‘ 1
Di
.

W...

 

 

w..1..._,ua.u.u..~_u.. .4 . y...) ...

 

Illlmjlliiii‘fiiilliiiIiiiiiillllllmli

1293 01771 8333

 

LIBRARY

Michigan State
University

Y

 

 

 

This is to certify that the

dissertation entitled

Camalexin biosynthesis in Arabidopsis: A study of putative

intermediates and of the effects of different pathogens
on its production

presented by

Isabelle Kagan

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Botany & Plant Pathology

 

Wm

Major professor

Date W

MS U it an Affirmative Action/Equal Opportunity Institution 0-12771

 

PLACE iN RETURN BOX to remove this checkout from your record.
TO AVOID FINB return on or before date due.
MAY BE RECAU£D with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

-— Y" "n
”"1121”‘1’03

 

 

 

 

 

 

 

 

 

 

 

 

 

use chIRC/DdeDuopGS—nu

CAMALEXIN BIOSYNTHESIS IN ARABIDOPSIS: A STUDY OF PUTATIVE
INTERMEDIATES AND OF THE EFFECTS OF DIFFERENT PATHOGENS ON ITS
PRODUCTION

by

Isabelle Ann Kagan

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Botany and Plant Pathology

1999

ABSTRACT
CAMALEXIN BIOSYNTHESIS IN ARABIDOPSIS: A STUDY OF PUTATIVE
INTERMEDIATES AND OF THE EFFECTS OF DIFFERENT PATHOGENS ON
ITS PRODUCTION

by

Isabelle Ann Kagan

The biosynthesis of camalexin was studied in order to
evaluate the role of this phytoalexin in resistance of
Arabidqpsis thaliana to disease. Mutant Arabidopsis
seedlings were screened for camalexin deficiency to search
for mutants with biosynthetic blocks and phenotypes that
would indicate the effects of camalexin deficiency on disease
resistance. Some plants, which were camalexin-deficient in
response to the fungus Cbchliobolus carbonum, produced wild-
type amounts of camalexin in response to the bacterium
Pseudomonas syringae pv. maculicola (PSm). Radiolabeling
studies with pad (phytoalexin-deficient) mutants suggested
that some of these mutants produced more camalexin in
response to C. carbonum than in response to Psm. Camalexin
accumulation was compared in wild-type and pad2.Arabidopsis
inoculated with different pathogens. Wild—type plants
produced comparable amounts of camalexin in response to both
pathogens, and pad2 plants produced little in response to
either pathogen. Camalexin usually accumulated more rapidly
in response to C. carbonum than in response to Psm, although
some exceptions reinforced the importance of evaluating

camalexin production at more than one point in time when

assessing the plant’s ability to synthesize it. No camalexin
was detected in plants inoculated with 106 to 107 colony—

forming units per milliliter of Pseudomonas syringae pv.
syringae. Attempts to characterize inducible compounds led
to the isolation of indole-3—carboxaldehyde, a putative
intermediate, in fungal—inoculated wild—type and pad2 plants.
Kinetic and radiolabeling studies indicated that it is a
possible biosynthetic intermediate but that a pathway
independent of indole—3fcarboxaldehyde may operate as well.
A study of twenty—four ecotypes of A. thaliana revealed
quantitative differences in camalexin production and in
susceptibility to the fungus Alternaria brassicicola.
However, no firm correlation between resistance and camalexin
production was found, possibly due to the relative

insensitivity of A. brassicicola to camalexin.

ACKNOWLEDGMENTS

I thank my adviser, Ray Hammerschmidt, for a great
deal of patience and encouragement. I also thank Dr. Alois
Ffirstner (Max Planck Institut ffir Kohlenforschung, Germany)
for a gift of synthetic camalexin; and Dr. W. Reusch (Dept.
of Chemistry, MSU) and Dr. J. Kagan (Dept. of Chemistry,
University of Illinois at Chicago) for letting me
synthesize some camalexin in their labs. Special thanks is
due to Dr. Michael Zook, visiting assistant professor in
the MSU Department of Botany and Plant Pathology, for
instruction of many of the techniques that I used in this
project and for unfailing willingness to help with
experiments and answer questions. I am also grateful to
Corey Sonnett for help with the mutant screen. Finally, I
thank my parents for a great deal of moral and scientific
support, and all of the many people at MSU who helped me
with various parts of this project and made me feel welcome

at MSU.

iv

TABLE OF CONTENTS

LIST OF TABLES....... ..... . ........ .............. ....... vii
LIST OF FIGURES.. ......................................... x
Chapter 1. General Introduction
References.........................................32
Chapter 2. Screen For Camalexin-Deficient Mutants Of

Arabidopsis And Radiolabeling Studies With Phytoalexin-
Deficient (pad).Mutants

Introduction.......... ........ .. ................... .38
Materials and Methods.. ....... .......... ......... ...42
Results....... ...... O OOOOOOOOOOOOOOOOOOOOO 00.0.0.0..51

DiscuSSionOOOOOOOOOOOOOOOOOOOOOOO...0..0.0.0.000000068

ReferenceSOOOOOOOOOOOOOI...OOOOOOOOOOCOOO0.0.0.0....76
Chapter 3. Comparison of Patterns of Camalexin Accumulation
in Response to the Fungal Pathogen Cochliobolus carbonum and
the Bacterial Pathogens.PseudOmonas syringae pv..maculicola
and Pseudomonas syringae pv. syringae

IntIOductionooooooooooo00000000000000oooo ...... 0.....78

Materials and Methods.. .......... . ................ ...80
Results 0 O O O O O O O ..... O ..... O OOOOOOOOOOOOOOOOOOOO O ..... 93
DiscuSSione O O O O O O O ..... O O O O O O ....... O O O O O O O O ........ O 117

References...........................................126

Chapter 4. Isolation of Indole-3-carboxa1dehyde from
Arabidopsis Leaves Inoculated with cechliabolus carbonum, and
Exploration of its Possible Role as an Intermediate in
Camalexin Biosynthesis

Introduction 128
materials and HethOds O O O O O O O O O ..... O O O O O O O O O O O O O O O O O O 1 3 1
Resu1ts O O O O O O O O O O O O O O O O O O O O O O O ...... O O ....... O O O O O O O O 142

References............................ ...... .........200
Chapter 5. Cdmparison of Camalexin Production and Resistance
to the Fungus Alternaria brassicicola Among Twenty-Four
Ecotypes of Arabidopsis thaliana.

Introduction............ ............ .. ...... .........202

Materials and methods ............ ..... ..............206

Results..............................................213

Discussion...........................................233

References... ........ . ..................... . ...... ...236

canoluSionSOCOOOOOOOOOOOOOOOOOOOOOOOOO0.0......0.0.00.0...239

”PEme 0. ..... ...OOOOOOOOOOCOO 0000000000 00.0.00000000000243

vi

LIST OF TABLES

Table 1. Camalexin in 2 putative camalexin-deficient
heterozygotes...0.0.0.0000...O. ....... 0.00.00.00.00000000054

Table 2. Camalexin in putative camalexin-deficient
mutants 4648 and 4512, and in padl and (padz.............56

Table 3. Comparison of camalexin concentrations in
inoculum droplets from fungal- and.water-inoculatsd
leaves of wild-type, padz, and pad3 p1ants................57

Table 4. Rf values and descriptions of inducible
bands in different TLC solvents...........................95

Table 5. Rf values of camalexin, indols-3-carbox-
aldehyde, indole-3-carboxylic acid, and anthranilic
acid in different TLC solvents............................96

Table 6. Time course of camalexin accumulation
in wild-type (wt) leaves inoculated with Cochliobolus
carbonum (inoc) or water (ctrl), 8/25/95-8/27/95..........100

Table 7. Time course of camalexin accumulation
in wild-type leaves inoculated with C. carbonum,
10/17/95-10/20/9590000000.000000000000000.00.0.00000000000101

Table 8. Effects of spore inoculum concentration on
camaler-n prMUCtion’ 11/3/9SOOOOOOOOOOO0.0.0.00.000000000104

Table 9. Effects of spore inoculum concentration on
cmleXin prOduCtion, 11/13/95.00000000000000000000000.0.0106

Table 10. Time course of camalexin accumulation in
wild-type and padz leaves inoculated.with C. carbonum

or Pseudomonas syringae pv..maculicola strain E54326,
12/31/95-1/3/96...........................................110

Table 11. Time course of camalexin accumulation in
wild-type and padz leaves inoculated with C. carbonum
or Pam E84326, 1/20/96-1/23/96....OOOOOOOOOOOOO...0.0.0.00112

Table 12. Time course of camalexin accumulation in
wild-type and padz leaves inoculated with C. carbonum
or Pam E84326’4/5/96-4/8/9600IOOOOOOOOOOOOOOOO0.0.0.000000114

Table 13. Time course of camalexin accumulation in
‘wild-type and padz leaves inoculated (inoc) with C.
carbonum or Pseudamonas syringes pv. syringes (Pss)
strain D20, 7/23/97-7/26/97...............................118

vii

Table 14. Time course of camalexin accumulation in
wild-type and pad2 leaves inoculated (inoc) with C.
Carmnum or P88 020' 9/21/97'9/25/970000000000.0.000...0.0120

Table 15. Recovery of indole-B-carboxaldehyde in
the absence of plant components...........................146

Table 16. Percent recovery of indole-3-carboxaldehyde
from.leaves spiked with a standard and extracted in the
manner of the nonradioactive studies of the kinetics of
indols-3-carboxaldehyde accumulation..................... 148

Table 17A. Percent recovery of indole-3-carboxaldehyde
from leaves fed a nonradioactive solution of indole-B-
carmxaldehYdeOOOOOOOOOOOO00.00.000.000.00....0.00.00.00.0149

Table 178. Recovery of indole-3-carboxaldehyde from
leaves spiked with aldehyde at time of extraction.........150

Table 18. Relative amounts of camalexin (expressed as
percentages of the total) in the phases obtained from the
liquid nitrogen extradtion method used for most
nonradioactive samples in this chapter. .................152

Table 19. Percent recovery (mean plus standard error) of
camalexin (cam.) from leaves spiked with a camalexin
standard and extracted in the manner of extractions done

in the radiolabeling experiment.......................... 153

Table 20A. Time course of accumulation of camalexin in
'wild-type and pad? leaves inoculated with water (ctrl)
or CO carwnum (inOC); 3/23/97-3/24/9700000000.0.0000....0155

Table 208. Time course of accumulation of indole—3-
carboxaldehyde wild-type and padz leaves inoculated

'with water (ctrl) or C. carbonum (inoc); 3/23/97-
3/24/97...................................................156

Table 21. Time course of accumulation of camalexin

(A) and indole-3-carboxaldehyds (B) in wild-type and

(pad? leaves inoculated with water (ctrl) or C. carbonum
(inoc); 5/31/97-6/1/97....................................159

Table 22. Accumulation of indole-3-carboxaldehyds

(ng/leaf, corrected for recovery) in wild-type and pad2
leaves inoculated with C. carbonum or water, 3/23/97-
3/24/97...................................................161

Table 23. Comparison of nanomoles of indole-B-

carboxaldehyde (corrected for recovery) and camalexin

(% recovery based on the results of Table 18) produced in
wild-type and pad2 leaves inoculated with C. carbonum.....16l

viii

Table 24. Time lapses during labeling of leaves..... ..... 164

Table 25. Percent incorporation (% inc) into chloroform
extracts of wild-type Arabidopsis leaves fed one of

solutions A, B, or C and then inoculated with water

(ctrl) or C. carbonum (inoc)..............................179

Table 26. Percent incorporation into camalexin in
wild-type Arabidopsis leaves fed one of solutions A, B,
or C and then inoculated with water or C. carbonum........182

Table 27. Percent incorporation into indole-3-carbox-
aldehyde in wild-type Arabidopsis leaves fed one of

solutions A, B, or C and then inoculated.with water or C.
carbonum..................................................186

Table 28. Average percent incorporation (mean plus
standard error of 2 replicates) into band 2 (Rf=0.81)

into leaves fed one of solutions A, B, or C and then
inoculated with water or C. carbonum......................189

Table 29. Average percent incorporation (mean plus
standard error of 2 replicates) into band 3 (Rf=0.71,

possibly indole) in leaves fed one of solutions A, B,
or C and then inoculated with water or C. carbonum........191

Table 30. Average percent incorporation (mean plus
standard error of 2 replicates) into band 3.5 (Rf-“0.67)

in leaves fed one of solutions A, B, or C and then
inoculated with water or C. carbonum......................193

Table 31. Average percent incorporation (mean plus
standard error of 2 replicates) into band 5 (Rf=0.46)

in leaves fed one of solutions A, B, or C and then
inoculated with water or C. carbonum.....................195

Table 32. Average percent incorporation (mean plus

standard error of 2 replicates) into band 7 (Rf=0.27)

in leaves fed one of solutions A, B, or C and then
inoculated with water or C. carbonum......................197

Table 33. Ecotypes studied for camalexin production
and A. brassicicola resistance: abbreviation, origin,
and disease phenOtYPeoooooeoooooo000.000.00.00...000.000.0217

Table 34. Camalexin production in response to
inmu1ation With C. carbOHMOOOOOOOOOOOOOOOOOI0.0.0.000000223

LIST OF FIGURES

Figure 1. A few cruciferous phytoalexins..... ...... .....10
Figure 2. Brassilexin...................................11
Figure 3. Camalexin.....................................13

Figure 4. Hypothetical relationship of the
camalexin and tryptophan biosynthetic pathways...........26

Figure 5. Two possible routes to camalexin..............29

Figure 6. Standard curve of camalexin concentration
versus fluorescence (excitation wavelength 330nm,
emission wavelength 393 nm)..............................53

Figure 7. Autoradiogram of TLC plate: tissue

extracts of wild-type (wt), padl (p1), pad? (p2),

and pad3 (p3) leaves, fed 14C-anthranilate 24 hours

after inoculation with water (c) or Cbchliobolus

carbonum (i), 6/27/94-6/28/94............................61

Figure 8. Autoradiogram of TLC plate: tissue

extracts of wild-type,‘pad1, and pad3 leaves fed
14C-anthranilate 24 hours after inoculation with

Co carbon“, 6/15/94-6/16/9400000000000000000..0.0000090062

Figure 9. Autoradiogram of TLC plate: tissue

extracts of wild-type and pad3 leaves fed 14C-

anthranilate 3 and 6 hours after inoculation with

C. carbonum, 10/9/94.....................................65

Figure 10. Autoradiogram of TLC plate: extracts of
inoculum draplets from wild-type leaves fed 358-

cysteine and then inoculated with C. carbonum,
3/29/95..................................................68

Figure 11. Standard curve of camalexin concentration
versus absorbancs at 215 nm, 12/15/95 (A)
and1/21/98 (8)0000000000000000000OOOOOOOOOOOOOOOOO0.0.0092

Figure 12. Time course of camalexin accumulation

in wild-type leaves inoculated with Cochliobolus

carbonum or water (control), 8/25/95-8/27/95 (A) and
10/17/95-10/20/95 (B)...................................102

Figure 13. Effect of concentration of Cochliobolus
carbonum spores on camalexin production, 11/3/95
(see also Table 8)..0....OOOOOOOOOOOOOOOOOOOO0.0.0.0....105

Figure 14. Effect of concentration of C. carbonum
spores on camalexin production, 11/13/95 (see also
Table 9,000.000000000000.00.00.00.00 ..... .00... ......... 107

Figure 15. Time course of camalexin accumulatin in
wild-type (wt) and pad2 leaves inoculated.with

C. carbonum (Cc, top graph), water (ctrl), or

Pseudomonas syringes pv. maculicola strain ES4326

(Psm, bottom graph), 12/31/95-1/3/96....................111

Figure 16. Time course of camalexin accumulation in
‘wild-type and pad? leaves inoculated with

C. carbonum (top graph) or Psm ES4326 (bottom graph),
1/20/96-1/23/96 (see Table 11 for data).................113

Figure 17. Time course of camalexin accumulation

in wild-type and pad? leaves inoculated with

C. carbonum (top graph) or Pam ES4326 (bottom graph),
4/5/96-4/8/96.......................................... 115

Figure 18. Camalexin accumulation in wild-type (wt)

and pad? leaves inoculated (inoc) with Pseudomonas

syringes pv. syringes, (Pss, top graph) or Cochliobolus
carbonum (Cc, bottom graph), 7/23/97-7/26/97 (see

Table 13 for data).................... ....... ....... ..... 119

Figure 19. Time course of camalexin accumulation in
wild-type and pad? leaves inoculated with Pss D20

(top graph) or C. carbonum (bottom graph),

9/21/97-9/24/97 (see Table 14 for data)................. 121

Figure 20. Mass spectrum of indole-3-carboxaldehyde ..... 143

Figure 21. Standard curve of peak area versus
micrograms of indole-B-carboxaldehyde, 12/17/96......... 144

Figure 22. Standard curve of peak area versus
micrograms of camalexin, 1/21/98.........................144

Figure 23A. Time course of accumulation of indole-B-
carboxaldehyde in wild-type and pad2 leaves inoculated

with C. carbonum (inoc) or water (ctrl)
3’23/97’3/24/9700000000eeeoooeeoeeeeooeeeo000000000000000157

Figure 238. Time course of accumulation of camalexin
in wild-type and pad2 leaves inoculated with C.
cmnm (iDOC), 3/23/97—3/24/9700000000.00....0......0.0158

Figure 24. Time course of accumulation of camalexin

(A) and indole-3-carboxaldehyde (B) in wild-type (wt)

and pad2.Arabidopsis leaves inoculated with C. carbonum,
5/31/97‘6/1/9700000009eeooeeeeeeoooeeeeeeoeoeee00.0.000000160

xi

Figure 25. Autoradiogram of TLC plate: tissue

extracts of wild—type Arabidopsis leaves, 0 and 3

hours after inoculation with C. carbonum (i) or

water (ct)......................................... ....... 166

Figure 26. Autoradiogram of TLC plate:extracts
of wild-type Arabidopsis leaves extracted 6 and 9
hours after inoculation with C. carbonum or water.... ..... 167

Figure 27. Autoradiogram of TLC plate: extracts
of wild-type Arabidopsis leaves extracted 6 and 9
hours after inoculation with water or C. carbonum.........168

Figure 28. Autoradiogram of TLC plate; tissue
extracts of wild-type Arabidopsis leaves, 24 hours after
inoculation with water or C. carbonum.....................169

Figure 29. Autoradiogram.of TLC plate: ethyl acetate
extracts of leaves extracted 0,3,6, and 24 hours after
inoculation with C. carbonum or water.....................170

Figure 30. Autoradiogram of TLC plate: camalexin
bands eluted and redeveloped in chloroform-acetic acid
94:6 (VIV)...0.0......OO...00.......OOOOOOOOOOOOOOOOIOO0.0171

Figure 31. Average percent incorporation into

chloroform extracts of leaves inoculated with water

(control) or C. carbonum after being fed labeled
anthranilate (graph A) or labeled anthranilate diluted

‘with cold anthranilate (graph B) or cold indole-B-
carboxaldehyde (graph C)178

Figure 32. Time course of incorporation of labeled
anthranilate into camalexin in leaves fed one of

solutions A, B, or C (see Figure 31) and then

inoculated with water (ctrl) or C. carbonum (inoc)........183

Figure 33. Time course of incorporation of labeled
anthranilate into camalexin in each of 2 replicates of

leaves fed one of solutions A, B, or C and then

inoculated with water (ctrl) or C. carbonum (inoc)........185

Figure 34. Time course of incorporation of 14-C

anthranilate into indole-B-carboxaldehyde in leaves

fed one of the indicated solutions (see Figure 31 for
solution preparation) and then inoculated with water
(control) or C. carbonum..................................188

Figure 35. Incorporation of 14-C anthranilate into
bandz (Rf=0081) over time....0.0......00.0.00000000000000190

Figure 36. Percent incorporation of l4-C anthranilate

xii

into band 3 (Rf=0.7) over time.... ........ ........... ..... 192

Figure 37. Percent incorporation of 14-C anthranilate
into band 3.5 (Rf=0067) over time....OOOOOOIOOOOOOOIOO0.00194

Figure 38. Percent incorporation into band 5 (Rf=0.45)
over timeOOOOOOIOOOI...OOICOOOOOOOOOOI.0... ..... 0.000.000.196

Figure 39. Percent incorporation into band 7 (Rf=0.27)
over time........ ....... 0.0... ....... 00.... ....... 0.. ..... 198

Figure 40. Relationship between log of micrograms of
camalexin and radius of zone of inhibition on TlC plates
bioassayed with Cladbsporium cucumerinum..................221

Figure 41. Light micrographs (400x magnification,

Nomarski optics) of ecotype Kas-l (low degree of

resistance) 19 hours (right) and 48 hours (left) after
inoculation...............................................227

Figure 42. Light micrographs (400x magnification,
Nomarski optics) of ecotypes Turk Lake and RLd (high
degree of resistance).....................................229

Figure 43a. Effects of Cochliobolus carbonum (neutral
inducer) on camalexin production..........................230

Figure 43b. Same TLC plate as in Figure 43a after a
bioassay with Cladosporium cucumerinum....................231

Figure 44. Effects of Alternaria brassicicola on
camleXi-n prwuctionOOOOOOOOOOOOOO0..0.0.00.00000000000000232

xiii

Chapter 1 . General Introduction

A pathogen’s entry into plant tissue has two possible
outcomes. One is a compatible interaction, in which
colonization is successful and the plant’s susceptibility is
manifested.macroscopically a few days later by pathogen
development-—with or without widespread necrosis--and,
eventually, by death of the plant. The other possible
outcome is an incompatible interaction, in which colonization
is unsuccessful and the plant’s resistance is manifested in
about 24 hours by a hypersensitive response, or HR, of which
the macroscopic evidence is small necrotic lesions at the
infection site (Hammond—Kosack and Jones, 1996; O’Connell et
al., 1985). Compatible and incompatible interactions are
accompanied by biochemical changes. These changes can
include the accumulation of salicylic acid, other phenolic
compounds, and pathogenesis—related (PR) proteins; production
of active oxygen species; calcium influx; and lignin
deposition (Bell, 1981; Ebel, 1986; Hammond-Kosack and Jones,
1996; Nicholson and Hammerschmidt, 1992; Nfirnberger et al.,

v 1994; Paxton and Groth, 1994).

Another change that frequently occurs in infected plants
is the production of phytoalexins, commonly defined as “low-
molecular weight, antimicrobial compounds synthesized by and
accumulating in plants following infection by microorganisms“
(Paxton, 1981). The concept of phytoalexins (from the Greek

,phyto, plant; + alexein, to push away) was first proposed by

Muller and Borger in 1940 to explain their finding that if a
potato tuber slice were inoculated with an incompatible race
of Rhytophthora infestans (the causal agent of potato
blight), a compatible race subsequently inoculated would not
grow. The incompatible race had elicited the production of
an antifungal compound (Deverall, 1982; Harborne, 1988).
Almost 20 years after the postulation of their existence, the
phytoalexin pisatin was isolated and crystallized from pea
pods (Cruickshank and Perrin, 1960). Since then, over 350
phytoalexins have been found in plants from about 30
different plant families, including the potato phytoalexins
(rishitin and lubimin) whose existence had been postulated by

Muller and Borger (Harborne, 1988; Kuc, 1995).

General Properties of Phytoalexins.

The chemical structures of phytoalexins vary and include
stilbenes, isoflavonoids, and sesquiterpenes. In general,
plants of a given family produce phytoalexins of the same
basic structure(s): members of the Solanaceae produce
sesquiterpenoid phytoalexins, while members of the Fabaceae
(Leguminosae) produce isoflavonoid or pterocarpan
phytoalexins (Ebel, 1986; Harborne, 1988; Kuc, 1995).

The biosynthetic pathways of phytoalexins originate from
the biosynthetic pathways of primary metabolites such as
carotenoids and amino acids (Kuc, 1995). The sesquiterpenoid
phytoalexins of the Solanaceae, for example, are synthesized

via the isoprenoid pathway responsible for sterol

biosynthesis. The phytoalexin pathway branches off from the
sterol pathway at farnesyl pyrophosphate, from which either
squalene (a sterol precursor) or sesquiterpenoid phytoalexin
precursors are made (Kuc, 1995).

This branching-off from primary biosynthetic pathways
occurs after infection. Phytoalexins are usually absent in
healthy plants or present only in trace amounts (Harborne,
1988). The presence of trace amounts in some uninfected
plants may indicate that the plants have been subjected to
some sort of stress, which is almost impossible for plants to
avoid in nature (Harborne, 1988).

As their presence in uninfected plants may indicate,
phytoalexins appear in some respects to be a fairly non-
specific response to stress. In addition to pathogens,
abiotic stress--heavy metals, cold stress, or ultraviolet
light--can elicit phytoalexin production and fall into the
category of “abiotic elicitors" (Ebel, 1986; Kuc, 1995).
Among pathogens (biotic elicitors), it is typical for many
kinds of fungi, bacteria, or viruses to be capable of
inducing the biosynthesis of a particular phytoalexin (Kuc,
1995). The wide range of elicitors suggests either that
phytoalexins are a very nonspecific response to stress, or
that these elicitors have a common mode of action. According
to Ebel (1986), it has been suggested that the different
elicitors may all cause cell death and, consequently,
responses similar to those following cell death in an HR. It

is also possible that phytoalexin synthesis is a general

response to alterations in metabolism after infection or
abiotic stress (Kuc, 1995). However, phytoalexin production
is not a completely non-specific response because not all
stresses induce phytoalexin production. Wounding of potato
tubers does not induce synthesis of rishitin or lubimin (Kuc,
1995).

Due to efforts to find a specific mechanism of action, a
major focus in the realm of biotic elicitors has been on
chemical constituents of pathogens. Efforts to characterize
the active components of elicitors have led to the
identification of oligosaccharides and peptides from fungi
(Ebel, 1986; Albersheim and valenti, 1978; Nurnberger et al.,
1994) and also from plants (Ebel, 1986).

Just as they can be induced by many different pathogens,
so can phytoalexins be effective against many pathogens. In
general, the antimicrobial spectrum of any given phytoalexin
is very broad. These compounds are usually active against
many kinds of fungi, bacteria, and viruses (Harborne, 1988).
However, they do not affect all pathogens, and those affected
differ in sensitivity (Harborne, 1988). Variations in
sensitivity were illustrated in a study by Cruickshank (1962)
on the antimicrobial activity of pisatin against 50 fungal
and 24 bacterial species. About half of the bacterial
species were highly sensitive to pisatin, while the others
were unaffected. Among the fungi tested, most were sensitive
to pisatin at fairly low concentrations, but a few

(Aspergillus nidulans and species of Pellicularia and

Fusarium) had intermediate sensitivity, and a few
(Mycosphaerellaipinodes, Pellicularia filamentosa, Fusarium
solani, and Ascochyta pinodella and A. pisi) were not
sensitive at all.

Sensitivity to phytoalexins in vivo is manifested by a
reduction in fungal spore germination, bacterial colony-
forming units, or virus titer and spread (Smith, 1982).
Sensitivity to phytoalexins in vitro typically is measured by
reduction of bacterial or fungal growth in liquid media
containing the phytoalexin, reduction of fungal mycelial
growth on agar impregnated with the phytoalexin, decrease in
spore germination, or decrease in germ-tube growth (Smith,
1982). The study by Cruickshank (1961), for example,
evaluated bacterial sensitivity on the basis of growth
inhibition in liquid cultures containing pisatin. Fungal
sensitivity was determined by reduction of mycelial growth on
agar containing pisatin. Because phytoalexins are small
organic molecules and usually are easily separated by thin-
layer chromatography (TLC), another common bioassay is to
spray a thin—layer chromatogram with fungal spores or
bacterial cells in a nutrient broth, incubate the
chromatogram in a humid environment to allow the spores/cells
to grow, and look for inhibition of growth at the site of the
putative phytoalexin (Harborne, 1988; Homan and Fuchs, 1970;
Smith, 1982).

Some phytoalexins are toxic to pathogens, while others

are inhibitory. Pisatin, for example, is fungistatic but not

fungitoxic: spores of Sclerotinia fructicola did not
germinate in the presence of 0.28 mM pisatin, but after being
thoroughly washed of the pisatin, some spores were able to
germinate (Cruickshank and Perrin, 1960).

Another characteristic of phytoalexins is the
localization of their production at the site of infection or
stress. According to Deverall (1982), Muller and Borger
found that if half a potato tuber slice were inoculated with
an incompatible race of P. infestans, a subsequent
inoculation of the entire slice with a compatible race would
lead to growth of the latter only on the half that had not
been previously inoculated. The unknown phytoalexin clearly
did not diffuse over long distances, although it did diffuse
into neighboring cells: if a thin layer of the tuber slice
were removed and the area below inoculated with a compatible
race, no growth occurred in the area just below the site of

inoculation with the incompatible race.

The Role of Phytoalexins in Disease Resistance.

The antimicrobial nature, localization, and de novo
synthesis of phytoalexins upon infection suggest that these
compounds may play a role in disease resistance and have led
to considerable interest in their potential use in
agriculture. Currently, with farmers trying to reduce
pesticide use and consumers expressing concern about
pesticides in their food, there is a growing interest in

protecting crops by exploiting the natural defenses of'

plants. If phytoalexins do help plants to resist various
diseases and stresses, perhaps some crops could become more
resistant to certain diseases if transformed with the genes
for biosynthesis of a phytoalexin. Also, as pathogens do
differ in their sensitivities to some phytoalexins
(Cruickshank, 1961), some of these compounds may prove to be
good selective fungicides, or their molecular structures may
provide clues as to the kinds of molecules that would be
effective against certain pathogens (Pedras et al., 1997).
However, because some phytoalexins are phytotoxic as well as
fungitoxic (Glazener and van Etten, 1978; Hargreaves, 1980),
genetic engineering of plants may be a safer way to use
phytoalexins in agriculture.

A transformation of tobacco with the genes required for
production of the grape phytoalexin resveratrol (Hain et al.,
1993) suggests that introducing genes for phytoalexin
biosynthesis into plants is feasible. Transformed plants
were more resistant to the fungus Botrytis cinerea than wild-
type plants or transformants with a plasmid lacking the
resveratrol biosynthesis genes (Hain et al, 1993). However,
one success story (of questionable success, since the
severity of disease in nontransformed plants was low enough
that the large amount of disease reduction in transformed
plants would have been unnoticeable in the field) is hardly
enough to convince the general public to purchase genetically
engineered produce, or to convince grant-giving agencies to

fund the cloning of phytoalexin biosynthetic genes and

transformation of vegetables. More evidence of the
importance of phytoalexins in disease resistance is
necessary, and so far, the evidence has been inconclusive.

If phytoalexins are important in disease resistance, a
pathogen's ability to colonize a plant should depend partly
on its ability to detoxify or metabolize the plant's
phytoalexin(s). A compatible pathogen should be able to
detoxify the phytoalexin(s), while an incompatible pathogen
should not. This colonization/detoxification relationship
has been demonstrated in at least one case: the maize
pathogen Cbchliobolus heterostrophus, which normally cannot
infect pea, was able to do so after it had been transformed
with a gene for pisatin demethylase (pdm), which allows
detoxification of pisatin (Schafer et al., 1989). Isolates
of the pea pathogen Nectria haematococca , after being
transformed with the pdm gene, also became more virulent on
pea (Ciufetti and VanEtten, 1996). However, additional work
(Wasmann and VanEtten, 1996) demonstrated that after
disruption of the inserted gene, the gene-disruption mutants
were still more virulent than the original pdm-deficient
isolates, suggesting that successful infection depended on
something other than pisatin detoxification.

Timing of synthesis is another indicator of whether
phytoalexins have a role in disease resistance, since they
should be produced shortly after infection in order to stop
the spread of a pathogen. In the resveratrol-transformed

tobacco mentioned previously, the leaves with the fewest

lesions, besides having higher levels of resveratrol than
leaves with more lesions, accumulated the phytoalexin more
rapidly than the more—diseased leaves (Hain et al., 1993).
In soybean roots infected with Phytophthore megasperma f.sp.
glycinea, the phytoalexin glyceollin accumulated to lower
concentrations in response to an incompatible race than in
response to a compatible one, but it accumulated more rapidly
(Hahn et al., 1985). In sorghum resistant to Cblletotrichum
graminicola, phytoalexin accumulation begins a few hours
after fungal penetration, suggesting that the sorghum
phytoalexin does play a role in disease resistance (Snyder
and Nicholson, 1990). The timing of phytoalexin production
in some other plants, however, is too late to suggest a
critical role in disease resistance. In parsley cell
suspension cultures, mRNA for parsley phytoalexins was first
detected 12 to 18 hours after inoculation, well after
initiation of mRNA transcripts for other defense responses
such as the pathogenesis—related protein PR-l (Schmelzer et
al., 1989). The variation in timing of phytoalexin
production suggests that the relative importance of a
phytoalexin in disease resistance varies with the

phytoalexin, and perhaps with the plant as well.

Phytoalexins of the Brassicaceae: General Properties
and Role in Disease Resistance.
This thesis describes work on camalexin, a phytoalexin

of the Brassicaceae (Cruciferae). The cruciferous

phytoalexins are relatively new to the world of plant
research. The first report of these compounds was in 1986 by
Takasugi and co-workers, who isolated four compounds
(methoxybrassinin, brassinin, cyclobrassinin, and
methoxybrassitin, shown in Figure 1) from Brassica

.pekinensis, or Chinese cabbage (Gross, 1993).

NH
s/\SCH3

R=H, brass inin; R=OCH3 ,methoxybrassinin

NH
©‘I S SCI-l3 ©‘I C/\SCH3

«1'13 “3‘13
cyclobrassinin methoxybrassitin

 

Figure 1. A few cruciferous phytoalexins.

These compounds had typical characteristics of phytoalexins,
as they were induced by abiotic and biotic elicitors
(ultraviolet light and the bacterial pathogen Pseudomonas
cichorii), and they had a broad antimicrobial spectrum, being
antifungal to over 31 plant pathogenic fungi (Takasugi et
al., 1988). As shown in Figure 1, they had similar
structures consisting of an indole ring with a sulfur-
containing moiety. These were the first sulfur-containing
phytoalexins isolated. About 24 phytoalexins from the
Brassicaceae have now been isolated from a number of plants,
including Raphanus sativus var. hortensis (Japanese radish),

Brassica campestris ssp. rapa (turnip), and Brassica napus

10

(canola) (Gross, 1993; Pedras et al., 1997). All consist of
an indole ring with a sulfur-containing substituent (Gross,
1993; Pedras et al., 1997).

Within the Brassicaceae, no phytoalexins have
demonstrated a clear role in disease resistance. Pedras et
al. (1997) found a correlation between the virulence of
Lsptosphaeria maculans (the causal agent of blackleg disease
of crucifers) and the extent to which it could metabolize
brassinin. Since brassinin is a precursor of spirobrassinin
and cyclobrassinin, the ability to detoxify brassinin could
be a strategy to stop phytoalexin biosynthesis before
encountering a triple phytoalexin threat (Pedras et al.,
1997).

Studies on the role of brassilexin (Figure 2) in disease
resistance revealed that when Brassica species were
inoculated with L. maculans, the resistant species Brassica
juncea, Brassica carinata, and Brassica nigra accumulated

more brassilexin than did the susceptible B. napus (Rouxel et

Cram

Figure 2. Brassilexin

al., 1990).

 

In response to cupric chloride treatment, B. juncea
accumulated more brassilexin than B. napus, and the onset of
accumulation in B. juncea (6 hours) was earlier than in B.

napus (18 hours) (Rouxel et al., 1989). Assuming similar

11

kinetics of brassilexin accumulation in response to fungal
inoculation, the more rapid accumulation of brassilexin in
the resistant species suggests that this phytoalexin could
have a role in disease resistance and that resistance depends
partly on rapid accumulation of brassilexin. When resistant
and susceptible species were crossed, resistance in the F2
generation correlated well with high brassilexin accumulation
after cupric chloride treatment (Rouxel et al., 1990).
Further studies (Rouxel et al., 1991) demonstrated, however,
that some resistant species accumulated very little
brassilexin, indicating that brassilexin was not the key

determinant in disease resistance.

General Properties of Camalexin.

Camalexin, which has the structure 3-thiazol—2'-yl-
indole (Figure 3), was first detected by Conn et a1. (1988)
in leaves of camelina sative inoculated with the fungus
Alternaria brassicae. A further study by Jejelowo et
al.(1991) of the isolated compound, whose structure was
unknown at the time, demonstrated that it had the typical
properties of a phytoalexin: it could be extracted with
organic solvents (methanol and chloroform) and separated by
TLC; it was produced following inoculation of leaves with A.
bressicae; and production of the antimicrobial compound was
restricted to the area underneath and immediately surrounding
the droplets of inoculum. On leaves, antimicrobial activity

was demonstrated by a decrease in fungal germination and

12

 

 

germ-tube growth. Bioassays of thin-layer chromatograms
revealed zones of inhibition at bands found later to

correspond to camalexin and methoxycamalexin.

OE 9

Figure 3. Camalexin

The structure of camalexin was determined by Browne et
al. (1991). The structure of the phytoalexin produced by
Arabidopsis thaliana was subsequently determined by Tsuji et
al. (1992) to be the same compound. Recently, camalexin has
been isolated from two other cruciferous plants, capsella
bursafipastoris (shepherd's purse; Jimenez et al., 1997), and
Arabis lyrata (M. Zook, in press). It is easily recognized
on thin-layer chromatograms by its purple fluorescence under

long-wave and short—wave ultraviolet (UV) light at an Rf of

about 0.6 in chloroform-methanol (9:1 or 19:1, v/v).

Like other phytoalexins, camalexin is induced by many
microorganisms and abiotic elicitors. The microorganisms
that induce camalexin production include the fungi Alternaria
brassicae (Conn et al., 1988; Jejelowo et al., 1991),
Rhizoctonia solani (Conn et al., 1994), and Cbchliobolus
carbonum (Glazebrook et al., 1997); the bacteria xanthomonas
campestris pv. campestris (Zhao and Last, 1996; Zhou et al.,

1998), Pseudomonas syringes pv. syringes (Tsuji et al.,

13

 

 

1992), P.syringae pv. maculicola (Glazebrook and Ausubel,
1994; Glazebrook et al., 1997; Zhao and Last, 1996; Zhou et
al., 1998); and P.3yringae pv. tomato (Glazebrook and
Ausubel, 1994; Zhou et al., 1998). Among viruses, turnip
crinkle virus (Dempsey, 1996; Dempsey et al., 1997) and
cauliflower mosaic virus (Callaway et al., 1996) elicit
camalexin production. Camalexin is also induced by silver
nitrate (Tsuji et al., 1992; Zhao and Last, 1996; Zhou et
al., 1998), starvation for certain amino acids (methionine or

branched—chain amino acids), the herbicide acifluorfen, and

a-aminobutyric acid, a putative chemical inducer of

resistance (Zhao et al., 1998). Little is known on the
mechanisms of camalexin induction. Despite the wide range of
biotic and abiotic elicitors, camalexin biosynthesis is not a
general response to all forms of stress, since heat shock
does not induce camalexin (Zhao et al., 1998). Salicylic
acid (SA), a key component in the plant signal transduction
pathway, is required, because SA-deficient Arabidopsis plants
produced less camalexin than wild-type (Zhao and Last, 1996).
The different elicitors may somehow stimulate SA production,
or activate a signal that in turn activates SA. Another
possibility, since one inducer, acifluorfen, generates free
radicals, is that oxidative stress is a key player in
camalexin elicitation (Zhao et al., 1998). Active oxygen
species (peroxides, hydroxy radicals, or superoxide anions)

frequently are produced in response to infection

14

 

 

(Hammond—Kosack and Jones, 1996; Zhao et al., 1998), and so
it is possible that the different elicitors all act by
stimulating the production of such molecules, which in turn
induce camalexin biosynthesis. Amino acid deprivation may
induce camalexin for the same reason, since such treatment
may cause chloroplast damage, with consequent oxidative
stress due to the lack of molecules to absorb and dissipate
light energy (Zhao st al., 1998).

Camalexin has a broad but not entirely non-specific
antimicrobial spectrum. Fungi seem to be more sensitive than
bacteria to camalexin: in liquid cultures, the threshold
concentration toxic to Pseudomonas syringes pv. phassolicola,
P. syringes pv.maculicola, and xanthomonas campsstris pv.
campsstris was 250-500 ug/ml, while the threshold
concentration for Fusarium oxysporum and saccharomyces
csrsvisias was 20-50 ug/ml (Rogers et al., 1996). The lower
inhibitory concentrations for fungi corroborate the findings
of Jejelowo st a1. (1991) that the hyphal tips of Alternaria
brassicae swelled and burst in aqueous solutions containing
20 ug/ml or more of camalexin. Similar differences in fungal
and bacterial sensitivities have been found with TLC plate
bioassays: the minimum amount of camalexin to inhibit P.
syringes pv. syringes was 4 times the minimum amount required
to inhibit the fungus Cladosporium cucumsrinum (Tsuji et
al., 1992). The antimicrobial properties of camalexin seem
to be due to membrane disruption, since addition of

inhibitory concentrations of camalexin to bacterial cultures

15

 

 

caused electrolyte leakage (Rogers et al., 1996). Recently,
camalexin was found to be highly toxic to a line of breast
cancer cells (Moody et al. 1997); its toxicity in this latter

case may be due again to membrane disruption.

The role of camalexin in disease resistance.

As with the other cruciferous phytoalexins, the role of
camalexin in disease resistance is unclear. Many studies
have demonstrated a relationship between camalexin
accumulation and resistance. Arabidopsis plants transformed
with a fusion product of the reporter gene B-glucuronidase
(GUS) and the promoter for the tobacco Tntl retrotransposon,
whose expression has been correlated with disease resistance
in tobacco, expressed the GUS gene (an indication of Tntl
expression) in response to the abiotic elicitor cupric
chloride, and these plants also produced camalexin under
these conditions (Mhiri st al., 1997). Camalexin production
was correlated.with resistance of Arabidopsis to turnip
crinkle virus, or TCV (Dempsey, 1996; Dempsey et al., 1997).
In genetic crosses with a resistant line of the Dijon ecotype
of Arabidopsis, camalexin production consistently segregated
with the HRT (hypersensitive response to TCV) locus, which is
required for TCV resistance, and after inoculation with TCV,
resistant lines of the Dijon ecotype produced significantly
more camalexin than susceptible lines (Dempsey, 1996; Dempsey
et al., 1997). Inoculation of the Columbia ecotype of

Arabidopsis with cauliflower mosaic virus (CaMV) or cucumber

16

mosaic virus (CMV), both compatible pathogens, did not lead
to camalexin production (Zhao and Last, 1996). However, the
En-2 ecotype of Arabidopsis, which is resistant to CaMV, did
produce camalexin following inoculation (Callaway et al.,
1996).

Although the En—2 ecotype produced camalexin without
displaying a visible hypersensitive response (HR), most
studies have demonstrated a strong correlation between
camalexin production and an HR. Since the HR is a sign of an
incompatible interaction, the observation of this
relationship supports a role for camalexin in disease
resistance. The studies by Dempsey et a1. (1996, 1997) just
described, indicated a link between camalexin production,
resistance, and the ability to produce an HR. The
accumulation of camalexin was correlated with an HR in
response to infection by the incompatible pathogen
Pseudomonas syringes pv. syringes (Pss) (Tsuji st al., 1992).
Bacterial growth reached a maximum before camalexin levels
reached a maximum, and bacterial growth then decreased while
camalexin levels continued to increase. Camalexin was not
produced upon inoculation with mutants of Pss, which were
nonpathogenic due to loss of a hrp (hypersensitive response
and pathogenicity) gene, which is required for bacteria to
cause an HR and disease on plants.(Tsuji et al., 1992).

Further strengthening the correlation between camalexin
production and the HR is the finding that regulation of

camalexin production appears to be controlled by the ACD2

l7

 

 

(accelerated cell death) locus, which controls the onset of
the HR and the size of HR-type lesions (Greenbsrg st al.,
1994). After bacterial inoculation or mechanical wounding
(which usually does not induce high production of camalexin),
acd2 mutants had camalexin concentrations identical to those
of bacterial—inoculated wild—type plants, as well as high
concentrations of salicylic acid and mRNA’s for a number of
defense-associated genes (Greenbsrg st al., 1994).

Studies with the bacterial pathogen xanthomonas
campsstris pv. campsstris (XCC), to which Arabidopsis is
resistant, further support the connection between camalexin
production and an HR. On Arabidopsis, XCC does not cause an
HR and can grow to relatively high titers in plants;

resistance is judged by the absence of symptoms (Tsuji et
al., 1991). Inoculation with XCc at 108 colony—forming units

per milliliter (cfu/ml)--a much higher inoculum concentration

than would be found in nature--elicits no camalexin (Tsuji et
al., 1991), although inoculation at 109 cfu/ml does (Zhao and

Last, 1996). These results suggest that camalexin is not
always produced in an encounter with a pathogen to which
Arabidopsis is resistant. Without a hypersensitive response
to the pathogen, camalexin is not produced, unless other
stressing conditions (such as unusually high inoculum levels)
are involved. The results both strengthen the camalexin/HR
correlation and weaken the case for a primary role for
camalexin in disease resistance.

As with most phytoalexins (Kuc, 1995), camalexin

18

accumulation is not associated exclusively with incompatible
interactions. Inoculation with Pseudomonas syringes pv.
maculicola, a compatible pathogen, leads to high camalexin
production (Glazebrook and Ausubel, 1994; Glazebrook et al.,
1997; Rogers et al., 1996; Zhao and Last, 1996; Zhou et al.,
1998).

It is difficult to determine the relative contribution
of camalexin to inhibiting pathogen spread in plants. Since
Tsuji st al. (1992) found that Pss growth decreased before
camalexin levels peaked, it is reasonable to assume that the
bacterial growth was inhibited in part by the increasing
concentration of camalexin. However, the fact that bacterial
growth did not increase as camalexin concentration decreased
indicates that camalexin is not the sole factor controlling
Pss spread in plants. It is not impossible that degradation
products of camalexin are antibacterial and that the
metabolism of camalexin contributes to the ability of
Arabidopsis to prevent further infection.

If camalexin contributes significantly to inhibiting
pathogen spread in a plant, the pathogens capable of
detoxifying camalexin should spread and grow more readily
than the pathogens that cannot detoxify camalexin, just as N:
hasmatococca transformed with a pisatin demethylase gene
caused limited infection on pea (Ciufetti and VanEtten,
1996). In the case of camalexin, however, the opposite has
been demonstrated in a couple of cases. Mycelial growth of

Lsptosphasria maculans was not inhibited on agar containing

19

camalexin, and the camalexin was not metabolized, suggesting
that it is so nontoxic toward L. maculans that the fungus
does not need camalexin detoxification mechanisms to be a
successful pathogen. The fungus Rhizoctonia solani, although
inhibited in vitro by 50 ug/ml camalexin and capable of
metabolizing camalexin to 3 non-inhibitory compounds (Pedras
and Khan, 1997), was a poor colonizsr of Camslina sative
roots in vivo (Conn et al., 1994). The fact that camalexin,
along with three other antimicrobial compounds, was isolated
from the roots in that study, indicates that R. solani did
not metabolize significant amounts of camalexin. Perhaps
other factors--PR proteins, accumulation of phenolics, or the
other antimicrobial compounds--stopped the spread of R.
solani before detoxification of camalexin became a
requirement for being able to infect.

Despite the fact that camalexin is not required for
resistance to some pathogens and that its production is not
sufficient to stop the spread of other pathogens, some of the
studies just mentioned suggest that it may be part of the
plant defense arsenal. Why is it produced in some
incompatible interactions, along with all of the other
responses of the HR? Do its antimicrobial properties aid in

inhibiting pathogen growth?

20

 

 

The Use of Camalexin-Deficient Mutants to Study the
Role of Camalexin in Disease Resistance.

A classic way to determine the function of a compound in
an organism is to find mutants deficient in that compound and
compare their phenotypes to those of wild-type organisms.

The isolation of camalexin-deficient mutants of Arabidopsis
by Glazebrook and Ausubel (1994) made this type of analysis
possible for camalexin in Arabidopsis. To date, five of
these pad (phytoalexin-deficient) mutants, each representing
a mutation in a different gene, have been found (Glazebrook
and Ausubel, 1994; Glazebrook et al., 1997).

One aspect of disease resistance examined with these
mutants was the effect of camalexin on gene-for-gene
resistance to bacterial pathogens, which requires recognition
of an avirulence (avr) gene due to a corresponding resistance
gene in the plant. The plant response usually involves an
HR, biochemical changes typical of incompatible interactions,
and restriction of pathogen growth (Whalen et al., 1988).

The pad mutants demonstrated that accumulating little or no
camalexin does not affect gene—for—gene resistance to
avirulent bacteria. In all five mutants, after inoculation
‘with the compatible pathogen Pseudomonas syringes pv.
.maculicola or with an isogenic counterpart bearing an avr
gene, the difference between final population densities of
virulent and avirulent pathogens was about the same as in
'wild-type plants (Glazebrook and Ausubel, 1994; Glazebrook et

al., 1997). Camalexin accumulation, therefore, does not

21

affect the ability of resistance genes to function. These
results help to explain why inoculation with ch, for which
resistance is controlled by a single gene (Tsuji et al.,
1991), elicits no camalexin production in wild—type plants.
Resistance genes and camalexin can function independently of
one another.

.Although the pad mutants still responded to avr—gene—
bearing bacteria by restricting growth in plants, three of
the mutants—spadl, pad2, and pad4—-allowsd more growth of the
compatible bacteria than did the wild-type plants, suggesting
that camalexin does somehow help to control the extent to
which a compatible pathogen can grow in plants (Glazebrook et
al., 1997). Camalexin may play a more important role in
resistance to eukaryotic than to prokaryotic pathogens, since
four of the pad mutants were more susceptible than wild-type
plants to the oomycete Psronospore paresitica (Glazebrook st

al., 1997).

Biosynthesis of Camalexin and Other Cruciferous
Phytoalexins.

Another way to determine the role of a compound in an
organism is to manipulate the genes for its biosynthesis and
then to observe the effect of inactivating or overexpressing
key biosynthetic enzymes or of introducing those enzymes into
another organism. This approach, however, requires knowing
the biosynthetic pathway of the compound. Since none of the

biosynthetic pathways of cruciferous phytoalexins were known

22

until 1994 (Monde et al., 1994; Gross et al., 1994), and many
have not yet been elucidated, camalexin was a potential model
system for the biosynthesis of phytoalexins in the
Brassicaceae. These would all be expected to have similar
biosynthetic origins, due to their similarity of structure.

Determining the biosynthesis of camalexin could have
additional benefits for agriculture. If camalexin could aid
in disease resistance, introducing the genes for its
biosynthesis into other plants--probably other plants in the
same family, of which there are many economically important
members-—might make the plants more resistant to certain
diseases. Also, if camalexin is truly an anticancer compound
(Moody et al., 1997), it might be possible to clone the genes
for its biosynthesis, overexpress them in plants, and thus
facilitate isolating large quantities of this compound for
medicinal purposes. These possibilities are a bit far-
fetched, but understanding basic mechanisms in biology can
lead to many benefits besides intellectual satisfaction, and
the results are sometimes far more astonishing than imagined,
as demonstrated by the development of Agrobacterium-based
genetic engineering from studies of crown galls (Chilton,
1983).

Mutants, besides providing phenotypic evidence of the
role of the compound that they lack, can be useful in
determining biosynthetic pathways. The inability to
synthesize a compound, if due to the lack of an enzyme

catalyzing a certain step of the pathway, can lead to the

23

accumulation of intermediates prior to that step (Swain;
1965). With a large number of mutants, each representing a
mutation in a different gene, all steps of the pathway could
conceivably be determined, if different intermediates were
found to accumulate in each mutant. The pad mutants
initially seemed like potential sources of such information,
and work with some other mutants (Tsuji st al., 1993; Zhao
and Last, 1996) has helped to understand the biosynthesis of
camalexin (see below).

What, then, is known about the biosynthetic pathways of
camalexin or the other cruciferous phytoalexins? The indole
ring suggests that these compounds originate from tryptophan.
In fact, the radish phytoalexins brassinin, spirobrassinin,
and cyclobrassinin do seem to originate from tryptophan
(Monde et al., 1994). The more immediate precursor is
glucobrassicin, a tryptophan-derived glucosinolate (Rausch et
al., 1983). Spirobrassinin in kohlrabi (Brassica oleracsa
var. gongyiodss) also was found to originate from tryptophan
and methionine (Gross et al., 1994). The phytoalexins
cyclobrassinone and 1-methoxyspirobrassinin, also isolated
from kohlrabi, seem to originate from tryptophan and
methionine as well (Gross et al., 1994). In addition, two
other minor phytoalexins of kohlrabi, methoxybrassitin and
methoxybrassinin, incorporated labeled tryptophan and
methionine (Gross et al., 1994).

Camalexin, however, appears to be synthesized by a

somewhat different pathway. Because anthranilate (the

24

product of the first committed step toward tryptophan
biosynthesis in the shikimic acid pathway) was incorporated
more efficiently into camalexin than tryptophan, it appeared
that the camalexin biosynthetic pathway branched off from an
intermediate between anthranilate and tryptophan (Tsuji et
al., 1993). These results were further supported by the
finding that of three tryptophan-deficient (trp) mutants
tested, the only one that did not synthesize wild-type levels
of camalexin was a mutant deficient in anthranilate synthase

(trpl—lOO); those deficient in tryptophan synthase (trp3-l,

deficient in tryptophan synthase a; and trpZ-l, deficient in

tryptophan synthase 8) were not affected (Tsuji et al.,
1993). The absence of indole glucosinolates in etiolated
seedlings of Camslina sative (Schraudolf, 1968) also suggests
that camalexin is not synthesized along the same pathway as
the radish phytoalexins. It may be that indole
glucosinolates were not detected in C. sative in Schraudolf’s
(1968) study because of the developmental stage of the plant,
or because they were being rapidly converted into something
else. However, it may be that camalexin biosynthesis is
partly the result of an inability of C. sative to make other
phytoalexins, due to the lack of indole glucosinolates.

It seemed possible that the indole ring of camalexin
could be formed from a pathway branching off from a
tryptophan pathway intermediate such as indole-3-glycerol

phosphate (Tsuji et al., 1993), as shown in Figure 4.

25

.masznucm owumnucxmown condoummuu
one saxoamemo on» no ownmcowumamu Hmowumnuooar .e Tasman
ousaoswmoosam oaooca

 

so .swowmmmunoosam sonooudzuu
A
_ is: bl AHV
mOm II R {/oaoocfl
\ 2 _ ©
been owahxonuoormlwaoocw .moum q
mohcmoaoROQMMOImthoosa .mum .\R\\\

ensconced

HousewamrmthOCCH

 

\ 4a,

cwxwaoeoo Aswxsm .«dH.
been owuoomtmtoaoocw .\\

 

 

26

This hypothesis was quite reasonable in light of the findings
by Wright et al. (1991) that indole-3-acetic acid (IAA) in
maize is not synthesized from tryptophan. In that study,
tryptophan-deficient mutants, although they produced much
less tryptophan than wild-type seedlings, produced much more
IAA, and they contained higher concentrations of anthranilate
and indole than did wild-type plants. These results suggest
that the IAA biosynthetic pathway, at least in maize,
branches away from the tryptophan biosynthetic pathway at or
before indole. The possible independence of this pathway
from the tryptophan pathway suggested a few possibilities for
tryptophan-independent intermediates of camalexin. A number
of indole derivatives have been found in plants, including
indole-B-carboxaldehyde and indole-B-carboxylic acid (Figure
4) (Muller, 1961; Melchior, 1957).

Browne et al. (1991) proposed that camalexin could be
produced by a condensation reaction between indole-3-
carboxaldehyde and cysteine. The former has been isolated in
fairly high quantities in some plants, including cabbage,
another crucifer (Devys and Barbier, 1991; Jones and Taylor,
1957). It could be formed in several ways. Indole-3-
carboxaldehyde can be formed by oxidation of IAA (Stutz,
1958), photolysis of IAA (van Denffer and Fischer, 1952;
.Melchior, 1957; Meyer and Pohl, 1956; Ray and Curry, 1958),
and also from photolysis or oxidation of tryptophan (Fischer,
1954; Melchior, 1957; Muller, 1961).It could also be formed

from indole glucosinolates (Devys and Barbier, 1991).

27

Although C. sative does not produce indole glucosinolates,
Arabidopsis does, and it may be that the camalexin pathway in
Arabidopsis differs from the pathway in C. sative, or that
more than one pathway to camalexin exists. The latter
possibility is described by Bu'lock's (1965) model of the
“metabolic grid,” which depicts multiple pathways to the same
compound.

Given the number of sources for the biogenesis of
indole-3-carboxaldehyde, it seemed that a possible
biosynthetic pathway for camalexin could involve the
intermediates shown in Figure 5. Nucleophilic attack of
indole-B-carboxaldehyde by cysteine would lead to an indole-
cystsine adduct. Under low pH conditions, the oxygen bonded
to the carbonyl carbon could be removed by protonation and
subsequent loss of a water molecule, which would allow the
carbocation to bind to the nitrogen atom in cystsine--thus
effecting closure of a precursor to the thiazole ring. This
indole—3-thiazoline carboxylic acid could be oxidized to
indole—B-thiazolidine carboxylic acid, and oxidized and
decarboxylated to produce camalexin (Zook and Hammerschmidt,
1997). Decarboxylation of the cysteine carboxyl group could
also occur prior to the formation of the double bonds in the

thiazole ring (J. Kagan, pers. comm.).

Project Description.
This project began as an effort to find intermediates in

camalexin biosynthesis by characterizing the pad mutants.

28

    

indole-B- cysteine
carboxaldehyde

indole—cysteine adduct

-H20 ‘

 

indole-3-
thiazolins

 

 

carjpxylic acid

CCK)

 

O. /
| -C02
}

 

indole-3- resonance—
thiazolidine stabilized
carboxylic acid ‘ intirmediate

1’ \

    

camalexin

Figure 5. Two possible routes to camalexin.

29

 

Radiolabeling and fluorimetry studies of leaves elicited with
the ascomycete Cochliobolus carbonum revealed no obvious
intermediates accumulating in the pad’mutants; however, they
did reveal that some of these supposedly camalexin-deficient
mutants produced fairly high amounts of camalexin. As these
mutants had been identified by inoculation with the bacterial
pathogen Pseudomonas syringes pv. maculicola, these results
suggested that camalexin accumulation in Arabidopsis might
vary with the infecting microorganism. Consequently, the
project turned into an effort to compare patterns of
camalexin accumulation in wild-type and pad? plants
inoculated with Cochliobolus carbonum or the bacterial
pathogens Pseudomonas syringes pv. maculicola or P. syringes
pv. syringes. In the process of isolating camalexin (by
thin-layer chromatography) from samples used for these
studies, additional inducible compounds were seen on thin-
layer chromatograms. Attempts to Characterize these
compounds led to the identification of indole-3—
carboxaldehyde in fungal-inoculated leaves of wild-type and
.padZ plants. Radiolabeling studies with labeled anthranilate
were then done to determine whether this putative
biosynthetic intermediate really was on the pathway between
anthranilate and camalexin. Finally, as an alternative to
studying the phenotypes of mutants in order to understand the
role of camalexin in disease resistance, a number of ecotypes
of wild-type Arabidopsis were screened for differences in

ability to produce camalexin, and efforts were made to

30

correlate those differences with the ability to respond to

the fungal pathogen Alternaria brassicicola.

31

REFERENCES

Bell AA. 1981. Biochemical mechanisms of disease
resistance. Annual Review of Plant Physiology 32: 21-81.

Browne LM, Conn KL, Ayer WA, Tewari JP. 1991. The
camalexins: new phytoalexins produced in the leaves of
camelina sativa (Cruciferae). Tetrahedron 47 (24): 3909-
3914.

Bu’Lock JD. 1965. The Biosynthesis of Natural Products.
London: McGraw-Hill.

Callaway A, Liu WN, Andrianov V, Stenzler L, Zhao JM,
Wettlaufer S, Jayakumar P, Howell SH. 1996.
Characterization of cauliflower mosaic virus (CaMV)
resistance in virus-resistant ecotypes of Arabidopsis.
Mblscular PlantaMicrobs Interactions 9(9): 810—818.

Chilton MD. 1983. A vector for introducing new genes into
plants. Scientific American June 1983: 50-59.

Ciufetti LM, VanEttsn HD. 1996. Virulence of a pisatin
demethylase-deficient Nectria haematococca MPVI isolate is
increased by transformation with a pisatin demethylase gene.
Mblscular PlantaMicrobs Interactions 9: 787—792.

Conn KL, Browne LM, Tewari JP, Ayer WA. 1994. Resistance to
Rhizoctonia solani and presence of antimicrobial compounds in
Gamelina sativa roots. JCurnal of Plant Biochemistry and
Biotechnology 3: 125-130.

Conn KL, Tewari JP, Dahiya JS. 1988 Resistance to
Alternaria brassicae and phytoalexin-elicitation in rapeseed
and other crucifers. Plant Science 56: 21-25.

Cruickshank IAM. 1962. Studies on phytoalexins. IV. The
antimicrobial spectrum of pisatin. Australian JCurnal of
Biological Sciences 15: 147-159.

Cruickshank IAM, Perrin DR. 1960. Isolation of a
phytoalexin from Pisum sativum L. Nature 187: 799-800.

Cruickshank IAM, Perrin DR. 1961. Studies on phytoalexins.
III. The isolation, assay, and general properties of a
phytoalexin from Pisum sativum L. Australian JOurnal of
Biological Sciences 14:336—348.

Dempsey DA. 1996. Characterization of the Responses of
Arabidopsis_thaliana to Turnip Crinkle Virus. Ph.D.

32

dissertation. Rutgers State University, New Jersey.

Dempsey DA, Pathirana MS, Wobbe KK, Klessig DF. 1997.
Identification of an Arabidopsis locus required for
resistance to turnip crinkle virus. Plant JCurnal 11(2):
301-311.

Deverall BJ. 1982. In: Bailey JA , Mansfield JW, eds.
Phytoalexins. Glasgow, UK: Blackie, 1-20.

Devys M, Barbier M. 1991. Indole—3-carboxaldehyde in the
cabbage Brassica oleracsa: a systematic determination.
Phytochemistry 30: 389-391.

Ebel J. 1986. Phytoalexin synthesis: the biochemical
analysis of the induction process. Annual Review of
Phytopathology 24:235—264.

Glazebrook J, Ausubel FM. 1994. Isolation of phytoalexin-
deficient mutants of Arabidopsis thaliana and
characterization of their interactions with bacterial
pathogens. Proceedings of the National Academy of Sciences
USA 91: 8955-8959.

Glazebrook J, Zook M, Mert FM, Kagan I, Rogers EE, Crute IR,
Holub EB, Hammerschmidt R, Ausubel FM. 1997. Phytoalexin-
deficient mutants of Arabidopsis reveal that PAD4 encodes a
regulatory factor and that four PADrgenes contribute to downy
mildew resistance. Genetics 146: 381-392.

Glazensr JA, Van Etten HD. 1978. Phytotoxicity of
phaseollin to, and alteration of phaseollin by, cell
suspension cultures of Phaseolus vulgaris. Phytopathology 68:
111-117.

Greenberg JT, Guo A, Klessig DF, Ausubel FM. 1994.
Programmed cell death in plants: a pathogen-triggered
response activated coordinately with multiple defense
functions. cell 77:551-563.

Gross D. 1993. Phytoalexins of the Brassicaceae.
Zeitschrift far Pflanzsnkrankhsitsn und Pflanzenschutz 100:
433-442.

Gross D, Porzel A, Schmidt J. 1994. Phytoalexins mit
Indolstruktur aus Kohlrabi (Brassica oleracsa var.
gongylodss). Zeitschrift far.Naturforschung 49c: 281-285.

Hahn MG, Bonhoff A, Grisebach H. 1985. Quantitative
localization of the phytoalexin glyceollin I in relation to
fungal hyphae in soybean roots infected with Phytophthora
megaspsrma f. sp. glycinsa. Plant Physiology 68: 1161-1169.

33

Hain R, Rsif H-J, Krause E, Langebartels R, Kindl H, Vornam
B, Wiese W, Schmelzsr E, Schreier PH, Stocker RH, Stenzel K.
1993. Disease resistance results from foreign phytoalexin
expression in a novel plant. Nature 361: 153-156.

Hammond-Kosack KE, Jones JDG. 1996. Resistance gene-
dependent plant defense responses. Plant Cell 8: 1773—1791.

Harborne JB (1988) Introduction to Ecological Biochemistry,
3rd edition. London: Academic Press, 315-327.

Hargreaves JA. 1980. A possible mechanism for the
phytotoxicity of the phytoalexin phaseollin. Physiological
Plant Pathology 16: 351-347.

Homan AL, Fuchs A. 1970. Direct bioautography on thin-layer
chromatograms as a method for detecting fungitoxic
substances. JOurnal of Chromatography 51: 327-329.

Jejelowo OA, Conn KL, Tewari JP. 1991. Relationship between
conidial concentration, germling growth and phytoalexin
production by camslina sativa leaves inoculated with
Alternaria brassicae. Mycological Research 95: 928-934.

Jimenez LD, Ayer WA, Tewari JP. 1997. Phytoalexins produced
in the leaves of capsella bursa-pastoris (shepherd’s purse).
Phytoprotsction 78: 99-103.

Jones ERH, Taylor WC. 1957. Some indole constituents of
cabbage. Nature 179: 1138.

Kuc J. 1995. Phytoalexins, stress metabolism, and disease
resistance in plants. Annual Review of Phytopathology 33:
275-97.

Manning DT, Galston AW. 1955. On the nature of the
enzymatically catalyzed oxidation products of indoleacetic
acid. Plant Physiology 30: 225-231.

Melchior GH. 1957. Uber den Abbau von Indolderivaten I.
Photolyse durch ultraviolettes Licht. Plants 50: 262-290.

Meyer J, Pohl R. 1956. Neue Erkenntnisse zum Problem dsr
phototropen Krummung bsi hohsren Pflanzen.
Naturwissenschaften 43: 114-115.

Mhiri C, Morel JB, Vernhettes S, Casacuberta JM, Lucas H,
Grandbastien MA. 1997. The promoter of the tobacco TNTl
retrotransposon is induced by wounding and by abiotic stress.
Plant.Mo1ecular Biology 33(2): 257-266.

Monde K, Takasugi M, Ohnishi T. 1994. Biosynthesis of
cruciferous phytoalexins. JOurnal of the American Chemical

34

Society 116: 6650—6657.

Moody CJ, Roffey JRA, Stephens MA, Stratford IJ. 1997.
Synthesis and cytotoxic activity of indolyl thiazoles. Anti-
cancer Drugs 8(5): 489-499.

Muller F. 1961. Uber den Abbau von synthetischen
Indolverbindungen durch Blumenkohlrosengewebe. Planta 57:
481-502.

Nicholson RL, Hammerschmidt R. 1992. Phenolic compounds and
their role in disease resistance. Annual Review of
Phytopathology 30:369-389.

Nurnberger T, Nennstiel D, Jabs T, Sacks WR, Hahlbrock K,
Scheel D. 1994. High affinity binding of a fungal
oligopeptide elicitor to parsley plasma membranes triggers
multiple defense responses. Cell 78: 449-460.

Paxton JD. 1981. Phytoalexins--a working redefinition.
Phytopathologischs Zeitschrift 101:106-109.

Paxton JD, Groth J. 1994. Constraints on pathogens
attacking plants. Critical Reviews in Plant Sciences 13(1):
77-95.

Pedras MSC, Khan AQ. 1997. Unprecedented detoxification of
the cruciferous phytoalexin camalexin by a root
phytopathogen. Bioorgenic and.Medicinal Chemistry Letters
7(17): 2255-2260.

Pedras MSC, Khan AQ, Taylor JL. 1997. Phytoalexins from
brassicas: overcoming plants’defenses. In Hedin PA,
Hollingworth RM, Masler EP, Miyamoto J, Thompson DG, eds.
Phytochemicals for Pest Control. ACS Symposium Series 658,
155-166.

Rausch T, Butcher DN, Hilgenberg W. 1983. Indols-3-
methylglucosinolate biosynthesis and metabolism in clubroot
diseased plants. Physiologia Plantarum 58: 93-100.

Ray PM, Curry GM. 1958. Intermediates and competing
reactions in the photodestruction of IAA. Nature 181: 895-
96.

Rogers EE, Glazebrook J, Ausubel FM. 1996. Mode of action
of the Arabidopsis thaliana phytoalexin camalexin and its
role in Arabidopsis—pathogen interactions. .Molscular Plant-
Microbe Interactions 9(8): 748-757.

Rouxel T, Kollmann A, Boulidard L, Mithen R. 1991. Abiotic
elicitation of indole phytoalexins and resistance to

Lsptosphasria maculans within Brassicsas. Planta 184: 271-
278.

35

Rouxel T, Renard M, Kollmann A, Bousquet J-F. 1990.
Brassilexin accumulation and resistance to Lsptosphasria
maculans in Brassica species and progeny of an interspecific
cross B. juncsa x B. napus. Euphytica 46: 175-181.

Rouxel T, Sarniguet A, Kollman A, Bousquet, J-F. 1989.
Accumulation of a phytoalexin in Brassica spp. in relation to
a hypersensitive reaction to Lsptosphasria maculans.
Physiological and.Molscular Plant Pathology 34: 507-517.

Schafsr W, Straney D, Ciuffetti L, Van Etten HD, Yoder CC.
1989. One enzyme makes a fungal pathogen, but not a
saprophyte, virulent on a new host plant. Science 246: 247-
249.

Schmelzer E, Kruger-Lebus S, Hahlbrock K. 1989. Temporal and
spatial patterns of gene expression around sites of attempted
fungal infection in parsley leaves. Plant Cell 1: 993-1001.

Schraudolf H. 1968. Untersuchungsn zur Verbreitung von
Indolglucosinolatsn in Cruciferen. EXpsrisntia 24: 434-435.

Smith DA. 1982. Toxicity of phytoalexins. In: Bailey JA,
Mansfield JW, eds. Phytoalexins. Glasgow, UK: Blackie, 218-
252.

Smith OH, Yanofsky C. 1963. Intermediates in the
biosynthesis of tryptophan. .Msthods of Ehzymology 6: 590-
597.

Snyder BA, Nicholson RL. 1990. Synthesis of phytoalexins in
sorghum as a site—specific response to fungal ingress.
Science 248: 1637-1638.

Stutz RE. 1958. Enzymatic formation of indole-3-
carboxaldehyde from indole-3-acetic acid. Plant Physiology
33:207.

Swain T. 1965. Methods used in the study of biosynthesis.
In: Pridham JB, Swain T, eds. Biosynthetic Pathways in Higher
Plants. London: Academic Press, 9-36.

Tsuji J, Somervills SC, Hammerschmidt R. 1991.
Identification of a gene in Arabidopsis thaliana that
controls resistance to xanthomonas campestris pv. campsstris.
Physiological and.Molecular Plant Pathology 38: 57-65.

'Tsuji J, Jackson EP, Gage DA, Hammerschmidt R, Somervills SC.
1992. Phytoalexin accumulation in Arabidopsis thaliana
during the hypersensitive reaction to Pseudomonas syringae
pv. syringes. Plant Physiology 98: 1304-1309.

36

van Denffer D, Fischer A. 1952. Papierchromatographischer
Nachweis des S-Indolaldehyds in photolytisch zersstzer IES-
Losung. Naturwissenschaften 39: 549-550.

Wasmann CC, VanEtten HD. 1996. Transformation-mediated
chromosome loss and disruption of a gene for pisatin
demethylase decrease the virulence of Nectria haematococca on
pea. .Mblscular Plant-Microbe Interactions 9: 793-803.

Whalen MC, Stall RE, Staskawicz BJ. 1988. Characterization
of a gene from a tomato pathogen determining hypersensitive

resistance in non-host species and genetic analysis of this

resistance in bean. Proceedings of the National Academy of

Sciences, USA: 6743-6747.

Wright AD, Sampson B, Neuffer MG, Michalczuk L, Slovin JP,
Cohen JD. 1991. Indole-B-acetic acid biosynthesis in the
mutant maize orange psricarp, a tryptophan auxotroph.
Science 254: 998-1000.

Zhao J, Last RL. 1996. Coordinate regulation of the
tryptophan biosynthetic pathway and indolic phytoalexin
accumulation in Arabidopsis. Plant Cell 8: 2235-2244.

Zhao J, Williams CC, Last RL. 1998. Induction of
Arabidopsis tryptophan pathway enzymes and camalexin by amino
acid starvation, oxidative stress, and an abiotic elicitor.
Plant Cell 10: 359-370.

Zhou N, Tootle TL, Tsui F, Klessig DF, Glazebrook J. 1998.
PAD4 functions upstream from salicylic acid to control
defense responses in Arabidopsis. Plant Cell 10: 1021-1030.

Zook M, Hammerschmidt R. 1997. Origin of the thiazole ring

of camalexin, a phytoalexin from Arabidopsis thaliana. Plant
Physiology 113: 463—468.

37

Chapter 2. Screen for Camalexin-Deficient Mutants of
Arabidopsis and Radiolabeling Studies With

Phytoalexin-Deficient (pad) Mutants

Introduction

The initial approach to determining the biosynthetic
pathway of camalexin and the role of camalexin in disease
resistance was to study mutants. A simple way to assess the
importance of a compound to an organism is to study the
phenotype of that organism lacking that particular compound.
The change in phenotype can be used to deduce the effect(s)
of having the compound. For example, the cprl mutant of
Arabidopsis, which constitutively expresses the pathogenesis—
relatsd protein PR-l, is slightly stunted and has smaller
leaves, suggesting that in the absence of a means to turn off
the PR-l gene, normal growth is reduced (Bowling et al.,
1994).

Mutants also frequently accumulate biosynthetic
precursors of the compound, if the mutation blocks a step of
the biosynthetic pathway such that the precursors cannot be
converted into the end product (Bu’lock, 1965; Pridham and
Swain, 1965; Smith and Yanofsky, 1963). These accumulating
precursors can be detected by supplying plants with labeled
precursors (either heavy isotopes or radioisotopes), and
studying the compound(s) into which the isotopic label was

incorporated. This approach was used by Wright et al. (1991)

38

to determine that indole-B-acstic acid (IAA, or auxin) in
maize was synthesized from indole and not tryptophan. Maize

tryptophan auxotrophs, when fed deuterated water (D20),

incorporated much more label into anthranilate than into
tryptophan. In addition, these mutants produced more IAA
than wild-type plants.

Besides acting as a source of accumulating precursors,
mutants can sometimes help to support or reject hypotheses
about whether a compound X'iS a precursor of another compound
.y. If a mutant is deficient in an enzyme needed to make
compound x, but that mutant can still make compound y; then
compound x is unlikely to be an intermediate in the
biosynthetic pathway to y. This approach was used to help
determine that camalexin is not synthesized from tryptophan.
Since tryptophan-deficient mutants produced wild-type amounts
of camalexin (Tsuji et al., 1993), tryptophan appears not to
be a precursor of camalexin.

Given the hypothetical pathway of camalexin biosynthesis
(Figure 5) and its hypothetical relationship to the
tryptophan biosynthetic pathway (Figure 4), some of the
compounds that might be expected to accumulate in camalexin-
deficient mutants would be indole-3-glycerol phosphate,
indole-3-carboxylic acid, indole-3-carboxaldehyde, and the
indole-3-thiazoline and indole-B-thiazolidine carboxylic
acids. The initial approach to finding these mutants was to
screen for camalexin deficiency in an ethyl methane sulfonate

(EMS)—mutagenized population of Arabidopsis seeds, and to

39

identify accumulating intermediates in those mutants by
feeding radioactive anthranilate, a known camalexin
precursor, and characterizing the compounds into which it was
incorporated. The agent used to elicit camalexin production
was the ascomycets Cochliobolus carbonum, a maize pathogen
that elicits an incompatible response (although it does not
cause a visible hypersensitive response) on Arabidopsis. The
choice of elicitor was based on the results of Samantha
Teplitsky, a high school student who worked in the
Hammerschmidt lab in the summer of 1991 and found that C.
carbonum induced higher amounts of camalexin than previously-
used elicitors (silver nitrate and Pseudomonas syringae pv.
syringes, an incompatible bacterial pathogen).

The mutant screen was facilitated by the fact that
camalexin has a very specific fluorescence spectrum, and that
it diffuses into droplets of spore inoculum directly above
the tissue in which it is made. The ability of camalexin to
diffuse into inoculum droplets was noted by Conn et al.
(1988), when inoculating camslina sativa and capsella bursa-
,pastoris with Alternaria brassicae. Phytoalexins from other
plants, such as pisatin in Pisum sativum (pea) (Cruickshank
and Perrin, 1961) and medicarpin in Medicago sativa (alfalfa)
(Harborne, 1988) have also been isolated from inoculum
droplets. This method simplifies extraction and separation
of phytoalexins, as it is not necessary to separate them from
pigments, and fewer compounds are present in droplets than in

leaf tissue (Cruickshank and Perrin, 1961; Harborne, 1988).

40

In the case of the mutant screen, the ability of camalexin to
diffuse into inoculum droplets made it possible to collect
those droplets and evaluate the mutants’ ability to produce
camalexin by measuring their fluorescence. This method was
simple and bypassed the more time-consuming procedure of leaf
tissue extraction and thin-layer chromatography.

Under normal circumstances, radiolabeling studies with
mutants would have been greatly delayed by the need to
backcross putative mutants to wild—type plants to ensure
that camalexin-deficient offspring were isogenic with the
wild-type plants, differing only in genes for camalexin
biosynthesis. However, near the start of this project, three
camalexin—deficient mutants were isolated by Dr. Jane
Glazebrook through a screen for plants that produced little
camalexin when inoculated with Pseudomonas syringae pv.
maculicola, a compatible pathogen (Glazebrook and Ausubel,
1994). Dr. Glazebrook kindly provided seed from the pad
(phytoalexin—deficient) mutants and agreed to test putative
mutants from this project for camalexin deficiency in her
screen, and to backcross confirmed mutants to wild-type
plants for genetic analysis. As the pad mutants had already
been backcrossed, the search for biosynthetic intermediates
by radiolabeling began with these, of which one (pad3) was a
null mutant producing no camalexin, and two (padl and pad2)
‘were leaky mutants producing 20 % and 10 % of wild—type
levels of camalexin, respectively (Glazebrook and Ausubel,

1994).

41

Materials and Methods

Reagents and Chemicals .

Camalexin cams primarily from a supply isolated from
Arabidopsis by Jun Tsuji (Tsuji et al., 1992). Some of the
camalexin used for standard curves in fluorimetric analysis
was synthesized according to the method of Ayer et al. (1991)

and purified as explained in Chapter 3. Radioactive
anthranilate (1W:,'uniformly labeled on the benzene ring.) was

purchased from Sigma. The specific activity indicated on the
label was 1-25 mCi/mmol, and the exact specific activity was
never determined. All other chemicals used for radiolabeling

and mutant screening were of reagent grade or better.

Plant Material and Growth Conditions.

The M2 generation of ethyl methane sulfonate (EMS)-
mutagenized seeds of the Columbia ecotype of Arabidopsis
(marker: glabrous; Col-g1), obtained from Lehle, were
screened for camalexin deficiency. For radiolabeling
studies, seeds of the third backcross of the padl, pad2, and
.pad3 mutants were generously provided by Dr. Glazebrook.
These mutants were compared initially to the Columbia-0 (Col-
0) ecotype. About halfway through these studies, a change
*was made to the Col-g1 ecotype, which was used because it was

a better point of comparison for the seeds being planted for

42

the mutant screen, which were grown from mutagenized Col—gl
seeds. The Col-O and Col—g1 ecotypes produced comparable
amounts of camalexin.

Seeds were grown in Clay pots 14.5 cm in diameter,
containing a mixture of perlite and Baccto® High Porosity
Professional Planting Mix on a 2-cm-thick layer of perlite.
The soil was wetted with fertilizer (half-strength Hoagland’s
solution, Appendix A) and hand-compacted to provide a level
surface, and a 1-2 mm layer of fine vermiculite was sprinkled
over the surface.

Seeds were washed in 5-ml glass screwcap tubes with
water plus a small drop of Triton X-100 detergent, rinsed (by
adding water to the tubes and removing it by aspiration)
until no foaming was seen, and distributed over the
vermiculite in a quasi-grid pattern with a Rainin P200
automatic pipettor. The plastic tip used on the pipettor was
cut to make an opening wide enough to take up seeds. To
ensure that many plants would grow in each pot,2 or 3 seeds
were planted in each spot, either by ejecting 2 or 3 seeds at
a time with the pipettor, or by distributing seeds twice over
the vermiculite. The latter approach, although generally
less desirable because it was more time-consuming, was used
for growing putative mutants, as it was easier to identify
the plant from which leaves cams if the plants were not
clustered too closely together. Pots were covered with
plastic wrap to maintain humidity; the trays holding the pots

were filled with a layer of water 3 cm deep; and the pots

43

were kept in a growth chamber with a 16- to 18-hour light
regime, at about 21 °C.

Seeds usually germinated 2 to 3 days after planting, at
which time the plastic wrap over the pots was slit. After
another 2 to 4 days, the plastic was removed, and about 20 ml
of half-strength Hoagland's solution was poured onto the
surface of the soil to fertilize the seedlings.
Fertilization was repeated every week, with the amounts
increasing to about 150 ml per pot as the seedlings grew. A
water height of about 3 cm was maintained in the trays for
about the first 10 days to ensure that the seedlings had
plenty of water. As the seedlings grew, the amount of water
in the trays was reduced, since the roots were bigger and
more capable of obtaining water and nutrients (and of
becoming oxygen-depleted in an overly moist environment).
Three- or four-week—old plants were watered only when the
water in the trays had almost disappeared. water was then

added only to cover the bottom of the tray.

Fungal cultures and inoculation procedures.
Cochliobolus carbonum was grown on V-8 agar (per liter:
200 ml V-8 juice, 2 g calcium carbonate, 14 g agar) on an 18-
to 24-hour light regime. Spores 1 to 2 weeks old were used
to inoculate Arabidopsis leaves. The inoculum was prepared
by flooding a plate of spores twice with non-sterile
deionized water and filtering the spores through two layers

of cheesecloth. Few precautions with sterility were taken at

44

this point, as the concentration of spores was far greater
than the concentration of any contaminant was likely to be.
The main objective was to find plants that could not produce
camalexin, regardless of the eliciting conditions. In
addition, the leaves to be inoculated were not sterile.
Inoculum concentration was not measured in mutant
screening, but as the amount of camalexin produced appeared
to increase with increasing spore concentration, based on
casual observations and published results (Jejelowo et al.,

1991), the inoculum was made as thick as possible and
probably ranged from5x10S to 1x106 spores/ml. For the
radiolabeling studies, spore concentrations ranged from
3.6x105 to 9.6x105 spores/ml.

Rosette leaves from 3- to 4-week-old plants were excised
with a razor blade and placed in covered 15-cm diameter Petri
dishes lined with Whatman filter paper (#1 or #4) moistened
with water to maintain a humid environment. Leaves were
placed with the adaxial side down and inoculated on the
abaxial side. The spore suspension was stirred frequently
during the inoculation, as the spores rapidly settled to the
bottom of the container. For the mutant screen, as the
objective was simply to determine if camalexin could be
produced, leaves were inoculated with the maximum amount of
inoculum that would stay on the leaf. For the radiolabeling
studies, leaves were inoculated with 0.1 ml of spores,
distributed with a P1000 automatic pipettor whose plastic tip

‘was cut to facilitate taking up spores. Inoculating with a

45

fixed volume ensured greater reproducibility of results. The
amount of camalexin produced in leaves is greatest directly
beneath the inoculum droplet (Jejelowo et al., 1991).
Consequently, if the same amount of inoculum was used on each
leaf and the covered surface of each leaf was roughly the
same, the amounts of camalexin and precursors produced should

have been fairly similar from one experiment to the next.

Extraction of camalexin.

For the mutant screen, camalexin did not need to be
extracted from inoculum droplets. However, when putative
mutants were found, the droplets from a few pairs of leaves
were pooled and extracted with ethyl acetate (see Chapter 3)
and compared on thin-layer chromatograms to extracts from
wild-type leaves. The general method of leaf tissue
extraction consisted of boiling leaves for 20 minutes in 80 %
methanol or 100 % methanol, a method frequently used for
plant tissue extraction because methanol penetrates tissues
quickly (Harborne, 1973). In the radiolabeling studies,6
leaves per sample were placed in 12 ml of 80 % methanol/20 %
water and heated in a water bath until the volume was reduced
by about one-half. The sample was filtered through
cheesecloth to remove leaves and boiling chips. The volume
of extract was sometimes increased by 50 % by adding water
(final volume approximately 9 ml), and it was then extracted
3 times with an equal volume of chloroform. The chloroform

layers were pooled and dried over sodium sulfate, and the

46

chloroform was removed by drying at 45-50 ’C on a rotary
evaporator (Buchi). Samples were transferred to small test
tubes by rinsing the round bottom flasks with two 250-ul
aliquots of methanol, dried under nitrogen, and redissolved

in 50 ul of methanol for thin-layer chromatography.

Thin-layer chromatography (TLC) of extracts.

Samples (10 or 20 ul) were loaded onto a 20x20 cm glass-
backed silica-gel TLC plate (Fisher dei-Plate). Although
initially it was not done, eventually the plates were always
activated by drying at 80-100 'C for 20 minutes. Plates were
pre-developed in chloroform-methanol (3:2 or 1:1, v/v) if
precleaning was needed. Samples were loaded as thin bands
with 10-u1 micropipets (VWR Scientific). Standards of
camalexin, indole-3-carboxaldshyde, and indole-3-carboxylic
acid were also loaded. Plates were developed in chloroform-
methanol (9:1 or 19:1, v/v). This development was sometimes
preceded by development with chloroform to help separate the

less-polar leaf pigments from other compounds.

Mutant Screen .

Plants were screened for camalexin deficiency when 3 to
4 weeks old: at or near the time of flower production, but
before the leaves began to senesce and turn purple. In each
pot, 10 to 15 plants were chosen at random and marked with a
numbered tag on a wooden toothpick. Two leaves were excised

from each plant and inoculated. After 36 to 48 hours,

47

droplets were collected with a Pasteur pipet and transferred
to 13x100 mm glass tubes. Deionized water (2 ml) was added
to each tube, and the samples were mixed by vortexing.
Fluorescence was measured on a Hitachi F-2000 fluorimeter
(wavelength of excitation 330 nm; wavelength of emission 393
nm). The wavelengths of excitation and emission were
supposedly fairly specific to camalexin (M. Zook, pers.
comm.). Thus, other fluorescent compounds that might have
been present, such as sinapic acid, were not likely to create
false positive results. Readings were given, as is common in
fluorescence spectroscopy, in relative intensities, based on

the following formula:

F=K Io c l s 0

where F=relative fluorescence intensity; K=a constant defined

for the instrument; Io=intensity of the light entering the

flow cell; c=sample concentration; l=path length of flow

cell; e=extinction coefficient of sample; and ®=quantum yield

(ratio of optical energy absorbed to total fluorescent energy
emitted).

No standard curve of fluorescence versus camalexin
concentration was prepared at the time of the mutant
screening. The screen was intended to be qualitative, and
plants were considered putative mutants if relative
intensities were below 200. Wild-type relative intensities
were typically over 1000 (Tables 1-3). Eventually, a few

samples from pad3 and wild-type plants were always inoculated

48

and evaluated with the samples from putative mutants, to
determine the range of values that represented camalexin
deficiency and wild-type camalexin production.

If samples had readings in the pad3 range, the plants
from which the leaves had been taken were saved and allowed
to go to seed. (As Arabidopsis is self-pollinating, the seed
contained only the genes of the original plant of interest.)
Eventually, plants with readings between pad 3 and wild-type
readings were saved for seed, in case the intermediate
readings were due to a heterozygous trait for camalexin
deficiency. The seed was collected by shaking the siliques
over a wire mesh, retrieving the seeds that fell through onto
paper, and storing them in vials or microfuge tubes over
desiccant at 4 °C. Seeds were planted (sometimes by the
method described previously, but more frequently by
sprinkling them over the pot surface without washing, as
aspirating water off of seeds could cause a high loss of
seed) and grown under the conditions previously described.
Ten to forty of these M3 seedlings were chosen at random and
numbered, and 2 leaves from each seedling were screened for
camalexin deficiency as previously described. If all or most
of the samples had low readings, the plants were considered
to be camalexin-deficient mutants, and the M3 or M4 seed (the
latter if no M3 seed were left over from planting) was sent
to Dr. Glazebrook for further analysis. If the M3 seedlings
had readings that ranged from low to high, the mutant was

considered a potential camalexin-deficient heterozygote, and

49

the seed from plants giving low readings was saved for
replanting and retesting, or sent directly to Dr. Glazebrook
if the readings fell into a convincing 1:2:1 ratio of low to
intermediate to high readings (a good indication, based on
Classical Mendelian genetics, that recombination between 2
genes Aa and Aa had occurred and given rise to a 1:2:1 ratio

of aa:Aa:AA).

Radiolabeling studies .

Leaves of 3- to 4-wesk-old wild-type and pad seedlings
were excised, placed in Petri dishes lined with moistened
filter paper, and inoculated with 0.1 ml of water or C.
carbonum spores. After the desired incubation time, the

leaves were blotted dry with a Kimwipe and leaves placed in
trimmed 0.7-ml microfuge tubes containing 50 ul of 1%}-

anthranilic acid diluted with water so that the activity in
each tube was 0.035-0.058 uCi (0.21-0.35 uCi per 6-leaf
sample). Initially, labeled anthranilate was fed to leaves
after a 24—hour incubation period. That was the time of
maximum camalexin accumulation in leaves inoculated with
Pseudomonas syringes pv. syringes (Tsuji et al., 1991), and
high concentrations of camalexin had been found at 24 hours
in preliminary non-radioactive extractions of C. carbonum-
inoculated leaves. It seemed, according to standard
procedures for biosynthetic studies (Bu’lock, 1965), that the
time of maximum accumulation would be the best time to detect

incorporation into camalexin and intermediates in wild-type,

50

and into biosynthetic blocks in the mutants. To ensure that
the leaves would be able to take up solution, the petiole was
cut under water just before placement in the tube. After the
solution was taken up (2-4 hours), the tube was marked with a
dot to indicate that uptake was completed, and 50 ul of water
were added to prevent the leaf from becoming desiccated.

When all leaves had finished taking up solution, they were
extracted as described above. After separation by TLC, the

Rf's of the standards were recorded; and the plates were

sprayed with ENTHANCE spray (a fluor from NEN Research
Products that aids in visualizing low-activity bands on
film), wrapped in Bordenn‘pflastic wrap, and laid onto film

for 7-10 days.

Results

Mutant screen.

Corey Sonnett, an undergraduate assistant in the lab,
and Dr. Michael Zook isolated mutant 2120. This mutant was
found by Dr. Glazebrook to be heterozygous for camalexin
deficiency, with the homozygous form of the gene being a
recessive trait. The camalexin-deficient homozygous
recessive progeny from this mutant were bred for further
analysis, and this mutation has been named pad5 (Glazebrook
et al., 1997). Due to the realization that some mutants

might be heterozygous for genes involved in camalexin

51

production, the checks for putative heterozygotes mentioned
in the Materials and Methods were implemented.

As a result of this more careful screening, two putative
heterozygotes, mutants 4420 and 4440, were identified (Table
1). These were identified by comparing their relative
fluorescence intensities to those of pad3 and wild-type, and
by grouping them according to the different ranges of
relative intensities noted. Later, to quantitate those
values, a standard curve of relative intensity versus
camalexin concentration was determined (Figure 6). For both
mutants, the average amount of camalexin in plants with
intermediate relative intensities was about twice the amount
in plants with low intensities and about half that in plants
with high intensities. The number of plants in each
category, for mutant 4440, was 10 low to 18 intermediate to
10 high readings: almost a 1:2:1 Mendelian ratio. Even the
plants with high relative intensities produced less than half
the amount of camalexin produced by wild-type. Although the
segregation ratios for mutant 4420 were less impressive (5
low to 10 intermediate to 14 high readings), the categories
were reasonably distinct, and it is possible that the number
of plants screened was too small to observe Mendelian
patterns of segregation. However, when these mutants
(identified by their response to inoculation with C.
carbonum), were inoculated with Pseudomonas syringes pv.
maculicola strain ES4326 (Psm ES4326), they produced wild-

type amounts of camalexin (J. Glazebrook, pers. comm.).

52

60001

5000'

4000‘ R‘Z I 1.000

3000-
20004

1000‘

d

o T U l I W I l I I l’ U V l

00 (L2 0A, (L6 03 11) L2
camalexin (ug/ml)

fluorescence (relative

intensity)

 

 

'fi

Figure 6. Standard curve of camalexin concentration
versus fluorescence (excitation wavelength 330 nm,
emission wavelength 393 nm). Data represent means
plus standard errors of relative intensities
(duplicate readings) of pure synthetic camalexin
dissolved in methanol and diluted with water
(maximum methanol concentration 0.5%).

53

iii flflflflnflflnflflfl Eli
,
I

HN.OHmN.H moo.OHmbN.o mHo.0Hmmm.o boo.oummd.o moao.o#mmbo.o

 

 

 

 

Aeuzv .oauzv Aoanzv Amauzv Aoauzv I

mbaawwmae h.NHHv.meH moauovom mmswmoa m.Hme.Hme oeuvm

. HN.0HNNZH mmoo.oambmo.o mmo.osmom.o mao.osmam.o mwoo.ouammo.o I
I Amnzv Aoauzv Avauzv Aoauzv Amuzv

w vaHHmmwm b.mvse.mbe Hmmaemmm mmsemaa m.besv.emm Axes

. oaks mmagfiom mmamfimm I
“ IUHwS Ga mme. snows; twee; mmamaom .BoHs

 

 

 

 

 

GflXOHMEMU CH CflXOHMfiMU «Hm GMXGHMEMU Gun GHXOHMEMU Gun CHXOHMEMU UnhfifiudflU

 

 

.musousfi m>fiuousm was no DEAD Tenn ozu so ombudsoosa mm>mmH mbmm.oso mmhuuoafi3 Ge
deaumuusmosoo saxwamamu wuocmo mnfidaoo auxflw can Aswan .m madmam as m>uso pudendum
was so Demon .moma mom mEoHGOHOflE cw coauouusoosoo saxmamsoo no sound pudendum mafia
some mmuosoo 30H Usage one nomswmuom wmcaaomom mo Henson mmuosmo 30H osoomm “huwmcwusfi
mucmommuosam m>flumeH «0 sound Choosoum moan some ascendancy HHTO some Ga muonfiss

mo 30H umnwm .wmuomhnoumuon snowbammonsflxmaoemu o>wumusm N Ca cfixwamemu .H wanna

54

Two putative camalexin-deficient mutants that were not
heterozygotes--plants 4648 and 4512—-were also isolated.
Table 2 demonstrates that mutant 4648 produced about 55 % as
much camalexin as pad3 produced on the day that 4648 was
screened, and about as much as pad3 produced on other
occasions. mutant 4512, although it produced 4-5 times more
camalexin than pad3, was considered a mutant because it
produced only 14 % as much camalexin as wild-type.

It should be noted that the amount of camalexin produced
by wild-type and pad3 leaves varied at each screening (Tables
1-3). Camalexin concentrations varied by a factor of 2 to 3
in wild-type leaves, and by a factor of 2 to 5 in pad3
leaves. It is rather interesting that pad3, supposedly a
null mutant, produced small amounts of camalexin. This
result was confirmed by Kaitlyn Hwang, a high school student
who worked in the Hammerschmidt lab during the summer of 1995
and found that pad3, 48 hours after inoculation with the
fungus Collstotrichum lagenarium, produced small amounts of
camalexin that could be detected by high—performance liquid
chromatography. Perhaps the fluorescence in pad3 was partly
due to other fluorescent compounds whose spectra slightly
overlapped the camalexin spectrum. The readings were not due
to instrument background, as inoculation with water alone

gave even lower readings (Table 3).

55

Table 2. Camalexin in putative camalexin-deficient mutants
4648 and 4512, and in padl and pad2. Third and fourth
columns represent camalexin concentrations for wild—type and
lpad3 leaves inoculated at the same time as the putative
mutants. See Table 1 for definitions of numbers in each row.

 

 

 

 

 

 

camalexin in. camalexin in. camalexin in
mutant wild-type 'pad3
4648 237.0:10.2 >9999 (out of 431.0116.5
fluorimeter
range)
(N=31) (N=10) (N=7)
0.0450r0.0018 >1.78 0.0795i0.0029
4512 449.9i4l.6 3225:757 97.41141.55
(N=39) (N=4) (N=10)
0.0828i0.0074 0.57710.135 0.0201:0.074
(pad2 524.9120.8 81071900 229.9i35.8
(N=5) (N=4) (N=3)
0.0962r0.0037 1.44:0.16 0.0437i0.0064
Ipadl 9724:1101 22240i3166 596.9:24.3
-. (N=6) (N=3) (N=10)
1.73:0.20 3.96:0.56 0.10910.004

 

 

 

 

56

Table 3. Comparison of camalexin concentrations in inoculum
droplets from fungal- and water-inoculated leaves of wild-
type, pad2, and pad3 plants. The mean plus or minus the
standard error of relative fluorescence intensities is shown
for the number of seedlings tested (N).

 

 

 

    
   
 

 

 

 

 

sample camalexin in camalexin in
droplets of fungalr- droplets of water-
inoculated inoculated
sample(relative sample(relative
intensity) intensity)
wild-type 8107:900 (N=4) 16.48r10.29 (N=3)
pad2 524.9:20.8 (N=5) 21.46r8.73 (N=4)
pad3 229.9:35.8 (N=3) 17.40r8.50 (N=3)

 

 

Mutant 4648 produced no camalexin in response to Psm
ES4326 (Glazebrook et al., 1997). Complementation tests
demonstrated that this mutant was in the same complementation
group as pad3, and it was named pad3-2 (Glazebrook et al.,
1997).

Mutant 4512, however, like mutants 4420 and 4440,
produced wild-type amounts of camalexin in response to Psm
ES4326. Apparently, some of the mutants were responding
differently to the different pathogens. This suspicion had
already arisen with the pad mutants, due to the suggestion of
undergraduate assistant Corey Sonnett to screen those
fluorimetrically. An initial check of several inoculated
,padl leaves indicated that they produced wild-type amounts of
camalexin; a more careful screen with more samples (Table 2)
reinforced those findings. Only one sample had a relative
intensity that did not exceed the range that the fluorimeter

could read (0-9999); other samples had to be diluted by a

57

factor of 1 or 2 to obtain a measurable intensity. The
average amount of camalexin produced by padl on that occasion
was higher than the amount produced by wild-type on other
days when plants were screened, and higher than the maximum
produced by wild-type in some of time courses described in
Chapter 3. In contrast, pad2 produced roughly 15 % of wild-
type amounts of camalexin (Table 2), an amount similar to
what had been found upon inoculation with Psm ES4326

(Glazebrook and Ausubel, 1994).

Radiolabeling studies .

The focus of the radiolabeling studies was on putative
intermediates in the organic (chloroform) layer of leaf
tissue extracts. The hypothetical camalexin intermediates
could be in either the aqueous or organic phase, depending on
the pH of the extract. The amino group on the indole moiety,
if it became protonatsd, would cause the compound to
partition to some extent into the aqueous phase; or if the pH

were above the pKa of the compounds, the carboxyl groups would

become deprotonated, and the compounds would partition into
the aqueous phase. At a lower pH, however, indole-derived
compounds would be more likely to partition into the organic
phase. Also, if any camalexin intermediates were
constitutive and stored as glycosides or other conjugates,
these would be in the aqueous phase. Because anthranilate is
a precursor of many polar compounds, which would be in the

aqueous phase and would generate very complex chromatograms,

58

it was simpler, and possibly less misleading, to focus on the
organic phases. A few attempts were made to chromatograph
the aqueous phases in a couple of different solvent systems
(butanol-acetic acid-water 4:1:1, v/v; and chloroform-ethyl
acetate-formic acid 35:55:10, v/v). However, the resolution
was poor, and the chromatograms contained numerous bands.

Analysis of the aqueous phases was quickly abandoned, apart
from extracts of inoculum droplets on 35’S—Cysteine-labeled

leaves that will be described later.

Since the camalexin biosynthetic pathway is pathogen—
induced, putative intermediates were expected to appear as
bands on autoradiograms of TLC plate analyses of extracts of
fungal-inoculated samples. It was also possible that
constitutive plant compounds would be biosynthetic
intermediates, since phytoalexin biosynthetic pathways are
derived from primary metabolic pathways (Kuc, 1995).
Evidence of a constitutive intermediate could be a band that
would be present in a water—inoculated control but absent or
very faint in the fungal-inoculated sample due to its being
used in camalexin biosynthesis. Other intermediates could be
constitutive but have their synthesis stimulated in response
to infection. In that case, a band would be present in both
water-and fungal-inoculated samples but darker in the
inoculated sample. Still other intermediates could be
nonconstitutive and formed de novo in response to infection,
in which case the bands would be absent in water-inoculated

and present in fungal-inoculated samples.

59

In two radiolabeling experiments (Figures 7 and 8), no
obvious intermediates were detected in padl, pad2, or pad3.
In other words, no compounds consistently behaved like one
of the types of intermediates described above. In most of
the samples extracted in late June of 1994 (Figure 7), a band

with an R.f of 0.13 was present. In wild-type and pad3, this

band was darker in the control than in the inoculated
samples, suggesting that it was a constitutive intermediate
converted into camalexin after infection. In padl, however,
this band was darker in the inoculated than in the control
sample, as if it were a constitutive intermediate whose
production was stimulated after infection. The relative
darkness of the band in control and inoculated samples varied
between experiments as well. In samples extracted in mid-

Juns of 1994 (Figure 8), a band with an R.f of 0.13-0.15

(presumably the same compound as the one seen at that
approximate position in Figure 7) was less dark in the padl
control than in the inoculated sample. It was absent in
lpad3, suggesting this time that if it were an intermediate,
it was formed ds novo instead of being constitutive.

Although it was again darker in the wild-type control than in
the inoculated sample, the majority of the information from

the two experiments was contradictory. The compound at the Rf

of 0.13-0.15 seemed likely to be unrelated to camalexin
biosynthesis because the amount of label incorporated varied

so much.

60

 

' pl-c pl-i p2-c

0.69 ’

0.49
(cam)->

0.13-p

OR -§

 

Figure 7. Autoradiogram of TLC plate: tissue
extracts of wild-type (wt), padl (p1), pad2 (p2),
and pad3 (p3) leaves, fed 1“C-anthranilate 24 hours
after inoculation with water (c) or Cochliobolus
carbonum (i), 6/27/94-6/28/94. Samples were
dissolved in methanol and loaded onto a

20x20 cm glass—backed silica TLC plate, which was
developed in chloroform followed by chloroform-
methanol (19:1, v/v). Arrows at left denote the
origin (OR), solvent front (SF), and bands discussed
in the text. Numbers at left denote Rfvalues of
indicated bands. The total distance traveled by the
solvent front was 15 cm. Other abbreviations:
cam=camalexin.

61

wt-c wt-i pl—c pl—i p3-c p3-i

SF

0.65
-0.67

0.47
(cam)

OR ->

 

Figure 8. Autoradiogram of TLC plate: tissue
extracts of wild—type, padl, and pad3 leaves fed
14C-anthranilate 24 hours after inoculation

with C. carbonum, 6/15/94—6/16/94. Abbreviations
and TLC conditions are as indicated in Figure 7.
Note the intensity of the camalexin(cam) band in the
extract of inoculated padl (p—l i) leaves.

62

A band with an R.f of 0.65-0.67 (the approximate Rf of

indole in this solvent) was darker in control than in
inoculated wild-type and padl in Figure 8, again suggesting a
constitutive intermediate. However, since in Figure 7 this

band (Rf=0.69) was darker in inoculated padl than in the

control, the results seemed again too variable to be related
to camalexin biosynthesis.

Interestingly, relatively high incorporation of
anthranilate into camalexin was observed in fungal-inoculated
leaves of padl (Figures 7 and 8) and pad2 (Figure 7). These
results suggested that the padl and pad2 mutants might be
mutated not in camalexin biosynthetic enzymes but in the
ability to recognize certain pathogens. Although they
synthesized 30 % and 10 % of wild-type levels of camalexin,
respectively, in response to Psm ES4326 infection (J.
Glazebrook, pers. comm.), they appeared to synthesize more in
response to C. carbonum infection.

Since padl and pad? appeared to be regulatory mutants,
the focus of the labeling studies turned to pad3, in which no
camalexin had been detected in previous experiments.

Previous experiments (Figures 7 and 8) indicated no
incorporation of anthranilate into camalexin in pad3. No
intermediates were detected in pad3 in those experiments
(Figures 7 and 8), but it seemed possible that they might
accumulate prior to the onset of camalexin production and

then be diverted into other biosynthetic pathways. Searching

63

for intermediates at various points in time is a technique
frequently used to establish a biosynthetic pathway. One of
the best-known examples is the work of Melvin Calvin to
establish the path of carbon in photosynthesis (Taiz and
Zeiger, 1992). Time course analysis was also used to
determine the biosynthesis of indole alkaloids in periwinkle
(Vince rosee)(Scott et al., 1971). To search for

intermediates that might accumulate before camalexin was
produced, wild-type and pad3 leaves were fed 1“C-anthranilate

3-18 hours after inoculation. In one such experiment (Figure
9), label was incorporated at 3 hours into a green

fluorescent band with an R.f of 0.50. This band was darker in

inoculated pad3 than in inoculated wild-type leaves, which
suggested that it was a camalexin intermediate accumulating
due to a blocked biosynthetic pathway. In wild-type leaves,
the band disappeared at 6 hours, when a camalexin band was
first detectable, which suggested that it was being converted
into camalexin. However, the results for these time courses
were again quite variable. In the first one done (results
not shown), the band seen at 3 hours was about equally
intense in control and inoculated wild-type samples, and it
was absent in pad3. In a third time course, camalexin was
present at the earliest time point (3 hours) and the sought-
after compound was not. This inducible compound was then not
investigated further.

The unusually early appearance of camalexin could have

64

wt-c wt-i p3—c p3—i wt-c wt—i p3—c p3-i
3h 3h 3h 3h 6h 6h 6h 6h

SF—>

0.50

0.45
(cam)

 

OR—> ..

Figure 9. Autoradiogram of TLC plate: tissue
extracts of wild—type and pad3 leaves fed 1"C-
anthranilate 3 and 6 hours after inoculation
with C. carbonum, 10/9/94. Abbreviations and
TLC conditions are as indicated in Figure 7.

65

wt-c wt—i
7h 7h

SF—+>

0.23
(cam)

OR—>

 

Figure 10. Autoradiogram of TLC plate: extracts of
inoculum droplets from wild-type leaves fed 35S-
cysteine and then inoculated with C. carbonum,
3/29/95. Droplets were extracted 7 hours after
inoculation. Samples were dissolved in methanol and
loaded onto a 20x20 glass-backed silica plate, which
was developed in chloroform—ethyl acetate-formic
acid (35:55:10, v/v/v). Abbreviations are as
indicated in Figure 7.

66

been due to the inoculated leaves synthesizing camalexin
during the time when they were being fed 14C-anthranilate.
Since the feeding period lasted about 4 hours, the total time
after incubation would have been 7 hours. This would have
been enough time for camalexin synthesis to occur. From then
on, radioactive solutions were fed to leaves prior to
inoculation.

Although it might have been reasonable to resume pursuit
of the putative intermediate after changing the sequence of
leaf feeding and inoculation, this question was temporarily
abandoned to test a hypothesis that pad3 might be mutated in
the ability to incorporate sulfur-containing compounds, which
would not be detectable by 14C-labeling. Wild-type and pad3
leaves‘were fed either 1’iC-anthranilic acid or 35S-cysteine
and inoculated with C. carbonum. Inoculum.droplets were
extracted 0, 3, and 7 hours after inoculation. One inducible
band near the origin was detected after a 7-hour incubation
in cysteine-fednwild-type leaves. The fact that the wild-
type plants appeared to be a better source of intermediates
than a null mutant was one of the final factors in the
decision to abandon the labeling studies with the mutants.

To see whether the inducible band might represent several
poorly-resolved.compounds, wild-type leaves were again fed
iBS-cysteine. Inoculum droplets were extracted 7 hours after
inoculation. The extracts were divided.between 2 TLC plates,

of which one was developed in ethyl acetate-methanol

67

(24:1,v/v) and one was developed in chloroform-ethyl acetate-
formic acid (35:55:10, v/v). Several inducible bands were
present on both plates, but more bands were present on the
plate developed in chloroform—ethyl acetate-formic acid
(Figure 10). One attempt was made to characterize these
compounds in a similar, nonradioactive experiment with larger
quantities of leaves. When high-performance liquid
chromatography (HPLC) did not resolve the compounds well
enough for characterization, the labeling project was

abandoned.

Discussion

Both the mutant screen and the radiolabeling studies had
unexpected and rather disappointing results: putative mutants
that were camalexin-deficient in response to C. carbonum
inoculation were not camalexin-deficient in response to Psm
ES4326 inoculation, and the pad mutants did not contain
obvious biosynthetic blocks based on TLC analysis. To
confuse matters further, padl appeared to produce wild-type
amounts of camalexin based on fluorescence intensity (Table
2) and radiolabeling data (Figures 7 and 8), and pad2,
although it produced low amounts of camalexin based on
fluorescence intensity (Table 2), seemed to incorporate
fairly high amounts of 14C-anthranilate into camalexin (Figure
7).

Several explanations are possible for the discrepancies

in results for the mutants that appeared camalexin-deficient

68

in response to C. carbonum inoculation and not in response to
Psm ES4326 inoculation. One is that the C. carbonum screen
was based on the camalexin content of inoculum droplets,
while the Psm ES4326 screen was based on the camalexin
content of leaf tissue. There may have been more camalexin
in the C. carbonum-inoculated leaves than in the droplets.

It was determined later, through dose-response analyses (see
Chapter 3), that although leaves inoculated with 10S spores/ml
of C. carbonum contained about the same total amount of
camalexin as leaves inoculated with 106 spores/ml, the
partitioning of camalexin between leaf tissue and droplets

differed greatly: at 105 spores/ml, most of the camalexin was

in the droplet, while at 106 spores/ml, most of the camalexin

was in the leaf. The reason for this difference is
uncertain, but it may be due to non-specific binding of
camalexin to spores on the leaf surface (R. Hammerschmidt,
pers. comm,) The inoculum concentration used in the mutant
screen was never measured, but the spore suspensions were

always very dark and probably (based on subsequent
experience) closer to 106 than to 105 spores/ml. Perhaps the

camalexin produced after inoculation remained in the leaf
tissue and was not detected in the screen. This possibility
helps to account for the many M2 plants that appeared to be
camalexin-deficient but produced M3 seedlings with wild—type
concentrations of camalexin.

It is also possible that the kinetics of camalexin

69

production varied between wild-type plants and some of the
putative mutants. It was determined later (see Chapters 3
and 4) that in wild-type plants alone, the time at which
camalexin concentrations start to increase after inoculation
with C. carbonum, and the time of maximum accumulation, can
vary from 12 to 48 hours. Perhaps some of the putative
mutants produced camalexin more rapidly than wild-type in the
C. carbonum screen and were examined after the camalexin was
metabolized to other compounds. Alternatively, they produced
camalexin more slowly and were examined before camalexin
concentrations increased significantly. This hypothesis is
supported by more recent studies of camalexin production in
tryptophan-deficient (trp) mutants (Zhao and Last, 1996).
Whereas the trp1-100 mutant previously was found to produce
significantly less camalexin than wild-type plants 18 hours
after inoculation (Tsuji et al., 1993), it produced as much
camalexin as the wild-type 24 hours after inoculation (Zhao
and Last, 1996).

Another possible explanation for discrepancies between
results of the C. carbonum and the Psm ES4326 screens (one
that would also provide an alternative explanation for the
results with the trp1-100 mutant just mentioned) is that some
of the mutants were mutated not in camalexin biosynthetic
genes, but in genes required for recognition of certain
pathogens. As mentioned in Chapter 1, the mechanism for
induction of phytoalexin synthesis may be a signal or signals

from oligosaccharides or polypeptides from plants (Ebel,

70

1986) or from fungal cell walls (Ebel, 1986; Albersheim and
Valenti, 1978; Nurnbergsr et al., 1994). These signals
appear to be very specific: in parsley cell suspension
cultures, phytoalexins were elicited by a 13-amino acid
polypeptide from cell walls of Phytophthora.msgasperme f. sp.
glycinse, and a putative receptor binding site was found
(Nurnbergsr st al., 1994). Mutants 4440, 4420, and 4512 may
have had an intact camalexin biosynthetic pathway but had a
mutation in a gene involved in recognizing and responding to
a signal from C. carbonum. This gene may not have been
required for responding to Psm ES4326--which would not be
surprising, as bacteria, having modes of infection different
from those of fungi, would be expected to produce different
signals that would be recognized by different genes or gene
products. Zhao and Last's (1996) results for the trp1-100
mutant may be explainable by this reason, as much as by the
consideration of kinetics of accumulation, as the trpl-lOO
mutant in their studies was inoculated with Psm ES4326,
whereas in the study by Tsuji et al. (1993), it was
inoculated with silver nitrate. It may be that the trpl-lOO
mutant responds more rapidly to bacterial infection than to
abiotic elicitation.

Both the fluorimetric and the radiolabeling studies
suggest that padl may be a regulatory mutant, although this
suggestion has not been verified. A more-recently isolated
,ped:mutant,.ped4, has demonstrated that at least one PAD gene

has a regulatory and not a biosynthetic function. In

71

response to Psm ES4326, ped4 produced 10 % as much camalexin
as wild-type, whereas it produced wild-type amounts of
camalexin in response to C. carbonum (Glazebrook et al.,
1997). Recent work by Zhou et al. (1998) demonstrated that
.ped 4 also produced wild-type levels of camalexin in response
to Kanthomonas campsstris pv. campsstris strain BP109, silver
nitrate, and Psm ES4326 carrying the avirulence gene eerpt2.
A role for the PAD4 gene was proposed, based on the fact that
the ped4 mutation led to low salicylic acid production in Psm
ES4326-inoculated plants. Salicylic acid (SA) plays an
important role in the induction of plant responses to
infection (Vernooij et al., 1994; Zhou et al., 1998). The
PAD4 gene may affect camalexin production by encoding a gene
needed to activate SA production. Since SA concentrations
were normal in plants inoculated with Psm ES4326 bearing avr
Rpt2, it appears that more than one signal transduction
pathway to camalexin production exists in Arabidopsis.
Following abiotic elicitation or infection by fungi or
avirulent pathogens, a signal transduction pathway not
involving PAD4 is activated. The genes that are activated
can, in turn, activate production of SA, which leads to
camalexin production (Zhou et al., 1998). In contrast,
infection by Psm ES4326 requires PAD4 in order for SA
production to be activated.

In light of the findings of Zhou et al.(1998), it is
easier to understand the ambiguous results in camalexin

production with padl in the fluorimetric and radiolabeling

72

analysis, and with pad2 in the radiolabeling analysis. These
mutants may respond differently to signals from C. carbonum
than to signals from Psm ES4326. It is also easier to find
reasons for why mutants 4420, 4440, and 4512 appeared
camalexin-deficient in response to C. carbonum but not in
response to Psm ES4326. The fact that camalexin is produced
in response to chemicals causing amino acid starvation or
oxidative stress (Zhao et al., 1998), suggests that many
regulatory genes could be involved in its biosynthesis, in
which case a number of mutations, unrelated to camalexin
biosynthesis, could lead to camalexin deficiency.

As far as the radiolabeling studies are concerned, it
may be unfair to invoke only regulatory mutations to explain
the absence of obvious intermediates in the pad mutants. A
number of other reasons may be involved, not the least of
which would be inexperience on the part of the author. One
possible reason is that the solvent system used in the thin-
layer chromatography did not separate bands well enough to
allow identification of inducible bands. In light of the
labeling experiments described in Chapter 4, in which
camalexin was poorly separated from another compound, this
explanation is not unreasonable. A single experiment with
.ped3 also supports this explanation: an inducible compound
was detected in a large-scale extract (60 leaves) of pad3
extracted 54 hours after inoculation. On a TLC plate
developed in a more polar solvent (chloroformrmethanol 9:1,

v/v; followed by ethyl acetate—methanol 24:1, v/v) the

73

compound appeared as a fluorescent blue band just below
camalexin. It was not characterized. Further studies of
‘pad3 may therefore yield some useful results.

Another possibility is that the inducible compounds
were in the aqueous and not the organic phase, and that they
were overlooked because only the organic phase was studied.
The need to study the correct phase was demonstrated by
Jimenez et al. (1997), who found camalexin and other
phytoalexins in Capsslla bursejpastoris (shepherd’s purse),
but only in the basic fraction of ethyl acetate extracts.

Considering the variable kinetics of camalexin
accumulation, it is also possible that some of the
information on the autoradiograms was misinterpreted. The

band with an R.f of 0.13 that was seen in Figures 7 and 8, and

whose intensity varied between the two experiments
represented in those figures, may have been a constitutive
intermediate. In one experiment (Figure 7), camalexin
accumulated slowly in padl. Consequently, the majority of
the compound was not converted into camalexin when the leaves
were extracted. In the second experiment (Figure 8),
camalexin accumulated quickly, and the majority of the
compound was converted into camalexin at the time of
extraction. This could explain why the camalexin band is
darker in Figure 8 than in Figure 7. A similar argument can
be made for the wild-type leaves in these figures.

Given all the possible explanations for the lack of

biosynthetic intermediates found in the pad mutants, it may

74

have been a mistake not to pursue the search with revised
tactics such as different TLC solvents, HPLC, or studies of
different phases. All of these tactics would be worth
attempting in future work. It is possible that the inducible
band found in pad 3, if studied further, will prove to be a
biosynthetic intermediate. Considering the ambiguous results
of the mutant screen and radiolabeling studies, however, the
most likely explanation for the lack of intermediates was
that the mutants were not biosynthetic but regulatory
mutants. Possibly the inducible band in pad3 had nothing to
do with camalexin biosynthesis. It seemed time to abandon
the search for biosynthetic intermediates and mutants, and to
examine more closely the differences in camalexin production
between fungal- and bacterial-inoculated plants, the results

of which are described in Chapter 3.

75

REFERENCES

Bowling SA, Guo A, Cao H, Gordon AS, Klessig DF, Dong X.
1994. A mutation in Arabidopsis that leads to constitutive
expression of systemic acquired resistance. Plant Cell 6:
1845-1857.

Bu’Lock JD. 1965. The Biosynthesis of Natural Products.
London: McGraw-Hill.

Conn KL, Tewari JP, Dahyia JS. 1988. Resistance to
Alternarie brassicae and phytoalexin—elicitation in rapeseed
and other crucifers. Plant Science 56: 21-25.

Cruickshank IAM, Perrin DR. 1961. Studies on phytoalexins.
III. The isolation, assay, and general properties of a
phytoalexin from Pisum sativum L. Australian Journal of
Biological Sciences 14:336-348.

Ebel J. 1986. Phytoalexin synthesis: the biochemical
analysis of the induction process. Annual Review of
Phytopathology 24:235—264.

Glazebrook J, Ausubel FM. 1994. Isolation of phytoalexin-
dsficient mutants of Arabidopsis thaliana and
characterization of their interactions with bacterial
pathogens. Proceedings of the National Academy of Sciences
USA 91: 8955-8959.

Glazebrook J, Zook M, Mert FM, Kagan I, Rogers EE, Crute IR,
Holub EB, Hammerschmidt R, Ausubel FM. 1997. Phytoalexin-
deficient mutants of Arabidopsis reveal that PAD4 encodes a
regulatory factor and that four PAD’genes contribute to downy
mildew resistance. Genetics 146: 381-392.

Harborne JB. 1973. Phytochemicel methods. London: Chapman
and Hall, 1-87.

Harborne JB. 1988. Introduction to Ebological Biochemistry,
3rd edition. London: Academic Press, 315-327.

Jejelowo 0A, Conn KL, Tewari JP. 1991. Relationship between
conidial concentration, germling growth and phytoalexin
production by camslina sativa leaves inoculated with
Alterneria brassicae. Mycologicel Research 95: 928-934.

Jimenez LD, Ayer WA, Tewari JP. 1997. Phytoalexins produced
in the leaves of Capsslla bursa-pastoris (shepherd’s purse).
Phytoprotsction 78: 99—103.

Kuc J. 1995. Phytoalexins, stress metabolism, and disease

76

resistance in plants. Annual Review of Phytopathology 33:
275-97.

Scott IA, Reichardt PB, Slaytor MB, Sweeny JG. 1971.
Mechanisms of indole alkaloid biosynthesis. Recognition of
intermediacy and sequence by short-term incubation.
Bioorgenic Chemistry 1: 157-173.

Smith OH, Yanofsky C. 1963. Intermediates in the
biosynthesis of tryptophan. Methods of Ehzymology 6: 590-
597.

Swain T. 1965. Methods used in the study of biosynthesis.
In: Pridham JB, Swain T, sds., Biosynthetic Pathways in
Higher Plants. London: Academic Press, 9-35.

Taiz L, Zeigsr E. 1991. Plant Physiology. Redwood City,
California: Benjamin/Cummings, 226.

Tsuji J, Jackson EP, Gage DA, Hammerschmidt R, Somervills SC.
1992. Phytoalexin accumulation in Arabidopsis theliene
during the hypersensitive reaction to Pseudomonas syringes
pv. syringes. Plant Physiology 98: 1304-1309.

Tsuji J, Zook M, Somervills SC, Last RL, Hammerschmidt R.
1993. Evidence that tryptophan is not a direct biosynthetic
intermediate of camalexin in Arabidopsis thaliana.
Physiological and.Molecular Plant Pathology 43: 221-229.

Vernooij B, Friedrich L, Morse A, Reist R, Kolditz-Jawhar R,
Ward E, Uknes S, Kessmann H, Ryals J. 1994. Salicylic acid
is not the translocatsd signal responsible for inducing
systemic acquired resistance but is required for signal
transduction. Plant Cell 6: 959-965.

Wright AD, Sampson B, Neuffer MG, Michalczuk L, Slovin JP,
Cohen JD. 1991. Indole-3-acetic acid biosynthesis in the
mutant maize orange pericerp, a tryptophan auxotroph.
Science 254: 998-1000.

Zhao Jianmin, Last RL. 1996. Coordinate regulation of the
tryptophan biosynthetic pathway and indolic phytoalexin
accumulation in Arabidopsis. Plant Cell 8: 2235-2244.

Zhao J, Williams CC, Last RL. 1998. Induction of
Arabidopsis tryptophan pathway enzymes and camalexin by amino
acid starvation, oxidative stress, and an abiotic elicitor.
Plant Cell 10: 359-370.

Zhou N, Tootle TL, Tsui F, Klessig DF, Glazebrook J. 1998.

PAD4 Functions upstream from salicylic acid to control
defense responses in Arabidopsis. Plant Cell 10: 1021-1030.

77

Chapter 3. Comparison of Patterns of Camalexin
Accumulation in Response to the Fungal Pathogen
Cochliobolus carbonum and the Bacterial Pathogens
Pseudomonas syringes pv. maculicola and Pseudomonas

syringes pv. syringes.

Introduction

The studies described in Chapter 2 suggested that some
camalexin-deficient mutants of Arabidopsis were not
biosynthetic mutants, but regulatory mutants that produced
different amounts of camalexin in response to cochliobolus
carbonum and Pseudomonas syringes pv. maculicola (Psm) strain
ES4326. It seemed possible that the differences in camalexin
production were due to different responses to fungal and
bacterial infection. Fungi and bacteria are distinct
organisms with distinct modes of plant infection. Whereas
fungi can infect plant cells by direct penetration as well as
by entry through wounds or natural openings, bacteria enter
cells only through wounds or natural openings (Agrios, 1997).
Bacteria are also much smaller than fungi and have chemically
different cell membranes. Therefore, it would be expected
that the signals or elicitors produced by fungi and bacteria
would differ and that in the infected plant, different genes
would be required to recognize and respond to the various
signals. The recent studies on the ped4 Arabidopsis mutant

support this model. As described in Chapter 1, the PAD4 gene

78

affects camalexin biosynthesis by controlling production of
salicylic acid, which is in some way necessary for initiating
camalexin biosynthesis (Zhou et al., 1998). Because more
than one signal transduction pathway appears to be present,
some elicitors may activate a signal transduction pathway
that does not go via PAD4 when triggering salicylic acid
production (Glazebrook et al., 1997; Zhou et al., 1998).

To determine whether some of the mutants might be
signal transduction mutants, camalexin accumulation over a 3-
day period was compared in fungal- and bacterial-inoculated
wild-type and pad2 plants. As it was not certain if pad2 was
a biosynthetic or a regulatory mutant, it was possible that
closer study of its patterns of camalexin accumulation in
response to different inducers would lead to finding
biosynthetic intermediates in large-scale extracts. These
experiments thus had the potential to provide information on
both the biology and the chemistry of camalexin biosynthesis.

Optimum conditions for measuring camalexin levels in Psm
ES4326-inoculated leaves, such as inoculum concentration and
time of maximum accumulation of camalexin, have been reported
(Glazebrook and Ausubel, 1994). As these conditions were not
previously established for C. carbonum, this part of the
project began by studying the kinetics of C. carbonum-
elicitsd camalexin accumulation in order to determine the
time of maximum.accumulation. Optimal spore inoculum
concentration was also determined. Camalexin accumulation

was then compared in wild-type and pad2 leaves inoculated

79

with C. carbonum or Psm ES4326. To determine if the patterns
of accumulation reflected differences between responses to
fungi and bacteria in general, or between responses to
compatible and incompatible pathogens (Psm ES4326 being a
compatible pathogen, and C. carbonum being an incompatible
pathogen), camalexin accumulation was also compared in wild-
type and pad2 plants inoculated with C. carbonum or the
incompatible bacterial pathogen Pseudomonas syringae pv.

syringes (Pss).

Materials and Methods

Reagents and chemicals .

Camalexin was synthesized according to the method of
Ayer et al. (1991) in the laboratory of Dr. W. Reusch at
Michigan State University. Some synthetic camalexin was also
generously provided by Dr. Alois Furstner of the Max-Planck-
Institut far Kohlenforschung (MUlheim/Ruhr, Germany).
Indole-3-carboxaldehyde was purchased from Aldrich or Sigma,
and anthranilic acid and indole-3-carboxylic acid were
purchased from Aldrich. Solvents used for high-performance
liquid chromatography (HPLC) were of HPLC grade. All other

chemicals were of reagent grade or better.

Synthesis and purification of camalexin.
The yield of camalexin in the synthesis (approximately

0.5 %) was considerably lower than the 68-76 % yield reported

80

by Ayer et al. (1991). When the reaction mixture was checked
by thin-layer chromatography (TLC) in chloroform-methanol

(9:1, v/v), the major band had an Rf value higher than that of

camalexin, although the latter was also present. The major
band, although never characterized, was probably
acetylcamalexin, which Ayer et al.(1991) found to comprise
about 6 % of the products. The predominance of
acetylcamalexin was probably due to failure to quench the
reaction mixture with water prior to work-up with ethyl
acetate (Ayer et al., 1991). Another factor may have been
the water bath temperatures of 45-50 °C used for rotary
evaporation during work-up. At such temperatures, excess
indole may have reacted with ethyl acetate to produce
acetylcamalexin (J. Kagan, pers. comm.). (When this
synthesis was attempted by other workers, acetylcamalexin was
also the major product, and the yield of camalexin was higher
when work-up was done in ether without heating [J. Kagan,
pers. comm.).)

The synthetic camalexin was purified by preparative TLC
on Fisherbrand 1000 um silica plates in chloroform-methanol
(9:1 or 19:1, v/v). The camalexin band was eluted with
acetone or ethyl acetate. The solvent was removed under
vacuum at 45-50 °C. The residue was transferred to a vial,
dried, and weighed. For TLC, the residue was redissolved in
methanol and used without further purification. For high—
performance liquid chromatography (HPLC), the residue was

redissolved in isopropanol-hexans (7:93, v/v), combined with

81

some plant-produced camalexin, and purified on an Alltech
Econosphsre silica column (pore size: 5 um, column
dimensions: 4.6 mm inner diameter and 250 mm length) attached
to a Waters HPLC (pump model 501, injector model U6K, UV
detector model 486). A wavelength of 215 nm was used for
detection because camalexin has an absorbence maximum at this
wavelength (Hammerschmidt st al., 1993). In a mobile phase
of isopropanol-hexane (7:93, v/v) with flow rate of l
ml/minute, camalexin had a retention time of 9 to 9.5 minutes
(retention times varied on different days, due to slight
variations in solvent composition). Fractions corresponding
to camalexin peaks were collected. The solvent was removed
under vacuum. The residue was weighed, redissolved in a
known volume of isopropanol-hexane (7:93, v/v), and aliquoted
into vials. These aliquots were used for standard curves in

HPLC and fluorimetric analysis.

Plant material and growth conditions.

Wild-type and pad2.Arabidopsis plants were grown as
described in Chapter 2. For wild-type plants, the glabrous
strain of the Columbia ecotype (Col-g1) was used, since this
strain had also been used for the radiolabeling studies
(Chapter 2). For the time courses done with Pss, the Col-0

ecotype was used.

82

Fungal and bacterial cultures.

C. carbonum was grown as described in Chapter 2. For
the time courses comparing camalexin accumulation in C.
cerbonum- and and Pss-inoculated leaves, the fungus was grown
on V-8 agar of a slightly different composition (per liter:
160 ml V-8 juice, 1 g calcium carbonate, 14 g agar). Psm
ES4326, a strain pathogenic on many ecotypes of Arabidopsis
(Dong et al., 1991), was provided by Dr. Glazebrook
(University of Maryland, College Park). The culture was
transferred from a plate or a slant to a plate of King’s B
agar (per liter: 1.5 g dibasic potassium phosphate, 1.5 g
magnesium sulfate hsptahydrate, 20 g peptone, 10 ml glycerol,
15 g agar), an indicator of the presence of pseudomonads
because of the production of fluorescent pigments. Cells
from this plate were used to inoculate a 50-ml flask
containing about 15 ml of LB (Luria-Bsrtani medium: 10 g
tryptone, 5 g yeast extract, 10 g sodium chloride per liter),
which was then incubated at room temperature on a shaker
(Lab—Line Instruments, Model 3590) at about 120 rpm. Pss
strain D20 (Tsuji et al., 1992) was grown on LB agar amended
with nalidixic acid, and cells from these plates were used to
inoculate 50-ml flasks containing about 15 ml of LB broth.
Fourteen-hour-old broth cultures of Psm ES4326 and P55 D20
were centrifuged on a benchtop centrifuge (International
Chemical Centrifuge, Schaar & Co.) at about two-thirds of the
maximum.spesd. Cells were washed 2 or 3 times with 10 mM

magnesium sulfate. After resuspension of the washed cells in

83

10 mM magnesium sulfate, the optical density at 600 nm (ODmm)

was measured on a Zeiss spectrophotometer. Cells were
diluted (again in 10 mM magnesium sulfate) so that the final

concentration would correspond to an OD600 of 0.02. To

determine the number of colony-forming units per ml of
inoculum (cfu/ml), a portion of the final cell suspension was

serially diluted, and dilutions were plated onto LB agar.

The number of cfu/ml ranged from 1.1x106 to 1.2x107.

Leaf inoculation .

For the time courses establishing approximate kinetics
of camalexin accumulation, one fungal-inoculated sample of 30
or 45 leaves and one water-inoculated control, with an equal
number of leaves, was prepared for each timepoint. For
subsequent experiments, only 15-20 leaves per sample were
inoculated, and all samples were prepared in duplicate or
triplicate. In these latter experiments, water controls were
extracted only at 72 hours, or at 24 and 72 hours. Three—
to four-week-old leaves were excised and in covered Petri
dishes, as described in Chapter 2. An effort was made to use
leaves from several pots of plants for each timepoint. and to
ensure that a fungal-inoculated sample and the corresponding
water-inoculated control contained leaves from the same
plants. Each leaf was inoculated with 0.1 ml of water or a
suspension of 8- to 14-day old C. carbonum spores, prepared

as described in Chapter 2 (see Materials and Methods). Spore

84

inoculum concentration (determined with a Bausch & Lomb
hemacytomster) ranged from 6x105-7x105 spores/ml in the

studies of the approximate kinetics of camalexin

accumulation. For dose-response experiments, spore
concentrations of 0, 2x104, 2x105, and 2x106 spores/ml were

prepared by serial dilution of a suspension of the highest
concentration. For subsequent experiments, the spore
concentration was 1.4-2.0 x10s spores/ml. Following
inoculation, leaves were incubated at room temperature in the
covered Petri dishes.

For bacterial inoculation, the cells, grown and diluted
as described above, were infiltrated into intact leaves with
the blunt end of a 3 ml plastic syringe. Some leaves were
inoculated with 10 mM magnesium sulfate as controls. For
each sample, 15-20 leaves were inoculated. Plants were
covered with wet sheets of plastic or with plastic bags about
12 hours prior to inoculation, to cause the stomata to open

and make the leaves easier to infiltrate.

Extractions.

On fungal-inoculated leaves (15-45 per sample), droplets
of water or inoculum were collected with a Pasteur pipet and
extracted by vortexing in a test tube 3 times with equal
volumes of ethyl acetate (the volume of the aqueous phase was
increased by adding about 50 % the original volume of water,

to reduce the percentage of sample lost to flask walls). The

85

organic phases were pooled, evaporated to dryness at 40 'C on
a Buchi rotary evaporator, and redissolved in 1 ml methanol
(two 0.5-ml aliquots) to transfer to lO-ml glass tubes. The
tubes were capped with aluminum foil, sealed with Parafilm,
and then stored at -20 °C.

The leaves were extracted by a method similar to that
used by Hammerschmidt et al. (1993). They were boiled for 20
minutes in 80 % methanol (about 1.5 ml per leaf, which seemed
enough to ensure an efficient extraction without generating
too much solvent waste), and then stored at 4'C in foil-
capped flasks sealed with Parafilm until a convenient time
for extraction with chloroform. In the time courses
comparing camalexin accumulation in response to Psm ES4326
and C. carbonum, or in response to Pss D20 and C. carbonum,
fungal-inoculated leaves and inoculum droplets were combined
and boiled together in 80 % methanol. The intention was to
make the extraction of these samples as similar as possible
to the extraction of the bacterial-inoculated leaves which
did not contain inoculum droplets.

The methanol extract was concentrated on the rotary
evaporator at 40-45 'C until the sample looked turbid (an
indication that the majority of the methanol was removed),
resuspended in water so that the final volume was 1.5 times
the volume obtained by concentration, and then extracted in a
separatory funnel 3 times with an equal volume of chloroform.
When the approximate kinetics of camalexin accumulation were

being determined, the aqueous phases were then extracted

86

three times with equal volumes of ethyl acetate to look for
residual camalexin and potential camalexin precursors.

After drying with anhydrous sodium sulfate, the organic
phases were evaporated to dryness on the rotary evaporator,
redissolved in 1-1.5 ml of methanol, and transferred to 10—ml
tubes and stored at -20'C, as described for droplet extracts.

This extraction procedure was time-consuming, and it was
not possible to process all samples in 1—2 days. Therefore,
an effort was made to standardize the time frame within which
samples were extracted. For example, in the time courses
comparing camalexin accumulation in C. carbonum- and Psm-
inoculated leaves, fungal- and bacterial-inoculated samples
from a given timepoint and replicate were extracted with
chloroform on the same day. Also, the period between the
initial methanol extraction and the work-up was kept roughly
the same for all timepoints, so that if replicate “B" of the
24-hour samples was extracted 3 days after boiling in
methanol, replicate “B" of the 48-hour samples was also

extracted 3 days after boiling in methanol.

Determination of the presence of camalexin and other
inducible compounds.

Camalexin was separated by TLC as described in Chapter
2. Extracts were dissolved in 30-45 ul of methanol.
Usually, 20 ul of the extract (two 10-ul aliquots with a
glass capillary micropipet) were loaded. For the dose-

response experiments, the entire sample was loaded, and for

87

heavily-pigmented leaf tissue extracts, 10 ul were loaded. A
camalexin standard (2.5-5 mg) was loaded onto each plate.
For TLC of extracts from the preliminary time courses of
camalexin accumulation (Tables 5 and 6, Figures 12A and 12B),
standards of indole-3-carboxaldehyde (0.07 mg/ml in
methanol), indole-3-carboxylic acid (0.67 mg/ml in methanol),
and anthranilic acid (0.5 or 0.2 mg/ml in methanol) were
loaded as well. Extracts from those experiments consisted of
more leaf tissue than extracts from previous experiments
(fivefold increase) or later ones (two- to threefold
increase). They consequently had the potential to contain
detectable amounts of camalexin intermediates. Therefore, it
was of interest to see whether any bands comigrated with
known or hypothetical intermediates.

Plates were usually developed in chloroform-methanol
(9:1 or 19:1, v/v), sometimes preceded by a development in
chloroform to improve separation from pigments. The large—
scale extracts were separated in several different TLC
solvent systems in order to find one that separated camalexin
well from pigments and one that resolved polar putative
intermediates.

Camalexin was visualized under long-wave and short-wave

ultraviolet (UV) light as a fluorescent purple band with an Rf

of 0.3-0.5 in chloroform-methanol (19:1, v/v) and 0.5-0.7 in
chloroform-methanol (9:1, v/v). Indole-B-carboxaldshyds was
visualized under short-wave UV light as a dark, light-

absorbing band that was best seen by shining a hand-held UV

88

lamp onto the front of the TLC plate and viewing the plate
from the back while in a dark room. Indole-3—carboxylic acid
appeared under short-wave UV light as a fluorescent purple
band, and anthranilic acid fluoresced purple under both

short-wave and long-wave UV light. The Rf'values of all

standards used are summarized for different solvent systems

in Table 5. Plates were photographed as described in Chapter

5.

Preparation of samples for HPLC analysis.
The camalexin bands on the TLC plates were eluted with 3
ml of ethyl acetate, and the eluates were dried under

nitrogen at 40-50 °C. To sluts samples, bands at the R.f of

the camalexin standard were scraped with a spatula onto
weighing paper and transferred to 20-ml glass scintillation
vials, which were stored at -20 °C if they could not be
eluted within several hours of scraping. Samples were stored
in this manner for only 1-3 days before being eluted, as
earlier attempts to scraps and quantitate samples had
demonstrated that after 2 months, the majority of the
camalexin on scraped silica gel degraded. In contrast, about
80 % of the camalexin remained on TLC plates stored for the
same length of time. The sample was transferred to a 15-ml
sintered glass funnel (fine frit) set in a 250-ml filtering
flask. The vial was rinsed with 1.5 ml of ethyl acetate; and

the rinsate was then added to the funnel. After the flask

89

had been gently swirled to mix silica and solvent, vacuum was
applied, and the solvent was collected in a 10—ml tube set
directly underneath the funnel inside the sidearm flask. The
vial was rinsed again with 1 ml of ethyl acetate, which was
added to the funnel and collected in the same tube after
being swirled in the funnel. The funnel was rinsed with 0.5
ml of ethyl acetate, to retrieve any camalexin that might
have been splashed near the upper portion of the funnel
during the elution, and this rinsate was also collected. The
tube of eluate was dried under nitrogen (Meyer N-Evap) at 40-
50 °C. The sintered glass funnel was rinsed with 7-15 ml of
ethyl acetate between samples.

After drying, samples were sealed and stored at -20 °C
until analysis. They were quite stable under those
conditions: Samples analyzed once and then dried and stored
yielded very similar results in HPLC analysis 2 months later

(8-15 % decrease in ODE”).

Quantitation of camalexin by HPLC.

The HPLC analysis was done under the conditions
described for camalexin purification. Samples were dissolved
in 100-200 ul of mobile phase, and 10-20 ul were injected
with a Hamilton 25-ul syringe. The syringe was rinsed with
methanol and mobile phase between injections, and mobile
phase was injected between each sample to ensure that no
residue from the previous sample remained. Because the

retention time of camalexin varied between 9 and 9.5 minutes

90

on different days, due to slight variations in mobile phase
composition, a standard was injected at the start of each
analysis to verify the retention time. The putative
camalexin peak from the standard, and from 1-3 injected
samples, were collected, concentrated, and co-spotted with a
standard on 4x5 cm TLC plates to verify the identity of the
peak. Samples usually were injected only once. The
injection error was checked on 2 occasions. For samples
dissolved in 200 ul or more of mobile phase, peak heights
differed by 0-9 % between 2 injections. For samples
dissolved in 100 pl of mobile phase, peak heights differed by
0-27 % between 2 injections. The greater variation in the
second case may be due to rapid evaporation of smaller
volumes, since peak heights were always greater on the second
injection.

Camalexin was quantitated by the equation obtained from
the standard curve in Figure 11A. For the time courses
comparing camalexin accumulation in response to C. carbonum
and P.syringas pv. syringes, camalexin was quantitated by the
equation obtained from a standard curve run 2 years later to
correct for decreased intensity of the detector bulb (Figure

118).

91

.0078341 + 0.5696:
0.992

215 nm

 

 

abs

 

' t
0 1

micrograms camalexin

   

y - 0.0075563 + 0.33445:
R‘2 I 0.998

215 nm

abs.
0
O
N

l

 

 

 

0 1 2 3
micrograms camalexin

Figure 11. Standard curve of camalexin concentration versus

absorbence at 215 nm, 12/15/95 (A) and 1/21/98 (B). Data
represent means of duplicate HPLC injections of dilutions of

pure synthetic camalexin in mobile phase.

92

Results

Determination of a standard curve of camalexin
concentration versus absorbance .
The standard curves obtained for camalexin concentration

versus OD215 are shown in Figure 11. The “known" amount of

camalexin in the standards was based on an aliquot from 600
ug of HPLC-purified camalexin. Since the analytical balance
used to weigh the standard was accurate only to 4 decimal
places (0.0001 g, or 100 pg) the accuracy of the
concentrations in the standard curve is questionable, and
numerical discrepancies with the results of other workers may
be due partly to initial inaccuracy in weighing the standard.
However, the precision of the standard curves was
satisfactory, judging by comparison of Figure 11A to a
standard curve done 2 months later with a separate aliquot of
that standard (y=0.03074+0.48916x; data not shown). As
mentioned in the Materials and Methods, the equation obtained
from Figure 11A was used to calculate camalexin concentration
for most of the experiments described in this chapter, and
the equation obtained from Figure 118 was used for the time

courses of camalexin accumulation in Pss D20 and C. carbonum.

Determination of optimal separation of camalexin and
putative intermediates by TLC.
The time courses depicted in Figures 12A and 12B were

done, in part, to look for biosynthetic intermediates. A few

93

inducible bands were seen on TLC plates in various solvent

systems. The Rf values of those bands in some of the solvent

systems tried are listed in Table 4. Most were not studied

further. Attempts to characterize one of these (Rf=0.4-0.5 in

chloroformrmethanol 9:1, [v/v]) led to the experiments
described in Chapter 4.

The Rf'values of the standards in those solvent systems

are listed in Table 5. These should be taken only as an

approximation. Rf‘values varied with the extent to which the

plate was activated prior to use, usually being higher on
less-activated plates. They could vary on a single plate,
with camalexin bands running higher at the edges than at the
center. The type of silica plate (glass- or plastic-backed)

could cause the relative Rf'values of standards to be reversed

with respect to each other or increased (see also Chapter 4).

The amount of standard loaded also made a difference. Rf

values were usually higher when more standard was loaded.

Finally, the size of the TLC plate affected R.f calculations:

those obtained from small (5x5 cm) TLC plates were
particularly rough approximations, since the error in
distance measurement increased as the distance measured
decreased.

Separation of camalexin from pigments was fairly good in

ethyl acetate-hexane (1:1, v/v), but the best separation was

94

Table 4. Rf values and descriptions of inducible bands in

different TLC solvents. Abbreviations: SWUV=short~wave
ultraviolet light; LWUV=long-wave ultraviolet light, h=hours;
CHC13=chloroform, MeOH=methanol, HOAc=acetic acid, EtOAc=ethyl

acetate .

Rf values and descriptions of inducible

   
        
     
     
     
 

TLC solvent

  

 

(V/V bands

composition)

CHCl3-MeOH 0.15 at 12 and 24 hours post-
'9;1 inoculation

fluorescent purple under LWUV and SWUV

  

0.1 at 24, 48, and
inoculation
fluorescent purple under LWUV

 
           
 

72 hours post-

 

0.4-0.5 at 24, 48, and 72 hours post—
inoculation
fluorescent purple under LWUV

  
    
     
      
   
       
     
    
       
 

 

CHC13 0.3-0.4 at 24, 48, and 72 hours post-

followed by inoculation
iCHCl3+MeOH fluorescent purple under LWUV and SWUV

[19:1

  

0.2 at 24, 48, and 72 hours post-
inoculation
fluorescent purple under LWUV

 

‘toluene 0.43 at 12 hours post-inoculation
lfollowed by appearance not recorded; probably
32 fluorescent purple

(developments
Pin CHCl3-H0Ac
'9:1 (1

  

 

95

Table 5. Rf values of camalexin, indole—3-carboxaldehyde,

indole-3-carboxylic acid, and anthranilic acid developed on
silica TLC plates in various solvent systems. Some Rf'values

obtained from experiments in Chapter 4 are included to make
the table more useful. Rf’s are listed with only 1

significant digit if they varied greatly among experiments.

Abbreviations are as in Table 4.

ITLC solvent
(v/v
composition)

Rf:
camalexin

Rf:
indole-3-
carbox—
aldehyde

Rf: indole-
3-carbox-
ylic acid

 

CHC l 3 -MeOH
9:1

0.5-0.6
(just
below
camalex-
in)

0.09-0.2

 

CHC13

ifollowed by
'CHC139MEOH

.19:1

0.3-0.5
(just
below
camalex-
in)

0.04-0.05

 

lCHCl3-H0Ac
:9:1, 2
(developments
(1
lexperiment)

0.45

 

1CHC13-EtOAc-

0 .2-0 . 513:9

 

j CHCl3-MeOH
(19:1
*followed by
CHCl3-EtOAC-
iHOAc
[35:55:10 (1
(experiment)

0.58a and
0.67a
(possibly
protonated
and un-
protonated
forms)

 

FEtOAc-MeOH

(24:1 (1
(experiment)

 

 

 

 

 

aglass-backed 20x20 cm.plate used; hplastic-backed 5x5cm.plate
used; cstandards well-separated from each other on a plate

96

‘TLC solvent : Rf: : indole-
!(V/V . . camalexin indole-3- 3—
[comp031tion) carbox- carboxylic
' aldehyde acid

 

EtOAc-hexane . 0.27 0.07
1:1 (1
experiment)

 

ICHCl3-EtOAc- 0.33b 0.42b
iHOAc
(35:55:10
’followed by
i CHC13 -HOAC
9:1 (1
experiment)

 

 

 

 

 

aglass-backed 20x20 cm.plate used; #plastic-backed 5x5cm.plate
used; con a single plate, standards well-separated from each
other

97

achieved by development in chloroform followed by chloroform-
methanol (19:1, v/v). This solvent was used to separate leaf
extracts in all subsequent experiments. Occasionally, if
camalexin and pigments were poorly separated, the camalexin
band was eluted and developed again by TLC.

For ethyl acetate extracts of droplets and leaf
tissue, which contained more polar compounds than the
chloroform extracts, fairly good separations were achieved in
chloroform-acetic acid (9:1, v/v) and chloroform—ethyl
acetate-acetic acid (35:55:10, v/v). In both solvent
systems, the relative polarities of camalexin, indole-3-
carboxaldehyde, and indole-B-carboxylic acid were reversed
from what was observed in chloroformrmethanol (19:1, v/v; see
Table 5). All 3 standards were well separated. Ethyl
acetate-methanol (24:1, v/v) resolved the standards poorly:

indole-3-carboxaldehyde and camalexin had nearly-identical Rf

values. Chloroform—ethyl acetate-formic acid (35:55:10, v/v)
yielded many bands, but as this solvent system can break
indole rings (M. Zook, pers. comm.), those bands may have
represented solvent-induced artefacts.

The most-commonly used solvent in the subsequent
experiments was chloroformrmethanol (9:1, v/v) because the
migration of camalexin and the other standards was similar to
what was seen for leaf tissue extracts in chloroformrmethanol

(19:1, v/v).

98

Preliminary time courses of camalexin accumulation in
wild-type Arabidopsis.

The other purpose of the time courses just described was
to establish the optimum time of camalexin accumulation in C.
carbonum-inoculated leaves, in order to know when to extract
leaves in the future studies of camalexin accumulation in C.
carbonum- and Psm-inoculated leaves. Figure 12, and tables 6
and 7, demonstrate that the kinetics of camalexin
accumulation were quite variable. In one experiment (Figure
12A), camalexin concentrations increased over 100 % between
24 and 48 hours after inoculation, and in the second
experiment (Figure 128), concentrations reached a maximum at
24 hours and then decreased about 33 % by 72 hours. In both
experiments, however, camalexin concentrations at all time
points were high enough to be easily detected by HPLC. About
5 % to 10 % of a sample eluted from a TLC plate (1.25 % to 5
% of the original extract) gave large peaks at a sensitivity
of 0.25 absorbance units full scale (AUFS). Therefore, any
time from 12-72 hours post-inoculation was suitable to
extract leaves in a time course. Since camalexin
concentrations in Psm ES4326-inoculated leaves peaked at
about 40 hours and remained fairly high through 72 hours
(Glazebrook and Ausubel, 1994), it was possible to compare
the kinetics of camalexin accumulation within a time frame of
24 to 72 hours. Also, for about the same amount of work
required to prepare a single 45-leaf sample, it would be

possible to prepare three lS-leaf samples, which would yield

99

Table 6.

Time course of camalexin accumulation in wild-type

(wt) leaves inoculated with Cochliobolus carbonum (inoc) or

water (ctrl).

8/25/95-8/27/95.

Leaves of pad2 were extracted

at one timepoint to compare camalexin concentrations to those
The inoculum concentration was 6.9x105
spores/ml, and 45 leaves per sample were used for wild—type

of wild-type.

 

 

 

 

 

 

 

 

 

samples. The pad2 samples consisted of 20 leaves each.
Standard errors are not shown because the samples were not
replicated.
[hours after ug pg total ug ug
{inoculation camalexin camalexin camalexin camalexin
in leaves in in sample per leaf
droplets
12-wt ctrl 0 0.54 0.54 0.012
12-wt inoc 14.0 21.9 35.9 0.798
24-wt ctrl 0 0.079 0.079 0.0018
24—wt inoc 11.6 24.6 36.2 0.805
48-wt ctrl 1.8 0.81 2.6 0.058
48-wt inoc 24.5 47.0 71.5 1.59
12-pad2 0.19 0.064 0.25 0.012 ’
ictrl
112-pad2 0.870 1.39 2.26 0.113
inoc

 

 

100

 

 

Table 7. Time course of camalexin accumulation in wild—type
leaves inoculated with C. carbonum, 10/17/95-10/20/95.
Abbreviations are as for Table 6. Leaves were inoculated with
6x105 spores/ml, and 30 leaves per sample were used.

ug ug
camalexin camalexin camalexin camalexin
in leaves in per leaf
droplets

24-wt-ctrl 0.26
.24 -wt-inoc . 23.0

 

.0088
.22
.0093
.05
.012

 

 

36- wt-ctrl 0.28

 

36- wt-inoc . 18.9
‘48- wt-ctrl . 0.23
48- wt-inoc . 13.9
72-wt-ctrl . 0
9.41 . .

 

 

 

 

OOHOHOI—‘O

 

 

 

 

 

101

2' A. 8/95

__g__ control

1. .__.__ inoculated

camalexin (no/leaf)

 

01_a—i~ I a. I:—...====1-:===::E..l
10 20 30 40 50

hours after
inoculation

 

2' B. 10/95

——o—— control

e~“\“_____.~‘--‘~“‘* .__.__ inoculated

o-W
20 30 40 50 60 70 80
hours after
inoculation

camalexin (ug/leaf)
H

 

Figure 12. Time course of camalexin accumulation in
wild-type leaves inoculated with Cochliobolus
carbonum or water (control), 8/25/95-8/27/95 (A)
and 10/17/95-10/20/95 (B).

102

measurable amounts of camalexin and provide the statistical
reliability of replication.

Some camalexin was present in the controls (Tables 6 and
7). It was barely detectable in 45-leaf samples, which
explains why no camalexin had been detected in the 6- to 15-
leaf samples used in the past, or by fluorimetric analysis of
pairs of leaves (Table 3). It may be that inoculation with
water creates a certain amount of stress, as no camalexin was
detected in one experiment in which 60 non-inoculated leaves
were extracted 54 hours after being excised (data not shown).

Interestingly, the amount of camalexin in spore droplets
was sometimes higher than the amount in leaves (Tables 6 and
7). This result was noted again in the dose-response

experiments described in the next section.

Effects of C. carbonum inoculum concentration on

camalexin production (dose-response studies).
.At 2x104 spores of C. carbonum per ml, little or no

camalexin was produced (Tables 8 and 9, Figures 13 and 14).

The maximum amount of camalexin was produced following

inoculation with 2x10S spores/ml in one experiment (Figure 13,

Table 8) and with 2x106 spores/ml in the other experiment

(Figure 14, Table 9). The yields of camalexin at both
inoculum.concentrations were high enough to be detected
easily by HPLC and differed by only 20 %. Apparently,

camalexin production approaches or attains a saturation point

103

Table 8. Effects of spore inoculum concentration on camalexin
Numbers represent means plus standard

production,

11/3/95.

errors of 3 replicates, consisting of 15 leaves each.

inoculum
concentra-
tion
(spores/m1)

camalexin

(pg/leaf) in
leaves

camalexin
(pg/leaf) in
droplets

camalexin

(pg/leaf) in
leaves and
droplets
combined

 

0.0l4i0.004

0.0l4i0.004

 

0.003i0.003

0.019i0.003

0.022i0.005

 

0.24310.003

0.602i0.054

0.845i0.056

 

 

0.56710.018

 

O.116:0.012

104

 

 

O.683i0.029

1.0-

 

 

A
A 0.8-
H I:
3 .
.4 0.6"
\ at
u .
n
V 0.4-
G I
g 0.2“
F. .
a 0.0-
0 0 2x10"4 2x10"5 2x10"6
inoculum.concentration
(spores/ml)
1.
SI
8 0.
p:
B
‘3 0.
ll
'5 0.
F:
a 0.
u
0.

 

 

0 2x10"4 2x10“5 2X10"6

inoculum.concentration

(spores/ma;
Figure 13. Effect of concentration of Cbchliobolus
carbonum spores on camalexin production, 11/3/95 (see
also Table 8). Standard error bars are shown.
Leaves of wild-type Arabidopsis were inoculated with
12-day-old spores at the concentrations indicated on
the horizontal axis, and inoculum droplets and leaves
were collected 40 hours after inoculation and
extracted separately. Graph A displays the total
camalexin in leaves and droplets, and graph B
displays the relative contributions of leaves (L) and
droplets (D) to the total.

105

Table 9. Effects of spore inoculum concentration on camalexin
production, 11/13/95. Numbers represent means plus standard
errors of 3 replicates (15 leaves each), except that only 2
samples with 2.5x104 spores/ml were extracted.

   

 

 

 

 

 

 

 

 

 

         

inoculum camalexin camalexin camalexin
concentra- (pg/leaf) in. (HQ/leaf) in (pg/leaf) in
tion leaves droplets leaves and
(spores/m1) droplets
combined
ll0 (water- 0 0 0
inoculated
control)
2.5x104 0 0.0020:0.0006 0.0020i0.0006
2.5x105 0.411:0.004 0.743:0.324 1.15:0.33
2.5x106 0.896i0.044 0.533i0.338 1.43:0.35

106

 

  

 

 

 

 

 

 

 

;: 2 j
3 A
F:
\ 1
D
3
a -
”g 1
":7
o
C l . l s
0 2x10"4 2x10"5 2x10"6
lnoculum.concentration
(spores/ma)
2.
to; B
0
r4
\
o
3
i 1' L
F:
a D
o
0 u T u .
0 2x10“4 2x10“5 2x10“6

inoculum.concentration
(spores [1111)

Figure 14. Effect of concentration of C. carbonum
spores on camalexin production, 11/13/95 (see also
Table 9). Graphs A and B are labeled as in Figure 13.
Experimental procedures were identical to those
described for Figure 13. Standard error bars are
shown.

107

between 105 and 106 spores/m1 of inoculum. A concentration of

1‘2x105 spores/ml was used in subsequent experiments, as it

provided the most camalexin for the least work. Spores

obtained from 1 or 2 plates of C. carbonum cultures had a

concentration of about 2x105 spores/ml, while obtaining a

concentration of 2x106 spores/ml required using more plates
and often centrifuging to concentrate the spores.

Although the total amount of camalexin produced was
similar in response to 105 and 106 spores/ml, the relative
amount in leaves and inoculum droplets was very different. At
105 spores/ml, 29 % to 36 % of the camalexin was in the
leaves, while 67 % to 71 % was in the droplets (Tables 8 and
9; Figures 13 and 14). At 106 spores/ml, 63 % to 83 % of the
camalexin was in the leaves, and only 17 % to 37 % was in the
droplets. The reason for this inoculum-dependent
partitioning of camalexin was not investigated. As mentioned
in Chapter 2 (see Discussion), it may be that camalexin binds
nonspecifically to spores and that at higher spore

concentrations with many spores available for binding, less

diffuses into the inoculum droplets.

Effects of C. carbonum and Pseudomonas syringae pv.
maculicola (Psm 884326) on camalexin accumulation in
wild-type and pad2 Arabidopsis.

In 3 time courses, Psm-inoculated,pad2 plants

accumulated 13-21 % as much camalexin as Psm—inoculated

108

wild-type plants (Tables 10-12, Figures 15—17). These
results agree fairly well with the results of other workers
that pad2 accumulates about 10 % of wild-type levels of
camalexin (Glazebrook and Ausubel, 1994; Glazebrook et al.,
1997). The different values may reflect differences in
methods of quantitation.

In pad2 leaves inoculated with C. carbonum, maximum
levels of camalexin were 7 % to 12 % of wild-type maximum
levels in 2 experiments (Tables 10 and 11, Figures 15 and
16), indicating that this mutant responded similarly to Psm
ES4326 and C. carbonum. In the third time course, pad2
leaves accumulated 85 % as much camalexin as wild-type leaves
(Table 12, Figure 17). Since the maximum amount produced by
‘pad2 in this case was less than the amount produced by Psm-
inoculated pad2 in the previous time course (Table 11, Figure
16), pad2 did not accumulate unusually high amounts of
camalexin on this occasion. Rather, the wild-type
accumulated unusually low amounts.

In wild-type leaves, camalexin accumulated to higher
concentrations in response to Psm ES4326 than in response to
C. carbonum (Tables 10-12, Figures 15-17). This difference
may reflect a difference in response to compatible and
incompatible pathogens, since a similar trend has been
observed in other plant—pathogen systems (Hahn et al., 1985;
Keen and Kennedy, 1974; Storck and Sacristan, 1994). The
.pad2 leaves also tended to accumulate more camalexin in

response to Psm than in response to C. carbonum, with one

109

Table 10. Time course of camalexin accumulation in wild—type
and pad2 leaves inoculated with C. carbonum or Pseudomonas
syringae pv. maculicola strain ES4326 (Psm ES4326),
12/31/95-1/3/96. Leaves and droplets of the C. carbonum-
inoculated plants were extracted separately, and the eluates
from TLC plates were combined for HPLC analysis. Not all of
the C. carbonum leaf extract was used; therefore, results
for C. carbonum are an underestimate of the total amount of
camalexin in the samples. Results for C. carbonum-inoculated
samples, except for the 72-hour pad2 samples, are from 1
replicate, and so standard errors are not shown. Results for
Psm ES4326—inoculated samples represent means plus standard
errors of 2 replicates. Inoculum concentrations were 1.4x105
spores/ml for C. carbonum, and 5.8x106 cfu/ml for Psm ES4326.
Controls (ctrl) for the C. carbonum treatments consisted of
water-inoculated leaves; controls for the Psm treatments
consisted of leaves infiltrated with 10 mM ma esium sulfate.

 

 

 

 

 

 

ours Camalexin produced in. Camalexin produced
after response to C. carbonum in response to Psm
inocu- (Hg/leaf) ES4326 (pg/leaf)
lation . .
Wl ld- type pad2 w1 ld- type pad2
24 0.320 0.019 0.0992 0.0096
10.0328 i0.0054
48 0.292 0.040 0.637 0.114
\ 10.115 10.039
72 0.206 0.004 0.892 0.072
$0.004 $0.064 $0.022
72-ctrl 0.015 sample lost 0.007 0

 

 

 

 

110

0A} CC

03'

wt
-—+—- pad2
+ wt ctrl

02'

01‘

-
o.o-—r. . . .m

20 30 40 50 60 70 80

camalexin (pg/leaf)

 

 

hours after

 

 

inoculation
A 1'0' Psm
9' 1
‘ ll
,3 0.8-
\ d
u I
3 0.6-
. -G-' wt
‘3 . —0— pad2
g 0.4-
-' 1
a
o ‘ «1— p__;
0.0‘K1 ' 1 ' l ' l

 

20 30 40 50 60 70 80

hours after
inoculation

Figure 15. Time course of camalexin accumulation in
wild-type (wt) and pad2 leaves inoculated with C.
carbonum (Cc, top graph), water (ctrl), or Pseudomonas
syringae pv. maculicola strain ES4326 (Psm, bottom
graph), 12/31/95-1/3/96. No camalexin was detected in
controls for Psm—inoculated samples. Standard error
bars are shown for Psm-inoculated samples. See Table
10 for data.

111

Table 11. Time course of camalexin accumulation in wild-type
and pad2 leaves inoculated with C. carbonum or Psm ES4326,
1/20/96-1/23/96. Leaves and droplets of the C. carbonum-
inoculated plants were combined for extraction. Numbers
represent means plus standard errors of 3 replicates except
for the controls (ctrl), which were prepared as described in
Table 10 and consisted of 1 sample each. Inoculum
concentrations were 2x10S spores/ml for C. carbonum, and
1.1x107 cfu/ml for Psm ES4326.

Camalexin produced in
response to C.
carbonum (Hg/leaf)

response to Pam
inocula- ES4326 (pg/leaf)

tion

 

wild—type

wild-type

“pad2

 

0.002

0

0

 

0.454
$0.009

0.056
$0.001

0.0062
$0.0005

 

0.589

0.597

0.16

$0.098
0.350

$0.078
0.763

$0.03

0.066
$ 0.011

 

 

 

 

 

 

 

112

—*— pad2

 

camalexin ug/leaf)

 

 

 

0 20 40 60 80
hours after
08 inoculation

 

 

 

4:
d
0
p:
\
o
a -——a—— wt
_§ -—*—' pad2
I3
0
I ‘ U U I I
0 20 40 60 80
hours after
inoculation

Figure 16. Time course of camalexin accumulation in
wild-type and pad2 leaves inoculated with C.
carbonum (top graph) or Psm ES4326 (bottom graph),
1/20/96—1/23/96 (see Table 11 for data).
Abbreviations are as in Figure 15. Controls

contained no detectable camalexin. Standard error
bars are shown.

113

Table 12. Time course of camalexin accumulation in wild—type
and pad2 leaves inoculated with C. carbonum or Psm ES4326,
4/5/96—4/8/96. Numbers represent means plus standard errors
of 3 replicates for the C. carbonum-inoculated samples, and 2
replicates for the Psm-inoculated samples. Inoculum
concentrations were 2.0x105 spores/ml for C. carbonum and
1.2x107 cfu/ml for Psm ES4326. Controls (ctrl) were prepared
as described for Table 10.

response to C. carbonum response to Psm ES4326
inocu- (pg/leaf) (pg/leaf)
lation

 

 

wi ld- type pad2 wi ld- type pad2
0 0 ' 0

 

.072$0.008 0.137$0.017 0.00210.002 0.06410.063

.16i0.06 0.093$0.020 0.37$0.08 0.02610.001

.1210.05 0.077i0.017 0.3610.22 0.029$0.003
0 0 0

 

 

 

 

 

 

 

 

114

 

 

 

 

 

 

 

 

 

"‘ 1 0-
w:
s Cc
'2 ..
\ 0.8"
u 1
3 .
s 0.6: _a__ wt
3 0.4:1 —e—pad2
0
U I I I I I I
0 20 40 60 80
hours after inoculation
. 1.0: Psm
‘: :
o
H 0.81)
\
u cl
3 0.6-
d —Ht
d
g —-—pad:
I?
o
I I I 1— "—I
0 20 40 60 80

hours after inoculation

Figure 17. Time course of camalexin accumulation
in wild—type and pad2 leaves inoculated with C.
carbonum (top graph) or Psm E84236 (bottom graph),
4/7/96-4/10/96 (see Table 12 for data).
Abbreviations are as in Figure 15. Water—
inoculated controls contained no detectable
camalexin. Standard error bars are shown.

115

exception (Table 12, Figure 17).

The maximum amount of camalexin accumulated, for Psm-
inoculated leaves, ranged from 0.37—0.89 ug/leaf (Tables 10—
12). These numbers correspond to 35-53 pg per gram of fresh
weight (gfw), which is within almost the same range (22—50
ug/gfw) as those determined in similar experiments by
Glazebrook et al. (1997). The fact that the numbers are
lower than those determined fluorimetrically by Zhao and Last
(1996) (70-80 ug/gfw) may, again, reflect differences in
method of quantitation.

Some patterns in the kinetics of camalexin accumulation
were observed. In general, camalexin accumulated more
rapidly in the C. carbonum— than in the Psm-inoculated leaves
(Tables 10 and 11, Figures 15 and 16). Twenty-four hours
after inoculation, concentrations in C. carbonum—inoculated
leaves were at or near their maximum, whereas they were
barely detectable in Psm—inoculated leaves until 48 hours.
This pattern was not invariable, and the kinetics of
accumulation in wild-type and pad2 leaves sometimes differed.
There was a case in which C. carbonum-inoculated wild-type
leaves produced very little camalexin until 48 hours after
inoculation (Figure 17). In contrast, pad2 concentrations
peaked at 24 hours, so that pad2 appeared to produce more
camalexin than wild-type. This variability in kinetics
underlined the need to evaluate camalexin production at

several timepoints.

116

Effects of C. carbonum and Pseudomonas syringae pv.
syringae on camalexin accumulation in wild-type and
pad2 Arabidopsis.

Very little camalexin was produced in response to Pss
D20 in 1 time course (Table 13, Figure 18) and none was
produced in response to Pss D20 in another time course (Table
14, Figure 19). This result was completely unexpected because
Pss D20 had been found by Tsuji et al. (1992) to elicit high
concentrations of camalexin.
In C. carbonum-inoculated leaves, patterns of camalexin
accumulation were similar to those observed for C. carbonum-
inoculated leaves in the studies with Pam ES4326. Wild-type
leaves accumulated more camalexin than pad2 leaves (Tables 13
and 14, Figures 18 and 19), and concentrations were
appreciably high 24 to 30 hours after inoculation. The
concentration of camalexin in wild-type decreased
dramatically in one experiment (Figure 19), which could raise

questions about its metabolic fate.

Discussion

In summary, these time courses and dose-response experiments
helped to clarify some biological aspects of camalexin
accumulation, and to explain some of the ambiguous results of
the mutant screen and radiolabeling studies described in
Chapter 2. The inoculum-dependent partitioning of camalexin

between leaves and inoculum droplets at different spore

117

Table 13. Time course of camalexin accumulation in wild-type
and pad2 leaves inoculated (inoc) with C. carbonum or
Pseudomonas syringae pv. syringae (Pss) strain D20, 7/23/97-
7/26/97. Numbers represent means plus standard errors of 3
replicates, except for the 24- and 72-hour C. carbonum-
inoculated samples,which consisted of 2 replicates. Inoculum
concentrations were 2.0x10S spores/ml for C. carbonum and
1.2x107 cfu/ml for Pss D20. Controls (ctrl) were prepared as
described for Table 10.

    
   

  

  

   
   

 

 

 
 

 

 

 

 

 

 

Camalexin produced in. Camalexin produced in
response to C. response to Pss D20

inocula- carbonum (pg/leaf) (ug/leaf)

tion wildrtype pad2 wildrtype ,pad2

“0 0 0 0 0

24 ctrl 0.0001 0.0001 0 0
$0.0001 $0.0001

24 inoc 0.387 0.0606 0.0068 0.0005
$0.031 $0.0067 $0.0008 $0.0005

48 inoc 0.453 0.0460 0.0023 0
$0.051 $0.0101 $0.0021

72 ctrl 0.0001 0 0 0
$0.0001

72 inoc 0.557 0.0418
$0.095 $0.018

 

 

 

118

 

 

 

 

 

 

 

I; Pss
d
0
Fl
\
o
3
1L —l!|— wt-ctrl
fl .
a 0 I —9— wt-inoc
g + pad2 ctrl
.3 —°— pad2 inoc
0
l
-1 . . . . . , . . , . . .
0 20 40 60 80
hours after
inoculation

—¢— wt ctrl

—"— wt inoc
+pad2 ctrl

—°— pad2 inoc

 

camalexin (ug/ leaf)

 

 

 

 

hours after
inoculation

Figure 18. Camalexin accumulation in wild-type (wt)
and pad2 leaves inoculated (inoc) with Pseudomonas
syringae pv. syringae, (Pss, top graph) or
Cochliobolus carbonum (Cc, bottom graph) 7/23/97-
7/26/97 (see Table 13 for data). Standard error bars
are shown. Controls (ctrl) were prepared as described
for Table 10.

119

Table 14. Time course of camalexin accumulation in wild-type
and pad? leaves inoculated (inoc) with C. carbonum or Pss
D20, 9/21/97-9/25/97. Numbers represent means plus standard
errors of 3 replicates. Inoculum concentrations were 2.0x105
spores/ml for C. carbonum and 1.1x106 cfu/ml for Pss D20.
Controls (ctrl) were prepared as described for Table 10.

 

 

 

 

 

 

 

 

 

 

 

 

I ours Camalexin produced in Camalexin produced in
after response to C. response to Pss D20
inocu- carbonum (pg/leaf) (pg/leaf)
lation . .
Wild—type pad2 Wild-type pad2
0 0 0 0 0
24 ctrl 0 0 0 0
24 inoc 0.280 0.0374 0 0
$0.030 $0.0101
48 inoc 0.111 0.0239$ 0 0.0008
$0.011 0.0096 $0.0008
72 ctrl 0 0 0 0.0002
$0.0002
72 inoc 0.117 0.0162 0 0
$0.0370 $0.00741

 

120

 

 

 

 
 
  

 

 

 

 

1'1
Pss
'73 +wt
8 ‘ —0— pad2
r1
\
6
3E oh or are I
n
‘5 .
'3
0 '1 I I I I U I t t I T I j
0 20 40 60 80
hours after
inoculation
“0.3“ CC
4:
s
o
p:
Boz-
.5 .
q +Wt
-g '-+ - pad2
0.1-
'3
U 1
fl
0.0 . - I . . u
0 20 40 60 80
hours after
inoculation

Figure 19. Time course of camalexin accumulation in
wild-type and pad2 leaves inoculated with Pss D20
(top graph) or C. carbonum (bottom graph), 9/21/97
-9/24/97 (see Table 14 for data). Abbreviations
are as in Figure 18. No camalexin was detectable in
controls. Standard error bars are shown.

121

concentrations provided a possible explanation for why some
putative camalexin-deficient mutants proved not to be mutants

when retested. If the initial inoculum concentration was
near 106 spores/ml, the majority of the camalexin would have

remained in the leaves, making relative fluorescence
intensities of the droplets misleadingly low. These results

could explain in part the variable results of the
radiolabeling studies, since leaves were fed 1"C—anthranilate

after being inoculated, and the inoculum droplet was removed.
In general, the spore concentration was high enough that most
of the camalexin probably remained in the leaf. However, the
low incorporation of anthranilate into camalexin on some
occasions may have been partly the result of low amounts of
camalexin in the leaf. These dose-response analyses
demonstrated that to trust the results of only droplets or
leaves, inoculum concentration had to be standardized. Had
this been known at the time of the mutant screen, some false
leads might have been avoided.

The variability in the kinetics of camalexin
accumulation demonstrated the need to evaluate camalexin
production at several timepoints in order to determine
whether a plant was camalexin-deficient. Had camalexin in C.
carbonum-inoculated leaves been measured only 24 hours post-
inoculation in the experiment depicted in Figure 17, pad2
would have seemed to be an overproducer, and the wild-type
would have appeared camalexin-deficient. As discussed in

Chapter 2, this variability in kinetics may explain the

122

results of Zhao and Last (1996), who found that the trpl-lOO
mutant, which was found by Tsuji et al. (1993) to be
camalexin-deficient when extracted 18 hours after elicitation
with silver nitrate, produced wild-type amounts of camalexin
24 hours after inoculation with Psm ES4326 (Zhao and Last,
1996). The difference in results may be due to differences
in response to biotic and abiotic elicitors of phytoalexin
synthesis, but it may also be that the silver nitrate-
elicited leaves would have accumulated more camalexin over a
longer incubation period.

The time courses comparing the responses of wild-type
and pad2 to different pathogens confirmed that pad2 responded
similarly to C. carbonum and Psm ES4326. Whereas pad4, as
explained in Chapter 2, is a signal transduction mutant
affected in the ability to trigger salicylic acid production
(Glazebrook et al., 1997; Zhou et al., 1998), pad2 may be a
biosynthetic mutant, albeit a somewhat leaky one (Figure 17).

The tendency for camalexin to accumulate more rapidly in
response to C. carbonum may again be due to differences in
response to compatible and incompatible pathogens, rather
than differences in response to fungal and bacterial
pathogens. It was hoped that by directly comparing the
kinetics of accumulation in C. carbonum and Pss, an
incompatible bacterial pathogen, it would be possible to
determine whether the plants responded similarly to
incompatible fungal and bacterial pathogens, or whether they

responded differently to fungal and bacterial infection. The

123

lack of camalexin production in response to Pss prevented
making such comparisons. However, studies by Zhou et al.
(1998) with Psm ES4326 transformed with the avirulence gene
aerptZ demonstrated that camalexin did accumulate more
rapidly in response to this incompatible bacterial pathogen
than in response to the compatible one. In light of those
results, it seems likely that the more rapid accumulation of
camalexin in response to C. carbonum was a response to an
incompatible pathogen.

The lack of camalexin production in response to Pss D20

may reflect the problem of using a low inoculum concentration
(106-10.7 cfu/ml, compared to 108 used by Tsuji et al. [1992]) .

No sign of a hypersensitive response (HR) was seen on the
inoculated leaves, although there was some yellowing. It may
be that when the inoculum concentration is too low to cause
visible necrosis, no camalexin is produced. The inoculum
concentration was kept low because in trial extractions with

leaves inoculated with Pss cells diluted to an OD600 of 0.1 or

0.2 (five- to tenfold greater concentration than what was
used in these experiments), very little camalexin was
detected. It was thought at that time that the leaves had
mounted such a rapid HR (demonstrated by large necrotic
spots), that the cells died before they could produce
camalexin, and the spread of the pathogen was stopped before
more leaf cells could respond with phytoalexin synthesis. A

lower inoculum concentration was hopefully a way to

124

circumvent the problem associated with a rapid HR. Perhaps a
visible HR of some intermediate magnitude was necessary for
camalexin production. It is also possible that camalexin
does not play an important role in resistance to Pss.

This part of the project demonstrated that camalexin
deficiency, and discrepancies in the results of inoculation
with pathogens, could be due to many factors besides
regulatory mutations. It also provided, via TLC analysis,
evidence for inducible compounds which led to the study

described in the next chapter.

125

REFERENCES

Agrios GN. 1997. Plant Pathology, 4th edition. New York:
Academic Press.

Ayer WA, Craw PA, Ma Y, Miao S. 1991. Synthesis of
camalexin and related phytoalexins. Tatrahedron 48(14):
2919—2924.

Conn KL, Browne LM, Tewari JP, Ayer WA. 1994. Resistance to
Rhizoctonia solani and presence of antimicrobial compounds in
Camslina sativa roots. JOurnal of Plant Biochemistry and
Biotechnology 3: 125-130.

Conn KL, Tewari JP, Dahyia JS. 1988. Resistance to
Alternaria brassicae and phytoalexin-elicitation in rapeseed
and other crucifers. Plant Science 56: 21-25.

Cruickshank IAM, Perrin DR. 1961. Studies on phytoalexins.
III. The isolation, assay, and general properties of a
phytoalexin from Pisum sativum L. Australian Jburnal of
Biological Science 14: 336-348.

Dong X, Mindrinos M, Davis KR, Ausubel FM. 1991. Induction
of Arabidopsis defense genes by virulent and avirulent
Pseudomonas syringae strains and by a cloned avirulence gene.
Plant Cell 3, 61-72.

Ebel J. 1986. Phytoalexin synthesis: the biochemical
analysis of the induction process. Annual Review of
Phytopathology 24:235-264.

Glazebrook J, Ausubel FM. 1994. Isolation of phytoalexin-
deficient mutants of Arabidopsis thaliana and
characterization of their interactions with bacterial
pathogens. Proceedings of the National Academy of Sciences
USA 91: 8955-8959.

Glazebrook J, Zook M, Mert FM, Kagan I, Rogers EE, Crute IR,
Holub EB, Hammerschmidt R, Ausubel FM. 1997. Phytoalexin-
deficient mutants of Arabidopsis reveal that PAD4 encodes a
regulatory factor and that four RAD genes contribute to downy
mildew resistance. Genetics 146: 381-392.

Hahn MG, Bonhoff A, Grisebach H. 1985. Quantitative
localization of the phytoalexin glyceollin I in relation to
fungal hyphae in soybean roots infected with Phytophthora
megasperma f. sp. glycinea. Plant Physiology 68: 1161-1169.

Hammerschmidt R, Tsuji J, Zook M, Somerville S. 1993. A
phytoalexin of Arabidopsis thaliana and its relationship to
other phytoalexins of crucifers. In: Davis K, Hammerschmidt

R, eds. Arabidgp§i§_thaliana as a.Mbdel for Plant-Pathogen

126

Interactions. St. Paul, MN: American Phytopathological
Society.

Jejelowo 0A, Conn KL, Tewari JP. 1991. Relationship between
conidial concentration, germling growth and phytoalexin
production by camelina sativa leaves inoculated with
Alternaria brassicae. .mycological Research 95: 928-934.

Keen NT, Kennedy BW. 1974. Hydroxyphaseollin and related
isoflavonoids in the hypersensitive resistance reaction of
soybeans to Pseudomonas glycinea. Physiological Plant
Pathology 4: 173-185.

Storck M, Sacristan MD. 1994. The role of phytoalexins in
the seedling resistance to Leptosphaeria maculans in some
crucifers. Zeitschrift ffir.Naturforschung 50c: 15.

Tsuji J, Somerville SC, Hammerschmidt R. 1991.
Identification of a gene in Arabidopsis thaliana that
controls resistance to xanthomonas campestris pv. campestris.
Physiological and.Mblecular Plant Pathology 38: 57-65.

Tsuji J, Jackson EP, Gage DA, Hammerschmidt R, Somerville SC.
1992. Phytoalexin accumulation in Arabidopsis thaliana
during the hypersensitive reaction to Pseudomonas syringae
pv. syringae. Plant Physiology 98: 1304-1309.

Zhao J, Last RL. 1996. Coordinate regulation of the
tryptophan biosynthetic pathway and indolic phytoalexin
accumulation in Arabidopsis. Plant Cell 8: 2235-2244.

Zhou N, Tootle TL, Tsui F, Klessig DF, Glazebrook J. 1998.

PAD4 functions upstream from salicylic acid to control
defense responses in Arabidopsis. Plant Cell 10: 1021-1030.

127

Chapter 4. Isolation of Indole-3-Carboxaldehyde from
Arabidopsis Leaves Inoculated with Cochliobolus
carbonum, and Exploration of its Possible Role as an

Intermediate in Camalexin Biosynthesis

Introduction

The time courses described in Chapter 3 had a dual
purpose: to look for differences in camalexin production in
response to fungal or bacterial infection, and to look for
camalexin biosynthetic intermediates. When thin-layer
chromatograms were examined under long-wave ultraviolet (UV)
light to locate camalexin bands, other purple- or blue-
fluorescent bands were also observed (Chapter 3, Table 4).
Since these bands were present in extracts of C. carbonum-
and Psm ES4326-inoculated leaves but not in the mock-
inoculated controls, they were potential biosynthetic
intermediates. One attempt was made to characterize one
inducible compound that was present in extracts of C.

carbonum-inoculated leaves and had an Rf‘value slightly less

than that of camalexin. When this compound was eluted from
several TLC plates and analyzed by HPLC, no peaks of
noteworthy size were detected. Sample degradation on TLC
plates was not a likely reason since preliminary TLC of the

eluates had confirmed that the Rf'values were unchanged.

However, as the compound was obtained from several samples,

128

each equivalent to about 7 leaves, it was probable that more
sample was needed to characterize the compound.

Consequently, fresh fungal-inoculated plant material was
extracted on a larger scale. The extraction procedure
described in Chapters 2 and 3 was altered because of a
growing concern that if camalexin biosynthesis did involve
volatile compounds like indole or aldehydes, extracting at
high temperatures could lead to loss of compounds or
condensations between compounds. The new extraction
procedure led to the isolation of indole-B-carboxaldehyde
(Figures 4 and 5) from fungal-inoculated wild-type and pad2
leaves.

This result added support to the hypothesis that
camalexin is formed by a condensation between indole—3-
carboxaldehyde and cysteine (Browne et al., 1991; see also
Figure 5). Because of the recent confirmation that cysteine
is an intermediate (Zook and Hammerschmidt, 1997), the
discovery of indole-3-carboxaldehyde in inoculated leaves
provided support for this step in the pathway. Indole-3-
carboxaldehyde is present in many plants as an oxidation or
photolysis product of indole-3-acetic acid (IAA) and
tryptophan (see Chapter 1), and so it could be a constitutive
intermediate.

If indole-3-carboxaldehyde were an intermediate in
camalexin biosynthesis, the concentration of indole-3-
carboxaldehyde in inoculated plants was likely to increase

soon after infection, reach a maximum, and then decrease as

129

it was converted into camalexin. Time course studies with
wild-type and pad2 leaves confirmed that the kinetics of
indole—3-carboxaldehyde accumulation followed this pattern.

These time courses, however, provided only correlative
evidence. Without following the fate of radioactive indole-
3-carboxaldehyde in inoculated leaves, it was not possible to
know whether the aldehyde was being converted into camalexin
or was being used for other, unrelated compounds. Because
radioactive indole-3-carboxa1dehyde was not, to our
knowledge, commercially available, a different approach was
necessary.

One method sometimes used to determine whether a
compound is a biosynthetic intermediate in a pathway, when a
radiolabeled form of the compound is not available, is to
feed the nonradioactive (cold) form along with a radiolabeled
precursor and see whether specific incorporation (ratio of
moles of product to moles of radioactive compound
administered) into products of the pathway decreases (Wolf,
1964). Radiolabeling studies with the pad mutants (Chapter

2) had reaffirmed the findings of Tsuji et al. (1993) that
14C-anthranilate was incorporated into camalexin.
Incorporation of 14C-anthranilate into camalexin was expected

to decrease in leaves fed 1“C-anthranilate mixed with cold

anthranilate, since the leaves would make camalexin from both
the radioactive and nonradioactive molecules. If indole-3-

carboxaldehyde were an intermediate between anthranilate and

130

camalexin, the incorporation onMC-anthranilate into

camalexin should decrease even more in leaves fed.1%}-

anthranilate mixed with cold indole-3-carboxaldehyde, since
those leaves would have a more immediate precursor from which
to make camalexin and should be less likely to make it from
the more remote precursor. This method has been used to
verify the role of putative intermediates in the biosyntheses
of other compounds, including ethylene (Adams and Yang,
1979), dimethylsulfonio-propionate (DMSP) (Hanson et al.,
1994) and quercetin (Watkin et al., 1957). Therefore, an
attempt was made to see whether dilution of radioactive
anthranilate with cold indole-3-carboxaldehyde decreased the

efficiency of incorporation of anthranilate into camalexin.

Materials and Methods.

Plant material, fungal cultures, and inoculations.
Wild-type plants (Columbia-0 ecotype) and pad? plants

were grown as described in Chapter 2. Leaves of 3- to 4-
week-old seedlings were used in all experiments. Both wild-
type and pad2 plants were used for the initial searches for
intermediates and kinetic studies. Only wild-type plants
were used for the radiolabeling studies. Cochliobolus
carbonum was grown as described in Chapter 2, on V-8 agar
(per liter: 160 ml V-8 juice, 1 g calcium carbonate, 14 g

agar). Plants were inoculated as described in Chapter 2.

131

Reagents and chemicals.
14C-anthranilic acid (uniformly labeled on the ring),

with a specific activity of 60 mCi/mmol, was purchased from
ARC. Indole-3-carboxaldehyde was purchased from Sigma.
Indole-3-carboxylic acid and anthranilic acid were purchased
from Aldrich. The anthranilic acid was recrystallized from
ethanol for the radiolabeling studies. Even after
recrystallization, some material remained at the origin on
TLC plates; however, the size and intensity of the material
at the origin were greatly reduced.

Camalexin was purified by preparative TLC (glass-backed
silica plates from Whatman, 1000 um thickness) of a mixture
obtained from Arabidopsis and from two syntheses: one done in
January 1994 in the laboratory of Dr. W. Reusch at Michigan
State University, and one done in May 1996 in the laboratory
of Dr. J. Kagan at the University of Illinois at Chicago.
Both syntheses followed the method of Ayer et al. (1992),
except that in the later synthesis, tetrahydrofuran was used
as a solvent instead of benzene. Some synthetic camalexin
was also generously provided by Dr. Alois Farstner of the
Max-Planck-Institut fur Kohlenforschung (Mfllheim/Ruhr,
Germany). Small amounts of these various sources of
camalexin were prep TLC-purified at various stages of this
project. The camalexin used in the radiolabeling study was
isolated from preparative TLC plates and recrystallized from
hexanes-acetone (4:1, v/v) to give a final yield of 20 mg.

Solvents used for HPLC were of HPLC grade. All other

132

chemicals were of reagent grade or better.

Sample size and incubation.

Confirmation of the presence of indole-3-carboxaldehyde
in inoculated leaves came from three experiments. Sample
size and duration of incubation varied with each experiment
as information was gathered from the previous one. For the
extractions that led unexpectedly to isolation of indole-3-
carboxaldehyde, 78 leaves per sample were used, and leaves
were extracted 55 hours after inoculation. In a later
attempt to confirm the presence of indole-3-carboxaldehyde,
leaves were extracted 24 to 29 hours after inoculation, to
see whether the aldehyde was present when camalexin
concentrations normally were increasing. About 400 C.
carbonum-inoculated leaves (11.7 g) were used for one wild-
type sample, and about 300 leaves (5.8 g) were used for one
,pad2 sample. As a rough estimate of the amount of aldehyde
produced without infection, water-inoculated controls of 50
wild-type and 60 pad2 leaves were prepared. In a third
experiment to compare results between replicates and to
assess better the concentrations of aldehyde in control
tissue, wild-type and pad2 fungal- and water-inoculated
samples, prepared in triplicate, all consisted of 75 to 98
leaves. Twenty-four hours after inoculation, droplets were
collected from leaves and frozen, and leaves were frozen in
liquid nitrogen and stored at -20 °C. Samples were extracted

4 days later.

133

The yield of indole-B-carboxaldehyde for some samples in
the latter experiment was high enough that it was considered
sufficient to use 40 leaves per sample for nonradioactive
kinetic studies. Leaves were taken from several pots to
minimize the chance that variation between pots would
influence results. All samples were prepared in triplicate.
Inoculum droplets and leaves were collected and frozen 3, 6,
12, 24, and 36 hours after inoculation. Water-inoculated
controls were collected at 6 and 24 hours. Non-inoculated
leaves were collected for extraction immediately after
inoculation of the other leaves to determine aldehyde
concentrations just before infection. Samples were stored
dry at -20 °C up to 6 weeks prior to extraction.

For radiolabeling studies, samples (prepared in
duplicate) consisted of 5 leaves each, except for two 6-leaf
samples (collected before it was realized that there would
not be enough leaves to use 6 per sample throughout the
experiment) and four 4-leaf samples (the result of running
out of leaves). Leaves were inoculated after taking up one
of three radioactive solutions (see below). Samples were
frozen 3,6,9 and 24 hours after inoculation. Non-inoculated
zero-hour samples were frozen with as much water as would
have been on inoculated leaves. Water-inoculated controls
were collected and frozen at every timepoint. Samples were

stored up to 3 weeks at -80 °C before being extracted.

134

Radiolabeling of leaves.

The accumulation of indole-3-carboxaldehyde and
camalexin was followed over a 24-hour period in water- or C.
carbonum-inoculated leaves to which one of the following

solutions had been fed:

a) 14C-anthranilate (1 . 7nmol/1eaf)

b)14C-anthranilate (1.7 nmol/leaf) plus unlabeled

anthranilate (17 nmol/leaf; 336 uM, added as a 4.28

mg/ml solution in ethanol)
cfllAC-anthranilate (1.7 nmol/leaf) plus unlabeled
indole-3-carboxaldehyde (l7 nmol/leaf; 336 uM, added as

a 5 mg/ml solution in dimethylsulfoxide)

A tenfold dilution of radioactive precursor seemed likely to
ibe enough to note changes ir114C-anthranilate incorporation.

without encouraging aberrant biosynthetic pathways due to an
unexpectedly large pool of a compound, as has been documented
in other cases (Leete, 1991). To determine whether the final
concentrations of ethanol and dimethylsulfoxide (DMSO) were
phytotoxic, cold versions of solutions a, b, and c were fed
to leaves, as well as solutions containing twice those
volumes of anthranilate and indole-3-carboxaldehyde. No

macroscopic signs of phytotoxicity were observed.
The 1‘IC-anthranilate in all three solutions was diluted

with sterile deionized water to an activity of about 0.1 uCi
per 50 ul (the amount of solution fed to each leaf). The

activity of each solution was determined by counting two

135

10—ul aliquots in 5 ml of Safety-Solve scintillation fluid on

a Packard 1500 Tri-Carb scintillation counter (95 %
efficiency for 183).

Feeding of the radioactive solutions was as described in
Chapter 2. Leaves were inoculated following solution uptake.
To ensure that the light regime was the same for each
treatment, lights were left on during most of the experiment.
The 3, 6, and 9-hour timepoints were constantly in the light.
The 24-hour timepoints were exposed to about 18 hours of

light.

Extractions.

All steps of extractions were done at temperatures below
35 °C. Leaves were extracted according to a modification of
the Bligh-Dyer technique for extracting lipids (Bligh and
Dyer, 1959), using the proportions of chloroform, methanol,
and water determined by Kates (1972). In the kinetic
studies, to compensate for smaller sample size, all volumes
were doubled. For the initial aldehyde-seeking experiments,
leaves were frozen in liquid nitrogen and ground with a
mortar and pestle in chloroform-methanol (1:2 v/v; 1 ml
chloroform/l g tissue). Water (0.8 ml/g tissue) was added
upon or after transfer to a beaker..The homogenate was
vacuum-filtered through a Whatman #1 or #4 filter, and the
residue was again ground with chloroform-methanol-water
1:2:0.8 and filtered. The residue in the funnel was washed

with half as much chloroform-methanol 1:2 (v/v) as was used

136

in tissue homogenization. The filtrate was transferred to a
separatory funnel, to which chloroform and water were added
to give a final chloroform-methanol-water ratio of 1:1:0.9
(v/v/v). The chloroform layer was discarded. The aqueous
phase was extracted with 2 equal volumes of ethyl acetate.
The ethyl acetate layers were combined, dried under vacuum or
nitrogen, and stored at -20 °C. In the radiolabeling study,
to compensate for small sample size and keep the final
volumes small, leaves were ground in lS-ml plastic centrifuge
tubes with a glass rod. The second homogenization in
chloroform-methanol-water was omitted as well.

Inoculum droplets, if extracted separately from leaves
(as was done for the latter two aldehyde-isolation
experiments and for the kinetic studies), were extracted
essentially according to the procedure for inoculum droplet
extraction described in Chapter 3.

For the radiolabeling study, and for the experiment in
which indole-3-carboxaldehyde was first isolated, the
extraction procedure was similar to the leaf tissue
extraction procedure described above. The main difference
was that for the radiolabeling study and the first aldehyde-
isolation experiment, inoculum droplets were extracted.with
the leaves. Chloroform and methanol were added so that the
volume ratios of chloroform, methanol, and inoculum droplets
were 1:2:0.8 (v/v/v). The leaf mass was not taken into
account because the mass of the droplets (0.1 g per leaf for

0.1-m1 droplets) far exceeded the mass of the leaves. The

137

chloroform layer, instead of being discarded, was dried at
30-50 °C and saved for TLC, in addition to the ethyl acetate

extract of the aqueous layer.

Identification and quantitation of camalexin and
indole-3-carboxaldehyde.

Extracted samples were separated by TLC, as described in
Chapter 2. The majority of the separations were done on
glass-backed silica plates (Analtech) 250 um thick. Cold
samples extracted to calculate recovery of camalexin and
indole-3-carboxaldehyde (see below) were separated on
plastic-backed plates 200 um thick. The plastic-backed
plates were not used for the radioactive samples because
separation of camalexin, indole-3-carboxaldehyde, and
anthranilate was not as good as on glass-backed plates.
Plates were developed in chloroform-methanol 9:1 or 19:1,
with a prior development in chloroform if samples contained
large amounts of pigment. Camalexin and indole-3-
carboxaldehyde were visualized under UV light (see Chapter
3). Plates were photographed under long-wave and short-wave
UV light (see Chapter 5).*

Indole-3-carboxaldehyde was quantitated by HPLC after
eluting from TLC plates with ethyl acetate, (described in
Chapter 3). HPLC operating conditions were as described in
Chapter 3, except that the mobile phase was isopropanol-
hexane (8:92, 9:91, or 10:90 [v/v]). Composition was

adjusted to minimize peak tailing, which varied with

138

analyses. The typical retention time of indole-3-
carboxaldehyde was 13 to 14 minutes on an old column and 20
to 22 minutes on a new column. The identity of the peak was
confirmed by injection of a standard and by TLC of fractions
collected at the retention times of the standard. Because
the aldehyde had a low solubility in the mobile phase, as the
long retention times indicated, samples were dissolved not in
mobile phase but in hexane-isopropanol 70:30 (v/v) or 80:20
(v/v). Aldehyde concentrations in samples were calculated by
measuring the area of the triangulated peaks (Johnson and
Stevenson, 1978) and calculating concentrations from a
standard curve of peak area versus concentration (Figure 21).
For kinetic studies and recovery determinations,
camalexin and indole-3-carboxaldehyde were eluted for HPLC as
one band. The identity of each peak was verified by TLC of
collected, concentrated peaks (see Chapter 3). Camalexin had
a retention time of 6.5 to 8.5 minutes, depending on the day
and the polarity of the mobile phase. The camalexin peak was
broader in isopropanol-hexane 8:92 or 10:90 (v/v) than in
isopropanol-hexane 7:93 (v/v). Consequently, camalexin was
quantitated by measuring peak area and comparing to the
results of the standard curve used for the C. carbonum/P.s.
syringae time courses (Chapter 3, Figure 11B). The standard
curve for concentration of camalexin was recalculated to

quantitate camalexin based on peak area (Figure 22).

139

Determination of percent incorporation of 1‘C-

anthranilate into: camalexin, indole-3-carboxaldehyde,
and other compounds.

After separation of samples by TLC, the radioactive
plates were wrapped in plastic wrap (Borden), and laid onto
8x10” X-ray film (Kodak). After 4 weeks in a drawer, plates
were developed. Bands on the film were located on the TLC
plates by laying plastic wrap over the film, setting the film
over a light source, and tracing over the bands. The traced
reproductions of the film were then laid over the TLC plates,
and the bands were outlined on the plates, scraped, and
counted for 2 minutes in 5 ml of Safety-Solve scintillation

fluid on a Packard Tri-Carb 1500 scintillation counter
(efficiency for 18:: 95 %). Incorporation of anthranilic acid

into those compounds was calculated by converting the number
of counts per minute (cpm) to degradations per minute (dpm,
determined by efficiency of counting instrument) and dividing
this number by the initial dpm in the administered solutions.
Camalexin and indole-3-carboxaldehyde bands, which were
visible on the back of plates illuminated by short-wave UV
light, were outlined on the back of the plate with a
permanent marker. The outlines made it possible, once the
radioactive bands on the plate had been scraped, to see
whether the putative camalexin and aldehyde bands in the

autoradiogram coincided with the bands seen under UV light.

140

Calculation of recovery of camalexin and indole-3-
carboxaldehyde.

Because aldehydes are generally unstable and readily
undergo aldol condensations, the reliability of the
extraction procedure was tested by extracting samples
containing known amounts of indole-3-carboxaldehyde. To
determine recovery of the aldehyde in the absence of plant
components, trial extractions were done with aqueous
solutions of indole-3-carboxaldehyde (7.5 pg per 29.7 ml,
prepared by adding 30 ul of a 0.25 mg/ml stock in isopropanol
to 29.7 ml of water. Trial extractions were also done with
leaf tissue and with droplets spiked with 40 ul (0.4 ug) of
an aldehyde solution (a 1 mg/ml stock in tetrahydrofuran that
was diluted 1:100 in water).

For the radiolabeling studies, samples collected during
the time course were spiked with both camalexin and indole-3-
carboxaldehyde, to minimize the loss of compounds to tube
walls and leaf debris. As the extraction procedures were
different from what had been done for the other experiments,
preliminary nonradioactive extractions were done with leaves
spiked with camalexin and indole-3-carboxaldehyde, to
determine the percent recovery of both compounds.

Leaves were not spiked with indole-3-carboxaldehyde if
fed the solution of anthranilate diluted with aldehyde, since
preliminary extractions demonstrated that such leaves
contained easily-detectable amounts of aldehyde (Table 17A).

Leaves not fed aldehyde were spiked with 10 pg per 5-leaf

141

sample, added as a 0.5 mg/ml solution in ethanol (prepared
from a 5 mg/ml stock in dimethylsulfoxide). Camalexin was
added as a 0.56 mg/ml solution in methanol (16.8 ug/S

leaves).

Results

Isolation of indole-3-carboxaldehyde from wild-type
and pad2 leaves.

Indole-3-carboxaldehyde (molecular weight 145.0 g) was
isolated on two occasions from wild-type but not pad2 leaves,
and on one occasion from inoculum droplets on pad2 leaves.
The putative aldehyde was visualized under short-wave

ultraviolet light as a dark, UV light-absorbing band at the Rf

of a standard. The eluted compound was purified by HPLC.

The putative indole-3-carboxaldehyde peak comigrated with an
aldehyde standard on TLC (chloroform-methanol 9:1, v/v). The
aldehyde peak in droplets of pad2 leaves was collected from
repeated injections of sample on the HPLC (mobile phase:
isopropanol-hexane 8.5:91.5, v/v), dried under nitrogen, and
sent for mass spectroscopic analysis.

The mass spectrum of the putative indole-3-
carboxaldehyde (Figure 20) contained the expected peak at m/z
(charge-to-mass ratio)=145. The peak at m/z=144 indicated
the loss of a hydrogen atom from the carbonyl group, which is
characteristic of the mass spectrum of aldehydes (Hill, 1966;

D. Gage, pers. comm.) The peak at m/z=116 suggested the loss

142

CHO

\
N/

Indole-B-carbomaldehyde (M.W.=145.0g)

 

 

 

 

 

100‘ 571

i .

_ BB:

2' J 59 145(-

i .

V .

c 60-

n i

b d

:1 40- 7‘

d 1

.3 :

n

5 1166-
h‘ I i' . 192
I 1 H l .,.( iLULU

 

100

150 200

Figure 20. Mass spectrum of indole-3-carboxa1dehyde
isolated from pad2 inoculated with Q. carbonum. Arrows

denote characteristic peaks.

143

 

 

 

.7 30‘ y a — 0.0074856 + 2.8428x
e R‘2 I 0.996

6'

3

s

o

u

s

'5

so
9 () 2 4 6 8

micrograms aldehyde injected

Figure 21. Standard curve of peak area versus
micrograms of indole-3-carboxaldehyde, 12/17/96.
Points are means of 2 HPLC injections of a pure
standard. Peak area is based on measurement of peak
height and width at a recording chart sensitivity of
1.0 AUFS.

y I 0.22492 + 6.86713
R‘2 I 0.993

 

 

 

8
6
E.
a
o
H
I!
25.
Q
micrograms camalexin injected
Figure 22. Standard curve of peak area versus

micrograms of camalexin, 1/21/98. Points
represent means of 3 HPLC injections of a pure
camalexin standard. Area is based on peak height
and width at a recording chart sensitivity of 0.25
absorbance units full scale (AUFS).

144

of the carbonyl group (CHO, 29 mass units), which is also
typical of aldehydes (Hill, 1966). The presence of the
characteristic peaks seemed sufficient proof of the presence
of indole-3-carboxaldehyde and provided further support for
the results of HPLC and TLC analysis. The fact that the
aldehyde was sometimes detected only in wild-type, and
sometimes only in pad 2, was attributed to instability of the

aldehyde.

Recovery of camalexin and indole-3-carboxaldehyde.
The average percent recovery of indole-3-carboxaldehyde
under various extraction conditions is presented in Tables
15-17. Table 15 demonstrates that in the absence of plant
material, indole-3-carboxaldehyde is quite stable. Recovery
was relatively high (over 60 %), and for the 0.25 ug/ml
solutions tested, recovery may have been a bit higher because
the aldehyde had a very low solubility in isopropanol and may
not have been completely dissolved. The high recovery in a
solution kept at room temperature for 9 hours indicated that
degradation was unlikely to occur while leaves were being fed
radiolabeled solutions. No other compounds were seen under
UV light, and nothing was seen under UV light in TLC's of the
lyophilized aqueous phases. Hence, it did not seem likely
that the aldehyde was very prone to oxidation to indole-3-
carboxylic acid, or to decarbonylation to indole. Also, it
would not be necessary to study the aqueous phases to find

residual aldehyde.

145

Table 15.

was extracted.

replicates were extracted, and means plus standard errors are
shown.

sample

 

Recovery of indole-3-carboxaldehyde in the absence
of plant components. Only one replicate of the first sample

For the second sample (0.25ug/ml),

Pg
aldehyde
recov-
ered

ug
aldehyde
added

 

% recovery

 

queous
solution of
indole-3—
carboxaldehyde
(33.9 ug/ml)
extracted after
9 hours at room
temperature

66.3

80.8

82.1

 

"aqueous
solution of
indole-3-
carboxaldehyde
(0.25
ug/ul)extracted
with ethyl
acetate

 

0.5871
0.102

 

7.5

 

146

64.6$ 11.2

 

 

The results of Table 16 were obtained under the
extraction conditions for the nonradioactive studies (40-leaf
samples homogenized in a mortar and pestle); and the results
of Tables 17A and 17B were obtained under the extraction
conditions for the radiolabeling studies. The low recoveries
in Table 16 may have been partly due to spiking with a small
amount of standard (2.8-5 % of what was used to spike samples
in Table 178), but as the concentration of indole-3-
carboxaldehyde in plants is also low, the small amount of
spike probably provided a more accurate picture of recovery
in extractions of unspiked leaves. The samples may have
degraded partially while frozen, since the results for
samples analyzed seven months later were more variable and
cold be quite low. In calculating recovery of indole-3-
carboxaldehyde, values from the most recent recovery-
determination analyses were used.

One very unexpected result of the trial extractions for
radiolabeling was that the majority of the aldehyde was in
the chloroform fraction, although it was evaporated at high
temperatures conducive to condensation reactions. This
result suggested that the aldehyde was more stable than
expected. In that case, the absence of indole-3-
carboxaldehyde in the studies described in chapters 2 and 3,
thought to be due to work done at high temperatures (boiling
in 80 % methanol, or concentrating at 45 °C), may have been
due to enzymatic degradation. Since the leaves were put into

cold 80 % methanol and gradually heated up to boiling, there

147

Table 16. Percent recovery of indole-3-carboxaldehyde from
leaves spiked with a standard and extracted in the manner of
the nonradioactive studies of the kinetics of indole-3-
carboxaldehyde accumulation. Means plus standard errors of 2
or 3 replicates (number indicated by N) are shown.

sample mean ug ug % recovery

 

aldehyde aldehyde
recovered added
“wild-type 0 . 145:0 . 015 0.4 36$3
leaves (no (N=3)
droplets),

extracted 5/97
and analyzed

 

 

6/97
ild-type 0.0690i0.0240 0.4 11:6
leaves, (N=2)

extracted 5/97
and analyzed
12/97

. oplets, 0.146$ 0.032 0.4 34$14
extracted 5/97 (N=2)
and analyzed
12/97

 

 

 

 

148

mm\mm\m.

NN.0Hom.m bN0.0+Obb.o mmmHIm wanum

 

 

 

 

mm\m\m
.mmHQEMm
m.mHm.Hm m.Hv vo.m . mmoHum 03»
mm\m\m .3658A
m.mm . om.m mmmHnm mso
omsflnsoo
mommsm wmmsm mmmnm

soon as Macao an own macro as mmmad
hhm>oowu hhw>oowu mohnmoam mohnwoam macaw Ga .
w w on on mohsmon on mHQEMm.

 

 

 

 

 

1' l I
I I I

.EH0m0H0H50umaumU .mumumom
ahsumnoaoum uncowumfi>muonm .momsmoamxonumonmuwfioosfl mo sowusHOm w>fioomoaomusoc
0 flow mm>mwa Bonn momnwoamxonumoumuwaooafl mo hum>oomn osmoumm .mba wanna

 

 

149

 

 

 

 

 

 

 

1
m mm\HH\m
_ .mmHQEMm
_ H.0HHm.mm «.65 N.¢H Hm.ouhm.b mv.onbm.a mamaum mmhnu
m mm\m\m .mmassmm_
“ >.HHH.mm o.Hm m.¢H ma.oamm.> ha.ouhb.a umwaum woman
A mm\mm\m
m .mmHQEmm
m o.nav.mv b.mm o.w mm.o«oa.m vuo.o«mmm.o mmwauw 03»
1 mm\mm\m
.meQEMm“
_ m.mHH.mm b.mm o.oa mm.o Hum.m va.OHovw.o mmmHnm 03u_
_ Uwsflneoo mmmsm mxwmm mmmam _
1 mmmmaa Macao an as 06>: Macao as mamas _
_ soon Ga >Mm>oowu (moan mohnmoam odOum ca
>Hm>oomuw w on a: mohnmoam on wHQEmm_

 

.dba magma as no mcoflumfl>wunn<

 

 

.sofiuomuoxm mo mafia

um oomswoam sues omxflmm mw>me Eouw momnmoamxonumonmnmaoosfl mo mum>oomm .mba manna

150

may have been time for stress-related responses to occur,
such as the release of degradative enzymes that could oxidize
the aldehyde. Putting leaves directly into boiling 80 %
methanol would have quickly inactivated all enzymes ( R.
Hammerschmidt, pers. comm.; Harborne, 1973). The
effectiveness of a liquid nitrogen extraction was perhaps due
not to the low temperatures, but to the prevention of
enzymatic degradation by immediate freezing of leaves.

Table 17A demonstrates that the recovery of aldehyde was
lower in leaves fed the solution than in leaves spiked with
the solution (45-51 % vs. 63 %; see Table 16). This
difference in yield suggests that some of the aldehyde is
utilized by the leaves when it is fed to them. Despite the
_lower recovery, the aldehyde was easily detected on thin-
layer chromatograms of extracts of wild-type plants.

The organic (chloroform) phase obtained in extraction
‘was heavily pigmented, making it very difficult to separate
camalexin and indole-3-carboxaldehyde from pigments on TLC.
To see whether the amount of camalexin in the chloroform
phase might be negligible, two analyses were made of the
amount of camalexin in each phase. The results are listed in

Table 18 (below).

151

Table 18. Relative amounts of camalexin (expressed as
percentages of the total) in the phases obtained from the
liquid nitrogen extraction method used for most
nonradioactive samples in this chapter. For each sample,
5-7 g of leaves per sample were extracted. Camalexin was

agantitated by HPLC (isopropanol-hexane 7:93 V/V .

 
      
  
    
   
  
    

 

 

phase from % % % %

extraction recovery recovery recovery recovery
wild-type wild-type pad2 pad2
10/1 10/4 10/1 10/4

chloroform 13.2 8.57 36.7 22.9

phase

Ichloroform 6.68 4.54 5.35 14.3

extract of

aqueous

phase

 

 

 

 

 

 

Although pad2 contained 20 % of its camalexin in the

chloroform phase, separating pigments from.compounds of
interest was so difficult that a 20 % error due to lack of
chloroform analysis seemed about as accurate as the amount of
error involved in trying to separate camalexin from pigments
and assuming high recovery when half the camalexin might
remain hidden by chlorophylls. Therefore, the chloroform
phases in the nonradioactive studies were not analyzed.
Considering the relatively low amount (<25%) of
camalexin found in the chloroform phases at that time (Table
18), it was surprising to find the majority of the camalexin
in the chloroform phase in trial extractions done just prior
to radiolabeling (Table 19). The difference may be due to
extracting droplets separately from leaves in the extractions

described in Table 18. Had the droplets been ground in

152

Table 19. Percent recovery (mean plus standard error) of
camalexin (cam.) from leaves spiked with a camalexin standard
and extracted in the manner of extractions done in the
radiolabeling experiment. In columns 2-4, each row of
numbers denotes a separate replicate. Abbreviations are as
in Tables 17A and 17B.

 

 

 

 

 

 

 

m
Ikample ug cam. ug cam. in ug cam. mean%
in EtOAc CHCl3 phase added recovery
phase
two 5-1eaf 0.201 7.12 14.0 60.2$8.0
samples, 0.0292 9.52 14.0
5/25/98
two 5-leaf 0 10.53 16 8 58.0$4.8
samples, 0 8.94
5/6/98
three 6- 0 5.4 8.0 65$11
leaf 0 3.6 8.0
samples, 0.024 6.6 8 0
2/98
two 4—leaf 0.094 2.90 11.2 33.3$6.6
samples, 0.035 4.44 11.2
5/25/98

 

 

 

 

   

153

chloroform and methanol, the camalexin that diffused into the

inoculum droplet might have dissolved in the organic phase.

Kinetics of accumulation of indole-B-carboxaldehyde
and camalexin.

Indole-3-carboxaldehyde, although present prior to
inoculation, increased after inoculation with C. carbonum
(Figure 23A, Table 238). Concentrations reached a maximum at
6 hours, when camalexin was first detectable (Figure 23B,
Table 23A), and decreased rapidly afterwards. The kinetics
were as expected for an intermediate (Hanson et al., 1994),
since it appeared to accumulate prior to the onset of
camalexin accumulation, and then to decrease--presumably
because it was being converted into camalexin. Patterns of
accumulation were similar in two time courses (Figures 23 and
24). However, the maximum amount of indole-3-carboxaldehyde
produced in the first time course, when corrected for
recovery (Table 22), was only 10 % of the maximum amount of
camalexin produced (Table 23). Unless recovery of indole-3-
carboxaldehyde was extremely poor, or conversion into
camalexin was very rapid, it seems that not enough indole-3-
carboxaldehyde was produced to account for the amount of
camalexin produced. In pad2, the ratio of nanomoles of
camalexin to nanomoles of aldehyde was not as high (Table
23), but the amount of camalexin produced still exceeded the
amount of indole-3-carboxaldehyde produced. However, the

lower ratio of nanomoles of camalexin to nanomoles of

154

  

 

  

 

 

  

 

   

  

 

   

 

  

 

  

 

 

 

 

 

 

 

 

Hmoo.o mmaoo.o mmaoo.o mva.o mva.o mmaoo.o
Hoaao.o Hmamoo.o Hmmaoo.o Hbmm.o Hmmm.o Hmmaoo.o conflumm
moao.o moao.o mmm.o omm.o mmooo.o
memo.o Hmmmo.o o «Hom.o Hmmm.o Hmmooo.o oosflnom
mwmoo.o mmmoo.o
o o o “Nmmoo.o “Nmmoo.o o Huuonvm
mwooo.o Nmooo.o mmo.o mmo.o
“memoo.o Hm>moo.o o HNHN.0 HNHN.o o oocfluma
bmmoo.o bmmoo.o
o o o Habomooo o «bomoo.o confine
o o o o o o HHUOIm
o o o o o o cousin
0 o o o o o o
Hmuou mumamoufi. mm>mwa Hauou mumamouo mw>me

 

  

mafia msmwz

Ammma\mnv meQ.Gfl wamamemo

.bm\om\mubm\mm\m

 

Ammma\mnv mmhunbafi3 GM wamamsmo

 

     

sofiumHsoosfl
nouns musos

.gOd—m QHM mHOHHw CHMUGMUW

a 8063 £52880 .o no 3.33 0063 sues "0003509:

mm>me meQ.Usm mmhuubHflS Ga wawamamo mo coaumHafidoum wo wmudoo mafia .dom mHQMB

155

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

m o o o mM.oamm.H o "unmm.oamm.s ooze-mm=
Hmm.ouamm.o o HmN.OHHmm.o m¢.oaom.s o mv.o«ow.a ooze-¢m
F Hm.owfiq.a o HN.OHH¢.H mm.HHsm.m mmm.o«mmm.o HN.HHHm.m Huuo-¢m
= om.oumv.m mm.oasm.m sa.ommfl.e as.oasm.v mm.oamfi.m mm.oawszfl cone-mfl
= mm.ouwo.m ma.ousm.m mm.OHHm.m amm.ommm.m sw.o «mm.s ma.oasm.m ooze-m
mm.oumm.e o mm.00mm.a amm.o«nos.o mmm.oummm.o mmm.o4mmm.o Huuo-m

— ma.oamm.v mm.00ma.a ss.aamv.m em.oaes.m ma.oumm.a om.oasv.a_ coca-m=
_ so.HHmH.m --- so.HHmH.m o --- o o
COHU

— asuoo mDmHQoub mm>me Houou mumHQoub mo>bma .Lmdmmmww
Whom. wmhuubaflz mason

 

.bm\om\mnbm\mm\m

“Avocflv Esconumo .0 no AHHDUV Hmuwz
meQ_Usm mahoubafl3 mbhnmbamxonumonmuwHOQCw Mo coaumH58300m mo

£Da3 bobcadoosfl mw>mma
wmhsou mafia .mom canoe

156

.CBOEm mum mama Hounw onmocmum .bm\om\mubm\mm\m AHHDUV
Hmum3 Hoxooaflv Esconumo .9 Spas bonsadoocfl mm>mma meQ.on mmhunoaflz
as mbhnmoamxonumoumanObcfl mo :oflumHzfidoom mo mmnsoo mafia .mmm musmflm

nadusHuOOCa
nouns unson nadusdsooufi

ow om om as o nouns mason
. . . ow om om OH o

'

          
  
 
  

‘

 

o
(J‘PI/bu)

(EFFI/fiu)
Opflqoptvxoqzta—c-GIOPUI

ep£q0ptvxoqxvo-c-°topnt

onhunwadz

\OIDV'MNI-IO

 

HHUU 1'!
002mm IT

157

.Czosm mum mama Honnw UHMUGMDm .mHonucoo bmumHooonflunmums

one a“ Uwuomomb mm3 cfixmamemo mo ocsoem DGMOMMfismflm oz .bm\vm\m
(bm\mm\m .Aoosfiv Enconhmo .U nufl3 boomHsoocfi mm>mma meQ.me mmxu
uUHH3 cw Gflxmamaoo mo CofiumHDEdoom mo wmusoo mafia .mmm wusmflm

 

 

 

 

 

coausasoosd
nouns mason douusfisuoad
S. cm cm 3 o no»? mason
- p h I b n n 00.0
m ow om om OH o
I I
:Ho.o m l
1
v u tr
\1 UOCflI-IBI. .
mod N ..
I I
r 1
a an .
. o n B
no.0 u. u 03

 

\OLOV'MNHO
OOOCOOO

(JPOI/bn) urxetsmso

158

Table 21. Time course of accumulation of camalexin (A) and
indole-3-carboxaldehyde (B)
inoculated with water (ctrl) or C.

in wild-type and pad2 leaves
carbonum (inoc); 5/31/97-

 

 

 

 

 

 

 

 

 

 

 

 

6/1/97. Leaves and droplets were extracted separately and

combined for TLC.

A.
ours camalexin in camalexin in
after wild-type .pad2 (HQ/leaf)
inocula- (pg/leaf)

tion

0 0 0

3-inoc 0 0

6-ctrl 0 0

6-inoc 0.0103$ 0.0035 0

12-inoc 0.0254$0.00654 0.00189$0.00159
24-ctrl 0 0

24-inoc 0.0220$0.0035 0

“36-inoc 0.0173$0.0032 0.00159$0.00125 "

 

 

 

 

 

 

 

 

 

 

 

B.

Hours aldehyde in aldehyde in
after wild-type ,padZ (ng/leaf)
tion
0 0 0
3-inoc 3.71$1.25 0.940$0.484
6-ctrl 0 0
6-inoc 3.41$0.467 0.997$0.501
12-inoc 1.43$0.325 0
24—ctr1 0 0
24-inoc 0 0
36-ctrl 0 0

 

 

 

159

 

 

camalexin (pg/leaf)

0 10 20 30 40

hours after
inoculation

+Wt
-—+—- pad2

 

 

 

indole-B-carboxaldehyde

E:

d

0

r1

\

S -.--.

" 0 10 20 30 40
hours after
inoculation

Figure 24. Time course of accumulation of camalexin (A)
and indole-3-carboxaldehyde (B) in wild-type (wt) and
pad2.Arabidopsis leaves inoculated with C.carbonum,
5/31/97-6/1/97 (see Table 21 for data). Standard error
bars are shown. No camalexin or aldehyde were detected
in water-inoculated controls.

160

Table 22. Accumulation of indole-3-carboxaldehyde (ng/leaf,
corrected for recovery) in wild-type and pad2 leaves
inoculated with C. carbonum or water, 3/23/97-3/24/97.
Abbreviations are as in Tables 20A and 20B. Calculations for
recovery were based on the results of Table 16.

time wild-type
(hours)

 

leaves droplets leaves
0 8.9

12.4 4.4
0.90
19.6

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1
6
4. 8.4
9
5
3

6-inoc

 

Table 23. Comparison of nanomoles of indole-3-carboxaldehyde
(corrected for recovery) and camalexin (% recovery based on
the results of Table 18) produced in wild—type and pad2
leaves inoculated with C. carbonum. NUmbers for camalexin
are based on the results of Table 18A for total camalexin
(droplets+leaves). Numbers for indole-3-carboxaldehyde are
based on the results of Table 16.

ilours wild-type ,padZ

 

 

 

 

 

 

 

 

 

 

 

iafter

(inocula- nmol nmol . nmol nmol .
ltion aldehyde camalex1n aldehyde camalex1n
i0 0 0 0.061 0
!3-inoc 0.116 0 0.0869 0
16-ctrl 0.014 0 0.038 0
i'6—inoc 0.179 0.0284 0.0959 0
:12-inoc 0.0910 1.19 0.065 0.0480
124-ctrl 0.0737 0 0158 0.027 0
’24-inoc 0 0345 3.03 0.011 0.183
.36-inoc 0 0262 1 84 0 0 078

 

 

161

aldehyde suggests that the aldehyde and camalexin do have a
precursor-product relationship that is stoichiometrically
fairly sound in pad2. It may be that more than one pathway
to camalexin exists, and that pad? is camalexin-deficient
because it lacks that alternative pathway. Possibly the
wild-type plants utilize several camalexin biosynthetic
pathways simultaneously, in a manner similar to what is
proposed in Bu’Lock‘s (1965) discussion of the “metabolic
grid” by which some compounds come from.many different
pathways. Examples of compounds synthesized in this manner
are the tryptamine alkaloids of the grass Phalaris tuberosa
(Baxter and Slaytor, 1972). Based on the incorporation of
various putative precursors, at least 5 different
biosynthetic routes were possible.

The fact that pad2 accumulated two-thirds of wild-type
amounts of indole-3-carboxaldehyde, but only one-tenth as
much camalexin (Tables 22 and 23), suggests that indole-3-
carboxaldehyde is not a key regulatory step in the pathway,
if it is a step at all. Possibly, too, the pad? mutant is
iblocked at a biosynthetic intermediate between indole-3-

carboxaldehyde and camalexin (M. Zook, pers. comm.).

Potential phytotoxicity of radioactive solutions.

No visible signs of phytotoxicity were observed in trial
feedings of nonradioactive solutions. However, the rapid
*wilting of some of the leaves during the radioactive feeding

many indicate that the solutions were somewhat phytotoxic, and

162

that the effect was more pronounced at the unexpectedly

higher temperature of the room on that occasion.

Radiolabeling.

Because anthranilate is a precursor of many compounds in
addition to camalexin (Dewick, 1995), an attempt was made to
standardize the amount of time that elapsed between
completion of solution uptake and leaf inoculation.
Otherwise, there would be a risk that the anthranilate would
be shunted into primary metabolic pathways, such as
tryptophan biosynthesis, and that less would be available for
camalexin biosynthesis. However, it was not possible to
inoculate leaves a fixed number of hours after they had taken
up the radioactive solution. The large number of leaves to
check for completion of solution uptake (378 total), and the
fact that the leaves took up the solutions more rapidly than
usual (perhaps due to larger size and a warmer room) made it
too difficult to inoculate before all leaves had finished
taking up solution. The time lag between leaf feeding and
inoculation is depicted in Table 24. It is not impossible
that variations in time lapses prior to inoculation affected
pools of available anthranilate and anthranilate-derived
compounds, which, in turn, may have introduced unexpected

‘variables in the patterns of incorporation seen.

163

Table 24.

Time lapses during labeling of leaves.

 

hours

 

 

 

 

 

 

after feeding and removal from solution and inoculation
inocu- solution (hours) (hours)

lation

solu- a b c a b c

tion

0 3-6 6-8 6-9 0 o 0 fl
"3 6-9 8-10 6-9 6-7 6-8 8 fl
“6 6-9 8-10 6-9 6-7 6-8 8

"9 7-12 8-10 6-9 4 6-8 8

24 6 9 4 8

 

 

 

 

 

 

 

Many leaves wilted during the feeding period, due to not
taking up the solution or taking it up rapidly and then
wilting before water was added to the tube. Consequently,

sample size was reduced from 6 to 5 leaves per sample, which

meant that yields of compounds were lower than anticipated.

Appearance of thin-layer chromatograms and
incorporation of 1‘C-anthranilate into camalexin and

indole-B-carboxaldehyde.

Thin-layer chromatograms of both the ethyl acetate and
the chloroform extracts contained many bands per sample
(Figures 25-30). Camalexin was easily identified in the
samples, due to the spike added to each sample at the time of
extraction. Indole-3-carboxaldehyde was identifiable in most

samples for the same reason. However, in samples fed
solution (c) (”C-anthranilate diluted with nonradioactive

indole-B-carboxaldehyde), the aldehyde was detectable only in

164

the zero-hour samples (one of which had been spiked by
mistake). The absence of aldehyde in the other “c" samples
was surprising because in trial feedings, when leaves were
fed nonradioactive anthranilate diluted with indole-3-
carboxaldehyde, the aldehyde was detectable on thin-layer
chromatograms of the extracts, and the recovery was 45-50 %
(Table 17A). However, in those trial feedings, leaves were
always extracted immediately after solution uptake.
Therefore, the fate of the fed aldehyde over a 24-hour
incubation period was not known. It may be that the aldehyde
was used in other metabolic pathways, or that it was oxidized
to indole-3-carboxylic acid (Muller, 1961).

On autoradiograms, bands at the Rf's of camalexin and

indole-3-carboxaldehyde were present in some samples,
indicating that anthranilate had been incorporated into both

compounds. Surprisingly, a band at the R.f of camalexin was

present in the zero- and three-hour samples (Figure 25) and
in some controls (Figure 26). Because a band just below
camalexin was observed in some samples (Figure 26, sample
6b-i, and Figures 27 and 28, controls), it seemed possible
that this latter compound sometimes comigrated with
camalexin. To determine whether a different TLC solvent
would resolve the compounds, the unused portions of some
extracts (one replicate of each of treatments a, b, and c)
were first developed on TLC plates in the usual solvent

system. The camalexin bands were eluted with ethyl acetate,

165

 

Figure 25. Autoradiogram of TLC plate: tissue
extracts of wild-type Arabidopsis leaves, 0 and 3
hours after inoculation with C. carbonum (i) or water
(ct). Before inoculation, leaves were fed one of the
following solutions: 1‘C-anthranilate (a), 1"C-anthra-
nilate+cold anthranilate (b), or 1“C-anthranilate+cold
indole-3—carboxaldehyde. Bands are numbered as in
Figures 31-38 (see those figures for Rf-values. Other
abbreviations: SF=solvent front (total distance
traveled by solvent was 15 cm); OR=origin; A=indole—3—
carboxaldehyde.

166

 

Figure 26. Autoradiogram of TLC plate:extracts of
wild-type Arabidopsis leaves extracted 6 and 9 hours
after inoculation with C. carbonum or water.
Abbreviations are as in Figure 25. C. carbonum-
inoculated 9-hour samples are shown in Figure 27.

167

9a 9b 9c 24a 24a 24b 24b 24c 24c
i i i ct i ct i ct i

 

Figure 27. Autoradiogram of TLC plate: extracts
of wild-type Arabidopsis leaves extracted 9 and 24
hours after inoculation with water or C. carbonum.
Abbreviations are as in Figure 25.

168

24a 24a 24b 24b 24c 24c
ct i ct 1 ct i

 

Figure 28. Autoradiogram of TLC plate; tissue
extracts of wild-type Arabidopsis leaves, 24 hours
after inoculation with water or C. carbonum.
Abbreviations are as in Figure 25. These extracts
were from a separate replicate of the experiment
shown in Figures 25-27.

169

24a 24a 24c

3a 3a 3b 3b 3C BC
03 Ob 00 ct 1 ct i ct '

 

Figure 29. Autoradiogram of TLC plate: ethyl acetate
extracts of leaves extracted 0, 3, 6, and 24 hours
after inoculation with C. carbonum or water. Bands
are not numbered because none were analyzed, due to
the complexity of the autoradiogram. Abbreviations

are as in Figure
one replicate of
samples are from
was accidentally
was consequently

25. The 24-hour samples are from
the experiment, and the other

a separate replicate. Sample 24a
loaded onto 2 lanes, and sample 24b
spotted on a separate plate.

170

0b 3b 3b 6b 6b 9b 9b 24b

 

Figure 30. Autoradiogram of TLC plate: camalexin bands
eluted and redeveloped in chloroform-acetic acid 94:6
(v/v). Rfvalues of camalexin (cam) bands and of the
other band (possibly corresponding to band 5 in Figures
25-28) are indicated next to the arrows. Samples were
from leaves fed 1‘C-anthranilate diluted with cold
anthranilate. Abbreviations are as in Figure 25.

171

as described in Chapter 3. The eluted samples were dissolved
in 45 ul of ethyl acetate, and the entire sample was loaded
onto glass-backed silica TLC plates 250 um thick. Plates
were developed in chloroform-acetic acid 94:6 (v/v) and laid
onto film for 3 weeks. As Figure 30 demonstrates, two
compounds were present in the 6- and 9-hour C. carbonum-
inoculated samples. In the other samples, only one compound

was present, and it had a much higher Rfvelue than camalexin.

Therefore, it seemed safe to conclude that the putative
camalexin bands in controls, and at very early timepoints,
were due to comigration of another compound (possibly band 5)
with camalexin.

For indole-3-carboxaldehyde, co-chromatography with the
aldehyde spike, or (in the case of the aldehyde samples) co-
migration with spiked samples, was considered sufficient

proof that the band at that R.f was indeed indole-3-

carboxaldehyde, since in the kinetic studies, no other

compound had been detected by HPLC at the R.f of the aldehyde.

It was probable that some of the other bands consisted
of more than one superposed band. .Anthranilate derivatives
are generally polar, and in a nonpolar solvent like
chloroform-methanol 19:1 (v/v), those compounds would tend to
stay near the origin and be poorly separated from one
another. They could have represented many compounds
‘unrelated to camalexin biosynthesis. Possibilities would

include indole derivatives like tryptamine, indole-3-butyric

172

acid, indole-3-glyoxylic acid, indole-3-pyruvic acid, or
ascorbigen (Muller, 1961; Robinson, 1962). These compounds,
due to their relatively high polarity, were more likely to be
present in the ethyl acetate than in the chloroform phases of
leaf extracts. The large number of faint bands on one
autoradiogram of a TLC of ethyl acetate extracts (Figure 29)
demonstrates that anthranilate was incorporated into many
compounds. Because the chromatograms of ethyl acetate
extracts were very complex and total incorporation into the
ethyl acetate phases varied greatly, no bands were counted
from these plates.

On TLC plates of chloroform extracts, 7 different bands

were scraped in each sample: camalexin (Rf=0.51), indole-3-
carboxaldehyde (Rf=0.37-0.39), and 5 bands with Rf-values of

0.81, 0.71, 0.67, 0.46, and 0.27, respectively. These bands
were numbered 2, 3, 3.5, 5, and 7, respectively. Band 1 and

4, which had an R.f value of about 0.88, was so faint and

incorporated so few cpm of radioactivity that it was not
analyzed. Band 4 occasionally appeared as a very faint band
below band 3.5 (so called because it was so close to band 3),
but the appearance was so sporadic that it was not analyzed.
Band 6 corresponded to indole-3-carboxaldehyde.

About 5 other bands closer to the origin appeared
consistently in the chloroform extracts (Figures 25-31), but
these were not further analyzed because they were not as well

resolved as bands 1-7.

173

Total percent incorporation is given in Table 25 and
Figure 31. The incorporation of L4C-anthranilate into these
different compounds is depicted in Figures 32-39 and Tables
26-33.

An unexpected result of this study was that out of 3

treatments, incorporation into camalexin was highest in
leaves fed 1"C-anthranilic acid diluted with cold anthranilic

acid (Table 26, Figures 32 and 33). Based on the results of
other biosynthetic studies (Adams and Yang, 1983),
incorporation into camalexin was expected to decrease in
leaves fed that solution, since leaves would be making
camalexin from both radioactive and non-radioactive
anthranilate. However, higher incorporation into diluted
compounds is not unprecedented. Similar results were found
in studies with the flavonoid.phloridzin, in which feeding a
(mold form of a putative precursor led to higher overall
.tncorporation (Hutchinson et al., 1959). It may be that the
iJuareased supply of anthranilate to leaves caused an overall
increase in metabolism. Thus, the uptake of radioactive
anthranilate was more efficient than it would have been
otherwise. Because more anthranilate was available, more
camalexin was made. If more camalexin were made due to the
iJuzreased anthranilate pool, the overall recovery of
cammalexin in the extraction may have been better. Thus, it
is grossible that not much more radioactive camalexin was made
in these leaves than in the leaves fed the other solutions,

but that the recovery of radioactive camalexin was improved.

174

Incorporationiof3MC-anthrani1ate into camalexin was

lowered slightly by dilution with unlabeled indole-3-
carboxaldehyde. However, as the decrease was within the span
of the error bars for leaves fed undiluted anthranilate, it
is difficult to conclude that indole-3-carboxaldehyde
significantly lowered the percent incorporation into
camalexin.

Analysis of the effects of indole-3-carboxa1dehyde was
complicated by the fact that the total incorporation into
aldehyde-fed leaves was less than in leaves fed the other two
solutions (Table 25). This observation suggests that the
aldehyde may indeed be phytotoxic at the concentrations in
which it was fed, despite a lack of macroscopic symptoms. If
so, differences in incorporation into camalexin may reflect
not differences in the availability of biosynthetic
precursors, but differences in the solutions which were fed
to the leaves.

The phytotoxicity hypothesis aside, when the two
replicates are considered separately (Table 26, Figure 33),
differences in kinetics of camalexin accumulation are
apparent. In one replicate, camalexin reached a maximum 9
hours after inoculation in leaves fed undiluted.1%}-
anthranilic acid. At this time, the percent incorporation
into camalexin in aldehyde-fed leaves was slightly lower
(0.287 % versus 0.296 %). At 24 hours post-inoculation,
incorporation into camalexin was lower than at 9 hours in

leaves fed only anthranilic acid, but in leaves fed

175

anthranilic acid diluted with aldehyde, incorporation into
camalexin increased between 9 and 24 hours. Perhaps
incorporation into camalexin decreased in leaves fed only
anthranilic acid, due to the radioactive camalexin being
diluted by nonradioactive camalexin synthesized from
endogenous precursors. However, because the results for a
second replicate are clearly different (incorporation at 24

hours being lower in aldehyde-fed leaves being lower than in
leaves fed undiluted 14C-anthranilate) , this conclusion is not

well supported. The differences observed between treatments
may simply reflect the fact that total incorporation into
aldehyde-fed leaves was lower. Again, it may be that
variation in incubation times prior to inoculation created
differences in metabolite pools that caused artificial
differences between treatments. Also, the variations may
reflect differences in rate of penetration of C. carbonum on
individual leaves. Leaves on which penetration was slower
would begin to produce camalexin later than leaves on which
penetration was rapid. With such small sample sizes (one-
eighth the number of leaves used for the kinetic studies),
slight variations in camalexin concentrations among leaves
could be a significant percentage of the total amount of
camalexin produced.

The patterns of incorporation of anthranilate into
indole-B-carboxaldehyde (Figure 28, Table 33) do not provide

definitive answers on its role in camalexin biosynthesis.

Incorporation was very low. In leaves fed 1“C-anthranilate

176

alone, there was a small peak 6 hours after inoculation,
suggesting that aldehyde had accumulated to a maximum before
being converted into camalexin, as had been seen in the
kinetic studies. The decrease between 0 and 3 hours could be
due to the aldehyde being used in other pathways or being
oxidized to indole-3-carboxylic acid (Mfiller, 1961).

However, the difference in incorporation between controls and
fungal-inoculated samples is so small that it would be
difficult to conclude that the kinetics of accumulation in
the fungal-inoculated samples represent the accumulation of a
compound in an inducible pathway.

It is also possible, as discussed earlier, that indole-
3-carboxaldehyde is one possible intermediate of camalexin
biosynthesis, but that other pathways exist. A possible
intermediate is indole-3-carboxylic acid (Figure 4). A band
at the approximate Rf of the carboxylic acid was present on
autoradiograms, but it was not analyzed because it was poorly
resolved. Incorporation of indole-3-carboxylic acid into
camalexin may be worth examining in the future.

None of the other bands examined appeared to be obvious
precursors of camalexin. The incorporation into bands 2 and
7 is very low (Tables 28 and 32, Figures 34 and 38), and the
incorporation into bands 3 and 3.5 (Tables 29 and 30, Figures
35 and 36) varied so much between replicates that it is
difficult to draw conclusions from those results. The
results of incorporation into band 5 (Table 31, Figure 37),

the band that ran just below camalexin or sometimes

177

comigrated with it, are incomplete because the camalexin band
may have contained some of that compound. If so, the eluted
band 5 samples did not contain the entire yield of the
compound, and the differences in the amount of compound
obtained from the TLC plate may have varied erratically
between samples.

In summary, although the kinetic studies of accumulation
of indole-3-carboxaldehyde and camalexin supported the role
of indole-3-carboxaldehyde as an intermediate, the
radiolabeling studies did not provide an unambiguous answer.
The slight decrease in incorporation into camalexin could
simply be a reflection of the lower overall incorporation
into leaves fed indole-3-carboxaldehyde. Clear resolution of
this quesiton may require the synthesis of radioactive
indole-3-carboxa1dehyde to feed as a precursor. It may be
‘worthwhile to examine other putative intermediates instead,
such as indole-3-carboxylic acid. Since the kinetics of
accumulation support the role of the aldehyde as a camalexin
precursor but stoichiometry and radiolabeling data do not
support it strongly, it may be that more than one pathway to
camalexin operates. The carboxylic acid is one possible
alternative precursor. Unlike camelina sativa, another
camalexin-producing crucifer, Arabidopsis makes indole
glucosinolates, which have been shown to be precursors of
other cruciferous phytoalexins (Monde et al., 1994). Perhaps
.Arabidopsis can utilize glucobrassicin (Figure 4) as a

camalexin precursor.

178

 

Table 25. Percent incorporation (% inc) into chloroform
extracts of wild-type Arabidopsis leaves fed one of solutions
A, B, or C and then inoculated with water (ctrl) or C.
carbonum (inoc) . Leaves were fed 14C-anthranilate alone or
with a tenfold higher concentration (on a per mole basis) of
nonradioactive (cold) anthranilate or indole-3-
carboxaldehyde. Results for the individual replicates are
shown in the “% inc" columns; the mean and standard error are
shown in the “mean % inc.” columns.

 
  
 

     

     

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ours A. 1“C- B. 1‘4C- C. 1“C-
after anthranilate anthranilate anthrani-
inocu— +cold late+cold
lation anthranilate indole-3-
carboxaldehyde
8 mean % 8 mean % 8 mean %
inc. inc. inc. inc. inc. inc.
0 4.547 4.48$ 5.535 5.43$ 3.406 3.06$
4.397 0'08 5.321 0'11 2.711 0'35
3 ctrl 2.114 2.40$ 2.286 2.521 2.763 2.60$
2.691 0'29 2.744 0'225 2.426 0'165
3 inoc 2.778 2.40$ 2.790 2.67$ 2.508 2.98$
2.006 0'38 2.547 0'120 3.436 0'46
6 ctrl 2.968 2.99$ 2.425 2.66$ 2.910 3.36$
“ 3.007 0'02 2.902 0'24 3.818 0'46
H6 inoc 2.281 2.44$0.16 4.050 4.02: 2.114 2.40:
2.602 4.001 0'02 2.688 0'29
9 ctrl 2.359 2.421 1.991 2.441 2.087 2.34$
2.489 0'06 2.889 0’45 2.605 0'26
9 inoc 2.742 2.72$ 3.031 3.841 1.877 l.86$
2.695 0'02 4.660 0'82 1.828 0'02
24 3.199 3.07$ 2.319 2.21$ 2.802 2.68$
“ctrl 2.942 0.13 2.101 0.11 2.568 0.12
24 3.027 3.42$ 3.230 3.34$ 2.434 2.60$
Iii": 3.817 0'40 3.445 0'10 2.764 0'16

 

179

Figure 31. Average percent incorporation into
chloroform extracts of leaves inoculated with water
(control) or C. carbonum after being fed labeled
anthranilate (graph A) or labeled anthranilate diluted
with cold anthranilate (graph B) or cold indole-3-

carboxaldehyde (graph C). See Table 25 for data and
solution preparation. Means plus standard errors are
shown.

180

 

% incorporation

% incorporation

% incorporation

 

A. 14-C-anthranilate

 

 

 

 

 

 

  

 

 

—""— control
——-——inoculated
1‘
O I ' l ' fl
0 10 20 30
hours after inoculation
6 B. 14-C-anthranilate+ cold
anthranilate
—fil— control
‘-l-inoculated
1...
0 ' I ' I ' I
0 10 20 30
hours after inoculation
C. 14-C-anthranilate + cold
4 indole-B-carboxaldehyde
3 I
‘ --fl—- control
2.
- -4P-inoculated
1.
0 ‘ ' ' ' ' '
0 10 20 30

hours after inoculation

Figure 31 (caption on facing page).

181

Table 26. Percent incorporation into camalexin in wild-type
Arabidopsis leaves fed one of solutions A, B, or C and then
inoculated with water or C. carbonum. Abbreviations and
solution preparations are as in Table 25. Zeroes indicate
that the zone scraped at the Rf of camalexin had an activity
less than 100 cpm (roughly twice background on the
scintillation counter).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

“hours solution A solution B (14C- solution C (“C-
after UAC- anthranilate anthranilate
inocu- anthranilate) +cold +cold indole-3-
lation anthranilate) carboxaldehyde)

% inc. mean % % inc. mean % % inc. mean %
inc. inc. inc.

N0 0.0742 0.0371$ 0.0815 0.0714$ 0.0748 0.0374$

0 0.0371 0.0613 0.0101 0 0.0374

3 ctrl 0 0 0 0 0.0450 0.0225$

0 0 0 0.0225
3 inoc O 0 0 0 0 0
O 0 0
6 ctrl 0 0 0 0 0 0.0229$
0 0 0.0459 0'0229
"6 inoc 0 0 0.241 0.186$ 0 0.02253:
0 0.131 0'055 0.0450 0'0225
9 ctrl 0 0 0 0 0 0
0 0 0
9 inoc 0.462 0.296: 1.09 1.28$ 0.287 0.201$
0.130 0.166 1.46 0.18 0.116 0.086

24 0 0 0 0 0 0

I“ o o 0 ||
24 0.230 0.59l$ 3.05 3.00$ 0.399 O.338$
m“ 0.952 2. 0. 0'

 

 

 

 

 

     

 

 

182

ll
«‘3 3
u -
1":
-4P- B inoc
E' 24 -+— A inoc
8 -4F- C ctrl
_°_ .
E, C anC
x

 

  

 

   

 

O I I .‘__I
0 10 20 30
hours after
inoculation

Figure 32. Time course of incorporation of labeled
anthranilate into camalexin in leaves fed one of
solutions A, B, or C (see Figure 31) and then
inoculated with water (ctrl) or C. carbonum (inoc).
No incorporation into camalexin was detected in the
controls of leaves fed solutions A and B. Standard
error bars are shown.

183

 

Figure 33. Time course of incorporation of
labeled anthranilate into camalexin in each of 2
replicates of leaves fed one of solutions A, B,
or C and then inoculated with water (ctrl) or C.
carbonum (inoc). Standard error bars are shown.
See Figure 31 for solution preparations.

184

 

 

 

4‘ A. l4-C—anthranilate
a . -—4P- inoc-l
.3 3' -<P- inoc-2
u .
3
E?
O
81-
w-l M
a?
0 l ' l ' 1
0 10 20 30
hours after inoculation
4- B. 14-C anthranilate + cold

anthranilate

-4F- inoc-l

inoc-2

 

% incorporation
"3

    

 

O I ' I ' I
0 10 20 30
hours after inoculation

 

 

 

 

4-
a C. 14-C anthranilate + cold
0 1 indole-3-carboxaldehyde
“ —EI— ctrl-1
E —0— ctrl-2
o 2‘ -4F- inoc-l
g 1 -4P- inoc-2
w-I
a: 1‘

lF-fli-F.‘E=::

o . . . 1-——1
0 10 20 30

hours after inoculation

Figure 33 (caption on facing page).

185

Table 27. Percent incorporation into indole-3-carboxaldehyde
in wild-type Arabidopsis leaves fed one of solutions A, B, or

C and then inoculated with water or C. carbonum.

Abbreviations and solution preparations are as in Table 25.

Zeroes indicate that the zone scraped at this R; had an
activity below 100 cpm (roughly twice background on the

scintillation counter).

  

solution

  

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A (14C- solution B (14C- solution C (“C-
after anthranilate) anthranilate+cold. anthranilate+cold
inocu- anthranilate) indole-3-
lation carboxaldehyde)
% inc. mean % % inc. mean % % inc. mean %
inc. inc. inc.

0 0.244 0.259$ 0.240 0.191$ 0.343 0.201$
0.275 0'016 0.142 0°049 0.0592 0'142

3 ctrl 0.0800 0.0766$ 0.0777 0.0712$ 0.147 0.150$
0.0732 0'0034 0.0646 0'0066 0.153 0'003

3 inoc 0.0786 0.0797$ 0.0884 0.0668$ 0.136 0.170$

1 0.0809 0.0012 0.0451 0.0217 0.204 0.034

6 ctrl 0.0730 0.0678$ 0 0.0308$ 0 0.0991$
0.0626 0.0052 0.0617 0.0308 0.198 0.0991

6 inoc 0.0822 0.0932$ 0 0.0436$ 0 0.0648$
0.104 0.0111 0.0872 0.0436 0.130 0.0648

9 ctrl 0.0619 0.0548$ 0 0 0.0916 0.0458$

fl 0.0477 0.0071 0 0 0.0458

“9 inoc 0.0504 0.0500$ 0.0809 0.0765$ 0.0801 0.0401$
0.0496 0.0004 0.0721 0.0044 0 0.0401

24 0.0566 0.0530$ 0 0 0 0

trl 0.0494 0.0036 0 0 “

“24 0.0656 0.0645$ 0.0990 0.106$ 0.0775 0.0819$ “

inoc 0.0634 0.0011 0.112 0.007 0.0863 0.0044
__A

186

 

Figure 34. Time course of incorporation of 14-C
anthranilate into indole-3-carboxaldehyde in leaves
fed one of the indicated solutions (see Figure 31 for
solution preparation) and then inoculated with water
(control) or C. carbonum. Standard error bars are
shown.

187

 

0,3 A. 14-C anthranilate

 

 

n
O
...1
3
N 0-27 -¢- control
é -*- inoculated
8 .
n 0.1
..4
N
0.0 ‘ '

 

0 10 ' 2o ' 30
hours after inoculation

 

 

 
     

" B. 14-C anthranilate+ cold

d . anthranilate
o 0.2
-d
3 I
g “ -4+- control
e -—*—- inoculated
8 0 1‘
fl
'7 1
R

0 . I ' 1 "—-'l

10 20 30

hours after inoculation

C. 14-C anthranilate+ cold indole-3-

 

 

% carboxaldehyde
'3
d
g 0.2
0
g 0 1. -—°—- control
.fi ’ -—+- inoculated
R
0.0 -

0 . 10 20 30
hours after inoculation

Figure 34 (caption on facing page).

188

Table 28. Average percent incorporation (mean plus standard
error of 2 replicates) into band 2 (Rf=0.81) into leaves fed

one of solutions A, B, or C and then inoculated with water or
C. carbonum. Abbreviations and solution preparations are as
in Table 25. Zeroes indicate that the zone scraped at this
Rf had an activity below 100 cpm (roughly twice background on
the scintillation counter).

a“
% inc. % inc. % inc
solution A solution B solution C

      
      
 
  
 
 
 
 
 
 
   
 
 
 
  
 

 

(14C- (14C- (14C-
lation. anthranilate) anthranilate + anthranilate +
cold cold indole-3-
anthranilate) carboxalde-
hyde)
0 0.117$0.030 0.127$0.0098 0.0817$0.0157

 

 

 

 

3 inoc 0.0734i0.0124 0.0658$0.0006 0.0602$0.0102

 

3 ctrl 0.0981$0.0003 0.0716$0.0004 0.0654$0.0090
6 ctrl 0.0787$0.0030 0.0576$0.0023 0.0681$0.0002

 

 

inoc 0.0624$0.0060 0.0734$0.0015 0.0695$0.0040
ctrl 0.0594$0.0055 0.063110.0002 0.0587$0.0112

 

9 inoc 0.0898$0.0020 0.0691:0.0015 0.0469$0.0052

 

24 ctrl 0.0589$0.0154 0.044610.00146 0.0427$0.0110

0.0489$0.003l 0.0518$0.0052 I

 

.061010.

   

0003

 

 

 

189

 

0'2. A. 14-C anthranilate

...4}__control
-———e—-inoculated

 

 

% incorporation

 

 

°'°o 1 io' 20 7 30

hours after inoculation
0.2

   
 
 
  

B. 14-C anthranilate +
cold anthranilate

-{h—- control
—0— inoculated

% incorporation
O
H

 

 

 

 

0.0 - . - T - '
0 10 20 30
hours after inoculation
0.21
C. l4-C anthranilate + cold
3 indole-3-carhoxaldehyde
..4
g 4 —a— control
:4 001‘ —0- inoculated
a .
8
s
-4
fl 1 i f I V I
0.00 10 20 30

hours after inoculation

Figure 35. Incorporation of 14-C anthranilate into band 2
(Rf=0.81) over time. Graphs labeled as in Figure 31.

190

 

Table 29. Average percent incorporation (mean plus standard
error) into band 3 (Rf=0.7l, possibly indole) in leaves fed

one of solutions A, B, or C and then inoculated with water or

C. carbonum. Abbreviations and solution preparations are as
in Table 25.

 

 

 

 

 

 

 

 

  
 

 

 

 

 
 
 

 

 

 

 

191

   
 

 

hours % inc. % inc % inc.

after' solution A solution B solution C

inocu- (14C- (14C- (14C-

lation. anthranilate) anthranilate + anthranilate +
cold cold indole-3-
anthranilate) carboxalde-

hyde)

0 0.202$0.131 0.0740$0.0305 0.0538$0.0538

3 ctrl 0.169$0.016 0.0731$0.0003 0.0906$0.0283

3 inoc 0.337$0.198 0.0837$0.0107 0.202$0.092

6 ctrl 0.330$0.167 0.0763$0.0173 0.147$0.032

6 inoc 0.114$0.021 0.149$0.024 0.0708$0.0029

9 ctrl 0.144$0.0358 0.0506$0.0028 0.0735$0.0295

9 inoc 0.259$0.094 0.146$0.026 0.0810$0.0084

24 ctrl 0.263$0.151 0.0600$0.0110 0.168$0.045

24 inoc 0.317$0.0664 0.0877$0.0184 .216$0.086

O
o
b

    

AL
V

  

A. l4-C anthranilate + cold
anthranilate

control (SE

 

 

 

 

 

 

 

 

 

 

s
o a
'3 0.3. ’ at 6h=0.167)
3 J
E. inoculated
o 0.2 (SE at
g 3h=0.198)
d 0 e 1"
*’ 1r
0.0 . - , _ 11
0 10 20 30
hours after inoculation
8 O 4‘
'3 ' B. l4-C anthranilate + cold
d ‘ anthranilate
u
E‘ 0 . 3 '
8 O 2' ' control
a ° qh—-
'7 ‘ inoculated
#
30
hours after inoculation
0'47 C-l4 anthranilate + cold
s indole-3-carboxaldehyde
3. 0 . 3 -
u .
8
fi- 0 . 2 5
o ' +— control
0 # -IF-inoculated
s 0.1
or!
“’ 0.0 . . - 1

 

 

10 ' 20 30

hours after inoculation

Figure 36. Percent incorporation of 14-C
anthranilate into band 3 (Rf=0.7) over time.

192

Table 30. Average percent incorporation (mean plus standard
error of 2 replicates) into band 3.5 (Rf=0.67) in leaves fed

one of solutions A, B, or C and then inoculated with water or
C. carbonum. Abbreviations and solution preparations are as
in Table 25. Zeroes indicate that the zone scraped at this
Rf had an activity below 100 cpm (roughly twice background on
the scintillation counter).

 

 

 

 

hours % inc. % inc. % inc.

after 14C- 14C-anthra- 1"“C-anthra-

inocu- anthranilate nilate+cold nilate+cold

lation anthranilate indole-3-
carboxaldehyde

0 0.0733$0.0419 0.0372$0.0054 0.0158$0.0158

 

3 ctrl 0.0705$0.0138 0.0360$0.0016 0.0355$0.0083 "

 

13 inoc 0.0782$0.0208 0.0428$0.0140 0.0693$0.0361 “
I

 

6 ctrl 0.12430.053 0.0290$0.0008 0.0595:0.0230 “

 

6 inoc 0.0670$0.0170 0.082110.0381 0.0269i0.0269

 

 

“9 ctrl 0.0427 $0.0009 0 0.026410.0264

 

“9 inoc 0.0847$0.0015 0.0551i0.0073 0.0347$0.0056

 

24 ctrl 0.154$0.086 0.0548$0.0069 0.0913:0.0311 "

24 inOC 0.013910.0029 0.0640i0.0127 0.101i0.024
gm

 

 

 

 

 

193

A. 14-C
anthranilate

0.2” 47'
control (SE at

'-‘-- 24h=0.086)

+

 

 

inoculated

 

‘

% incorporation

{0 20 30
hours after inoculation

B. l4-C anthranilate +
cold anthranilate

O
o
...-

—II— control
'——*—'inoculated

 

 

t incorporation

     

 

'o 10 720 77” '30
hours after inoculation

c. ll-C anthranilate + cold
‘ indole-3-carboxaldehyde

control

(58 at 9h-0.0264)

 

-—o—inoculated
(SE at 0h=0.0158
* v - w .. - SE at 6h=0.0269)
0 10 20 30
hours after inoculation

 

 

o % incorporation

Figure 37. Percent incorporation of 14-C
anthranilate into band 3.5 (Rf=0.67) over time.
Graphs labeled as in figure 31.

194

Table 31. Average percent incorporation (mean plus standard
error of 2 replicates) into band 5 (R;=0.46) in leaves fed one
of solutions A, B, or C and then inoculated with water or C.
carbonum. Abbreviations and solution preparations are as in
Table 25. Zeroes indicate that the zone scraped at this Rf
had an activity below 100 cpm (roughly twice background on
the scintillation counter).

 

m

 

hours % inc. % inc. % inc.

after solution A solution B solution C

inocula- (14C- (14C- (14C-

tion anthrani- anthrani- anthrani-
late) late+ cold late)

anthrani-
fl late)
0 0.174$0.008 0.228$0.023 0.125$0.035

 

 

 

 

 

3 inoc 0.122$0.0018 0.12610.012 0.0848i0.0036

 

3 ctrl 0.127$0.002 0.118$0.008 0.0837$0.0019
0.14110.002 0.090510.0274 0.0838$0.0076

 

0.0896$0.0371 0.124$0.062 0.0695$0.0082"

 

 

0.149$0.004 0.085710.0003 0.0689i0.0005
0.0817$0.0358 0.0321i0.0321 0

 

0.103$0.001 0.0776$0.0029 0.0564$0.0105

 

0.0327i0.0051 0 0.0142i0.0142

 

 

 

 

195

0
fi
_l

A. 14-C anthranilate

 

 

s
O
I: 0.3
a
E 0.2 1 .._control
0 : ‘*——inoculated
3 0.1:
d 4
#

0.0 T . - - . 111

0 10 20 30

hours after inoculation

  
 
 
 
   

 

 

 

s 0.4
.3 B. 14-C anthranilate +
g 0 3 cold anthranilate
u . -O— inoculated
E: —.——control
0 0 . 2
o
a
...g
“ 0.1
0 e 0 T 1
0 10 20 30
hours after inoculation

a 0 o 4 n
.3 C. 14—C anthranilate +cold
“ 1 indole-3-carboxaldehyde
a 0.3
g (
o 0.27 .._.inoculated
g 1' *F-control
"' o . 1
“’ ——l

0.0

0 10 20 30

hours after inoculation

Figure 38. Percent incorporation into band 5
(Rf-0.45) over time. Graphs labeled as in Figure 31.

196

Table 32. Average percent incorporation (mean plus standard
error of 2 replicates) into band 7 (Rf=0.27) in leaves fed
one of solutions A, B, or C and then inoculated with water or
C. carbonum. Zeroes indicate that the zone scraped at this Rf

had an activity below 100 cpm (roughly twice background on

the scintillation counter).

 

 

Abbreviations and solution

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

re arations are as in table 25.

lhours % inc. % inc. % inc.

after' solution A solution B solution C

inocu— (14C— (14C- (14C-

lation. anthranilate) anthranilate + anthranilate +
cold cold indole—3-
anthranilate) carboxalde-

hyde)

0 O.177i0.073 O.215i0.089 0.0596i0.0596

3 ctrl 0.0835i0.0010 0.0733i0.0102 0.086510.0050

3 inoc 0.0769i0.0096 0.0735$0.0123 0.120i0.028 ll

6 ctrl 0.0536¢0.0174 0.0390i0.0390 0.0628i0.0269

6 inoc 0.0284i0.0284 0.0351i0.0351 0.0422i0.0422J

9 ctrl o o 0 ll

9 inoc O 0 0 n

24 ctrl O O O

24 inoc 0.0137i0.0137 O 0.015410.0154

 

 

197

 

 

 

 

 

 

 

 

 

a 0'3- A. 14-C anthranilate
o d}
3 +— control
3 -I— inoculated
:3.
#4
fl .
0 10 20 30
hours after inoculation
0’3 B. 14-C anthranilate + cold
3 anthranilate
3
a 4— control
E. 4F-inoculated
§
..g
#
. - , 4—
0 10 20 30
hours after inoculation
0.3m
C. 14-c anthranilate + cold
a w indole-3-carboraldehyde
.9.
u 0 O 2 -
s
u i +— control
E‘ -I— inoculated
8 0 e 1 d
s
as!
fl

0.0
O 10 20 30

hours after inoculation

Figure 39. Percent incorporation into band 7
(Rf=0.27) over time. Graphs labeled as in Figure 31.

198

This part of the project demonstrated the difficulty of
determining biosynthetic precursors in radiolabeling
experiments, and perhaps above all else, the difficulty of
balancing the need to work with small amounts of
radioactivity and yet obtain measurable amounts of product.
Given the difficulty of obtaining and interpreting the
results, it was a pleasure to turn to the work described in
the following chapter, partly because it provided hope for

finding biosynthetic precursors.

199

REFERENCES

Adams DO, Yang SF. 1979. Ethylene biosynthesis:
Identification of l-aminocyclopropane—l—carboxylic acid as an
intermediate in the conversion of methionine to ethylene.
Proceedings of the National Academy of Sciences USA 76(1):
170-74.

Ayer WA, Craw PA, Ma T, Miao S. 1992. Synthesis of
camalexin and related phytoalexins. Tetrahedron 48(14):
2919-2924.

Baxter C, Slaytor M. 1972. Biosynthesis and turnover of N,N-
dimethyltryptamine and S-methoxy—N,N-dimethyltryptamine in
Phalaris tuberosa. Phytochemistry 11: 2767-2773.

Bligh EG, Dyer WJ. 1959. A rapid method of total lipid
extraction and purification. Canadian JOurnal of
Biochemistry and Physiology 37(8): 911-917.

Browne LM, Conn KL, Ayer WA, Tewari JP. 1991. The
camalexins: new phytoalexins produced in the leaves of
camelina sativa (Cruciferae). Tbtrahedron 47 (24): 3909-
3914.

Bu'Lock JD. 1965. The Biosynthesis of.Natural Products.
London: McGraw—Hill, 82-83.

Dewick PM. 1995. The biosynthesis of shikimate metabolites.
.Natural Product Reports 1995, 579-607.

Gross D, Porzel A, Schmidt J. 1994. Phytoalexine mit
Indolstruktur aus Kohlrabi (Brassica oleracea var.
gongylodes). Zeitschrift fur.Naturforschung 49c: 281-285

Hanson AD, Rivoal J, Paquet L, Gage D. 1994. Biosynthesis
of 3-dimethylsulfoniopropionate in Wollastonia biflora (L.)
DC. Plant Physiology'los: 103-110.

Herbert RB. 1989. The biosynthesis of secondary
.metabolites. London: Chapman and Hall, 231 pp.

ZHill HC. 1966. Introduction to Mass Spectroscopy. London,:
Heyden and Son, 133pp.

IHutchinson A, Taper CD, Towers GHN. 1959. Studies of

Phloridzin in Malus. canadian Journal of Biochemistry and
.Physiology 37: 901-910.

200

Johnson E, Stevenson B. 1978. Basic Liquid Chromatography.
Varian Press, 354 pp.

Kates M. 1972. Tachniques of Lipidology: Isolation,
Analysis, and Identification of Lipids. Amsterdam: North
Holland Publishing Co., 349-350.

Leete E. 1991. Biosynthesis of Some Natural Products.
Minneapolis, MN: The Eleete Press, 132 pp.

Monde K, Takasugi M, Ohnishi T. 1994. Biosynthesis of
cruciferous phytoalexins. JOurnal of the American Chemical
Society'll6: 6650-6657.

Mflller F. 1961. fiber den Abbau von synthetischen
Indolverbindungen durch Blumenkohlrosengewebe. Planta 57:
481-502.

Robinson T. 1962. The Organic Constituents of Higher
Plants. Minneapolis, MN: Burgess Publishing Co., 306 pp.

Watkin JE, Underhill EW, Neish AC. 1957. Biosynthesis of
quercetin in buckwheat. Part II. canadian JOurnal of
Biochemistry and Physiology 35:229—237.

Wolf G. 1964. Isotopes in Biology; New York: Academic
Press, 173 pp.

Wright AD, Sampson B, Neuffer MG, Michalczuk L, Slovin JP,
Cohen JD. 1991. Indole-3—acetic acid biosynthesis in the
mutant maize orange pericarp, a tryptophan auxotroph.
Science 254: 998-1000.

Zook M, Hammerschmidt R. 1997. Origin of the thiazole ring

of camalexin, a phytoalexin from.Arabidopsis thaliana. Plant
Physiology 113: 463-468.

201

Chapter 5. Comparison of camalexin production and
resistance to the fungus Alternaria brassicicola among

twenty-four ecotypes of Arabidopsis thaliana.

Introduction

As explained in Chapter 1, the pad mutants initially
seemed ideal plants to obtain information on camalexin
biosynthesis and its role in disease resistance. The pad
phenotypes would illustrate the relative importance of
camalexin in disease resistance. The biosynthetic blocks due
to the pad mutations would yield camalexin intenmediates.
However, no putative intermediates had been found in the pad
mutants, except for indole-3-carboxa1dehyde, whose kinetics
of accumulation were similar in wild-type and in pad2
plants.

The latest studies of pad4 (Glazebrook et al., 1997;
Zhou et al., 1998) demonstrated that the PAD4 gene controlled
the ability of Arabidopsis to produce salicylic acid, a
compound necessary for potentiating camalexin production
(Zhao and Last, 1996; Zhou et al., 1998) in one of at least
two signal transduction pathways. A study of the pad mutants
indicated that camalexin plays a part in restricting growth
of some bacterial pathogens and in imparting resistance to
the oomycete Peronospora parasitica (Glazebrook and Ausubel,
1994; Glazebrook et al., 1997). It was possible that the
other pad mutations might be regulatory as well, and that

conclusions being drawn about the role of camalexin in

202

disease resistance might reflect the effects not of
biosynthetic blocks, but of mutations in other genes
regulating signal transduction. What other methods, or other
sources of biosynthetic blocks, might be available to
evaluate the role of camalexin in disease resistance?

It seemed possible that natural variation in Arabidopsis
populations would provide answers. Many ecotypes of
Arabidopsis thaliana (plants of the same species grown in
different environments) have been collected and made
commercially available. Although these ecotypes are similar
enough to be considered the same species, they differ
genetically in some respects because of selection pressures
imposed by the different environments. It is generally
expected that plants of the same species in different
locations will differ genetically (Gerson and Kelsey, 1998).
These differences frequently include variation in resistance
or susceptibility to certain pathogens. For example,
different ecotypes of Arabidopsis differ in their ability to
resist infection by the slime mold Plasmodiophora brassicae
(Fuchs and Sacristan, 1996), cauliflower mosaic virus
(Callaway et al., 1996; Leisner and Howell, 1992) turnip
crinkle virus (Dempsey, 1996; Dempsey et al., 1997), the
bacterium xanthomonas campestris pv. campestris (Tsuji et
al., 1991) the oomycete Albugo candida (Holub et al., 1995),
and the oomycete Peronospora parasitica (Mauch-Mani et al.,
1993). In those cases, resistance to a pathogen was due to

the presence of one or more genes, which some ecotypes had

203

and others did not.

Plant species in different environments also differ in
the kinds of compounds that they produce. The same pine
species, in different locations, will qualitatively and
quantitatively produce different piperidine alkaloids (Gerson
and Kelsey, 1998). Some oat species do not produce
avenacins, which are saponins that provide resistance to the
wheat pathogen Gaeumannomyces graminis var. tritici, and
these avenacin-deficient species are susceptible to that
pathogen (Osbourn et al., 1994). Since camalexin is produced
by other genera in the Brassicaceae, the ability to produce
it is a relatively ancestral trait, and an absence of
camalexin among some ecotypes of Arabidopsis thaliana could
represent a relatively recent evolutionary divergence within
the species. The absence of camalexin would conceivably
reflect differences in selection pressures of each habitat.
Since camalexin is not effective against all pathogens
(Rogers et al., 1996; Pedras et al., 1997), and sulfur is a
scarce element in plants, it is reasonable to think that
genes for camalexin production would be lost in environments
where it served no purpose. However, genes for production of
precursors might remain if any conferred a selective
advantage in a habitat.

It seemed reasonable, therefore, to look for variation
in camalexin production among Arabidopsis ecotypes and to see
whether any differences would be correlated with differences

in response to a given pathogen. A few outcomes were

204

possible, assuming variation in response to the pathogen.
Resistant and susceptible plants might produce similar
amounts of camalexin, or differing amounts that did not
correlate with any type of response. Alternatively, there
might be a pattern of either the resistant or the susceptible
plants accumulating more camalexin. Those plants
accumulating less camalexin might accumulate precursors--
antimicrobial ones providing resistance, or non-antimicrobial
ones that would not impede or contribute to resistance.

For this study, 24 ecotypes of Arabidopsis thaliana were
screened for the ability to produce camalexin when inoculated
with Cbchliobolus carbonum. This fungus was chosen because
it induced high concentrations of camalexin in the Columbia
ecotype (see Chapter 2, tables 2 and 3; and Chapter 3, Tables
6-11). Because C. carbonum is an incompatible pathogen of
Arabidopsis, and phytoalexin production frequently occurs in
incompatible interactions, it seemed like a good neutral
inducer to use in testing the general ability of other
ecotypes to make camalexin. A compatible pathogen might not
induce phytoalexin production, or induce a slower initial
rate of accumulation. At the same time that plants were
tested for their general ability to produce camalexin, they
were inoculated with the fungus Alternaria brassicicola, a
common crucifer pathogen that causes disease primarily on
species of the genus Brassica but also on other genera
(Ellis, 1968). Symptoms of the different ecotypes were

evaluated, and correlations were sought between resistance to

205

A. brassicicola and ability to produce camalexin.

Materials and Methods

Reagents and Chemicals .

Camalexin was from the source used for the radiolabeling
experiments described in Chapter 4. The concentration of a
sample was checked by HPLC, using the parameters described in
Chapter 3, and the sample was redissolved in methanol to give
a 0.5 mg/ml solution. The solution was stored at -20 °C to
prevent evaporation or degradation. Some of the indole-3-
carboxaldehyde used was purchased from Sigma Chemical Company
and used without further purification. Some was purchased
from Aldrich Chemical Company and recrystallized from 95 %
ethanol. Indole-3-carboxylic acid was from Aldrich.

Solvents used for HPLC were of HPLC grade. The methanol used
for TLC was of HPLC grade, since a casual check demonstrated
that separation of compounds was better if the methanol
component of the TLC solvent was of HPLC grade. All other

chemicals were of reagent grade or better.

Plant Material and Plant Growth Conditions.

The following Arabidopsis ecotypes were purchased from
Lehle Seed Company (Tucson, AZ): Aua/Rhon, Bensheim, C24,
Cape verde, Columbia-0, Columbia-glabrous, Dijon G, Estland,
Greenville, Kendalville, Landsberg erecta, Muhlen,

Niederzenz, Nossen, RLD, RLDl, S96, Turk Lake, Wassilewskija.

206

Ecotypes Ksk—l, Kas-l, Wei-1, and WA-l were provided by
S. Somerville (Carnegie Institute, Washington). Ecotype Anna
was provided by Jeff Conner (Kellogg Biological Station,
Hickory Corners, Michigan). Seeds were stored in desiccant
at room temperature, except for Columbia—0 and Columbia-g1,
which were stored at 4 °C.

Seeds were grown in clay pots (14.5 cm or 10 cm in
diameter) containing Baccto® High Porosity Professional
Planting Mix. The soil was wetted with Peters Professional
Fertilizer (2.49 g/l, as specified on the package) and hand-
compacted to provide a level surface, and a l- to 2-mm layer
of fine vermiculite was sprinkled over the surface. Seeds
were transferred to small dishes of water and distributed
over the vermiculite surface with a Rainin P200 automatic
pipettor, using pipet tips whose ends had been cut to make
them wide enough to take up seeds. Pots were covered with
plastic wrap until the seeds germinated, as described in
Chapter 2. Plants were grown in a growth chamber kept at 21

°C day and night, with a l6-hour light regime.

Fungal cultures and inoculations.

Cochliobolus carbonum and Alternaria brassicicola were
grown on V-8 agar (see Chapter 4) at room temperature under a
24-hour light regime. Inoculations with these cultures were
always done on the same day, using 7- to l4-day-old spores.
For C. carbonum inoculations, leaves were excised with a

razor blade and laid, adaxial side down, in Petri dishes

207

lined with moistened filter paper. Leaves were inoculated
with 0.1 ml of spores (2.0x10S spores/ml). For A.

brassicicola inoculations, a spore suspension was sprayed
onto intact plants with a commercial mister. The suspension
was prepared in water containing 0.05 % Tween-20, which acted
as a surfactant and created finer and more uniform droplets.
Each pot received roughly the same amount of inoculum. To
maintain a humid environment conducive to fungal infection,
pots were then covered with plastic bagsthat had been wetted
on the inside with water and held over a steam bath to ensure
that the inside was uniformly moist. The plastic bags were
arranged to cover the pots fairly snugly, again to maintain a
humid environment. Plants were left covered in this way for
24 hours to allow the fungus ample time to infect (minimum of
18 hours required [R. Hammerschmidt, pers. comm.]). A 3-cm
slit was then made in each bag, to allow the plants to adjust
gradually to the outside environment. The slit was widened,
usually 17 hours later, to about 7 cm. The 7-cm slit was
widened 4-7 hours later (44-48 hours after inoculation) to a
hole the size of the pot diameter. Symptoms were first
recorded about 60 hours (2.5 days) after inoculation. Bags
were usually removed at this time. Symptoms were then
recorded once a day for an additional 3 to 7 days, so that
the record of symptoms spanned 2.5-9.5 days post-inoculation

(dpi).

208

Cytological analysis of A. brassicicola infection.
Leaves of one susceptible and two resistant ecotypes,
and of one ecotype with an intermediate degree of resistance,
were excised 19, 48, and 72 hours after inoculation. Leaves
were placed in glass Petri dishes lined with 2 layers of
Whatman #1 or #4 filter paper saturated with a fixative
consisting of ethanol-acetic acid 3:1 (v/v). This solution,
by diffusing into the leaves, usually removed pigments in 2
to 5 days. Additional solution was pipetted over leaves once
or twice a day to keep them from drying out. Excess
solution, usually quite green, was sometimes removed prior to
adding fresh solution. Doing so helped to speed the pigment-
removal (clearing) process, probably because it prevented the
equilibration of pigments remaining in the leaves with
pigments in the surrounding solution. After a few days,
ethanol alone was added to the leaves, as the concentration
of residual acetic acid in the filter paper and leaves was
quite high by then. Prior to being mounted for microscopy,
leaves were rinsed with 3-5 aliquots of deionized water
(changed every 3-12 hours over a 1- to 2 -day period) to
remove the ethanol and acetic acid. Leaves were then
transferred to glass slides, blotted dry with tissue paper,
mounted in glycerol, and stored at 4 °C until photographed
for microscopy. Light micrographs were done on a Leica
microscope at a magnification of 400x, using Nomarski

differential interference contrast (DIC) optics.

209

Extraction of camalexin from inoculated leaves.

For each sample of leaves inoculated with C. carbonum,
inoculum droplets from 10 leaves were pooled and extracted as
described in Chapter 3. Samples were stored dry at 4 °C or
-20 °C until analysis.

To examine camalexin production in response to A.
brassicicola, 0.1-0.7 g of leaves (6-75 leaves) were excised
48 hours after inoculation (a time when camalexin was almost
always present in high concentrations in C. carbonum-
inoculated plants [Chapter 3]). Because the leaves were
sprayed with spores instead of being inoculated with a known
droplet volume, the area of infected tissue was less
consistent, and consequently, it seemed reasonable to
quantitate camalexin based on micrograms of camalexin per
gram of fresh weight (ug/gfw) instead of micrograms per leaf.
The leaves were weighed and boiled in 80 % methanol (added
such that the ratio of solvent volume to leaf weight was
150:1) for 20 minutes. Samples were usually stored at 4 'C
for 24-60 hours and then extracted in the manner of leaf

tissue extractions described in chapter 3.

Thin-layer chromatography (TLC) of extracts.

Droplet extracts were redissolved in 35 ul of ethyl
acetate, and the entire sample was loaded (in S-ul aliquots)
onto plastic- or glass-backed silica-gel TLC plates (silica
thickness: 200 um on plastic-backed plates [Selecto

Scientific], and 250 um on glass-backed plates [Analtech]).

210

In one case, leaf extracts were dissolved in 50 ul of
chloroform, and the entire sample was loaded. Because so
many pigments dissolved in the chloroform, it was not certain
that the camalexin was well-separated from pigments.
Consequently, in subsequent experiments, samples were
dissolved in 35 to 60 ul of methanol, depending on the amount
of leaf tissue from which the extract was made.

Known amounts of standards of camalexin, indole—3-
carboxaldehyde, and indole-3-carboxylic acid (usually 5 ug,
1.25-5 ug, and 2.5—5 pg, respectively) were also loaded.
Plates, which had been washed and activated prior to use,
were developed once in chloroform and then twice in
chloroformrmethanol 19:1. Camalexin, indole—3—
carboxaldehyde, and indole—3-carboxylic acid were identified
in samples by seeing which bands comigrated with the
appropriate standards. The plates were photographed with
Kodak slide film (100 ASA) under long-wave or short-wave
ultraviolet light, with an Olympus OM-l camera. A Kodak
Wratten filter for blue light was taped over the camera lens.
The camera was mounted onto a camera stand or a round ring
stand clamp and held about 52 cm above the plate. A hand-
held UV lamp was attached with a clamp about 7 cm above the
TLC plate. To obtain a sharper image of bands, the camera
was focused on the scored lanes on TLC plates, and the lens
was then raised 4 mm for photographing under long-wave UV
light, and 20 mm for photographing under short-wave UV light.

An exposure time of about 1 minute was used for long-wave UV

211

light exposures. This exposure time was sufficient for
short-wave UV exposures of fluorescent-backed TLC plates.
For non-fluorescent-backed plates, an exposure time of 2.5

minutes was necessary .

Bioassays for evaluation of camalexin production in
response to C. carbonum and A. brassicicola.

To compare camalexin production among the different
ecotypes, and to identify other antimicrobial compounds, TLC

plates were sprayed with a suspension of Cladosporium
cucumerinum spores (usually’3-3.5x106 spores per ml) in half-

strength potato-dextrose broth (per liter: 100 g potatoes,

5 g dextrose) and incubated in a humid chamber for 3 days, so
that spores grew everywhere except in fungal-inhibitory areas
such as camalexin bands. TLC plates containing extracts of

A. brassicicola-inoculated leaves were sprayed with A.
brassicicola (5x105 to 1x106 spores/ml). Cultures used for

bioassays were 7-11 days old, but an effort was made to use
7-day-old cultures, as the younger cultures contained more

viable spores (R. Hammerschmidt, pers. comm.).

Quantitation of camalexin. .
Camalexin was quantitated from the bioassayed plates

according to the method of Lazarovits et a1. (1982), who

found that the radii of zones of inhibition at the site of a

compound on a TLC plate were directly proportional to the

logarithm of the mass of the compound present. The zones of

212

inhibition at the Rf of camalexin were traced onto white paper

and scanned with a Hewlett—Packard ScanJet 4c scanner
(resolution of 150 dots per inch in both the horizontal and
vertical directions). The area in pixels was determined with

the Jandel Sigma-Scan program. These numbers were converted
into square centimeters (cm?) based on the number of pixels in
standards of known size. Since the area of a circle is equal

to In."2 (n83.14159), the radius (r) of the zone of inhibition

was calculated by the formula r=(A/n) “7. The mass of

camalexin was determined from a standard curve (Figure 40) of

inhibitory zone radius versus log10 of mass of a known amount

of camalexin.

RESULTS

Phenotypic characterisation of responses to Alternaria
brassicicola.

In general, symptoms of A. brassicicola infection
consisted of circular brown or gray lesions (1-20 per leaf),
ranging in size from less than 0.5 mm to 1.5 mm in diameter.
The larger lesions were sometimes irregularly shaped and
possibly consisted of several smaller lesions that had
coalesced. Leaves with lesions were sometimes chlorotic in
the entire area in which the lesions were found, but
sometimes the chlorosis was confined to a yellow halo around

the lesion, which appears to be fairly typical of A.

213

brassicicola infection (Ellis, 1968; Conn et al., 1988). A
condition termed “spreading chlorosis” was sometimes noted on
susceptible plants, in which a chlorotic zone began around
the leaf midrib and radiated outward, with no fixed
boundaries. Sometimes leaves did not become chlorotic around
the lesions, and the lack of chlorosis was considered a sign
of resistance. Leaves without lesions were sometimes
chlorotic as well. This might be partly due to the
inoculation procedure, as control plants inoculated with
water + 0.05 % Tween-20 sometimes had some chlorotic leaves.
However, most chlorosis was from.the infection, since there
was more chlorosis in infected than in control leaves.

On some plants, leaves with lesions wilted and became
necrotic 3 to 7 days post inoculation (dpi). Even on plants
that appeared to have no symptoms of infection, it was common
to find that at least one of the oldest leaves (usually close
to the ground, and often closely surrounded by neighboring
plants and hence easily overlooked) was dead 3 to 5 dpi.
Apparently, the older leaves are more susceptible than the
younger leaves to A. brassicicola infection. Another symptom
of infection was the formation of a lesion spanning the
cross-sectional diameter of the petiole. Leaves with such
lesions usually died. Since A. brassicicola is a
necrotrophic fungus (Boyd et al., 1994), quickly strangling
the plant would speed up its ability to acquire nutrients.

Few of the ecotypes tested had specific characteristic

disease phenotypes. Leaves could have spreading chlorosis

214

and large lesions in one experiment, and small lesions with
chlorosis only around the lesions in another experiment.
Consequently, the ecotypes in Table 33 were grouped by degree
of resistance (high, intermediate, or low). Any consistent
phenotypic traits were noted, but as these were fairly rare,
it seemed artificial to use those as categories.

For each ecotype, the severity of infection and degree
of resistance were judged partly by the percentage of leaves
with visible symptoms. The number of plants was counted at
the start of symptom observation, and the total number of
leaves was calculated based on the assumption that each plant
had at least five obvious rosette leaves. Since plants
seldom bolted before being inoculated, cauline leaves were
not counted. Disease severity was also judged by whether the
number of diseased leaves increased over time. Plants that
had symptoms on 15 % of the leaves at the start of the
observation period and did not develop more symptoms could
have been responding with an HR and stopping the spread of
the fungus. In contrast, if the number of diseased leaves
increased, it seemed that the plants might be less capable of
stepping the spread of the pathogen.

Plants with a high degree of resistance tended to have
few chlorotic leaves, if any, and fewer than 10 % of the
leaves had lesions. Sometimes the only symptoms were the
dead rosette leaves. Plants with a moderate degree of
resistance had lesions on 10 % to 40 % of leaves. The

symptoms sometimes looked rather severe. That is, the

215

lesions and chlorotic/necrotic areas could be quite large.
However, if relatively few leaves displayed symptoms, it did
not seem right to rate the plant as not resistant. The
“intermediate” category included plants with gradually
increasing numbers of diseased leaves, and plants with
symptoms that varied greatly from one experiment to the next.
Since those plants had a wide range of possible responses to
infection, it was difficult to rate them as resistant or
susceptible.

Plants with a low degree of resistance had large lesions
on many leaves at the start of the observation period. An
increase in the number of diseased leaves (up to 50-100 %)
was common. Necrosis, wilting, and death would begin 3-4
days after inoculation. Leaves often had the “spreading
chlorosis" mentioned before.

The disease phenotypes of the 24 ecotypes are shown in
Table 33, with additional comments about symptoms unique to
certain ecotypes. For ecotypes with greatly varying degrees
of susceptibility, those variations are noted.

The degree of resistance or susceptibility in each
ecotype depended partly on the light regime and the
concentration of inoculum used. When plants were grown under
less light, some of the normally resistant ecotypes, such as
Bensheim, had severe symptoms. These plants were somewhat
stunted at the outset and probably did not have enough

reserves for normal metabolism, let alone infection

conditions. At a higher concentration of inoculum (3.5x105

216

Table 33.

Ecotypes studied for camalexin production and A.

brassicicola resistance: abbreviation, origin, and disease

 

 

 

 

 

 

 

 

 

 

 

 

 

phenotype.
EEcotype name Place of degree of additional ;
)and common origin, if resistance traits
abbreviation , known :
)if known I
i l
Anna USA (MI) high displayed §
symptoms in ‘
only 1 out of
7 6 experi-ments;
Aua/Rhfin (Aa) Germany intermediate --- ‘
Bensheim (Be) Germany high -.._
?C24 --- low large lesions,l
‘ usually at
margins
Cape Verde Cape Verde intermediate large lesions
§(Cvi—0) Islands and spreading :
chlorosis .
Columbia-0 USA intermediate variable j
(Col—0; no symptoms (<58 ;
genetic to 15% 1
markers) diseased 5
‘ leaves) :
(Columbia- USA low large lesions
rglabrous (Col— at margins and}
391; glabrous rapid ;
gleaves used as development of;
*a genetic wilting and 5
‘ marker ) necros is
Dijon G (Di-G) France low ---
‘ (natural ;
habitat: ?
botanic ,
garden) .
{Estland (Est) Estland, intermediate -_-
‘ former ;
. USSR
iGreenville USA.(MI) high ---
(Gre)
‘Kas-l India low ---
aKendalville USA (MI) high ---
gKBk-l high "‘ :

 

 

 

 

217

Table 33 (cont'd).

 

Landsberg
erecta (La—er)

Germany

intermediate

 

Muehlen (Mh)

Poland

high

 

Niederzenz
(Nie)

Germany

intermediate
to low

lesions
usually small
(<0.5-1 mm in I
diameter) and 1
numerous (5-10)
per diseased !
leaf)

 

Nossen

high

 

l

Russia

high

leaves
sometimes
developed
lesions but
seldom became I
chlorotic

 

RLD 1

Russia

intermediate
to high

large necrotic;
zones )
following *
lesion
development

 

S96

Koornneef
Group,
Nether-
lands

high

leaves
sometimes
appeared more ‘
chlorotic than?
on most ;
ecotypes due !
to leaf g
morphology
(leaves more
spread out
than on most

 

Turk Lake

USA (MI)

high

ecotypes) i
very seldom 7
had any i
symptoms '

 

WA-l

Poland

intermediate

;
I
large circular}
lesions 1

 

Wassilewskija
(W8)

Russia

intermediate

 

Wei-1

 

Switzerland

 

intermediate
to low

218

 

round lesions
about 1.5 mm }
in diameter }

spores/ml), almost all of the ecotypes tested (MUhlen, RLDl,
RLD, Columbia-0, Greenville, Kendalville, Estland, S96,

Aua/Rhén, Nossen, and Anna) with the exception of Anna, had
severe symptoms. At a lower inoculum concentration (1.0x10S

spores/ml), all plants tested (Anna, Columbia-0, Columbia—
glabrous, Cape Verde, C24, Ksk-l, Landsberg erecta,
Niederzenz, Wei-1, Wa-l, Wassilewskija, and Kas-l) appeared
resistant except for Niederzenz, Wei-1, and Wassilewskija.
Symptoms could vary considerably from one experiment to the
next. Dijon G, although it consistently had lesions and
chlorotic leaves, sometimes had large lesions within
chlorotic areas and sometimes had many small necrotic lesions

less than 0.5mm in diameter.

Camalexin production in response to C. carbonum.

All ecotypes were capable of producing camalexin (Table
34). The Dijon G ecotype appeared to produce less than the
others, and on some occasions, it produced none. However, no
ecotype consistently failed to produce camalexin. Although
the amount produced was not always enough to inhibit
Cladosporium cucumerinum in a bioassay, it could usually, in
those cases, be detected under ultraviolet light (Figure
43A). The Col-0 ecotype sometimes produced very little
camalexin. Given the results of the time courses described
in Chapter 3, this variability in the amount of camalexin
produced may reflect variation in the kinetics of camalexin

accumulation from one experiment to the next. Inoculum

219

droplets were always collected 24 hours after inoculation,
and as was seen in one time course (Chapter 3, Figure 12),
camalexin concentrations are sometimes still low then.
Because the spore concentration was one at which the majority
of the camalexin should have diffused into the droplets
(Chapter 3, Tables 8 and 9), differences in camalexin
production were not likely to be due to variable partitioning
between droplets and leaves.

A few zones of inhibition besides camalexin were
observed on some TLC plates after bioassaying with C.
cucumerinum. However, these were not consistently present in
all experiments, and even between replicates, their presence
was inconsistent.

Although its identity was not verified, a compound at

the approximate R.f of indole-3-carboxylic acid was observed in

most extracts. Whether this compound was present as a
precursor or metabolite of camalexin, or as a degradation
product of other indole compounds in Arabidopsis, is unknown.

No zones of inhibition were formed at the R.f of that compound,

demonstrating that if it was indole-3-carboxylic acid and a
precursor of camalexin, it was not antimicrobial at the
concentrations in which it was present. The amount of
indole-3-carboxylic acid standard loaded (2.5 ug) was also

not inhibitory.

220

 

 

fl

0

3 y = 0.26054 + 0.28798):

up. 0.6- RAZ = 00997

.9

ad

'5 1

..1

'3 0.5-

0

a

O

I

$0 0.4-

O

U

5

0".

30.3 '1I"I"l*'r"l
H 0.2 0.4 0.6 0.8 1.0 1.2

log micrograms of
camalexin

Figure 40. Relationship between log of micrograms of
camalexin and radius of zone of inhibition on TLC
plates bioassayed with Cladosporium cucumerinum,. The
radii are the mean radii of zones of inhibition from
two separate TLC plates onto which known amounts of
camalexin were loaded.

221

Table 34. Camalexin production in response to inoculation
with C. carbonum. Droplets of inoculum (2x105 spores/ml)
were removed from leaves 24 hours after inoculation and
extracted with ethyl acetate. Camalexin on TLC plates was
quantitated by measuring the area of the zone of fungal
inhibition after bioassay with Cladosporium cucumerinum.
Numbers represent means plus standard errors of three
computer scans of a traced zone of inhibition. The last
column in the table denotes the amount of camalexin produced
relative to the amount produced by the Col-0 ecotype
extracted and bioassayed on the same dates. Where two
extracts of Col-0 were extracted and bioassayed on the same
date, the average of the results was used. Samples of Col-0
that were extracted on the same day and bioassayed on two
different days were not averaged.

Symbols and abbreviations: A “*" denotes that the amount of
camalexin in the Col-0 ecotype could not be used for
comparison, due to not being extracted or to the lack of a
measurable zone of inhibition. 'The letters “ND" indicate
that the amount of camalexin produced was not detectable by
a TLC plate bioassay (no zone of inhibition was formed). A
”---" indicates that the amount of camalexin produced
relative to Columbia-0 was not calculated because the amount
produced was too low to quantitate.

222

  

date of

 

date of

       

micrograms

% of ;

 

 

 

 

 

 

 

inocula- bioassay camalexin per amount of 3
tion 10 leaves camalexin i
produced
by Col-0
;Anna 7/7/98 7/21/98 0.65910.000 *
0.8641 0.000 *
7/24/98 8/23/98 7.1610.01 81.9
8/6/98 9/1/98 1.3510.00 *
9/21/98 10/8/98 5.9210.00 331
5.1810.01 289
ua/Rhon 7/24/98 8/23/98 4.1510.01 47.5
2.5410.00 29.1
8/15/98 9/28/98 0.85110.00 *
10/24/98 11/11/98 87.010.0 82.3
12110 114
Bensheim 8/8/98 9/1/98 3.7010.13 243
ND ---
8/19/98 9/28/98 4.5910.00 425
4.1210.00 381
10/4/98 10/16/98 48910 91.2
: 56510 105
C-24 7/24/98 8/23/98 2.8010.00 32.0
3.5810 41.0
8/8/98 9/17/98 3.1810.00 209
9/1/98 1.3810.01 *
Cape Verde 7/24/98 8/23/98 2.56 10.00 29.3
0.78810.001 9.02
8/8/98 9/17/98 3.5610.00 234
9/1/98 12.510.0 *
Col-0 7/24/98 8/23/98 8.7410.01 N.A.
8/8/98 9/17/98 1.5210.00
8/8/98 9/1/98 ND
8/15/98 9/28/98 ND
8/19/98 9/28/98 0.90710.000
8/19/98 1.2610.00
9/5/98/ 9/28/98 5.8810.02
4.8110.00
9/21/98 10/8/98 1.7610.00
1.8210.00
10/4/98 10/16/98 55110
52110
10/24/98 11/11/98 11310
98.310.l

 

 

 

 

 

223

Table 34 (cont'd).

 

 

 

 

 

 

 

 

 

 

 

 

 

Columbia-g1 7/24/98 8/23/98 1.0110.00 11.6
2.4210.00 27.7
8/19/98 9/28/98 1.0910.00 101
9/21/98 10/8/98 2.7410.00 153
ND _--
Dijon G 7/7/98 7/21/98 0.58510.000 *
0.80410.000 *
7/24/98 8/23/98 ND -—-
8/8/98 9/17/98 ND ---
9/1/98 1.5310.00 *
9/21/98 10/8/98 ND ---
Estland 8/19/98 9/28/98 0.74710.000 69.2
ND _--
9/21/98 10/8/98 2.1010.00 119
1.9210.00 107
Greenville 7/7/98 7/21/98 2.9510.00 *
7/24/98 8/23/98 8.7510.01 100
8/15/98 9/28/98 ND --—
9/21/98 10/8/98 2.9410.00 164
3.6210.00 202
10/4/98 10/16/98 34710 64.7
22410 41.8
Kas-l 7/24/98 8/23/98 5.3110.01 60.8
8/8/98 9/1/98 3.3610.00 *
9/17/98 4.7610.01 313
10/4/98 10/16/98 53712 100.
31310 58.4
Kendalville 7/24/98 8/23/98 2.0510.01 23.4
3.0610.00 35.0
8/8/98 9/1/98 2.9510.00 *
ND ---
Ksk-l 8/8/98 9/1/98 1.8710.00 *
8/19/98 9/28/98 3.0610.00 289
9/21/98 10/8/98 3.7010.00 207
7.1310.00 398
Landsberg 7/24/98 8/23/98 1.8510.00 21.2
erecta 3.3310.01 38.1
8/8/98 9/17/98 ND ---
9/1/98 2.9610.00 *
Mfihlen 7/24/98 8/23/98 10.410.0 119
8/19/98 9/28/98 3.1410.00 291
0.72810.000 67.4
9/21/98 10/8/98 1.9710.00 110.
1.6110.00 89.9

 

224

 

Table 34 (cont'd).

 

  

 

 

 

 

 

 

 

 

 

 

 

225

 

 

gNiederzenz 7/24/98 8/23/98 4.5110.00 51.6
‘ 2.7910.00 34.0
8/8/98 9/1/98 1.7610.00 *
9/21/98 10/8/98 2.1910.00 122
Nessen 7/24/98 8/23/98 1.5610.01 17.8
1.7610.00 20.1
8/8/98 9/17/98 l.6610.00 109
9/1/98 0.40510.000 *
iRLD 7/7/98 7/21/98 1.8410.00 *
2.1610.00 *
' 7/24/98 8/23/98 6.5610.01 75.0
; 8/8/98 9/1/98 1.3010.00 *
. 9/17/98 ND ---
; 9/21/98 10/8/98 3.8610.00 188
; 2.77:0.00 155
:RLDl 7/7/98 7/21/98 3.5310.00 *
. 2.0410.00 *
7/24/98 8/23/98 8.9510.01 102
4.4810.00 512
9/21/98 10/8/98 2.3810.00 133
3896 7/24/98 9/1/98 3.2410.00 37.1
' 5.6510.01 64.6
8/8/98 9/17/98 2.0910.00 202
9/1/98 5.0310.00 *
Turk Lake 8/8/98 9/1/98 ND ---
9/5/98 9/28/98 5.3110.00 99.4
‘ s.79::0.00 108
5 6.7710.02 127
;WA-1 7/24/98 8/23/98 2.7810.00 31.8
' 8/19/98 9/28/98 ND ---
9/21/98 10/8/98 2.6010.00 145
assilew- 7/24/98 8/23/98 1.8810.00 21.5
skija ND ---
8/15/98 9/28/98 ND ---
10/4/98 10/16/98 560.10 104
410.10 76.5
)wei-l 7/24/98 8/23/98 5.3010.00 60.6
‘ 2.8010.00 32.0
; 8/8/98 9/17/98 0.84110.000 55.3
' 8/19/98 9/28/98 2.0610.00 191

Camalexin production in response to A. brassicicola.
Of ten ecotypes extracted after inoculation with A.
brassicicola, only two (Kas-l and RLD) produced enough
camalexin to be detected on TLC plates. No zones of
inhibition were formed in TLC plate bioassays. When the
experiment was repeated with nine of those ecotypes, using
about five times more leaves than in the first experiment,
only Kas-l, Anna, and RLD produced enough camalexin to detect
under UV light. After bioassay, faint zones of inhibition at

the R! of camalexin were detectable in extracts from.Anna and

Kas-l, but not from any other ecotype. Anna contained two
other weakly antimicrobial compounds. The absence of zones
of inhibition may reflect not only a lack of camalexin, but a
lack of sensitivity of A. brassicicola to camalexin. The
zone of inhibition around the standard was smaller and less
clearly defined than it was on a plate containing the same

concentration of standard and sprayed with C. cucumerinum.

Cytological analysis of infection in resistant and
susceptible ecotypes.

About the same number of conidia per square centimeter
were present on the surfaces of resistant and susceptible
ecotypes (Figures 41 and 42). Therefore, differences in
response to A. brassicicola were not due to differences in
the number of spores available to cause infection. On the

resistant ecotype Turk Lake, few signs of infection were seen

226

.Haoo ecu mo mucmummdmm ooNHcmmuomHo ecu ocm
we um demo 0;» as was»; was 6602
lucmauv mason as

.coeumHSUOCH Houwm
Amocmumfimwu wo woumoo zoavalwmx mm>uoow Mo
wcmmumouoflfi unwed

Amofiumo exhumaoz .cofiumoflmecmmfi xoovv

 

227

Figure 42. Light micrographs (400x magnification,
Nomarski optics) of ecotypes Turk Lake and RLD (high
degree of resistance). Figure 42A, Turk Lake, 72
hours after inoculation with A. Brassicicola. Note
the short germ tubes. Figure 42B, RLD, 19 hours
after inoculation. Germ tubes are longer than germ
tubes on Turk Lake, but few signs of penetration are
visible. Figure 42C, RLD, 72 hours after
inoculation. Penetration has occurred in the cell
beneath the topmost conidium (upper middle, near the
right), and the infected cell appears collapsed.

228

.Nv museum

 

229

.oflom 0fla>x0bumoumuoaoocflumo “woxcooameQHMUum

nmaooceud "wcoaumfl>wubbm uwcuo .3ouum ecu >o czocm we ocmn AOL saonmEmo
one .szv Hiawz ocm .Axv Hummx .Auwv oHHH>cwwuo .Afimv Hoax .A>UV oouw>
mono txmv mam .Abzv swans: .Aouv oanonHou .Amzv mnflxmzmaemmmz .Am<v
cocm\m:¢ momxuooo owumHDUOCfltfizconumo .0 mo muomuuxw uoamouo ansooca

mo .ucmfla >3 o>m3|uuocm noon: owzmfl>v wumao DAB .c0fluosooua sewamEmo

co Auouooce Hmuuoocv Esconamo mswonoHHaUOU wo muowwmm .mm¢ gunmen

o

 

mo< 0 $3 xuw Hm >0 m :2 OUmB ad

230

.mmq ouomem ca mm mum mCOHHMH>mHQn¢

.mxmo m Mom Hmbfimco peas: m CH owumnoocfl pom AHE\o0me DDOQMV

monomm Esseuossoso .U cue: om>mumm mm: wumam .Ezcwumfisoso Ezwuomwoommo
cue: xmmmmoHQ m Hound mm¢ wusmwm CH no woman DAB mEmm .bmq madman

A.mo

 

.AThm

 

mo <0 mzx swam >0 mczo. Ga

231

.mv mudofim GM mm mc0flumfl>wunn< .Bouum man he

owumofiodfl we cflxwamsmu .mhmo m How Hwnamso umHoE M GA owumnoosfl
ocm mmuomm mHoowowmmmHQ .< SDMB owmmumm msfiwn uwumm wumHQ DAB
meow wan .uflmflm .unmfla >5 w>o3nuuozm noon: UmBmfl> .Ncwnumowflz
one .0 Cohen .wxmq Ruse .Huds .Hnmmm mmmmuoow omumHDUOCfi
amHoofiowWMMHQ .< mo muomuuxm mo madam Ode .ummq .coHDUDooum
cflxmamfino so mHoowowmmmHQ MHMMCHmuH< mo muowwwm .vv wusmflm

 

zoqsmooaqes M 2 o nemouaqes x

232

in leaf tissue. Conidia formed unusually short germ tubes,
and penetration events were rare. On the resistant ecotype
RLD, conidial germ tubes were about as long as on the
susceptible ecotype Kas-l. However, fewer lesions and signs
of browning were seen in RLD than in Kas-l cells.

Micrographs of RLD leaves viewed under UV light revealed the
presence of a fluorescent orange material permeating the
tissue, which may indicate the presence of phenolic compounds

(R. Hammerschmidt, pers. comm.).

Discussion

These studies provided no earth-shattering conclusions
about the role of camalexin in disease resistance. All
ecotypes had the ability to produce camalexin. However, they
produced little in response to A. brassicicola. The only
ecotypes producing detectable amounts of camalexin in
response to A. brassicicola were a highly susceptible one
(Kas-l) and two highly resistant ones (RLD and Anna).
Camalexin does not seem, therefore, to have a significant
role in resistance to A. brassicicola. However, it is
possible that it has a role in resistance to other pathogens.

The variations in symptoms demonstrated the need to
consider the model of the disease triangle (Agrios, 1997)
when evaluating disease resistance. Even if a plant has the
capacity to resist infection, disease can develop if the

environmental conditions are right and the pathogen is

233

aggressive enough. Similarly, in the case of the ecotypes
studied here, differences in light regime or inoculum
concentration led to aberrant disease phenotypes.

When disease phenotypes were evaluated under similar
conditions between experiments, some general trends emerged,
such that it was possible to classify plants in terms of
high, intermediate, or low resistance. These categories
encompassed a range of symptoms, which varied within each
ecotype from one experiment to the next.

Microscopic analysis of infection revealed differences
in resistance responses. Resistance in Turk Lake appeared to
be due to inhibition of fungal growth at or near the leaf
surface. In contrast, resistance in RLD appeared to be the
result of cell response after penetration. It may be that the
orange fluorescent material seen in RLD tissue represents
phenolics or other defense-associated compounds forming in
response to infection.

Similar differences in resistance responses were noted
for the Col-0 and RLD ecotypes in response to Peronospora
,parasitica (Mauch-Mani et al., 1993). In RLD, an HR occurred
near the site of penetration, and no spread of disease
occurred. In Col-0, some spread of the pathogen occurred,
demonstrated by a trail of necrotic flecks across the leaf
(Mauch-Mani et al., 1993). These variations were thought to
be due toimultiple resistance genes. The variations in

response to A. brassicicola seen in these studies may also be

234

the result of differential activation of multiple resistance
genes.

Because no antimicrobial compounds besides camalexin
were detected in TLC plate bioassays of A. brassicicola-
inoculated leaves, it is difficult to determine the source of
disease resistance in resistant ecotypes. It may be that the
antimicrobial compounds in leaves were too labile to be
detected, or that they were too polar to be extracted with
chloroform. It is also possible that resistance to A.
brassicicola depends primarily on the accumulation of the
defensins which have been found in Arabidopsis inoculated
‘with A. brassicicola (Penninckx et al., 1996) and which,
being proteins, would not be isolated by the methods used in
these studies.

This part of the project demonstrated the importance of
studying more than one ecotype to evaluate resistance of a
plant to a pathogen. It also demonstrated that camalexin is
not a reliable marker for resistance to A. brassicicola,
although it may be for resistance to other pathogens. The
variations in camalexin production in response to C. carbonum
demonstrated, as did the kinetic studies of chapter 3, that
patterns of accumulation are not set in stone and that it is
wise to study such patterns on more than one occasion before

drawing conclusions about camalexin deficiency.

235

REFERENCES

Agrios GN. 1997. Plant Pathology, 4th ed. New York:
Academic Press.

Boyd Rs, Shaw JJ, Martens SN. 1994. Nickel
hyperaccumulation defends Streptanthus polygaloides
(Brassicaceae) against pathogens. .American JOurnal of Botany'
81(3): 294-300.

Callaway A, Liu WN, Andrianov V, Stenzler L, Zhao JM,
wettlaufer S, Jayakumar P, Howell SH. 1996.
Characterization of cauliflower mosaic virus (CaMV)
resistance in virus-resistant ecotypes of Arabidopsis.
Molecular.PlanteMicrobe Interactions 9(9): 810-818.

Conn KL, Tewari JP, Dahyia JS. 1988. Resistance to
Alternaria brassicae and phytoalexin-elicitation in rapeseed
and other crucifers. Plant Science 56: 21-25.

Dempsey DA, Pathirana MS, WObbe KK, Klessig DF. 1997.
Identification of an Arabidopsis locus required for
resistance to turnip crinkle virus. Plant Jburnal 11(2):

Dempsey DA (1996) Characterization of the Responses of
' ' to Tarnip Crinkle Virus. Ph.D.
dissertation. Rutgers State University, New Jersey.

Ellis MB. 1968. Alternaria brassicicola. CHI Description
of Pathogenic FUngi and Bacteria, no. 163.

Fuchs H, Sacristan MD. 1996. Identification of a gene in
Arabidopsis thaliana controlling resistance to clubroot
(Plasmodiophora brassicae) and characterization of the
resistance response. .Molecular’Planthicrobe Interactions
9(2): 91-97.

Gerson EA, Kelsey RG. 1998. Variation of piperidine
alkaloids in ponderosa (Pinus ponderosa) and lodgepole pine
(P. contorta) foliage from central Oregon. aburnal of
Chemical Ecology 24(5): 815-827.

Glazebrook J, Ausubel FM. 1994. Isolation of phytoalexin-
deficient mutants of Arabidopsis thaliana and
characterization of their interactions with bacterial
pathogens. Proceedings of the National.Academy of Sciences
USA 91: 8955-8959.

236

Glazebrook J, Zook M, Mert FM, Kagan I, Rogers EE, Crute IR,
Holub EB, Hammerschmidt R, Ausubel FM. 1997. Phytoalexin-
deficient mutants of Arabidopsis reveal that PAD4 encodes a
regulatory factor and that four PAD genes contribute to downy
mildew resistance. Genetics 146: 381-392.

Holub EB, Brose E, Tor M, Clay C, Crute IR, Beynon JL. 1995.
Phenotypic and genotypic variation in the interaction between
.Arabidopsis thaliana and Albugo candida. .Mblecular.Plant-
.Microbe Interactions 8(6): 916-928.

Lazarovits G, Brammall RA, Ward EWB (1982) Bioassay of
fungitoxic compounds on thin-layer chromatograms with Pythium
and Phytophthora species. Phytopathology 72(1): 61-63.

Lehle Seed Company Arabidopsis Catalogue. [Online] Available
http://www.arabidopsis.com/maillost/seeds/wildtypes

Leisner SM, Howell SH. 1992. Symptom.variation in different
Arabidopsis thaliana ecotypes produced by cauliflower mosaic
virus. Phytopathology 82: 1042-1046.

McRoberts N, Lennard JH. 1996. Pathogen behaviour and plant
cell reactions in interactions between Alternaria species and
leaves of host and nonhost plants. Plant Pathology 45: 742-
752.

Mauch-Mani B, Croft KPC, Slusarenko A. 1993. The genetic
basis of resistance to Arabidopsis thaliana L. Heyhn to
Peronospora parasitica. In: Davis KR, Hammerschmidt R, eds.
AEQQLQQE§i§_£fldlidnd as a.Model for'Plant-Pathogen
Interactions. St. Paul, MN: The American Phytopathological

Nottingham Arabidopsis Seed Catalogue. [Online] Available
http://nasc.nott.ac.uk/

Osbourn AE, Clarke BR, Lunness P, Scott PR, Daniels MJ.
1994. An oat species lacking avenacin is susceptible to
infection by Gaeumannomyces graminis var. tritici.
Physiological and Mblecular’Plant.Pathology 45: 457-467.

Pedras MSC, Khan AQ, Taylor JL. 1997. Phytoalexins from
brassicas: overcoming plants'defenses. In Hedin PA,
Hollingworth RM, Masler EP, Miyamoto J, Thompson DG, eds.
Phytochemicals far Pest Cbntrol. AC8 Symposium Series 658,
155-166.

Penninckx IAM, Eggermont K, Terras FRG, Thomma BPH, de
Samblanx GWD, Buchala A, Metraux J-P, Manners JM, Broekaert
WF. 1996. Pathogen-incuced systemic activation of a plant
defensin gene in Arabidopsis follows a salicylic acid-
independent pathway. Plant cell 8: 2309-2323.

237

Rogers EE, Glazebrook J, Ausubel FM. 1996. Mode of action
of the Arabidopsis thaliana phytoalexin camalexin and its
role in Arabidopsis-pathogen interactions. .Molecular.Plant-
.Microbe Interactions 9(8): 748-757.

Tsuji J, Somerville SC, Hammerschmidt R. 1991.
Identification of a gene in.Arabidopsis thaliana that
controls resistance to xanthomonas campestris pv. campestris.
Physiological and.Mo1ecular.Plant Pathology 38: 57-65.

Zhao Jianmin, Last RL. 1996. Coordinate regulation of the
tryptophan biosynthetic pathway and indolic phytoalexin
accumulation in Arabidopsis. Plant Cell 8: 2235-2244.

Zhou N, Tootle TL, Tsui F, Klessig DF, Glazebrook J. 1998.

PAD4 functions upstream from.salicy1ic acid to control
defense responses in Arabidopsis. Plant cell 10: 1021-1030.

238

Conclusions

The biosynthetic intermediates of camalexin, apart from
anthranilate and cysteine, are still unknown. Indole-3-
carboxaldehyde may be a precursor, but the radiolabeling data
do not seem too supportive. The attempt to seek biosynthetic
intermediates in camalexin-deficient mutants demonstrated
that apparent camalexin deficiency can occur even in wild-
type plants, depending on inoculum concentration, the use of
leaves or droplets alone, and unknown factors causing
camalexin to accumulate more slowly. The work with
Arabidopsis ecotypes demonstrated that the choice of pathogen
affects camalexin production, and it reaffirmed the variable
kinetics of camalexin accumulation found in work with the pad
mutants.

A.few questions remain as a result of this project. The
question of biosynthetic intermediates is an obvious one.
Since indole-3-carboxaldehyde does not seem a strong
candidate for an intermediate, it may be time to look more
closely at the role of indole-3-carboxylic acid in camalexin
biosynthesis. On the autoradiograms shown in chapter 4

(Figures 25-30) , bands at the approximate Rf of indole-3-

carboxylic acid were present, and the relative darkness of

those bands did vary over time. It is possible that indole-
3-carboxylic acid leads to camalexin by reduction to indole-
3-carboxaldehyde or formation of an acetyl CoA thioester, as

occurs in lignin biosynthesis. Indole is another possible

239

intermediate. Given the volatility of indole, the extraction
procedure may need to be modified extensively. Indole-3-
carboxaldehyde may be worth reinvestigating if the
radiolabeled form can be made.

Both the time courses described in Chapter 3 and the
large-scale extractions described at the beginning of Chapter
4 generated several inducible bands besides indole-3-
carboxaldehyde on TLC plates. Characterizing those may
reveal biosynthetic intermediates.

Since pad 2 does not appear to be a regulatory mutant,
it may be a leaky biosynthetic mutant. Radiolabeling with pad
2 was only done with a 24-hour incubation period after

feeding and inoculation. Perhaps a time course of camalexin
accumulation in pad 2 leaves fed 14C-anthranilate would yield

intermediates at early time points.

Since camalexin concentrations in leaves sometimes
decreased during 72-hour time courses, it would be
interesting to determine the metabolic fate of camalexin.
Some of the inducible bands on TLC plates may be metabolites
and not biosynthetic precursors. Knowing the metabolites of
camalexin would help to distinguish between potential
precursors and metabolites in further radiolabeling studies.
Also, those metabolites may have some role in disease
resistance.

Ultimately, the purpose of determining the camalexin
biosynthetic pathway is to understand its role in disease

resistance. The work with Arabidopsis ecotypes may provide

240

answers without a firm.knowledge of the biosynthetic pathway.
Little correlation was found between camalexin production and
resistance to Alternaria brassicicola, but it is possible
that A. brassicicola is not a pathogen of those ecotypes in
their natural habitat. It would be interesting to find out
what pathogens--compatible and incompatible--are present in
the natural habitats of the ecotypes studied, and to
determine if resistance or susceptibility are correlated with
camalexin production. If little camalexin is produced in
response to incompatible pathogens, perhaps a search for
other antimicrobial compounds, like those that appeared on
some TLC plate bioassays, would help to explain resistance.
The presence of a fluorescent orange compound in infected
leaves of the RLD ecotype suggests that other compounds are
being produced.

The relative insensitivity of A. brassicicola to
camalexin raises the question of whether resistance to a
pathogen is correlated with the pathogen's sensitivity to
camalexin. This question could also be examined in a study
of ecotype resistance to native pathogens.

It may also be helpful to study other compounds that
Arabidopsis produces during infection. Looking only at
camalexin production leads to a narrow picture of the
responses of Arabidopsis to infection. A more complete view
‘would require studying the types of other compounds produced
(phenolics, defensins, or PR proteins) and the relationship

between the timing of their production and camalexin

241

production. Eventually, it would be interesting to see
compare results of such studies on Arabidopsis and on other
camalexin-producing genera (Camelina sativa and capsella

.bursafipastoris).

242

Appendix A. Recipe for half-strength Hoagland's solution used
to fertilize Arabidopsis plants in the experiments described
in Chapters 2-4 (courtesy of J. Klug, Michigan State
University).

243

Half-strength Hoagland's Solution
The following solutions are prepared.
Solution A:

calcium nitrate (Ca[N03]-4H20), 295.0 g/l
sequestrene (DTPA, 10% Fe), 38.44 g/l

Solution B:

potassium phosphate (KH2P04), 34.25 g/l
potassium nitrate (KNO3), 126.65 g/l
magnesium sulfate (Mgso4o7H20), 62.5 g/l
zinc sulfate (ZnSO4-7H20), 0.056 g/l
manganous sulfate (MnSO4:H20), 0.391 g/l
copper sulfate (CuSO4 :4H20), 0.021 g/l
boric acid (H3B03), 0.725 g/l

molybdic acid (M003- ZHZO), 0.0059/1

The pH of solution B is adjusted to about 4.6.
For 1 l of nutrient solution, 2 ml of solution A and 2 ml of
solution B are mixed in 996 ml of water. The pH is then
adjusted to 6.0-6.4 with 1N KOH (about 0.25 ml of KOH
required to adjust pH for 1 liter of solution).

Stock solutions A and B should not be mixed unless diluted.

244

      

   

      

  

HICHIGQN STRTE UNI V IBIERRR

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII IIIIIII