Generation of Self-Incompatible Arabidopsis thaliana by Transfer of Two S Locus Genes from A. lyrata

See allHide authors and affiliations

Science  12 Jul 2002:
Vol. 297, Issue 5579, pp. 247-249
DOI: 10.1126/science.1072205


Transitions from cross-fertilizing to self-fertilizing mating systems have occurred frequently in natural and domesticated plant populations, but the underlying genetic causes are unknown. We show that gene transfer of the stigma receptor kinase SRK and its pollen-borne ligand SCR from one S-locus haplotype of the self-incompatible and cross-fertilizing Arabidopsis lyratais sufficient to impart self-incompatibility phenotype in self-fertile Arabidopsis thaliana, which lacks functional orthologs of these genes. This successful complementation demonstrates that the signaling cascade leading to inhibition of self-related pollen was maintained in A. thaliana. Analysis of self-incompatibility will be facilitated by the tools available in this species.

Self-incompatibility (SI) is a widespread mechanism assuring cross-fertilization (outbreeding) in plant species. The family Brassicaceae (crucifers) includes self-incompatible and self-fertilizing species and has emerged as a model for investigating the evolution of mating systems in plants. In this family, SI is controlled by haplotypes of the S locus, whereby cells of the stigma epidermis recognize and inhibit self-related but not genetically unrelated pollen. Mathematical modeling (1), phylogenetic data (2), and mutational (3) studies argue that SI is the ancestral condition and that loss-of-function mutations in genes required for SI have occurred repeatedly to generate self-fertile (autogamous) species and inbreeding populations within a species. As a primary determinant of the outbreeding mating habit in crucifers, the S locus is key to understanding the evolutionary switch from SI to self-fertility in this family. Indeed, inactivation of S-locus genes may be the principal mutation underlying the switch to autogamy in the model plant A. thaliana (4). To test this hypothesis, we attempted to complement the self-fertile phenotype ofA. thaliana by transformation with S-locus genes from A. lyrata, an obligate outbreeder that diverged fromA. thaliana ∼5 million years ago (2).

The S locus of crucifers is complex. In Brassicaspecies (5) and in A. lyrata(4), the SI-specificity encoding region of this locus contains two genes, the SRK (S-locus receptor kinase) and SCR (S-locus cysteine rich protein) genes, which are the determinants of SI in stigma and pollen, respectively (6, 7). A third gene found in the majority of Brassica S haplotypes (5) but absent in A. lyrata (4) is the SLG(S-locus glycoprotein) gene, which encodes a secreted protein that enhances the activity or stability of SRK (5,6). The highly polymorphic SRK and SCR proteins function as receptor and ligand (8, 9) whose allele-specific binding triggers an intracellular cascade in the stigma epidermal cells that culminates in the failure of self pollen to hydrate and germinate on the stigma epidermis.

We transformed A. thaliana ecotype Columbia, which contains defective SRK and SCR genes (4), with the SRKb and SCRballeles isolated from the A. lyrata S-locus haplotypeSb (4). Eight independent SRKband nine independent SCRb transformants were generated (10). Primary (T1) transformants expressing eitherSRKb or SCRb were self-fertile. We chose threeSRKb transformants that showed fewer pollen tubes (10) when pollinated with pollen from the transgenicSCRb plants, reasoning that these SRKbtransformants had near-adequate levels of SRKbexpression. Plants homozygous for each of SRKb andSCRb were obtained by self-pollination (10) and were crossed to generate plants harboring both the SRKb andSCRb transgenes, which we designate SRKb/SCRbplants (Fig. 1A).

Figure 1

Generation of SRKb/SCRb transformants and expression of SI in A. thaliana. (A) Generation and RNA analysis of SRKb/SCRb transformants starting from T1A. thaliana ecotype Columbia plants transformed withSRKb (col::SRKb) or SCRb (col::SCRb). Results of RT-PCR (10) for representative T2 plants in each of two SRKb (plants 2-6 and 4-2) and two SCRb (plants 6-5 and 9-7) transgenic families, and for two SRKb/SCRb plants are shown. Controls: untransformed A. thaliana ecotype Columbia (col); A. lyrata (A.l.) RNA and genomic DNA (G). Amplification with (+) and without (–) reverse transcriptase. The dot and arrowhead mark the SRKb and SCRb RT-PCR products, respectively. The asterisks indicate PCR products derived from contaminating genomic DNA (10). The RNA gel blot showsSRKb transcripts in the pistils of SRKb/SCRbplants and A. lyrata SaSb. Hybridization with an actin probe served as a loading control (10). (B) Developmental regulation of SI in SRKb/SCRb stigmas from stage-12, -13, early -14 (14E), and late -14 (14L) buds. Numbers of pollen tubes produced per stigma in self- and cross-pollinations are the means (±SE) determined from analysis of two inflorescences in each of six SRKb/SCRb plants.

The presence of the transgenes was confirmed, and their expression was demonstrated by reverse transcriptase–polymerase chain reaction (RT-PCR) in T2 plants and in the SRKb/SCRb plants (10) (Fig. 1A). Pollination phenotype was determined at two stages of floral bud development (11), mature buds (stage 13) and young flowers (early stage 14), before anthers had extended above the stigma and deposited their pollen on the stigma epidermis (10). The inhibition of pollen from SCRbtransformants on SRKb stigmas was recapitulated in the T2 generation (Table 1). In theSRKb/SCRb plants, a robust SI response was manifested by complete or near-complete inhibition of self pollen (Table 1,Figs. 1B and 2A). The selective inhibition of self pollen was clearly demonstrated by simultaneous selfing and crossing of individual stigmas (Fig. 2B), a method we used to monitor the response of a single stigma to differing pollen genotypes (10). Furthermore, the pollen of SRKb/SCRbplants failed to grow on the stigmas of SRKb-expressing plants, and their stigmas inhibited pollen from SCRbtransformants (Table 1). The specificity of these interactions was further demonstrated by interspecies pollinations withA. lyrata (10). Whereas untransformed A. thaliana stigmas allowed the development of A. lyratapollen tubes irrespective of their S-locus genotype,SRKb/SCRb stigmas inhibited A. lyrata pollen fromSb homozygotes but not from plants homozygous for theS-locus haplotype Sa (Table 1, Fig. 2C), demonstrating that these transgenic plants had acquired Sbspecificity.

Figure 2

Pollination assays of SRKb/SCRbstigmas. (A) Pollination test of an early stage-14SRKb/SCRb stigma. The stigma was pollinated with pollen from untransformed A. thaliana on its left side and self pollen on its right side. The two images of the same stigma were taken with different filters and show aniline blue fluorescence of pollen tubes (left) and autofluorescence of pollen grains (right). Incompatible pollen grains adhere poorly and are normally dislodged from the stigma surface upon processing for microscopy (Fig. 1B), unless mild conditions are used (10), as was the case for this stigma. Pollen tubes grow on the cross-pollinated side, but no tubes are formed on the self-pollinated side. Some pollen grains (arrows) are visible in both images and provide landmarks for comparison of the images. Bar, 30 μm. (B) Single-stigma self- or cross-pollination test ofSRKb/SCRb stigmas from stage-12, 13, early -14 (14E), and late -14 (14L) buds. The left half of each stigma was self-pollinated, and the right half was cross-pollinated with pollen from untransformed A. thaliana. The SI response is manifested by absence of pollen tubes on the self-pollinated half of stage-13 and stage-14E stigmas. Bar, 50 μm. (C) Interspecies pollination of A. thaliana stigmas with pollen from an A. lyrata SbSb plant. Left, compatible response of an untransformed control. Bar, 50 μm. Right, incompatible response of an SRKb/SCRb transformant, in which short pollen tubes (arrows) that do not invade the stigma epidermis were produced. Bars, 20 μm.

Table 1

Pollination responses in crosses among transgenic lines (col::SRKb, col::SCRb, col::SRKb/SCRb), with untransformed A. thalianaecotype Columbia (col), and with A. lyrata Saand Sb homozygotes. +, >100 pollen tubes per stigma; –, 0 to 10 pollen tubes per stigma. Based on pollinations of stage-13 and early stage-14 stigmas (10).

View this table:

A hallmark of SI in crucifers is that the ability of the stigma to reject self pollen is developmentally regulated and is first manifested in mature buds just before flower opening. Immature bud stigmas are self-compatible, which allows the generation ofS-locus homozygotes. In Brassica, this regulation reflects the progressive increase in the amount of SRK to maximal levels in mature buds and young flowers (12). In olderBrassica flowers, SRK transcript and protein levels decline, although they usually remain above the threshold required for maintenance of the SI response (12). SI phenotype was also developmentally regulated in the stigmas of A. thaliana SRKb/SCRb transformants (Figs. 1B and 2B), albeit over a narrower window than Brassica due to the accelerated developmental program of this species. Stigmas of immature buds were about equally receptive to self and cross (unrelated) pollen, whereas stage-13 to early stage-14 stigmas inhibited self pollen but remained receptive to cross pollen. Unlike Brassica, however, olderSRKb/SCRb flowers regained a measure of receptivity to self pollen (Figs. 1B and 2B).

The developmental weakening of SI in SRKb/SCRb transformants resembles the pseudo-compatible response observed in some self-incompatible crucifer cultivars (4, 13).SCRb transcript levels were about equivalent in A. thaliana SRKb/SCRb transformants and A. lyrata SaSbplants (data not shown). However, transgene-encoded SRKbtranscript levels were only ∼40% of those in A. lyrata SaSb, as determined by RNA gel blot analysis (Fig. 1A) (10), and may be reduced below a critical threshold in older flowers. In the future, it might be possible to increase the strength of the SI response in transgenic A. thaliana by using highly active stigma-specific promoters, a strategy that proved effective for expressing SRK in transgenic Brassica(6).

Plants offer several examples of dramatic interspecific trait differences that result from mutations at a small number of major genes (14, 15). Our results show that acquisition of just two key S-locus recognition genes is sufficient to reverse the self-fertile phenotype of A. thaliana to the ancestral state of SI in mature buds and young flowers. The acquisition of Sb specificity in A. thaliana SRKb/SCRb transformants demonstrates that the shift to autogamy inA. thaliana resulted from inactivation of S-locus recognition genes. It appears that all other components of the SRK-mediated signal transduction pathway have been maintained in this species, possibly because they have essential functions unrelated to SI. In view of this result, the failure of a previous attempt at conferring a SI specificity on A. thaliana stigmas by transformation with a Brassica SRK gene (16) is difficult to explain. It might have been due to factors related to the evolutionary divergence of Brassicaand Arabidopsis, such as aberrant maturation of theBrassica SRK protein in A. thaliana stigmas or its inability to interact productively withArabidopsis-derived downstream targets.

The large number of genetically well-characterized Shaplotypes that are available in Brassica species has been critical for identification of the SRK and SCR SI recognition proteins. However, the relatively laborious transformation methods and rudimentary state of genome studies in Brassica make further studies of the SI response difficult. The availability of A. thaliana strains that express SI provides new opportunities for exploiting the tools of this tractable model plant for structure-function studies of SRK and SCR as well as for the genetic and molecular dissection of the SRK-mediated signal transduction pathway.

  • * To whom correspondence should be addressed. E-mail: men4{at}


View Abstract

Navigate This Article