Download COMMENTARY: Why do pathogens carry avirulence genes?

Physiological and Molecular Plant Pathology (1999) 55, 205–214
Article No. pmpp.1999.0230, available online at http:\\www.idealibrary.com on
COMMENTARY
Why do pathogens carry avirulence genes ?
D. W. G A B R I E L
Plant Molecular and Cell Biology Program, Plant Pathology Department, UniŠersity of Florida, GainsŠille, FL 32611-0680,
U.S.A.
(Accepted for publication June 1999)
In the gene-for-gene hypothesis, Flor [19 ] proposed the existence of avirulence alleles of virulence genes,
and that the virulence genes actively conditioned cultivar-specific virulence on hosts. Yet cultivar-specific
virulence genes have not been found, and many of the avirulence (aŠr) genes cloned to date (over 40) do
not appear to condition pathogenicity in general, virulence or anything else of value to the microbe. There
is strong indirect evidence that many, if not nearly all, Avr proteins are secreted from pathogenic bacteria,
enter plant cells and signal a response phenotype, usually associated with a hypersensitive defense response
(HR). Why the bother for avirulence ?
Evidence is accumulating that most aŠr genes are, or once were, pathogenicity ( pth) genes found in
biotrophic pathogens that determine (d ) host range, not in a cultivar-specific manner, but in a host speciesspecific manner. At least some of these genes appear to function for pathogenicity by encoding protein
signals that are ‘‘ injected ’’ into plant cells by the hrp system, resulting in programmed host cell death, a
characteristic normally associated with necrotrophs. A growing body of evidence indicates that most
microbial genes conditioning pathogenicity, including the hrp, pth and aŠr genes, are present because of
horizontal gene transfer, often involving movement of large gene clusters on ‘‘ pathogenicity islands ’’.
Since horizontal transfer is a stochastic process, many aŠr genes are likely to be maladapted pth genes,
following their horizontal transfer to strains in which their function may be either gratuitous or
detrimental. The structure of some of these genes may allow rapid adaptive selection for pathogenic
function.
# 1999 Academic Press
Keywords : avirulence ; avr ; pathogenicity ; virulence ; harpins ; hrp ; pathogens ; gene-for-gene ;
Xanthomonas ; Pseudomonas ; Ralstonia ; Erwinia ; horizontal transfer ; biotroph ; necrotroph.
INTRODUCTION
Plant breeders look for genes by which plants can defend
themselves against microbial attack, and microbial
pathologists look for genes in microbes that determine
virulence. The breeding work resulted in a large number
of studies documenting intraspecific variation in plant
resistance, and the classical microbiology research resulted
in more limited studies on intraspecific variation in
microbial virulence. All classical genetic analyses were by
necessity limited to intraspecific variation analysed by
meiotic recombination. Until the advent of molecular
genetics, a comprehensive and systematic approach for
understanding why certain plant species and not others
are hosts of certain pathogens, and how microbes become
pathogens in the first place, was impossible. Is it due to a
failure of resistance (pathogen recognition) or the acquisition of pathogenicity genes ? Can saprophytes become
pathogens quickly, and can pathogens change host range
or become much more pathogenic quickly, or do these
events require complex combinations of adaptations,
requiring long evolutionary periods ? Such questions are
0885–5765\99\100205j10 $30.00\0
now being addressed experimentally and new paradigms
are emerging to explain the origins of new disease
epidemics and to better control plant disease.
CLASSICAL TERMINOLOGY, NEODARWINISM AND THE AVIRULENCE GENE
DILEMMA
Naturally-occurring variation in degree of susceptibility
or resistance in a particular host species to a particular
pathogen has been shown to be controlled, in nearly all
cases, by single resistance (R) genes. R genes trigger
programmed plant defense responses, including cell death
[21, 11, 75 ], most often resulting in the appearance of a
plant HR. Naturally-occurring variation in the phenotype
of microbial virulence on crop plants was also found to be
determined by single genes. Harold Flor’s classic papers
[19, 20 ] using flax rust (Melampsora lini) demonstrated
that ‘‘ pairs of factors ’’ (specific avirulence and virulence
alleles) in the pathogen determined virulence against each
particular host (cultivar-specific) R gene, thus establishing
# 1999 Academic Press
D. W. Gabriel
the gene-for-gene hypothesis. The dilemma posed by the
concept of dominant avirulence genes is implicit in Flor’s
writing. Although Flor first used the designation A to
"
refer to the dominant avirulent allele of the ‘‘ pair of rust
conditioning factors ’’ specific for the cognate host gene R ,
"
his emphasis was always on the recessive virulence alleles
as being the active determinants of virulence and
pathogenicity. Compounding the conceptual dilemma,
Flor also demonstrated that avirulence\R gene interactions were phenotypically epistatic over any and all
virulence\susceptibility gene interactions.
Since that time, many ‘‘ pairs of factors ’’ (still usually
described as virulence genes) were discovered by classical
genetic analyses, and the gene-for-gene hypothesis was
demonstrated or suggested in many, primarily biotrophic,
parasite\host systems [12 ]. Despite the fact that avirulence
alleles were dominant and epistatic, Ellingboe’s idea that
pathogens carried a number of genes where the active
allele conditioned avirulence ‘‘ superimposed upon a basic
compatibility between host and parasite ’’ [17 ], while the
inactive allele allowed (but did not condition) the virulent
state was not widely accepted until the first avirulence
genes were cloned in the 1980s. The primary objections
raised were based on neo-Darwinian selection arguments :
there could not be selection for avirulence, therefore the
genes must function for virulence. For example, Person &
Mayo writes : ‘‘ Although avirulence usually segregates as
a dominant character, microevolutionary history suggests
that in its primary function an A gene associates not with
avirulence but with virulence …’’ [50 ]. In the only and
classic textbook on the subject, Day’s ‘‘ Genetics of HostParasite Interactions ’’, avirulence as a phenotype is
explicitly compared to microbial auxotrophy and to
sensitivity to a poison [12 ]. In this widely accepted view,
the avirulent pathogen must be lacking something a
virulent pathogen has, exactly analogous to lacking an
appropriate allele to grow on minimal medium or to
break down a poison.
The cloning of the first avirulence gene by Staskawicz
et al. [60 ] from a readily cultured bacterial pathogen
(Pseudomonas syringae), overcame the objections caused by
the difficulty of accepting the idea that microbial
pathogens carry genes that primarily and simply function
to limit their virulence. One of the hallmarks of that paper
was the deliberate search not only for an avirulence gene,
but also for a virulence gene sensu Flor, a gene that
actively conditioned virulence against a specific resistant
host cultivar. Cultivar-specific virulence genes were not
found. In fact, not even a DNA fragment hybridizing to
the aŠr gene was found in other P. syringae strains in that
study that might have indicated a corresponding ‘‘ allele ’’
in the species. Since then, many avirulence (aŠr, following
bacterial nomenclature) genes (over 40) have been cloned
from a variety of different bacterial plant pathogens
[20, 42, 70 ] with similar results. Furthermore, many gene
Host cultivar with
Pathogen strain with
206
R1–
r1r1
avr1
–
+
∅
+
+
F. 1. The gene-for-gene ‘ quadratic check ’, as observed with
phytopathogenic bacteria. The host R gene is genetically
"
specific for the pathogen aŠr gene, usually resulting in a
"
hypersensitive defense response, indicated as an incompatible
(k) interaction. Cultivar-specific avirulence appears to be
superimposed on a basic pathogenic ability (on hosts) and
results in a compatible (j) interaction. The pathogen may have
strains with an alternate aŠr allele, but molecular data from
phytopathogenic bacteria indicate that this is not necessarily the
case (refer text). The shaded area is to emphasize the fact that
the alternate allele, called the ‘‘ virulence ’’ allele and assumed to
play a central role to condition virulence in classical genetic
analyses, may not exist.
disruption experiments of bacterial aŠr genes in several
labs failed to reveal any effect on growth or pathogenicity
of the microbial strain on (or off) hosts, other than the
expected loss of R gene-specific avirulence. That is, the
majority of bacterial aŠr genes appear to be dispensable.
Although only a few fungal avirulence genes have been
cloned to date [38 ], similar data emerges. With many
microbial aŠr genes, there is no evidence for alleles or
analogues that condition anything of value to the
pathogen. Such aŠr genes appear to be not only
dispensable, but also gratuitous. At the very least, cultivarspecific virulence in many gene-for-gene interactions is
determined by the mere absence of aŠr genes ; virulence alleles are not necessary for cultivar-specific
virulence (Fig. 1).
It became clear that use of the term ‘‘ virulence gene ’’
to refer to the hypothetical alternate allele of cultivarspecific aŠr genes was a misnomer ; in R gene for aŠr gene
interactions, a ‘‘ virulence ’’ allele often does not exist, and
if it does, does not condition the cultivar-specific state it
describes. Unfortunately, there is considerable confusion
in the literature brought about by use of the term
‘‘ virulence ’’ as synonymous with ‘‘ pathogenicity ’’. For
example, in their review of the terms virulence and
pathogenicity, Shaner et al. [56 ] recount that three wellknown scientists gave three different definitions of the
term ‘‘ virulence ’’ at a 1989 international congress. With
the exception of the virulence (Šir) genes of Agrobacterium
tumefaciens, formal use of the term ‘‘ virulence ’’ as
synonymous with ‘‘ pathogenicity ’’ is incorrect. The Šir
genes of A. tumefaciens control pathogenicity and host
range, but in other plant pathogenic microbes, these
AŠirulence genes
would be called pathogenicity genes, since these Šir genes
are not cultivar-specific. The term ‘‘ pathogenicity ’’ is
generally used with reference to genes that function to
condition that state, and this terminology will be used
throughout this paper. The term ‘‘ virulence ’’ will only be
used to refer to cultivar-specific situations, keeping in
mind that there are no known ‘‘ Šir ’’ genes by this
definition.
WHY ARE THERE SO MANY avr GENES ?
Some avr genes are pth genes
There appear to be several valid explanations for the
enigma of why pathogens carry aŠr genes, and the
explanations are not mutually exclusive. First, some aŠr
genes have selective value in terms of pathogenicity
[22, 38, 42 ]. That is, they function pleiotropically to
condition pathogenicity on all hosts tested (usually with
increased pathogen growth in planta) in the absence of a
cognate (‘‘ recognized ’’ in a formal, gene-for-gene sense)
R gene. It is important to emphasize that the pathogenicity, as opposed to the avirulence, encoded by these
genes, appears from limited data to include entire host
species. For example, pthA, the first member of the
Xanthomonas aŠrBs3\pthA gene family to be recognized as
being essential for pathogenicity, is required for the
production of necrotic cankers on all species of citrus
attacked by X. citri [62 ]. Similarly, but to a quantitatively
lesser extent than pthA on citrus, aŠrA and aŠrE of
Pseudomonas syringae pv. tomato are important for pathogenicity on tomato [44 ]. Some aŠr genes, such as many of
those found in X. campestris pv. malvacearum, do not
exhibit obvious selective value when examined individually, but when studied together, exhibit weak,
additive and synergistic effects on pathogenicity on cotton
[76 ]. Altogether, aŠr genes with a confirmed or suspected
pleiotropic selective value on susceptible host species may
account for about a third of the total cloned and
investigated to date. Yet it remains striking that the
majority of aŠr genes cloned and reported to date appear
to be dispensable.
Stochastic eŠents, and particularly horizontal gene transfer
Another explanation for the enigma is that many aŠr genes
may be present because of a variety of stochastic events,
including horizontal gene transfer, duplication and
mutation. Horizontal gene transfer is becoming recognized
as a major epidemiological factor in new disease outbreaks
[48 ], and there is good reason to believe that many aŠr
genes are gratuitously present in a variety of bacterial
strains following horizontal transfer from other species.
For example, the % GC content of every one of the aŠr
genes cloned from P. syringae is low for the species [70 ].
What if pthA from X. citri is moved by conjugation into X.
207
compestris pv. malvacearum ? The result is not the
production of necrotic cankers on cotton, but instead is
gene-for-gene avirulence and elicitation of the HR on
cotton [63 ]. One probable example of a natural horizontal
gene transfer scenario involves aŠrBs3, the first member
cloned of the aŠrBs3\pthA gene family. This gene, which
was isolated as an aŠr gene from an X. campestris pv.
vesicatoria strain and is evidently not needed for
pathogenicity on any host by any strain of this pathovar,
resides on a self-mobilizing plasmid [7 ].
Interspecific and intergeneric transfers of aŠr genes
could occur simply because of coincidental linkage with
another factor on the same plasmid that has selective
value, for example copper resistance. Copper sprays are
widely used in some areas of the country, and copper
resistance has been found linked to aŠr genes on an X.
campestris pv. vesicatoria self mobilizing plasmid recovered
from strains in Florida [59 ].
Other stochastic events may account for the presence of
dispensable aŠr genes. Redundancy through duplication is
especially indicated for members of the aŠrBs3\pthA gene
family. Sixty-two bp terminal inverted repeats mark the
boundaries of homology among all members of this gene
family, and the terminal 38 bp of these inverted repeats
are highly similar to the 38 bp consensus terminal sequence
of the Tn3 family of transposons [14 ]. It is therefore
possible that these genes can, or once could, transpose. A
different stochastic event might be loss of host range on a
particular wild plant species due to mutation and
adaptation to new plant species, leaving behind functional
but unnecessary genes that once assisted in conditioning
pathogenicity on the original plant species. Evidence for
nonfunctional relics of aŠr genes in both Pseudomonas and
Xanthomonas have been reported, for example, [39, 76 ]. Of
course, experiments designed to reveal potential pathogenicity function(s) of aŠr genes are conducted on known
hosts of the pathogen in which the genes currently reside.
Recent advances in our analyses of microbial genomes
has revealed that they are much more fluid than previously
thought, and horizontal gene transfer between species is
thought to occur to a greater extent than generally
assumed [57 ]. In fact, one of the primary surprises to
come as a result of recent extensive genomic sequencing
analyses is the discovery that both prokaryotic and
eukaryotic genomes are chimeras, with large blocks of
genes from a variety of sources, sometimes even mixtures
from prokaryotic and eukaryotic sources [41 ]. It seems
that horizontal transfer of clusters of genes conferring
adaptive phenotypes in bacteria is the rule, not the
exception [53 ]. Horizontal transfer of gene clusters is not
limited to prokaryotes : horizontal transfer of an entire
supernumerary chromosome has been demonstrated in
Colletotricum [30 ] ; supernumerary chromosomes are known
to carry genes that condition growth and\or pathogenesis
in some filamentous plant pathogenic fungi [10, 67 ].
208
D. W. Gabriel
Evidence for the horizontal transfer of many aŠr genes
is indicated by their function (avirulence), structural
features, % GC content, residency on plasmids, lack of
conservation within pathogenic groups and the fact that
so many appear to be dispensable. However, as several
authors have noted, the fact that nearly all bacterial aŠr
genes require the specific action of the hrp genes (discussed
below) in order to express the avirulence phenotype, leads
back to the conclusion that most of these genes must have
originally been elaborated for pathogenic purposes.
Even the hrp system itself seems to be acquired by
horizontal gene transfer. The function of most hrp genes is
to provide for a specialized protein secretion system called
type III [31, 68 ]. This type III secretion system is a
generally conserved mechanism found in both plant and
animal pathogens for the export, secretion and often
delivery of specific proteinaceous effector molecules (virulence or pathogenicity factors) directly and\or indirectly
into host cells in a contact-dependent manner [24, 31, 46 ].
The Salmonella typhimurium secretion system appears to form
a hollow, needle-like structure [40 ], that functions as a
toxic protein injection system [58 ]. The hrp pili may form
channels through which macromolecules travel to reach
the plant cell cytoplasm [43 ]. Interestingly, widely diverse
bacteria probably acquire the entire export\injection
system on a ‘‘ pathogenicity island ’’ in a single horizontal
gene transfer event [48 ]. The evidence for this is that the
genes encoding these protein secretion systems in various
animal and plant pathogens are all clustered and the
codon usage and % GC content of the genes encoding
these type III systems are typically different from that of
the rest of the genome.
Pathogenicity islands are not limited to phytopathogenic bacteria. A pathogenicity island consisting of at
least six genes and differing in codon usage and % GC
content from that of other chromosomal genes in the
fungal pea pathogen, Nectria haematococca, has been found
on a dispensable, supernumerary chromosome [37 ].
Interestingly, the genes of this cluster appear to be
involved in both offense and defense : some function to
condition pathogenicity by killing host cells, while others
function to degrade pisatin, a phytoalexin elicited in
response to the fungal attack [29 ]. Also found on the same
dispensable chromosome a gene allowing utilization of
homoserine as a sole carbon and nitrogen source [54 ].
Homoserine is one of the nutrients found in relative
abundance in pea root exudates and also released from
dying pea cells. As mentioned earlier, at least one fungal
supernumerary chromosome is known to be capable of
horizontal transfer [30 ] ; the large number of genes related
to pathogenicity found on the N. haematococca chromosome
indicates that many, if not all of the tools required for
pathogenicity and host range might be transferred to a
nonpathogenic saprophyte in a single event. Evolution
from saprophyte to pathogen might not always involve a
long and complex series of incremental improvements in
pathogenicity [32 ] as is commonly assumed, but rather
could be achieved in a single horizontal transfer events
(acquisition of a pathogenicity island or two) followed by
a ‘‘ clean up ’’ operation, involving mutational shedding of
maladapted aŠr genes that are on the island.
WHAT DO avr GENES DO BIOCHEMICALLY ?
The capacity of aŠr genes to trigger programmed cell
death (PCD) and defense responses, plus their dependence
on the hrp secretion system, indicates that most aŠr genes
encode plant cell signal proteins that are perceived by
receptors and trigger signal transduction pathways
[8, 11, 69, 75 ]. The hypothesis that aŠr genes either are, or
once were, pth genes indicates that pathogenicity function
is also by way of plant cell signal proteins that are
perceived by receptors and trigger signal transduction
pathways. The fact that the defense responses triggered by
aŠr genes appear to be programmed responses strongly
indicates that, if they also have a pathogenic purpose, it
must be to trigger (a) programmed event(s) in the plant
cell other than defense. Since aŠr genes are found primarily
in those pathogens thought to be biotrophic, their purpose
must be related to the pathogenicity lifestyle unique to
biotrophs.
WHAT IS REQUIRED FOR PATHOGENICITY ?
Once it is realized that aŠr\pth genes encode signals
affecting plant cell programs, it is obvious to ask what
kinds of programs might enhance a biotrophic pathogenic
lifestyle. This involves a consideration of two related
questions : What adaptations are required on the part of
microbes to become biotrophic pathogens, and what is
required by a biotroph to attack a specific plant species,
i.e. to make the species a host ? Biotrophic pathogens,
whether of plants or animals, appear able to cause
changes in host cell functions that increase both the
pathogen’s ability to reproduce and to survive. Ability to
reproduce involves deriving nutrition from plants, considered in the next section, and ability to survive involves
ability to move to new infection sites, considered in the
following section.
Ability to deriŠe nutrition
The essential difference between parasites and saprophytes
is that parasites derive nutrition from a living plant at its
expense. The essential difference between pathogens and
parasites is that pathogens cause overt disease symptoms.
In many cases, the disease symptoms appear to directly
favour the pathogen, at least in terms of pathogen
dispersal. Plant pathogens have traditionally been roughly
divided into two groups : those which are necrotrophic
AŠirulence genes
and those which are biotrophic. The distinction is thought
to be useful in determining the primary means by which
nutrition is derived. For example, necrotrophs are easy to
culture, produce abundant degradative enzymes and
toxins, tend to have limited host cell contact, hardly
stimulate host protein synthesis, and cause early necrosis,
seemingly in advance of colonization. By contrast,
biotrophs are often difficult to culture, produce few
degrading enzymes or toxins that are important in
pathogenicity, have extensive contact with living host
cells, greatly stimulate host protein synthesis and cause
necrosis only after extensive colonization [35 ].
Most pathogenic bacterial biotrophs kill plant cells and
thereby derive major nutritional benefits. An exception is
Agrobacterium, which ‘‘ engineers ’’ its hosts to produce
novel compounds that only it can catabolize [36 ]. Note
that A. tumefaciens does not utilize type III secretion as a
primary means of pathogenicity and that there are no aŠr
genes in A. tumefaciens. Another possible exception is
Rhizobium, which can be pathogenic under high nitrogen
conditions, but which normally confers an advantage to
its hosts, and is more properly considered as a mutualistic
parasite. As a general rule, the essential difference between
necrotrophs and biotrophs that kill host cells is that
necrotrophs kill at a distance, while pathogenic biotrophs
kill after intimate cellular contact. Both types of pathogen
derive nutritional benefits from dead or dying cells.
Interestingly, it is the contact-dependent hrp system that
provides the ability for biotrophs to kill both host and
nonhost cells.
Of all phytopathogenic bacteria that require hrp genes,
members of the genus Xanthomonas are by far the most
biotrophic. Unlike Pseudomonas, Ralstonia and Erwinia,
members of the genus Xanthomonas are all plant-associated,
nearly all growth is endophytic, and each strain is highly
host-specific. In fact, among prokaryotic plant pathogens,
Xanthomonas exhibits by far the highest known level of
plant specialization, with over 120 pathovars. The second
most biotrophic group is P. syringae, with over 40 pathovars
and with strains typically exhibiting less host specificity
than a typical xanthomonad. Is there a significant
difference in the nature of the effector molecules secreted
by Xanthomonas and P. syringae strains that reflects a more
biotrophic lifestyle ?
It is clear that one of the most significant classes of effect
or molecule delivered by type III secretion is the harpins.
Harpins are glycine rich, cysteine-lacking proteins that
are secreted in culture media under appropriate growth
conditions. Externally applied harpins from E. amyloŠora,
E. chrysanthemi, R. solanacearum and P. syringae have the
capacity to elicit a rapid K+ efflux and H+ influx exchange
reaction (XR) that results in the alkalinization of the
apoplast, sucrose leakage and also PCD [1 ]. Harpins are
therefore toxin-like. Harpins are clearly important pathogenicity factors for E. amyloŠora (fire blight) [71 ] and E.
209
chrysanthemi (soft rot) [6 ], and are also found in phytopathogenic ralstonias and pseudomonads [1 ]. Their role
in the latter two genera is unclear [1, 2 ]. The fact that
harpins cause alkalinization of the apoplast, sucrose efflux
and death of host cells indicates a primary role in
nutrition acquisition. Although individual and double
harpin gene knockout mutations in P. syringae did not
affect pathogenicity, P. syringae appears to produce
multiple harpins [9 ], which might compensate the loss of
one or more (but not all) of the genes. Do even the more
biotrophic pathogens, such as P. syringae, have to kill cells
for nutrition ? The surprising answer seems to be ‘‘ yes ’’ ;
growth of P. syringae pv. tomato in tomato plants is
significantly inhibited in plants in which PCD is inhibited
[25, 26 ].
It is possible that acquisition of nutrients from living
cells by P. syringae is a two stage process : one stage that
can involve harpins and a later stage that must involve
PCD. Some nutrition can be derived from plant cells by
way of harpins without killing plant cells via an exchange
reaction (XR) that results in sucrose leakage from the
plant cell. P. syringae induces an XR and increased efflux
of sucrose from host cells within 2–4 h after inoculation
[4 ]. This XR exchange precedes onset of P. syringae
population growth [5 ], which strongly indicates that it is
a required step for growth.
An XR reaction (and resulting nutrient efflux) is also
induced by Xanthomonas, and possibly by way of (a)
completely different mechanism(s). No harpins have been
reported to date from Xanthomonas, and by contrast with
induction in a matter of hours by Pseudomonas and Erwinia,
the XR reaction induced by Xanthomonas campestris pv.
malvacearum (bacterial blight of cotton) does not occur
until 1–2 days after inoculation [65 ]. Since Xanthomonas is
well into logarithmic population growth by this time [49 ],
the XR does not appear to precede the onset of population
growth. The Xanthomonas XR may be induced by
ammonium ion accumulation in the intercellular space of
the apoplast [65 ]. Ammonium ions do not kill green plant
cells grown in light, because they are rapidly recycled to
amino acids by way of the GS-GOGAT pathway. It is
reasonable to assume that the nutrient release induced by
the XR contributes to endophytic growth of Xanthomonas.
However, as with P. syringae, it is unclear if the XR is
essential for early stage growth of Xanthomonas. There are
also at least two other potential mechanisms that may
cause nutrient efflux at an early stage of infection :
extracellular proteases [15 ] and additional effectors
secreted by the hrp system.
It seems likely that there are effector molecules secreted
by the Xanthomonas hrp system that are not yet discovered.
X. campestris pv. malvacearum is found close to host cell
walls as early as 12 h after inoculation [3 ] and in contact
with host cell walls by 24 h after inoculation [49 ]. Thus,
there is the potential for hrp system involvement in
210
D. W. Gabriel
pathogenesis even before the XR. Not all xanthomonads
kill host cells [45 ], however, the xanthomonads that cause
the major disease problems (cankers, blights and rots) do
kill host cells. For example, cotton blight causes vesiculation between the plasmalemma and cell wall by 48 h
after inoculation, and major degenerative changes resulting in cell death occur over the next 3 days [3 ]. Rapid
host cell death per se does not inhibit X. campestris pv.
malvacearum growth in cotton, since plants held in
continuous darkness became necrotic, while permitting
logarithmic bacterial growth for 5 days [49 ].
The ability of a pathogen to produce an externally
applied, toxic compound that results in cell death is a
characteristic of necrotrophs : they kill in advance of
colonization in order to obtain nutrition. If P. syringae
strains do not use harpins to kill plant cells and Xanthomonas
strains do not even have harpins, what do they use ?
Necrotrophs do not carry gene-for-gene aŠr genes, whereas
biotrophs do. Unlike harpins, the protein products of aŠr
genes fail to elicit any symptoms at all when artificially
introduced or infiltrated as purified or crude extracts into
the spongy mesophyll of plants. Could some aŠr genes,
functioning pleiotropically in a pathogenic role, be
injected into plant cells and signal susceptible hosts to give
up nutrients by causing PCD ?
Several Xanthomonas and P. syringae aŠr genes elicit an
HR when expressed inside plant cells carrying specific R
genes [13, 16, 28, 55, 61, 64, 66 ]. More importantly, both
aŠrB from P. syringae pv. glycinea and aŠrRpt2 from P.
syringae pv. tomato cause cell death and necrosis when
expressed in plants lacking cognate R genes, and
suggesting a possible role for these genes in pathogenicity
[28, 47 ]. A role for a known pathogenicity determinant,
pthA from X. citri, in programmed killing of host cells were
recently established. Expression of pthA in citrus cells
was sufficient to cause symptoms diagnostic of the disease
caused by the pathogen : division, enlargement and death
of host cells [16 ]. Since PthA is not a toxin and does not
appear to be an enzyme, the cell death that is signalled
must be programmed. Furthermore, the protein signal is
specific for citrus.
As mentioned earlier, pthA is a member of the largest
group of aŠr genes cloned and sequenced to date : the
Xanthomonas aŠrBs3\pthA gene family [22, 42 ]. Members of
the gene family, all of which are gene-for-gene aŠr genes,
are either known or thought be to be required for
pathogenicity of xanthomonads causing cotton blight,
Asiatic citrus canker, false citrus canker, Mexican lime
cancrosis, common bean blight and rice blight [22 ].
Among other symptoms, the primary disease symptom
elicited in common by those members appears to be PCD.
Such PCD would likely have to occur without the
oxidative burst and phytoalexin induction. Evidence to
support this idea has been found using X. campestris pv.
malvacearum, causing cotton blight. Cotton phytoalexins
have been convincingly shown to be sufficient for the
resistance observed in cotton against X. campestris pv.
malvacearum [52 ]. However, on susceptible cotton, host
cell death occurs without the production of phytoalexins
[49 ]. This host cell death on susceptible cotton is caused
by multiple members of the aŠrBs3\pthA gene family in X.
campestris pv. malvacearum [76 ]. As discussed above, pthA
appears to cause PCD, and it is likely that other members
of the gene family that function for pathogenicity also
cause PCD. In fact, there is strong evidence that the
oxidative burst and initiation of the HR are independent
of PCD in a number of systems [27, 33, 51 ].
Ability to moŠe to new infection sites
Since there is a common export mechanism for PCD
signals shared by animal and plant bacterial pathogens,
how do they cause such diverse diseases ? Since most of the
type III secretion system genes are involved in the export
machinery per se, the answer may again lie in the effector
molecules. The results of disease phenotype screens with
several xanthomonads yielded only a relatively few genes
that determine specific disease phenotypes [21, 23 ].
Examples are pthA of X. citri (hyperplastic cankers) and
aŠrb6 and aŠrb7 (water soaking and blight) of X. campestris
pv. malvacearum. These are all bona fide, gene-for-gene aŠr
genes.
The PthA protein signals citrus plant cells to divide. In
a natural infection, enough plant cells are affected by X.
citri to result in tissue hyperplasia, eventually resulting in
rupturing the leaf epidermis and allowing egress of the
canker pathogen to the leaf surface. This is important
because the pathogen is spread by wind-blown rain. A
similar story emerges from studies of some other aŠr genes,
affecting other pathogenicity phenotypes (reviewed in
[22, 42 ]). In a well studied case, aŠrb6 and other members
of the aŠrBs3\pthA gene family found in X. campestris pv.
malvacearum cause water soaked, necrotic lesions on
cotton, a phenotype which results in 1600 times more
bacteria available on the leaf surface than if the genes
were not present [76 ]. Once again, since this pathogen is
also primarily spread by wind-blown rain, availability of
infection units in quantity on the leaf surface would have
epidemiological significance. Therefore the host–specific
disease phenotype may serve an epidemiological purpose
independently of multiplication within the host, to enable
more efficient movement to uninfected hosts.
The ability to elicit the division of differentiated plant
cells, the ability to water soak cotton and the ability to
elicit PCD are radically different phenotypes. Yet all
three phenotypes are determined by genes which are 97 %
identical in DNA sequence (pthA, aŠrb6 and aŠrBs3) [74 ].
The dozen or more 102 bp direct tandem repeats found in
the middle of these genes are known to facilitate intragenic
recombination, which generates mutant variants of these
211
AŠirulence genes
genes with altered specificity and with altered pathogenicity phenotypes [74 ]. Reshuffling of repetitive tandem
repeats has become well recognized as one of several
mechanisms used by organisms to achieve genetic adaptation [57 ]. If one of these aŠr\pth genes moves
horizontally into a new bacterial pathogenic strain to
which they were not previously adapted, the most likely
result would be to confer either cultivar-specific avirulence
or no phenotype at all. Like aŠrBs3 in X. campestris pv.
vesicatoria, the gene would be both dispensable and
gratuitous. However, intragenic or intergenic recombination might generate a variant aŠr\pth gene with
selective value in the pathogen, one that enhanced
pathogenicity. Such a change would require that the
Avr\Pth protein interact not with a resistance factor in
the host, but rather with a ‘‘ susceptibility ’’ factor. The
fact that nearly identical genes from the aŠrBs3\pthA gene
family can cause radically different diseases on plants in
different families indicates that this has already happened
several times and will likely happen again.
THE CONVERSE OF R GENES : COGNATE
PLANT SUSCEPTIBILITY GENES ?
There is strong evidence that aŠr genes encode protein
signals that are delivered into the plant cell in order to
cause disease. Some are ‘‘ intercepted ’’ by R genes and
result in an HR, while others are nonfunctional because of
various stochastic processes, but the original purpose
appears to be for pathogenicity. These conclusions are
based primarily on : (1) the hrp dependency of all aŠr
genes ; (2) the discovery of functional nuclear localization
signal (NLS) sequences on some of the genes [21, 73 ] ; (3)
the ability of many tested aŠr genes to elicit an HR when
expressed inside resistant plant cells and (4) the ability of
pthA to elicit all symptoms diagnostic of citrus canker
disease when expressed inside citrus cells [16 ]. Such
protein signals must interact with plant receptors. A
direct physical binding of the avirulence protein AvrPto
from P. syringae with the tomato resistance protein Pto has
been demonstrated [64 ]. The specificity of the aŠrBs3\pthA
gene family has been shown to reside in the leucine rich
repeats (LRRs) of the protein products [34, 72 ], and the
specificity of 13 alleles of the flax rust resistance gene L
was recently demonstrated to reside in the LRRs of the
protein products [18 ]. LRRs are thought to be involved in
direct protein-protein interactions. The fact that pthA and
aŠrb6 elicit pathogenic phenotypes specific for the entire
range of citrus and cotton hosts, respectively, provides
evidence that the receptors in citrus and cotton must be
conserved among respective citrus and cotton hosts, and
that they are specific cognate receptors.
The idea that aŠr\pth genes encode signal proteins that
bind and activate plant receptors may be generally true
for all biotrophic pathogens that utilize type III secretion
as a primary means of pathogenicity. If so, it might not be
the race-specific R genes and an active defense response
(so useful in breeding resistant cultivars) that are generally
responsible for limiting the host range of biotrophic plant
pathogens, as has often been suggested. In fact, several
labs have attempted to find evidence for this by mutational
analysis of aŠr genes, including pthA, but with negative
results, reviewed by [22, 42 ]). Rather, it might be a failure
of the pathogenicity signal to bind an appropriate cognate
plant response receptor (to condition disease) that would
limit host range. The host range of biotrophic plant
pathogenic bacteria would in part be determined by
possession of a type III protein injection system and
appropriate protein effectors capable of binding receptors
that trigger host cell death and\or other plant responses
useful to the pathogen, such as cell division.
CONCLUSIONS
Plant cells are killed by both necrotrophic and biotrophic
pathogens, most likely primarily for nutritional purposes,
but also for dissemination. The bacterial hrp genes, which
encode a contact-dependent (Type III) secretion system,
appear to deliver signal proteins directly to plant cells to
help condition pathogenicity, and involving host cell
death. Secreted harpins are utilized by the more necrotrophic bacterial plant pathogens to kill cells, while the
more biotrophic pathogens appear to utilize protein
signals encoded by pth genes to kill cells. Some pathogens
use both. The limited number of microbial pathogenicity
signals that have been identified implies a limited number
of plant receptors that are responsive to such signals.
These receptors appear to determine host range and
specific susceptibility to biotrophs.
By contrast with pth genes, a relatively large number of
microbial aŠr genes have been identified, and the majority
appear to be dispensable or gratuitous. Many aŠr genes,
and even the entire hrp system, appear to have transferred
horizontally among gram negative bacteria. The aŠr genes
that are dispensable require the same contact-dependent
secretion system as pth genes, and the hypothesis is
advanced that most aŠr genes : (a) are, or once were, host
species-specific pth genes ; (b) encode protein signal
molecules that must bind plant response receptors in or on
the plant cell, and (c) often possess a physical structure
that enhances recombination and genetic adaptability.
Some of these may be able to evolve new host specificities
and pathogenicity phenotypes.
I thank Al Ellingboe, Margaret Essenberg and Sheng
Yang He for valuable insights and perspectives.
212
D. W. Gabriel
REFERENCES
1. Alfano JR, Collmer A. 1996. Bacterial pathogens in
plants : life up against the wall. Plant Cell 8 : 1683–1698.
2. Alfano JR, Collmer A. 1997. The type III (Hrp) secretion
pathway of plant pathogenic bacteria : trafficking harpins,
Avr proteins and death. Journal of Bacteriology 179 :
5655–5662.
3. Al-Mousawai AH, Richardson PE, Essenberg M,
Johnson WM. 1982. Ultrastructural studies of a compatible interaction between Xanthomonas campestris pv.
malŠacearum and cotton. Phytopathology 72 : 1222–1230.
4. Atkinson MM, Baker CJ. 1987. Alteration of plasmalemma sucrose transport in Phaseolus Šulgaris by Pseudomonas
syringae pv. syringae and its association with K+\H+
Exchange. Phytopathology 77 : 1573–1578.
5. Atkinson MM, Baker CJ. 1987. Association of host plasma
membrane K+\H+ exchange with multiplication of
Pseudomonas syringae pv. syringae in Phaseolus Šulgaris.
Phytopathology 77 : 1273–1279.
6. Bauer DW, Wei ZM, Beer SV, Collmer A. 1995. Erwinia
chrysanthemi HarpinEch : an elicitor of the hypersensitive
response that contributes to soft-rot pathogenesis. Molecular
Plant-Microbe Interactions 8, 484–491.
7. Bonas U, Stall RE, Staskawicz B. 1989. Genetic and
structural characterization of the avirulence gene, avrBs3
from Xanthomonas campestris pv. vesicatoria. Molecular
and General Genetics 218 : 127–136.
8. Cameron RK, Dixon RA, Lamb CJ. 1994. Biologically
induced systemic acquired resistance in Arabidopsis thaliana.
The Plant Journal 5 : 715–725.
9. Charkowski AO, Alfano JR, Preston G, Yuan J, He SY,
Collmer A. 1998. The Pseudomonas syringae pv. tomato
HrpW protein has domains similar to harpins and pectate
lyases and can elicit the plant hypersensitive response and
bind to pectate. Journal of Bacteriology 180 : 5211–5217.
10. Covert SF. 1998. Supernumerary chromosomes in filamentous fungi. Current Genetics 33 : 311–319.
11. Dangl JL, Dietrich RA, Richberg MH. 1996. Death don’t
have no mercy : cell death programs in plant-microbe
interactions. Plant Cell 8 : 1793–1807.
12. Day PR. 1974. Genetics of host-parasite interaction, W. H.
Freeman and Co., San Francisco. 238 pp.
13. De Feyter R, McFadden H, Dennis L. 1998. Five
avirulence genes from Xanthomonas campestris pv. malvacearum cause genotype-specific cell death when expressed transiently in cotton. Molecular Plant-Microbe
Interactions 11 : 698–701.
14. De Feyter R, Yang Y, Gabriel DW. 1993. Gene-for-genes
interactions between cotton R genes and Xanthomonas
campestris pv. malvacearum avr genes. Molecular PlantMicrobe Interactions 6 : 225–237.
15. Dow JM, Clarke BR, Milligan DE, Tang JL, Daniels
MJ. 1990. Extracellular proteases from Xanthomonas
campestris pv. campestris, the black rot pathogen. Applied
and EnŠironmental Microbiol. 56 : 2994–2998.
16. Duan YP, Castaneda A, Zhao G, Erdos GW, Gabriel
DW. 1999. Expression of a single, host-specific, bacterial
pathogenicity gene in citrus cells elicits division, enlargement and cell death. Molecular Plant-Microbe Interactions 12 : 556–560.
17. Ellingboe AH. 1996. Genetics of host-parasite interactions.
Pages 761–778 in : Encyclopedia of plant pathology, new
series, Vol 4 : Physiological Plant Pathology. R. Heitefuss
and P. H. Williams, eds. Springer-Verlag, Heidelberg.
18. Ellis JG, Lawrence GJ, Luck JE, Dodds PN. 1999.
Identification of regions in alleles of the flax rust resistance
gene L that determine difference in gene-for-gene
specificity. Plant Cell 11 : 495–506.
19. Flor HH. 1946. Genetics of pathogenicity in Melampsora lini.
Journal of Agricultural Research 73 : 335–357.
20. Flor HH. 1947. Inheritance of reaction to rust in flax.
Journal of Agricultural Research 74 : 241–262.
21. Gabriel DW. 1997. Targeting of protein signals from
Xanthomonas to the plant nucleus. Trends in Plant Science 2 :
204–206.
22. Gabriel DW. 1999. The Xanthomonas aŠr\pth gene family.
Pages 39–55 in : Plant-Microbe Interactions, Vol 4. G.
Stacey and N. T. Keen, eds. APS Press, St. Paul.
23. Gabriel DW, Kingsley MT, Yang Y, Chen J, Roberts P.
1993. Host-specific virulence genes of Xanthomonas. Pages
141–158 in : Molecular mechanisms of bacterial virulence.
C. I. Kado and J. Crosa, eds. Kluwer Academic Publishers, Dordrecht.
24. Galan JE. 1996. Molecular genetic bases of Salmonella entry
into host cells. Molecular Microbiology 20 : 263–271.
25. Gilchrist DG. 1998. Programmed cell death in plant
disease : the purpose and promise of cellular suicide.
Annual ReŠiew of Phytopathology 36 : 393–414.
26. Gilchrist DG. 1998. Blocking of programmed cell death in
tomato. Abstracts of IBC’s 3rd International Conference
on Transgenic Plants, Lake Buena Vista, FL, November
2–3.
27. Glazener JA, Orlandi EW, Baker CJ. 1996. The active
oxygen response of cell suspensions to incompatible
bacteria is not sufficient to cause hypersensitive cell death.
Plant Physiology 110 : 759–763.
28. Gopalan S, Bauer DW, Alfano JR, Loniello AO, He SY,
Collmer A. 1996. Expression of the Pseudomonas syringae
avirulence protein AvrB in plant cells alleviates its
dependence on the hypersensitive response and pathogenicity (Hrp) secretion system in eliciting genotypespecific hypersensitive cell death. Plant Cell 8 : 1095–1105.
29. Han Y. 1999. Identification of a pea pathogenicity gene
cluster on a dispensable chromosome in Nectria haematococca
MPVI. Ph.D. Dissertation, University of Florida 127 pp.
30. He C, Rusu AG, Poplawski AM, Irwin JAG, Manners
JM. 1998. Transfer of a supernumerary chromosome
between vegetatively incompatible biotypes of the fungus
Colletotrichum gloeosporioides. Genetics 150 : 1459–1466.
31. He SY. 1998. Type III protein secretion systems in plant
and animal pathogenic bacteria. Annual ReŠiew of Phytopathology 36 : 363–392.
32. Heath MC. 1997. Evolution of plant resistance and
susceptibility to fungal parasites. Pages 257–276 in : The
Mycota, Vol. V, Part B. Carroll and Tudzynski, eds.
Springer-Verlag, Berlin.
33. Heath MC. 1998. Involvement of reactive oxygen species in
the response of resistant (hypersensitive) or susceptible
cowpeas to the cowpea rust fungus. New Phytologist 138 :
251–263.
34. Herbers K, Conrads-Strauch J, Bonas U. 1992. Racespecificity of plant resistance to bacterial spot disease
determined by repetitive motifs in a bacterial avirulence
protein. Nature 356 : 172–174.
35. Keen N. 1986. Pathogenic strategies of fungi. Pages 171–188
in : Recognition in Microbe-Plant Symbiotic and Pathogenic Interactions. B. Lugtenberg, eds.
AŠirulence genes
36. Kemp JD. 1982. Plant pathogens that engineer their hosts.
Pages 443–457 in : Phytopathogenic Prokaryotes. M. S.
Mount and G. H. Lacy, eds. Academic Press, New York.
37. Kistler HC, Han Y, Benny U, Liu X, Temporini E,
VanEtten HD. 1999. A pathogenicity island in Nectria
haematococca. Fungal Genetics Newsletter 46 (Suppl.) : 33–33.
(Abstract).
38. Knogg W. 1996. Fungal infection of plants. Plant Cell 8 :
1711–1722.
39. Kobayashi DY, Tamaki SJ, Trollinger DJ, Gold S,
Keen NT. 1990. A gene from Pseudomonas syringae pv.
glycinea with homology to avirulence gene D from P.s. pv.
tomato but devoid of the avirulence phenotype. Molecular
Plant-Microbe Interactions 3 : 1103–111.
40. Kubori T, Matsushima Y, Nakamura D, Uralil J,
Lara-Tejero M, Sukhan A, Galan JE, Aizawa S. 1998.
Supramolecular structure of the Salmonella typhimurium
Type III protein secretion system. Science 280 : 602–605.
41. Lake JA, Jain R, Rivera MC. 1999. Mix and match in the
tree of life. Science 283 : 2027–2028.
42. Leach JE, White FF. 1996. Bacterial avirulence genes.
Annual ReŠiew of Phytopathology 34 : 153–179.
43. Llosa M, Zambryski P. 1998. On the origin and function
of pili. Trends in Microbiology 6 : 98–99.
44. Lorang JM, Shen H, Kobayashi D, Cooksey D, Keen
NT. 1994. aŠrA and aŠrE in Pseudomonas syringae pv. tomato
PT23 play a role in virulence on tomato plants. Molecular
Plant-Microbe Interactions 7 : 508–515.
45. Mass JL, Finney MM, Civerolo EL, Sasser M. 1985.
Association of an unusual strain of Xanthomonas campestris
with apple. Phytopathology 75 : 438–445.
46. Marenda M, Brito B, Callard D, Genin S, Barberis P,
Boucher C, Arlat M. 1998. PrhA controls a novel
regulatory pathway required for the specific induction of
Ralstonia solanacearum hrp genes in the presence of plant
cells. Molecular Microbiology 27 : 437–453.
47. McNellis TW, Mudgett MB, Li K, Aoyama T, Horvath
D, Chua NH, Staskawicz BJ. 1998. Glucocorticoidinducible expression of a bacterial avirulence gene in
transgenic Arabidopsis induces hypersensitive cell death.
Plant Journal 14 : 247–257.
48. Mecsas J, Strauss EJ. 1996. Molecular mechanisms of
bacterial virulence : type III secretion and pathogenicity
islands. Emerging Infectious Diseases 2 : 271–288.
49. Morgham AT, Richardson PE, Essenberg M, Cover
EC. 1988. Effects of continuous dark upon ultrastructure,
bacterial populations and accumulation of phytoalexins
during interactions between Xanthomonas campestris pv.
malŠacearum and bacterial blight susceptible and resistant
cotton. Physiological and Molecular Plant Pathology 32 :
141–162.
50. Person C, Mayo GME. 1974. Genetic limitations on
models of specific interactions between a host and its
parasite. Can. J. Bot. 52 : 1339–1347.
51. Piedras P., Hammond-Kosack KE, Harrison K, Jones
JDG. 1998. Rapid, Cf-9- and Avr9-dependent production
of active oxygen species on tobacco suspension cultures.
Molecular Plant-Microbe Interactions 11 : 1155–1166.
52. Pierce ML, Cover EC, Richardson PE, Scholes VE,
Essenberg M. 1996. Adequacy of cellular phytoalexin
concentrations in hypersensitively responding cotton
leaves. Physiological and Molecular Plant Pathology 48 :
305–324.
213
53. Preston GM, Haubold B, Rainey PB. 1998. Bacterial
genomics and adaptation to life on plants : implications for
the evolution of pathogenicity and symbiosis. Current
Opinion in Microbiology 1 : 589–597.
54. Rodriguez MC, Temporini ED, Wasmann CC, VanEtten HD. 1999. The homoserine utilization gene from a
CD chromosome of Nectria haematococca : a habitatdefining gene ? Fungal Genetics Newsletter 46 (Suppl) :
114–114. (Abstract).
55. Scofield SR, Tobias CM, Rathjen JP, Chang JH, Lavelle
DT, Michelmore RW, Staskawicz BJ. 1996. Molecular
basis of gene-for-gene specificity in bacterial speck disease
of tomato. Science 274 : 2063–2065.
56. Shaner G, Lacy GH, Stromberg EL, Barker KR, Pirone
TP. 1992. Nomenclature and concepts of pathogenicity
and virulence. Annual ReŠiew of Phytopathology 30 : 47–66.
57. Shapiro JA. 1999. Views about evolution are evolving.
ASM News 65 : 201–207.
58. Silhavy TJ. 1997. Death by lethal injection. Science 278 :
1085–1086.
59. Stall RE, Loschke DC, Jones JB. 1986. Linkage of copper
resistance and avirulence loci on a self-transmissible
plasmid in Xanthomonas campestris pv. vesicatoria. Phytopathology 76 : 240–243.
60. Staskawicz BJ, Dahlbeck D, Keen NT. 1984. Cloned
avirulence gene of Pseudomonas syringae pv. glycinea determines race-specific incompatibility on Glycine max (L.)
Merr. Proceedings of the National Academy of Sciences, USA 81 :
6024–6028.
61. Stevens C, Bennett MA, Athanassopoulos E, Tsiamis
G, Taylor JD, Mansfield JW. 1998. Sequence variations
in alleles of the avirulence gene aŠrPphE.R2 from Pseudomonas syringae pv. phaseolicola lead to loss of recognition of
the AvrPphE protein within bean cells and a gain in
cultivar-specific virulence. Molecular Microbiology 29 : 165–
177.
62. Swarup S, De Feyter R, Brlansky RH, Gabriel DW.
1991. A pathogenicity locus from Xanthomonas citri enables
strains from several pathovars of X. campestris to elicit
cankerlike lesions on citrus. Phytopathology 81 : 802–809.
63. Swarup S, Yang Y, Kingsley MT, Gabriel DW. 1992. A
Xanthomonas citri pathogenicity gene, pthA, pleiotropically
encodes gratuitous avirulence on nonhosts. Molecular PlantMicrobe Interactions 5 : 204–213.
64. Tang X, Frederick RD, Zhou J, Halterman DA, Jia Y,
Martin GB. 1996. Initiation of plant disease resistance by
physical interaction of AvrPto and Pto kinase. Science 274 :
2060–2063.
65. Ullrich WR, Kunz G, Stiller A, Novacky AJ. 1993. Role
of ammonium accumulation in bacteria-induced hypersensitive and compatible reactions of tobacco and cotton
plants. Physiologia Plantarum 89 : 644–652.
66. Van den Ackerveken G, Marois E, Bonas U. 1996.
Recognition of the bacterial avirulence protein AvrBs3
occurs inside the host plant cell. Cell 87 : 1307–1316.
67. VanEtten HD, Funnell-Baerg D, Wasmann C, McCluskey K. 1994. Location of pathogenicity genes on
dispensable chromosomes in Nectria haematococca MPVI.
Antonie Šan Leeuwenhoek 65 : 263–267.
68. van Gijsegem F, Genin S, Boucher C. 1993. Conservation
of secretion pathways for pathogenicity determinants of
plant and animal bacteria. Trends Microbiol. 1 : 175–180.
69. Vivian A, Gibbon MJ. 1997. Avirulence genes in plantpathogenic bacteria : signals or weapons ? Microbiology
143 : 693–704.
214
D. W. Gabriel
70. Vivian A, Gibbon MJ, Murillo J. 1997. The molecular
genetics of specificity determinants in plant pathogenic
bacteria. Pages 293–328 in : Gene-for-gene relationships in
plant-parasite interactions. I. R. Crute and E. B. Holub,
eds. CAB International, Oxford.
71. Wei ZM, Laby RJ, Zumoff CH, Bauer DW, He SY,
Collmer A, Beer SV. 1992. Harpin, elicitor of the
hypersensitive response produced by the plant pathogen
Erwinia amyloŠora. Science 257 : 85–88.
72. Yang Y, De Feyter R, Gabriel DW. 1994. Host-specific
symptoms and increased release of Xanthomonas citri and
X. campestris pv. malvacearum from leaves are determined by the 102 bp tandem repeats of pthA and
avrb6, respectively. Molecular Plant-Microbe Interactions 7 :
345–355.
73. Yang Y, Gabriel DW. 1995. Xanthomonas avirulence\
pathogenicity gene family encodes functional plant nuclear
targeting signals. Molecular Plant-Microbe Interactions 8 :
627–631.
74. Yang Y, Gabriel DW. 1995. Intragenic recombination of a
single plant pathogen gene provides a mechanism for the
evolution of new host specificities. Journal of Bacteriology
177 : 4963–4968.
75. Yang Y, Shah J, Klessig DF. 1997. Signal perception and
transduction in plant defense responses. Genes & DeŠelopment 11 : 1621–1639.
76. Yang Y, Yuan Q, Gabriel DW. 1996. Watersoaking
functions of XcmH1005 are redundantly encoded by
members of the Xanthomonas avr\pth gene family. Molec.
Plant-Microbe Interact. 9 : 105–113.