Vensel, W.H., and Creene, F.C., Purification
and characterization of wheat a-gliadin synthesized in the yeast,
Saccharomyces cerevisiae, Gene, 116, 119-127, 1992.

Chaumont, F., O’Riordan, V., and Boutry, M., Protein transport into
mitochondria is conserved between plant and yeast species, J. Biol. Chem. ,
265, 16856-16862, 1990.

Huang, J., Struck, F., Matzinger, DE, and Levings, C.S., Functional
analysis in yeast of cDNA coding for the mitochondrial Rieske iron-sulfur
protein of higher plants, Prac. Natl. Acad. Sci. USA, 88, 10716-10720, 1991.

Hirt, H., Pay, A., Gyorgyey, J ., Bako, L., Nemeth, K., Bogre, L.,
Schweyen, R.J., Heberle-Bors, E., and Dudits, D., Complementation of a
yeast cell cycle mutatn by an alfalfa cDNA encoding a protein kinase
homologous to p34°°°2, Prac. Natl. Acad. Sci. USA, 88, 1636-1640, 1991.

Tinland, B., Koukolikova-Nicola, 2., Hall, M.N., and Hohn, B., The T-
DNA-linked VirD2 protein contains two didtinct functional nuclear localization
signals, Prac. Natl. Acad. Sci. USA, 89, 7442-7446, 1992.

Anderson, J.A., Huprikar, S.S., Kochian, L.V., Lucas, W.J., Gaber, R.F.,
Functional expression of a probable Arabidapsis thaliana potassium channel in
Saccharomyces cerevisiae, Proc. Natl. Acad. Sci. USA, 89, 3736-3740, 1992.

Sentenac, H., Bonneaud, N., Minet, M., Lacroute, F., Salmon, J.-M.,
Gaymard, F., and Grignon, C., Cloning and expression in yeast of a plant

30.

31.

81
potassium ion transport system, Science, 256, 663-665, 1992.

Villalba, J .M., Palmgren, M.G., Berberian, G.E., Ferguson, C., and Serrano,
R. , Functional expression of plant plasma membrane H+-ATPase in yeast
endoplasmic reticulum, J. Biol. Chem., 267, 12341-12349, 1992.

Bozak, K.R., O’Keefe, DR, and Christoffersen, R.E., Expression of a
ripening-related avocado (Persea americana) cytochrome P450 in yeast, plant
Physiol., 100, 1976-1981, 1992.

CHAPTER 4
EXPERIMENTS USING DESATURASE GENES IN A'T'TEMPTS TO ISOLATE

THE HYDROXYLASE GENE

ABSTRACT

It is hypothesized that oleate-12-hydroxylase is homologous to the microsomal
fatty acid desaturases, based upon three arguments. These are (a), that ricinoleic acid
is found in isolated taxa of the plant kingdom, suggesting that oleate-12-hydroxylase
has evolved recently and rapidly; (b), that biochemical studies identify similarities
between the hydroxylase and desaturases; and (c), that the active site of the
membrane-bound desaturases may contain a p-oxo bridged diiron cluster, capable of
catalysing hydroxylation as well as desaturation.

On this hypothesis, two approaches were used in attempts to clone the
hydroxylase gene using desaturase genes. The Brassica napus fad3 gene was used to
screen a castor developing md cDNA library at low stringency. One class of clones
was isolated, encoding the castor fad 7 desaturase, for which a complete sequence was
obtained.

Polymerase chain reaction was used to amplify sequences from castor
developing seed mRNA which encode a motif conserved among desaturase proteins
(GHDCGH). Fad 7 was the only desaturase or desaturase-like sequence isolated by

this approach.

82

83

It is concluded that the oleate-12—hydroxylase gene, if homologous to the
desaturase genes, is too divergent from them to be isolated using direct screening and

PCR approaches.

INTRODUCTION

A feature of the biochemistry of fatty acid modifying enzymes involved in the
biosynthesis of usual fatty acids (such as the desaturaSes) and of unusual fatty acids
(such as castor oleate-12-hydroxylase) is that they catalyse reactions in which an
unactivated C-H bond is cleaved. This cleavage is energetically demanding, and these
fatty acid modifying enzymes utilise the high oxidising power of molecular oxygen.
There are presently two known classes of enzyme cofactors capable of this type of Oz-
dependent chemistry. The haem-containing oxygenases including cytochromes P450
are one class. These enzymes have been extensively characterised in animal and
bacterial systems.

The mechanism of a hydroxylation reaction by a cytochrome P4501 might be
summarised as follows. Two electrons must be transferred to the cytochrome P450 to
effect the reaction, but this occurs as two discrete single electron transfers. Binding
of substrates to the resting state iron(III) enzyme increases the oxidation potential of
the cofactor, .potentiating the first electron transfer, leading to the iron(II) state.
Oxygen binds to the iron(II) state, giving rise to an oxygenated intermediate described

as a coordinated superoxide, Fe(III)-Oz’. Alternatively, carbon monoxide can bind to

84

the iron(Il) state, leading to the diagnostic inhibition of P450 enzymes. The second
electron transfer reduces the superoxide form of the cofactor to a peroxide form,
Fe(m-OOZ'. Catalysis is now initiated, the peroxide form undergoing spontaneous
heterolytic cleavage to generate water and the reactive intermediate Fe(V) =O. It is
thought that this reactive intermediate is stabilised by an oxidisable ligand provided by
the porphyrin ring or a tryptophan residue. Such stabilisation would be essential for
the specificty of the reaction; in its absence the highly reactive intermediate would
likely oxidise the nearest available atom in an uncontrolled fashion. Thus stabilisation
of the Fe(V) =O intermediate allows the abstraction of a specific hydrogen atom from
the substrate followed by a recombination that results in hydroxylation of the substrate
and completion of the reaction cycle.

Research leading to this understanding of the mechanism of cytochrome P450
catalysis has provided an appreciation of the thermodynamic stability of unactivated
aliphatic C-H bonds and the highly reactive oxoiron intermediates necessary for their
cleavage. Non-haem cofactors performing similar chemistry will require similarly
reactive intermediates which will also require stabilisation for a controlled reaction.

In addition, the common intermediate used by different P450 enzymes indicates that
reaction specificity must reside in the electronic and structural properties of the active
site of the protein. In other words, the differing protein environments of different
enzymes using the same cofactor results in their specific biochemistry.

We now consider the second class of cofactor known to be capable of this type

of 02-dependent chemistry. This class is less well characterised, and is typified by

85

the bacterial enzyme methane monooxygenase.1 The cofactor in the hydroxylase
component of methane monooxygenase is termed a p-oxo bridged diiron cluster
(F eOFe). The two iron atoms of the FeOFe cluster are liganded by protein-derived
nitrogen or oxygen atoms, and are tightly redox-coupled by the covalently-bridging
oxygen atom. The catalytic cycle of methane monooxygenase is not so well
understood as that of the P450 oxygenases, but there are known differences and
similarities. Rather than two discrete single-elctron reductions of the haem cofactor,
the FeOFe cluster accepts two electrons, reducing it to the diferrous state, before
oxygen binding. Upon oxygen binding, it is likely that heterolytic cleavage also
occurs, leading to a high valent oxoiron reactive species that is very smilar to that of
the haem cofactor, but stabilised by resonance rearrangements possible within the
tightly coupled FeOFe cluster, rather than through a porphyrin- or protein-derived
ligand. The stabilised high-valent oxoiron state of methane monooxygenase is capable
of proton extraction from methane, followed by oxygen transfer, giving methanol.
The FeOFe cofactor has been shown to be directly relevant to plant fatty acid
modifications by the demonstration that castor stearoyl-ACP desaturase contains this
type of cofactor.2 This desaturase is unique among the usual plant fatty acid
desaturases in being a soluble enzyme, whereas the other usual desaturases are
membrane-bound. The protein was purified from avocado fruit and a cDNA was
isolated from deve10ping castor endosperm.3 Putative iron-binding motifs have been
identified in the stearoyl-ACP desaturase primary structure by comparison to other

soluble enzymes containing the FeOFe cluster.2 These similar motifs, (D/E)-E-X-R-

86

H, are characteristically spaced approximately 90 residues apart in a number of
soluble diiron-oxo proteins, including methane monooxygenase. Since the work
described in chapter 3 was done, the first cDNA clone for a plant membrane-bound
desaturase was isolated in our laboratory,‘ encoding the microsomal w-3 desaturase of
Brassica napus. Of great interest is the identification of a similarly repeated motif in
both this sequence, the membrane—bound rat stearoyl-CoA desaturase,’ and in two
membrane-bound monooxygenasesfi7 This motif, H-X-X-H-H in the desaturases and
H-X-X-X-H-H in the monooxygenases, may be the functional equivalent in
membrane-bound FeOFe proteins of the (D/E)-E-X-R-H motif in the soluble FeOFe
proteins. Evidence in support of this view is that replacement with asparagine of any
of the histidine residues of these motifs in the rat stearoyl-CoA desaturase, abolishes
enzyme activity (J. Shanklin, personal communication). This suggests that the plant
membrane bound desaturases may also accomplish oxygen-dependent fatty acid
desaturation through an FeOFe cofactor. All membrane-bound plant desaturases
subsequently sequenced also contain the conserved histidine-rich repeats (below, and
references 8,9).

Of the well-characterised FeOFe—containing enzymes, methane monooxygenase
catalyses a reaction involving oxygen-atom transfer (CH. -. CH3OH), while the
FeOFe cluster of ribonucleotide reductase catalyses the oxidation of tyrosine to form a
tyrosyl cation radical without oxygen-atom transfer. However, site—directed
mutagenesis of Phe208 to Tyr resulted in the conversion of this enzyme to an oxygen

transfer catalyst, Tyr208 being hydroxylated and shown to be acting as a ligand to

87

one iron of the FeOFe cluster.

The argument made for the P450 oxygenases catalysing a range of reactions
through the use of the same reactive intermediate modulated by the electronic and
structural environment provided by the protein, might also be applied to FeOFe-
containing enzymes. Modifications of the active site of plant fatty acid oxidising
enzymes containing FeOFe clusters could thus alter the outcome of the reaction,

including whether oxygen-atom transfer occurs or not.

Hypothesis For the Origin of Oleate-IZ-Hydraxylase

An hypothesis is now proposed for the origin of castor oleate-12-hydroxylase,
based upon three arguments. The first argument involves the taxonomic distribution
of plants containing ricinoleic acid, mentioned in chapter 1. Ricinoleic acid has been
found in 12 genera of 10 families of higher plants.”21 Thus, plants in which
ricinoleic acid occurs are found throughout the plant kingdom, yet close relatives of
these plants do not contain the unusual fatty acid. This pattern suggests that the
ability to synthesize ricinoleic acid has arisen several times independently, and is
therefore a quite recent divergence. In other words, the ability to synthesize
ricinoleic acid has evolved rapidly, suggesting that a relatively minor genetic change
was necessary to accomplish it. Two mechanisms for such facile evolution of a new
enzyme activity are envisaged. One mechanism would be for the modification of a
gene normally encoding a fatty acid hydroxylase active in the epidermis and involved

in the synthesis of a hydroxy-fatty acid cutin monomer. The other mechanism would

88

be for modification of a gene encoding a microsomal fatty acid desaturase, such that
instead of performing one type of oxidation reaction (desaturation) it now performs
another (hydroxylation).

The second argument is that many biochemical properties of castor oleate-l2-
hydroxylase are similar to those of the microsomal desaturases, as discussed in
chapter 2. Furthermore, biochemical studies of the hydroxylase also suggest that it is
not a P450 enzyme. The fatty acid hydroxylases known to be involved in synthesis of
cutin monomers are cytochromes P450.22

The third argument stems from the discussion of oxygenase cofactors above, in
which it is suggested that the plant membrane bound fatty acid desaturases may have a
p-oxo bridged diiron cluster-type cofactor, and that such cofactors are capable of
catalysing both fatty acid desaturations and hydroxylations, depending upon the
electronic and structural properties of the protein active site.

Taking these three arguments together, it is now hypothesized that oleate-12-
hydroxylase of castor endosperm is homologous to the microsomal oleate A12
desaturase found in all plants. This is in accord with the apparently rapid evolution
of the hydroxylase gene. However, it should be noted that this hypothesis is proposed
specifically for the castor enzyme, for which biochemical studies have shown the
similarity of many properties of the hydroxylase and desaturases. Other species
accumulating ricinoleic acid, for which no biochemical studies are available, may
have evolved an hydroxylase by a different route, such as by modification of an

existing cytochrome P450 gene.

89
The above hypothesis proposes that the oleate-l2-hydroxylase gene is

homologous to the gene of the biochemically most-similar desaturase, the microsomal
A12 desaturase. However, it was thought at this time that this gene, in turn, was
probably homologous to the other microsomal (to-3) desaturase in particular, and
possibly also to the equivalent desaturases of the chloroplast inner envelope. A gene
(fad3) had recently been isolated in our laboratory for the microsomal w—3 desaturase
from Brassica napus,4 hence I sought to use this clone in various attempts to isolate
homologous genes from castOr, one of which might encode the oleate-12-hydroxylase.
These experiments are the subject of this chapter. First, a castor genomic Southern
blot is probed with the fad3 clone to confirm that a number of homologous sequences
exist in castor. Attempts are then made to isolate these by direct screening and

polymerase chain reaction (PCR) approaches.

MATERIALS AND NIETHODS

Southern Analysis of Castor Genomic DNA Using the fad3 Probe

The Brassica napus fad3 clone was labelled by random priming23 and purified
of unincorporated nucleotides by ethanol precipitation in the presence of ammonium
acetate.23 A Southern blot of digested castor genomic DNA was generously provided
by Dr. J. Shanklin’. The blot was prehybridised at 52°C in a solution containing 4 X

SET (0.6 M NaCl, 0.12 M Tris—HCl pH 7.4, 8 mM EDTA), 0.1% sodium

90
pyrophosphate, 0.2% SDS, and 100 pg ml" heparin.“ The probe was hybridised to

the blot at 52°C overnight in the same solution, except for the addition of 10%
dextran sulphate. The blot was washed three times in 2 X SSC, 0.1% SDS at 52°C,

then exposed to X-ray film.

Screening of cDNA Library with fad3 Probe

Five plates of each of the four pools of the pYES2.0 cDNA library (chapter 3)
were grown until small colonies (~105 total) were visible. A nitrocellulose filter
(Schleicher & Schiill BA85) was laid on each plate, its position marked, and lifted off
to a fresh plate, the adhering colonies now facing upwards. Care was necessary that
both plate and filter were not too moist, to avoid smearing of the colonies. The
original plate was incubated for a few h at 37°C to recover colonies, while the filters
were processed as folllows. Each filter was sequentially placed, colony side up, on
Whatman 3MM paper moist with 10% SDS (3 min), denaturing solution (0.5 M
NaOH, 1.5 M NaCl; 5 min), neutralising solution (0.5 M Tris-Cl pH 7.4, 1.5 M
NaCl; 5 min), and 2 X SSC (0.3 M NaCl, 0.03 M Na-Citrate, pH 7.0). The filters
were then air-dried for ca. 1 min before pressing twice between sheets of filter paper
to remove cell debris. After air-drying a further 30 min, DNA was fixed to the
filters by baking in vacuo at 80°C for 1-2 h.

A probe was prepared in the same manner as for the Southern blot described

above, and hybridised to the filters overnight at 55°C in a solution with the same

91

composition as the prehybridisation solution described above for the Southern blot.
The filters were washed at 55 °C in the solution described above and exposed to X-ray
film. Hybridizing colonies were picked and replated for a second round of
hybridization, yielding positive clones. Minipreparations of DNA from these clones
was simultaneously digested with BamHl , EcoRl , and Xhal , releasing and possibly
fragmenting the insert of each clone, and electrophoresed through agarose gels. A
Southern blot was prepared from these gels as follows. DNA in the gels was
denatured by incubation for ~1 h in 0.5 M NaOH, 1.5 M NaCl, and then the gels
were neutralised by two 30 min incubations in 0.5 M Tris-Cl pH 7, 1.5 M NaCl.
DNA from the gels was transferred overnight onto a nylon membrane (Hybond N,
Amersham) by elution from the gel in 10 X SSC (1.5 M NaCl, 0.15 M Na-citrate,
pH 7.0). After air-drying for 1 h, DNA was fixed to the membrane by exposing to a
UV light source for 2 min (based upon an empirical calibration). The fad3 probe was

prepared and hybridised (at 55°C) to the Southern blot as described above.

Generation of Nested Deletions and DNA Sequencing

Plasmid DNA was prepared from large (500 ml) cultures of E. coli harbouring
pFLl or pFLlr by the alkaline lysis technique and purified by banding in CsCl
density gradients as described.23 Nested sets of deletions were generated from these
clones by the exonuclease III protocol.23 The enzymes used to digest the plasmid

DNA were BamHI and Kpnl . Deletion clones varying in size by approximately 250

92
bp were sequenced. Plasmid DNA for sequencing was prepared from overnight

cultures (5 ml, LB medium containing 100 mg l" ampicillin) using "Magic
Minipreps" (Promega) according to the instructions of the manufacturer, and
submitted for automated cycle sequencing on Applied Biosystems 373A instruments
using the fluorescent primer T7. Sequence data was aligned, assembled and analysed

with the DNASIS and PROSIS computer programs.

Reverse Transcription-Polymerase Chain Reaction

Poly(A)+ RNA from developing castor endosperm was that used for
construction of the pYE82.0 cDNA library (chapter 3). Leaf poly(A)+ RNA was
purified by the same technique (chapter 3) from small true leaves of seedling plants,
from which the midribs were removed.

Oligonucleotide primers used in these experiments were as follows:

RlTls: 5’ GAC ATC GAT AAT ACT 'I'IT TIT T'l’l‘ T'l'l‘ 'IT 3’
R1: 5’ GAC ATC GAT AAT AC 3’
GO: 5’ GGN CA(C/T) GA(C/T) TG(C/T) GGN CA 3’.

Reverse transcription reactions were as follows. RNA (2 pg, 10.25 pl) was
heated to 65°C for 3 min, then cooled on ice, and added to a tube containing the
other reaction components:

4 p15x reverse transcriptase buffer (BRL)

2 pl 100 mM dithiothreitol

93
0.25 pl (~10 units) RNAsin (Promega)

0.5 pl (2.5 pmol) RlTls primer
2 pl 10 mM each dNTP.
The reaction was initiated by addition of 1 pl (200 units) reverse transcriptase (BRL,
from murine Maloney leukemia virus) and incubated at 37°C for 2 h. The reaction
was then diluted with 980 p1 TE and stored at 4°C.

PCR reactions contained 250 ng each primer (R1, GO), 0.2 mM each dNTP, 7
mM MgC12, Taq polymerase buffer (Promega). Taq polymerase (2.5 units, Promega)
and first-strand cDNA template (5 pl) were added, bringing total volume to 50 p1, at
75°C to prevent mis-priming at low temperature. Cycling was essentially according
to Frohman et al”: first cycle: 95°C for 5 min, 50°C 2 min, 72°C 40 min; second
cycle: 95°C 45s, 50°C 25 s, 72°C 3 min (with addition of 2.5 units additional Taq
polymerase); followed by twenty-nine identical cycles (no further enzyme addition)

and a final 15 min extension at 72°C.

RESULTS AND DISCUSSION

Southern Analysis of Castor Genomic DNA Using the fad3 Probe

A genomic castor Southern blot was probed with the fad3 clone from Brassica
napus, at moderately low stringency (52°C). Approximately 9 hybridizing bands
could be detected in an EcaRl digest (Figure 9), indicating that at least several genes

in the castor genome share homology with the fad3 probe, as expected.

94

 

 

' 23 .—
34 _ _,
.6 — 532.“.
4.4 _ w
2.3 _ ‘
2.0 —
0.56 —

 

 

 

Figure 9. A Southern blot of castor genomic DNA digested with restriction enzymes
EcaRI (E) or HindIII (H), was hybridised at moderately low stringency (52°C, 4 X
SET) with the 32P-labelled Brassica napus fad3 cDNA. Migration of DNA standards
(kb) is shown to the left.

95
Screening of cDNA Library with fad3 Probe

Approximately 105 colonies of the pYES2.0 cDNA library, constructed in
chapter 3, were screened with the Brassica napus fad3 clone. The hybridization
temperature was slightly higher (55 °C) than that used for the Southern blot (52°C).
Low stringency screening of E. coli colonies was frequently problematic, binding of
the probe to bacterial debris remaining on the filters frequently leading to
unacceptably high background levels. Twenty-five apparent positives were re-
screened, giving seven apparently true positives. However, when the plasmid DNA
of these clones was digested and analysed by Southern blot, the fad3 probe hybridised
specifically to the insert DNA of only three clones. These three clones gave the same
restriction pattern but varied in total length. The longest clone, isolate 3ei,b, was
approximately 2.0 kb, designated pFLl , and selected for full-length sequencing. The
insert of pFLl was excised with Eagl , which cuts on both sides of the polylinker, and
religated. A clone in which the insert was now in the opposite orientation to pFLl
was selected and designated pFLlr. A deletion series was generated from both pFLl
and pFLlr, such that by sequencing with the T7 primer the complete sequence of both
strands was obtained. The clone is 1958 nucleotides in length and consists of a 344
nucleotide 5’ untranslated region, a 1380 nucleotide open reading frame, and a 234
nucleotide 3’untranslated region. The possibility that the unusually large 5’
untranslated region is due to some cloning artifact was not investigated. The open

reading frame encodes a 460 amino acid protein with a calculated M, of 52558.

96

While this work was in progress, additional desaturase clones were being isolated by
other members of the laboratory, including the fad 7 gene of Arabidapsis thaliana.
This gene encodes a plastid 00-3 desaturase.9 The gene of a membrane-bound A12
desaturase (desA) from the cyanobacterium Synechacoccus was also known.”5
Comparison of the pFLl sequence with these other sequences is shown schematically
in Figure 10 (see also Figure 11). At the nucleotide level, pFLl shares the highest
sequence similarity to a central region of the fad 7 sequence, and respectively lower
similarities to the fad3 and desA sequences in approximately the same region. There
is an additional region of moderate similarity between the pFLl sequence and the 5’
end of the fad 7 sequence, not found for fad3 or desA. These results are obtained also
by comparison at the amino acid level, similarity being highest to fad 7 and also
extending further toward the amino terminus than for either the fad3 or desA
sequences. Thus, the amino-terminus of the predicted castor protein has an extension
similar to that of the fad 7 protein but not found in the fad3 protein. This amino-
terminal extension is rich in the hydroxy amino acids serine and threonine (23% of
the first 78 residues), a characteristic feature of the transit peptide of plastid
proteins.27 These results suggest that pFLl encodes a plastid w—3 desaturase
homologous to fad 7. This sequence has been deposited in Genbank (accession
L25 897), and a brief report submitted for publication.” Other 03-3 desaturase genes
have been reported recently.8 i

The sequence similarity between the w-3 desaturases demonstrates the high

degree of conservation between the endoplasmic reticulum and plastid forms. This

97

 

desA

 

 

hut?

 

 

 

 

 

Fad7

 

 

 

Fad3

 

 

DesA

 

 

 

Figure 10. Sequence similarity between pFLl (centre) and other membrane-bound
desaturases at the nucleotide (above) and amino acid (below) levels. The pFLl coding
sequence and untranslated regions are 'shown as solid and checked bars, respectively.
The percent sequence similarity between pFLl and the other genes is shown in the
shaded regions over which the similarity was averaged.

98

was not the expected result when this experiment was embarked upon. The same
result has also been obtained by other members of the laboratory, that is, that
desaturases acting at the w-3 position are conserved among themselves, but do not
serve as heterologous probes for those acting at the A12 position, even when probing
within a species (A rabidapsis) (S. Gibson, personal communication). The evolution
of the desaturases, therefore, appears to have involved an earlier differentiation
between the A12 and co-3 enzymes, and a subsequent differentiation for each type
giving rise to the microsomal and plastid forms. Since it is hypothesized that the
castor oleate-l2-hydroxylase gene is homologous to the microsomal A12 desaturase,
only a desaturase probe capable of isolating the microsomal A12 desaturase gene
should also be a possible probe for the hydroxylase. Our experiments show that fad3
is not such a probe. In addition, one screening experiment was done with the
Arabidapsis clone GO313. This desaturase homologue was isolated by Dr. S. Gibson
in our laboratory, and is tentatively assigned as the gene (fad6) for the plastid A12
desaturase. Of 80 000 lambda phage clones screened from the XZAPST library
(described in chapter 5) at 52 °C, no duplicate positives were obtained. These results
indicate that an alternative strategy is required for cloning the hydroxylase on the

basis of possible homology to desaturases.

 

99

 

10 20 30 40 50
RCFAD7 1 MAAGUVLSEC GLRPLPRIYS RPRIGFTSKT TNLLKLRELP DSKSYNLCSS 50
ATFADT -1 .H‘NL****‘ *1'******T T**SN*L*NN NKFRPSLSSS SY*TSSSPL* 49
“F”; -n 0.0.0.0... 00.0.00... 0.0.0.0... 0.00.00... I... ...... -30
SDESA -86 .................................................. -37
60 70 80 90 100
RCFADT 51 FKVSSUSNSK OSNUALNVAV PVNVSTVSGE DDREREEFNG IVNVDEGKGE 100
ATFAD7 50 *GLN*RDGFI R-*****‘ST* ----------- LTTP1*EE -SPLE*DNKO 99
BNFAD3 -29 .......... . ............ ......H VVAHDORS*V NflflSGAR*E* 21
”E“ -“ 0.0.0.0... 0.00.09... 0.0.0.0... OIOOIOHIA1*PPL1PTVTP 1‘
110 120 130 140' 150
RCFADT 101 FFDAGAPPPF ILADIRAAIP KHCHVKNPHR SHSYVLRDVV VVFGLAAVAA 150
ATFAD7 10° Reapaataaa "aaaaaaaaa aanaaaaaar oLattvtaaA laaAaaaGat 149
BNFAD3 22 Gaepstgaac (Icoaaaaaa ititttstLt anteater}; A'AA‘*MA*V 71
SDESA 15 SNPDRPIADL K*O**1KTL* *E'FE*KASK AHAS**ITLG Al-AVGYLGI 64
160 170 190 200
RCFAD? 151 YFNNHVAHPL YHFCOGTHFH ALFVLGHDCG HGSFSNNPKL NSVVGHLLHS 200
ATFAD7 150 *Laaalvaaa «*LAaaaaat fl aaaaaaoaaa ***a****** 199
BNFAD3 72 aaosafLaaa aaVAaaaLaa ’ aaaaojaLa eternal... 121
SDESA 65 lYLP*YCL*1 T*lHT**ALT *R**AKKRUV *0L***1AFA 114
210 220 230 240 250
RCFAD7 201 SILVPYHGHR GHVENDESHH PLSEKIFKSL DNVTKTLRFS 250
ATFAD7 200 aaaaaaaaaa aaaaeaaaat antaaayuTa *KP*RFF**T 249
BNFAD3 122 Fataaaatat taaaataaav tapaaLyaua paganflaay1 171
SDESA 115 PL1Y*F*S** NK1*V*NA*D *H'VEA*0AS PAl---V*-- 164
260 270 280 290 300
RCFAD? 251 LPFPHLAYPF YLHSRSPGKK GSHFHPDSGL FVPKERKDI- ITSTACHTAH 300
ATFAD7 250 «*Lvaaaaaa aaaAaaaawa taayaaaaoa «Laaaaaav- Laaaaaaaaa 299
BNFAD3 172 viLtittttI aaaytaaaafi *titutyist .AasaaaLa- AaaaTaasja 221
SDESA 165 *FYRAIRG** --*HTGSIFH H*LH*FKLSN *AORD*NKVK LSIAVVFLFA 214
310 320 330 340 350
RCFAD7 301 AALLVYLNFS HGPVOMLKLY GlPYUlFVHw LDFVTYLHNN GHEDKLPHYR 350
ATFAD7 300 aaaaacaaaT [atlaaataa *tttttuttt tttttttttt *ttttttttt 349
BNFAD3 222 Lataaaasal VD**TV**V* *vitlittit aaAaaaaata aaosaaaaat 271
SDESA 215 *1AFPA’11T T*VUGFV*FU LH*ULVYHF* HSTF*IV**1 lP'lRF--RP 264
360 370 380 ................. 390 400
RCFADT 351 GKAHSYLRGG L-TTLDRDYG -H1NN1HHDI GlhfijflfiLFP OIPHYHLVEA 400
ATFAD7 350 **E******* a-aaaaaaaa -Laaaataat tagggggaaa aaaattaaae 399
BNFADS 272 aaeaaaaaaa a.aalaaaaa -1faaataaa aafffi§§taa ********D* 321
suesa 265 MD**MEAO *NG*VHC**P n*vsvrc*** uv‘ififfisv A**s*rt*m.* 311.
410 420 430 440 450
RCFADT 401 TEAAKPVHGK YYREPKKSGP LPLHLLGSLV RSHKEDHYVS DTGDVVYYOK 450
ATFAD7 400 aatnaaaLaa aaaaaoaaaa *aaaaasltA Kalaaaaaaa aEaEaaaa(A 449
BNFAD3 322 aaataflaLan watercraaA It‘ttvettt Aalaxaaaat *titltFtET 371
SDESA 315 NGSL*ENH*P FLY*R1FNHO *HOOIS*OCN LYDP*HG*RT FGSLKKV*.. 364
460 470 480 490 500
“CF”? ‘51 DPKLSGIGGE KTE. oooooooooooooooo coon-.0... no ..... .0. 5m
AtFm7 ‘50 *'“.Y*EVK. 0.0.0.0... 000......- OOLIOOI... .......... ‘”
BNFAD3 372 **D*YVYASO *SKll ................................... 421
”BSA “5 0.0.0.0... 0.0.0.0... 00.0.00... 0 IIIIII IO. ..... .0... ‘1‘

Figure 11. Comparison of the deduced amino acid sequences of four membrane
bound desaturase genes: RCFAD7 Ricinus communis Fad7 (pFLl , above); ATFAD7
Arabidapsis thaliana Fad7;9 BNFAD3 Brassica napus Fad3;‘ SDESA Synechacaccus
DesA.26 The conserved GHDCGH and HXXHH motifs are shaded.

100
Amplification of Sequences With Conserved Desaturase Motifs by Polymerase

Chain Reaction

A comparison of different desaturase genes (Figure 11) reveals a short stretch
of perfectly conserved amino acids (GHDCGH). This sequence is found in the
sameposition in the cyanobacterial A12 desaturase desA as well as the plant w-3
desaturases fad3 (from Brassica napus) and fad 7 (from both Arabidapsis thaliana and
Ricinus communis). This suggested that this sequence may be directly involved in an
aspect of the enzyme function (such as substrate binding), and might, therefore, be
conserved among all desaturases and functional homologues, including the oleate-12-
hydroxylase. An experiment was, therefore, designed to isolate sequences expressed
in the developing castor endosperm that contain this conserved motif, using the
polymerase chain reaction (PCR). Furthermore, ricinoleic acid, typical of unusual
fatty acids (chapter 1), is specific to the seed tissue of castor, and is not found in
vegetative tissues or synthesized in germinating seeds.”32 By comparing sequences
amplified from both leaf and seed tissue, I sought to identify the hydroxylase in
particular. Since it was known from the preceding experiment that the castor fad 7
homologue is expressed in the seed and contains this motif, a control for the
experiment was that it should be possible at least to re-isolate this sequence.

Poly(A)+ RNA purified from leaves and developing endosperm of castor was
used as a template for reverse transcription. The reaction was primed with the

Oligonucleotide RlT1 5. A degenerate Oligonucleotide ("Golden Oligo" , GO) was

101
designed corresponding to all possible codons for the conserved motif GHDCGH.

PCR reactions were then initiated using the first-strand cDNAs as templates and R,
and GO as primers. The expected product size for an amplified desaturase fragment
is 700-900 bp. Reaction conditions, particularly annealing temperature and
magnesium concentration, were optimised to give a comprehensive range of products
when analysed by agarose gel electrophoresis, and dependency of amplification upon
both primers. Two particular difficulties were encountered. Gels of the PCR
products were invariably smeared due to an unknown problem, making it difficult to
identify individual bands. Secondly, where putative seed-specific bands were
identified and isolated from the gel, they were completely resistant to cloning; all
clones isolated from these ligations being of a different size than expected from the
band size isolated from the gel, and invariably containing extraneous (bacterial or
lambda phage) DNA. Hence an alternative approach was taken, simply cloning the
seed-RNA PCR products directly, followed by analysis of the clones obtained.
Difficulties were also encountered with these ligations and transformations, but a total
of eleven clones apparently containing bona fide plant DNA could be isolated and
were submitted for terminal sequencing, summarised in Table 6. The primer
sequence used for 3 ’ end amplification could be identified in eight of the eleven
clones, and a partial primer sequence (matching 14 nucleotides of the 3’ end of the
l7-nucleotide primer) was identified in one further sequence. Since the PCR products
were not directionally cloned, were sequenced only from one end, and in some cases

appear to be too large to sequence through in a single sequencing run, it was not

102

Table 6. Identification (blastx database search result) of sequences cloned by PCR
amplification using a sequence conserved among desaturases.

 

Clone

 

pIIIH6
pIB9
pIEl
pIF7
pIH2
pIIC6
pIIF2
pIIIA4

pIBS

p12-c

Sequence #

 

CRS 107

CRS 108

CRS109

CRSllO

CR8112

CRSl 14

CRSll6

CRS238

CRS243

CR8247

CRS254

Highest Blast Score

 

52

59

269

Non-structural polyprotein, rabbit hemorrhagic
disease virus.

35 kDa major secreted protein precursor,
vaccinia virus.

Stress-inducible protein ST'135, Fusariurn
salani.

Same sequence as above (CRSlO9).

77

367

ORF515 gene product, Pinus cantarta
chloroplast.
60S ribosomal protein L5A, Xenapus laevis.

Same sequence as CR8108, above.

172

49

61

59

Linoleic acid desaturase, Brassica napus.

N ADH-ubiquinone oxidoreductase chain 4,
Paramecium tetraurelia.

Respiratory nitrate reductase 2 gamma chain,
Escherichia coli.

DL-hydantoinase (hyuR) gene, Pseudamanas

sp.

 

103
expected that the primer sequence would be identified in every case. The

identification of the primer sequence in many of the clones suggests that castor
cDNAs containing the primer sequence were indeed amplified.

DNA sequences were compared in all reading frames to the non-redundant
translated-nucleotide and protein sequence databases (swissprot, PIR, GenBank, and
EMBL) by the program blastx.33 Blast scores lower than 80 are considered here to be
insignificant, the corresponding sequences therefore having no matches in the
databases. Most of the clones thus contain unidentified DNA. Three clones
contained sequences similar to some other sequence in the databases. One of these,
pIIIA4, was similar to the fad3 desaturase of Brassica napus. In fact, this sequence is
identical to a portion of the fad 7 sequence isolated from castor in the preceding
experiment.

Thus, it is possible, as predicted in the introduction to this experiment, to
isolate at least the fad7 desaturase sequence from castor by this PCR approach using a
conserved desaturase motif. However, no new desaturase or desaturase-like sequence
was obtained in this experiment. The developing castor endosperm contains relatively
small amounts of linolenic acid, the product of the fad 7 desaturase, hence high mRN A
levels for this particular desaturase are not expected in this tissue. Indeed, in the
preceding experiment, only three fad 7 clones were isolated from approximately 10’
colonies of the cDNA library made from the same RNA. Based on the much higher
activity of oleate-lZ-hydroxylase in the developing endosperm, and upon the

hypothesis that hydroxylase and desaturase turnover numbers should be similar (see

104
chapter 3), it was expected that mRNA for oleate-12-hydroxylase should be more

abundant than that for fad 7. Thus the fact that only the desaturase was cloned by the
PCR approach suggests that the conserved motif (GHDCGH) may not be highly
conserved in the hydroxylase even if the hydroxylase is homologous to the
desaturases.

Sequencing of other desaturase genes in our and other laboratories following
the completion of this experiment supports this possibility. The Arabidapsis clone
GO313 isolated by Dr. S. Gibson, tentatively identified as the plastid A12 desaturase,
contains instead the divergent sequence GHDCAH, while an unidentified desaturase-
homologue cloned from parsley in the laboratory of Dr. K. Hahlbrock, ELI72,
contains instead the divergent sequence GHECDH (I. Sommsich, personal

communication).
CONCLUSIONS

Attempts described here to isolate a clone for oleate-l2-hydroxylase based
upon hypothesized homology between this enzyme and microsomal desaturases,
resulted instead in the isolation of a plastid desaturase gene. The sequence of this
gene, fad 7, demonstrated that genes encoding the same enzyme are strongly conserved
between species (such as between castor and Arabidapsis). However, sequence
similarity between different desaturases appears too weak in some cases to allow

direct screening approaches. Thus it has not been possible to use the desaturase gene

105
fad3 as a probe for the isolation of the gene fad2. Likewise, in the experiments

described here, it was not possible to isolate a clone for oleate-12-hydroxylase using
fad3 as a probe. On the basis of these results, it is not possible to distinguish whether
oleate-lZ-hydroxylase is not homologous to the desaturases (contrary to the
hypothesis), or whether they are homologous but considerably divergent. On the

basis of the divergence observed among desaturases, the latter possibility is favoured.

REFERENCES

1. van de Loo, F.J., Fox, B.G., and Somerville, C., Unusual fatty acids, in
Lipid Metabolism in Plants, T.S. Moore Jr. , Ed. , CRC Press, Boca Raton,
1993, chap 3.

2. Fox B.G., Shanklin, J ., Somerville, C., and Munck, E., Stearoyl-acyl carrier

protein A9 desaturase from Ricinus communis is a diiron-oxo protein, Proc.
Natl. Acad. Sci, 90, 2486-2490, 1993.

3. Shanklin, J. and Somerville, C. , Stearoyl-acyl-carrier—protein desaturase
from higher plants is structurally unrelated to the animal and fungal
homologs, Proc. Natl. Acad. Sci. USA, 88, 2510, 1991.

4. Arondel, V., Lemieux, B., Hwang, 1., Gibson, 8., Goodman, H.M., and
Somerville, C.R. , Map-based cloning of a gene controlling omega-3 fatty acid
desaturation in Arabidapsis, Science, 258, 1353-1355, 1992.

5. Thiede, M.A., Ozols, J., and Strittmatter, P., Construction and sequence of
cDNA for rat liver stearoyl coenzyme A desaturase, J. Biol. Chem. , 261,
13230-13235, 1986. '

6. Kok, M., Oldenhuis, R., van der Linden, M.P.G., Raatjes, P., Kingma, J .,
van Lelyveld, P.H. , and Witholt, B. , The Pseudamanas aleavarans alkane

hydroxylase gene: sequence and expression, J. Biol. Grem. , 264, 5435-5441 ,
1989.

7. Suzuki, M., Hayakawa, T., Shaw, J.P., Rekik, M., Harayama, 8., Primary
structure of xylene monooxygenase: similarities to and differences from the

10.

ll.

12.

13.

14.

15.

16.

17.

18.

106
alkane hydroxylation system, J. Bacterial., 173, 1690—1695, 1991.

Yadav, N.S., Wierzbicki, A., Aegerter, M., Caster, C.S., Perez-Gran, L.,
Kinney, A.I., Hitz, W.D., Booth, R., Schweiger, B., Stecca, K.L., Allen,
S.M., Blackwell, M., Reiter, R.S., Carlson, T.J., Russell, S.H., Feldmann,
K.A., Pierce, J ., and Browse, J., Cloning of higher plant w3 fatty acid
desaturases, Plant Physiol., 103, 467-476, 1993.

Iba, K., Gibson, 8., Nishiuchi, T., Fuse, T., Nishimura, M., Arondel, V.,
Hugly, S. , and Somerville, C., A gene encoding a chloroplast omega-3 fatty
acid desaturase complements alterations in fatty acid desaturation and
chloroplast c0py number of the fad 7 mutant of Arabidapsis thaliana, J. Biol.
Chem., in press.

Brieskom, C.H., and Kabelitz, L., Hydroxyfettsauren aus dem cutin
des blattes von Rosmarinus officinalis, Phytochemistry, 10, 3195, 1971.

Green, A.G., The occurrence of ricinoleic acid in Linum seed oils, J.
Am. Oil Chem. Soc., 61, 939, 1984.

Garg, S. P. ,,Sherwani M. R. K. ,A,rora A. ,Agarwal, R. ,and Ahmad,
M. Ri,cinoleic acid 1n Vinca rosea seed oil, J. Oil Technol. Assoc.
India, 19, 63, 1987.

Ahmad, M.U., Husain, S.K., and Osman, S.M., Ricinoleic acid in
Phyllanthus niruri seed oil, J. Am. Oil Chem. Soc., 58, 673, 1981.

Daulatabad, C.D., and Mirajkar, A.M., Ricinoleic acid in Artocarpus
integrifolia seed oil, J. Am. Oil Chem. Soc., 66, 1631, 1989.

Daulatabad, C.D., Desai, V.A., Hosamani, K.M., and Jamkhandi,
A.M., Novel fatty acids in Azima tetracantha seed oil, J. Am. Oil
Chem. Soc., 68, 978, 1991.

Daulatabad, C.D. and Hosamani, K.M., Unusual fatty acids in

Brunfelsia americana seed oil: a rich source of oil, J. Am. Oil Chem.
Soc., 68, 608, 1991.

Yunusova, S.G., Gusakova, S.D., Glushenkova, A.I., Lipids from the
seeds of Securinega suffruticosa, [Grim Prir. Soedin. (Tashk.), 3, 277,
1986.

Kleiman, R., Spencer, G.F., Earle, F.R., Nieschlag, H.J., and
Barclay, A.S. , Tetra-acid triglycerides containing a new hydroxy

19.

20.

21.

22.

23.

24.

25 .

26.

27.

28.

29.

30.

107

eicosadienoyl moiety in Lesquerella auriculata seed oil, Lipids, 7, 660,
1972.

Sreenivasan, B., Kamath, N.R., and Kane, J.G., Studies on castor oil.
1. Fatty acid composition of castor oil, J. Am. Oil 01cm Soc., 33, 61,
1956.

Badami, R.C., and Kudari, S.M., Analysis of Hiptage madablota seed
oil, J. Sci. Food Agric., 21, 248, 1970.

Nikolin, A., Nikolin, B., and Jankovic, M., Ipopurpuroside, a new
glycoside from Ipomoea purpurea, Phytochemistry, 17, 451, 1978.

von Wettstein-Knowles, P.M., Waxes, cutin, and suberin, in Lipid Metabolism
in Plants, T.S. Moore Jr., Ed., CRC Press, Boca Raton, 1993, 127-166.

Sambrook, J ., Fritsch, E.F., and Maniatis, T., Molecular Cloning: a
Laboratory Manual, 2nd ed., Cold Spring Harbor Laboratory Press, 1989.

Singh, 1., and Jones, K.W., The use of heparin as a simple and cost-effective
means of controlling background in nucleic acid hybridization procedures,
Nucleic Acids Res., 12, 5627-5630, 1984.

Frohman, M.A., Dush, M.K., and Martin, G.R., Rapid production of full-
length cDNAs from rare transcripts: amplification using 8 single gene-specific
Oligonucleotide primer, Prac. Natl. Acad. Sci. USA, 85, 8998-9002, 1988.

Wada, H., Gombos, Z., and Murata, N., Enhancement of chilling tolerance of

a cyanobacterium by genetic manipulation of fatty acid desaturation, Nature,
347, 200-203, 1990.

Keegstra, K.,. Olsen, L.J., and Theg, S.M., Chloroplastic precursors and their
transport across the envelope membranes, Ann. Rev. Plant Physiol. Plant Mol.
Biol., 40, 471-501, 1989.

van de Loo, EL, and Somerville, C., Plastid w-3 fatty acid desaturase cDNA
from Ricinus communis, Plant Physiol. , submitted.

Coppens, N., Biosynthesis of fatty acids by seeds of Ricinus communis,
Nature, 177, 279, 1956.

Canvin, D.T., Formation of oil in the seed of Ricinus communis L., Can. J.
Biochem. Physiol., 41, 1879-1885, 1963.

31.

32.

33.

108

Yamada, M., and Stumpf, P.K., Enzymic synthesis of ricinoleic acid by
extracts of developing Ricinus communis L. seeds, Biochem. Biophys. Res.
Cammun., 14, 165-171, 1964.

James, A.T. Hadaway, H.C., and Webb, J.P.W., The biosynthesis of
ricinoleic acid, Biochem. J., 95 , 448-452, 1965 .

Altschul, S.P., Gish, W., Miller, W., Myers, E.W., and Lipman, D.J., Basic
loeal alignment search tool, J. Mol. Biol., 215, 403-410, 1990.

CHAPTER 5

LARGE SCALE SEQUENCING OF SEED-SPECIFIC CASTOR CLONES

ABSTRACT

Clones of a castor developing seed library were differentially screened for
those not expressed in leaves, and not expressed in seeds at very high levels.
Terminal (5’) sequence data were obtained by automated DNA sequencing for 526
such clones. Of these ESTs generated, 468 contained informative sequence, and 'a
putative identification could be made for 46% of them by comparison to sequence
databases using the program Blastx.

Among the clones identified were two clones of a fatty acid desaturase
homologue, which appeared to encode a novel desaturase or desaturase-like protein.
These are considered putative oleate-12-hydroxylase clones.

In addition, a range of other novel plant genes were identified, and the

possible functions of these genes in the developing castor seed are discussed.

INTRODUCTION

Recent advances in technology accruing from the increased interest in genome
sequencing have made it possible to sequence DNA rapidly and voluminously using

automated instruments. By automation, it has become possible to do experiments

109

110

involving large amounts of DNA sequencing, where previously these experiments
were prohibitively demanding of investigator time. This technology was exploited in
the continuing quest for a gene encoding the oleate-12-hydroxylase of castor.

The design of this large-scale DNA sequencing project was based upon three
considerations. Firstly, ricinoleic acid is specific to the seed tissue of castor, and is
not found in vegetative tissues or synthesized in germinating seeds.“ A differential
screening approach was therefore used to sequence only clones that were seed-
specific. Differential screening has been used with success to isolate induced genes,
particularly genes expressed in a given tissue following treatment with a drug,
environmental variable, etc, but not expressed in the same tissue at the same
developmental stage without treatment (for example, see Haj ela et als). In these cases
a small number of induced sequences can be detected which may have direct
relevance to the treatment. The differential screen used here is qualitatively different
from the usual approach, in that the genes expressed in two completely different
organs are compared, and a much larger proportion of sequences should differ
between the two (discussed further below).

Secondly, it is hypothesized that mRNA for the oleate-12-hydroxylase is
moderately abundant in the mRN A pool in developing endosperm. This is based
upon the same considerations outlined above (chapter 3), namely that the enzyme is
very active in this tissue, and probably does not have an extraordinarily high turnover
number. In fact it is estimated that, to a first approximation (disregarding possible

subtleties of gene regulation), the mRN A abundance should be similar to that for the

lll

stearoyl-ACP-desaturase. These enzymes are involved in the same pathway, perform
similar fatty acid oxidations and, it is hypothesized (chapter 4), that they do so with
the same cofactor, hence might have similar turnover numbers. Clones for the
stearoyl-ACP desaturase have been isolated from developing castor endosperm cDN A
libraries at frequencies of 1/3000 (chapter 3) to l/ 1250 (below). Furthermore, in this
experiment identification of cDNA clones was made on the basis of their sequence,
not by functional expression as in chapter 3. This considerably reduces the number of
clones which must be examined in this experiment, since they nwd not be full-length
or in a particular orientation. Indeed, partial cDNA clones were preferable, ensuring
that DNA sequence obtained was from the coding sequence and hence more readily
identified by comparison to databases, since coding sequences of genes are more
highly conserved in evolution than non-coding sequences. The cDNA library to be
used for the experiment was therefore first examined for the relative abundance of
partial cDNA clones. In addition to moderately-abundant seed-specific clones, some
seed-specific clones (viz. seed storage proteins) were expected to be highly abundant.
These clones were eliminated from the experiment.

The third consideration was whether the oleate-12-hydroxylase DNA sequence
would be recognisable as such. In chapter 4 it was hypothesized that this gene is
homologous to those of the fatty acid desaturases, particularly fad2, the microsomal
A12 desaturase gene. However, attempts to isolate the oleate-l2-hydroxylase gene by
probing with the fad3 gene were unsuccessful, and others have found that fad2 could

also not be isolated by this approach. Furthermore, it was also not possible to isolate

112

the oleate-l2-hydroxylase gene using an apparently conserved desaturase motif in a
PCR approach. It is still hypothesized (for the reasons described in chapter 4) that
the oleate-12-hydroxylase gene is homologous to those of the fatty acid desaturases,
however some of the desaturases, including fad3 and fad 7 which were available for
these experiments, are too divergent from the oleate-l2-hydroxylase to be used as
probes for direct gene isolation or gene isolation by conserved-sequence PCR.
Nevertheless, sequence similarity should be obvious if the sequences were to be
compared.

Another aspect of this large-scale sequencing experiment was the likelihood of
isolating other genes of interest to our laboratory or other laboratories, regardless of
the outcome with respect to the oleate-12-hydroxylase gene. For example, the
microsomal acyltransferases are active in this lipid-synthesizing tissue, including the
seed-specific diacylglycerol acyltransferase, and it was of interest to isolate any of the

genes encoding these (and many other) enzymes.

MATERIALS AND METHODS

Screening of JtZAPST Library

Duplicate filters (Hybond N" , Amersham) were lifted from plaques of the

AZAPST library. DNA was fixed to the filters by placing them on filter paper moist

with denaturing solution (0.5 M NaOH, 1.5 M NaCl; 5 min), neutralising solution

113
(0.5 M Tris-Cl pH 7.4, 1.5 M NaCl; 5 min), and 2 X SSC (0.3 M NaCl, 0.03 M

Na-Citrate, pH 7.0). The filters were then air-dried, with no further fixation of the
DNA. A 32P—labelled probe was prepared from the insert of the stearoyl-ACP-
desaturase clone pRch6 by random priming7 and purified of unincorporated
nucleotides by ethanol precipitation in the presence of ammonium acetate.7 The filters
were prehybridised at 65°C for ~1 h in the hybridization solution (4 X SET (0.6 M
NaCl, 0.12 M Tris-HCl pH 7.4, 8 mM EDTA), 0.1% sodium pyrophosphate, 0.2%
SDS, 5 % dextran sulphate, and 0.1% heparin) before addition of the probe and
hybridization overnight at 65°C. The filters were washed three times in 2 X SSC,

0.1% SDS at room temperature, then exposed to X-ray film.

Differential Screening

Phage in nine separate 96-well plates were replicated onto a single bacterial
lawn the size of one 96-well plate using a 96-prong device which could be lowered
onto the lawn through a 3 x 3 array of guides. The blunt ~1 mm diameter prongs
carried sufficient phage to give plaques of consistent size, without significant
encroachment between neighbouring plaques. Triplicate filters were lifted from these
plaques and screened as described above, except that for plates 1-9, polyadenylic acid
(1 pg ml") was added to the hybridization solution, and results for plates 1-9 were
obtained from a phosphor-imager (Molecular Dynamics) rather than from

autoradiographs. Exposure times were: plates 1-9, 21 h (note that phosphor imaging

114
is several-fold more sensitive than autoradiography); plates 28—36 leaf probe 3 days,

seed probe 24 h, redundant-clone probe 1.5 h.

Probes for screening plates 1-9 were prepared as follows.
Poly(A)+ RNA (1 pg) from seed or leaf (chapter 3) in a volume of 17 pl was heated
to 70°C for 5 min, then chilled in ice-water, and added to the reaction tube, to a final
volume of 50 pl:
RNasin (Promega), 1 U pl"
reverse transcriptase buffer (Boehringer-Mannheim), 1 x
oligo(dT),2,,3, 20 ng pl"
dGTP, dATP, dTI‘P, 1 mM each
dCTP (unlabelled), 4.8 pM
a-32P dCTP (3000 Ci mmol"), 100 pCi
Avian Myeloma Virus Reverse Transcriptase, 40 U.
The reaction was incubated at 42 °C for 60 min. The reaction was stopped and RNA
removed by addition of EDTA (to 16 mM), SDS (to 0.4%), NaOH (to 0.4 M) and
incubation at 65°C for 30 min. The probe was neutralised with 6 pl 2 M HCl and 20
pl 1 M Tris-Cl, pH7.4, then precipitated with 375 p1 EtOH in the presence of 0.7 M
ammonium acetate and 10 pg denatured carrier (salmon sperm) DNA. After
incubation at -20°C for ~ 3 h, DNA (~ 60% of total radioactivity) was pelleted by
centrifugation for 15 min, and resuspended in 200 pl water and added to the filters.

For plates 28-36, first-strand cDNA was made using the same RNA (0.5 pg

seed, 1.2 pg leaf) in a reverse transcription reaction similar to that described above,

115

but using unlabelled nucleotides and all other components from a reverse transcription
kit (Promega). The RNA was hydrolysed and the cDNA was neutralised as described
above, and then purified by batch chromatography on glass (GeneClean, BiolOl).
The cDNA was then used as a template for random priming,7 using 100 pCi cit-”P
dCTP. The probes were precipitated as described above, and heated to 100°C (5
min) before addition to the filters. Incorporation of radioactivity was ~ 60% (leaf
probe) or ~30% (swd probe).

A probe was made from redundant clones as follows. Clones (Table 7) were
digested with BamHI and KpnI and the inserts purified from agarose gels. DNA of
these inserts was pooled and ~ 600 ng labelled with 100 pCi a-32P dCTP (~ 80%

incorporation) by random priming as described above.

DNA Sequencing

Plasmid DNA was prepared from Escherichia coli cultures (5 ml, LB medium
containing 100 mg l" ampicillin) using "Magic Minipreps" (Promega) according to the
instructions of the manufacturer, and submitted for automated cycle sequencing on

Applied Biosystems 373A instruments using the fluorescent primer T3 (occasionally

r7).

l 16
RESULTS AND DISCUSSION

Characterization of AZAPST cDN A Library

Bacterial colony hybridizations frequently give high background signal due to
the bacterial debris adhering to the filter. Since a differential screen was planned, it
was imperative to minimise background signal, and so a lambda phage library was
chosen. The library used in this experiment (AZAPST) was constructed in AZAP II
(Stratagene) by Dr. S. Turner, from developing endosperm and embryo of castor, as
described above for the pYES2.0 library (chapter 3). Construction of the KZAPST
library included directional cloning, so that 5’ ends of the inserts should be found at
the T3 side of the polylinker. This library had not yet been characterised, so an
investigation was made similar to that described in chapter 3. The castor stearoyl-
ACP-desaturase° probe was used to isolate 16 primary duplicated positive clones from
a single filter bearing ~20 000 plaques. All 16 of these clones re-screened,
indicating an abundance of about 1 per 1250, which is the same order of magnitude as
for the pYES2.0 library (1 per 3000). DNA of 14 of the clones was analysed by
restriction digestion and agarose gel electrophoresis. Full-length clones should
contain an insert of ~ 1.6 kb giving a BamHI fragment of ~0.7 kb. Three of the 14
clones analysed appeared to be close to full-length. Inserts of the remaining clones

varied between 0.5 kb and 1.0 kb.

117

Table 7. Redundant sequences obtained in batch 1 sequences, used to make a probe
for screening of batch 2. Frequency of these sequences in batch 2 (after the screen)
is compared with batch 1 (before the screen) to indicate effectiveness of the screen.

 

 

 

 

Redundant Sequence Frequency in Batch Frequency in Batch
Ribosomal proteins1 12 (6.6%) 23 (8.0%)
128 seed storage protein2 10 (5.5%) 8 (2.8%)
28 seed storage protein3 6 (3.3%) 5 (1.7%)
Heat shock proteins4 4 (2.2%) 5 (1.7%)
Enolase’ 4 (2.2 %) 0

 

‘ Clones pCRS262, pCRS312, pCRS356, pCRS358, pCRS377,
pCRS396, pCRS407, pCRS409, pCRS416, pCRS426, pCRS432,
pCRS442, pCRS446.

2 Clones pCRS267, pCR8269, pCRS298, pCRS404, pCRS405,
pCRS408, pCRS434, pCRS443, pCRS453, pCRS454.

3 Clones pCRS28l, pCRS328, pCRS337, pCRS362, pCRS375,
pCRS43l.

“ Clones pCRS264, pCRS348, pCRS397.

5 Clones pCRS330, pCRS380, pCRS415, pCRS439.

 

1 18
Differential Screening with Seed and Leaf First-Strand cDNA Probes

Plaques of the XZAPST library were picked from fresh, low-density plates (to
avoid cross-contamination) into 96-well plates. They were then plated onto bacterial
lawns with the addition of several control clones. These included non-recombinant
phage (which gave blue plaques on plates containing IPTG and XGal indicating that
the lacZ gene was uninterrupted) and clones of the stearoyl-ACP-desaturase, acyl
carrier protein, 28 seed storage protein, and 12S seed storage protein, the latter three
clones being provided by Dr. S. Turner. By using an offsetting device, plaques from
nine 96-well plates could be represented on a filter the size of one plate. Multiple
filters, each representing 864 identifiable clones, were lifted from the resulting
plaques and screened with 32P-labelled first-strand cDN A probes reverse transcribed
from leaf or developing endosperm/ embryo poly(A)+ RNA. Only those clones were
selected which gave no detectable signal with the leaf probe, and did not give a very
strong signal with the md probe. These results were obtained by on-screen analysis
of images from a phosphor-imaging system (Molecular Dynamics). Plates 1-9 were
processed in this manner, from which the first batch of sequences were obtained.

Of 864 possible plaques from plates 1—9, 10 did not appear and 15 were
observed to be occluded by bubbles separating the plaque and filter, leaving 839
clones with DNA on the filter. Of these, 162 (19.3%) were scored as having a strong
seed signal, while 280 (33.4%) gave no detectable signal with the leaf probe. Of

these 280, 222 were not among the previous category and were selected for

119
sequencing. These results therefore indicated that 222 of 839, or 26.5% of clones,

were in the category ”seed-specific and not highly abundant". Of the 162 clones
having a strong seed signal, only 58 appeared to be seed specific.

Some changes were made when screening plates 10-54 for the second
sequencing batch. The seed and leaf probes were made by random priming using
first-strand cDNA as a template, in an attempt to gain maximum incorporation of
radioactivity into less-abundant sequences. In addition, a probe was made from the
pooled insert DNA of clones frequently sequenced in the first batch (Table 7) so that
fewer redundant sequences would be obtained. Screening results were obtained
directly from autoradiograms. The most clearly-interpretable data were for plates 28-
36, from which clones for the second sequencing batch were isolated. These
autoradiograms are presented in Figures 12-14.

For plates 28-36, 851 of a possible 864 plaques were represented on the filter,
and of these 851, 370 (43.5%) gave a strong seed signal, 512 (60.2%) gave no
detectable leaf signal, and 141 (16.6%) gave a signal with the probe made from
redundant sequences (the effectiveness of screening with this particular probe is
discussed below). This resulted in the selection of 348 (40.9% of 851) clones to be
sequenced.

Considerable differences are evident among the results of the first and second
screening batches. More (43.5% compared with 19.3%) clones were classed as
having a strong seed signal in the second batch than in the first. This may be

attributable to the subjective nature of determining when the signal was strong as

120

 

de-hUIONQCD

 

 

 

 

 

ABCDEFGH 363534
333231
302928

 

 

 

 

Figure 12. Autoradiogram of plates 28-36 probed with 32P-labelled cDNA from
developing seeds. The positions of the wells in the original 96-well plates is indicated
(borders), and the position of clones from each 96-well plate relative to other 96-well
plates is indicated in the box.

121

 

d-l—l
—I~

HNw-FUIQNQOO

 

 

 

 

 

36 35 34
33 32 31
30 29 28

 

 

 

Figure 13. Autoradiogram of plates 28-36 probed with 32P-labelled cDNA from
developing leaves. The positions of the wells in the original 96-well plates is
indicated (borders), and the position of clones from each 96-well plate relative to
other 96-well plates is indicated in the box.

 

 

122

 

 

 

 

ABCDEFGH 363534
333231
302928

 

 

 

 

#

Figure 14. Autoradiogram of plates 28-36 probed with 32P-labelled DNA from
redundant clones sequenced in batch 1. The positions of the wells in the original 96-
well plates is indicated (borders), and the position of clones from each 96-well plate
relative to other 96-well plates is indicated in the box.

123

opposed to merely detectable, and the inevitable differences in background obtained
with different probes and filters in separate hybridisations. In particular, probes for
the second batch were made by random priming off first strand cDNA on the premise
that DNA polymerase I (Klenow) is less sensitive to low concentrations of the labelled
nucleotide than is reverse transcriptase, and would therefore give greater
incorporation of label. This may have been a successful strategy for the seed probe,
but apparently not for the leaf probe, with which only 39.8 % of clones gave a
detectable signal in the second batch, compared with 66.6% in the first batch.

Despite the differences between the two batches, these results are in general
agreement with those of RNA excess/single-copy DNA hybridization experiments in
tobacco."’9 These experiments show that the total complexity of mRNA or nuclear
RNA in different organs of the tobacco plant is very similar. The organs analysed in
these experiments were leaf, root, stem, ovary and anther, but the similarity of the
results for these organs suggest that it is possible to extrapolate to other organs also,
such as the seed. Approximately 25000 mRNAs were expressed in each organ. Of
these, approximately 25-30% were ubiquitous, that is, expressed in all organs, while
an additional 30-40 % were expressed in another organ (but not all other organs). The
remaining 30-45 % of messages expressed in a particular organ were unique to that

organ. “

124

Sequencing and Sequence Analysis

The differential screens described above gave a total of 570 lambda phage
clones selected for sequencing. These were individually excised in viva to yield the
corresponding Bluescript plasmid. DNA was prepared from each plasmid, analysed
spectrophotometrically for DNA concentration, and submitted to the PRL Plant
Biochemistry Facility for automated sequencing. The T3 primer was used because in
the directional library, this should give sequence from the 5’ end of the clones, and
since most of them are not full-length, this should frequently yield protein—coding
sequence which is most readily identified by database searches. In some sequences
(e.g. CRS262, CR8263, CR8269), a poly(A) tail was obvious, indicating that the
clone was small enough to sequence through in one run. In fewer cases (e. g CRS270,
CRS294, CRS315), a poly(T) tract was observed at the 5’ end of the insert DNA,
indicating that in some cases directional ligation fails during library construction. For
this reason, sequences which apparently gave database matches in a negative reading
frame were not ignored. Sequence data was edited to remove vector/ linker
sequences, and truncated at the point where sequence quality declined substantially.
These edited sequences (typically 400-500 nucleotides) were submitted electronically
to the National Center for Biotechnology Information, Bethesda, MD. DNA
sequences were compared in all reading frames to the non-redundant translated-
nucleotide and protein sequence databases (Swiss-Prot 24.0 or 25 .0 plus weekly

updates; PIR 35.0, 36.0, or 37.0; GenBank Release 75.0, 76.0, or 77.0, plus daily

125
updates; and EMBL Release 34.0 or 35 .0, plus daily updates) by the program

blastx,lo in the months March-July, 1993. In general, a blastx score of 80 is used as
the cutoff for determining whether a sequence has a match in the databases. An
exception is made for repetitive sequences, which may give scores higher than 80
without meaningful identification.

In addition to the 570 clones selected for sequencing on the basis of the
differential screen, five clones were sequenced which gave a strong signal with the
seed probe and would therefore normally have been excluded. Four encoded 2S seed
storage protein, and the fifth encoded 12S seed storage protein. This confirmed that
the clones excluded on the basis of strong seed signal were not of interest in
identifying novel genes.

Data in Table 7 indicate that screening the second batch to remove redundant
clones obtained in the first batch was partially successful. The frequency of md
storage protein clones sequenced in the second batch is lower, but not zero, probably
due in part to the weak signal given by clones with very short inserts. No enolase
clones were sequenced in the second batch. There may have been a small reduction
in the frequency of heat shock protein clones, but there was no effect on the
frequency of clones for ribosomal proteins. This is not surprising since these were
generally not truly redundant clones, most clones corresponding to different heat
shock proteins or ribosomal proteins.

Figure 15 is a flow diagram depicting the various fates of the total of 570

clones selected for sequencing. Insufficient DNA of 44 clones was obtained (in a

126

Figure 15. Fates of clones selected for sequencing by differential screening.

127

 

570
Selected for Sequencing

 

 

/\

 

 

526

_ ‘44-,

Sequenced Nat Sequenced "

 

 

 

Failed excision,
insufficient. DNA ;

 

/\U

 

 

468

 

Sequence: Blastx run Sequence: Blastx not run

 

 

 

 

 

 

 

/ e A

 

 

 

 

 

 

 

 

 

 

 

Entirely I Poly(A) ‘ Sequence
231 i VeCtOr only _‘ too poor
Score < 80 Score > 80
19 218 18 213
  Poor i . Good . Score dueto "Hit":
sequence sequenCe, repetitive sequence
maybe Cause nomatch in: sequence ' identified
~ of low score database . *

 

 

 

 

 

 

 

 

 

 

 

 

 

128

single attempt) for sequencing. Of the remaining 526 clones sequenced, 58 gave
sequence data which was not considered informative, containing, for example,
extraneous DNA. Sequence from the 468 informative clones was analysed by the
blastx program, leading to the putative identification of 213 (46%) of them. This is a
high level of identification, compared to other similar sequencing projects. Of 2375
human brain cDNAs, 17% were identifiable by comparison to the databases.11 In an
earlier report of human brain cDNA sequences where less stringent criteria were used
for the assignment of identity, 52% of 475 cDNAs were identified.‘2 Of 130 random
maize leaf cDNAs sequenced, 20% were identified.” The identification of only 8%
of 830 cDNAs from rice suspension culture cells was reported,“ but a much higher
alignment score (160, compared with 80 used in this and other projects) was used as
the criterion for identification. Differential screening approaches have also been used
previously in this type of sequencing project. In one study,” mouse testis cDNAs
were screened with probes made from testis and from a pool of other tissues. The
majority (90%) of 51 abundant clones which hybridised with both probes were
identified, but only 25% of 120 rare clones, which hybridised with neither probe,
i were identified.

i Sequences for which an alignment with database sequences suggests an identity
(blastx score 2 80) are presented in four tables. In Table 8 are listed those clones
for which the most similar match in the databases was a prokaryotic gene. Similarly,
Table 9 lists clones for which the most similar match was from a eukaryote other than

a higher plant. Clones for ribosomal protein genes have been excluded from this

129
table and are listed in Table 10. In Tables 8 and 9, the possible function of each

clone is discussed, in an attempt to reveal new opportunities arising from the
identifieation of genes apparently not previously sequenced from higher plants.

The fourth table, Table 11, lists most of the clones in the remaining category,
by the other higher plant gene to which they were found to be most similar. Clones
for ribosomal proteins are listed separately, in Table 10. Clones in Table 11 are
grouped according to the general area of metabolism or cell function with which they
are most likely to be involved, and these categories are discussed below, rationalising
the sequences obtained in the light of the major metabolic or cellular processes known
to be active in the developing seed tissue from which the cDNA library was made,

and for which the differential screen employed might be expected to enrich clones.

130

Table 8. Clones for which the most similar match in the databases was from a
prokaryote. The clones are listed by the suffix of their clone number (pCRS265 etc.).
Information for each clone is structured as follows: blastx score, (probability of
random alignment), database (G: GenBank; S: SwissProt; P: PIR), accession number,
protein description, organism. This is followed by a discussion of the possible
function of each clone. Information for these discussions is drawn from annotations
to the database entries, as well as standard textbooks, and original references where
cited.

 

265 Blast 82 (0.00019) S P04990: threonine synthase, Bacillus subtilis.
18/ 39 (46%) residues identical between CRS265 and B. subtilis protein, 28/39
(71%) when allowing for conservative substitutions. Residues conserved
between the E. coli and B. subtilis enzymes“ are generally also conserved in
the castor sequence, suggesting that the marginal score reflects a true
identification. Enzymes of amino acid biosynthesis must be active in the
castor endosperm for storage protein synthesis.

296 Blast 418 (1.3e-55) S P17242: asparaginyl-tRNA synthetase, E. coli.
Also a weaker match to the yeast homologue (43/78 = 55% identical
compared with 81/125 = 64% identical in E. colt). Tyr-426 of the E. coli
protein is involved in ATP binding17 and is conserved in the castor sequence.
This is in a region which is strongly similar between the two proteins, and is
also conserved between other aminoacyl-tRNA synthetases for amino acids
with an XAX codon (includes Tyr, stop, His, Gln, Asn, Lys, Asp, and Glu)."’
Castor 28 seed storage protein is notably rich in glutamine (44 residues of
258), raising the possibility that this clone is in fact a glutaminyl-tRNA
synthetase.

332 Blast 119 (1.4e-09) S P26242: hypothetical 67.9 kDa protein in pufx
region (photosynthetic gene cluster), Rhadabacter capsulatus.
No other matches. Rhadabacter sequence has similarity with enzymes
requiring thiamine pyrophosphate, such as acetolactate synthase, pyruvate
decarboxylase and indolepyruvate decarboxylase, and transketolases. the
predicted Rhadabacter protein has a slight mean hydrophilicity, suggesting that
it is not a membrane protein. Only mutations affecting the differentiation of
the photosynthetic aparatus are known to map to this gene cluster in
Rhadabacter, suggesting that the predicted protein has some role in
bioenergetics. ‘9

336 Blast 153 (9.0e-14) S P22256: 4-aminobutyrate aminotransferase
(GABA transaminase), E. coli.
Other matches are also bacterial amino acid metabolism enzymes (2,2-

Table 8 (cont’d). 131

435

463

519

dialkylglycine decarboxylase, 3-isopropylmalate dehydrogenase, acetylornithine
aminotransferase). In each case, sequence similarity is found in the same 2 or
3 small clusters, presumably of functional significance. The same applies also
for two weaker matches in the same reading frame to the yeast enzymes
omithine aminotransferase and 4-aminobutyrate aminotransferase. Comparison
of the GABA transaminase with several other aminotransferases also revealed
these conserved domains, but their functional significance has not been
elucidated.20 Note: CRS 841 has strong homology to alanine aminotransferase
of millet, human and rat, as well as some similarity to aspartate
aminotransferase of Rhizobium. CR8856 has strong homology to aspartate
aminotransferase from higher plants and other organisms.

Blast 106 (1.4e-07) G D10483: ORF, E. coli.

Also weaker matches in the same reading frame to rat and human short-chain
specific acyl-CoA dehydrogenases, and the residues which match between
these and castor are also those most strongly conserved amongst various rat
and human acyl-CoA dehydrogenase isozymes,21 and are generally the same as
residues matching with the E. coli sequence. Hence probably a true match for
an acyl-CoA dehydrogenase. The rat and human proteins are mitochondrial
flavoprotein enzymes for fatty acid B-oxidation (butanoyl-CoA » 2-butenoyl-
CoA). The developing castor seed may be ”preparing” for germination by
making glyoxisomes (cf. CRS266, Table 9) and the enzymes of B-oxidation
they require.

Blast 117 (4.3e-09) 8 P08534: NIFM protein, Klebsiella pneumoniae.

Also matches to NIFM of Azatabacter spp. , and to protein export protein
PRSA of Bacillus subtilis, in the same reading frame and at the same clusters
of conserved residues, suggesting that NIFM and PRSA and CRS463 all have
a related function. NIFM is probably involved in a processing step necessary
to generate an active nitrogenase iron protein. PRSA is involved in a late
stage of protein export and is probably located on the outer membrane. An
hypothetical role of CRS463 might involve storage protein import into the
vacuole.

Blast 118 (4.4e-21) S P00968: carbamoyl phosphate synthase large

(ammonia) chain, E. coli. ’

Also matches to other bacterial, mammalian and yeast carbamoyl-phosphate
synthases, some the ammonia-dependent enzyme involved in arginine
biosynthesis, some the glutamine-dependent enzyme involved in pyrimidine
biosynthesis. In liver mitochondria, carbamoyl phosphate synthase I (ammonia
dependent) is tightly associated with ornithine carbamoyl transferase
(CRS772), forming a multifunctional protein which can most efficiently use the
labile carbamoyl phosphate for synthesis of citrulline (precursor to arginine).
Presumably the arginine biosynthetic pathway is active in the developing castor

Table 8 (cont’d). 132

559

631

640

673

seed for storage protein synthesis. Alternatively, arginine synthesis may be
active in the developing castor md to support synthesis of polyamines
(discussed in the text).

Blast 134 (5 .0e-12) 8 P23523: hypothetical 31.0kDa protein in mpB 3’

region (ORF2), E. coli.

No other matches. The function of this E. coli ORF is unknown, but it shows
week (approx 24 %) amino acid sequence similarity with gntZ of Bacillus
subtilis, which affects the utilization of gluconate, and gnd of E. coli, which
encodes 6-phosphogluconate dehydrogenase.22 No plant gene for this enzyme
was found by searching the databases. 6-Phosphogluconate dehydrogenase is
an enzyme of the pentose phosphate pathway, which may"3 or may not be“
important in providing NADPH for fatty acid synthesis.

Blast 162 (4.1e-17) G X51510: riboflavin synthase, B. subtilis.

Matches also to other bacterial riboflavin synthase B subunits. Riboflavin
synthase is itself a flavoprotein, involved in the last steps of riboflavin
synthesis. (Mammals cannot make riboflavin).

Blast 100 (1.6e-06) P 818956: fix23-4 protein, Rhizabium melilati.

Fix23 is a multi-ORF locus bearing homology to fatty acid synthases and
polyketide synthases, involved in synthesis of a secreted lipopolysaccharide.”
Other matches are to mycocerosic acid synthase, a multifunctional protein of
Mycabacterium tuberculosis which catalyses fatty acid elongation,2° the
polyketide synthase multifunctional protein of Saccharapalyspora erythraea
involved in erythromycin biosynthesis,27 and the type-I fatty acid syunthases of
chicken and rat. Fix23-4 has homology to enoyl reductase, and CRS640
matches the conserved residues in the NAD(P)H binding site. There is also
homology to alcohol dehydrogenases and zeta-crystallins (quinone
oxidoreductase homologues) in the same region. CRS640 appears to align to
the enoyl-reductase domain of the other matches also, but has no homology to
the enoyl-ACP reductase cloned from rapeseed that is a component of the plant
plastid type-II fatty acid synthase. No genes for plant type-I fatty acid
synthases are yet cloned. Potentially CRS640 is from a cytoplasmic fatty acid
elongase or polyketide synthase. Further sequence has been obtained: from
the 3’ end, and by sequencing further from the 5’ end, thus covering the entire
insert. The assembled sequence now has a blast score for fix23-4 of 113
(8.0e-08). The lack of sequence matching a different domain of a FAS
suggests that this cDNA is from a gene encoding only one domain/protein of a
putative FAS (though other genes may be found in the same region of the
chromosome).

Blast 226 (1.4e-27) 8 P21773: hypothetical 17.4kDa protein in firA
(ssc)-lpr intergenic region, Salmonella typhimurium.

Table 8 (cont’d). 133

Match also to the homologous ORF in E. coli, and a shorter alignment (hence
much weaker score: 61 , 13/ 29 identical residues) in the same reading frame to
3-hydroxydecanoyl—ACP-dehydratase also of E. coli. However, the residues
identical between CRS673 and the 3-hydroxydecanoyl-ACP-dehydratase are
notably the same as those conserved between CRS673 and the Salmonella and
E. coli ORFs in the same region, suggesting that these ORFs and CRS673
have something functional in common with the FAS enzyme.

The E. coli 17.4 kDa ORF is found in an operon which includes other
ORFs and the two known genes of lipid A synthesis, lpr and lpr.2”° The
17.4kDa ORF is immediately upstream of lpr, with overlapping start and
stop codons, suggesting cotranscription and translational coupling.” Lipid A
is an essential component of the outer leaflet of the outer membrane of most
gram-negative bacteria.31 It has a glucosamine backbone rather than the
glycerol backbone of typical membrane lipids, and acyl chains are attached by
ester and amide linkages.”32 The primary acyl chains are 3-hydroxy-myristic
acid, to the hydroxyl groups of which are esterified the secondary acyl chains,
lauric or myristic acids. In the firA (omsA) mutant of E. coli, the secondary
acyl groups include palmitic acid, but the function of the firA gene product is
unknown.33 The first enzyme in lipid A synthesis, UDP-N-Acetylglucosamine
3-O-Acyltransferase, is encoded by lpr.”34 The gene lpr encodes lipid A
disaccharide synthase?” The genes encoding other enzymes of lipid A
synthesis have yet to be identified?"33 though the ORFs associated with Ipr
and lpr are obvious candidates. Other enzymes required for lipid A
synthesis31 include a deacetylase (hydrolysing acetate from the glucosamine
group); a 3-hydroxymyristoyl-ACP N-acyltransferase; and the ”late”
acyltransferases, specific for lauroyl- and myristoyl-ACPs, which esterify these
acyl groups to the hydroxyl groups of the previously esterified 3-hydroxy-
myristate moieties. These "late” acyltransferases have been assayed in crude
E. coli fractions, but have not been purified or well characterized.“'37

The FAS enzyme, 3-hydroxyacyl-ACP-dehydrase, is involved (possibly
together with an isomerase) in the introduction of a double bond in the
growing acyl chain, enabling the production of unsaturated fatty acids under
anaerobic conditions. A single histidine (His7°) is the only active site residue
that is directly involved in the reaction of this enzyme,38 and this residue is
also found in CRS673 and the two E. coli and Salmonella l7 .4 kDa
hypothetical proteins. 3-hydroxyacyl-ACP-dehydrase has been purified from
spinach and has two histidine residues per molecule”, as does the E. coli
enzyme, but has not been cloned from higher plants.

The possible homology between the 3-hydroxyacyl-ACP-dehydrase and
the 17.4 kDa ORFs suggests a possible function for the ORFs. The "late”
acyltransferases produce estolide-linkages, where one acyl group is esterified
to the hydroxyl-group of a substituted acyl chain. In Claviceps, such estolide
linkages are formed by addition of the secondary acyl group to a double bond
of the primary acyl group (see chapter 2). It is possible that a component of

Table 8 (cont’d). 134

706

756

772

798

the “late“ acyltransferases is a 3-hydroxyacyl-dehydrase, giving the 2,3-
unsaturated primary acyl group, to which the secondary acyl group is then
added.

A report”, of which I was only able to obtain the abstract at time of
writing, suggests that the 17 kDa ORF described here was purified as a
component of an "actomyosin' complex of E. coli.

Blast 266 (1.7e-32) G Ll4862: biotin carboxylase, Anabaena sp. (PCC
7120).

Biotin carboxylase is one of the two subunits of the Anabaena acetyl-CoA-
carboylase (ACCase), carboxylating the carrier protein, before the carboxyl is
transferred to acetyl-CoA. There are also matches to the biotin carboxylase
subunits of other bacterial ACCases, and weak matches to the biotin
carboxylase domain of yeast pyruvate carboxylase and various mammalian
carboxylases. The Anabaena match (50/73 = 68% identical) is considerably
stronger than the next strongest, E. cali'(34/72 = 47% identical). This makes
it tempting to suggest that this is a plastid enzyme. Dr. Ohlrogge compared
CRS706 to their plant ACCase sequences, and found that this clone is
significantly different, and so would be some other biotin carboxylase.
However, it is possible that the previously cloned plant eukaryotic-type
ACCase genes are for a cytoplasmic isozyme, since no transit peptide was
identified. Such a cytosolic enzyme. may produce malonyl-CoA for the
synthesis of flavonoids, phytoalexins, ethylene, or fatty acid elongation (see
CRS640).23 Dr Ohlrogge is interested in the possibility that pCRS706 encodes
a subunit of the plastid form of ACCase, responsible for provision of malonyl
CoA for fatty acid synthesis, and will characterise this clone in more detail.
There is evidence for a pokaryotic-type ACCase in plastids, and one subunit
(of the carboxyltransferase) has been identified in pea chloroplasts.“

Blast 144 (7.0e-l3) 8 P05054: ribokinase, E. coli.

Weaker matches to tagatose-6-phosphate kinase of Staphylococcus aureus, 1-
phosphofructokinase of Rhadabacter capsulatus and E. coli and Xanthamanas
campestris, a probable ribokinase of yeast, and ketohexokinase of rat. Clearly
CR8756 encodes some carbohydrate kinase. Note also CR8517: very strong
homology to 6-phosphofructokinase (B-subunit) from potato; CRS792:
fructokinase, potato; CRS304: pyruvate kinase.

Blast 189 (1.4e-2l) S Q02095: omithine carbamoyltransferase,
anabolic, Mycabacterium bavis.
Similar matches to a number of other bacterial ornithine transferases, and

weaker matches to the enzyme from various mammals. Arginine biosynthesis.
See CR8519.

Blast 178 (2.7e-l7) P A42653: pyruvate dehydrogenase El-a subunit,

Table 8 (cont’d). 135

828

852

Achaleplasma laidlawii.

Matches also to homologues from bacteria and mammals (Achaleplasma is a
mycoplasma). Part of PDH complex. There are E1 or and B subunits, these
are distinguished from the dihydrolipoamide acetyltransferase and
dihydrolipoamide dehydrogenase components of the complex. CR8504 and
CR8844 are homologous to dihydrolipoamide dehydrogenase (see text). In the
developing castor seed, the source of carbon for fatty acid biosynthesis has
been proposed to be from imported sucrose via glycolysis to malate in the
cytoplasm, followed by the action of malic enzyme and pyruvate
dehydrogenase in the plastid, yielding acetyl-CoA and NAD(P)H.23
Components of pyruvate dehydrogenase are therefore logically abundant
sequences expressed in the developing castor seed.

Blast 189 (5.0e-20) 8 P31038: succinate dehydrogenase flavoprotein

subunit, Rickettsia prowazekii.

Also matches to the same protein from yeast, other bacteria, and mammals, as
well as weaker similarity to bacterial fumarate reductase flavoprotein. In
bacteria, succinate dehydrogenase is a complex containing the flavoprotein, an
iron-sulphur protein, and cytochrome bm. In mammals, the enzyme has two
subunits, the larger being the flavoprotein. Succinate dehydrogenase oxidises
succinate to fumarate in the tricarboxylic acid cycle.

Blast 145 (6.4e-l4) 8 P80016: nitrogen-regulatory protein P-II,
Synechacaccus sp. (PCC6301).

Also matches to the homologous protein from a range of bacteria. Involved in
deadenylation (activation) of glutamine synthase under nitrogen-limiting
conditions (when ratio of glutamine to 2-ketoglutarate decreases). P-II is also
involved in the regulation of the transcription of the glutamine synthase gene.
P-II is controlled by photosystem 11.“2 Regulation of glutamine synthase in
plants by the same mechanisms as in these bacteria is hitherto unknown.‘3
CR8852 may be involved in regulation of amino acid/nitrogen metabolism in
the castor seed, where amino acid and protein synthesis has to be coordinated
with nitrogen import. P-II is a small (112 amino acid) protein, and the
alignment with CR8852 extends from the N—terminus, suggesting that
pCR8852 may be a full length clone. A collaboration has been initiated with
Dr. Gloria Coruzzi, New York University.

 

136

Table 9. Clones for which the most similar match in the databases was from a
eukaryote other than a higher plant. Ribosomal proteins are not included. The clones
are listed by the suffix of their clone number (pCR8266 etc.). Information for each
clone is structured as follows: blastx score, (probability of random alignment),
database (G: GenBank; 8: SwissProt; P: PIR), accession number, protein description,
organism. This is followed by a discussion of the possible function of each clone.
Information for these discussions is drawn from annotations to the database entries, as
well as standard textbooks, and original references where cited.

 

266

282

Blast 151 (5 .6e-15) S P14292: 20kDa peroxisomal membrane protein, Candida
baidinii.

Membrane-associated protein of methanol-induced peroxisomes. ”Function
very likely to be related to the metabolism of methanol "“ presumably would
not explain its occurrence in castor seed. Targetting of proteins to
peroxisomes appears to be broadly conserved throughout the eukaryotes, and
the C-terminal tripeptide (S/A/C)-(K/R/H)-L may be sufficient to direct
peroxisomal protein import in plants, yeast, insects and mammals.“ This
consensus sequence is common among peroxisomal proteins, including plant
peroxisomal soluble proteins,“5 but is not the only type of peroxisomal
targetting determinant.“7 The Candida protein has the C-terrninal tripeptide A-
K-L. The sequence CR8266 aligns to the central portion of the Candida
sequence; a second sequencing run, from the 3’ end of the clone, revealed that
the open reading frame terminates with the variant tripeptide K-A-L. Since
this is the first plant putatively peroxisomal membrane protein sequenced, it is
not clear if this variant tripeptide could function in targetting to the plant
peroxisome. A comparison of C-terminal sequences of a number of yeast
peroxisomal proteins revealed that they are rich in hydroxy amino acids at the
-8 and -9 positions;‘“ serine occupies both these positions in the castor
sequence. The developing castor seed may make glyoxisomes in ”preparation”
for germination when glyoxisomal function dominates metabolism (see also
CR8435). A collaboration has been initiated with Dr. Laura Olsen (University
of Michigan) who is interested in using the clone as a possible marker for
plant peroxisome biogenesis.

Blast 297 (8.2e-37) S P20659: trithorax, Drosaphila melanagaster.

A 404 kDa nuclear zinc-finger protein involved in segment determination by
interacting with the genes of bitharax and antennapedia.48 Homology is to the
carboxy terminus, not surprising for such a large protein/ message. Trithorax
mutants have homoeotic segmentation. Also a match for the human trithorax
homologue, the function of which is not known, but is implieated in
translocations resulting in leukemias," i.e. seems to have a DNA-binding role.
Possibility of controlling embryo-development in castor? CR8282 would be

Table 9 (cont’d). 137

304

314

347

349

interesting to map in Arabidapsis where several interesting mutants affecting
embryo development are mapped, and to try to correlate with a function by
antisense-inhibition in Arabidapsis upon isolating the Arabidapsis homologue.

Blast 165 (6.3e-16) S P00548: pyruvate kinase, chicken muscle.

Final step in glycolysis (phospho-enol-pyruvate - enol-pyruvate
espontaneously pyruvate), hence high level of expression in castor seed would
not be surprising, to maintain the acetate pool used for fatty acid synthesis.
However, it has been suggested that the major source of pyruvate in the
plastids of developing castor endosperm is not made via pyruvate kinase, but
rather from oxaloacetate which is converted to malic acid by malate
dehydrogenase (CR8503) and to pyruvate by malic enzyme.23 Nevertheless,
when measured in vitro, the leukoplasts of the developing castor endosperm
have high activities of pyruvate kinase (330 nmol/min/ mg protein) and some
phosphoglycerate kinase (9 nmol/ min/ mg protein). These enzymes are
proposed to participate in a triose-phosphate shuttle (also requiring malate
dehydrogenase, CR8503, and g1yceraldehyde-3-phosphate dehydrogenase,
CR8294) which contributes to the provision of ATP for fatty acid biosynthesis
in non-photosynthetic tissues such as the castor seed.24 This shuttle would be
particularly important if ATP is not derived from glycolysis in the plastid, as
in the case of the malic enzyme proposal, or imported from the cytoplasm (see
CRS692). A cytosolic pyruvate kinase has been cloned from potato, but this
was the weakest match against CRS304 compared to all the other non-plant
pyruvate kinases, so CRS304 is quite divergent from the potato clone. Dr.
David Dennis, Queens University (personal communication 4/8/93) has
isolated at least four pyruvate kinase genes from plants (not available in the
database), and believes there may be more, quite divergent ones. This clone
might be an example of such a gene.

Blast 84 (7.3e-05) P 821342: ring-infected erythrocyte surface antigen,
Plasmadium falciparum.
Alignment: 18/43 (41%) identities, 28/43 (65%) positives.

Blast 189 (4.9e-22) G 851858: M025 gene, putative calcium binding protein,
mouse embryo.

M025 was isolated from the 8-cell stage of mouse embryos by differential
screening, its level being increased relative to unfertilised eggs.’0 Function of
the gene product is unknown and it is dissimilar to other known proteins,
except the C-terminus of the predicted protein has similarity to the calcium
binding site of EF-hand proteins, such as a human plasma membrane calcium
pump (ATPase).

Blast 152 (3.4e-15) S P15170: GSTl-HS GTP-binding protein, human.
Also homologous to sup2 of yeasts. These are elongation factor 10:

Table 9 (cont’d). 138

401

438

455

495

homologues, involved in protein synthesis (EF 1 mediates interaction between
the ribosome and aminoacyl-tRNA) and in the Gl-to-S phase transition in
yeast‘1 and regulation of mammalian cell growth.’2 8up2 is essential for
viability. Match is in the C-terminus which is well conserved among the
mammalian and yeast proteins. Involvement in protein synthesis and cell
growth logical for the castor endosperm or embryo used to make the library,
which are actively growing and synthesizing storage proteins. Note also:
CRS680 is elongation factor 2.

Blast 89 (2.7e-14) same as for pCRS349, above.

Blast 110 (5.0e-20) G X14977: aldehyde dehydrogenase, rat.

Also match to the human homologue, which are both mitochondrial, but
matches to cytosolic forms are also observed. [e. g. acetaldehyde -l- NAD+ +
H20 —. acetic acid + NADH]

Blast 141 (3.7e-12) P 822439: NAM8 protein, yeast.

Encodes a protein with putative RNA binding motifs and acts as a supressor of
mitochondrial splicing deficiencies when overexpressed. Also a match for
NGRl , negative growth regulatory protein, of yeast, which encodes a putative
glucose repressible protein containing two RNA recognition motifs. There is
weaker similarity (in the same reading frame) to other RNA binding proteins.

Blast 154 (3.8e-14) S P23913: lamin B receptor, chicken.

Also 145 (7 .2e-13) G 849653: sterol C-14 reductase, Saccharomyces
cerevisiae; 90 (6.5e-05) S P25340: hypothetical transport protein YGIJO22,
Saccharomyces cerevisiae; and 84 (0.00047) P A43765: stsl + protein,
Schizasaccharamyces pambe. The match of the castor sequence to these four
apparently disparate sequences is at the same conserved residues, suggesting
that they all have something in common. The lamin receptor is on the inner
membrane of the nuclear envelope. Alignment of castor sequence and lamin
receptor includes parts of two potential transmembrane domains. The sterol
reductase is quite likely a membrane protein since its substrate is membrane-
soluble, but the cellular location of the enzyme has not been described.”
Alignment with the sterol reductase involves matches between similar residues
as for the lamin receptor. Alignment of castor sequence and YGL022 (which
is a different protein to sterol reductase) also includes part of one and the
whole of another potential transmembrane domain, but sequence similarity is
sparse (25/95 identities, 42/95 positives). Stsl”, a gene of unknown function
and pleiotrOpic effects when mutated has been suggested to be an integral
membrane protein on the basis of sequence similarity between stsl”, the lamin
B receptor and YGLOZZ,‘4 all three proteins having a similar hydrophobicity
pattern consisting of eight or nine putative transmembrane domains. It seems
that these four characterised genes and CRS495 may consitute a class of

Table 9 (cont’d). 139

545

548

592

633

636

similar membrane proteins with differing functions.

Blast 122 (1.5e-10) P 825013: vacuolar ATPase 14 kDa chain, Manduca sexta
(tobacco hornworm).

This is the only match found. The Manduca ATPase is unpublished. See also
CRS785, which is a clone from the same gene, and CR8512, which is a match
for the vacuolar H+ ATPase catalytic subunit of cotton, with a weak alignment
also to chain A of the Manduca ATPase. Rapid enlargement of the protein-
storing vacuoles in the developing castor seed would require the concomitant
synthesis of vacuole-associated proteins, such as the ATPase (which
furthermore may be involved in protein import), and tonoplast intrinsic protein
(see CRS279, Table 11).

Blast 413 (2.4e-56) G X56932: 23 kDa highly basic protein, human.
Also 411 (4.7e-56) S P19253: transplantation antigen P198, mouse. P198 is
also highly basic. The function of P198 is not known.”

Blast 100 (7.1e-12) P 827783: hypothetical protein 1 (adenylate cyclase
homologue), C. elegans.

Blast 186 (4.7e—20) P 827951: renal cortical Na/Pi cotransporter, rabbit.

A collaboration has been initiated with Dr. Yves Poirier, who has isolated and
sequenced two different cDNAs of Arabidapsis homologues, using pCRS633
as a probe, which are found to encode proteins having the characteristic
transmembrane domains of transporters.56 Preliminary data suggests that
neither of these cDNAs maps to the chromosomal location of the phoI
mutation involved in phosphate translocation to the shoot.‘7 Few genes
directly involved in nutrient uptake in plants have been cloned, and none
involved in phosphate nutrition. This clone may be a useful beginning.

Blast 132 (7.8e-12) S P13641: autoantigen small nuclear ribonuleoprotein 8M-
D, human.

Also a very similar match for the mouse homologue. Nucleic acid binding
protein involved in mRNA splicing. SM-D is one of the core proteins of the
small nuclear RNP complexes, suggesting that the castor clone is probably for
a nuclear-localised ribonucleoprotein. Both human and mouse proteins have a
lysine rich domain and Gly-Arg repeats at carboxy terminus, but their
functional significance is unclear.58 These are also found in the castor
sequence, and do not directly resemble any known plant or animal nuclear
localisation sequences.”

Blast 442 (6.6e-61) G 219599: epsilon isoform of 14-3-3 protein, mouse.
Similar matches to homologues from many other organisms including plants,
but the castor sequence is more similar to the mouse, sheep and yeast proteins

Table 9 (cont’d). 140

669

680

692

than the homologue from barley, suggesting that this isoform is not previously
sequenced from higher plants. The mammalian proteins are protein kinase II
and protein kinase C inhibitors. The barley homologue is induced upon
attempted penetration by powdery mildew fungi.

Blast 171 (2.8e-17) P A37863: ferrochelatase (haem synthetase), mouse.

Also similar matches to other mammals, bacteria, and yeast, but this is
presumably the first sequence from plants. This mitochondrial enzyme
catalyses the last step in haem biosynthesis, in which protoporphyrin IX binds
Fe“ . A collaboration has been initiated with Dr. Mary Lou Guerinot,
Dartmouth College.

Blast 502 (4.8e-65) 8 P28996: elongation factor 2, Chlorella kessleri.

As well as this lower-plant match, there, are matches to elongation factor 2 of
other eukaryotes, and elongation factor G of prokaryotes, but this appears to
be the first higher plant sequence. EF-2/EF-G is required for the GTP-
hydrolysing translocation of the ribosome from one codon to the next, during
which the nascent peptide chain is switched from the A to the P site and the
uncharged tRNA is released. This, along with the ribosomal proteins, is an
unsurprising sequence to find expressed in the developing castor seed where
seed storage protein synthesis is active. EF-2 (but not EF-G) is the target for
inactivation by diptheria toxin (by ADP-ribosylation of the novel amino acid
dipthamide) and ricin.60 Note also that CRS349, CRS401 are elongation factor
1a homologues.

Blast 89 (6.5e-15) S Q01888: mitochondrial solute carrier protein homologue,
bovine.

Similar score mitochondrial ATP/ ADP carrier protein, human, and many
homologues from other organisms. The human ATP/ADP carrier protein is in
mitochondrial inner membrane. Maize brittle-1 appears to be the highest-
scoring plant match, and may be an amyloplast homologue of the
mitochondrial carrier proteins, found in endosperm of developing kernels.
Brittle-1 mutants have severely reduced amounts of starch deposition in the
endosperm. Sequence analysis of the Brittle-1 gene revealed a plastid transit
peptide, and it was suggested that the Brittle-1 protein may be an amyloplast
envelope translocator, important for ATP import for starch synthesis."l Could
this be a leukoplast version in castor? Uptake of ATP by the plastids of
developing castor endosperm may be necessary for fatty acid synthesis (see
CRS304), as it is necessary for starch synthesis in maize. CR8523, CR8581,
and CR8630 appear to be quite different ADP/ATP carrier proteins, very
strongly homologous to other such proteins from higher plants (CR8523 giving
a score of 584 with a potato protein, and only 82 with brittle-1, CR8581 a
score of 348 with an Arabidapsis protein and 70 with brittle-1, CR8630: 474
for rice, 70 for brittle-l).

Table 9 (cont’d). 141

729

741

753

763

785

Blast 162 (1.9e-l8) G X15141: histone H4, Physarum palycephalurn (slime
mould).

Many very similar matches for other histone H4 genes, including wheat. Very
high conservation that is characteristic of histones: 32/ 38 identical residues
with the Physarum protein. Note: CR8808 is an histone H2A homologue.

Blast 224 (3.6e-30) S P13641: autoantigen small nuclear ribonucleoprotein
SM-D, human.

Also a very similar match for the mouse homologue. These are the same
proteins which match CRS636, but this sequence is substantially different to
CR8636 (approx 40% identity at the nucleotide level, and approx 25 % at the
amino acid level), suggesting that this is a second, quite different homologue
of SM-D. CR8636 has 27/ 86 (31%) identical amino acids to human SM-D,
while CR8741 has 45/68 (66%) identical to SM-D. CR8741 also has the
conserved lysines and Gly-Arg repeats, and also shows matches to the Gly-Arg
repeats of another nuclear protein, of human herpesvirus.

Blast 93 (1.3e-11) S P21187: poly(A) binding protein, Drosaphila
melanagaster.

Also matches yeast poly(A) binding protein, human initiation factor eIF-4B, as
well as weaker matches to other higher plant RNA-binding proteins, e. g.
Nicotiana, Arabidapsis. The castor sequence is most similar to the Drosaphila
protein in a short domain identified to be involved in RNA binding (5 identical
residues of 8 (63%) compared to 19/66 (28%) overall).

Blast 94 (7.4e-14) S P19623: spermidine synthase (putrescine
aminopropyltransferase), human.

Last step in the biosynthesis of spermidine from arginine and methionine:
SAMamine + putrescine » methylthioadenosine + spermidine, i.e. transfers
amino propyl group (HzN-[CHZ]3). Also a weaker match to the same enzyme
from E. coli, and also a much weaker match in the same reading frame to
aspartate aminotransferase of Sulfitlabus salfataricus. Other enzymes involved
in polyamine biosynthesis were also identified. These are enzymes of arginine
biosynthesis (CR8519 and CR8772, see Table 8), S—adenosylmethionine
synthetase (CR8554), and S-adenosylmethionine decarboxylase (pCR8555).
Polyamine metabolism is discussed further in the text.

Blast 275 (3.3e-34) P 825013: vacuolar ATPase 14 kDa chain, Manduca sexta
(tobacco hornworrn).

Match to the same protein as CR8545, except CRS785 is from a much longer
clone. There are few nucleotide sequence differences between CR8545 and
CR8? 85 in the region of overlap, probably ascribable to sequencing errors,
thus CR8545 and CRS785 probably represent clones originating from the same
gene. Again, this was the only match in the databases.

Table 9 (cont’d). 142

865 Blast 144 (5.4e-13) S P12613: T-complex protein 1 homologue, Drosaphila
melanagaster.
Cytosolic molecular chaperone, involved in folding of actin and tubulin. An
Arabidapsis homologue has been cloned, but is a slightly weaker match to the
castor sequence than are the mammalian and yeast proteins, suggesting that
this may be a new isoform for plants.

 

143

Table 10. Clones homologous to ribosomal proteins. Those having homology to
ribosomal proteins for which higher plant sequences are already available, and those
having homology to ribosomal proteins apparently previously unsequenced from
higher plants (see text), are listed separately. The clones are listed by the suffix of
their clone number (pCR84l6 etc.). Information for each clone is structured as
follows: blastx score, (probability of random alignment), database (G: GenBank; 8:
SwissProt; P: PIR) , accession number, protein description, organism.

 

From Higher Plants

416 Blast 242 (6.1e-30) G 217784: 408 ribosomal protein 813, Arabidapsis
thaliana. ‘

432 Blast 373 (4.1e-51) 8 Pl9950: 408 ribosomal protein 814, Zea mays.

497 Blast 275 (7.2e—36) P 825550: ribosomal protein 86, Nicatiana tabacurn.

446 Blast 412 (4.5e-53) P 821519: acidic ribosomal protein P0 homologue,
Chenopodium rubrum.

546 Blast 466 (3.9e-62) P 819893: cytosolic ribosomal protein L8, Lycapersican
esculentum.

571 Blast 86 (6.7e-l4) P 825983: mitochondrial ribosomal protein 811, Marchantia
palymarpha.

595 Blast 299 (6.9e-40) G 217767: 608 ribosomal protein L27A, Arabidapsis
thaliana.

684 Blast 348 (2.4e-44) S P17094: 608 ribosomal protein L3, Arabidapsis
thaliana.

707 Blast 263 (4.0e-34) 8 Q00332: ribosomal protein 815A, Brassica napus.

710 Blast 227 (8.2e-28 ) 8 P29766: 608 ribosomal protein L8, Lycapersican
esculentum.

766 Blast 508 (7.3e-68) 8 P17094: 608 ribosomal protein L3, Arabidapsis
thaliana.

837 Blast 452 (5.7e-62) 8 Q00332: ribosomal protein 815A, Brassica napus.

From Other Eukaryotee

262 Blast 230 (5 .0e-28) 8 P08636: 408 ribosomal protein 817, chicken.

312 Blast 476 (4.8e-65) G M96570: ribosomal protein 828, Saccharomyces
cerevisiae.

346 Blast 303 (1.0e-37) 8 P19889: acidic ribosomal protein PO/DNAse,
Drosaphila melanagaster.

356 Blast 279 (3.4e-34) 8 P02362: 408 ribosomal protein 88, Xenapus laevis.

358 Blast 424 (6.2e-56) P 825633: ribosomal YLlO protein homologue,
Chironamus tentans.

377 Blast 266 (3.4e-33) G X68202: ribosomal protein L27, Chlamydabotrys

Table 10 (cont’d). 144

407
409
426
442
522
643
655
683
690
701
708
709
744
8 14

825
860

stellata.

Blast 102 (4.1e-09) S P29314: 408 ribosomal protein 89, rat.

Blast 139 (3.7e-12) G M77233: ribosomal protein, human.

Blast 342 (7.8e-46) G X68202: ribosomal protein L27, Chlamydabotrys
stellata.

Blast 276 (8.4e-36) G X68202: ribosomal protein L27 , Chlamydabatrys
stellata.

Blast 122 (2.0e-22) 8 P05387: 608 acidic ribosomal protein P2, human.
Blast 178 (1.9e-19) 8 P18124: 608 ribosomal protein L7, human.

Blast 248 (8.7e-29) 8 P26321: 608 ribosomal protein 1.5, Saccharomyces
cerevisiae.

Blast 81 (0.00029) G M76718: 608 ribosomal protein rp129, Tetrahymena
thermaphila.

Blast 295 (4.5e-39) P 824989: ribosomal protein L31, Chlamydomonas
reinhardtii.

Blast 170 (2.5e-20) G M96570: ribosomal protein 828, Saccharomyces
cerevisiae .

Blast 267 (3.4e-34) S P05744: 608 ribosomal protein YL37, Saccharomyces
cerevisiae.

Blast 221 (4.2e-26) 8 P05387: 608 acidic ribosomal protein P2, human.
Blast 367 (4.9e-49) 8 P25121: 608 ribosomal protein L11, rat.

Blast 80 (0.00030) G 221487: ribosomal protein L21, Saccharomyces
cerevisiae.

Blast 258 (8.1e-33) 8 P25111: 408 ribosomal protein 825, human.

Blast 107 (5.8e-10) 8 P24050: 408 ribosomal protein 85, rat.

 

145

Table 11. Clones for which the most similar match in the databases was from a
higher plant, other than ribosomal proteins. Clones are grouped according to the
general area of metabolism or cell function with which they may be involved (see text
for discussion). The clones are listed by the suffix of their clone number (pCRS792
etc.). Information for each clone is structured as follows: blastx score, (probability of
random alignment), database (G: GenBank; 8: SwissProt; P: PIR), accession number,
protein description, organism.

 

Lipid Synthesis

792 Blast 105 (5 .1e-08) G 212823: fructokinase, Salanum tuberosum.

517 Blast 501 (7 .0e-66) S P21343: 6-phoshofructokinase, Solanum tuberasurn.

427 Blast 531 (3.7e-68) 8 P08440: fructose-bisphosphate aldolase, Zea mays.

823 Blast 572 (3.8e-76) G D13512: cyt0plasmic aldolase, Oryza sativa.

294 Blast 208 (1.3e-24) G X59517: glycolytic glyceraldehyde-3-phosphate
dehydrogenase, Antirrhinum majus.

330 Blast 420 (5. 8e-53) P J Q1 186: enolase, Lycapersican esculentum.

380 Blast 573 (2.4e-76) P J Q1 186: enolase, Lycapersican esculentum.

415 Blast 346 (3.2e-44) P JQ1186: enolase, Lycapersican esculentum.

439 Blast 371 (2.5e-46) P JQ1186: enolase, Lycapersican esculentum.

509 Blast 561 (2.2e-73) P J Q1186: enolase, Lycapersican esculentum

503 Blast 354 (7 .8e-47) G X17362: mitochondrial malate dehydrogenase, C‘itrullus
vulgaris. ~

504 Blast 127 (8.4e-11) P 818152: dihydrolipoamide dehydrogenase precursor,
Pisum sativum.

844 Blast 153 (2.6e-34) P 822384: dihydrolipoamide dehydrogenase, Pisum
sativum.

544 Blast 317 (6.1e-42) S P08817: acyl carrier protein 11 precursor, Hardeum
vulgare.

601 Blast 176 (2.3e-22) S P07854: acyl carrier protein 1, Spinacia aleracea.

507 Blast 554 (7.1e-75) G L13242: beta-ketoacyl-ACP synthase, Ricinus
communis.

794 Blast 155 (1.4e-16) S P28643: 3-oxoacyl-ACP reductase precursor, Cuphea
lanceolata.

291 Blast 292 (2.7e-38) 8 P22243: stearoyl-ACP desaturase, Carthamus tinctarius.

667 Blast 255 (2.8e-3l) G M91238: stearoyl ACP desaturase, Salanum tuberasum.

858 Blast 163 (1.9e-17) 8 P22243: stearoyl-ACP desaturase, Carthamus tinctarius.

677 BLAST 110 (2.7E-16) G D14410: ORF, VIGNA RADIATA; BLAST 112
(3.2E-l3) G 101418: LINOLEIC ACID DESATURASE, BRASSICA NAPUS.

834 Blast 126 (6.8e-21) P A44227: OMEGA-3 FATTY ACID DESATURASE,
BRASSICA NAPUS; 131 (2.7e-20) G D14410: ORF, VIGNA RADIATA.

698

Blast 134 (1.3e-12) G M87514: cytochrome b5, Brassica aleracea.

 

Table

384
191

293

494

Table 11 (cont’d). 146

384 Blast 120 (1.2e-10) 8 P29111: major oleosin nap-II, Brassica napus.

398 Blast 256 (2.4e-29) P J Q0986: lipid body-associated membrane protein,
Daucus carota.

293 Blast 482 (2.0e-66) G 870711: protein disulfide isomerase homologue,
Medicaga sativa.

494 Blast 517 (4.9e-67) P 822479: protein disulphide isomerase precursor,
Medicaga sativa.

728 Blast 384 (9.2e-51) 8 P29828: protein disulphide isomerase precursor,
Medicaga sativa.

Protein Synthesis

299 Blast 222 (3.5e-27) G M92353: anthranilate synthase alpha subunit,
Arabidapsis thaliana.

841 Blast 267 (1.4e-33) P 828429: alanine transaminase, Panicurn miliaceum.

856 Blast 406 (9.4e-55) G L09702: aspartate aminotransferase, Glycine max.

267 Blast 199 (1.2e-22) G X59802: cruciferin, Raphanus sativus. Shouldn’t have
been sequenced - seed positive.

269 Blast 130 (4.8e-12) 8 P09800: legumin precursor, Gassypium hirsutum.

298 Blast 326 (9.5e-42) S P09800: legumin precursor, Gassypium hirsutum.

404 Blast 259 (2.7e-31) S P09800: legumin precursor, Gassypium hirsutum.

405 Blast 105 (3.5e-O8) 8 P09800: legumin precursor, Gassypium hirsutum.

408 Blast 169 (6.0e-19) 8 P09800: legumin precursor, Gassypium hirsutum.

434 Blast 134 (1.4e-13) G X59802: cruciferin, Raphanus sativus.

443 Blast 99 (1.1e-06) 8 P09800: legumin precursor, Gassypium hirsutum.

453 Blast 276 (1.1e-32) S P09800: legumin precursor, Gassypium hirsutum.

454 Blast 182 (1.8e-19) P A35540: cruciferin precursor, Brassica napus.

520 Blast 154 (1.2e-15) G X14393: preproglutelin, Oryza sativa.

540 Blast 163 (6.2e-20) G X59804: cruciferin, Raphanus sativus.

584 Blast 400 (3.5e-52) 8 P13744: 118 globulin beta subunit precursor, Cucurbita
pepo. -

641 Blast 215 (4.4e-28) G X57849: cruciferin cru2/3 subunit, Brassica napus.

686 Blast 343 (1.4e-44) G X59807: cruciferin, Raphanus sativus.

691 Blast 253 (3.3e-32) G X14312: CRAl 128 md storage protein, Arabidapsis
thaliana.

739 Blast 252 (1.2e-32) G X14312: CRAl 128 md storage protein, Arabidapsis
thaliana.

811 Blast 117 (5.9e-lO) S P09800: legumin precursor, Gassypium hirsutum.

813 Blast 372 (2.1e-46) S P09800: legumin precursor, Gassypium hirsutum.

281 Blast 183 (1.4e-20) G X54158: 28 albumin, Ricinus communis.

328 Blast 183 (2.2e-20) G X54158: 28 albumin, Ricinus communis.

337 Blast 113 (2.9e-18) G X54158: 28 albumin, Ricinus communis.

362 Blast 245 (3.1e-32) G X54158: 28 albumin, Ricinus communis.

375 Blast 521 (2.7e-70) S P01089: 2S albumin precursor, Ricinus communis.

Table 11 (cont’d). 147

431
549
735

780
807
816
424
467
826
385
359
264

348
382

397
598
657
759
778
820
833
275
279
360
663
749
765
83 l

512

Blast 319 (1.4e-39) G X54158: 28 albumin, Ricinus communis.

Blast 82 (3.7e-05) G X54158: 28 albumin, Ricinus communis.

Blast 96 (2.8e-08) P B25 802: 28 seed storage protein large chain, Bertholletia
excelsia.

Blast 78 (1.4e-05) G X54158: 28 albumin, Ricinus communis.

Blast 339 (5.8e-50) G X54158: 28 albumin, Ricinus communis.

Blast 243 (7.0e-31) G X54158: 28 albumin, Ricinus communis.

Blast 440 (1.9e-60) P A24010: ricin D chain B, Ricinus communis.

Blast 200 (9.4e-22) P A24210: agglutinin B chain, Ricinus communis.

Blast 570 (1.4e-75) P A24210: agglutinin chain B, Ricinus communis.

Blast 188 (4.3e-21) S P24922: initiation factor 5A, Nicatiana plurnbagintfalia.
Blast 93 (3.2e-05) G L00623: opaque2 heterodimerizing protein, Zea mays.
Blast 324 (1.6e-42) S P29357: chloroplast membrane 70kDa heat shock
protein, Spinacia aleracea.

Blast 245 (7.2e-29) P 825005: heat shock protein 70kDa, Phasealus vulgaris.
Blast 544 (2.1e-72) P JQ1360: luminal binding protein BLP-4 precursor,
Nicatiana tabacum.

Blast 363 (8.4e-48) P J 80710: low molecular weight heat shock protein, Oryza
sativa. .

Blast 407 (7 .4e-54) G M87646: chaperonin 10, Spinacia aleracea.

Blast 206 (2.1e-23) P 825005 : heat shock protein 70K, Phasealus vulgaris.
Blast 382 (6.5e-50) P J80710: low molecular weight heatr shock protein,
Oryza sativa.

Blast 324 (5 .7e-42) G X67695: DNA] -1 (mitochondrial heat shock protein),
Cucumis sativus.

Blast 295 (2.1e-37) 8 P29357: chloroplast envelope membrane 70 KD heat
shock-related protein, Spinacia aleracea.

Blast 211 (1.3e-26) 8 P27879: 18.1 KD class 1 heat shock protein, Medicaga
sativa.

Blast 255 (1.4e-30) G M95796: 8t12p protein, Arabidapsis thaliana.

Blast 86 (1.6e-05) G 852690: nodulin-26 homologue, Pisum sativum.

Blast 81 (9.2e-05) G 852690: nodulin-26 homologue, Pisum sativum.

Blast 517 (3.7e-72) 8 P265 87: tonoplast intrinsic protein alpha, Arabidapsis
thaliana.

Blast 230 (1 .3e-28) 8 P2395 8: tonoplast intrinsic protein alpha, Phasealus
vulgaris.

Blast 438 (2.2e-59) P J G1 106: tonoplast intrinsic protein alpha, Phasealus
vulgaris. -

Blast 396 (3.5e-53) P J Q1 106: tonoplast intrinsic protein alpha, Phasealus
vulgaris.

Blast 279 (6.0e-35) G L03186: vacuolar H+-ATPase catalytic subunit,
Gassypium hirsutum.

Table 11 (cont’d). 148

Ricinine Biosynthesis

696 Blast 214 (3.7e-25) S Q00763: bispecific caffeic acid/5-hydroxyferulic acid 0-
methyltransferase, Papulus tremulaides.

738 Blast 140 (1. 6e-12) G X72593. naringenin, 2-oxoglutarate 3-dioxygenase,
Callistephus chinensis.

, Polyamine Biosynthesis

554 Blast 458 (2.3e—62) G M62758: S-adenosylmethionine synthetase, Petraselinum
crispum.
555 Blast 136 (2.2e-12) P 828047: TUB13 protein, Salanum tuberasum.

Organ-Specific Expression

371 Blast 80 (0.001) G 218891: BP8 gene product, Betula pendula.

565 Blast 120 (3.3e-09) 8 P21746: embryonic abundant protein precursor, Vicia
faba.

589 Blast 301 (5 .2e-39) P J Q1 107: 18.3k protein precursor, pollen, Zea mays.

406 Blast 172 (8.5e-18) 8 P16148: pplz02 protein, Lupinus palyphyllus.

430 Blast 131 (1.8e-11) 8 Pl6146: pplz02 protein, Lupinus palyphyllus.

Defense-Related Proteins

413 Blast 180 (8.1e-23) G X69139: protease inhibitor 11, Arabidapsis thaliana.

648 Blast 200 (4.6e-26) G X69139: protease inhibitor 11, Arabidapsis thaliana.

653 Blast 198 (1.5e-26) G X69139: protease inhibitor II, Arabidapsis thaliana.

754 Blast 215 (2.7e-29) P 824965: p322 protease inhibitor, Glycine max.

580 Blast 137 (2.4e-13) P 628027: protein P10, Nicatiana sp.

711 Blast 109 (3.0e-08) 8 P28493: thaumatin-like protein precursor, Arabidapsis
thaliana.

Dessication-Related Proteins

521 Blast 137 (1.8e-13) P 825121: dehydrin-cognate, Pisum sativum.
557 Blast 245 (8.2e-28) 8 P22242: dessication-related protein clone PCCl3-62
precursor, Craterastigma plantagineum.

Other

Protein Kinases

280 Blast 483 (6. 2e-67) G D10152: protein tyrosine-serine-threonine kinase,
Arabidapsis thaliana.

Table 11 (cont’d). 149

436 Blast 231 (1.4e-26) G 1108789: protein kinase, Arabidapsis thaliana.

731 Blast 428 (9.6e-57) G L05562: protein kinase, Arabidapsis thaliana.

733 Blast 550 (2.2e-73) P 827760: protein kinase 6, Glycine max.

817 Blast 308 (1.1e-38) 8 P28583: calcium-dependent protein kinase, Glycine max.
836 Blast 509 (2.6e-67) P 826627: serine-protein kinase, Arabidapsis thaliana.

ADP/ATP Carrier Proteins

581 Blast 348 (3.4e-47) G X68592: adenosine nucleotide translocator, Arabidapsis
thaliana.

523 Blast 584 (2.6e-79) P 817917: ADP, ATP carrier protein precursor, Salanurn
tuberasum.

630 Blast 474 (2.6e-63) P 180711: ATP/ADP translocator protein, Oryza sativa.

Ubiquitin-Related Proteins

342 Blast 382 (3.9e-53) G L06967: ubiquitin carrier protein, Medicaga sativa.
505 Blast 564 (3.5e-76) G M74100: ubiquitin fusion protein, Nicatiana sylvestris.
387 Blast 343 (4.1e-47) G M74100: ubiquitin fusion protein, Nicatiana sylvestris.
672 Blast 236 (9.5e-33) G L06967: ubiquitin carrier protein, Medicaga sativa.

Actin Depalymerizing Factor
313 Blast 158 (4.8e-16) P 825061: actin depolymerizing factor, Lilium langiflarum
363 Blast 212 (7 .3e-25) P 825059: actin depolymerizing factor, Brassica napus.
524 Blast 320 (4.9e-44) P 825061: actin depolymerizing factor, Lilium langiflarum.

Tltbulin
558 Blast 578 (6.8e-78) P 823221: tubulin alpha chain, Prunus amygdalus.
781 Blast 626 (5 .7e—85) P 823221: tubulin alpha chain, Prunus arnygdalus.

Ras/G-Prateins

322 Blast 502 (9.3e-69) G 876123: ras-related ypt family cDNA yptm2, Zea mays.

343 Blast 344 (3.7e-46) P 825543: RAS-related GTP-binding protein, Pisum
sativum.

Metallathianein
444 Blast 423 (6.3e-60) G 102306: RCMTl metallothionein, Ricinus communis.
746 Blast 99 (1.3e-07) 8 P30564: metallothionein I homologue, Ricinus communis.

Miscellaneous

305 Blast 249 (3.9e-51) P 822499: hypothetical protein, Bramus secalinas.

743 Blast 171 (7.0e-19) P 822499: hypothetical protein, Bramus secalinas.

447 Blast 150 (7 . 1e-20) G X60391: proline-rich protein, Phasealus vulgaris.

499 Blast 242 (1.7e-32) 8 P10973: non-specific lipid transfer protein A, Ricinus
communis.

511 Blast 466 (1.9e-62) 8 P12412: legumain endopeptidase, Vigna munga.

Table 11 (cont’d). 150

596
700

720

779
805
808
818
842

Blast 89 (8.0e-06) P 819253: gene le25 protein, Lycapersican esculentum.
Blast 522 (4.0e-72) P J Q0939: cyc07 protein, S-phase specific, Catharanthus
roseus.

Blast 179 (1.5e-l9) S P19595: UTP-glucose-l-phosphate uridylyltransferase,
Salanum tuberasurn.

Blast 256 (3.2e-34) 8 P28756: superoxide dismutase (Cu-2n), Oryza sativa.
Blast 557 (2.7e-77) G M90504: RNA polymerase H, Glycine max.

Blast 312 (2.2e-41) G X67819: histone H2A, Picea abies.

Blast 134 (4.0e-12) G L20864: ascorbate peroxidase, Spinacia aleracea.
Blast 148 (1.1e-17) S P13089: aux28 protein, Glycine max.

 

151
Lipid Synthesis

Lipid synthesis is a dominant area of metabolism in the developing castor
seed, such that at maturity, oil constitutes 50% of the seed dry weight. Not
surprisingly therefore, many genes with a likely or certain role in some aspect of lipid
synthesis were identified.

We begin by considering the source of carbon for fatty acid and lipid
synthesis. Sucrose imported by the seed is inverted to glucose and fructose, which
enter the glycolysis pathway. Both are converted to their 6-phosphates; pCRS792
encodes fructokinase. Glucose-6-phosphate is isomerised to fructose-6—phosphate, and
fructose-6-phosphate from both sources is phosphorylated to fructose-1,6-diphosphate
by 6-phosphofructokinase (pCR8517). Fructose-1,6-diphosphate is then cleaved to
dihydroxyacetone phosphate and 3-phosphoglyceraldehyde by aldolase (pCRS427,
pCR8823).

At this point we are interested in a branch of the pathway, in which
dihydroxyacetone phosphate (directly from the aldolase reaction or isomerized from 3-
phosphoglyceraldehyde) is converted to glycerol-3-phosphate by glycerol-3-phosphate
dehydrogenase (pCRS294). Glycerol-3-phosphate is the carbon backbone for the
synthesis of glycerolipids, including the major storage lipid, triacylglycerol.

Several steps further in the glycolysis pathway, 2-phosphoglyceric acid is
dehydrated to phosphoenol pyruvate by enolase (pCRS330, pCRS380, pCRS415,
pCRS439 and pCR8509; see also Table 7). Phosphoenol pyruvate is converted to

enol pyruvate (and thus spontaneously to pyruvate) by pyruvate kinase (pCRS304,

152
Table 9), completing glycolysis.

An alternative source of pyruvate has been proposed to be active in the
developing castor seed.62 Rather than being converted to pyruvate by pyruvate
kinase, phosphoenol pyruvate would be carboxylated to oxaloacetate in the cytoplasm,
and then converted to malate by malate dehydrogenase (pCR8503). Malate is then
imported by the plastid and converted to pyruvate by malic enzyme.

Pyruvate, derived from either source, is converted to acetyl-CoA by the
pyruvate dehydrogenase complex (PDC). The PDC has three subunits: E1
(pCRS798, Table 8), dihydrolipoamide acetyltransferase, and dihydrolipoamide
dehydrogenase (pCR8504, pCRS844; these two clones share almost no similarity at
the nucleotide level, and probably represent the plastid and mitochondrial forms,
respectively, since pCR8844 has much higher amino acid identity (86%) with the
previously-sequenced pea mitochondrial protein than does pCR8504 (45%)). A
collaboration has been initiated Dr Christoph Benning, IGF Berlin, who will use the
three PDC clones to characterise the genes of plastid PDC, and their possible
manipulation for increasing seed oil production;

Acetyl-CoA is carboxylated to malonyl-CoA, which is used for fatty acid
synthesis. This is accomplished by acetyl-CoA carboxylase, of which pCRS706 (see
Table 8) may encode a subunit. Fatty acid synthesis occurs in the plastid with the
growing acyl chain esterified to acyl carrier protein (ACP). Two ACP clones were
identified, pCRSS44 and pCR8601 , which appear to be independent clones derived

from the same gene, since they show 100% nucleotide identity in the region of

153 '

overlap. The fatty acid synthase cycle involves condensation of the acyl-ACP with
malonyl-ACP, catalysed by 3—ketoacyl-ACP synthase (pCR8507, which shows 100%
identity to a gene previously sequenced from castor, but which has not been assigned
to one of the three possible isozymes), reduction to 3-hydroxyacyl-ACP by 3-
ketoacyl-ACP reductase (pCRS794), and dehydration to enoyl-ACP by 3-
hydroxyacyl-ACP dehydratase (the putative identity of pCRS673; see Table 8). The
cycle is completed by enoyl-ACP reductase.

In the castor seed, approximately 90% of the fatty acid synthesized is modified
by the action of two enzymes, producing ricinoleic acid. The first is the soluble
stearoyl-ACP desaturase (pCR8291, pCRS667, and pCRS858) which introduces a
double bond between carbons 9 and 10 (counting from the carboxyl end). The second
enzyme, oleate-12-hydroxylase, is the principle enzyme for which a clone was sought
in this project. As described above, it was hypothesized that this gene is homologous
to those of the microsomal fatty acid desaturases, and it was hoped that the oleate-12-
hydroxylase gene could be identified on this basis. Two clones, pCRS677 and
pCRS834, gave such a match to the fad3 gene, encoding the microsomal (0'3
desaturase, the only known higher plant membrane-bound desaturase which was in the
databases. These two clones were therefore considered putative oleate-12-hydroxylase
clones and their analysis is the subject of chapter 6.

In addition, electron transfer to the microsomal desaturases and oleate-12-
hydroxylase requires cytochrome b, (chapter 2), and a cytochrome b, clone

(pCRS698) was also identified.

154
Fatty acids are sequentially esterified to glycerol-3-phosphate to form the

storage lipid, triacylglycerol. Triacylglycerol accumulates in the cell in spherosomes
(oil bodies), which are bounded by a phospholipid monolayer studded with
characteristic proteins, oleosins (pCR8384, pCRS398; these are independent clones,
but further sequence data is required to determine whether they represent the same or
different genes). The mechanism of transport of triacylglycerol from the site of
synthesis in the endoplasmic reticulum to the spherosome is unknown. In this light, it
is interesting to note that protein disulphide isomerase (PDI; pCR8293, pCRS494,
and pCRS728) is an essential subunit of the mammalian liver triacylglycerol-specific
microsomal lipid transfer protein (MTP).‘53 In liver, MTP may mediate the transport
of triacylglycerol to nascent very low density lipoprotein particles in the lumen of the
endoplasmic reticulum and golgi aparatus. The other subunit of MTP has no activity
when separated from PDI, and has no homology to other known proteins.“ A role
for PDI in triacylglycerol transfer in the developing castor seed is purely speculative,
and the abundance of PDI clones can also be rationalised by its known role in protein

folding (below).

Protein Synthesis

In addition to storing lipid, the castor seed accumulates storage proteins, and
these are synthesized at similar stages of seed development to those selected (on the
basis of active lipid synthesis) for construction of the cDNA library. All aspects of

storage protein synthesis are considered here, from nitrogen import into the

155

developing seed, to storage of the synthesized protein in the vacuole.

Import of nitrogen by the developing seed must be co-ordinated with its
assimilation into amino acids, and this may suggest a role for a nitrogen regulatory
protein PII homologue (pCR8852, see Table 8).

Various clones identified may encode enzymes of amino acid biosynthesis.
These are threonine synthase (pCR8265, see Table 8); anthranilate synthase, involved
in tryptophan biosynthesis (pCR8299); two enzymes of arginine biosynthesis:
carbamoyl-phosphate synthase and omithine carbamoyltransferase (pCR8519 and
pCRS772 respectively, see Table 8); and three aminotransferases: alanine
transaminase (pCRS841), aspartate aminotransferase (pCR8856), and 4-
aminobutyrate aminotransferase (pCRS336, see Table 8).

The most abundant transcripts in the developing seed are probably those
encoding the seed storage proteins themselves. Attempts were made, by differential
screening, to eliminate these from the pool of clones to be sequenced. This was
largely successful, but some such clones were sequenced nevertheless, and are listed
in Table 11. Short clones gave a weaker hybridisation signal, making them more
difficult to resolve from the less abundant clones which targetted for sequencing. In
addition to the 28 and 12S seed storage proteins, the castor seed accumulates the toxic
protein ricin, to about 1.5 % of the total protein content. Several clones (pCRS424,
W7, and pCRS826) encoding ricin or the closely related agglutinin were
sequenced.

A large number of clones encoding a great variety of ribosomal proteins were

156
sequenced, and are listed in Table 10. These include those previosly sequenced from

higher plants, and those for which the strongest match was not to a higher plant gene,
and therefore have presumably not previously been sequenced from higher plants. It
should be noted that the databases used for clone identification in this project do not
include dbEST, in which expressed sequence tags (ESTs) from projects similar to this
one are deposited. Ribosomal protein genes are also a prominent class of clones
obtained in the other plant EST projects, and there may be some overlap between
those identified in this and the other projects.

A number of clones were identified that encode components of protein
synthesis other than ribosomal proteins. These components are: asparaginyl-tRNA
synthetase (pCR8296, see Table 8); initiation factor 5A (pCRS385); elongation factor
la (pCRS349 and pCR8401, see Table 9); and elongation factor 2 (pCRS680, see
Table 9).

Some proteins are known which control storage protein gene expression in
maize, including Opaque2 and Opaque2 heterodimerizing protein (OHPl), which are
components of a heterodimeric transcriptional activator of the zein seed storage
proteins.“ This type of regulation is not yet understood in any oilseed, but is an
important biotechnological target for increasing the ratio of oil to protein in oilseeds.
A putative castor homologue for OHPl (pCRS359) was identified and is being further
characterised by Dr. James Zhang of our laboratory.

A number of clones were identified which may have a role in protein

processing and transport. These include protein disulphide isomerase (above) which

157

is an endoplasmic reticulum lumen protein facilitating the correct folding of proteins
in the secretory pathway, including the seed storage proteins and ricin which are
transported to the vacuole. Similarly, heat shock proteins (pCR8264, pCRS348,
pCRS382, pCRS397, pCRSS98, pCR8657, pCRS759, pCRS820, pCRS833; see
also Table 7) are molecular chaperones involved in protein folding during normal cell
metabolism and may be abundantly expressed in the developing castor seed in
association with active storage protein synthesis. Storage protein transport via the
golgi aparatus presumably requires quite active golgi function, and a protein involved
in endoplasmic reticulum to golgi vesicle traffic (Sec12p) was homologous to one of
the castor clones (pCR8275). Lastly, a clone was identified (pCRS463, see Table 8)
which has homology to bacterial protein processing and/or export proteins. It is
speculated that this castor clone might have a role in vacuolar protein import.

Growth of the developing castor seed tissue and storage protein deposition in
the vacuoles of the growing cell requires enlargement of the vacuoles themselves.
This suggests a reason for the abundance of tonoplast proteins among the clones
sequenced, namely tonoplast intrinsic protein (pCR8279, pCRS360, pCR8663,
pCRS749, pCR8765, pCRS83l) and vacuolar ATPase (pCR8512, and pCRSS45,

pCRS785-see Table 9).

Ricinine Biosynthesis
Ricinine (N-methyl—4-methoxy-3-cyano—2-pyridone) accumulates in various

parts of the castor plant,“ and is derived from quinolinic acid.°7'°8 or possible

158

relevance to the formation of the methoxy and keto substituents are the identification

of a putative O-methyltransferase (pCRS696) and a putative dioxygenase (pCRS738),

respectively.

Polyamine Biosynthesis

A number of enzymes involved in polyamine biosynthesis“9 were identified.
These are S-adenosylmethionine synthetase (pCR8554), a stolon tip-induced protein
from potato tentatively identified70 as S-adenosylmethionine decarboxylase
(pCRSSSS), spermidine synthase (pCRS763, see Table 9), and enzymes of arginine
biosynthesis (pCR8519 and pCRS772, see Table 8). Polyamines are essential
components of all cells, though their functions in viva remain unclear". Increased
synthesis of polyamines appears to correlate with active growth ,‘9'7"72 and the putative
S-adenosylmethionine decarboxylase of potato was identified by differential screening
for clones induced in the stolon tip at the onset of tuberisation, which undergoes rapid
growth."0 Polyamines such as putrescine are precursors of some alkaloids, for
example nicotine is synthesised from putrescine and nicotinic acid.73 The castor seed
accumulates the alkaloid ricinine, but this is a pyridine alkaloid, the biosynthesis of

which probably does not involve polyar'rlines.““'67

Organ-Specific Expression
Several clones appear to be homologous to known embryo- or seed-specific

proteins. These matches are to trithorax (pCR8282, see Table 9), M025 (pCRS347,

159
see Table 9), embryonic gene BP8 of birch (pCR837l), and embryonic abundant

protein of faba bean (pCRSS65). However, in the latter case, there is also a weaker,
but apparently still homologous (since residues conserved are a subset of those
conserved with the bean protein) match with "shoot-specific protein" of pea.
Contradictions of supposed specificity are observed in two other cases.
"Pollen-specific” proteins of maize, rice, tomato and olive are homologous to
pCR8589. A clue to this apparent contradiction is that the olive protein is also
described as the major pollen allergen, while the castor seed is known to contain
allergenic proteins. In the second example of an apparently contradictory match, two
clones (pCR8406, pCR8430) had homology to "lupin-specific” clone pplz02. Not

surprisingly, the evidence for specificity to the lupin plant was very feeble."

Defense-Related Proteins

Seeds typically contain various proteins presumed to have a role in defense
against microbial or insect attack, and the castor seed appears to be no exception, its
other defenses (such as ricin and ricinine) notwithstanding. A number of protease
inhibitor clones (pCRS413, pCRS648, pCRS653, pCRS754) and thaumatin-like

clones (pCR8580, pCR8711) were identified. '

Dessicatian-Related Proteins
Two clones (pCRSSZl, pCR8557) identify sequences characteristic of

dessication-induced proteins in other plants, typical of the later stages of seed

160

development.

Database Accession Numbers of Sequence Data

DNA sequences generated in this study have been deposited in the NCBI
database, dbEST (database for Expressed Sequence Tags), as identification numbers
39704-40169, and in GenBank, as accession numbers T14820-T15266.
Correspondence between the clone numbers used here, dbEST identification numbers,

and GenBank accession numbers is depicted in Table 12.

161

Table 12. Database (dbEST, GenBank) accession numbers of sequences obtained in
this study, corresponding to clone numbers used in the text.

 

 

dbEST Clone GenBank dbEST Clone GenBank
39704 crs262 Tl4820 39741 crs304 T14857
39705 cr8263 T14821 39742 crs305 T14858
39706 crs264 T14822 39743 crs309 T14859
39707 crs265 T14823 39744 crs310 T14860
39708 crs266 Tl4824 39745 crs311 T14861
39709 crs267 T14825 39746 crs312 Tl4862
39710 crs268 Tl4826 39747 crs313 Tl4863
39711 crs269 T14827 39748 crs314 T14864
39712 crs270 T14828 39749 crs315 T14865
39713 crs273 T14829 39750 crs316 T14866
39714 crs274 Tl4830 39751 crs317 T14867
39715 crs275 T14831 39752 crs319 T14868
39716 crs279 T14832 39753 crs320 T14869
39717 crs280 T14833 39754 crs32l T14870
39718 cr8281 Tl4834 39755 crs322 T14871
39719 crs282 T14835 39756 crs323 T14872
39720 crs283 Tl4836 39757 crs324 Tl4873
39721 crs284 T14837 39758 crs325 T14874
39722 crs285 Tl4838 39759 crs328 T14875
39723 crs286 T14839 39760 crs330 Tl4876
39724 cr8287 T14840 39761 crs331 T1487?
39725 cr8288 T14841 39762 crs332 T14878
39726 crs289 T14842 39763 crs333 T14879
39727 cr8290 Tl4843 39764 crs334 T14880
39728 cr8291 T14844 39765 crs335 T14881
39729 cr8292 Tl4845 39766 crs336 T14882
39730 cr8293 T14846 39767 crs337 T14883
39731 crs294 T1484? 39768 crs339 T14884
39732 crs295 Tl4848 39769 crs340 Tl4801
39733 crs296 T14849 39770 crs341 T14802
39734 crs297 T14850 39771 crs342 T14803
39735 crs298 T14851 39772 cr8343 T14804
39736 crs299 T14852 39773 crs345 Tl4805
39737 crs300 T14853 39774 crs346 T14806
39738 crs301 T14854 39775 cr5347 Tl4807
39739 crs302 T14855 39776 crs348 T14808
39740 crs303 Tl4856 39777 crs349 T14809

 

Table 12 (cont’d).

 

dbEST Clone GenBank
39778 crs350 T14810
39779 crs351 T14811
39780 crs352 T14812
39781 crs353 T14813
39782 cr8354 T14814
39783 crs355 T14815
39784 crs356 T14816
39785 crs35 8 T14817
39786 crs359 T14818
39787 crs360 T14819
39788 crs361 T14885
39789 crs362 T14886
39790 crs363 T14887
39791 crs364 T14888
39792 crs366 Tl4889
39793 crs368 T14890
39794 crs369 Tl4891
39795 crs370 T14892
39796 crs371 T14893
39797 crs372 T14894
39798 crs373 Tl4895
39799 crs374 Tl4896
39800 crs375 T14897
39801 crs376 T14898
39802 crs377 T14899
39803 crs378 T14900
39804 crs380 T14901
39805 c3382 T14902
39806 crs384 T14903
39807 crs385 T14904
39808 crs386 T14905
39809 crs3 87 T14906
39810 crs388 T14907
39811 crs389 T14908
39812 crs39l T14909
39813 crs392 T14910
39814 crs393 T1491 1
39815 crs394 T14912
39816 crs395 T14913
39817 crs396 T14914
39818 crs397 T14915

 

dbEST Clone GenBank
39819 crs398 T14916
39820 crs399 T14917
39821 crs400 T14918
39822 ch01 Tl4919
39823 crs403 T14920
39824 crs404 T14921
39825 crs405 Tl4922
39826 crs406 T14923
39827 crs407 T14924
39828 crs408 Tl4925
39829 crs409 T14926
39830 crs410 Tl4927
39831 crs411 T14928
39832 crs412 T14929
39833 crs413 Tl4930
39834 crs414 T1493l
39835 crs415 T14932
39836 crs416 T14933
39837 crs417 T14934
39838 ch19 T14935
39839 crs420 Tl4936
39840 ch21 Tl4937
39841 crs424 T14938
39842 ch25 T14939
39843 crs426 T14940
39844 crs427 T14941
39845 crs428 T14942
39846 crs430 T14943
39847 crs431 Tl4944
39848 crs432 T14945
39849 crs433 Tl4946
39850 crs434 T14947
39851 crs435 T14948
39852 crs436 T14949
39853 crs437 Tl4950
39854 crs438 T14951
39855 crs439 Tl4952
39856 ch40 Tl4953
39857 crs441 T14954
39858 crs442 T14955
39859 crs443 T14956

Table 12 (cont’d).

 

dbEST Clone GenBank
39860 crs444 T14957
39861 crs445 T1495 8
39862 ch46 T14959
39863 crs447 Tl4960
39864 ch48 T14961
39865 crs449 Tl4962
39866 crs450 T14963
39867 crs45 1 T14964
39868 crs452 T14965
39869 crs453 T14966
39870 crs454 Tl4967
39871 crs455 T14968
39872 crs45 6 Tl4969
39873 crs457 T14970
39874 crs45 8 T14971
39875 crs459 T14972
39876 crs460 T14973
39877 crs461 T14974
39878 crs463 T14975
39879 crs464 Tl4976
39880 crs466 T14977
39881 crs467 T14978
39882 ch69 T14979
39883 crs470 T14980
39884 crs493 T14981
39885 crs494 Tl4982
39886 crs495 T14983
39887 crs496 T14984
39888 crs497 Tl4985
39889 crs498 Tl4986
39890 crs499 T14987
39891 crs500 T14988
39892 crs501 T14989
39893 crs502 T14990
39894 crs503 T14991
39895 crs504 T14992
39896 crs505 T14993
39897 crs506 T14994
39898 crs507 T14995
39899 crs509 T14996
39900 crsSll T14997

 

dbEST Clone GenBank
39901 crs512 Tl4998
39902 crs513 T14999
39903 crs514 T15000
39904 cr8515 T15001
39905 crs516 T15002
39906 crs517 T15003
39907 ch8 T15004
39908 ch9 T15005
39909 crs520 T15006
39910 cr852l T15007
39911 cr8522 T15008
39912 cr8523 T15009
39913 crs524 T15010
39914 crs539 T15011
39915 crs540 T15012
39916 crs541 T15013
39917 crs542 T15014
39918 crs543 T15015
39919 CI'8544 T15016
39920 crs545 T15017
39921 crs546 T15018
39922 crs547 T15019
39923 crs548 T15020
39924 crs549 T15021
39925 crs550 T15022
39926 cr8551 T15023
39927 cr8552 T15024
39928 crs553 T15025
39929 cr8554 T15026
39930 crs555 T15027
39931 crs556 T15028
39932 crs557 T15029
39933 crs558 T15030
39934 crs559 T1503]
39935 crs560 T15032
39936 crs562 T15033
39937 cr8563 T15034
39938 cr8564 T15035
39939 crs565 T15036
39940 crs566 T15037
39941 crs567 T15038

Table 12 (cont’d).

 

dbEST Clone GenBank
39942 cr8568 T15039
39943 crs569 T15040
39944 cr8570 T15041
39945 crs571 T15042
39946 crs578 T15043
39947 crsS79 T15044
39948 crs580 T15045
39949 crs581 T15046
39950 crs583 T1504?
39951 cr8584 T15048
39952 cr8585 T15049
39953 cr8586 T15050
39954 crs587 T15051
39955 crs588 T15052
39956 crs589 T15053
3995? crs590 T15054
39958 crs592 T15055
39959 crs593 T15056
39960 cr8594 T1505?
39961 crs595 T15058
39962 crs596 T15059
39963 crs598 T15060
39964 cr8599 T15061
39965 crs700 T15062
39966 crs701 T15063
39967 crs702 T15064
39968 crs703 T15065
39969 crs704 T15066
39970 crs705 T1506?
39971 crs706 T15068
39972 crs707 T15069
39973 crs708 T150?0
39974 crs709 T15071
39975 crs710 T15072
39976 crs711 T15073
39977 crs712 T15074
39978 crs713 T15075
39979 crs7l4 T15076
39980 crs? 15 T1507?
39981 crs717 T15078
39982 crs718 T15079

 

dbEST Clone GenBank
39983 crs720 T15080
39984 crs721 T15081
39985 crs722 T15082
39986 crs724 T15083
39987 crs725 T15084
39988 crs727 T15085
39989 crs728 T15086
39990 crs729 T1508?
39991 crs730 T15088
39992 crs731 T15089
39993 ‘ crs733 T15090
39994 crs734 T15091
39995 crs735 T15092
39996 crs736 T15093
3999? crs737 T 15094
39998 crs738 T15095
39999 crs739 T15096
40000 crs74l T1509?
40001 crs742 T15098
40002 crs743 T15099
40003 crs744 T15100
40004 crs?45 T15101
40005 crs746 T15102
40006 crs747 T15103
40007 crs748 T15104
40008 crs749 T15105
40009 crs750 T15106
40010 crs751 T1510?
40011 crs752 T15108
40012 crs753 T15109
40013 crs756 T15110
40014 crs75? T15111
40015 crs758 T15112
40016 crs759 T15113
40017 crs763 T15114
40018 crs764 T15115
40019 crs765 T15116
40020 crs766 T151 17
40021 crs767 T15118
40022 crs768 T15119
40023 crs770 T15120

Table 12 (cont’d).

 

dbEST Clone GenBank
40024 crs?71 T15121
40025 crs772 T15122
40026 crs773 T15123
40027 crs774 T15124
40028 crs775 T15125
40029 crs776 T15126
40030 crs778 T1512?
40031 crs779 T15128
40032 crs780 T15129
40033 crs?81 T15130
40034 crs782 T15131
40035 crs783 T15132
40036 crs784 T15133
4003? crs?85 T15134
40038 crs786 T15135
40039 crs787 T15136
40040 crs788 T1513?
40041 crs?89 T15138
40042 crs790 T15139
40043 crs791 T15l40
40044 crs792 T15141
40045 crs793 T15142
40046 crs794 T15143
40047 crs796 T15144
40048 crs797 T15145
40049 crs798 T15146
40050 crs601 T15147
40051 crs602 T15148
40052 crs628 T15149
40053 crs629 T15150
40054 crs630 T15151
40055 crs63l T15152
40056 crs633 T15153
40057 crs634 T15154
40058 crs635 T15155
40059 crs636 T15156
40060 crs637 T15157
40061 crs640 T15158
40062 crs641 T15159
40063 crs642 T15160
40064 crs643 T15161

 

dbEST Clone GenBank
40065 crs645 T15162
40066 crs64? T15163
40067 crs648 T 15164
40068 crs649 T15165
40069 crs650 T15166
40070 crs651 T1516?
40071 crs652 T15168
40072 crs653 T15169
40073 crs654 T15170
40074 crs655 T15171
40075 crs656 T15172
40076 crs657 T15173
4007? crs658 T15174
40078 crs659 T15l75
40079 crs660 T15176
40080 crs661 T1517?
40081 crs662 T15178
40082 crs663 T15179
40083 crs666 T15180
40084 crs667 T15181
40085 crs668 T15182
40086 crs669 T15183
40087 crs670 T15184
40088 crs67l T15185
40089 crs672 T15186
40090 crs673 T1518?
40091 crs674 T15188
40092 crs675 T15189
40093 crs676 T15190
40094 crs677 T15191
40095 crs678 T15192
40096 crs679 T15193
4009? crs680 T15194
40098 crs68l T15195
40099 crs682 T15196
40100 crs683 T1519?
40101 crs684 T15198
40102 crs685 T15199
40103 crs686 T15200
40104 crs689 T15201
40105 crs690 T15202

Table 12 (cont’d).

 

dbEST Clone GenBank
40106 crs691 T15203
40107 crs692 T15204
40108 crs693 T 15205
40109 crs694 T 15206
401 10 crs696 T1520?
401 l 1 crs69? T15208
40112 crs698 T15209
40113 crs699 T15210
40114 crs803 T15211
40115 crs804 T15212
40116 crs805 T15213
40117 crs806 T15214
40118 crs807 T15215
40119 crs808 T15216
40120 crs810 T1521?
40121 crs8ll T15218
40122 cr8812 T15219
40123 cr8813 T15220
40124 crs814 T15221
40125 crs815 T15222
40126 cr8816 T15223
40127 crs817 T15224
40128 crs8l8 T15225
40129 crs820 T15226
40130 crs821 T1522?
40131 crs822 T15228
40132 crs823 T15229
40133 crs824 T1523O
40134 crs825 T15231
40135 crs826 T15232
40136 crs827 T15233
40137 crs828 T15234
40138 crs829 T15235
40139 crs831 T15236
40140 crs832 T1523?
40141 crs833 T15238
40142 crs834 T15239
40143 crs835 T15240

 

dbEST Clone GenBank
40144 crs836 T15241
40145 crs837 T15242
40146 crs838 T15243
40147 crs839 T15244
40148 crs840 T15245
40149 crs841 T15246
40150 crs842 T1524?
40151 crs843 T15248
40152 crs844 T15249
40153 crs848 T15250
40154 crs849 T 15251
40155 crs852 T15252
40156 cr8853 T15253
40157 crs854 T15254
40158 crs855 T15255
40159 cr8856 T15256
40160 crs857 T1525?
40161 crs858 T15258
40162 cr8859 T15259
40163 crs860 T15260
40164 crs86l T1526l
40165 cr8862 T15262
40166 crs863 T15263
4016? crs864 T15264
40168 crs865 T15265
40169 crs754 T15266

 

16?
CONCLUSIONS

The objective of this experiment was to identify a putative oleate-12-
hydroxylase cDNA, and this was achieved. The two clones pCRS677 and pCRS 834
have sequence similarity with plant membrane-bound desaturase genes, but do not
encode a previously identified desaturase. These clones are examined in detail in the
subsequent chapter.

The secondary objective of this experiment was to isolate other ovel plant
genes, and this also was achieved. A large proportion of the clones sequenced could
be identified by comparison to sequence databases, which may reflect the relatively
high levels of expression in the developing castor md of relatively well characterised
enzymes, such as those of glycolysis. Among the sequences identified were a
coniderable number which appear to be previously unknown or uncharacterised in

higher plants, and these are expected to provide new oportunities in plant biology.

REFERENCES
1. Coppens, N. , Biosynthesis of fatty acids by seeds of Ricinus communis, Nature,
177, 279, 1956.

2. Canvin, D.T., Formation of oil in the seed of Ricinus communis L., Can. J.
Biochem. Physiol., 41, 1879-1885, 1963.

3. Yamada, M., and Stumpf, P.K., Enzymic synthesis of ricinoleic acid by
extracts of developing Ricinus communis L. seeds, Biochem. Biophys. Res.
Commun., 14, 165-171, 1964.

4. James, A.T. Hadaway, H.C., and Webb, J.P.W., The biosynthesis of ricinoleic

10.

11.

12.

l3.

14.

15.

168
acid, Biochem. J., 95, 448-452, 1965.

Hajela, R.K., Horvath, D.P., Gilmour, S.J., Thomashow, M.F., Molecular
cloning and expression of car (cold-regulated) genes in Arabidapsis thaliana,
Plant Physiol., 93, 1246-1252, 1990.

Shanklin, J. and Somerville, C. , Stearoyl-acyl-carrier-protein desaturase
from higher plants is structurally unrelated to the animal and fungal
homologs, Prac. Natl. Acad. Sci. USA, 88, 2510, 1991.

Sambrook, J ., Fritsch, E.F., and Maniatis, T., Molecular Cloning: a
Laboratory Manual, 2nd ed., Cold Spring Harbor Laboratory Press, 1989.

Kamalay, J .C. , and Goldberg, R.B. , Regulation of structural gene expression
in tobacco, Cell, 19, 935-946, 1980.

Kamalay, J .C., and Goldberg, R.B., Organ-specific nuclear RNAs in tobacco,
P.N.A.S. USA, 81, 2801-2805, 1984.

Altschul, 8.F., Gish, W., Miller, W., Myers, B.W., and Lipman, D.J., Basic
local alignment search tool, J. Mal. Biol., 215, 403-410, 1990.

Adams, M.D., Dubnick, M., Kerlavage, A.R., Moreno, R., Kelley, J.M.,
Utterback, T.R., Nagle, J .W., Fields, C., and Venter, J .C., Sequence
identification of 2375 human brain genes, Nature, 355, 632-634, 1992.

Adams, M.D., Kelley, J.M., Gocayne, J.D., Dubnick, M., Polymeropoulos,
M.H., Xiao, H., Merril, C.R., Wu, A., Olde, B., Moreno, R.F., Kerlavage,
A.R., McCombie, W.R., and Venter, J .C., Complementary DNA sequencing :
expressed sequence tags and human genome project, Science, 252, 1651-1656,
1991.

Keith, C.S., Hoang, D.O., Barrett, BM, Feigelman, B., Nelson, M.C., Thai,
H., and Bayersdorfer, C. , Partial sequence analysis of 130 randomly selected
maize cDNA clones, Plant Physiol., 101, 329-332, 1993.

Uchimiya, H., Kidou, 8., Shimazaki, T., Aotsuka, 8., Takamatsu, 8., Nishi,
R., Hashimoto, H., Matsubayashi, Y., Kidou, N., Umeda, M., and Kato, A.,
Random sequencing of cDNA libraries reveals a variety of expressed genes in
cultured cells of rice (Oryza sativa L.), Plant J., 2, 1005-1009, 1992.

Hoog, C., Isolation of a large number of novel mammalian genes by a
differential cDNA library screening strategy, Nucleic Acids Research, 19,
6123-6127, 1991.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

169

Parsot, C. , Evolution of biosynthetic pathways: a common ancestor for
threonine synthase, threonine dehydratase and D-serine dehydratase, EMBO J. ,
5, 3013-3019, 1986.

Anselme, J. , and Hartlein, M., Tyr-426 of the Escherichia coli asparaginyl-
tRNA synthetase, an amino acid in a conserved C-terminal motif, is involved in
ATP binding, FEBS Lett., 280, 163-166, 1991.

Anselme, J ., and Hartlein, M., Asparaginyl—tRNA synthetase from Escherichia
coli has significant sequence homologies with yeast aspartyl-tRN A synthetase,
Gene, 84, 481-485, 1989.

Youvan, D.C., Bylina, E.J., Alberti, M., Begusch, H., and Hearst, J.E.,
Nucleotide and deduced polypeptide sequences of the photosynthetic reaction-
center, B870 antenna, and flanking polypeptides from R. capsulata, Cell, 37,
949-957, 1984.

Bartsch, K. , Johnn-Marteville, A. von, and Schulz, A., Molecular analysis of
two genes of the Escherichia coli gab cluster: nucleotide sequence of the
glutamate:succinic semialdehyde transaminase gene (gabT) and characterisation
of the succinic semialdehyde gene (gabD), J. Bacterial, 172, 7035-7042,
1990.

Matsubara, Y., Indo, Y., Naito, E., Ozasa, H., Glassberg, R., Vockley, J.,
Ikeda, Y., Kraus, J ., and Tanaka, K., Molecular cloning and nucleotide
sequence of cDN As encoding the precursors of rat long chain acyl-coenzyme
A, short-chain acyl-coenzyme A, and isovaleryl-coenzyme A dehydrogenases;
sequence homology of four enzymes of the acyl-CoA dehydrogenase family, J.
Biol. Chem., 264, 16321-16331, 1989.

Komine, Y., and Inokuchi, H., Precise mapping of the mpB gene encoding the
RNA component of RNaseP in Escherichia coli K-12, J. Bacterial, 173, 1813-
1816, 1991.

Ohlrogge, J .B., Jaworski, LG, and Post-Beittenmiller, D., De nova fatty acid
biosynthesis, in Lipid Metabolism in Plants, Moore, T.S., Ed., CRC Press,
Boca Raton, 1993, chap. 1.

Sparace, 8.A., and Kleppinger-Sparace, K.F., Metabolism in
nonphotosynthetic, nonoilseed tissues, in Lipid Metabolism in Plants, Moore,
T.S., Ed., CRC Press, Boca Raton, 1993, chap. 17.

Petrovics, G., Putnoky, P., Kim, J., Tharp, T.A., Noel,. K.D., Carlson,
R.W. , and Kondorosi, A. , The presence of a novel type of surface

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

170

polysaccharide in Rhizabiurn melilati requires a new fatty acid synthase-like
gene cluster involved in symbiotic nodule development, Molecular
Microbiology, 8, 1083-1094, 1993.

Mathur, M. , and Kolattukudy, P.E. , Molecular cloning and sequencing of the
gene for mycocerosic acid synthase, a novel fatty acid elongating
multifunctional enzyme, from Mycobacterium tuberculosis var. bavis Bacillus
Calrnette-Guerin, J. Biol. Chem., 267, 19388-19395, 1992.

Donadio, 8., Staver, M.J., McAlpine, J.B., Swanson, S.J., Katz, L., Modular
organization of genes required for complex polyketide biosynthesis, Science,
252, 675-679, 1991.

Coleman, J ., and Raetz, C.R.H., First committed step of lipid A biosynthesis
in Escherichia coli: sequence of the lpr gene, J. Bacteriol, 170, 1268-1274,
1988.

Crowell, D.N., Reznikoff, W8, and Raetz, C.R.H., Nucleotide sequence of
the Escherichia coli gene for lipid A disaccharide synthase, J. Bacterial, 169,
5727-5734, 1987.

Tomasiewicz, H.G. , and McHenry, C.8. , Sequence analysis of the Escherichia
coli dnaE gene, J. Bacterial, 169, 5735-5744, 1987.

Raetz, C.R.H., Biochemistry of endotoxins, Annu. Rev. Biochem., 59, 129-
170, 1990.

Holst, 0., Muller-Loennies, 8., Lindner, B., and Brade, H., Chemical
structure of the lipid A of Escherichia coli J-5, Eur. J. Biochem., 214, 695-
701, 1993.

Helander, l.M., Lindner, B., Seydel, U., and Vaara, M., Defective
biosynthesis of the lipid A component of temperature-sensitive firA (omsA)
mutant of Escherichia coli, Eur. J. Biochem., 212, 363-369, 1993.

Anderson, MS, and Raetz, C.R.H., UDP-N-Acetylglucosamine 3-O-
Acyltransferase from Escherichia coli, Meth. Enzymal. , 209, 449-454, 1992.

Raetz, C.R.H., Lipid A disaccharide synthase from Escherichia coli, Meth.
Enzymal., 209, 455-466, 1992.

Brozek, K.A., and Raetz, C.R.H., Biosynthesis of lipid A in Escherichia coli:

acyl carrier protein-dependent incorporation of laurate and myristate, J. Biol.
Chem., 265, 15410-15417, 1990.

37.

38.

39.

40.

41.

42.

43.

45.

46.

47.

171

Brozek, K.A., and Raetz, C.R.H., 3-Deoxy-D-manno-octulosonate transferase
and late acyltransferases of lipopolysaccharide biosynthesis, Meth. Enzymal. ,
209, 476-485, 1992.

Cronan, J .E., Jr., Li, W.-B., Coleman, R., Narasimhan, M., Mendoza, D. de,
and Schwab, J .M. , Derived amino acid sequence and identification of active
site residues of Escherichia coli B-hydroxydecanoyl thioester dehydrase, J.
Biol. Chem., 263, 4641-4646, 1988.

Shimakata, T., and Stumpf, P.K., Purification and characterizations of B-
ketoacyl-[acyl carrier protein] reductase, B-hydroxyacyl-[acyl carrier protein]
dehydrase, and enoyl-[acyl carrier potein] reductase from Spinacia aleracea
leaves, Arch. Biochem. Biaphys., 218, 77-91, 1982.

Foster, 8.]. , Purification and characterization of an ”actomyosin" complex
from Escherichia coli W3110, FEMS Microbiol. Lett., 110, 295-298, 1993.

Sasaki, Y., Hakamada, K., Suama, Y., Nagano, Y., Furusawa, I., and
Matsuno, R. , Chloroplast-encoded protein as a subunit of acetyl-CoA
carboxylase in pea plant, J. Biol. Chem., 268, 25118-25123, 1993.

Tsinoremas, N.F., Castets, A.M., Harrison, M.A., Allen, J .F., and Tandeau
de Marsac, N. , Photosythetic electron transport controls nitrogen assimilation
in cyanobacteria by means of posttranslational modification of the glnB gene
product, Prac. Natl. Acad. Sci. USA, 88, 4565-4569, 1991.

Edwards, J .W. , and Coruzzi, G.M., Photorespiration and light act in concert to
regulate the expression of the nuclear gene for chloroplast glutamine synthetase,
Plant Cell, 1, 241-248, 1989.

Garrard, L.J., and Goodman, J .M., Two genes encode the major membrane-
associated protein of methanol-induced peroxisomes from Candida baidinii, J.
Biol. Chem., 264, 13929-13937, 1989. '

Gould, S.J., Keller, G.-A., Schneider, M., Howell, S.H., Garrard, L.J.,
Goodman, J .M., Distel, B., Tabak, H., and Subramani, 8., Peroxisomal

protrin import is conserved between yeast, plants, insects and mammals, EMBO
J., 9, 85-90, 1990.

Gould 8.1 ., Keller, G.-A., Hosken, N., Wilkinson, J ., and Subramani, 8., A
conserved tripeptide sorts proteins to peroxisomes, J. Cell Biol. , 108, 1657-
1664, 1989.

Olsen, L.J., Ettinger, W.F., Damsz, B., Matsudaira, K., Webb, M.A.,

48.

49.

50.

51.

52.

53.

54.

55.

56.

172

Harada, J .J . , Targetting of glyoxisomal proteins to peroxisomes in leaves and
roots of a higher plant, Plant Cell, 5, 941-952, 1993.

Mazo, A.M., Huang, D., Mozer, B.A., and Dawid, I.B., The trithorax gene, a
trans-acting regulator of the bithorax complex in Drosaphila, encodes a protein
with zincbinding domains, Prac. Natl. Acad. Sci. USA, 87, 2112-2116, 1990.

Tkachuk, D.C., Kohler, 8., and Cleary, M.L., Involvement of a homolog of
Drosophila trithorax by 11q23 chromosomal translocations in acute leukemias,
Cell, 71, 691-700.

Miyamoto, H., Matsushiro, M. , and Nozaki, M. , Molecular cloning of a novel
mRNA sequence expressed in cleavage stage mouse embryos, Molecular
Reproduction and Development, 34, 1-7, 1993.

Kushnirov, V.V., Ter-Avanesyan, M.D., Didichenko, 8.A., Smirnov, V.N.,
Chernoff, Y.O., Derkach, I.I., Novikova, O.N., Inge-Vechtomov, G.I.,
Neistat, M.A. , and Tolstorukov, I.I. , Divergence and conservation of sup2
(sup35) gene of yeasts Pichia pinus and Saccharomyces cerevisiae, Yeast, 6,
461-472, 1990. '

Hoshino, 8., Miyazawa, H., Enomoto, T., Hanaoka, F., Kikuchi, Y., Kikuchi,
A., and U1, M., A human homologue of the yeast GSTI gene codes for a GTP-
binding protein and is expressed in a proliferation-dependent manner in
mammalian cells, EMBO J., 8, 3807-3814, 1989.

Lorenz, R.T., and Parks, L.W., Cloning, sequencing, and disruption of the
gene encoding sterol C- 14 reductase in Saccharomyces cerevisiae, DNA Cell
Biol, 11, 685-692, 1992.

Shimanuki, M., Goebl, M., Yanagida, M., and Toda, T., Fission yeast stsl+
gene encodes a protein similar to the chicken lamin B receptor and is impicated
in pleiotropic drug-sensitivity, divalent cation-sensitivity, and osmoregulation,
Mol. Biol. Cell, 3, 263-273, 1992.

Sibille, C., Chomez, P., Wildmann, C., Van Pel, A., De Plaen, E.,
Maryanski, J.L., Bergeyck, V de, and Boon, T., Structure of the gene of tum‘
transplantation antigen P198: a point mutation generates a new antigenic
peptide, J. Exp. Med, 172, 35-45, 1990'.

Werner, A., Moore, M.L., Mantei, N., Biber, J ., Semenza, G., and Murer,
H. , Cloning and expression of cDNA for a Na/Pi cotransport system of kidney
cortex, Prac. Natl. Acad. Sci., 88, 9608-9612, 1991.

57.

58.

59.

61.

62.

63.

65.

67.

173

Poitier, Y, Thoma, 8., Somerville, C., and Schieffelbein, J ., A mutant of
Arabidapsis deficient in xylem loading of phosphate, Plant Physiol. , 97, 1087-
1093, 1991.

Rokeach, L.A., Haselby, J .A., and Hoch, 8.0., Molecular cloning of a cDNA
encoding the human Sm-D autoantigen, Proc. Natl. Acad. Sci. USA, 85 , 4832-
4836, 1988.

Raikhel, N., Nuclear targetting in plants, Plant Physiol., 100, 1627-1632,
1992.

Lord, J .M., Lamb, I.F., and Roberts, L.N., Ricin: structure, biological
activity and synthesis, Oxford Surveys Plant Malec. Cel Biol, 1, 85-101, 1984.

Sullivan, T.D., Strelow, L.I., Illingworth, C.A., Phillips, R.L., and Nelson,
O.E., Jr., Analysis of maize Brittle-I alleles and a defective suppressor-
mutatar-induced mutable allele, Plant Cell, 3, 1337-1348, 1991.

Smith, R.G., Gauthier, D.A., Dennis, D.T., and Turpin, D.H., Malate- and
pyruvate-dependent fatty acid synthesis in leucoplasts from developing castor
endosperm, Plant Physiol., 98, 1233, 1992.

Wetterau, J.R., Combs, K.A., Spinner, SN, and Joiner, B.J., Protein
disulfide isomerase is a component of the microsomal triglyceride transfer
protein complex, J. Biol. Chem., 265, 9800-9807, 1990.

Sharp, D., Blinderman, L., Combs, K.A., Kienzle, B., Ricci, B., Wager-
Smith, K., Gil, C.M., Turck, C.W., Bouma, M.-E., Rader, D.J., Aggerbeck,
L.P., Gregg, R.E., Gordon, D.A., and Wetterau, J .R., Cloning and gene
defects in microsomal triglyceride transfer protein associated with
abetalipoproteinaemia, Nature, 365, 65-69, 1993.

Pysh, L.D., Aukerrnan, M.J., and Schmidt, R.J., OHPl: a maize basic
domain/leucine zipper protein that interacts with Opaque2, Plant Cell, 5, 227-
236, 1993.

Rizk, A.-F.M., and El-Missiry, M.M., Non-diterpenoid constituents of
Euphorbiaceae and Thymelaeaceae, in Naturally Ocurring Pharbol Esters,
Evans, P.J., Ed., CRC Press, Boca Raton, 1986, chap 5.

Hiles, R.A., The chemistry and biochemistry of ricinine and some related
pyridine compounds, PhD Thesis, Department of Biochemistry, Michigan State
University, 1970.

68.

69.

70.

71.

72.

73.

74.

174

Yang, K.S., and Waller, G.R., Biosynthesis of the pyridine ring of ricinine
from quinolinic acid glycerol and aspartic acid, Phytochem. , 4, 881-889, 1965.

Galston, A.W. and Kaur-Sawhney, R. , Polyamines as endogenous growth
regulators, in Plant Hormones and Their Role in Plant Growth and
Development, Davies, P.J., Ed., Martinus Nijhof, Dordrecht, 1987, 280-295.

Taylor, M.A., Mad Arif, 8.A., Kumar, A., Davies, H.V., Scobie, L.A.,
Pearce, S.R. , and Flavell, A.J. , Expression and sequence analysis of cDNAs
induced during the early stages of tuberisation in different organs of the potato
plant (Salanum tuberasurn L.), Plant Mal. Biol, 20, 641-651, 1992.

Smith, T.A., Polyamines, Ann. Rev. Plant Physiol., 36, 117-143, 1985.

Evans, P.T. , and Malmberg, R.L., Do polyamines have roles in plant
development?, Ann. Rev. Plant Physiol. Plant Mol. Biol, 40, 235-269, 1989.

Hamill, T.D., Robins, R.J., Parr, A.J., Evans, D.M., Furze, J.M., and
Rhodes, M.J .C. , Over-expressing a yeast omithine decarboxylase gene in
transgenic roots of Nicatiana rustica can lead to enhanced nicotine
accumulation, Plant Mol. Biol, 15, 27-38, 1990.

Perrey, R. , Warskulat, U. , and Wink, M. , Molecular cloning of a lupin-
specific gene from a cDN A library of suspension-cultured cells of Lupinus
palyphyllus, Plant Mal. Biol, 15, 175-176, 1990.

CHAPTER 6

ANALYSIS OF PUTATIVE OLEATE 12-HYDROXYLASE CLONES

ABSTRACT

Two putative oleate-12-hydroxylase clones, pCRS677 and pCR8834, identified
in chapter 5, were compared and found to contain identical sequences. A full length
clone, pFL2, was isolated using pCRS677 as a probe, and sequenced. An 1161 bp
open reading frame encoding a 4440? Da protein had limited sequence similarity
(37% at the amino acid level) to plant membrane-bound desaturases. Northern blot
analysis revealed that pFL2 corresponds to a strongly-expressed seed-specific
transcript in castor. A Southern blot showed that this transcript originates from a
single-copy gene. The clone pFL2 was engineered for expression in transgenic yeast
and plants. Three plant transformation systems were employed: tobacco leaf explants,
tobacco cultured cells, and carrot root disks were each transformed by co-cultivation
with Agrabacterium tumefaciens. No ricinoleic acid could be detected in any of the
transgenic tissues, nor were there any consistent changes in the relative proportions of
the usual fatty acids. It is concluded that the sequence and pattern of expression of
pFL2 are consistent with the possibility that it encodes oleate-12-hydroxylase, but this
was neither supported nor refuted by attempts at expression in transgenic yeast and

plants.

175

176
INTRODUCTION

In the preceding chapter, an experiment is described in which seed-specific

‘ clones expressed at moderate levels of abundance were selected by differential
screening, and a large number of these clones were partially sequenced. The
principle objective of this experiment was to isolate a gene encoding oleate-12-
hydroxylase. Since it was hypothesized that this gene is homologous to those of the
microsomal fatty acid desaturases, it was hoped that the oleate-12-hydroxylase gene
could be identified on this basis. Two clones, pCRS677 and pCRS834, gave such a
match to the Brassica napus fad3 gene, encoding the microsomal w-3 desaturase, the
only known higher plant membrane-bound desaturase which was in the databases.
These two clones were therefore considered putative oleate-12-hydroxylase clones and
their analysis is the subject of this chapter.

The partial clone pCRS677 was used to isolate full-length cDN A clones, one
of which was completely sequenced. Since ricinoleic acid is only produced in the
seed,“ it was expected that a hydroxylase clone should only be expressed in seed
tissues, and this was examined by northern analysis. In order to test whether the
isolated clone encodes a protein with oleate-12-hydroxylase activity, the clone was

expressed in both yeast and plants.

17?
MATERIALS AND METHODS

Screening of pYE82.0 cDNA Library

Three plates of each of the four pools of the pYE82.0 cDNA library (chapter
3) were screened by the same method described above (chapter 4). The pCRS677
insert was excised with BamHI and ApaI, gel-purified, 32P-labelled by random
priming‘ and purified of unincorporated nucleotides by ethanol precipitation in the
presence of ammonium acetate.‘ This probe was hybridised to the filters overnight at
65°C, and unhybridised probe was removed by washing at room temperature (using

the method and solutions described in chapter4).

DNA Sequencing

Terminal sequencing of various clones in pYE82.0 was as described above
(chapter 4) using the T? fluorescent primer, or the primer F1 (5’ AGC GTG ACA
TAA CTA ATP 3’) with fluorescent terrninators. Sequencing of the entire pFL2 gene
employed, in addition to F1 and T7, the following Oligonucleotide primers, in
combination with fluorescent terminators:

HF2: 5’ GCT CTT TI‘G TGC GCT CAT TC 3’
HF3: 5’ GTC CAT TCT GCA CTI‘ CTG GT 3’

HF4: 5’ ACG ATC GCT 'l'I‘G CTT GCC AT 3’

178
HFS: 5’ GGA GCA ATG GTG ACT GTC GA 3’

HF6: 5’ CAA GGC G'I'T TI‘C TGG TAC CG 3’
HRl: 5’ CGG TAC CAG AAA ACG CCT TG 3’
HR2: 5’ TCG ACA GTC ACC A'IT GCT CC 3’
HR3: 5’ ATG GCA AGC AAA GCG ATC GT 3’
HR4-2: 5’ GAG AAT GCA GCC 'ITG GAA GA 3’
HRS: 5’ GAA TGA GCG CAC AAA AGA GC 3’
HR6: 5’ CTT CAA GCG GGA G'IT CAA CC 3’.

Sequence data was analysed using the programs DNASIS and PROSIS.

Northern Blot Analysis

Poly(A)+ RNA prepared (chapter 3) from leaves and developing seeds was
electrophoresed through an agarose gel containing formaldehyde.’ An equal quantity
(3 pg) of RNA was loaded in both lanes, and RNA standards (0.16-1.77 kb ladder,
Gibco-BRL) were loaded in a third lane. Following electrophoresis, RNA was
transferred from the gel to a nylon membrane (Hybond N, Amersham) and fixed to
the filter by exposure to UV light for 2 min (based on an empirical calibration). A
32P-labelled probe was prepared from insert DNA of clone pCRS677 as above, and
hybridised to the membrane overnight at 65 °C, after it had been prehybridised for
~1 h. The hybridization solution contained 4 X SET (0.6 M NaCl, 0.12 M Tris-HCl

pH 7.4, 8 mM EDTA), 0.1% sodium pyrophosphate, 0.2% SDS, 0.1% heparin, and

179
5% dextran sulphate. The blot was washed three times in 2 X SSC, 0.1% SDS at

room temperature, then exposed to X-ray film, and to a phosphor-imaging screen

(Molecular Dynamics). A probe was subsequently made from the Calletatrichum

graminicala B-tubulin gene TUBZ‘ and hybridised to the same blot under the same
conditions, except that the hybridization temperature was reduced to 58°C, and

exposed to X-ray film.

Southern Blot Analysis

Genomic Arabidapsis DNA (1 pg, courtesy of Dr. Yves Poirier) and genomic
castor DNA (2 pg, courtesy of Dr. John Shanklin) were digested with EcoRI, BamI-II,
or HindIII, in reactions supplemented with spermidine (1 mM), and separated in 0.7%
agarose gel. A Southern blot was prepared as described in chapter 4, and hybridised

and washed in the solutions described for the northern blot, above.

Yeast Strains and Transformation

Yeast strain CGY2557 is described in chapter 3. The strain J0522 was
obtained from Dr. Joe Ogas: MATa, GAL+, ura3-52, trp1A65, leu2, his3AI, pm]-
1122, pep4-3, prcI-407.

A method for electroporation of yeast was used that was simpler, but less

efficient, than that used in chapter 3. Yeast were grown to OD600 = 1.0 in 100 ml

180
YPD medium (10 g l" yeast extract, 20 g " bacto peptone, 20 g l" dextrose). Cells

were pelleted (SS-34 rotor, 5000g, 5 min, 4°C) and resuspended in ~200 pl of the
supernatant. These cells (40 p1 per transformation) were used for electroporation with
a minimal volume of DNA in cold, 2 mm-gap, cuvettes with a BioRad instrument set
at 600 V, 200 0, 25 pF. The cells were diluted with 160 p1 YPD medium and plated

on selective medium (chapter 3), without sorbitol.

Gas Chromatography

Fatty acid methyl ester standards of ricinoleate (Sigma), Rapeseed Reference
Mixture (Supelco), and an equal-mass mixture of 16:0, 18:0, 18:1, 18:2 and 18:3
(Sigma), were generally injected at 0.125-0.25 mg ml".
Yeast Samples

Yeast cells (from ~1 cm2 patches) were scaped from the plate and transferred
to a glass scew-cap tube. 1.0 M methanolic HCl (1.5 ml) was added, the tube capped
with a teflon-lined cap, and heated to 80°C for 1 h. Upon cooling, 1 ml
hexane:isopropanol (3:2) and 0.5 ml 0.2 M Na2804 were added and the FAMEs
removed in the hexane phase, which could be stored at -20°C. The samples were
analysed with a Hewlet-Packard 5890 series 11 gas chromatograph equipped with a
SPB-l thin film fused silica capillary column (15 m long, 0.53 mm internal diameter,
0.10 pm film thickness; Supelco) with helium as carrier gas. Methyl-ricinoleate
chromatographs well on this column without TMS derivatization. The injected sample

was split between the column and the purge valve. The column was held at 150°C

181

for four minutes and then programmed to 245°C at 4°C min".
Plant Samples

Experiment I: The oldest part (1-2 cm) of carrot roots were used, so as to
allow continued growth of the meristematic part. FAMEs were prepared as above,
dried under nitrogen, redissolved in hexane (2 x 25 p1), and transferred to a gas
chromatograph vial, where trimethylsilylimidazole (TriSilZ, Pierce; 0.5 pl) was
added, to derivatize any hydroxyl groups in a rapid room temperature reaction. The
samples were analysed on a SPB-l fused silica capillary column (30 m long, 0.25 mm
internal diameter, 0.25 pm film thickness; Supelco) using a Hewlet-Packard 5971
series mass selective detector, in place of the flame ionization detector used in other
experiments. The detector scanned masses between 40 and 650. The injected sample
was not split; the oven was held at an initial temperature of 150°C for 7 min, then
programmed to 250°C at 5°C min".

Experiments 2 and 3: Samples were prepared as for experiment 1, except the
final sample was dissolved in 2 x 50 p1 hexane, analysed by gas chromatography once
before derivatization with TMS, and then injected a second time. SP2330 glass
capillary columns (30 m, 0.75 mm ID, 0.20 pm film, Supelco) were used, the
samples were not split, the temperature program was 150°C (6 min) to 215°C (4°C

min"), and flame ionization detectors were used.

1 82
Plant Transformation

Constructs for expression of the pFL2 insert in plants using the vector pBIl217
were prepared by two independent routes. The use of this vector, in which the only
3’ cloning site is SacI, was complicated by the presence of a Sad site in the coding
region of the pFL2 insert. In the first route, pFL2 was linearised with Xbal (which
cuts at the 3’ region flanking the insert), blunt-ended with the Klenow fragment of
DNA polymerase I, then digested with BamHI (which cuts at the 5’ end of the insert),
releasing the insert, which was gel-purified. The vector pBI_121 was digested with
SacI and blunt-ended with T4 DNA polymerase, then cut with BamHI and treated
with calf intestinal phosphatase to prevent religation with the excised B-glucuronidase
fragment. The pFL2 insert was ligated to this pBIl21 vector and used to transform
Escherichia coli DH5a cells to kanamycin resistance. Plasmid DNA of transformants
was digested with XbaI and Sacl, and two clones (A4, B6) were chosen that had the
1.3 kb fragment indicating that the pFL2 cDNA was correctly inserted into the
pBIl21 vector. This was confirmed by the fact that SnaBI did not cut these clones
(SnaBl cuts the B—glucuronidase gene), and EcaRI/HindIII released a band of
appropriate size ( ~ 2.5 kb).

In the second route, clone pFL2 was digested with XbaI and then partially
digested with SacI. A band of ~ 1.45 kb representing the entire insert was isolated
from a gel. The vector 8LJ4K1 (J. Jones, unpublished: see Figure 16) was digested

with WI and SacI, and the vector fragment was gel-purified. The pFL2 insert was

183

     
    
   
   
 

Hindlll 6720

SLJ4K1

6720 bp pUC1 1 8

Clal 4606
Sacl 4538
Xhol 4530

  

358 Promoter

EcoRl 3200

Figure 16. The vector SIJ4K1 constructed by J. Jones, and used for preparation of
constructs for expression of pFL2 in plants.

184

ligated to this vector, transformed into DH5a, and checked for the presence of the
1.3 kb SacI insert fragment. Such a clone was then digested with EcoRI and Hindlll,
and this DNA was ligated to the large EcaRI/HindIII fragment of pBIl2l,
transformed into DHSa and selected for both kanamycin resistance and ampicillin
sensitivity. By this procedure, the entire (358 promoter)-(pFL2 insert)-(nos
terminator) fragment derived from 8L14K1 was used to replace the (358 promoter)-
(B-glucuronidase)-(nos terminator) fragment of pB1121. The clones obtained were
digested with Sacl, and one clone (9/18 3) which gave the appropriate 1.3 kb
fragment was selected.

The three clones (A4 and B6 prepared by the first route, and 9/18 3 prepared
by the second), plus the unmodified vector pB1121, were transformed into
Agrabacterium tumefaciens strains GV3101, R1000 and LBA4404 by electroporation.
Cells for electroporation were prepared as follows. GV3101 and R1000 were grown
in LB medium with reduced NaCl (5 g l"), and LBA4404 was grown in TY medium
(5 g l" bacto-tryptone, 3 g l" yeast extract, pH 7.5). A 500 m1 culture was grown to
OD600 = 0.6, then centrifuged at 4000 rpm (GS-A rotor) for 5 min. The supernatant
was aspirated immediately from the loose pellet, which was gently resuspended in 500
ml ice-cold water. The cells were centrifuged as before, resuspended in 30 ml ice-

_ cold water, transferred to a 30 ml tube and centrifuged at 5000 rpm (SS-34 rotor) for
5 min. This was repeated three times, resuspending the cells consecutively in 30 ml
ice-cold water, 30 ml ice-cold 15 % dimethyl sulfoxide (DMSO), and finally in 4 ml

ice-cold 15 % DMSO. These cells were aliquoted, frozen in liquid nitrogen, and

185

stored at -80°C. Electroporations employed a BTX instrument using cold 1 mm-gap
cuvettes containing 40 pl cells and a minimal volume of DNA, a voltage of 1.44 KV,
and 129 0 resistance. The electroporated cells were diluted with 1 ml SOC medium5
and incubated at 28°C for 1-2 h before plating on medium containing kanamycin (50
mg l".

Plasmid DNA was purified from Agrabacterium cells using ”Magic Minipreps"
(Promega) according to the instructions of the manufacturer (for E. call), with the
exception that lysozyme (2.7 mg ml") was added to the resuspension solution in
which cells were incubated for 10 min at room temperature, and a lysis time of 5 min
was strictly observed. Plasmid yields from cells grown in TY medium were better
than from cells grown in LB medium. Plasmid DNA from Agrabacterium cultures
used to transform plant tissues was digested with restriction enzymes and compared to
the plasmid DNA isolated from E. coli, and was in all cases identical.

Fresh carrots were obtained from the garden of Drs Susan Gibson, Deane
Falcone and Joe Ogas, or from a local store, and transformed using strain R1000
according to Petit et al,8 with the modifications of Yadav et al.9 Peeling of the
carrots (with a sterile scalpel) was essential; surface sterilization treatments with
bleach alone were insufficient to prevent infection of the disks.

Nicatiana tabacum SR-l leaf explants were transformed according to Newman
et al,‘° except that leaves were maintained on No. 3 medium for 3 days prior to
inoculation.

N. tabacum NT-l cells were transformed according to Newman et al‘°

186
RESULTS AND DISCUSSION

Isolation and Sequencing of cDN A Clone pFL2

Comparison of the initial sequence data of pCRS677 and pCR8834 obtained
with the T3 primer, indicated that these are independent clones derived from the same
gene. Clone pCRS677 has a small (14 bp) deletion in the vector/ linker sequence
immediately preceding the cDNA sequence, but is otherwise identical to pCR8834.

The insert of pCRS677 (~7OO bp) was [used as a probe to screen the pYE82.0
library (chapter 3), by colony hybridization at high stringency. In the primary screen
of 47 000 colonies, 84 hybridizing colonies were obtained. The first 28 of these
positive colonies were screened again, to obtain pure positive clones. All 28 of the
primary positives were positive in the secondary screen, indicating an overall
frequency of one positive clone per 560 clones in the library.

DNA prepared from the 28 purified clones was digested with restriction
enzymes and analysed by agarose gel electrophoresis. The enzymes BarnIII and Xhol
cut the vector on either side of the cloning site, and therefore should excise the
inserted DNA when used together. With one exception, all clones had a single
fragment smaller than ~ 800 bp, or an ~ 800 bp fragment plus one or two additional
fragments. Clone 4avi did not fit this pattern. A double-digest with XbaI and HindIII
should, similarly, excise the insert. All clones analysed yielded only one fragment,

ranging in size between ~ 700 bp and ~ 2.2 kb, except clone 4avi, which had an

187

insert of ~ 4 kb. Due to minor technical difficulties, however, clones 2ci, 3cv, 4cii,
4aiii, 4ci, 4aii, 4ai, and 3ciii, were not analysed by digestion with m1 and HindIII.
The majority of clones had one HincII site in the insert, with the exception of clones
3cv, 4cii, 4aiii, 4ci, 3cii, 4aii, 3cvii, and 3cvi, which either lacked this site or had an
additional site. Taken together, these results indicate that most of the 28 clones
purified have a similar restriction pattern, with 9 possible exceptions. This is
compatible with the possibility that most, if not all, represent the same gene. Those
that appeared to have different restriction patterns were not analysed further. Of the
majority of clones, which appeared to have similar restriction patterns but varying
insert sizes, 14 were used to obtain sequence data (below). This data supports the
conclusion that these 14 clones were derived from the same gene. It is concluded that
this one class of clones is present in the pYE82.0 library at a frequency between
1/560 and 1/1120. This frequency is compatible with the possibility that these clones
encode oleate-12-hydroxylase, which is actively expressed in the developing castor
seed.

As already mentioned, 14 of the clones were used to obtain sequence data, by
sequencing from one or both ends of each clone, in addition to the data previously
obtained from clones pCRS677 and pCR8834. These sequence data were used to
assemble a contiguous sequence of ~ 1.7 kb. The longest clone, 3cvii, was 113 bp
longer than the next longest, 3civ-l. However, the first 305 bp of 3cvii showed no
similarity to the overlapping portion of 3civ-1 or several other clones of similar

length, which were, however, all identical in sequence to each other (Figure 17).

188

210 220 230 240 250
3cvii 234 ACTIGGTGAT GAIAGITCCG GTTATAGCAA ATCCGACCAA AAACGGCCAG 283
3civ-1 79 CACACTTGGT GACCTCAAAT CAAACACCAC ACCTTATAAC TTAGTCTIAA 128
4cvii-1 -6 ........................................ CTAAAGTTAA 44
3cx-3 -S6 ............................................ . ..... -7
Sci -19 .................................................. 31
4ev 5 .................................................. S4
2cii 3 .................................................. 52
3cix -40 .................................................. 10
4avii -31 .................................................. 19
2ciii-4 4 .................................................. 53

260 270 280 290 300
3cvii 284 TTACGGTTGA ACTCCCGCTI GAAGAACACG GGCCATGGAT CGAACCACCT 333
3civ-1 129 GAGAGAGAGA GAGAGAGAGG AGACAITTCT CTTCTCTGAG ATAAGCACTI 178
4cvii-1 45 GAGAGAGAGA GAGAGAGAGG AGACATTTCT CTTCTCTGAG ATAAGCACTT 94
3cx-3 -6 ........................................ .ICTAAAGTT 44
3ci 32 .................................... CTCT AAAGGCACTT 81
48V 55 ..CTCTAAAG GAGAGAGAGG AGACATTTCT CTTCTCTGAG ATAAGCACTT 104
2:11 53 .......... .CTCTAAAGG AGACACTTCT CTTCTCTGAG ATAAGCACTT 102
3cix 11 ................................................ CT 60
4avii 20 ...................................... CT CTAAAGACTT 69
2ciii-4 54 ............. CTCTAAA GGACATTTCT CTTCTCTGAG ATAAGCACTT 103

310 320 330 340 350
3cvii 334 TTTCATCTTT TCTCGAAGCC TCAGGAAAGT GTTTAAAAAA GAGCTTTAGA 383
3civ-1 179 CTCTTCCAGA CATCGAAGCC TCAGGAAAGT GCTTAAAAAG AGCTTAAGAA 228
4cvii-1 95 CTCTTCCAGA CATCGAAGCC TCAGGAAAGT GCTTAAAAAG AGCTTAAGAA 144
3cx-3 45 CTCTICCAGA CATCGAAGCC TCAGGAAAGT GCTTAAAAAG AGCTTAAGAA 94
3ci 82 CTCITCCAGA CAICGAAGCC TCAGGAAAGI GCTTAAAAAG AGCTTAAGAA 131
4av 105 CTCTICCAGA CATCGAAGCC ICAGGAAAGT GCITAAAAAG AGCTTAAGAA 154
2cii 103 CTCTTCCAGA CATCGAAGCC TCAGGAAAGT GCTTAAAAAG AGCTTAAGAA 152
3cix 61 CTAAAGCAGA CATCGAAGCC TCAGGAAAGT GCTTAAAAAG AGCTIAAGAA 110
4avii 70 CTCTTCCAGA CATCGAAGCC TCAGGAAAGT GCITAAAAAG AGCTTAAGAA 119
2ciii-4 104 CTCJJCCAGA CATCGAAGCC ICAGGAAAGT GCTTAAAAAG AGCTTAAGAA 153

Figure 17. Alignment of the 3’ sequences of various castor cDNA clones isolated
with pCRS677. Note that the linker sequences (CTCTAAAG) have not been
removed. Similarity between 3cvii and the other clones begins at position 313 of this
alignment.

189
It was concluded that the first 305 bp of the cDNA in clone 3cvii contained

extraneous DNA, not related to pCRS677 (nor any other known sequence). Further
sequence data was obtained only from clone 3civ-1, hereafter designated pFL2.

Sequencing primers were designed from the assembled contiguous sequence,
and used to sequence both strands of the entire clone pFL2 (Figure 18). The clone
encodes a 186 bp 5’ untranslated region (i.e. before the first ATG codon), an 1161 bp
open reading frame, and a 101 bp 3’ untranslated region, including a short (9 bp)
poly(A) tail. The open reading frame encodes a 387 amino acid protein with a
predicted molecular weight of 44406.8. The amino terminus lacks features of a
typical signal peptide.“ The predicted sequence of the Brassica napus fad3
microsomal desaturase also lacks a typical signal peptide.12

Comparison of the pFL2 nucleotide and deduced amino acid sequences with
sequences of membrane-bound desaturases (Table 13) indicates that pFL2 is
homologous to these genes. Arabidapsis fad6 and fad8 sequences used for
comparison in Table 13 were unpublished data of 8. Gibson and V. Arondel.
Sequence similarity between pFL2 and these desaturase genes is considerably weaker
than similarities among the desaturase genes (see for example Figure 10, chapter 4).
The most similar sequence to pFL2 is that of aidesaturase-homologue of unknown
function cloned from Petraselinum crispum (parsley) in the laboratory of Dr. K.
Hahlbrock (I. Sommsich, personal communication). An alignment of the deduced
amino acid sequences of pFL2, Brassica napus fad3 and pFLl (the castor fad 7 cDN A

isolated in chapter 4, above) is shown in Figure 19. The sequence of fad2, the

190

Figure 18. DNA and deduced amino acid sequence of the clone pFL2. Motifs
conserved among membrane-bound desaturases are shaded.

61
121

181

19
241

39
301

59
361
421

481

119
541

139
601

159
661

179
721

199
781

219
841

239
901

259
961

279
1021
1081

319
1141

339
1201

359
1261

379
1321

1381
1441

GCC ACC

ACA CCA

CTC TGA

TTA AGA
GGA
GAC CTC
TAT GTT
CCT TAC
TGC ATT
TAT CAG
TTT TCA
GTG
CCG

GCT

 

TGG

V
GTT

GTA

AAA AAA

TTA AGC
CAC CTT
GAT AAG

H
ATG

G
GGA

G S
GGA AGC
AAG AGA
GCC TAT
ATC TCT
CTC ACT

CTG

GCT

TGG

TTC G

GGT

AGG GAG

F H
TTC TGG

ATT

AA

GAG
ATA

CAC

GGT

AGC

GCC

GCA

Y
TAC

CGC
ACT

TTC

GGT

TTG

GGT

AGA

TAT

CTT

AAG

R
CGG

CGC ACA
TAG TCT
TCT TCC

G R
GGT CGC

L K
CTT AAG

CCA

CTT

CCT

AGG

GCT

GTT

TAT

GGG

GCT

G E
GAG

E C
GAG TGC

N K
AAC AAG

AGA ACA AGA AGG AAT

CGA AGC
TAA GAG
AGA CAT

H S
ATG TCT

ATG
TGT
TCA
GTG TTG
ACA GTG
TAT TAC
TTG TTC

TAT TAA

GTG TGT

191

CTC CTT TCA

CAC TTG

AGA GAG AGA GAG AGA

CGA AGC CTC

T
ACT

p
CCG

F
TTT

L
CTT

A
GCT

GTC

GTA

V

GTC ATA ACC

CAC ACG AAG

E ,
GAA

p
TTC

CAT

TAT

GAG

AGT

GTG

T

CGC

Y
TAC

L
CTG

“first

GTT

TGG

GTA

GTC

TAA

AGG AAA

T S
AGC

K P
CCT

S F
TCT TTT

S I
TCG ATC

V Y
GTT

 

TCT

GGC

CAT

GCT

GAC

TTG

GCT

ATG

ATC

GAT

TGG

TTC

CAT

GCA

ACC

ATG TAG

TGT GTT

GTG ACC
GGA GAC
GTG CTT

I I
AAC AAC

P F
CCT TTC

GTG CGG

 

TGG

TAT

CTC

TGG

GTA

TAC

CGG

ATT

GAG

TTT

G A

GCT

CCT GTT

TCA AAT

ATT TCT

AAA AAG

AGT

GAG

ACA

CTT

TCA

TTC

AAC

TTC

TTC

TTT

GTT

GAG

TCA

TTA

TAC

CCC

TAT

ATC

ATG

TTG

GGA

GCA

GAC

GCC ACT

Y K
TAC AAG

P T
CCT ACA

TCT TTA

CTT

AGC

AAG

GGT

TCC

TTC

GGC

GAG

TAT

GAC

AAC

TTA

>
-s
a

>
n-fl
--0

2° E> E's

AGA

CTA ATA AAG AAG GCA

192

Table 13. Sequence similarity at the nucleotide (NT) and amino acid (AA) levels
between pFL2 and membrane-bound desaturase genes. Nucleotide sequence of the
clone ELI72 was not available for comparison, and the deduced amino acid sequence
used here is considered preliminary data only.

 

 

Sequence Similarity
Gene Organism Ref. NT AA
ELI72 Petraselinum crispum ' text 56
fad7 Ricinus communis chap 4 47.1 38.6
fad3 Brassica napus 12 46.5 37.4
fad8 Arabidapsis thaliana text 46.5 36.2
fad7 Arabidapsis thaliana 13 47.4 35.5
fad6 Arabidapsis thaliana text 46.5 23.5

desA Synechoccocus sp. 14 45.5 22.6

 

pFL2
pFL1

pFL2
pFL1

pFL2
Fod3
pFL1

pFL2
Fad3

pFL1

pFL2
Fad3
pFL1

pFL2
pFL1

pFL2
Fad3
pFL1

pFL2
Fad3
pFL1

pFL2
Fad3

pFL1

pFL2
Fad3
pFL1

Figure 19. Alignment of the deduced amino acid sequence of pFL2 with pFLl (Fad7

10

20

1593

30 40

50

-75 .......................................... ........
-79 ..................

1

~25
-29
51

26
22
101

76
151

126
122
201

176
172
251

226
222
301

276
272
351

326
322
401

376
372
451

HAAGHVLSEC

GLRPLPRIYS

RPRIGFTSKT TILLKLRELP DSKSYNLCSS

80 90
..... HGGGG RHSTVITSIN

100
SEKKGGSSHL

............................. I VVAHDOR**V NGDS*ARKEE

FKVSSHSNSK

110
KRAPHTKPPF
GFD*SAO***
FFDAGAP***

160
NFFPYISSPL
vv* ------ o
---A*FNN--

210
VGLIVHSALL

.*H*L**FI§
**“LL**S!*

260
SNNPPGRVLT
P--HST*H*R
D*VTKTLRFS

310
ERLGIYIADL
**KL*ATSTT
**KD*ITSTA

360
lTYL-OHTHP
v***uu*c*o
V***HH*G*E

410
AHHLFATVPH

l**fl*P°l**
IQQOOPOIii

460
PDEGAPTOGV
K*HYVSDT*D
E*HYVSDT*D

GSNHALNVAV

120
TLGDLKRAIP

K1.*l“***
**‘*IRAQ.*

170
SYVAULVYUL
*HFL‘PL**V
-Htttthtf

220
VPYFSHKYSH

***HG*R3**
iiiuciaiti

270
LAATL-LLGU
YTVP*PN*AY
*PFP--H*AY

320
"GIFATTFV
CHS*HLA*L*
-'CHT*HAAL

370
A-IPRYGSSE
EKL'H'RGK*
DKL*H*RGKA

420
YHAHEATKAI
**LVD**R*A
*‘LV***E*A

470

PVNVSTVS‘E DDREREEF*G

130 140
PHCFERSFVR SFSYVAYDVC
Kttuvxtplt ontoatka1-
K**UVKNPU* *H***LR**V

180 190
FGGClLTGLH VIGHECGHHA
A**TLFHAlF *L**D***GS
C**THFHA*F *L**D***GS

230
RRHHSNIGSL
*T**0*H*HV
*T**o*u*uv

280
PLYLAFNVSG
*l**UYRSP*
*F**----HS

330
LYGATHAKGL

240
eaoevrvpxs
*l**SH**LP
*"**S""*L*

29o
vaoasAcnv
----KEGS*F
*spcxxcs*r

'340
AHVHRIYGVP

-*LSFLVDPV T-*LKV****
*VYLNFSM'P VOHLKL**I*

380 390
UDULRGAHVT VDRDYGVLNK
*SY***GLT* iteratiftu
*SY***GLT* L*****UI*N

430 440
KPlMGEYYRY DGTPFYKALU
1'IIVL"R"'**E PK*SGAlPlH
**V**K***E PKKSGPLP'H

480 490

IVNVDEGKGE

150
LSFLFYSIAT
FAVAALAH‘A
VV*GLAAV‘-

200
FSEYOLADDI
**OIP*LNSV
**NNPKLNSV

250
KSKISUYSKY
-E*L--*KNL
-E**--FKSL

300
DPYGPTFSER
N**SSL*APS
H*DSGL*VPK

350
LLlVNCFLVH
Y1*FVHU*0A
YU‘FVHH*DF

400
vruutAotuv
1u*o*-c***
tu*o*-c***

450
REAKECLFVE
LVESLVASTK
LLGSLVRSHK

500

FHYRNKY .................................

IVFYETDPOL
VV*YG*DPKL

YWASDKSKI N .........
SGlGGEKTE.

from castor: chapter 4), and Brassica napus Fad3.12

-26
-30
50

21
100

71
150

125
121
200

175
171
250

225
221
300

275
271
350

325
321
400

375
371
450

425
421
500

194
desaturase gene hypothesized to be most similar to the oleate— 12-hydroxylase gene,

was not available for comparison.

The deduced amino acid sequence of pFL2 (Figures 18, 16) contains the
conserved histidine-rich repeats (HXXHH) also found in other desaturases (chapter 4).
These motifs are separated by 169 intervening amino acid residues. The motifs are
separated by only 122 residues in the deduced amino acid sequence of the Arabidapsis
thaliana fad2 gene (John Browse, personal communication to Sue Gibson). This
information suggests that pFL2 does not encode the microsomal A12 desaturase.

A divergent form (GHECGH) of the conserved desaturase motif (GHDCGH) used for
priming PCR reactions (chapter 4), was found in the pFL2 sequence. Since the
codons for glutamic acid (E) differ from the codons for aspartic acid (D) only at the
third position, priming in the PCR reaction would have resulted from only a single-
base mis-match. Despite attempts to select cycling conditions (viz. annealing
temperature) at which some mis—match could occur, the clones analysed in chapter 4
almost all contained the authentic primer sequence, indicating that mis-match did not

occur, and hence the pFL2 sequence was not amplified.

Expression of pFL2 in Castor

Since oleate-12-hydroxylase activity is only found in the developing seeds of

castor, '4 it was of great interest to determine whether pFL2 is also expressed only in

seeds, or is also expressed in other tissues. This question was addressed by testing

195
for hybridization of pFL2 to RNA purified from developing seeds and from leaves.

A northern blot of RNA from leaves and developing seeds (stage III to stage
V”) of castor was probed with the 32P-labelled insert of clone pCRS677, which
corresponds to ~ 700 bp of the 3’ end of pFL2. Brief (30 min) exposure of the blot
to X-ray film revealed that the probe hybridised to a single band of ~ 1.67 kb, only
in the seed RNA lane (Figure 20, panel A). Upon overexposure (16 h) of the film, a
band of similar size was detected in the leaf RNA lane, in addition to a second,
larger, band in the seed RNA lane (Figure 20, panel B). The blot was also exposed
to a phosphor-imagin g screen, for quantitation of probe hybridisation. Total exposure
to this screen in an area covering the band in the leaf lane was 4.36 x 10‘ units above
background. Total exposure in an area of equal size over the major band in the seed
lane was 1.17 x 107 units above background, 268-fold more than in the leaf lane.

The blot was re-probed with a B-tubulin gene, which gave bands of equal intensity in
the md and leaf lanes (Figure 20, panel C), verifying that equal quantities of
undegraded RNA were loaded in the two lanes.

These results show that pFL2 is highly and specifically expressed in seed of
castor. Over-exposure of the northern revealed a 268-fold weaker band of similar
size in leaf RNA, but also a second band in seed RNA, suggesting that these bands
are due to weak hybridization of pFL2 to related sequences, such as desaturases.
Strong md-specific expression of pFL2 is compatible with the possibility that it
encodes oleate- 12-hydroxylase. N o desaturase or other putative desaturase-homologue

which would have these characteristics is known to be expressed in castor.

196

 

 

 

—1.77
—1.52

—1.28

 

— 0.78

—O.53

—O.40
—O.28

 

—0.16

 

 

 

 

 

 

 

Cl!- 91

Figure 20. Northern blot analysis of pFL2 expression in castor. A 32P-labelled probe
corresponding to ~ 700 bp of the 3’ end of clone pFL2 was hybridised to poly(A)+
RNA from leaves (L) and developing seeds (8) of castor. Panel A: the blot was
exposed to film for 30 min. The migration of RNA standards (kb) is shown to the
right. Panel B: the same blot was exposed for 16 h. Panel C: the same blot was
hybridised to a 32P-labelled probe made from the Colletotrichum graminicola B-tubulin
gene TUBZ.6

 

 

197
Southern Analysis with pFLZ

Southern analysis was used to examine the copy number of genes in the castor
genome corresponding to clone pFL2, and to examine whether related sequences
could be detected in the castor genome, and in the genome of a different plant, in
which oleate-lZ-hydroxylase is absent.

Genomic DNAs of castor and of Arabidapsis thaliana were digested with
restriction enzymes, separated by agarose gel electrophoresis and transferred to a
nylon membrane. Arabidapsis was chosen for the negative control DNA because it
has no known oleate-12-hydroxylase, and the DNA was readily available. The
membrane was hybridised with the 32P-labelled insert of clone pFL2 at 65 °C, and
exposed to X-ray film. The probe hybridised with a single band in each digest of
castor DNA, but did not hybridise to the Arabidapsis DNA (Figure 21), indicating
that the gene from which pFL2 was transcribed is present in a single copy in the
castor genome, and is not present in the Arabidapsis genome. The blot was then
hybridised again, with an identical probe, but at less stringent hybridization conditions
(52°C) (Figure 22). This revealed additional weakly-hybridising bands. In castor
DNA, a total of four bands were detected in both the EcoRI digest and the BamHI
digest. In Arabidapsis DNA, four bands (EcoRI), five bands (BamHI), or possibly
three bands (Hindlll) were detected. These results suggest that two or more genes
with sequence similarity to pFL2 occur in both the castor and Arabidapsis genomes,

in agreement with sequence similarity between pFL2 and desaturase genes (above).

198

Arabidapsis Castor

EBH EBH

 

23
9.4 —
6.6 —

4.4 —

2.3 —
2.0 —

 

0.56 —

 

 

 

 

Figure 21. A Southern blot of genomic DNA from Arabidapsis thaliana and castor
(Ricinus communis) digested with restriction enzymes EcoRI (E), BamHI (B), or
Hindlll (H), was hybridised at high stringency (65°C) with the 32P-labelled insert of
clone pFL2. Migration of DNA standards (kb) is shown to the left.

Arabidapsis Castor

EBH EBH

 

23

9.4: '1'

6.6 —
4.4 —

2.3 —
2.0 —

 

0.56 —

 

 

 

 

Figure 22. The Southern blot shown in Figure 21 was hybridised again with the
pFL2 probe under low stringency (52°C). (See legend to Figure 21 for details).

200
Expression of pFL2 in Transgenic Yeast

The experiments described above indicate that pFL2 has features expected of
an oleate-lZ-hydroxylase gene, but do not provide convincing evidence that pFL2 is
in fact the oleate-12-hydroxylase gene. The function of the protein encoded by pFL2
was, therefore, examined by expression in transgenic cells. Since the isolation of
clones using pCRS677 as a probe, described above, made use of the pYE82.0 library,
designed for expression of the cloned cDNA in yeast (chapter 3), it was a simple
experiment to express pFL2 in transgenic yeast.

In addition to pFL2, two other clones, 2cii and 3ci, were transformed into
yeast cells by electroporation. Sequence data and digestion with the enzyme KpnI
(which cuts near the 3’ end of pFL2 and also in the cloning site of the vector directly
downstream of the Gal 1 promoter (Figure 5), and is therefore diagnostic for the
orientation of the insert DNA) indicated that 2cii and 3ci contained the full coding
sequence, but were in the reverse orientation for expression in pYE82.0, while pFL2
was in the forward orientation. Clones 2cii and 3ci could therefore be used as
negative controls. Two yeast strains were used in the experiment, CGY2557 (chapter
3) and a protease-deficient strain, J 0522. Transformants were patched onto selective
medium (lacking uracil) with either glucose or galactose as carbon source. Under the
control of the GAL promoter, the castor DNA would only be transcribed in cells
grown on the galactose-containing medium. Fatty acid methyl esters (FAMEs)

prepared from the harvested yeast cells were analysed by gas chromatography.

201

FAMEs from each of the three clones grown on both carbon sources were
analysed for the host strain CGY2557, and FAMEs from duplicate patches for each of
these combinations were analysed for the host strain J 0522. Growth of yeast
harbouring pFL2 on medium containing galactose did not result in the appearance of
any material chromatographing at the position of authentic methyl-ricinoleate.
Furthermore, no novel peaks were detected in any of the chromatograms. A typical
chromatogram is shown in Figure 23, with the authentic methyl-ricinoleate standard
for comparison. Expression of the castor clones also caused no changes in the
relative proportions of yeast FAMEs, though it should be noted that the unsaturated
18—carbon fatty acids were poorly resolved on the column used in this experiment.

The results of this experiment neither support nor contradict the possibility that
pFL2 encodes oleate-12-hydroxylase. If pFL2 does encode oleate-l2-hydroxylase, the
lack of expression observed in this experiment indicates a possible reason why pFL2

was not isolated in the experiments reported in chapter 3.

Expression of pFL2 in Transgenic Plants

The insert of clone pFL2 was ligated between the 358 promoter and nos
terminator of the plant expression binary vector pB1121,7 in the correct orientation for
expression of the open reading frame, by two independent cloning strategies. These
constructs were transformed into Agrobacterium tumefaciens strains R1000, GV3101,

and LBA4404. R1000 is a C58—derivative cured of the Ti plasmid and harbouring the

202

7000i
6000{
5000€
4000€

mitt

2000-

 

 

 

 

 

0 5 10 15 20
70004

6000;
50004

4000;

 

8000— L L

2000-. .. I I
. . ... . ...,. ...,.
0 5 10 15 20

 

 

 

Figure 23. Typical gas chromatogram (above) of yeast fatty acid methyl esters
(CGY2557 harbourrng pFL2, grown on glucose medium), and chromatogram of

methyl-ricinoleate standard (below). Detector signal is plotted against retention time
(min).

203

A. rhizogenes plasmid pRiA4b,‘° and was used to produce transgenic roots on carrot
disks. GV3lOll7 and LBA4404“ contain disarmed Ti plasmids, and were used to
transform both tobacco leaf explants and tobacco suspension culture cells. Transgenic
tissues or callus were analysed by gas chromatography for fatty acid composition in

three experiments.

Experiment 1. Roots produced on the inoculated carrot disks were excised and tested
for growth in medium containing kanamycin (50 mg 1"). Transformation of carrot
roots using the strain R1000 involves cotransformation of the binary vector and the Ri
plasmid. Seven of the eight roots analysed in this experiment were kanamycin
resistant, indicating they were cotransformed. These represented construct B6 (roots
8-1, 11-1, 22-1), construct 9/18 3 (roots 20—1, 20-2, 20-3), and the intact vector
pBIl2l (root 23-1).

Root FAMEs were treated with silylating reagent and analysed by gas
chromatography using a mass-selective detector (electron impact-mass spectrometry).
No peaks were observed in the chromatograms at the elution time of authentic
trimethylsilyloxy-methyl-ricinoleate. A typical chromatogram is shown in Figure 24,
and a chromatogram of the standard is shown in Figure 25. Examination of the ions
present in the background signal in this part of the chromatogram failed to detect the
major ion (m/z 187) produced from the authentic standard. Mass spectra taken at
similar retention times are shown in Figure 26 (corresponding to the chromatogram

shown in Figure 24) and Figure 27 (corresponding to the standard, Figure 25).

204

Figure 24. Typical gas chromatogram (experiment 1) of transformed carrot root
TMS-derivatised fatty acid methyl esters (root 20-1). Detector signal is plotted
against retention time (min). Peaks identified by co-chromatography with standards,
and by comparison of mass spectra to a mass-spectral data library: 17.38 min, 16:0;
20.88 min, 18:2; 21.04 min and 21.19 min, 18:1; 21.69 min, 18:0.

 

 

 

205

 

TIC: 1101001.D

 

 

 

 

 

 

 

 

17.38 21069
28.12
24.53
200813
20.30
‘ .‘9
a]
21 80
3 ‘04 25.18
2'074
26.97!
30.82
2c.'4 23-35
19. 3; I 24.63
19 SI .
II I I ‘ Iii .46
" ‘1n2 .7! 2’-fl

 

 

 

 

 

 

 

 

8'“ ’3 19 I T 3.1.. 27
“I185: I IIIIIIIII IIIIIfiII III III II II III III 3 |I72I

 

10. 00 15.00 20. 00 25.00 30. 00

206

Figure 25. Gas chromatogram of trimethylsilyloxy-methyl-ricinoleate (experiment 1).
Detector signal is plotted against retention time (min). Retention time of standard:
25.1 min.

 

 

207

 

TIC: 0301003.D

 

 

A———J‘-‘~ A~~ lJW

 

 

I l I I I I I T j I Y I I T T l l T I I I I j I
10.00 15.00 20.00 25.00 30.00

 

208

Figure 26. Mass-spectrum of ions derived from material eluting from the column at
25 .165 min, in the chromatogram shown in Figure 24.

209

 

 

”/2

Abundance

50000—

45000€
40000€
35000?
30000€
25000€
200003
15000€

1
10000-

5000‘

*

 

o-
->

43

 

 

40

 

 

60

Scan 1320 (25.165 min):

85

99

113

 

 

 

I

80 100 120 140 160 180 200

141
155

MIII t JWIML.M

183

197

1101001.D

239

225

 

2672
l

220 240 260 2:

210

Figure 27. Mass-spectrum of trimethylsilyloxy-methyl-ricinoleate, taken from the
chromatogram shown in Figure 25 at 25.120 min.

211

 

 

”/2

fibundance

9000005
850000{

800000{

750000

ALLIJJL

7000003

4

650000;
600000;
550000;
500000;
450000;
400000;

1
1

350000

LJLLI

3000001

4
q

zsoooof

1
2000001

.1
.1

1500005
100000{

soooof
t

 

1
o-
->

55

50

73

 

103

100

Scan 1330 (25.120 min): 0301003.D

 

 

 

1 7
270
299
159
[2A.1L 245‘ 337‘ 383
Tji'lIT' I'III I—I ITIUIUIIIIUIIIIIIII
150 200 250 300

350 400 450 5(

 

212

The mass spectrum of TMS-methyl-ricinoleate shown in Figure 27 is the same as that
previously established for this compound (Appendix A, which includes an explanation
of the fragmentation pattern resulting in the ions observed). It was concluded that the
samples contained no ricinoleic acid. The relative content of the abundant fatty acids
in each sample is presented in Table 14. Octadecadienoic and octadecatrienoic acids
were not resolved on the column used in this experiment. Though there was some
variability in the combined content of these particular fatty acids, there appeared to be

no correlation with expression of pFL2.

Experiment 2. Transformed carrot roots which were growing in medium containing
50 mg 1'1 kanamycin were analysed for fatty acid composition. Eighteen
independently-transformed roots were analysed, including the seven roots previously
analysed in experiment 1. These represented the four constructs used for
transformation: 9/ 18 3 (5 roots), B6 (7), A4 (1) and pB112] (5). Also analysed was
shoot tissue from nine transformed tobacco shoots derived from the tobacco leaf
explant transformation. Eight of these nine shoots subsequently rooted in medium
containing 100 mg 1'1 kanamycin; it is not certain that the ninth shoot is truly
transformed.

FAMEs from these samples were first analysed by gas chromatography
without TMS derivatization, since the TMS reagent elutes as a broad peak which
comigrates approximately with stearic acid. After the relative composition of usual

fatty acids had been examined, samples were derivatised with TMS and re-injected.

213

Table 14. Relative content (96) of the abundant fatty acids in transgenic earrot roots.
Fatty acid methyl ester abbreviations: 16:0 hexadecenoic acid, 18:2 octadecadienoic
acid, 18:3 octadecatrienoic acid, 18:1 octadecenoic acid (two isomers were resolved
but not identified; carrot contains petroselenic acid, A6-18:1, in addition to oleic acid,
A9-18:1), 18:0 stearic acid.

Fatty Acid Methyl Ester

 

16:0 18:2 18:1 18:1 18:0
+18:3

Retention Time (min) 17.28 ' 20.81 20.97 21.11 21.60

 

constr. kanR root 1

B6 + A8-1 28.2 7.0 9.9 23.9 26.7

B6 + All-1 30.1 3.8 10.0 25.0 27.5

B6 + 221 30.5 22.1 7.9 17.6 19.6
9/18 3 + 201 28.7 20.7 6.9 16.5 24.4
9/18 3 + 202 N 29.3 15.5 7.9 18.5 25.7
9/18 3 + 203 ' 30.1 7.4 8.8 21.0 29.4
pB1121 + 234 : 29.4 9.5 9.1 22.5 26.3
p131121 - A3—1 27.9 12.0 9.2 22.7 24.6

 

 

214

The relative content of the abundant fatty acids in each sample is presented in Table
15 . Considerable variation was again observed in the relative contents of
octadecadienoic and octadectrienoic acids, but there was no clear correlation with
expression of pFL2. It appears that samples may differ considerably from each other
simply due to differences in the age of the tissue, its physiological state, or particular
environment. When the samples were re-inj ected following TMS derivatization, no
peaks were found in the position of authentic trimethylsilyloxy-methyl-ricinoleate in

any of the chromatograms.

Experiment 3. A gene involved in the biosynthesis of one other unusual fatty acid,
petroselinic acid, has been cloned. ‘9 This gene encodes an acyl—ACP desaturase, and
resulted in the production of petroselinic acid and hexadec-4-enoic acid when
expressed in transgenic tobacco callus. These unusual fatty acids, normally absent
from tobacco, comprised between 1% and 4% of total fatty acids in the transgenic
samples. ‘9 However, differentiated plant tissues regenerated from these transgenic
calli, contained little or no detectable levels of these unusual fatty acids (E. Cahoon,
personal communication).

In case the results of experiments 1 and 2, above, were due to a similar
phenomenon whereby no detectable ricinoleic acid accumulated in the transgenic
carrot root or tobacco shoot tissue, this experiment focused upon analysis of

transgenic tobacco calli. Calli from both the tobacco leaf explant transformation

215

Table 15. Relative content (%) of the abundant fatty acids in transgenic carrot roots
and tobacco shoots. Fatty acid methyl ester abbreviations: 16:0 hexadecenoic acid,
18:0 stearic acid, 18:1 octadecenoic acid, 18:2 octadecadienoic acid, 18:3
octadecatrienoic acid.

Fatty Acid Methyl Ester

 

16:0 18:0 18:1 18:2 18:3
Retention Time (min) 7.16 10.12 10.83 12.06 13.55

 

 

 

Tobacco
constr. kan" shoot
B6 + 103 26.9 9.8 6.4 14.1 38.8
B6 + 10-1 31.1 7.4 8.9 13.3 36.3
B6 + 6—1 31.8 5.4 1.2 22.0 38.0
B6 + 10—2 28.9 6.5 1.9 16.6 44.4
A4 + 9-3 32.9 4.9 3.0 24.8 32.8
A4 + 9-1 28.9 7.4 4.3 18.7 37.1
A4 - 9-2 28.8 8.7 4.7 10.1 44.9
9/18 3 + 8 29.9 8.5 4.7 9.9 43.0
pB1121 + 4/12-1 29.2 7.5 4.7 11.7 43.6
Carrot
constr. lcanR root
B6 + 18-5 34.2 9.2 15.5 22.4 14.4
B6 + 184 33.4 9.8 18.5 17.3 15.2
B6 + All-6 33.2 ~ 18.0 22.5 1.5 17.0
B6 + All-4 31.9 19.0 18.4 3.5 17.8
B6 + 22-1 35.8 18.0 18.7 9.9 14.6
B6 + All—1 33.4 18.2 21.4 5.4 18.0
B6 + A8-1 36.4 11.8 11.5 24.6 12.0
A4 + A4-4 29.8 15.5 19.3 7.6 21.4
9/18 3 + 201 30.9 11.7 16.2 18.3 16.0
9/18 3 + 20-5 32.8 19.1 19.0 5.0 18.1
9/18 3 + 20-10 34.3 11.3 14.6 21.4 13.7
9/ 18 3 + 20-3 34.6 11.7 14.0 20.9 14.7
9/18 3 + 20—2 34.4 8.7 10.4 30.6 13.1
pB1121 + 19-4 35.2 11.4 18.8 16.9 13.0
pB1121 + 23-6 33.7 17.5 22.6 1.6 19.2
pB1121 + 11-1 30.3 14.8 22.3 2.5 20.5
pB1121 + 23-5 33.4 16.0 21.5 3.7 17.9
pB1121 + 23-1 34.9 18.0 17.4 10.7 15.1

216

system and the tobacco suspension culture (NT-1) cell transformation system were
analysed in this experiment. Some additional transgenic tobacco shoot samples were
also analysed. In the tobacco leaf explant transformation system, callus may form at
the margins of the explants, in addition to callus formed at the points of inoculation
(by stabbing the leaf with a needle bearing Agrabacterium cells). Only callus clearly
arising from these inoculation sites was analysed in this experiment, and only shoots
derived from such calli were considered transgenic.

FAMEs were prepared from 7 calli and.7 shoots (2 of which included some
callus tissue in the sample analysed, and all of which had rooted-—or subsequently
rooted--in kanamycin-oontaining medium) from the leaf explant system, and 4 calli
from the NT-l system. In addition, samples of each type were analysed that had been
transformed with the intact pB1121 vector, as negative controls.

The relative content of the abundant fatty acids in each sample is presented in
Table 16. As in the previous experiments, considerable variation was seen in the
relative proportions particularly of octadecadienoic and octadecatrienoic acids, but this
did not correlate with expression of pFL2. When the samples were re-injected
following TMS derivatization, no peaks were found in the position of authentic

trimethylsilyloxy-methyl-ricinoleate in any of the chromatograms.

217

Table 16. Relative content (%) of the abundant fatty acids in transgenic tobacco

shoots (S), calli generated on leaf explants (C), and calli obtained by NT-l cell
transformation (NT). Fatty acid methyl ester abbreviations as for Table 15.

 

 

Fatty Acid Methyl Ester

16:0 18:0 18:1 18:2 18:3

constr. tissue identity
B6 S 18-1 26.1 8.8 5.9 9.1 35.5
B6 S 10-3 25.7 9.2 3.7 8.1 39.0
B6 S 6-1 27.3 7.3 3.7 8.9 40.3
B6 S(c) 6-1 34.2 4.5 1.8 23.4 26.7
A4 S(c) 9-3 27.6 4.3 2.1 13.9 32.8
A4 S 9-3 25.2 9.4 7.0 6.7 33.4
A4 S 9-1 24.9 9.2 3.7 10.9 36.1
pB1121 S(c) 4/12-1 33.8 4.0 2.0 27.3 23.1
pB1121 S 4/12-1 25.3 7.9 5.6 7.8 36.8
pB1121 S 4/12-1 23.9 . 10.0 6.0 8.6 33.1
B6 C 10-4 31.2 5.1 2.9 34.9 18.3
B6 C 10-4 31.9 5.1 2.2 30.0 23.3
B6 C 105 30.1 7.1 6.0 28.6 19.5
A4 C 1-1 33.6 4.2 3.1 35.7 16.2
A4 C 9-4 30.6 5.0 3.6 36.5 15.0
A4 C 17-1 28.2 11.1 10.6 28.1 13.0
9/18 3 C 19-1 27.1 7.1 8.4 31.1 12.8
pB1121 C 4/12-2 33.0 4.8 3.3 34.0 18.9
B6 NT 1 29.0 9.7 16.7 31.4 8.5
B6 NT 2 22.4 8.1 8.6 15.0 10.1
A4 NT 3 25.8 8.6 17.7 25.4 9.4
A4 NT 4 23.7 7.9 11.1 18.7 10.9
pB1121 NT 5 28.8 7.8 15.3 37.1 7.5
pB1121 NT 6 36.8 5.8 8.2 36.1 7.7
pB1121 NT 7 34.2 6.4 7.6 36.5 6.2
pB1121 NT 8 34.4 6.1 8.5 32.8 6.2

 

 

2 1 8
CONCLUSIONS

As with expression of pFL2 in yeast, attempts to express pFL2 in plants have
not yet yielded information supporting or contradicting the possibility that pFL2
encodes oleate-12-hydroxylase. So far, these transgenic plants have only been
examined for the accumulation of ricinoleate, or other changes in the equilibrium fatty
acid composition. A next experiment would be to determine whether pFL2 is in fact
being expressed at the RNA level, by northern blot analysis. It would also be a high
priority to assay the transgenic tissues for oleate-12-hydroxylase activity in vitro. The
sensitivity of the radioisot0pic in vitro assay may detect hydroxylase activity not
detected by gas chromatography. Furthermore, expression of an oleate-12-
hydroxylase in transgenic plant tissue may result in targetting of the ricinoleate
produced to a degradation pathway. Plant microsomes contain a phospholipase with a
preference for oxygenated acyl groups such as ricinoleic acid.20 This may be part of
a pathway fortumover of oxygenated fatty acids (such as those associated with
membrane damage), and it is possible that the in vitro assay would separate

hydroxylation from degradation, allowing ricinoleic acid to be detected.

REFERENCES

l. Coppens, N ., Biosynthesis of fatty acids by mds of Ricinus communis,
Nature, 177, 279, 1956.

2. Canvin, D.T., Formation of oil in the md of Ricinus communis L., Can. J.
Biochem. Physiol., 41, 1879-1885, 1963.

10.

11.

12.

13.

219

Yamada, M., and Stumpf, P.K., Enzymic synthesis of ricinoleic acid by
extracts of developing Ricinus communis L. seeds, Biochem. Biophys. Res.
Commun., 14, 165-171, 1964.

James, A.T. Hadaway, H.C., and Webb, J.P.W., The biosynthesis of
ricinoleic acid, Biochem. J., 95 , 448-452, 1965 .

Sambrook, J., Fritsch, E.F., and Maniatis, T., Molecular Cloning: a
Laboratory Manual, 20d ed., Cold Spring Harbor Laboratory Press, 1989.

Panaccione, D.M. , and Hanau, R.M. , Characterization of two divergent B-
tubulin genes from Colletotrichum graminicola, Gene, 86, 163-170, 1990.

Jefferson, R.A., Kavanagh, T.A., and Bevan, M.W., GUS fusions: B-
glucuronidase as a sensitive and versatile gene fusion marker in higher plants,
EMBO J., 6, 3901-3907, 1987.

Petit, A. , Berkaloff, A. , and Tempe, J. , Multiple transformation of plant cells
by Agrobacterium may be responsible for the complex organisation of T-DNA
in crown-gall and hairy root, Mol. Gen. Genet., 202, 388-393, 1986.

Yadav, N.S., Wierzbicki, A., Aegerter, M., Caster, C.S., Perez-Gran, L.,
Kinney, A.J., Hitz, W.D., Booth, R., Schweiger, B., Stecca, K.L., Allen,
S.M., Blackwell, M., Reiter, R.S., Carlson, T.J., Russell, S.H., Feldmann,
K.A., Pierce, J., and Browse, J., Cloning of higher plant 013 fatty acid
desaturases, Plant Physiol., 103, 467-476, 1993.

Newman, T.C., Ohme-Takagi, M., Taylor, C.B., and Green, P.J., DST
sequences, highly conserved among plant SA UR genes, target reporter
transcripts for rapid decay in tobacco, Plant Cell, 5, 701-714, 1993.

Alberts, B., Bray, D., Lewis, J ., Raff, M., Roberts, K., and Watson, J.D.,
Molecular Biology of the Cell, 2nd ed.,.Garland, New York, 1989, pp 414-
415.

Arondel, V., Lemieux, B., Hwang, 1., Gibson, 8., Goodman, H.M., and
Somerville, C.R. , Mapcbased cloning of a gene controlling omega-3 fatty acid
desaturation in Arabidopsis, Science, 258, 1353-1355, 1992.

Iba, K., Gibson, 8., Nishiuchi, T., Fuse, T., Nishimura, M., Arondel, V.,
Hugly, S., and Somerville, C., A gene encoding a chloroplast omega-3 fatty
acid desaturase complements alterations in fatty acid desaturation and
chloroplast copy number of the fad 7 mutant of Arabidopsis thaliana, J. Biol.
Chem., in press.

14.

15.

16.

17.

18.

19.

20.

220

Wada, H., Gombos, Z., and Murata, N., Enhancement of chilling tolerance of
a cyanobacterium by genetic manipulation of fatty acid desaturation, Nature,
347, 200-203, 1990.

Greenwood, J.S. , and Bewley, J.D. , Swd development in Ricinus communis
(castor bean). 1. Descriptive morphology, Can. J. Bot., 60, 1751-1760, 1982.

Moore, L., Warren, G., and Strobel, G. , Involvement of a plasmid in the
hairy root disease of plants caused by Agrobacterium rhizogenes, Plasmid, 2,
617-626, 1979.

Koncz, C., and Schell, J., The promoter of TL-DNA gene 5 controls the
tissue-specific expression of chimaeric genes carried by a novel type of
Agrobacterium binary vector, Mol. Gen. Genet., 204, 383-396, 1986.

Ooms, G., Hooykaas, P.J.J., van Veen, R.J.M., van Beelen, P., Regensburg-
Tuink, T.J.G. , and Schilperoort, R.A. , Octopine Ti-plasrrrid deletion mutants
of Agrobacterium tumefaciens with emphasis on the right side of the T-region,
Plasmid, 7, 15-29, 1982.

Cahoon, E.B., Shanklin, J., and Ohlrogge, J.B., Expression of coriander
desaturase results in petroselinic acid production in transgenic tobacco, Proc.
Natl. Acad. Sci. USA, 89, 11184-11188, 1990.

Banas, A., Johansson, 1., and Stymne, 8., Plant microsomal phospholipases
exhibit preference for phosphatidylcholine with oxygenated acyl groups, Plant
Sci., 84, 137-144, 1992.

CHAPTER 7

CONCLUSIONS AND PERSPECTIVES

Different species of flowering plants synthesize varying amounts of
triacylglycerol in their developing seeds. The majority of these species, including
major oilseed crops, synthesize triacylglycerol in which the component fatty acids are
similar to those found in other plant tissues, such as oleic, linoleic and linolenic acids.
However, a number of plant species have been identified, as reviewed in chapter 1,
which accumulate an unusual fatty acid in the md triacylglycerols. Little is known
about any function for such unusual fatty acids, and it is not a structural one, since
these unusual fatty acids are generally excluded from membrane lipids. As in other
oilseeds, one of the major metabolic activities of the developing endosperm of castor
(Ricinus communis) seeds is the synthesis from imported sucrose of triacylglycerol.
Approximately 90% of the component fatty acids of this storage lipid are the
hydroxylated fatty acid, ricinoleic acid. This fatty acid is derived from oleic acid by
hydroxylation at carbon 12. Ricinoleic acid is not produced by plants more
commonly used in biological research, such as the model organism Arabidopsis
thaliana, hence its biosynthesis must be studied in a plant such as castor. All
previous studies of ricinoleic acid biosynthesis in plants have employed castor. Not
surprisingly (given the amount of ricinoleic acid produced), the hydroxylase
responsible for the transformation of oleic acid to ricinoleic acid is readily assayed in

castor endosperm extracts. However, the nature of this hydroxylase was poorly

221

222

understood at the biochemical or molecular level. Characterization of the
hydroxylase, and particularly, isolation of a gene (or genes) encoding the hydroxylase
enzyme were the goal of this work.

Assays of oleate-12-hydroxy1ase were used in chapter 2 of this study to gain
some new insight into the enzyme. The pattern of induction of the enzyme during
seed ontogeny was investigated, and it was found that hydroxylase activity is
detectable from the time of differentiation of the cellular endosperm tissue in which
lipid is stored. It was conclusively shown for the first time that oxygen is a necessary
cofactor for hydroxylation, and an approach employing stable isotopes and mass
spectrometry was used to demonstrate that molecular oxygen is the immediate source
of the hydroxyl oxygen atom at C 12 of ricinoleic acid (Appendix A). Furthermore, it
could be shown that antibodies against cytochrome b, completely inhibit oleate
hydroxylation by castor endosperm extracts, indicating that cytochrome b, is the
electron donor to the hydroxylase. The involvement of molecular oxygen and
cytochrome b5, and other properties of the hydroxylase reported in the literature, such
as the identification of oleate esterified at sn-2 of phosphatidylcholine as the
immediate substrate,‘ are similar to the biochemical properties of the microsomal fatty
acid desaturases.

Despite being readily assayed in crude endosperm extracts, oleate-12-
hydroxylase was found to be labile during various attempts at fractionation of these
extracts. This lability thwarted attempts at partial purification of the enzyme, and

prompted the exploration of a number of alternative approaches to isolation of an

223

oleate-12-hydroxylase gene. These included attempts to obtain polyclonal antibodies
against castor proteins which would inhibit hydroxylase activity (chapter 2), and
screening yeast cultures expressing random castor endosperm cDNAs for ricinoleic
acid production (chapter 3). Neither of these approaches was successful.

The results of some of the experiments described above, and a consideration of
the possible iron—containing cofactors which could be involved in the hydroxylation
mechanism, led to the development of a hypothesis concerning the molecular nature
and evolutionary origin of the oleate-12-hydroxylase of castor. It appears that the
ability to synthesize ricinoleic acid is found in isolated taxa of the plant kingdom, and
there are no reports of the occurrence of ricinoleic acid in species closely related to
castor. Therefore, it appears that the hydroxylase is a recently-evolved enzyme.
Furthermore, the biochemical similarities betweeen the hydroxylase and the
microsomal fatty acid desaturases of all plants suggests that the desaturases could be
the progenitors of the hydroxylase. Recently, it has become known that the active
site cofactor of at least one of the plant fatty acid desaturases is the diiron cluster
previously characterised in methane monooxygenase’. This cofactor is likely to be
found in all the plant fatty acid desaturases, and a discussion of the properties of this
cofactor (chapter 4) reveals that, from a chemical viewpoint, it could also be the
cofactor of a fatty acid hydroxylase, such as the oleate-12-hydroxylase of castor.
Therefore, it is hypothesized that the castor oleate-12-hydroxylase is homologous to
the plant rrricrosomal fatty acid desaturases.

This hypothesis was used as the rationale for further experiments directed

224

toward isolation of an hydroxylase gene. Experiments described in chapter 4 involved
use of one of the microsomal fatty acid desaturase genes, fad3 of Brassica napus, in
direct sceening of a castor developing-seed cDNA library. An apparently-conserved
amino acid motif among widely-divergent fatty acid desaturases was also used to
amplify sequences from castor endosperm mRN A using the polymerase chain
reaction. In both these approaches, clones for what is concluded to be the castor fad 7
fatty acid desaturase were isolated, but no other desaturase-homologues (putative
oleate-12-hydroxylase genes) were identified.

These and other results suggested that plant fatty acid desaturase genes were in
some cases too divergent for such direct gene-isolation techniques. Still
hypothesizing, however, that the hydroxylase is a desaturase-homologue, an
experiment was designed that exploited recent advances in automated DNA
sequencing technology. The partial sequencing of approximately 500 moderately
abundant and seed-specific castor cDNAs allowed the identification of many genes
expressed in castor endosperm. In particular, I was interested to identify clones with
sequence similarity to desaturases, as candidates clones for oleate- lZ-hydroxylase.
This is the first use of large-scale DNA sequencing as a strategy for the isolation of a
particular gene. The use of this strategy was based on the expectation that oleate-12-
hydroxylase transcripts should be moderately abundant in the developing endosperm,
and that these could be further enriched by differential screening for md-specific
clones. Two of the clones sequenced, pCRS677 and pCRS834, had homology to

plant membrane-bound fatty acid desaturases. In addition, a number of the other

225

genes identified in this experiment appear to encode proteins not previously isolated
or characterized from higher plants, and will be a valuable resource for the work of
numerous laboratories.

The clones pCRS677 and pCR8834 were characterized in experiments
described in chapter 6, investigating the possibility that they encode oleate-12-
hydroxylase. The two clones contained identical sequences and therefore appear to be
derived from the same gene. The presence in the castor genome of a single copy of
these sequences was confirmed by Southern analysis. The clone pCRS677 was used
to isolate a class of abundant (1/560 to 1/1120) cDNAs, including the full-length
cDNA clone pFL2, of which both strands were, completely sequenced. This 1448 bp
cDNA contains an 1161 bp open reading frame, encoding a 387 amino acid
polypeptide with a calculated molecular weight of 44407. The nucleotide and
deduced amino acid sequences have limited, but significant (ca. 47% and 38%,
respectively) identity to known plant membrane-bound fatty acid desaturase genes.
This indicated that pFL2 encodes either a desaturase-like protein, as oleate-12-
hydroxylase is hypothesized to be, or a previously uncharacterized class of desaturase.

The abundance of clones isolated using pCRS677 as a probe was in agreement
with the expression of pFL2 mRNA, as revealed by northern blot analysis.
Transcripts hybridising to pFL2 are strongly expressed in developing endosperm, and
very weakly expressed in leaf tissue, where ricinoleic acid is not synthesized, but
where fatty acid desaturases are active. These data support the possibility that pFL2

encodes oleate-12-hydroxylase.

226

The clone pFL2, in a vector suitable for galactose-inducible expression in
yeast, was transformed into yeast cells. These yeast cultures were analysed for fatty
acid content, but no differences could be detected between those expressing the pFL2
cDN A and wild—type. However, control experiments demonstrating that pFL2 was
being transcribed in the transformed yeast cells have not yet been done. The pFL2
cDNA was also inserted into a vector designed for expression in plants, and
introduced into plant cells by three separate transformation procedures, all of which
employed Agrobacterium tumefaciens. Carrot roots were cotransformed with a rooty-
tumour inducing (Ri) plasmid, while tobacco leaf explants and tobacco cultured cells
were transformed using A. tumefaciens strains harbouring disarmed Ti plasmids. An
analysis of the fatty acid composition of the transgenic carrot root, tobacco shoot, and
tobacco callus tissues obtained from these transformations failed to detect any
ricinoleic acid, and no other changes in fatty acid composition (consistent, for
instance, with pFL2—derived fatty acid desaturase activity) could be detected. Control
experiments have not yet been done to confirm accumulation of RNA transcribed
from pFL2 in the transgenic plant tissues.

It is concluded that the clone pFL2 has sequence and expression characteristics
consistent with the possibility that it encodes oleate-12-hydroxylase, but this has not
yet been shown. Despite having sequence similarity to fatty acid desaturases (and
being strongly expressed in developing castor endosperm), pFL2 had no detectable
effect on fatty acid content of transformed yeast or plants. Since ricinoleic acid is an

unusually polar fatty acid, and not normally produced in vegetative tissues, or

227

incorporated into membrane-forming lipids, it is likely that production of large
quantities of ricinoleic acid in tissues other than the developing castor endosperm
could be detrimental to the cell. Poorly-understood mechanisms apparently operate in
the castor endosperm to sequester ricinoleic acid in triacylglycerol, preventing its
possible disruption of membrane structure. Vegetative plant tissues and growing yeast
cells do not synthesize significant quantities of triacylglcerol, and may not have a
mechanism for exclusion of unusual fatty acids, such as ricinoleic acid, from
membrane-forming lipids. There are numerous possible reasons why ricinoleic acid
was not detected in the yeast or plants transformed with pFL2, including the
possibilities that pFL2 does not encode oleate-12-hydroxylase or that it was not
expressed due to problems at the level of mRNA accumulation, translation or post-
translational modification or targetting, or protein-protein interaction. However,
supposing for a moment that expression of pFL2 resulted in the introduction of oleate-
lZ-hydroxylase activity into the transgenic cells, then this was not associated with any
obvious inhibition of growth, and it is an interesting possibility that these cells have
mechanisms for recognition of the oxygenated fatty acid as alien, and for targetting
this fatty acid for degradation, precluding its accumulation to detectable levels. A
number of future experiments are suggested below which may help to differentiate
between these possibilities.

Two obvious initial experiments are of the highest priority. These are to
examine the transgenic tissues or cells for the accumulation of RNA transcribed from

the pFL2 cDNA; and to assay the tissues or cells for 14C-oleate-12-hydroxylase

228

activity using the in vitro assay described in chapter 2, or even simply by feeding “C-
oleate in viva for a brief period, after which the tissue could be killed and analysed
for any labelled ricinoleic acid. A failure to accumulate RNA transcribed from the
pFL2 cDNA would indicate that there was some problem at the level of transcription
or transcript stability, and these areas would be the logical targets of further
experiments designed to improve expression, such as the use of a different promoter,
or a transcriptional enhancer.

Alternatively, detection of pFL2 transcripts, but no detectable enzyme activity
would suggest two directions for investigation. It should be mentioned that the
membrane-bound fatty acid desaturases of plants appear to be encoded by single
genes, and the available genetic evidence in Arabidopsis thaliana gives no indication
that more than one gene product is necessary for the desaturase component of the
NADH reductase-cytochrome b5-desturase system. However, none of the membrane-
bound desaturases have been expressed in heterologous systems, and the possibility
cannot be ruled out that expression of a single castor cDNA in another plant or other
organism is sufficient for formation of a functional hydroxylase enzyme. This must
be kept in mind, and re-evaluated as experimental characterization of the desaturases
and of pFL2 proceeds.

In the first direction of investigation, expression of pFL2 could be attempted in
E. coli, with the objective of purification of the encoded protein and preparation of
polyclonal antibodies against this protein. Such antibodies could be used to determine

whether pFL2 translation products accumulate in the transgenic tissues, and to

229

confirm that they accumulate only in the developing seeds of castor, and not in other
parts of the plant. The developmental induction of such seed-specific expression in
castor might be compared to the developmental induction of oleate-12-hydroxy1ase
activity characterised in chapter 2. Furthermore, the ability of these antibodies to
inhibit oleate-hydroxylation assays could be tested.

Relevant to the second direction of investigation, preliminary information not
reported above, reveals that pFL2 is more closely related to the fad2 desaturase than
other desaturases. This is the expected result if pFL2 encodes oleate-12-hydroxylase,
since fad2 encodes the microsomal A12 desaturase, biochemically most similar to the
hydroxylase. However, it is also possible that pFL2 encodes a castor fad2 desaturase,
not expressed, for some reason, in transgenic yeast, tobacco, or carrot. It is possible
however that A12 desaturase activity could be detected in a tissue with a very low
background level of this activity, and pFL2 might therefore be tested for
complementation of the Arabidopsis thaliana fad2 mutant.

If the experiments suggested earlier do detect oleate-12—hydroxylase activity in
tissues expressing pFL2, and it can be established that this clone does indeed encode
the hydroxylase, then an assortment of interesting experiments await the investigator.
One would be to isolate the castor fad2 gene, the presumed progenitor of the
hydroxylase gene. As structure-function relationships of the fatty acid desaturases
become better characterised, a comparison of the primary and predicted secondary
structures of a desaturase and its derived hydroxylase homologue would be extremely

interesting, and might provide valuable insight into the nature of the reaction

230

mechanisms of these enzymes, and suggest strategies for protein engineering.

Other interesting experiments would obviously be to work towards the
envisioned future goal of this work, the production of ricinoleic acid in a plant not
previously synthesizing this compound. Since it appears that one of the hurdles to be
crossed may be to overcome endogenous mechanisms for degradation of alien fatty
acids, this would be an interesting area for study. One experiment that might be
suggested would be to transform Arabidopsis with the hydroxylase gene, and then to
screen for mutants with a defect in the process of recognition or degradation of the
alien ricinoleic acid. Another experiment might exploit an interesting technique
which has not yet been used to its potential for the study of unusual fatty acids. This
is the technique of Terzaghi,3 in which exogenous fatty acids, as their Tween
(polyoxyethelyenesorbitan) esters, painted on or otherwise applied to plant tissues, are
extensively incorporated into membrane lipids. What would be the fate of ricinoleic
acid applied in this way?

In any case, whether it is found that pFL2 encodes oleate-12-hydroxylase, or
possibly some new class of desaturase, it seems that this clone will provide useful
opportunities in the study of plant fatty acid metabolism. Furthermore, the many
other clones, some with homology to genes previously uncharacterised in plants,
identified in the course of isolating pFL2, should provide plant biologists with a

number of other new opportunities.

231 '
REFERENCES

1. Bafor, M., Smith, M.A., Jonsson, L., Stobart, K., and Stymne, S.,
Ricinoleic acid biosynthesis and triacylglycerol assembly in microsomal
preparations from developing castor—bean (Ricinus communis)
endosperm, Biochem. J., 280, 507, 1991.

2. Fox B.G., Shanklin, J., Somerville, C., and Munck, E., Stearoyl-acyl carrier
protein A9 desaturase from Ricinus communis is a diiron-0x0 protein, Proc.
Natl. Acad. Sci., 90, 2486-2490, 1993.

3. Terzaghi, W.B. , Manipulating membrane fatty acid compositions of
whole plants with tween-fatty acid esters, Plant Physiol. , 91 , 203,
1989.

APPENDIX

APPENDIX

MASS SPECTRONIETRIC STUDIES: POSITION OF HYDROXYLATION AND

SOURCE OF HYDROXYL OXYGEN

Note: the work presented in this appendix is largely that of Dr B.W. Underhill, my
contributions being relatively minor. We acknowledge the assistance of Drs Tim

Heath and Doug Gage with mass spectral analyses.

INTRODUCTION

Ricinoleic acid is formed by two pathways; by hydration of linoleic acid as is
the case in the fungus Claviceps purpurea and by the hydroxylation of oleic acid as
occurs in castor (see chapter 2). Yamada and Stumpt1 reported ricinoleic acid was
formed from l“C-oleoyl-CoA by an NADPH-requiring soluble fraction (105 000 x g
supernatant) isolated from extracts of the developing castor bean. However, in a
following publication’ it was claimed that the hydroxy product isolated above was not
ricinoleic acid and, without reporting experimental data, indicated the product "has
now been shown to be B-hydroxyoleic acid". Galliard and Stumpf2 found that
ricinoleic acid was derived from oleoyl-CoA by hydroxylation, that the oxidation
required NADH and that the enzyme was present in the microsomal fraction of

developing seeds of castor. In both papers, the methods used to identify the product

232

233

of the enzyme reaction were rigorous and appeared convincing.

Several methods are available which may be used to determine if ricinoleic
acid is the product formed in a cell-free reaction mixture. Most frequently, chemical
degradation followed by thin-layer and/ or gas chromatographic separation of the
resulting degradation products has been used for the identification of hydroxy fatty
acids, including ricinoleic acid.“5 In the studies of Galliard and Stumpt2 a very
elegant system was employed in which both 1‘C- and 3H-1abeled oleoyl-CoA were
utilized. A similar method was considered for the studies reported here and an
attempt was made to obtain 3H-labeled oleic acid from the same source as had been
used by Galliard and Stumpf. However, the material was no longer available and its
synthesis is not a trivial matter.

An alternative and unequivocal method of identifying the hydroxylated product
formed in these studies was investigated and it was decided to employ mass
spectrometry. The use of mass spectrometry for reaction product analysis is an
attractive method for a number of reasons. These include: (1) The electron impact
mass spectra (EI/ MS) of the trimethylsilyloxy (TMS) derivatives of several
hydroxylated fatty acid methyl esters, including ricinoleic acid, have been reported.7
It has been shown that the position of the hydroxyl group within the chain may be
readily ascertained from the fragmentation ions, produced. (2) Mass spectrometry,
particularly when combined with the high resolution capability of capillary GC, has
high sensitivity and can be used for reaction product identification as opposed to

product confirmation as is often the case with chemical degradation studies. (3)

234
Chemical degradation is, at best, a confirmatory procedure that offers little if the

product of the enzyme reaction is other than that expected. In addition, chemical
degradation is often followed by thin-layer and/or packed column separation of the
reaction products, procedures that have relatively low resolving power by current
standards. (4) In addition to providing structural identification of the product formed,
mass spectrometry in combination with the use of stable isotopes may also be used to

determine the source of oxygen in the hydroxyl group of ricinoleic acid.

EXPERINIENTAL DESIGN

The EI mass spectrum of the TMS-derivative of methyl-ricinoleate (mol. wt.
384) contains a number of fragmentation ions which characterize the compound.
These include the ions at m/z 369 [M-15]+, 353 [M-(CH3O)]+, 187 (base peak) and
299 (the last 2 are cleavage ions alpha to the TMS group) and the rearrangement ion
at 270 [CH3O(O'I‘MS)C(CH2)-,CH=CH-CH2]+. The ions at m/z 187, 299 and 270
unequivocally establish the hydroxyl at C-12. Accordingly, ricinoleic acid or any
other hydroxylated product formed from a specifically deutero-labeled oleoyl—CoA,
would be recognisable in a reaction mixture from the resulting fragmentation ions
present in the EI mass spectrum of the TMS-methyl ester(s). In addition, since the
hydroxyl oxygen in the product is not lost as a result of TMS—derivatization, it should
be possible to determine if the hydroxyl oxygen is derived from molecular oxygenby

conducting the experiment in the presence of oxygen-18. Since the ion at m/z 187

235

constitutes the base peak in the spectrum a search was made for a source of oleic acid
containing deuterium at carbons 12 through 18. Thus, for ricinoleic acid formed
from an oleoyl-CoA labeled with deuterium at carbons 12 through 18, the mass of the
fragmentation ion corresponding to the ion at m/z 187 [CH3(CH2)5CHOTMS]+ would
be increased by (1) the number of deuteriums in the substrate and by (2) two

additional mass units if the oxygen was derived from "‘02.

MATERIALS AND NIETHODS

Chemicals

1-"C-oleoyl-C0A (52 pCi/pmol) was prepared using acyl-CoA synthetase
(Sigma) and 1-“C-oleic acid (1 pmol) according to the method of Taylor et al.8 The
product was formed in > 95 % yield and had a radiopurity of 98 % . Trideuterooleoyl-
CoA was synthesized in a similar manner starting from 18-trideuterooleic acid (1
pmol, >98% isotope purity): the deutero-fatty acid, synthesized by Dr. A.P. Tulloch
(deceased), was provided by the Plant Biotechnology Institute, Saskatoon, SK.

Canada. All other chemicals were obtained from commercial sources.

Enzyme Preparation

Developing castor endosperm plus embryo tissue at stages IV to VI’ was

236

collected into liquid nitrogen. This material was powdered with a mortar and pestle,
and stored at -70°C. This powdered endosperm (1.41 g) was extracted by grinding
(Ten Broek homogenizer) with 1.4 ml of extraction buffer consisting of 0.05 M
PIPES pH 7.1, 10% (w/v) glycerol, leupeptin (5 ug/ml) and 1 mM cysteine. The
residue was extracted a second time with a similar volume of extraction buffer. The
enzyme extracts were combined and centrifuged at 13 600 x g for 15 min in a
microfuge. The clear supernatant (~ 3 ml) was decanted leaving behind the floating
pad of fat and residual solids. MgCl2 was added to give a final concentration of 50
mM and the mixture centrifuged at 13 600 x g for 10 min. The supernatant was used
as the source of enzyme. The enzyme was used immediately, or was dropped into

liquid nitrogen and stored at -70°C.

Enzyme Assays

Reaction mixtures contained 50 ul enzyme, 10 pl 20 mM NADH in Tris-HCl
pH 7.1, 5 ul of substrate [1-“C—oleoyl-C0A (15 950 dpm/ul; 52 Ci/mol) or 18-
trideuterooleoyl-CoA (0.2 nmol/ul)] and 435 pl of 50 mM PIPES pH 7 .1 containing
10% (w/v) glycerol. For reactions carried out in the presence of “Oz, substrate,
cofactor and buffer were first added, and immediately following the addition of
enzyme, the reaction tubes were alternately evacuated and filled with one of the
following: (1) a mixture of 20% ”02 plus 80% nitrogen, (2) nitrogen, or (3) air.

Replicated reactions were carried out at 30°C and were terminated by addition of 0.5

237
ml of 15% KOH in methanol. Mixtures were saponified at 80°C for 30 min, cooled,

acidified with 0.5 ml 2.5 M HCl and the liberated acids extracted using 3.5 ml
hexane:ispropanol (3:2) and 2.5 ml 0.2 M Na2804‘°. The organic phase was

separated and concentrated under nitrogen.

Quantitation of Ricinoleic Acid

Ricinoleic acid was determined by GC using heptadecanoic acid as an internal
standard; for routine analyses the C-17 internal standard was added to the reaction
mixtures immediately prior to saponification. Methyl esters of fatty acids were
prepared after reducing the organic phase to dryness by the addition of 0.5 ml of
anhydrous 1 M HCl in methanol and heating to 80°C for l h. Methyl esters were
recovered from the cooled reaction mixtures by addition of 1.5 ml 0.9% NaCl and
extraction into 200 pl hexane.

TMS-derivatives of hydroxy fatty acids were prepared by adding an excess (~
5 ul) of trimethylsilylimidazole (Pierce) to a dried (N a280,) solution of methyl esters
in hexane. Derivatization was carried out at room temperature for a period of at least
2 h. Excess derivatizing reagent was removed under a stream of dry nitrogen and the

TMS-methyl esters dissolved in hexane.

238
Thin Layer Chromatography

Fatty acids were spotted (alongside authentic oleic acid and ricinoleic acid
standards) on a silica TLC plate (Baker Si250). The plate was developed in a paper-
lined tank containing benzenezethyl etherzethanol (100:30:2).u Fatty acids were
detected by staining with iodine vapour. Radioactivity could be quantitated by
autoradiography followed by scraping and scintillation-counting of the oleic acid and
ricinoleic acid spots, but equivalent quantitation was routinely obtained using a

BioScan 2000 scanner.

Mass Spectrometry

GC-MS of TMS derivatives of fatty acid methyl esters was performed on a
Finnigan TSQ 70 mass spectrometer equipped with a Varian 3400 gas chromatograph.
The column used was a DB-% capillary (30 m x 0.25 mm: J & W Scientific Inc.,
Ranchero Cordova, CA), programmed from 150 to 280°C at 12°C min". Helium
was used as carrier gas (~ 1 ml min"). Spectra were aquired every 0.5 s over the
range m/z 50 to 400.

Mass spectral analyses were carried out on duplicate samples of enzyme
incubations containing lS-trideuterooleoyl-CoA and each sample was assayed in

triplicate by GC-MS. The results were averaged and standard deviations calculated.

239
RESULTS AND DISCUSSION

The crude enzyme preparation used contained appreciable quantities of
ricinoleic acid; the amount of ricinoleic acid (calculated as methyl ricinoleate) was
5.88 :1; 0.56 ug(50 ul)‘l (n=6).

The enzyme preparation was active. Bioscan data, obtained from three
incubations using 1-“C-oleoyl-COA as substrate, showed 20.15% (20.66, 20.60 and
19.19) of the activity supplied had been converted into a component co-migrating with
ricinoleic acid. Based on the total activity and the specific activity of 1-"C-oleoyl-
CoA employed, the amount of labelled product (ricinoleic acid) formed in these
calculations was calculated to be 0.14 nmol or 41.5 ng. (Note: labelled ricinoleate
constitutes ~ 0.7 % of the total ricinoleate in the sample).

Mass spectral data obtained from incubations containing 18-trideuterooleoyl-
CoA unequivocally demonstrated that the substrate was converted to ricinoleic acid.
As indicated above, alpha-fragmentation to the TMS group at the 12-position of
ricinoleic acid gives rise to the ion at m/z 187, the base peak in the mass spectrum.
The corresponding fragmentation ion which could be derived from ricinoleic acid
formed from the trideuterated substrate would occur at m/z 190. From the % relative
abundance of ions in the range m/z 187 to 192 (Table 17) the ion at m/z 190 is
significantly greater than that found in ”blank” incubations (1.356 vs 0.631). From
this data the amount of the deutero-labeled substrate converted can be estimated,

namely (1.356 - 0.631)(5880)/ 100 = 42.6 ng, in close agreement with the calculated

240

conversion of the “C-labeled substrate.

The aquired mass-spectral data was searched using single ion monitoring to
determine if another mono-trimethylsilyloxy, monounsaturated C-18 methyl ester
could be detected. Those searched included ions at m/z 384 [M+], 369 [M+], as well
as ions corresponding to alpha cleavages of 2- and 3—trimethylsi1yloxy-methyl oleate,
with and without a trideuterated carbon. No evidence was obtained in support of
these components. In the course of these searches, data was obtained supporting
the presence of 2-hydroxystearic; co-incident retention time and matching spectrum
with authentic 2-trimethylsilyloxy-methy stearate.

Data obtained from experiments employing 1-"C- and 18-D3-oleoyl-C0A
confirmed and extended findings reported previously.2 labeled carbon from 1-“C-
oleoyl-CoA was extensively incorporated into ricinoleic acid (or a component 00-
rnigrating with ricinoleic acid) when reactions were carried out in air or in a mixture
of 20% "‘02 plus 80% nitrogen. In air, 16.6% (18.7, 15.4, 15.7) of 14C co-migrated
with authentic ricinoleic acid. In 20% “‘02, 15.0% (15.4, 15.1, 14.6) of 1“C co-
migrated with authentic ricinoleic acid. There was no conversion of 1‘C into the
product in the absence of oxygen.

The results obtained from reaction mixtures containing 18-trideuterooleoyl-
CoA, incubated in the presence of “Oz, conclusively demonstrated that the hydroxyl
oxygen of ricinoleic acid was derived from the “02. Data obtained from these
incubations (Table 17) show a significant increase in the relative abundance of the ion

at m/z 192, as would be expected from a trimethylsilyloxy alpha fragmentation ion

 

 

241

containing an 18O and 3 deuterium atoms. The slightly elevated relative abundance of
the ion at ml 2 189 is likely a result of 1“O incorporation into a ricinoleic acid

intermediate present in the enzyme preparation.

Table 17. Formation of ricinoleic acid from oleoyl-CoA by cell-free extracts of
developing castor endosperm. Extracts were incubated with 18-trideutero-oleoyl-COA
in the presence of air or 1802. The position at which oleic acid was hydroxylated as
well as the source of the hydroxyl oxygen was determined by GC-MS analysis of
fragmentation ions (m/z 187 to 192) derived from the recovered trimethylsilyl ether of
methyl ricinoleate. Each treatment was replicated two times and each replicate was
analyzed three times by GC-MS.

 

Experiment % Relative Abundance of Ions‘ at m/z

187 188 189 190 191 192

 

l-“C-oleoyl-CoA 100 14.859 4.722 0.631 0.075 0.011
(0.227)” (0.144) (0.076) (0.062) (0.014)

D3-oleoyl-COA (air) 100 15.302 4.773 1.356 0.183 0.050
(0.405) (0.086) (0.059) (0.013) (0.024)

D3-oleoyl-C0A (“02) 100 15.094 5.687 0.880 0.130 0.525
(0.263) (0.082) (0.122) (0.061) (0.069)

 

‘ mean values, n=6

" standard error of the mean

242

REFERENCES

10.

11.

Yamada, M. , and Stumpf, P.K. , Enzynric synthesis of ricinoleic acid
by extracts of developing Ricinus communis L seeds, Biochem.
Biophys. Res. Commun., 14, 165-171, 1964.

Galliard, T., and Stumpf, P.K., Fat metabolism in higher plants XXX.
Enzymatic synthesis of ricinoleic acid by a microsomal preparation
from developing Ricinus communis mds, J. Biol. Chem. , 241, 5806,
1966.

Barucha, K.E., and Gunstone, F.D., Vegetable oils VI: The component
acids of ergot oil, J. Chem. Soc., 610-614, 1957.

Ahmad, M.S., Ahmad, M.U., and Osman, S.M., A new
hydroxyolefinic acid from Plantago major seed oil, Phytochem. , 19,
2137-2139, 1980.

Husain, S., Ahmad, M.U., and Osman, S.M., New hydroxy fatty acid
from seed oil of Baliospennum axillare, Phytochem., 19, 75-77, 1980.

Ahmad, M.U., Husain, S.K., and Osman, S.M., Ricinoleic acid in
Phyllanthus niruri seed oil, J. Am. Oil Chem. Soc., 58, 673, 1981.

Kleiman, R. , and Spencer, G.F., Gas chromatography-mass
spectrometry of methyl esters of unsaturated oxygenated fatty acids, J.
Am. Oil Chem. Soc., 50, 31-38, 1973.

Taylor, D.C., Weber, N., Hogge, L.R., and Underhill, B.W., A
simple enzymatic method for the preparation of radiolabeled erucoyl-
CoA and other long—chain fatty acyl-CoAs and their characterization by
mass spectrometry, Anal. Biochem., 184, 311-316, 1990.

Greenwood, J.S. , and Bewley, J.D. , Seed development in Ricinus communis
(castor bean). 1. Descriptive morphology, Can. J. Bot., 60, 1751-1760, 1982.

Moreau, R.A., and Stumpf, P.K., Recent studies of the enzymic
synthesis of ricinoleic acid by developing castor beans, Plant Physiol. ,
67, 672, 1981.

Donaldson, R.P. , Accumulation of free ricinoleic acid in germinating
castor bean endosperm, Plant Physiol., 59, 1064-1066, 1977.