Growth rate of axillary buds of plants grown under different N concentrations
To explore the influence of different N concentrations on axillary bud growth, rice seedlings were grown in hydroponic solutions with six concentrations of ammonium nitrate as the N source. We defined 2 mM N supplies by 1 mM ammonium nitrate as the optimal concentration for the growth rice seedlings, which was revealed in a previous report [38]. Under LN concentrations (0.5 and 1.0 mM), the lengths of the first and second axillary buds of 18- to 33-day-old seedlings were shorter than those under the optimal N concentration (Fig. 1a, b), whereas when the seedlings were grown under 5 mM and 10 mM N concentrations, the lengths of the first and second axillary buds were longer. Remarkably, when the N concentration reached 15 mM, the rice seedlings presented axillary buds that were shorter than those of seedlings grown under 5 mM and 10 mM (Fig. 1a, b). These results indicate that different N concentrations significantly influence the outgrowth of axillary buds.
Transcriptomic profiles of basal parts and axillary buds of plants grown under different N concentrations
To investigate the regulatory mechanisms underlying the axillary bud growth of rice plants grown under different N concentrations, we collected a mixture of the first and second axillary buds and adjacent basal parts of 30-day-old rice seedlings grown under six different N concentrations for RNA-seq analysis (Fig. 1c). In total, 39.6 to 64.6 million clean reads were generated in the libraries, and more than 90% of these reads were mapped to the reference genome (Additional file 1: Table S1). The gene expression levels were estimated using the fragments per kilobase of transcript per million reads (FPKM), and the clean reads were aligned to 31,937 genes whose FPKM was greater than 1 for at least one sample, including 25,378 protein-coding genes, 5,028 non-coding RNAs, and 1,530 novel transcripts. The principal component analysis (PCA) results showed good consistency between two replicates (Additional file 1: Figure S1a). Remarkably, the expression profiles of the basal parts and axillary buds were highly separated by principal component 1 (PC1), and the expression profiles of the axillary buds of plants grown under different N concentrations were separated by principal component 2 (PC2). Together with the results of Pearson’s correlation coefficients (PCCs) (Additional file 1: Figure S1b), overall, the results indicated that the transcriptomic profiles are diverse enough to identify genes responsible for N effects on axillary bud growth.
Expression profiling of genes involved in N transport and axillary bud outgrowth
To investigate the genome-wide responses of genes associated with N utilization and the axillary bud growth of plants grown under different N concentrations, we analysed the expression patterns of related genes whose FPKM value was greater than 1 for at least one out of 12 samples. In both the basal parts and axillary buds, the expression levels of these genes significantly changed in response to the different N concentrations. Of the genes involved in nitrate uptake, transport and assimilation, OsNPF7.1, OsNRT2.3, OsNPF2.4, OsNRF6.5, and OsNPF7.7 were differentially expressed in both the basal parts and axillary buds (Fig. 2a). In the basal parts, the expression of OsNRF6.5 and OsNPF7.1 was induced under both LN and HN conditions; however, the expression of OsNRT2.3 and OsNPF2.4 was induced only under HN conditions, while that of OsNPF7.7 was induced only under LN conditions (Fig. 2a), indicating that these five genes play crucial roles in promoting or inhibiting axillary bud growth in plants grown under different N concentrations. In addition, the expression of OsNR1 was significantly induced in the basal parts under HN concentrations, indicating that nitrate was assimilated under high concentrations to adapt to the high-nitrate environment. Remarkably, the expression of OsNRF6.5, OsNPF7.1, OsNPF7.2, OsNPF7.3, and OsNPF7.7 was induced at a concentration 0.5 mM N (Fig. 2a), suggesting that these genes might function in activating dormant axillary buds to grow.
Of the genes involved in ammonium transporter, the expression of OsAMT1;1, OsAMT2;1, OsAMT3;1, and OsAMT3;2 was induced at certain LN concentrations in both the basal parts and axillary buds (Fig. 2b), indicating that these genes play roles in ammonium uptake to meet the basic needs for rice growth. The expression of OsAMT1;1, OsAMT2;1 and OsAMT3;2 were also induced under HN conditions in the basal parts to take up additional ammonium to adapt to the high-ammonium environment. The expression of genes involved in ammonium assimilation was induced in basal parts and/or axillary buds of plants grown under certain HN concentrations. Specifically, the expression of OsGS1;3 and OsNADH-GOGAT1 was induced under 10.0 mM and 15.0 mM in the basal parts; the expression of OsNADH-GOGAT2 and OsGS2, at both 5.0 mM and 10.0 mM in the axillary buds; and the expression of OsGS1;2 under 10.0 mM in the basal parts and 5.0 mM in the axillary buds. These results indicate that the weakened ammonium assimilation may be a reason for the inhibited outgrowth of the axillary buds of plants grown under LN concentrations. The increased ammonium assimilation under 5.0 mM and 10.0 mM may be a reason for the accelerated outgrowth of the axillary buds, but ammonium toxicity under 15.0 mM inhibited axillary bud growth. Additionally, with the exception of that of OsAPP13 and OsAPP14, the expression of most OsAPPs was induced in the basal parts under LN and HN conditions (Fig. 2c), indicating that the induction of OsAPP gene expression may be an adaptation to ammonium concentrations in plants and the external environment but that this induction is not the main cause of axillary bud outgrowth.
Many genes have been reported to regulate axillary bud formation and outgrowth, but whether and how these genes respond to different N concentrations are unclear. The expression of all these genes involved in axillary bud formation and outgrowth was induced at certain concentrations in the basal parts and axillary buds of rice plants (Fig. 2d). Genes involved in SL biosynthesis, including OsD27, OsD17, and OsD10, were highly expressed under LN concentrations in the basal parts and axillary buds, indicating that LN concentrations can promote SL biosynthesis, resulting in the suppression of axillary bud growth. However, genes involved in SL signalling, such as OsD3, OsD14 and OsD53, were highly expressed under 15.0 mM N in the basal parts, indicating that enhanced SL signalling might play a role in ammonium toxicity to inhibit axillary bud growth. OsMOC1 was highly expressed under 0.5 and 1.0 mM N in the axillary buds, which is consistent with its function in activating axillary meristem activity [2]. In addition, the expression of OsCKX4 was induced at LN conditions in both the basal parts and axillary buds and under 15.0 mM N in the basal parts, indicating that the upregulation of OsCKX4 plays a negative role in axillary bud growth, which is consistent with the results of a previous study [19]. OsDLT, which is regulated by the BRs pathway, coinciding with its role in branching in Arabidopsis [39], was highly expressed under 10.0 mM N in the basal parts and axillary buds. Overall, the expression profiles of these genes under different N concentrations are responsible for the N response and the growth rate of axillary buds, which is consistent with previous studies, confirming that our transcriptomic data reflect the dynamic and complex developmental process of axillary bud outgrowth of plants grown under different N concentrations.
DEGs and their expression patterns in the basal parts and axillary buds of plants grown under different N concentrations
To explore the dynamic expression patterns of genes in response to different N concentrations, we identified DEGs between the optimal and each LN or HN concentration in the basal parts and axillary buds and identified the DEGs whose expression differed specifically between those two tissues at each N concentration. We found that some DEGs were specific and common between LN and HN concentrations and between both tissues (Additional file 1: Figure S2). In total, 10,221, 12,180, and 21,284 DEGs that respond to N concentrations were identified in the basal parts, in the axillary buds, and between those two tissues, respectively (Figs. 3a, 4a, Additional file 1: Figure S5a). We then clustered these DEGs into 9, 10, and 13 coexpression modules based on their expression patterns under the different N concentrations (Figs. 3b, 4b, Additional file 1: Figure S5b). To investigate the processes associated with N concentrations and axillary bud growth, we conducted a Gene Ontology (GO) enrichment analysis for each module.
With respect to the 10,221 DEGs in the basal parts, we found that, compared with the numbers under NN conditions, more of the DEGs were downregulated at each N concentration than that upregulated (Fig. 3a). Based on the 9 coexpression modules (B01-B09) in the basal parts, the genes could be divided into LN-responsive genes (B01), NN-responsive genes (B08), and HN-responsive genes (B03, B04, and B06). B01, B08, B03, and B04 were specific to 1.0, 2.0, 5.0, and 15.0 mM N conditions, respectively, and B06 was specific to both 10.0 mM and 15.0 mM N conditions. The LN-responsive genes in B01 were involved mainly in the response to auxin, hormone-mediated signalling pathways (auxin), cellular response to chemical stimulus/stimulus, response to endogenous stimulus/light stimulus/chemical, regulation of cellular process/transcription/metabolic process, RNA metabolic process, and gene expression. These results indicate that auxin plays important roles in the LN response to inhibit axillary bud outgrowth, which is consistent with the shortened axillary buds under LN conditions. The NN-responsive genes in B08 were enriched in the oxidation-reduction process, response to oxidative stress/stress/stimuli, apoptotic process, death, and cell death (Fig. 3c). Unfortunately, the HN-responsive genes in B03, B04, and B06 were not enriched in biological processes at a cutoff of a false discovery rate (FDR) < 0.05 (Fig. 3c). In contrast, the expression of genes in B05 was upregulated under both 2.0 and 10.0 mM conditions, and these genes were related to protein modification, cellular macromolecule biosynthetic and metabolic processes, phosphorus (P)-containing compound metabolic process, P metabolic process, RNA biosynthetic process, and glucan metabolic process (Fig. 3c). In addition, the enriched GO terms within the molecular function category in B05 included acid amino acid ligase activity and ubiquitin-protein transferase activity (Additional file 1: Figure S2). These results indicate that the promotion of axillary bud growth of plants grown under HN conditions may be due to enhanced metabolic processes involving P and sugars, which coincides with their functions in axillary bud growth [40, 41].
With respect to the 12,180 DEGs in the axillary buds of plants grown under N treatments, there were more downregulated genes than upregulated genes under the 0.5 and 1.0 mM N conditions, but the opposite occurred under the 5.0, 10.0 and 15.0 mM N conditions (Fig. 4a). All 10 coexpression modules (A01-A10) under N treatments were discernibly separated according to each N concentration (Fig. 4b). The LN-responsive genes (represented by modules A01, A02, and A05) were highly expressed under the 0.5 and 1.0 mM N conditions and were related to primary and secondary metabolism, including P, lipids, terpenoids, diterpene phytoalexins, biogenic amines, amino acids, nitrogen compounds, organic acids, cellular ketones, monocarboxylic acids, and glucose, indicating that ammonium uptake and assimilation are active and mainly utilized by plants grown under LN conditions. In addition, the LN-responsive genes were also involved in protein modification, protein phosphorylation, glycolytic process, carbohydrate catabolic process, defense response, response to stimulus, apoptotic process, death, RNA metabolic process, metal ion transport. The NN-responsive genes (represented by module A03) were highly expressed under 2.0 mM N. The overrepresented terms were associated with metabolic process of various substances, including cell wall macromolecule, polysaccharide, chitin, aminoglycan, amine, and other processes, such as oxidation-reduction, response to abundant production of amino acids in the process of nitrogen metabolism. The HN-responsive genes (represented by modules A04 and A06) were highly expressed under 5.0 mM N and were involved in photosynthesis-related processes, heterocycle biosynthetic process, cellular nitrogen compound metabolic process, oxidation-reduction process, and response to oxidative stress. Remarkably, the expression of genes involved in microtubule-based process, cell cycle, cell division, cellular component biogenesis, DNA packaging, DNA replication, nucleosome organization, M phase, DNA and RNA metabolic process, and organelle fission and organization were upregulated under 10.0 mM N in module A09, indicating that the accelerated axillary bud growth was mainly due to enhanced cell division. In addition, the expression of genes involved in the regulation of transcription, developmental processes, and multicellular organism development was upregulated under 15.0 mM N in the A07 module.
With respect to the 21,284 tissue-specific DEGs (Additional file 1: Figure S5a), 13 coexpression modules were clustered (Additional file 1: Figure S5b) and were discernibly separated into two groups: a basal tissue-specific group (BA01-BA06) and an axillary-specific group (BA07-BA13). Moreover, the axillary-specific modules (BA07-BA13) showed relatively clear expression patterns of genes in response to LN and HN conditions (Additional file 1: Figure S5b), which is consistent with the patterns of DEGs in the axillary buds (Fig. 4b). In addition, the genes enriched in various biological processes also coincided with the DEGs in the basal parts (Fig. 3c) and axillary buds (Fig. 4c); these processes include P metabolic process and response to auxin in BA04, photosynthesis-related process in BA07, and cell cycle-related processes in BA13 (Additional file 1: Figure S5c). Therefore, these results revealed not only the gene clusters that aggregated in certain tissues at certain N concentrations but also the underlying molecular mechanisms that regulate the responses of axillary bud growth to different N concentrations.
Expression profiles of genes related to cell division and expansion in the basal parts and axillary buds of plants grown under different N concentrations
The growth rate of the first and second axillary buds increased with increasing N concentrations ranging from 0.5 to 10.0 mM N but decreased under 15.0 mM. Based on the enriched processes associated with A09 in the axillary buds, we inferred that cell division and cell expansion are the predominant processes that determine the growth rate of the axillary buds. The expression profiles of genes involved in the cell cycle and cell division displayed tissue and N-response specificity (Fig. 5a). Most cell division-related genes were specifically expressed in the axillary buds (Fig. 5a). Remarkably, the axillary bud-specific cell division-related genes were highly expressed at HN concentrations, displaying A07, A08, A09, and A10 patterns, especially under 10.0 mM, indicating that cell division is active at HN concentrations to promote axillary bud growth.
Cell wall relaxation and loosening determine the extent of cell expansion. Similar to cell division-related genes, genes encoding cell wall relaxation- and loosening-related proteins, such as xyloglucan endotransglucosylase/hydrolase (XTH) and expansins (EXPs), exhibited tissue and N-response specificity (Fig. 5b, c) as well. Importantly, most of the OsEXPs were highly expressed in the axillary buds at all N concentrations except 15.0 mM, displaying A01-A06 and A08-A10 patterns (Fig. 5b). Furthermore, most of the OsXTHs were highly expressed in the axillary buds under 5.0 and 10.0 mM N, displaying A06 and A09 patterns (Fig. 5c). These findings indicate that cell expansion is active under 0.5 to 10.0 mM N to regulate axillary bud growth. Taken together, these results indicate that the accelerated growth of the axillary buds is made possible by cell division and expansion and that the low expression of cell expansion-related genes resulted in the suppression of the axillary bud growth of plants grown under 15.0 mM N.
Expression dynamics of TFs in the basal parts and axillary buds of plants grown under different N concentrations
Because TFs are the main regulators that alter the expression of transcripts, to identify which TF families play a more important role in axillary bud growth in response to different N concentrations, we analysed the expression profiles of TF genes in detail. Of the total 1,822 TF genes in rice on the Plant Transcription Factor Database (PlantTFDB), 908 TF genes within 52 families were differentially expressed; among them, 599 TF genes were classified as belonging to the B01 to B09 clusters (Additional file 1: Figure S7a), and 716 were classified as belonging to A01-A10 (Additional file 1: Figure S7b). Additionally, in the basal parts and axillary buds, more than half of these TF genes (52.1% and 55.9%, respectively) were expressed at the highest levels in response to HN stimuli, and the expression levels of only 16.2% and 22.8%, respectively, peaked in response to LN stimuli. To investigate the expression levels of members of TF families in the basal parts and axillary buds of plants grown under N stress, we calculated the total FPKM values of all TF members within a family (Additional file 1: Figure S8). We also explored the expression trends of TF families via the proportion of TF members within a family relative to the total TF members within a TF family, the information of which provided by PlantTFDB (Fig. 6).
For the basal parts, the TF gene in the LSD family was induced specifically under LN stress (Fig. 6a), which might indicate involvement in cell death under oxidative stress [42]. The expression of many TF family members was induced under HN stress, including RAV, YABBY, HD-ZIP, ZF-HD, ERF, MYB, AP2, TCP, bHLH, GATA, HSF, Trihelix, Dof, GRF, DBB, NAC, TALE, WOX, B3, and GRAS family members (Fig. 6a). Together with the expression levels of TF family members (Additional file 1: Figure S8), YABBY, HD-ZIP, ZF-HD, TCP, and GRF TFs were expressed at low levels under HN conditions, whereas TALE and WOX TFs, which are involved in meristem formation or maintenance [43] and embryonic patterning [44], respectively, were moderately highly expressed under HN conditions. GATA, NAC, and B3 TFs, which are involved in auxin signalling [45] and axillary bud development [46], were highly expressed under 5.0 mM N. Trihelix TFs were highly expressed under 5.0 and 10.0 mM N. ERF, MYB, bHLH, HSF, GRAS TFs, which are involved in stress and ethylene responses [46], phosphate-starvation responses [47], and abiotic stress responses [48], were highly expressed under 10.0 and 15.0 mM N. RAV, AP2, Dof, and DBB TFs were highly expressed under 15.0 mM N, indicating their roles in inhibiting the axillary bud growth of plants grown under HN stress.
With respect to axillary buds, the expression of members of the SRS, AP2, ARR-B, B3, YABBY, and ARF TF families was induced under LN conditions (Fig. 6b). Of these TFs, the SRS and AP2 TFs were highly expressed under 0.5 mM, and the expression of the ARR-B and ARF TFs was high and low under 1.0 mM N, respectively, indicating that they play various roles in axillary buds under N stress. Members of the CO-like, GATA, NF-YB, HD-ZIP, Trihelix, NAC, BES1, GeBP, and C3H TF families were highly expressed under 5.0 mM N, whereas TCP TFs under 10.0 mM and GRF TFs were expressed at low levels under 10.0 and 15.0 mM. Among these TFs, the TCP and GRF TFs are involved in cell proliferation [49, 50], supporting a role in cell division in the axillary bud growth of plants grown under HN conditions.
Global analysis of phytohormone signals triggered by different N concentrations
To reveal the functions of phytohormones involved in the axillary bud growth of plants grown under different N concentrations, we identified 242 DEGs associated with eight types of hormones (auxin, CK, SLs, BRs, abscisic acid (ABA), ethylene, gibberellic acid (GA), and jasmonic acid (JA); Figs. 2d,. 7, Additional file 1: Figure S9). We classified the hormone-related genes into four types on the basis of their function: 1-biosynthesis, 2-degradation, 3-transport, and 4-signalling and response. The expression levels of these hormone-related genes displayed tissue- and N concentration-specificity. Additionally, more genes were highly expressed in the basal parts than in the axillary buds, suggesting that different N concentrations might regulate axillary bud growth mainly by altering hormone signals in basal parts.
SLs, auxin, CK, and BRs are well known to regulate axillary bud formation and outgrowth: SLs and auxin act as inhibitors, whereas CK and BRs act as promoters [51–53]. Consistent with SLs as inhibitors of axillary bud growth, biosynthesis-related genes (OsD27, OsD17, and OsD10) were highly expressed under LN conditions, whereas signalling-related genes, such as OsD3 and OsD14, were highly expressed under 15 mM N (Fig. 2d). For auxin, the genes related to biosynthesis and transport, such as OsYUCCAs and OsPINs, presented relatively high expression levels in the basal parts under 1.0 mM N (Fig. 7a), suggesting that enhanced auxin biosynthesis in the shoot meristem and the oscillation of the auxin gradient are involved in axillary bud development under LN stress. These results coincide with previously reported results in which enhanced biosynthesis and transport of auxin negatively affect rice tillering [51]. Moreover, consistent with the enriched GO terms associated with B01 (Fig. 3b, c), most of the AUX/indole-acetic acid (IAA) genes were expressed at relatively high levels under 1.0 mM N, but the expression of ARFs exhibited various changes, suggesting that auxin plays complex and crucial roles in regulating the axillary bud growth of plants grown under different N concentrations. The expression of genes related to CK biosynthesis exhibited tissue specificity in the basal parts and axillary buds under both LN and HN conditions (Fig. 7b). Of which, OsIPT4 had peak expression level at 10.0 mM N and moderate expression levels at 2.0 and 5.0 mM in basal part, and the expression of OsCKX2 and OsCKX4 was high under 15.0 mM N, which is consistent with CK acting as a promoter of axillary bud growth [51, 52]. Of the CK signalling- and responsive-related genes, the expression of the type A-ARRs (OsRR1/4/7/6/9/10) and type B-ARRs (OsRR24/26) was upregulated and downregulated, respectively, under 15.0 mM N in the basal parts, indicating the roles of CK signalling in inhibiting the axillary bud growth of plants grown under HN stress. For BRs, the biosynthesis and signaling related genes displayed deverse response to different N concentrations (Fig. 7c). Of these genes, OsDLT was reported to positively regulate rice tiller number [53], and its expression was triggered in response to 10.0 and 15.0 mM N, suggesting that BRs play a role in the N response to regulate axillary bud growth.
Additionally, ABA, GA, ethylene, and JA also play crucial roles in plant development and the abiotic stress response. Under HN conditions, ABA biosynthesis-related genes were highly expressed in the axillary buds, especially under 5.0 mM N, but they were expressed at low levels in the basal parts (Additional file 1: Figure S9a). In contrast, under HN conditions, degradation-related genes were expressed at low levels in the axillary buds but were expressed at high levels in the basal parts (Additional file 1: Figure S9a), indicating that the amount of ABA may be increased in the axillary buds under HN stress to promote their growth. ABA signalling-related genes were also highly expressed under HN stress, suggesting that ABA may preferentially function under HN stress to be involved in axillary bud growth. OsGA20ox1, which encodes the rate-limiting enzyme in GA biosynthesis, was expressed at high levels in the axillary buds of plants grown under LN conditions (Additional file 1: Figure S9b), whereas the expression of GA degradation-related genes (OsGA2ox3/4/5) was low in the axillary buds (Additional file 1: Figure S9b). These results suggest that the upregulated biosynthesis genes and downregulated degradation genes may result in increased GA levels in the axillary buds under LN conditions. The expression of the ethylene biosynthesis-related genes OsACS2 and OsACO7 was upregulated in the axillary buds under 1.0 mM N, and the expression of OsACS2, OsACO5, and OsACO6 was upregulated in the axillary buds under 15.0 mM N. In contrast, ethylene degradation-related genes were expressed at low levels under 15.0 mM N (Additional file 1: Figure S9c), indicating that ethylene levels might increase in axillary buds and function under 1.0 and 15.0 mM N conditions to suppress axillary bud growth. JA biosynthesis genes, such as OsDAD1;2, OsLOX2;1, OsLOX2;3, OsAOS1, OsAOS2, OsAOC, and OsJAR1;2, were highly expressed in the basal parts under 15.0 mM N, suggesting that JA-isoleucine (JA-Ile) levels increased in the basal parts under 15 mM N to repress axillary bud growth.
Dynamic expression of genes involved with potassium (K) and phosphate in plants grown under different N concentrations
Mineral nutrients are required for plant development, especially N, P, and K. To determine the interactions among N, P, and K involving axillary bud growth, we examined the expression profiles of genes associated with P and K under different N concentrations. Overall, the expression patterns showed that genes associated with P and K displayed distinct expression patterns under different N concentrations (Additional file 1: Figure S10), suggesting that different N concentrations affect the uptake and localization of P and K. The expression of the K transporter-associated OsHAK genes showed tissue specificity; for example, OsHAK12/10/4/23/7/9/27 were expressed specifically in the basal parts, and OsHAK16/17/1/19/13/22/24/18/11 were expressed specifically in the axillary buds (Additional file 1: Figure S10a). The expression of several OsHAKs and OsHKTs was induced specifically in the axillary buds under 5.0 mM N (Additional file 1: Figure S10a), indicating that increased N concentration may activate the uptake and transport of K to coordinately promote axillary bud growth. With respect to P transporters, OsPT1 was highly expressed under LN stress, increasing P uptake and remobilization. In contrast, the expression of vacuolar P transporters such as OsSPX-MSF2 and OsSPX-MSF3 displayed tissue specificity; these genes were highly expressed in the basal parts under 15.0 mM N, whereas OsVPE1 and OsVPE2 were highly expressed in the axillary buds under 5.0 mM N (Additional file 1: Figure S10b). The expression of most of the genes involved in P signalling and response was induced with increasing N concentration ranging from 2.0 to 10.0 mM N (Additional file 1: Figure S10c), suggesting that the increased N concentrations may activate new signalling pathways to promote axillary bud growth. In summary, increased N concentrations promote the uptake and localization of nutrients and trigger their signalling pathways to promote axillary bud growth.
Identification of ammonium assimilation genes associate with altered tiller numbers in rice
Transcriptome analysis of basal parts and axillary buds under different N concentrations can provide important clues for the identification of novel genes that regulate axillary bud growth and tiller number in rice. In the present study, the genes OsGS1;2 (Os03g0223400) and OsGS2 (Os04g0659100) are involved in ammonium assimilation and were found to be highly expressed in the axillary buds under 5.0 mM N (Fig. 8a, b), indicating that they may play roles in promoting axillary bud growth and, hence, increased tiller numbers at harvest. We generated transgenic rice lines that overexpressed and downregulated the expression of OsGS1;2 and OsGS2, respectively. Three overexpression (OE) lines (OE1, OE2, OE3) and RNA-interference (Ri) lines (Ri1, Ri2, Ri3) with increased and reduced expression levels of OsGS1;2 and OsGS2, respectively, were selected for study (Fig. 8c-f). We observed that the tiller numbers at the heading stage of the OE lines were significantly greater than that of wild type (Fig. 8g), whereas the Ri lines had significantly fewer tillers (Fig. 8g), which coincides with the reported phenotype of the OsGS1;2 knockout mutant [54]. Thus, these results indicate that OsGS1;2 and OsGS2 regulate axillary bud growth and tiller number via ammonium assimilation in rice, confirming the value of our transcriptomic data.