# Source paper archive

- Source title: LORELEI sequence variation modulates lineage-dependent pollen tube reception in Brassicaceae
- Author(s): Jong, E., Noble, J. A., Alcaraz, M. J., Hafidh, S., Varghese, I., Honys, D., Lisi, G. P., Beilstein, M. A., Palanivelu, R.
- Source landing URL: https://www.biorxiv.org/content/10.64898/2026.11.18.767685v1
- PDF URL: https://www.biorxiv.org/content/10.64898/2026.11.18.767685v1.full.pdf
- PDF SHA-256: 69b9242e401cd462ce0eac3a3fc3585081bf1be18ab36c68be6aa1769ececc00
- Converter: pymupdf4llm 0.0.27
- Limits: Automated PDF text extraction; reading order, equations, tables and formatting may be incomplete or incorrect. Images are omitted, and no OCR was performed. Markup and links are neutralized for safe downloading. Consult the original PDF for the authoritative version.

---

&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

# 1 LORELEI sequence variation modulates lineage

2
# dependent pollen tube reception in Brassicaceae

3

4
#### 5 Emma Jong &#91;1&#93;, Jennifer A. Noble &#91;1 &#93;, Matthew J. Alcaraz &#91;1&#93;, Iz Varghese &#91;3&#93;, Said 6 Hafidh &#91;2&#93;, David Honys &#91;2&#93;, George P. Lisi &#91;3&#93;, Mark A. Beilstein &#91;1*&#93;, Ravishankar 7 Palanivelu &#91;1*&#93;


8


9 1 The School of Plant Sciences, University of Arizona, Tucson, AZ 85721, USA


10


11 2 Institute of Experimental Botany of the Czech Academy of Sciences, Rozvojová 263, 165


12 02 Prague 6, Czech Republic


13


14 3 Department of Molecular Biology, Cell Biology, and Biochemistry, Brown University,


15 Providence, RI 02903, USA


16


17 *Authors for correspondence: Mark A. Beilstein (mbeilstein@arizona.edu) and


18 Ravishankar Palanivelu (rpalaniv@arizona.edu)


19


20 Short title: LORELEI and interspecific compatibility


21


22 Keywords: pollen tube reception; prezygotic reproductive barrier; LORELEI (LRE);


23 Brassicaceae; reproductive isolation


24


25


26


1


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

### 27 Abstract


28
29 Pollen tube reception (PTR) is essential for fertilization and can act as a prezygotic barrier


30 in interspecific crosses. While core female components such as LORELEI (LRE) and


31 FERONIA (FER) mediate conspecific PTR, their roles in heterospecific compatibility


32 remain unclear. Here, we investigated how LRE sequence variation influences PTR


33 outcomes with heterospecific pollen in Brassicaceae. In _Arabidopsis thaliana_, PTR


34 frequently failed when ovules were pollinated with _Arabidopsis lyrata_ or _Sisymbrium irio_,


35 and loss of LRE or FER further increased pollen tube overgrowth in a pollen-species

36 dependent manner. Complementation of _lre_ ovules with LRE orthologs showed distinct,


37 donor-dependent effects: _AlyLRE_ and _SiLRE_ produced modest and variable


38 improvements with _A. lyrata_ pollen, whereas SiLRE consistently and markedly improved


39 reception of _S. irio_ pollen. Evolutionary analyses indicated relaxed purifying selection in


40 the LRE lineage relative to its paralog _LLG1_ and nominated residues 47 and 120 as


41 candidates for functional divergence. Consistent with this, the Lineage II-associated


42 R120T substitution enhanced reception of _S. irio_ pollen while partially reducing


43 conspecific PTR. In addition, the ancestral-state E47P–R120G combination also


44 improved heterospecific PTR without measurably compromising conspecific function.


45 Together, these results indicate that LRE sequence diversification can modulate female

46 side compatibility with pollen tubes from different evolutionary lineages, potentially by


47 altering how pollen tube-derived signals are sensed and transduced during reception.


48

49


2


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

### 50 Introduction


51
52 Reproductive isolation results from both prezygotic and postzygotic barriers that


53 prevent successful hybridization between species. Prezygotic barriers act before


54 fertilization, either via disrupting pollen-pistil interactions or by preventing zygote


55 formation, while postzygotic barriers act after fertilization by affecting hybrid seed viability


56 (Baack et al. 2015). In angiosperms, prezygotic barriers can occur before pollination,


57 because of traits such as flower color and pollinator preference, or after pollination, during


58 processes such as pollen hydration, germination, pollen tube growth, guidance, and


59 reception in the pistil. Understanding the molecular mechanisms underlying these post

60 pollination barriers is important for explaining how species-specific reproductive


61 compatibility is achieved and for informing breeding programs designed to generate novel


62 hybrids.


63 When fertilization occurs, the pollen grain germinates on the stigma and forms a


64 pollen tube, which then grows through the style and transmitting tract to reach an ovule.


65 Once the pollen tube enters the micropyle, it contacts a synergid cell, which halts its


66 growth and triggers sperm release in a process called pollen tube reception (PTR). Each


67 step in this pathway, from pollen germination to PTR, can serve as a prezygotic barrier to


68 fertilization. Studies of pollen-pistil interactions within a species (conspecific interactions)


69 have identified several molecular components necessary for effective fertilization, and


70 variation in these components can result in reproductive incompatibility. Consequently,


71 molecular mechanisms that facilitate compatible pollen-pistil interactions may also act as


72 barriers when pollen and pistils originate from different species (heterospecific


73 interactions).


74 In _Arabidopsis thaliana_, pollen hydration and germination during compatible


75 pollination are regulated by a receptor signaling module that includes the CrRLK1L


76 receptor kinases FERONIA (FER) and ANJEA (ANJ), the co-receptor LORELEI-LIKE


77 GPI-ANCHORED PROTEIN 1 (LLG1), and peptide ligands. Before pollination, stigma

78 derived RAPID ALKALINIZATION FACTOR 23 and 33 (RALF23 and RALF33) peptide


79 ligands maintain FER–ANJ-dependent production of reactive oxygen species (ROS) in


80 papilla cells. During compatible pollination, pollen coat-derived POLLEN COAT PROTEIN


3


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


81 B-class peptides (PCP-Bs) compete with RALF23 and RALF33 for binding to the receptor


82 complex, reducing stigmatic ROS and promoting pollen hydration and germination (Liu et


83 al. 2021). This ligand switch also promotes water availability to the pollen by activating


84 plasma-membrane H+-ATPases and opening PIP aquaporin channels (Liu et al. 2025).


85 FER-dependent ROS signaling is also recruited during self-incompatibility in _Brassica_


86 _rapa_, where self-pollination increases stigmatic ROS, whereas compatible pollination


87 reduces it. Disruption of FER–Rac/Rop signaling, RBOH activity, or ROS accumulation


88 compromises self-pollen rejection (Zhang et al. 2021). Subsequent work linked this


89 pathway to the classical self-incompatibility recognition system by showing that binding


90 of the pollen ligand SCR/SP11 to SRK promotes SRK association with FER, thereby


91 linking self-pollen recognition to FER-dependent ROS signaling (Huang et al. 2023).


92 Related signaling mechanisms add to the rejection of pollen from other species. In


93 _A. thaliana_, stigma-derived RALFs act through CrRLK1L receptors and LRX proteins to


94 maintain a barrier to pollen tube penetration, while RALFs from compatible pollen


95 compete with these stigma-derived ligands and weaken the barrier. Disrupting this


96 pathway, or applying synthetic pollen-derived RALFs, can allow otherwise incompatible


97 pollen tubes to penetrate the stigma and, in some cases, result in hybrid embryo formation


98 (Lan et al. 2023). Additional mechanisms reinforce this interspecific barrier: the stigma

99 localized transmembrane protein STIGMATIC PRIVACY 1 (SPRI1) restricts penetration


100 by pollen from distantly related Brassicaceae species independently of the self

101 incompatibility pathway (Fujii et al. 2019), and _SPRI2/SRS7_ and its paralog _SPRI2_ 

102 like/ _SRS5_ further strengthen this barrier by regulating SPRI1 and the cell-wall-associated


103 genes _TBL40_ and _TBL45_ (Fujii et al. 2023). Recently, pollen-derived SRK-interacting


104 interspecific pollen signals (SIPS) were shown to activate an interspecific stigma barrier


105 by engaging SRK, promoting its association with FER, and stimulating FER-dependent


106 ROS production; loss of these pollen signals weakens rejection of heterospecific pollen


107 (Cao et al. 2025). Thus, signaling systems that regulate compatible pollen acceptance


108 can be differentially engaged during self-incompatible and heterospecific interactions,


109 converting components of normal pollen-stigma communication into checkpoints for


110 reproductive compatibility (Zhang et al. 2026).


4


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


111 After germination, the pollen tube must traverse the stigma and grow through the


112 pistil's transmitting tissues to reach the ovary, providing additional opportunities to reject


113 pollen from other species. Researchers have identified several molecular components of


114 this stylar barrier, particularly in _Solanum_ . Pistil-expressed S-RNases and High Top-band


115 (HT) proteins contribute to heterospecific pollen rejection, while pollen-side factors such


116 as S-locus F-box proteins and CULLIN1 can confer resistance to these barriers.


117 Suppression of HT proteins can weaken or eliminate stylar rejection of heterospecific


118 pollen, showing the contribution of multiple female factors to this barrier (Tovar-Mendez


119 et al. 2017). Other extracellular components, including stigma/style cysteine-rich


120 adhesins and related lipid-transfer proteins, facilitate pollen tube adhesion and growth


121 through the transmitting tract and may likewise influence species-dependent compatibility


122 (Wang and Filatov 2023). Thus, both female stylar factors and pollen-expressed


123 components determine whether a heterospecific pollen tube can successfully navigate


124 the pistil, making pollen–style interactions an additional molecular checkpoint in post

125 pollination reproductive isolation.


126 An additional critical post-pollination checkpoint is pollen tube guidance to the


127 ovule. Peptides produced by the female gametophyte attract conspecific pollen tubes


128 from the transmitting tract toward the ovular micropyle, and some of these guidance


129 signals display species-specific preferences. For example, in maize, the synergid

130 expressed peptide EGG APPARATUS 1 (ZmEA1) attracts pollen tubes in a species

131 preferential manner; when ZmEA1 is expressed in _A. thaliana_ ovules, it can direct maize


132 pollen tubes toward the micropyle (Márton et al. 2012). Similarly, in Torenia and _A._


133 _thaliana_, synergid-secreted LURE peptides preferentially attract pollen tubes from the


134 same species (Higashiyama et al. 2006; Takeuchi and Higashiyama 2012). In _A. thaliana_,


135 the receptor kinase PRK6, located at the pollen tube tip, detects AtLURE1 and is required


136 for efficient LURE1-mediated guidance (Takeuchi and Higashiyama 2016).


137 The discovery of different types of attractants has further refined this concept. The


138 AtLURE1-PRK6 signaling pathway promotes guidance of conspecific pollen tubes and


139 accelerates the emergence of _A. thaliana_ pollen tubes from the septum, providing a


140 competitive advantage over heterospecific pollen tubes, such as those from _A. lyrata_


141 (Zhong et al. 2019). In contrast, the evolutionarily conserved XIUQIU peptides attract


5


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


142 pollen tubes from a wider range of Brassicaceae species and function independently of


143 PRK6 (Zhong et al. 2019). Thus, ovule guidance utilizes both species-preferential and


144 broadly acting signals, assuring efficient targeting of conspecific pollen while retaining


145 some attraction to heterospecific pollen. Consequently, pollen tube guidance represents


146 another molecular checkpoint that can restrict reproductive compatibility.


147 Even if heterospecific pollen tubes reach the ovule, they may face another


148 prezygotic barrier at the synergid cells during PTR, the process culminating in pollen tube


149 rupture and sperm release for double fertilization. Successful PTR involves a coordinated


150 sequence: after the pollen tube contacts the receptive synergid cell, its growth slows, the


151 synergid undergoes degeneration, and the pollen tube ultimately ruptures to release its


152 sperm cells (Sandaklie-Nikolova L, et al. 2007; Leydon et al. 2015; Zhong et al. 2025).


153 Before this transition, an autocrine signaling system maintains pollen tube integrity. Pollen


154 tube-derived RALF4 and RALF19 signal through the ANXUR1/2 (ANX1/2) and


155 BUDDHA'S PAPER SEAL1/2 (BUPS1/2) receptor complexes, along with the GPI

156 anchored co-receptors LLG2/3, to prevent premature rupture during pollen tube growth


157 (Ge et al. 2017; Ge et al. 2019). As the pollen tube approaches the female gametophyte,


158 ovule-derived RALF34 is proposed to compete with RALF4/19 for the ANX–BUPS


159 receptor system, shifting the pollen tube from an integrity-maintaining state toward rupture


160 (Ge et al. 2017; Feng et al. 2019). Thus, similar to the stigma, a signaling state maintained


161 by an autocrine ligand can be redirected by a signal from the reproductive partner.


162 In parallel, the arriving pollen tube activates a female signaling pathway in the


163 synergid cell. The CrRLK1L receptors FER, HERCULES RECEPTOR KINASE 1


164 (HERK1), and ANJEA (ANJ), along with the GPI-anchored proteins LORELEI (LRE) and


165 ENODLs (ENs), function at the filiform apparatus and are required for compatible PTR


166 (Escobar-Restrepo et al. 2007; Capron et al. 2008; Tsukamoto et al. 2010; Hou et al.


167 2016; Galindo-Trigo et al. 2020). Pollen tube-derived RALFs signal to this female


168 machinery. RALF6, RALF7, RALF16, RALF36, and RALF37, regulated by the pollen


169 tube-expressed transcription factors MYB97, MYB101, and MYB120, are essential for


170 normal PTR; their combined loss leads to pollen tube overgrowth. These peptides interact


171 with FER, ANJ, and HERK1, promoting receptor-complex assembly (Zhong et al. 2022).


172 RALF4 and RALF19 can also function as pollen tube-derived ligands for FER–LRE,


6


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


173 linking pollen tube arrival to downstream Ca²⁺ signaling (Gao et al. 2022). Pollen tube


174 arrival induces FER-dependent relocalization of NORTIA (NTA) to the filiform apparatus,


175 and FER, LRE, and NTA form a receptor–channel module through which pollen tube

176 derived RALFs stimulate Ca²⁺ responses in the synergid cell (Ju et al. 2021; Gao et al.


177 2022). Additionally, female-derived signals can act directly on the pollen tube to promote


178 its rupture. In maize, the synergid-secreted defensin-like peptide ZmES4 activates the


179 pollen tube K+ channel KZM1, causing membrane depolarization and ionic changes that


180 rapidly lead to pollen tube rupture (Amien et al. 2010). Collectively, these signaling events


181 couple perception of the arriving pollen tube to synergid responses and ultimately to


182 pollen tube rupture and sperm release.


183 Despite this coordinated signaling process, heterospecific pollination can still fail.


184 In interspecific crosses within _Rhododendron_, pollen tube overgrowth has been observed


185 (Williams et al. 1982), and a similar phenotype occurs when _A. thaliana_ ovules are


186 pollinated with _A. lyrata_ or _Cardamine flexuosa_ pollen (Escobar-Restrepo et al. 2007;


187 Müller et al. 2016). Several female-side factors have been implicated in this barrier. FER


188 was initially associated with heterospecific PTR when _A. lyrata_ and _C. flexuosa_ pollen


189 tubes exhibited a _fer_ -like overgrowth phenotype in wild-type _A. thaliana_ ovules (Escobar

190 Restrepo et al. 2007). Subsequent genetic evidence demonstrated that loss of FER


191 further compromises reception of _A. lyrata_ pollen tubes, supporting a direct role for FER

192 dependent signaling in heterospecific PTR (Ju et al. 2021). Notably, heterospecific pollen


193 tubes can trigger FER/LRE-dependent Ca²⁺ oscillations in synergid cells yet still fail to


194 complete PTR, indicating that activation of this signaling response alone is insufficient for


195 successful reception (Ponvert and Johnson 2024). Species-dependent compatibility can


196 also act at the final rupture step: ZmES4 rapidly induces rupture in maize pollen tubes,


197 acts more slowly on pollen tubes of the related grass _Tripsacum dactyloides_, and fails to


198 trigger rupture in more distantly related species, suggesting that compatibility between


199 female-derived rupture signals and their pollen tube targets contributes to prezygotic


200 reproductive isolation (Amien et al. 2010). ARTUMES (ARU), an OST3/6 subunit of the


201 oligosaccharyltransferase complex involved in protein N-glycosylation, is particularly


202 required for efficient reception of _A. lyrata_ pollen tubes by _A. thaliana_ ovules, with only a


203 minor effect on conspecific PTR (Müller et al. 2016). Thus, core components of the


7


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


204 conspecific reception machinery, female factors with more pronounced roles in


205 heterospecific interactions, and species-dependent compatibility between female-derived


206 rupture signals and their pollen tube targets can influence heterospecific PTR.


207 In spite of these advances, it remained unclear whether evolutionary variation in a


208 key female PTR component could influence its compatibility with pollen tubes from other


209 species. LRE provides an ideal system to investigate this question, as it is essential for


210 conspecific PTR and belongs to a family of signaling co-receptors that has undergone


211 evolutionary diversification. Our previous studies established LRE's role in synergid

212 mediated PTR and provided the genetic and molecular tools necessary to examine its


213 function (Tsukamoto et al. 2010; Liu et al. 2016; Noble et al. 2022).


214 Gene duplication, such as that resulting from whole-genome duplication, can allow


215 duplicated genes to evolve differences in regulation or function because selective


216 constraints change after duplication (Ramsey and Schemske 1998; Panchy et al. 2016).


217 Extensive evolutionary analysis of the _LLG_ family further supports focusing on LRE


218 (Noble et al. 2022). The _A. thaliana_ genome contains four _LLG_ family members, each


219 having distinct expression patterns: _LRE_ is primarily expressed in the female


220 gametophyte, _LLG2_ and _LLG3_ in the male gametophyte, and _LLG1_ mainly in sporophytic


221 tissues (Tsukamoto et al. 2010; Feng et al. 2019; Ge et al. 2019). Phylogenetic studies


222 indicate that _LRE_ and _LLG1_ are closely related paralogs within Brassicaceae, and


223 functional complementation experiments have shown a high degree of interchangeability


224 among LLG-family proteins when expressed in the same cellular context. Thus, these


225 genes are retained primarily because of expression differences rather than distinct protein


226 functions (Noble et al. 2022).


227 Our previous phylogenetic analysis also revealed an interesting feature: the branch


228 leading to _LRE_ exhibited greater sequence divergence than that leading to _LLG1_ (Noble


229 et al. 2022). This raised the possibility that, despite maintaining its conserved role in PTR,


230 _LRE_ may have acquired sequence changes with additional functional effects. However,


231 it remained unclear whether _LRE_ and _LLG1_ experienced different selective constraints


232 after duplication, which sequence changes were associated with this divergence, and


233 whether variation within _LRE_ influences its function during interactions with heterospecific


234 pollen tubes.


8


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


235 In this study, we examine these questions by investigating LRE function during


236 heterospecific PTR within an evolutionary framework. We assess whether LRE and its


237 co-receptor FER contribute to heterospecific PTR, whether LRE orthologs from different


238 Brassicaceae species differ in their ability to support heterospecific PTR, and whether


239 these functional differences are linked to evolutionary diversification within the LRE


240 lineage. Additionally, we examine the effects of variation at evolutionarily significant LRE


241 residues and determine whether variants with different PTR effects share predicted


242 conformational characteristics. Collectively, these approaches permit us to evaluate


243 whether diversification of a conserved female PTR component underlies species

244 dependent reproductive compatibility.


245

### 246 Results


247


248 **Loss of LRE and FER negatively impacts heterospecific pollen tube reception in a**


249 **pollen-species-dependent manner**


250 The PTR constitutes a prezygotic barrier in interspecific crosses between _A._


251 _thaliana_ and either _A. lyrata_ or _Cardamine flexuosa_ (Escobar-Restrepo et al. 2007; Müller


252 et al. 2016), both of which belong to Lineage I of the Brassicaceae (Beilstein et al. 2006).


253 Whether a similar PTR barrier occurs when _A. thaliana_ is pollinated with the more distantly


254 related Lineage II Brassicaceae species _Sisymbrium irio_ has not been examined. To


255 compare reproductive outcomes among these crosses, we pollinated _A. thaliana_ pistils


256 with _A. thaliana_, _A. lyrata_, or _S. irio_ pollen and assessed pollen tube behavior using aniline


257 blue staining (Fig. 1A-F).


258 We classified ovules into four categories based on pollen tube behavior. Types 1


259 and 2 represented defects occurring before pollen tube entry into the female gametophyte


260 (Fig. 1D and 1E). Pollen tubes that entered the micropyle but continued to grow and coil


261 within the female gametophyte were classified as Type 3, yielding clear morphological


262 evidence of failed PTR (Fig. 1F). Pollen tubes that entered the female gametophyte and


263 terminated near the expected position of the synergid cells without further growth or


9


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


264 coiling were classified as Type 4, consistent with a reception-like outcome (Fig. 1B and


265 1C; see Methods).


266 The distribution of these four pollen tube outcomes differed markedly among the


267 three crosses (Figure S1). In the conspecific _A. thaliana × A. thaliana_ cross, 75% of ovules


268 exhibited a Type 4 reception-like outcome, whereas only 1% exhibited Type 3 PTR failure.


269 Consistent with previous reports (Escobar-Restrepo et al. 2007; Müller et al. 2016), the


270 _A. thaliana × A. lyrata_ cross showed a pronounced PTR barrier: 66% of all ovules


271 exhibited PTR failure, whereas only 21% exhibited a Type 4 outcome. In the previously


272 uncharacterized _A. thaliana × S. irio_ cross, _S. irio_ pollen encountered considerable


273 barriers before entry into the female gametophyte, with 61% and 14% of ovules classified


274 as Types 1 and 2, respectively. Of all ovules, 17% exhibited PTR failure, and 8% exhibited


275 a Type 4 outcome. Thus, both _A. lyrata_ and _S. irio_ pollen encounter reproductive barriers


276 in _A. thaliana_, although the relative frequencies of defects occurring before ovule entry


277 and during PTR differ substantially between the two heterospecific crosses.


278 Having established PTR failure in both heterospecific crosses, we next asked


279 whether LRE and FER, core components of conspecific PTR, influence defect severity.


280 Following Müller et al. (2016), we quantified PTR failure as the percentage of pollen tube

281 entered ovules (PEO) exhibiting overgrowth (OG) &#91;%OG/PEO = Type 3/(Type 3 + Type


282 4) × 100&#93;. This metric specifically reports morphologically evident PTR failure among


283 ovules that received a pollen tube and does not include defects occurring before entry.


284 Because Type 3 classification required clear continued growth and coiling after entry,


285 %OG/PEO provides a stringent and conservative estimate of PTR failure among pollen


286 tube-entered ovules (see Methods).


287 As a conspecific control, we used _A. thaliana_ pollen to pollinate wild-type (Col-0),


288 _lre-7_ single mutant, _fer-4_ single mutant, _lre-7 fer-4_ double mutant, and _lre-7_ single mutant


289 pistils complemented with _AtLRE_ _pro_ _:AtLRE-cYFP_ (Liu et al. 2016; Figure S2). As


290 expected, wild-type (1.7%) and _AtLRE_ _pro_ _:AtLRE-cYFP_ complemented line showed


291 efficient PTR (2.1%), whereas the single mutants and the _lre-7 fer-4_ double mutant


292 exhibited elevated PTR failure (92.6%, 90.6%, and 90.4%, respectively; Fig. 1G),


293 consistent with the established roles of LRE and FER in conspecific PTR. For these


10


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


294 experiments, statistical comparisons were made to Col-0 wild type because the objective


295 was to determine how loss of endogenous LRE and FER function affected PTR.


296 When _A. thaliana_ pistils were pollinated with heterospecific _A. lyrata_ pollen,


297 approximately 73% of pollen tube-entered ovules exhibited PTR failure, consistent with


298 previous reports (Escobar-Restrepo et al. 2007; Müller et al. 2016; Fig. 1G). Loss of LRE


299 further increased this defect, with _lre-7_ pistils exhibiting 93.3% pollen tube overgrowth,


300 significantly higher than in wild-type pistils. _fer-4_ and _lre-7 fer-4_ pistils also showed


301 significantly elevated pollen tube overgrowth with _A. lyrata_ pollen, at 93.5% and 97%,


302 respectively. The AtLRE-complemented _lre-7_ line showed a lower frequency of pollen


303 tube overgrowth than _lre-7_ (82.5% OG/PEO), reversing the effects of the _lre-7_ mutation,


304 although the complemented line remained significantly different from wild type (Fig. 1G).


305 PTR failure was also frequent when _A. thaliana_ pistils were pollinated with _S. irio_


306 pollen. In this cross, pollen tube overgrowth was modestly higher in _lre-7_ (85%) and _fer-_


307 _4_ (87.1%) pistils than in wild type (73.7%), but these differences were not statistically


308 significant. The _lre-7 fer-4_ double mutant, however, exhibited significantly greater pollen


309 tube overgrowth than wild type, with 90.2% OG/PEO (Fig. 1G). The _lre-7_ complemented


310 with AtLRE-cYFP showed overgrowth levels similar to wild type (78%); however, because


311 _lre-7_ itself was not significantly different from wild type in this cross, we did not interpret


312 this as evidence of significant complementation. Together, these results show that LRE


313 and FER limit heterospecific PTR failure, although the magnitude of their contribution


314 depends on the pollen species.


315


316 **LRE ortholog identity influences heterospecific pollen tube reception**


317 We next asked whether replacing AtLRE in _A. thaliana_ with an LRE ortholog from


318 the pollen-donor species could improve heterospecific PTR. We generated


319 _AtLRE_ _pro_ _:AlyLRE-cYFP_ and _AtLRE_ _pro_ _:SiLRE-cYFP_ constructs (Figure S2) and introduced


320 them into the _A. thaliana lre-7_ background. For each construct, three independent


321 homozygous lines containing a single transgene copy were analyzed, as determined by


322 segregation analysis (Methods), and confirmed by ddPCR (Figure S3). In all transgenic


323 lines, cYFP fluorescence localized to the filiform apparatus of synergid cells, like full

11


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


324 length AtLRE-cYFP (Figure S4; Liu et al. 2016). Thus, both orthologs localized to the


325 expected site of LRE function during PTR.


326 Before testing these lines in heterospecific crosses, we determined whether


327 AlyLRE and SiLRE retained the LRE activity required for conspecific PTR. When


328 pollinated with _A. thaliana_ pollen, all AlyLRE and SiLRE lines complemented the _lre-7_


329 PTR defect and supported PTR at levels comparable to AtLRE (Fig. 2A; Liu et al. 2016).


330 For analyses of these transgenic lines, we compared them to _lre-7_ complemented with


331 _AtLRE_ _pro_ _:AtLRE-cYFP_ rather than to Col-0, which was used as the reference in Figure 1.


332 This transgenic line provides the appropriate matched control for the _lre-7_ genetic


333 background, AtLRE promoter, and cYFP-tagged protein used in the test constructs.


334 These results established that both transgenes are functional and that both orthologs


335 retained substantial basal LRE activity, allowing us to assess whether they differed


336 specifically in their ability to support heterospecific PTR.


337 We first tested whether AlyLRE improved reception of _A. lyrata_ pollen. Ovules


338 expressing AtLRE and pollinated with _A. lyrata_ pollen exhibited approximately 88.6%


339 OG/PEO, whereas those expressing AlyLRE showed a modest but significant decrease


340 to approximately 67.7% and 84.8% OG/PEO in two out of three lines tested (Fig. 2B).


341 Ovules expressing SiLRE exhibited approximately 86.5%, 83.9%, and 79.1% OG/PEO in


342 three lines when crossed to _A. lyrata_ pollen. Although the first SiLRE line was comparable


343 to AtLRE, the other two showed significantly lower OG/PEO when pollinated with A _. lyrata_


344 pollen. Thus, some lines expressing either AlyLRE or SiLRE showed reduced OG/PEO,


345 indicating that improved PTR was not restricted to the species-matched AlyLRE ortholog.


346 In contrast, we observed a substantially larger effect with _S. irio_ pollen. Ovules


347 expressing AlyLRE did not significantly improve PTR relative to ovules expressing AtLRE


348 and pollinated with _S. irio_ pollen, with pollen tube overgrowth frequencies from 72.3% to


349 74.4% OG/PEO (Fig. 2C). By contrast, all three independent SiLRE lines showed


350 markedly reduced pollen tube overgrowth (39.7%, 39.9%, and 36.8%) when pollinated


351 with _S. irio_ pollen (Fig. 2C). The improvement conferred by SiLRE also exceeded the level


352 of PTR observed in wild-type _A. thaliana_ pistils pollinated with _S. irio_ (77.1%) (Fig. 2C).


353 Independent experiments performed at the Institute of Experimental Botany, Czech


354 Academy of Sciences, reproduced the SiLRE effect (Figure S5). Thus, expression of the


12


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


355 species-matched SiLRE ortholog markedly improved reception of _S. irio_ pollen among


356 pollen tube-entered _A. thaliana_ ovules.


357 We next asked whether this enhancement was a general property of _LRE/LLG_


358 family members (Tsukamoto et al. 2010; Noble et al. 2022). We used previously


359 described _A._ _thaliana_ _lre-7_ lines expressing full-length _A._ _thaliana_ _LLG1_


360 ( _AtLRE_ _pro_ _:AtLLG1-cYFP_ ) or the single-copy _Cleome violacea LRE/LLG1_ ortholog


361 ( _AtLRE_ _pro_ _:Clevi-LRE/LLG1-cYFP_ ) (Noble et al. 2022). Both fusion proteins localize


362 primarily to the filiform apparatus when expressed in synergid cells and retain substantial


363 activity during conspecific PTR (Noble et al. 2022). When these lines were pollinated with


364 _S. irio_, none of the three independent lines expressing either construct significantly


365 reduced pollen tube overgrowth relative to the AtLRE-complemented control (61.6%–


366 80.9% OG/PEO; Fig. 2D). Thus, the ability of SiLRE to improve reception of _S. irio_ pollen


367 is not a general property of the _LRE/LLG_ family or of the ancestral single-copy LRE/LLG1

368 like protein, but instead depends on features that arose within the LRE lineage.


369


370 **The** _**LRE**_ **clade, but not the** _**LLG1**_ **clade, experienced relaxed selective pressures**


371 **following duplication**


372 AlyLRE and SiLRE affect heterospecific PTR modestly and variably with _A. lyrata_


373 pollen, but strongly and consistently with species-matched SiLRE and _S. irio_ pollen,


374 suggesting that sequence changes acquired during LRE diversification may contribute to


375 species-dependent heterospecific PTR. To identify candidate amino acid residues that


376 might contribute to the species-matched enhancement of heterospecific PTR, we placed


377 these functional differences in an evolutionary framework by reconstructing an expanded


378 _LRE/LLG1_ phylogeny and testing whether the two clades experienced different selective


379 pressures following duplication.


380 We identified _LRE/LLG1_ homologs from two Malvales and 25 Brassicales species,


381 with broader sampling of non-Brassicaceae Brassicales, early-diverging Brassicaceae,


382 and Lineage II Brassicaceae than in our previous analysis (Noble et al. 2022) (Table S1


383 and Table S2). The resulting phylogeny recovered two strongly supported Brassicaceae


384 clades corresponding to _LRE_ and _LLG1_ (bootstrap support = 100% and 99%,


385 respectively) (Figure S6). In contrast, the single-copy _LRE/LLG1_ orthologs from non

13


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


386 Brassicaceae Brassicales fell outside both clades (Fig. 3; Figure S6). This pattern is


387 consistent with _LRE_ and _LLG1_ having arisen from the Brassicaceae alpha whole-genome


388 duplication (Noble et al. 2022). Both paralogs were retained in all 17 Brassicaceae


389 species included in our sampling.


390 The branch leading to the _LRE_ clade was longer than the branch leading to _LLG1_,


391 suggesting that the _LRE_ clade experienced greater sequence divergence after duplication


392 (Fig. 3). In the branch-model analysis, the _LLG1_ clade did not differ significantly from the


393 background lineages in selective pressure (p = 0.3853). In contrast, the _LRE_ clade had a


394 significantly higher nonsynonymous-to-synonymous substitution-rate ratio than the


395 background (ω₁ = 0.37659 versus ω₀ = 0.28972; p = 0.0079; Table S3). Because both ω


396 values remained below 1, the _LRE_ clade was still under purifying selection, but that


397 selection was significantly weaker than in the background lineages.


398 Site-model analyses also showed that selective constraint differed among sites


399 within LRE/LLG1 proteins ( _p_ &lt; 0.0001), with one class of sites under strong purifying


400 selection (ω₀ = 0.19513; Table S3; Figure S7; Table S5). Likewise, Model C test identified


401 a subset of sites that evolved faster in the _LRE_ clade (ω₂,LRE = 0.53707) than in the


402 background lineages (ω₂, background = 0.31928; _p_ = 0.0077). However, the branch-site


403 tests provided no clear evidence for positive selection on either the _LRE_ post-duplication


404 branch ( _p_ = 0.0838) or branches within the _LRE_ clade ( _p_ = 1; Table S3). These results


405 suggest that, after the duplication that gave rise to _LRE_ and _LLG1_, selective constraint


406 was reduced at a subset of sites in the _LRE_ lineage, allowing greater sequence


407 divergence while the protein remained under purifying selection. In contrast, _LLG1_


408 showed no comparable shift in selective pressure and appears to have remained under


409 constraints similar to those acting on the background lineages.


410


411 **Evolutionary analyses identify E47 and R120 as candidates for LRE functional**


412 **diversification**


413 Although the branch-site tests did not detect significant positive selection, the


414 associated Empirical Bayes analyses identified two LRE sites of evolutionary interest.


415 Glutamic acid at position 47 (E47) was assigned to an elevated-rate class along the LRE


416 post-duplication branch and within the LRE clade, whereas arginine at position 120


14


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


417 (R120) was identified only within branches of the LRE clade (Table S4; Figure S7; Table


418 S5). Because the branch-site tests were not significant, we considered these residues


419 candidates for functional diversification rather than sites under positive selection. Notably,


420 the residue homologous to LRE R120 in LLG1, glycine 120 (G120), contributes to RALF


421 23 binding specifically to LLG1 but not LRE (Xiao et al. 2019).


422 At the position corresponding to E47 in AtLRE, proline (P) predominated in the


423 single-copy _LRE/LLG1_ grade and the _LLG1_ clade, indicating that P47 is the likely


424 ancestral state (Fig. 3). By contrast, glutamic acid was present in nearly all sampled _LRE_


425 orthologs, suggesting that a P-to-E substitution occurred before diversification of the _LRE_


426 clade and was subsequently strongly conserved.


427 At the position corresponding to R120, glycine (G) predominated in the single-copy


428 _LRE/LLG1_ grade and the _LLG1_ clade, indicating that G120 was likely ancestral. Within


429 the LRE clade, however, glycine was largely replaced by arginine or threonine. Arginine


430 predominated among Lineage I orthologs, whereas threonine predominated among


431 Lineage II orthologs and occurred in selected additional lineages (Fig. 3).


432 E47 and R120 therefore represent two distinct patterns of LRE evolution: an


433 ancestral substitution (E47) that became conserved throughout the LRE clade and a site


434 (R120) that diversified in association with certain lineages. We therefore tested whether


435 variation at these positions alters LRE's ability to support heterospecific PTR.


436


437 **LRE variants retain basal activity, but R120T modestly compromises conspecific**


438 **pollen tube reception**


439 To test whether sequence changes associated with LRE divergence from _LLG1_


440 contributed to LRE functional diversification, we replaced the _LRE_ clade residues E47


441 and R120 with the corresponding ancestral, _LLG1_ -like residues, P47 and G120. We


442 generated the single-variant constructs _AtLRE_ _pro_ _:AtLRE(E47P)-cYFP_ and


443 _AtLRE_ _pro_ _:AtLRE(R120G)-cYFP_ and the double-variant construct _AtLRE_ _pro_ _:AtLRE(E47P–_


444 _R120G)-cYFP_ (Figure S2). We also replaced the Lineage I-associated arginine at position


445 120 in AtLRE with threonine (T), the residue found predominantly at this position in


446 Lineage II LRE orthologs, including in SiLRE, generating _AtLRE_ _pro_ _:AtLRE(R120T)-cYFP_ .


15


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


447 This variant allowed us to test whether lineage-associated variation at position 120 affects


448 AtLRE in _A. thaliana'_ s ability to support PTR with pollen from _S. irio_, a Lineage II species.


449 We introduced all four AtLRE variants into the _A. thaliana lre-7_ background and


450 selected three independent transgenic lines for each construct. cYFP fluorescence


451 confirmed that all four variants were expressed in synergid cells and localized to the


452 filiform apparatus, as in wild-type AtLRE-cYFP (Figure S4B to C, F to I). We first pollinated


453 these lines with _A. thaliana_ pollen to determine whether the substitutions affected the


454 basal LRE activity required for conspecific PTR.


455 Lines expressing AtLRE-E47P or AtLRE-R120G exhibited low levels of pollen tube


456 overgrowth (1.4%–10% OG/PEO) that were not significantly different from the AtLRE


457 control, indicating that either ancestral-state substitution alone retained substantial


458 conspecific PTR activity (Fig. 4A). The AtLRE-E47P–R120G double variant behaved


459 similarly: all three independent lines exhibited very low levels of pollen tube overgrowth


460 and were not significantly different from the AtLRE control (0.1%-0.7% OG/PEO; Fig. 4A).


461 Thus, replacing E47 and R120 individually or together with their inferred ancestral, LLG1

462 like residues did not measurably compromise AtLRE function during conspecific PTR.


463 By contrast, substitution of R120 with the Lineage II-associated threonine


464 significantly reduced conspecific PTR efficiency. All three AtLRE-R120T lines exhibited


465 significantly higher pollen tube overgrowth than the AtLRE control, ranging from 8.3% to


466 23.5% OG/PEO (Fig. 4C). Nevertheless, these values remained substantially below the


467 84.6% overgrowth observed in _lre-7_, demonstrating that AtLRE-R120T retained


468 considerable LRE activity. Thus, unlike the ancestral R120G substitution, introducing the


469 Lineage II-associated R120T substitution partially compromised AtLRE's ability to support


470 reception of _A. thaliana_ pollen tubes.


471


472 **Distinct LRE sequence changes improve reception of** _**S. irio**_ **pollen**


473 Because all four AtLRE variants retained substantial LRE activity during


474 conspecific PTR, although R120T partially reduced reception of _A. thaliana_ pollen, we


475 next tested whether these substitutions altered heterospecific PTR in the _A. thaliana × S._


476 _irio_ interlineage cross. In heterospecific crosses, SiLRE produced a substantially greater


477 improvement with _S. irio_ pollen than AlyLRE did with _A. lyrata_ pollen (Fig. 2C). We


16


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


478 therefore focused the current analysis of the LRE variants on the _A. thaliana × S. irio_


479 cross, where the larger phenotypic effect provided greater resolution for determining


480 whether candidate residues affect heterospecific PTR.


481 Reverting either residue individually to its inferred ancestral, LLG1-like state did


482 not consistently improve reception of S _. irio_ pollen. None of the three R120G lines differed


483 significantly from _lre-7_ complemented with wild-type AtLRE, and two of the three E47P


484 lines likewise showed no improvement (Fig. 4B). Although one E47P line exhibited


485 significantly less pollen tube overgrowth, the effect was modest and was not reproduced


486 in the other independent lines. Thus, neither the E47P nor the R120G substitution alone


487 was sufficient to enhance the ability of AtLRE expressed in _A. thaliana_ ovules to support


488 the reception of _S. irio_ pollen.


489 By contrast, replacing the Lineage I-associated arginine at position 120 with the


490 threonine present in SiLRE and predominant among Lineage II LRE orthologs (R120T)


491 significantly improved PTR with _S. irio_ pollen. All three R120T lines exhibited significantly


492 less pollen tube overgrowth than the wild-type AtLRE control (69.5% OG/PEO), with the


493 strongest line exhibiting 42.1% OG/PEO (Fig. 4D). This effect is particularly notable


494 because the same R120T substitution significantly reduced PTR efficiency with


495 conspecific _A. thaliana_ pollen. Thus, R120T did not simply increase overall LRE activity;


496 rather, it altered LRE function in a pollen-species-dependent manner, reducing


497 compatibility with _A. thaliana_ pollen while improving compatibility with _S. irio_ pollen. These


498 results show that changing a single AtLRE residue to the predominant state in the pollen

499 donor lineage is sufficient to shift heterospecific PTR toward greater compatibility with _S._


500 _irio_ pollen.


501 Unexpectedly, the combined ancestral-state substitutions E47P–R120G also


502 markedly improved reception of _S. irio_ pollen. All three independent lines expressing the


503 double variant exhibited significantly less pollen tube overgrowth than the wild-type AtLRE


504 control, with values of 36.7%, 39.0%, and 21.2% OG/PEO (Fig. 4B). This reproducible


505 effect was not observed when either ancestral-state substitution was introduced


506 individually, indicating that the effects of residues 47 and 120 depend on sequence


507 context. Notably, unlike R120T, the E47P–R120G variant improved reception of _S. irio_


508 pollen without measurably compromising conspecific PTR with _A. thaliana_ pollen.


17


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


509 Together, these results show that distinct sequence changes in AtLRE can


510 improve reception of _S. irio_ pollen. The Lineage II-associated R120T substitution


511 produces a pollen-species-dependent shift in PTR, whereas the E47P–R120G


512 combination improves heterospecific reception without a detectable loss of conspecific


513 PTR activity. Thus, residues 47 and 120 contribute to AtLRE's ability to support


514 heterospecific PTR and identify sequence context within LRE as an important determinant


515 of incompatibility with _S. irio_ pollen.


516


517 **Pollen tube reception-enhancing LRE variants share predicted conformational**


518 **features**


519 Because R120T and E47P–R120G both improved the reception of _S. irio_ pollen


520 despite containing distinct mutations, we asked whether these variants shared structural


521 conformational features that distinguished them from variants that did not consistently


522 improve heterospecific PTR. We first generated standard AlphaFold 3 predictions for wild

523 type AtLRE, E47P, R120G, R120T, and E47P–R120G using a randomized seed


524 approach. These predictions did not display a consistent structural difference that clearly


525 delineated the variants from wild-type AtLRE or from one another (Figure S8). Prediction


526 confidence was lowest in the N- and C-terminal regions, prompting us to examine whether


527 conformational sampling differed among variants in these regions.


528 We therefore used subsampled AlphaFold, which increases the diversity of


529 predicted conformations by reducing the depth of the multiple-sequence alignment used


530 for modeling (Monteiro da Silva et al. 2024). Across all variants, the central α-helical


531 bundle remained well conserved, whereas the N- and C-terminal regions displayed


532 greater conformational variability and lower prediction confidence, likely due to poor


533 sequence coverage in the multiple-sequence alignment (Figure S9).


534 Comparison of the highest-ranked structures revealed that R120T and E47P–


535 R120G adopted predicted conformations in which the N- and C-terminal regions were


536 farther apart. By contrast, wild-type AtLRE and R120G adopted more compact structural


537 ensembles in which the termini were closer together, while E47P exhibited an


538 intermediate arrangement. The R120T and E47P–R120G (EPRG in Fig. 5 and S10)


18


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


539 ensembles also sampled broader conformational spaces than those of wild-type AtLRE,


540 R120G, or E47P (Fig. 5A).


541 RMSD analysis within each ensemble showed similar results. R120T and E47P–


542 R120G had the widest spread relative to their top-ranked structures, showing more


543 variability across the predicted models. E47P showed less variation, and R120G was


544 closest to wild-type AtLRE (Fig. 5; Figure S10).


545 The shared predicted features of R120T and E47P–R120G parallel their shared


546 ability to improve the reception of _S. irio_ pollen. Although these computational evaluations


547 do not establish the underlying molecular mechanism, they suggest that distinct amino


548 acid substitutions at positions 47 and 120 may relieve the heterospecific PTR barrier by


549 altering the conformational states of LRE samples. Together, our genetic, evolutionary,


550 and structure prediction results implicate these residues in limiting heterospecific _S. irio_


551 PTR by _A. thaliana_ ovules and provide a testable structural hypothesis for how LRE


552 sequence variation can modulate prezygotic reproductive compatibility.


553

### 554 Discussion


555


556 **LRE contributes to lineage-associated compatibility during heterospecific pollen**


557 **tube reception**


558 Our results show that LRE, a core component of conspecific PTR, also influences


559 compatibility between heterospecific pollen tubes and the female reception machinery,


560 thus affecting whether PTR succeeds or fails. This indicates that LRE can influence


561 heterospecific compatibility. Several female components have previously been implicated


562 in heterospecific PTR. FERONIA (FER), a core component of conspecific PTR, was


563 initially linked to heterospecific reception when _A. lyrata_ and _Cardamine flexuosa_ pollen


564 tubes were found to overgrow within wild-type _A. thaliana_ female gametophytes,


565 producing a phenotype resembling that of _fer_ mutants (Escobar-Restrepo et al. 2007).


566 This observation led to the proposal that divergence in FER-dependent pollen tube–


567 synergid signaling could contribute to reproductive isolation. More direct genetic evidence


568 was subsequently provided by Ju et al. (2021), who showed that reception of _A. lyrata_


19


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


569 pollen was severely compromised in _fer-1_ ovules and that the enhancement of


570 heterospecific reception by filiform apparatus-targeted NTA remained dependent on FER.


571 Thus, components of the core machinery that mediate conspecific PTR could also


572 influence the outcome of heterospecific interactions.


573 ARU, an OST3/6 subunit involved in protein N-glycosylation, provides a different


574 example. Loss of ARU strongly increases overgrowth of _A. lyrata_ pollen tubes while


575 leaving conspecific PTR largely unaffected (Müller et al. 2016). The greater sensitivity of


576 heterospecific reception to ARU function suggests that proper configuration of the female


577 reception machinery is especially important when interacting male and female


578 components are less well matched. Thus, the heterospecific PTR barrier can arise


579 through altered performance of molecular components whose functions are embedded


580 within the normal female reception machinery rather than through a dedicated pathway


581 for rejecting foreign pollen.


582 Our results expand this framework by showing that sequence variation within LRE


583 modulates heterospecific compatibility in a manner that appears to reflect evolutionary


584 lineage rather than strict species matching. AlyLRE did not uniquely improve reception of


585 _A. lyrata_ pollen; both AlyLRE and SiLRE produced modest and variable improvements in


586 some lines. By contrast, with the more distantly related Lineage II species _S. irio_, SiLRE


587 consistently and markedly improved reception, whereas AlyLRE did not improve


588 reception relative to AtLRE. Together with the effects of substitutions at residues 47 and


589 120, this evidence indicates that lineage-associated sequence diversification within LRE


590 can alter the compatibility of the female reception machinery with pollen tubes from


591 different evolutionary lineages. The results therefore do not support a simple species

592 specific recognition system in which each LRE ortholog preferentially receives pollen from


593 its own species. Rather, variation within this conserved PTR component appears to


594 modify the compatibility of the female machinery with heterospecific pollen in a lineage

595 associated and sequence-context-dependent manner. Collectively, these results indicate


596 that diversification of LRE following duplication generated protein-level variation that


597 contributes to lineage-associated reproductive compatibility.


598 These conclusions should, however, be interpreted considering what our assay


599 directly measures. PTR success was inferred from pollen tube morphology in aniline blue

20


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


600 stained pistils. This readout provides a strong and tractable means of distinguishing PTR


601 failure, indicated by continued pollen tube growth and coiling after micropyle entry, from


602 a reception-like outcome, indicated by pollen tube termination near the synergid cells


603 region in ovules. However, it does not directly demonstrate downstream hallmarks of


604 productive reception, including sperm discharge and initiation of double fertilization. This


605 distinction may be especially relevant in heterospecific pollinations, in which pollen tube


606 termination could, in principle, reflect growth arrest without sperm discharge. We


607 attempted to establish _S. irio_ lines carrying fluorescent pollen- and sperm-cell markers


608 that would permit direct visualization of these events, but despite repeated efforts, we


609 were unable to transform _S. irio_ successfully. Consequently, although the reduced


610 overgrowth observed with SiLRE and the AtLRE variants R120T and E47P–R120G


611 demonstrates that LRE sequence variation shifts heterospecific pollen tube behavior


612 toward a PTR-like outcome, whether these changes also increase sperm discharge and


613 productive fertilization remains to be determined. Future development of transformation


614 methods or alternative live-cell markers for _S. irio_ will be important for directly linking these


615 changes in pollen tube behavior to fertilization outcomes. Nevertheless, the consistent


616 reductions in pollen tube overgrowth seen across independent transgenic lines provide


617 strong comparative evidence that LRE sequence variation is associated with protein-level


618 variation that may contribute to lineage-associated reproductive compatibility.


619


620 **Evolutionary diversification of LRE identifies residues 47 and 120 as candidates for**


621 **functional divergence**


622 Consistent with the idea that diversification of LRE contributed to its species-dependent


623 behavior during heterospecific PTR, our expanded evolutionary analyses, building on our


624 previous work (Noble et al. 2022), revealed greater sequence divergence in _LRE_ than in


625 its paralog _LLG1_ following their duplication. Our earlier studies suggested that retention


626 of the LLG family was driven largely by regulatory subfunctionalization: family members


627 retained substantial functional interchangeability when expressed in the same cellular


628 context but acquired distinct expression domains (Noble et al. 2022). The present results


629 suggest an additional layer of divergence at the protein level. Although both _LRE_ and


630 _LLG1_ have been broadly retained across Brassicaceae, purifying selection was


21


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


631 significantly relaxed in the _LRE_ clade, particularly at a subset of sites, whereas selective


632 pressure on _LLG1_ remained like that on the background lineages (Table S3). One


633 possible explanation is that the relatively restricted expression of _LRE_ in the female


634 gametophyte reduced pleiotropic constraints on its sequence, permitting greater


635 evolutionary flexibility, whereas the varied developmental functions of _LLG1_ may have


636 imposed stronger constraints on sequence divergence. Thus, our data extend the


637 previous model of predominantly expression-based divergence within the LLG family by


638 showing that LRE also underwent protein-level functional diversification that contributes


639 to species-dependent compatibility during heterospecific PTR.


640 Within this pattern, evolutionary analyses highlighted E47 and R120 as residues


641 of particular interest (Table S4; Table S5; Figure S7). Our results showed that E47 shows


642 a sequence change associated with the early divergence of LRE from LLG1, whereas


643 R120 captures subsequent lineage-associated diversification within the LRE clade (Fig.


644 3). The structural and functional contexts of these residues provided additional reasons


645 to consider their contributions to LRE function. E47 lies immediately after the first cysteine


646 in LRE and is predicted by AlphaFold to reside in a loop following an α-helix (Jumper et


647 al. 2021; Fleming et al. 2025). R120 has an independently established functional context:


648 the homologous residue in LLG1 contributes to RALF peptide-binding specificity, and


649 substitutions at this position can alter RALF binding (Xiao et al. 2019). Thus, both residues


650 occupy positions that could influence LRE function, although their separate evolutionary


651 histories suggest they may have played different roles in its diversification.


652 The functional consequences of variation at these positions were also informative.


653 Replacing E47 or R120 individually with their ancestral, LLG1-like residues did not


654 substantially impair conspecific PTR, and the combined E47P–R120G variant likewise


655 retained normal basal LRE activity. This result agrees with our previous observation that


656 LLG1, which has proline and glycine at the equivalent positions, can rescue the _lre-7_ PTR


657 defect (Noble et al. 2022). In contrast, changing the Lineage I-associated R120 residue


658 to the Lineage II-associated threonine caused only a modest decrease in conspecific


659 reception of _A. thaliana_ pollen, indicating that the R120T substitution retains much of LRE


660 function. These data suggest that considerable sequence variation at positions 47 and


22


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


661 120 can be accommodated without loss of core LRE function, although the Lineage II

662 associated R120T substitution modestly reduces conspecific PTR efficiency.


663


664 **Effects of E47 and R120 substitutions on heterospecific pollen tube reception**


665 Although substitutions at E47 and R120 retained substantial basal LRE function,


666 their effects during heterospecific pollination suggest that these residues contribute to


667 reproductive compatibility rather than to LRE function alone. The strongest evidence for


668 a lineage-associated effect was obtained with the R120T substitution, which substantially


669 enhanced the PTR of _S. irio_ pollen. Nevertheless, position 120 cannot by itself explain


670 species-dependent compatibility. Full-length SiLRE produced a greater improvement in


671 reception of _S. irio_ pollen than the R120T substitution alone, indicating that additional


672 sequence features of SiLRE add to its species-matched activity. This interpretation is also


673 supported by the conspecific pollination results. The R120T substitution modestly


674 reduced reception of _A. thaliana_ pollen, whereas full-length SiLRE supported conspecific


675 PTR as well as AtLRE. This suggests that the effect of T120 depends on the sequence


676 context in which it occurs, with other residues in SiLRE likely modifying or compensating


677 for its effect.


678 The E47 results add an important dimension to this interpretation. E47P alone did


679 not consistently improve reception of _S. irio_ pollen, nor did the ancestral R120G


680 substitution. Remarkably, however, combining these two individually ineffective


681 substitutions in E47P–R120G substantially improved heterospecific PTR in all three


682 independent lines while leaving conspecific reception essentially unaffected. The strong


683 phenotype of the double variant therefore implicates E47 in a context-dependent


684 interaction with residue 120. It demonstrates that improved heterospecific PTR cannot be


685 explained solely by matching residue 120 to the pollen-donor Lineage. Instead, positions


686 47 and 120 together influence the incompatibility imposed by wild-type AtLRE during


687 reception of _S. irio_ pollen. Together, these data indicate that LRE-dependent reproductive


688 compatibility is influenced by residue identity and sequence context. In contrast, LRE


689 sequence variation represents only one component of a wider species-dependent


690 signaling interaction.


691


23


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


692 **LRE may mediate heterospecific compatibility through interactions with pollen**


693 **tube-derived signals**


694 Although changes in LRE substantially weakened the heterospecific PTR barrier, they did


695 not fully restore reception, suggesting that LRE-dependent compatibility operates within


696 female signaling machinery and that matching additional female signaling components to


697 the pollen donor could further improve PTR. FER is an obvious candidate because it


698 functions with LRE during PTR, and our genetic results show that loss of FER further


699 compromises reception of _A. lyrata_ pollen. In _A. thaliana_ synergid cells, the CrRLK1L


700 receptors ANJEA and HERK1 also function redundantly with FER, and LRE, FER,


701 ANJEA, and HERK1 can associate in a multimeric receptor complex (Galindo-Trigo et al.


702 2020). Additional female components that could be tested include the GPI-anchored


703 EARLY NODULIN-LIKE proteins (ENODLs), which interact with FER and are also


704 required for efficient PTR (Hou et al. 2016).


705 Compatibility is also likely to depend on the pollen tube-derived signals to which


706 this machinery responds. In our experiments, the female genetic background remained


707 _A. thaliana,_ while pollen identity varied, and changing LRE sequence altered PTR in a


708 pollen-species-dependent manner. We therefore propose that LRE sequence variation


709 may influence how effectively the female LRE-associated receptor machinery responds


710 to signals from the incoming pollen tube (Fig. 6). In this model, one or more _S. irio_ -derived


711 factors interact inefficiently with AtLRE, resulting in ineffective activation of the female


712 reception machinery, whereas replacement of AtLRE with SiLRE improves compatibility


713 with these signals and promotes productive PTR (Fig. 6).


714 RALF peptides are particularly strong candidates for these pollen tube-derived


715 factors. A biochemical precedent comes from RALF23 signaling through LLG1: RALF23


716 binds LLG1, which contains glycine at the position homologous to LRE R120, whereas


717 wild-type LRE, which contains arginine at this position, shows little or no detectable


718 RALF23 binding; substitution of LRE R120 with the LLG1-like glycine restores


719 measurable RALF23 binding (Xiao et al. 2019). Thus, the amino acid at this homologous


720 position can contribute strongly to RALF-binding specificity within the LLG protein family,


721 although the difference in affinity between LLG1 and LRE-R120G also indicates that


722 additional sequence context is important. This relationship is especially relevant to our


24


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


723 findings because variation at the same LRE position alters PTR in a pollen-species

724 dependent manner.


725 Independent genetic evidence further identifies biologically relevant pollen tube

726 derived RALFs that function during PTR. The pollen tube-expressed transcription factors


727 MYB97, MYB101, and MYB120 are required for normal reception, and five of their


728 downstream targets - RALF6, RALF7, RALF16, RALF36, and RALF37 - are required for


729 efficient PTR. Their combined loss causes pollen tube overgrowth, and these peptides


730 can interact with FER, ANJEA, and HERK1 and promote receptor-complex interactions


731 (Zhong et al. 2022). These data make the five RALFs strong candidates for the pollen


732 tube-derived factors represented schematically in Fig. 6. Together, this evidence supports


733 a model in which LRE sequence diversification modulates the response of the female


734 reception machinery to pollen tube-associated signals (Fig. 6). In this view, the prezygotic


735 PTR barrier is not controlled by a single molecular gatekeeper but emerges from


736 compatibility among interacting male and female signaling components.


737


738 **Reconstituting lineage-matched female receptor complexes to further enhance**


739 **heterospecific pollen tube reception**


740 A further prediction of this model is that LRE sequence variation is only one


741 contributor to heterospecific compatibility, and that reception may improve further when


742 additional interacting components of the female receptor machinery are also matched to


743 the pollen-donor lineage. In our experiments, SiLRE and AtLRE variants weakened the


744 _A. thaliana × S. irio_ PTR barrier in an otherwise _A. thaliana_ female background, implying


745 that incompatibility can be partially relieved by tweaking a single component of the


746 complex. However, because LRE functions within a multimeric receptor module that


747 includes FER and the related CrRLK1L receptors ANJEA and HERK1 (Galindo-Trigo et


748 al. 2020) and is probably shaped by additional associated factors such as ENODLs (Hou


749 et al. 2016), it will be informative to ask whether “reconstituting” a more lineage-matched


750 complex in _A. thaliana_ improves compatibility beyond what is achieved by LRE alone. For


751 example, expressing _S. irio_ orthologs of FER and/or other core receptors together with


752 SiLRE in _A. thaliana_ synergid cells could test whether co-adapted receptor components


753 provide additive or synergistic improvements in heterospecific PTR. If such a lineage

25


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


754 matched complex reconstitution yields a stronger reduction in pollen tube overgrowth


755 and/or increases downstream indicators of successful PTR, it would support a model in


756 which species-dependent PTR outcomes emerge from cumulative matching effects


757 across male-female signaling components rather than from a single gatekeeping factor.


758

### 759 Conclusions


760


761 Our results show that a component of the conserved machinery required for normal


762 pollen tube reception can also shape reproductive compatibility between species. Rather


763 than acting as a dedicated barrier to heterospecific pollen, LRE may influence how


764 effectively the female reception machinery functions with pollen tubes from different


765 evolutionary lineages. Importantly, this property can be altered by relatively limited


766 changes in LRE sequence without eliminating its core role in conspecific PTR. The


767 opposing effects of LRE orthologs and of substitutions at residues 47 and 120 additionally


768 indicate that heterospecific compatibility is determined by sequence context rather than


769 by a simple species-specific recognition code.


770 These results suggest that prezygotic barriers at the PTR stage may emerge


771 through evolutionary divergence of otherwise conserved male-female signaling


772 components. We propose that LRE sequence variation may modify the compatibility


773 between pollen tube-derived signals and the female receptor machinery, with successful


774 reception ultimately depending on both signal recognition and productive transmission


775 through downstream PTR pathways. In this view, heterospecific PTR failure may reflect


776 imperfect molecular matching across the signaling system rather than the action of a


777 single barrier factor. Defining the pollen tube-derived signals involved and determining


778 how their interactions with LRE-associated receptor complexes differ across lineages will


779 be important to understand how reproductive barriers evolve and how they might


780 eventually be modified to broaden interspecific compatibility.


781


782


26


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

### 783 Methods


784


785 **Plant materials and growth conditions**


786 _A. thaliana_, _Arabidopsis lyrata,_ and _Sisymbrium irio_ seeds were sterilized as


787 follows: 0.05 - 0.1 mL of seeds were added to a 1.5 mL microcentrifuge tube and placed


788 in a desiccator that contained an open beaker filled with 95 mL bleach and 5 mL HCl.


789 After 3 hours, sterilized seeds were plated on MS plates with 2% sucrose and appropriate


790 antibiotics for selection. _A. thaliana_ seeds were stratified for 3 days in the dark at 4ºC and


791 moved to a Percival growth chamber maintained at 21ºC with constant light. _S. irio_ seeds


792 were stratified for 10 days in the dark at 4ºC and moved to a Percival or Phytotron growth


793 chamber maintained at 21ºC with constant light. 10–14-day-old seedlings were


794 transplanted to soil and grown under long-day conditions: 16 hours of light at 21ºC and 8


795 hours of dark at 18ºC.


796 In the case of _A. lyrata_, sterilized seeds were suspended in sterile deionized water


797 and placed in the dark at 4ºC for 4 weeks. After 4 weeks, _A. lyrata_ seeds were dried and


798 plated on MS plates with 2% sucrose, stratified for 10 days in the dark at 4ºC, and moved


799 to a Percival growth chamber that was maintained at 21ºC with constant light. 10–14-day

800 old seedlings were then transplanted to soil and grown under long-day conditions: 16


801 hours of light at 20ºC and 8 hours of dark at 16ºC. Plants were kept in these conditions


802 for 2 months before vernalization at 4ºC with 20 hours of light and 4 hours of dark for 9


803 weeks. After vernalization, plants were moved back to the growth chamber with 16 hours


804 of light at 21ºC and 8 hours of dark at 16ºC.


805 All _A. thaliana_ seeds used in this study are the _Columbia_ ( _Col-0_ ) ecotype. The


806 following constructs were transformed into the homozygous _lre-7/lre-7_ mutant


807 background: _AtLRE_ _pro_ _:AtLRE-cYFP,_ _AtLRE_ _pro_ _:AtLRE(E47P)-cYFP,_


808 _AtLRE_ _pro_ _:AtLRE(R120G)-cYFP, AtLRE_ _pro_ _:AtLRE(R120T)-cYFP, AtLRE_ _pro_ _:AtLRE(E47P-_


809 _R120G)-cYFP, AtLRE_ _pro_ _:AlyLRE-cYFP,_ and _AtLRE_ _pro_ _:SiLRE-cYFP._ Transgenic plants


810 were selected on plates containing both hygromycin B and glufosinate ammonium.


811


812 **Cloning transgenic constructs**


27


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


813 The _AtLRE_ _pro_ _:AtLRE(E47P)-cYFP,_ _AtLRE_ _pro_ _:AtLRE(R120G)-cYFP,_


814 _AtLRE_ _pro_ _:AtLRE(E47P-R120G)-cYFP,_ and _AtLRE_ _pro_ _:AtLRE (R120T)-cYFP_ transgenic


815 lines were created by introducing nucleotide changes that will cause E to P, R to G, and


816 R to T changes, respectively, in the LRE protein sequence in the previously published


817 construct: _AtLRE_ _pro_ _:LRE-cYFP_ (Liu et al. 2016). The _AtLRE_ _pro_ _:AlyLRE-cYFP_ and


818 _AtLRE_ _pro_ _:SiLRE-cYFP_ transgenic lines were created following the protocol established


819 with the previously published _AtLRE_ _pro_ _:LRE-cYFP_ construct (Liu et al. 2016). Mutations


820 were introduced by PCR with PrimeSTAR® GXL DNA Polymerase (TaKaRa Bio Inc.;


821 Catalog # R050A) using primers and DNA templates listed in Table S6. Overlap PCR was


822 performed for _AtLRE_ _pro_ _:AtLRE(E47P-R120G)-cYFP_, _AtLRE_ _pro_ _:AlyLRE-cYFP_ and


823 _AtLRE_ _pro_ _:SiLRE-cYFP_ as indicated in Table S6prior to cloning. The amplified PCR


824 products were cloned into _AtLRE_ _pro_ _:LRE-cYFP_ plasmid linearized with _SpeI-HF_ (NEB,


825 Catalog # R3133S) and _AscI_ (NEB, Catalog # R0558S) by using the In-Fusion HD Cloning


826 Plus system (Clontech, Catalog # 639645). The recombinant plasmids were transformed


827 into Stellar™ Competent Cells (Clontech, Catalog # 636763), and positive colonies were


828 selected on LB plates containing Spectinomycin (100 µg/mL, Sigma-Aldrich, Catalog #


829 85555).


830 All constructs generated were sequence-verified for the introduced point mutations


831 (Eton Bioscience, Inc.) before transforming into _Agrobacterium tumefaciens_ (GV3101


832 pMP90 strain). The positive colony selected for transformation into _A. thaliana_ was also


833 verified by colony PCR for the presence of the transgene.


834


835 **Plant transformation**


836 Transformation solution containing _Agrobacterium tumefaciens_ (GV3101 pMP90


837 strain) harboring the desired transgene was sprayed onto _A. thaliana_ inflorescences


838 (Chung et al. 2000). Hygromycin-resistant T1 transformants were selected as described


839 (Harrison et al. 2006). Briefly, T1 seeds were plated and stratified in the dark at 4°C for


840 2-3 days. Plates were then placed in a Percival growth chamber set at 21°C with


841 continuous light (70-100 µmol·m &#91;-2&#93; - s &#91;-1&#93; ) for 5-6 hours, then placed in the dark at room


842 temperature for 3-4 days. Plates were then returned to the Percival growth chamber, and


28


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


843 transformants were selected based on the presence of true leaves, which were present


844 only in hygromycin-resistant plants.


845


846 **Identification of single-loci insertion lines**


847 For each construct, at least 15 T1 hygromycin-resistant transformants were


848 transplanted to soil. Among these lines, candidate single insertion lines were identified


849 based on segregation of resistance to hygromycin B among T2 progeny. T1 plants are


850 expected to be heterozygous for the transgene at the insertion locus. T1 plants that gave


851 rise to progeny with a 3:1 ratio of resistant to sensitive plants were considered as


852 candidate single locus insertion lines. 15 T2 plants of candidate lines were transplanted


853 to soil, and T3 selfed seeds were collected to identify homozygous lines in T3 populations.


854 Those T2 plants that gave rise to T3 progeny that segregated 100% resistance to


855 hygromycin B were homozygous for the transgene.


856


857 **Genomic DNA isolation from** _**A. thaliana**_ **seeds**


858 Genomic DNA was extracted from dry _A. thaliana_ seeds (50 mg) using a CTAB

859 based method adapted from Dellaporta et al. (1983). Seeds were frozen in liquid nitrogen,


860 ground to a fine powder, and suspended in CTAB B solution (1.4 M NaCl, 0.1 M Tris–HCl


861 (pH 8.0), 2% (w/v) CTAB). After incubation at 65°C, the homogenate was treated with a


862 1:1 mixture of CTAB B and CTAB C (50 mM Tris–HCl (pH 8.0), 10 mM EDTA (pH 8.0),


863 1% (w/v) CTAB) and centrifuged to remove debris. The supernatant was extracted twice


864 with chloroform solution (chloroform: isoamyl alcohol, 24:1), followed by precipitation with


865 CTAB C. The resulting DNA pellet was resuspended in CTAB D (50 mM Tris–HCl (pH


866 8.0), 10 mM EDTA (pH 8.0), 1 M CaCl, 1% (w/v) CTAB) containing RNase A and


867 incubated at 37°C. DNA was subsequently precipitated with ethanol, washed with 75%


868 ethanol, and resuspended in TE buffer. DNA yield and purity were assessed


869 spectrophotometrically at 260 nm, typically ranging between 200 ng/mL and 1 mg/mL,


870 and these samples were subsequently used to determine transgene copy number using


871 digital droplet PCR.


872


873 **Digital droplet PCR**


29


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


874 Genomic DNA from selected transgenic lines that were designated as single-locus


875 insertion lines was isolated and tested to check whether they had any variation in the


876 copy number of the transgene inserted in that locus using digital droplet PCR (ddPCR)


877 (Figure S3) as follows: ddPCR was performed using QX200TM ddPCR EvaGreen


878 Supermix on a Bio-Rad QX200 Droplet Digital PCR system. The single-copy gene PMR4


879 was used as a reference (Dong et al. 2015). Two sets of specific oligonucleotides were


880 designed to target the PMR4 gene or the cYFP coding sequence. For the ddPCR reaction


881 mix, a 20 µL ddPCR supermix was made for each gene using 10ng template genomic


882 DNA and 50nm oligos mix. Each reaction was loaded into 8-well cartridges for droplet


883 generation and then transferred to the QX200 Droplet Digital PCR system, with a


884 standard ddPCR reaction step of 10 minutes at 95°C, 40 cycles of 94°C for 30 seconds


885 and an annealing temperature of 58 °C for 60 seconds, and a final 10 minutes at 98°C.


886 The ddPCR data were analyzed using QuantaSoft analysis (Bio-Rad) and plotted using


887 gPlot database.


888


889 **Citrine YFP expression in ovules**


890 Expression of the cYFP in the FA of synergid cells was examined and imaged with


891 a Zeiss LSM880 inverted confocal microscope. For cYFP imaging, samples were excited


892 with a 488-nm laser line, and emission was collected between 510 and 550 nm. YFP


893 images were processed with ZEN Black and ImageJ software (http://imagej.nih.gov/ij/;


894 Schindelin et al. 2012). ImageJ was used to insert scale bars and assemble the final


895 images.


896


897 **Standardization of inter-lineage crosses**


898 Plants that were 4 or 5 weeks old after they were transplanted to soil were used


899 as male and female parents. Stage 12c buds were emasculated as previously described


900 (Smyth et al. 1990). Twenty-four hours after emasculation, pistils were manually and


901 profusely pollinated with _A. thaliana, A. lyrata,_ or _S. irio_ pollen. Pistils were collected 24


902 hours after pollination for aniline blue staining.


903


904 **Aniline blue staining**


30


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


905 Aniline blue staining was performed as described (Mori et al. 2006; Noble and


906 Palanivelu 2020). Pistils were stained for 24 hours and mounted in decolorized aniline


907 blue solution (DABS) containing 50% glycerol, then sealed with clear nail polish (Sally


908 Hansen, Hard as Nails). Pollen tube-ovule interactions were scored under a Zeiss


909 Axiovert 100 fluorescent microscope. Images were captured using Metamorph (Version


910 7; Molecular Devices) with an IR filter and auto-exposure settings. The number of pistils


911 and ovules scored is indicated in each figure.


912


913 **Classification and quantification of pollen tube reception outcomes**


914 Pollen tube behavior in aniline blue-stained pistils was classified into four


915 categories. Type 1 comprised ovules for which no pollen tube was observed at the


916 micropyle. Type 2 comprised ovules in which a pollen tube reached the micropyle but did


917 not enter the female gametophyte. Type 3 comprised ovules in which a pollen tube


918 entered the micropyle but continued to grow and coil within the female gametophyte,


919 providing morphological evidence of failed pollen tube reception. Type 4 comprised


920 ovules in which a pollen tube entered the female gametophyte and terminated near the


921 expected position of the synergid cells without visible continued growth or coiling.


922 All four categories were scored in each cross. For comparisons of PTR after pollen


923 tube entry, failure was calculated as the percentage of pollen tube-entered ovules


924 exhibiting overgrowth &#91;%OG/PEO = Type 3/(Type 3 + Type 4) × 100&#93;. Types 1 and 2 were


925 excluded from this calculation because the pollen tubes had not entered the female


926 gametophyte and PTR therefore could not be assessed. Consequently, %OG/PEO


927 measures PTR failure conditional on pollen tube entry rather than the total reproductive


928 barrier, which may also include defects occurring before entry.


929 Classification as Type 3 required unequivocal evidence of continued pollen tube


930 growth or coiling within the female gametophyte. Type 4 classification was based on the


931 lack of apparent overgrowth and did not directly establish pollen tube rupture, sperm


932 release, or fertilization. Some Type 4 outcomes could therefore represent pollen tube


933 arrest without productive reception. Accordingly, %OG/PEO should be interpreted as a


934 conservative estimate of morphologically evident PTR failure among pollen tube-entered


935 ovules.


31


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


936


937 **Statistical analysis**


938 For each crossed pistil, the proportion of ovules that received an abnormal pollen


939 tube was calculated by dividing the number of ovules with abnormal pollen tube behavior


940 (type 3) by the total number of ovules that received a pollen tube (both abnormal (type 3)


941 and normal (type 4); Fig. 1B to F). This proportion was then subjected to an arcsine


942 square root transformation using Microsoft Excel. Transformed values were used in two

943 sample _t_ -tests comparing each test line to the _AtLRE_ _pro_ _:AtLRE-cYFP_ control line, except


944 in Fig. 1G, where values were compared to wild-type _A. thaliana_ values, with a


945 significance level of 0.05. Significant values are indicated in the graphs as follows: where


946 ns &gt; p= 0.05, *p &lt; 0.05, **p &lt; 0.01, ***p &lt; 0.001, ****p &lt; 0.0001. Graphs show the average


947 percentage of pollen tube overgrowth per pollen tube-entered ovule (OG/PEO) across all


948 siliques with pollen tube reception for each cross. Data were plotted using Prism 10


949 (https://www.graphpad.com/).


950 Control genotypes (Col-0, _lre-7_, and _lre-7_ complemented with _AtLRE_


951 ( _AtLRE_ pro: _AtLRE_ -cYFP)) were included in each experimental series; therefore, control


952 values reported in each figure/graph represent the contemporaneous controls measured


953 in the same experimental block as the corresponding test lines, rather than values pooled


954 across experiments.


955


956 **Identifying orthologs of** _**LRE**_ **and** _**LLG1**_


957 Genomes were obtained for _Arabidopsis thaliana,Brassica rapa_, _Carica papaya_,


958 _Capsella rubella_, _Camelina sativa_, _Cleome violacea_, _Descurainia sophioides_, _Eutrema_


959 _salsugineum_, _Euclidium syriacum_, _Gossypium raimondii_, _Lunaria annua_, _Lepidium_


960 _sativum_, _Malcolmia maritima_, _Myagrum perfoliatum_, _Rorippa islandica_, _Sinapis alba_,


961 _Sisymbrium irio_, _Schrenkiella parvula_, _Stanleya pinnata_, _Thlaspi arvense_, and


962 _Theobroma cacao_ from Phytozome, v13 (Goodstein et al. 2012); _Aethionema arabicum_


963 from _Ae. arabicum_ DB (Fernandez-Pozo et al. 2021); and for _Tarrenaya hassleriana_


964 (Cheng et al. 2013), _Bretschneidera sinensis_ (Liu et al. 2022), _Moringa oleifera_ (Chang et


965 al. 2022), _Capparis spinosa_ (Wang et al. 2022), and _Tropaeolum majus_ (Friedhoff et al.


966 2024) from their respective publications (Table S1). If available, coding sequence (CDS)


32


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


967 genomes were obtained. For _B. sinensis_ and _T. hassleriana_, whole-genome and


968 annotation files were used as input into GFFRead, v0.12.7, to generate the CDS genome,


969 using the command: gffread -g genome.fa annotation.gff3 -x cds.fa (Pertea and Pertea


970 2020). Alternative transcripts were removed using primary_transcript.py from Orthofinder,


971 v2.5.5 (Emms and Kelly 2015; Emms and Kelly 2019). CDS genomes were translated


972 into peptide genomes using SeqKit2, v2.9.0, with the command: seqkit translate cds.fa &gt;


973 pep.fa (Shen et al. 2016; Shen et al. 2024). Using previously published phylogenies of


974 Brassicaceae and Brassicales for reference (Hall et al. 2004; Beilstein et al. 2006;


975 Beilstein et al. 2008; Edger et al. 2018; Hendriks et al. 2023), a Newick tree, newick.txt,


976 was made for the selected species within Mesquite, v3.81 (Maddison and Maddison


977 2023). Homologs of the selected genomes were identified and assigned to Orthogroups


978 using Orthofinder, v2.5.5, with the command: orthofinder -f /path/to/peptide_genomes -s


979 newick.txt -y -X -M msa -t 64 (Emms and Kelly, 2015; Emms and Kelly, 2019). _LRE_ and


980 _LLG1_ homologues were found in two Orthogroups: one containing all Brassicaceae _LRE_


981 orthologs and the other containing all Brassicaceae _LLG1_ orthologs and non

982 Brassicaceae _LRE/LLG1_ homologues. Together, these two Orthogroups contained all


983 previously identified Brassicales and Malvales _LRE/LLG1_ homologues from Noble et al.


984 2022 in the species selected (Table S2).


985


986 _**LRE**_ **and** _**LLG1**_ **CDS alignments and phylogenies**


987 _LRE_ and _LLG1_ peptide sequences were aligned using the PPP algorithm in


988 MUSCLE, v5.2.linux64, using the command: muscle -align pep_sequences.fa -output


989 pep_alignment.afa -threads 20 (Edgar 2022). The peptide alignment was reverse

990 translated into a nucleotide alignment using PAL2NAL, v14, with the command: pal2nal.pl


991 pep_alignment.afa nuc_sequences.fa -output fasta &gt; nuc_alignment.afa (Suyama et al.


992 2006). The nucleotide alignment was then trimmed at the codon level with the minimum


993 block length set to 5 and allowing half of the gap positions and otherwise default


994 parameters using Gblocks, v0.91b, yielding trimmed_nuc_alignment.afa (Talavera and


995 Castresana 2007). Phylogeny of the _LRE_ and _LLG1_ homologues was then inferred in


996 RAxML-ng, v1.2.2-master, starting with 10 random trees and 10 parsimony trees using


997 the GTR+G model of evolution and 1000 Felsenstein bootstrap replicates with the


33


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


998 command: raxml-ng -all -msa trimmed_nuc_alignment.afa -model GTR+G -bs-trees 1000


999 (Kozlov et al. 2019). The best tree was then visualized using iTOL, v7 (Letunic and Bork


1000 2024) and rooted by the Malvales _LRE/LLG1_ clade, which served as an outgroup.


1001


1002 **PAML Analysis**


1003 PAML analyses were performed with PAML, v4.10.9 (Yang 2007), using the


1004 alignment, trimmed_nuc_alignment.afa, and best tree from RAxML-ng rooted by the


1005 Malvales _LRE/LLG1_ clade, as described above. The guidance of Álvarez-Carretero et al.


1006 2023 was followed for running the site, branch, and branch-site models. Alternative


1007 models were compared against their corresponding null model using the likelihood-ratio


1008 test (LRT). To test for multiple site classes, M1a (Nearly Neutral) was compared against


1009 M0 (one-ratio) (Goldman and Yang 1994; Nielsen and Yang 1998; Wong et al. 2004;


1010 Yang et al. 2005). The branch tests were performed four times, with different branches


1011 highlighted as the foreground: the post-duplication branch (PDB) or all clade branches,


1012 including the PDB, in either _LRE_ or _LLG1_ were highlighted as the foreground. For the


1013 branch tests, the alternative models were compared against M0 (Yang 1998; Yang and


1014 Nielsen 1998). For the Model C clade tests, all branches within the _LRE_ clade, including


1015 the PDB, were highlighted as foreground, and the alternative model was compared


1016 against M2a_rel (Bielawski and Yang 2004; Weadick and Chang 2012). The Model A


1017 branch-sites test was performed on either the PDB or the entire clade, including the PDB,


1018 for _LRE_ . For the Model A branch-site tests, the alternative model, where ω was set as a


1019 free parameter starting at a value of 0.5, was compared against the null model, where ω


1020 was fixed at a value of 1 (Yang et al. 2000; Yang and Nielsen 2002; Yang et al. 2005;


1021 Zhang et al. 2005). Posterior probabilities for site classes were determined using Bayes


1022 empirical Bayes (BEB) analysis (Yang et al. 2005). Results from PAML, including


1023 parameter estimates, likelihood scores, likelihood-ratio tests, and BEB results for Model


1024 A branch-site tests, are recorded in Table S5.


1025


1026 **Structural Modeling and Analysis**


1027 AlphaFold2 was used with 32 seeds to generate 160 models for analysis


1028 consistent with established protocols (Monteiro Da Silva et al. 2024; Galeazzi et al. 2025).


34


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


1029 Multiple Sequence Alignments for each variant were generated using the JackHMMR


1030 method and respective sequences as queries (Finn et al. 2011). After various


1031 subsampling conditions were tested by modifying the _max_seq_ and _extra_seq_ for WT and


1032 each AtLRE variant, a _max_seq_ : _extra_seq_ of 128:256 was selected for downstream


1033 analyses. Any structures with pTM &lt;50 and pLDDT &lt;70 were filtered out in accordance


1034 with established protocols (Jumper et al. 2021; Abramson et al. 2024).


1035 The predictions were then loaded as trajectories into MDAnalysis in Python 3.12.3


1036 (Michaud-Agrawal et al. 2011; Gowers et al. 2016). Structures were first superimposed


1037 and aligned to the conserved core region using the AlignTraj function in MDAnalysis. Root


1038 mean square deviation (RMSD) with an all-atom Cα selection was calculated relative to


1039 the top-ranked structure for each variant. Plots were generated using Matplotlib (Hunter


1040 2007). Visual rendering of proteins was generated using PyMOL (version 2.3.0).


1041


1042
### Acknowledgments


1043


1044 E.J. and M.J.A. were supported by funding from the U.S. National Science


1045 Foundation awarded to RP (IOS-1146090) and MAB (DBI-2019674), respectively, and by


1046 the School of Plant Sciences. In addition, M.J.A. received funding from the U.S. National


1047 Institutes of Health Initiative to Maximize Student Development (T32GM139779). J.A.N.


1048 was supported by the following: IGERT Comparative Genomics Program at the University


1049 of Arizona (Award ID: 0654435); NSF Graduate Research Fellowship: Grant DGE

1050 1143953; the Boynton Graduate Fellowship; and the University of Arizona Graduate


1051 College Office of Diversity and Inclusion. We thank Dr. Zuzana Gadiou and Dr. Vinod


1052 Kumar for assistance with interspecific crosses done at the Institute of Experimental


1053 Botany of the Czech Academy of Sciences. We thank Dr. Dean Billheimer, Zuckerman


1054 College of Public Health, University of Arizona, for advising us on statistical analysis of


1055 the data presented in this study. We thank Dr. Mark Johnson, Dr. Ueli Grossniklaus, Dr.


1056 Steven Smith, Dr. Ramin Yadegari, and Dr. Rebecca Mosher for helpful discussions.


1057


35


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

### 1058 References


1059
1060 Abramson, J., Adler, J., Dunger, J., Evans, R., Green, T., Pritzel, A., Ronneberger, O.,


1061 Willmore, L., Ballard, A. J., Bambrick, J., Bodenstein, S. W., Evans, D. A., Hung, C.-C.,


1062 O’Neill, M., Reiman, D., Tunyasuvunakool, K., Wu, Z., Žemgulytė, A., Arvaniti, E., …


1063 Jumper, J. M. (2024). Accurate structure prediction of biomolecular interactions with


1064 AlphaFold 3. _Nature_, _630_ (8016), 493–500. https://doi.org/10.1038/s41586-024-07487-w


1065 Álvarez-Carretero, S., Kapli, P., &amp; Yang, Z. (2023). Beginner’s Guide on the Use of


1066 PAML to Detect Positive Selection. _Molecular Biology and Evolution_, _40_ (4), msad041.


1067 https://doi.org/10.1093/molbev/msad041


1068 Amien, S., Kliwer, I., Márton, M. L., Debener, T., Geiger, D., et al. (2010). Defensin-like


1069 ZmES4 mediates pollen tube burst in maize via opening of the potassium channel


1070 KZM1. _PLOS Biol_ . 8:e1000388https://doi.org/10.1371/journal.pbio.1000388


1071 Baack, E., Melo, M. C., Rieseberg, L. H., &amp; Ortiz-Barrientos, D. (2015). The origins of


1072 reproductive isolation in plants. _New Phytologist_, _207_ (4), 968–984.


1073 https://doi.org/10.1111/nph.13424


1074 Beilstein, M. A., Al-Shehbaz, I. A., &amp; Kellogg, E. A. (2006). Brassicaceae phylogeny and


1075 trichome evolution. _American Journal of Botany_, _93_ (4), 607–619.


1076 https://doi.org/10.3732/ajb.93.4.607


1077 Beilstein, M. A., Al-Shehbaz, I. A., Mathews, S., &amp; Kellogg, E. A. (2008). Brassicaceae


1078 phylogeny inferred from phytochrome A and _ndhF_ sequence data: Tribes and trichomes


1079 revisited. _American Journal of Botany_, _95_ (10), 1307–1327.


1080 https://doi.org/10.3732/ajb.0800065


1081 Bielawski, JosephP., &amp; Yang, Z. (2004). A Maximum Likelihood Method for Detecting


1082 Functional Divergence at Individual Codon Sites, with Application to Gene Family


1083 Evolution. _Journal of Molecular Evolution_, _59_ (1). https://doi.org/10.1007/s00239-004

1084 2597-8


36


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


1085 Cao, Y., Cui, X., Yang, Y., Pan, L., Yang, F., Li, S., Wu, D., Ding, Y., Chen, R., Wang,


1086 N., et al. (2025). Pan-family pollen signals control an interspecific stigma barrier across


1087 Brassicaceae species. _Science_ 391, Article eady2347. https://10.1126/science.ady2347


1088 Capron, A., Gourgues, M., Neiva, L. S., Faure, J.-E., Berger, F., Pagnussat, G.,


1089 Krishnan, A., Alvarez-Mejia, C., Vielle-Calzada, J.-P., Lee, Y.-R., Liu, B., &amp; Sundaresan,


1090 V. (2008). Maternal Control of Male-Gamete Delivery in _Arabidopsis_ Involves a Putative


1091 GPI-Anchored Protein Encoded by the _LORELEI_ Gene. _The Plant Cell_, _20_ (11), 3038–


1092 3049. https://doi.org/10.1105/tpc.108.061713


1093 Chang, J., Marczuk-Rojas, J. P., Waterman, C., Garcia-Llanos, A., Chen, S., Ma, X.,


1094 Hulse-Kemp, A., Van Deynze, A., Van De Peer, Y., &amp; Carretero-Paulet, L. (2022).


1095 Chromosome-scale assembly of the _Moringa oleifera_ Lam. Genome uncovers polyploid


1096 history and evolution of secondary metabolism pathways through tandem duplication.


1097 _The Plant Genome_, _15_ (3), e20238. https://doi.org/10.1002/tpg2.20238


1098 Cheng, S., Van Den Bergh, E., Zeng, P., Zhong, X., Xu, J., Liu, X., Hofberger, J., De


1099 Bruijn, S., Bhide, A. S., Kuelahoglu, C., Bian, C., Chen, J., Fan, G., Kaufmann, K., Hall,


1100 J. C., Becker, A., Bräutigam, A., Weber, A. P. M., Shi, C., … Schranz, M. E. (2013). The


1101 _Tarenaya hassleriana_ Genome Provides Insight into Reproductive Trait and Genome


1102 Evolution of Crucifers. _The Plant Cell_, _25_ (8), 2813–2830.


1103 https://doi.org/10.1105/tpc.113.113480


1104 Chung, M.-H., Chen, M.-K., &amp; Pan, S.-M. (2000). Floral Spray Transformation Can


1105 Efficiently Generate Arabidopsis. _Transgenic Research_, _9_ (6), 471–486.


1106 https://doi.org/10.1023/A:1026522104478


1107 Dellaporta, S. L., Wood, J., &amp; Hicks, J. B. (1983). A plant DNA minipreparation: Version


1108 II. _Plant Molecular Biology Reporter_, _1_ (4), 19–21. https://doi.org/10.1007/BF02712670


1109 Dong, L., Meng, Y., Sui, Z., Wang, J., Wu, L., &amp; Fu, B. (2015). Comparison of four


1110 digital PCR platforms for accurate quantification of DNA copy number of a certified


1111 plasmid DNA reference material. _Scientific Reports_, _5_ (1), 13174.


1112 https://doi.org/10.1038/srep13174


37


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


1113 Edgar, R. C. (2022). Muscle5: High-accuracy alignment ensembles enable unbiased


1114 assessments of sequence homology and phylogeny. _Nature Communications_, _13_ (1),


1115 6968. https://doi.org/10.1038/s41467-022-34630-w


1116 Edger, P. P., Hall, J. C., Harkess, A., Tang, M., Coombs, J., Mohammadin, S., Schranz,


1117 M. E., Xiong, Z., Leebens-Mack, J., Meyers, B. C., Sytsma, K. J., Koch, M. A., Al

1118 Shehbaz, I. A., &amp; Pires, J. C. (2018). Brassicales phylogeny inferred from 72 plastid


1119 genes: A reanalysis of the phylogenetic localization of two paleopolyploid events and


1120 origin of novel chemical defenses. _American Journal of Botany_, _105_ (3), 463–469.


1121 https://doi.org/10.1002/ajb2.1040


1122 Emms, D. M., &amp; Kelly, S. (2015). OrthoFinder: Solving fundamental biases in whole


1123 genome comparisons dramatically improves orthogroup inference accuracy. _Genome_


1124 _Biology_, _16_ (1), 157. https://doi.org/10.1186/s13059-015-0721-2


1125 Emms, D. M., &amp; Kelly, S. (2019). OrthoFinder: Phylogenetic orthology inference for


1126 comparative genomics. _Genome Biology_, _20_ (1), 238. https://doi.org/10.1186/s13059

1127 019-1832-y


1128 Escobar-Restrepo, J.-M., Huck, N., Kessler, S., Gagliardini, V., Gheyselinck, J., Yang,


1129 W.-C., &amp; Grossniklaus, U. (2007). The FERONIA Receptor-like Kinase Mediates Male

1130 Female Interactions During Pollen Tube Reception. _Science_, _317_ (5838), 656–660.


1131 https://doi.org/10.1126/science.1143562


1132 Feng, H., Liu, C., Fu, R., Zhang, M., Li, H., Shen, L., Wei, Q., Sun, X., Xu, L., Ni, B., &amp;


1133 Li, C. (2019). LORELEI-LIKE GPI-ANCHORED PROTEINS 2/3 Regulate Pollen Tube


1134 Growth as Chaperones and Coreceptors for ANXUR/BUPS Receptor Kinases in


1135 Arabidopsis. _Molecular Plant_, _12_ (12), 1612–1623.


1136 https://doi.org/10.1016/j.molp.2019.09.004


1137 Fernandez-Pozo, N., Metz, T., Chandler, J. O., Gramzow, L., Mérai, Z., Maumus, F.,


1138 Mittelsten Scheid, O., Theißen, G., Schranz, M. E., Leubner-Metzger, G., &amp; Rensing, S.


1139 A. (2021). Aethionema arabicum genome annotation using PacBio full-length transcripts


1140 provides a valuable resource for seed dormancy and Brassicaceae evolution research.


38


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


1141 _The Plant Journal: For Cell and Molecular Biology_, _106_ (1), 275–293.


1142 https://doi.org/10.1111/tpj.15161


1143 Finn, R. D., Clements, J., &amp; Eddy, S. R. (2011). HMMER web server: Interactive


1144 sequence similarity searching. _Nucleic Acids Research_, _39_ (suppl), W29–W37.


1145 https://doi.org/10.1093/nar/gkr367


1146 Fleming, J., Magana, P., Nair, S., Tsenkov, M., Bertoni, D., Pidruchna, I., Lima Afonso,


1147 M. Q., Midlik, A., Paramval, U., Žídek, A., Laydon, A., Kovalevskiy, O., Pan, J., Cheng,


1148 J., Avsec, Ž., Bycroft, C., Wong, L. H., Last, M., Mirdita, M., … Velankar, S. (2025).


1149 AlphaFold Protein Structure Database and 3D-Beacons: New Data and Capabilities.


1150 _Journal of Molecular Biology_, _437_ (15), 168967.


1151 https://doi.org/10.1016/j.jmb.2025.168967


1152 Friedhoff, R., Wittstock, U., &amp; Pucker, B. (2024). _Genome sequence of the medicinal_


1153 _plant_ Tropaeolum majus _provides insights into flavonoid biosynthesis_ . Plant Biology.


1154 https://doi.org/10.1101/2024.11.01.621629


1155 Fujii, S., Tsuchimatsu, T., Kimura, Y., Ishida, S., Tangpranomkorn, S., Shimosato

1156 Asano, H., Iwano, M., Furukawa, S., Itoyama, W., Wada, Y., Shimizu, K. K., &amp;


1157 Takayama, S. (2019). A stigmatic gene confers interspecies incompatibility in the


1158 Brassicaceae. _Nature Plants_, _5_ (7), 731–741. https://doi.org/10.1038/s41477-019-0444-6


1159 Fujii, S., Yamamoto, E., Ito, S., Tangpranomkorn, S., Kimura, Y., et al. (2023). SHI


1160 family transcription factors regulate an interspecific barrier. _Nat. Plants_ 9, 1862–73.


1161 https://doi.org/10.1038/s41477-023-01535-5


1162 Galeazzi, F. M., Monteiro Da Silva, G., Arantes, P., Varghese, I., Shukla, A., &amp;


1163 Rubenstein, B. M. (2025). _FastConformation: A Standalone ML-Based Toolkit for_


1164 _Modeling and Analyzing Protein Conformational Ensembles at Scale_ . Biophysics.


1165 https://doi.org/10.1101/2025.05.09.653048


1166 Galindo-Trigo, S., Blanco-Touriñán, N., DeFalco, T. A., Wells, E. S., Gray, J. E., Zipfel,


1167 C., &amp; Smith, L. M. (2020). _Cr_ RLK1L receptor-like kinases HERK1 and ANJEA are


39


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


1168 female determinants of pollen tube reception. _EMBO Reports_, _21_ (2).


1169 https://doi.org/10.15252/embr.201948466


1170 Gao, Q., Wang, C., Xi, Y., Shao, Q., Li, L., &amp; Luan, S. (2022). A receptor–channel trio


1171 conducts Ca2+ signalling for pollen tube reception. _Nature_, _607_ (7919), 534–539.


1172 https://doi.org/10.1038/s41586-022-04923-7


1173 Ge, Z., Bergonci, T., Zhao, Y., Zou, Y., Du, S., Liu, M.-C., Luo, X., Ruan, H., García

1174 Valencia, L. E., Zhong, S., Hou, S., Huang, Q., Lai, L., Moura, D. S., Gu, H., Dong, J.,


1175 Wu, H.-M., Dresselhaus, T., Xiao, J., … Qu, L.-J. (2017). _Arabidopsis_ pollen tube


1176 integrity and sperm release are regulated by RALF-mediated signaling. _Science_,


1177 _358_ (6370), 1596–1600. https://doi.org/10.1126/science.aao3642


1178 Ge, Z., Zhao, Y., Liu, M.-C., Zhou, L.-Z., Wang, L., Zhong, S., Hou, S., Jiang, J., Liu, T.,


1179 Huang, Q., Xiao, J., Gu, H., Wu, H.-M., Dong, J., Dresselhaus, T., Cheung, A. Y., &amp; Qu,


1180 L.-J. (2019). LLG2/3 Are Co-receptors in BUPS/ANX-RALF Signaling to Regulate


1181 Arabidopsis Pollen Tube Integrity. _Current Biology_, _29_ (19), 3256-3265.e5.


1182 https://doi.org/10.1016/j.cub.2019.08.032


1183 Goldman, N., &amp; Yang, Z. (1994). A codon-based model of nucleotide substitution for


1184 protein-coding DNA sequences. _Molecular Biology and Evolution_ .


1185 https://doi.org/10.1093/oxfordjournals.molbev.a040153


1186 Goodstein, D. M., Shu, S., Howson, R., Neupane, R., Hayes, R. D., Fazo, J., Mitros, T.,


1187 Dirks, W., Hellsten, U., Putnam, N., &amp; Rokhsar, D. S. (2012). Phytozome: A


1188 comparative platform for green plant genomics. _Nucleic Acids Research_, _40_ (D1),


1189 D1178–D1186. https://doi.org/10.1093/nar/gkr944


1190 Gowers, R., Linke, M., Barnoud, J., Reddy, T., Melo, M., Seyler, S., Domański, J.,


1191 Dotson, D., Buchoux, S., Kenney, I., &amp; Beckstein, O. (2016). _MDAnalysis: A Python_


1192 _Package for the Rapid Analysis of Molecular Dynamics Simulations_ . 98–105.


1193 https://doi.org/10.25080/Majora-629e541a-00e


40


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


1194 Hall, J. C., Iltis, H. H., &amp; Sytsma, K. J. (2004). Molecular Phylogenetics of Core


1195 Brassicales, Placement of Orphan Genera Emblingia, Forchhammeria, Tirania, and


1196 Character Evolution. _Systematic Botany_, _29_ (3), 654–669. JSTOR.


1197 Harrison, S. J., Mott, E. K., Parsley, K., Aspinall, S., Gray, J. C., &amp; Cottage, A. (2006). A


1198 rapid and robust method of identifying transformed Arabidopsis thaliana seedlings


1199 following floral dip transformation. _Plant Methods_, _2_ (1), 19. https://doi.org/10.1186/1746

1200 4811-2-19


1201 Hendriks, K. P., Kiefer, C., Al-Shehbaz, I. A., Bailey, C. D., Hooft Van Huysduynen, A.,


1202 Nikolov, L. A., Nauheimer, L., Zuntini, A. R., German, D. A., Franzke, A., Koch, M. A.,


1203 Lysak, M. A., Toro-Núñez, Ó., Özüdoğru, B., Invernón, V. R., Walden, N., Maurin, O.,


1204 Hay, N. M., Shushkov, P., … Lens, F. (2023). Global Brassicaceae phylogeny based on


1205 filtering of 1,000-gene dataset. _Current Biology_, _33_ (19), 4052-4068.e6.


1206 https://doi.org/10.1016/j.cub.2023.08.026


1207 Higashiyama, T., Inatsugi, R., Sakamoto, S., Sasaki, N., Mori, T., Kuroiwa, H., Nakada,


1208 T., Nozaki, H., Kuroiwa, T., &amp; Nakano, A. (2006). Species Preferentiality of the Pollen


1209 Tube Attractant Derived from the Synergid Cell of _Torenia fournieri_ . _Plant Physiology_,


1210 _142_ (2), 481–491. https://doi.org/10.1104/pp.106.083832


1211 Hou, Y.,Guo, X., Cyprys, P., Zhang, Y., Bleckmann, A., et al. (2016). Maternal ENODLs


1212 are required for pollen tube reception in Arabidopsis. _Curr. Biol._ 26:2343–50


1213 Huang, J., Yang, L., Yang, L., Wu, X., Cui, X., Zhang, L., Hui, J., Zhao, Y., Yang, H.,


1214 Liu, S., Xu, Q., Pang, M., Guo, X., Cao, Y., Chen, Y., Ren, X., Lv, J., Yu, J., Ding, J., …


1215 Duan, Q. (2023). Stigma receptors control intraspecies and interspecies barriers in


1216 Brassicaceae. _Nature_, _614_ (7947), 303–308. https://doi.org/10.1038/s41586-022-05640

1217 x


1218 Hunter, J. D. (2007). Matplotlib: A 2D Graphics Environment. _Computing in Science &amp;_


1219 _Engineering_, _9_ (3), 90–95. https://doi.org/10.1109/MCSE.2007.55


41


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


1220 Ju, Y., Yuan, J., Jones, D. S., Zhang, W., Staiger, C. J., &amp; Kessler, S. A. (2021).


1221 Polarized NORTIA accumulation in response to pollen tube arrival at synergids


1222 promotes fertilization. _Developmental Cell_, _56_ (21), 2938-2951.e6.


1223 https://doi.org/10.1016/j.devcel.2021.09.026


1224 Jumper, J., Evans, R., Pritzel, A., Green, T., Figurnov, M., Ronneberger, O.,


1225 Tunyasuvunakool, K., Bates, R., Žídek, A., Potapenko, A., Bridgland, A., Meyer, C.,


1226 Kohl, S. A. A., Ballard, A. J., Cowie, A., Romera-Paredes, B., Nikolov, S., Jain, R.,


1227 Adler, J., … Hassabis, D. (2021). Highly accurate protein structure prediction with


1228 AlphaFold. _Nature_, _596_ (7873), 583–589. https://doi.org/10.1038/s41586-021-03819-2


1229 Kozlov, A. M., Darriba, D., Flouri, T., Morel, B., &amp; Stamatakis, A. (2019). RAxML-NG: A


1230 fast, scalable and user-friendly tool for maximum likelihood phylogenetic inference.


1231 _Bioinformatics_, _35_ (21), 4453–4455. https://doi.org/10.1093/bioinformatics/btz305


1232 Lan, Z., Song, Z., Wang, Z., Li, L., Liu, Y., Zhi, S., Wang, R., Wang, J., Li, Q.,


1233 Bleckmann, A., Zhang, L., Dresselhaus, T., Dong, J., Gu, H., Zhong, S., &amp; Qu, L.-J.


1234 (2023). Antagonistic RALF peptides control an intergeneric hybridization barrier on


1235 Brassicaceae stigmas. _Cell_, _186_ (22), 4773-4787.e12.


1236 https://doi.org/10.1016/j.cell.2023.09.003


1237 Letunic, I., &amp; Bork, P. (2024). Interactive Tree of Life (iTOL) v6: Recent updates to the


1238 phylogenetic tree display and annotation tool. _Nucleic Acids Research_, _52_ (W1), W78–


1239 W82. https://doi.org/10.1093/nar/gkae268


1240 Leydon, A. R.,Tsukamoto, T., Dunatunga, D., Qin, Y., Johnson, M. A., Palanivelu, R.


1241 (2015). Pollen tube discharge completes the process of synergid degeneration that is


1242 initiated by pollen tube–synergid interaction in Arabidopsis. _Plant Physiol_ . 169:485–96.


1243 https://doi.org/10.1104/pp.15.00528


1244 Leydon, A. R., Beale, K. M., Woroniecka, K., Castner, E., Chen, J., Horgan, C.,


1245 Palanivelu, R., &amp; Johnson, M. A. (2013). Three MYB transcription factors control pollen


1246 tube differentiation required for sperm release. _Current Biology: CB_, _23_ (13), 1209–1214.


1247 https://doi.org/10.1016/j.cub.2013.05.021


42


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


1248 Liu, C., Shen, L., Xiao, Y., Vyshedsky, D., Peng, C., Sun, X., Liu, Z., Cheng, L., Zhang,


1249 H., Han, Z., Chai, J., Wu, H.-M., Cheung, A. Y., &amp; Li, C. (2021). Pollen PCP-B peptides


1250 unlock a stigma peptide–receptor kinase gating mechanism for pollination. _Science_,


1251 _372_ (6538), 171–175. https://doi.org/10.1126/science.abc6107


1252 Liu, H.-L., Harris, A. J., Wang, Z.-F., Chen, H.-F., Li, Z.-A., &amp; Wei, X. (2022). The


1253 genome of the Paleogene relic tree _Bretschneidera sinensis_ : Insights into trade-offs in


1254 gene family evolution, demographic history, and adaptive SNPs. _DNA Research_, _29_ (1),


1255 dsac003. https://doi.org/10.1093/dnares/dsac003


1256 Liu, X., Castro, C., Wang, Y., Noble, J., Ponvert, N., Bundy, M., Hoel, C., Shpak, E., &amp;


1257 Palanivelu, R. (2016). The Role of LORELEI in Pollen Tube Reception at the Interface


1258 of the Synergid Cell and Pollen Tube Requires the Modified Eight-Cysteine Motif and


1259 the Receptor-Like Kinase FERONIA. _The Plant Cell_, _28_ (5), 1035–1052.


1260 https://doi.org/10.1105/tpc.15.00703


1261 Liu, Z., Chu, X., Ren, W., Cheng, L., Liu, C., Wang, C., Gao, S., Dai, S., Li, C. 2025.


1262 PCP-B peptides and CrRLK1L receptor kinases control pollination via pH gating of


1263 aquaporins in Arabidopsis. _Developmental Cell_, 60, 1336–1347.


1264 https://doi.org/10.1016/j.devcel.2024.12.026


1265 Maddison, W. P., &amp; Maddison, D. R. (n.d.). _Mesquite: A modular system for evolutionary_


1266 _analysis_ (Version 3.812023) &#91;Computer software&#93;. Retrieved


1267 http://www.mesquiteproject.org


1268 Márton, M. L., Fastner, A., Uebler, S., &amp; Dresselhaus, T. (2012). Overcoming


1269 Hybridization Barriers by the Secretion of the Maize Pollen Tube Attractant ZmEA1 from


1270 Arabidopsis Ovules. _Current Biology_, _22_ (13), 1194–1198.


1271 https://doi.org/10.1016/j.cub.2012.04.061


1272 Michaud-Agrawal, N., Denning, E. J., Woolf, T. B., &amp; Beckstein, O. (2011). MDAnalysis:


1273 A toolkit for the analysis of molecular dynamics simulations. _Journal of Computational_


1274 _Chemistry_, _32_ (10), 2319–2327. https://doi.org/10.1002/jcc.21787


43


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


1275 Monteiro Da Silva, G., Cui, J. Y., Dalgarno, D. C., Lisi, G. P., &amp; Rubenstein, B. M.


1276 (2024). High-throughput prediction of protein conformational distributions with


1277 subsampled AlphaFold2. _Nature Communications_, _15_ (1), 2464.


1278 https://doi.org/10.1038/s41467-024-46715-9


1279 Mori, T., Kuroiwa, H., Higashiyama, T., &amp; Kuroiwa, T. (2006). GENERATIVE CELL


1280 SPECIFIC 1 is essential for angiosperm fertilization. _Nature Cell Biology_, _8_ (1), 64–71.


1281 https://doi.org/10.1038/ncb1345


1282 Müller, L. M., Lindner, H., Pires, N. D., Gagliardini, V., &amp; Grossniklaus, U. (2016). A


1283 subunit of the oligosaccharyltransferase complex is required for interspecific


1284 gametophyte recognition in Arabidopsis. _Nature Communications_, _7_ (1), 10826.


1285 https://doi.org/10.1038/ncomms10826


1286 Nielsen, R., &amp; Yang, Z. (1998). Likelihood Models for Detecting Positively Selected


1287 Amino Acid Sites and Applications to the HIV-1 Envelope Gene. _Genetics_, _148_ (3), 929–


1288 936. https://doi.org/10.1093/genetics/148.3.929


1289 Noble, J. A., Bielski, N. V., Liu, M.-C. J., DeFalco, T. A., Stegmann, M., Nelson, A. D. L.,


1290 McNamara, K., Sullivan, B., Dinh, K. K., Khuu, N., Hancock, S., Shiu, S.-H., Zipfel, C.,


1291 Cheung, A. Y., Beilstein, M. A., &amp; Palanivelu, R. (2022). Evolutionary analysis of the


1292 _LORELEI_ gene family in plants reveals regulatory subfunctionalization. _Plant_


1293 _Physiology_, _190_ (4), 2539–2556. https://doi.org/10.1093/plphys/kiac444


1294 Noble, J. A., &amp; Palanivelu, R. (2020). Workflow to Characterize Mutants with


1295 Reproductive Defects. In A. Geitmann (Ed.), _Pollen and Pollen Tube Biology_ (Vol. 2160,


1296 pp. 109–128). Springer US. https://doi.org/10.1007/978-1-0716-0672-8_8


1297 Panchy, N., Lehti-Shiu, M., &amp; Shiu, S.-H. (2016). Evolution of Gene Duplication in


1298 Plants. _Plant Physiology_, _171_ (4), 2294–2316. https://doi.org/10.1104/pp.16.00523


1299 Pertea, G., &amp; Pertea, M. (2020). GFF Utilities: GffRead and GffCompare.


1300 _F1000Research_, _9_, 304. https://doi.org/10.12688/f1000research.23297.2


44


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


1301 Ponvert, N., &amp; Johnson, M. A. (2024). Synergid cell calcium oscillations refine


1302 understanding of FERONIA/LORELEI signaling during interspecific hybridization. _Plant_


1303 _Reproduction_, _37_ (1), 57–68. https://doi.org/10.1007/s00497-023-00483-6


1304 Ramsey, J., &amp; Schemske, D. W. (1998). PATHWAYS, MECHANISMS, AND RATES OF


1305 POLYPLOID FORMATION IN FLOWERING PLANTS. _Annual Review of Ecology and_


1306 _Systematics_, _29_ (1), 467–501. https://doi.org/10.1146/annurev.ecolsys.29.1.467


1307 Sandaklie-Nikolova, L., Palanivelu, R., King, E. J., Copenhaver, G. P., Drews, G. N.


1308 (2007). Synergid cell death in Arabidopsis is triggered following direct interaction with


1309 the pollen tube. _Plant Physiol._ 144:1753–62. https://doi.org/10.1104/pp.107.098236


1310 Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T.,


1311 Preibisch, S., Rueden, C., Saalfeld, S., Schmid, B., Tinevez, J.-Y., White, D. J.,


1312 Hartenstein, V., Eliceiri, K., Tomancak, P., &amp; Cardona, A. (2012). Fiji: An open-source


1313 platform for biological-image analysis. _Nature Methods_, _9_ (7), 676–682.


1314 https://doi.org/10.1038/nmeth.2019


1315 Shen, W., Le, S., Li, Y., &amp; Hu, F. (2016). SeqKit: A Cross-Platform and Ultrafast Toolkit


1316 for FASTA/Q File Manipulation. _PLOS ONE_, _11_ (10), e0163962.


1317 https://doi.org/10.1371/journal.pone.0163962


1318 Shen, W., Sipos, B., &amp; Zhao, L. (2024). SeqKit2: A Swiss army knife for sequence and


1319 alignment processing. _iMeta_, _3_ (3), e191. https://doi.org/10.1002/imt2.191


1320 Smyth, D. R., Bowman, J. L., &amp; Meyerowitz, E. M. (1990). Early flower development in


1321 Arabidopsis. _The Plant Cell_, _2_ (8), 755–767. https://doi.org/10.1105/tpc.2.8.755


1322 Suyama, M., Torrents, D., &amp; Bork, P. (2006). PAL2NAL: Robust conversion of protein


1323 sequence alignments into the corresponding codon alignments. _Nucleic Acids_


1324 _Research_, _34_ (Web Server), W609–W612. https://doi.org/10.1093/nar/gkl315


1325 Takeuchi, H., &amp; Higashiyama, T. (2012). A Species-Specific Cluster of Defensin-Like


1326 Genes Encodes Diffusible Pollen Tube Attractants in Arabidopsis. _PLoS Biology_,


1327 _10_ (12), e1001449. https://doi.org/10.1371/journal.pbio.1001449


45


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


1328 Takeuchi, H., &amp; Higashiyama, T. (2016). Tip-localized receptors control pollen tube


1329 growth and LURE sensing in Arabidopsis. _Nature_, _531_ (7593), 245–248.


1330 https://doi.org/10.1038/nature17413


1331 Talavera, G., &amp; Castresana, J. (2007). Improvement of Phylogenies after Removing


1332 Divergent and Ambiguously Aligned Blocks from Protein Sequence Alignments.


1333 _Systematic Biology_, _56_ (4), 564–577. https://doi.org/10.1080/10635150701472164


1334 Tovar-Méndez, A., Lu, L., &amp; McClure, B. (2017). HT proteins contribute to S- RN ase

1335 independent pollen rejection in _Solanum_ . _The Plant Journal_, _89_ (4), 718–729.


1336 https://doi.org/10.1111/tpj.13416


1337 Tsukamoto, T., Qin, Y., Huang, Y., Dunatunga, D., &amp; Palanivelu, R. (2010). A role for


1338 LORELEI, a putative glycosylphosphatidylinositol-anchored protein, in Arabidopsis


1339 thaliana double fertilization and early seed development: LORELEI functions in


1340 fertilization and seed development. _The Plant Journal_, _62_ (4), 571–588.


1341 https://doi.org/10.1111/j.1365-313X.2010.04177.x


1342 Wang, L., Fan, L., Zhao, Z., Zhang, Z., Jiang, L., Chai, M., &amp; Tian, C. (2022). The


1343 Capparis spinosa var. Herbacea genome provides the first genomic instrument for a


1344 diversity and evolution study of the Capparaceae family. _GigaScience_, _11_, giac106.


1345 https://doi.org/10.1093/gigascience/giac106


1346 Wang, L., &amp; Filatov, D. A. (2023). Mechanisms of prezygotic post-pollination


1347 reproductive barriers in plants. _Frontiers in Plant Science_, _14_, 1230278.


1348 https://doi.org/10.3389/fpls.2023.1230278


1349 Weadick, C. J., &amp; Chang, B. S. W. (2012). An Improved Likelihood Ratio Test for


1350 Detecting Site-Specific Functional Divergence among Clades of Protein-Coding Genes.


1351 _Molecular Biology and Evolution_, _29_ (5), 1297–1300.


1352 https://doi.org/10.1093/molbev/msr311


1353 Williams, E. G., Knox, R. B., &amp; Rouse, J. L. (1982). Pollination sub-systems


1354 distinguished by pollen tube arrest after incompatible interspecific crosses in


46


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


1355 _Rhododendron_ (Ericaceae). _Journal of Cell Science_, _53_ (1), 255–277.


1356 https://doi.org/10.1242/jcs.53.1.255


1357 Wong, W. S. W., Yang, Z., Goldman, N., &amp; Nielsen, R. (2004). Accuracy and Power of


1358 Statistical Methods for Detecting Adaptive Evolution in Protein Coding Sequences and


1359 for Identifying Positively Selected Sites. _Genetics_, _168_ (2), 1041–1051.


1360 https://doi.org/10.1534/genetics.104.031153


1361 Xiao, Y., Stegmann, M., Han, Z., DeFalco, T. A., Parys, K., Xu, L., Belkhadir, Y., Zipfel,


1362 C., &amp; Chai, J. (2019). Mechanisms of RALF peptide perception by a heterotypic receptor


1363 complex. _Nature_, _572_ (7768), 270–274. https://doi.org/10.1038/s41586-019-1409-7


1364 Yang, Z. (1998). Likelihood ratio tests for detecting positive selection and application to


1365 primate lysozyme evolution. _Molecular Biology and Evolution_, _15_ (5), 568–573.


1366 https://doi.org/10.1093/oxfordjournals.molbev.a025957


1367 Yang, Z. (2005). Bayes Empirical Bayes Inference of Amino Acid Sites Under Positive


1368 Selection. _Molecular Biology and Evolution_, _22_ (4), 1107–1118.


1369 https://doi.org/10.1093/molbev/msi097


1370 Yang, Z. (2007). PAML 4: Phylogenetic Analysis by Maximum Likelihood. _Molecular_


1371 _Biology and Evolution_, _24_ (8), 1586–1591. https://doi.org/10.1093/molbev/msm088


1372 Yang, Z., &amp; Nielsen, R. (1998). Synonymous and nonsynonymous rate variation in


1373 nuclear genes of mammals. _Journal of Molecular Evolution_, _46_ (4), 409–418.


1374 https://doi.org/10.1007/PL00006320


1375 Yang, Z., &amp; Nielsen, R. (2002). Codon-Substitution Models for Detecting Molecular


1376 Adaptation at Individual Sites Along Specific Lineages. _Molecular Biology and Evolution_,


1377 _19_ (6), 908–917. https://doi.org/10.1093/oxfordjournals.molbev.a004148


1378 Yang, Z., Nielsen, R., Goldman, N., &amp; Pedersen, A.-M. K. (2000). Codon-Substitution


1379 Models for Heterogeneous Selection Pressure at Amino Acid Sites. _Genetics_, _155_ (1),


1380 431–449. https://doi.org/10.1093/genetics/155.1.431


47


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.18.767685; this version posted November 20, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.18.767685)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


1381 Zhang, J., Nielsen, R., &amp; Yang, Z. (2005). Evaluation of an Improved Branch-Site


1382 Likelihood Method for Detecting Positive Selection at the Molecular Level. _Molecular_


1383 _Biology and Evolution_, _22_ (12), 2472–2479. https://doi.org/10.1093/molbev/msi237


1384 Zhang, L., Huang, J., Su, S., Wei, X., Yang, L., Zhao, H., Yu, J., Wang, J., Hui, J., Hao,


1385 S., Song, S., Cao, Y., Wang, M., Zhang, X., Zhao, Y., Wang, Z., Zeng, W., Wu, H.-M.,


1386 Yuan, Y., … Duan, Q. (2021). FERONIA receptor kinase-regulated reactive oxygen


1387 species mediate self-incompatibility in Brassica rapa. _Current Biology_, _31_ (14), 3004

1388 3016.e4. https://doi.org/10.1016/j.cub.2021.04.060


1389 Zhang T., Li S., Dou S., Duan Q. (2026). From recognition to response: integrated


1390 signaling pathways determining pollen acceptance and rejection in Brassicaceae. _New_


1391 _Phytologist_, 250, 788–798. https://doi.org/10.1111/nph.70991


1392 Zhong, S., Li, L., Wang, Z., Ge, Z., Li, Q., Bleckmann, A., Wang, J., Song, Z., Shi, Y.,


1393 Liu, T., Li, L., Zhou, H., Wang, Y., Zhang, L., Wu, H.-M., Lai, L., Gu, H., Dong, J.,


1394 Cheung, A. Y., … Qu, L.-J. (2022). RALF peptide signaling controls the polytubey block


1395 in _Arabidopsis_ . _Science_, _375_ (6578), 290–296. https://doi.org/10.1126/science.abl4683


1396 Zhong, S., Liu, M., Wang, Z., Huang, Q., Hou, S., Xu, Y.-C., Ge, Z., Song, Z., Huang, J.,


1397 Qiu, X., Shi, Y., Xiao, J., Liu, P., Guo, Y.-L., Dong, J., Dresselhaus, T., Gu, H., &amp; Qu, L.

1398 J. (2019). Cysteine-rich peptides promote interspecific genetic isolation in _Arabidopsis_ .


1399 _Science_, _364_ (6443), eaau9564. https://doi.org/10.1126/science.aau9564


1400


48


#### **Figure 1 (legend in the next page)**


**Figure** **1.** Visualization of progamic phase and associated defects, and quantification
of pollen tube overgrowth mutant crosses to interspecific pollen
(A) Pollen tube growth in an _A._ _thaliana_ pistil visualized by staining it with aniline
blue. (Left) Diagram of _A._ _thaliana_ pistil with pollen tube shown in red and different
parts of the pistil are marked. (Right) An epifluorescent image of an aniline bluestained pollinated pistil. Scale bar= 0.5 mm.

(B and C) Diagram (B) and an aniline-blue stained epifluorescent image of pollen
tube reception by an _A._ _thaliana_ ovule (C, type 4). In the diagram (B), the pollen tube
is marked in red and the persistent synergid cell is marked in green. The receptive
synergid cell is filled with red, pointing to the successful discharge of the pollen tube
cytoplasm into it and marking the completion of pollen tube reception by that ovule.
In C, successful pollen tube reception and burst is inferred when the callose-rich
pollen tube enters the ovule at the micropylar end and ceases growth in the area of
the ovule proximal to the micropyle just after it enters the ovule. Neither the pollen
tube cytoplasm or sperm cells are stained by DABS and so the bolus left by the burst
pollen tube in the receptive synergid cell is not observed in this assay.
(D-F) Defects in pollen tube-ovule interactions. Abnormal pollen tube growth can be
divided into three other categories: the pollen tube fails to exit the transmitting tract
and target an ovule (D, type 1), the pollen tube targets an ovule but does not enter
the ovule (E, type 2), and the pollen tube targets the ovule and enters but does not
stop to burst at the synergid cells; instead, it continues to grow and coils inside the
female gametophyte (F, type 3). (C-F) Scale bar= 0.01 mm.
(G) Quantification of the average pollen tube overgrowth per pollen tube-entered
ovule (%OG/PEO) in crosses between female parents listed on the X-axis and _A._
_thaliana,_ _A._ _lyrata,_ _or_ _S._ _irio_ pollen. The average % OG/PEO for each female was
compared to the % OG/PEO of wild type Columbia (Col-0) ecotype. % OG/PEO
indicates the percentage of ovules in a silique that experienced pollen tube
overgrowth in the female gametophyte out of all the ovules that received a pollen
tube. The number of pistils crossed, and ovules scored is listed for each female
parent. Statistical significance was determined by first using an arcsine
transformation of percentage data, followed by pairwise _t-_ tests between Col-0
and tested mutants; ns&gt; 0.05, *p&lt;0.05, **p&lt;0.01, ***p&lt;0.001, and ****p&lt;0.0001.
In each experimental series, Col-0, _lre-7_, and _lre-7_ + _AtLRE_ _pro_ _:AtLRE-cYFP_ were
crossed and scored in parallel with the indicated test lines; thus, the control values
shown in this panel are the contemporaneous controls from the same experimental
block and may differ from control values shown in other figures.


|A|B|
|---|---|
|**C**|**D**|

#### **Figure 2 (legend in the next page)**


**Figure** **2:** **Full-length** **SiLRE** **significantly** **improves** **pollen** **tube** **reception** **in** **inter-**
**lineage crosses to** _**S. irio**_ **pollen**


Quantification of the average pollen tube overgrowth per pollen tube-entered ovule
(%OG/PEO) in crosses between female parents listed on the X-axis and _A._ _thaliana_ pollen
(A), _A. lyrata_ pollen (B), and _S. irio_ pollen (C, D).


The average % OG/PEO for each female was compared to the % OG/PEO of _lre-7/lre-7_ _+_
_AtLRE_ _pro_ _:AtLRE-cYFP_ . In each construct, three independently transformed lines were
crossed to _A._ _thaliana_ pollen _._ % OG/PEO indicates the percentage of ovules in a silique
that experienced pollen tube overgrowth in the female gametophyte out of all the ovules
that received a pollen tube. The number of pistils crossed, and ovules scored is listed for
each female parent. Statistical significance was determined by first using an arcsine
transformation of percentage data, followed by pairwise _t-_ tests between _lre-7/lre-7+_
_AtLRE_ _pro_ _:AtLRE-cYFP_ and tested lines; ns&gt; 0.05, *p&lt;0.05, **p&lt;0.01, ***p&lt;0.001, and
****p&lt;0.0001.


In each experimental series, Col-0, _lre-7_, and _lre-7_ + _AtLRE_ _pro_ _:AtLRE-cYFP_ were crossed
and scored in parallel with the indicated test lines; thus, the control values shown in this
panel are the contemporaneous controls from the same experimental block and may differ
from control values shown in other figures.


## **A)**

### **B)**





**E47** **R120**



**��**
**region** **GAS**




##### * *



KCEVN

KCNVN

KCEVN

KCEVN

KCEVN

KCEVN

KCEVN

KCEVN

KCEVN

KCEVN
QCEVN
ECEVN

KCEVN

KCEVN

KCEVN

KCEVN

ECEVN

KCEVN
QCEVD
QCEVD
KCEVN

----
KCEVN

ECEVN

KCKVN

KCEVS

KCEVN

TCPVN

PCSVN

PCPVN

TCPVN

PCPVN

PCPVN

PCPVN

TCPVN

TCPVN

PCPEK

ACRVN

ACRVN

ACPVN

ACPVN

ACPVK

ACPVN

ACPVN

ACPVK

ACPVN

----
KCSLN

KCSVN

KCSVN

TCPVN

TCPVN

TCPVN

ACPVN

ACPVN

ACPVN

ACPVN

ACPVN

ACPVN

PCPVS

TCPVN

PCPVN

SCPVN

SCPVN

PCPVN

PCPVN

SCPVN

PCPVN

PCTVN

PCTVK

ACPVN

PCPVN

ACPVN

ACPVN

ACPVN

ACPIK

ACPVN

ACPVN







RERKD

KERKD

KEKKD

KERKD

RERKD

RETKD

RERRD

RERKD

RERKD

RERKD

KETKT

KETKT

KERKD

EETKD

KEMKD

LS--
KETKN

KDTKD

KETKV

QDTKE
KDTKD

KDTKD

KDTKD

KDTKY

KERKD

KDTKY

KETRH

KEGKE

KEGKE

KEGKE

KEGKE

KEGKE

KEGKE

KEGKE

KRGKE

KDGKE

KEGIE

KEGKE

KEGKD

KEGKE

KEGKE
QEGKG
QE-KG
QR-KG
KEGKE

TRVKG

NEGKG

HGNKI
QEGKY
QEGKY
KEGKE

KGSKE

KGGKE

KEGKE

KEGKE

KDGKE
QEGKE
KEGKY
QEGKE
QEGKE
KEGKK

SESKD

TEGKE

KEGKE

KESKE

KEGKE

KEGKE

KEGKL

REGKE

REGKK

REGKE

REGKE

LEGKL

REGKL

REGKL
REGKQ
HEGKE

REGKE



_**LRE**_


_**LLG1**_


_**LRE/LLG1**_



**Brassicaceae**


**��������������**

**Brassicales**


**Malvales**


#### **Figure 3 (legend in the next page)**


**Figure 3: The** _**LRE**_ **clade experienced a relaxation of purifying selection post-gene**
**duplication in Brassicaceae.**
A) A diagram of the nascent _A. thaliana LRE_ polypeptide with arrows indicating residues
E47 and R120. The pre-processed protein contains a secretion peptide (SP), an
extracellular domain (black) containing a modified eight-cysteine motif (M8CM; denoted
by C1-8), a proline rich unstructured ω-11 region, ω, and a GPI-attachment signal
(GAS) region.
B) Maximum-likelihood phylogram inferred from 80 _LRE_ and _LLG1_ homologues
retrieved from a set of 25 Brassicales and two Malvales species. The tree was rooted
on Malvales _LRE/LLG1_ homologues. _LRE_ and _LLG1_ clades from Lineage I and Lineage
II Brassicaceae are marked with grey arrows. _LRE_ and _LLG1_ post-duplication branches
are marked in red. _LRE_ and _LLG1_ from _A._ _thaliana_ are highlighted with purple. _LRE_ and
_LLG1_ duplicated early in the evolution of Brassicaceae. Sites potentially responsible for
this signature of selection are shown at two short regions of the aligned polypeptide
(corresponding to sites 45-49 and 118-122 of _A._ _thaliana_ _LRE_ ); red asterisks denote
E47 and R120.


|A|B|
|---|---|
|**C**|**D**|


**Figure 4: Both** **R120T** **single point mutant and** **E47P,** **R120G** **double point mutant** **improve**
**pollen tube reception in inter-lineage cross to** _**S.irio**_ **pollen.**


Quantification of the average pollen tube overgrowth per pollen tube-entered ovule (%OG/PEO)
in crosses between female parents listed on the X-axis and _A._ _thaliana_ pollen (A, C) and _S._ _irio_
pollen (B, D). The average % OG/PEO for each female was compared to the % OG/PEO of _lre-_
_7/lre-7_ _+_ _AtLRE_ _pro_ _:AtLRE-cYFP_ . In each construct, three independently transformed lines were
crossed to _A._ _thaliana_ pollen _._ % OG/PEO indicates the percentage of ovules in a silique that
experienced pollen tube overgrowth in the female gametophyte out of all the ovules that
received a pollen tube. The number of pistils crossed, and ovules scored is listed for each
female parent. Statistical significance was determined by first using an arcsine transformation of
percentage data, followed by pairwise _t-_ tests between _lre-7/lre-7+_ _AtLRE_ _pro_ _:AtLRE-_
_cYFP_ and tested lines; ns&gt; 0.05, *p&lt;0.05, **p&lt;0.01, ***p&lt;0.001, and ****p&lt;0.0001.


In each experimental series, Col-0, _lre-7_, and _lre-7_ + _AtLRE_ _pro_ _:AtLRE-cYFP_ were crossed and
scored in parallel with the indicated test lines; thus, the control values shown in this panel are
the contemporaneous controls from the same experimental block and may differ from control
values shown in other figures.


**Figure** **5:** Structural models from subsampled AlphaFold suggest distinct N- and C-terminal

ensembles between LRE variants. ( **A** ) Renderings of the fifty highest ranked structures

(shown in grey) from each conformational ensemble. The highest ranked structures are
highlighted in color for WT (blue), R120G (green), E47P (grey), R120T (brown) and E47P
R120G (EPRG, cyan). ( **B** ) RMSD plots of each variant, calculated using the highest ranked
structural model as a reference. Variants are also compared to WT _At_ LRE using same

coloring as ( **A** ).


**Figure** **6:** A working model of the interaction between _S._ _irio_ pollen tubes and
_A. thaliana_ synergid cells in ovules with and without _S. irio_ LRE.


In a wild-type cross between _A._ _thaliana_ and _S._ _irio_, we observe that pollen
tube reception fails, possibly due to incompatible paternal factors (F) that the
AtLRE- AtFER- AtHERK1/AtANJ complex does not recognize and therefore
does not bind. When AtLRE is replaced with SiLRE in the synergid cells, we
observe an increase in normal pollen tube reception with _S._ _irio_ pollen tubes,
potentially due to pollen tube recognition by SiLRE. In this scenario, the _S._ _irio_
paternal factors (F) bind SiLRE and initiate SiLRE- AtFER- AtHERK1/AtANJ
complex formation for downstream signaling leading to successful pollen tube
reception.


