STENOFOLIA Regulates Blade Outgrowth and Leaf Vascular Patterning in Medicago truncatula and Nicotiana sylvestris C W OA

Dicot leaf primordia initiate at the ﬂanks of the shoot apical meristem and extend laterally by cell division and cell expansion to form the ﬂat lamina, but the molecular mechanism of lamina outgrowth remains unclear. Here, we report the identiﬁcation of STENOFOLIA (STF), a WUSCHEL-like homeobox transcriptional regulator, in Medicago truncatula , which is required for blade outgrowth and leaf vascular patterning. STF belongs to the MAEWEST clade and its inactivation by the transposable element of Nicotiana tabacum cell type1 ( Tnt1 ) retrotransposon insertion leads to abortion of blade expansion in the mediolateral axis and disruption of vein patterning. We also show that the classical lam1 mutant of Nicotiana sylvestris , which is blocked in lamina formation and stem elongation, is caused by deletion of the STF ortholog. STF is expressed at the adaxial–abaxial boundary layer of leaf primordia and governs organization and outgrowth of lamina, conferring morphogenetic competence. STF does not affect formation of lateral leaﬂets but is critical to their ability to generate a leaf blade. Our data suggest that STF functions by modulating phytohormone homeostasis and crosstalk directly linked to sugar metabolism, highlighting the importance of coordinating metabolic and developmental signals for leaf elaboration.


INTRODUCTION
Leaves are the principal organs for photosynthetic carbon assimilation. Independent origins of the flat lamina in different plant families provide evidence that it is an important evolutionary adaptation of land plants for efficient capture of solar energy and gaseous exchange. Dicot leaf primordia initiate at the flanks of the shoot apical meristem (SAM) and extend laterally during primary and secondary morphogenesis in which growth occurs predominantly by cell division and cell expansion, respectively (Sussex, 1955;Poethig, 1997;Scarpella et al., 2010). As the primordium extends, it asymmetrically differentiates into distinct upper (adaxial) and lower (abaxial) surfaces, forming a flattened lamina.
The first described dorsoventral polarity mutant in Antirrhinum, phantastica (phan), displays radialized abaxial leaves in severely affected cases and at restrictive growth temperatures (Waites and Hudson, 1995;Waites et al., 1998). PHAN encodes an MYB transcription factor required for adaxial identity and blade outgrowth (Waites et al., 1998). This was the first demonstration that the adaxial/abaxial identity sets a positional landmark for lateral outgrowth of the blade. However, loss-of-function mutants of PHAN homologs in other species have variable and less dramatic phenotypes, including crispa in pea (Pisum sativum) (Tattersall et al., 2005) and roughsheath2 in maize (Zea mays) (Timmermans et al., 1999). The Arabidopsis thaliana homolog in particular, the as1 mutant (Byrne et al., 2000), displays no obvious narrowing of lamina except occasionally in combination with the as2 mutant and in certain genetic backgrounds (Xu et al., 2003). AS2 encodes a LOB domain protein containing Leu-zipper motif and forms a complex with AS1 to promote adaxial cell fate (Lin et al., 2003;Xu et al., 2003) and KNOX1 repression (Guo et al., 2008;Jun et al., 2010).
In Nicotiana sylvestris, a unique lamina deletion mutant called lam1 has been described (McHale, 1992). Morphological and cellular studies demonstrated that mutant leaf primordia lacking LAM1 establish normal polarity, and blade founder cells are recruited in the correct position at the adaxial/abaxial boundary, but lateral outgrowth of an organized lamina fails (McHale, 1992(McHale, , 1993. This demonstrated that blade formation occurs in two distinct phases. The first step involves recruitment of a loosely organized group of blade founder cells at the adaxial/abaxial boundary of the primordium, a process that does not require LAM1. The second phase, lateral outgrowth and layer organization, however, is entirely dependent on LAM1 function. Later work on periclinal chimeras revealed that wild-type LAM1 cells in the internal L3 domain could nonautonomously establish blade organization and lateral outgrowth (McHale and Marcotrigiano, 1998), and that this internal organizing influence of LAM1 is an ongoing requirement during blade expansion.
Our most recent insights on blade formation have come from analysis of genes in the WUSCHEL (WUS)-RELATED HOMEOBOX (WOX) family. WUS, the founding member of WOX, is required for stem cell maintenance in shoot and floral meristem (Laux et al., 1996;Mayer et al., 1998). WOX genes also play important roles in lateral organ development. Two mutants in maize, narrowsheath1 and 2 (ns1 and ns2) lead to a loss of leaf blade in the ns1 ns2 double mutant (Scanlon et al., 1996). NS1 and NS2 are duplicated WOX genes (Nardmann et al., 2004) related to the Arabidopsis gene PRESSED FLOWER (PRS/WOX3), the knockout of which leads to defects in lateral sepals, petals, and stipules, but not in the leaf blade (Matsumoto and Okada, 2001;Shimizu et al., 2009). Interestingly, expression of WUS using the PRS promoter rescues the prs phenotype, indicating some common mechanism in WOX gene function (Shimizu et al., 2009). A related Arabidopsis gene, wusrelated homeobox 1 (wox1), has no visible mutant phenotype (Haecker et al., 2004;Vandenbussche et al., 2009), but in Petunia, mutation in WOX1-like gene, maewest (maw), has been shown to cause a narrow lamina and defective petal fusion phenotype (Vandenbussche et al., 2009). The maw mutation leads to stronger lamina reduction when combined with the chsu mutation (Vandenbussche et al., 2009), suggesting redundancy in regulating lamina expansion. Likewise, a prs wox1 double mutant in Arabidopsis has been shown to cause lamina reduction (Vandenbussche et al., 2009), although the phenotype is weaker than the maw chsu double mutant. The molecular mechanism by which the phan single mutant or ns1 ns2 or wox1 prs or maw chsu double mutants restrict(s) blade outgrowth remains unknown.
A gain-of-function mutation in blade-on-petiole (bop) or loss-offunction in bop1 bop2 double mutants leads to formation of bladelike outgrowth on the petiole (Ha et al., 2007;Jun et al., 2010). BOP1 and BOP2 encode BTB/POZ domain proteins that repress YABBY and KNOX1 genes, and activate AS2 in leaf primordia (Jun et al., 2010), suggesting that the BOP/YABBY/KNOX module is important for leaf elaboration. LYRATE, the tomato homolog of JAGGED, is shown to coordinate lateral outgrowth by interacting with KNOX1 genes and the auxin transcriptional network (David-Schwartz et al., 2009). Modulation of KNOX1 and auxin activity also appears to be a key component in dissected leaf morphogenesis in Cardamine hirsuta and other leaf forms (Barkoulas et al., 2008;Hay and Tsiantis, 2010). Together, these observations suggest a role for auxin in lamina expansion in simple-and dissected-leafed species. Although distinct aspects of leaf development, including phylotactic arrangement, margin serration, and vein patterning, are well documented to involve auxin (Reinhardt et al., 2003;Scarpella et al., 2006;Bilsborough et al., 2011), direct evidence for auxin control of lamina outgrowth and its genetic regulation at the leaf primordial margins is lacking.
Here, we report the identification and characterization of a novel leaf blade mutant in Medicago truncatula called stenofolia (stf) and a classical mutant called bladeless lam1 in N. sylvestris with very severe defects in blade outgrowth and vein patterning. We show that STF is critical for lamina outgrowth and leaf vascular patterning in simple leaf (N. sylvestris) and compound leaf (M. truncatula) species. Our data suggest that STF is a modulator of auxin and cytokinin homeostasis and hormonal crosstalk that coordinates developmental signals at the adaxial-abaxial boundary region of leaf primordia.

RESULTS
The stf Mutant Displays Multiple Defects in Leaf Lamina, Leaf Vasculature, and Flower Development Seven narrow leaf blade mutants with identical phenotypes were identified in forward genetics screens for bladeless mutants in 2 of 18 The Plant Cell (Tnt1) retrotransposon tagged lines of M. truncatula genotype, R108 .These mutants were named stf from Greek stenos for narrow. In all seven stf mutants, lamina growth is initiated and some blade tissue is formed, but further growth in the mediolateral axis is arrested, while growth in the proximodistal axis is virtually unaffected (Figures 1A to 1C). Mature leaves of R108 have regularly serrated margins along two-thirds of the blade distal to the petiole ( Figure 1C). In stf leaves, margin serrations are absent and adaxial-abaxial differentiation is drastically reduced ( Figures 1D and 1E) except in the unifoliate leaf (first leaf), which is only partially affected ( Figure 1F). The lack of lateral expansion is also evident in stf flowers in which the outer petal is narrow ( Figures 1I and 1J), and fails to enclose the anthers and stigma ( Figures 1G and 1H), while the ovary wall fails to close leaving ovules exposed ( Figure 1K), resulting in female sterility.
Disruption of vein patterning is another significant phenotype of stf leaves. In R108 leaves, lateral veins radiate from the midvein at regular intervals, extending laterally to the margin ( Figure 2B). The tip of each lateral vein corresponds to one serration in the margins and is open ended ( Figure 2D). Numerous minor veins make a complex network of branches all over the blade (Figures 2B and 2D). In stf leaves, however, the midvein is thin and less prominent and lateral veins are poorly developed, few in number, and do not extend to the margin ( Figures 2C and 2E). There appears to be another major vein-like structure (marginal vein), one on either side of the midvein near the margin, extending from the base of the blade to the tip ( Figures 2C and 2E). Minor veins are reduced in number ( Figure 2E).
Examination of blade epidermal cells by light and scanning electron microscopy (SEM) showed that stf leaves are affected in cell division and cell expansion. Under the light microscope, epidermal cells in stf appear slightly elongated and thinner, with  Transverse sections through the leaf blade revealed that the mesophyll tissue is not well differentiated in stf, especially on the adaxial side. In the wild type, palisade mesophyll cells (adaxial side) are uniformly cylindrical in shape, whereas spongy mesophyll cells (abaxial side) are compact and irregular in shape ( Figure 2H). In stf, although cells on either side are not identical, the contrast between palisade and spongy mesophyll cells is diminished ( Figure 2I). Transverse sections through the midvein showed that the xylem and phloem cells maintain their relative adaxial and abaxial positions, respectively, but fail to differentiate further and appear difficult to distinguish in stf ( Figures 2J and 2K).
Together, these observations suggest that stf retains polarity but has defects in cell division, cell expansion/differentiation, and vascular patterning that severely curtail lamina outgrowth. Unlike other lamina mutants caused by polarity defects, the stf polarity defect appears to be a consequence rather than the cause of the lamina phenotype since adaxial and abaxial cell types are correctly specified both in the mesophyll and vasculature of stf, but further differentiation of both cell types is compromised.

STF Encodes a WUS-Like Homeodomain Protein Conserved in Dicots
We cloned STF by PCR-based genotyping of flanking sequence tags (FST) in segregating populations  and we confirmed that the seven independent lines (stf1-1 to stf1-7) are allelic, having Tnt1 insertions in exons one, two, and four ( Figure 3A). This was further confirmed by complementation of the stf1-2 mutant phenotypes with a 5.3-kb genomic fragment of STF ( Figure 3B). STF encodes a 358-amino acid homeodomain transcriptional regulator with 38% amino acid identity to Arabidopsis WOX1 and 45% amino acid identity to Petunia MAW. We identified STF-like (STL) sequences from alfalfa and tobacco by RT-PCR and from other dicot genomes by BLAST search (see Supplemental Figures 2 and 3 online). In most of these genomes, STF is represented by a small gene family with one or two members. In the sequenced part of the M. truncatula genome, there is a second partial STF with predicted 165 amino acids but its functionality has not yet been tested. At the sequence level, STF, MAW, and STL share conserved amino acid motifs at the N-and C-terminal regions, in addition to the homeodomain and WUS box, including a highly conserved 10-amino acid motif at the 39 end (QFI/FEFLPLKN), which we named the STF box (see Supplemental Figure 3B online). The region corresponding to the STF box in WUS has been recognized as an EAR-like motif (Ikeda et al., 2009), but the STF box shows more similarity to the WUS box than the WUS EAR-like motif. The WUS box confers repressive functions to Arabidopsis WUS (Ikeda et al., 2009).
Phylogenetic analysis of a subset of related WOX and STL proteins using full-length and homeodomain region amino acid sequences showed that STF and STL group together in a separate subclade that encompasses WOX1 but distinct from the closely related PRS/WOX3 and WUS subclades (see Supplemental Figure 2 online), in agreement with previous studies (Vandenbussche et al., 2009;Zhang et al., 2010). In addition to M. truncatula, tobacco, petunia, and Arabidopsis mentioned above, for which functionality has been tested, STL sequences were found in all eudicot species sequenced to date, suggesting that STF is conserved in both dicot classes: rosids and asterids. By contrast, STF homologs were not found in the genomes of rice (Oryza sativa), maize, sorghum (Sorghum bicolor), foxtail millet (Setaria italic), or purple false brome (Brachypodium distachyon), or in the wheat (Triticum aestivum) and barley (Hordeum vulgare) database of 1.5 million expressed sequence tags, as well as the banana (Musa spp) NCBI database of 89,151 expressed se-quence tags. On the other hand, the related gene, PRS/WOX3, is widely conserved in angiosperms and gymnosperms (see Supplemental Figures 2B and 4 online). The molecular mechanism by which PRS or maize NS1 and NS2 orchestrate their functions is unknown, but the stf mutant phenotypes suggest that STF may not be functionally redundant with PRS/NS because the strong stf phenotype prevails in the presence of wild-type Medicago PRS (Mt WOX3). All the WUS, PRS, and MAW clades have strong conservation in the WUS box, whereas the STF box is absent from PRS/WOX3 genes and is modestly conserved in the WUS clade (see Supplemental Figure 5 online). Further investigation will be required to determine if the STF box is a functional domain and if STLs play specific roles in dicot lamina evolution.

STF Is Expressed at the Adaxial-Abaxial Boundary Layer in Leaf Primordia and Shows Developmental Regulation
To investigate tissue-specific expression patterns of STF, we performed RNA in situ hybridization in vegetative shoot apex and flower, where the highest expression was detected by quantitative RT (qRT)-PCR. In young leaf primordia at the P1 and P2 stages, STF is adaxially expressed in few cells, but absent from the SAM ( Figure 4A). In older primordia after stage P2, strong STF expression was detected in more cells and localizes in the central region of the adaxial-abaxial boundary layer extending from the distal tip to the proximal base ( Figure 4B). In the flower, STF expression was detected in petal primordia and developing petals (Figures 4C and 4E); in carpels, expression was observed in developing margins ( Figure 4C), in the placenta, at the base of ovules ( Figure 4D) and in the central region. No expression was observed in stamens, in sepals ( Figures 4D and 4E), in inflorescence meristem ( Figure 4F), or in the floral meristem ( Figure 4G), compared with controls ( Figures 4H and 4I). Fusing a 2.6-kb region of STF promoter with a b-glucuronidase (GUS) reporter showed a developmental gradient of GUS expression in leaves. In very young trifoliate leaves, GUS staining was intense at the leaf margin in the distal half, including the leaf tip (see Supplemental Figure 6B online). As the leaf grows, expression moves to the proximal half and becomes progressively weaker, to an undetectable level, in mature leaves (see Supplemental Figures 6B to 6E online). GUS staining was also detected in the unifoliate leaf, roots, cotyledons, petals, stigma, and pods (see Supplemental Figures 6A and 6F to 6H online).

The lam1 Mutation of N. sylvestris Is Caused by Deletion of the STF Homolog
In the classical bladeless lam1 mutant of N. sylvestris, leaf blades are reduced to vestigial strips lacking mesophyll differentiation (McHale and Marcotrigiano, 1998). The lam1 mutant strongly resembles stf except that lam1 phenotypes are stronger and lack stem elongation ( Figure 5A). Based on morphological characters, we hypothesized that stf and lam1 may be caused by mutations in homologous genes. To test this, we cloned an STF-like gene (Ns STF1) from wild-type N. sylvestris by RT-PCR. We also cloned the full-length Ns STF1 genomic sequence with its promoter by thermal asymmetric interlaced (TAIL)-PCR. Ns STF1 is similar to Mt STF in gene structure (see Supplemental Figure 2A online) and shares 45% amino acid identity. PCR experiments demonstrated that Ns STF1 was deleted in the lam1 mutant. Using primers specific to Ns STF1 (see Supplemental Table 1 online), including 2.65 kb of the promoter, we determined that at least a 5.67-kb region of the Ns STF1 locus was deleted in lam1 ( Figure 5B).
To confirm that the lam1 phenotype is due to lack of Ns STF1 function, a 5.3-kb genomic fragment of Mt STF was introduced into lam1. Mt STF fully complemented the lam1 phenotypes ( Figures  5C and 5D), confirming that STF function is indeed absent in the lam1 mutant and that STF and LAM1 are functional homologs.

Transcript Profiling Identified Auxin and Multiple Hormone-Associated Changes in stf Mutant
To identify potential targets of STF and to gain insight into the mechanism of its function, we performed transcript profiling analysis using the Medicago Affymetrix GeneChip containing 61,278 probe sets. We compared gene expression in three independent stf mutant lines (stf1-1/2/3) with their segregating wild types in 0.5-to 0.8-mm shoot apices of 4-week-old seedlings, and identified 241 probes differentially expressed with a twofold or more difference in the mutant; 105 probes upregulated and 136 probes downregulated. This analysis identified genes that are known to be involved in leaf development such as SCARECROW (SCR)-like, TCP3, and indole-3-acetic acid (IAA) amidosynthase (GH3) as downregulated, and BOP1/2 and KNAT2, KNAT6-related KNOX1 genes as upregulated, as well as changes in several phytohormone-associated genes, especially auxin ( Figure 6A, Table 1). The PHAN and HD-ZIP III-type adaxial polarity determinants were not detected despite the presence of probes on the array. We assessed the expression level of PHAN, PHABULOSA (PHB), and CORONA together with eight upregulated and 16 downregulated genes by qRT-PCR. Consistent with the microarray data, polarity genes were found to be unaffected (PHAN and CORONA) or modestly reduced (PHB), while the other genes differentially expressed in the microarray were clearly induced or repressed in qRT-PCR ( Figure 6B), confirming the notion that polarity disruption is not the primary defect in stf. The microarray data rather showed overrepresentation of genes related to phytohormones, including auxin, cytokinin, brassinosteroid, ethylene, gibberellic acid (GA), abscisic acid, and jasmonic acid (Table 1). After removing 11 redundant probes, at least 17.5% of the downregulated and 17.3% of the upregulated probes were (I) RNA in situ hybridization in the inflorescence apex using STF sense probe as negative control.

of 18
The Plant Cell found to be directly related to genes associated with hormones, especially auxin (Table 1). This represents a 10-fold enrichment (17.4% compared with a random 0.74%) for the seven hormones mentioned here. This enrichment is 17-fold for auxin-associated genes alone just by querying for the word "auxin." For example, for most altered auxin genes in the array, 11 of the 48 small auxinupregulated RNA (SAUR) probes present on the GeneChip were induced (Table 1), accounting for 4.78% of the altered genes instead of the expected 0.08% for unbiased change. Small auxinupregulated RNA genes are known to negatively regulate auxin biosynthesis and auxin response (Kant et al., 2009). This transcriptional change in auxin-associated genes is consistent with the stf mutant phenotype and the anticipated role of auxin in lamina outgrowth because auxin has been implicated in mechanisms controlling leaf vascular patterning, leaf margin serrations, reproductive organ development, and blade outgrowth although the association with blade outgrowth is circumstantial. These observations prompted us to think that the function of STF may be connected to auxin, and the other hormone responses may have been altered as a result of hormone crosstalk. We followed this lead and evaluated the auxin connection further.
stf and lam1 Mutants Accumulate Less Free Auxin in Leaves To investigate the role of auxin in stf/lam1 phenotype, we measured free auxin (IAA) directly by gas chromatographymass spectrometry (GC-MS) in 4-week-old leaves of lam1 and two alleles of stf (stf1-2 and sft1-3). We found that stf accumulates ;68% and lam1 accumulates 50% of the respective wild-type IAA levels ( Figure 6C), indicating that the mutants have an auxin deficiency. Because N. sylvestris is faster to transform and is known to be more sensitive to hormones than M. truncatula, we focused the next experiments solely on lam1.To independently confirm the reduced auxin content by a different method, we transformed lam1 and its wild type with the auxin responsive DR5:GUS construct. GUS staining in young leaves from 3-to 4-week-old plants was faintly detectable in DR5:GUS-expressing lam1 leaves, whereas strong GUS staining was observed in DR5: GUS-expressing wild-type leaves ( Figures 6D and 6E), confirming that lam1 indeed accumulates reduced auxin. The reduced steady state level of free auxin could be caused by a defect in auxin biosynthesis, signaling, transport, homeostasis, or a combination thereof. To distinguish among these possibilities, we applied exogenous auxins to lam1 shoots and roots. Foliar spraying of 2-week-old seedlings with naphthalene acetic acid (NAA) showed that mutant and wild-type leaves respond in the same way by epinastic curling and tolerated up to  Figure 7F online), although both roots tolerated much higher levels of the natural auxin IAA compared with the synthetic auxins. NAA and 2,4-D are known to be substrates of efflux and influx carriers, respectively (Yamada et al., 2009). The absence of a differential response to both or to NPA suggests that efflux and influx processes and transport are not significantly affected in lam1. However, growing mutant and wild-type plants on MS medium containing the auxin biosynthetic intermediate Trp, indicated that lam1 roots elongated slightly better (P < 0.05) than wild-type roots (see Supplemental Figure 7F online), suggesting that lam1 could be partly affected in auxin biosynthesis upstream of Trp.
We next performed metabolite profiling in leaf extracts by GC-MS to identify any accumulated intermediates of auxin biosynthesis that would indicate a metabolic blockage point in lam1. We found that metabolites of the shikimate pathway, including shikimic acid, Tyr, and Phe, as well as five-and six-carbon sugars of sugar metabolism, including Glc and Fru were significantly reduced in the mutant (P < 0.001), indicating a defect in sugar metabolism. No significant change (P < 0.05) in 12-carbon sugars, including Suc and galactinol were observed, however, glucitol, Gln, Ser, Gly, and Pro were increased (P < 0.001) by 2-to 15-fold in lam1 (Figure 7), which also indicated a defect in sugar metabolism. Hexoses limit erythrose and pyruvate availability and could account for the downregulation of the shikimate pathway, including auxin biosynthesis. These observations suggest that the reduced auxin levels measured in lam1 leaves may be caused by a broader defect in sugar metabolism rather than a single block in the IAA biosynthetic pathway.

Ectopic Expression of STF Leads to Auxin and Cytokinin Overproduction Phenotypes
To evaluate if increased STF activity leads to auxin overproduction, we ectopically expressed Mt STF and Nicotiana benthamiana STF (Nb STF1) in N. sylvestris using the 35S promoter. IAA was increased by approximately twofold in transgenic lines compared with controls ( Figure 8A), indicating a direct relationship between STF expression and steady state level of free auxin accumulation. Transgenic plants expressing either of the genes showed a range of auxin and cytokinin overproduction phenotypes that have been previously observed in tobacco (Eklö f et al., 2000), including leaf upward curling ( Figures 8B and 8C). Eight of the strongest STF expressers among 27 independent transgenic lines displayed extreme dwarf and nonflowering phenotypes with deformed shoots, roots, and leaves ( Figure 8E; see Supplemental Figure 8A online). Interestingly, these severely affected lines produced one or more tumors on the roots or at the crown ( Figure 8F; see Supplemental Figure 8A online). Because tumor formation requires overproduction of auxin and cytokinin (Zambryski et al., 1989), these results suggest that STF expressors are also overproducing cytokinin. Consistent with this, exogenous application of cytokinin alone with lanolin to lam1 shoot apex frequently produced leaf branches, and rarely shoots, while application of auxin and cytokinin together partially rescued the lam1 blade phenotype ( Figures 9B, 9D, and 9F), suggesting that STF may directly or indirectly affect cytokinin biosynthesis. The application of cytokinin or cytokinin and auxin together to wild-type shoots has an inhibitory  Figures 9A, 9C, and 9E), but the application of cytokinin alone rarely induces shoot regeneration on leaves similar to its effect on lam1 shoots. These data suggest that the function of STF in regulating lamina outgrowth may be connected to the auxin: cytokinin ratio. WUS activates cytokinin signaling by A-type ARR repression (Leibfried et al., 2005). To explore the possibility that cytokinin signaling is disrupted in lam1 mutant primordia, we transformed lam1 explants with STF:WUS. We observed complementation of the lamina and venation defects of lam1 (Figure 10), with this construct suggesting that cytokinin signaling is in fact a major component of LAM1 function. This is consistent with the STF overexpression phenotypes and the combined auxin and cytoki-nin treatments. Together, our data support a model in which auxinand cytokinin-mediated signaling modulated by STF/LAM1 at the leaf primordial margins regulates vein patterning and blade outgrowth in the simple-and dissected-leafed eudicots.

STF Controls Blade Outgrowth at the Adaxial-Abaxial Boundary Layer
We have described the leaf lamina mutants, stf and lam1, representing rosids and asterids, respectively. In several adaxial polarity mutants, including phan (Waites and Hudson, 1995) and phb/phv , there is not only loss of adaxial identity but also the appearance of abaxial characteristics in the adaxial domain. The abaxialization of severely affected phan mutants led Waites and Hudson (1995) to propose that the juxtaposition of adaxial and abaxial cells is a prerequisite for blade outgrowth. In the stf mutant, such complete loss of adaxial domain or appearance of abaxial characters in the adaxial surface has not been observed, and domain specification is intact. Unlike most polarity factors that exhibit domain-specific expression in leaf primordia, such as HD-ZIP III and KAN family genes (Kerstetter et al., 2001;McConnell et al., 2001;Emery et al., 2003), STF expression is not axially confined ( Figure 4B). The expression domains of STF and the presence of intact dorsoventral polarity in the stf mutant are consistent with recent observations reported for maw (Vandenbussche et al., 2009). Our work on STF/LAM1, along with the studies on MAW, suggests that polarity governs only the initial phase where founders are recruited at the adaxial-abaxial boundary. Subsequent assembly of specialized cell layers and outgrowth of flattened lamina is dependent on WOX function. We propose that STF function at the adaxial-abaxial boundary is required for cell proliferation, cellular differentiation, and expansion controlling lamina elaboration in the mediolateral axis. In addition to STF, LAM1, MAW, NS1, and NS2, which are WOX genes, class I HD-ZIP genes have been reported to control tendril and bract morphogenesis without necessarily disrupting polarity. For example, the tendril-less (tl) mutation in pea revealed that tendrils are modified leaves where lamina outgrowth is inhibited by repressive function of TL (Hofer et al., 2009), and a related gene in Arabidopsis, LATE MERISTEM IDENTITY1 (LMI1), is required for suppression of bract formation (Saddic et al., 2006). It remains to be shown if class I HD-ZIP genes target or interact with WOX genes in leaf primordia.

STF Controls Blade Outgrowth and Vein Patterning
Auxin has been described as the key factor that modulates growth and pattern formation during vascular morphogenesis Representative common metabolites are shown. Colors indicate downregulated (green) and upregulated (red) metabolites in lam1 mutant compared with the wild type (WT). The black color shows metabolites that are unchanged. The numbers 1 through 6 at the top indicate replicates of lam1 and wildtype samples each from six individual plants. Statistical significance was calculated using Student's t test (***P < 0.001, **P < 0.01, and *P < 0.05).

of 18
The Plant Cell (Scarpella et al., 2006;Cañ o-Delgado et al., 2010;Scarpella et al., 2010). Recently, Scarpella et al. (2010) have suggested that auxin internalization from PIN1 convergence points in the epidermal layer induces formation of veins and that auxin maxima generated by PIN1 polarity is the major factor that orchestrates vascular patterning and blade outgrowth. However, auxin response factors that specifically and differentially regulate auxin signaling in the outer epidermal layer have not been identified, nor is there a mechanism for local auxin biosynthesis that could account for continued generation of PIN1 polarity as the leaf expands. The pattern of vein disruption and lamina growth arrest in the stf mutant fits the mechanism of blade outgrowth and venation coupling described by Scarpella et al. (2010) and suggests that the STF regulation of vascular patterning and lamina outgrowth may be connected to auxin signaling at the leaf margins.

STF May Regulate Blade Outgrowth and Vascular Patterning by Modulating Auxin and Cytokinin Homeostasis
Our microarray analysis identified a strong connection between the stf mutant and transcriptional changes in hormone-responsive genes, especially genes associated with auxin (Table 1). Consistent with this, stf and lam1 mutants accumulate less free IAA in their leaves (Figures 6C to 6E). The reduced auxin cannot be simply a consequence of the narrow leaf mutant phenotype because ectopic expression of STF in N. sylvestris results in a smaller leaf, yet there is a twofold increase in free IAA compared with controls ( Figure 8A). Auxin is a multifunctional phytohormone required for cell division and cell expansion and has been  associated with a plethora of developmental programs, including leaf margin serration, reproductive development, leaf vascular patterning, and blade outgrowth (Scarpella et al., 2006(Scarpella et al., , 2010Koenig et al., 2009;Pagnussat et al., 2009;Perrot-Rechenmann, 2010;Bilsborough et al., 2011), indicating that the stf phenotypes are subsets of auxin-mediated developmental programs. It has been proposed that intra-and intercellular auxin gradients are critical in conveying positional information during auxin signal transduction (Benková et al., 2003;Vanneste and Friml, 2009). It is conceivable that a twofold reduction in total free auxin content in the mutant could disrupt auxin gradients that convey developmental signals. Under the model of Scarpella et al. (2010), STF/ LAM1 could function by controlling the spatial and temporal availability of auxin at the leaf margin. Normal auxin response and transport, but reduced free IAA associated with low metabolic flux in lam1 mutant, suggests that STF/LAM1 may be involved in the modulation of auxin homeostasis. However, auxin is not the only hormone altered in stf and lam1. Abscisic acid is also lower in lam1 mutants and higher than wild type in STF-overexpressing N. sylvestris (see Supplemental Figure 8B online). Moreover, the stem elongation defect in the lam1 mutant can be rescued by GA application (McHale, 1992), which together with the downregulation of GA biosynthetic enzymes in stf microarray, suggests that STF has a pleiotropic effect on GA biosynthesis. The microarray analysis also revealed that 1-aminocyclopropane-1-carboxylic acid oxidase (a key enzyme in ethylene biosynthesis), and His-containing phosphotransmitter protein1 (a key element in two-component cytokinin signal transduction phosphorelay) are strongly downregulated (Table 1). More importantly, the highest STF overexpressor transgenic plants developed typical auxin and cytokinin overproduction phenotypes, including tumors ( Figure 8). lam1 is sensitive to cytokinin application and responded by forming leaf branches similar to auxin treatment, but the application of auxin and cytokinin together partially rescued the lam1 lamina ( Figure  9). Together, these data suggest the integration of multiple hormones (Jaillais and Chory, 2010) during leaf morphogenesis coordinated by auxin:cytokinin ratio in leaf primordia.

Possible Mechanisms for STF Function
We envision three scenarios for how STF may function in modulating hormone homeostasis and integration. In the first scenario, STF regulates tissue-specific hormone homeostasis by controlling one or more hormone-conjugating enzymes. In support of this hypothesis, jasmonic acid carboxyl methyltransferase and putative auxin and cytokinin glucosyltransferases are upregulated in the stf mutant (Table 1). It is known that activation of auxin-conjugating enzymes leads to auxin depletion phenotypes and impairs auxin-mediated plant growth and environmental responses (Jackson et al., 2002;Qin et al., 2005;Park et al., 2007;Tognetti et al., 2010). In this scenario, failure of STF to control IAA/cytokinin glucosyltransferase in the stf mutant could lead to the buildup of IAA/cytokinin-glucoside conjugate, which signals excess IAA/cytokinin and this feeds back to sugar metabolism to slow down biosynthesis by reducing or rechanneling hexoses (see Supplemental Figure 9 online). Decreased Glc and Fru, but increased Suc have been shown to be associated with reduced auxin biosynthesis in the maize cell wall invertase mutant mn1 (LeClere et al., 2010). This decrease in hexoses, in turn, could downregulate the entire shikimate pathway, leading to reduced auxin biosynthesis, and could also affect other hormones (see Supplemental Figure 9 online).
In the second scenario, STF modulates the auxin:cytokinin ratio by regulating cytokinin signaling through activation of a histidinecontaining phosphotransmitter protein, which is strongly downregulated in stf or by repressing A-type two-component response regulators, RR9 and RR15, which are modestly upregulated in stf, analogous to repression of ARR7 and ARR15 by WUS in Arabidopsis (Leibfried et al., 2005). Complementation of lam1 with At WUS expression (Figure 10) favors this possibility. Cytokinin The Plant Cell signaling is known to affect auxin biosynthesis, and although the interaction could be antagonistic in roots (Dello Ioio et al., 2008;Mü ller and Sheen, 2008;Moubayidin et al., 2009), auxin-cytokinin crosstalk in shoots is thought to be synergistic (Pernisová et al., 2009;Zhao et al., 2010). In this hypothesis, cytokinin response is the primary target of STF, whereas auxin and other hormones are altered as a result of hormone crosstalk or a change in the auxin: cytokinin ratio.
In the third scenario, STF regulates other transcription factors that may be directly or indirectly connected to auxin/cytokinin activity. Two SHORT ROOT/SCR-related GRAS family transcription factors and one D-type cyclin (CYCD6;1 homolog) are downregulated in stf. SHORT ROOT and SCR are regulators of cell proliferation in leaves (Dhondt et al., 2010) and directly target D-type cyclins to control cell cycle progression (Sozzani et al., 2010). The combined downregulation of GRAS genes and CYCD6 could lead to premature exit from the cell cycle in stf. Other possible targets include HOMEOBOX PROTEIN 1, which is abscisic acid responsive and a repressor of auxin-responsive lateral organ boundary domain protein in M. truncatula roots (Ariel et al., 2010), BOP1/2, LOBD38, TCP3, and KNAT2/KNAT6related KNOX genes, which all have the potential to coordinate or interact with the auxin/cytokinin signaling machinery.
However, the above three scenarios are not mutually exclusive, and it remains to be shown if they act in concert. For example, STF may repress auxin/cytokinin-glucosyltransferase and also activate cytokinin by activating a His-containing phosphotransmitter protein at the same time to maintain the right auxin:cytokinin ratio and homeostasis at the leaf margins. The synergistic interaction of auxin and cytokinin could then presumably deliver the instructive spatiotemporal signal to TCPs, GRAS genes, and the cell cycle to execute morphogenic functions. In this context, STIMPY/WOX9mediated coordination of sugar and cytokinin signaling to the cell cycle has been proposed for shoot meristem establishment in Arabidopsis (Skylar et al., 2010(Skylar et al., , 2011. Identifying the direct target (s) of STF will enlighten our understanding of the mechanism of this fundamental process.

Mutant Screening and Cloning of STF
Insertional mutagenesis in Medicago truncatula genotype R108 using Tnt1 retrotransposon and screening conditions in the greenhouse have been previously described (Tadege et al., , 2009. Forward genetics screening of 5600 Tnt1-tagged lines under standard conditions (16 h/8 h and 248C/198C day/night cycles) in the greenhouse for leaf blade mutants have identified seven mutants with identical phenotypes of stf designated stf1-1 to stf1-7. The cloning of STF by genotyping of FST was performed as previously described .
The stf mutant phenotype segregates as a recessive mutant (95 wild types to 30 mutants in stf1-2 heterozygous) in the seven independent lines. Forty-one FST were recovered from stf1-1 by TAIL-PCR using a combination of Tnt1-specific primers (Tnt1-F1, Tntail 1, 2, 3, and LTR3 or LTR5) and arbitrary primers AD1 or AD2 (see Supplemental Table 1 online). Of 22 FST genotyped in a segregating population of stf1-1 mutants using FST-specific primers, two FST were identified with homozygous Tnt1 insertions that cosegregated with the narrow leaf mutant phenotype. Only one of the two FST (75insert1) was also found to be tagged in the other six stf mutant lines and cosegregated with the mutant phenotype. The gene corresponding to this FST was designated STF and the full-length sequence, including its promoter, was amplified from genotype R108 using primers STF1F and STF1R from the sequenced M. truncatula A17 genome. All primers are listed in Supplemental Table 1 online.

Cloning of Medicago sativa, Nicotiana benthamiana, and Nicotiana sylvestris STL Sequences
Ms STL1, Ms STL2, and partial Nb STL1 sequences were isolated by RT-PCR from shoot apex-enriched tissue of seedlings using primers STFcd1F and STFcd1R from conserved STF regions. A partial Ns STL1 sequence was amplified by RT-PCR from wild-type N. sylvestris using Nb STL1-derived primers NbSTLfd and NbSTLrs, and a full-length sequence, including its promoter and 39 untranslated region (UTR), was isolated by TAIL-PCR (Liu et al., 1995) using primers NsSTL391F, NsSTL392F, and NsSTL591R, NsSTL592R in combination with arbitrary primers AD1 and AD2. Full-length Nb STL1 was amplified using Ns STL1-derived primers NsSTLfrd and NsSTLrvs. All other STLs were identified by BLAST search from NCBI or the respective genome databases. After functional confirmation, Nb STL1 and Ns STL1 were renamed Nb STF1 and Ns STF1, respectively. Deletion of Ns STF1 in the lam1 mutant was identified using Ns STF1-specific primers; F1 and R1 for the middle of the genomic sequence, F1 and R2 for coding sequence (CDS) extending to the 39 UTR, and F2 and R3 for the promoter region and 59 UTR (see Supplemental  Table 1 online for primer sequences).

Transgene Construction and Plant Transformation
A 5.3-kb genomic fragment of the M. truncatula STF gene was amplified using primers STFgattB1 and STFgattB2 and was cloned in gateway vector pMDC99 for plant transformation (see Supplemental Table 2 online). A genomic DNA fragment of STF containing a 2663-bp region immediately upstream of the translation start was cloned into pMDC162 vector for the STF:GUS construct using primers pSTF1F and pSTF1R. The DR5 promoter was cloned into pMDC162 vector for the DR5:GUS construct using primers DR5gusattB1p and DR5gusattB2p. For overexpression constructs, the Mt STF CDS and Nb STF1 CDS were cloned into pMDC32 using primers STFcdattB1 and STFcdattB2 for STF and NbSTF1attB1 and NbSTF1attB2 for Nb STF1. For the STF:WUS construct, the At WUS1 CDS was amplified from PRS:WUS plasmid using WUSgattB1 and WUSgattB2 primers and was cloned in front of the STF promoter in pMDC162 vector (see Supplemental Table 2 online). Constructs were introduced into Agrobacterium tumefaciens by electroporation. A. tumefaciens strain AGL1 was used for M. truncatula transformation as described (Cosson et al., 2006) and strain GV2260 was used for N. sylvestris transformation.

Phylogenetic Analysis
Phylogenetic analysis was performed using full-length and homeodomain region amino acid sequences (see Supplemental Data Set 1 online). Sequences were aligned using Clustal W, and a neighbor-joining phylogenetic tree was constructed using MEGA 4 software. The most parsimonious trees with bootstrap values from 1000 trials were shown.

Tissue Fixation and Embedding
Leaf samples from 4-week-old seedlings of M. truncatula were cut into small pieces and collected directly into 4% formaldehyde made in PHEM buffer (60 mM Pipes, 25 mM HEPES, 2 mM MgCl, and 10 mM EGTA, pH 6.9). Samples were vacuum infiltrated for 30 min and then left in the fixative solution for an additional 2 h. After fixation, samples were washed three times in PHEM buffer and dehydrated by passing through a graded ethanol Regulation of Lamina by STENOFOLIA 13 of 18 series of 25%, 50%, 75%, 95%, and 100% ETOH, each step lasting for 1 h at room temperature and repeating three times for the 100% ETOH step. Ethanol was then replaced with a graded series of low-melting-temperature Steedman's wax in ethanol (WAX:ETOH). The Steedman's wax was prepared by melting 900 g of polyethylene glycol 400 distearate (Sigma-Aldrich) and 100 g of 1-hexadecanol (Sigma-Aldrich) and stirring at 658C, aliquoted in 50-mL volumes and stored at 2208C. The wax was melted at 388C before use. Samples were treated with 1ETOH:1WAX once for 3 h, 1ETOH:3WAX once for 3 h, and 100% WAX three times for 3 h. Tissues were then embedded in plastic molds and after hardening at room temperature, blocks were stored at 48C prior to sectioning. Embedded tissues were sectioned at 15 to 20 mm using a Leica RM 2235 rotary microtome (Leica Microsystems). The wax was removed by treating sections with absolute ethanol. After air drying, sections were stained with Toluidine Blue (Sigma-Aldrich) for visibility and were viewed under an Olympus BX-51 compound microscope (Hitschfel Instruments).

Tissue Preparation for Light Microscopy
Equivalent leaves from 4-week-old R108 and stf mutant plants were fixed in 6:1 ethanol:glacial acetic acid overnight at room temperature and were washed with 95% ethanol twice for 10 min. Samples were then cleared with 85% lactic acid over 24 h until the tissue became transparent. Cleared tissues were mounted with 30% glycerol on glass slide and were viewed under anOlympus BX-51 compound microscope equipped with a digital camera.

SEM
Leaf tissue from 4-week-old R108 and stf1-2 plants were vacuum infiltrated in fixative solution (3% glutaraldehyde in 25 mM phosphate buffer, pH 7.0) for 1 h. Samples were further fixed with 1.0% osmium tetroxide overnight, dehydrated in graded ethanol series, critical point dried, coated with Gold as previously described (Wang et al., 2008), and viewed using a ZEISS DSM-960A scanning electron microscope (Carl Zeiss MicroImaging).

RNA in Situ Hybridization
RNA in situ hybridization with digoxigenin-labeled STF-specific probe was performed on shoot apices of M. truncatula plants grown for 12 d after germination or on inflorescence apices as described (Ferrá ndiz et al., 2000). RNA antisense and sense probes were generated with the T7 and SP6 polymerases, respectively, using a 780-bp STF cDNA template, which included the last 431 bp of CDS plus 355 bp of the 39 UTR.

Histochemical GUS Staining
GUS staining assay was performed as described (Zhao et al., 2001) and images of GUS staining patterns of tissues were collected with a digital camera mounted on an Olympus BX-51 compound microscope (Hitschfel Instruments) or an Olympus SZX-16 Stereoscope (Hitschfel Instruments).

Microarray Analysis
Microarray was performed on RNA extracted from 0.5-to 0.8-mm shoot apices of 4-week-old seedlings. Three independent stf alleles (stf1-1, stf1-2, and stf1-3) and their corresponding wild type in segregating F2 populations were used. The three lines were treated as three biological replicates, and for each replicate, a pooled tissue collected from 20 plants was used to make RNA preparation. Total RNA was extracted using RNeasy Plant Mini Kit (Qiagen). The microarray analysis was performed using Medicago Affymetrix GeneChip. Probe labeling, hybridization, and scanning were conducted according to the manufacturer's instructions (Affymetrix). For each microarray sample, the .CEL file was exported from GeneChip Operating System software (Affymetrix) and imported into robust multi-chip average for normalization. The presence/absence call for each probe set was obtained from dCHIP. Differentially expressed genes between the stf mutant and the wild type were selected based on "associative analysis" (Dozmorov and Centola, 2003) using Matlab (MathWorks). In this method, the background noise presented between replicates and technical noise during microarray experiments was measured by the residual presented among a group of genes whose residuals are homoscedastic. Genes whose residuals between the compared sample pairs that are significantly higher than the measured background noise level were considered to be differentially expressed. A selection threshold of 2 for transcript ratios and a Bonferroni-corrected P value threshold of 8.15954E-07 were used. The Bonferroni-corrected P value threshold was derived from 0.05/N in these analyses, where N is the number of probe sets (61,278) on the chip.

Quantitative RT-PCR
Total RNA from 0.5-to 0.8-mm shoot apices of 4-week-old plants was extracted using RNeasy Plant Mini Kit (Qiagen) and the RNA was treated with Turbo DNase (Ambion) to remove contaminating genomic DNA. Two micrograms of total RNA was used for cDNA synthesis using the Omniscript Kit (Qiagen). Primers were designed to anneal near the 39 end or at the 39 UTR (see Supplemental Table 4 online). Three stf mutants (stf1-1, stf1-2, and stf1-3) and their segregating wild type were used as biological replicates for the analysis and three technical replicates were run for each. The qRT-PCR analysis was performed as described in Pfaffl (2001) using Power SYBR Green PCR master mix (Applied Biosystems) in an ABI Prism 7900 HT sequence detection system (Applied Biosystems). Gene expression was normalized using the expression of the EF1a and with UBQ5 used as housekeeping genes. Relative gene expression for each gene in the mutant plants was compared with that obtained for wild type, which was arbitrarily set to 1.0.

Quantification of Auxin
Quantification of free IAA was performed using GC-selective ion monitoring-MS essentially as described (Chen et al., 1988) using fresh tissue from equivalent leaves of 4-week-old wild-type, mutant, and transgenic plants with the following modifications. 0.5 g of fresh tissue was ground and extracted with 2 mL of 65% isopropanol in 200 mM imidazole buffer, pH 7.0. d 7 -IAA (100 pM) (CDN) was added as an internal standard and was equilibrated in the extract for 1 h at 48C. The extract was centrifuged and the supernatant was diluted to 12.5 mL using water. The diluted extract was then applied to a conditioned amino anion exchange column (BAKER-10 SPE 3 mL). After the diluted extract passed through the column, aspiration was continued for 30 s and the column was washed sequentially with 2.0 mL of each of hexane, ethyl acetate, acetonitrile, and methanol. The IAA was eluted from the amino column using 3.0 mL of 2% acetic acid in methanol. The acidic methanol eluent was evaporated to near dryness and the residue was redissolved in 10% aqueous methanol and applied to a C18 SPE column. The column was washed with 10% aqueous methanol containing 1% acetic acid, and IAA was eluted from the column with 2% acetic acid in methanol. The acidic methanol was evaporated to dryness, and the residue was redissolved in 50 mL of methanol. IAA was methylated by the addition of 2 mL of 2.0 M trimethylsilyldiazomethane (Sigma-Aldrich), the reaction was allowed to go for 30 min at room temperature and excess trimethylsilyldiazomethane was quenched by the addition of 2 mL of 2.0 M acetic acid in hexane. One microliter of methylated IAA solution was injected in an Agilent 6890 GC connected to 5973 MS detector with electron ionization source (Agilent Technologies). The injector was at 2808C and in splitless mode. The oven temperature was initially at 708C for 2 min and then ramped to 3158C at 58C/min. The monitored ions were mass-to-charge 14 of 18 The Plant Cell ratio 130 and 137, which correspond to the quinolinium ions from IAA and d 7 -IAA, respectively, as well as the mass-to-charge ratio 189 and 196 for the corresponding molecular ions. Dwell times were 100 ms for each ion. Abscisic acid was quantified by reanalyzing the same samples using abscisic acid internal standard.

Metabolite Profiling
Leaves from equivalent positions of lam1 and wild-type plants were collected from 4-week-old plants grown in growth cabinets. Metabolite analysis by GC-MS was performed as described (Broeckling et al., 2005). The GC system used was an Agilent 6890 GC coupled to a 5973 mass spectrometry detector. One microliter of samples was injected at a 15:1 split ratio and the injector was held at 2808C. Separation was achieved on DB-5MS column (J & W Scientific; 60 m, 0.25 mm i.d., and 0.25-mm film). Helium was used as carrier gas at constant flow of 1 mL/min. The temperature program was 2 min at 808C followed by a 58C/min ramp to 3158C and this was held at 3158C for 12 min. Mass spectra were scanned from m/z 50 to 650 with the acquisition rate of 2 spectra/s. Acquired mass spectra were deconvoluted using AMDIS software, and metabolite identifications were achieved by mass spectral matching to the Noble Foundation's in-house EIMS spectral library of authentic compounds, the publicly available GOLM library (http://csbdb.mpimpgolm. mpg.de/ csbdb/dbma/msri.html), and the NIST08 library. Peak selection and alignment were performed using MET-IDEA software (Broeckling et al., 2006). The area of each peak was normalized against the area of the internal standard, and absolute quantification for selected metabolites was achieved using authentic standard calibration curves.

Application of Auxin and Cytokinin to Shoot Apex
For foliar spray of auxin, 10 mL of 1, 10, 50, or 100 mM NAA was sprayed per plant to shoots of 2-week-old seedlings daily for 10 d. For local treatments of apices with hormones, 10 mM IAA, 10 mM NAA, 10 mM IAA plus 10 mM BAP, or 1 mM IAA plus 10 mM BAP were dissolved in a prewarmed (508C) lanolin paste. The paste was manually administered to shoot apices of 2-week-old N. sylvestris and lam1 seedlings or 6-weekold lam1 plants with pipette tips.

Root Elongation Assay
To measure root elongation in the presence of auxin, auxin polar transport inhibitor, or auxin biosynthetic intermediates, seeds were germinated on 0.5 MS plates and 9-d-old seedlings were transferred to 0.5 MS plates containing different concentrations of IAA, NAA, 2,4-D, Trp, or NPA. Root length was measured 4 or 5 d after transfer. Auxins and all other chemicals used in this assays were obtained from Sigma-Aldrich.

Accession Numbers
Sequence data from this article can be found in the EMBL/GenBank data libraries under accession numbers STF, JF276252; Ns STF1, JF276252; Ms STL1, JF276252 and Ms STL2, JF276252 (see also Supplemental Table 3 online). Microarray data from this manuscript can be found in ArrayExpress under accession number E-MEXP-3187.

Supplemental Data
The following materials are available in the online version of this article. Supplemental Table 1. Primers Used for Recovering FST, Genotyping FST, and Cloning of STLs.

Supplemental
Supplemental Table 2. Primers Used for Plasmid Construction.
Supplemental Table 4. Accession Numbers or Gene Identifiers of Sequences Used for Multiple Sequence Alignment and Phylogenetic Tree Construction.
Supplemental Data Set 1. Text File of Alignment Corresponding to Supplemental Figure 2 Online.