Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Jul 17, 2007; 104(29): 12211–12216.
Published online Jul 3, 2007. doi:  10.1073/pnas.0705186104
PMCID: PMC1924578
Plant Biology

Structural modules for receptor dimerization in the S-locus receptor kinase extracellular domain


The highly polymorphic S-locus receptor kinase (SRK) is the stigma determinant of specificity in the self-incompatibility response of the Brassicaceae. SRK spans the plasma membrane of stigma epidermal cells, and it is activated in an allele-specific manner on binding of its extracellular region (eSRK) to its cognate pollen coat-localized S-locus cysteine-rich (SCR) ligand. SRK, like several other receptor kinases, forms dimers in the absence of ligand. To identify domains in SRK that mediate ligand-independent dimerization, we assayed eSRK for self-interaction in yeast. We show that SRK dimerization is mediated by two regions in eSRK, primarily by a C-terminal region inferred by homology modeling/fold recognition techniques to assume a PAN_APPLE-like structure, and secondarily by a region containing a signature sequence of the S-domain gene family, which might assume an EGF-like structure. We also show that eSRK exhibits a marked preference for homodimerization over heterodimerization with other eSRK variants and that this preference is mediated by a small, highly variable region within the PAN_APPLE domain. Thus, the extensive polymorphism exhibited by the eSRK not only determines differential affinity toward the SCR ligand, as has been assumed thus far, but also underlies a previously unrecognized allelic specificity in SRK dimerization. We propose that preference for SRK homodimerization explains the codominance exhibited by a majority of SRKs in the typically heterozygous stigmas of self-incompatible plants, whereas an increased propensity for heterodimerization combined with reduced affinity of heterodimers for cognate SCRs might underlie the dominant–recessive or mutual weakening relationships exhibited by some SRK allelic pairs.

Keywords: self-incompatibility, ligand-independent receptor dimerization, yeast two-hybrid

The S-locus receptor kinase (SRK) is a single-pass integral plasma membrane protein of the stigma epidermis that functions in the recognition and rejection of self-related pollen in self-incompatible members of the Brassicaceae (reviewed in refs. 1 and 2). The SRK receptor and its pollen-coat localized ligand, the S-locus cysteine-rich protein (SCR) (1), also designated SP-11 (2), are encoded by tightly linked and highly polymorphic genes within the S-locus haplotype. Specificity in the self-incompatibility response derives from allele-specific SRK–SCR interaction (3, 4). On self-pollination, the binding of SCR to the extracellular domain of its cognate SRK activates the receptor and triggers an intracellular signaling cascade that leads to arrest of “self” pollen tubes. In contrast, pollen tube growth in cross-pollinations proceeds unimpeded because “nonself” SCR cannot bind SRK. In view of this high degree of specificity in the affinity of SRK for its SCR ligand, the extensive polymorphisms exhibited by the SRK extracellular region (eSRK) have heretofore been considered in the context of specificity in the SRK–SCR interaction.

The eSRK does more than interact with SCR, however. SRK forms oligomers in vivo in the absence of ligand (5, 6). Similar ligand-independent self-association into dimeric or oligomeric inactive complexes via “preligand association domains” (PLADs) has been reported for several receptors in animal and plant systems (7, 8). These observations are inconsistent with the classical view of ligand-activated receptors, whereby receptor homodimerization or oligomerization is induced by ligand binding and only serves to bring the receptor intracellular domains in close proximity for transphosphorylation and recruitment of effector cytoplasmic proteins (9). Rather, receptor oligomerization is now viewed as insufficient for receptor activation and as having additional roles in receptor signaling (10). In the case of SRK, ligand-independent dimerization might provide a “primed” condition that allows rapid recruitment and activation of the receptor on ligand binding. Indeed, recent evidence indicates that SRK dimerization is critical for high-affinity SCR binding and is mediated at least partly by the eSRK (6).

It is not known which specific domains or residues in eSRK mediate receptor dimerization, how these residues are arranged in the folded protein, nor whether the carbohydrate moiety of this highly glycosylated extracellular region is important for self-interaction. To address these issues, we assessed the role in ligand-independent SRK dimerization of different eSRK subregions, each of which is predicted to represent a distinct structural module on the basis of homology modeling/fold recognition techniques. Assays for self-interaction activity in yeast, sequence alignments of eSRK variants, and analysis of chimeric eSRKs generated by swapping domains between SRK variants identified specific domains that mediate SRK self-interaction and a small region that confers a previously unsuspected allelic specificity in ligand-independent eSRK dimerization.

Results and Discussion

Preliminary Mapping of PLADs to the C Terminus of eSRK Using Yeast Surface Display.

We first investigated self-interaction of eSRK by using yeast surface display (11) on the assumption that this method would be suitable for analysis of extracellular proteins. We cloned eSRK6 and eSRK13, the extracellular regions of the Brassica oleracea SRK6 and SRK13 variants, into the yeast surface display vector pYD1, which allows galactose-inducible expression of recombinant proteins as V5- and 6×His-tagged C-terminal fusions with the a-agglutinin subunit Aga2p. To detect self-interaction, we assayed for yeast agglutination, a technique that was initially developed for the study of yeast mating proteins (12). We reasoned that interactions between proteins displayed on the surface of yeast cells would promote cell agglutination and precipitation of cell aggregates relative to single cells, and that the extent of agglutination (estimated by the OD600 absorption of resuspended cell pellets) would provide a semiquantitative measure of eSRK interactions (see Methods). We found that yeast cells displaying eSRK6 or eSRK13 exhibited enhanced agglutination relative to cells containing vector control [Fig. 1 and supporting information (SI) Fig. 7], demonstrating that the eSRK exhibits ligand-independent self-association in this yeast expression system.

Fig. 1.
Yeast surface display and cell agglutination assays. Absorbance (OD600) of pellets formed by yeast cultures expressing full-length eSRK6, the lectin-rich region (LRR) (amino acids 33–192), the CTR (amino acids 193–446), the SLG region ...

To obtain a preliminary assessment of the location of the self-interaction domain(s) within eSRK, we examined the domain arrangement predicted for eSRK by using the BLAST program. This arrangement is conserved among available SRK alleles and is shown in Fig. 2A for SRK6. Within eSRK6 (residues 33–446, following a signal peptide), residues 50–200 (FRAG1) are recognized as part of a d-mannose-binding lectin-like domain for which structural information is available in the Protein Data Bank (PDB) (13). A BLAST search with FRAG1 of eSRK6 produced two alternative low E value alignments that map the sequence of this fragment to the bulb-type mannose-specific lectin domain, with the first alignment spanning residues 48–161 and the second alignment spanning residues 79–192. The remainder of the eSRK may be divided roughly into two regions: one region, from residue 200 to 295 (FRAG2), contains a signature sequence of the S-locus glycoprotein family (designated “SLG consensus”), for which structural information is lacking; a second region, from residue 296 to 450 (FRAG3), contains 12 cysteines and a sequence (residues 345–430) that matches the PAN_APPLE domain consensus sequence.

Fig. 2.
Domain architecture and 3D models of predicted structural domains in eSRK6. (A) Arrangement of domains in the full-length SRK6 protein as assigned by National Center for Biotechnology Information (Upper) and eSRK6 generated by 3D modeling (Lower). SP, ...

On the basis of this predicted domain assignment, we divided the eSRK6 into an N-terminal segment containing the lectin-rich region (amino acids 33–192) and a segment consisting of the remaining C-terminal region (CTR) (amino acids 193–446). We found that only yeast cells displaying the CTR exhibited self-agglutination, whereas yeast cells that displayed the lectin-like region did not (Fig. 1), indicating that the PLADs responsible for eSRK self-interaction are located in the CTR. Further subdivision of the eSRK6 CTR showed that the region spanning residues 349–446, which contains the PAN_APPLE domain, caused yeast agglutination at a rate equivalent to that caused by the full-length eSRK6, whereas the region spanning residues 193–349, which was used to assay the region containing the SLG consensus sequence, effected a small increase in agglutination relative to vector control (Fig. 1). These results implicate the PAN_APPLE domain as a major mediator of ligand-independent eSRK dimerization.

Identification of Structural Modules Within eSRK by Homology Modeling/Fold Recognition Techniques.

Individual structural modules within proteins typically have an intrinsic capacity to assume a stable conformation and maintain their interactive properties even when taken out of the context of the larger protein that contains them. Consequently, functional studies are often guided by 3D structural information and make use of isolated structural modules produced in heterologous systems (14). In the case of eSRK, a 3D crystal structure is not available, however. Therefore, we used homology modeling and threading techniques to refine the BLAST-generated domain assignment and obtain a more precise delineation of individual structural domains for further functional studies.

A search of PDB with eSRK6 (Fig. 2A, FRAG1 to FRAG3) failed to identify a suitable template for homology modeling of the entire extracellular region. However, it was possible to build structural models for individual domains within eSRK6, as described in detail in SI Text and outlined briefly here. For FRAG2, a BLAST search of the PDB database did not retrieve any sequence with a significantly low E value. However, threading tools associated this sequence with a lectin-like domain, thus predicting the existence of two contiguous lectin-like domains in eSRK. A subsequent search of the PDB database identified a single structure that showed a reasonable match to the query and contained two contiguous lectin domains, namely the fetuin-binding protein SCAfet from bulbs of the bluebell Scilla campanulata (PDB ID code 1DLP). By using the lectin domain arrangement in 1DLP (molecule A), we generated 3D models for the FRAG1 and FRAG2 portions of SRK6. The most probable model (Fig. 2B) shows two lectin-like domains, designated LLD1 (amino acids 33–145) and LLD2 (amino acids 154–283), linked by a loop region of nine residues (Fig. 2A). Each of these domains consists of a 12-stranded β-prism II fold similar to that reported in other single-domain mannose-binding lectins (1517), although LLD1 and LLD2 share only 18% amino acid sequence identity compared with 55% identity between the two lectin-like domains of SCAfet (18). In support of this predicted structure, the region linking LLD1 and LLD2 corresponds to a “deletable region” previously identified (18) from sequence alignments of various SRK variants (Fig. 2A and SI Fig. 8). The variability in this “linker” region among SRK variants (19, 20) suggests that the exact spacing of LLD1 and LLD2 is not critical for SRK function.

For FRAG3, secondary structure prediction methods indicate that the first portion of the sequence (amino acids 293–346), which includes six cysteine residues (C299, C305, C311, C319, C321, and C342), has a low propensity to form an ordered secondary structure. Threading tools return with low score alignments to EGF modules (PDB ID code 1KLO), but there is no strong consensus from the different contributing servers on the structure of this fragment. It should be noted, however, that EGF modules are often difficult to identify (21) because highly diverged amino acid sequences can adopt an EGF-type protein fold (22) (SI Fig. 9B). In view of the overlap between the EGF-like domain with the SLG consensus sequence, which is characteristic of members of the S-gene family, it is possible that this region might assume an EGF-type conformation unique to this family of proteins (SI Text and Fig. 10). Although acknowledging that this subregion might be either unstructured or forced into an undetermined structure by disulfide bonds, we will henceforth refer to this subregion as the “EGF-like” domain.

The second subregion of FRAG3 (amino acids 347–433 in eSRK6) is identified by BLAST with a PAN_APPLE domain, and the Structure Prediction Meta Server (SPMS) assigns the highest scores to alignments of this region with structures containing PAN_APPLE domains. Our 3D model for the SRK6 PAN_APPLE domain (Fig. 2C) consists of a five-stranded antiparallel β-sheet, β1–β5, and an α-helix connected to the core by two disulfide bonds linking C380 to C405 and C384 to C390. The strands in the central β-sheet are arranged in the spatial order β1–β5–β3–β4–β2. Whereas the connections between the antiparallel strands β3–β4 are relatively short, the β1–β2 and β4–β5 connections form two long loops that lie on the face of the β-sheet opposite to the α-helix, which is formed by the β2–β3 connection. Analysis of PAN_APPLE domain structures shows that only two disulfide bonds are shared by the four available PDB structures 1I8N, 1HKY, 1GMO, and 2HGF, and this condition is satisfied by our 3D model (Fig. 2C and SI Fig. 9C). A third disulfide bond observed in some experimental structures is not present in others, such as 1GMO and 2HGF. According to our model, the two cysteine residues in this subregion of eSRK6, C350 and C388, cannot form a disulfide bridge. They would either be free or bound to cysteine residues outside the PAN_APPLE domain.

Identification of Ligand-Independent Dimerization Domains in eSRK.

Each of the eSRK structural modules inferred from homology modeling and fold recognition techniques was tested for the presence of PLADs. For this analysis, we used the yeast two-hybrid system because, unlike our application of yeast surface display, this system can provide information on the strength of interactions (23), because the expression of three reporters, ADE2, HIS3, and LacZ, may be monitored by assessing the extent of cell growth on selective media under different stringency conditions and by measuring β-galactosidase activity (see Methods). Furthermore, this system has been particularly useful for investigating the role of individual domains or combination of domains in intermolecular interactions (24, 25), and it is increasingly being used to investigate interactions between extracellular proteins (2628).

We first confirmed that the full-length eSRK6 and eSRK13, when fused to the Gal4 DNA binding domain (Gal4BD, bait) or activation domain (Gal4AD, prey), did not exhibit autoactivation and that each variant showed self-interaction when cotransformed as Gal4BD and Gal4AD plasmids into the PJ69-4A yeast strain (Fig. 3). This result established the usefulness of the yeast two-hybrid system for analysis of ligand-independent eSRK interaction and further demonstrated that this self-interaction does not require protein glycosylation because it occurred in the yeast nucleus.

Fig. 3.
Self-interaction activity of eSRK6 and eSRK13 (A) and various subregions of eSRK6 (B) in yeast two-hybrid assays. PJ69-4a yeast cells expressing the indicated AD and BD fusion protein were tested for histidine and adenine prototrophy on plates containing ...

We next assayed eSRK6 fragments corresponding to the structural domains defined by homology modeling. After ascertaining that AD and BD fusions of individual domains accumulated in yeast cells (SI Fig. 11), the LLD1, LLD2, and PAN_APPLE (PAN) fragments, which did not exhibit autoactivation, were tested individually (Fig. 3). However, the EGF-like fragment exhibited autoactivation of ADE2 and was assayed in conjunction with the LLD2 domain because a bait construct containing the LLD2 and EGF-like domains did not show autoactivation. Whereas neither LLD1 nor LLD2 exhibited self-interaction activity, LLD2+EGF exhibited relatively weak self-interaction and PAN showed strong self-interaction activity (Fig. 3). These results demonstrate that the major PLADs of eSRK are located within the PAN_APPLE domain, and they provide indirect evidence that the EGF-like domain might also contribute to eSRK self-interaction. The data support our yeast surface display results (Fig. 1) and are consistent with the known roles of PAN_APPLE and EGF modules in mediating protein–protein interactions (21, 29). In the case of SRK, these two domains might form two dimer interfaces that act synergistically to mediate ligand-independent dimerization.

Allelic Specificity in eSRK Self-Interaction.

In the course of our yeast two-hybrid assays, we noted that full-length eSRK6 and eSRK13 had the propensity for heterotypic (i.e., eSRK6–eSRK13) interaction (Fig. 3). However, this heterotypic interaction was evident only under relatively low-stringency selection conditions (Fig. 3), indicating that it is much weaker than the homotypic eSRK6–eSRK6 and eSRK13–eSRK13 interactions. In addition, coprecipitation experiments using FLAG-tagged eSRK6 and eSRK13 produced in Nicotiana benthamiana (3, 30) showed that eSRK6, but not eSRK13, could pull down SRK6 from S6 stigma extracts (Fig. 4A). Thus, SRK exhibits, in addition to its ligand-binding specificity, a second previously unreported facet of specificity that causes it to interact more strongly with an identical SRK molecule derived from the same SRK allele than with a variant SRK derived from another SRK allele. Interestingly, the PAN_APPLE domain is a major contributor to this preference for homotypic over heterotypic eSRK self-interaction. We found that PAN_APPLE domains derived from all four tested eSRK variants (B. oleracea SRK6 and SRK13, and Arabidopsis lyrata SRKa and SRKb) exhibited strong homotypic interactions (i.e., PAN6–PAN6, PAN13–PAN13, PANa–PANa, and PANb–PANb) but only weak heterotypic interactions with PAN_APPLE domains derived from different alleles (i.e., PAN6–PAN13 and PANa–PANb) (Fig. 4B).

Fig. 4.
Allelic specificity in eSRK self-interaction and the role of the PAN_APPLE domain. (A) Allele-specific interaction of recombinant eSRK-FLAG with stigma SRK. The binding of stigma SRK6 was detected by using antibodies that react with the kinase domain ...

The eSRK PAN_APPLE domain is contained within the so-called C-terminal variable region (VR) (19, 20), which is one of four regions that exhibit extensive variability among SRK alleles (SI Fig. 8). For example, PAN6 and PAN13 differ by 12 substitutions and three insertions, whereas PANa differs from PANb by 23 substitutions (Fig. 5). To identify residues responsible for specific eSRK self-interaction from the total number of varying sites within the PAN_APPLE domain, a multiple sequence alignment of PAN_APPLE domains from various SRK variants was subjected to correlation analysis of amino acid substitutions by using the CRASP package (http://wwwmgs.bionet.nsc.ru/mgs/programs/crasp/). This analysis is based on the assumption that, during evolution, substitutions at residues whose integral physicochemical characteristics are important for the overall conformation of a protein or protein domain occur in a coordinated manner (31). Thirteen covarying amino acid residues (Fig. 5) were identified by CRASP analysis of the eSRK. Interestingly, most of these residues do not vary among individual alleles; rather, they appear to be specific to classes of alleles (e.g., compare B. oleracea class I and class II), apparently reflecting the evolutionary relatedness of alleles within each class. Furthermore, these residues map to the buried core of the PAN_APPLE structural model (Fig. 2C), as expected for residues important for the structural integrity of the domain. Therefore, these covarying residues are unlikely to contribute to self-interaction specificity of individual eSRKs.

Fig. 5.
Multiple sequence alignment of the PAN_APPLE domain of various SRK alleles from B. oleracea and A. lyrata. Numbering of residues is based on the SRK6 sequence. B. oleracea SRK alleles are grouped into two previously identified allelic classes. Shaded ...

Inspection of the remaining variable residues in the multiple sequence alignment reveals a concentration of variability in a small region, corresponding to residues 410–420 in the eSRK6 sequence (Fig. 5), which is predicted to reside within a surface-exposed loop in the PAN_APPLE structural model (Fig. 2C), and might therefore be available for intermolecular interactions. Amino acid residues within this VR might represent flexible sites on which selection might act without disrupting the structural integrity of the domain. Consistent with the notion that this VR contributes to specificity in eSRK self-interaction, the chimeric PAN_APPLE domain, PAN13(VR6), created by replacing the VR of PAN13 with that of PAN6, exhibited a clear preference for interaction with PAN6 over PAN13 (Fig. 4C). Furthermore, the composition of the VR alone can allow interaction between PAN_APPLE domains that do not normally interact. For example, simply replacing the VRa region of PANa, which does not interact with PAN6, with the VRb region of PANb, which shows strong interaction with PAN6, converts PANa into a PAN6 interactor (Fig. 4C). Similarly, replacing the VR of PAN13, which does not interact with PANb, with the VR of PAN6, which interacts with PANb, confers on PAN13 the ability to interact with PANb (Fig. 4C). Although interactions of A. lyrata eSRKs with B. oleracea eSRKs do not occur in nature, these results serve to underscore the importance of the PAN_APPLE VR in determining specificity in eSRK self-interaction.

In conclusion, our results show that self-association is an inherent property of the eSRK. Of the four eSRK structural modules predicted by homology modeling/fold recognition techniques, only the PAN_APPLE domain and to a lesser extent, the EGF-like domain exhibited self-interaction in yeast. Because the SRK kinase domain does not self-interact in the yeast two-hybrid system (data not shown) and SRK dimerizes in the absence of the kinase domain (6), our data are consistent with these eSRK domains being the major PLADs that mediate ligand-independent SRK–SRK and SRK–eSRK dimerization in unstimulated stigma epidermal cells. However, we cannot rule out the possibility that additional PLADs might occur in the SRK transmembrane region.

Our results also warrant a reevaluation of the functional significance of SRK polymorphisms, which have previously been considered only in relation to specificity in the interaction of SRK with its SCR ligand. The observations that eSRK exhibits a marked preference for homotypic interactions over heterotypic interactions with eSRKs from other alleles and that this previously unknown specificity in eSRK self-interaction is mediated largely by variable residues in the PAN_APPLE domain demonstrate that some of the allelic variability in eSRK underlies specificity in receptor self-association, rather than specificity in the eSRK–SCR interaction.

Self-incompatible individuals are typically heterozygous and therefore the potential exists for the two coexpressed SRK variants to form heterodimers. Because receptor heteromers and homodimers can have different specificities (32, 33), the relative proportions of SRK homo- and heterodimers that occur at the stigma surface are likely to have important implications for SRK function. A well known but unexplained feature of crucifer self-incompatibility relates to the complex allelic interactions exhibited by SRKs in heterozygous stigmas, including codominance, nonlinear dominant-recessive relationships, and mutual weakening (34, 35). It is tempting to speculate that these differences in dominance relationships are based on differences in the propensity of particular pairs of SRKs to form heterodimers or in the affinity or responsiveness of these heterodimers toward SCRs. Under this hypothesis, codominance, which characterizes a large proportion of SRK alleles, would reflect the preference for homodimerization revealed by our analysis and would be associated with the predominant or exclusive formation of SRK homodimers (Fig. 6). In contrast, dominant-recessive and mutual weakening relationships might result from increased propensity for heterodimerization, with heterodimers having reduced affinity for one or both SCRs, respectively.

Fig. 6.
Preference for SRK homodimerization as a possible basis for codominant SRK allelic interactions. The example shows SRK dimers displayed at the surface of a stigma heterozygous for the codominant S-locus variants S1 and S2, with SRK1–SRK2 heterodimers ...


Yeast Surface Display and Two-Hybrid Analysis.

The B. oleracea SRK6 and SRK13 and A. lyrata SRKa and SRKb were used in this study. For yeast interaction assays, full-length eSRKs lacking signal peptides and eSRK subfragments were amplified by PCR from appropriate plasmid templates. Chimeric PAN_APPLE domains were generated by using recombinant PCR. Amplified fragments were cloned either into the pYD1 yeast surface display vector (Invitrogen, Carlsbad, CA) or the yeast two-hybrid vectors: the activation domain (AD) vector pGAD-C1, which contains the LEU2 gene for selection, and the binding domain (BD) vectors pGBD-C1or pGBDU-C1, which carry the TRP1 and URA3 marker genes, respectively (23). All constructs were sequenced before yeast transformation.

Each of the pYD1-based constructs encoded a chimeric protein carrying the AGA2 signal sequence and three tags: the Xpress epitope at the N terminus, and the V5 and 6×His tags at the C terminus. Protein expression was confirmed by using anti-V5 antibody and agglutination assays were performed essentially as described (ref. 12; SI Text). Similarly, the AD and BD fusions carried a C-terminal FLAG tag and their expression was confirmed by using anti-FLAG antibody (SI Fig. 11).

To test for interactions in the yeast two-hybrid system, a pair of plasmids containing sequences for the proteins of interest was transformed into the yeast two-hybrid strain PJ69-4a (23), and double transformants were selected on synthetic complete (SC) medium lacking specific nutrient(s) (Bio 101, Vista, CA). Activation of the HIS3 and ADE2 reporter genes was tested as previously described (36) by growth of the double transformants on SC-histidine containing 2.5–5 mM 3-aminotriazole (3AT) and SC-adenine, respectively. Each construct was tested for ability to activate reporter genes in the absence of interacting partner. All assays were repeated at least twice. In this system, activation of the ADE2 reporter gene results only from strong protein–protein interactions. In contrast, the HIS3 reporter gene can detect both strong and weak interactions, with interaction strength being proportional to the intensity of HIS3 gene expression (estimated by extent of cell growth in the presence of progressively higher and more stringent concentrations of 3AT).

Quantitation of β-galactosidase activity was performed by using o-nitrophenyl-β-d-galactopyranoside as substrate (37). One Miller unit of β-galactosidase is defined as the amount of enzyme that hydrolyzes 1 μmol of o-nitrophenyl-β-d-galactopyranoside to o-nitrophenol and d-galactose per minute per cell (38), which was calculated as follows: units = 1,000 × OD420/[volume of cells assayed × incubation time (minutes) × OD600].

Pull-Downs of SRK from Stigma Extracts.

Full-length eSRK6 and eSRK13 were expressed as soluble secreted glycoproteins carrying a C-terminal FLAG tag in N. benthamiana leaves as previously described (3, 30), and expression of the eSRKs was confirmed by protein immunoblot analysis by using mAbH8, an antibody that recognizes eSRK6 and eSRK13 (38). Anti-FLAG M2-agarose (Sigma, St. Louis, MO) was mixed with eSRK-FLAG in a solution containing 30 mM Tris (pH 7.4), 150 mM NaCl, and protease inhibitor mixture (Sigma), and incubated at 4°C for 1 h. For each precipitation reaction, a 5-μg aliquot of stigma protein extract was added and incubation was continued for another hour at 4°C. After washing the agarose beads in the same buffer (three washes, 15 min each), proteins were eluted with SDS gel loading buffer and subjected to electrophoresis on 4–20% (wt/vol) polyacrylamide gels. A polyclonal antibody that reacts with the kinase domain of SRK6 but not that of SRK13 (39) was used to probe the blots.

Supplementary Material

Supporting Information:


We thank Nathan Boggs for the PANa-AD, PANb-AD, and PANb-BD clones. This work was supported by National Institutes of Health Grant GM057527 (to J.B.N.). The structural modeling was carried out by using the resources of the Computational Biology Service Unit from Cornell University, which is partially funded by Microsoft (Redmond, WA).


S-locus receptor kinase
S-locus cysteine-rich
SRK extracellular region
preligand association domain
S-locus glycoprotein consensus sequence
C-terminal region
activation domain
binding domain
variable region
synthetic complete


The authors declare no conflict of interest.

This article contains supporting information online at www.pnas.org/cgi/content/full/0705186104/DC1.


1. Nasrallah JB. Trends Immunol. 2005;26:412–418. [PubMed]
2. Takayama S, Isogai A. Annu Rev Plant Biol. 2005;56:467–489. [PubMed]
3. Kachroo A, Schopfer CR, Nasrallah ME, Nasrallah JB. Science. 2001;293:1824–1826. [PubMed]
4. Takayama S, Shimosato H, Shiba H, Funato M, Che FS, Watanabe M, Iwano M, Isogai A. Nature. 2001;413:534–538. [PubMed]
5. Giranton JL, Dumas C, Cock JM, Gaude T. Proc Natl Acad Sci USA. 2000;97:3759–3764. [PMC free article] [PubMed]
6. Shimosato H, Yokota N, Shiba H, Iwano M, Entani T, Che F-S, Watanabe M, Isogai A, Takayama S. Plant Cell. 2007;19:107–117. [PMC free article] [PubMed]
7. Chan FK, Chun HJ, Zheng L, Siegel RM, Bui KL, Lenardo MJ. Science. 2000;288:2351–2354. [PubMed]
8. Wang X, Li X, Meisenhelder J, Hunter T, Yoshida S, Asami T, Chory J. Dev Cell. 2005;8:855–865. [PubMed]
9. Schlessinger J. Cell. 2002;110:669–672. [PubMed]
10. Jiang G, Hunter T. Curr Biol. 1999;9:R568–R571. [PubMed]
11. Boder ET, Wittrup KD. Nat Biotechnol. 1997;15:553–557. [PubMed]
12. Zou W, Ueda M, Murai T, Tanaka A. Yeast. 2000;16:995–1000. [PubMed]
13. Berman HM, Westbrook J, Feng Z, Gilliland G, Bhat TN, Weissig H, Shindyalov IN, Bourne PE. Nucleic Acids Res. 2000;28:235–242. [PMC free article] [PubMed]
14. Daly R, Hearn MT. J Mol Recognit. 2005;18:119–138. [PubMed]
15. Hester G, Kaku H, Goldstein IJ, Wright CS. Nat Struct Biol. 1995;2:472–479. [PubMed]
16. Chandra NR, Ramachandraiah G, Bachhawat K, Dam TK, Surolia A, Vijayan M. J Mol Biol. 1999;285:1157–1168. [PubMed]
17. Sauerborn MK, Wright LM, Reynolds CD, Grossmann JG, Rizkallah PJ. J Mol Biol. 1999;290:185–199. [PubMed]
18. Wright LM, Reynolds CD, Rizkallah PJ, Allen AK, Van Damme EJ, Donovan MJ, Peumans WJ. FEBS Lett. 2000;468:19–22. [PubMed]
19. Nishio T, Kusaba M. Ann Bot. 2000;85:141–146.
20. Sainudiin R, Wong WS, Yogeeswaran K, Nasrallah JB, Yang Z, Nielsen R. J Mol Evol. 2005;60:315–326. [PubMed]
21. Tagaki J, Beglova N, Yalamanchili P, Blacklow SC, Springer TA. Proc Natl Acad Sci USA. 2001;98:11175–11180. [PMC free article] [PubMed]
22. Wouters MA, Rigoutsos I, Chu CK, Feng LL, Sparrow DB, Dunwoodie SL. Protein Sci. 2005;14:1091–1103. [PMC free article] [PubMed]
23. James P, Halladay J, Craig EA. Genetics. 1996;144:1425–1436. [PMC free article] [PubMed]
24. Koizumi K, Naramoto S, Sawa S, Yahara N, Ueda T, Nakano A, Sugiyama M, Fukuda H. Development (Cambridge, UK) 2005;132:1699–1711. [PubMed]
25. Koster J, Geerts D, Favre B, Borradori L, Sonnenberg A. J Cell Sci. 2003;116:387–399. [PubMed]
26. Park AR, Cho SK, Yun UJ, Jin MY, Lee SH, Sachetto-Martins G, Park OK. J Biol Chem. 2001;276:26688–26693. [PubMed]
27. Tang W, Ezcurra I, Muschietti J, McCormick S. Plant Cell. 2002;14:2277–2287. [PMC free article] [PubMed]
28. Tang W, Kelley D, Ezcurra I, Cotter R, McCormick S. Plant J. 2004;39:343–353. [PubMed]
29. Cheng Q, Sun MF, Kravtsov DV, Aktimur A, Gailani D. J Thromb Haemost. 2003;1:2340–2347. [PubMed]
30. Chookajorn T, Kachroo A, Ripoll DR, Clark AG, Nasrallah JB. Proc Natl Acad Sci USA. 2004;101:911–917. [PMC free article] [PubMed]
31. Afonnikov D, Kolchanov NA. Nucleic Acids Res. 2004;32:64–68. [PMC free article] [PubMed]
32. Akira S. J Biol Chem. 2003;278:38105–38108. [PubMed]
33. Janeway CA, Medzhitov R. Annu Rev Immunol. 2002;20:197–216. [PubMed]
34. Hatakeyama K, Takasaki T, Suzuki G, Nishio T, Watanabe M, Isogai A, Hinata K. Plant J. 2001;26:69–76. [PubMed]
35. Thompson KF, Taylor JP. Heredity. 1966;21:345–362.
36. Naithani S, Saracco SA, Butler CA, Fox TD. Mol Biol Cell. 2003;14:324–333. [PMC free article] [PubMed]
37. Miller JH. Experiments in Molecular Genetics. Cold Spring Harbor, NY: Cold Spring Harbor Lab Press; 1972.
38. Kandasamy MK, Paolillo DJ, Faraday CD, Nasrallah JB, Nasrallah ME. Dev Biol. 1989;134:462–472. [PubMed]
39. Stein JC, Nasrallah JB. Plant Phys. 1993;101:1103–1106. [PMC free article] [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...