The MYB98 subcircuit of the synergid gene regulatory network includes genes directly and indirectly regulated by MYB98

Authors


*(fax +1 801 581 4668; e-mail drews@bioscience.utah.edu).

Summary

The female gametophyte contains two synergid cells that play a role in many steps of the angiosperm reproductive process, including pollen tube guidance. At their micropylar poles, the synergid cells have a thickened and elaborated cell wall: the filiform apparatus that is thought to play a role in the secretion of the pollen tube attractant(s). MYB98 regulates an important subcircuit of the synergid gene regulatory network (GRN) that functions to activate the expression of genes required for pollen tube guidance and filiform apparatus formation. The MYB98 subcircuit comprises at least 83 downstream genes, including 48 genes within four gene families (CRP810, CRP3700, CRP3730 and CRP3740) that encode Cys-rich proteins. We show that the 11 CRP3700 genes, which include DD11 and DD18, are regulated by a common cis-element, GTAACNT, and that a multimer of this sequence confers MYB98-dependent synergid expression. The GTAACNT element contains the MYB98-binding site identified in vitro, suggesting that the 11 CRP3700 genes are direct targets of MYB98. We also show that five of the CRP810 genes, which include DD2, lack a functional GTAACNT element, suggesting that they are not directly regulated by MYB98. In addition, we show that the five CRP810 genes are regulated by the cis-element AACGT, and that a multimer of this sequence confers synergid expression. Together, these results suggest that the MYB98 branch of the synergid GRN is multi-tiered and, therefore, contains at least one additional downstream transcription factor.

Introduction

The angiosperm female gametophyte is composed of one egg cell, two synergid cells, one central cell and three antipodal cells (Drews and Yadegari, 2002; Yadegari and Drews, 2004). The synergid cells lie adjacent to the egg cell and play a role in many steps of the fertilization process, including pollen tube guidance, arrest of pollen tube growth within the female gametophyte, pollen tube discharge and sperm cell migration. At their micropylar poles, the synergid cells contain a thickened and elaborated cell wall, the filiform apparatus, which is thought to function in pollen tube reception and in the secretion of the pollen tube attractant(s) (reviewed in Higashiyama, 2002; Punwani and Drews, 2008).

The molecular processes by which the synergid cells acquire their unique physiological and structural features during cell differentiation are not yet understood. Understanding these processes requires mechanistic insight into the gene regulatory networks (GRNs) that control synergid-specific gene expression. As a first step towards dissecting the synergid GRN, we identified a collection of genes expressed exclusively or predominantly in the synergid cells (Steffen et al., 2007). One of these, MYB98, encodes an R2R3 MYB transcription factor (Kasahara et al., 2005). MYB98 is localized to the nuclei of the synergid cells, and binds to a specific DNA sequence in vitro (Punwani et al., 2007). myb98 synergid cells have defects in filiform apparatus structure and pollen tube guidance, but are otherwise normal (Kasahara et al., 2005). Together, these data suggest that MYB98 controls a subcircuit of the synergid GRN, and functions to activate the expression of genes required for filiform apparatus formation and pollen tube guidance.

To identify MYB98-regulated genes, we (Punwani et al., 2007) and others (Jones-Rhoades et al., 2007) performed differential expression screens and identified 83 genes exhibiting reduced expression in myb98 ovules, relative to wild-type ovules. Sixteen of these genes were directly tested for expression in wild-type and myb98 synergid cells, and all were found to be downregulated in the mutant cells (Punwani et al., 2007). We refer to these genes collectively as DIM (downregulated in myb98) genes. Features of the DIM genes are summarized in Table S1.

Most (59/83) of the DIM genes encode cysteine-rich peptides (CRPs; Table S1). The CRP genes in the Arabidopsis genome have been classified into subgroups by Silverstein et al. (2007). Of these, the CRP810, CRP3700, CRP3730 and CRP3740 subgroups appear to be especially important for the synergid cell, because all or most genes within these subgroups exhibit the DIM expression pattern, and these subgroups contribute most (48/59) of the synergid CRPs. These genes encode defensin-like (CRP810) and lipid transfer protein (LTP)-like (CRP3700, CRP3730 and CRP3740) proteins (Silverstein et al., 2007). Defensins and LTPs have been shown to function as antimicrobial peptides (Dresselhaus, 2006; Ganz, 2003; Garcia-Olmedo et al., 1995; Molina et al., 1993; Terras et al., 1992), suggesting that the genes in these four families function in defense, and that an important function of the synergid cell, and of the MYB98 subcircuit in particular, is defense.

Thus, MYB98 controls an important subcircuit of the synergid GRN, and understanding how this subcircuit is regulated is essential for understanding synergid differentiation. To this end, we extend our analysis of the MYB98 subcircuit, and focus on the regulation of two important gene groups: CRP810 and CRP3700. We show that some of the DIM genes are directly regulated by MYB98 through the cis-element GTAACNT, and that other DIM genes are indirectly regulated by MYB98 through the cis-element AACGT. These results indicate that the MYB98 subcircuit of the synergid GRN is multitiered, and, therefore, contains at least one additional transcription factor downstream of MYB98.

Results

The DD11 promoter contains a GTAACNT element necessary for MYB98-dependent synergid expression

We previously dissected the DD11 promoter, and showed that the region between −174 and −139 bp (GAAATGTGACTTAATTATTAATCAAGTAAGTAACAT), relative to the translational start sequence, is necessary for synergid expression (Punwani et al., 2007). This region contains a TAAC element at position −144 bp that binds MYB98 in vitro, and is necessary for synergid expression (Punwani et al., 2007). However, these studies did not exclude the possibility that this region contains additional sequences necessary for synergid expression. To determine whether this is the case, we generated a series of small (8–10 bp) 5′ deletions of this region. We fused these promoter fragments with a GFP coding region, introduced these promoter:GFP constructs into wild-type Arabidopsis plants and scored the expression in the synergid cells.

Figure 1(a) shows that a promoter fragment containing 146-bp upstream of the translational start site was sufficient to drive expression of GFP in the synergid cells. By contrast, the removal of an additional 9 bp from the 5′-end abolished expression. These data indicate that sequences required for synergid expression lie within the 8-bp region between −146 and −139 bp, relative to the translational start site.

Figure 1.

 The DD11 promoter contains a GTAACNT element necessary for synergid expression.
(a) Expression of the promoter:GFP constructs. Schematics of the constructs are shown to the left. The promoter-deletion constructs contain 5′ endpoints from −174 to −137 bp, relative to the translational start site, a 3′ endpoint of +21 bp, relative to the translational start site, and are fused with a GFP coding region. The 7xMS1 construct contains seven copies of the sequence TAAGTAACATAT (open circles) fused with mpGFP. The 7xMS1m construct contains seven copies of the sequence TAAGTGGGATAT fused with mpGFP. Expression levels of the corresponding constructs in transgenic plants are shown to the right.
(b) Expression of ProDD11:GFP constructs containing single base-pair mutations in the 8-bp region identified in (a). These constructs contain point mutations in the 8-bp region, and are otherwise identical to the −174-bp 5′ promoter-deletion construct shown in (a). The wild-type sequence is shown in the top row. Mutated nucleotides are highlighted in gray. Expression levels of the corresponding constructs in transgenic plants are shown to the right.
(c) Expression of the 7xMS1 construct in wild-type and myb98 ovules. The bar graph represents the percentage of synergid cells expressing GFP in MYB98/MYB98, myb98-1/MYB98 and myb98-1/myb98-1 pistils.
In (a) and (b), expression levels are indicated as follows: + indicates that a strong and consistent (detected in >50% of ovules) GFP signal was detected in the synergid cells, (+) indicates that an extremely weak and inconsistent (detected in <25% of ovules) GFP signal was detected in the synergid cells, and − indicates that GFP was not detected in the synergid cells in any transgenic plants, or that an extremely weak and inconsistent GFP signal was detected in only one transgenic plant. The ratio to the far right indicates the number of transgenic lines expressing GFP (numerator) and the total number of transgenic lines analyzed (denominator).

This 8-bp region consists of the sequence AGTAACAT, and contains the TAAC element that we previously showed to be necessary for synergid expression. We previously showed that mutating the core AAC nucleotides abolishes both MYB98 binding in vitro and synergid expression in vivo (Punwani et al., 2007), suggesting that this region contains an in vivo MYB98-binding site. To further define this site, we mutated the nucleotides surrounding the core AAC residues. Figure 1(b) shows that mutations in the T residues at positions 3 and 8, and in the G residue at position 2, abolished or dramatically reduced synergid expression. By contrast, mutations in the A residues at positions 1 and 7 had no effect on synergid expression. Together, these data suggest that the in vivo MYB98 DNA-binding site is GTAACNT (where N = any nucleotide).

A multimer of the GTAACNT element confers MYB98-dependent synergid expression

To determine whether the GTAACNT element is sufficient to direct expression to the synergid cells, we fused seven copies of this sequence with the mpGFP construct, which contained a GFP coding region fused at its 5′ end with a minimal promoter derived from the cauliflower mosaic virus 35S promoter (Benfey and Chua, 1990), and introduced this construct (Pro7xMS1:mpGFP, where MS1 is the multimerized sequence 1) into wild-type plants. We also analyzed the expression of a control construct (Pro7xMS1m:mpGFP) containing seven copies of a mutated version of MS1 (MS1m). Figure 1(a) shows that Pro7xMS1:mpGFP, but not Pro7xMS1m:mpGFP, was expressed in the synergid cells.

To determine whether expression of Pro7xMS1:mpGFP is MYB98 dependent, we analyzed expression of this construct in myb98-1 female gametophytes. Figure 1(c) shows that myb98 synergid cells did not express Pro7xMS1:mpGFP, indicating that MYB98 is required for the expression of this construct. Together, these data show that multimerization of the GTAACNT element confers MYB98-dependent synergid expression.

To determine whether MYB98 is sufficient to activate the expression of the DIM genes, we fused the MYB98 coding region with the constitutive UBIQUITIN 10 promoter (Sun and Callis, 1997), and used real-time RT-PCR to assay the expression of the DIM genes in transgenic seedlings containing this construct. Table S2 shows that strong MYB98 expression was detected in these plants. By contrast, no expression of any of the DIM genes tested was detected, indicating that ectopic expression of MYB98 in seedlings is not sufficient to activate the expression of the DIM genes in these cells.

The genes within the CRP3700 subgroup contain a GTAACNT element necessary for synergid expression

DD11 is a member of the CRP3700 subgroup, which contains 11 genes (Silverstein et al., 2007). All 11 genes exhibit the DIM pattern, and we previously showed that three of these (DD11, DD18 and DD31) are expressed in the synergid cells (Punwani et al., 2007; Steffen et al., 2007). These observations suggest that the CRP3700 subgroup comprises a group of co-expressed genes.

Figure S1 shows that 10 of the 11 genes within the CRP3700 subgroup (all except DD11) have promoters that are highly similar within a region of ∼300-bp upstream of the translational start site. Upstream of position −300 bp, the promoters are much less similar. Furthermore, the highest similarity occurs within a region of ∼200-bp upstream of the translational start site.

These observations suggest that these 10 genes may be regulated by common cis-elements, and that these cis-elements lie within the 200–300-bp region of sequence similarity. To test this, we generated promoter:GFP constructs for nine of these genes, in which promoter fragments containing ∼200- and/or ∼300-bp upstream of the translational start site were used. Figure 2 shows that all of these constructs were expressed in the synergid cells. Furthermore, in the three cases (DD18, DD31 and At2g4037) in which both ∼200- and ∼300-bp promoter fragments were tested, differences in GFP expression levels were not detected. Together, these data indicate that the sequences necessary for synergid expression lie within ∼200-bp upstream of the translational start site, which is the region of maximal sequence similarity.

Figure 2.

 The CRP3700 promoters contain GTAACNT elements necessary for synergid expression.
Expression of promoter:GFP constructs. Schematics of these constructs are shown to the left. The 5′ endpoints (relative to the translational start site) are shown. In the DD18m1, DD31m, At2g04031m and At5g35405m constructs, the downstream GTAACNT element is mutated to GTGGGNT. Expression levels of the corresponding constructs in transgenic plants are shown to the right. Expression levels are indicated as follows: + indicates that GFP was detected in the synergid cells in some transgenic plants; − indicates that GFP was not detected in the synergid cells in any transgenic plants; the numerator of the ratio indicates the number of transgenic lines expressing GFP, and the denominator of the ratio indicates the total number of transgenic lines analyzed.

Figure S1 shows that these 10 genes contain one or two GTAACNT elements within their promoters, suggesting that they are regulated by this element. Furthermore, the conservation of the downstream GTAACNT element suggests that it is of primary importance. The absence of the upstream GTAACNT element in some genes suggests that it is non-functional; consistent with this, deletion of this element in the promoters of DD31, At2g04031, At2g04037 and At5g35405 did not cause a loss of synergid expression (Figure 2).

To test the functionality of the conserved (downstream) GTAACNT element, we mutated this element in the promoters of DD18, DD31, At2g04031 and At5g35405. Figure 2 shows that mutating the conserved GTAACNT element in these genes abolished expression, indicating that this element is essential for synergid expression.

In summary, 10 of the 11 genes within the CRP3700 subgroup have promoters that are highly similar, these 10 genes have GTAACNT elements in a conserved position and the conserved GTAACNT element is essential for synergid expression in all four genes tested. These data suggest that the conserved GTAACNT element is functional in all 10 genes. The eleventh gene of the CRP3700 group is DD11, which is also regulated by a GTAACNT element. Together, these data suggest that the 11 CRP3700 genes are co-regulated by MYB98 binding to a common cis-element: GTAACNT.

The DD2 promoter contains an AACGT element necessary and sufficient for synergid expression

Figure S2 shows that the CRP810 subgroup contains a group of DIM genes with highly similar promoters. Within a region of ∼210-bp upstream of the translational start site, the promoters of these five genes are nearly identical. Upstream of position −210 bp, the promoters are much less similar.

These observations suggest that the cis-elements required for synergid expression lie within the ∼210-bp region of sequence similarity. To test this, we chose one of these genes, DD2, and generated a promoter:GFP construct using 210-bp upstream of the translational start site. Figure 3(a) shows that this construct was expressed in the synergid cells, indicating that sequences necessary for synergid expression lie within the region of high sequence similarity.

Figure 3.

 The DD2 promoter contains an AACGT element necessary for synergid expression.
(a) Expression and of the DD2m1 construct. Schematics of these constructs are shown to the left. The 5′ endpoints (relative to the translational start site) are shown. In the DD2m1 construct, the GTAACNT element is mutated to GTGGGNT and is otherwise identical to the −210-bp promoter-deletion construct. Both promoter fragments have 3′ endpoints of +21-bp, relative to the translational start site, and are fused with a GFP coding region. Expression levels of the corresponding constructs in transgenic plants are shown to the right.
(b) and (c) Expression of the 5′ (b) and 3′ (c) promoter-deletion constructs. Schematics of the constructs are shown to the left. The 5′ promoter-deletion constructs (b) contain 5′ endpoints from −200 to −68 bp, relative to the translational start site, a 3′ endpoint of +21 bp, relative to the translational start site, and are fused with a GFP coding region. The 3′ promoter-deletion constructs (c) contain a 5′ endpoint of −250 bp, relative to the translational start site, 3′ endpoints of between −1 and −199 bp, relative to the translational start site, and are fused with mpGFP. Expression levels of the corresponding constructs in transgenic plants are shown to the right.
(d) Sequences of the regions identified as necessary for synergid expression in the 5′ (top) and 3′ (bottom) deletions of the DD2 promoter. These regions overlap by 5 bp (positions between −163 and −159 bp), identifying AACGT as the sequence necessary for synergid expression.
(e) Expression of the DD2m2 and Pro6xMS2:mpGFP constructs. Schematics of these constructs are shown to the left. In DD2m2, the AACGT element is mutated to AGGGG, and is otherwise identical to the −170-bp 5′ promoter-deletion construct in (b). The Pro6xMS2:mpGFP construct contains six copies of the sequence AAAACGTTTAAATG fused with mpGFP. Expression levels of the corresponding constructs in transgenic plants are shown to the right.
In (a), (b), (c) and (e), expression levels are indicated as follows: + indicates that GFP was detected in the synergid cells in some transgenic plants; − indicates that GFP was not detected in the synergid cells in any transgenic plants; the numerator of the ratio indicates the number of transgenic lines expressing GFP, and the denominator of the ratio indicates the total number of transgenic lines analyzed.

Figure S2 shows that the five CRP810 genes contain a conserved GTAACNT element within the promoter regions of high sequence similarity. To determine whether this GTAACNT element is functional, we tested the effect of mutating it in DD2. Figure 3(a) shows that mutating the GTAACNT element in DD2 did not affect expression, indicating that this element is non-fuctional.

These results suggest that DD2 is not directly regulated by MYB98, and that other cis-elements are required for synergid expression. To identify these elements, we generated a series of 5′ (Figure 3b) and 3′ (Figure 3c) deletions of the DD2 promoter, fused these with a GFP coding region, introduced these promoter:GFP constructs into wild-type Arabidopsis plants and analyzed their expression in the synergid cells.

With the 5′ deletion series, a promoter fragment containing 163-bp upstream of the translational start site was sufficient to drive expression of GFP in the synergid cells, but removal of an additional 13 bp (to position −150 bp) abolished this expression (Figure 3b). With the 3′ deletion series, a promoter fragment containing from −250 to −159 bp, relative to the translational start site, was sufficient to drive the expression of GFP in the synergid cells, but removal of an additional 20 bp from the 3′ end (to position −179 bp) abolished this expression (Figure 3c).

The regions identified in the 5′ and 3′ deletions are compared in Figure 3(d). These regions overlap by 5 bp, at positions from −163 to −159 bp, suggesting that this 5-bp region contains sequences necessary for synergid expression. To confirm that this 5-bp sequence is necessary for DD2 expression, we mutated this sequence in the context of a promoter fragment containing 170-bp upstream of the translational start site, and introduced this construct into wild-type plants. Figure 3(e) shows that this mutation abolished synergid expression, indicating that the AACGT element is necessary for DD2 expression.

To determine whether the AACGT element is sufficient to direct expression to the synergid cells, we fused six copies of this sequence to the mpGFP construct, and introduced this construct (Pro6xMS2:mpGFP) into wild-type plants. Figure 3(e) shows that Pro6xMS2:mpGFP was expressed in the synergid cells, indicating that multimerization of AACGT confers synergid expression.

In summary, five genes within the CRP810 subgroup have promoters that are nearly identical, the promoters of these five genes have GTAACNT and AACGT elements in a conserved position, and in one of these genes, DD2, the GTAACNT element is non-functional and the AACGT element is functional. These data suggest that all five of these genes are indirectly regulated by MYB98 through the cis-element AACGT.

The AACGT element is necessary for synergid expression in the DD18 gene

Figure S1 shows that most genes within the CRP3700 subgroup have an AACGT element at a conserved position. To evaluate the role of the AACGT element in this gene group, and to identify any additional sequences necessary for synergid expression, we chose one gene, DD18, and dissected its promoter using a strategy similar to that described previously for DD2 (Figure 3) and DD11 (Figure 1). Figure 4(a) shows that a promoter fragment containing 186-bp upstream of the translational start site was sufficient to drive strong expression of GFP in the synergid cells. Removal of an additional 8 bp (to position −178 bp) from the 5′ end caused reduced expression. Further deletion to position −151 bp caused no further reduction in expression. However, removal of an additional 10 bp (to position −141 bp) abolished expression (Figure 4a).

Figure 4.

 The DD18 promoter contains an AACGT element necessary for synergid expression.
(a) Expression of the 5′ promoter-deletion constructs. Schematics of the constructs are shown to the left. These constructs contain 5′ endpoints from −193 to −100 bp, relative to the translational start site, a 3′ endpoint of +18 bp, relative to the translational start site, and are fused with a GFP coding region. Expression levels of the corresponding constructs in transgenic plants are shown to the right.
(b) Sequence of the regions identified as necessary for synergid expression in the 5′ deletions in (a). The region from −186 to −178 bp contains an AACGT element (bold), and the region from −151 to −141 bp contains a GTAACNT element (bold).
(c) Expression of the DD18m2 construct. A schematic of this construct is shown to the left. In DD18m2, the AACGT element is mutated to AGGGG and is otherwise identical to the −193-bp 5′ promoter-deletion construct in (a). Expression levels of the corresponding constructs in transgenic plants are shown to the right.
In (a) and (c), expression levels are indicated as follows: + indicates that a strong and consistent (detected in >50% of ovules) GFP signal was detected in the synergid cells, (+) indicates that an extremely weak and inconsistent (detected in <25% of ovules) GFP signal was detected in the synergid cells, and − indicates that GFP was not detected in the synergid cells in any transgenic plants. The ratio to the far right indicates the number of transgenic lines expressing GFP (numerator), and the total number of transgenic lines analyzed (denominator).

These data identify two regions of the DD18 promoter that are required for synergid expression. Figure 4(b) shows that the deletion from −186 to −178 bp removed the AACGT element identified as necessary for synergid expression of DD2, and that the deletion from −151 to −141 bp removed most of the GTAACNT element identified as necessary for synergid expression of DD18 (Figure 2).

To confirm that the AACGT element is necessary for DD18 expression, we mutated this sequence in the context of a promoter fragment containing 193-bp upstream of the translational start site, and introduced this construct into wild-type plants. Figure 4(c) shows that the wild-type promoter fragment was sufficient to drive expression within the synergid cells. By contrast, the same construct containing a mutation in AACGT had dramatically reduced expression in the synergid cells (Figure 4c). These data indicate that the AACGT element is necessary for DD18 expression.

Discussion

The synergid cells are important and unique cells within the plant (Punwani and Drews, 2008). They are part of the haploid gametophyte generation, are structurally specialized and control many steps of the angiosperm fertilization process, including pollen tube guidance, pollen tube growth arrest and discharge, and sperm cell migration (Punwani and Drews, 2008).

We are dissecting the synergid GRN to understand how the synergid cells acquire their unique features and functions. GRNs generally are modular in structure and are composed of subcircuits that perform specific developmental or functional tasks (Oliveri and Davidson, 2007). We previously showed that myb98 synergid cells are largely normal (Kasahara et al., 2005), and that they express non-MYB98-regulated genes (Punwani et al., 2007), suggesting that MYB98 does not lie at the top of the synergid GRN, and that instead it controls a subcircuit within the synergid GRN.

The MYB98 subcircuit is large: consisting of at least 83 genes (Table S1), and presumably including more. In prior studies, synergid expression was demonstrated with 20 (Table S1) of the DIM genes (Jones-Rhoades et al., 2007; Punwani et al., 2007). Here, we analyzed the expression of seven additional genes, and showed that all are expressed in the synergid cells (Figures 2c and 3c). Thus, of the 83 DIM genes, 27 have been tested and all are expressed in the synergid cells, suggesting that the remaining 56 also exhibit synergid expression.

Based on the myb98 phenotype (Kasahara et al., 2005), it is likely that the MYB98 subcircuit functions to coordinate the expression of genes required for filiform apparatus formation and pollen tube guidance. An additional function of the MYB98 subcircuit may be to activate genes required for defense. Of the 83 DIM genes identified (Table S1), over half (48 genes) belong to four gene families (CRP810, CRP3700, CRP3730 and CRP3740) that encode CRPs (Table S1), similar to known antimicrobial peptides (Silverstein et al., 2007). Antimicrobial peptides may be necessary to defend against pathogens associated with pollen tube growth through the female tissue (Dresselhaus, 2006).

The data presented here allow us to begin to assemble the MYB98 subcircuit of the synergid GRN (Figure 5). Our analysis has identified two cis-elements – GTAACNT and AACGT – that are important for synergid expression. Our analysis also indicates that these cis-elements can act together (DD18) or independently (DD2 and DD11) to direct expression to the synergid cells.

Figure 5.

 Model for the MYB98 subcircuit of the synergid gene regulatory network.
Arrows indicate that the expression of the downstream gene is induced by the transcription factor. MYB98 acts through the sequence GTAACNT. An unknown transcription factor acts through the sequence AACGT.

Several observations suggest that the GTAACNT element comprises the in vivo MYB98-binding site. First, the GTAACNT element is similar to the in vivo DNA-binding sequences of other MYB proteins in both plants and animals. Mammalian c-Myb binds to the consensus sequence YAACNG (where Y = C or T, and N = any nucleotide) (Oda et al., 1998). In addition, similar in vivo sites are bound by several plant R2R3-MYB proteins, including WEREWOLF (Koshino-Kimura et al., 2005; Ryu et al., 2005), MYB2 (Abe et al., 1997), GAMyb (Gubler et al., 1995), MYBGA (Chen et al., 2006), C1 (Roth et al., 1991) and P (Grotewold et al., 1994). Second, the GTAACNT element in the DD11 promoter binds MYB98 in vitro (Punwani et al., 2007). Third, mutations throughout the GTAACNT element abolish both synergid expression of DD11 (Figure 1b) and MYB98 binding in vitro (Punwani et al., 2007). Finally, multimerization of the GTAACNT element confers MYB98-dependent synergid expression (Figure 1c). Together, these data suggest very strongly that the in vivo MYB98-binding site is GTAACNT, although we cannot eliminate the possibility that MYB98 activates the expression of another MYB protein that binds to GTAACNT in the DD11 promoter.

Although multimerization of the GTAACNT element is able to confer MYB98-dependent synergid expression (Figure 1a,c), ectopic expression of MYB98 in seedlings is not sufficient to activate the DIM genes (Table S2). These results suggest that MYB98 activation of target genes may require co-factors that are present in the synergid cells, but not in seedling cells.

Currently, the transcription factor that binds the AACGT element is unknown. The AACGT element resembles the core binding site of bZIP transcription factors (Jakoby et al., 2002), suggesting that a bZIP transcription factor may play a role within the MYB98 subcircuit. However, at the present time, a synergid-expressed bZIP gene has not been identified. Multimerization of the AACGT element confers synergid expression (Figure 3e), indicating that this element binds a synergid-specific factor. Furthermore, it is likely that the gene encoding the AACGT binding factor is downstream of MYB98 within the synergid GRN, based on the observation that DD2 expression, which is regulated by the AACGT element, but not by the GTAACNT element, is abolished by the myb98 mutation (Punwani et al., 2007).

Together, these observations lead to the subcircuit structure shown in Figure 5. It is likely that additional DIM genes are regulated by the GTAACNT and AACGT elements. Of the 83 DIM genes identified, 46 have GTAACNT elements and 63 have AACGT elements within 500-bp upstream of the ATG. However, as shown with DD2(Figures S2,Figures 3a) and with several CRP3700 genes (Figure 2), not all GTAACNT and AACGT elements are functional. Further dissection of the MYB98 subcircuit will require identification of the cis-elements required for synergid expression in all of the DIM genes, and of the transcription factors that bind to these cis-elements.

Experimental procedures

Plant materials and growth conditions

Arabidopsis thaliana seeds were surface sterilized with chlorine gas and were then germinated on plates containing 0.5×Murashige and Skoog salts (M-9274; Sigma-Aldrich, http://www.sigmaaldrich.com), 0.05% 2-(N-morpholino)-ethane-sulfonic acid (MES), 1.0% sucrose and 0.8% Phytagar (Life Technologies, http://www.invitrogen.com). Ten-day-old seedlings were transferred to Sunshine Mix #2 supplemented with MiracleGro, and were grown under 24-h illumination.

Plant transformation

T-DNA constructs were introduced into Agrobacterium tumefaciens strain LBA4404 by electroporation. Arabidopsis plants (accession Columbia-0) were transformed using a modified floral-dip procedure (Clough and Bent, 1998). Transformed progeny were selected by germinating surface-sterilized T1 seeds on growth medium containing antibiotics (30 μg ml−1 kanamycin sulfate), supplemented with 15 μg ml−1 cefeotaxime. Resistant seedlings were transplanted to soil 10 days after germination. Transgene identity was verified by PCR using a gene-specific primer and a T-DNA-specific primer.

Promoter alignments

Promoter sequences were obtained from The Arabidopsis Information Resource (http://www.arabidopsis.org) and were aligned using ClustalW (http://www.ebi.ac.uk/clustalw) with manual adjustments.

Analysis of GFP expression patterns

Plants containing GFP constructs were analyzed using a Zeiss Axioplan compound microscope. GFP was excited using a UV lamp and was then detected with the 38 HE EGFP filter set. Images were captured using an Axiocam MRm REV2 camera with the AxioVision software package v4.5 (Zeiss, http://www.zeiss.com). Analysis was performed at 24 h after emasculation of stage-12c flowers (Christensen et al., 1997). For analysis of GFP fluorescence in the synergid cells, we followed the dissection procedure outlined by Steffen et al. (2007).

Promoter fusion constructs

Promoter fragments were generated by PCR using the primers listed in Tables S3 and S4. The primers introduced unique restriction enzyme sites. The PCR fragments were digested and ligated with pBI101.GFP (for 5′ deletions) or 35SmpGFP (for 3′ deletions; Yadegari et al., 2000). All constructs were sequence verified.

Mutant promoter reporter constructs were made either by introducing mutations into the forward primer, or by a two-step PCR mutagenesis. PCR mutagenesis was carried out using two primary reactions and a secondary annealing reaction. The first primary reaction used a wild-type forward primer and a mutation-bearing reverse primer; the second used a mutation-bearing forward primer with a wild-type reverse primer. These two PCR products were mixed and annealed in a PCR reaction using the outermost primers. Mutant PCR products were cloned into pBI101.GFP as described above.

Multimers were made by annealing oligonucleotides containing the multimerized sequence flanked by a restriction site compatible overhang. Annealed oligonucleotides were cloned into 35SmpGFP as described above. For Pro7xMS1:mpGFP, the oligonucleotides used were: DD11-7xTRF (5′-AGCTTAAGTAACATATTAAGTAACATATTAAGTAACATATTAAGTAACATATTAAGTAACATATTAAGTAACATATTAAGTAACATAT-3′) and DD11-7xTRR (5′-TCGAATATGTTACTTAATATGTTACTTAATATGTTACTTAATATGTTACTTAATATGTTACTTAATATGTTACTTAATATGTTACTTA-3′). For Pro7xMS1m:mpGFP, the oligonucleotides used were: DD11-7xTR(mut)F (5′-AGCTTAAGTGGGATATTAAGTGGGATATTAAGTGGGATATTAAGTGGGATATTAAGTGGGATATTAAGTGGGATATTAAGTGGGATAT-3′) and DD11-7xTR(mut)R (5′-TCGAATATCCCACTTAATATCCCACTTAATATCCCACTTAATATCCCACTTAATATCCCACTTAATATCCCACTTAATATCCCACTTA-3′). For Pro6xMS2:mpGFP, the oligonucleotides used were DD2-SE26xTR-F (5′-AGCTAAAACGTTTAAATGAAAACGTTTAAATGAAAACGTTTAAATGAAAACGTTTAAATGAAAACGTTTAAATGAAAACGTTTAAATG-3′) and DD2-SE26xTR-R (5′-TCGACATTTAAACGTTTTCATTTAAACGTTTTCATTTAAACGTTTTCATTTAAACGTTTTCATTTAAACGTTTTCATTTAAACGTTTT-3′).

Analysis of Pro7xMS1:mpGFP expression in MYB98/MYB98, MYB98/myb98 and myb98/myb98 pistils

We first identified a single-locus line exhibiting strong expression. T2 seed collected from individual T1 plants was plated on growth media containing kanamycin sulfate (30 μg ml−1), the resistant:sensitive (R:S) ratio was scored and lines with an R:S ratio of 3:1 were transplanted to soil. Eight plants from each line were scored for the percentage of synergid cells expressing GFP, to confirm that the reporter construct was present at only a single locus, and to identify homozygous plants. Transgenic line #1 was used in the crosses outlined below.

Plants homozygous for the Pro7xMS1:mpGFP transgene were used in a cross with myb98-1/myb98-1 pollen. The resulting F1 plants were hemizygous at the promoter:GFP locus and were heterozygous at the MYB98 locus. F1 plants were used as pollen donors in a second cross with myb98-1/myb98-1 females. The resulting F1 plants were genotyped by PCR to identify plants hemizygous for the promoter:GFP locus and homozygous for the myb98-1 mutation, using primers described in Kasahara et al. (2005).

When scoring the percentage of female gametophytes with GFP fluorescence in the synergid cells, at least three pistils were analyzed from at least three individuals per generation.

Acknowledgements

We thank Ramin Yadegari and Josh Steffen for their critical review of this manuscript. We thank Ramin Yadegari for providing the pBI101.GFP and 35SmpGFP vectors. We thank Judy Callis for providing a plasmid with the UBQ10 promoter (p3573). This work was supported by a National Science Foundation grant (grant no. IOB-0542953) to GND and a National Institutes of Health Developmental Biology Training Grant (grant no. 5T32 HD07491) appointment to JAP.

Ancillary