Genome-Wide Identification of the Highly Conserved INDETERMINATE DOMAIN (IDD) Zinc Finger Gene Family in Moso Bamboo (Phyllostachys edulis)

INDETERMINATE DOMAIN (IDD) proteins, a family of transcription factors unique to plants, function in multiple developmental processes. Although the IDD gene family has been identified in many plants, little is known about it in moso bamboo. In this present study, we identified 32 PheIDD family genes in moso bamboo and randomly sequenced the full-length open reading frames (ORFs) of ten PheIDDs. All PheIDDs shared a highly conserved IDD domain that contained two canonical C2H2-ZFs, two C2HC-ZFs, and a nuclear localization signal. Collinearity analysis showed that segmental duplication events played an important role in expansion of the PheIDD gene family. Synteny analysis indicated that 30 PheIDD genes were orthologous to those of rice (Oryza sativa). Thirty PheIDDs were expressed at low levels, and most PheIDDs exhibited characteristic organ-specific expression patterns. Despite their diverse expression patterns in response to exogenous plant hormones, 8 and 22 PheIDDs responded rapidly to IAA and 6-BA treatments, respectively. The expression levels of 23 PheIDDs were closely related to the outgrowth of aboveground branches and 20 PheIDDs were closely related to the awakening of underground dormant buds. In addition, we found that the PheIDD21 gene generated two products by alternative splicing. Both isoforms interacted with PheDELLA and PheSCL3. Furthermore, both isoforms could bind to the cis-elements of three genes (PH02Gene17121, PH02Gene35441, PH02Gene11386). Taken together, our work provides valuable information for studying the molecular breeding mechanism of lateral organ development in moso bamboo.


Introduction
During the development of multicellular organisms, transcription factors (TFs) are responsible for dictating the fate of individual cells by affecting gene expression. Among them, the INDETERMINATE DOMAIN (IDD) family is a class of TFs found only in higher plants. It was first identified through transposon insertion and acts non-cell-autonomously to regulate the production of a transmissible signal in the leaf that elicits the transformation of the shoot apex to reproductive development in maize (Zea mays) [1].
IDDs are also essential for plant growth and development in Gramineae species, including rice, maize, and barley (Hordeum vulgare). For example, rice OsID1 and maize ZmID1 both play essential roles in flowering; they regulate the transition from vegetative to reproductive growth [16][17][18][19][20]. Ghd10, the rice ortholog of ZmID1, plays an important role in regulating yield component traits by increasing rice plant height and the number of primary branches under short-day (SD) conditions [21]. OsIDD3/ROC1 positively regulates cold stress responses by directly binding to the promoter of the gene that encodes dehydration-responsive element-binding protein (DREB)/CBF1, which is involved in cold tolerance [22]. A duplicated gene pair, ZmIDDveg9/NAKED ENDOSPERM1 (NKD1) and ZmIDD9/NKD2, regulates aleurone cell specification and differentiation [23]. BROAD LEAF1 (BLF1), which is homologous to AtIDD14, AtIDD15, and AtIDD16, functions during the outgrowth of leaf primordia to limit cell proliferation in the leaf-width direction [24]. In-depth studies have shown that IDD functions by binding to the core cis-element TTTGTC in maize, Arabidopsis, and rice [15,[25][26][27].
Moso bamboo, one of the fastest growing plants on earth, is an important source of income in large areas of Asia and Africa. It is primarily propagated asexually from lateral buds on rhizomes that develop into new shoots, which grow to an average culm height of 13 m. Therefore, the development of lateral buds is an essential step in the entire growth and development process of moso bamboo.
With the rapid development of sequencing technology, the identification and specific molecular functions of IDDs have been validated in the model plants Arabidopsis and rice and the C 4 plant maize. Release of the moso bamboo chromosome-level reference genome provides a great opportunity for understanding this gene family [36].
In the present work, we identified 32 PheIDD members in moso bamboo and comprehensively analyzed their phylogenetic relationships, sequence features, interacting proteins, and scaffold distribution. We systematically characterized their expression patterns in underground bamboo buds/shoots of different sizes, aboveground branches of different sizes, and different tissues. We also tested their responses to plant hormone treatments. Finally, we examined the interactions of PheIDD21 isoforms with GRAS proteins as well as their target genes. This comprehensive analysis of PheIDD members provides valuable information for further functional characterization.

Identification and Characterization of PheIDD Family Genes in Moso Bamboo
To identify IDD genes in moso bamboo, we performed local blastn and blastp searches using IDD sequences from Arabidopsis and riceagainst the moso bamboo reference genome (V2.0) [36]. After manually removing sequences with incomplete conserved IDD domains, 32 candidate PheIDD genes were obtained (Table S1). All these 32 PheIDD proteins contained two C2H2 and two C2HC zinc finger (ZF) motifs.
We then amplified the coding sequences of the PheIDDs and analyzed their gene structures. The genes were named PheIDD1 to PheIDD32 based on their homology to ZmID1. Most PheIDD genes contained three exons and two introns with normal GT-AG splicing sites ( Figure 1). The predicted chemical characteristics of all PheIDDs are summarized in Table S1. The 32 PheIDDs were distributed unevenly on 16 scaffolds, and scaffolds 13 and 14 each contained four PheIDDs ( Figure S1). searches using IDD sequences from Arabidopsis and riceagainst the moso bamboo reference genome (V2.0) [36]. After manually removing sequences with incomplete conserved IDD domains, 32 candidate PheIDD genes were obtained (Table S1). All these 32 PheIDD proteins contained two C2H2 and two C2HC zinc finger (ZF) motifs.
We then amplified the coding sequences of the PheIDDs and analyzed their gene structures. The genes were named PheIDD1 to PheIDD32 based on their homology to ZmID1. Most PheIDD genes contained three exons and two introns with normal GT-AG splicing sites (Figure 1). The predicted chemical characteristics of all PheIDDs are summarized in Table S1. The 32 PheIDDs were distributed unevenly on 16 scaffolds, and scaffolds 13 and 14 each contained four PheIDDs ( Figure S1).
Multiple sequence alignment revealed that all PheIDD proteins contained two C2H2 (ZF1 and ZF2) and two C2HC (ZF3 and ZF4) ZFs, as well as a putative N-terminal nuclear localization signal (NLS, KK/RK/RR) (Figure 2), similar to other IDD proteins from Arabidopsis, rice, and maize. Their amino acid sequences displayed an overall identity of 32.83%, and the highly conserved IDD domain exhibited an identity of 75.41%. We also identified two C-terminal domains: the TR/L/QDFLG domain that is characteristic of IDDs from higher plants and an MSATALLQKAA domain that is present in some IDDs but absent from others ( Figure S2). In addition, a spacer region longer than those between ZF2 and ZF3 and between ZF3 and ZF4 was identified between ZF1 and ZF2: 37 aa in PheIDD20, 34 aa in PheIDD23, and 29 aa in PheIDD1 and PheIDD2 ( Figure 2).  Multiple sequence alignment revealed that all PheIDD proteins contained two C2H2 (ZF1 and ZF2) and two C2HC (ZF3 and ZF4) ZFs, as well as a putative N-terminal nuclear localization signal (NLS, KK/RK/RR) (Figure 2), similar to other IDD proteins from Arabidopsis, rice, and maize. Their amino acid sequences displayed an overall identity of 32.83%, and the highly conserved IDD domain exhibited an identity of 75.41%. We also identified two C-terminal domains: the TR/L/QDFLG domain that is characteristic of IDDs from higher plants and an MSATALLQKAA domain that is present in some IDDs but absent from others ( Figure S2). In addition, a spacer region longer than those between ZF2 and ZF3 and between ZF3 and ZF4 was identified between ZF1 and ZF2: 37 aa in PheIDD20, 34 aa in PheIDD23, and 29 aa in PheIDD1 and PheIDD2 (Figure 2). Phosphorylation site analysis identified two conserved amino acid residues, Ser73 and Ser182 among all the PheIDDs ( Figure S3). This result suggests that these two amino acids may be necessary for PheIDD function. Phosphorylation site analysis identified two conserved amino acid residues, Ser73 and Ser182 among all the PheIDDs ( Figure S3). This result suggests that these two amino acids may be necessary for PheIDD function.

Phe-Phe
Phe-Os To reveal the evolutionary relationships between IDD genes from bamboo and other plant species, an unrooted NJ phylogenetic tree was constructed by aligning full-length IDD proteins from moso bamboo, rice, maize, and Arabidopsis. The IDD proteins were resolved into 15 distinct groups, each containing an OsIDD ( Figure 3). In most cases, the orthologous groups were monophyletic, and we therefore could not determine one-to-one orthologous relationships between Arabidopsis and moso bamboo IDD family members. By contrast, one-to-two rice/moso bamboo IDD gene pairs could be identified. A comparative synteny analysis was performed to explore the origin and evolution of the PheIDD genes. Orthologous relationships were detected between 30 PheIDD genes and 15 OsIDD genes, most of which were located in syntenic loci on the rice chromosomes and moso bamboo scaffolds ( Figure 4). Two PheIDD genes were unique to moso bamboo. Consistent with the phylogenetic tree, multiple PheIDD genes were identified as putative orthologs of a single OsIDD gene. These results indicated that expansion of PheIDD genes may have occurred after the divergence of the lineages leading to rice and moso bamboo.
of the PheIDD genes. Orthologous relationships were detected between 30 PheIDD g and 15 OsIDD genes, most of which were located in syntenic loci on the rice chro somes and moso bamboo scaffolds ( Figure 4). Two PheIDD genes were unique to m bamboo. Consistent with the phylogenetic tree, multiple PheIDD genes were identifie putative orthologs of a single OsIDD gene. These results indicated that expansio PheIDD genes may have occurred after the divergence of the lineages leading to rice moso bamboo.   with colored boxes. The accession numbers of the genes used for sequence analysis and tree construction are shown in Table S4. In addition, tandem and segmental duplication events were investigated to provide additional insight into PheIDD gene duplication. No PheIDD genes were identified as tandem duplicates, whereas 22 segmental duplication events were identified ( Figure S4). These results indicated that expansion of the PheIDD genes may have occurred primarily through segmental duplication.

Expression Profiles of PheIDDs in Different Tissues
The IDD genes have been reported to display obvious tissue-specific expression patterns. For example, some IDDs are preferentially expressed in mature leaves [37], some in immature leaves [16,20] or roots [15]. To determine the tissue-specific expression patterns of the PheIDDs, roots, stems, mature leaves, immature leaves, and tillers of 2-month-old moso bamboo seedlings were used for qRT-PCR analysis. With the exception of PheIDD21 and PheIDD25, other PheIDDs exhibited very low transcript levels in all tested tissues ( Figure 5). Most PheIDDs showed an obvious organ-specific expression pattern. For example, PheIDD21 was exclusively expressed in tillers, whereas PheIDD25 accumulated preferentially in roots. Only two genes (PheIDD2 and PheIDD21) had their highest expression level in tillers. Ten PheIDDs had relatively higher transcript levels in leaves, including mature and immatureleaves: five (PheIDD3, PheIDD13, PheIDD15, PheIDD17, and PheIDD26) were preferentially expressed in mature leaves and four (PheIDD19, PheIDD22, PheIDD31, and PheIDD32) in immature leaves. PheIDD6 had similar transcript levels in mature and immature leaves. Seventeen genes had relatively higher expression levels in roots than in other tissues, and seven of these (PheIDD1, PheIDD4, PheIDD7, PheIDD12, PheIDD24, PheIDD25,and PheIDD30) were preferentially expressed in roots. Collectively, these data suggest that most PheIDDs display strong tissue-specific expression patterns, and five pairs of paralogous genes, such as PheIDD7/PheIDD16 that are highly expressed in roots, share similar expression patterns. In addition, tandem and segmental duplication events were investigated to provide additional insight into PheIDD gene duplication. No PheIDD genes were identified as tandem duplicates, whereas 22 segmental duplication events were identified ( Figure S4). These results indicated that expansion of the PheIDD genes may have occurred primarily through segmental duplication.

Expression Profiles of PheIDDs in Different Tissues
The IDD genes have been reported to display obvious tissue-specific expression patterns. For example, some IDDs are preferentially expressed in mature leaves [37], some in immature leaves [16,20] or roots [15]. To determine the tissue-specific expression patterns of the PheIDDs, roots, stems, mature leaves, immature leaves, and tillers of 2-month-old moso bamboo seedlings were used for qRT-PCR analysis. With the exception of PheIDD21 and PheIDD25, other PheIDDs exhibited very low transcript levels in all tested tissues ( Figure 5). Most PheIDDs showed an obvious organ-specific expression pattern. For example, PheIDD21 was exclusively expressed in tillers, whereas PheIDD25 accumulated preferentially in roots. Only two genes (PheIDD2 and PheIDD21) had their highest expression level in tillers. Ten PheIDDs had relatively higher transcript levels in leaves, including mature and immatureleaves: five (PheIDD3, PheIDD13, PheIDD15, PheIDD17, and PheIDD26) were preferentially expressed in mature leaves and four (PheIDD19, PheIDD22, PheIDD31, and PheIDD32) in immature leaves. PheIDD6 had similar transcript levels in mature and immature leaves. Seventeen genes had relatively higher expression levels in roots than in other tissues, and seven of these (PheIDD1, PheIDD4, PheIDD7, PheIDD12, PheIDD24, PheIDD25,and PheIDD30) were preferentially expressed in roots. Collectively, these data suggest that most PheIDDs display strong tissue-specific expression patterns, and five pairs of paralogous genes, such as PheIDD7/PheIDD16 that are highly expressed in roots, share similar expression patterns.

Dynamic Expression Profiles of PheIDDs in Underground Lateral Buds/Shoots at Different Developmental Stages
In Arabidopsis, IDD family members perform diverse functions, including lateral shoot morphogenesis and cellular patterning [11,12]. To further characterize their func-

Dynamic Expression Profiles of PheIDDs in Underground Lateral Buds/Shoots at Different Developmental Stages
In Arabidopsis, IDD family members perform diverse functions, including lateral shoot morphogenesis and cellular patterning [11,12]. To further characterize their functions during the development of underground lateral buds/shoots in bamboo (Figure 6a), we analyzed the expression patterns of all PheIDDs by qRT-PCR. Twenty of the 32 PheIDD genes were upregulated in underground buds upon dormancy awakening (from stage S1 to S2), and the expression of 14 genes reached the highest level at the S2 stage. The expression of PheIDD1 showed the greatest increase (39.4-fold) (Figures 6b and S5). By contrast, transcript levels of the PheIDD11/PheIDD12 gene pair were reduced from dormant buds (S1) to 16-cm buds (S4), suggesting that they may be related to the maintenance of bud dormancy. In addition, a paralogous gene pair (PheIDD21/PheIDD25) displayed opposite expression patterns in underground bamboo buds/shoots of different sizes. tions during the development of underground lateral buds/shoots in bamboo (Figure 6a), we analyzed the expression patterns of all PheIDDsby qRT-PCR. Twenty of the 32 PheIDD genes were upregulated in underground buds upon dormancy awakening (from stage S1 to S2), and the expression of 14 genes reached the highest level at the S2 stage. The expression of PheIDD1 showed the greatest increase (39.4-fold) (Figures 6b and S5). By contrast, transcript levels of the PheIDD11/PheIDD12 gene pair were reduced from dormant buds (S1) to 16-cm buds (S4), suggesting that they may be related to the maintenance of bud dormancy. In addition, a paralogous gene pair (PheIDD21/PheIDD25) displayed opposite expression patterns in underground bamboo buds/shoots of different sizes. The color scale indicates the relative expression level. Green in color scale erected vertically at the right side of the picture indicates lower transcript abundance compared to the relevant control, and red indicates higher transcript abundance. The PheNTB gene was used as a reference. The qRT-PCR results were obtained from three biological and three technical replicates.

Dynamic Expression Profiles of PheIDDs in Aboveground Branches of Different Sizes
To further characterize their functions during the development of branches in moso bamboo, expression patterns of all PheIDDswere analyzed in aboveground branches of different sizes (Figure 7a) by qRT-PCR. Twenty-three PheIDDswereexpressed at a very

Dynamic Expression Profiles of PheIDDs in Aboveground Branches of Different Sizes
To further characterize their functions during the development of branches in moso bamboo, expression patterns of all PheIDDs were analyzed in aboveground branches of different sizes (Figure 7a) by qRT-PCR. Twenty-three PheIDDs were expressed at a very low level in the dormant bud; their transcript levels peaked in 1-cm branches (L2) and then declined when branch length reached 7 cm (L4) (Figures 7b and S6). The absolute tran-script levels of 23 PheIDDs increased more than two-fold in 1-cm branches compared with dormant buds, and PheIDD1 showed the greatest increase (39.3-fold). By contrast, the expression levels of five PheIDDs (PheIDD5, PheIDD25, PheIDD26, PheIDD31, and PheIDD32) were very high in dormant buds and then gradually decreased with bud development. A total of 28 PheIDDs were expressed at high levels in aboveground branches; these included PheIDD2 and PheIDD21, which had higher expression levels in tillers than in other tissues. Several paralogous gene pairs, including PheIDD1/PheIDD2, showed a similar expression pattern, whereas others exhibited different or even opposite expression patterns. These observations reflect the different evolutionary fates of duplicated genes. low level in the dormant bud; their transcript levels peaked in 1-cm branches (L2) and then declined when branch length reached 7 cm (L4) (Figures 7b and S6). The absolute transcript levels of 23 PheIDDs increased more than two-fold in 1-cm branches compared with dormant buds, and PheIDD1 showed the greatest increase (39.3-fold). By contrast, the expression levels of five PheIDDs (PheIDD5, PheIDD25, PheIDD26, PheIDD31, and PheIDD32) were very high in dormant buds and then gradually decreased with bud development. A total of 28 PheIDDs wereexpressed at high levels in aboveground branches; these included PheIDD2 and PheIDD21, which had higher expression levels in tillers than in other tissues. Several paralogous gene pairs, including PheIDD1/PheIDD2, showed a similar expression pattern, whereas others exhibited different or even opposite expression patterns. These observations reflect the different evolutionary fates of duplicated genes.

Transcript Profiling of PheIDDs under IAA and 6-BA Treatments
Previous studies have reported a link between IDD members and hormones. IDD members were documented to mediate hormone signaling to regulate plant development

Transcript Profiling of PheIDDs under IAA and 6-BA Treatments
Previous studies have reported a link between IDD members and hormones. IDD members were documented to mediate hormone signaling to regulate plant development [5,12,38,39]. A cis-element analysis showed that each PheIDD promoter contained at least one cis-element implicated in the response to hormones. To investigate whether PheIDDs are involved in hormone signaling, we analyzed the expression patterns of each PheIDD under hormone treatments by qRT-PCR.
As shown in Figure 8, 21 PheIDDs showed strong responses to exogenous 6-BA. The expression levels of eight PheIDDs (PheIDD10, PheIDD12, PheIDD13, PheIDD20, PheIDD23, PheIDD25, PheIDD27, and PheIDD28) increased significantly, whereas those of thirteen PheI-DDs (PheIDD1, PheIDD2, PheIDD3, PheIDD6, PheIDD16, PheIDD17, PheIDD18, PheIDD19, PheIDD22, PheIDD26, PheIDD30, PheIDD31, and PheIDD32) were lower at four or more time points in treated compared with untreated seedlings. PheIDD25 expression was either unchanged or higher in the 6-BA treatment than in the control at all time points, whereas PheIDD22 displayed the opposite expression pattern in the 6-BA treatment. The expression levels of 16 PheIDDs were highest 2 h after 6-BA treatment, indicating that they respond rapidly to 6-BA.  [5,12,38,39]. A cis-element analysis showed that each PheIDD promoter contained at least one cis-element implicated in the response to hormones. To investigate whether PheIDDs are involved in hormone signaling, we analyzed the expression patterns of each PheIDD under hormone treatments by qRT-PCR. As shown in Figure 8, 21 PheIDDs showed strong responses to exogenous 6-BA. The expression levels of eight PheIDDs (PheIDD10, PheIDD12, PheIDD13, PheIDD20, PheIDD23, PheIDD25, PheIDD27, and PheIDD28) increased significantly, whereas those of thirteen PheIDDs (PheIDD1, PheIDD2, PheIDD3, PheIDD6, PheIDD16, PheIDD17,  PheIDD18, PheIDD19, PheIDD22, PheIDD26, PheIDD30, PheIDD31, and PheIDD32) were lower at four or more time points in treated compared with untreated seedlings. PheIDD25 expression was either unchanged or higher in the 6-BA treatment than in the control at all time points, whereas PheIDD22 displayed the opposite expression pattern in the 6-BA treatment. The expression levels of 16 PheIDDs were highest 2 h after 6-BA treatment, indicating that they respond rapidly to 6-BA.  Fourteen PheIDDs responded strongly to exogenous IAA application (Figure 9). Specifically, the expression levels of nine PheIDDs (PheIDD3, PheIDD9, PheIDD10, PheIDD12, PheIDD13, PheIDD20, PheIDD24, PheIDD26, and PheIDD30) increased significantly, whereas those of five PheIDDs (PheIDD2, PheIDD6, PheIDD16, PheIDD17, and PheIDD31) decreased in IAA-treated samples compared with the control at four or more time points. The transcript levels of PheIDD25, PheIDD28, and PheIDD30 were either unchanged or increased upon IAA treatment compared with the control at all eight time points. One paralogous gene pair, PheIDD5/PheIDD6, displayed similar expression patterns inIAA-treated samples. Taken together, these results revealed different responses of many duplicated PheIDD gene pairs to hormone treatments.
The qRT-PCR results were obtained from three biological and three technical replicates. Five 0.6-cm sections of the stem base were harvested per replicate. Values are means ± SD. At a given time point, asterisks indicate a significant difference in expression between 6-BA treatment and Mock treatment based on LSD multiple comparisons. *p< 0.05; **p< 0.01.
Fourteen PheIDDs responded strongly to exogenous IAA application ( Figure 9). Specifically, the expression levels of nine PheIDDs (PheIDD3, PheIDD9, PheIDD10, PheIDD12, PheIDD13, PheIDD20, PheIDD24, PheIDD26, and PheIDD30) increased significantly, whereas those of five PheIDDs (PheIDD2, PheIDD6, PheIDD16, PheIDD17, and PheIDD31) decreased in IAA-treated samples compared with the control at four or more time points. The transcript levels of PheIDD25, PheIDD28, and PheIDD30 were either unchanged or increased upon IAA treatment compared with the control at all eight time points. One paralogous gene pair, PheIDD5/PheIDD6, displayed similar expression patterns inIAA-treated samples. Taken together, these results revealed different responses of many duplicated PheIDD gene pairs to hormone treatments.

PheIDD21Generated Two Transcripts by Alternative Splicing
During cloning of the PheIDD genes, we unexpectedly found that two PCR products were generated by the PheIDD21 primer pair, suggesting AS of PheIDD21 in moso bamboo. By comparing the two cDNA sequences, we discovered that one splicing product was missing 9 nucleotides at the 3 end of the first exon compared with the full-length transcript (Figure 10a). The AS type was therefore an alternative donor site (selection of 5 splice sites). We named the full-length transcript PheIDD21α and the shorter transcript PheIDD21β. A protein sequence alignment showed that the three missing amino acids were located next to the NLS but not within the conserved domain. The qRT-PCR results were obtained from three biological and three technical replicates. Five 0.6-cm sections of the stem base were harvested per replicate. Values are means ± SD. At a given time point, asterisks indicate a significant difference in expression between IAA treatment and Mock treatment based on LSD multiple comparisons. * p< 0.05; ** p < 0.01.

PheIDD21Generated Two Transcripts by Alternative Splicing
During cloning of the PheIDD genes, we unexpectedly found that two PCR products were generated by the PheIDD21 primer pair, suggesting AS of PheIDD21 in moso bamboo. By comparing the two cDNA sequences, we discovered that one splicing product was missing 9 nucleotides at the 3′ end of the first exon compared with the full-length transcript (Figure 10a). The AS type was therefore an alternative donor site (selection of 5′ splice sites). We named the full-length transcript PheIDD21α and the shorter transcript PheIDD21β. A protein sequence alignment showed that the three missing amino acids were located next to the NLS but not within the conserved domain.

Interaction of PheIDD21 with RGAS Proteins and Its Binding to Target Genes
IDD proteins have been reported to form homo-and heterodimers with proteins from different families, especially the GRAS TFs [4]. To identify GRAS proteins that interact with IDDs, we searched a protein-protein network using STRING 10 software. Given that none of the 16 Arabidopsis IDDs were homologous to ZmID1/OsID [16], we used the full-length sequences of 30 PheIDD proteins (PheIDD3-PheIDD32) to search for candidate interacting proteins based on their interologs in the Arabidopsisinteractome ( Figure S7). A total of 11 high-confidence interacting proteins were identified for the IDDs, four of which, SCARECROW-LIKE3 (SCL3), SCR, SHR, and REPRESSOR of ga1-3 (RGA), belonged to the GRAS family. All 30 PheIDD proteins, including PheIDD21, were predicted to interact with AT2G20050, a PP2C phosphatase. In addition, five PheIDDs, including PheIDD21, were shown to interact with SCL3 and RGA1/DELLA. We then performed yeast two-hybrid (Y2H) experiments to test the predicted interactions. Yeast cells were co-transformed with PheIDD21 isoforms and GRAS proteins, including PheSCL3 and PheDELLA. All yeast cells grew normally on SD/-Trp-Leu medium.
The positive control and various tested protein combinations grew normally on SD/-Trp-Leu-His-Ade medium, whereas the negative control did not (Figure 10b). The results showed that both PheIDD21α and PheIDD21β could interact strongly to form homodimers and could also form heterodimers. Both isoforms could form heterodimers with PheDEL-LAa, PheDELLAb, and PheSCL3a. The data therefore confirmed that both PheIDD21 isoforms could function as homodimers and heterodimers.
IDD members act as TFs to regulate a variety of plant developmental and physiological processes [3], and we therefore investigated their downstream target genes. First, we constructed a gene co-expression network with PheIDD21 based on the transcriptome data (Table S2). Previous studies have shown that IDD TFs directly bind to the TTTGTC core element [15,25,26], and we next analyzed cis-elements in the genome sequences of the coexpressed genes. Four putative elements in the promoter and intron of gene PH02Gene17121, three putative elements in the promoter and intron of gene PH02Gene35441, and two putative elements in the promoterof gene PH02Gene11386 were identified (Figure 11a). We next performed yeast one-hybrid (Y1H) assays with 220-350-bp fragments to test the binding affinity of PheIDD21 isoforms to the above three genes. Both PheIDD21α and PheIDD21β bind to three (A, B, and D) of the four elements of gene PH02Gene17121, one (A) of the three elements of gene PH02Gene35441, and one (A) of the two elements of gene PH02Gene11386 (Figure 11b), suggesting that PheIDD21 directly binds to these three genes to regulate their expression.

Discussion
The IDD family is a plant-specific zinc finger subfamily. Previous reports

Discussion
The IDD family is a plant-specific zinc finger subfamily. Previous reports haveshown that several IDD members play essential roles in the development of Arabidopsis, rice, and maize; however, little is known about IDD members in moso bamboo. In the current study, we performed a genome-wide investigation of the IDD gene family in moso bamboo.

The IDDGene Family in Moso Bamboo
Prochetto and Reinheimer reported that the IDD TF originated in the last common ancestor of Charophytic algae and land plants and underwent a gene duplication event 470 million years ago (MYA) [2]. The IDDs were then duplicated several times during the emergence and diversification of land plants. Therefore, the monocot and dicot genomes contain similar numbers of IDD genes (16)(17)(18)(19)(20)(21)(22)(23) [16,40,41]. However, monocots have a higher number of IDD clusters than dicots, most of which are specific to grass families; one example is the OsID1/RID1/Ehd2 cluster [2]. This suggests that whole-genome duplication (WGD) may have given rise to the grass IDD gene families [42,43]. Here, we identified 32 PheIDD genes in moso bamboo (Figures 1 and 3), suggesting that additional duplications, especially segmental duplication events, have occurred in the IDD family during bamboo evolution ( Figure S4). Gene duplication and diversification are important for plant evolution and have been reported to contribute to the morphological evolution of moso bamboo [44]. Although the NLS motif was reported to be absent from some IDD proteins [2], all PheIDDs contained this motif ( Figure 2). Therefore, it is very likely that PheIDDs are located in the nucleus and function astranscriptional regulators. Furthermore, all PheIDDs are likely to be functional because, like Arabidopsisand maize IDD proteins, they each possess all four intact ZFs (two C2H2 ZFs and two C2HC ZFs) [1,37]. In addition, they all contained two conserved amino acid residues ( Figure S3): Ser73 has been reported to be modified by MPK6 in several independent phosphoproteomic studies [45], and Ser182 is known to be phosphorylated by AKIN10 [38]. Although the PheIDDs share high similarity in their N-terminal ID domains (75.41% identity), the remaining portions of the sequences vary substantially. An exception occurs in two conserved regions at the C terminus of 21 PheIDDs (Figure S2), which are exclusively found in the IDDsof higher plants [16].
Both the IDD phylogenetic tree and the full-length sequence alignment suggest that PheIDD1 is closely related to ZmID1 (64.57% identity) and OsID1 (61.59% identity) [16]. Like ZmID1 and OsID1, PheIDD1 also contains a long spacer region of 29 aa between the first two ZFs (Figure 2) [16]. However, the tissue-specific expression pattern of PheIDD1 differed from those of OsID1 and ZmID1, indicating thatit may function differently. Whether this feature of PheIDD1 is related to the irregular flowering of moso bamboo is worthy of in-depth exploration.
Three of the 16 AtIDDs (AtIDD14, AtIDD15, and AtIDD16) and three of the 15 OsIDDs (OsIDD12, OsIDD13, and OsIDD14/Loose Plant Architecture1, LPA1) shared a unique conserved coiled-coil domain at the Cterminus but had lost two other conserved domains (MSATALLQKAA and TR/LDFLG), at least one of which is present in other IDD members [16]. The six PheIDDs that are homologous to the above proteins (PheIDD20, PheIDD23, PheIDD27, PheIDD28, PheIDD29, and PheIDD30)-as well as two other PheI-DDs (PheIDD10 and PheIDD14)-have all lost these two conserved C-terminal domains, suggesting that these 14 proteins may have similar functions ( Figure S2).

PheIDDs Respond to Exogenous Hormone Treatments
Previous studies have shown that some IDDs are involved in hormone signaling as well as hormone homeostasis to regulate plant development [5,38,45]. AtIDD14, AtIDD15, and AtIDD16 are known to cooperatively regulate lateral organ morphogenesis and gravitropism by promoting auxin biosynthesis and transport [12]. Conversely, some IDDs are subject to regulation by hormones [22,46]. Gamuyao et al. have reported that in addition to auxin, CTK and GA also accumulate in the moso bamboo shoot apical meristem (SAM), and several PheIDD genes are highly expressed in the SAM, suggesting a possible link between IDDs and these hormones [47]. Our data showed that the expression of 24 of the 32 PheIDDs was induced by 6-BA, and 22 reached the maximum fold-change in less than 6 h ( Figure 8). The expression of 19 of the 32 PheIDDs was induced by IAA, and the expression of eight PheIDDs peaked less than 6 h after hormone application (Figure 9). Collectively, our data indicate that PheIDDs respond rapidly to hormone treatments and that they may play roles in the development of bamboo shoots.

Expression Profile Analysis of PheIDDGenes in the Development of Aboveground Branches, and Underground Buds/Shoots
All IDDs have undergone extensive gene duplications in monocots during evolution and have acquired diverse expression patterns [2]. Most IDDs, including duplicated IDDs, show different expression levels during different stages of vegetative organ development [2]. Here, PheIDD25 exhibited the highest expression level in all tissues tested, especially in roots, whereas its paralog, PheIDD21, had a moderate expression level in all tissues tested. The remaining 30 PheIDD genes were expressed at low levels in the tested tissues ( Figure 5). In addition, more than half of the PheIDDs had a high expression level in roots, and ten and three PheIDDs accumulated primarily in leaves and stems, respectively. The remaining two genes, PheIDD2 and PheIDD21, were highly expressed in tillers, whereas their paralogs PheIDD1 and PheIDD25 were expressed exclusively in roots. These data suggest that the functions of each pair of parologous genes may have diverged during the evolution of grass species.
AtIDD4 and several other IDDs have been implicated in the establishment of adaxial/abaxial patterns and cell fates in leaves and other lateral organs [11]. Mutation of IDD4 and IDD4 overexpression have been reported to affect the shoot and root growth of Arabidopsis [45]. AtIDD14, AtIDD15, and AtIDD16are known toplay critical roles in lateral shoot morphogenesis and gravitropism by regulating spatial auxin accumulation [12]. Therefore, we assume that some PheIDDs may be related to the development of bamboo buds/shoots in moso bamboo. Consistent with this hypothesis, the expression levels of 20 PheIDDs (62.5%) were higher in awakening buds compared with dormant buds (Figure 6). The above data suggest that several PheIDD TFs may be involved in the early development of underground bamboo buds/shoots. Given the changes in PheIDD expression during the transition from underground dormant buds to awakened buds, as well as the conserved C-terminal motifs shared by PheIDD25 and PheIDD26, we speculate that these two proteins may function redundantly in the development of underground bamboo buds/shoots. This possibility remains to be investigated in future studies. Consistent with the results obtained by Gamuyao et al. [47], PheIDD18 expression increased from S1 to S4 and was highest in S4, suggesting a possible role in the later growth and development of underground bamboo shoots (Figure 6b). Expression of the PheIDD27 gene was similar. However, we did not identify PheIDDs whose expression increased in the L4 stage, suggesting that PheIDDs were not involved in the late stages of aboveground branch development.
It is worth noting that although PheIDD1 is most closely related to rice OsID1 and maize ZmID1, it was primarily expressed in roots and showed the greatest increase in expression compared to other PheIDDs during the development from dormant buds to awakening buds. Similar to previous findings for maize ZmID1, Wu et al. reported that OsID1 is exclusively expressed in immature leaves but not in roots, mature leaves, the SAM, or other tissues [20]. Given that OsID1 functions as a master switch that controls the transition from vegetative growth to floral development, our results suggest that the regulation of PheIDD1 may be different from that of OsID1 and that PheIDD1 may be involved in the regulation of lateral shoot development.

IDD Interacting Proteins and Target Genes
Based on phylogenetic analysis, PheIDD21 has the highest homology with rice Os-IDD10 and with a maize duplicated gene pair, ZmIDDveg9 and ZmIDD9. OsIDD10 is expressed in the primordia of lateral roots [26], and ZmIDDveg9 and ZmIDD9 are re-quired for cell patterning and differentiation, cell growth and division, and hormone pathways [23,39]. In addition, several Arabidopsis IDDs, including AtIDD10 and AtIDD3, are required for cell division in ground tissues through direct protein-protein interactions with the GRAS proteins SHR, SCR, and SCL3 [9,48,49]. Both isoforms of AtIDD14 interact with each other and participate in starch metabolism [4]. Our work suggests that PheIDD21 generates two transcripts by AS. Both the isoforms interact with each other to form homodimers and interact with GRAS proteins to form heterodimers ( Figure 10).
Multiple studies have reported that IDD TFs bind to cis-elements to regulate plant growth and development. AtIDD8 contributes to photoperiodic flowering by binding directly to the CTTTTGTCC sequence motif within the SUS4 gene promoter [15]. OsIDD10 influences root ammonium uptake and nitrogen metabolism by activating AMT1;2 and GDH2, binding to a specific cis-element TTTGTC in the AMT1;2 promoter and toa similar TTTGTCC/G element in the fifth intron of GDH2 [26]. OsIDD13 and OsIDD3 regulate defense to sheath blight disease by binding to the TTTGTCG cis-element in the PIN1a promoter [27]. The maize ID1 protein also binds to the TTTTGTCG/C sequence [25]. Similarly, PheIDD21 in this study bindsto the core cis-elements of three genes (Figure 11), suggesting that PheIDD21 may play a role in regulating their expression.

Plant Materials and Treatments
The moso bamboo plantation was located in Qingshan, Lin'an District, Hangzhou City, Zhejiang Province, China (30 • 14 N, 119 • 42 E). Underground buds or shoots at four different developmental stages (S1-S4) in this area were sampled as described in a previous study from our lab [50]. Buds at the S1 stage were dormant,~0.2 cm in length, and close to the groove of the rhizome. In stage S2, the angle between the tip of the awakening bud (2 cm in length) and the rhizome was~45 • . In stage S3, shoots were 8 cm in length and more robust, and the angle between the tip of the shoot and the rhizome was 60 • . Shoots at the S4 stage were~16 cm in length. At least three buds or shoots with similar size and shape were harvested from the rhizome at each stage. Bamboo sheaths were carefully removed, and all buds or shoots were collected separately.
Developing branches were harvested from one-year-old moso bamboo plants which had just ended their height growth. Branches at three different developmental stages (1, 4, and 7 cm in length) on the same plant were harvested. The base of an internode groove that was beneath an internode with a branch was collected as the dormant bud (referred to as 0-cm bud in this study). At least three branches of similar size were harvested in each stage.
The seeds of moso bamboo used in this study were collected from Guilin, Guangxi. The seeds were rinsed and soaked in water at room temperature for 24 h and then germinated in Petri dishes on moist, sterile filter paper at room temperature. After growing for 10 days, some seedlings were moved to Hoagland's nutrient solution in containers. The nutrient solution was changed every two days to ensure seedling growth.
Tissue-specific expression of the PheIDDs: Various vegetative tissues, including roots, stems, mature leaves, immature leaves, and tillers, were collected from 2-month-old seedlings. Three biological replicates were analyzed, with five samples per biological replicate.
Expression patterns of the PheIDDs under exogenous plant hormones: Moso bamboo seedlings were sprayed with 500 µM IAA or 500 µM 6-BA, and control seedlings were sprayed with water. A 0.6-cm section of the stem base, where tiller buds emerged, was harvested from treated and control plants at 0, 2, 4, 6, 12, 24, 48, and 72 h after plant hormone treatments. Three biological replicates were analyzed, with five sections per replicate.
All samples were immediately frozen in liquid nitrogen and stored at −80 • C until further use.

Identification of PheIDDFamilyGenes in the Moso Bamboo Genome
The nucleotide and protein sequences of 16 AtIDD, 15 OsIDD, and 23 ZmIDD geneswere downloaded from the Phytozome v12.1 database (https://phytozome.jgi.doe.gov/pz/ portal.html, accessed on 30 May 2019). These were used as query sequences to search against moso bamboo databases downloaded from GigaDB [36] (https://trace.ncbi.nlm. nih.gov/Traces/sra/sra.cgi, accessed on 31 May 2019) and BambooGDB [31,51] (http: //server.ncgr.ac.cn/bamboo/down.php and http://www.bamboogdb.org/, accessed on 31 May 2019) with cutoffs of "e-value ≤ 0.001" and "alignment-length ≥ 100" using Blast (version 2.2.25). Redundant sequences and sequences that lacked the complete IDD domain were removed after a similarity comparison. The resulting sequences were used to search the National Center for Biotechnology Information (NCBI) database to confirm their identities as IDDs. Several candidate PheIDDs were randomly selected and amplified from the cDNA of moso bamboo, cloned into the pMD18 vector (Takara, Japan), and Sanger-sequenced in both directions to verify their sequences. Gene-specific primers were designed using the online tool Primer3 (http://bioinfo.ut.ee/primer3-0.4.0/, accessed on 16 June 2019) and are listed in Table S3. Putative PheIDD genes were named PheIDD1 to PheIDD32 based on their homologyto the maize ZmID1.

Analysis of Chemical Characteristics
The chemical characteristics of all predicted proteins, including CDS length, amino acid number, and physicochemical parameters, were analyzed using the online analysis tool ProtParam (http://web.expasy.org/protparam, accessed on 18 August 2019).

Gene structures and Scaffold Locations
Structures of the PheIDDs were generated using the redraw gene structure function in TBtools (V1.068), and a scaffold distribution map was generated using the amazing gene location function in TBtools [52].

Screening of Gene Pairs and Calculation of Ka/Ks Ratios
PheIDD paralogous in moso bamboo and orthologous IDD gene pairs between moso bamboo and rice were identified as described in a previous study [53]. Ka/Ks ratios were calculated using the Simple Ka/Ks Calculator function of TBtools.

Synteny Analysis and Gene Duplication
The moso bamboo protein sequences were aligned to the protein sequences from rice using TBtools software. Duplication events of the PheIDD genes were confirmed as described previously [54]. The figures were visualized using Circos and Dual Synteny Plot in TBtools [52].

Prediction of Protein-Protein Interactions
Protein-protein interactions were predicted using the STRING online platform (http: //string-db.org/, accessed on 18 March 2020) based on interologs from the Arabidopsis database.

Multiple Sequence Alignment and Phylogenetic Analyses
Multiple sequence alignment of the 32 PheIDD proteins and the maize ZmID1 protein was performed using DNAMAN software (Version 6). Sequences of the rice, maize, and Arabidopsis IDD proteins were downloaded from the Phytozome v12.1 database (https: //phytozome.jgi.doe.gov/pz/portal.html, accessed on 18 November 2019).The accession numbers of the 86 IDDs used in this study are provided in Table S4. The amino acid sequences of IDD proteins were aligned using Clustal W software. A phylogenetic tree was constructed using the NJ method implemented in MEGA6 with 1000 bootstrap replicates.

Quantitative Real-Time PCR (qRT-PCR)
Total RNA was extracted using the TRIzol reagent (Invitrogen, Waltham, MA, USA), and RNase-free DNase I (Promega, Madison, WI, USA) was used to remove genomic DNA following the manufacturer's instructions. cDNA was synthesized with 2 µg of RNA using a PrimeScript RT kit (Takara, Shiga, Japan) following the manufacturer's recommendations. Gene-specific primers for the 32 PheIDD genes were designed using the online Primer3 tool. Due to the high similarity of PheIDD N-terminal sequences, we avoided these conserved regions when designing primers. Primers were designed in the 5 and 3 UTR regions to ensure specificity when necessary. PCR products obtained using each primer pair were validated by Sanger sequencing.
The qRT-PCR experiment was performed using the SYBR Color qPCR Master Mix reagent (Vazyme, Nanjing, China) and the CFX 96 Real-Time System (Applied Biosystems, Waltham, CA, USA). The 15-µL qRT-PCR assay mix contained 1.2 µL cDNA (diluted 1:8), 7.5 µL 2 × SYBR Color qPCR Master Mix reagent (Vazyme), 0.5 µL of each primer (100 µM), and 5.3 µL distilled deionized H 2 O. The reaction procedure was as follows: initial denaturation at 95 • C for 3 min, followed by 39 cycles of 95 • C for 10 s and 58 • C for 30 s. The specificity of the amplified products was verified by a melting-curve analysis. The PheNTB gene (GenBank accession No. gi|242381788) was used as the internal control to normalize the transcript levels of target genes [55]. The relative transcript levels of target genes were calculated and analyzed using the 2 −∆∆Ct method [56]. Three biological replicates and three technical replicates were analyzed for each sample.The primers are listed in Table S3.

Yeast Two-Hybrid Assay (Y2H)
The CDSs of PheIDD21α and PheIDD21β were cloned into the pGBKT7 vector, and the CDSs of three predicted target genes (PheDELLAa, PheDELLAb, and PheSCL3) were cloned into the pGADT7 vector. The pGBKT7 vector and pGADT7 vector were co-transformed into the yeast strain AH109. The co-transformed yeast was cultured in SD/-Trp/-Leu (-DDO) agar medium at 30 • C for 3 days and then transferred to SD-Trp-Leu-His-Ade (-QDO) for 3-5 days at 30 • C to identify the interaction in yeast. Y2H assays were performed according to the Clontech Matchmaker GAL4 Two-Hybrid System 3 & Libraries User Manual instructions. The positive and negative controls were from the kits. The primers are listed in Table S3.

Yeast One-Hybrid Assay (Y1H)
The CDSs of PheIDD21α and PheIDD21β were cloned into the prey vector (pGADT7), and the 230-340-bp promoter fragments of IDD-binding sites (four putative elements in PH02Gene17121, three putative elements in PH02Gene35441,and two putative elements in PH02Gene11386) were amplified and ligated into the pHIS2 vector (Clontech). The corresponding pGADT7 and pHIS2 constructs were co-transformed into yeast strain Y187 following the manufacturer's instructions for the yeast one-hybrid library screening system (Clontech). Interaction of pHIS53 with pGADT753 was used as a positive control, and the empty AD and pHIS-pro-frag were used as negative controls. The transformed strains were cultured at 30 • C for 2-3 days on SD/-Trp/-Leu media and were used for spot assays on SD/-Trp/-Leu/-His media with 10 mM 3-AT.The primers are listed in Table S3.

Statistical Analysis
The statistical analysis of all data in this study was carried out in SPSS 11.5 software (Chicago, IL, USA). Analysis of variance (the GLM procedure) was used to analyze the data.
Data are presented as means ± standard deviation (SD). Graphs were generated using Sigmaplot 12.5 (Systat Software Inc., San Jose, CA, USA).

Conclusions
In this study, we identified and characterized 32 PheIDD proteins that contained a complete IDD domain. They were unevenly distributed among 16 scaffolds. Based on their amino acid sequences, the 32 PheIDDs were divided into nine groups. Synteny analysis indicated that 30 PheIDD genes were orthologous to those in rice. During the evolution of PheIDD genes, segmental duplication events played an important role. Expression analysis revealed the diverse expression patterns of PheIDDs in different tissues. In addition, PheIDDs responded differently to exogenous applications of the plant hormones IAA and 6-BA, and their expression patterns were different during the development of aboveground branches and underground buds/shoots. A few PheIDD genes were predicted as candidate genes to study their functions during bamboo shoot development. Finally, we determined that one candidate gene, PheIDD21, underwent AS and generated two transcripts. Both isoforms formed homodimers and heterodimers and bound to the cis-elements of three genes. These findings increase our understanding of the characteristics of the PheIDD gene family and provide valuable information for further functional characterization of PheIDDs in the shoot development of moso bamboo.