.. 27,2; 7

. 2,-2-222".
I'T'nw-
(..'

.
.2- . ,. ..
Aug-52‘ .m. '.1..
I" '3
\

.mnhrp
(2.

n

.ffi'miw

.2.«_.\.2.2.'...L.....2 ..2
.__.,. :..«.2- m7.
.“ '.:

Y ' J
m- :..2".;-'.:.5
22,23“- u-4 :.,»......,. .-

22“.}!
-E.‘

a
2.

"IF" 0 .
1,.

 

 

 

 

"'2 '-n-'I vrv
2 “-2. 2 v2 (m; -.,_.,\n;.,- , ,
J. ,._, 7:1: wk. .2. I .. ,u
-u2-‘. Jug «11;;
4 ‘

7... g.

.22
A},
II

24...... .2‘4- . A
n ‘r'

I
..
. Jlftt'Y'JZ'T
_-u'¢

w! _
wd 2
V‘TI; up: J? v 'm'
r gang-pm -
rulvrm . ‘rrrs
3m. ., . .33

5.21“"
'2-pn
.

{2 3.2.72? .
.3... 2 r
,.,...,r..,....,.,.,.
”my.”
. 2.622
’.

$112721"?!
fl.

. v-v'r nwmmm 90-.
n a"... I...» a?" 72..., .1.
5'53

;
I'rri-(lev' .
w. . . ...
."L 1.1.:‘11rv',\",">

4,... (-

15‘ r'
qu'qY-xpn
Y'" r, r- X.

3-:

'2
“0.5.”.-.
‘ not, r- I.

I":

..
'.‘" :Eu'.'.:..=m
. .

"7:2—
:v'

 

1
3,75“!!!
9— ummvr u —:
Vrlv‘n! “KP;
-.2(-

.- 'r!‘
,r. ar.7r;-.,'¢ ”4 r
.2. .2“. :5..."- n
m r-

r
“ 2-221 ~' .
ummré-wvim: r~~-l:»'r
w...»,-..,—.2,.,2...........,...,.,,. N
22v...
..

,.,.

.2.- .r - . .
’f‘lvw—l" ”J"

_.;_....,..I_an-,., 3.... -— .

2-,. ”fly-re” _

vf‘F 2n 1"! '52:.
{.23 .2 :. -- , ,, 'Jar’w'
w. "2.3%.. Wan-r. Mfr ‘
439' «1"? 172-:1mr42. 1-:- (- 4
”2;" :g r rv'vL-w-wu- 3
.., ,.

.' ', . .2
«3—01.1321. 1.1-!
m7“ ~02
thy-2'. "Hawk-1;...- _
... w"... r.
mans:
"' 2 Cla'g"
4..

5.9,...»
.,
.7

.aww'iu’} -,-.
fir!"

{my-rs”... .

.2. .Y . ..

.Ji— .iqr‘ m
,...

r2 - =_ r-
3221.2: 3;"... '— Q’T; p: '3 f" . 1..-...- ‘
PE‘}: 3::- . r — I 33:1 1-2.
1.....2‘." ”""'r.':‘::r:'.":';. .."- "
_" “1::
.30....“ 3,
7‘“ "rm. " .L‘..3:;,::..;..,:...
'1‘” ”1‘": "Iv-w" r»-
-v: 'vu'IP-W'rr" '- v i
r. ..- .2...

r).
V

41.53;; "-

#3:”...- “13:57.51 -.. 752.2: w—Tswr.

3.22.3.7 f:

”we

2‘3: 1 ILJE'Ji- W

3,.

,9

-r ....,.

.
If .l H
"‘ 5'" '1’?! 751......7ra..,.. .9. m
' ’3," ’l' I" .. ' ,1.” V, 7.: .
V: . _ . ..
3 . . a... . y, .. . ,:
x L' 47 ,...

II V.
.,.g;.-;
.mfi 222:2
.3332?" ”$4.3
. ...

.Y. ..,.

II>’. --‘/.;_ul.|<.-2..-n..
....'.. ~v n
.\. v .’. ..

...

., .F- ’2“...

... ....,.. ..
' . 4.;

 

worry—4:.
4 fay:

. ... .. ,
' ':,x.v,7:rg..:;fi;g ,
v " 4 r
7m' '13:;
' t"

3 v I ‘5

. I". T,L..f.._.

. 37,. .,.2. .-

wig» C—fnr—Fv'thrn.
.. .13-..

HIGAN STATE

llllfllll lzll

 

 

 

 

 

 

 

 

 

 

llllllllllllllllllllllllll!lllllllll

93 00917 3554

This is to certify that the

dissertation entitled

GENETIC ANALYSIS OF THE PATHWAY FOR THE BIOSYNTHESIS
OF THE PLANT SULFOLIPID IN THE PURPLE BACTERIUM
RHODOCATER SPHAEROIDES

presented by

CHRISTOPH BENNING

has been accepted towards fulfillment
of the requirements for

PH.D. degree in GENETICS

 

 

f\

:'

.,.. .f‘“ \ U
L Rdllifitul k

Major professor

Date MARCH 14, l99l

 

MSU is an Affirmative Action/Equal Opportunity Institution 0712771

 

LIBRARY
Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE
WAR 3 9 Z
i 1 T AUG 2 2 2&0

 

 

a
wile u:

 

4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution
C1\CllC\dfl1“LflIW3‘p.1

GENETIC ANALYSIS OF THE PATHWAY FOR THE BIOSYNTHESIS OF THE
PLANT SULFOLIPID IN THE PURPLE BACTERIUM
HHODOBACTEH SPHAEHO/DES

By

Christoph Benning

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Genetics Interdepartmental Doctoral Program
MSU-DOE Plant Research Laboratory

1991

37

Viet- 07>?

I

ABSTRACT
GENETIC ANALYSIS OF THE PATHWAY FOR THE BIOSYNTHESIS OF THE
PLANT SULFOLIPID IN THE PURPLE BACTERIUM
RHODOBACTER SPHAEROIDES
BY

Christoph Benning

Approximately thirty years ago, Benson and collaborators discovered
the plant sulfolipid and determined its structure to be 6-sulfo-a-D-quinovosyl
diacylglycerol. Since then, relatively little progress has been made towards the
elucidation of the biosynthetic pathway and the function of sulfolipid.

To solve this classical problem, the purple non-sulfur bacterium
Rhodobacter sphaeroides was used as a model organism. The presence of
authentic sulfolipid in this bacterium was confirmed by comparative mass
spectrometry of bacterial and spinach sulfolipid. Mutants of Ft. sphaeroides
were isolated that were deficient in sulfolipid biosynthesis. The analysis was
focused on three mutants carrying defects in the squ, squ, and squ
genes, respectively. The squ and squ mutants were found to contain
35S-sulfate labelled, novel compounds, possibly precursors of sulfolipid
biosynthesis.

The squ, squ, and squ genes were isolated using a
complementation assay. The squ mutant was complemented by a different
cosmid than the squ and squ mutants. Small subclones of the cosmids

complementing the squ and squ mutants were sequenced, and the

putative coding sequences for the two genes were identified. The squ gene
encodes a 33.6-kD protein, not homologous to any known protein. The squ
gene codes for a 46-kD protein. Its N—terminal amino acid sequence shows
homology to the N-terminus of UDP-glucose epimerase from bacteria and
fungi, suggesting that the squ gene product is involved in sulfolipid
biosynthesis on the level of UDP-sulfoquinovose or some other sugar
nucleotide. The squ gene was found to be located directly downstream of
the squ gene, suggesting that squ and squ are organized in an operon.

To analyze putative precursors of sulfolipid biosynthesis, an in vitro
sulfolipid synthesis system was developed. The incorporation of 35S-PAPS into
authentic sulfolipid by extracts of Ft. sphaeroides wild type cells was observed.

Based on the data provided in this work, definitive experiments can be
designed, which will provide a solution to the problem of sulfolipid

biosynthesis and function in the near future.

To

My Parents

 

ACKNOWLEDGMENTS

I thank Chris Somerville for his support throughout my work. He provided
me with excellent advice, and taught me how to approach difficult, scientific
problems. I am grateful for the support and advice I received from the members
of my advisory committee: Hans Kende, John Ohlrogge, Peter Wolk, and Lee
McIntosh. I would also like to thank all the members of the lab and all my friends
in the PRL for their advice and their helpful discussions. I thank Doug Gage and
Dr. Huang for analyzing samples by mass-spectrometry, and for their help
concerning the interpretation of the spectra.

I thank my parents, who gave me the freedom to find my own way, and
who supported all my decisions regardless of the hardships these brought upon
them. I would also like to thank Susi and Urs for their love and patience they
gave during the preparation of this work.

I am grateful for a stipend awarded to me by the German Scholarship

Foundation, which made it possible for me to get started in E. Lansing.

TABLE OF CONTENTS

Page
LIST OF TABLES .......................................... ix
LIST OF FIGURES .......................................... x
LIST OF ABBREVIATIONS .................................. xiv
CHAPTER 1: INTRODUCTION ................................. 1

DISCOVERY AND STRUCTURE OF THE PLANT SULFOLIPID . . . . 1
OCCURRENCE AND DISTRIBUTION OF THE PLANT

SULFOLIPID .................................... 3
PROPOSED BIOSYNTHEI'IC PATHWAYS FOR THE PLANT
SULFOLIPID ......................................... 5
PROPOSED FUNCTION OF THE PLANT SULFOLIPID .......... 12
ABOUT THE PRESENTED WORK ......................... 14
REFERENCES ....................................... 15

CHAPTER 2: COMPARISON OF SULFOLIPID FROM SPINACH AND R.
SPHAEROIDES BY MASS SPECTROMETRY ................. 20
ABSTRACT ......................................... 20
INTRODUCTION ...................................... 21
MATERIALS AND METHODS ............................ 23
RESULTS AND DISCUSSION ............................ 25
CONCLUSIONS ...................................... 37

vi

REFERENCES ....................................... 38
CHAPTER 3: H. SPHAEROIDES AS A MODEL SYSTEM FOR THE

BIOSYNTHESIS OF THE PLANT SULFOLIPID ................ 39
ABSTRACT ......................................... 39
INTRODUCTION ...................................... 40
MATERIALS AND METHODS ............................ 44
RESULTS AND DISCUSSION ............................ 56
CONCLUSIONS ...................................... 74
REFERENCES ....................................... 77

CHAPTER 4: ISOLATION AND CHARACTERIZATION OF A CLONE
COMPLEMENTING THE SULFOLIPID-DEFICIENT squ MUTANT

OF H. SPHAEFIOIDES ................................. 78
ABSTRACT ......................................... 78
INTRODUCTION ...................................... 79
MATERIALS AND METHODS ............................ 80
RESULTS AND DISCUSSION ............................ 93
CONCLUSIONS ..................................... 1 14
REFERENCES ....................................... 1 16

CHAPTER 5: ISOLATION AND CHARACTERIZATION OF A CLONE
COMPLEMENTING THE SULFOLIPID-DEFICIENT squ AND squ

MUTANTS OF Fi. SPHAEHOIDES ......................... 119
ABSTRACT ........................................ 1 19
INTRODUCTION ...................................... 120

vii

 

MATERIALS AND METHODS ............................ 121

RESULTS AND DISCUSSION ............................ 126
CONCLUSIONS ...................................... 153
REFERENCES ....................................... 155
CHAPTER 6: IN VITRO SULFOLIPID BIOSYNTHESIS FROM “S-PAPS IN
EXTRACTS FROM Fi. SPHAEFiO/DES ...................... 157
ABSTRACT ......................................... 157
INTRODUCTION ...................................... 158
MATERIALS AND METHODS ............................ 159
RESULTS AND DISCUSSION ............................ 167
CONCLUSIONS ...................................... 183
REFERENCES ....................................... 185
CHAPTER 7: CONCLUSIONS AND PERSPECTIVES ................ 187
REFERENCES ....................................... 195
APPENDIX A: THE EFFECT OF SULFOLIPID DEFICIENCY ON GROWTH
OF R. SPHAERO/DES ................................. 196
APPENDIX B: CONSTRUCTION OF THE EXPRESSION VECTOR
pCHB$00 ........................................... 199
REFERENCES ....................................... 203

viii

LIST OF TABLES

Page

Table 3-1: Stock solutions for OrmerOd’s medium ................... 45

Table 3-2: Ingredients Of Sistrom’s 10X medium. ................... 46

Table 3-3: Trace element solution for Sistrom’s medium. ............. 46

Table 3-4: Vitamin solution for Sistrom’s medium. .................. 47
Table 3-5: Effect of duration Of exposure to nitrosoguanidine on survival

and on induction Of mutations. ........................... 57

Table 3-6: Preliminary characterization Of putative sulfolipid mutants. . . . . 63
Table 3-7: Polar lipid composition (weight %) Of wild type and sqd mutant
strains Of R. sphaeroides grown under Chemoheterotrophic

conditions. .......................................... 72

Table 4-1: Bacterial strains. ................................... 81

Table 4-2: Plasmid sources and derivation. ....................... 82
Table 4-3: List Of synthetic oligonucleotide primers used to sequence

pCHB17820 ........................................ 103

Table 5-1: Plasmids used in the experiments described in Chapter 5. . . . . 122
Table 5-2: List Of synthetic oligonucleotide primers used for sequencing

of the pCHB16014 insert. ............................... 134
Table 6-1: Effect of sulfur-containing compounds and aspartic acid on the

incorporation Of 35S-PAPS into sulfolipid. .................... 181

LIST OF FIGURES

Page
Figure 1-1: Structure Of the plant sulfolipid, 6-sulfo—a-D-quinovosyl diacyl-
glycerol. ............................................
Figure 12: Proposed pathways for the biosynthesis of the plant sulfolipid
(adopted from 15). .................................... 7

Figure 2-1: Negative FAB mass spectra Of sulfolipid samples from spinach
(top) and R. sphaeroides (bottom). ........................ 26

Figure 2-2: B/E linked FAB-CAD-MS/MS scans focusing on the most
abundant molecular ion Of the spinach (top) and the bacterial
sample (bottom). ..................................... 28

Figure 2-3: B/E linked FAB-CAD-MS/MS scans focusing on monoacyl
species of sulfolipid from spinach (top) and from Fi. sphaeroides
(bottom). ........................................... 30

Figure 2-4: B/E linked FAB-CAD-MS/MS scans focusing on the head
group key ion at m/z 241 from spinach (top) and bacterium
(bottom). ........................................... 31

Figure 2-5: Proposed fragmentation pathway showing the decomposition
of the key ion at m/z 241. ............................... 32

Figure 2-6: B/E linked FAB-CAD-MS/MS scans of the Spinach sulfolipid,
focusing on the "daughter ions" with the mass 299 (top), and 283
(bottom). ........................................... 34

Figure 2-7: B/E linked scans Of the spinach sulfolipid, focusing on the
"daughter ions" with the mass 225 and 80. .................. 35

Figure 2-8: Alternative fragmentation pathway showing the decomposition
of the key ion at m/z 299. ............................... 36

Figure 3-1: Structures Of the most abundant polar lipids found in
H. sphaeroides. ...................................... 41

Figure 3-2: Separation of in viva labelled lipids from R. sphaeroides on
ammonium sulfate plates. ............................... 59

Figure 3-3: Separation of lipid samples prepared from twelve mutagenized
cell lines on ammonium sulfate plates ....................... 61

Figure 3-4: Separation Of water-soluble, a“‘S-labelled compounds contained
in the acidic fraction of wild type and mutant extracts by 20
TLE/T LC. ........................................... 65

Figure 3-5: Separation of water-soluble, 35S-labelled compounds contained
in the basic fraction of wild type and mutant extracts by 20
TLE/T LC. ........................................... 66

Figure 3-6: Time course for the incorporation Of asS-labelled sulfate into
protein (upper panel) and sulfolipid (lower panel) by wild type and
mutant strains CHB16 (squ), CHB17 (squ), and CHB18

(squ). ............................................ 69
Figure 3-7: Separation of lipids isolated from wild type (A) and three sqd

mutants by 20 TLC. ................................... 71
Figure 4-1: Detailed restriction map Of pLA2917 (taken from ref. 1). ..... 94
Figure 4-2: Complete restriction map for the cosmid clone pCHB1701. . . 96
Figure 4-3: Complete restriction map Of two subclones, derived from

pCHB1701, which complement the squ mutant. ............. 98
Figure 4-4: Southern blot Of pCHB17011 and pCHB17012 probed with a

0.4 kb Pstl fragment of pCHBl7012. ...................... 100
Figure 4-5: Southern blot Of pCHB1701 probed with a 0.4 kb Pstl

fragment from pCHB17012. ............................ 101
Figure 4-6: Strategy for sequencing pCHB17820. ................. 104

xi

 

Figure 4-7: DNA sequence of pCHB17320. ...................... 105

Figure 4-8: Comparison of the putative squ ribosome binding site to
binding sites from various H. sphaeroides genes (A) and to the
3’-end Of the 168 rRNA from R. sphaeroides (B) .............. 107

Figure 4-9: Complementation analysis using derivatives Of pCHB17B20
cloned into the expression vector pCHB500. ................ 111

Figure 4-10: Hydropathy plot of the predicted squ amino acid
sequence. ......................................... 113

Figure 5-1: Lipid phenotype of the squ mutant lacking or containing the

cosmid pCHB1601. ................................... 127
Figure 5-2: Simple restriction maps for pCHB1601, the two subclones
pCHB1611 and pCHB1612, and pLA2917. .................. 129
Figure 5-3: Detailed restriction maps for the pCHB1601 subclones
pCHB1611 and pCH81621. ............................. 131
Figure 5-4: Sequencing strategy used to sequence the 2.4 kb insert Of
pCHB16014 .......................................... 133
Figure 5-5: Nucleotide sequence Of the insert found in pCHB16014. . . . . 135

Figure 5-6: Comparison Of the putative squ ribosome binding site to
binding sites from various Ff. sphaeroides genes (A) and to the 3’-

end Of the 168 rRNA from R. sphaeroides (B). ............... 137
Figure 5-7: Complementation analysis using various subclones Of

pCHB16014 .......................................... 141
Figure 5—8: Cat-gene insertion clones of pCHB16014 and CAT-activity

Observed in cells containing these Clones .................... 143
Figure 5-9: Hydropathy plot for the putative squ gene product. ....... 146

xii

Figure 5-10: Comparison Of the amino acid sequences Of the putative
squ gene product and UDP-glucose epimerases from S.
cerevisiae and E. coli. .................................. 147

Figure 5-11: Location and orientation Of the squ and squ genes on
pCHB1611. ......................................... 150

Figure 6-1: Incorporation of 35S-PAPS into sulfolipid by broken and intact
wild type cells of Ft. sphaeroides. ......................... 168

Figure 6-2: Dependence Of the incorporation of 35S-PAPS into sulfolipid on
the amount Of extract present in the reaction mixture. .......... 169

Figure 6-3: Comparison Of extracts from wild type and sulfolipid deficient
squ mutant cells for their capability to label sulfolipid from
35S-PAPS. ........................................... 171

Figure 6-4: Cochromatography of deacylated sulfolipid labelled in vitro
and in vivo. ......................................... 172

Figure 6-5: Substrate saturation for the incorporation Of 35S-PAPS into
sulfolipid. ........................................... 174

Figure 66: Effect Of the addition of various compounds on the
incorporation Of asS-PAPS into sulfolipid. .................... 177

Figure 6-7: Effect Of sulfite on the incorporation of 35S—PAPS into
sulfolipid. ........................................... 180

Figure A-1: Comparison of two isogenic lines of R. sphaeroides,
containing different amounts Of sulfolipid, for their heterotrophic
growth. ............................................ 197

Figure B-1: Sequence of the cycA gene from R. capsu/ata, Upstream Of
the initiation codon (taken from 2). ........................ 200

Figure B-2: Sequence Of the two PCR primers used to Clone the cycA
promotor from Fl. capsu/ata. ............................. 200

Figure 8-3: Circular restriction map Of pCHBSOO .................... 201

xiii

LIST OF ABBREVIATIONS

ADP Adenosine diphosphate

APS Adenosine phosphO-sulfate
ATP Adenosine triphosphate

CAD Collision activated dissociation
CDP Cytosine diphosphate

DGD Digalactosyl diacylglycerol
EDTA (Ethylenedinitrilo)tetraacetic aid
FAB Fast atom bombardement

FAM Fatty acid methylester

GDP Guanosine diphosphate

MS Mass spectrometry

M-H Molecular ion minus one proton
PAPS Phosphoadenosine phosphosulfate
pfu plaque-forming unit

TLC Thin layer Chromatography

TLE Thin layer electrophoresis

xiv

 

CHAPTER 1

INTRODUCTION

DISCOVERY AND STRUCTURE OF THE PLANT SULFOLIPID
Approximately thirty years ago, Benson and collaborators discovered
that 35804-labelled extracts from various higher plants, algae, and a purple
non-sulfur bacterium harbored large amounts of a sulfur containing lipid (2).
They determined the structure to be 1’-O-(6’-deoxy-a-D-glucosyl-6’-sulfonic
acid)-3-O- diacylglycerol (3). The 6-deoxy-6-sulfO-glucose is also known as

sulfoquinovose (Figure 1-1).

0
(”2503' cnzoym
o O
HOsz
OH
HO O—CHz
OH

6-Sulpho- a-D-Quinovosyl Diacylglycerol
(Plant Sulpholipid)

Figure 1-1: Structure Of the plant sulfolipid, 6-sulfo-a-D-quinovosyl diacyl-
glycerol.

 

2

Since the elucidation of the structure Of the sulfolipid, most studies
concerning the function and mechanism Of synthesis have been carried out
with higher plants. Thus, much Of the discussion that follows concerns the
plant work Since virtually nothing has been done with bacteria. The two major
fatty acids found in sulfolipid in higher plants are palmitic acid and linolenic
acid (4). In Rhodobacter sphaeroides, a purple non-sulfur bacterium, the place
Of linolenic acid is taken by vaccenic acid, a monounsaturated 18-carbon fatty
acid typically found in bacteria (40). Positional analysis of the fatty acid
composition of sulfolipid in higher plants reveals an interesting difference in
comparison tO the galactolipid MGD, the most abundant chloroplast lipid. In
Spinach, the predominant 16-Carbon fatty acid found in MGD is
hexadecatrienoic acid, which is confined to the sn-2 position of the
diacylglycerol moiety, while the most abundant 16-Carbon fatty acid in spinach
sulfolipid is palmitic acid, which can be found in the sn-2 as well as in the sn-1
position Of the glycerol backbone. This unique fatty acid composition Of
sulfolipid raises a question about the origin of its diacylglycerol moiety. It is
generally accepted that in some higher plants, including spinach, the
diacylglycerol moiety Of the galactolipids is derived from two distinctive pools
as a result of the eucaryotic pathway, located in the endoplasmatic reticulum,
and the procaryotic pathway, Operating inside the chloroplast (33). Due to
different substrate specificities Of the acyltransferases belonging to one or the
other pathway, diacylglycerol derived from the Chloroplast pOOl contains 16-

carbon fatty acids predominantly in the sn-2 position, while those molecular

3

species derived from the endoplasmatic reticulum pool contain 18-carbon fatty
acids in the sn-2 position. Higher plants, such as cucumber, which have a
reduced procaryotic pathway, virtually lack galactolipids with 16—carbon fatty
acids in the sn-2 position of the glycerol backbone. Therefore, by analysis Of
the fatty acids found in the sn-2 position of the chloroplast lipids, the presence
of the two pathways in a particular plant species, as well as their contribution
tO the synthesis of each lipid, can be estimated. In analogy to the
galactolipids, attempts have been made to determine the origin Of the
diacylglycerol found in sulfolipid, isolated from various higher plants, by
positional analysis Of the fatty acid composition (4). Based on these data, it
was concluded that in spinach the majority Of the diacylglycerol in sulfolipid is
derived from the chloroplast pool. However, the molecular mechanisms that
determine the unique fatty acid composition Of each chloroplast lipid are not

understood.

OCCURRENCE AND DISTRIBUTION OF THE PLANT SULFOUPID
Sulfoquinovosyl diacylglycerol occurs in all higher plants, ferns and
mosses. Although it seems that the sulfolipid is mainly associated with
photosynthetically active membranes, it can be found in small amounts (pre-
sumably in plastid membranes) in nonphotosynthetic tissue, such as roots
(38), potato tubers (7), or apples (8). It is also present in substantial amounts
in etiolated tissue (39). In addition it has been found in various algae such as

hemofiagellata, phytoflagellata, red algae, brown algae, and green algae. The

4

sulfolipid has also been reported to be a constituent of cyanobacteria and
some species of purple non-sulfur bacteria. A compilation Of the distribution Of
the sulfolipid in plants and bacteria has been presented by Haines (11).

The purple non-sulfur bacteria can be divided into two groups, depend-
ing on whether or not they contain sulfoquinovose diacylglycerol. Species
known to contain the sulfolipid are Fihodobacter sphaeroides (40) and Rhodo-
microbium vannielii (29). Benson et al. also reported the sulfolipid to be
present in labelled extracts from Rhodospiri/Ium rubrum (2). This finding could
not be confirmed by Wood at al. (40), who were also not able to measure
detectable levels Of the sulfolipid in Ft. gelatinosus, Ff. capsu/atus, and H.
palustris. However, Wood et al. did not use isotope labelling experiments and
might have missed trace amounts of sulfolipid in these bacteria (11). Thus, it is
not certain whether all photosynthetic organisms have the sulfolipid. However,
no organism with oxygenic photosynthesis has been found to lack it.

Various photosynthetic organisms have been compared with respect to
their sulfolipid content in a careful study by Radunz (31). The amounts,
measured in percent Of ether-extractable lipids, vary from 2.6% in
Ft. sphaeroides Up to 18.3% in the brown alga Fucus vesicu/osus. Leaves Of
higher plants usually contain around 5% sulfolipid. Within photosynthetic tissue
sulfolipid is found exclusively in the Chloroplast membranes (28). The detailed
subchloroplastic localization Of the sulfolipid remains unclear. It has been
found in association with the granal Iamellae Of the Chloroplast as well as with

the envelope of the chloroplast (6). In a study by Radunz and Berzborn (32),

5

who employed polyclonal antibodies against Fucus sulfolipid, it was found that
the antibodies, which were also reactive against Antirrhinum sulfolipid, were
able to agglutinate the lamellar membranes isolated from Antirrhinum
chloroplast only indirectly, while mechanically disrupted Iamellae could be
directly agglutinated. However, the conclusion drawn by the authors, based
on these data, that the sulfolipid is predominantly located on the inner surface
of the thylakoids needs to be supported by more data. The intracellular
localization of sulfoquinovosyl diacylglycerol in the bacterium R. sphaeroides
has not been investigated. It would be interesting to know whether the
occurrence of the sulfolipid is confined to the intracellular membranes that
carry the photosynthetic apparatus, or whether It can also be found in the
plasma membrane or even in the outer membrane.

In a photosynthetically active leaf of a higher plant, approximately the
same amounts Of sulfur are bound in sulfolipid as in proteins (14). This implies
that a major proportion of organic sulfur on earth exists as sulfolipid, which
therefore plays an important role in the global sulfur cycle. In view of the
abundance of this compound in plant membranes, it is a quantitatively
important component of human and animal diets. Various aspects of

degradation of plant sulfolipid have been reviewed by Harvvood (15).

PROPOSED BIOSYNTHEITC PATHWAYS FOR THE PLANT SULFOUPID
Although the plant sulfolipid was discovered over 30 years ago, the

biosynthetic pathway remains unknown. Studying sulfur metabolism in general

6

seems to be very complicated, due to the diversity and reactivity Of sulfur
containing compounds. The biosynthesis Of the sulfoquinovose head group
brings together two plant-specific processes: first, photosynthetic sugar
biosynthesis and, second, photosynthetic sulfur assimilation. At least by now,
we have quite a deep understanding Of sugar biosynthesis in plants, and
many Of the ideas about sulfolipid biosynthesis were derived from our
knowledge about this process. The elucidation Of the pathway for sugar
biosynthesis took many years and was favored by the fact that most of the
intermediates are highly abundant in photosynthetically active leaves.
Intermediates Of sulfolipid biosynthesis, on the other hand, are not quite as
abundant.

Several possible pathways for the biosynthesis Of sulfoquinovosyl
diacylglycerol have been proposed on the basis Of "biosynthetic parsimony"
rather than the availability Of any direct evidence (15, 27). A summary Of the
most popular ideas is shown in Figure 1-2. The experimental evidence on
which Fig. 1-2 is based can be summarized as follows:

Shibuya et al. (34) discovered various sulfoniC acids in extracts from
35S—sulfate-labelled Chlorella cells, including sulfoacetic acid, sulfolactic acid,
cysteic acid, and a nucleoside diphosphate sulfoquinovose. Later, in an in vivo
labelling study employing Eug/ena graci/is, Davis et al. (5) showed that
labelled sulfate and L-cysteic acid were incorporated by this organism into
sulfolipid. However, no methods were employed in these studies to prevent

incorporation from degradation products of cysteic acid. In addition,

PROPOSED PATHWAYS
FOR THE BIOSYNTHESIS OF SQD

Sulfate
APS \w‘\’
I SulfIte
PAPS ——-———/
pEp , " ‘2 O-Actylserine
2-Phospho-3l-Sulfolactate Cysteic Acid

3-Sulfo-L-Lactate +————3-Sulfopyruvate
s

4
3-Sulfo-D-Lactaldehyde

DHAP —\ 5
6-Sulfo-6-Deoxy-FFIIctose-1-P hosphate

 

UDP-Sulfoquinovose

meal '

800

Figure 1-2: Proposed pathways for the biosynthesis of the plant sulfolipid
(adopted from 15). The numbers refer to reactions mentioned in the text.

 

8

molybdate, which is known to inhibit ATP-sulfurylase, reduced the
incorporation of labelled sulfate into sulfolipid, suggesting that adenosyl-1-
phosphosulfate (APS) or 3-phosphoadenosyl-1-phosphosulfate (PAPS) might
be precursors of sulfolipid biosynthesis. Based on these data, Davis et al. (5)
outlined a pathway (Fig.1-2, steps 1-7) starting with the reaction Of phospho-
enolpyruvate and PAPS leading tO the formation Of 3-L—sulfolactic acid, which
might be converted to 3-sulfO-D-lactaldehyde via a proposed epimerase and a
dehydrogenase. Sulfolactaldehyde and dihydroxyacetone phosphate could
then condense via an aldolase reaction, forming 6—sulfO-6-deoxy-D-fructose-1-
phosphate, which might be a direct precursor of 6-sulfoquinovose. It was first
suggested by Benson (3) that activated 6-sulfoquinovose in the form of a
nucleoside diphosphate sulfoquinovose and diacylglycerol might react to give
sulfoquinovose diacylglycerol (Figure 1-2, step 7). The pathway outlined
above has some similarity with glycolysis, and therefore the term
"sulfoglycolytic pathway" was coined by Benson (3), implying that the sulfonic
acid group is equivalent to a phosphate ester. This could be misleading, Since
the sulfonic acid group might be more similar to a carboxyl group, as pointed
out by Mudd et al. (27).

Davies et al. (5) assumed that the incorporation of cysteic acid intO
sulfolipid by Euglena graci/is was due tO a side reaction. Haines proposed that
cysteic acid feeds into the pathway on the level Of 3-sulfO-L-lactic acid (steps 8
and 9, Figure 1-2) via a transaminase reaction (11). He suggested some

alternative, initial reactions that would give cysteic acid a more central role in

9
the pathway (Figure 1-2, steps 10,11, and 12). In that case, PAPS or APS

would be the precursor Of sulfite, which could displace O-acetyl from O-acetyl
serine analogous to the reactions of thiOlS with O-acetyl serine (18). The
proposal that cysteic acid is a direct precursor Of sulfolipid biosynthesis has
been controversial. Harwood reported that labeled cysteic acid can be
incorporated into sulfolipid by intact alfalfa plants (13), while Mudd et al. could
not find evidence for incorporation of labeled cysteic acid into sulfolipid by
spinach seedlings (26).

Isolated chloroplasts from spinach are capable of incorporating
35S-SUlfate into sulfoquinovosyl diacylglycerol, as was first Shown by Haas et
al. (10). Rates Of 1.7 meI/mg chlorophyll per hour were Observed in the light.
Later, Kleppinger-Sparace et al. demonstrated rates Of up to 700 pmol/ mg
chlorophyll per hour for the incorporation of 3"S-sulfate into sulfolipid by
isolated spinach chloroplast in the light, which approached the rates Observed
for phosphatidyl glycerol synthesis in the same system (21). The incorporation
of “S-sulfate into sulfolipid was Observed in the dark, when the isolated
spinach chloroplast was supplied with ATP either directly, or indirectly via the
triose phosphate shuttle, which provided dihydroxyacetone phosphate as the
substrate for ATP biosynthesis inside the chloroplast (22). The general
conclusion drawn from these results is that all the enzymes required for the
biosynthesis of the sulfoquinovose head group are located inside the chloro-
plast. Kleppinger-Sparace et al. (21) measured the effect Of various

compounds on the incorporation of 35S-sulfate into sulfolipid by intact spinach

10

chloroplasts in the light. ATP slightly increased the incorporation, while the
addition of cysteic acid, cysteine, and other sulfhydryl compounds had no
effect. In labelling experiments, using 1“C-acetate as a substrate for sulfolipid
biosynthesis in isolated spinach chloroplasts, Joyard et al. (20) showed that
these chloroplasts were capable Of synthesizing the diacylglycerol moiety as
well as the sulfoquinovose head group de novo. Positional analysis of the fatty
acid composition Of the labeled sulfolipid revealed that palmitic acid was
preferentially located in the sn-2 position Of the glycerol backbone, typical for
sulfolipid synthesized by the procaryotic pathway inside the chloroplast (see
above; and reference 4). During the same study it was found that
UDP-galactose, the precursor Of galactolipid biosynthesis, strongly inhibited
the incorporation of labelled sulfate into the sulfolipid. This has been
interpreted as competition by UDP-galactose with a nucleoside diphosphate
sulfoquinovose for diacylglycerol. Labelling of isolated chloroplasts with 1“C-
acetate in the presence Of UDP-galactose also resulted in a slower rate Of
incorporation Of label into sulfolipid as compared to the controls without UDP-
galactose, while at the same time the biosynthesis of galactolipids was
Stimulated (20).

Direct evidence for the involvement Of a nucleoside diphosphate as the
last precursor Of sulfolipid biosynthesis, has recently been obtained by Heinz
et al., who synthesized 6-sulfoquinovosyl-1-phosphate and UDP-
sulfoquinovose (16, 17). Chloroplast envelopes were prelabeled with 1‘0-

glycerol to form radioactive diacylglycerol. Addition Of UDP-sulfoquinovose

11

stimulated the incorporation of label into sulfolipid. The GDP derivative were
only half as active as UDP-sulfoquinovose, and the ADP and GDP derivatives
were inactive.

Unfortunately, chloroplasts have several limitations as a system for
investigating the earlier steps in sulfolipid biosynthesis. First, intact
chloroplasts are quite selective in their uptake of small molecules, making
precursor and inhibitor studies quite difficult or impossible; second, it has not
been possible to observe the biosynthesis Of sulfolipid from labeled sulfate or
PAPS in preparations of lysed chloroplasts (B. Mudd, personal
communication). However, Hoppe and Schwenn (19) reported the incorpo-
ration Of labeled sulfate, sulfite, and PAPS into sulfoquinovosyl diacylglycerol
by cell-free extracts from Chlamydomonas reinhardii. Incorporation Of labelled
sulfate required ATP and magnesium, and was stimulated by the addition Of
glutathione, while addition of cysteic acid did not affect the incorporation.
PAPS behaved as a legitimate substrate, since the reaction followed a
saturation curve with increasing amounts of PAPS. On the other hand, the
incorporation of sulfite into sulfolipid could not be saturated, even with very
high amounts of sulfite, suggesting a non-enzymatic artifact.

In conclusion, the biosynthesis of sulfoquinovose diacylglycerol
probably starts with activated sulfate, either APS or PAPS. It seems to be fairly
clear that the last step involves the reaction between UDP-sulfoqulnovose and
diacylglycerol to yield the sulfolipid. Every other step in between (see Figure

1-2) is not sufficiently supported by experimental evidence to allow any

 

12

conclusions to be drawn. It will be very interesting to see at what level the
sulfonate group is introduced into the molecule. It is entirely possible that the

pathways outlined in Figure 1-2 are wrong.

PROPOSED FUNCTION OF THE PLANT SULFOUPID

The exclusive occurrence of sulfoquinovosyl diacylglycerol in
photosynthetic organisms and the unique structure of the sulfonic acid head
group, suggests that this compound might play a critical role in
photosynthesis (for reviews see ref. 1, 15, 27). This idea is supported by a
study in which correlations were Observed between the amounts Of sulfolipid
and the levels Of chlorophyll in Chlore/la protothecoides during cycles of
bleaching and greening, induced by changes in the growth conditions (35).
Cells grown in the light, in medium with high urea and low glucose content,
were green and contained high amounts of chlorophyll and sulfolipid , while
cells grown in medium rich in glucose and poor in urea were bleached,
independent of the illumination, resulting in low chlorophyll and sulfolipid
contents. Similarly, Sinensky was able to lower the levels of sulfolipid in
Chlore/la pyrenoidosa by growing the cells on cysteine as the sole sulfur
source (37). The cells retained the ability to evolve oxygen, but were unable to
fix carbon dioxide, and triacylglycerols accumulated to high levels. However, a
specific function for the plant sulfolipid cannot be inferred from these
experiments, because it is impossible to evaluate the importance Of the

non-specific side effects of the growth conditions. A correlation between the

13

amount Of sulfolipid and photosynthetic activity was also Observed in a study
employing the different developmental stages of plastids along maize leaves.
The levels of chlorophyll and sulfolipid were found to increase in parallel with
increased maturation of the plastids (24). Correlations have also been
Observed in higher plants between increased levels of sulfolipid and cold
hardiness (23). Unfortunately, such correlative approaches seem suggestive at
best, since many cellular constituents may differ as a result of the contrasting
treatments.

Reconstitution experiments employing protein complexes common to
the thylakoid membrane have been used to examine the function of the
sulfolipid. It has been found in very tight association with isolated
photosystem II apoproteins (9, 36). Sulfolipid was also present in
stoichiometric amounts in highly purified preparations of the ATPase CFO-CF1
complex (30). Removal Of the tightly bound sulfolipid from the ATPase
complex resulted in loss of activity. Enzymatic activity was restored by
readdition of sulfoquinovosyl diacylglycerol and other chloroplast lipids in a
fixed ratio. It was concluded that the sulfolipid is either required as boundary
lipid, or that it might even participate directly in the enzymatic reaction (1).

Haines proposed a model in which the charged, anionic head groups
of phospholipids, or the sulfated lipids of Ochromonas, could act as a proton
conducting pathway in proton pumping membranes (12). This model implies
that sulfolipid located in the inner leaflet of the thylakoid membrane might play

an essential and direct role in photosynthetic electron transport. It has been

14

shown that antibodies, raised against sulfolipid, can inhibit photosynthetic
electron transport (25). The inhibition was Observed for a photosystem I
mediated reaction between the electron donor 2,6-dichlorophenol indophenol
and the electron acceptor methylviologen. However, the pH-dependence for
the inhibitory reaction correlated with the pH-dependence for the interaction of
the sulfolipid with polypeptides in vitro, and it was concluded that the
antibodies indirectly affected the function Of proteins Of the electron transport
chain, to which sulfolipid was tightly bound.

In summary, although several experiments suggest a role for the plant
sulfolipid in photosynthesis, the evidence is indirect and not sufficient to

propose a specific model for its function.

ABOUT THE PRESENTED WORK

Although sulfolipid is one Of the most abundant sulfur-containing mole-
cules in the biosphere, the biosynthetic pathway is not known. Because it has
not been possible to identify intermediates in the pathway between activated
sulfate in the form of APS or PAPS and sulfoquinovose diacylglycerol by
conventional biochemical approaches, a genetic approach to this thirty-year-
Old problem has been applied. The first step Of this approach is to isolate
mutants deficient In sulfolipid biosynthesis in the purple non-sulfur bacterium
Hhodobacter sphaeroides. Specific blocks in the pathway in different mutants
provide the tools to dissect the sequence Of reactions in the formation of the

sulfolipid. Bacterial mutants allow the isolation Of the respective genes by

15

complementation. Mutants may serve as a source of precursors, which build
up due to a specific metabolic block in the pathway. In addition, these
mutants may help to determine the function of the plant sulfolipid.

In Chapter 2 the identity Of purified sulfolipid from spinach and R.
sphaeroides will be tested by mass spectrometry, since the occurrence Of
authentic plant sulfolipid in R. sphaeroides is critical for the described work
and its presumptive relevance to higher plants. In Chapter 3, the isolation of
mutants deficient in plant sulfolipid as well as their basic characteristics will be
described. Chapters 4 and 5 provide information about the genetic
complementation of three sulfolipid mutants and a detailed analysis of two Of
the genes directly involved in the biosynthesis of the plant sulfolipid in Ft.
sphaeroides. The versatility Of this purple bacterium for the study Of sulfolipid
biosynthesis is further demonstrated in Chapter 6, where an in vitro system for
the biosynthesis of sulfolipid in extracts of R. sphaeroides is described. Future
prospects for using the genetic and biochemical tools provided in the
presented work, to solve the questions about the biosynthesis and function of

the plant sulfolipid, will be discussed in Chapter 7.

REFERENCES
1. Barber J., Gounaris K. (1986). What role does sulpholipid play within
thylakoid membranes? Photosynthesis Res. 9:239-249

2. Benson A.A., Daniel H., WIser R. (1959). A sulfolipid in plants. Proc.
Natl. Acad. Sci. USA 45:1582-1587

10.

11.

12.

13.

14.

15.

16
Benson AA. (1963). The plant sulfolipid. Adv. Lipid Res. 1:387-394

Bishop D.G., Sparace S.A., Mudd B.J. (1985). Biosynthesis of sulfo-
quinovosyldiacylglycerol in higher plants: The origin of the
diacylglycerol moiety. Arch. Biochem. Biophys. 240:851-858

Davies W.H., Mercer E.I., Goodwin T.W. (1966). Some Observations on
the biosynthesis Of the plant sulfolipid by Euglena gracilis. Biochem J.
98:369-373

Douce R., Holtz R.B., Benson AA. (1973). Isolation and properties of
the envelope of spinach chloroplast. J. Biol. Chem. 248:7215-7222

Galliard T. (1968). Aspects of lipid metabolism in higher plants-1:
Identification and quantitative determination Of the lipids in potato
tubers. Phytochem. 7:1907-1914

Galliard T. (1968). Aspects of lipid metabolism in higher plants-2: The
identification and quantitative analysis Of lipids from the pulp of pre- and
post-climacteric apples. Phytochem. 7:1915-1922

Gounaris K., Barber J. (1985) Isolation and characterization Of a photo-
system ll reaction center lipoprotein complex. FEBS Letter 188:68-72

Haas R., Siebertz H.P., Wrage K., Heinz E. (1980). Localization Of
sulfolipid labeling within cells and chloroplast. Planta 148:238-244

Haines TH (1973). Sulfolipids and halosulfolipids. In Lipids and Bio-
membranes of Eukaryotic Microorganisms, Ed Erwin J.A., Academic
Press, NY, pp 197-232

Haines TH. (1982). Anionic lipid head groups as a proton conducting-
pathway along the surface of membranes: a hypothesis. Proc. Natl.
Acad. Sci. USA 80:160-164

Harwood J.L. (1975). Synthesis Of sulfoquinovosyl diacylglycerol by
higher plants. Biochim. Biophys. Acta 398:224-230

Harwood J.L., Nicholls R.G. (1979). The Plant sulfolipid - a major
component of the sulphur cycle. Biochem. Soc. Trans. 72440-447

Harwood J.L. (1980). Sulfolipids. In The Biochemistry of Plants, Vol 4,
Ed Stumpf P.K., Academic Press, NY, pp 301-320

16.

17.

18.

19.

20.

21.

23.

24.

25.

26.

17

Heinz E., Schmidt H., Hoch M., Jung K.H, Binder H., Schmidt RR.
(1989). Synthesis Of different nucleoside 5’-diphospho-sulfoquinovoses
and their use for studies on sulfolipid biosynthesis in chloroplast. Eur. J.
Biochem. 184:445-453

Hoch M., Heinz E., Schmidt RR. (1989). Synthesis of 6-deoxy-6-sulfO-a-
D-glucopyranosyl phosphate. Carbohyd. Res. 191 :21 -28

Hodson R.G., Schiff J.A., Mather JP. (1971). Studies on the sulfate
utilization by algae. Plant Physiol. 47:306-311

Hoppe W., Schwenn JD. (1981). In vitro biosynthesis Of the plant
sulpholipid: on the origin Of the sulphonate head group. Z.
Naturforsch. 360:820-826

Joyard J., Blee E., Douce R. (1986). Sulfolipid synthesis from 35802',
and [1-“C]acetate in isolated spinach chloroplast. Biochim. Biophys.
Acta 879:78-87

Kleppinger-Sparace K.F., Mudd B.J., Bishop D.G. (1985). Biosynthesis
of sulfoquinovosyl diacylglycerol in higher plants: the incorporation of
3PS0, by intact chloroplast. Arch. Biochem. Biophys. 240:859-865

Kleppinger-Sparace K.F., Mudd J.B. (1987). Biosynthesis of sulfo-
quinovosyldiacylglycerol in higher plants. The incorporation of 3”SO, by
intact chloroplast in darkness. Plant Physiol. 84:682-687

Kuiper P.J.C. (1970). Lipids in Alfalfa leaves in relation to cold
hardiness. Plant Physiol. 45:684-686

Leech R.M., Rumsby M.G., Thomson WW. (1973). Plastid differentiation,
acyl lipid, and fatty acid changes in developing green maize leaves.
Plant Physiol. 52:240-245

Menke W., Radunz A., Schmid G.H., Koenig F., Hirtz RD. (1976). Inter-
molecular interactions of polypeptides and lipids in the thylakoid mem-
brane. Z. Naturforsch. 31cz436-444

Mudd J.B., Dezacks R., Smith J. (1980). Studies on the biosynthesis Of
sulfoquinovosyldiacylglycerol in higher plants. In Biogenesis and Func-
tion of Plant Lipids, Eds Mazliak P., Beneviste P., Costes C., Douce R.,
Elsevier, Amsterdam pp 57-66

 

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

18

Mudd B.J., Kleppinger-Sparace K.F. (1987). Sulfolipids. In The Biochem-
istry of Plants, Vol 9, Eds Stumpf PK and Conn E.E., Academic Press,
NY, pp 159-173

Ongun A., Thomson W.W., Mudd J.B. (1968). Lipid composition of
chloroplast isolated by aqueous and nonaqueous techniques. J. Lipid
Res. 92409-415

Park 0., Berger LR, (1967). Complex lipids Of Rhodomicrobium
vannie/ii. J. Bacteriol. 93:221-229

Pick U., Weiss M., Gounaris K., Barber J. (1987). The role of different
thylakoid glycolipids in the function Of reconstituted chloroplast ATP
synthase. Biochim. Biophys. Acta 891 :28-39

Radunz A. (1969). Ueber das Sulfochinovosyl-diacylglycerin aus
hoeheren Pflanzen, Algen und Purpurbakterien. Hoppe-Seylers Z.
Physiol. Chem. 350241 1-417

Radunz A., Berzborn R. (1970). Antibodies against sulfoquinovosyl-
diacylglycerol and their reactions with chloroplast. Z. Naturforsch.
25b:412-419

Roughan G., Slack, R. (1984). Glycerolipid synthesis in leaves.
Trends Biochem. Sci. 92383-386

Shibuya l., Yagi T., Benson AA. (1963). Sulfonic acids in Algae. In
Microalgae and Photosynthetic Bacteria, Ed Jpn. SOC. Plant Physiol.,
Univ Of Tokyo Press, Tokyo, pp 627-636

Shibuya |., Hase E. (1965). Degradation and formation of sulfolipid
occurring concurrently with de- and regeneration Of chloroplast in the
cells Of Chlore/la protothecoides Plant Cell Physiol. 6: 267-283

Sigrist M., Zwillenberg C., Giroud C.H., Eichenberger W., Boschetti A.
(1988). Sulfolipid associated with the light harvesting complex
associated with photosystem II apoproteins Of Chlamydomonas
reinhardir). Plant Science 58:15-23

Sinensky M. (1977). Specific deficit in the synthesis Of 6-sulfoquinovosyl
diglyceride in Chlore/Ia pyrenoidosa. J. Bacteriol. 129:516-524.

Stuiver C.E.E., Kuiper P.J.C., Marschner H. (1978). Lipids from barley
and sugar beet in relation to salt resistance. Physiol. Plant. 42:124-128

39.

40.

19

Tevini M. (1977). Light, function and lipids during plastid development.
In Lipids and Lipid Polymers in Higher Plants, Eds Tevini M. and
Lichtenthaler H.K, Springer Verlag, Berlin pp 121-145

Wood B.J.B., Nichols B.W., James A.T. (1965). The lipids and
fatty acid metabolism of photosynthetic bacteria. Biochim.
Biophys. Acta 106:261-273

 

CHAPTER 2
COMPARISON OF SULFOLIPID FROM SPINACH AND Ff. SPHAEROIDES BY
MASS SPECTROMETRY

ABSTRACT

To test the identity of sulfolipid isolated from the purple non-sulfur
bacterium R. sphaeroides with the plant sulfolipid sulfoquinovosyl
diacylglycerol, the structures Of sulfolipid from spinach and Ff. sphaeroides
were compared by negative mode FAB-MS and FAB-CAD-MS/MS tandem
mass Spectrometry. Utilizing these techniques, it was demonstrated that the
sulfur containing head group from the plant and bacterial lipids produced
identical fragmentation patterns in agreement with sulfoquinovose. However, it
should be mentionend that the employed method did not allow to distinguish
between various stereo isomeres. The overall differences in the samples were
based on different fatty acid composition of the sulfolipid isolated from the two
organisms. Fragmentation pathways were postulated, to interpret head group
specific fragments found at m/z 299, 283, 241, 225, 161 and 80 in the two
samples. The two predominant molecular species Of Spinach sulfolipid
contained either palmitic and linolenic acids (m/z 815) or two molecules of
linolenic acid (m/z 837), while the two predominant species of bacterial
sulfolipid contained palmitic and vaccenic acids (m/z 819) or stearic and

vaccenic acids (m/z 847).

20

21
INTRODUCTION

The plant sulfolipid was discovered by Benson et al. (1), who also
provided the first structural data about it. By chromatography of the hydrolysis
products of uniformly 14C-labeled sulfolipid, it was established that the
molecule contained glycerol, two fatty acids and a 6-carbon compound, which
could also be labeled with 3E5S-sulfate. The sulfur containing, 6-carbon
compound was recognized as a sulfonic acid by its complete resistance to
acid hydrolysis. The presence of hydroxyl groups, as well as its reaction with
2,4-dinitrophenyl- hydrazine suggested a aldohexose structure. By analogy to
the galactolipids, and based on the fact that the glycosidic linkage was not
cleaved by a-galactosidase from yeast, a structure was initially proposed, that
featured a basic diacylglycerol structure, with a 6-sulfo-6-deoxy-galactose
head group linked in a 143-glycosidic linkage to the glycerol moiety (1).
However, further experiments, which included elemental analysis, proton
magnetic resonance spectrometry, infrared spectrometry, Optical rotary
dispersion Spectrometry, and X-ray crystallography of the sulfolipid from
Chlorella and alfalfa, required modifications of the originally proposed
structure. The sulfolipid was thus correctly recognized as 6-sulfO-a-D-
quinovosyl diacylglycerol (Figure 1-1). Benson summarized the data on which
the final structural elucidation of the plant sulfolipid was based in a review (2).

The occurrence Of the plant sulfolipid in the purple bacterium
R. sphaeroides was originally demonstrated by Wood et al. (8), and later by

Radunz (7). Both groups identified the sulfolipid by cochromatography with

22
standards. This finding Offered the opportunity to study the biosynthesis of the

plant sulfolipid in a bacterial system. However, since it was crucial to the study
that the bacterial and the plant sulfolipid were both sulfoquinovosyl
diacylglycerol, sulfolipids isolated from spinach and R. sphaeroides were
compared by mass spectrometry to confirm their identity. This technique is
very sensitive, requiring minimal amounts of substance (1 nmol, equivalent to
0.1-1.0 pg depending on the molecular weight) to be analyzed. Fast Atom
Bombardment (FAB) employing a high energy (6 keV) Xenon atom beam is
used to desorb and ionize nonvolatile analytes from a matrix. The ions
produced are then accelerated in an electric field (10 kV), passed through an
electric sector (E), which acts to narrow the energy spread of the accelerated
ions, and then mass analyzed by their momentum in a magnetic field (B). TO
acquire a spectrum the magnetic field strength is scanned so that ions Of
different momenta are sequentially allowed to reach the detector. In the case
Of sulfolipids the instrument is run in the negative ion mode, since these
compounds readily form anions. The FAB ionization method is very "soft",
resulting in a relatively low amount of fragmentation and a high percentage of
molecular ions ([M-H]'), although sulfolipids do fragment tO some extent. The
negative mass spectrum of the sulfolipids displays prominent peaks,
representing the different molecular species present in the sample.

For structural information other than molecular weight, diagnostic
fragmentation patterns are Observed in a FAB-CAD-MS/MS experiment, which

is based on an extension of the technique described above. This allows

23

"daughter ions" of one particular "parent ion" to be selectively observed. To
generate "daughter ions", a helium gas filled chamber (0.1 Torr) is placed in
the first field-free region of the instrument resulting in the collision of the
accelerated ions with helium gas atoms. Through "Collision Activated
Dissociation" (CAD), "daughter ions" are generated. Fragments or “daughter
ions" created in this process are formed with an energy that is proportional to
their mass in a linear fashion. The value of the energy/mass ratio is dependent
upon the mass of the "parent ion". To detect "daughter ions" of a particular
"parent ion", the instrument is scanned at a constant ratio of B/E (magnetic
field Strength/electrical field strength in the electrical sector), typical for the
respective parent ion, permitting the observation of only those "daughter ions"
generated from that particular "parent ion" without interference from matrix
related ions or ions from other compounds in the sample (for more detailed
physical theory, see reference 4).

In this chapter, a diagnostic fragmentation pattern for sulfoquinovosyl
diacylglycerol will be postulated, and the identity of the plant and bacterial

sulfolipid will be confirmed by the MS/MS techniques described above.

MATERIALS AND METHODS

Plant and bacterial material. Spinach leaves were obtained from a local
grocery store. The wild type R. sphaeroides strain 2.4.1 used, was kindly
provided by Dr. Uffen’s lab at Michigan State University. Bacterial 4-l cultures

were grown photoheterotrophically in Sistrom’s medium, as will be described

 

24

in detail in the methods section of chapter 3.

Large scale preparation of sulfolipid. Spinach leaves (500 g) were
blended in 1.2 l of chloroform-methanol-formic acid (10:10:1, v/v). Solids were
separated by filtration and reextracted with 440 ml chloroform-methanol-water
(5:5:1, v/v). The two combined filtrates were extracted with 600 ml 1 M KCI,
0.2 M phosphoric acid. The dried chloroform phase yielded about 4.5 g of
lipids. This material was purified by column chromatography according to
O’Brien and Benson (6). Similarly, sulfolipid was extracted and purified from 20
g of R. sphaeroides (2.4.1) cells, except that volumes were cut in half. As a
final purification step, TLC on ammonium sulfate plates was employed to
remove contaminating phosphatidyl glycerol. (The TLC system and its
characterization will be described in detail in chapter 3). The sulfolipid band
was scraped off the plate after staining with iodine vapor, and the material was
eluted using chloroform-methanol (1:1, v/v). In addition, this TLC system was
used to monitor the purity of the sulfolipid throughout the purification
procedure.

Mass spectrometry. Approximately 10 pg of the sulfolipid sample
dissolved in chloroform/methanol (1:1, v/v) were mixed with 2 pl of
triethanolamine matrix on the FAB probe tip. Ions were produced by
bombardment with a beam of Xe atoms (6 keV) or Cs+ ions (8 keV) in a JEOL
HX-110 double focusing mass spectrometer (EB) configuration operating in
the negative ion mode; no differences were observed in spectra acquired by

neutral beam (Xe°) or charged cesium beam (Cst) bombardment. The

 

25

accelerating voltage was 10 kV and the resolution was set at 1000. For FAB-
CAD-MS/MS, helium was used as the collision gas in a cell located in the first
field-free region. The helium pressure was adjusted to reduce the abundance
of the parent ion by 50%. A JEOL DA-5000 data system generated linked
scans at a constant ratio of magnetic to electrical fields (B/E). The instrument
was scanned in 2 minutes from m/z 0-1500. Data presented were acquired in

a single scan and were found to be reproducible.

RESULTS AND DISCUSSION

Comparison of molecular species of sulfolipid found In samples of
spinach and R. sphaemldes. Negative FAB mass scans made it possible to
detect the most abundant molecular species contained in the two sulfolipid
samples obtained from spinach and Ft. sphaeroides. Two predominant
molecular species were found in the spinach sample at m/z 815 and 837,
which corresponded to sulfoquinovosyl diacylglycerol ([M-H]‘) containing
palmitic and linolenic acids, or two linolenic acids respectively. This result was
in agreement with data obtained by conventional analysis of the fatty acid
composition of Spinach sulfolipid (3). Loss of one fatty acid in the form of a
ketene (R=C=O) resulted in ions of m/z 555 and 577, depending on which
fatty acid was lost from which molecular ion (Figure 2-1, top). The equivalent
negative FAB mass scan for bacterial sulfolipid showed two predominant
species of sulfoquinovosyl diacylglycerol at m/z 847 and 819 ( Figure 21,

bottom), which corresponded to two molecular species containing either

 

 

26

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

100 1
815 [M-H]
$3123 16:0/169
‘ ' Spinach-8L
80-
1 .
60- ”[503]
‘ 2 -
1 225 241 837 [M-l-I)
g; 40 « 16:311623
V 1
L” 1 1 1 - 1 2 .
O 204 [M-H-R-C-O] 555 577 [M-H-R-C-O]
. 2 1 -
E q [M-H-R-C-Oj
D
Z 0‘ ' *— T T T T fi 1 f fl T ' v '
:3 o 200 400 600 600 1
m
< 50
. [SOJ' . 1 .
I; 1 80 “9 "5“" R. sphaerOIdes-SL 847 [NI-HI
— ‘ 18:0]18.1
'_ 4o 4
4
...l 1
1
LIJ so 4 2 ,
CC I [M-H] 819
‘ 1610/18:1
.I
I 2 2 ' R1: 18:0
20 -1 241 W'H'R'C‘OI R2: 16:0
‘ R3: 18:1
225 , , _
[M-I-I-R-C-O]
‘ 581
‘0‘ 2 3 -
1 [M-H-R-C-Oj 555
0"
0 200 400 600 600 1

Figure 2-1: Negative FAB mass spectra of sulfolipid samples from spinach
(top) and R. sphaeroides (bottom). R, residues characteristic for various fat
acids. [M-H]', molecular ion following the loss of a proton. [TEA-H]; triethaI
amine matrix peak. The mass of characteristic fragments is indicated.

27

stearic and palmitic acids, or palmitic and vaccenic acids respectively. The
most unsaturated fatty acid commonly found in bacteria was vaccenic acid, a
cis-A11 mono-unsaturated 18-carbon fatty acid. The observed fatty acid
composition was in agreement with data obtained by Radunz (7). Loss of one
fatty acid moiety (as the ketene analog) resulted in ions of m/z 581 and 555,
depending on the fatty acid and the molecular species. Both spectra from
spinach and bacterial sulfolipid showed peaks at m/z 241, 225, and 80, which
were suspected to be characteristic for the head group.

B/E linked scans of molecular spades of sulfolipid. Two B/E linked
scans focusing on the major species found in the spinach sample (m/z 815)
and the bacterial sample (m/z 847) are shown in Figure 2-2. Due to collision
activated dissociation as described above, there was a high abundance of
fragments at m/z 600-800 at intervals of about 14 mass units. These
corresponded to fragmentation along the fatty acid chains, generating
fragments with increments of one methylene (CH2). In the low mass range
fragments characteristic for the lipid head group were found. Both spectra
showed a comparable pattern of fragments with peaks at m/z 283, 241, 225,
and 80. (It must be emphasized that the relative abundance of peaks in mass
spectra generated by the use of FAB and CAD is not always reproducible.
The relative abundance of "daughter ions" seems to be influenced by sample
concentration and purity.) Similar, low mass daughter ions were detected
when B/E linked scans were performed on the parent ions of m/z 555 for

spinach and of m/z 581 for the bacterial sample (Figure 2-3), which represent

28

 

 
 
 
   

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

    

 

 

 

 

 

 

 

 

 

 

 

 

 

1.0
Spinach-8L
Parent Ion m/z 815
0.5 J
A
o\°
v
I.I.l
O 283
225
Z 241
< /
D 80
Z
_u-_‘.‘.---t.‘m_wu-u- ...__..._.L -...-_.. .- .... .. -L—AL.‘ - L-A..J 4.....5 .LN‘J-A.LL
g o ' Yfi T I I V Y Ir 1 I V V V V v I V Y r V V T V I Y T V V r V V V 7’1
0 100 200 300 400 500 600 700 900
<
LLI 4-
> .
j.— I R. sphaeroides-8L
< i
I Parent Ion m/z 847
_J 3 1
“J 1
[I .
283
2 .1
225
581
/
1-1
. 241
j 50 299
I l
o 'U "TV""Tfi'"l'ffifi"'f1""l“"I"
1 00 zoo 300 400 500 6% 700 800

Figure 2-2: B/E linked FAB-CAD-MS/MS scans focusing on the most abundant
molecular ion of the spinach (top) and the bacterial sample (bottom). The
mass of characteristic head group fragments is indicated. In the high mass

region the abundant peaks are due to fragments formed from cleavage along
the fatty acyl chains.

29

the respective molecular species following the loss of one fatty acid (compare
Figure 2-1).

BIE linked scans of sulfolipid head group specific fragments. The
sulfoquinovose sugar moiety has a molecular weight of 243. Consistently in all
the scans discussed, an ion was observed at mlz 241 and seemed to be a
key ion. It was proposed that this ion represents a stable oxidation product of
the ion at mlz 243, which has lost two hydrogens. The same phenomenon has
been observed for sulfated sugar head groups from animal lipids (5). B/E
linked scans focusing on the ion at m/z 241 from the spinach and the
bacterial samples again showed a comparable pattern (Figure 2-4). The
"daughter ion" at m/z 225 was proposed to be the epoxy derivative of the
sulfoquinovose following the loss of water, while the ion at m/z 80
corresponded to the sulfonic acid group 803' (Figure 2-5). The most abundant
ion at m/z 161 resulted from the loss of the sulfonic acid group. Based on the
limited available information, it was not attempted to propose a precise
structure for this fragment.

BIE linked scans focusing on the 299 mlz key Ion and Its derivatives.
The B/E linked scan for the"parent ion" at m/z 299 from the spinach sample
showed fragments at m/z 225, 165, and 80, but non at m/z 241 ( Figure 26,
top). This suggested that the m/z 241 key ion was preferentially derived from
the molecular ion, but that its formation from subfragments was not favored.
Therefore, an alternative fragmentation pathway was proposed, which started

with the formation of the key ion at m/z 299 (Figure 2-8). The ion at m/z 225

30

 

Spinach-8L
Parent Ion m/z 555

 

 

 

 

 

I R. sphaeroides-8L
1 Parent Ion m/z 581

RELATIVE ABUNDANCE (%)

 

 

 

 

 

Figure 2-3: B/E linked FAB-CAD-MS/MS scans focusing on monoacyl spe
of sulfolipid from spinach (top) and from H. sphaeroides (bottom). The m:
of important fragments is indicated.

31

 

Spinach-6L
‘ Parent Ion m/z 241

1 161

80

225

 

 

 

 

 

 

R. sphaeroides-3L
Parent Ion m/z 241

RELATIVE ABUNDANCE (%)

d

4

1

I
161

1

225
80

 

 

 

50 100 150 200

Figure 2-4: B/E linked FAB-CAD-MS/MS scans focusing on the head grOI.

key ion at m/z 241 from spinach (top) and bacterium (bottom). The mass
important fragments is indicated.

32

so;
0
OH 1 1
no I o OCOR R =C17 H29
2
no OCOR2 R =(315331

m/z 31519—91

_-(R2CO—H) —(R CO— a) so—
so3
o
1Mon
o M ocoa
ocon
HO

HO m/z 555
m/z 577 "/////mn
._\\W///\\\OCOR so3 "
OH

OH o
I on
no ' on
no —H20 _
_ /// \\\\\ 0
ml: 241 OH
no 0
-so3
,/// \\\\ l m/z 225
so: m/z 161
m/z 80 so;
m/z 80

Figure 2-5: Proposed fragmentation pathway showing the decomposition of
the key ion at m/z 241.

33

was interpreted as epoxy sulfoquinovose, while the S03; group was
represented by a peak at m/z 80. The structure of the ion at m/z 165 could
not be proposed on the basis of the available data. This ion was not observed
in the B/E linked scan focusing on the ion at m/z 241 (Figure 2-4). The B/E
linked scan for the ion m/z 283 was very similar to the B/E linked scan of the
ion at m/z 299 (Figure 2-6). The proposed structure for this ion is shown in
Figure 2-8. The Carboxylate ion of stearic acid [RCOO]' has a mass of 283.
Since there were considerable amounts of stearic acid in the bacterial
sample, B/E linked scans focusing on the m/z 283 ion of the bacterial sample
could not be interpreted.

To complete the series of B/E linked scans the fragmentation patterns
for the ions m/z 225 and m/z 80 from the spinach sample were obtained
(Figure 2-7). The m/z 225 fragment, as part of the sulfoquinovose head group,
produced the 803‘ fragment at m/z 80. In addition, several other fragments,
including that at m/z 165, were generated, for which no structure could be
proposed. The ion at m/z 80 gave rise to the SO2 ion with a mass of 64,
confirming that this ion represents the sulfonate group (Figure 2—7, bottom).
B/E linked scans for the ions at m/z 299, 283, 225, and 80 for the bacterial
sample were not run and could therefore not be compared with the equivalent

ions of the spinach sample.

.34..

an:

 

i Spinach-8L
Parent Ion m/z 299

225

1 so

 

 

 

 

.4 1.1-1.41“; pull-IA

 

 

[ V V r V U V I '7 I I

0 100 200 31

 

Spinach-8L
Parent Ion m/z 283

RELATIVE ABUNDANCE (%I

225
80

 

 

 

 

 

 

0 100 150 200 250

M/Z

Figure 2-6: B/E linked FAB-CAD-MS/MS scans of the spinach sulfolipid,
focusing on the "daughter ions" with the mass 299 (top). and 283 (bottomj
The mass of important ions is indicated.

 

 

35

 

 

 

 

 

 

 

 

10
I
1 Spinach-6L
1 Parent Ion m/z 225
8 ‘ so
1 \ 165
6 -I
4 -I
A ‘
°\° .
V 2 1
m J
O ..._La-L. . J.L....“ i-hL4-- “..I...“
Z 0 I
< 'TT"'T"'I"TIT"U'TTTfT'U'V'I'r'I
D so so 100 120 140 160 180 zoo 221
2
D 1
m 1
< . Spinach-8L
LLI 1 Parent Ion m/z 80
> .
f.-
<
..I
LU 64
Ct

 

 

 

 

 

 

T f V f j V V V V ' V V V V I Y Y Y Y ' i V V V I

50 80 70

Figure 2-7: B/E linked scans of the spinach sulfolipid, focusing on the
"daughter ions" with the mass 225 and 80. The mass of important ions is iI

 

36

1 1
/\/\OCOR R =C17 Hzg

2
OCOR2 R =C15331

m/z 815(M-H)

;-(R 2Cyl“) \RICO-o H) 80;
on
oo/Yo\ OCOR1 3°C :oO/Y\2
OCOR

m/z 555

Hm/z 57\ )“/////
-R c R -R CO, H
0‘ so3

m/z 299 AA

 

-16 HO
so;
I //“\<¢9
O
m/z 283 lift/z 225
m/z 165 _
503’
m/z 80

Figure 2-8: Alternative fragmentation pathway showing the decomposition of
the key ion at m/z 299.

37
CONCLUSIONS

Based on the observed data, two diagnostic "daughter ions" of
sulfolipid were postulated: first, the sulfoquinovose head group specific ion
with a mass of 241, and second, the sulfoquinovose glycerol specific
deacylation product with a mass of 299. Since the “daughter ion" at m/z 299
did not give rise to the ion at m/z 241, and since the B/E linked scans
focusing on these two ions revealed distinct fragmentation patterns, it was
postulated that both ions were produced in Simultaneously occurring reactions
from the molecular ion, and that further fragmentation occurred in a pattern
specific for each of the two "daughter ions". Two alternative fragmentation
pathways were postulated featuring the m/z 241 and the m/z 299 ions as key
ions (Figures 2-5 and 2-8).

Comparison of linked B/E scans of the high mass ions and in particular
of the head group Specific ion at m/z 241, revealed similar fragmentation
patterns for the spinach and bacterial sulfolipid, strongly suggesting that the
sulfolipid found in both organisms is sulfoquinovosyl diacylglycerol. The data
were supported and extended earlier findings obtained by Wood et al. (8) and
Radunz (7), who identified the sulfolipid in H. sphaeroides by cochromato-
graphy with standards. However, it Should be pointed out that the mass-
spectrometry technique use did not allow to distinguish between the sulfo-
quinovose sugar and its stereo isomers. Although the data were consistent
with the presence of the sulfoqinovose In the bacterial lipid, it could not be

excluded that the sugar head group is a Stereo isomere of sulfoquinovose.

 

38

REFERENCES

1.‘

Benson A.A., Daniel H., Wiser R. (1959). A sulfolipid in plants. Proc.
Natl. Acad. Sci. USA 45:1582-1587

Benson AA. (1963). The plant sulfolipid. Adv. Lipid Res. 1:387-394

Bishop D.G., Sparace S.A., Mudd B.J. (1985). Biosynthesis of sulfo-
quinovosyldiacylglycerol in higher plants: The origin of the
diacylglycerol moiety. Arch. Biochem. Biophys. 2401851 -858

Jennings K R., Mason R. S. (1983). Tandem mass spectrometry
utilizing linked scanning of double focusing instruments. In:
Tandem mass spectrometry; Ed. McLafferty; John Wiley & Sons,
New York pp. 197-222

Kushi Y., Handa S., lshizuka I. (1985). Secondary ion mass
spectrometry for sulfoglycolipids: Application of negative ion
detection. J. Biochem. 97:419-428

O'Brien J.S., Benson AA. (1964). Isolation and fatty acid
composition of the plant sulfolipid and galactolipids J. Upid Res.
5:432-436

Radunz A. (1969). Ueber das SulfochinovosyI-diacylglycerin aus
hoeheren Pflanzen, Algen und Purpurbakterien. Hoppe-Seylers Z.

Physiol. Chem. 350:41 1-417

Wood B.J.B., Nichols B.W., James A.T. (1965). The lipids and
fatty acid metabolism of photosynthetic bacteria. Biochim.
Biophys. Acta 106:261-273

 

CHAPTER 3
R. SPHAEROIDES AS A MODEL SYSTEM FOR THE BIOSYNTHESIS OF
THE PLANT SULFOLIPID

ABSTRACT

A TLC screening method for the isolation of chemically induced
mutants of Ft. sphaeroides deficient in sulfolipid biosynthesis has been
developed. Initially, 15 putative mutants were isolated from 1512 mutagenized
cell lines examined. Following the preliminary analysis of all 15 putative
mutants, further analysis was confined to three strains. Two of these, CHB16
and CHB17, were found to contain about 10% Of the amount Of sulfolipid as
compared to the wild type, while in the mutant strain CHB18 the sulfolipid
levels were reduced by about 50% with respect to the wild type. All mutants
showed greatly reduced rates Of incorporation of labelled sulfate into sulfolipid,
while proteins were labelled at a normal rate compared to the wild type,
indicating that the general sulfate assimilation pathway was not affected.
Mutants CHB17 and CHB18 accumulated a"’S-labelled water-soluble
compounds, which are potential intermediates in the pathway of sulfolipid

biosynthesis.

39

40
INTRODUCTION

The four major polar lipids found in R. sphaeroides are phosphatidyl
glycerol, phosphatidyl ethanolamine, phosphatidyl choline, and sulfoquinovosyl
diacylglycerol (17). In addition, the presence of cardiolipin (17) as well as the
occurrence of two unusual lipids, the ornithine lipid (4) and
N-acylphosphatidylserine (2) has been reported. The structures of the
R. sphaeroides lipids are given in Figure 3-1.

The occurrence of the plant sulfolipid in the purple bacterium
Ff. sphaeroides was originally demonstrated by Wood et al. (17) and was later
confirmed by Radunz (11). Both groups identified the sulfolipid by
cochromatography with standards. The structural identity of the plant and the
bacterial sulfolipid was further confirmed by mass spectrometry, as was
described in the previous chapter. This finding offered the opportunity to Study
the biosynthesis Of sulfoquinovosyl diacylglycerol in a bacterial system, which
has several advantages as a genetic model system as compared with a higher
plant system such as Arabidopsis thaliana. There are two main reasons for the
choice of Ft. sphaeroides as a model system for the study of sulfolipid
biosynthesis.

First, there is circumstantial evidence that the plant sulfolipid, which is
predominantly found in association with photosynthetic membranes, might
play an essential role in photosynthesis. The chemical properties of
sulfoquinovosyl

diacylglycerol, such as its low pK, make this lipid unique amongst the other

 

41
O
O CHz—O—(fli—R
R—g—O—CH O
(IIHz—O—II—O—CHz—CHZN (CH3);
0.
PC

( Phosphatidyl Choline )

0
II
I I II cI-Iz—o—c-Rl
Rz-C-O-C-H o
CH " T
2 O-P-OCHZCHZNHS
0" 0'

PG PE

 

( Phosphatfdyl Glycerol )

(3
("2503' 0120981
C) (J
HOQRz
II
HO O—CI-Iz
(3H
SL

( 6-Sulphoquinovosyl Diacylglycerol )

( Phosphatidyl Ethanolamine )
(cries-NH.
R, COOH—NH—CH—CO—O R2

OL

( Ornithine Lipid )

Figure 3-1: Structures of the most abundant polar lipids found in

H. sphaeroides.

42

membrane lipids in the chloroplasts of higher plants. In addition, the sulfolipid
is found in tight association with intrinsic protein complexes of the thylakoid
membrane, leading to the suggestion that it might be required for the catalytic
activity of these protein complexes (1; see also discussion about the function
of sulfolipid in chapter 1). There are sufficient reasons to assume that the loss
of sulfolipid by an obligate photoautotrophic organism such as a higher plant
may not be tolerated.

Purple non-sulfur bacteria, including R. sphaeroides, can grow under a
wide range of growth conditions. The photosynthetic activity of the
membranes is dispensable when the bacterium is grown chemohetero-
trophically under aerobic conditions. The bacterium ceases production of the
photosynthetic apparatus in certain growth regimes (e.g. in a high oxygen
atmosphere in the presence of organic substrate). If the atmospheric pressure
of oxygen is lowered below a certain point, de novo synthesis of the
photosynthetic membranes is induced. R. sphaeroides is therefore ideally
suited for the dissection of photosynthetic membrane biogenesis by classical
and molecular genetic studies (8). Thus, if the plant sulfolipid is required for
photosynthetic growth, it will still be possible to identify mutations that
inactivate sulfolipid biosynthesis under conditions in which the bacterium can
grow chemoheterotrophically.

A second reason for the use of R. sphaeroides Is that all the genetic
techniques required for modern genetic analysis of the pathway for sulfolipid

biosynthesis have previously been applied to Ft. sphaeroides. Mutants

43

affecting complex biochemical pathways have been generated by chemical
mutagenesis (e.g. reference 10); genes have been Cloned by
complementation of mutants using cosmid libraries, prepared from wild type
DNA (e.g. reference 16); and gene replacement techniques, commonly used
in other bacteria, have been used to construct specific mutations (e.g.
reference 3). In addition, a physical map of the whole genome of

R. sphaeroides has been recently constructed (15). This map might become
useful for the linkage analysis of the various genes involved in sulfolipid
biosynthesis.

Because of the genetic tools and the rapid life cycle, it is much more
straightforward to isolate a gene by genetic complementation in the bacterial
system, than in a higher plant.

It is expected that the information gained by Studying sulfolipid
biosynthesis in R. sphaeroides will be applicable to higher plants. Any
hypothesis about sulfolipid biosynthesis in Ff. sphaeroides can be tested in
higher plants using the tools developed during the course of this work.
However, there is the possibility that the biosynthesis of sulfolipid proceeds
along different routes in the bacteria compared to higher plants.

In this chapter, the feasibility of using mutants of H. sphaeroides to
study the biosynthesis of the plant sulfolipid is reported. A TLC screening
procedure for the isolation of polar lipid mutants and the primary phenotypes

of three mutants deficient in sulfoquinovosyl diacylglycerol are described.

 

MATERIALS AND METHODS

Bacterial strain. The wild type strain Hhodobacter sphaeroides 2.4.1.
used throughout this study was kindly provided by Dr. Uffen’s lab at MSU.

Growth conditions for R. sphaeroides. Cell cultures were initially grown
in the malate-basaI-salt medium described by Ormerod et al. (9). The
ingredients of this medium are listed in Table 3-1. For aerobic growth, stock
solutions (I, II, III, IV, V, VI) were diluted into water according to the dilution
factor given in Table 3-1. The pH was adjusted to 7.0 with NaOH following
addition of 3 g D,L—malate per liter. Sterile solutions V and VI were added after
autoclaving. For photoheterotrophic growth, sterile solutions VII and VIII were
also added. In labelling experiments using 32P-phosphate, solution I was
replaced by solution l-P, and in labelling experiments involving 35S-sulfate
solutions I and II were replaced by l-S and HS respectively.

In later experiments, Ormerod’s medium was replaced by Sistrom’s
succinate-basal-salt medium A (13, 14), which was specifically adapted to the
growth requirements of Ff. sphaeroides. The stock solutions required for
Sistrom’s medium are listed in Tables 3-2, 3-3, and 3-4. The 10X stocks were
kept at -20°C. Following dilution of the stock solution, the pH was adjusted to
7.0 with KOH and the solution was autoclaved. Agar plates (1.5% agar) were
either incubated in a dark incubator at 30°C or in a plexiglass box, which was
flushed at a flow rate of 1 ml/min with a gas mixture containing 5% CO2 and
95% N2. Light and heating were provided by a 60 Watt incandescent light

bulb placed in such a distant from the box to sustain a temperature of about

Table 3-1: Stock solutions for Ormerod's medium.

45

 

Stock# (x) Ingredients g/I
I (10): KHZPO, 0.6
KZHPO4 0.9
(NH,)ZSO4 1.0
l-P (10): KCI 1.0
KH2P04 0.06
KZHPO4 0.09
(NH,,)2SO4 1.0
l-S (10): KH2P04 0.6
KZHPO4 0.9
NH,CI 0.75
ll (10): MgSO4-7H20 0.2
"S (10): MgCl2-6HZO 0.14
Ill (10): Fe-EDTA 0.03
IV (1000): H3803 4.3
MnCl2 2.7
CUSO4 0.125
ZnSO4 0.288
NaCI 0.584
CoCl2 0.0024
V (1000): Biotin 0.15
Thiamine 1.0
Nicotinic acid 1.0
VI (100): CaCl2 8.0
VII (50): NaHCO3 84.0
VIII (100): Ascorbic acid 50.0

 

46
Table 3-2: Ingredients of Sistrom’s 10X medium.

 

Components g/l

1. KZHPO, 34.8

2. (NH,)ZSO, (NH4CI)' 5.0 (1 .95)’
3. Succinic acid 40.0

4. Glutamic acid 1.0

5. Aspartic acid 0.4

6. NaCl 5.0

7. Nitrilo acetic acid 2.0

8. MgSO4-7HZO (MgCl2-6H20)' 3.0 (2.44)‘
9. CaCI2-2H20 0.334

10. FeSOflHZO 0.02

11. (NH4)6MO7OZ,-4HZO 0.002

12. Trace element solution (ml) (1.0)

13. Vitamin solution (ml) (1.0)

 

(' chloride salts to replace sulfate salts for 35S-sulfate label experiments)

Table 3-3: Trace element solution for Sistrom’s medium.

 

Components g/100 ml
NaZEDTA'ZHZO 1 .41
ZnSO.-7HZO 10.95
FeSO,,-7HZO 5.0
MnSO,‘H20 1.54
CUSO4'5H20 0.392
Co(N03)2-6H20 0.248

H31303 0.1 14

 

47

Table 3-4: Vitamin solution for Sistrom’s medium.

 

Component g/100 ml
Nicotinic acid 1.0
Thiamine-HCI 0.5

Biotin 0.010

 

30-35°C and a light irradiance of approximately 100 pE/s per rn2 at the level of
the plates. Liquid cultures were inoculated with a single colony grown on an
agar plate containing Sistrom’s medium. Aerobic Chemoheterotrophic, liquid
cultures were incubated at 30°C under shaking in Erlenmeyer flasks closed
with a foam stopper. Cells grown in these aerobic shaker cultures had a
tendency to clump. Anaerobic photoheterotrophic cultures were grown in
tightly closed, filled 200 ml bottles, which were kept under a 60 Watt
incandescent lamp. This lamp was placed at such a distance as to provide a
temperature of 30°C inside the bottle and a light irradiance of approximately
100 pE/s per m2 at the surface of the bottle. The bottles were mixed once or
twice a day by shaking. The generation time for photoheterotrophic cultures
was about 4 h. For the large scale preparations, photoheterotrophic cultures
were grown in 4-l Erlenmeyer flasks, with two 60 Watt bulbs placed on the
outside (approximately 100 pE/s per m2 at the surface of the flask). A 200-ml
preculture was used for inoculation. Adequate mixing was provided by placing

a magnetic stir bar inside the flask and the flask was placed in a water bath

 

- “A‘—

 

48
kept at 30°C.

For growth of mutagenized cells on complete medium, Z-broth (14)
was used. One liter of aqueous solution contained 10 g tryptone, 5 g yeast
extract, 10 g NaCl, 1 g glucose, 0.4 g CaClz-2H20 (pH 6.8-6.9). Plates,
containing solidified medium were incubated in a dark incubator at about
35°C, liquid cultures were grown in a dark shaker at 35°C.

Mutagenesis of H. sphaeroides. A 50 ml culture (Z-broth) was grown
from a single colony to a density of ODm=0.5 (mid-log phase). TO 20 ml of
this culture 2 mg of nitrosoguanidine crystals were added (0.1 mg/ml), and the
culture was incubated at room temperature in the fume hood with vigorous
shaking. At various times, aliquots of 3 ml were removed, and the cells were
immediately spun down and the supernatant was removed. The cells were
washed once with 3 ml of Z-broth. Small aliquots were plated directly onto
agar-solidified Z-broth to establish a killing curve, while the bulk of the culture
was incubated for 17 h with Shaking at 35°C to allow the cells to replicate two
to three times. Aliquots of these cultures were plated onto Z-broth plates
containing 100 pg/ml rifampicin or 100 pg/ml streptomycin to test for the
degree of mutagenesis, by scoring the frequency of antibiotic resistant
mutants. Further aliquots were plated onto plain Z-broth plates to determine
the total amount of cells in the culture, and to initiate the screening for
mutants. TO the remainder of the cultures DMSO was added to a final

concentration of 20% (v/v), to allow storage of the mutagenized cells at -70°C.

49
Screening for polar lipid mutants. Randomly chosen colonies from a

mutated population of Ft. sphaeroides cells were picked to patches on a fresh
Z-broth plate (24 per plate). Lipids were isolated from these patches by taking
cells onto the wide end of a flat tooth pick and swirling the material in 75 pl
chloroform/methanol (1:1, v/v) contained in polypropylene microfuge tubes.
Following the addition of 25 pl 1 N KCI, 0.2 M H3PO4 the tubes were vortexed
and centrifuged to separate the organic and the aqueous phases. Using a
glass capillary micropipet, 10 pl from the lipid-containing lower phase were
withdrawn and directly spotted onto an activated ammonium sulfate plate (12
samples per plate). The plates were developed and the lipids were visualized
by exposure to iodine vapor as described below. Using this procedure, it was
possible to examine about 144 samples per day.

Preparation of lipids labelled with various Isotopes. A 500 pl sample of a
stationary, photoheterotrophic R. sphaeroides culture was centrifuged, and the
cells were resuspended in 5 ml of the respective Ormerod’s labelling medium
for aFP—phosphate or 35S-sulfate labelling (see above). For 1“C-acetate labelling
experiments, regular Ormerod medium was used. The cultures, contained in
closed 8-ml screw cap tubes, were incubated for 6 h under a 60 Watt light
bulb (irradiance approximately 100 pE/s per m2) prior to addition of 50 pCi
carrier- free a2P-phosphate, 50 pCi carrier-free 35S-sulfate, or 1 pCi 14C-acetate
with a specific activity of 56.6 mCi/mmole. Following further incubation for 18 h
under the same conditions, the cells were Spun down in a clinical centrifuge

and resuspended in 0.5 ml water. The cells were extracted by vortexing with 4

50
ml chloroform/methanol (1:1, v/v). Addition of 1.3 ml 1 M KCI, 0.2 M H3P04,

vortexing and centrifugation resulted in phase partitioning of the lipids into the
lower chloroform phase. This phase was removed and concentrated down to
0.2 ml by evaporation under a stream of N2. The lipid extracts were analyzed
by TLC as described below.

Time course for the incorporation of S‘s-sulfate Into proteins and
sulfolipid by the wild type and mutants of Ft. sphaeroides. Cells (grown
photoheterotrophically in Ormerod’s medium) from a 2 ml late log phase
culture were spun down and resuspended in 8 x 2 ml low sulfate Ormerod’s
medium containing 10 pCi 35S-sulfate each (specific activity adjusted to 0.5
pCi/nmole). These cultures, contained in microfuge tubes (2 ml), were grown
in the dark at 30°C. One of each of the 8 parallel cultures was analyzed at 3,
6, 9, 12, 15, 18, 24, and 30 hours after addition of the label. Following
centrifugation in a microfuge, the cells were extracted with 0.5 ml
methanol/chloroform/formic acid/water (12:5:1 :2, v/v). The extracts were stored
at -20°C. Proteins and cell debris were pelleted by centrifugation. The
supernatant was transferred to a new tube and the pellet was reextracted with
0.25 ml methanol/formic acid/water (40:1 :59, v/v). Following centrifugation, the
supernatants from the two extractions were combined and phase partitioned
by addition of 0.15 ml chloroform and 0.05 ml water. The chloroform phase
contained the labelled sulfolipid, which was further analyzed by TLC on
ammonium sulfate plates (see below). The sulfolipid band was scraped from

the plate and the radioactivity was determined by scintillation counting. The

 

51

pellets, containing most of the protein and cell debris, were resuspended in
0.05 ml 0.1 N NaOH and incubated at 37°C for 30 min. The samples were
diluted with 0.45 ml H20 and 0.1 ml were used for protein determination using
the Pierce BCA protein assay according to the manufactor’s instructions. The
radioactivity in 0.01 ml samples was measured by scintillation counting.
Preparation of s“S-labelled, water-soluble compounds and sulfolipid
from wild type and mutant strains. Cells grown aerobically with shaking in
Ormerod’s medium in the dark at 30°C were spun down, and resuspended in
20 ml of low- sulfate labelling medium containing 0.5 mCi 35S-sulfate (specific
activity 2.5 pCi/nmole). The cultures were incubated for 17 h under the same
conditions as the precultures. The cells were harvested by centrifugation and
extracted with 5 ml methanol/chloroform/formic acid/H20 (12:5:122, v/v).
Following centrifugation in a clinical centrifuge, the supernatant was saved and
the pellet was reextracted with 2.5 ml methanol/formic acid/H20 (40:1:59, v/v).
The sample was centrifuged again. To the combined supernatants from the
two extractions, 1.5 ml chloroform and 0.5 ml H20 were added, resulting in the
partitioning of water soluble compounds into the aqueous phase and lipids
into the chloroform phase. The labelled lipids were analyzed by TLC (see
below), while the aqueous phase was evaporated at 45°C under a stream of
N2 and the residue was resuspended in 0.3 ml water. Insoluble material was
removed by centrifugation. To separate basic and acidic compounds, the
solution was applied to a 0.25 ml Dowex 50H’r column which was prepared in

a 1 ml plastic pipet tip. The column was eluted with 0.5 ml water followed with

52
0.5 ml 2 N NH,OH resulting in the acid and basic fractions respectively. This

material was dried under a stream of N2 at 45°C prior to further analysis.
Quantitative lipid analysis of mutants and wild type by gas

chromatography of fatty acid methyl esters. For each strain three 50 ml
cultures inoculated with single colonies were grown aerobically with shaking at
32°C using Sistrom’s medium. Lipid extracts were prepared from pelleted cells
as described for the preparation of labelled lipids (see above) using the same
volumes of solutions. After concentration of the chloroform phase, the material
was spotted onto two activated silica TLC plates (Si250, Baker). The plates

were developed in two dimensions using the system described below. Lipids

were visualized with iodine vapor and the spots were individually scraped into “

8 ml screw cap tubes. To the samples 5 pg myristic acid methyl ester in 0.1
ml hexane was added as an internal standard. Fatty acid methyl esters were
prepared by addition of 1 ml of 1 N methanolic HCI followed by incubation at
80°C for 1 h. Following the addition of 1 ml 0.95% (w/v) KCI, the fatty acid
methylesters were extracted into 1 ml hexane. The hexane phase was dried
down to 0.1 ml. Samples of 2 pl were injected into the gas chromatograph
(Varian 2000), which was equipped with a 2.4 m column (2 mm inner
diameter), packed with 3% SP-2310/2% SP-2300 on 100/120 Chromosorb W
AW (Supelco). The carrier gas (N2) flow rate was adjusted to 20 mI/min, and
the column temperature was set for 2 min at 180°C, in 10 min to 200°C, and 4
min at 200°C. The fatty acid methyl esters were detected by a flame ionization

detector, and the data were integrated by a Spectra Physics integrator. To

 

53

calculate the relative amounts of the eight polar lipids included in the analysis,
the amount of fatty acids contained in each lipid was calculated. The validity of
calculation was based on the assumption that each of the lipids, including the
unidentified lipids, contained two fatty acids per molecule, and that the
different lipids had a similar fatty acid composition. The amount (pg) of fatty
acids in each lipid of a particular sample was calculated from the
chromatograms as following:
[(total area under all peaks - standard peak area)/standard peak area] X 5 pg.
The relative amount for each lipid in the sample was expressed as percent of
all lipids analyzed. It should be pointed out that the data were not corrected
for the slightly different response of the different fatty acids in the flame
ionization detector. However, the resulting error was considered to be very
small compared to the experimental error occurring during sample preparation
and handling. A total of 6 samples (2 for each of the three cultures) per strain
was analyzed, and means and standard deviations were calculated. The data
were subjected to an analysis of variance, to identify significant differences
between the lipid composition of the different Strains at the 0.05% confidence
level.

Quantitative analysis of fatty acid methylesters was also used to
estimate the total amount of lipids in 35S-labelled lipid samples.

Analysis of lipids by TLC. One dimensional separation of polar lipids
from R. sphaeroides was accomplished by using ammonium sulfate treated

silica plates (modified according to reference 7). Baker Si250 silica plates with

 

54

preadsorbent layer were soaked in 0.15 M ammonium sulfate for 30 s and air
dried completely for a minimum of 2 days. Prior to use, the plates were
activated for 2.5 h at 120°C. This process drives off the ammonia and leaves
behind sulfuric acid. The amount of acidification of the plate depends on the
activation time and is crucial for successful separation of phosphatidyl glycerol
and sulfoquinovosyl diacylglycerol. The plates were developed in
acetone/benzene/water (91 :30:8, v/v). This qualitative system was employed
for screening of sulfolipid mutants.

For quantitative lipid analysis, a two dimensional TLC-system was used
(6). Lipid samples were spotted in the lower right corner of Baker Si250 silica
plates (without preadsorbent layer) activated for 30 min at 120°C. The plates
were air-dried for 10 min following development of the first dimension with
chloroform/methanol/water (65:25:4, v/v). The second dimension was
developed with chloroform/acetone/methanol/acetic acid/water (50:20:10:10:5,
v/v). Lipids were nondestructively visualized by exposing the plates to iodine
vapor, or by spraying the plates with 50% sulfuric acid followed by heating at
160°C for 10-15 min to char the lipids (destructive, permanent). Radioactive
labelled lipids were visualized by autoradiography. The Ornithine lipid and
phosphatidyl ethanolamine were specifically detected due to their free amino
group by spraying the plate with ninhydrin solution (0.2% w/v in sec-butanol).
Following heating at 120°C for 10—15 min the respective spots turned pink.

Analysis of asS-labelled compounds by 20 TLE/TLC. Water-soluble,

3°S-labelled compounds extracted from wild type and mutant strains were

55

analyzed employing a two dimensional separation system which was originally
designed to separate the intermediates of the Calvin cycle (12). For
electrophoresis in the first dimension a modified BioRad isoelectric focusing
apparatus was used. The apparatus was equipped with two electrode
compartments large enough to accommodate paper wicks of 20 cm width. A
0.5 cm thick aluminum plate (20X25 cm) was placed on top of the cooling
table, which was flushed with a 20% (v/v) ethylene glycol mixture. The solvent
was cycled through a Lauda water bath with the temperature set at -10°C. The
aluminum plate was necessary to accommodate a 20 X 20 cm TLC plate. Due
to this setup the temperature at the level of the TLC plate was kept at about
+10°C throughout the electrophoresis run. Samples (5 pl) were spotted into
the lower right corner of a MN300 cellulose TLC plate (Analtech), about 2 cm
from each edge. Prior to sample application the plate was sprayed with
electrophoresis buffer (955 ml H20, 35 ml acetic acid, 10 ml pyridine pH 4.0)
and dried in air for about 10 min. A picric acid standard (2.5 pl of 1:10 diluted
water saturated solution) was spotted into the lower left corner. The plate was
sprayed again with electrophoresis buffer avoiding excessive liquid on the
plate to prevent smearing of the sample. The plate was Immediately placed
onto the precooled aluminum plate, which was covered with plastic wrap to
avoid electric shorts. The orientation was such that the spotting area was near
the cathode and migration of anions towards the anode could be observed.
Two 20 X 7 cm Whatman 3MM paper wicks were soaked in buffer and placed

onto the edges of the TLC plate connecting the two electrode compartments.

 

56

A sequencing power pack was used to run the plate at a constant voltage of
800 V. The run was finished when the picric acid standard had migrated 5.5
cm towards the anode. Following air drying, the plate was developed in the
second dimension by conventional TLC (solvent: 2-butanol/formic acid/H20;

60:10:20, v/v). Labelled compounds were detected by autoradiography.

RESULTS AND DISCUSSION

Isolation of mutants deficient in the biosynthesis of the plant sulfolipid in
Ft. sphaeroides. To enhance the mutation rate, wild type Ff. sphaeroides cells
were treated with the chemical mutagen nitrosoguanidine as described in the
methods section of this chapter. This compound causes clustered point
mutations, since it acts preferentially at the replication fork (5). The degree of
mutagenesis was optimized by incubating the cells for various lengths of time
in the presence of 100 pg/ml nitrosoguanidine. Survivor ratios were
determined by plating a small aliquot of the cultures immediately following
exposure to the mutagen. Since actively growing cells during the exponential
phase of growth contain at least two copies of the genome due to continuous
replication, the cultures were allowed to grow for 18 h before plating, to
ensure segregation of the mutagenized chromosomes and genotypic purity of
the resulting colonies. The degree of mutagenesis was monitored by the
frequency of mutations for resistance to the antibiotics rifampicin and
streptomycin. In addition, the occurrence of pigment mutations (e.g. dark red

colonies as opposed to wild type pink colonies) provided a visual indication of

57

Table 3-5: Effect of duration of exposure to nitrosoguanidine on
survival and on induction of mutations.

 

Exposure time (min)

 

 

0 1 0 20 30 45
Survivors (%) 100.0 10.0 1.3 0.7 0.7
rifR1 2.5X10'° 1 .0X10‘5 2.3X10'5 5.0X10'4 2.5X10"
strepR1 5.0X10'7 2.5X10'° 7.0X10'° 4.0X10'5 7.0X10'5

 

1- The frequency of antibiotic resistance is expressed as resistant
colonies per total number of colonies.

the extent of the mutagenesis. Table 3-5 shows the results for a typical
mutagenesis experiment.

The exact measurements of the frequencies by plating was adversely
affected by the fact that cells treated with the mutagen had a tendency to
clump. Therefore, the values given in Table 3-5 can only be considered
approximate. However, the results suggested that in the most heavily
mutagenized population (45 min exposure) the occurrence of antibiotic
resistance is increased approximately 100-fold compared to spontaneous
frequencies in wild type cultures (0 exposure). The number of survivors fell

below 1% for the longest exposures to the mutagen. Upon visual examination

 

m allowed for the separation in one
dimension of all the major polar lipids

found in R. sphaeroides. Figure 3
shows the separation of 353-,

-2
32R, and 1‘C-

labelled lipids on these plates.

identified. The three lanes shown i

-2, U) were not
n Figure 3

-2 were run on different TLC

 

 

 

59

35S-Sulfate 32 P-Phosphate 14C-Acetate
Front - Front - Front -

U- "

U- SL-

PE- '
U- PE-

PC- . PC-

Origin- . Origin- I Origin-

Figure 3-2: Separation of in vivo labelled lipids from R. sphaeroides on
ammonium sulfate plates. The labelling substrate is indicated on top of the
lanes. The three autoradiograms shown were derived from three different
experiments. OL, ornithine lipid; PC, phosphatidyl choline; PE, phosphatidyl
ethanolamine; PG, phosphatidyl glycerol; SL,sulfolipid; U, unidentified
phospholipids.

60

activation of ammonium sulfate-treated plates at 120°C produces sulfuric acid,
which protonates phosphatidyl glycerol, making it less polar. Less polar lipids
run in this TLC system closer to the front. Sulfolipid on the other hand is not
being protonated, and its migration was therefore not affected by the
activation process. However, given that the plate was sufficiently activated
prior to the run, phosphatidyl glycerol was readily separated away from the
sulfolipid, permitting easy scoring of mutants deficient in the biosynthesis Of
the sulfolipid or of any other lipid.

During the screening procedure, 1512 cell lines, the equivalent of 126
TLC plates, were analyzed for their lipid composition. The plates were visually
examined following staining of the lipids with iodine. Figure 3-3 shows a
picture of a developed TLC plate as it appeared during the screening
procedure. One of the samples on this plate contained no visible sulfolipid
band, identifying the cell line from which this extract was derived as a
probable mutant. A total number of 15 putative mutants showing very low to
moderately low levels of sulfolipid in relation to the other lipids was saved. In
addition, 6 putative mutants with dramatic changes in lipid composition other
than sulfolipid were isolated. These included two mutants seemingly lacking
phosphatidyl choline.

For practical reasons, it was necessary to focus further work on a
selected number of mutants, which were most likely to provide some clues
about the biosynthetic pathway for sulfolipid. Two criteria were used to select

the mutants for further study. First, the mutant must grow on Ormerod’s

61

—PG
——OL
—SL

—PE

—PC
12345678910111?

Figure 3-3: Separation of lipid samples prepared from twelve mutagenized cell
lines on ammonium sulfate plates. One sample (lane 8) does not show a
sulfolipid band indicating a putative sulfolipid mutant. OL, ornithine lipid; PC,

phosphatidyl choline; PE, phosphatidyl ethanolamine; PG, phosphatidyl
glycerol; SL, sulfolipid.

Haul—V

62

minimal medium to exclude the possibility that it was not deficient in sulfolipid
due to a defect in the general sulfate assimilation pathway. Second, it was
decided to focus on mutants that accumulated labelled compounds, which
might represent putative precursors of sulfolipid biosynthesis.

To obtain a broader knowledge of the phenotype of the putative
mutants, the 15 putative sulfolipid mutants were rescreened using a labelling
procedure. Those which grew on Ormerod’s minimal medium were labelled
with 35S-sulfate. Labelled cells were extracted, and the lipids and water soluble
compounds were analyzed. Equal amounts of lipids from each mutant sample,
as determined by fatty acid methyl ester quantification by gas
chromatography, were loaded onto the TLC plate. Following development,
the radioactivity in the sulfolipid band was determined and expressed as
pCi/mg fatty acid methyl esters. The water soluble compounds were
separated into basic and acidic compounds, and analyzed by 2D TLE/T LC.
The patterns observed in mutant samples were compared to that obtained
from a wild type sample following autoradiography. The data obtained in this
screening experiment are shown in Table 3-6. They have to be considered
preliminary, since only one labelling experiment for each putative mutant was
performed.

Based on the screening experiment described above, the three mutant
lines CHB16, CHB17, and CHB18 were selected for further analysis. The
selection was based on the fact that CHB16 and CHB17 Showed very small,

almost undetectable amounts of labelled sulfolipid, while all three of these

63

Table 36: Preliminary characterization of putative sulfolipid mutants. Growth
refers to visible colony formation after 2 weeks Of anaerobic, photoheterotrophic
growth on Ormerod’s minimal medium. The incorporation of asS-Sulfate into
sulfolipid (SL) is expressed as radioactivity in the sulfolipid band per total fatty
acid methyl esters (FAM) in the lipid sample (pCi/mg FAM). The mean value for
three independent wild type labelling experiments was 22.2 i 6.7 (pCi/mg FAM).
The column "New Compounds" indicates the presence of additional 35S-sulfate-
labelled, water-soluble compounds as compared to the wild type.

 

Strain Growth SL(pCi/mg FAM) New Compounds
CHB2 + 23.3 —
CHB5 + 7.9 +
CHB7 — n.d. n.d.
CHB8 — n.d. n.d.
CHB9 + 8.1 -
CHB10 + 36.2 —
CHB12 + 15.9 —
CHB13 — n.d. n.d.
CHB14 + 11.3 —
CHB15 + * —
CHB16 + 0.3 (+)
CHB17 + 0.2 +
CHB18 + 1.3 +
CHB19 + 0.2 —
CHB21 + 4.2 —

 

* The sample was lost prior to scintillation counting. However, examination of the
autoradiogram showed no clear reduction of labelled sulfolipid compared to wt
in this mutant.

64

mutants seemed toaccumulate water-soluble, “S-sulfate—labelled, compounds.
In particular CHB18 showed three very prominent additional spots on the 2D
chromatograms compared to the wild type (see further below). However, the
additional spot initially seen upon analysis of the extract from mutant CHB16
could not be reproducibly detected in later experiments. Several of the
putative mutants clearly did not rescreen (CHBZ, CHB10, CHBiZ), while
others (CHB5, CHBQ, CHB14, CHB15), showing slight reductions of labelled
sulfolipid compared to the wild type, need to be retested in the future to be
certain. Two certain candidates for sulfolipid mutants, on which no further
analysis has been performed to date, are CHB19 and CH821. It will be
interesting to know whether these two mutants carry mutations in genes other
than the three genes mutated in CHB16, CHB17, and CHB18.

The primary phenotype of three selected sulfolipid mutants. As will be
seen in chapters 4 and 5, the three mutants CHB16, CHB17, and CHB18
represent three different genes involved in sulfolipid biosynthesis. These genes
were given the designation squ (CHB17), squ (CHB16), and squ (CHB18)
and the mutants will be referred to by these designations throughout this text.

To confirm the accumulation of novel 3°S-labelled, water-soluble
compounds in extracts from the three mutants, three individual labelling
experiments for each of the mutants and the wild type were performed. The
labelled, water-soluble compounds were analyzed as described in the
methods section. The results were found to be reproducible, and the 20

chromatograms from one representative experiment are shown in Figures 3-4

 

65

TLE

TLC

Figure 3-4: Separation of water-soluble, a"’S-labelled compounds contained in
the acidic fraction of wild type and mutant extracts by 2D TLEfT LC. Samples
containing approximately 0.5-1.0 pCi were applied to each plate, and the film
was exposed for 3 days. A: wild type; B: squ mutant; C: squ mutant; D:
squ mutant. Spots which reproducibly appeared in the mutant samples but
not in the wild type are indicated by arrows. The origin is marked with an o.

66

o
. O O
.4 T 0.
° P
c D
‘
/ 1
fl
\ .' ‘.
b . '
' O ,' ' O .5,
/ ii.
00 ‘9‘
TLC

Figure 3-5: Separation of water-soluble, 35S-labelled compounds contained in
the basic fraction of wild type and mutant extracts by 2D TLE/T LC. Samples
containing approximately 0.5-1.0 pCi were applied to each plate, and the film
was exposed for 3 days. A: wild type; B: squ mutant; C: squ mutant; D:
squ mutant. Spots which reproducibly appeared in the mutant samples but
not in the wild type are indicated by arrows. The origin is marked with an o.

67

and 3-5. Based on the assumption that some of the precursors might contain
sulfonic acid or phosphate ester groups, a separation system was chosen,
which is commonly used for the separation of photosynthetic intermediates of
the reductive pentose phosphate cycle (12). The electrophoresis was run in
such a way as to observe anions, which are charged at the pH (4.0) of the
running buffer. These included all the phosphate esters and sulfonic acids
harboring a net negative charge at pH (4.0). However, ions which were
positively charged at pH (4.0) such as certain amino acids migrating towards
the cathode, were lost for the analysis, since they ran off the plate into the
paper wick. Aqueous extracts from the various cell lines were first separated
into acid and basic compounds using a Dowex 50H+ column. The aqueous
'flow through" contained the acidic compounds while the ammonium
hydroxide eluate made up the basic fraction, containing mostly amino acids.
The compounds contained in the acid fraction (Figure 3-4) and those
contained in the basic fraction (Figure 3-5) showed characteristic and
reproducible patterns. By spotting a sample into the middle of the plate
instead of into the lower right corner, it could be estimated that about 5% of
the radioactivity in the acidic fractions and about 30-40% of the radioactivity in
the basic fractions was lost due to migration towards the cathode into the
paper wick.

As Figure 3—4 shows (indicated by arrows), the squ mutant
accumulated 3 compounds visible in the acid fraction, which could not be

detected in the wild type sample. The squ mutant accumulated one

68

compound compared to wild type, while no reproducible difference between
the wild type and the squ mutant could be detected. In the basic fraction
(Figure 3-5) only the squ mutant showed the accumulation of new
compounds compared to the wild type. It was assumed that the appearance
Of these compounds in the mutants was somehow related to a block in the
pathway for the biosynthesis of the plant sulfolipid. Due to this specific block
caused by the mutation, precursors of sulfolipid biosynthesis accumulated,
which were visualized by the experiment described above. The squ mutant is
clearly deficient in sulfolipid biosynthesis, yet did not show the accumulation of
35S-labelled compounds in the experiment described above. There are three
possible explanations: first, such compounds accumulated but were missed
because they ran off the plate during electrophoresis; second, the respective
precursor was shunted into a different pathway and did not accumulate; or
third, the block in the pathway occurred at a very early stage in sulfolipid
biosynthesis prior to the incorporation of sulfur into a precursor molecule.
Therefore, no a'~"S-labelled compound would be expected to accumulate.

To test whether the three sqd mutants harbor defects in the general
sulfur assimilation pathway which, in turn, could affect protein biosynthesis,
incorporation rates of 35S-sulfate into proteins and the sulfolipid were
compared as described in the method section. As can be seen in Figure 3-6,
all three mutants and the wild type incorporated labelled sulfate into protein at
a Similar rate, while the capability of the mutants to label sulfolipid was greatly

reduced. Only in the squ mutant was the sulfolipid labelled at a detectable

Incorporation of Sulfate (nmol)

lncorportlon of Sulfate (pmol)

 

69

Sulfate Incorporation into Protein

 

 

 

 

 

 

 

 

 

 

5 - WT .
A—ACHBIS
4 . o—OCHB17
a—ACHBIS
3 . Q
0/
2 ' a
‘ ' 3,9
35
o _ o__.-§“
o s 12 15 24 30
Time (h)

Sulfate Incorporation into Sulfolipid
120 .
100 - °

A—A CHBIS
so - o—o CH817
A—A CHBIB
60 - o
40 - /
/:
20 - 0,0
A
o - o-=-a—-n—a-e-==r—"‘A a
o 6 12 1a 24 30

Time (h)

Figure 3-6: TIme course for the incorporation of 35S-labelled sulfate into protein
(upper panel) and sulfolipid (lower panel) by wild type and mutant strains
CHB16 (squI. CHB17 (squ), and CHB18 (squ).

70

rate. The increase in labelled protein was paralleled by the increase of total
protein as determined by a colorimetric protein assay (data not shown). The
result was in agreement with the fact that all three mutants can grow on
Ormerod’s malate-basal-salt medium, which does not contain any reduced
sulfur compounds such as cysteine. Therefore, it was assumed that the
defects in the three mutants must be specific for sulfolipid biosynthesis, since
the biosynthesis of cysteine and its methionine were not affected in the three
mutants.

In the experiments described above, the capability of the mutants to
incorporate 35S-sulfate into sulfolipid was used as a measure of their ability to
synthesize sulfolipid de novo. This did not necessarily reflect the steady state
levels of sulfolipid in the mutants compared to the wild type. To determine the
lipid composition of the wild type and the mutants in a quantitative way, lipid
extracts from the wild type and the sqd mutants were separated in two
dimensions on silica TLC plates. In Figure 3-7 the typical lipid pattern for the
wild type and the three sqd mutants is shown. Lipids were identified by
isolation Of the spots from the plates, followed by cochromatography of the
lipids with standards on ammonium sulfate plates. NO sulfolipid spot could be
detected by Charring Of the separated lipids from the squ and squ mutant
samples, while a weak sulfolipid spot was visible in the squ mutant sample.
To quantify the lipids, the iodine stained spots were isolated from the plates
and fatty acid methyl esters were prepared and quantified by gas

chromatography using an internal standard. Myristyl methyl ester was used as

71

PE—
OL— . -UL' C
C
ULII— /‘ 9—5L . 0
PG '7": P c
.9
ID
5
o. o. .E
. o
. C o .9
C u.
’ I
' ' . O 0 ‘ O
. o
g A
O 0.

Second Dimension

Figure 3-7: Separation of lipids isolated from wild type (A) and three sqd mutants
by 2D TLC. A, wild type; B, squ; C, squ; D, squ. OI, ornithine lipid; PC,
phosphatidyl choline; PG, phosphatidyl glycerol; PE, phosphatidyl ethanolamine;
SL, sulfolipid; ULI and UL", two unidentified lipids. The position for the sulfolipid

in panels 3,0, and D are circled. The origin is indicated with O. The lipids on the
plates were visualized by charring.

72

Table 3-7: Polar lipid composition (weight %) of wild type and sqd mutant strains
of R. sphaeroides grown under Chemoheterotrophic conditions. OI, ornithine lipid;
PC, phosphatidyl choline; PG, phosphatidyl glycerol; PE, phosphatidyl
ethanolamine; SL, sulfolipid; ULI and ULII, two unidentified lipids; WT, wild type.
For each Strain two samples were analyzed from each of three independent
cultures. These data were subjected to analysis of variance. Means with identical

superscripts are not significantly different at the 0.05% level.

 

 

Lipid WT squ squ squ pooled SD
SL 54. i 0.4 07" i 0.2 04" i 0.2 2.6° i 1.0 0.6
PC 154‘ i 0.8 20.0b : 1.5 17.2“ i 0.7 21 .2b i 1.9 1.3
PG 18.7‘ i 0.9 242° i 1.4 29.6b i 1.6 231‘ i 2.8 1.8
PE 44.8‘I i 0.9 38.4b i 2.1 42.7‘ i 1.4 35.4b i 2.6 1.8
OL 18‘ i 0.5 3.2°b i 1.4 2.4‘” i 0.4 3.5b i 0.5 0.8
ULI 1.7‘ i 0.3 2.0” i 0.4 2.1‘b i 0.5 2.7'3 i 0.4 0.4
ULII 12.0‘: 1.1 11.6‘ i1-3 5.3b 111.0 11.4“: 1.0 1.2

 

73

standard, since only negligible amounts of endogenous myristic acid were
found in the bacterial lipids. The lipids included in the analysis were
phosphatidyl choline, phosphatidyl ethanolamine, phosphatidyl glycerol,
ornithine lipid, sulfolipid, and two unidentified lipids (Figure 3-7, panel A).
These lipids include more than 95% of the polar lipids. The large spot in the
Upper left corner represents neutral lipids, in particular carotenoids. Two
aliquots, each from three independent aerobic cultures grown in Sistrom’s
medium, for each of the three sqd mutants and the wild type were analyzed,
and the data were submitted to an analysis of variance. From the result shown
in Table 3-7, it became clear that the sulfolipid content in the squ and squ
mutants was reduced to about 10% of the wild type level.

Based on the labelling experiments described above, in which very
small amounts of labelled sulfolipid were detected (Table 3-6), as well as
based on the result of the lipid analysis, it was concluded that the squ and
squ mutants are slightly leaky, and still retained about 10% of their capacity
to synthesize sulfolipid. The squ mutant shows a reduction of the steady
state levels of sulfolipid down to little less than 50% of the wild type levels
(Table 3-7). This value seemed to be surprisingly high since incorporation of
labelled sulfate into sulfolipid occurred at much lower rates than 50% of the
wild type rate (Figure 3-6, and Table 3-6). As mentioned above, incorporation
rates of labelled sulfate into sulfolipid may not necessary reflect steady state
levels. The pool size of the various intermediates, the uptake of the labelled

substrate, as well as the turnover rates may influence the level of labelled

74

sulfolipid detected in a labelling experiment. Unless the system was completely
equilibrated with the labelled precursor, the degree of label found in the
sulfolipid may not necessary reflect the steady state level of sulfolipid as
determined by quantification of the total mass of the lipid. It was most likely
that the system had not been in equilibrium during the labelling time course
described in Figure 3-6, explaining the differences between the label
experiment and the lipid analysis experiment. Therefore, it is suggested that
the values presented in Table 3-7 reflect more realistically the steady state
levels for each lipid than those observed in the labelling experiments.

It should be mentioned that the squ mutant showed significantly
higher levels of phosphatidyl glycerol, and lower levels of the unidentified lipid
ll (ULII; see Table 3-7) than the wild type. The squ and squ mutants on the
other hand, showed significantly higher levels of phosphatidyl choline and
phosphatidyl glycerol than the wild type. It was not clear, whether this
observation was related to the defect in sulfolipid biosynthesis, or whether

mutations in the background resulted in these changes.

CONCLUSIONS

The experiments described in this chapter were designed to establish
the purple bacterium R. sphaeroides as a model system for the study of the
biosynthetic pathway for the plant sulfolipid. This goal was clearly achieved. A
TLC screening procedure was developed, facilitating the isolation of a series

of sulfolipid deficient mutants from a chemically mutagenized population of

75

H. sphaeroides. Some of the mutants accumulate asS-Iabelled, water-soluble
compounds as detected by ZD-TLE/T LC. These compounds may be
precursors which accumulate due to specific blocks in the pathway for
sulfolipid biosynthesis.

The sulfolipid mutants described in this work are the first isolated from
any organism. Because it was not efficient to perform a simultaneous
characterization of all mutants, the following work was focused on three sqd
mutants: squ (CHB17), squ (CHB16), and squ (CHB18). These three
mutants were selected, because they showed a pronounced deficiency in their
sulfolipid levels, and because they revealed three distinctive patterns for the
accumulation of 35S-labelled compounds. Based on this result, it was assumed
that these three mutants might mark three different genes involved in sulfolipid
biosynthesis. It should be noted that the mutant collection described above
(Table 3-6) contains at least two more mutants with significant sulfolipid
deficiencies (CHB19, CHB21). Once the genes for the three sqd mutants have
been isolated, a complementation analysis can be performed to test whether
these mutants carry mutations in different genes involved in sulfolipid
biosynthesis, or whether they represent alleles of the three sqd mutants.

Based on the results described above, it was concluded that the next
logical step for the characterization of these mutants would be the isolation of
the wild type genes complementing these mutations. First of all, any
characterization of the phenotype of the mutants, could be misleading due to

background mutations. Since the cells were quite heavily mutagenized, it is

76

certain that the described strains contain more than one mutation. Thus, in
principle, the accumulation of labelled compounds seen in the squ and squ
mutants may be due to a second mutation in the background and therefore
unrelated to the sulfolipid phenotype. To avoid this problem, isogenic lines
have to be constructed. In order to accomplish this, a series of experiments
was undertaken to clone the genes, which complement the various sqd
mutants. The next two chapters of this work deal with the isolation and
characterization of the genes complementing the squ, squ,and squ

m utants.

REFERENCES

1. Barber J., Gounaris K. (1986). What role does sulpholipid play within
thylakoid membranes? Photosynthesis Res. 9:239-249

2. Donahue T.J., Cain B.D., Kaplan S. (1982). Purification and
Characterization of an N-acylphosphatidylserine from R.
sphaeroides. Biochemistry 21:2765-2773

3. Falcone D.L., Quivey R.G. JR, Tabita FR. (1988). Transposon
mutagenesis and physiological analysis of strains containing inactivated
form | and form ll ribulose bisphosphate carboxylase/oxygenase genes
in Ft. sphaeroides. J. Bacteriol. 1705-11

4. Gorchein A. (1967). The separation and identification of the lipids
of Ft. sphaeroides. Proc. Roy. Soc. N. 170:279-297

5. Guerola N., lngraham J.L., Cereda-Olmedo E. (1971). Induction
of closely linked multiple mutations by nitrosoguanidine. Nature
New Biology 230:122-125

6. Joyard J., Blee E., Douce R. (1986). Sulfolipid synthesis from 35802;,
and [1-“C]acetate in isolated spinach chloroplasts. Biochim. Biophys.
Acta 879:78-87

 

10.

16.

17.

77

Kahn M.-U., Willams J. P. (1977). Improved thin-layer
chromatographic method for the separation of major
phospholipids and glycolipids from plant lipid extracts and
phosphatidyl glycerol and bis(monoacylglyceryl) phosphate from
animal lipid extracts. J. Chromato. 140:179-185

Kiley P.J., Kaplan S. (1988). Molecular genetics Of photosynthetic
membrane biosynthesis in R. sphaeroides. Microbiol. Rev.
52:50-69

Ormerod J.G., Ormerod K.S., Gest H. (1961). Light-dependent
Utilization of organic compounds and photoreduction of
molecular hydrogen by photosynthetic bacteria; relationships
with nitrogen metabolism. Arch. Biochem. Biophys. 94:449-463

Pemberton J.M., Bowen A.R.S.G.(1981). High-frequency
chromosome transfer in R. sphaeroides promoted by broad-host
range plasmid RP1 carrying mercury transposon Tn501. J.
Bacteriol. 147:110-117

Radunz A. (1969). Ueber das SulfochinovosyI-diacylglycerin aus
hoeheren Pflanzen, Algen und Purpurbakterien. Hoppe-Seylers Z.
Physiol. Chem. 350:411-417

Schurmann P. (1969). Separation of phosphate esters and algal
extracts by thin-layer electrophoresis and chromatography. J.
Chromatog. 39:507-5-9

Sistrom W.R. (1960). A requirement of sodium in the growth of R.
sphaeroides. J. Gen. Microbiol. 222778-785

Sistrom W.R. (1962). The kinetics Of the synthesis of
photopigments in R. sphaeroides. J. Gen. Microbiol. 28:607-616

Suwanto A., Kaplan S. (1989). Physical and genetic mapping of the R.
sphaeroides 2.4.1. genome: presence of two unique circular
chromosomes. J. Bacteriol. 171:5850-5859

Weaver K.E., Tabita FR. (1985). Complementation of a Fi. sphaeroides
ribulose bisphosphate carboxylase-oxygenase regulatory mutant from a
genomic library. J. Bacteriol. 164:147-154

Wood B.J.B., Nichols B.W., James A.T. (1965). The lipids and
fatty acid metabolism of photosynthetic bacteria. Biochim.
Biophys. Acta 106:261-273

 

CHAPTER 4
ISOLATION AND CHARACTERIZATION OF A CLONE COMPLEMENTING THE
SULFOLIPID-DEFICIENT squ MUTANT OF H. SPHAERO/DES

ABSTRACT

A cosmid library was prepared from R. sphaeroides wild type DNA.
Individual cosmids were mated into the sulfolipid squ mutant, and lipid
extracts from the exconjugants were analyzed by TLC for the presence of the
sulfolipid. Two cosmid clones were isolated which were able to restore the
biosynthesis of the plant sulfolipid in the squ mutant. A physical map for
one of the cosmids was constructed, and two 5-6 kb subfragments derived
from this cosmid were found to complement the mutant. A region of 2.1 kb
common to both of these subclones was sequenced and found to contain an
Open reading frame of 295 amino acids preceded by a ribosome binding site.
Deletion analysis in combination with Cloning of the Open reading frame into a
novel expression vector, constructed during this study, suggested that the
protein encoded by the Open reading frame can restore the wild type
phenotype Of the squ mutant in a complementation assay. By searching two
protein banks, no known proteins with striking homology to the putative
protein encoded by the squ gene were found, indicating that this gene

codes for a novel protein involved in plant sulfolipid biosynthesis.

78

79
INTRODUCTION

In the previous chapter the isolation and characterization of three
sulfolipid-deficient mutants was described. It was noted that further
biochemical or physiological analysis is hampered by the fact that the mutants
contain secondary mutations in their genetic background, making it difficult to
link the observed aspects of the phenotype directly to the mutation causing
the sulfolipid deficiency. lsogenic lines can be created by cloning the gene,
which complements the mutation, and using it, either to construct a null
mutation in the wild type background, by gene replacement, or by introducing
the wild type gene carried on a single copy plasmid into the mutant, restoring
the wild type phenotype. Such a line, which would be wild type with respect to
sulfolipid biosynthesis, could be compared with the mutant carrying a control
plasmid lacking the complementing gene. Because both lines carry the same
background mutations, any differences observed between the lines can be
directly attributed to the mutation.

Analysis of the DNA sequence of a gene which complements can also
provide clues about the nature of the protein product of the gene. Given the
sequence, it is possible to determine by hydropathy analysis according to the
method of Kyte and Doolittle (16) whether a protein is membrane bound or
soluble (16). Also, protein-sequence data bases can be searched to identify
similarities with other proteins, which may provide clues about the type of
enzymatic reaction catalyzed by the protein.

By expression of the cloned gene in E. coli, it is frequently possible to

8O

purify large amounts of the protein. The purified protein can either be studied
directly, or it can be used to raise specific antibodies. These antibodies can
serve as tools to study the protein and may be useful in identifying a similar
protein involved in sulfolipid biosynthesis in other photosynthetic organisms.
For the reasons described above, the next logical step in the

establishment of a model system for the study of the biosynthesis of the plant
sulfolipid was the isolation of the genes complementing the mutations in the
squ, squ, and squ mutants. In this chapter, the isolation and

characterization of the gene complementing the squ mutant is described.

MATERIALS AND METHODS

Bacterial strains and plasmids. The bacterial strains used in the
experiments described in this chapter and in chapter 5 are shown in Table 4-
1. The source of the plasmids used in this chapter, and the construction of
novel plasmids, is presented in Table 4-2.

Growth media. R. sphaeroides strains were grown in Sistrom’s medium
as described in the methods section of chapter 3. When required, 0.8 ug/ml
tetracycline were added to agar plates containing Sistrom’s medium.

E. coli strains were grown on LB medium containing 10 g bacto-
tryptone, 5 g bacto-yeast extract, and 10 9 NaCl per liter (pH 7.5). For the
preparation of single stranded DNA for sequencing purposes, the cells were
grown in 2X YT medium containing 16 g bacto-tryptone, 10 g yeast extract,

and 10 9 NaCl per liter (pH 7.4). When required, 10 pg/ml tetracycline,

81

Table 4—1: Bacterial strains.

 

Bacterial Source or
Strains Relevant Characteristics Reference
R. sphaeroides:
2.4.1 wild type R. L. Uffen
MSU
CHB17 sulfolipid deficient, this study,
carrying a mutation chapter 3
in squ
CHB16 sulfolipid deficient, this study,
carrying a mutation in chapter 3
squ
CHB18 sulfolipid deficient, this study,
carrying a mutation in chapter 3
squ
E. coli:
HB101 F- proA2 recA13 mch (17)
DH5aF’ F’/endA1 recA1
A(lacZYA-arg F) U1 69
(¢80dlacA(lacZ)M15) (20)
MM294 F- endA1 hdei17 thi-1 (2)
817-1 RP4-2-Tc::Mu-Km::Tn7
(integrated) pro, res A. Puehler

mod+

(22)

 

82

Table 4—2: Plasmid sources and derivation.

 

Plasmid Description or Source or
Construction Reference
pBS-KS”, Bluescript Vector Stratagene
pBS-KS'
pRK2013 Km' self-transmissible
RK2 derivative containing
ColE1 replicon and
transfer functions
to mobilize RK2
derivatives (1 0)
pLA2917 Cosmid Vector (T 0') RS. Hanson
derived from RK2 (1)
pCHBSOO Expression vector for
R. sphaeroides Appendix B
pCHB1701 pLA2917 derived
Cosmid clone complemen- Figure 4-2
ting the squ mutant this study
pCHB17011, 2 subclones of pCHB1701
pCHB17012 in pLA2917 complementing Figure 4-3
the squ mutant this study
pCHB17H3B 10.5 kb Hindlll fragment Figure 4-2
of pCHB1701 in pBS-KS+ this study
pCHB17BZB 10.5 kb Bglll fragment Figure 4-2
of pCHB1701 in pBS-KS+ this study
pCHB17BZO-D“, 2.1 kb BamHl fragment
pCHB17BZO-A‘, of pCHB17012 in BamHl
of pBS—KS+ and -KS' Figure 4-3
(same orientation) this study
pCHB17BZO-A+ (as above, but

opposite orientation)

83
Table 4—2 (oont'd)

 

Plasmid Description or Source or
Construction Reference
pCHB1741, 2.1 kb Xbal, EcoRl
pCHB1742 fragment of pCHB17BZO—D+
and -A+ in Xbal, EcoRl Figure 4-9
of pCHB500 this study
pCHB1751, 1.2 kb BamHl fragment of
pCHB1752 pCHB17BZO—A+ in both
orientations in BamHl of Figure 4-9
pCHB500 this study
pCHB17NS-l, 1.74 kb Stul, Nael
pCHB17NS-ll (partial) fragment of

pCHB17BZO in both
orientations in

EcoRV of pBS-KS+ this study
pCHB1761, 1.8 kb Kpnl, EcoRl
pCHB1762 fragment of pCHB17NS-l
and -II in Kpnl, EcoRl Figure 4—9
of pCH8500 this study

 

50 ug/ml kanamycin, or 100 ug/ml ampicillin were added to the medium. In
order to score the lacZ phenotype of clones carrying pBluescript derivatives,
80 pl of a solution of 2% (w/v) X-gal in dimethyl formamide were spread onto
the agar plates prior to plating of the cells.

Construction of the cosmid library. High molecular weight DNA from
R. sphaeroides wild type cells grown photoheterotrophically in Sistrom’s

medium (chapter 3), was prepared using a modified procedure originally

84

designed for yeast chromosomes (7). Pelleted cells from a 200 ml culture were
resuspended in 13 ml 8% (w/v) sucrose, 50 mM Tris-Cl, 50 mM EDTA (pH 8.0)
buffer. Following the addition of 3.2 ml of the same buffer containing 10 mg/ml
lysozyme (final concentration 2 mg/ml), the suspension was incubated at
30°C. Lysis was monitored by mixing 100 pl of the solution with 100 pl of 1%
SDS. Sufficient lysis capability was generally observed after about 30 min.
Four ml aliquots of the sphaeroplast solution were lysed by pipeting the
solution very slowly under gentle shaking into a 6 ml solution of 3% (w/v) Na
Lauryl Sarcosine, 0.5 M Tris-Cl, 0.2 M EDTA (pH 9.0). The solution was heated
for 15 min at 65°C and subsequently cooled on ice. Aliquots (10 ml) of this
DNA solution were loaded onto a sucrose step gradient (bottom to top: 10 ml
50% w/v sucrose cushion, 10 ml 20% w/v sucrose. and 10 ml, 10% w/v
sucrose), and the samples were centrifuged at 27,000 rpm for 3 h at 15°C in a
SW27 rotor (Beckman). Brown, flooculent genomic DNA floating just inside the
50% sucrose cushion, was removed from the tubes and dialyzed overnight at
4°C in 1.5 l 10 mM Tris-Cl, 1 mM EDTA (pH 8.0) using dialysis tubing (10,000
MW cut- off). Aliquots of 0.5 ml were distributed into microfuge tubes using a
sterilized pair of scissors. Following soaking of the DNA for 10 min in 0.5 ml
buffer-equilibrated phenol and 0.5 ml chloroform, the DNA was dialyzed again
for two days with 4 changes of buffer in 1 | of 10 mM Tris-Cl, 1 mM EDTA (pH
8.0) at 4°C. The DNA, which appeared quite green, was stored at 4°C.
Aliquots of 40 pl (approximately 5-10 pg) from this DNA preparation

were partially digested for 1 h at 37°C with Sau3Al in a total volume of 80 pl of

85

the recommended enzyme buffer using between 10 and 0.001 units of the
enzyme. The solution was stirred periodically using a closed plastic pipet tip.
The reaction was stopped by addition of 20 pl of 50 mM EDTA solution. The
DNA was separated on a 0.4% (w/v) agarose gel as described further below.
intact lambda phage (48 kb), phage cut with Sell or with Hindlll were used as
size markers. Despite its "dirty" appearance, the DNA was readily digested.
Fragments in the size range from 20 to 35 kb were isolated by electroelution
from the respective gel slice and purified using standard DNA procedures
(17). About 150 ng of genomic DNA fragments were ligated into the Bglll site
of the cosmid vector (500 ng of prepared pLA2917). The vector DNA was
purified twice on a cesium chloride gradient, overdigested with Bglll, and
treated with calf intestinal phosphatase using standard procedures (17).
Ligation tests using the vector prepared in this way, revealed only very small
amounts of self-ligation. The resulting cosmids were packaged in vitro using
commercially available (Promega) lambda phage packaging extracts, and the
resulting phage were used to infect E. coli (HB101 or 817-1). Cell lines
containing the cosmids were selected on LB plates containing 10 pg/ml
tetracycline. Individual colonies were transferred to 96-well plates (200 pl LB
tetracycline medium per well). Following the addition of 20% (v/v) DMSO, the
plates were kept as frozen stocks at -70°C.

Triparental mating and complementation assay. Individual cosmids
carried by H8101 host cells were introduced into the H. sphaeroides squ

mutant using a triparental mating procedure. The squ mutant cells were

86

grown photoheterotrophically in 200 ml Sistrom’s medium as described in
chapter 3, until the culture reached mid to late log phase. An aliquot of 5 ml
from this culture was diluted with 20 ml Sistrom's medium, and 100 pl of a
concentrated suspension of MM294 cells containing the helper plasmid
pRK2013 were added. The helper strain was grown in 50 ml LB kanamycin
medium to midlog phase, and was resuspended in 1 ml LB following
centrifugation. This mixture was poured over a sterile cellulose acetate filter
(MSI, 9 cm diameter) under suction. Following the complete removal of the
liquid, the underside of the filter was blotted on sterile Whatman #2 filter paper
and placed onto a LB agar plate. Using a sterilized pronging device matching
48 wells of a 96-well microtiter plate, frozen material from the -70°C cosmid
library stocks was directly transferred onto the filter containing the lawn of
squ mutant and helper cells. Mating efficiency increased slightly when freshly
grown cultures inoculated from the stock cultures were used as donors.
However, since there was a concern that any additional growth of the cosmid
clones increases the chance for deletions, direct transfer from frozen stocks
was preferred. The plates, containing the filters carrying the three parent cell
lines, were incubated for 6 h at 30°C in the dark prior to transfer of the filters
onto agar plates containing Sistrom‘s medium with 0.8 pg/ml tetracycline.
These plates were incubated in a plexiglass box, which was illuminated using
a 60 Watt incandescent bulb with an irradiance of 50-100 pE/s per m2 at the
surface of the plates. The box was continuously flushed with a gas mixture of

95% N2 and 5% CO2 (flow rate approximately 20 ml/min). The selective growth

 

 

 

87

conditions permitted only the growth of exconjugants (squ mutant cells
containing cosmids), which became visible after about three to four days of
incubation at the positions where the pronging device had touched the filter.
These colonies were restreaked in patches onto Sistrom’s media agar plates
containing 0.8 pg/ml tetracycline. The sulfolipid phenotype of the exconjugants
was tested by extraction of the lipids, followed by TLC analysis on ammonium
sulfate-impregnated plates as described for the screening procedure used for
the isolation of the sulfolipid mutants in chapter 3.

Isolation of small, complementing subclones from cosmids. Purified
DNA from the complementing cosmid was partially digested with Sau3Al, and
the fragments were separated by gel electrophoresis as described above for
the preparation of the cosmid library. Fragments of the 2-3 kb, 3-4 kb, and 4-6
kb range were isolated and cloned separately into the Bglll site of pLA2917,
which was prepared as described above. Following ligation, the resulting
plasmids (23-27 kb) were used to transform E. co/i HB101. The cells were
plated onto LB/tetracycline plates, and the resulting colonies were restreaked
onto LB/tetracycline and LB/kanamycin plates to test for the presence of
inserts (loss of kanamycin resistance). Clones containing inserts were
transferred to microtiter plates, which were kept as frozen sublibrary stocks as
described above for the cosmid library. Complementing clones were identified
using the complementation assay described in the previous section.

Preparation of cosmid and plasmid DNA. Cosmid DNA was isolated on

a large scale using the alkaline lysis method followed by purification on a

 

88

vertical CsCI gradient (standard protocol in reference 17). Preparations
starting from 1 l cultures yielded about 200-400 pg DNA for low copy plasmids
(pLA2917 and pCHB500 derivatives), and about 1-2 mg for multicopy
plasmids (Bluescript derivatives) as determined by UV absorption.

Plasmid and cosmid minipreps were made using a rapid alkaline lysis
method (26) with the exception that the samples were never vortexed, and
that the DNA suspension was extracted once with 1 volume chloroform prior
to precipitation of the DNA. Preparations of single copy plasmids were started
from 3 ml overnight cultures, while preparations of multicopy plasmids were
started from 1.5 ml overnight cultures. The DNA was resuspended in 30 pl
10 mM Tris-Cl, 1 mM EDTA (pH 8.0). For cosmids, the complete sample was
digested with restriction enzymes and analyzed by gel electrophoresis. For
smaller single-copy plasmids 1/3, and for multi-copy plasmids 1/10 of the DNA
was analyzed in one reaction.

Restriction analysis and agarose gel electrophoresis. For detailed
restriction maps, CsCl gradient-purified DNA was used. The DNA (1 pg) was
digested in 20 pl of the recommended buffer for 1-2 h at the recommended
temperature. Following the addition of loading dye (dye #3, reference 17) the
samples were directly loaded into a well of an 0.8% (w/v) agarose gel.

DNA from minipreps was generally digested in 100 pl of the
recommended buffer. Following digestion, the DNA was precipitated by
adding 1/10 volume of 10 M LiCl and 2 volumes of cold ethanol. To remove

salts, the samples were washed once with 200 pl 80% (v/v) ethanol. The dried

89

DNA was resuspended in an appropriate amount of 10 mM Tris-Cl, 1 mM
EDTA (pH 8.0) and loading buffer prior to loading of the gel.

Gel electrophoresis was generally performed using 0.8% (w/v) agarose
gels in combination with Tris-Borate-EDTA buffer as described in (17).

Southem analysis. DNA fragments separated on agarose gels were
blotted onto Hybond-N+ (Amersham) nylon membranes using the alkaline
transfer method. Following soaking of the gel in 0.25 N HCI for 15 min and
rinsing with water, the DNA was denatured in 1.5 M NaCl, 0.5 N NaOH for 30
min and was then allowed to equilibrate in the transfer buffer (1.5 M NaCl,
0.25 N NaOH) for 15 min. The gel was placed into a plastic tray equipped
with a snap lid, and the bottom of the tray was covered with transfer buffer.
The filter, soaked in transfer buffer, was placed onto the gel and was covered
with 3 layers of whatman 3MM paper. Paper towels, out to size, were stacked
onto the 3MM paper to a height such that upon closing of the lid a slight
pressure was exerted onto the gel on the bottom of the tray. Following the
transfer of the DNA for about 5 hours, the filter was prehybridized in 50 ml of 1
mM EDTA, 0.5 M NaH2P04, 7% (w/v) SDS (pH 7.2) at 65°C for 15 min, prior to
addition of the labeled probe in a small volume, just enough to cover the filter,
of the same buffer inside tightly closed plastic boxes. The filters were
incubated in the presence of the probe overnight at 65°C with shaking. To
remove background label, it was usually sufficient to wash the filters 2X for 10
min at 65°C in 50-100 ml of 1 mM EDTA, 40 mM NaH2P04, 5% (w/v) SDS (pH

7.2). On some occasions the filters were reprobed after removal of the label

 

90

by repeated boiling in water.

Probes were prepared by labelling DNA fragments excised from low-
melting agarose gels using the random priming method (9).

Cloning of DNA fragments. DNA restriction fragments were usually
isolated from agarose gels using NA45 paper strips (Schleicher & Schflll). The
DNA fragment was electrophoresed into a small piece of NA45 paper placed
into an incision in the gel in front of the band. The paper was rinsed with
10 mM Tris-Cl, 1 mM EDTA (pH 8.0) and incubated in 400 pl of 1.5 M NaCl at
70°C for 1 h to elute the DNA. Following the precipitation of the DNA by
addition of 1 ml of cold ethanol, subsequent washing with 80% ethanol, and
drying, the fragment was directly ligated without further purification to the
prepared vector using a standard protocol (17). This method was efficient
enough to yield the desired clones, following transformation of competent cells
(CaCl2 method, reference 17) and analysis of the transformants. Clones
containing inserts were identified as colonies lacking B-galactosidase activity
on plates containing X-gal, prepared as described above.

Preparation of single stranded DNA for sequencing. E. coli DH5aF’cells
containing derivatives of the bluescript plasmid were grown on LB/ampicillin
plates at 37°C overnight. Three to five colonies were used to inoculate 25 ml
of prewarmed 2X YT medium containing 100 pg/ml ampicillin (see above) in a
250 ml flask. The cells were grown for about 2 h at 37°C under vigorous
shaking until a OD.50° of about 0.1 was reached. At this stage, 20 pl of VCS

phage (Stratagene; 5 x 1011 pfu/ml) were added, and incubation was

91

continued for about 5-6 h. Cells were removed by two consecutive
centrifugations at 12,000 rpm for 10 min in a Sorvall $834 rotor (Dupont). To
20 ml of the supernatant 5 ml of 20% (w/v) PEG°°° , 2.5 M NaCI were added,
and the samples were incubated for a minimum of 30 min at 4°C. To pellet the
phage, the samples were centrifuged at 15,000 rpm in an 8834 rotor for

20 min. The supernatant was completely removed under suction followed by
wiping the tube with a cotton tipped applicator. Following the resuspension of
the phage pellet in 200 pl 10 mM Tris-Cl, 1 mM EDTA, 300 mM Na acetate
(pH 8.0), the suspension was extracted twice with 1 volume of buffered phenol
and 1 volume of chloroform. The DNA was precipitated by addition of 3
volumes of cold ethanol, and washed and dried as usual. The DNA yield was
approximately 12 pg as estimated by agarose gel electrophoresis. This was
sufficient for about 540 sequencing reactions.

Sequencing. DNA sequences were determined by employing the chain
termination reaction method (21). In reaction mixtures, single stranded
templates ( about 200 ng per reaction) were sequenced using [a-“S1-dATP as
labelled substrate (6). To reduce sequencing artifacts due to the high GC
content of Ft. sphaeroides DNA, GTP was replaced by 7-deaza-dGTP (3)
provided in a sequencing kit purchased from United States Biochemical
Corporation. Modified T7 DNA polymerase (Sequenase, reference 23)
obtained along with the sequencing kit from the same supplier, was used as
the sequencing enzyme. The reactions were done according to the

instructions provided by the company. The labelled DNA fragments were

92

separated on 6% polyacrylamide, 8 M urea slab gels. Gels were double
loaded and run according to reference (19).

Preparation of synthetic oligonucleotides. Oligonucleotides used as
sequencing primers were synthesized on an Applied Biosystems 380A DNA
synthesizer using the phosphoramidite method (4) according to the
manufacturer’s instructions. Following I‘trityl-on'I synthesis, the oligonucleotides
were cleaved from the support with concentrated NH40H and deprotected by
incubation for 8 h at 55°C. The material was diluted with one third volume of
H20 and purified using a SepPac C13 cartridge (Waters) attached to a 5 ml
syringe. Prior to loading of the sample, the cartridge was flushed with 5 ml
acetonitrile followed by 5 ml of 2 M triethylamine acetate (pH 7.0). The flow
rate was always kept at 1-2 drops per second. The sample was pushed twice
through the cartridge, and the resulting flow through, still containing
oligonucleotides, was stored at -20°C. The cartridge was flushed 3 X with 5 ml
10% ammonium hydroxide and 2 X with 5 ml H20. The oligonucleotide was
detritylated with 2 X 1 min treatments of 5 ml 2% trifluoroacetic acid. Following
flushing of the cartridge with 2 X 5 ml H20, the oligonucleotides were eluted
with 2 ml 20% acetonitrile in water. The sample was evaporated to dryness,
and the DNA was resuspended in 50-200 pl of 10 mM Tris-Cl, 1 mM EDTA
(pH 8.0). The concentration was determined by measuring the absorption at

260 nm.

93
RESULTS AND DISCUSSION

Isolation of a cosmid complementing the squ mutant. A cosmid library
from Ff. sphaeroides wild type DNA was constructed using the 21 kb cosmid
vector pLA2917 (Figure 4-1). This vector contains the A phage cos site, which
is sufficient for the packaging of approximately 39-51 kb DNA fragments into
A phage particles by in vitro packaging extracts. The Bglll cloning site is
located in between the promotor and the coding region of the neomycin
phosphotransferase gene located on pLA2917 (5). Upon insertion of a DNA
fragment into this site, expression of the gene is prevented, allowing for the
identification of insert-containing clones as kanamycin-sensitive colonies. Two
libraries were prepared by infecting two different E. coli hosts, 817-1 (1728
clones) and HB101 (1920 clones). It was noted that upon infection with equal
amounts of phage under comparable conditions, infection of HB101 con-
sistently yielded about 4 X more colonies than S17—1. It was suspected that
S17-1 discriminated against certain sequences as compared to HB101. HB101
is partially deficient in the methyl-cytosine specific restriction system (mchC),
while the genotype of 817-1 with respect to this locus is not known. The
presence of the methylation specific restriction system has been known to
cause discrimination of methylated DNA during the construction of libraries
(18). Therefore, it was concluded that the library in HB101 was more represen-
tative for the H. sphaeroides genome and should be used for the screening
procedure. To test for the presence of inserts in this library, 768 clones were

replica-plated onto LB/kanamycin plates, and 113 (14.7%) were found to be

 

94

 

 

 

Tc Km
H 9
Sp P RV c Sp RV RI Sp RV SP Sp
2:) g : : 3 : t
I i
u“ .-‘-~
pii-.. ._m_... “““ ~
:— . _ .. . . l
b) , , w - 1 w
RV 5 R1 c 85 Hp Sp P‘ \P H RV RI
Ps
Tc __> I Km .
C) : , J. 3 ' 5 5 1 J
Sall cos BstEll Hpal Pstl 89111 Hindlll

a) <—> lkbp
b) ‘———-> lkbp

c) ‘——-—D lkbp

Figure 4-1: Detailed restriction map of pLA2917 (taken from ref. 1). The vector
is shown linearized at one of the Sphl sites. Abbreviations: Bg, Bglll; Bs,
BstEll; 0, cos site; H, Hindlll; Hp, Hpal; Km, kanamycin; Ps, Pstl; P, Pvull; R1,
EcoRl; RV, EcoRV; S, Sall; Sp, Sphl; and To, tetracycline. (a) pLA2917; (b)
expanded view of the region between the Pvull site in the tetracycline gene
and one of the EcoRI sites. (0) same as (b) showing antibiotic resistance
determinants, cos site, and unique restriction sites.

 

95

resistant, indicating that they might not contain inserts. Given the genome size
of R. sphaeroides of 5 X 10°bp (15) and the number of 1920 clones with an
average insert size of 25 kb, the HB101 library was expected to cover the
genome about four to five times. It was assumed that on average about one
positive clone could be found in about 400 clones tested.

As opposed to the 817-1 library, which does not require a third helper
strain for the transfer of cosmids into the squ mutant (2), cosmids
contained in HB101 had to be transferred using the helper plasmid pRK2013
in MM294. Triparental matings were performed as described in the methods
section. About 60-70 % of the cosmid clones were transferred (formation of
exconjugants) in this particular experiment. Testing of about 450 exconjugants
for restoration of the sulfolipid deficiency in the squ mutant by TLC of the
lipids yielded two complementing cosmid clones, of which one (pCHB1701)
was further analyzed.

Restriction analysis of cosmid pCHB1701. To establish a physical map
for the complementing cosmid pCHB1701, cesium chloride gradient-purified
DNA from this cosmid was cut with a series of restriction endonucleases.
Simple restriction patterns were obtained for the restriction enzymes Hindlll,
EcoRV, Bglll, EcoRI, and Pstl, listed in order of increasing complexity. EcoRV
had no site internal to the insert, while the other five enzymes cut the insert at
least twice. By double digesting with all possible combinations for the five
enzymes and applying Southern analysis to the resulting fragment patterns, it

was possible to establish the restriction map for pCHB1701 as shown in

96

 

 

2kb pCHBT7OT

R P B H 3 H3 .HR R R

I 1 I J I ll ' n _
PEP B P E PPP H B EHEB P P P P B ‘H
EU I l J In I J [11] l l l l l _l

*

L-o
IL;

E 9 pp H a pCHB‘l7H38

I 111 I

For

pCHB17B2B

2kb

 

Figure 4-2: Complete restriction map for the cosmid clone pCHB1701. (A)
overview showing insert and vector. (B) expanded view of the insert of
pCHB1701, and the two subclones pCHB17B2B and pCHB17H33. The
location of the smallest subclones complementing the squ mutant
(pCHB1751, pCHB1752) is indicated (cross hatched bar). The pLA2917 vector
part was drawn with thicker line. Restriction sites: H, Hindlll; R, EcoRV; E,
EcoRl; B, Bglll; and P, Pstl. (*) location of the 0.4 Pstl fragment,
crosshybridizing with the comparable fragment from the subclones
pCHB17011 and pCHB17012 .

 

97

Figure 4-2. One filter was prepared, which was probed with the complete
pLA2917 vector, and stripped and reprobed with the internal 3.5 kb Hindlll
fragment. Use of the vector as a probe allowed the assignment of the
vector/insert boarder fragments, as well as the vector internal fragments.
Similarly, by using the Hindlll fragment as a probe the central region of the
cosmid could be mapped. This information was sufficient to establish a
complete map for all the enzymes listed above except for Pstl. To obtain the
map for this enzyme, an approximately 10.0~kb Bglll fragment from the left
side of the cosmid, and a 10.5-Rb Hindlll fragment from the right side were
subcloned into the Bluescript vector pBS-KS“, giving rise to clones
pCHB17B2B and pCHB17H3B respectively (Figure 4-2). Comparative
restriction analysis of these two clones and pCHB1701 confirmed the map for
Hindlll, Bglll, and EcoRI. In addition, the Pstl map for the left half of the cosmid
could be solved, and the number of Pstl sites located on the 10.5-Rb Hindlll
fragment from the right side of the cosmid were determined. However, to
establish the order of the Pstl sites located on the right side of the cosmid, the
10.5-kb Hindlll fragment was excised from pCHB17H3B and partially digested
with Pstl. The separated fragments from the partial digest were blotted onto
nitrocellulose filters and probed with the 0.35-kb Bglll, Hindlll fragment of
pLA2917 (Figure 4—1). This fragment contained the right border fragment of
the pLA2917 cosmid vector, which was present in the 10.5-kb Hindlll fragment
of pCHB17H3B. This strategy allowed the identification of the right end of the

insert. Therefore, the sequential order of the Pstl sites beginning from the right

98

pCHB17012

R E P A P PAP AA H H! E
' l l I III II 1 n a

9

W pCHB17320

p

Fl'l'l

RH R A P PAP R
J 1 1 III I

*

1kb pCHB17011

 

Figure +3: Complete restriction map of two subclones, derived from
pCHBt701, which complement the squ mutant. The two clones pCHB17011
and pCHB17012 were aligned to match the common region around the 0.4 kb
Pstl fragment marked with (*). The location of the pCHB17012 BamHl
fragment giving rise to the pCHB17820 subclone series is indicated (cross
hatched box). Vector sequences were drawn with thick line. Restriction sites:
R, EcoRV; E, EcoRl; A, BamHl; P, Pstl; H, Hindlll.

99

end of the insert could be established. Summing up the size of each individual
Pstl fragment, the total length of the cosmid clone including the vector was
estimated to be about 44.5 kb. Given the low accuracy to which the size of
longer fragments could be determined on agarose gels, this value might have
been underestimated by 1-2 kb. It was possible that sites closer than

0.2-0.3 kb outside of the sequenced subclone containing the squ gene (see
below) were missed, as well as sites generating fragments of similar size
below 0.5 kb.

Isolation and characterization of two subclones from pCHB1701
complementing the squ mutant. To isolate smaller subclones from the
cosmid pCHB1701, three sublibraries with insert sizes of 2-3 kb, 3-4 kb, and
4-6 kb were prepared as described in the method section. Testing of 57
clones from the sublibrary containing the 2-3 kb inserts for complementation
of the squ mutant yielded one positive clone (pCHB17011). In addition, one
positive clone (pCHB17012) was found in the sublibrary (87 clones),
containing 3-4 kb inserts. Restriction analysis of purified DNA from both clones
indicated that pCHB17011 with its insert size of 5.2 kb as well as pCHB17012
(insert size 7 kb) obviously contained two independent “inserts (Figure 4-3).
They both shared a common region centering around a 0.4—kb Pstl fragment.
The 0.4-kb Pstl fragment isolated from pCHB17012 was used as probe
against restricted DNA from pCHB17011 and pCHB17012 (Figure 4-4), as well
as against DNA from pCHB1701 (Figure 4-5). The Southern hybridization

experiment confirmed that this fragment was common to all three clones, and

100

123456

0.6 -

Figure 4-4: Southern blot of pCHB17011 and pCHB17012 probed with the
0.4 kb Pstl fragment of pCHB17012. Lanes 1-3, pCHB17011; lanes 4-6,
pCHB17012. The DNA was restricted with BamHl (lanes 1 and 4), Pstl (lanes 2

and 5) and BamHl/Pstl (lanes 3 and 6). The position of size markers (kb) is
indicated on the left.

 

101

123456789101112131415

 

 

wells - _ . E-
23.1 - .

9.4 - .

6.6 - . . b .. .

4.3 - .

2.3 -

2.0 -

0.6-

Figure 4—5: Southern blot of pCHB1701 probed with the 0.4 kb Pstl fragment
from pCHB17012. The DNA was restricted with Bglll (1), Hindlll (2), Pstl (3),
EcoRl (4), EcoRV (5), Bglll/Hindlll (6), Bglll/Pstl (7), Bglll/EcoRl (8),
Bglll/EcoRV (9), Hindlll/Pstl (10), Hindlll/EcoRl (11), Hindlll/EcoRV (12),

Pstl/EcoRl (13), Pstl/EcoRV (14), and EcoRl/EcoRV (15). The position of size
markers (kb) is indicated on the left.

 

102

established the location of the 0.4-kb Pstl fragment on the map of cosmid
pCHB1701 (Figure 4-2). The fusion of the two inserts in both clones was
assumed to occur between the left BamHl site of pCHB17012 and the
neighboring Pstl site to the right (Figure 4-3). A 2.1-kb BamHl fragment
obtained in a partial digest from pCHB17012 (Figure 4-3) was cloned in both
orientations into the BamHl site of pBS-KS+ (pCHB17B20-N, -D+), and into
pBS-KS‘ (pCHB17B20-A, same orientation as pCHB17B20-0+). Based on the
information obtained from comparative restriction analysis and Southern
hybridization, this fragment was thought to contain the stretch of DNA
common to both subclones, and being responsible for the complementation of
the squ mutant.

Sequence analysis of pCHB17B20. The various pCHB17BZO Bluescript
clones were used to prepare single stranded templates for each strand. Using
the universal M13 primers, initial sequencing runs were done from both ends
of the complete clone as well as from a small number of deletion clones
created by using internal restriction sites. Based on the sequence information
obtained from these runs, oligonucleotides were synthesized and used as
sequencing primers (Table 4-3). The bulk of the sequence was obtained by
using these synthetic oligonucleotide primers (Figure 4-6).

The sequence (Figure 4-7), confirmed by sequencing of both strands,
had a GC-content of 70%, which is typical for R. sphaeroides DNA (15).
Analyzing the sequence for open reading frames initially revealed three larger

frames (Figure 4-6), which could encode proteins of about 33 kD (ORF1),

 

103

Table 43: List of synthetic oligonucleotide primers used
for sequencing of pCHB17B20.

 

Primer Sequence

C824 5' - CGT, GGA, ACA, GAG, TGA, TGT -3'
C825 5' - GAT, CGA, GAT, GCT, GCG, CGA —3'
C826 5' - GCC, AAC, ATG, GAA, GGC, GAA -3'
C828 5' - GGC, CTC, CTG, CTC, TCG, CAT -3'
C829 5' - TGC, CAC, GGC, GTG, CGG, CGA -3'
C830 5' - GAT, CCA, TGC, GTG, CGC, ACT -3'
C831 5' - GGC, CAT, TTT, CCT, CAT, GCT -3'
C832 5' - CAC, AAG, GTG, CCC, TAC, CAG -3'
C833 5' - ACA, CGG, CTC, ATG, TCA, TGC -3'
C834 5' - CTC, CAC, CAC, ATC, CTC, GCG -3'
C835 5' ‘- AAG, TTC, GAC, CTG, CCT, GTG —3'
C836 5' - GTG, CAA, GAG, CAT, GAC, GAA -3'

C838 5' - TGA, AGG, ACG, ATC, CGA, CGC -3'

 

104

Sequencing pCHB17820

_ he“. he
_.__he —

 

 

 

 

 

 

onFsr C>
ORF1L C>
A N O P NPANP A UA
I_ l l I lLlll I [I
<3 . ORF2

 

 

Figure 4-6: Strategy for sequencing pCHB17820. Only the 2.1 kb BamHl
fragment from pCHB17012 is shown. Sequencing runs (closed arrows) primed
with synthetic oligonucleotides are preceded by an open circle. The location
and orientation of three open reading frames (ORF1, ORF2, and ORF3) are
indicated by open arrows. Restriction sites: A, BamHl; N, Neel; O, Xhol; P,

Pstl; U, Stul.

 

 

105

Figure 4—7: DNA sequence of pCHB17B20. Only the proposed coding strand
for the squ gene is shown. The protein sequence predicted by translation of
the proposed squ coding sequence, is given below the sequence. The
locations of stop codons (8”) in frame with the coding sequence, a potential
ribosome binding site (SD), 4 possible initiation codons (l"‘), and restriction
sites, used to construct deletion clones are marked above the sequence by a
line.

1 GGATCCGGCTCGACCCGCTGAACCCGCAGACGGCGGCGCGCATGTCGACCGCCTTCGAGA
61 CCCGGGCCCGCTACGATGCGGACCGGCAGGGCATGATACTGGCCGAACTCGACCGCATCC
1 21 TCGCCACGCCGGGCCTCAGCCGCGATCTGGGCGAGATGGCGGGGCGGATGAGGCACGCAT
181 GAGCGGGTGGGAACAGTTGGACGGGAAGCGGCGTTCT—IGACCGGCGGATGACTGCCGAGCG
241 CAGAngGTGCAAAGG—ilcGGCGCTTiin—riGéCCGGGACAACCGGGGCTCATGTCATGCGAG

 

301 AGCAGGAGGCCCCGGGTGTCGCTCATGATGCAGTCTCGAATCGATCCGCTCATCGAAGAG

MSLMMQSRIDPLIEE
361 CGTGCCCCCTGGCTCTTCTCCGGCCGTCCCCATCATGAAGCCTCGCGCAGACTTCTGTGC
16 RAPWLFSGRPHHEASRRLLC

421 CGCGTTCTGGGCTACGACCGCACCGTCGAGATCGGAGAGCAGCTGAAGGACGATCCGACG
36 RVLGYDRTVEIGEQLKDDPT

481 CCCGAGATCATGCGCCGCGTGGCGCAGATGATCGCGCGGGACGTAGAGGTCTCGGGTCTC
56 PEIMRRVAQMIARDVEVSGL

541 GAGAACCTGCCGCGCAACGGCCCGGCGCTGGTCGTCTGCAACCATCCGACCGGCATCGCC
76 ENLPRNGPALVVCNHPTGIA

601 GACGGGGTGATCCTCCACCACATCCTCGCGCCCCTGCGGCCCGATCTCTTCTTCTACGCC
95 DGVILHHILAPLRPDLFFYA

661 AACCGCGACATTCTTCGCGTGCTGCCGCAGATGGAGGACCTGATCTGCCCGGTCGAATGG
116 NRDILRVLPQMEDLICPVEW

721 CGGCAGGAGAAGCGGAGCCTGCAGAAGACGCGCGAGACGATGGACTACACCCGCCGCGCC
136 RQEKRSLQKTRETMDYTRRA

781 ATCGAGGCCGGGCGGCTGGGCGTAATCTTCCCCTCCGGGCGGCTGGCCAAGCGGACGGGA
156IEAGRLGVIFPSGRLAKRTG

841 CTGAAGCTCATGGAGCGGCCGTGGATGGCCTCCGCTGCGATGATCGCGCGGAAGTTCGAC
176LKLMERPWMASAAMIARKFD

901 CTGCCTGTGGCGCCGGTGAACATCCGGGCGCGCAACTCGACGCTGTTCTACCTCTTCGAC
196 LPVAPVNIRARNSTLFYLFD

 

106

Figure 4—7 (oont'd)

961
216

1021
236

1081
256

1141
276

1201
1261
1321
1381
1441
1501
1561
1621
1681
1741
1801
1861
1921
1981
2041

CTGATCCACCCCACGCTGCGGGACATCACTCTGTTCCACGAGGTGCTGAACAAGCATCGC
L I H P T L R D I T L F H E V L N K H R

CAGCCGTTCCGACTGACGATGGGCCGGCCGATCCCGGCCGAGCGTCTGCCGTCGAAGTCC
Q P F R L T M G R P I P A E R L P S K S

GAAGAGGGGATCGAGATGCTGCGCGAGGCGACGCTCGCGCTCGGCGGGCAATATTCGCCT
E E G I E M L R E A T L A L G G Q Y 5 P2
_§L
GCCGTCAGCATCCTGCAGATGACGAGCCTGCGCGGCGCGCTCCGCTACGCACGCGGGTAG
A V S I L Q M T S L R G A L R Y A R G
BamHI
GGGATCCCGGAGCGCGCGGCTGGCCGGACAGGCGGCCGCGCCTCCTCCCCACCCCTCCGC
GGCCGGCAGGAGCTGTCGCCGCTGCGTTCCGGCAGCCGCGGGCGGCCATTTTCCTCATGC
TCCTGCAGTGACCGGAGGACCCGTCGCCGCCGCGTCCAACGGCAGGCGCAGGACGCCCCC

TTCGTCATGCTCTTGCACGGCGGCGCGGGATGGCGTCAACTCGCGCCATGACCGGACCGC

 

GCTACALLLLBL I U: I LhAGAbLL I (11.1.1. I LLALLbI 1 Lu. I HA: I LGGCCCCGAGACGC
AGGAGCGTGCGCGTGGCCGCCCGTTCCGCGCCCGCTTGGGCGCGAACGAGAGCCTCTTCG
GCCCCTCGCCGCACGCCGTGGCAGCCATGGCCGCGGCCGCCGCCGAGGTCTCGCTGTATG
GCGATCCAGAGATCCATGCGCTGCGCACTGCCATCGCGGCCCATCACGGCGTGGCGCCGG
ATCAGGTGACGGTGGGCGAGGGCGTCGACGGGCTTCTGGGCAACCTCGTCCGGCTGCTTG
TCGACGAGGGCGCGCCGGTCGTGACCTCGGAGGGCGCCTATCCGACCTTCGCCTTCCATG
TGGCGGGGTTCGGCGGCCGCCTGCACAAGGTGCCCTACCAGGGCGATCACGAGGATCCCG
AGGCGCTGGCGGCCGAGGCCCGGAGGTCGGGGCCCGGATCGTCTATTTCGCCAATCCGGA
CAATCCGATGGGGTCGCACCATCCTGCCGCCTCCGTCCGCGCCCTGATCGAGGCGCTGCC
CGAGGGCACGCTGCTCGCGCTCGACGAgggéTATGTGGAACTGGCCCCCGAGGGCACGGC
GCCCGAGATCGCCCCCGAGGATCC

 

  
   
  
   

107
A
-15 -5 O

squ A G C %C C C C G G G T G
pucB T T G A C G A C A C A G T G
pucA A A G A C T G A C A T G
puhA C ' A C C C C T T A A A T G
pqu AGGGAAGCATG
pufM TCCCGI-SI‘ES "ICATCAATG
pufB GATCCG: 'TAGCATG
puffl G A C A A G A T C A T G
B

1653' HO-UCUUUCCUCCUCCACUA 5'
rRNA ******
squ 5'AGC

 

:CCCCGGGTG 3'

Figure +8: Comparison of the putative squ ribosome binding site to binding
sites from various Fl. sphaeroides genes (A) and to the 3’-end of the 168 rRNA
from Ff. sphaeroides (B). Bases matching the complementary sequence of the
16S rRNA 3’-end representing the ribosome binding site are shaded.
Sequences were taken from: pucB and pucA (14), puhA (8), pqu (25), pufM
(24), pufB and pqu (13), and H. sphaeroides 16$ rRNA 3’-end (11).

108
26 kD (ORF2) and 32 kD (ORF3). However, complementation analysis of

deletion clones, as described below, excluded two of the three frames (ORF2
and ORF3). The maximum size of the third reading frame (ORF1) was found to
be determined by the position of two in-frame stop codons (Figure 4-7, S""‘).
Downstream of the first stop codon, four possible in-frame initiation codons
were located, of which only the second, a GTG codon (Figure 4-7, I2), was
preceded by a putative ribosome binding site. This putative ribosome binding
site showed perfect sequence match complementary to the 3’ end of the
R. sphaeroides 16$ rRNA, as well as correct spacing as compared to ribosome
binding sites from other H. sphaeroides genes (Figure 4-8). It has been shown
that the presence of a ribosome binding site consensus sequence, correctly
spaced upstream of the initiation codon, can greatly affect the translation of the
respective mRNA (12). The presence of a sequence satisfying the requirements
for a correctly spaced ribosome binding site only upstream of the second
initiation codon, suggested that a GTG codon 7 bp into the 3’-direction was the
best candidate for the translational initiation site of the squ gene. Although the
usual initiation codon is ATG, the use of GTG has been observed previously.
One example for a GTG initiation codon found in R. sphaeroides is the pucB
gene (Figure 4-8).

Complementation analysis using various DNA fragments of pCHB17820
cloned into an expression vector. To directly determine, which of the open
reading frames in the 2.1-kb BamHl fragment of pCHB17012 represented the

squ gene, the complete insert as well as deletion fragments of pCHB17B20

109

were cloned in both orientations behind a promotor to create transcriptional
fusions, which would result in transcription of the various fragments in

R. sphaeroides. The clone pCHB17B20 by itself could not be employed for
complementation analysis, since Bluescript derivatives can not replicate in

R. sphaeroides. Instead, the vector used was pCH8500, which was constructed
by inserting the promotor region of the cytochrome-c2 gene (cycA) from
thodobacter capsu/ata into the poly-linker of pRK415 (construction and testing
of pCH8500 are described in Appendix B). Insertion of the complete fragment
of pCHB17820, corresponding to the 2.1-kb BamHl fragment of pCHB17012,
into the BamHl site of pCH8500 in both orientations, gave rise to clones
pCHB1741 and pCHB1742 (Figure 4-9), both of which restored the wild type
phenotype when present in the sulfolipid deficient squ mutant. This result
confirmed that the 2.1 kb fragment of pCHB17012 was indeed capable of
complementing the squ mutant. In addition, these data suggested that the
fragment contained a promotor, making the complementation independent of
the insert orientation and therefore independent of the presence of the external
cycA promotor provided by pCH8500. There were two possible origins for a
promotor present in the 2.1 kb BamHl fragment of pCHB17012: first, the
endogenous promotor naturally occurring in front of the gene was located on
this fragment; or second, the squ open reading frame was part of an operon,
and the promotor necessary for complementation was provided by the second
inserts found in the complementing subclones pCH817011 and pCHB17012

(Figure 4-3). The fact that the only two positive subclones found, both

 

110

contained two inserts, supported the second possibility.

A 1.2 kb BamHl subfragment of pCHB17B20 extending from the left end
eo3of the insert to the BamHl site directly downstream of the stop codon of the
putative squ gene (Figure 4-7), was cloned in both orientations into the BamHl
site of pCH8500 giving rise to subclones pCHB1751 and pCHB1752
(Figure 4-9). Both clones transferred into the squ mutant were able to restore
the wild type phenotype, suggesting that the 1.2 kb BamHl subfragment of
pCHB17820 contained a promotor preceding the squ coding sequence.
These two subclones were the smallest clones capable of complementing the
squ mutant. The location of this fragment on the cosmid clone pCHB1701 is
shown in Figure 4-2. The 3’-end of one of the three potential open reading
frames ((Figure 4-9, ORF3) was not included in this fragment, making it unlikely
to represent the squ gene. However, there were still two intact reading frames
(ORF1, and ORF2; Figure 4-9) located on opposite strands of this fragment.

An experiment was designed to determine the coding strand for the
squ gene, and thereby the correct open reading frame by testing the
orientation of the insert with respect to the cycA promotor, necessary for
complementation. To observe orientation dependent complementation using the
vector pCH8500, the squ coding sequence lacking a promotor had to be
cloned in proper orientation behind the cycA promotor provided by pCH8500.
A Nael site upstream of the putative initiation codon of ORF1 was used to
remove most of the upstream sequence of the putative squ gene (Figure 4-7).

Using this site, a 1.8-kb partial Nael/Stul fragment from pCHB17B20 was cloned

 

 

 

 

 

111
Clone SL
ORF3 :2) __
ORF1:{>
A N 0 P NPANP A UA
; I I I I Illl I II 17320 /
<=ORF2
Pcyc XA N O P NPANP A UAE
_LI_I I I I I IIII I 1].] 1741 +
EA N O P NPANP A UAX Pcyc
Ll I I I IIIII I IUA— 1742 «I-
‘—
Pcyc A N O P NPA
1751 +
A N O P NPA POYO
T752 *
t It
Pcyc KN O P NPANP A UE
—.LLI I I I [III I 1.1 1761 +
i w
EN 0 P NPANP A UK POYO
L.I_ I I I III I Li 1762 -

 

 

Figure 4—9: Complementation analysis using derivatives of pCHB17820 cloned
into the expression vector pCH8500. The 2.1 kb BamHl insert of pCHB17B20 is
shown at the top with the orientation and location of three open reading frames
(ORF1, ORF2, and ORF3) indicated by open arrows. For each clone shown,
three independent constructs were prepared, which gave identical results in
complementation assays. SL: presence (+) or absence of the sulfolipid in lipid
extracts of the sulfolipid deficient squ mutant containing the indicated clone
(clones are listed without the pCHB prefix); pCHB17B20 can not be tested (/).
Pcyc: cycA promotor (filled arrows), the direction of transcription is indicated by
the arrow head. Small stretches of Bluescript poly-linker region flanking the
inserts are indicated with thicker line. Restriction sites: A, BamHl; K, Kpnl; N,
Nael; O, Xhol; P, Pstl; U, Stul; X, Xbal. Locations of Nael and Stul sites, no
longer cleavable due to the cloning process, are marked with an asterix (*).

 

112

in both orientations into the EcoRV site of the Bluescript vector. From there, the
fragment was cloned in both orientations into the pCH8500 poly-linker using
Kpnl and EcoRI sites, giving rise to pCHB1761 and pCHB1762 (Figure 4-9).
Due to the cloning procedure, the Neel/Stul fragment was flanked by small
stretches of Bluescript poly-linker sequence in pCH8500. In a complementation
assay, only pCHB1761, as opposed to pCHB1762, was able to restore the wild
type phenotype. In pCHB1761 the cycA promotor was located upstream of
ORF1 reading in the direction of the putative squ gene (shown in Figure 4—7),
while in pCHB1762 the situation was reversed. Therefore, only ORF1 but not
ORF2 could represent the squ gene.

The complementation analysis of the pCHB17820 derivatives, cloned into
the expression vector pCH8500 as described above, clearly supported the
findings of the sequence analysis data, and localized the squ coding
sequence. The orientation dependence of complementation shown by clones
pCHB1761 and pCHB1762 indicated the direction of gene expression, which
coincided with the orientation of the putative squ coding sequence shown in
Figure 4-7. The results of the deletion and complementation experiments
excluded two alternative open reading frames as possible candidates for the
coding sequence of the squ gene.

Amino acid sequence analysis of the putative squ gene product. The
putative squ coding sequence, as suggested by the results discussed above,
was translated into the corresponding amino acid sequence (Figure 4-7)

consisting of 295 amino acids. A molecular weight of 33.6 kD could be

113

 

 

 

 

5
4 -
3 -
ti
o 2 ‘
Z
i ‘ '
E 0
8 —1
9 —2 -
I
-3 _
-4 _
_5 I I I I I I
O 50 1 00 150 200 250 300

AMINO ACID #

Figure 4-10: Hydropathy plot of the predicted squ amino acid sequence.
Amino acids (304) from the start of the first initiation codon to the second stop
codon (see Figure 4-7) were included in the analysis. The analysis was done
according to Kyte and Doolittle (16). The segment size used was 10 amino
acids.

 

114

calculated for the resulting protein. A hydropathy analysis was done on the
predicted protein according to the Kyte and Doolittle method (16) to determine
whether it might be membrane bound or soluble. Every amino acid is given a
value according to its hydrophobicity. Using a moving-segment approach the
average hydropathy within a given segment of predetermined length, usually
between 7 and 11 amino acids, is determined along the amino acid sequence,
and the consecutive scores are plotted along the sequence. A midpoint line is
indicated that corresponds to the grand average of the hydropathy of the
amino acid composition found in most of the sequenced proteins. Regions of
the sequence scoring above the mid point line are hydrophobic, while those
regions scoring below the midpoint line are hydrophilic. The analysis result for
the putative squ gene product is given in Figure 4-10. The lack of extensive
regions of continuous hydrophobicity, which are indicative of the membrane
spanning regions of membrane proteins, suggested that the protein is soluble,
rather than membrane bound.

To investigate the possible function of the predicted squ protein, the
amino acid sequence was used to screen two protein data banks for
homologous proteins of known function. No known protein showed sufficient

homology to the putative squ protein, to draw any conclusions.

CONCLUSIONS
Two cosmids complementing the sulfolipid deficient squ mutant were

isolated from a H. sphaeroides wild type DNA cosmid library. Complementation

115

was also observed with two smaller subclones, derived from one of the
complementing cosmids. A fragment covering a region common to both of the
two clones was sequenced. Analysis of the smallest fragment capable of
complementing the squ mutation permitted the identification of a single open
reading frame. The maximum size of the putative squ coding sequence was
found to be determined by two flanking in-frame stop codons. Of the four
possible initiation codons, only one was preceded by a potential ribosome
binding site, satisfying the required criteria concerning complementarity to the
3’-end of the 168 rRNA as well as correct spacing. The respective GTG codon
was the only putative initiation codon of any of the observed reading frames,
preceded by a perfect ribosome binding site. Based on this circumstantial
evidence, the protein sequence outlined in Figure 4-7 was proposed to be the
squ gene product responsible for the complementation of the sulfolipid
deficient squ mutant of Ff. sphaeroides. However, it should be pointed out that
only purification of the protein from wild type Fisphaeroides cells, followed by
sequencing of the N-terminus of the protein can clarify the exact location of the
initiation codon.

The negative result obtained from screening protein data banks for
homology with the putative squ gene product indicated that the squ gene
encodes a novel protein involved at one, as yet, unknown level in sulfolipid
biosynthesis. It is expected that the analysis of the 35S-Iabelled, water-soluble
compounds found to accumulate in the squ mutant (see chapter 3) will

provide some clues about the function of squ gene product. With the squ

116

gene in hand isogenic lines can be constructed for comparative biochemical
and physiological experiments. The squ coding sequence can be expressed
using E. coli expression vectors, and antibodies raised against the product will
provide valuable tools to study the squ gene product.

In summary, based on the data provided in this chapter, experiments
can be designed to further confirm the identity of the squ gene as well as to
elucidate the function of the squ gene product. It is expected that these
experiments will provide an answer to the remaining questions about the exact
role of the squ gene in sulfolipid biosynthesis in R. sphaeroides in the near

future.

REFERENCES

1. Allen L.N., Hanson RS. (1985). Construction of broad-host-range
cosmid cloning vectors: Identification of genes necessary for
growth of Methylobacterium organophi/um on methanol. J.
Bacteriol. 161:955-962

2. Bachmann B.J. (1987) In Escher/aha coli and Salmonella
typhimurium, Cellular and Molecular Biology. Ed. Neidhardt F.C. et
al. ASM, pp 1190-1219

3. Barr P.J., Thayer R.M., Laybourn P., Najarian R.C., Seela F., Tolan
DR. (1986). 7-Deaza-2 ’-Deoxyguanosine-5 LTriphosphate:
Enhanced Resolution in M13 Dideoxy Sequencing. Bio Techniques
4:428-432

4. Beaucage S.L., Caruthers, M.H. (1981). Deoxynucleoside
Phosphoramidites - A new class of key intermediates for
deoxypolynucleotide synthesis. Tetrahedron Lett. 22:1859-1862

5. Beck E., Ludwig G., Auerswald E.A., Reiss 8., Schaller H. (1982).
Nucleotide sequence and exact localization of the neomycin
phosphotransferase gene from transposon Tn5. Gene 192327-336

11.

12.

13.

14.

117

Biggin. M.D., Gibson T.J., Hong G.F, (1983). Buffer gradient gels
and 35$ label as an aid to rapid DNA sequence determination.
Proc. Natl. Acad. Sci. USA 80:3963-3965

Carle G.F, Olson M.V. (1984). Separation of DNA molecules from
yeast by orthogonal-field-alternation gel electrophoresis. Nuc. Acid
Res. 12:5647-5664

Donohue T.J., Hoger J.H., Kaplan S. (1986). Cloning and
expression of the Ff. sphaeroides reaction center H gene. J.
Bacteriol. 168:953-961

Feinberg A.P., Vogelstein B. (1984). Addendum: A technique for
radiolabeling DNA restriction endonuclease fragments to high
specific activity. Anal. Biochem. 137: 266-267

Figurski D., Helinski D.R.(1979). Replication of an origin-containing
derivative of plasmid RK2 dependent on a plasmid function in
trans. PNAS 76:1648—1652

Gibson J., Stackebrandt E., Zablen L.8., Gupta R., Woese CR.
(1979). A phylogenetic analysis of the purple photosynthetic
bacteria. Cur. Microbiol. 3:59-64

Gold L., Pribnow D., Schneider T., Shinedling 8., Swebilius Singer
8., Stormo G. (1981). Translational initiation in prokaryotes. Ann.
Rev. Microbiol. 352365-403

Kiley P.J., Donohue T.J., Havelka W.A., Kaplan S. (1987). DNA
sequence and in vitro expression of the 8875 light-harvesting
polypeptides of R. sphaeroides. J. Bacteriol. 169:742-750

Kiley P.J., Kaplan S. (1987). Cloning, DNA sequence, and
expression of the R. sphaeroides light-harvesting 8800-850-a and
8800-850-8 genes. J. Bacteriol. 169:3268-3275

IGley P.J., Kaplan S. (1988). Molecular genetics of photosynthetic
membrane biosynthesis in R. sphaeroides. Microbiol. Rev.
52:50-69

Kyte J., Doolittle R. F. (1982). A simple method for displaying the
hydropathic character of a protein. J. Mol. Biol. 157:105-132.

17.

18.

19.

20.

21.

23.

24.

25.

26.

118

Maniatis T., Fritsch E.F., Sambrook J. (1982). Molecular Cloning, A
Laboratory Manual. Cold Spring Harbour Laboratory, New York.

New England BioLabs. Restriction of foreign DNA by E. coli.
Catalog (1990-1991), pp 143-144

Parkinson C., Cheng S.Y. (1989). A convenient method to increase
the number of readable bases in DNA sequencing. Bio
Techniques 72828-829

Raleigh E.A, Lech K., Brent R. (1989). In: Current Protocols in
Molecular Biology. Eds. Ausubel PM et al. Publishing Associates
and Wiley lntersience; New York, Unit 1.4

Sanger F., Nicklen S., Coulson AR. (1977). DNA sequencing with
chain termination inhibitors. Proc. natl. Acad. Sci. USA 74:5463-
5467

Simon R., Priefer U., Puehler A. (1983). A broad host range
mobilization system for in vivo engineering: transposon
mutagenesis in gram negative bacteria. Bio/Technology 1:37-45

Tabor S., Richardson CC. (1987). DNA sequence analysis with a
modified bacteriophage T7 DNA polymerase. Proc. Natl. Acad.
Sci. USA 84:4767-4774

Williams J.C., Steiner L.A., Feher G., Simon MI. (1984). Primary
structure of the L subunit of the reaction center from R.
sphaeroides. Proc. Natl. Acad. Sci. USA 81:7303-7307

Williams J.C., Steiner L.A.. Ogden R.G., Simon M.l., Feher G.
(1983). Primary structure of the M subunit of the reaction center
from Fl. sphaeroides. Proc. Natl. Acad. Sci. USA 80:6505-6509

Zhou, C., Yang Y., Jong A.Y. (1990). Mini-prep in 10 minutes. Bio
Techniques 8:172-173

 

 

CHAPTER 5

ISOLATION AND CHARACTERIZATION OF A CLONE COMPLEMENTING THE

SULFOLIPID-DEFICIENT squ, AND squ MUTANTS OF H. SPHAEHOIDES

ABSTRACT

Using a complementation assay, a cosmid (pCHB1601) was isolated
which was able to restore the wild type phenotype when present in the squ,
and squ mutants, but not when present in the squ mutant. A detailed
restriction map for this clone was prepared. From this cosmid a 2.4 kb
subfragment containing clone (pCHB16014) was isolated which
complemented the squ mutant. Sequence analysis revealed an open reading
frame encoding a putative protein of 404 amino acids. The N-terminus of this
putative protein showed striking homology to the N-terminal amino acid
sequence of UDP-glucose epimerase, encoded by the bacterial gaIE gene
and the ga/10 gene of yeast. Further deletion and complementation analysis
of the pCHB16014 subclone, as well as insertion of a promotor-less
chloramphenicol-acetyl—transferase gene into the clone to determine the
predominant direction of expression, supported the idea that the observed
open reading frame might be the coding sequence of the squ gene.

A second subclone (pCHB16018) was isolated from the cosmid
pCHB1601, which complemented the squ mutant. Restriction and Southern
analysis revealed that this clone was different from the subclone

complementing the squ mutant. Cloning of a restriction fragment from the

119

120
cosmid pCHB1601, which contained sequences common to the pCHB16018

subclone, into an expression vector and subsequent complementation
analysis, revealed the location and orientation of the putative squ gene. It
was proposed that the squ gene is located downstream of the squ gene
and is transcribed in the same direction. This raises the possibility that squ
and squ are two genes involved in sulfolipid biosynthesis, which might be

organized in an operon.

INTRODUCTTON

In the previous chapter the isolation and analysis of a cosmid clone
(pCHB1701) was described, which complemented the squ mutant. For the
same reasons described in the introduction to the previous chapter, it was
desired to isolate the genes complementing the squ and squ mutants. It
was hoped that detailed physical mapping of the DNA sequences
complementing all three mutants might reveal that they affect three different
genes involved in sulfolipid biosynthesis. In addition, it was assumed that the
complementing clones for the mutants will provide the basis for the
construction of isogenic lines, which are important for further biochemical and
physiological analysis.

In this chapter, the analysis of a cosmid (pCHB1601) complementing
the squ and squ mutants is described. The strategy employed was, in
principle, the same as for the identification of the squ gene. However, the

detailed analysis of the putative squ gene was complicated by the fact that

121

relatively few suitable restriction sites for the construction of deletion clones
were present, and that stop codons rarely occurred in some of the reading
frames, making it difficult to identify the coding sequence of the putative squ
gene. The location and orientation for the squ gene with respect to the squ
gene were determined, but attempts to find the putative coding sequence

were not made during the course of this study.

MATERIALS AND METHODS

Bacterial strains and plasmids. The bacterial strains used for the
experiments described in this chapter were the same as those described in
Table 4-1 of the previous chapter, except that Bluescript plasmids were
propagated in DH5a, and that for the preparation of single stranded DNA,
DH5aF’ was replaced by XL1-Blue ([ F’::TN10 proA+B+ laclq A(lacZ)M15 ]
recA1 endA1 gyrA96(nal') thi hst17 (rK-mK+) supE44 relA1 lac; reference 3).

In addition to the vectors pBS-KS“, pBS-KS‘, pLA2917, pRK2013, and
pCH8500 described in Table 4-2 in the previous chapter, the plasmids listed in
Table 5—1 were used in the experiments described in this chapter.

Growth and media. Cultures of Ft. sphaeroides were grown as
described in the method section of chapters 3 and 4. Growth conditions and
media for E. oo/i were described in chapter 4.

Triparental mating and complementation assay. Isolation of the clones
complementing the squ and squ mutants was done using triparental

matings, followed by analysis of lipid extracts obtained from the exconjugants

122

Table 5—1: Plasmids used in the experiments described in chapter 5.

 

Plasmid Description or Source or

Construction Reference
pBR328 pBR322 derivative containing the

CAT gene (1)
pCHB1601 cosmid clone complementing the this study

squ and squ mutants Figure 5-2
pCHB1611 pCHB1601 digested with this study

Hindlll and recircularized Figure 5-3
pCHB1621 11 kb Hindlll fragment of pCHB1621 this study

in Hindlll of pLA2917 Figure 5-3
pCHB16014 2.4 kb fragment of pCHB1601 in pLA2917, this study

complementing the squ mutant Figure 5—7
pCHB100, partial 2.9 kb Pstl, Hindlll fragment
pCHB200 from pCHB16014 containing the insert

and flanking pLA2917 sequences in the

Pstl, Hindlll sites of pBS-KS+ an -KS' this study
pCHB16014D-B pCHB16014 cut with Bglll and this study

recircularized Figure 5—7
pCHB16014D-AB pCHB16014 cut with BamHl and Bglll, this study

and recircularized Figure 5-7
pCHB16201, 2 kb Sall, EcoRl partial fragment of this study
pCHB16202 pCHB16014 first cloned into the Sall, Figure 5-7

EcoRl, sites of pBS-KS'; excised from

there with Kpnl and EcoRI, and cloned

into the Kpnl, EcoRl sites of pCH8500

to give pCHB16201; or excised with

Kpnl and Xbal, and cloned into the Kpnl,

Xbal sites of pCH8500 to give pCHB16202 this study

(opposite orientation)

Figure 5-7

 

Table 5-1 (oont'd)

123

 

pCHB16151,
pCH816152

pCHB16014CAT-1,

pCHB16014CAT-2

pCHB16018

pCHB181,
pCHB182

1.5 kb Sall, EcoRI fragment of

pCHB16014 first cloned into the

Sall, EcoRl sites of pBS-KS';

excised from there with Kpnl and

EcoRI, and cloned into the Kpnl,

EcoRl sites of pCH8500 to give

pCHB16151; or excised with Kpnl and

Xbal, and cloned into the Kpnl,

Xbal sites of pCH8500 to give this study
pCHB16152 (opposite orientation) Figure 5-7

1 kb Sau3Al fragment of pBR328
pBR328, containing the CAT gene
coding sequence, cloned in both

orientations into the BamHl site this study
of pCHB16014 Figure 5-8
9.5 kb subclone of pCHB1601 this study
complementing the squ mutant Figure 5-11

4.2 kb Bglll fragment of pCHB1601
cloned in both orientations into this study
the BamHI site of pCH8500 Figure 5-11

 

124

as described in the method section of chapter 4.

DNA manipulations and analysis. The construction of small subclones
by partial Sau3AI digestion from the complementing cosmid, as well as
general methods for DNA manipulations such as preparation of plasmid DNA,
cloning, restriction analysis, and Southern analysis were done following the
protocols outlined in detail in chapter 4.

Sequence analysis. Single stranded + and - strands for sequencing
were prepared from the Bluescript clones pCHB100 and pCHB200
respectively and their deletion derivatives as described in chapter 4, except
that XL1-Blue was used as host strain instead of DH5aF’. All growth media for
XL1-Blue contained 10 pg/ml tetracycline to maintain the F-episome. An
incomplete, unidirectional series of deletions from one end of pCHB100 and of
pCHB200 (for sequencing of the other strand) was prepared by exonuclease
Ill/mung bean nuclease treatment as described by Henikoff (7). (Some
restriction enzymes produce 3’-overhanging ends, which are resistant to
exonuclease Ill attack, while 5’-overhangs are readily digested. The unique
arrangement of restriction sites for enzymes creating the two kinds of sticky
ends on the polylinker of the Bluescript vector allows the controlled
unidirectional digestion of the insert with exonuclease lll. Subsequent
treatment of the ends with mung bean nuclease creates blunt ends, which can
be readily ligated).

Sequencing reactions on single stranded templates were performed

using the conditions described in the method section of chapter 4. Utilizing the

 

125

obtained sequence information, the remaining gaps were closed using

synthetic oligonucleotides, which were prepared as described in chapter 4.
Detennlnation of CAT-activity. R. sphaeroides cells harboring various

plasmid constructs containing the chloramphenicoI-acetyl-transferase (CAT)

gene were grown photoheterotrophically in 200 ml Sistrom’s medium

 

containing 0.8 pg/ml tetracycline as described in chapter 3. Cells were
harvested by centrifugation and resuspended in 5 ml enzyme buffer (0.5 M
sucrose, 5 mM ascorbic acid, 0.1 M Tris-Cl (pH 7.8), 5 mM DTI’). Following
the addition of 0.5 mg DNAse l, the cells were broken in a French press at
18,000 psi. Aliquots of 1 ml were frozen in liquid N2 and stored at -70°C. As
needed, the extracts were thawed on ice, and the debris was removed by
centrifugation for 5 min in a microfuge. The protein concentration in the
supernatant was determined using the Bradford micro assay (2). For the
determination of CAT-activity in the supernatants, 40 pl of enzyme (diluted
appropriately with enzyme buffer) were mixed with 5 pl of 0.25 M Tris-Cl

(pH 8.0), 1 pl of 10 mM acetyI-CoA, and 0.5 pl of 14C-chloramphenicol

(57 pCi/mmol; 25 pCi/ml). Following incubation at 37°C for 1 hour, the reaction
mixture was extracted with 300 pl ethyl acetate. The ethyl acetate phase was
dried down under a stream of N2, and the material was resuspended with 10
pl ethyl acetate. Following spotting of the material onto a Si250 Silica plate
(Baker), activated for 15 min at 120°C, the plates were developed with
chloroform/methanol (95:5, v/v). Radioactivity was determined by scanning

individual lanes using a System 200 Imaging Scanner (Bioscan). In addition,

126

autoradiograms of the TLC plates were obtained to confirm the location of the
bands as detected by the scanner. The enzyme concentration was adjusted,
such that only monoacetylated product was formed and not more than 80% of

the substrate was converted.

RESULTS AND DISCUSSION

Isolation of a cosmid complementing the squ and squ mutants.
Individual cosmids containing R. sphaeroides wild type sequences were mated
from E. coll (HB101) into the squ mutant, and the exconjugants were tested
for their ability to synthesize sulfolipid. The cosmid library used was
identical to the library described in chapter 4. Approximately 800 exconjugants
were screened to isolate the first cosmid complementing the squ mutant
(pCHB1601). Two other complementing cosmids were isolated by screening
an additional 100 cosmid clones. Figure 5-1 shows the lipid phenotype of the
squ mutant alone and carrying the cosmid pCHB1601. It was not clear
whether all three cosmids were independent, since at least two were found in
the same microtiter plate of the cosmid library at fairly close positions,
suggesting the possibility of cross contamination. However, all further
experiments were done focusing on the cosmid pCHB1601. To test whether
this cosmid could complement not only the squ mutant, it was mated into
the squ and squ mutants. Testing of the exconjugants for restoration of the
ability to synthesize sulfolipid revealed that the cosmid pCHB1601 could not

only complement the squ mutant, but also the squ mutant. However, the

 

127

1 2 3
m
PG” - -
SL'
PE" ~ .
PC- ‘

Figure 5-1: Lipid phenotype of the squ mutant lacking or containing the
cosmid pCHB1601. The following lipid extracts were separated on an
ammonium sulfate impregnated silica plate. (1) wild type, (2) squ mutant
containing pLA2917, (3) squ mutant containing pCHB1601. PG,
phosphatidyl glycerol; SL sulfolipid; PE, phosphatidyl ethanolamine; PC,
phosphatidyl choline. The lipids were visualized by charring.

128

squ mutant was not complemented. This result indicated that the squ
mutant did not belong to the same complementation group as the other two
mutants and, therefore, had to be affected in a different gene. At that stage of
the analysis, the question had to be raised whether the squ and squ
mutants were allelic. This question could be answered through the isolation of
two distinctive classes of subclones of pCHB1601 complementing the two
mutants as described below.

Construction of a physical map for the cosmid pCHB1601. To
construct a restriction map for the cosmid pCHB1601, DNA purified on a CsCl
gradient was cut with a series of restriction enzymes. The Enzymes Kpnl,
Hindlll, and EcoRV had only one or two restriction sites internal to the insert.
These enzymes were employed in various combinations to establish a simple,
but complete map of pCHB1601 (Figure 5-2). The length of the cosmid was
estimated to be about 48 kb. A centrally located Hindlll site allowed the
isolation of the right half of the cosmid as an 11-kb Hindlll fragment, which
was subcloned into the Hindlll site of pLA2917 to give pCHB1612 (Figure 5-2).
Recircularization of the remaining plasmid gave rise to pCHB1611, which
contained the left half of pCHB1601 (Figure 5-2). These two subclones could
be tested for complementation and allowed the mapping of multiple restriction
sites for the enzymes Pstl, Bglll, and EcoRI. Mapping of these sites for clone
pCHB1612 was unambiguously possible due to the central location of a
unique Kpnl site and a single EcoRV site internal to the insert and did not

require further Southern analysis (Figure 5-3, 8). However, because no unique

129

 

 

2kb pCHB1601

R R H K R HR R

h I I I I .1 _ .I

R R HR R R

— . __ - - pCHB1611
pCHB1621 Li” I‘ 5' ”L7 "

 

DLA2917 U31“ '3 '3

Figure 5-2: Simple restriction maps for pCHB1601, the two subclones
pCHB1611 and pCHB1612, and pLA2917. Vector sequences are drawn with
thick lines. Restriction sites: H, Hindlll; K, Kpnl; R, EcoRV. The plasmids are
shown linearized at one of the EcoRV sites.

130

restriction sites in pCHB1611 were found, Southern analysis was used to
establish the map shown in Figure 5-3 (A). Several combinations of enzymes
were used to prepare double digests of the pCHB1611 DNA, which was
transferred to nylon membrane filters following electrophoresis. These filters
were probed and repeatedly stripped using the vector as well as various
subfragments of pCHB1611 as indicated in Figure 5-3 (A). The order of the
relatively small number of Bglll fragments could be established based on
identification of the left and right border fragments by hybridization with the
vector probe, and due to the fact that one of the internal Bglll fragments was
cut with EcoRV, resulting in an unambiguous restriction pattern (Figure 5-2, A).
The relatively high number of Pstl and EcoRI sites located on the left half of
pCHB1611 were mapped by probing the filters with two Bglll fragments (83
and 84; Figure 5-3, A) spanning the left half of the insert. The remaining Pstl
and EcoRI sites located on the right side of the insert could be deduced
without further Southern analysis. Restriction maps of subclones of pCHB1611
(see further below) provided additional information confirming the map
outlined in Figure 5-3 (A). It should be pointed out that sites closer together
than 0.3 kb might have been missed, due to the resolution limits of the
restriction mapping technique ( e.g., two Bglll sites about 50 bp apart were
discovered by sequencing of one of the subclones).

Isolation and sequence analysis of a small subclone of pCHB1601,
complementing the squ mutant. To isolate smaller subclones of pCHB1601,

the purified DNA from pCHB1601 was partially digested, and fragments of the

 

 

 

 

 

131
A
2kb pCHB1611
R R HR R R
I— | l I
33 34
2: -__§
R E PBE EPEBPE PEPB E ERFB EB HR E R
#1; H I I] II [I ll 1] l I II I l I ”*1 1
2kb
8
2kb pCHB1621
R BH K R HR R R
hi I l J_—l
R E PBH EP PEKPB HR E R
g; l l I [III II I —L;
2kb

 

Figure 5-3: Detailed restriction maps for the pCHB1601 subclones pCHB1611
and pCHB1621. Vector sequences are drawn with thicker line. (A) Map for
pCHB1611. Open boxes indicate the location of fragments (83, and 84) used
as probes in DNA hybridization experiments as mentioned in the text.(8) Map
for pCHB1621. Restriction sites: 8, Bglll; E, EcoRl; H, Hindlll; K, Kpnl; P, Pstl;
and R, EcoRV.

132
size ranges 2-3 kb, 3-4 kb, and 4-6 kb were ligated into the Bglll site of

pLA2917 providing the material for three pCHB1601 sublibraries. Out of 96
clones from the 2-3 kb size range library, five could restore the wild type
phenotype when mated into the squ mutant. The smallest, positive clone
containing an insert of about 2.4 kb (pCHB16014) was further analyzed. A
2.9 kb fragment containing the insert as well as flanking pLA2917 sequences
was excised using a Hindlll, Pstl partial digest, and was subsequently cloned
into the Bluescript vectors pBS-KS+ and pBS-KS' cut with Hindlll and Pstl. A
restriction map of the 2.9 kb fragment is given in Figures 5-4 and 5-7. The two
resulting Bluescript clones pCH8100 and pCHB200 were sequenced using a
exonuclease Ill/mung bean nuclease deletion strategy as described in the
methods section (Figure 5-4). Gaps in the sequence were filled using the
synthetic oligonucleotide primers listed in Table 5—2.

Analysis of the sequence (Figure 5-5) for open reading frames revealed
that two reading frames, located on opposite strands, contained only one or
no stop codon, while the other four possible reading frames were blocked
with stop codons allowing the presence of only relatively small open reading
frames. Due to the high GC-content of 67 % of the examined sequence, which
is typical for H. sphaeroides, one would expect a statistical
underrepresentation of the 3 stop codons TGA, TAA, and TAG with their
overall GC-content of only 22 %. The observed low number of stop codons in
at least two of the six possible reading frames did not allow for a

straightforward identification of the coding sequence for the .9qu gene.

133

pCHB16014

0‘ ~ ~ fl 9

D “ W‘o"

H A esp EEPB P
C—o fi— fi—o ‘_°‘—° ‘-

h—o 4—0‘—‘_‘—

250 hp

Figure 5—4: Sequencing strategy used to sequence the 2.4-kb insert of
pCHB16014. The Hindlll, Pstl fragment containing the insert of pCHB16014
(thin line) as well as flanking pLA2917 vector sequences (thick line) shown,
was cloned into the Bluescript vectors as described in the text. Sequencing
runs (arrows) preceded by an open circle were performed using synthetic
oligonucleotide primers. Restriction sites: A, BamHl; 8, Bglll; E, EcoRl; H,
Hindlll; P, Pstl.

134

Table 52: List of synthetic oligonucleotide primers used
for sequencing of the pCHB16014 insert.

 

Primer Sequence

C83 5'-GAC,AGG,TTG,TCG,ACG,ATA-3'
C85 5'-AGA,GGC,ATC,GAG,ATC,GTC-3'
C86 5'-CCG,CAG,CCC,GTC,GTT,CTG-3'
CB7 5'-ATC,CAG,TCG,GCG,ATC,GGT-3'
C88 5'-TGC,CAG,ATC,CAA,TCA,GGA-3'
C89 5'-TCC,TGA,TTG,GAT,CTG,GCA—3'
C811 5'-TCT,TGC,CGC,CAA,GGA,TCT-3'
C812 5'-CGA,CGC,AGC,GCA,CCG,AAT-3'

C813 5'-CAC,GTT,CCG,CAG,CAC,CAG-3'

 

135

Figure 5-5: Nucleotide sequence of the insert found in pCHB16014. Only the
coding strand for the putative squ gene is shown. The amino acid sequence
deduced by translation of the open reading frame is given beneath the
nucleotide sequence. Restriction sites which were used for subcloning are
listed above the sequence. Three putative initiation codons (F‘s), as well as a
putative ribosome binding site (SD) are indicated above the sequence.

1 CCCGGCGGCACGGTGGGTGCCGACAAGATCCGGTCGAGCGAGGAGGCCGTGGCCTTCGTG
61 CGCGGCTTCGTCAGTGCGGGCAAGCCGGTGGCTGCCATCTGCCATGGCCCCTGGGCGCTG
121 GTTGAGGCGGACGTGCTCAAAGGGCGCGAAGTGACCTCCTACCCCTCGCTCGCCACCGAC
181 ATCCGCAACGCAGGGGGCCGCTGGGT%GACCGCGAGGTCGTTGTCGATTCGGGGCTCGTC
241 ACCAGCCGCAAGCCGGACGGATCTCGATGCCTTCTGCGCCAAGATGATCGAGGAATTTGC
301 CGAAGGCGTGCACGACGGCCAGCGCCGCAGCGCCTGAAAAAGCAGGCCCGCCATCGGCAC
361 AGCCTIGCCGAGCGTCGTGAAAAGTTCATTCGACCGGTGGACTTGCAGCCGATACGCACC
421 AACATEGTCACGAGGACGTGTGCCGCCCGTGAGGCGGCACGCGCGATTCACGAAG%%GGC
481 TTCATGCGCATCGCAGTTCTGGGTGGGGACGGGTTTGTCGGCTGGCCGACGGCGCTCCAC
1 M R I A V L G G D G F V G W P T A L H

541 CTGTCGGACCTCGGACACGAGATCCATATCGT%g%CAACCTGTCCCGGCGCTGGATCGAT
20 L S D L G H E I H I V D N L S R R W I D

601 ACCGAGCTGGGCGTGCAGTCCCTGACCCCGATGGACTCGATCCAGGAGCGCTGCCGGATC
40 T E L G V Q S L T P M D S I Q E R C R I

661 TGGCATCAGGAGACCGGCCAGCGGCTGCATTTCCACCTGCTCGACCTGCGCGAATACGAC
60 N H Q E T G Q R L H F H L L D 0

Ba nfU
721 CGGATCCGCGCCTGGCTCGCCGAATACCGCCCCGAGGCCATCATCCACTTCGCCGAGCAG
80 R I R A W L A E Y R P E A I I H F A E Q

781 CGCGCCGCGCCCTATTCGATGAAGTCGGACCGCCACAAGGTCTACACCGTCAACAACAAC
100 R A A P Y S M K S D R H K V Y T V N N N

841 GTCAATGCGACCCACAACCTGTTGGCCGCCATGGTCGAGACCGGCATCGATGCCCATCTC
120 V N A T H N L L A A M V E T G I D A H L

901 GTCCATCTCGGCACCATGGGGGTCTATGGCTATTCGACCGTGGGCGCGCCCATCCCCGAG
140 V H L G T G G Y S T V G A P I P E

961 GGTTATCTCGACGTCTCGGTCGAGACCCCGGCGGGCCCGAAAGAGCTCGAGATCCTCTAT
160 G Y L D V S V E T P A G P K E L E I L Y

 

136

Figure 5-5 (oont'd)

1021
180

1081
200

1141
220

1201
240

1261
260

1321
280

1381
300

1441
320

1501
340

1561
360

1621
380

1681
400

1741
1801
1861
1921
1981
2041
2101

CCGACGCGGCCGGGCTCGGTCTATCACATGACCAAGAGCCTCGATCAGATCCTCTTCCAG
P T R P G S V Y H M T K S L D Q I L F Q

TATTATGCTCAGAACGACGGGCTGCGGATCACCGACCTGCATCAGGGCATCGTCTGGGGC
Y Y A Q N D G L R I T D L H Q G I V W G
ACGCACACGAACCAGACCCGCCGCCATCCGCAGCTCATCAACCGCTTCGACTATGACGGC
T H T N Q T R R H P Q L I N R F D Y D G

GACTATGGCACCGTCCTGAACCGCTTCCTGATCCAGTCGGCGATCGGTTATCCGCTGACC
D Y G T V L N R F L I Q S A I G Y P L T

GTGCACGGCACCGGCGGGCAGACCCGCGCCTTCATCCATATCCAGGATTCGGTGCGCTGC
V H G T G G Q T R A F I H I Q D S V R C
BglII
GTCGAGCTCGCGCTGTCGGATGCGCCCAAGGCCGGCGAACGGGTGAAGATCTTCAACCAG
V E L A L S D A P K A G E R V K I F N 0
ATGACCGAGACGCACCGCGTGCGAGATCTGGCCGAACTGGTGGCGAAGATGACCGGCGCC
M T E T H R V R D L A E L V A K M T G A
AAGGTCTCGTTCCTGCCCAACCCCCGCAAGGAGGCCGACGAGAACGAGCTCGTGGTCAGG
K V S F L P N P R K E A D E N E L V V R
AACGACCAGTTCCTCGCGCTGGGGCTGAAGCCCATCACGCTGCAGGAAGGGCTGCTGGGC
N D 0 F L A L G L K P I T L Q E G L L G
GAGGTGGTCGATGTGGCGAAGAAGTTCGCCCACCGGATCGACCGCAGCCGCGTGCCCTGC
E V V D V A K K F A H R I D R S R V P C
GTCTCGGCCTGGACGAAGGACATCGCGCAGCGGGTCGAGCACGATCCCGAGGGGCGGCGG
V S A w T K D I A Q R V E H D P E G R R
CTGCGCTCGGTGTCCTGATTGGATCTGGCAGCCGAAGGACCGGCTGGGTCCGAGCGGGCC
L R S V S
TATGTCACGCTCGTGACCAATGCCGATTATGCGCTGGGGGCACGGGCGCTGCTGCGCTCG
CTGGCCCTCAGCGGCACCACGGCCGACCGGGTGGTGCTGCATACCGACGTGCCCGAGGAG
GCGCTGGCGCCGCTCCGGGCGCTCGGCGCGCGGCTCGTGCGGGTGGAGCTTCTGCCCACC
TCGCCCGAGTTCAACGCGGCCCATGCGCGAGAGGCGCTTCATGCGCGCGCGGCCTTCACC
AAGGGCGGCAAGCCGCCCTTCCACACGCCGCTCGATAATTTCGCCAAGCTCCGGCTCTGG
CAGCTCGTGGACTATCGCTCGGTCGTCTTCATCGACGCGGATGCGCTGGTGCTGCGGAAC
EcoRI
GTGGACCGGCTGTTCGACTATCCCGAATTCTGCGCCGCGCCGAATGTCTACGAGAGCCTG

 

137

Figure 5-5 (oont'd)

2161
2221
2281
2341

TCGGACTTTCACCGGATGAATTCGGGCGTCTTCACCGCCCGGCCCTCGACGGACACCTAC
GCCCGGATGCTCGAAGCACTCGACGTGCCGGGGGCCTTCTGGCGGCGGACCGACCAGAGC
TTCCTGCAGCAGTTCTTCCCCGACTGGCAGGGCCTGCCGGTCTTCTGCAACATGCTCCAG
TATGTCTGGTTCGCGATGCCCGAGCTCTGGAGCTGGGAGC

Figure 5-6: Comparison of the putative squ ribosome binding site to binding
sites from various R. sphaeroides genes (A) and to the 3’-end of the 168 rRNA
from R. sphaeroides (8). Bases matching the complementary sequence of the
16S rRNA 3’-end representing the ribosome binding site are shaded.
Sequences were taken from: squ (this study), pucB and pucA (10), puhA (5),
pqu (13), pufM (14), pufB and pqu (9), and R. sphaeroides 16$ rRNA 3’-end

(6).

s qu
s qu
pu c3
pucA
puhA
pu fL

pu fM
pu fB
pu 1‘34

165 3'
rRNA
squ

 

-15 0
CA TTCATG

AG CGGGTG

TT ACAGTG

TA GACATG

cg: TAAATG

AG" AGCATG

TC TCAATG

GA AGCATG

GK. ATCATG
HO-UCUUUCCUCCUCCACUA 5'

******

5'CACGGA’ CTTCATG3'

  

138

However, the reading frame containing one stop codon looked promising,
because one of the in-frame ATG codons (Figure 5-5, P) was preceded by a
putative ribosome binding site showing the AGGAGG motif complementary to
the 3’ end of the 16S rRNA of R. sphaeroides (Figure 5-6, B). The spacing
between the putative ribosome binding site and the ATG initiation codon fell into
the range observed for other genes of Ff. sphaeroides (Figure 5-6, A). None of
the other possible initiation codons showed a comparable ribosome binding
site. Beginning at the initiation codon described above (Figure 5-5, I“), the open
reading frame could encode a 46-kD protein. However, the absence of a stop
codon upstream of the respective initiation codon inside the pCHB16014 insert
did not allow the determination of an upper limit for the length of the coding
sequence. Since at least one stop codon is required for the termination of
translation, it seemed unlikely that the other reading frame, that lacked a stop
codon, could represent the coding sequence. Given the indicated uncertainties
about the location of the squ gene on the pCHB16014 sequence, additional
experiments were required to determine the correct open reading frame.
Complementation analysis using various deletion clones derived from
pCHB16014. Two simple deletion clones of pCHB16014 were constructed. First,
digestion of pCHB16014 with Bglll followed by recircularization resulted in a
1 kb deletion expanding from the right end of the insert into the 3’-end of the
putative open reading frame for the squ gene (Figure 5—7, pCH816014D-8).
Second, double digestion of pCHB16014 with Bglll and BamHl followed by

recircularization gave rise to clone pCH816014D-AB, which carried a deletion of

139
1.7 kb beginning at the right insert and (Figure 5-7). Both deletion clones were

unable to restore the wild type phenotype when present in the squ mutant,
indicating that parts of the open reading frame or an essential promotor
sequence were lost. As described for the analysis of the putative open reading
frame for the squ gene in chapter 4, various fragments of the pCHB16014
insert were cloned behind a cytochrome-c promotor from R. capsu/ata located
on the pCH8500 vector (see Appendix B for the construction and analysis of
this vector). Two Sall sites, one upstream of the putative initiation codon (Figure
5-5, I3) and one downstream in the putative coding region for the squ gene
were particularly useful for the analysis. A 1.9 kb fragment spanning the region
from the upstream Sall site to the nearest EcoRI site downstream of the putative
open reading frame was obtained by a partial Sall, EcoRl digest and cloned
into the Bluescript vector cut with Sall and EcoRI. From the resulting clone, the
fragment was excised with Kpnl and EcoRI, and cloned into pCH8500 cut with
the same enzymes to construct pCHB16201 (Figure 5-7). To obtain the
opposite orientation, the fragment was excised from the Bluescript clone with
Kpnl and Xbal, and cloned into the respective sites of pCH8500 giving rise to
clone pCHB16202 (Figure 5-7). For the construction of the two clones
pCH816151 and pCHB16152 (Figure 5-7), a shorter fragment of 1.55 kb from
the second Sall site to the EcoRI site nearest to the 3’-end of the putative open
reading frame was cloned in both orientations into pCH8500 using the same
strategy as described above. In a subsequent complementation experiment,

both of the clones containing the larger insert in either orientation (pCHB16201,

 

140
and pCHB16202) were able to restore the capability of the squ mutant to

synthesize sulfolipid, while neither of the smaller clones (pCHB16151, and
pCHB16152) could restore the wild type phenotype when present in the squ
mutant. Therefore, it was concluded that the squ gene must be located within
the 1.9 kb Sall/EcoRl fragment of pCHB16014. In addition, this fragment must
contain an endogenous promotor driving the expression of the squ gene,
since complementation was independent of the orientation of the insert relative
to the external promotor provided by the pCH8500 vector. The fact that the
smaller Sail, EcoRI fragment cloned in both orientations into pCH8500 failed to
provide sufficient sequence for complementation, was consistent with the idea
that this fragment did not include the N-terminal portion of the putative coding
sequence for the squ gene as shown in Figure 5-7, and that it was missing a
putative promotor proposed to be located on the 350 bp fragment between the
two Sall sites. An open reading frame analysis performed on the sequence
contained in the 1.9 kb Sall/EcoRl fragment of pCHB16014 revealed only one
large open reading frame identical to the one described above (Figure 5-5),
with no or relatively small open reading frames, as defined by the presence of
an initiation as well as a stop codon, located on the opposite strand. From the
deletion and complementation experiment described above, it was concluded
that the initiation codon had to be located between the two Sall sites. It was
proposed that only one of the three initiation codons located on this fragment
(Figure 5-5, l3), was a good candidate, since it was preceded by a ribosome

binding site (Figure 5-6).

 

141

 

 

 

 

 

 

ORF Clone SL
# ——
II.— T T T ? BIIBT EIIEIP. F 16014 1’
a P
'L_ ? ‘1’ I ‘3 3 ._, 16014D-B -
' e' P
”_ 8. 3. A. .._I 16014D-AB -
”"5118 I I ‘3 ”.77 E. 16201 +
—’r
“3 i3 I ‘3 3..” 7 TLE 16202 +
32“.? I ‘3 an”? I 16151 -
”8 I ‘1’ 3.? 7 TLE: 16152 -
500 bp

Figure 5-7: Complementation analysis using various subclones of pCHB16014.
the putative open reading frame (ORF) for the squ gene is indicated above
the pCHB16014 sequence (the pCHB prefix of the clone names was omitted).
The complementation result is indicated as absence (-) or presence (+) of the
sulfolipid (SL). Three independently constructed clones for each of the
constructs shown were tested for complementation, and gave identical results.
Vector sequences derived from pLA2917 are drawn with a thicker line, while
sequences derived from the Bluescript polylinker are shown as open boxes.
The location and direction of the cytochrome-c promotor (Pcyc) is indicated by
a thick arrow. Restriction sites: A, BamHl; 8, Bglll; E, EcoRl; H, Hindlll; K, Kpnl;
O, Xhol; P, Pstl; S, Sail; X, Xbal. Sites which were lost during the cloning
procedure are marked (*).

 

 

142

However, due to the lack of suitable restriction sites, it was not possible
to isolate a fragment which would contain the putative squ coding sequence
without being preceded by an internal promotor. It was therefore not feasible to
confirm the orientation of complementation by cloning the promotor-less coding
sequence in both orientations behind the promotor provided by pCH8500. In
the case of the squ gene (chapter 4), orientation dependent complementation
was observed by cloning the putative coding sequence lacking a promotor into
pCH8500, thereby confirming the coding strand. A different approach was
chosen to confirm the predominant orientation of gene expression in the insert
of pCHB16014 containing the squ gene.

Expression of a promotor-less CAT-gene Inserted into pCHB16014. To
determine the predominant orientation of gene expression and, thereby, the
orientation of the squ gene on pCHB16014, a 0.9 kb Sau3Al fragment
containing the chloramphenicol acetyl transferase coding sequence but not a
promotor was excised from pBR328 and cloned in both orientations into the
unique BamHl site of pCHB16014. The two resulting clones pCHB16014CAT-1
and -2 (Figure 5-8, A) were mated into the squ mutant along with the pLA2917
vector and pCHB16014. Upon testing of the lipid composition of the resulting
strains, it was noted that both clones containing the insertions lost their ability
to complement the defective squ gene, indicating that the insertion resulted in
the disruption of the squ gene located on pCHB16014. Measurements of the
CAT-activity in extracts obtained from the four cell lines described above

revealed that the presence of the clone pCHB16014CAT-1 in the squ mutant

 

 

 

 

 

 

A ORF
fl
L— ? 3‘.‘ ‘3 '.."3 SELL} 16014CAT-1
V E V
—
——-§
CAT
ORF
fl
L— 3 3‘.‘ ‘3 '..".' ‘UELEJ 16014CAT-2
—
_
CAT 500 bp
B
2.0
.73
I)
§ 1.6»
O.
3
> 1.2-
3
5
E 0.8-
U
‘f 0.4-
I-
6 I_-—I
0.0
t If N '-
f 8 '5 5
5 6
'0 'D

Figure 5-8: Cat-gene insertion clones of pCHB16014 and CAT-activity observed
in cells containing these clones. (A) Location and orientation of the fragment
containing the CAT coding sequence (filled box) in clones pCH816014CAT—1
and -2. The exact location and orientation of the putative squ open reading
frame (ORF) and the coding sequence for the CAT-gene (CAT) are indicated by
arrows. Vector sequences derived from pLA2917 are drawn with a thicker line.
Restriction sites: A, BamHl; B, Bglll; E, EcoRI; H, Hindlll; O, Xhol; P, Pstl; V,
Sau3Al. (8) Cat-activity measured in extracts from squ mutant cells containing
the plasmids pLA2917, pCHB16014, pCH816014CAT-1, and pCH816014CAT-1.
Three replicates were measured and averaged for each of the extracts, and the
calculated standard errors are indicated.

144
resulted in ten-fold higher activity compared to pCHB16014CAT-2. No

CAT-activity could be detected in extracts obtained from cells harboring
pLA2917 or pCHB16014 (Figure 5-8, 8). Since in pCHB16014CAT-1 the
orientation of the CAT coding sequence was the same as for the putative squ
coding sequence, it was concluded that the main orientation of gene
expression coincided with the orientation of the putative squ gene and its
promotor, supporting the idea that the coding sequence as depicted in Figure
5-5 indeed represents the squ gene. However, since small but significant
amounts of CAT-activity were observed when the CAT coding sequence and
the putative squ coding sequence were facing in opposite direction in
pCHB16014CAT-2 (Figure 5-8), one or the other strand could not be
definitively excluded as the location for the squ gene coding sequence.
Analysis of the amino acid sequence encoded by the putative squ
gene. The open reading frame, which was proposed to contain the squ
coding sequence based on the experiments described above, was translated
into the amino acid sequence (Figure 5-5). The resulting sequence contained
404 amino acids and could give rise to a 46-kD protein. The amino acid
sequence was subjected to a hydropathy analysis according to the method
described by Kyte and Doolittle (11). The result (Figure 5-9) led to the
conclusion that the putative protein is probably soluble rather than membrane
bound, because no major hydrophobic domains typical for membrane
spanning regions could be detected. Most proposed reactions involved in

sulfolipid biosynthesis (see Figure 1-2, chapter 1), would not require membrane

 

145

bound proteins, except for the enzyme catalyzing the last reaction.

Using the amino acid sequence deduced from the putative coding
sequence of the squ gene in a homology search involving two protein data
bases, it was discovered that the amino acid sequence for the enzyme
UDP-glucose epimerase from various fungi and bacteria showed homology to
the N-terminal region of the putative squ gene product (Figure 5-10). Out of
100 amino-terminal amino acid residues 13 were identical in all three sequences
(Figure 5-10, shaded amino acids). Given the degree of identity of about 25%
observed in the stretches of homology, and given the number of conservative
substitutions, which did not change the overall character of the protein, the
pattern of homology between the UDP-glucose epimerase genes from various
organisms and the putative squ gene product opened the possibility for
speculation about the function of the putative squ gene. The enzyme
UDP-glucose epimerase catalyzes the conversion of UDP-glucose to UDP-
galactose (8). The fact that a class of enzymes involved in sugar nucleotide
metabolism showed homology to the putative squ gene product, thought to
be involved in the biosynthesis of the plant sulfolipid, was very exciting,
because UDP-sulfoquinovose belonging to the same class of compounds was
proposed to be a precursor of the plant sulfolipid (chapter 1). It is therefore
conceivable that the squ gene product may be involved in sulfolipid
biosynthesis at the level of UDP-sulfoquinovose or some other sugar nucleotide.
In addition, this finding of homology supported the idea that the coding

sequence proposed for the squ gene (Figure 5-5), actually constituted the

 

 

 

 

 

146
5
4-
5-
if;
a 2-
E
E 1
1;: O
O.
o —1-
0C
9 —2-
I
-3-
_4_
_5 I I J I I I I I

O 50 100 150 200 250 300 350 400

AMINO ACID #

Figure 5-9: Hydropathy plot for the putative squ gene product. Included in the
analysis were 404 amino acids, beginning at the initiation codon I3 as shown in
Figure 5-5.

 

147

Figure 5-10: Comparison of the amino acid sequences of the putative squ gene
product and UDP-glucose epimerases from S. cerevisiae and E. coli. (A) The
amino acid sequences were translated from the nucleotide sequence of the ga/10
gene of S. cerevisiae (4), and (B) from the galE gene of E coli (12). Single dots
indicate conservative substitutions, while colons represent perfect identity. Amino
acids conserved in all three sequences are shaded.

A
s qu
93710

squ
galIO

squ
galIO

s qu

galE

5 (MB

galE

   
  

29

100 110

KRELKIAIQTELQV SCKDLWKNTT PFGYQLRGVEARFSAEDMRYDARFVTI GAGTR
8 90 100 110 120
120 130 140 150

VNAT- HNLLAAMVETGI DAHLVHLGTMGVYGY

13FQATFANIJIASI VDLKVNGQSVVLGYENEEGY
140 160

 

70 809000
HQE- TGQRLHFHLLEIE- LREYDRI RAWLAEYRPEAIIHFAEQRAAPYSMKSDRH

 

oooooooooooooooooo

NEALMTEI LHDHAI III HFAGLKA-VGESVQKPLEYYDNNVNGTLRLI SAMR
70 80 100 110

 

148

squ gene, since it seemed unlikely that a random sequence could show the
observed degree of homology.

Localization and orientation of the squ and squ genes relative to each
other on pCHB1601. To determine the position of the putative squ gene on
the map of the original cosmid pCHB1601, an 0.8-kb BamHl, Pstl fragment
internal to the squ coding region (fragment PA; Figure 5-11, A) was used to
probe DNA from pCHB1611 and pCH81621 digested with Pstl and Bglll,
containing the left and right half of the cosmid pCHB1601 (Figure 5-2). Only
fragments of pCHB1611 hybridized to this squ specific probe. From the
results of this Southern experiment, it was concluded that the putative squ
gene was located in the center of pCHB1611, and by comparison of the
restriction maps of pCHB16014 (Figure 5-7) and pCH81611, its orientation was
deduced (Figure 5-11).

From initial experiments it was known that the squ mutant was
complemented by the cosmid pCHB1601. Subclones pCHB1611 and
pCHB1621 (Figure 5—2, 5-3) were tested for their ability to restore the wild type
phenotype when present in the squ mutant. Only pCHB1611 which contained
the left half of pCHB1601 on which the putative squ gene was also located,
was capable of complementing the squ mutant. At this stage of the analysis, it
was not clear that the two mutants carried mutations in two different genes.
However, screening the sublibrary of pCHB1601, containing 2-3 kb fragments,
did not yield a positive clone which could complement the squ mutant. This

was the first genetic evidence that the two mutants, indeed, represented two

149

different genes, since the same sublibrary did yield five positive clones
complementing the squ mutant (see above). Upon screening of the sublibrary
of pCHB1601 that contained the 4-6 kb fragments, two clones could be isolated
which were able to restore the wild type phenotype of the squ mutant.
Restriction analysis of these two clones revealed that they were identical,
possibly clonal. Further analysis was, therefore, restricted to one of the clones,
pCHB16018. Detailed restriction mapping and Southern analysis revealed that
this clone contained two independent fragments resulting in an insert size of
about 10 kb (Figure 5-11, C). Using the DNA of the 2.6-kb Pstl band of
pCHB1611, which contained two comigrating Pstl fragments, one extending into
the left vector border and one designated P26 (Figure 5-11, B) as a probe in a
Southern experiment, the central 2.6-kb Pstl fragment of pCHB16018 showed a
positive signal. The P26 fragment was thought to be located directly
downstream of the squ gene, and its neighboring 0.8-kb Pstl fragment P08
(Figure 5-11, B) was thought to be contained in pCHB16014 extending into

the putative coding sequence of the squ gene (compare Figure 5-11, A).
Probing restricted DNA of pCHB16018 using the 0.8-kb Pstl fragment of
pCHB16014 (identical to P08; Figure 5-11, B), gave a positive result suggesting
that a contiguous stretch of DNA covering the central 2.6-kb Pstl fragment of
pCHB16018 and portions of its flanking sequences (indicated as open box in
Figure 5-11, C) were located on pCHB1611 directly downstream from the squ
gene. Insert sequences of pCHB16018 located to the left of this central region

(Figure 511, C) were found to hybridize to the 5.3-kb Bglll fragment

 

150

Figure 5—11: Location and orientation of the squ and squ genes on
pCHB1611. (A) Restriction map of pCHB1611; the location of the two subclones
pCHB16014 and pCHB181 that complement the mutants squ and squ,
respectively, is indicated as cross hatched arrows. The presumed orientation of
gene expression is indicated by the arrow heads. PA (open box) depicts the
location of a 0.8-kb Pstl, BamHI fragment from pCHB16014 as determined by
Southern analysis. (B) Restriction map of pCHB1611 showing the location of
various fragments (B4, BS, P26, and P08; open boxes) as described in the text.
(C) Restriction maps of the three subclones pCHB16018, pCHB181, and
pCHB182 containing the squ gene. Indicated above the pCH816018 map
(jagged, open bar) is the approximate extent of the DNA stretch common to
pCHBi81 and pCHB182, and pCHB16018. The three maps were aligned at the
sites located in this region. The orientation and location of the external
cytochrome-c promotor, located on the pCHBSOO vector, is indicated by filled
arrows. Restriction sites: B, Bglll; E, EcoRl; H, Hindlll; P, Pstl; R, EcoRI. Sites
marked (*) were lost due to the cloning procedure.

151

 

 

 

 

 

 

 

 

 

 

 

 

 

2kb pCH31611
H R HR 3 H
h ' h
PA ............
. as 34 =, .........
E FEE EPEBPE PEPE E EHPB EB HR E T}!
1 u 1 it n u 1 n 1 Li h
«Mum pCH816014
pCHB181/182 W
2kb
pCHB1611
B4 35
R E PBE EPEBPE PEPE E ERPB EB HR E
hi I 1' l H 1L “ l l '1 ' ' ' h
:25
2kb P28 P08
pCHB16018
) A
P 3 PRP E P a P E B H
L; I I 11 I l I J I :1
O E P O
E '7 '3 1' . 3‘, pCHB181
1kb I}? 7 T I T I? pCHB182

 

 

 

 

 

152
(BS; Figure 5-11, B) of pCH81611. To test the possibility that the squ gene

was located directly downstream from the squ gene, a 4.3-kb Bglll fragment
(B4; Figure 5-11, B) of pCH81611 beginning inside the squ coding
sequence, and containing the two Pstl fragments designated P26 and P08
(Figure 5-11, B), was cloned in both directions into the BamHl site of the
pCH8500 expression vector, giving rise to clones pCHB181 and pCHB182
(Figure 5-11, C). Mating these two clones into the squ mutant indicated that
only pCHB181, but not pCHB182 could restore the wild type phenotype. This
result clearly located the squ gene on a defined fragment of pCHB1611
which overlapped with the putative squ gene. The orientation dependence of
the fragment relative to the cytochrome-c promotor in pCH8500 suggested
that the squ gene was expressed in the same orientation as had been
determined for the squ gene (Figure 5-11, A). In addition, the result indicated
that the squ gene was not directly preceded by a promotor, and therefore
the ability of small DNA fragments to restore the wild type phenotype of the
squ mutant depended on an external promotor as in pCHB181. The fact that
using the sublibrary approach the only complementing clone (pCHB16018),
that could be isolated from pCHB1611 contained two large fragments, was
consistent with this idea, because the second insert might have brought a
promotor sequence close to the squ coding sequence in pCHB16018 to
allow for complementation.

Location of the squ and squ genes relative to the squ gene. As

described above, the cosmid pCHB1601 did not complement the squ

153

mutant described in the previous chapter. However, the reciprocal experiment,
mating the cosmid pCHB1701 complementing the squ mutant into the squ
and squ mutants and testing for complementation has not been done. Using
the left border fragment of pCHB1611 ( fragment B3; Figure 5-3, A) to probe
pCHBi701 DNA in a Southern experiment, no hybridizing fragments could be
detected. The right end was not tested. Despite this incomplete Southern and
complementation analysis, a comparison of two distinct restriction maps of
pCHB1701 and pCHB1601, made it appear unlikely that the two cosmids
share common DNA sequences, suggesting that the three genes belong to
two relatively distant gene clusters or operons. However, further analysis is

required to carefully test this idea.

CONCLUSIONS

Using a complementation assay, a cosmid complementing the squ
and squ mutants was isolated. Two subclones containing the left half
(pCHB1611) and the right half (pCHB1612) of pCHB1601 were constructed,
and a physical map was prepared by restriction and Southern analysis (Figure
5-3). Further subcloning allowed the isolation of a 2.4-kb subclone
(pCHB16014), which could complement the squ mutant. The location and
orientation of the coding sequence for the squ gene was identified using
sequence, deletion, and insertion analysis as described above. Based on
these data, an open reading frame for the squ gene (Figure 5-5) preceded

by a ribosome binding site (Figure 5-6) was proposed. The proposed open

154

reading frame could encode a 46-kD protein. The N-terminal quarter of the
amino acid sequence showed homology to the N-terminal portion of the UDP-
glucose epimerase from the yeast and E. coli (Figure 5-10). This finding
suggested the possibility that the product of the squ gene is involved in
sulfolipid biosynthesis at the level of UDP-sulfoquinovose or some other sugar
nucleotide.

A subclone containing two fragments derived from pCHB1601 was
isolated (pCHB16018) which was able to complement the squ mutant.
Southern and restriction analysis of this clone suggested that the central
portion contained a sequence, which was located directly downstream of the
squ gene on the physical map of pCHB1611 (Figure 5-11). A fragment of
pCHB1611 partially overlapping the squ gene was able to complement the
squ mutant in an orientation-dependent fashion, when cloned into the
pCH8500 expression vector. The orientation of complementation coincided
with the proposed orientation of the squ gene. Based on the results
discussed above, it is proposed that the squ gene and the squ genes may
be organized in an operon which is transcribed from a promotor upstream of
the squ gene. Experiments can be designed to test this idea. Deletion and
sequence analysis should allow the delineation of the location of the squ
coding sequence with respect to the squ gene. Analysis of the RNA species
hybridizing to the two genes as well as mapping of the transcriptional start
site(s) should clarify whether the squ and squ genes are present in a

single operon. In addition, other genes involved in sulfolipid biosynthesis may

155

be located nearby. To find these genes, the construction of a deletion
mutation of R. sphaeroides lacking the region spanned by pCHB1611 will be
useful. This would allow the construction of defined mutations (deletion,
insertions) in the clone pCHB1611, which could be tested in trans.
Alternatively, a series of gene replacements spanning the region of the cosmid
could be constructed.

Based on the available evidence, the squ mutant described in the
previous chapter may mark a second operon or gene cluster involved in
sulfolipid biosynthesis in R. sphaeroides. Further experiments will be required
to test this idea. Some of the mutants described in chapter 3, which have not
been analyzed, might possibly carry mutations in additional genes. The fact
that all three mutants analyzed in detail are affected in different genes,
suggests that the experiments did not saturate the pathway for sulfolipid

biosynthesis with mutations.

REFERENCES

1. Balbas P., Soberon X. Merino E., Zurita M., Lomeli H., Valle F.,
Flores N., Bolivar F. (1986). Plasmid vector pBR322 and its
special-purpose derivatives - a review. Gene 5013-40

2. Bradford MM. (1976). A rapid and sensitive method for the
quantitation of microgram quantities of protein utilizing the
principle of protein-dye binding. Anal. Biochem. 722248-252

3. Bullock W.O., Fernandez J.M, Short J.M (1987). XL1-Blue: A high
efficiency plasmid transforming recA E. co/i strain with beta-
galactosidase selection. Bio Techniques 5;376-378

10.

11.

12.

13.

 

156

Citron. B.A., Donelson J.E. (1984). Sequence of the
Saccharomyces GAL region and its transcription in vivo. J.
Bacteriol. 158:269-278

Donohue T.J., Hoger J.H., Kaplan S. (1986). Cloning and
expression of the R. sphaeroides reaction center H gene. J.
Bacteriol. 168:953-961

Gibson J., Stackebrandt E., Zablen L.B., Gupta R., Woese OR.
(1979). A phylogenetic analysis of the purple photosynthetic
bacteria. Cur. Microbiol. 3:59-64

Henikoff Steven (1984). Unidirectional digestion with
exonuclease Ill creates targeted breakpoints for DNA
sequencing. Gene 28: 351 -359

Kalckar HM. (1958). Uridine diphosphogalactose: metabolism,
enzymology, and biology. Adv. Enzym. 202111-134

lfiley P.J., Donohue T.J., Havelka W.A., Kaplan S. (1987). DNA
sequence and in vitro expression of the 3875 light-harvesting
polypeptides of Ff. sphaeroides. J. Bacteriol. 169:742-750

Kiley P.J., Kaplan S. (1987). Cloning, DNA sequence, and
expression of the R. sphaeroides light-harvesting 3800-850-a
and 8800-8508 genes. J. Bacteriol. 169:3268-3275

Kyte J., Doolittle R. F. (1982). A simple method for displaying the
hydropathic character of a protein. J. Mol. Biol. 157:105-132.

Lemaire H.G., Muller-Hill B. (1986). Nucleotide sequences of the
ga/E gene and the ga/T gene of E. coli. Nuc. Acid Res. 14:7705-
7711

Williams J.C., Steiner L.A., Ogden R.C., Simon M.l., Feher G.
(1983). Primary structure of the M subunit of the reaction center
from Ff. sphaeroides. Proc. Natl. Acad. Sci. USA 80:6505-6509

Williams J.C., Steiner L.A., Feher G., Simon MI. (1984). Primary
structure of the L subunit of the reaction center from Ff.
sphaeroides. Proc. Natl. Acad. Sci. USA 81 :7303-7307

CHAPTER 6
IN VITRO SULFOLIPID BIOSYNTHESIS FROM 35S-PAPS IN EXTRACTS FROM

R. SPHAEHOIDES

ABSTRACT

Crude extracts of R. sphaeroides cells were tested for their capability to
convert 35S—PAPS into sulfolipid. lncubating labelled PAPS with crude extracts
from wild type or squ mutant cells resulted in the production of several
labelled, chloroform-soluble compounds, which were usually not found in
extracts from cells labelled in vivo. However, a labelled compound comigrating
with sulfolipid in different TLC systems was observed in extracts from wild type
but not from squ mutant cells. The identity of the sulfolipid was confirmed by
comigration of the deacylation product with deacylated sulfolipid prepared in
vivo. Intact cells, which cannot take up PAPS, were not able to label sulfolipid
from PAPS, indicating that the conversion of PAPS into sulfolipid by extracts
was truly an in vitro reaction. Maximal rates of sulfolipid synthesis from PAPS
in vitro were about 30 pmol/h mg protein. Addition of water soluble
compounds extracted from wild type cells, as well as a mixture of ATP, MgClz,
NADPH, 3-phosphoglycerol, 3—phosphoglycerate, and glucose-6—phosphate
did not stimulate the synthesis of sulfolipid in the extracts. In a preliminary
series of experiments it was determined that possible intermediates or
substrate analogues such as cysteic acid or DL-glyceraldehyde did not inhibit

in vitro sulfolipid biosynthesis, but that addition of UDP-glucose had a small

157

158

inhibitory effect. Reduced sulfur compounds such as DTT and sulfite had a

more pronounced inhibitory effect.

INTRODUC110N

As a first step for the initiation of a biochemical analysis of the enzymes
involved in sulfolipid biosynthesis, a cell-free system to study the process in
vitro is required. Such a cell-free system allows the analysis of the effects of
potential inhibitors and precursors of sulfolipid biosynthesis, without regard to
whether the respective compound is taken up into the cell or not. Separating
the enzymes catalyzing the various reactions into membrane-bound and
soluble proteins and dissecting the reaction sequence with inhibitors and
substrates can then eventually lead to the identification and purification of
individual enzymes. Although several investigators have attempted to study the
reactions involved in sulfolipid biosynthesis in extracts of broken chloroplasts,
no success has been reported using chloroplast extracts. However, there is
one published study by Hoppe and Schwenn (3), who reported the
incorporation of labelled sulfate, sulfite, and PAPS into sulfolipid by cell-free
extracts from Chlamydomonas re/nhardii. Incorporation of labelled sulfate
required ATP and magnesium, and was stimulated by the addition of
glutathione, while addition of cysteic acid did not affect the incorporation.
PAPS behaved as a legitimate substrate, since the reaction followed a
saturation curve with increasing amounts of PAPS. On the other hand, the

incorporation of sulfite into sulfolipid could not be saturated, even with very

159

high amounts of sulfite, suggesting a non-enzymatic artifact.

Although mutants of R. sphaeroides deficient in sulfolipid biosynthesis,
and their complementing genes, were isolated by genetic and molecular-
biological techniques as described in the previous chapters, the function of
the genes has to be elucidated with biochemical means to determine the
pathway for sulfolipid biosynthesis. The accumulation of water-soluble, 3“S-
labelled compounds in some of the mutants provides an opportunity for such
a biochemical analysis since elucidation of the structure of the compounds
may reveal the enzymatic reaction affected in the respective mutant. However,
the confirmation of their precursor nature has to be by feeding these
compounds to cell extracts from R. sphaeroides. To complete the analysis, it
was, therefore, necessary to develop an in vitro sulfOIipid biosynthesis system
using extracts of wild type Ff. sphaeroides cells. In this chapter, exploratory
experiments are described concerning the development of a cell-free system

to study sulfolipid biosynthesis.

MATERIALS AND METHODS
Bacterial strains. The experiments described in this chapter involved
two strains of Fl. sphaeroides, the wild type 2.4.1 obtained from R.L. Uffen’s
lab at MSU, and the sulfolipid-deficient squ mutant described in chapter 3.
Preparation of the sulfur-activating enzymes from yeast. A crude
enzyme preparation from yeast was employed to prepare 3583’-

phosphoadenosine-5’- phosphosulfate (PAPS) using a modified method

 

160

according to Robbins (5). Three enzymes are required to prepare PAPS from
ATP and “S-sulfate. ATP-sulfurylase catalyzes the reaction:
ATP+SO4+H’r -> adenosine phosphosulfate (APS) + pyrophosphate (PP).
The energetically unfavored synthesis of APS is driven by the removal of
pyrophosphate from the equilibrium by pyrophosphatase, and by the
concomitant conversion of APS to PAPS by APS-kinase. The yeast preparation
described below contained all three enzymes. As starting material 500 g dry
yeast, obtained from a bakery supplier, were resuspended in 1.5 I of 50 mM
KZPO4 (pH 9.2) and the suspension was poured into 4 I of liquid N2. After
thawing and addition of 0.5 l of 50 mM KZPO4 (pH 9.2), the suspension was
stirred at 4°C over night. All of the following steps were done at 4°C. The
debris was sedimented at 9000 rpm for 10 min using a Sorvall GS-3 rotor
(Dupont). The extract was acidified with 1 N acetic acid to pH 5.1 and, the
material centrifuged again as described above. To 1 l of supernatant 1.5 I of
saturated ammonium sulfate solution (1400 g of ammonium sulfate dissolved
in 2 l of water, 0.2 mM EDTA, pH 7.5) were added. Following equilibration, the
material was centrifuged as described above. The supernatant was discarded
and the pellet was resuspended in 25 ml of 20 mM Tris-Cl (pH 8.0). The
samples were frozen in 5 ml aliquots in liquid N2 and stored at -20°C. Prior to
use, the enzyme solution was dialyzed against 4 X 1 l of 20 mM Tris-Cl (pH
8.0) to remove ammonium sulfate.

Preparation of 3"S-PAPS using the sulfur activating enzyme isolated

from yeast. Reaction mixtures of 1 ml (usually 5 separate reactions) were set

 

161
up for the synthesis of 35S-PAPS from ass-sulfate. The reaction mixtures

contained (final concentrations): 100 mM Tris-Cl (pH 8.5), 10 mM ATP (1 M
stock solution in 0.1 M Tris; pH 8.0), 10 mM MgCl2, about 600 pl dialyzed
enzyme suspension, and 500 pCi 3“‘S-sulfate (approximate specific activity of
10 pCi/nmol, adjusted by the addition of cold sulfate to carrier free s“S-sulfate;
it was assumed that only negligible amounts of sulfate were carried over by
the dialyzed enzyme suspension). The mixtures were incubated for 2.5 h at
37°C and stopped by addition of 2 ml of ethanol. The proteins were allowed to
precipitate for 10 min on ice and then sedimented in a clinical centrifuge at top
speed. The conversion of 3‘58-sulfate and ATP into 3‘5S-PAPS was tested by
TLC on polyethylene-imine cellulose plates (see below), and was repeatedly
found to be about 60 %. Only two labelled compounds were detected,
assumed to be sulfate and PAPS as determined by comigration with
standards . The supernatant solutions were stored at -20°C until they were
used for the purification of labelled PAPS.

Analysis and purification of the 85S-PAPS by anion exchange
chromatography. To test the quality of the PAPS preparation, the material was
analyzed by anion exchange chromatography on polyethylene-imine plates
(0.1 mm thick on plastic sheets with fluorescence indicator, obtained from
Baker). The plates were soaked in methanol and air-dried prior to spotting,
and were developed using 0.9 M guanidine-HCI dissolved in water as running
buffer. Following development, the plates were soaked again in methanol to

remove excess salt from the plates prior to autoradiography. Unlabelled

162
standards were detected by placing the plate under a UV lamp. The Rrvalues

observed for various compounds were: ATP, 0.06; PAPS, 0.25; ADP, 0.31;
AMP, 0.72; and sulfate, 0.81.

For preparative purposes, labelled PAPS was separated from 35S-sulfate
by chromatography on ECTEOLA-cellulose (coarse; obtained from Sigma)
columns as described by Vargas (7). The column material was prepared by
sequential washes with 0.1 N HCI, H20. 0.1 N NaOH, H20, 0.5 N formic acid,
and H20 (25 ml/g material). Ethanolic supernatants from labelling reactions
(described above), which were stored at -20°C, were centrifuged in a clinical
centrifuge to remove precipitated material. The supernatants were dried to
500 pl under a stream of N2. Following the addition of 2 ml 10 mM (NH,)HCOa,
the sample was loaded onto a 2—ml column, prepared in a Pasteur pipet and
equilibrated with 20 ml of 10 mM (NH4)HCO;,. Sulfate was washed off the
column with 12 ml of 10 mM (NH4)HCO;., and PAPS was eluted with 2 ml of
0.5 M (NH.)HCOa. The buffer was removed under vacuum following freezing of
the sample in liquid N2. The material was resuspended in 300 pl of 50 % (v/v)
ethanol, dried down under a stream of N2, and was finally resuspended in
100 pl 0.1 M Hepes-Na buffer (pH 7.5). Since PAPS is very unstable in water,
the purified 3iss-PAPS was not stored longer than one week at -70°C. No other
contaminating, a“‘S-labelled compounds were detected by TLC immediately
following purification. For practical reasons, the specific activity of the “S-
PAPS was assumed to be the same as that of the substrate asS-sulfate (10

uCi/nmol). It was assumed that no significant amounts of non-radioactive

 

 

 

163

sulfate were carried over with the enzyme preparation, since the sulfate
concentration in the enzyme suspension was reduced about 109-fold by
repeated dialysis as described above. In principle, it would have been
possible to quantify the amount of PAPS in the sample using a very
cumbersome method involving an enzymatic transfer reaction described by
Vargas (7).

Preparation of concentrated extracts from R. sphaeroides. Cells of
Ft. sphaeroides were grown photoheterotrophically in 200 ml Sistrom’s
medium as described in chapter 3. Late log cultures were harvested by
centrifugation, and the cells were resuspended in 1.5 ml of 0.1 M Hepes (pH
7.2). Subsequent to the addition of 1000 units of DNAse l, the cells were
broken in a French press at 16,000 psi. Using the Bradford micro assay (1),
the protein concentration in the extracts was determined to be about
20-30 mg/ml, depending on the amount of cells in the 200 ml culture.

Assay for In vitro sulfolipid biosynthesis. To study sulfolipid
biosynthesis in extracts of H. sphaeroides, reaction mixtures (100 pl) were set
up in polypropylene microfuge tubes, containing 0.1 M Hepes (pH 7.2), 1 pCi
35S-PAPS (approximate specific activity 10 pCi/nmol) and 25 pl cell extract
(about 0.5 mg protein). Initially, the following compounds were added to
circumvent the possible dilution of precursors: 20 mM ATP, 1 mM NADPH,
5 mM MgCIZ, 1 mM 3-phosphoglycerol, 1 mM 6-phosphoglucose, and 1 mM
3-phosphoglycerate. The addition of these compounds was later omitted,

since they did not affect the synthesis of sulfolipid in the described

164

experiments. Several mixtures were incubated in parallel at 32°C. The
reactions were stopped at different times, usually not later than 12 min
following the addition of the substrate, by addition of 800 pl
chloroform/methanol (1:1, v/v) on ice. To the extract 260 pl of 0.2 M H3PO4, 1
M KCI were added. Mixing, followed by centrifugation, resulted in phase-
partitioning of the lipids into the lower chloroform phase. The chloroform
phase was spotted onto an ammonium sulfate-treated silica plate, which was
activated and run as described in chapter 3. Due to the high lipid
concentration in these extracts, the plates were overloaded on some
occasions, resulting in smearing or broadening of the bands. Labelled lipids
were detected by autoradiography, and the bands were transferred into
scintillation cocktail. The radioactivity was determined by scintillation counting.
Taking into account the counting efficiency, and based on the specific activity
of the substrate and the amount of protein in the sample, rates of
incorporation of PAPS into sulfolipid were expressed in some cases as

(pmol PAPS incorporated)/(h mg protein).

Preparation and TLC of labelled sulfolipid standard. To identify in vitro
synthesized sulfolipid on TLC plates, lipid extracts containing asS-labelled
sulfolipid prepared by labelling intact Ff. sphaeroides wild type cells, were run
on the same TLC plates as the lipid extracts from the in vitro reaction
mixtures. The labelled sulfolipid standard was essentially prepared as
described in chapter 3. During the preparation of water soluble, asS«-labelled

compounds from wild type cells (methods section of chapter 3), a large

165

amount of labelled sulfolipid could be obtained in the chloroform phase as
byproduct, which was used for this purpose. The two TLC systems used were
ammonium sulfate plates, prepared and run as described in chapter 3, and
regular Baker Si250 silica plates activated for 15 min at 120°C and developed
with chloroform/acetone/methanol/acetic acid/water (50:20:10:10:5, v/v).
Deacylation of sulfolipid and analysis of sulfoquinovose glycerol by TLC.
Following chromatography of the labelled sulfolipid on TLC plates and
detection by autoradiography, the respective bands (about 5000 cpm) were
scraped from the plates and the material was transferred to a Pasteur pipet
plugged with glass wool. Sulfolipid was eluted from the silica with 2 ml
chloroform/methanol (1 :1, v/v). The solvent was removed under a stream of N2
and the lipid was resuspended in 100 pl of the same solvent. Addition of 25 pl
0.2 M H3PO4, 1 M KCI resulted in phase-partitioning of the sulfolipid into the '
chloroform phase, which was transferred to a new tube, and dried with N2.
This material was resuspended in 100 pl chloroform/methanol (2:3, v/v), and
100 pl of 0.2 N NaOH in methanol were added. Following incubation at 37°C
for 20 min, 40 pl methanol, 160 pl chloroform, and 180 pl water were added.
The aqueous upper phase was transferred to a new tube containing about
100 pl Dowex 50H+ beads in methanol/water (10:9, v/v), to neutralize the
solution. The beads were spun down and the supernatant was transferred to a
new tube. To maximize the yield, the chloroform phase, obtained in the
previous step, was washed with 200 pl methanol/water (10:9 v/v), and the

resulting upper phase was used to wash the Dowex 50H+ beads. The

166

supernatant solutions from the first treatment and the wash were combined
and the solvent was evaporated under a stream of N2. The material was
resuspended with 10 pl methanol/water (10:9) and the sample was spotted
into the lower right hand corner of a cellulose plate (0.1 mm thick on plastic,
Merck). The plate was developed in the first dimension with water-saturated
phenol, and in the second dimension with methanol/formic acid/water
(80:13:7, v/v).

Extraction of possible precursors from wild type cells with hot ethanol.
To prepare an extract from wild type cells containing small molecules, which
could stimulate the in vitro sulfolipid biosynthesis, cells from a 200-ml
photoheterotrophically grown culture of R. sphaeroides (for growth conditions
see chapter 3) were sedimented and resuspended in 5 ml 80% (v/v) ethanol,
and the sample was incubated at 65°C for 5 min. Following centrifugation in a
clinical centrifuge, the supernatant solution was transferred to a new tube, and
the pellet was reextracted with 5 mL of 100% (v/v) ethanol for 5 min at 65°C.
Following removal of the supernatant the pellet was extracted again with 50%
(v/v) ethanol at 65°C for 5 min. The three supernatant solutions were
combined and dried to 1 ml at about 50°C under a stream of N2. Following the
addition of 3 ml of chloroform and 1 ml of water, the upper, aqueous phase
was transferred to a new tube and dried down under a stream of N2 at about
50 °C. The material was resuspended in 1 ml of water and stored at -20°C.

The pH value in the solution was determined to be pH 7.3.

1 67
RESULTS AND DISCUSSION

asS-PAPS is only incorporated into sulfolipid by broken cells, but not by
Intact cells. Labelled PAPS was chosen as a substrate, because it is the only
known precursor for sulfolipid biosynthesis besides sulfate, and has the
advantage that it can not be taken up by intact cells. Therefore, it was
assumed that any incorporation of label from PAPS into sulfate by extracts of
R. sphaeroides was truly in vitro and could not be explained as being due to
the presence of unbroken cells in the extract. This was demonstrated as
follows. A cell suspension of R. sphaeroides wild type cells was split and one
aliquot was run through the French press at 16,000 psi as described in the
methods section. Various amounts of broken and intact cells were incubated
in parallel with asS-PAPS. Only in lipid extracts obtained from incubation
mixtures containing broken cells was a labelled lipid detected that comigrated
with authentic sulfolipid (Figure 6-1). In addition, several labelled compounds
that did not comigrate with sulfolipid were detected in reaction mixtures
containing broken cells but not in those containing intact cells. The sulfolipid
standard shown in Figure 6-1 was actually not purified sulfolipid, but a
complete lipid extract obtained from cells labelled in vivo with 35S-sulfate. No
additional sulfur- labelled bands except the sulfolipid band were visible,
indicating that the additional bands observed in extracts from in vitro
reactions were artifacts. Two conclusions from this experiments were that,
since only broken cells could label chloroform-soluble material, including

sulfolipid, from PAPS, the labelling reactions must have been in vitro and

168

* -, «r it "‘
\— . ’
~
.- ’ -
~ 7‘
o e O 9 ° ‘. s t l' s

SL123456SL 789101112$L

Figure 6-1: Incorporation of 35S-PAPS into sulfolipid by broken and intact wild
type cells of Ff. sphaeroides. Shown are lipid extracts obtained from reaction
mixtures, separated on ammonium sulfate-treated plates. Labelled lipids were
detected by autoradiography. Various amounts of broken cells (lanes with
even numbers) or intact cells (lanes with odd numbers) were incubated for 12
min in the presence of asS-PAPS. Lanes marked with SL show lipid extracts
containing authentic sulfolipid obtained from cells labelled with sulfate. The
various protein concentrations used in the reactions were (in mg): (lanes 1, 2)
1.8; (lanes 3, 4) 1.5; (lanes 5, 6) 1.2; (lanes 7, 8) 0.9; (lanes 9, 10) 0.6; and
(lanes 11, 12) 0.3.

 

 

 

 

 

169
E 14- .
3
a 12- . -
O
o
E 10- -
m C
s 8- .
o.
H— 6" -
2 e
.2
15. 4' ‘
o
a 2- ~
0
o
E O 1 1 1 L
0.0 0.5 1.0 1.5 2.0

Amount of Protein in Assay (mg)

Figure 6-2: Dependence of the incorporation of a“‘S-PAPS into sulfolipid on the
amount of extract present in the reaction mixture. The radioactivity in the
putative sulfolipid bands observed in the experiment described in Figure 6-1,
was determined, and used to calculate the incorporation of 35S-PAPS into the
sulfolipid. The incorporation shown on the Y-axis was plotted against the
protein concentration in the reaction mixtures (X-axis).

170

contaminations of extracts with intact cells would probably not bias the
interpretation of results. Attempts to quantify the fraction of intact cells in
French press extracts by plating the extract onto plates containing Sistrom’s
medium and counting of the resulting colonies, were hampered by the fact
that the cells had a tendency to clump and colonies derived from single cells
could not be obtained, but it was obvious that there was still a substantial
amount of viable, unbroken cells present in the extracts.

When the amount of extract, expressed as mg of protein, used in the
experiment described above was plotted against the incorporation of PAPS
into the putative sulfolipid, a linear relationship was observed (Figure 62),
indicating that the system was not saturated under the chosen conditions. The
amounts of extract used in the reactions, ranging from 0.3 mg of protein to
1.8 mg (Figure 6-1), were relatively high. However, since it was expected that
several unknown enzymatic reactions and intermediates were involved to
catalyze the incorporation of PAPS into sulfolipid, the extracts were kept as
concentrated as possible to avoid the dilution of intermediates.

Characterization of the sulfolipid labelled In vitro. To confirm that the
lipid labelled from asS-PAPS by extracts of R. sphaeroides wild type cells,
which seemed to comigrate with authentic sulfolipid (Figure 6-1), indeed
represented sulfolipid, a series of experiments was performed. When extracts
obtained from the sulfolipid deficient squ mutant (see chapter 3) and extracts
from wild type cells were incubated in the presence of ai‘S-PAPS, no band

comigrating with the sulfolipid standard was observed in mutant extracts, while

171

~a~sasssss

A

SL > - -
. I . g o ' 4 O I .
st 2 4 6 8 10 St 2 4 6 8 10

Time (min) Time (min)

Figure 6-3: Comparison of extracts from wild type and sulfolipid deficient squ
mutant cells for their capability to label sulfolipid from at‘S-PAPS. (A) wild type
cell extracts, and (B) squ mutant cell extracts were used in the reaction
mixtures. Lipid extracts from in vitro reaction mixtures separated on
ammonium sulfate treated-plates are shown. The lipids were detected by
autoradiography. Lanes marked (st) show in vivo labelled sulfolipid. Lipids
were extracted from the reaction mixtures at various times as indicated.

172

-F2

F1-

uolsuaullo 19113

Second Dimension

Figure 6-4: Cochromatography of deacylated sulfolipid labelled in vitro and in
vivo. Presumptive sulfoquinovose-glycerol prepared from in vitro (2000 cpm)
and in vivo labelled sulfolipid (4000 cpm) were mixed together and analyzed in
a two dimensional TLC-system as described in the method section. The open
circle marks the origin. The two solvent fronts are indicated in the upper left
corner (F1, solvent front of the first dimension; F2, solvent front of the second
dimension).

 

173

the same number and intensity of artifactual bands was observed as for the
wild type extracts (Figure 6-3). This result provided strong evidence that wild
type extracts were indeed capable of synthesizing sulfolipid from labelled
PAPS, and that the labelled compounds, only observed in in vitro labelling
experiments, were most likely unrelated to sulfolipid biosynthesis. The nature
of these compounds, specific to the in vitro system, has not been further
analyzed. However, invoking the analogy to a different in vitro system for
studying sulfolipid biosynthesis, in lipid extracts from isolated spinach
chloroplasts, which were labelled with 35S-sulfate, a labelled, chloroform-
soluble compound was found in addition to sulfolipid (4). This compound was
not observed in extracts obtained from labelled leaves. Analysis of this
compound identified it as elemental sulfur which is chloroform-soluble (4). With
this result in mind, it was assumed that the labelled compounds observed
besides sulfolipid in lipid extracts which, were prepared from the in vitro
reaction mixtures employing broken cells of Ff. sphaeroides, represented
some form of reduced sulfur multimers. However, further experiments are
required to test this idea.

The identity of the putative sulfolipid labelled in in vitro reactions from
35S-PAPS was further confirmed by comigration of the isolated lipid from the
respective TLC-band with in vivo labelled sulfolipid in two different solvent
systems. In addition, in vivo and presumptive in vitro labelled sulfolipid were
deacylated, and the two resulting preparations, were found to comigrate in a

two dimensional TLC-system, when spotted together in one mixed sample

174

 

 

 

 

 

 

 

 

 

 

”9 IO - .
x
E a -
8
.J
(O 5 _ ‘
.E
z <
E 4-
‘5
m
.9 2 - 1
'3
a:
O:
O
C
Time (min)
B A 40
o
E
.c
> 30
o ..
3
‘3 e .
g 20 - /
c
VI); . Km "' 5 [J.M
d 10 —/
'5
3 D
o
m I I I I
O
O 20 4O 60 80 100

Concentration of PAPS (,uM)

Figure 6-5: Substrate saturation for the incorporation of 35S-PAPS into
sulfolipid. (A) Apparent rates of sulfolipid biosynthesis for various total
concentrations of PAPS as indicated. The radioactivity in the sulfolipid bands
as determined by scintillation counting was plotted against the incubation time.
The amount of protein used in each reaction mixture was 0.85 mg. (B)
Absolute rates of PAPS incorporation into sulfolipid, calculated on the basis of
data shown in (A), were plotted against the concentration of PAPS in the
reaction mixtures (Michaelis-Menten plot). The substrate concentration, at
which the process is half saturated, is indicated (kM).

175
(Figure 6-4).

Saturation of the incorporation of alsS-PAPS into sulfolipid by addition of
unlabelled PAPS. To determine, whether PAPS behaved as a legitimate
substrate for an enzymatic reaction in the in vitro label experiments, the rate of
incorporation of 3"Si-PAPS into sulfolipid was determined at a range of
substrate concentrations. The total amount of PAPS in the reaction mixture
was varied while the amount of label (1 pCi) per reaction was kept constant.
The apparent rates, measured as increase in radioactivity in the respective
sulfolipid bands isolated from TLC-plates over time, were found to be linear,
and stable PAPS could compete effectively with labelled PAPS under the
chosen conditions (Figure 6-5, A). From the apparent rates, absolute rates for
the incorporation of PAPS into sulfolipid were calculated and plotted against
the total concentration of PAPS present in the reaction mixtures (Michaelis-
Menten plot; Figure 6-5, B).

Based on this experiment, the maximal rate of sulfolipid biosynthesis
from PAPS in extracts from H. sphaeroides was estimated to be about 25-30
pmol/h mg protein, and the process was half saturated at PAPS
concentrations of about 5 pM (kM in Figure 6-5, B). Unfortunately, no data for
the rate of sulfolipid biosynthesis in vivo in R. sphaeroides were available for
comparison. The value of 5 pM for the substrate concentration at which the
process was found to be half saturated, fell well into the range of known kM
constants for various enzymes. This result confirmed that PAPS behaved as a

legitimate substrate for an enzymatic reaction. This result was not in

176

agreement with a nonenzymatic, artifactual incorporation of label from PAPS
into sulfolipid by extracts of Ff. sphaeroides.

Effect of various compounds added to the reaction mixtures. In a series
of exploratory experiments, various compounds were added to the in vitro
reaction mixtures, to screen for those which could either stimulate or inhibit
sulfolipid biosynthesis. First of all, to stimulate the reactions in the extract, a
basic mixture of metabolic intermediates commonly found in cells was added,
as well as a concentrated, aqueous extract from wild type cells, prepared by
extraction of cells with hot ethanol (see methods section), which was thought
to contain metabolic intermediates possibly required for the reactions involved
in sulfolipid biosynthesis. The basic metabolite mixture contained (final
concentrations in the reaction mixture): 1 mM ATP, 1 mM NADPH, 5 mM
MgCl2, 1 mM glycerol-3-phosphate, 1 mM glucose-6-phosphate, and 1 mM 3-
phospho-glycerate. Neither the addition of these compounds, the metabolite
extract from the wild type cells, or any combination of these was able to
stimulate the rate of incorporation of 3i‘S-PAPS into sulfolipid (Figure 6-6, A).
Additions of the listed compounds individually also had no effect. It was
concluded that either the availability of intermediates in the highly
concentrated, crude cell extracts used was not the limiting factor under the
chosen conditions, or that the enzymes involved in sulfolipid were organized in
a complex resulting in metabolite channeling. Some of the Calvin cycle
enzymes are thought to form such a complex (2). It is assumed that the

respective metabolites do not freely exchange with the surrounding medium,

177

Figure 6-6: Effect of the addition of various compounds on the incorporation
of asS-PAPS into sulfolipid. (A) Addition of a metabolite mixture as mentioned
in the text, and of a concentrated extract (25 pl/reaction) from wild type cells
prepared as described in the method section. Symbols: closed circles, no
additions; open circles, extract from wild type cells; closed squares, metabolite
mixture; open squares, metabolite mixture and extract. Aliquots were
incubated for 2, 4, and'8 min prior to analysis of the reaction products. (B)
Effect of different concentrations of DL-glycerate in the reaction mixture.
Symbols: closed circles, 0 mM; open circles, 1 mM; closed squares, 3 mM;
and open squares 10 mM of DL-glyceraldehyde present in the reaction
mixture. The reaction products were analyzed after 2, 4, and 6 min of
incubation. (C) Effect of the addition of UDP—glucose. The concentrations in
the reaction mixtures are indicated in the Figure. Reaction products were
analyzed after 2, 4, and 6 min of incubation. The amount of radioactivity found
in the sulfolipid band on TLC plates as determined by scintillation counting is
indicated on the y-axis for all three data sets. In all three experiments 10 pM
concentrations of PAPS were used per reaction mixture (specific activity about
1 pCi/nmol).

Radioactivity in SL (cpm)

2400

2000

1600

1200

800

400

1400
1200
1000
800
600
400

200

2000

1600

1200

800

400

178

 

I

I

I

I

 

.fll‘

 

 

 

l l L

 

2 4 6

 

I

I

 

UDP-Glucose (mM)
l.-—-.MLO

EEEi’i 5%
/

 

 

 

 

 

179

but instead are passed along from reaction to reaction by the enzymes in the
complex. Such a process would imply that intermediates could not be easily
fed to the enzymes. Based on this result, the metabolite mixture mentioned
above, which was added to the reaction mixtures in the initial experiments,
was omitted in later experiments.

Two compounds, DL—glyceraldehyde and UDP-glucose, were added to
the reaction mixtures, to test whether they could inhibit the incorporation of
label from PAPS into sulfolipid. DL-glyceraldehyde is a competitive inhibitor of
the aldolase in the Calvin cycle (6). This, or a similar enzyme was proposed to
catalyze the synthesis of sulfoquinovose-1-phosphate from sulfolactaldehyde
and dihydroxyacetone-3-phosphate (see chapter 1, Figure 1-2). However, in
one exploratory experiment, addition of DL-glyceraldehyde up to 10 mM did
not inhibit the incorporation of asS-PAPS into sulfolipid (Figure 6-6, B). On the
other hand, the addition of 3 mM UDP-glucose resulted in slight reduction of
the rate of incorporation of PAPS into sulfolipid (Figure 6-6,C). The sugar
nucleotide UDP-glucose was tested, because it resembles UDP-
sulfoquinovose, which was thought to be a precursor of sulfolipid biosynthesis
(chapter 1). No firm conclusions should be drawn from a single, exploratory
experiment. However, future experiments may show that UDP-glucose plays a
role as competitive inhibitor or substrate of sulfolipid biosynthesis in R.
sphaeroides.

In extracts of Chlamydomonas, the only other in vitro system described

until now (3), labelled sulfite could be incorporated into sulfolipid. However,

 

 

   

 

 

 

 

 

 

 

180
A
GE?» 0.30 . r
Sulfite (11M)
% 0—0: 0 Q q
E 0-25 ' e—e: 100
3 A—A: 250 O
m ' III—CI: 750 O A
3 I—I: 1000 ./A
.E O 15 " /
Q
g 0.05 - E
U)
9“ 0 00 ‘
CL ' 1 2 :5 4 5 5 7
Time (min)
B
‘C 3.0 . 4
0‘
E
_o 2.5 n g
E
3
.1 2.0 - \
m 0
O
E 1'5 ' \.\
c O
O
E 1.0 ~ ‘/ \
g. e
8 0.5 -
E
In
% 0.0 1 1 1 1 L g
“- -0.2 0.0 0.2 0.4- 0.6 0.8 1.0 1.2

Concentration of Sulfite in Assay (mM)

Figure 6-7: Effect of sulfite on the incorporation of 3“S-PAPS into sulfolipid. (A)
Rates of sulfolipid synthesis from PAPS in the presence of various
concentrations of sulfite as indicated in the Figure. The total concentration of
PAPS in the reaction mixtures was 10 pM (specific activa 1 pCi/nmol).
Samples were analyzed after 2, 4, and 6 min of incubation in the presence of
labelled substrate. The rates were calculated from the radioactivity found in the
sulfolipid bands upon separation of the reaction products by TLC. (B)
Inhibition of sulfolipid biosynthesis by sulfite. The rates of sulfolipid synthesis
calculated from the data shown in (A), were plotted against the concentration
of sulfite in the reaction mixtures.

181

Table 6-1: Effect of sulfur containing compounds and aspartic acid on the
incorporation of 35S-PAPS into sulfolipid. Samples were extracted after 2, 4,
and 6 min of incubation in the presence of ass-PAPS. The rates, presented as
pmol PAPS incorporated into sulfolipid per hour and mg protein, were
calculated from the radioactivity associated with the sulfolipid bands detected
upon separation of the reaction products by TLC. The concentration of PAPS
in the reaction mixtures was 10 pM (specific activity 1 pCi/ nmol). The results
shown were obtained from a single experiment, and therefore have to be
considered preliminary.

 

Compound Concentration Rate of PAPS Incorporation
Into (mM) Sulfolipid (pmol/h mg protein)
control 1 4.3

cysteine 1 3.2

cysteic acid 1 3.2

aspartic acid 1 2.1

reduced glutathione 1 3.7

dithiothreitol 1 1 .1

 

182

the process could not be saturated, indicating that sulfite was not behaving as
a legitimate substrate for an enzymatic reaction. To test the effect of sulfite in
the Fl. sphaeroides system, varying amounts of sulfite were added to the
reaction mixtures and its effect on the incorporation of ai‘S-PAPS into sulfolipid
was examined. Addition of 1 mM sulfite resulted in 5-fold reduction of the rate
of sulfolipid biosynthesis from PAPS. It was concluded that this inhibitory effect
of sulfite, which was observed repeatedly in different experiments, could be
due to the fact that sulfite was either an intermediate in the pathway of
sulfolipid biosynthesis from PAPS, or alternatively that sulfite was poisoning
the system in an unexplained way.

Addition of 1 mM DTT repeatedly inhibited the incorporation of PAPS
into sulfolipid, while 1 mM concentrations of cysteine, cysteic acid, and
reduced glutathione, tested only once in a preliminary experiment (T able 61),
showed no pronounced inhibitory effect. Cysteic acid was proposed to be a
possible intermediate of sulfolipid biosynthesis (compare chapter 1), but had
very little effect on sulfolipid biosynthesis in cell extracts. Based on this single
experiment, it seemed likely that cysteic acid was not a precursor of sulfolipid
biosynthesis in R. sphaeroides. However, further experiments are required to
confirm the preliminary data. Aspartic acid, which resembles cysteic acid in its
structure and physical properties, showed some inhibition of sulfolipid
biosynthesis in extracts (T able 6-1). Glutathione was reported to stimulate in
vitro sulfolipid biosynthesis in extracts from Chlamydomonas (3), but appeared

to have little effect in the R. sphaeroides system (T able 6-1).

1 83
CONCLUSIONS

The series of experiments described above demonstrated that crude
extracts from R. sphaeroides cells are capable of catalyzing the incorporation
of label from asS-PAPS into sulfolipid. Amongst other unidentified, chloroform-
soluble compounds, which were believed to be an artifact of the in vitro
system, sulfolipid could be identified. The evidence for the identity of sulfolipid
was threefold: first, extracts isolated from the sulfolipid deficient squ mutant
could not label the putative sulfolipid from a“S-PAPS (Figure 63); second, the
putative sulfolipid band comigrated with authentic sulfolipid labelled from 3:‘S-
sulfate in vivo in two different TLC-systems; and third, the deacylation product
of presumptive in vitro and in vivo labelled sulfolipid comigrated in a two
dimensional TLC-system (Figure 6-4). The fact that only broken but not intact
cells could label sulfolipid from asS-PAPS (Figure 6-1) confirmed that unbroken
cells contaminating the extracts could not be responsible for the observed
synthesis of sulfolipid, and supported the idea that the reactions indeed
proceeded in vitro. Dependence of the process on the extract (protein)
concentration (Figure 6-2) as well as substrate saturation (Figure 6-5; A, B)
indicated that asS-PAPS behaved as a legitimate substrate for a series of
enzymatic reactions involved in sulfolipid biosynthesis and excluded the
artifactual, chemical incorporation of label from asS-PAPS into sulfolipid.
Maximal rates of 20-30 pmol/h mg protein were observed for the synthesis of
sulfolipid from PAPS (Figure 6-5, B). Under suboptimal substrate

concentrations mostly used during the experiments to avoid the dilution of the

 

 

 

 

 

184

label with non-radioactive substrate, rates of about 3 pmol/h mg protein were
observed (Figure 6-7, B, Table 6-1). In a series of preliminary experiments, it
was determined that sulfite (Figure 6-7; A, B) and DTT (T able 6-1) inhibited
sulfolipid biosynthesis from PAPS. A substrate analog of aldolase,
DL-glyceraldehyde, had no effect (Figure 6-6, B), while addition of UDP-
glucose, a compound similar to UDP-sulfoquinovose, inhibited the process
slightly (Figure 6-6, C). Cysteic acid, proposed as a possible precursor of
sulfolipid, had little effect, as did cysteine and glutathione (T able 6—1). Aspartic
acid, on the other hand, which chemically resembles cysteic acid, appeared to
reduce the rate of sulfolipid biosynthesis. However, these preliminary should
be used only as guidance for future research, rather as a basis for
conclusions.

The synthesis of sulfolipid in cell extracts was not stimulated by the
addition of metabolites such as ATP , NADPH, and various sugar phosphate
esters (Figure 6-6, A). This result was not surprising, since highly concentrated
cell extracts were used in the reactions, which might have provided all the
required metabolites in non-limiting concentrations. Serious attempts to purify
individual components of the system have not yet been undertaken. Given the
limited information available about the nature of precursors of sulfolipid
biosynthesis, and the unavailability of labelled, putative precursors, the
resulting approach of trial and error seems to be cumbersome and time
consuming. However, with the isolation of sulfolipid deficient mutants of H.

sphaeroides described in this study, which were shown to accumulate

185

possible precursors, the analytical process of dissecting the components of
sulfolipid biosynthesis assumes a new dimension. The next step should be to
isolate and purify these compounds from the mutants and to test them as
precursors in the in vitro system described in this chapter. Information gained
from structural elucidation of these compounds will be useful to design
experiments involving new substrates and inhibitors of sulfolipid biosynthesis,
which can be tested in the in vitro system described. Therefore, the data
presented in this chapter provide a crucial cornerstone for further analysis of
the pathway of sulfolipid biosynthesis in R. sphaeroides, and ultimately in

higher plants.

REFERENCES

1. Bradford MM. (1976). A rapid and sensitive method for the
quantitation of microgram quantities of protein utilizing the
principle of protein-dye binding. Anal. Biochem. 72:248-252

2. Gontero M., Cardenas M.L., Ricard J. (1988). A functional five-
enzyme complex of chloroplasts involved in the Calvin cycle. Eur.
J. Biochem. 173:437-443

3. Hoppe W., Schwenn JD. (1981). In vitro biosynthesis of the plant
sulpholipid: on the origin of the sulphonate head group. Z.
Naturforsch. 360:820826

4. Joyard J., Forest E., Blee E., Douce R. (1988). Characterization
of elemental sulfur in isolated intact spinach chloroplasts. Plant
Physiol. 88:961-964

5. Robbins PW. (1962). Sulfate activating enzymes. Methods
Enzymol. 52964-977

186

Stokes D.M., Walker DA. (1972). Photosynthesis by isolated
chloroplasts. Inhibition by DL-glyceraldehyde of carbon dioxide
assimilation. Biochem. J. 128:1 147-1 1573.

Vargas F. (1988). Preparation and quantification of 3’-
phosphoadenosine 5’-phospho[35$]sulfate with high specific
activity. Anal. Biochem. 172:82-88

CHAPTER 7

CONCLUSIONS AND PERSPECTIVES

Since the discovery and structural elucidation of the plant sulfolipid by
Benson over thirty years ago (2), relatively little progress has been made
towards the elucidation of the biosynthetic pathway and the function of this
unusual lipid. This is surprising, since the occurrence of substantial amounts of
the plant sulfolipid in all photosynthetic organisms examined make it one of the
most abundant, sulfur-containing organic compounds on earth. Therefore, this
compound plays an important role in the global sulfur cycle and is a
substantial ingredient of the human and animal diet. Lack of progress in our
understanding of sulfolipid biosynthesis and function cannot be blamed entirely
on lack of interest by the scientific community, since a number of researchers
tried to solve the problem during the past thirty years. Among the reasons for
slow progress might have been the fact that our general knowledge about
sulfur metabolism in plants is not very deep. Not much knowledge can be
derived from animal systems, since mammals are not capable of assimilating
inorganic sulfur. This process is unique to plants and bacteria and must,
therefore, be studied in these organisms. Studying sulfur metabolism in general
seems to be very complicated, due to the diversity and reactivity of sulfur-
containing compounds. Sulfolipid biosynthesis brings together two plant-
specific processes: first, photosynthetic sugar biosynthesis and second,

photosynthetic sulfur assimilation. At present, we have quite a deep

187

188

understanding of sugar biosynthesis in plants, and many of the ideas about
sulfolipid biosynthesis, as outlined in chapter 1, were derived from our
knowledge about this process. The elucidation of the pathway for the sugar
biosynthesis took many years and was facilitated by the fact that most of the
intermediates are highly abundant in photosynthetically active leaves.
Intermediates of sulfolipid biosynthesis, on the other hand, are less abundant.
Given these facts, it becomes clear that the study of sulfolipid biosynthesis
provides an intellectual challenge, and that novel ideas are required to resolve
this problem.

The objective of the study presented here was to develop an
experimental system which would facilitate a solution of the classical problem of
sulfolipid biosynthesis. A genetic approach to dissect the components of
sulfolipid biosynthesis had never been tried before. In classical studies,
complicated biosynthetic pathways, such as the biosynthesis of histidine, were
elucidated with the help of bacterial mutants, which were blocked at various
steps of the pathway (1, 4). It was assumed that studying mutants that showed
blocks in the pathway of sulfolipid biosynthesis would provide the means to
elucidate the pathway for sulfolipid biosynthesis. The organism of choice for
such a study was the purple non-sulfur bacterium FI'. sphaeroides. It contained
the plant sulfolipid with its typical 6-sulfo-6-deoxy hexose head group, as was
confirmed by comparative mass spectroscopy of isolated sulfolipid from
spinach and R. sphaeroides. As described in chapter 2, a characteristic,

comparable fragmentation pattern for the sulfolipid head group from spinach

189
and R. sphaeroides could be observed employing negative FAB-MS/MS

techniques.

The purple non-sulfur bacterium R. sphaeroides had distinct advantages
over other organisms for a genetic approach, since common genetic
techniques were available for this bacterium, and, more importantly, this
organism was able to grow in the absence of photosynthesis. Although the
function of sulfolipid is not clearly established, there was considerable concern
that obligately photoautotrophic organisms, such as higher plants or many
cyanobacteria, would not be able to grow in the absence of sulfolipid, causing
lethality of possible sulfolipid-deficient mutants.

Using a novel screening procedure, the examination of lipid extracts from
chemically mutagenized R. sphaeroides cells by TLC made it possible to isolate
a series of mutants deficient in sulfolipid biosynthesis (chapter 3). The analysis
was focused on three mutants (squ, squ, and squ), two of which
accumulated 35S-labelled, water-soluble compounds that may be precursors of
sulfolipid biosynthesis. Structural analysis of these compounds as well as the
elucidation of their possible role as precursors in sulfolipid biosynthesis, will
provide the crucial data to establish a pathway.

However, the next logical step in developing the R. sphaeroides model
system for the study of sulfolipid biosynthesis, following the isolation and
phenotypic characterization of the sulfolipid deficient mutants (Chapter 3), was
thought to be the isolation and characterization of the genes complementing

the mutations. All strains containing the mutations causing the sulfolipid

190

deficient phenotype contained background mutations due to the extensive
chemical mutagenesis process. To determine that the observed phenotype of
the mutants such as accumulation of compounds or growth changes as
compared to wild type cells was due to the mutation causing the sulfolipid
deficiency, isogenic lines had to be created. This is normally accomplished
either by gene replacement of the respective wild type gene with a defective
copy, or by comparing the mutants to the same mutant strain carrying the
complementing gene on a single-copy plasmid. Both approaches required the
isolation of the respective genes. In addition, it was assumed that the isolation
and analysis of the genes would provide information about the number of
complementation groups involved and the structure and possible function of the
respective gene products.

To isolate the genes complementing the different mutants, a cosmid
library was constructed and a complementation assay was devised, through
which complementing cosmids were identified (Chapter 4). By screening the
cosmid library, two cosmid classes were identified, one of which was able to
complement the squ and squ mutants, while the second complemented the
squ mutant. From these two cosmid classes distinct subclones were isolated,
which were able to complement each of the three mutants. Sequence analysis
of the clone complementing the squ mutant revealed an open reading frame
which could encode a 33.6-kD protein. Further deletion and complementation
analysis, and localization of a ribosome binding site in front of the open reading

frame allowed the identification of the putative squ gene (Chapter 4). The

191

analysis was facilitated by the construction of an expression vector for R.
sphaeroides (Appendix B), which allowed the orientation-dependent expression
of reading frames to determine the coding strand for the squ gene. The role
of the squ gene product in sulfolipid biosynthesis could not be deduced from
homology of the translated protein sequence with protein sequences of known
enzymes. Further experiments will be required to elucidate the function of the
squ gene product. Analysis of the compounds accumulating in the squ
mutant seems to be the most promising strategy to accomplish this goal. The
availability of the squ gene will make it possible to construct isogenic lines,
which are crucial for the analysis of the precursor role of the accumulating
compounds, and the elucidation of the function of sulfolipid by comparative,
physiological experiments. In addition, expression of the squ gene in E. coli
and production of antibodies against the purified product will provide further
tools to study the role of the squ gene in sulfolipid biosynthesis.

The putative gene complementing the squ mutant which was found to
be located within a different set of cosmids than the squ gene, was similarly
analyzed. The putative squ open reading frame, identified by sequence
analysis of a small complementing subclone derived from the complementing
cosmid, could encode for a 46-kD protein (Chapter 5). The identity of this open
reading frame with the squ gene was further confirmed by deletion and
complementation analysis using a novel R. sphaeroides expression vector
(Appendix B). Most interestingly, The translated amino acid sequence for the

putative squ gene product showed N-terminal homology with UDP-glucose

192

epimerase from bacteria and fungi. This result was thought-stimulating, because
.UDP-glucose resembles UDP-sulfoquinovose, a proposed precursor of
sulfolipid biosynthesis. Based on this homology, it was suggested that the squ
gene product might be involved in sulfolipid biosynthesis at the level of UDP-
sulfoquinovose or another nucleotide sugar (Chapter 5).

The squ gene was found to be complemented by the same cosmid
that complemented the squ gene. A subclone derived from this cosmid which
contained two inserts was able to complement this mutant (Chapter 5).
Southern analysis indicated that parts of this subclone were located directly
downstream of the squ gene. A fragment of the cosmid downstream of and
partially overlapping the squ gene, was found to complement the squ
mutant in an orientation-dependent fashion when cloned in two orientations into
an expression vector. The conclusion drawn from this result was that the squ
gene is located directly downstream of the squ gene, and that its direction of
expression coincides with that assumed for the squ gene. Based on this
observation, it is proposed that the squ and squ genes are organized in an
operon, coding for genes involved in sulfolipid biosynthesis. However, this idea
.has to be confirmed by further experiments. The squ gene on the other hand,
does not seem to be closely linked to the other two genes and may therefore
belong to a second operon or gene cluster involved in sulfolipid biosynthesis.
Using a recently established, physical map of the complete R. sphaeroides
genome (5, 6), it should be possible to determine the relative location of the

two sets of genes on this map by Southern hybridization.

193

Most likely, there are more genes involved in sulfolipid biosynthesis than
the three which have been described in this study. Some of the putative
mutants which were isolated during this study (Chapter 3), but have not been
further analyzed, may mark additional genes involved in sulfolipid biosynthesis.
A detailed analysis of the DNA sequences surrounding the two gene clusters
may reveal novel sqd genes, since genes catalyzing a metabolic sequence
frequently occur in operons or clusters in bacteria. Using the cosmid DNA, it
should be possible to construct large deletions in the genome of
Ft. sphaeroides by gene replacement techniques, covering the sequences
surrounding the known sqd genes. Cosmids mutagenized in vitro could than be
analyzed in trans for their capability to complement the deleted genes, thereby
possibly revealing the location of novel sqd genes. Alternatively, a series of
gene replacements spanning the region of the cosmid could be constructed.

Finally, to elucidate the pathway for the biosynthesis of the plant
sulfolipid, the genetic analysis has to go hand-in-hand with biochemical
analysis. The clue to solving the pathway seems to be the analysis of the
compounds accumulating in the squ and squ mutants. To determine their
precursor function for sulfolipid biosynthesis, an in vitro sulfolipid synthesis
system using extracts from wild type Ff. sphaeroides cells was established
(Chapter 6). Although the system is still crude and not very well developed, it
can be useful to test the capability of the labelled compounds accumulating in
the mutants, thereby establishing their precursor function. Experiments with

intact cells would not necessarily be able to provide conclusive results, since it

194

is unknown whether these compounds can be taken up by the cells. In
addition, inhibitors and putative substrate metabolites can be tested in vitro,
unaffected by problems of uptake. The in vitro system described in Chapter 6,
provides the basis for the biochemical dissection of the pathway for sulfolipid
biosynthesis.

To summarize the findings of this study, one can say that
R. sphaeroides was established as a genetic model system for the study of
sulfolipid biosynthesis. First, sulfolipid deficient mutants were obtained and
characterized; second, three genes were isolated, which seem to be directly
involved in sulfolipid biosynthesis in Ff. sphaeroides, and third, the biochemical
basis for solving the pathway of sulfolipid biosynthesis has been provided by
establishing an in vitro sulfolipid synthesis system. Based on the data and
materials provided by this study, a series of experiments can be designed, as
pointed out throughout the text, which will provide avenues to elucidate the
pathway for sulfolipid biosynthesis in H. sphaeroides in the near future. Solving
the structure of the novel compounds accumulating in the mutants and
determining their role in sulfolipid biosynthesis seems to be the key to the
solution of the problem. Efforts to purify and characterize these compounds
have just started.

It is anticipated that exploration of the described genetic model system
will invigorate the field of sulfolipid research. The recent finding that
sulfoquinovose diacylglycerol isolated from cyanobacteria shows AIDS-antiviral

action (3) may increase interest in this compound in the future. It may turn out

195

that the study of plant sulfolipid not only represents an intellectual challenge for

scientists, but also has an application in everyday life.

REFERENCES

1. Ames B.N., Martin R.G., Garry B.J. (1961). The first step in
histidine biosynthesis J. Biol. Chem. 236:2019-2026

2. Benson AA. (1963). The plant sulfolipid. Adv. Lipid Res. 1:387-394

3. Gustafson K.R., Cardellina ll J.H., Fuller R.W., Weislow O.S., lfiser
R.F., Snader K.M., Patterson G.L.M., Boyd MR. (1989). AIDS-
antiviral sulfolipids from cyanobacteria. J. Natl. Canc. Inst.

81 :1 254-12583.

4. Smith D.W.E., Ames EN. (1964). Intermediates in the early steps
of histidine biosynthesis. J. Biol. Chem. 239:1848-1855

5. Suwanto A., Kaplan S. (1989). Physical and genetic mapping of
the Rhadabacter sphaeroides 2.4.1 genome: genome size,
fragment identification, and gene localization. J. Bacteriol.

171 :5840-5849

6. Suwanto A., Kaplan S. (1989). Physical and genetic mapping of
the Hhadabaater sphaeroides 2.4.1 genome: presence of two
unique circular chromosomes. J. Bacteriol. 171:5850-5859

APPENDIX A
THE EFFECT OF SULFOLIPID DEFICIENCY ON GROWTH OF

Ff. SPHAEFl’O/DES

With the isolation of sulfolipid-deficient mutants of R. sphaeroides and
their complementing genes, it became possible to create isogenic lines, that
would allow a direct test of the role of sulfolipid. In a preliminary experiment
two isogenic lines of R. sphaeroides containing different amounts of sulfolipid
in their membranes were compared for their ability to grow
photoheterotrophically. The sulfolipid-deficient squ mutant (strain CHB16)
carrying the wild type squ gene on a plasmid (pCHB16201; Chapter 5,
Figure 5-7), which restored the wild type phenotype with respect to the
sulfolipid content, was compared to the squ mutant carrying the vector
plasmid alone (pCH8500; Appendix B), resulting in a sulfolipid-deficient strain
(Figure A-1). These two strains were carrying the same mutations in their
genome, except that one of the strains contained the wild type squ gene in
trans, while the other did not. Hence any phenotypic differences observed,
should be due to the absence or presence of an intact copy of the squ
gene. To examine whether the plant sulfolipid plays a role in photosynthesis
(see Chapter 1 for a complete discussion of sulfolipid function) the two strains
were incubated under conditions which would allow them to grow
photoheterotrophically. For this purpose, 8 ml of Sistrom’s medium, containing

0.8 pig/ml tetracycline for selection of the plasmids, were inoculated with the

196

 

197

Photoheterotrophic Growth of CHB1 6/ 500 and CHB16/16201

ii
2

 

 

 

 

'— "0'“
I A .
E 100 O.CH816/500 /O _
c: .1CH816/16201 . 1 - ’ PG
lg Q/
l /
O C
8 SL
v 10 -
C
3
8
Q ~ - PC
1 C
I I l l I I origin
0 4 8 12 16 2O 24 28
Time (h)

Figure A-1: Comparison of two isogenic lines of R. sphaeroides, containing
different amounts of sulfolipid, for their heterotrophic growth. Two lines of the
sulfolipid deficient squ mutant were compared, one containing the pCH8500
vector (CHB16/500) and the other, carrying the wild type squ gene on the
plasmid pCHB16201 (CHB16/16201). Lipid extracts obtained from the two
strains, separated an ammonium sulfate-treated silica plates are shown on the
right. PG, phosphatidyl glycerol; SL, sulfolipid; PE, phosphatidyl ethanolamine;
PC, phosphatidyl choline. Growth curves for both strains, obtained during
logarithmic growth, are shown on the left. The cultures were grown as
described in the text.

 

 

198

two strains described above. Screw-cap tubes were used, which would fit into
a Klett photometer equipped with a filter of 660-700 nm cutoff. The tubes
were percolated with a steady stream of a gas mixture of 5% 002 and 95%
N2, and the cultures were incubated in a water bath at 30°C under a 60 Watt
incandescent bulb (100 pE/s per m2). The growth of the cultures was
monitored at various times by measuring the absorption with the aid of a Klett
photometer. Replicates for each strain were monitored, and were not found to
be very variable.

The result of the experiment described (Figure A-1) indicated that the
two strains showed indistinguishable growth rates under the chosen
conditions, despite drastic differences in their sulfolipid content. However, it
would be premature to conclude from this experiment that sulfolipid plays no
role in photosynthesis. As indicated in Chapter 3, the squ mutant (strain
CHB16) was found to be leaky, and the residual amounts of sulfolipid present
in the membranes of the squ mutant may be sufficient to maintain a vital role
of sulfolipid in the squ mutant. To clarify this point, it will be necessary to
create a null mutation by gene replacement. If the resulting mutant still grows
with the same rate as the wild type under the conditions described, this may
indicate that sulfolipid is not important for photosynthetic growth in
R. sphaeroides. If in such an experiment a difference between the wild type
and the mutant were to be observed, a comparative physiological and

biochemical study could be initiated to study the function of plant sulfolipid.

 

APPENDIX B

CONSTRUCTION OF THE EXPRESSION VECTOR pCHB500

Crucial for the analysis of the coding sequence for the squ and squ
genes as described in Chapters 4 and 5 was the construction of a vector
which would allow the construction of a transcriptional fusion between a
promotor functional in R. sphaeroides and a potential reading frame.
Promotors of E. ca/i were shown not to work in H. sphaeroides (F. Daldal,
personal communications), but endogenous promotors of Fl. sphaeroides
have not yet been described in detail, and were not available. F. Daldal
suggested the use of a fragment upstream from the cycA gene of Ff.
capsu/ata (2), which he knew was able to promote transcription in R.
sphaeroides. Based on this information, a 350 bp fragment beginning just
upstream of the ribosome binding site of the cycA gene of Ft. capsu/ata was
isolated by PCR (Figure B-1). The two 30-bp oligonucleotide PCR primers
were designed such that one end of the PCR product could be cut by Hindlll,
while the other and contained a Pstl site (Figure B-2). This design allowed the
directional cloning of the promotor fragment into the polylinker of the broad
host range vector pRK415 (3), obtained from N.T. Keen, giving rise to the
expression vector pCH8500 (Figure B-3), which was used throughout much of
this work. The PCR reactions were performed using the linearized plasmid
pSH3 (cut with EcoRI) as template, which was kindly provided by F. Daldal

and contained the cycA gene complete with its promotor sequences from

199

200

Figure B—1: Sequence of the cycA gene from Ff. capsulata, upstream of the
initiation codon (taken from 2). The location of the two PCR primers CB22 and
C823 is indicated.

 

1 GTCGACCCGGZATCTGAGCCATTTTTCCGGCATCGTGCCCTGCGGCATCCGGAGCATGGCG
61 TGACCTCGCTCGTCGATCTGGGGCTGCCGGTGACGATGGCCGATCTCGACCTGGCGCTGC
121 GGGCCCGGTTCGAGACCGATTTCACCCCGCGCCTGCGCGGCTGAGCCGCCGCCGCAAGCG
181 CTGCCGGAGGGCGGCATTTCATCTGGCATTCACTTTGCGTCCGGATTCTTGCCCCGCGGG
241 TCATGGAATTGAAAGCAAAGGGATATGCGCGACCTTTTGCCTngggTAAAGCGTGGACC

 

301 GGGTAAGGAATCTGCCCCTTGGCCCCGGCTCCAGACCGGGCAAGCAAACAGACGAAGGAT
361 ACAGCGATG

Figure B-2: Sequence of the two PCR primers used to clone the cycA
promotor from R. capsu/ata. The Hindlll (H3) and Pstl(P1) sites are indicated
above the sequence.

H3
CB22 5' -ATCGAAGCTTGTCGACCCGGATCTGAGCCA—3'

 

P1
CB23 5' -ATCGCTGCAGCGTCTGTTTGCTTGCCCGGT-3'

 

201

     

Smal/ADaI‘ pcuasoo

   

E°°RV~ 10.73 Kb

    

Hincl l/Sall'

 

5°" Pstl.Sall.HinclI.Accl.Xbal.BamHl.Smal.Kpnl.Sstl.EcoRl

Figure B-3: Circular restriction map of pCH8500. The orientation and location
of two promotors, Plac (promotor of the E. ca/i lac-operon) and Pcyc (putative
H. capsu/ata promotor of the cycA gene), are indicated by thick arrows. The
tetracycline gene (T cR) is indicated as thick a line. The locations of the origin
of transfer (oriT) and origin of replication (oriV) are shown. Restriction sites are
shown with their common names. The sites of the polylinker, downstream from
the Pcyc promotor, are shown in an expanded view.

 

202

H. capsu/ata. The PCR reactions were set up on a Perkin Elmer thermocycler
using the buffers, templates, and primers at concentrations recommended by
that company. Reactions of 100pl were run using the following temperature
program: 3 min at 95°C; 30 cycles of 1 min at 94°C, 3 min at 60°C, and 3 min
at 72°C; 7 min at 72°C. The reaction product was gel-purified and digested
with Hindlll and Pstl. Following ligation with pRK415 cut with the same
enzymes, and transformation of HB101 cells, several clones were identified in
DNA minipreps which contained the expected fragment, as was confirmed by
Southern analysis.

To test the capability of pCHBSOO to express open reading frames in an
orientation-dependent fashion, a 977 bp Sau3Al fragment obtained from
pBR328 (1) containing the open reading frame for the chloramphenicol acetyl
transferase (CAT) gene, was cloned in both orientations into the BamHI site of
the pCHBSOO polylinker. Upon transfer of the resulting clones into
R. sphaeroides and plating of the cells onto agar plates containing various
concentrations of chloramphenicol, it was observed that clones, which
contained the CAT gene in the same orientation as the cycA promotor, could
confer resistance in the presence of 10 pg/ml chloramphenicol, while clones
containing the CAT gene in the opposite orientation could not. A similar result
was observed , when a cassette containing the NPT open reading frame,
conferring resistance to kanamycin, was cloned in both orientations behind
the cycA promotor. Although the orientation-dependent expression was not

precisely quantitated, the data were in agreement with proper transcription

203

from the cycA promotor in R. sphaeroides. As described in Chapters 4, and 5,
the pCH8500 vector was successfully used to test the orientation of the
coding sequence for the squ and squ genes in complementation
experiments. The vector has been given to F. Daldal, who is interested in

analyzing the expression from the cycA promotor in more detail.

REFERENCES

1. Balbas P., Soberon X. Merino E., Zurita M., Lomeli H., Valle F.,
Flores N., Bolivar F. (1986). Plasmid vector pBR322 and its
special-purpose derivatives - a review. Gene 50:3-40

2. Daldal F., Chang S., Appelbaum J., Davidson E., Prince RC.
(1986). Cytochrome 02 is not essential for photosynthetic growth
of Fihadapseudamanas capsu/ata. Proc. Natl. Acad. Sci. USA
83:2012-2016

3. Keen N.T., Tamaki S., Kobayashi D., Trollinger D. (1988).
Improved broad-host-range vector for DNA cloning in Gram-
negative bacteria. Gene 702191-197

 

'IICHIGQN STQTE UNIV LIBRQRIES

lllllllll

31293009

 

_ l;i.,.'1,.'?:....,
"we.“ .

.... 1.2.. "2...

. ....~.-.,
‘y. .1 ...1-..,...

””13: 11.-
. ....

1:1..1'..~.1 -. 1......-
V. .,. .15.-.. p. a».
.. .1 .111. ~- ~ -'

. 3:1: 7:71:51!-

» 1

“burn.-

..1... ...
..11

1 v

. ....1

. e...- 1.1-nu...

1'.-

uo'y.
1.
.1 1.1 1... 11.... 1.1:-

‘ v
c. 1.1.
1v:~;1v1\0 7.19.1311;
q .1 1'" ..' 1.. 1 ..'
1.
'1

V..L~.

 

 

 

 

 

.. [a ‘9‘
..11..,. .. 1 1 1.1n1..~'-1>~11‘M--
.,-..-... .. ... ywlvan ..«1- Ks:-

1 .
'—_41-1‘-.
.". ;‘.l "'L .

: 1

V “.1
.1. ..1' 11.114»
..‘11'....111-~

.1.
1.,1-..1u'

.'.'-1.»11-..,..
..1 11"”

III II I I

173554