A subunit of the oligosaccharyltransferase complex is required for interspecific gametophyte recognition in Arabidopsis

Abstract

Species-specific gamete recognition is a key premise to ensure reproductive success and the maintenance of species boundaries. During plant pollen tube (PT) reception, gametophyte interactions likely allow the species-specific recognition of signals from the PT (male gametophyte) by the embryo sac (female gametophyte), resulting in PT rupture, sperm release, and double fertilization. This process is impaired in interspecific crosses between Arabidopsis thaliana and related species, leading to PT overgrowth and a failure to deliver the sperm cells. Here we show that ARTUMES (ARU) specifically regulates the recognition of interspecific PTs in A. thaliana. ARU, identified in a genome-wide association study (GWAS), exclusively influences interspecific--but not intraspecific--gametophyte interactions. ARU encodes the OST3/6 subunit of the oligosaccharyltransferase complex conferring protein N-glycosylation. Our results suggest that glycosylation patterns of cell surface proteins may represent an important mechanism of gametophyte recognition and thus speciation.

ARTICLE

Received 21 Aug 2015 | Accepted 21 Jan 2016 | Published 11 Mar 2016

Species-specic gamete recognition is a key premise to ensure reproductive success and the maintenance of species boundaries. During plant pollen tube (PT) reception, gametophyte interactions likely allow the species-specic recognition of signals from the PT (male gametophyte) by the embryo sac (female gametophyte), resulting in PTrupture, sperm release, and double fertilization. This process is impaired in interspecic crosses between Arabidopsis thaliana and related species, leading to PT overgrowth and a failure to deliver the sperm cells. Here we show that ARTUMES (ARU) specically regulates the recognition of interspecic PTs in A. thaliana. ARU, identied in a genome-wide association study (GWAS), exclusively inuences interspecicbut not intraspecicgametophyte interactions. ARU encodes the OST3/6 subunit of the oligosaccharyltransferase complex conferring protein N-glycosylation. Our results suggest that glycosylation patterns of cell surface proteins may represent an important mechanism of gametophyte recognition and thus speciation.

DOI: 10.1038/ncomms10826 OPEN

A subunit of the oligosaccharyltransferase complex is required for interspecic gametophyte recognition in Arabidopsis

Lena M. Mller1,w, Heike Lindner1,w, Nuno D. Pires1, Valeria Gagliardini1 & Ueli Grossniklaus1

1 Department of Plant and Microbial Biology and Zrich-Basel Plant Science Center, University of Zrich, Zollikerstrasse 107, 8008 Zrich, Switzerland.w Present addresses: Boyce-Thompson Institute for Plant Research, 533 Tower Road, Ithaca, New York 14853, USA (L.M.M.); Carnegie Institution for Science,

Department of Plant Biology, 260 Panama Street, Stanford, California 94305, USA (H.L.). Correspondence and requests for materials should be addressed to U.G. (email: mailto:[email protected]

Web End [email protected] ).

NATURE COMMUNICATIONS | 7:10826 | DOI: 10.1038/ncomms10826 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 1

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10826

Species evolve and are maintained by a variety of hybridization barriers that prevent interspecic gene ow and thus the formation of potentially unviable or sterile hybrids1. To

date, the molecular basis of hybridization barriers is still poorly understood. In plants, such barriers can either act before (pre-pollination barriers) or after pollination (post-pollination barriers). Pre-pollination barriers can be spatial or temporal patterns preventing plants from being pollinated by pollen from a different species, whereas post-pollination barriers come into play only after an interspecic pollination event occurs and can be further divided into pre- and post-zygotic barriers2. The latter usually act at the genomic level (for example, incompatibilities leading to hybrid lethality or sterility), while pre-zygotic barriers prevent the formation of a zygote and usually rely on direct cellcell communication between the male and female tissues. Most species pairs are isolated by a complex interplay of different types of isolation barriers. Whereas barriers that prevent fertilization (both pre-pollination and pre-zygotic barriers) often represent the most important means to reduce interspecic gene ow, post-zygotic hybridization barriers appear to contribute less to reproductive isolation in many species pairs3.

In plants, successful fertilization starts with the deposition of intraspecic (same-species) pollen (male gametophyte) onto the stigma of a gynoecium. The subsequent steps involve extensive communication between the male and female tissues, leading to pollen adherence, hydration, and the germination of a pollen tube (PT). Within its cytoplasm, the tip-growing PT transports the two sperm cells through the transmitting tract of the pistil to the embryo sac (female gametophyte), which is deeply embedded in the ovule, the precursor of seed. During its journey, the PT is guided towards the embryo sacs by attractants secreted by female tissues4. On arrival at the embryo sac, communication between the PT and the synergid cells of the female gametophyte is initiated (Fig. 1a). The two synergid cells are located at the micropylar end of the embryo sac and possess a secretory region characterized by membrane invaginations and thickened cell wall structures5. This so-called liform apparatus is the rst point of contact between the male and female gametophytes, which communicate in preparation for penetration of the receptive synergid cell by the PT, PT rupture, sperm release, and double fertilization6. While one sperm fuses with the egg cell to form the diploid zygote, the other fertilizes the homo-diploid central cell to produce the triploid endosperm, an embryo-nourishing tissue. The communication process between the male and female gametophytes leading to PT rupture and sperm cell discharge is known as PT reception, and its success or failure is under female gametophytic control6.

However, if a pollen grain originating from a different species (interspecic pollination) is placed on a plants stigma, all the communication processes described above have the potential to act as pre-zygotic post-pollination barriers. Several studies describe a species-preferential behaviour of molecular factors involved in pollen adherence to the stigma, PT growth, and PT guidance towards the ovules4,711. In interspecic crosses between closely related Ericaceae or Brassicaceae, respectively, hybridization barriers act at the stage of PT reception12,13. In such crosses, PTs are properly targeted to the female gametophyte but, upon arrival at the embryo sacs, interspecic PTs are not recognized and fail to arrest growth and discharge their sperm. Instead, they continue growing inside the embryo sacs (referred to as PT overgrowth) and cannot effect double fertilization. Therefore, we consider PT reception to be an integral part of the hybridization barrier in these species. Interspecic PT overgrowth phenocopies the female gametophytic mutants feronia/sirne (fer/srn), lorelei (lre), nortia (nta), turan (tun), evan (evn), and Zea mays embryo sac 4 (ZmES4) RNAi-lines1320, which are

defective in the reception of intraspecic PTs. In addition, FER has been proposed to be involved in interspecic PT recognition13, and there is evidence that ZmES4 is sufcient to trigger PT growth arrest and rupture in a species-preferential manner20.

Despite the rapid advance in our understanding of the molecular basis of intraspecic PT reception21, the genetic basis of post-pollination hybridization barriers remains largely unknown. All molecular factors that have so far been described to be involved in pre-zygotic species-discrimination, including species-preferential pollen adherence, PT guidance, growth, and reception, act primarily during intraspecic pollination and have additional species-preferential effects4,711. Here, we report the identication of the rst gene required exclusively for inter- but not for intraspecic pollination, thus likely representing a specic component for the establishment of a hybridization barrier. By making use of the striking natural variation of A. thaliana accessions in interspecic PT reception, we identied ARTUMES (ARU) as an indispensable factor for the recognition of interspecic A. lyrata PTs by A. thaliana embryo sacs. In contrast, aru mutants do not affect gametophytic communication in intraspecic A. thaliana crosses. ARU encodes the OST3/6 subunit of the oligosaccharyltransferase complex, which is known to regulate site- and substrate-specic N-glycosylation of proteins in yeast22,23 and a similar substrate specicity has been reported for A. thaliana OST3/6 (ref. 24). Thus, a possible mechanism for the discrimination of inter- and intraspecic PTs may depend on the species-specic glycosylation of proteins on the surface of the synergid cells of the female gametophyte.

ResultsPT overgrowth restricts gene ow between Arabidopsis species. Self-fertilizing A. thaliana and its outcrossing relative A. lyrata are separated by strong pre-pollination barriers due to their different mating systems1. In addition, they are isolated by post-pollination barriers based on direct malefemale interactions. Although A. thaliana (Col-0) pollen germination is inhibited at the A. lyrata stigma, A. lyrata PTs are guided towards A. thaliana embryo sacs, but PT reception fails (Fig. 1c, as opposed to Fig. 1b showing successful PT reception, Supplementary Fig. 1). Such unilateral incompatibility is similar to that observed in other crosses between self-compatible and self-incompatible species25. We also observed PT overgrowth in interspecic crosses betweenA. arenosa and A. lyrata (Supplementary Fig. 2), between which natural gene ow occurs26. This nding indicates that A. lyrata PT overgrowth in A. thaliana ovules does not only occur between species that do not interbreed in nature (A. thaliana A. lyrata)

but also between species that are only partially reproductively isolated and do interbreed (A. arenosa A. lyrata).

Natural variation in interspecic PT reception. To analyse interspecic hybridization barriers within the genus Arabidopsis, we assessed PT overgrowth in 86 A. thaliana accessions that were pollinated with A. lyrata pollen (Supplementary Table 1). PTs were visualized by staining callose in PT cell walls with aniline blue. We scored the proportion of ovules that failed to recognize interspecic PTsleading to PT overgrowthin relation to the total number of ovules that attracted a PT in a silique (overgrowth per silique, OG/S). We found a striking variation in the ability to recognize interspecic PTs between different A. thaliana accessions, with OG/S ranging from about 10 to 90% (Fig. 1d, broad-sense heritability H2 0.7). Examples

of accessions with extreme phenotypes are Lz-0 (10% OG/S, n 12 siliques) and Kz-9 (87.3% OG/S, n 10; Fig. 1e,f). There is

2 NATURE COMMUNICATIONS | 7:10826 | DOI: 10.1038/ncomms10826 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10826 ARTICLE

a

b c

Central cell

Antipodal cells

Egg cell

Synergid cells

Filiform apparatus

Pollen tube (PT)

d e f

A. thaliana accessionsA. lyrata

100

80

60

OG/S (%)

*

*

40

20

0

A. thaliana accessions

Lz-0 Kz-9

Figure 1 | PT reception in interspecic crosses. (a) Diagram of the female gametophyte with its cell types. The synergids with their membrane-rich liform apparatus are crucial for communication with the arriving PT. (b) Ovule with normal PT reception, visualized by callose staining of the PT cell walls with Aniline Blue. The PT stopped its growth and ruptured. Dashed line indicates outline of the ovule. (c) Ovule with PT overgrowth. The PT continues growing inside the female gametophyte. Dashed line indicates outline of the ovule. (d) Natural variation in the proportion of ovules with PT overgrowth per silique (OG/S) in 86 A. thaliana accessions that were pollinated with A. lyrata pollen. OG/S varies between 10% and more than 90%, depending on the genotype of the mother. (e) A silique of Lz-0 pollinated with A. lyrata pollen. Most of the ovules show normal PT reception. Ovules with PT overgrowth are marked with an arrowhead. (f) A silique of Kz-9 pollinated with A. lyrata pollen. Most of the ovules display PT overgrowth. Asterisks mark ovules with normal PT reception. Per accession and per pollen donor, 510 siliques were analysed. Scale bars, 50 mm (b and c), 250 mm (e and f).

Pollen donor: A. halleri

a

Pollen donor: A. lyrata

b

c Pollen donor: A. arenosa

100

80

60

OG/S (%)

40

20

0

Lz0

Kas1

Ga0

Lp26

Col0

Kz1

Nd1

Lz0

Kas1

Ga0

Lp26

Col0

Kz1

Nd1

Lz0

Kas1

Ga0

Lp26

Col0

Kz1

Nd1

Figure 2 | Natural variation in PT reception with different interspecic pollen donors. (a) A subset of A. thaliana accessions pollinated withA. lyrata pollen. Per accession, four to eight siliques were analysed. Box plots are ordered by the mean OG/S value and colour-coded to facilitate comparison with b and c. (b,c) The same subset pollinated with pollen from A. halleri (b) and A. arenosa (c). The accessions show comparable OG/S with all three interspecic pollen donors.

no obvious correlation between the geographical origin of the accessions and their phenotype (Supplementary Fig. 3).

To analyse whether the variation in the ability to recognize interspecic PTs is species-dependent, we pollinated a subset

of A. thaliana accessions with low or high OG/S in crosses withA. lyrata (Lz-0, Kas-1, Ga-0, Lp2-6 and Col-0, Kz-1, Nd-1, respectively) also with pollen of A. halleri and A. arenosa. Although OG/S in the accessions pollinated with A. lyrata or

NATURE COMMUNICATIONS | 7:10826 | DOI: 10.1038/ncomms10826 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 3

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10826

A. halleri pollen was highly comparable (Fig. 2a,b), the values were slightly lower for all accessions when pollinated byA. arenosa (Fig. 2c), indicating that A. thaliana recognizesA. arenosa PTs better than those of A. lyrata or A. halleri. However, accessions showing very low or high OG/S in crosses with A. lyrata, respectively, displayed a similar phenotype withA. halleri and A. arenosa as pollen donors, suggesting a common molecular PT reception mechanism for all three species. Thus, PT overgrowth is a hallmark of interspecic crosses with close Brassicaceae relatives and not a species-specic feature ofA. thaliana and A. lyrata.

To investigate whether intraspecic PT reception was affected

in accessions with high OG/S (Col-0, Kz-1, Kz-9, Nd-1, Fei-0, and Sq-8), we crossed them with A. thaliana pollen (from both low and high OG/S accessions). Intraspecic PT reception was normal in all the tested accessions (Supplementary Fig. 3), indicating that high OG/S frequencies result from a failure in the recognition of interspecic PTs only, and are not due to a general defect in PT reception.

ARTUMES regulates inter- but not intraspecic PT reception. To identify loci causing the variation in interspecic PT reception in A. thaliana, we used publicly available single-nucleotide polymorphism (SNP) data from the 86 accessions to perform a genome-wide association study (GWAS)27. To date, most GWAS in Arabidopsis have identied previously known candidate genes, with only a few studies identifying novel

regulatory genes in the respective pathways28,29. Applying the GLM function implemented in TASSEL30, we identied a region on chromosome 1 containing 8 of the top 20 SNPs with the highest correlation to the OG/S trait (Fig. 3a and Supplementary Table 2). This 28-kb region (positions 22,814,316 to 22,842,689) contains six genes and one pseudogene (Fig. 3b). Interestingly, calculation with mixed linear models that simultaneously correct for population structure and unequal genetic relatedness between individuals masked the peak, whereas it could be detectedalthough below the signicance threshold using a step-wise multi-locus mixed model (MLMM) specically designed for mapping complex traits31 (Supplementary Fig. 4). With each step of MLMM, new peaks appear, consistent with a multigenic basis for interspecic PT reception.

To narrow down the 28-kb candidate region to a single gene, we analysed OG/S in T-DNA insertion lines of three synergid-expressed genes32 in this region because the synergids control PT reception (At1g61780, At1g61790 and At1g61810; as well as At1g61795, for which no expression data was available; Supplementary Fig. 5). In a homozygous T-DNA insertion allele disrupting the coding sequence of At1g61790 (Fig. 4a), an average of 84.3% (n 18 siliques) of ovules

display A. lyrata PT overgrowth, signicantly more than in the Col-0 wild-type control (58.7% OG/S, n 28, Students t-test

Po0.001, Fig. 4b,d,e). We named the At1g61790 gene, which has previously been described as OST3/6 based on homology24,

ARTUMES (ARU) after the Etruscan goddess of night, nature, and fertility33. The T-DNA allele was denoted aru-1.

a

7

6

log(Pvalue)

5

4

3

2

Chromosome 1 Chromosome 2 Chromosome 3

Chromosome 4 Chromosome 5

7

6

5

PERL0198560

PERL0198559

PERL0198538

PERL0198933

PERL0198952 PERL0198924

PERL0198872

PERL0198834

log(Pvalue)

At1g61780 At1g61790 At1g61795 At1g61800 At1g61810 At1g61820 At1g61830

4

3

2

1

0

22,814,000 22,820,000 22,830,000 22,840,000

Position on Chromosome 1

b

Figure 3 | GWAS identies an associated region on chromosome 1. (a) Manhattan plot showing a peak on chromosome 1 (grey box) with its highest correlated SNP showing signicance at Po0.1 (after Bonferroni correction; dotted line). The peak corresponds to a 28 kb region spanning position 22,842,68922,814,316 (magnied in the second panel). The eight SNPs that were identied to be among the 20 most highly correlated ones in the GWAS are annotated with their PERL identiers. (b) Genes and pseudogenes (grey) in the 28 kb region. Genes expressed in synergids are marked in red, genes without available expression data are in orange32.

4 NATURE COMMUNICATIONS | 7:10826 | DOI: 10.1038/ncomms10826 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10826 ARTICLE

a

C388T (>>STOP) aru-2

PERL0198560

SALK_067271 aru-1

PERL0198559

PERL0198538

At1g61780 At1g61790

b

Col and aru mutantsA. lyrata

c

aru-1A. lyrata and A. thaliana

100

***

***

100

90

80

OG/S (%)

80

60

70

OG/S (%)

40

60

50

20

40

0

Col aru-2

aru-1

Mother plant

Pollen donor

d

e

f g

*

* *

A. lyrata Col Self

*

*

*

** *

*

* *

*

**

*

*

*

*

*

*

*

*

*

ColA. lyrata

*

aru-1A. lyrata

aru-2A. lyrata

aru-1Col

Figure 4 | ARTUMES mutants are impaired in interspecic PT reception. (a) Genomic region of ARTUMES (At1g61790) with the two mutant alleles aru-1 and aru-2, and the surrounding polymorphisms identied by GWAS. (b) PT overgrowth of Col-0 wild-type (n 28 siliques) and aru mutant plants in

interspecic crosses. Both aru mutant alleles show signicantly higher proportions of ovules with PT overgrowth per silique (OG/S, Students t-test ***Po0.001; n 18 and 20 siliques, respectively). (c) PTovergrowth of aru-1 in inter- and intraspecic crosses. The mutant is impaired in interspecic crosses

with A. lyrata pollen (n 9 siliques), but not in intraspecic crosses with Col-0 (n 6) or self pollen (n 9). (d) A silique of Col-0 pollinated with A. lyrata

pollen. Ovules with normal PT reception (marked with asterisks) and with PT overgrowth are visible. (e,f) aru-1 and aru-2 siliques pollinated with A. lyrata pollen. Both mutant alleles show high proportions of ovules with PT overgrowth in interspecic crosses. Ovules with normal PT reception are marked with asterisks. (g) A silique of aru-1 pollinated with intraspecic Col-0 pollen. All ovules display normal PT reception. Scale bars, 250 mm.

A second mutant allele, aru-2, carrying an EMS-induced premature stop-codon after aa residue 129 (ref. 34), also showed an increase in interspecic OG/S (96.1%, n 20,

Fig. 4a,b,f). Likewise, aru-1 mutants pollinated with A. arenosa pollen showed signicantly more ovules with PT overgrowth(67.6% OG/S, n 8) than the wild type (34.9% OG/S,

n 12, Students t-test Po0.001, Supplementary Fig. 6),

suggesting a common basis for interspecic PT recognition. In contrast, aru mutant ovules have no problem recognizing and receiving intraspecic PTs from A. thaliana (Fig. 4c,g), indicating that the PT reception pathway is fully functional within the species.

NATURE COMMUNICATIONS | 7:10826 | DOI: 10.1038/ncomms10826 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 5

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10826

a b pMYB98::ARU-GFP in aru/aru A. lyrata

100

***

***

90

80

OG/S (%)

70

60

50

40

30

aru1 Col

pMYB98::ARU-GFP in aru-1

Line 1

Line 2 Line 3

pMYB98::ARU:GFP autofluorescence

Mother plants

Figure 5 | Synergid-specic expression of ARTUMES complements the mutant phenotype. (a) An ovule expressing pMYB98::ARU-GFP in the synergids. Inset: ARU-GFP localizes to perinuclear structures resembling the ER. (b) PT overgrowth in interspecic crosses using aru-1 (n 22 siliques),

Col-0 (n 28), and three independent transformant lines of pMYB98::ARU-GFP in the aru-1 background as mother plants and A. lyrata as pollen donor.

All the three transformant lines complement the mutant phenotype (n 16, 22, and 27 siliques, respectively); line 2 shows even lower OG/S than the wild

type. Signicance levels in comparison with Col-0 (Students t-test ***Po0.001). Scale bar, 50 mm.

To conrm that ARU function is required for interspecic PT recognition in the synergids, we expressed an ARU-GFP translational fusion protein under the control of the MYB98 and FERONIA promoters (pMYB98::ARU-GFP and pFER:: ARU-GFP) in aru-1 mutants. These promoters are highly active in synergids13,35, and in ovules the strongest ARU-GFP signal was detected in these cells (Fig. 5a and Supplementary Fig. 7). ARU-GFP localized to perinuclear structures resembling the ER in synergids (Fig. 5a, inset), and co-localized with an ER-marker in transiently transformed onion epidermal cells (Supplementary Fig. 7). These results are consistent with the previously reported ER-localization of ARU-GFP in inltrated tobacco leaves24. Mutant aru-1 plants expressing a functional copy of ARU-GFP in their synergids displayed wild-type-like PT reception in interspecic crosses (Fig. 5b and Supplementary Fig. 7), indicating that ARU expression in synergid cells is sufcient to complement the aru mutant phenotype. Consistent with this, the ARU-GFP translational fusion protein driven by the endogenous promoter (pARU::ARU-GFP) is highly expressed in wild-type synergids (Supplementary Fig. 7), suggesting an important role for ARU in these cells. Functional complementation of the aru-1 mutant was also observed when ARU was driven by the endogenous promoter in a construct also containing 865 bp downstream sequence (pARU::ARU), although there was more line-to-line variability than when the pMYB98 promoter was used (Supplementary Fig. 7). Since ARU is not fused to a uorescent tag here, technical difculties in measuring synergid specic expression of the transgene make it hard to explore this difference experimentally. However, it is conceivable that additional regulatory sequences, up- or downstream of the ARU coding sequence that are present in the pARU::ARU construct but not in pMYB98::ARU-GFP or pFER::ARU-GFP, contributed to the observed phenotypic variability.

SNPs around ARTUMES correlate with phenotypic variation. We assessed the correlation of amino acid differences in the ARU coding sequence in all accessions to determine whether differences in the protein sequence could explain the phenotypic variation. Within this population, we detected a total of 10 amino acid differences, four of which are signicantly correlated to variation in OG/S (Pearsons correlation coefcient R, Po0.05,

Supplementary Data 1). However, differing residues had similar chemical properties, implying small, if any, differences in protein function. Alternatively, differential expression levels could cause the observed phenotypic variation. Therefore, we examined

alignments of 1,000 bp up- and downstream sequence of ARU. We found 9 of 49 and 39 of 79 upstream and downstream SNPs, respectively, to be correlated with phenotypic variation (Po0.05,

Supplementary Data 1), suggesting that phenotypic variation could be due to differences in gene expression. To investigate this further, we used RNA extracts from pistils and ovules collected 2 days after emasculation from selected accessions (Lz-0, Kas-1, Ga-0 and Col-0, Nd-1, Fei-0, Kz-1, Kz-9) for quantitative real-time PCR and digital droplet PCR, respectively36. We found ARU mRNA levels differed between accessions (Supplementary Fig. 8), but they did not correlate with the OG/S phenotype among the selected set of accessions. Because we used RNA from whole pistils and ovules, we cannot exclude the possibility that ARU is differentially expressed in synergids only, where it is required and sufcient for interspecic PT reception. In addition, post-transcriptional regulation of gene expression could play a role in mediating ARU protein levels. ARU has a 410 bp long 30-UTR (30-untranslated region)37 containing 17 SNPs that are correlated with the OG/S phenotype and might contribute to accession-specic differences in ARU protein levels.

To further investigate the role of ARU in different accessions, we transformed high OG/S accessions (Fei-0, Kz-1, Kz-9) with pMYB98::ARU-GFP to ensure strong expression in synergid cells and assessed interspecic PT reception in these lines. Of several independent transformants of all accessions, none showed a signicant reduction of OG/S (Supplementary Fig. 8), which would be expected if low expression of ARU alone would be the cause of impaired interspecic PT reception in these accessions. Similarly, the ARU allele (including 1,492 bp up- and 865 bp downstream sequence) from Ga-0 could not better complement the aru mutantwhich is in the Col-0 backgroundthan the Col-0 allele, as it would be expected if differences in ARU alone caused the phenotypic difference between Ga-0 and Col-0 (Supplementary Fig. 7). Although these experiments could not establish a mechanistic link between variation in ARU and variation in OG/S, they are not inconsistent with the hypothesis that ARU is differentially regulated in A. thaliana accessions. Because ARU is likely only one of several factors involved in interspecic PT reception, we cannot rule out that crucial epistatic interactions with other factors were missing in the particular accessions we tested. At present, it is conceivable that population structure or co-segregating SNPs, which do not directly inuence ARU expression or protein function, caused the GWAS peak on chromosome 1 and that ARU itselfalthough undoubtedly involved in interspecic PT receptionis not responsible for the observed natural variation among the 86

6 NATURE COMMUNICATIONS | 7:10826 | DOI: 10.1038/ncomms10826 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10826 ARTICLE

accessions tested. We consider co-segregating SNPs a highly unlikely explanation, however, because mutations disrupting other genes in the region identied by GWAS did not show a phenotype in interspecic crosses with A. lyrata pollen.

Signatures of selection at the ARTUMES locus. Genes involved in reproductive isolation and speciation are often subject to selective pressures driving rapid divergence38. We tested ARU plus 1,000 bp up- and downstream sequence for signatures of positive selection by estimating Tajimas D39 for a set of96 A. thaliana accessions40, including all accessions used in this study. A negative D is due to an excess of low frequency polymorphisms that can be caused by positive selection on the locus or by population expansion. Tajimas D was 2.07 for the

1,000 bp upstream of the translation start, 1.57 for the coding

sequence and 1.58 for the 1,000 bp downstream of the ARU

stop-codon. All values signicantly deviate from the neutral model (Po0.05), but do not fall into the 5% tail of the estimated genomic distribution of D in A. thaliana40 (cut-off value: 2.08).

Thus, although the 1,000 bp upstream region was very close to this cut-off, the observed negative values might be inuenced by demographic factors that shaped the entire genome rather than selective pressure acting on the ARU locus. In addition, we estimated Fay and Wus H, another test statistic to detect positive selection41, which is not as sensitive to demographic factors as Tajimas D42. All values for H were strongly negative (upstream region: 20.20, coding sequence: 21.79, downstream region: 20.73, Po0.02), and fall into the very extreme tails of both an empirical and a simulated distribution of H calculated with 12 accessions, which represent a world-wide distribution and are a subset of the accessions used in this study42. Such strongly negative values for H indicate that positive selection may indeed have acted on each of the regions of the ARU locus. The fact that D for the upstream region was very close to the 5% tail and the strongly negative values for H provide evidence for possible positive selection and indicate that ARU may have undergone a recent selective sweep. This is consistent with the lack of variation in ARU amino acid sequence and expression level but is difcult, although not impossible, to reconcile with the fact that we identied ARU in a GWAS for variation in the OG/S phenotype. As pointed out above, this may be due to epistatic interactions with additional factors involved in this complex process. Nevertheless, as in animal speciation genes, selective pressures appear to have contributed to shaping the genetic basis that underlies interspecic PT reception in A. thaliana.

Interspecic PT reception depends on protein N-glycosylation. ARU encodes the OST3/6 subunit of the hetero-oligomeric plant oligosaccharyltransferase complex (OST), which catalyses the co- or posttranslational transfer of pre-assembled carbohydrate oligomers (Glc3Man9GlcNAc2) to asparagine (N) residues of polypeptides43. N-glycosylation affects the substrate proteins folding, targeting, and/or processing through the ER. Subsequently, the N-linked glycan can be modied in the Golgi apparatus in a cell-type and species-specic manner, accounting for the functionality and binding specicity of the glycoprotein43. The yeast OST consists of eight subunits and the homologues of OST3/6, Ost3p and Ost6p, differ in their protein substrate and site-specic glycosylation efciency22,23.

In plants, OST3/6 confers similar substrate specicity since in the A. thaliana ost3/6 (aru) mutant only a subset of glycoproteins is misglycosylated and therefore non-functional24. Among these are the pathogen-associated molecular pattern (PAMP) receptor kinase EF-TU RECEPTOR (EFR), and KORRIGAN1, an endo-b-1,4-glucanase involved in cellulose biosynthesis. In line

with this, aru has previously been identied in an EMS-screen for cell wall mutants34. Some of the known members in the PT reception pathway, FER and NTA, have been implicated in the perception of cell wall perturbations, pathogen resistance, and innate immunity17,44,45. After pollination with A. lyrata, fer heterozygous mutants show higher OG/S (74.9%) than wild-type segregants (61.2%, Po0.01, Supplementary Fig. 9), suggesting that inter- and intraspecic PT reception both involve the FER pathway. Moreover, FER (a receptor-like kinase with an extracellular malectin-binding domain) and LRE (a glycosylphosphatidylinositol-anchored protein) are likely to be glycosylated13,16,46 and could be substrates of ARU. To test this, we analysed the expression and localization of uorescent FER and LRE fusion proteins in aru ovules. We included NTA, which itself does not contain any putative glycosylation sites but whose subcellular localization depends on FER signalling17. All fusion proteins displayed a wild-type-like subcellular localization in the synergids of aru embryo sacs: FER-GFP and LRE-Citrine were observed at the micropylar end of the synergids, and NTA-GFP was re-localized there upon PT arrival (Supplementary Fig. 10). These results indicate that, in the absence of functional ARU, these proteins are properly targeted to their subcellular compartment. However, we cannot rule out that the extracellular domain of FER is un- or misglycosylated at specic glycosylation sites in aru, which may allow the protein to recognize intra- but not interspecic PTs.

The synergid has a specialized secretory region at its micropylar end, the liform apparatus, which contains a large amount of secreted material and is believed to be the site of PT recognition6. It is possible that even subtle differences in ARU protein levels could lead to the misglycosylation of target proteins, including FER, such that a few specic glycosylation sites remain unglycosylated. Given the high secretory activity of the liform apparatus, such small changes could have a large effect on PT reception. Further work will be necessary to shed more light on the target proteins of ARU and to elucidate the role of specic glycosylated surface proteins in PT reception.

DiscussionA possible interpretation of our results is that FER, and/or yet unknown synergid (co-)receptors, bind putative ligands from intraspecic PTs both via specic interactions with carbohydrates on the receptor protein and via direct proteinprotein interactions, a mechanism similar to the proposed domain-specic model in mammalian sperm-egg binding47 (Fig. 6a). Ligands from interspecic PTs might not be able to sufciently interact via proteinprotein contacts alone but could still be recognized to some extent via the carbohydrate moieties, explaining the partial A. lyrata PT reception success in Col-0 ovules (Fig. 6b). It is conceivable that in aru, and potentially also in A. thaliana accessions with a similar phenotype, changes in the glycosylation status of the receptor could completely abolish the ability to recognize and receive interspecic PTs (Fig. 6d), while ligands from A. thaliana PTs are still efciently recognized via proteinprotein interactions, leading to normal PT reception (Fig. 6c).

The crosstalk between gametophytes constitutes a specic form of cellcell communication. Cellular interactions are often mediated by specic binding of an extracellular ligand to a receptor, triggering downstream signalling cascades in the recipient cell. Most extracellular ligands and receptors are heavily glycosylated48, which inuences their binding specicities and conformation, such that already the absence of a single glycosylation motif can reduce or abolish a receptors function and ligand-binding afnity49,50. Our results suggest that both proteinprotein interactions and

NATURE COMMUNICATIONS | 7:10826 | DOI: 10.1038/ncomms10826 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 7

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10826

a Col-0 synergid cell

ARU

Glycosylation

b Col-0 synergid cell

ARU

Glycosylation

Glycosylated receptor

Putative PT ligand

N

ER

A. lyrata PT

PT reception

PT overgrowth

c

aru synergid cell

No or misglycosylation

N

ER

Non- or misglycosylated receptor

Putative PT ligand

A. thaliana PT

PT reception

d

aru synergid cell

No or mis

glycosylation

N

ER

Non- or misglycosylated receptor

Putative PT ligand

A. lyrata PT

PT overgrowth

N

ER

Glycosylated receptor

Putative PT ligand

A. thaliana PT

PT reception

Figure 6 | Potential mechanisms of interspecic PT reception and the role of glycosylation. (a) In wild-type Col-0 synergid cells, a receptor is glycosylated by ARU and can bind putative signals from the A. thaliana PT with its carbohydrate moieties and via proteinprotein interactions. (b) A. lyrata PT signals might not be properly recognized by proteinprotein contacts but partially by carbohydrate interactions, thus leading to PT overgrowth in some ovules and normal PT reception in others. (c) In aru mutants, the synergid receptor is un- or misglycosylated; nevertheless, A. thaliana PTs can be efciently received through proteinprotein interactions of the receptor with the PT signals. (d) In aru mutants pollinated with A. lyrata pollen, PT reception is neither possible by carbohydrate, nor by proteinprotein interactions. Therefore, almost all ovules show PT overgrowth.

recognition mediated by carbohydrates may be crucial factors to ensure species-specic PT reception. Thus, divergent evolution of receptorligand pairs, as well as of the factors controlling their glycosylation status, could establish new species barriers. Deciphering the molecular basis of speciation in plants might enable us to overcome existing hybridization barriers, which could eventually be of great agronomic importance.

Methods

Plant material and growth conditions. The A. thaliana accessions were part of the Nordborg collection for GWAS27,40. Amplied seed stocks were kindly donated by Ortrun Mittelsten Scheid (Gregor Mendel Institute, Vienna). After stratication (2 days at 4 C), the seeds were allowed to germinate for 6 days on MS plates (22 C, 16 h light, MS from Carolina Biological Supply). Because some accessions require vernalization, all the seedlings were kept in a vernalization chamber for5 weeks (4 C, 16 h light) on MS plates before they were transferred to the soil (ED73, Universalerde).

The accessions were grouped into early- (four incomplete blocks A, B, C andD), mid-, and late owering plants (three complete blocks A, B, C each) according to the owering time27, and were grown in a greenhouse chamber (22 C, 16 h light) in an incomplete randomized block design. See Supplementary Table 1for the assignment of accessions to the blocks.A. lyrata13, A. halleri (a gift from Marcus Koch, University of Heidelberg), and A. arenosa (donated by Matthias Helling, University of Zurich) plants were stratied for 10 days and grown in the same greenhouse chamber. Plants were vernalized to induce owering (see above).

SALK-lines were obtained from NASC: SALK_067271 (At1g61790, aru-1), SALK_137883C (At1g61780), SALK_052207C and SALK_026074C (At1g61795), SALK_104077 (At1g61810). The EMS allele aru-2 (ref. 34) was a gift from Peter McCourt (University of Toronto). The plants were grown as described before17.

Crosses and aniline blue staining. Flowers were emasculated and the pistils were pollinated 2 days after emasculation (dae). Siliques were collected two days after pollination and xed for aniline blue staining in 9:1 ethanol:acidic acid. Aniline blue staining was performed as described previously17, and the samples were analysed with a Leica DM6000B microscope (Leica Microsystems). For GWAS phenotyping, 920 siliques of a minimum of three individuals were analysed for each accession (Exceptions: Zdr-1: seven siliques, Got-7: four siliques from two individuals).

GWAS analyses. Association mapping was conducted using the mean values of the proportions of ovules with PT overgrowth per silique (OG/S) as phenotypes. An A. thaliana 250 K Affymetrix SNP genotyping data set27 was downloaded from https://cynin.gmi.oeaw.ac.at/home/resources/atpolydb

Web End =https://cynin.gmi.oeaw.ac.at/home/resources/atpolydb . GWAS analyses were performed using a compressed mixed linear model, using population parameters previously determined51,52, and a kinship matrix to account for family relatedness, in the R package GAPIT53. The mixed linear models were run with and without

principal components as xed effects to correct for population structure. Multiple testing was controlled using the Bonferroni correction and false-discovery rate54. GWAS analyses were also run using a general linear model in the web-based interface TASSEL3.0 (ref. 30) and an accelerated mixed model with Box-Cox transformed phenotypes in GWAPP55. MLMM analysis was conducted as previously described31.

Constructs for stable plant transformation. For pMYB98/pFER::ARU-GFP the complete coding sequence of ARU without the stop-codon was amplied using gene-specic primers with attB-sites for Gateway cloning: 50-GGGGACAAGTT

TGTACAAAAAAGCAGGCTTCATGGCGCTCAAATCAAAACTCGTC-30 and 50-GGGGACCACTTTGTACAAGAAAGCTGGGTCACGCCAACTCGATGGC CAATACGGA-30. We introduced the PCR-fragment into pDONR207, and subsequently into the destination vector (using the E. coli strain DH5-alpha FIq from New England Biolabs). The destination vector was a modication of the plant Gateway vector pMDC83, which contains the 2x35S-promotor before and GFP after the Gateway cassette56. For our purpose, we exchanged the original 2x35S-promoter with the promoters of FER13 and MYB98 (ref. 35) to express ARU specically in synergid cells. The MYB98 promoter was amplied from Col-0 with primers 50-TTTAAGCTTATACACTCATTGTCCTTCG-30 and 50-CCCTCTAG

ATGTTTTGGAAAGGAGAAAAAA-30, introducing a HindIII and XbaI restriction site, respectively. The FER promoter was amplied from the pFER:: FER-GFP construct13 using specic primers 50-TTTGGTAAGCTTCGATT

TAAGCGAG-30 and 50-TTTTCTAGACGATCAAGAGCACTTCTCCGGG-30, which introduce HindIII and XbaI restriction sites as well. The 2x35S-promoter was cut out of pMDC83 (ref. 56) with HindIII/XbaI (New England Biolabs), and the PCR fragments were introduced into the dephosphorylated vector backbone by ligation. pDONR207 carrying the ARU coding sequence and the modied destination vector were combined in an LR reaction. The resulting vectors, pFER::ARU-GFP and pMYB98::ARU-GFP were transformed into Agrobacterium tumefaciens strain CV1310, and homozygous aru-1 plants were transformed by the oral dip method57. The complementation assays were conducted in the T2 and T3 generations with plants homozygous for the complementation construct and the aru-1 mutation. For experiments with pMYB98::ARU-GFP in accessions with high OG/S, the construct was transformed into Fei-0, Kz-1, and Kz-9, and OG/S was assessed in heterozygous plants (T1 generation).

For the constructs with ARU under the control of its endogenous promoter, the ARU fragment (including 1,492 bp of upstream, ARU CDS, and 865 bp of downstream sequence) was amplied from Col-0 and Ga-0 genomic DNA using primers 50-TTTTACTAGTAGGCAATTCCATCAGTTGTT-30 and 50-TTTTGGT

ACCGTTACTTCACTTTCTCGAGT-30, introducing a SpeI and a KpnI restriction site, respectively. The fragments were cloned into pMDC99 (ref. 56) using restriction-ligation and transformed into aru mutants (Col-0 background). pARU::ARU-GFP was cloned by amplication of a part of ARU coupled to GFP-tNOS from pFER::ARU-GFP with primers 50-GCGTTAACGCTTTACCT

CA-30 (including the natural HpaI site in ARU) and 50-TTTGGATCCAGTAAC ATAGATGACACCGCG-30 (introducing a BamHI site after tNOS). This fragment was introduced by ligation into the pMDC99 vector carrying the genomic fragment of ARU (Col-0). By cutting this vector with HpaI and BamHI, part of the ARU

8 NATURE COMMUNICATIONS | 7:10826 | DOI: 10.1038/ncomms10826 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10826 ARTICLE

coding sequence and the downstream sequence were removed and replaced with the respective fragment of ARU coupled to GFP-tNOS, resulting inpARU(1,492 bp)::ARU-GFP, which was transformed into Col-0. pLRE::LRE-Citrine was cloned with overlapping PCR fragments that were assembled using the Gibson cloning Master Mix from New England Biolabs according to the manufacturers recommendations. The 779-bp long promoter sequence with the predicted signal peptide from LRE16 was amplied with primers 50-GTGCTGCAAGGCGATTAAGTCCGTGTGCTCTGTCTGCATT-30 and 50-CACAGCTCCACCTCCACCTCCAGGCCGGCCTATGGAACTTGAAGAG GAGAGAGA-30, introducing an overhang complementary to the vector pMDC99 (ref. 56). Citrine was amplied from the transgenic line CS36962 (ordered from Arabidopsis Biological Resource Center, ABRC), using gene-specic primers with overhang primers for the signal peptide of LRE and overhang primers for the GPI-anchor of LRE: 50-GGCCGGCCTGGAGGTGGAGGTGGAGCTGTGAGCA

AGGGCGAGGAGCT-30 and 50-GGCCCCAGCGGCCGCAGCAGCACCAGCAG GATCCTTGTACAGCTCGTCCA-30. The GPI-anchor of LRE was amplied with overhang primers for pMDC99: 50-TGCTGGTGCTGCTGCGGCCGCTGGGGC CTCGGGTATGTCTTTTTGTTGTC-30 and 50-AGCTCCACCGCGGTGGCGG CCGCTCTAGAAGTCTCGCTTCTTCTTTTGT-30. pMDC99 was amplied with overhang primers for the LRE promoter and the GPI-anchor using primers: 50-ACTTAATCGCCTTGCAGCAC-30 and 50-TCTAGAGCGGCCGCCACC

GCGG-30. All the constructs were veried by sequencing. pFER::FER-GFP and pNTA::NTA-GFP were described previously13,17.

ARU-GFP subcellular localization. We used the pFER::ARU-GFP construct for microprojectile bombardment of onion epidermal cells and co-localized it with the ER-marker pER-rk (mCherry) obtained from ABRC58. Biolistic bombardment of onion epidermis was performed as described17.

For visualizing GFP expression in the synergids, owers were emasculated and pistils were dissected 2 dae to ensure the development of mature, unfertilized embryo sacs. The tissue was mounted on slides in 1 M glycine, pH 9.6. Images were captured on Leica Confocal Microscopes SP2 and SP5 (Leica Microsystems).

RNA extraction and reverse transcriptase PCR. RNA from pistils (25 pistils, 2 dae), inorescences, and ovules (extracted from 30 pistils, 2 dae) was extracted using the Trizol reagent (Invitrogen) according to the manufacturers recommendations. Pistil and inorescence cDNA was reverse transcribedusing Oligo-dT primers and Superscript II reverse transcriptase from Invitrogen. Ovule cDNA was amplied using the Ovation Pico SL WTA system V2 from Nugen.

Reverse transcriptase PCR (RTPCR) of ARU was done using primers 50-CAATGTGCTTGTTCGAGTG-30 and 50-ATCCAGTCTTCCAGTTATCCA-30.

For quantitative RT and digital droplet PCR of ARU in A. thaliana accessions, the primers 50-GTTTGTTACCAATGTGCTTGTTCG-30 and 50-TCCATATCC

AGTCTTCCAGTTATCC-30 were used and expression levels were normalized against UBIQUITIN C (UBC9, primers: 50-ATGCTTGGAGTCCTGCTTGG-30 and 50-TGCCATTGAATTGAACCCTCTC-30). For digital droplet PCR on ovule cDNA, the UBC9 assay was performed as an EvaGreen assay, whereas ARU transcripts were detected using a gene-specic probe (50-FAM-TACTGCAC

AAAGGTTG-MGB-30). The samples were analysed with the QX200 system from Bio-Rad.

Population genetic analyses and statistical tests. Determination of ARU gene structure and UTRs is based on annotations in the ARAMEMNON database37. Sequences of ARU and 1,000 bp up- and downstream anking regions were downloaded from http://signal.salk.edu/atg1001/3.0/gebrowser.php

Web End =http://signal.salk.edu/atg1001/3.0/gebrowser.php . For accessions for which no sequences or only sequences with missing data were available, we amplied the whole region from genomic DNA using primers 50-TTTGCTATAG

GCACATGTGT-30 and 50-GACCCGAAATTGTCAAATGA-30, and sequenced the resulting PCR products of Bay-0, Fab-2, Fab-4, Omo-2-3, Knox-10, Kz-1, LL-0, Lz-0, Mr-0, Mrk-0, Zdr-6. The upstream region was sequenced additionally from Got-7, Pu2-23, and Spr1-6 (primer 50-TTTGCTATAGGCACATGTGT-30 and 50-CGGAGGTTAGGAATTTTGAGA-30), and the downstream region from Got-7, Pu2-23, Kz-9, Mz-0, Pro-0, Van-0, and Var2-1 (primer 50-CAATGTGCTTGTTC GAGTG-30 and 50-GACCCGAAATTGTCAAATGA-30). Tajimas D and Fay and Wus H were calculated separately for the 1,000 bp up- and downstream as well as the coding sequence with the set of 96 accessions40 using DnaSP 5.10 (ref. 59). Several accessions that had big indels in the up- and downstream regions were left out from the analysis (Mr-0, Got-7, Pu2-23, and Spr1-6 for the upstream, Var2-1, Nok-3, and Got-7 for the downstream region). A. lyrata was used as outgroup.

P values against the null model were obtained by running 10,000 coalescent simulations and for Tajimas D, the 5% quantile was calculated using previously published estimates for D40.

References

1. Lowry, D. B., Modliszewski, J. L., Wright, K. M., Wu, C. A. & Willis, J. H. The strength and genetic basis of reproductive isolating barriers in owering plants. Philos. Trans. R. Soc. Lond. B Biol. Sci. 363, 30093021 (2008).

2. Rieseberg, L. H. & Blackman, B. K. Speciation genes in plants. Ann. Bot. 106, 439455 (2010).

3. Widmer, A., Lexer, C. & Cozzolino, S. Evolution of reproductive isolation in plants. Heredity 102, 3138 (2009).

4. Okuda, S. et al. Defensin-like polypeptide LUREs are pollen tube attractants secreted from synergid cells. Nature 458, 357361 (2009).

5. Huang, B. Q. & Russel, S. D. Female germ unit: organization, isolation, and function. Int. Rev. Cytol. 140, 233293 (1992).

6. Kessler, S. A. & Grossniklaus, U. Shes the boss: signaling in pollen tube reception. Curr. Opin. Plant. Biol. 14, 622627 (2011).

7. Zinkl, G. M., Zwiebel, B. I., Grier, D. G. & Preuss, D. Pollen-stigma adhesion in Arabidopsis: a species-specic interaction mediated by lipophilic molecules in the pollen exine. Development 126, 54315440 (1999).

8. Marton, M. L., Fastner, A., Uebler, S. & Dresselhaus, T. Overcoming hybridization barriers by the secretion of the maize pollen tube attractant ZmEA1 from Arabidopsis ovules. Curr. Biol. 22, 11941198 (2012).

9. Lausser, A., Kliwer, I., Srilunchang, K. O. & Dresselhaus, T. Sporophytic control of pollen tube growth and guidance in maize. J. Exp. Bot. 61, 673682 (2010).

10. Kanaoka, M. M. et al. Identication and characterization of TcCRP1, a pollen tube attractant from Torenia concolor. Ann. Bot. 108, 739747 (2011).11. Higashiyama, T. et al. Species preferentiality of the pollen tube attractant derived from the synergid cell of Torenia fournieri. Plant Physiol. 142, 481491 (2006).

12. Williams, E. G., Kaul, V., Rouse, J. L. & Palser, B. F. Overgrowth of pollen tubes in embryo sacs of Rhododendron following interspecic pollinations. Aust. J. Bot. 34, 413423 (1986).

13. Escobar-Restrepo, J. M. et al. The FERONIA receptor-like kinase mediates male-female interactions during pollen tube reception. Science 317, 656660 (2007).

14. Huck, N., Moore, J. M., Federer, M. & Grossniklaus, U. The Arabidopsis mutant feronia disrupts the female gametophytic control of pollen tube reception. Development 130, 21492159 (2003).

15. Rotman, N. et al. Female control of male gamete delivery during fertilization in Arabidopsis thaliana. Curr. Biol. 13, 432436 (2003).

16. Capron, A. et al. Maternal control of male-gamete delivery in Arabidopsis involves a putative GPI-anchored protein encoded by the LORELEI gene. Plant Cell 20, 30383049 (2008).

17. Kessler, S. A. et al. Conserved molecular components for pollen tube reception and fungal invasion. Science 330, 968971 (2010).

18. Lindner, H. et al. TURAN and EVAN mediate pollen tube reception in the synergid cells but play distinct roles in the male gametophyte. PLoS Biol. 13, e1002139 (2015).

19. Lindner, H. et al. SNP-ratio mapping (SRM): identifying lethal alleles and mutations in complex genetic backgrounds by next-generation sequencing. Genetics 191, 13811386 (2012).

20. Amien, S. et al. Defensin-like ZmES4 mediates pollen tube burst inmaize via opening of the potassium channel KZM1. PLoS Biol. 8, e1000388 (2010).

21. Qu, L.-J., Li, L., Lan, Z. & Dresselhaus, T. Peptide signalling during the pollen tube journey and double fertilization. J. Exp. Bot. 66, 51395150 (2015).

22. Schulz, B. L. & Aebi, M. Analysis of glycosylation site occupancy reveals a role for Ost3p and Ost6p in site-specic N-glycosylation efciency. Mol. Cell. Proteomics 8, 357364 (2009).

23. Schwarz, M., Knauer, R. & Lehle, L. Yeast oligosaccharyltransferase consists of two functionally distinct sub-complexes, specied by either the Ost3p or Ost6p subunit. FEBS Lett. 579, 65646568 (2005).

24. Farid, A. et al. Specialized roles of the conserved subunit OST3/6 of the oligosaccharyltransferase complex in innate immunity and tolerance to abiotic stresses. Plant Physiol. 162, 2438 (2013).

25. Lewis, D. & Crowe, L. K. Unilateral interspecic incompatibility in owering plants. Heredity 12, 233256 (1958).

26. Schmickl, R. & Koch, M. A. Arabidopsis hybrid speciation processes. Proc. Natl Acad. Sci. USA 108, 1419214197 (2011).

27. Atwell, S. et al. Genome-wide association study of 107 phenotypes in Arabidopsis thaliana inbred lines. Nature 465, 627631 (2010).

28. Verslues, P. E., Lasky, J. R., Juenger, T. E., Liu, T. W. & Kumar, M. N. Genome wide association mapping combined with reverse genetics identies new effectors of low water potential-induced proline accumulation in Arabidopsis thaliana. Plant Physiol. 164, 144159 (2013).

29. Meijon, M., Satbhai, S. B., Tsuchimatsu, T. & Busch, W. Genome-wide association study using cellular traits identies a new regulator of root development in Arabidopsis. Nat. Genet. 46, 7781 (2013).

30. Bradbury, P. J. et al. TASSEL: software for association mapping of complex traits in diverse samples. Bioinformatics 23, 26332635 (2007).

31. Segura, V. et al. An efcient multi-locus mixed-model approach for genome-wide association studies in structured populations. Nat. Genet. 44, 825830 (2012).

NATURE COMMUNICATIONS | 7:10826 | DOI: 10.1038/ncomms10826 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 9

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10826

32. Wuest, S. E. et al. Arabidopsis female gametophyte gene expression map reveals similarities between plant and animal gametes. Curr. Biol. 20, 506512 (2010).

33. Gerhard, E.ber die Gottheiten der Etrusker (Druckerei der Kniglichen Akademie der Wissenschaften, 1847).

34. Austin, R. S. et al. Next-generation mapping of Arabidopsis genes. Plant J. 67, 715725 (2011).

35. Kasahara, R. D., Portereiko, M. F., Sandaklie-Nikolova, L., Rabiger, D. S. & Drews, G. N. MYB98 is required for pollen tube guidance and synergid cell differentiation in Arabidopsis. Plant Cell 17, 29812992 (2005).

36. Hindson, B. J. et al. High-throughput droplet digital PCR system for absolute quantitation of DNA copy number. Anal. Chem. 83, 86048610 (2011).

37. Schwacke, R. et al. ARAMEMNON, a novel database for Arabidopsis integral membrane proteins. Plant Physiol. 131, 1626 (2003).

38. Swanson, W. J. & Vacquier, V. D. The rapid evolution of reproductive proteins. Nat. Rev. Genet. 3, 137144 (2002).

39. Tajima, F. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 123, 585595 (1989).

40. Nordborg, M. et al. The pattern of polymorphism in Arabidopsis thaliana. PLoS Biol. 3, e196 (2005).

41. Fay, J. C. & Wu, C. I. Hitchhiking under positive Darwinian selection. Genetics 155, 14051413 (2000).

42. Schmid, K. J., Ramos-Onsins, S., Ringys-Beckstein, H., Weisshaar, B. & Mitchell-Olds, T. A multilocus sequence survey in Arabidopsis thaliana reveals a genome-wide departure from a neutral model of DNA sequence polymorphism. Genetics 169, 16011615 (2005).

43. Aebi, M. N-linked protein glycosylation in the ER. Biochim. Biophys. Acta 1833, 24302437 (2013).

44. Keinath, N. F. et al. PAMP (pathogen-associated molecular pattern)-induced changes in plasma membrane compartmentalization reveal novel components of plant immunity. J. Biol. Chem. 285, 3914039149 (2010).

45. Cheung, A. Y. & Wu, H. M. THESEUS1, FERONIA and relatives: a family of cell wall-sensing receptor kinases? Curr. Opin. Plant Biol. 14, 632641 (2011).

46. Lindner, H., Mller, L. M., Boisson-Dernier, A. & Grossniklaus, U. CrRLK1L receptor-like kinases: not just another brick in the wall. Curr. Opin. Plant Biol. 15, 659669 (2012).

47. Clark, G. F. Molecular models for mouse sperm-oocyte binding. Glycobiology 21, 35 (2011).

48. Zielinska, D. F., Gnad, F., Schropp, K., Wisniewski, J. R. & Mann, M. Mapping N-glycosylation sites across seven evolutionarily distant species reveals a divergent substrate proteome despite a common core machinery. Mol. Cell 46, 542548 (2012).

49. van der Hoorn, R. A. et al. Structure-function analysis of Cf-9, a receptor-like protein with extracytoplasmic leucine-rich repeats. Plant Cell 17, 10001015 (2005).

50. Haweker, H. et al. Pattern recognition receptors require N-glycosylation to mediate plant immunity. J. Biol. Chem. 285, 46294636 (2010).

51. Zhang, Z. et al. Mixed linear model approach adapted for genome-wide association studies. Nat. Genet. 42, 355360 (2010).

52. Kang, H. M. et al. Variance component model to account for sample structure in genome-wide association studies. Nat. Genet. 42, 348354 (2010).

53. Lipka, A. E. et al. GAPIT: genome association and prediction integrated tool. Bioinformatics 28, 23972399 (2012).

54. Benjamini, Y. & Hochberg, Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. R. Statist. Soc. 57, 289300 (1995).

55. Seren, U. et al. GWAPP: a web application for genome-wide association mapping in Arabidopsis. Plant Cell 24, 47934805 (2012).

56. Curtis, M. D. & Grossniklaus, U. A Gateway cloning vector set for high-throughput functional analysis of genes in planta. Plant Physiol. 133, 462469 (2003).

57. Clough, S. J. & Bent, A. F. Floral dip: a simplied method for Agrobacterium-mediated transformation of Arabidopsis thaliana. Plant J. 16, 735743 (1998).

58. Nelson, B. K., Cai, X. & Nebenfuhr, A. A multicolored set of in vivo organelle markers for co-localization studies in Arabidopsis and other plants. Plant J. 51, 11261136 (2007).

59. Rozas, J. & Rozas, R. DnaSP, DNA sequence polymorphism: an interactive program for estimating population genetics parameters from DNA sequence data. Comput. Appl. Biosci. 11, 621625 (1995).

Acknowledgements

We thank Sharon A. Kessler and Aurlien Boisson-Dernier for critically reading the manuscript and helpful discussions; and Karl J. Schmid and Dounia Saleh for their help and insight with population genetic analyses; Ortrun Mittelsten Scheid, Marcus Koch, Peter McCourt, and Matthias Helling for providing seeds; Daniela Guthrl, Arturo Bolanos, Christof Eichenberger, and Peter Kopf for general lab support; and Christian Frei and Karl Huwiler for greenhouse maintenance. This work was supported by the University of Zrich, partial support for L.M.M. and H.L. through Research Modules of the SNF ProDoc Programs Plant Science and Policy and Molecular Life Science, respectively, to U.G., and grants from the Swiss National Science Foundation and the European Research Council to U.G.

Author contributions

U.G. initiated and supervised the project. L.M.M. and U.G. conceived the experiments and wrote the manuscript; H.L. and N.D.P. critically read and commented on the manuscript; L.M.M., H.L., N.D.P., and V.G. performed the experiments. All the authors were involved in data analysis and interpretation.

Additional information

Supplementary Information accompanies this paper at http://www.nature.com/naturecommunications

Web End =http://www.nature.com/ http://www.nature.com/naturecommunications

Web End =naturecommunications

Competing nancial interests: The authors declare no competing nancial interest.

Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/

Web End =http://npg.nature.com/ http://npg.nature.com/reprintsandpermissions/

Web End =reprintsandpermissions/

How to cite this article: Mller, L. M. et al. A subunit of the oligosaccharyltransferase complex is required for interspecic gametophyte recognition in Arabidopsis.

Nat. Commun. 7:10826 doi: 10.1038/ncomms10826 (2016).

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Web End =http://creativecommons.org/licenses/by/4.0/

10 NATURE COMMUNICATIONS | 7:10826 | DOI: 10.1038/ncomms10826 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications