Maternal Control of PIN1 Is Required for Female Gametophyte Development in Arabidopsis

Land plants are characterised by haplo-diploid life cycles, and developing ovules are the organs in which the haploid and diploid generations coexist. Recently it has been shown that hormones such as auxin and cytokinins play important roles in ovule development and patterning. The establishment and regulation of auxin levels in cells is predominantly determined by the activity of the auxin efflux carrier proteins PIN-FORMED (PIN). To study the roles of PIN1 and PIN3 during ovule development we have used mutant alleles of both genes and also perturbed PIN1 and PIN3 expression using micro-RNAs controlled by the ovule specific DEFH9 (DEFIFICENS Homologue 9) promoter. PIN1 down-regulation and pin1-5 mutation severely affect female gametophyte development since embryo sacs arrest at the mono- and/or bi-nuclear stages (FG1 and FG3 stage). PIN3 function is not required for ovule development in wild-type or PIN1-silenced plants. We show that sporophytically expressed PIN1 is required for megagametogenesis, suggesting that sporophytic auxin flux might control the early stages of female gametophyte development, although auxin response is not visible in developing embryo sacs.


Introduction
In flowering plants, the gametophytes are comprised of only a few cells embedded within the diploid sexual organs of the flower. The formation of the female gametophyte is divided into two main steps, megasporogenesis and megagametogenesis [1,2]. During megasporogenesis the Megaspore Mother Cell (MMC) undergoes meiosis and produces four haploid megaspores. Three of the megaspores are destined to degenerate, and just one will persist and become a functional megaspore (FM). This event marks the FG1 stage of megagametophyte development. Subsequently, the FM undergoes three consecutive mitotic divisions (FG1-FG4) that lead to the formation of the mature embryo sac (FG5) [1,2]. The embryo sac is surrounded by the two integuments and connected to the placenta through the funiculus.
Active cross-talk between the maternal tissues and the developing embryo sac has been revealed from the study of ovule sporophytic mutants that display defects in megagametogenesis (for a review see [3]), and very recently it has been shown that auxin and cytokinins play key roles in ovule pattern formation [4]. Auxin moves from the placental tissue to the ovule [5] but the expression of genes involved in the tryptophan-dependent pathway of auxin production, such as TAA (TRYPTOPHAN AMINOTRANSFERASE OF ARABIDOPSIS) and YUC (YUCCA), indicate that auxin is also synthesized in developing ovules. The TAA enzymes convert tryptophan into IPA (indole-3-pyruvate), whereas the YUCs catalyse the conversion of IPA to IAA (indole-3-acetic acid) [6].
According to the accepted and well-documented model, the directionality of intercellular auxin flux mainly depends on the polar localization of the PIN proteins [7]. In Arabidopsis there are eight PIN genes (PIN1-PIN8), all of which may have diverged from a single ancestral gene sequence [8,9].
In accordance with its function, the pin1 mutant is affected in auxin transport [10] and displays progressive defects in organ initiation and phyllotaxy, leading to a pin-shaped inflorescence meristem devoid of flowers [11,12]. A mutant allele of PIN1, pin1-5, develops flowers having abnormal pistils and containing only a few ovules, some of which display defective embryo sacs [13,4]. This observation suggests a pivotal role for auxin during the early phases of megagametogenesis. In mature embryo sacs, overexpression of the auxin-producing enzyme YUC1 and downregulation of auxin-dependent transcription factors (ARFs) alter embryo sac cell identity [14], demonstrating that auxin also play an important role in the later stages of megagametogenesis.
In this manuscript we explore the roles PIN1 and PIN3 in ovule development by analysing pin1-5 and pin3-4 mutants and by silencing the wild-type genes using the DEFIFICENS Homologue 9 ovule-specific promoter [15]. Our data indicate that PIN1 is required for correct female gametophyte development and we therefore propose that auxin directly or indirectly controls female gametogenesis and coordinates its developmental progression.

Quantitative Reverse Transcription PCR
Expression analyses of PIN1 and PIN3 were performed using the iQ5 Multi Color real-time PCR detection system (Bio-Rad). PIN1 and PIN3 specific primers are listed in Table S1; normalisation was performed using UBIQUITIN10 (UBI10), rRNA18 and PROTEIN PHOSPHATASE 2A SUBUNIT A3 (PP2A3) as internal standards.
Transcript abundances were confirmed by two independent biological experiments and four technical repetitions. Moreover, PIN1 was amplified with two different primer pairs.

GUS Assays and Whole-mount Preparation
GUS staining was performed as reported by Vielle-Calzada and collaborators [16]. Developing ovules were cleared according to Yadegari et al. [17] and observed using a Zeiss Axiophot D1 microscope (http://www.zeiss.com) equipped with differential interface contrast (DIC) optics. Images were recorded with an Axiocam MRc5 camera (Zeiss) using the Axiovision program (4.1).

Confocal Laser Scanning Microscopy (CLSM ) and Immunolocalization
Dissected carpels were mounted with glycerol 5% (v/v). CLSM analysis was performed using a Zeiss LSM510 Meta confocal microscope.
To detect the GFP signal a 488 nm wavelength laser was used for excitation and a BP 505-550 nm filter was applied for GFP emission. For FMH 4-64 FX dye application (Invitrogen) an additional BP 575-615 nm filter was used.
For ovule morphological analyses pistils were fixed as described by Braselton et al. (1996) [18]. Samples were excited using a 532 nm laser and emission was detected between 570 and 740 nm.
For the immune-detection experiment a monoclonal anti-PIN1 antibody (Nanotools GmbH) was used as the primary antibody. To detect the signal from the Alexa FluorH 555 secondary antibody (Invitrogen) a 561 nm laser was used for excitation and a LP 575 nm filter was applied for fluorophore emission.

Plasmid Construction and Arabidopsis Transformation
All constructs were verified by sequencing and used to transform wild-type (Col-0) plants using the 'floral-dip' method [19].

In situ Hybridisation
In situ hybridisations with digoxigenin-labelled antisense RNA were performed as previously described [22]. A specific PIN5 cDNA fragment was amplified with primers Atp1229 and Atp1240.

PIN1 and PIN3 are Expressed in Sporophytic Tissues during Ovule Development
We have investigated the expression of the eight members of the Arabidopsis PIN gene family in ovules. PIN2:PIN2-GFP, PIN4:-GUS, PIN6:GUS and PIN7:PIN7-GFP transgenic plants were analysed and revealed that PIN genes 2, 4, 6 and 7 are not expressed in developing ovules ( Figure S1). PIN5 also fails to be expressed in developing ovules, as shown by the in situ hybridisation performed with a PIN5 antisense specific probe ( Figure S1), and it has been demonstrated recently that PIN8 is expressed exclusively in pollen grains [23].
With regard to PIN1, we performed a detailed PIN1 expression analysis from the earliest stages of ovule primordium development (stage 1-I) [1]. Analyses of transgenic plants containing the PIN1:PIN1-GFP construct demonstrate that PIN1 protein is localized at the lateral-apical membranes of diploid cells in the nucellus of developing ovule primordia ( Figure 1A). From stage 1-II to stage 3-II, PIN1 was also detected in the inner integument and in the funiculus (Figures 1B). After stage 3-II the nucellus cells degenerate and PIN1 is restricted to the chalaza ( Figure 1C and 1D). These data were confirmed by immuno-localisation experiments using specific antibodies against PIN1 ( Figures 1E and 1F) [12]. PIN1 is expressed in the funiculus during all stages of ovule development.
Analysis of PIN3:PIN3-GFP plants [24] showed PIN3 to be present in the cells of the developing nucellus shortly after ovule primordium appearance (stage 1-II, Figure 1G). The PIN3-GFP chimeric protein persists and is detectable until stage 3-VI; PIN3 is also present in the funiculus starting from stage 2-III until late stages of ovule development ( Figures 1H and I).

Auxin Synthesis and Accumulation during Ovule Development
Since it is very difficult to measure cellular auxin levels, IAA distribution has often been inferred from the analysis of PIN protein orientation across tissues coupled to patterns of auxin response revealed using synthetic auxin-responsive reporter lines [25].
To investigate auxin accumulation in developing ovules we used plants carrying the auxin-responsive reporter DR5rev:GFP [26] which provides a convenient tool to study auxin distribution in developing organs [5,27]. In the wild-type background the GFP signal appears in the epidermal cell layer of the nucellus of the ovule primordium starting from stage 1-II ( Figure 2A). From stage 2-IV [1,2] the auxin response was also detected in the provascular cells of the funiculus ( Figure 2C), and was further confirmed in the sporophytic nucellus of stage 2-V ovules by analysis of DR5rev:VENUS-N7 transgenic plants ( Figure 2F) [28] in which the DR5rev synthetic promoter drives three tandem copies of VENUS (a rapidly folding YFP variant) fused to a nuclear localization sequence. These observations clearly show that the auxin response is restricted to the micropylar pole of the nucellus until FG3 (stage 3-II), however auxin accumulation inside the female gametophyte could not be detected. This pattern is maintained until the nucellus progressively degenerates and is substituted by the endothelium ( Figure S2).
To verify whether auxin is locally synthesized we have examined the expression of the genes TAA1 and YUC4 which are involved in the Trp-dependent auxin biosynthetic pathway [6,29], taking advantage of transgenic lines expressing reporter genes (GFP/GUS) under the control of the TAA1 or YUC4 promoters. The TAA1 promoter is active from stage 2-I ( Figure 2G), and at stage 2-III the GFP signal is also detected in inner integument primordia and in the funiculus ( Figure 2H). At stage 3-II the TAA1 promoter drives GFP expression in those cells surrounding the female gametophyte (FG3) and in the funiculus ( Figure 2I). The YUC4 promoter [29] starts to be active at stage 3-II when the female gametophyte is at stage FG3 ( Figure 2J) and remains active till stage 3-VI ( Figure S3). Taken together, the data regarding auxin biosynthesis along with that on the localisation of the PIN1 and PIN3 proteins are consistent with auxin accumulation in the nucellus tissues.

PIN1 is Required for Megametogenesis Progression
To unravel the importance of PIN1 and PIN3 in ovule development, we adopted a genetic approach.
In an initial step, we analysed the seed set of 10 PIN1pin1 heterozygous plants (Gabi-KAT line GK_051A10). All ovules in these plants, half of which contain a pin1 female gametophyte, were successfully fertilized by wild-type pollen in reciprocal crosses. This excludes the possibility that the pin1 mutation has gametophytic effects.
pin1 homozygous plants develop a pin-shaped inflorescence meristem devoid of flowers [11,12]. Thus we analysed the pin1-5 allele to elucidate the sporophytic role of PIN1 during ovule development [13,30]. pin1-5 plants develop flowers containing very few ovules compared to wild-type. In an analysis of 183 ovules, Bencivenga and co-workers [4] reported that 10% of them develop as finger-like structures, whilst 70% showed defects in megagametogenesis since most of the embryo sacs arrested in FG1 or FG3, corresponding to ovule developmental stages 3-I and 3-III ( Figures 3A and 3B, Table S2A) [1,4].
The phenotype described in the pin1-5 mutant could be due to the additive effect of reduced auxin flux in the developing pistil and reduced auxin flux during ovule development. To uncouple the two possible causes of the observed phenotype in pin1-5 female gametophytes, we silenced PIN1 with an artificial micro-RNA (amiPIN1; [20]) under the control of the ovule-specific promoter pDEFH9 [15]. DEFH9 is an Antirrhinum majus ovule-specific gene and its promoter is active in Arabidopsis ovules from stage 1-II ( Figure S4). Among 50 DR5rev:GFP T 1 plants containing the pDEFH9:amiPIN1 construct, 18 were characterised by an abnormal seed set with a significant percentage of their ovules (from 26% to 72%) unable to complete proper development (five siliques per plant were manually dissected). Analysis of the T 2 population (T 1 pollen was used to pollinate wild-type plants) using BASTA selection showed that 15 out of the 18 plants analysed had only one T-DNA copy (or multiple T-DNA copies in linkage segregating as a single locus -data not shown). Optical microscopic analysis showed that ovules from these 15 transgenic plants were blocked at either the FG1 stage (10 lines) or at the FG3 stage (5 lines) (Figures 3C and 3D) resulting in a high percentage of ovule abortions (Table S2B). These data in conjunction with the observation that PIN1pin1 heterozygous plants have normal seed sets clearly suggest that the observed gametophytic defects in pDEFH9:amiPIN1 lines are of sporophytic origin. This is also in accord with the nature of the pin1-5 ovule phenotype.
To further support the contention that ovule abortion was due to PIN1 down-regulation, we performed real-time PCR analysis on the carpels of the T 2 segregating plants using either UBIQUITIN10 or 18S RNA as reference genes ( Figure S5). This analysis showed that PIN1 transcript levels were reduced by 2 to 5 fold in those pDEFH9:amiPIN1 plants producing defective female gametophytes ( Figures 3C and D) in comparison to sibling wild-type plants. Interestingly, analysis of the DR5rev:GFP reporter in the PIN1 silenced lines showed that those ovules that were unable to complete megagametogenesis maintained an auxin maximum at the distal edge of the blocked embryo sacs. Such an auxin response is not observed in wild-type sister FG5 ovules ( Figure S6). The persistence of auxin activity was associated with the block of gametogenesis, however the sporophytic tissues of these ovules proceeded with development and reached maturity. These data suggest that PIN1 is involved in auxin flux towards the female gametophyte which is required for gametophyte progression.

PIN3 Function is not Required for Ovule Development
PIN3 is also expressed during ovule development as demonstrated in pPIN3:PIN3-GFP plants. To elucidate a potential role for PIN3 in ovule development we have analysed pin3-4 homozygous mutant plants. These plants do not show any ovule defects and are fully fertile. To assess whether PIN3 plays a role in ovule development in pin1 down-regulated plants, we transformed wildtype plants with a micro RNA against PIN3 and PIN1 under the control of the DEFH9 promoter (pDEFH9:amiPIN1-3). We selected 9 transgenic plants characterised by the presence of ovule defects (Table S2C). In these plants, both PIN1 and PIN3 were downregulated and the ovule arrest in, FG1 or FG3 ( Figure 3F) pin1-5 and pDEFH9:amiPIN1 plants. The percentage of ovule abortions (values varied greatly in different plants ranging from 45% to 70%) is not significantly different compared to the PIN1 silenced plants, indicating that PIN1 is the major player in female gametophyte progression.

PIN1 is Expressed in Ovule Sporophytic Cells and Promotes Female Gametophyte Development
Based on our data on local auxin synthesis, PIN localization and auxin response patterns we propose a dynamic model describing auxin flux during ovule development ( Figure 4A-D).
Shortly after ovule primordium formation, PIN1 is detectable in the nucellus cells. PIN1 membrane localisation clearly indicates the direction of auxin flux in the nucellus ( Figure 4A-C), leading to auxin accumulation in specific nucellus cells ( Figure 4A-C).
The location of PIN1 in the nucellus cells ( Figure 4A-C) suggests that either auxin could be transported into the forming embryo sacs and regulate gametophytic progression, or alternatively it might regulate the expression of unidentified factor/s in the nucellus that directly or indirectly promote megagametogenesis. Identification of the target genes regulated by the auxin response factors required for female gametophyte development will contribute to discriminate between the two possibilities.
The accumulation of auxin in the female gametophyte was described previously by Pagnussat and co-workers [14], however we did not detect an auxin responce in the MMC or in female gametophytes ( Figure 4A-D) despite the long half-life of the reporter protein GFP [31]. Nevertheless, our observations clearly demonstrate that sporophytic PIN1 expression in the ovule ( Figure 4A-C) is needed for the early phases of embryo sac formation since PIN1 down-regulation compromises embryo sac development. The importance of PIN1 in early and later stages of female gametophytic development has been proposed previously: the disruption of cytokinin signalling in the ovule's sporophytic tissue results in the down-regulation of PIN1 and the lack of embryo sac formation [4]. It has been shown that PIN1 expression is under the control of SPOROCYTLESS/NOZZLE (SPL/NZZ) [4]. Interestingly, while spl ovules do not develop the MMC [32], the majority of ovules in both pin1-5 and pDEFH9:amiPIN1 plants are able to [1]. Thus whilst PIN1 plays a key role in promoting megagametogenesis, this observation suggests that MMC formation does not require PIN1 expression.
Although PIN1 seems to be the only member of the PIN protein family to play an important role in ovule development, we cannot at present exclude the possibility that other auxin transporters are also required to dynamically coordinate auxin flux, such as the ABC transporters [33].

Auxin Might Play a Major Function in Haploid/diploid Generation Cross-talk
The importance of dynamic gradients is strengthened by the tight temporal and spatial regulation of auxin biosynthesis in ovules ( Figure 4A-D). For instance, the expression of both TAA1 and YUC4 genes is restricted to a small group of cells during ovule development. TAA1 is expressed within the ovule primordium and at later developmental stages in the chalaza region of stage 2-IV ovules and in the developing inner integuments ( Figure 4A-D). YUC4 is expressed at later stages than TAA1, starting at FG2 and being restricted to the tip of developing inner integument ( Figure 4A-D).
Several genetic screenings (for a review see [34]) have been performed to identify genes involved in megagametophyte development. To date all the gametophytic mutants described develop ovule sporophytic tissues correctly. In contrast, several ovule mutants with sporophytic defects show anomalies in gametophyte development. Therefore, it has been proposed that a hierarchy exists in the communication between ovule sporophytic tissues and the gametophytic haploid generation, and that greater importance is attributed to the sporophytic maternal tissues.
We propose that auxin (directly or indirectly) might have an important function in the communication between generations.
In conclusion, we propose that auxin (directly or indirectly) has an important function in the communication between generations. Previous work has shown that auxin is involved in several aspects of ovule development such as ovule primordium formation [35] and the determination of embryo sac cell identity [14]. The results presented in the current work highlight the involvement of auxin flux modulation in the progression of megagametogenesis. ovules at the corresponding stages. In C and D the endothelium begins to develop (see arrow) whereas the nucellar cells degenerate and accordingly the fluorescent nuclear auxin response is no longer detectable, though the auxin response is still visible in the funiculus (asterisk). Scale bars: 20 mm fg, female gametophyte; ii, inner integument; oi, outer integument; fun, funiculus; nu, nucellus; end, endothelium. (TIF) Figure S3 YUC4 is expressed in developing ovules. The YUC4 promoter (pYUC4) does not drive reporter gene expression (GUS) in developing ovules between stages 1-I (A) and 3-I (C). Reporter gene activity begins to be detected in ovules at stage 3-II (D) and it is maintained until stage 3-V (E). fg, female gametophyte; ii, inner integument; oi, outer integument; fun, funiculus Scale bars: 20 mm. (TIF) Figure S4 pDEH9 is an ovule-specific promoter also in Arabidopsis thaliana. The Antirrhinum majus DEF9 promoter (pDEH9) drives reporter gene expression (GUS) only in developing ovules, the promoter being active from stage II (A) to stage 3-V (E). fg, female gametophyte; ii, inner integument; oi, outer integument; fun, funiculus; nu, nucellus. (TIF) Figure S5 PIN1 expression in pDEFH9:amiPIN1 flowers. Realtime PCR to evaluate PIN1 expression in pDEFH9:amiPIN1 flowers. Two pairs of PIN1 specific primers were employed (orange and green). (TIF) Figure 4. Model of auxin flux and production during ovule development. The diagram summarises the auxin response (in green), auxin flux (red arrows) and auxin synthesis (grey) in developing ovules from stage 2-I (A) to stage 3-IV (D). At stage 2-I (A) the megaspore mother cell (MMC) is recognisable. At this stage the auxin response (DR5rev::GFP) is restricted to a few cells of the nucellus; PIN1 is also detected in the nucellar cells and has is a basal polar localization. At this stage pTAA1 drives reporter expression in the ovule chalaza region and in the funiculus. At stage 2-III (B) PIN1 is detected in the inner integument primordia and in the funiculus, comcometeley with the auxin response. At stage 2-IV (C) the TAA1 promoter reduces its expression in the inner integument but at the tip of the inner integument YUC4 begins to be expressed. At this stage PIN1 and the auxin response are restricted to the funiculus since the nucellus is degenerating. The position of PIN1 in the nucellus cells strongly indicates its importance in auxin exit from these cells. The exit of auxin stimulates megagametogenesis as shown by PIN1 silencing. fg, female gametophyte; ii, inner integument; oi, outer integument; fun, funiculus; chal, chalaza. doi:10.1371/journal.pone.0066148.g004 Figure S6 The auxin response in normal and mutated ovules of pDEFH9:amiPIN1 plants. (A) The DR5rev:GFP promoter is active in mutated ovules unable to complete megagametogenesis. fg, female gametophyte; ii, inner integument; oi, outer integument; fun, funiculus; nu, nucellus Scale bars: 20 mm.

(TIF)
Table S1 A Analysis of pin1-5 developing ovules. 20 pistils collected from 4 independent plants have been analysed. B and C Analysis of T2 pDEFH9:amiPIN1 and pDEFH9:amiPIN1-3 plants. The total number of ovules and the percentage of ovule abortion are reported. 5 siliques per individual were observed, and 15 independent lines have been analysed. Aberrant embryo sac arrest at FG1 or FG3 stages (Figure 3). D Analysis of pin3-4 developing ovules. 10 pistils collected from 4 independent plants have been analysed. (DOCX)