AHP2, AHP3, and AHP5 act downstream of CKI1 in Arabidopsis female gametophyte development.

Histidine phosphotransfer proteins (HPs) are key elements of the two-component signaling system, which act as a shuttle to transfer phosphorylation signals from histidine kinases (HKs) to response regulators (RRs). CYTOKININ INDEPENDENT 1 (CKI1), a key regulator of central cell specification in the Arabidopsis female gametophyte, activates the cytokinin signaling pathway through the Arabidopsis histidine phosphotransfer proteins (AHPs). There are five HP genes in Arabidopsis, AHP1-AHP5, but it remains unknown which AHP genes act downstream of CKI1 in Arabidopsis female gametophyte development. Promoter activity analysis of AHP1-AHP5 in embryo sacs revealed AHP1, AHP2, AHP3, and AHP5 expression in the central cell. Phenotypic studies of various combinations of ahp mutants showed that triple mutations in AHP2, AHP3, and AHP5 resulted in defective embryo sac development. Using cell-specific single and double markers in the female gametophyte, the ahp2-2 ahp3 ahp5-2/+ triple mutant ovules showed loss of central cell and antipodal cell fates and gain of egg cell or synergid cell attributes, resembling the cki1 mutant phenotypes. These data suggest that AHP2, AHP3, and AHP5 are the major factors acting downstream of CKI1 in the two-component cytokinin signaling pathway to promote Arabidopsis female gametophyte development.

Introduction

Two-component signaling systems were originally identified in bacteria, the simplest form of which employs a receptor kinase and a response regulator (RR). In response to an environmental stimulus, the histidine kinase (HK) perceives the signal and autophosphorylates itself on a conserved histidine residue, and the phosphate is then transferred to a conserved aspartic acid residue within another group of signal transducers called the RRs (Mizuno, 1997; Stock ). Plants have multistep phosphorelays involving HKs, histidine-containing phosphotransfer proteins (HPs), and RRs. The HPs are responsible for phosphate transfer from the HKs to the RRs, which could contribute to increasing diversity and complexity of the signal transduction. Multistep phosphorelays have been implicated in the growth and development regulation, hormone responses, and osmotic stresses of plants (Schaller, 2000; Maxwell and Kieber, 2010; Müller, 2011).

The five Arabidopsis HP genes, AHP1–AHP5, encode small proteins with putative histidine phosphotransfer activity similar to that of the yeast and prokaryotic histidine phosphotransfer domains. AHPs are localized in the cytoplasm and the nucleus, to transfer phosphorylation signals from AHKs to the ARRs (Arabidopsis RRs) (Hwang ). A sixth protein, AHP6, which lacks the conserved histidine residue, is an inhibitory pseudo-phosphotransfer protein that inhibits the phosphorelay from AHPs to ARRs (Mähönen ).

CYTOKININ INDEPENDENT 1 (CKI1) was identified as a HK gene, which could induce a typical cytokinin response in the absence of cytokinin when overexpressed (Kakimoto, 1996). CKI1 is primarily expressed in the micropylar end of embryo sacs, and loss-of-function mutants are semi-sterile and exhibit a block in megagametogenesis, mainly characterized by the abortion or degradation of embryo sacs (Pischke ; Hutchison ; Hejátko ). Research showed that the CKI1-induced cytokinin response is independent of cytokinin receptors but is dependent on AHP1–AHP5 (Deng ), and further studies on cki1/+ mutants and cytokinin receptor mutants indicated that the CKI1–AHPs–ARRs pathway, rather than the cytokinin receptor AHK–AHP–ARR pathway, is required for female gametophyte development (Pischke ; Hutchison ; Hejátko ; Deng ). Deng showed that the ahp1 ahp2 ahp3 ahp4 ahp5 quintuple mutants caused severe defects in female gametophyte development, the progression of female gametogenesis was disturbed, embryo sacs were arrested at the FG5 stage, or degraded to varying degrees, but the male gametophyte development was unaffected, which resembled the phenotypes described in cki1 mutants (Pischke ; Hejátko ; Rabiger and Drews, 2013; Yuan ). However, the exact functions of AHPs involved in the CKI1–AHPs–ARRs pathway remain unknown. Protein–protein interactions between CKI1 and AHPs were assayed by bimolecular fluorescence complementation (BiFC), and the results showed that CKI1 interacts with AHP1, AHP2, AHP3, and AHP5; although a weak interaction signal was recorded for AHP1 (Pekárová ). However, in yeast two-hybrid assays, CKI1 interacted with only AHP2, AHP3, and AHP5, and not AHP1 (Urao ; Pekárová ). Thus, there is good evidence for AHP2, AHP3, and AHP5 interacting directly with CKI1, and the evidence for AHP1 as a CKI1 interactor remains ambiguous.

In this study, the promoter activities of AHP1–AHP5 were primarily investigated in floral organs and embryo sacs. Various combinations of ahp mutants were constructed, and the morphological phenotypes were observed. Female gametophytic cell-specific markers were introduced to determine the cell fate switch in these mutants. Our results indicated that AHP2, AHP3, and AHP5 are downstream regulators of CKI1; these proteins act in an overlapping and redundant manner to regulate female gametophyte development, thereby providing more evidence to clarify the functions of the two-component system signaling network in plant reproduction.

Materials and methods

Plant materials and growth conditions

The ahp1-1 (Ws, CS860143), ahp2-2 (Col-0, SALK_019024), ahp3 (Col-0, SALK_041384), ahp5-2 (Col-0, SALK_079857), and cki1-9/+ (Col-0, SALK_057881) single mutants were obtained from the Arabidopsis Biological Resource Center (ABRC). The ahp1-1 ahp2-2 ahp3/+ ahp4 ahp5-2 quintuple mutant was kindly provided by Jianru Zuo (Chinese Academy of Sciences). Plants were grown in a growth chamber with a 16 h light/8 h dark cycle at 22 °C with 60% relative humidity.

Generation and screening of ahp multiple mutants

For the crosses, the flowers of the female parent were manually emasculated at 24 h before anthesis and cross-pollinated after 24 h. Double mutants were first generated by crossing single mutants with each other before double mutants were further crossed to generate triple mutants and quadruple mutants. Plants homozygous and heterozygous for insertions in all five loci were identified by PCR with gene-specific primers as described in Supplementary Table S1 at JXB online.

Histology and microscopy

To observe seed development, siliques were dissected with needles and checked under a dissecting microscope. For whole-mount clearing observations, pistils containing at least 20 ovules were dissected and cleared overnight in Hoyer’s solution (Liu and Meinke, 1998). The ovules were subsequently observed under differential interference contrast (DIC) optics with a Zeiss 2 Axioplan imaging microscope (Axioskop 2 plus). For fluorescence microscopy, individual ovules were dissected from the pistils in 0.1 M phosphate buffer (pH 7.0), and samples were observed under a Zeiss 710 confocal laser scanning microscope (LSM 710). For β-glucuronidase (GUS) staining, inflorescences and four-whorl flower organs were dissected and incubated in the GUS staining buffer as previously described (Pagnussat ). Samples were observed under a Zeiss 2 Axioplan imaging microscope (Axioskop 2 plus).

Constructs and plant transformation

For the AHP1, AHP2, AHP3, AHP4, AHP5, and ETR1 promoter–reporter constructs, putative promoter regions (~2 kb upstream of ATG) were separately amplified from Col-0 genomic DNA. The PCR fragments were cloned into pENTR/D-TOPO (Invitrogen) and subsequently cloned into the gateway vector pBGWFS7 and NLS-3xGFP-NOST-pMLBART by the LR clonase reaction (Invitrogen).

For the single cell-specific markers, the Nos terminator (NOST-HindIII-F/NOST-HindIII-R), H2B (H2B-XbaI-F/H2B-SalI-R), and eGFP or TagRFP (SalI-F/SalI-R) were amplified by PCR and inserted into the pCAMBIA1300 vector by a multiplestep process. The promoters DD22, EC1.1, DD31, and DD13 (BamHI-F/XbaI-R) were separately amplified from Col-0 genomic DNA and cloned into the H2B-eGFP/TagRFP-NOST-pCMBIA1300 vector to generate the promoter–reporter constructs. To generate double cell-specific markers, fragments of promoter–H2B-TagRFP-NOST were amplified by PCR and cloned into single green fluorescent protein (GFP) marker constructs at the BamHI cloning site.

Individual binary expression vectors were delivered into Agrobacterium tumefaciens AGL1, which was used to transform Col-0 wild-type plants or ahp multiple mutants. Transgenic plants were screened by spraying 1000-fold diluted glyphosate herbicide (Syngenta) or grown on Murashige and Skoog (MS) medium containing 25 mg l–1 hygromycin B (Invitrogen). At least 20 independent lines were assayed for each construct. For the marker transgenic plants, only lines showing Mendelian genetic segregation ratios were used.

Results

Expression of AHP genes in floral organs and embryo sacs

CKI1 expression has been reported in vascular tissues of floral organs and the chalazal domain of mature embryo sacs (Hejátko ; Yuan ). Genes acting downstream of CKI1 might be expected to exhibit similar expression profiles. To investigate the expression patterns of AHP1–AHP5, we determined the transcriptional activity of transgenic lines with the GUS marker gene. GUS staining results showed that all five AHP genes exhibited vascular tissue-specific expression levels in the four whorls of floral organs (Supplementary Figs S1–S5). The expression profiles of these genes were further determined in the mature embryo sac by employing transgenic lines with the nuclear eGFP marker gene. The fluorescence signal of AHP1 was detected in the central cell and the synergid cells; both AHP2 and AHP5 were expressed in all the female gametophytic cells; AHP3 was specifically expressed in the central cell, whereas no signals were detected for AHP4 in the embryo sac (Fig. 1).

Fig. 1.

Expression of AHP1–AHP5 in Arabidopsis mature embryo sacs. (A and B) AHP1pro::NLS-3XeGFP; (C and D) AHP2pro::NLS-3XeGFP; (E and F) AHP3pro::NLS-3XeGFP; (G and H) AHP4pro::NLS-3XeGFP; (I and J) AHP5pro::NLS-3XeGFP. CCN, central cell nucleus; SCN, synergid cell nuclei; ECN, egg cell nucleus; ACN, antipodal cell nuclei. Scale bars=20 µm.

Silique check of ahp multiple mutants

Promoter activity analysis suggested that AHP1, AHP2, AHP3, and AHP5 are candidate genes for regulation of female gametophyte development downstream of CKI1. As ahp single mutants are completely fertile, multiple mutant combinations were generated, and seed development was examined. As shown in Table 1 and Supplementary Fig. S6, the homozygous ahp2-2 ahp3 double mutant, the ahp2-2 ahp5-2 double mutant, and the ahp3 ahp5-2 double mutant were fully fertile, similar to wild-type plants, without the obvious seed development abortion phenotype. Seed development in the homozygous ahp1-1 ahp2-2 ahp3 triple mutant, the ahp1-1 ahp2-2 ahp5 triple mutant, and the ahp1-1 ahp3 ahp5-2 triple mutant was normal, also with no obvious differences from the control plant. However, the triple mutants with combinations of ahp2-2, ahp3, and ahp5-2 showed obvious seed development abortion, characterized by unfertilized ovules. No homozygous triple mutants could be obtained, thereby suggesting the redundant functions among AHP2, AHP3, and AHP5. The ratio of unfertilized ovules in ahp2-2 ahp3 ahp5-2/+ was 47.4% (n=430), whereas the ratios in ahp2-2 ahp3/+ ahp5-2 and ahp2-2/+ ahp3 ahp5-2 reached only 36.9% (n=452) and 36.3% (n=113), respectively. In addition, the ahp1-1 allele was used for the construction of the quadruple mutants. The results showed that the ratio of unfertilized ovules in ahp1-1 ahp2-2 ahp3 ahp5-2/+ was 47.2% (n=466), which was similar to that in ahp2-2 ahp3 ahp5-2/+. The ratios of unfertilized ovules in ahp1-1 ahp2-2 ahp3/+ ahp5-2 and ahp1-1 ahp2-2/+ ahp3 ahp5-2 were 35.5% (n=405) and 36.1% (n=432), respectively, which were similar to those in ahp2-2 ahp3/+ ahp5-2 and ahp2-2/+ ahp3 ahp5-2. These results suggest that AHP1 is not involved in the CKI1 signaling pathway to regulate female gametophyte development. Similar observations in siliques with quadruple mutant combinations of ahp4, namely ahp1-1 ahp2-2 ahp4 ahp5-2/+ and ahp1-1 ahp2-2 ahp3/+ ahp4 ahp5-2, also excluded the involvement of AHP4 in this biological process.

Table 1.

Silique check in cki-9/+ and ahp multiple mutants

Lines	Abnormal seeds	Normal seeds	Total
Wild type	6 (2.3%)	255 (97.7%)	261 (100%)
cki1-9/+	191 (48.8%)	200 (51.2%)	391 (100%)
ahp2-2 ahp3	5 (2.5%)	197 (97.5%)	202 (100%)
ahp2-2 ahp5-2	3 (1.3%)	230 (98.7%)	233 (100%)
ahp3 ahp5-2	7 (2.0%)	343 (98.0%)	350 (100%)
ahp1 ahp2-2 ahp3	10 (2.7%)	366 (97.3%)	376 (100%)
ahp1 ahp2-2 ahp5	5 (1.4%)	363 (98.6%)	368 (100%)
ahp1 ahp3 ahp5-2	8 (2.3%)	343 (97.7%)	351 (100%)
ahp2-2 ahp3 ahp5-2/+	204 (47.4%)	226 (52.6%)	430 (100%)
ahp2-2 ahp3/+ ahp5-2	167 (36.9%)	285 (63.1%)	452 (100%)
ahp2-2/+ ahp3 ahp5-2	41 (36.3%)	72 (63.7%)	113 (100%)
ahp1 ahp2-2 ahp3 ahp5-2/+	220 (47.2%)	246 (52.8%)	466 (100%)
ahp1 ahp2-2 ahp3/+ ahp5-2	144 (35.5%)	261 (64.4%)	405 (100%)
ahp1 ahp2-2/+ ahp3 ahp5-2	156 (36.1%)	276 (63.9%)	432 (100%)
ahp1 ahp2-2 ahp4 ahp5-2/+	7 (1.7%)	401 (98.3%)	408 (100%)
ahp1 ahp2-2 ahp3/+ ahp4 ahp5-2	98 (35.1%)	181 (64.9%)	279 (100%)

Ovule clearing analysis of ahp multiple mutants

To investigate further the underlying mechanism of the unfertilized ovules in ahp multiple mutants, ovules were cleared and observed. As shown in Fig. 2, Table 2, and Supplementary Fig. S7, compared with wild-type plants, 97.8% ovules showed normal embryo sacs without apparent phenotypic defects in the ahp2-2 ahp3 double mutant plants. However, 46.2% of ovules (n=208) in ahp2-2 ahp3 ahp5-2/+ showed abnormal embryo sacs; among which 32.2% of the polar nuclei were unable to fuse within the central cell, and 13.9% of the embryo sacs had degenerated by varying degrees. In ahp2-2 ahp3/+ ahp5-2, 31.6% of the ovules showed abnormal embryo sacs with unfused polar nuclei (18.7%) and degenerated embryo sacs (12.9%). Similarly, in ahp2-2/+ ahp3 ahp5-2, 30.6% of the ovules showed abnormal embryo sacs with unfused polar nuclei (18.5%) and degenerated embryo sacs (12.1%). Comparatively, these three mutants showed similar ratios of degenerated embryo sacs, but the ratio of unfused polar nuclei in ahp2-2 ahp3 ahp5-2/+ was considerably higher than that in ahp2-2 ahp3/+ ahp5-2 and ahp2-2/+ ahp3 ahp5-2 (32.3% versus 18.7% and 18.5%, respectively). Notably, unlike the cki1/+ mutants, the supernumerary nuclei phenotype was not observed in all three of these mutants. The ratios of abnormal embryo sacs by direct observations paralleled the unfertilized ovule ratios in corresponding ahp multiple mutants, indicating that unfertilized ovules might be attributable to the female gametophytic defects before fertilization.

Fig. 2.

DIC microscopy of the cleared ovules in ahp multiple mutants. (A) Normal embryo sacs. (B–D) Abnormal embryo sacs. (B) Polar nuclei fail to fuse. (C) Polar nuclei fail to fuse; three antipodal cells persist; antipodal cell nuclei move toward the micropylar end. (D) Late stage degenerated embryo sacs with invisible female gametophyte nuclei. The position and size of the nuclei in embryo sacs are manually marked with yellow dots. Ch, chalazal end; Mi, micropylar end; CCN, central cell nucleus; ECN, egg cell nucleus; SYN, synergid cell nuclei; ACN, antipodal cell nuclei; PN, polar nuclei. Scale bars=15 µm.

Table 2.

Classification of embryo sac phenotypes in cki-9/+ and ahp multiple mutants

Line	Abnormal embryo sacs				Normal embryo sacs	Total
	Unfused polar nuclei	Degaratedembryo sacs	Supernumerary nuclei	Total
Wild type	3 (2.1%)	2 (1.4%)	0 (0%)	5 (3.5%)	140 (96.5%)	145 (100%)
cki1-9/+	45 (33.8%)	16 (12.0%)	5 (3.8%)	66 (49.6%)	67 (50.4%)	133 (100%)
ahp2-2 ahp3	1 (0.4%)	4 (1.8%)	0 (0%)	5 (2.2%)	222 (97.8%)	227 (100%)
ahp2-2 ahp3 ahp5-2/+	67 (32.2%)	29 (13.9%)	0 (0%)	96 (46.2%)	112 (53.8%)	208 (100%)
ahp2-2 ahp3/+ ahp5-2	65 (18.7%)	45 (12.9%)	0 (0%)	110 (31.6%)	238 (68.4%)	348 (100%)
ahp2-2/+ ahp3 ahp5-2	55 (18.5%)	36 (12.1%)	0 (0%)	91 (30.6%)	206 (69.4)	297 (100%)

Female gametophytic cell fates are altered in ahp2-2 ahp3 ahp5-2/+ triple mutants

Since ahp2-2 ahp3 ahp5-2/+, ahp2-2 ahp3/+ ahp5-2, and ahp2-2/+ ahp3 ahp5-2 mutants showed similar phenotypes under microscopy, the triple mutant ahp2-2 ahp3 ahp5-2/+ was chosen as the representative mutant to study the female gametophytic cell fate specification.

Female gametophyte cell-specific single markers were primarily introduced into ahp2-2 ahp3 ahp5-2/+ by the Agrobacterium-mediated floral dipping method. Thus, all plants examined were heterozygous for the markers, in addition to being heterozygous for ahp5-2. As shown in Fig. 3A–D and Table 3, the ratio of embryo sacs with normal central cell-specific positive GFP signals in ahp2-2 ahp3 ahp5-2/+ was ~25% less than in the ahp2-2 ahp3 control plant (25.2% versus 48.1%), thereby indicating the loss of the central cell fate in ahp2-2 ahp3 ahp5-2/+. Likewise, embryo sacs with normal antipodal cell-specific positive GFP signals in ahp2-2 ahp3 ahp5-2/+ were also 25% less than in the ahp2-2 ahp3 control plant (25.7% versus 50.0%) (Fig. 3E–H; Table 3), thereby indicating the loss of antipodal cell fate in ahp2-2 ahp3 ahp5-2/+. For synergid cell-specific marker expression, only 32.1% of the embryo sacs in ahp2-2 ahp3 ahp5-2/+ showed normal GFP signals, whereas 16.5% of the embryo sacs showed that the two unfused polar nuclei also expressed the synergid cell-specific marker (Fig. 3I–L and Table 3), similar to the observed cki1-9/+ phenotype (Yuan ), indicating the adoption of the synergid cell attributes in the polar nuclei. For egg cell-specific marker expression, additional phenotypic manifestations were observed. Only 26.0% of the embryo sacs in ahp2-2 ahp3 ahp5-2/+ exhibited normal GFP signals; the remaining 21.9% of the embryo sacs exhibited the cki1-like phenotype of ectopic egg cell marker expression. Specifically, three antipodal cells expressed the egg cell-specific marker or three antipodal cells and two unfused polar nuclei expressed the egg cell-specific marker. Occasionally, the fused central cell also expressed the egg cell-specific marker (Fig. 3M–T; Table 3). These results demonstrated the adoption of the egg cell attributes in antipodal cells and unfused or fused polar nuclei.

Fig. 3.

Expression of female gametophyte cell-specific markers in ahp2-2 ahp3 ahp5-2/+ triple mutants. First and third column, nuclear-localized eGFP fluorescent signals; second column, merged images of the first column and the bright field images; fourth column, merged images of the third column and the bright field images. (A–D) Expression of a central cell-specific marker (DD22pro::H2B-eGFP/+). (A and B) Embryo sac with the central cell-specific marker; (C and D) embryo sac without the central cell-specific marker. (E–H) Expression of an antipodal cell-specific marker (DD13pro::H2B-eGFP/+). (E and F) Embryo sac with the antipodal cell-specific marker; (G and H) embryo sac without the antipodal cell-specific marker. (I–L) Expression of a synergid cell-specific marker (DD31pro::H2B-eGFP). (I and J) Embryo sac with the normal synergid cell-specific marker; (K and L) embryo sac with the abnormal synergid cell-specific marker; unfused polar nuclei also express the synergid cell-specific marker. (M–T) Expression of an egg cell-specific marker (EC1.1pro::H2B-eGFP). (M and N) Embryo sac with the normal egg cell-specific marker; (O–T) embryo sac with the abnormal egg cell-specific marker. (O and P) Three antipodal cells express the egg cell-specific marker. (Q and R) Three antipodal cells and two unfused polar nuclei all express the egg cell-specific marker. (S and T) Three antipodal cells and central cell express the egg cell-specific marker. Ch, chalazal end; Mi, micropylar end; CCN, central cell nucleus; ACN, antipodal cell nuclei; SCN, synergid cell nuclei; PN, polar nuclei; ECN, egg cell nuclei. Scale bars=20 µm.

Table 3.

Expression of female gametophyte cell-specific markers in ahp2 ahp3 ahp5/+ triple mutants

Lines	Positive GFP			Negative GFP	Total
	Positive GFP (wild type like)	Positive GFP (cki1-9/+ like)	Total
ahp2 ahp3; DD22pro::H2B-eGFP/+	90 (48.1%)	0 (0%)	90 (48.1%)	97 (51.9%)	187 (100%)
ahp2 ahp3 ahp5/+; DD22pro::H2B-eGFP/+	96 (25.2%)	0 (0%)	96 (25.2%)	285 (74.8%)	381 (100%)
ahp2 ahp3; DD13pro::H2B-eGFP/+	127 (50%)	0 (0%)	127 (50%)	127 (50%)	254 (100%)
ahp2 ahp3 ahp5/+; DD13pro::H2B-eGFP/+	83 (25.7%)	0 (0%)	83 (25.7%)	240 (74.3%)	323 (100%)
ahp2 ahp3; DD31pro::H2B-eGFP/+	138 (47.6%)	0 (0%)	138 (47.6%)	152 (52.4%)	290 (100%)
ahp2 ahp3 ahp5/+; DD31pro::H2B-eGFP/+	130 (32.1%)	67 (16.5%)	197 (48.6%)	208 (51.4%)	405 (100%)
ahp2 ahp3; EC1.1pro::H2B-eGFP/+	118 (47.2%)	0 (0%)	118 (47.2%)	132 (52.8%)	250 (100%)
ahp2 ahp3 ahp5/+; EC1.1pro::H2B-eGFP/+	88 (26.0%)	74 (21.9%)	162 (47.9%)	176 (52.1%)	338 (100%)

Similar to cki1 mutants (Yuan ), the central cell and antipodal cell fates might be expected to be transformed into the egg cell or synergid cell fates in ahp2-2 ahp3 ahp5-2/+ triple mutants. Female gametophyte cell-specific double markers were used to confirm this hypothesis. For the egg cell-specific and central cell-specific double marker in the ahp2-2 ahp3 control plants, 46.9% of the embryo sacs showed egg cell-specific red fluorescent protein (RFP) signals, and 47.7% of the embryo sacs showed central cell-specific GFP signals. These frequencies are as expected for normal embryo sac development. However, in the ovules of the ahp2-2 ahp3 ahp5-2/+ triple mutant plants, only 26.2% of the embryo sacs showed normal egg cell-specific RFP signals and 24.1% of the embryo sacs showed normal central cell-specific GFP signals (Fig. 4A–D; Supplementary Table S2). Ectopic egg cell marker expression with concomitant loss of central cell marker expression was observed in 22.1% of the embryo sacs (Fig 4E–P). The ectopic egg cell expression with the double marker resembled the previously described expression observed with the egg cell-specific single marker (Fig. 3O–T). For the synergid cell-specific and central cell-specific double markers, two unfused polar nuclei showed synergid cell-specific RFP signals, whereas the central cell-specific GFP signal was negative (Fig. 5). These results provided more evidence that the central cell and antipodal cell fates were completely lost in ahp2-2 ahp3 ahp5-2/+ embryo sacs, along with gain of egg cell or synergid cell fates. Thus, the ahp2-2 ahp3 ahp5-2/+ triple mutants reproduced the phenotypes that were observed in cki1 mutants, consistent with AHP2, AHP3, and AHP5 acting downstream of CKI1 to regulate cell fates in the chalazal domain of the embryo sac.

Fig. 4.

Expression of central cell-specific and egg cell-specific double markers (DD22pro::H2B-eGFP and EC1.1pro::H2B-TagRFP) in ahp2 ahp3 ahp5/+ triple mutants. First column, central cell-specific marker expression; second column, egg cell-specific marker expression; third column, merged images of first and second columns; fourth column, merged images of the third column and the bright field images. (A–D) Embryo sacs with normal central cell-specific and egg cell-specific markers; (E–H) embryo sacs with abnormal central cell-specific and egg cell-specific markers; no central cell-specific markers were expressed in the central cell; egg cells show normal egg cell-specific markers. (E–H) Three antipodal cells showing egg cell-specific markers, (I–L) Three antipodal cells and two unfused polar nuclei showing egg cell-specific markers. (M–P) Three antipodal cells showing egg cell-specific markers, (I–L) Three antipodal cells and a central cell showing egg cell-specific markers. Ch, chalazal end; Mi, micropylar end; CCN, central cell nucleus; ECN, egg cell nucleus; PN, polar nuclei. Scale bars=20 µm.

Fig. 5.

Expression of central cell-specific and synergid cell-specific double markers (DD22pro::H2B-eGFP and DD31pro::H2B-TagRFP) in ahp2 ahp3 ahp5/+ triple mutants. First column, central cell-specific marker expression; second column, synergid cell-specific marker expression; third column, merged images of first and second columns; fourth column, merged images of the third column and the bright field images. (A–D) Embryo sacs with normal central cell-specific and synergid cell-specific markers. (E–H) Embryo sacs with abnormal central cell-specific and synergid cell-specific markers; two unfused polar nuclei do not show central cell-specific markers but show synergid cell-specific markers. Ch, chalazal end; Mi, micropylar end; CCN, central cell nucleus; SYN, synergid cell nuclei; PN, polar nuclei. Scale bars=20 µm.

Discussion

Gene regulatory pathway for female gametophyte development

The supernumerary nuclei phenotype is characterized by the presence of 10–14 nuclei in embryo sacs, which exceeds the normal 8 nuclei. RETINOBLASTOMA RELATED (RBR) is known as a negative regulator of the cell cycle during mitosis; the loss of RBR perturbs the normal mitotic process, thereby leading to supernumerary nuclei because of the uncontrolled excessive nuclear division in embryo sacs (Ebel ; Johnston ). The supernumerary nuclei phenotype has also been identified in a small percentage of embryo sacs from the cki1/+ mutant (Pischke ; Hejátko ; Rabiger and Drews, 2013; Yuan ), but its mechanism is still to be determined. In this study, we set out to determine which specific AHPs out of the five known AHPs in Arabidopsis act downstream of CKI1 to regulate female gametophyte development. We determined that only AHP2, AHP3, and AHP5 appear to be involved. Specifically, the ahp2-2 ahp3 ahp5-2/+, ahp2-2 ahp3/+ ahp5-2, and ahp2-2/+ ahp3 ahp5-2 mutants all reproduced the cki1/+ phenotype of defective cell type specification in the chalazal domain (i.e. specification of the central cell and antipodal cells). However, the supernumerary nuclei phenotype identified in a small percentage of embryo sacs in the cki1/+ mutant (Pischke ; Hejátko ; Rabiger and Drews, 2013; Yuan ) was not observed with the ahp2 ahp3 ahp5 triple mutants. This was also the case with various ahp quadruple combinations and the quintuple mutants (data not shown). It is possible that a CKI1-related regulatory mechanism distinct from the AHP–ARR pathway is responsible for the nuclear proliferation defect observed in a small fraction of cki1 mutant embryo sacs.

Involvement of AHP1 in the two-component signal transduction pathway

Promoter fusions of AHP genes showed that AHP1, AHP2, AHP3, and AHP5 were expressed in embryo sacs, especially in the central cell. Functional analysis showed that only AHP2, AHP3, and AHP5 contributed to normal embryo sac development, and an ahp1 loss-of-function mutation did not enhance the ahp2 ahp3 ahp5 triple mutant phenotype. Interestingly, in protein–protein interaction studies by yeast two-hybrid analysis, the ethylene receptor protein ETR1 was found to interact with three AHPs, AHP1–AHP3 (Urao ). Transcriptome expression data reveal ETR1 expression in the central cell (Wuest ; Schmid ). The expression of ETR1 in embryo sacs was further checked by promoter activity analysis. GUS staining and fluorescence microscopy results indicated that ETR1 showed obvious expression in the central cell (Supplementary Fig. S8), which was consistent with transcriptome expression data (Wuest ; Schmid ). The ETR1–AHPs–ARRs ethylene signal transduction pathway seems independent of the classical ETR1–CTR1 ethylene signal transduction pathway (Chen ). Thus it is possible that AHP1 might be involved in ethylene signaling through the ETR1–AHPs–ARRs signal transduction pathway; however, its specific functions in female gametophyte development still need to be clarified.

Supplementary data

Supplementary data are available at JXB online.

Table S1. Summary of the primers used.

Table S2. Expresssion of the egg cell-specific and central cell-specific marker in ahp2 ahp3 ahp5/+ triple mutants.

Fig. S1. AHP1pro::GUS expression in Arabidopsis floral organs.

Fig. S2. AHP2pro::GUS expression in Arabidopsis floral organs.

Fig. S3. AHP3pro::GUS expression in Arabidopsis floral organs.

Fig. S4. AHP4pro::GUS expression in Arabidopsis floral organs.

Fig. S5. AHP5pro::GUS expression in Arabidopsis floral organs.

Fig. S6. Microscopy of seed development in siliques.

Fig. S7. DIC microscopy of the cleared ovules in ahp multiple mutants.

Fig. S8. Transgenic expression in mature Arabidopsis embryo sacs.

Author contributions

ZNL performed the experiments and analyzed the data, as well as drafting and writing the manuscript. YL and XYS contributed to mutant construction. VS and XLY proposed and supervised the research. All authors read and approved the final manuscript.

Click here for additional data file.