|
Plant Physiol, December 2001, Vol. 127, pp. 1367-1374
SCIENTIFIC CORRESPONDENCE
Sentinels of Disease. Plant Resistance Genes
Robert
Fluhr
Department of Plant Sciences, Weizmann Institute of Science,
Rehovot 76100, Israel
 |
INTRODUCTION |
Act 2, SCENE I. King Henry VI, William Shakespeare
(Enter a Sergeant with two Sentinels)
Sergeant speaks:
"Sirs! take your places and be vigilant:
If any noise or soldier you perceive
Near to the walls, by some apparent sign
Let us have knowledge at the court of guard."
| |
CONCEPTS IN IMMUNITY |
Successful defense against an enemy
requires perception of his whereabouts. In the last few years, much
progress has been made in delineating the plant molecular sentinels
that participate in pathogen identification. This ability is encoded by
genetically "hard-wired" information, and is called "innate
immunity". It draws its origins from a phylogenetically ancient form
of immunity that is common to all Metazoa and Viridiplantae. However,
the rapid evolution of plant innate immunity genes has led to massive gene diversification. The appreciation of this diversification offers
challenging prospects in understanding the forces that have shaped
multicellular innate immunity and underlines the necessity of
furthering our understanding by examining multiple plant systems.
A basic concept of innate immunity is the recognition of "non-self"
components accomplished by constitutive pattern recognition receptors.
Like sentinels, they alert the organism to activate defense genes. The
agents that plant sentinels detect are called avirulence factors. They
are produced by the pathogen and probably play a role in the process of
pathogen colonization of the host. In vertebrates, non-self recognition
is mediated by both innate and acquired immunity systems. Animal innate
immunity consists of a limited number of pattern recognition genes that
assist and prime the more complex acquired immunity response. In
contrast to innate immunity, acquired immune system consists of a huge array of somatically produced recombinant proteins. Upon pathogen encounter and by the process of clonal expansion, a particular subset
of recognition molecules is called into service. As acquired immunity
attains its complexity during the organism's somatic life, it offers
the benefit of very complete non-self coverage (approximately
107 different gene products/Mbp coding sequence)
without a huge penalty in storage of genetic information (Fig.
1). Innate systems are inherently more
costly in their "information content" (approximately 2 × 102 different gene products/Mbp coding sequence).
However, the developmental strategy of plants, with their rigid
cellular structures and absence of a circulatory system, would seem to
preclude potential evolutionary development of the sophisticated
acquired pattern recognition systems. Instead, plant sentinels are
required to be completely cell autonomous. There is no reason to
surmise that plants face a smaller number of potential pathogens than
their animal counterparts. How then does an organism that relies solely
on innate immunity cope? Plants have met the challenge by choosing as
their sentinels a versatile group of pattern recognition molecules
called resistance genes (R-genes), by greatly amplifying their numbers
and by positioning them in the genome in a manner that facilitates
their rapid evolution.
View larger version (39K):
[in this window]
[in a new window]
|
Figure 1.
Strategies of self and non-self recognition in
innate and adaptive immunity. The shaded boxes represent infected
cells. In animals, specialized cells are sources for acquired immunity
and they spread throughout the organism (arrows). The genomic
information content (bp) devoted to genes of acquired immunity is
represented by three recombinatorial systems as exemplified in humans
by the following: T-cell receptors (TCR), major histocompatibility
complex (MHC), and the different antibody classes (V). The major
potential pattern recognition genes in plants are shown in Figure 2.
Gene number estimates are for Arabidopsis and include NBS/LRR-type
(150), LRR kinase-type (174), Cf-type (30), and Pto-type (70).
Estimates of genomic sequence dedicated to each gene type are 5 × 103 for NBS/LRR and LRR-kinase and 2 × 103 for the rest.
|
|
| |
SENTINEL BUILDING BLOCKS |
Sentinel receptor units contain a pattern recognition domain
combined with accessory domains that participate in signal relay (Fig.
2). Leu-rich repeat (LRR) elements have
generally been implicated in the pattern recognition role, but coiled
coil (CC) and kinase domains may also be involved. Signal relay
elements include membrane anchor or transmembrane domains, Toll and
interleukin-1 receptor-like domains (TIR), and nucleotide binding
domains (NBD). Additionally, there are animal-specific domains
including caspase-activating recruitment domains and
plant-specific domains that involve CC and kinase domains. Innate
sentinel molecules are built up in different organisms by the
combinatorial use of these components, as shown in Figure 2. Thus,
evolution appears to have freely juggled with the building blocks of
innate immunity, and as such they can be presented as independent motif
entities.
View larger version (34K):
[in this window]
[in a new window]
|
Figure 2.
Pattern receptors are built from domains common to
plants and animals. Upper, Domains have been conceptually divided into
five functional categories. A, LRR, CC, and a kinase are domains that
can function as pattern receptors. B, TIR, CC and caspase-activating
recruitment domains (CARD) are likely involved in signal transduction
by homo- or heterodimerization with acceptor molecules. C, NBD common
to Nod and plant TIR/CC-NBD may serve a nucleotide-dependent switch
function. D, The transmembrane motif functions to anchor attached
domains or participate in transmembrane signal transfer and may also
contain endocytosis signals for signal attenuation as has been found in
the tomato Ve Verticillium R-gene (Kawchuk et al., 2001 ). E,
The kinase domain probably participates in signal transduction. Lower,
Known examples of domain combinations are illustrated.
|
|
| |
LRR DOMAIN |
Structure Function Relationships
What are the biophysical properties of the LRR domain that favored
its choice as the pattern receptor for sentinels? LRR structures mediate protein-protein interaction and are the major determinants of
recognition specificity. The protein structure solved by the crystal
structure of the porcine ribonuclease inhibitor (RI) has served as a
rough structural platform for conceptualizing what the distantly
related plant LRR may look like (Kobe and Deisenhofer, 1994). The 15 LRRs of RI are composed of an inner "solvent-exposed" surface rim
comprising  -sheets connected by an outer rim of  -helical segments. In RI, the -sheets are stabilized by a ladder of hydrogen bonds between conserved cysteines and Asn side chains. The -helical segments force a curvature on the molecule so that it comes nearly full
circle leaving a 60° opening. It is through this opening that the
ribonuclease interacts with the solvent surface. Plant LRR can contain
many fewer repeats; for example, seven as found in the sugar beet
(Beta vulgaris) nematode
Hs1pro1 R-gene (Cai et al., 1997), or many more
than 15 repeats, which is usually the case. Overabundant LRR repeats in
a domain would seemingly force complete circle closure. However, the
plant -helical regions tend to be much less conserved (they are
shorter or nonexistent). The LRR domain structure must therefore
deviate from RI in a manner that would generate a less constrained
flexible domain stabilized by the hydrogen bonds between the
-sheets.
Specificity between LRR domains and their potential ligands has been
inferred in animal TIR-LRR by the finding that Toll-Like Receptor 4 (TLR4) immunity genes from different species can each impart the
particular species-specific sensitivity to pharmacologically different
lipopolysaccharide structures (Poltorak et al., 2000). The tomato
(Lycopersicon esculentum) Cf-like R-genes
confer resistance to infection by the biotrophic leaf mold pathogen
Cladosporium fulvum and contains an extracellular LRR
domain. By precise domain swapping, a number of LRRs could be shown to
be essential for both Cf-4 and Cf-9 function (Van
Der Hoorn et al., 2001; Wulff et al., 2001). Direct evidence for
interaction of LRR domain with avirulence factors is based on the
finding that a single amino acid difference in the LRR domain
distinguished susceptible and resistant alleles of the fungal
Magnapporthe grisea rice R-gene (Bryan et al., 2000). In
this case, by using the yeast (Saccharomyces cerevisiae) two-hybrid system, the recombinant LRR domain
of the resistant allele could be shown to directly interact with its avirulence factor while the susceptible allele displayed a much weaker
interaction (Jia et al., 2000). Interestingly, this system gives us a
glimpse of the complex biology involved in plant receptor/avirulence factor interactions. The avirulence factor gene from M. grisea (AVR-Pita) encodes for a pre-propeptide that has the
features of a metalloprotease. The processed form is then transferred
by an unknown mechanism from the fungus into the plant cell.
Evolution
If solvent-exposed regions of the LRR domain play a role in
interaction with the pathogen avirulence factor, they may display high
mutability. Detecting adaptive evolution is carried out by estimating
base changes needed to generate amino acid changes, i.e. comparing the
number of substitutions per synonymous site (Ks) with the number of
substitutions per nonsynonymous site (Ka; Li, 1993). Ks is expected to
exceed Ka when mutations that generate amino acid change in a
particular gene are deleterious to an organism's fitness. However, if
mutations in an area of a particular gene are advantageous for the
organism, natural selection favors sequence diversification, and a
Ka/Ks ratio larger than 1 will emerge. A survey of plant R-genes shows
that, generally, the N-terminal CC structures as well as the NBD show
low average Ka/Ks ratio. In contrast, positive selection (high Ka/Ks
ratio) was detected in the predicted -strand region of the LRR
domain in many R-gene analogs (for review, see Bergelson et al., 2001).
What dictates the elevated Ka/Ks ratio? Is it entirely due to high
selection pressure applied to the basal mutation rate? Or do perhaps
the solvent-exposed LRR regions also have a propensity for
hypermutation, a phenomenon that has been detected for immunoglobulin genes?
| |
TIR DOMAINS |
Structure Function Relationships
TIR domains are a common link between animal and plant innate
immunity (Kimbrell and Beutler, 2001). In plant NBD-LRR R-genes, the
TIR motif appears at the N terminus, whereas in animals the TIR domain
is at the carboxyl end of a single-pass transmembrane receptor (Fig.
1). Structural analysis of the diverged TIR domains of human TLR2 and
TLR4 reveals a relatively conserved surface for binding to MyD88 (Xu et
al., 2000). MyD88, a signal adapter molecule, is essential for signal
transduction of the immune response. The idea of surface conservation
of TIR is important because it may explain how divergent TIR-containing
genes can signal through the apparently singular MyD88 adapter.
Database searches have yet to reveal a plant MyD88 homolog; however,
the finding that mutations in enhanced disease susceptibility
locus 1 (EDS1) compromise the articulation of distinct
R-genes of the TIR-NBD-LRR class argues for signal funneling
through a common intermediate (Aarts et al., 1998).
The TIR domain is essential for tobacco (Nicotiana
tabacum) N-gene-mediated tobacco mosaic virus resistance, and
amino acids that affect Drosophila melanogaster and
human TIR-dependent signaling cause either partial or full loss of
N-gene function (Dinesh-Kumar et al., 2000); for example, N-gene D46
resides in a position consistent with the conserved human TLR2 surface.
Mutation D46H completely eliminates function, whereas the D46Y
substitution that is the normal state for the human interleukin-1
receptor had no effect. Interestingly, partial loss-of-function
mutations can act as dominant negative mutations by promoting systemic
hypersensitive response in the wild-type N-gene background, a result
that also argues for the involvement of sentinels in higher order
complexes. Unexpectedly, the plant TIR domain may contribute to
determining R-gene specificity. In flax rust resistance, swapping the
TIR domains of L6 and L7
R-genes switches their specificity (Luck et al., 2000). This, together
with evidence for diversifying selection in the TIR region of the flax
rust R-genes, may also indicate that pattern recognition operates as a complex.
Evolution
Phylogenetic trees of TIR domains reveal a clear division between
animal and plant taxa (Kimbrell and Beutler, 2001). However, the low
identity of amino acid sequence between plant and animal TIR domains
(less then 20%) obviates facile structural comparison. The animal taxa
are further divided into at least two independent groups, Toll-like and
Interleukin receptor-like. The plant TIR family as potentially
represented in the Arabidopsis genome can likewise be divided into a
few TIR phylogenetic groups. One group of over 100 genes includes TIR
domains as part of NBD or as part of NBD-LRR sequences. Another group
of more than 30 genes is composed of solitary TIR domains or TIR
domains juxtaposed to other domains of unknown function (N. Kaplan-Levy
and R. Fluhr, unpublished data).
The sequence similarity between plant and animal TIR domains suggests a
common unicellular ancestor. Indeed, sensitive "SMART" searches of
current databases, which use reiterative sequence alignment together
with a broad definition of conserved polypeptide structural elements,
have revealed distant TIR homologies in prokaryotes as well
(http://smart.embl-heidelberg.de/; Schultz et al., 2000). Surprisingly,
despite their obviously ancient origin, TIR domains have not been
detected in any of the cereal genomes, their function apparently
replaced by CC domains (Meyers et al., 1999; Pan et al.,
2000b).
| |
CC DOMAIN |
CCs are an oligomerization motif of helical structures that are
made up of bundles containing two to five helices. A typical CC
structure shows a heptad repeat where the seven positions are labeled a
through g. Residues a and d tend to be hydrophobic, and the
residues at the e and g positions are charged or polar. A large subset
of eudicot and cereal NBD-LRR genes can be shown to contain general CC
domains in their N-terminal region with over 95% probability (Pan et
al., 2000b). In this context, they may serve the function of adapter
TIR-like motifs (Fig. 1). However, alternative functionality is
suggested in the case of the RPW8 Arabidopsis R-gene
(DA2-type, Fig. 2), responsible for
broad-spectrum resistance to mildew. It contains CC domains that appear
at the C-terminal of a predicted transmembrane domain or signal peptide domain (Xiao et al., 2001). In this case, the CC domain may be analogous to the function played by the LRR domain and could act directly in pattern recognition of an avirulence factor common to many
mildew races. Alternatively, it may play an accessory role facilitating
other pattern recognition molecules to function. RPW8 is
encoded by a small gene family of only five linked members, but due to
the nature of CC domains, true assessment of their potential numbers in
plant genomes will require sophisticated database mining.
| |
KINASE DOMAIN |
Structure Function Relationships
The tomato resistance protein kinase encoded by Pto
shows remarkable similarity to interleukin-1 receptor-associated
kinase and Pelle kinases that act downstream in animal TIR-LRR-directed immune response (for review, see Cohn et al., 2001).
Autophosphorylation competence and the Pto activation domain were
obligatory for interaction of Pto and its cognate avirulence factor
(Sessa et al., 2000). Importantly, the function of Pto requires the
presence of Prf gene, an NBD-LRR-type gene (Salmeron et al.,
1996). Conserved activation domains delineate numerous but distant
putative Pto homologs in Arabidopsis (Fig. 1); some of them
appear together with other functional domains in receptor-like
kinases. In functional analogy to Pto, the
Arabidopsis avrPphB-susceptible kinase is necessary for
Resistance to Pseudomonas syringae 5 function but not for
Resistance to P. syringae subsp. maculicola 1 or
Resistance to P. syringae 2-type NBD-LRR resistance function
(Swiderski and Innes, 2001).
Evolution
Sequence signatures of the Pto activation domain together with the
presence of sequence insertions/deletions helped define nine Pto-like
phylogenic families. The clades were made up of sequence from different
plant families suggesting ancient origin for Pto in the
genus Solanum (Vleeshouwers et al., 2001). Indeed, orthologous Pto genes of Lycopersicon
pimpinellifolium and Lycopersicon hirsutum interact
with the same avirulence factor and confer resistance (Riely and
Martin, 2001). Cross-species maintenance of R-gene functionality may be
due to the existence of persistent pathogens that help maintain ancient
disease lineages. Interestingly, L. esculentum accessions
examined contained no Pto ortholog showing that Pto kinases,
like their cognate NBD-LRR, are either part of metabolically
dispensable pathways or are functionally redundant.
Ser/Thr kinase domains are also found as part of
A1DE-type receptors (Fig. 2). These
domains play a role in signal transduction as was shown exquisitely by
switching the extracellular LRR and transmembrane domains of the
Arabidopsis brassinosteroid receptor kinase BRI1 with the
Ser/Thr kinase domain of the rice XA21 R-gene. The resultant
transgene promoted brassinosteroid-dependent resistance signaling (He
et al., 2000). Shiu and Bleecker (2001) have shown that all the kinase
domains of the A1DE-type receptors belong to a
single monophyletic group. The closest eukaryotic homolog to the plant
receptor kinase family was found to be the D. melanogaster kinase Pelle, similar to that found for Pto.
| |
NBD |
Structure Function Relationships
NBDs are characterized by several sequence motifs found in animal
ATP- and GTP-binding proteins including the Ras superfamily and the
caspase pathway-related Ced-4 and Apaf-1 animal
genes (Li et al., 1997). In animals, the genes regulate the activity of
proteases that can initiate apoptotic cell death. As defense mechanisms
in plants include apoptotic-like hypersensitive responses, the
appearance of these homologies in this context is particularly intriguing. By analogy to Ras, NBD may serve as a switch function moderating the inter- or intramolecular activity of the polypeptide. Alternative structural modeling of a subset of the NBD region revealed
homology to the receiver domain of His-Asp phosphoproteins typical of
prokaryotic signaling pathways (Rigden et al., 2000). This would argue
for NBD participation in phosphorelay as opposed to actual nucleotide
binding. Direct biochemical studies with this motif are lacking.
Interestingly, the Arabidopsis RPP5 NBD domain was shown to interact in
the yeast two-hybrid system with RelA/SpoT gene homologs. In
Escherichia coli these genes are involved in p/pppGpp stress
effector molecule signaling (van der Biezen et al., 2000). Whether this
implies a similar function in planta is not known.
A systematic approach to elucidate regions critical for activity was
carried out in the N-gene NBD region. Mutations in the conserved
nucleotide binding site, e.g. glycines or Lys (G216 and K222), led to
loss of N-gene action but also interfered in a negative dominant
fashion when the mutant transgene was present in the normal N-gene
background (Dinesh-Kumar et al., 2000). These results are consistent
with that detected in other NBD-containing proteins and point to
similar mechanisms of action. For example, mutations in the conserved
kinase 2 position at D301 leads to complete loss of function. The
equivalent mutations of the conserved Asp in CED4 disrupt the
oligomerization process that is necessary for activating downstream
caspases (Yang et al., 1998).
Evolution of Species-Specific R-Gene Diversity
In contrast to the highly diverged TIR/CC and LRR domains that
compose NBD-LRR resistance homologs, all NBD domains contain considerable conserved stretches of sequence homology. They can be
clearly divided into two distinct groups due to their different conserved subdomains (Meyers et al., 1999; Pan et al., 2000b). One
group is always found linked to resident TIR sequences in its
N-terminal region. This entire supergene group is notably absent from
cereal databases and cannot be amplified from cereal genomes (Pan et
al., 2000b). The other group, which also contains particular subdomain
signatures in the NBD, is linked to resident CCs in its N-terminal
region. This group is present as a superfamily of genes in both eudicot
and monocot species. As EDS1 and NDR1 were shown
to modulate TIR-type or CC-type multiple R-gene action, respectively,
in Arabidopsis (Aarts et al., 1998), the lack of a whole group of
R-gene analogs implies bifurcation of signal transduction pathways
between these two great plant classes.
The ancient origin of TIR domains indicates that cereal genomes have
lost or utterly corrupted their copies. The question is how and when
did this occur? TIR domains have been detected in Metazoans and
Gymnosperms, suggesting that the Angiosperm progenitor plant contained
TIR, as well. The loss of TIR sequences in cereals (and perhaps all
monocot genomes) must have occurred during monocot and eudicot
divergence over the last 100 million years. While mechanisms such as
unequal crossing-over can account for elimination of small multigene
families, no mechanism has been suggested for complete elimination of
dispersed supergene families that are detected in the genome of
today's modern plants. It is thus likely that in the ancient
angiosperm progenitor only a few germline R-genes existed similar to
the lower number of TIR-LRR detected in the Metazoan genomes (Fig.
3). If the great expansion in NBD-LRR gene number occurred after eudicot/monocot division, considerable diversity should be detected. Indeed, clustering of NBD sequences by
the neighbor-joining comparison method showed that Asteraceae, Brassicaceae, Linaceae and Poaceae yielded distinct family-level sequence clades (Pan et al., 2000a). Utilizing maximum parsimony and
likelihood methods plus a larger sequence database, Steven Cannon
working with Nevin Young and Georgiana May has confirmed and extended
these findings to detect ancient origin of some non-TIR-NBD. In their
analysis, taxa originating from monocots, eudicots, and gymnosperms can
appear together in the same ancient sequence clades (S. Cannon,
University of Minnesota, Minneapolis, personal communication; www.tc.umn.edu/~cann0010/CannonEtAl_R_genes.html). Thus, some non-TIR-NBDs were found to predate the angiosperm-gymnosperm radiation. Importantly, in any one sequence clade, particular species could be
dramatically over- or under-represented. Taken together, this argues
for intense diversification of NBD-LRR and indicates that no one
species will serve as a good sequence model for NBD-LRR homologs.
View larger version (29K):
[in this window]
[in a new window]
|
Figure 3.
Evolution of NBD-LRR-type disease R-gene homologs.
The diagram shows that stages in plant R-gene evolution involved early
novel domain accretion (NBD and CC) followed by massive gene
duplication and gene diversification. The two major divisions are
called Group I and II. The TIR-NBD-LRR-type genes
(B1CA, Fig. 2) were probably eliminated from
cereal genomes prior to their amplification.
|
|
| |
CLUSTERING OF R-GENES |
NBD-LRR sequences can reside in large extended arrays spanning
millions of bp that consist of dozens of R-gene homologs. In the
Arabidopsis genome, about 77% of the detected NBD sequence appear as
clusters of more than one gene (http://niblrrs.ucdavis.edu). To get
an idea of the global genome architecture of R-genes, random NBD
analogs have been isolated and mapped by high-resolution genetic mapping in soybean (Glycine max; Kanazin et al.,
1996), cereal genomes (Leister et al., 1998), and Solanaceae (Grube et
al., 2000; Pan et al., 2000a; Gebhardt and Valkonen, 2001). Clustering of NBD sequences was found to be evident. Juxtaposition of sequences in
the genome would serve as a ready source of new variance due to unequal
recombination and gene conversion events. R-gene evolution could follow
a pattern of rapid positive selection of successful stochastic
combinations of R-gene/avirulence factor. Strong selection pressure may
favor gene duplication, followed by new recurrent rounds of
recombination. It is of interest that highly related paralogs show the
highest Ka/Ks ratios, suggesting rapid selection in the newly
duplicated alleles (Bergelson et al., 2001).
| |
DO R-GENES SHOW GENOME SYNTENY AND FUNCTIONAL
CONSERVATION? |
Given the rapid evolution of R-genes due to
environmental pressure, it is of interest to compare the status of
these genes relative to general genome architecture. Global
conservation of synteny despite speciation has been observed in
Solanaceae together with some specific R-gene loss (Pan et al., 2000a).
Similar loss of genes, but together with the occurrence of more
frequent nonsyntenic map positions, was reported for cereal R-genes and
their homologs and may indicate more rapid sequence divergence in the
monocots (Leister et al., 1998). The fact that cereals completely lack the more conserved TIR-LRR-type R-genes is consistent with their rapid
global rate of R-gene evolution.
It is important to keep in mind that even when syntenic clusters are
maintained during speciation, there is no reason to expect functional
conservation. Examination of species within the Solanaceae family can
readily detect conserved structural positioning but show completely
different disease specificity. For example, resistance to
Fusarium spp. in tomato corresponds to resistance to tobacco mosaic virus in pepper on chromosome 11 and the tomato nematode Hero R-gene corresponds to a potato Phytophthora
infestens resistance loci on chromosome 4 (Grube et al., 2000;
Gebhardt and Valkonen, 2001). However, at the level of closely related
species, a degree of functional conservation can be detected. This
question was approached by directly examining a series of lines
carrying the complete Lycopersicon pennellii genome
introgressed into a L. esculentum background. Many
independent Fusarium spp. resistance loci were detected with
varying quantitative differences between them, and two out of six loci
show common evolutionary origins by appearing in the same chromosomal
location in more than one species (Sela-Buurlage et al., 2001). Whether
these loci represent true orthologous sequence is of interest and
remains to be seen. What then is the utility in comparative genomic
mapping of R-genes? Other than furthering our understanding of how
resistance strategy has been shaped on the species level, such physical
maps, especially at the family level, will offer predictive jump-off
points for map-based cloning and isolation of map-specific markers of
importance in breeding strategies.
| |
MORE FROM LESS: HOW DOES A LIMITED REPERTOIRE OF R-GENES DO
THE JOB? |
Simplifying the vast complexity of potential non-self
ligands to common pattern recognition elements is shown by the
Arabidopsis FLS2 LRR receptor kinase (Gomez-Gomez and Boller, 2000). In
this case, the pattern of a conserved short amino acid domain in
bacteria flagellin is sufficient to elicit an FLS2-dependent
response (Felix et al., 1999). Unfortunately, most of the known
avirulence factors characterized bear little resemblance to each other.
However, the fact that a single R-gene can recognize more than one
avirulence factor, as in the case of Resistance to P. syringae
subsp. maculicola 1 (Grant et al., 1995) and probably
Mi, which confers resistance to both nematodes and aphids in
tomato (Rossi et al., 1998), argues for structural commonalties that
may have been overlooked. Another hypothesis that would diminish the
need for requiring extreme R-gene complexity is the "guard"
hypothesis (Dangl and Jones, 2001). The concept is that avirulence
factors are actually meant to target and render inoperative the
cellular machinery that induces general plant defense pathways. R-genes
would protect these targets and activate defense when they disassociate
as a result of interaction with the avirulence factor. Again, one could
expect multiple R-gene specificity when different pathogen avirulence
genes target the same component.
Our perception of R-gene/avirulence factor interaction is biased by the
high affinities achieved in antibody/ligand interactions, but this may
be misleading. An alternative hypothesis explaining how R-genes cope
requires modifying the expectation that the receptor/ligand association
value must be high. For example, in higher vertebrates, only 500 to
1,000 odor receptors are enough to "sense" a great spectrum of
smells. How then does our nose cope with the plethora of environmental
smells? Lancet and coworkers (1993) described a probability model that
predicts the repertoire size of smell receptors that would ensure
proper representation of receptors with a specified affinity. At the
expense of reduced receptor affinity, the repertoire of potential
interactions can be economic (Fig. 4). By
setting an affinity value of 105
M 1, a number that has been
experimentally measured in smell receptors/ligand interactions, a
repertoire of smell receptors of 300 to 1,000 in number would suffice
for finding at least one receptor with the specified affinity. R-genes
numbers in the plant genome are in a similar range (Figs. 1 and 4). If
in an analogous manner their average affinity requirements are modest,
it would explain the robustness of the plant resistance response.
Indeed, the limited repertoire of R-genes in any one genome can display
multiple independent loci of varying quantitative efficiency
(Sela-Buurlage et al., 2001). In animals, additional higher order
integration of information input from different receptors yields the
final sensation of smell. The "equivalent" integration in plants
could imply groups of R-genes cooperating with each other perhaps in a
multicomponent sentinel network. Results of cross-immunoprecipitation
experiments showing one avirulence factor interacting with more than
one R-gene implies that such networks may exist (Leister and Katagiri,
2000).
View larger version (26K):
[in this window]
[in a new window]
|
Figure 4.
Distribution of receptor-ligand association
constants and probability function N. The figure conceptualizes the
discrete probability that among N receptors a specific receptor of a
particular association value will be found. The figure was adopted
using the reciprocal of the  function, which describes the average
of all individual affinity distributions (Lancet et al., 1993). In the
cases shown, immunoglobulin genes were set at 108.5
M1 which would require a receptor repertoire
of 107 size. Olfactory receptors are estimated to
show association constants of 105 for their
ligands and would require a receptor repertoire of 500 to
103 size. A similar estimation of affinity
constants for plant R-genes could account for plant R-gene family
sizes.
|
|
Genetic approaches have drawn a complex and as yet an incomplete
picture of the essence of disease sentinel function. Biochemical knowledge of what precisely makes up the sentinel complex as well as
understanding the interplay of pathogen colonization and pathogenesis response presents an intriguing challenge that will see continued multidisciplinary and multisystem approaches.
| |
FOOTNOTES |
Received August 21, 2001; accepted September 10, 2001.
1
This work was supported by the MINERVA
Foundation, Munich, Germany.
*
E-mail lpfluhr{at}weizmann.ac.il; fax 972-8-934-4181.
www.plantphysiol.org/cgi/doi/10.1104/pp.010763.
| |
LITERATURE CITED |
-
Aarts N, Metz M, Holub E, Staskawicz BJ, Daniels MJ, Parker JE
(1998)
Proc Natl Acad Sci USA
95: 10306-10311[Abstract/Free Full Text]
-
Bergelson J, Kreitman M, Stahl EA, Tian D
(2001)
Science
292: 2281-2285[Abstract/Free Full Text]
-
Bryan GT, Wu KS, Farrall L, Jia Y, Hershey HP, McAdams SA, Faulk KN, Donaldson GK, Tarchini R, Valent B
(2000)
Plant Cell
12: 2033-2046[Abstract/Free Full Text]
-
Cai D, Kleine M, Kifle S, Harloff HJ, Sandal NN, Marcker KA, Klein-Lankhorst RM, Salentijn EM, Lange W, Stiekema WJ
(1997)
Science
275: 832-834[Abstract/Free Full Text]
-
Cohn J, Sessa G, Martin GB
(2001)
Curr Opin Immunol
13: 55-62[CrossRef][Web of Science][Medline]
-
Dangl JL, Jones JDG
(2001)
Nature
411: 826-833[CrossRef][Medline]
-
Dinesh-Kumar SP, Tham WH, Baker BJ
(2000)
Proc Natl Acad Sci USA
97: 14789-14794[Abstract/Free Full Text]
-
Felix G, Duran JD, Volko S, Boller T
(1999)
Plant J
18: 265-276[CrossRef][Web of Science][Medline]
-
Gebhardt C, Valkonen JPT
(2001)
Annu Rev Phytopathol
39: 79-102[CrossRef][Web of Science][Medline]
-
Gomez-Gomez L, Boller T
(2000)
Mol Cell
5: 1003-1011[CrossRef][Web of Science][Medline]
-
Grant MR, Godiard L, Straube E, Ashfield T, Lewald J, Sattler A, Innes RW, Dangl JL
(1995)
Science
269: 843-846[Abstract/Free Full Text]
-
Grube RC, Radwanski ER, Jahn M
(2000)
Genetics
155: 873-887[Abstract/Free Full Text]
-
He Z, Wang ZY, Li J, Zhu Q, Lamb C, Ronald P, Chory J
(2000)
Science
288: 2360-2363[Abstract/Free Full Text]
-
Jia Y, McAdams SA, Bryan GT, Hershey HP, Valent B
(2000)
EMBO J
19: 4004-4014[CrossRef][Web of Science][Medline]
-
Kanazin V, Marek LF, Shoemaker RC
(1996)
Proc Natl Acad Sci USA
93: 11746-11750[Abstract/Free Full Text]
-
Kawchuk LM, Hachey J, Lynch DR, Kulcsar F, van Rooijen G, Waterer DR, Robertson A, Kokko E, Byers R, Howard RJ
(2001)
Proc Natl Acad Sci USA
98: 6511-6515[Abstract/Free Full Text]
-
Kimbrell DA, Beutler B
(2001)
Nat Rev Genet
2: 256-267[CrossRef][Web of Science][Medline]
-
Kobe B, Deisenhofer J
(1994)
Trends Biochem Sci
19: 415-421[CrossRef][Web of Science][Medline]
-
Lancet D, Sadovsky E, Seidemann E
(1993)
Proc Natl Acad Sci USA
90: 3715-3719[Abstract/Free Full Text]
-
Leister D, Kurth J, Laurie DA, Yano M, Sasaki T, Devos K, Graner A, Schulze-Lefert P
(1998)
Proc Natl Acad Sci USA
95: 370-375[Abstract/Free Full Text]
-
Leister RT, Katagiri F
(2000)
Plant J
22: 345-354[CrossRef][Web of Science][Medline]
-
Li P, Nijhawan D, Budihardjo I, Srinivasula SM, Ahmad M, Alnemri ES, Wang X
(1997)
Cell
91: 479-489[CrossRef][Web of Science][Medline]
-
Li WH
(1993)
J Mol Evol
36: 96-99[CrossRef][Web of Science][Medline]
-
Luck JE, Lawrence GJ, Dodds PN, Shepherd KW, Ellis JG
(2000)
Plant Cell
12: 1367-1377[Abstract/Free Full Text]
-
Meyers BC, Dickerman AW, Michelmore RW, Sivaramakrishnan S, Sobral BW, Young ND
(1999)
Plant J
20: 317-332[Web of Science][Medline]
-
Pan Q, Liu YS, Budai-Hadrian O, Sela M, Carmel-Goren L, Zamir D, Fluhr R
(2000a)
Genetics
155: 309-322[Abstract/Free Full Text]
-
Pan Q, Wendel J, Fluhr R
(2000b)
J Mol Evol
50: 203-213[Web of Science][Medline]
-
Poltorak A, Ricciardi-Castagnoli P, Citterio S, Beutler B
(2000)
Proc Natl Acad Sci USA
97: 2163-2167[Abstract/Free Full Text]
-
Riely BK, Martin GB
(2001)
Proc Natl Acad Sci USA
98: 2059-2064[Abstract/Free Full Text]
-
Rigden DJ, Mello LV, Bertioli DJ
(2000)
Proteins
41: 133-143[CrossRef][Web of Science][Medline]
-
Rossi M, Goggin FL, Milligan SB, Kaloshian I, Ullman DE, Williamson VM
(1998)
Proc Natl Acad Sci USA
95: 9750-9754[Abstract/Free Full Text]
-
Salmeron JM, Oldroyd GE, Rommens CM, Scofield SR, Kim HS, Lavelle DT, Dahlbeck D, Staskawicz BJ
(1996)
Cell
86: 123-133[CrossRef][Web of Science][Medline]
-
Schultz J, Copley RR, Doerks T, Ponting CP, Bork P
(2000)
Nucleic Acids Res
28: 231-234[Abstract/Free Full Text]
-
Sela-Buurlage MB, Budai-Hadrian O, Pan Q, Carmel-Goren L, Vunsch R, Zamir D, Fluhr R
(2001)
Mol Genet Genomics
265: 1104-1111[CrossRef][Medline]
-
Sessa G, D'Ascenzo M, Martin GB
(2000)
EMBO J
19: 2257-2269[CrossRef][Web of Science][Medline]
-
Shiu S-H, Bleecker AB
(2001)
Proc Natl Acad Sci USA
98: 10763-10768[Abstract/Free Full Text]
-
Swiderski MR, Innes RW
(2001)
Plant J
26: 101-112[CrossRef][Web of Science][Medline]
-
van der Biezen EA, Sun J, Coleman MJ, Bibb MJ, Jones JD
(2000)
Proc Natl Acad Sci USA
97: 3747-3752[Abstract/Free Full Text]
-
Van Der Hoorn RA, Roth R, De Wit PJ
(2001)
Plant Cell
13: 273-285[Abstract/Free Full Text]
-
Vleeshouwers VG, Martens A, van Dooijeweert W, Colon LT, Govers F, Kamoun S
(2001)
Mol Plant Microbe Interact
14: 996-1005[Medline]
-
Wulff BB, Thomas CM, Smoker M, Grant M, Jones JD
(2001)
Plant Cell
13: 255-272[Abstract/Free Full Text]
-
Xiao S, Ellwood S, Calis O, Patrick E, Li T, Coleman M, Turner JG
(2001)
Science
291: 118-120[Abstract/Free Full Text]
-
Xu Y, Tao X, Shen B, Horng T, Medzhitov R, Manley JL, Tong L
(2000)
Nature
408: 111-115[CrossRef][Medline]
-
Yang X, Chang HY, Baltimore D
(1998)
Science
281: 1355-1357[Abstract/Free Full Text]
© 2001 American Society of Plant Physiologists
This article has been cited by other articles:
|
|
|
|
 
N. R. Forsthoefel, K. Cutler, M. D. Port, T. Yamamoto, and D. M. Vernon
PIRLs: A Novel Class of Plant Intracellular Leucine-rich Repeat Proteins
Plant Cell Physiol.,
June 1, 2005;
46(6):
913 - 922.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
P. Veronese, M. T. Ruiz, M. A. Coca, A. Hernandez-Lopez, H. Lee, J. I. Ibeas, B. Damsz, J. M. Pardo, P. M. Hasegawa, R. A. Bressan, et al.
In Defense against Pathogens. Both Plant Sentinels and Foot Soldiers Need to Know the Enemy
Plant Physiology,
April 1, 2003;
131(4):
1580 - 1590.
[Full Text]
[PDF]
|
|
|
|
|
|
|
W. I. L. Tameling, S. D. J. Elzinga, P. S. Darmin, J. H. Vossen, F. L. W. Takken, M. A. Haring, and B. J. C. Cornelissen
The Tomato R Gene Products I-2 and Mi-1 Are Functional ATP Binding Proteins with ATPase Activity
PLANT CELL,
November 1, 2002;
14(11):
2929 - 2939.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
TOP
INTRODUCTION
CONCEPTS IN IMMUNITY