Over-expression of the photoperiod response regulator ZmCCT10 modifies plant architecture, flowering time and inflorescence morphology in maize

Maize originated as a tropical plant that required short days to transition from vegetative to reproductive development. ZmCCT10 [CO, CONSTANS, CO-LIKE and TIMING OF CAB1 (CCT) transcription factor family] is a regulator of photoperiod response and was identified as a major QTL controlling photoperiod sensitivity in maize. We modulated expression of ZmCCT10 in transgenic maize using two constitutive promoters with different expression levels. Transgenic plants over expressing ZmCCT10 with either promoter were delayed in their transition from vegetative to reproductive development but were not affected in their switch from juvenile-to-adult vegetative growth. Strikingly, transgenic plants containing the stronger expressing construct had a prolonged period of vegetative growth accompanied with dramatic modifications to plant architecture that impacted both vegetative and reproductive traits. These plants did not produce ears, but tassels were heavily branched. In more than half of the transgenic plants, tassels were converted into a branched leafy structure resembling phyllody, often composed of vegetative plantlets. Analysis of expression modules controlling the floral transition and meristem identity linked these networks to photoperiod dependent regulation, whereas phase change modules appeared to be photoperiod independent. Results from this study clarified the influence of the photoperiod pathway on vegetative and reproductive development and allowed for the fine-tuning of the maize flowering time model.

Introduction expressed in leaves with a diurnal pattern, peaking in the early morning under LDs. In contrast, its expression is low under SDs [41]. The GHD7 protein represses transcription of the floral inducer Early Heading Date1 (Ehd1) which promotes flowering under SDs [43]. EHD1 is a B-type response regulator with DNA binding properties suggesting it functions as a transcription factor [43]. Ehd1 is required for expression of the rice florigen genes Heading Date3a (Hd3a) [23] and Rice flowering locus T1 (RTF1), major floral activators under LD conditions [44,45]. The floral promoter Ehd1 and the floral repressor Ghd7 fine tune the expression of the Hd3a/RFT1 genes to recognize a critical day length for transition to reproductive development [46].
The Ghd7-Ehd1 genetic module seems to be conserved in other short day tropical grass crops as was demonstrated for sorghum [47,48] and also for maize [39,49]. In the long day winter crops wheat and barley, VERNALIZATION2 (VRN2), a homolog of Ghd7, represses FT-like genes prior to cold exposure [50,51]. But the Ghd7-Ehd1 module is not present in Arabidopsis or other eudicots [41,52].
Ghd7 over-expression or down-regulation in transgenic plants revealed its role as a central regulator of growth, development and stress response in rice [42]. Over-expression of Ghd7 affected plant architecture resulting in taller plants with thick stems, fewer tillers but increased panicle branching which led to more grain per plant [41,42]. Ghd7 is one of the major targets for increasing grain yield in rice breeding programs [53].
Our knowledge of how the ZmCCT10 gene functions at the molecular level is limited [39]. However, being a repressor of the photoperiod pathway, manipulation of ZmCCT10 expression provides an opportunity to study the role of this pathway in floral transition and other developmental processes. To investigate these roles, we over-expressed ZmCCT10 driven by constitutive promoters of different strengths in the day neutral early-flowering maize line Gaspé Flint. The resulting transgenic plants displayed dramatic modification of plant morphology producing tall, late flowering phenotypes with about 50% of the events showing vegetative reversion of the tassel to a branched, leafy structure (phyllody).

Plant materials
Extremely early temperate cultivar Gaspé Flint, temperate inbred line B73, tropical CML436 and CML311maize lines were used due to their distinct photoperiod sensitivities and differences in flowering times. Diurnal experiment using these lines was described in [30]. Teosinte lines were obtained from North Central Regional Plant Introduction Station, Ames, Iowa, USA. Teosinte accession# PI 441934 is Zea mays sp huehuetenangensis originally from Guatemala. Teosinte accession # PI 422162 is Zea luxurians originally from Mexico.

T-DNA constructs and plant transformation
GATEWAYTECHNOLGY (Invitrogen, CA) was used for vector construction. The co-integrated vectors were constructed, and maize Gaspé Flint plants were transformed as described previously [30,54]. Typically, 10 independent single copy events were generated for each construct. T1 seeds were generated by pollination with Gaspé Flint as a pollen donor.
CS184 (ft-2) and CS185(ft-3), the Arabidopsis later flowering mutants in the ectotype Lansberg erecta, were obtained from ABRC (Columbus, OH). Plants were grown under fluorescent lamps in growth chambers with 16 hr light period at 24˚C and 8 hr dark period at 23˚C and 50% relative humidity. Gateway technology (Invitrogen, CA) was used to construct transformation vectors. The binary vector, pBC.Yellow, was used for Arabidopsis transformation [55]. Using the flower dipping method, plants of Ler WT, CS-184 and CS-185 were inoculated by the Agrobacterium strain GV3101. Genes transformed into the binary vector was constitutively expressed under control of the promoter CaMV35S. Positive transgenic lines were identified by color selection from yellow fluorescent protein (YFP) and herbicide selection from the BAR gene. Seeds of at least three positive transgenic lines for every construct were saved for further study.

Phenotypic data collection
Vegetative growth stages (V stage) were defined according to the appearance of the leaf collar of the uppermost leaf [56]. The staging notes were taken twice a week. Using staging notes leaf appearance rate was calculated with linear regression models. Growth rate calculated as (H F -H 1 )/(DAP F -DAP 1 ) where H-height, F-final height, 1-first recorded measurement, where DAP stands for days after planting. Leaf length was measured on fully expanded leaves with a visible ligule as the distance from the leaf collar/ligule to the tip of the blade. Leaf width was measured at the widest point of the blade. The ratio of the length to the width was calculated. Nodes were identified by the leaf number originating from that node. Internode lengths were measured when fully mature plants were harvested and were calculated as the distance between nodes. Only above-ground internodes were measured.

Toluidine blue O staining of epidermal peels
To determine when plants switched from juvenile to adult growth, leaf sections were collected from all leaves starting with leaf # 2 until adult traits were detected. Leaf segments from margin to margin were collected from the base and the tip of each leaf. Segments were fixed in a mixture of 1-part ethanol to 3 parts acetic acid [57]. The abaxial epidermis and mesophyll were removed using abrasive techniques. Once cleared, the adaxial epidermis was stained with 0.05% toluidine blue O/acetic acid, pH 4.5 (TBO solution) [7,58] for 30 seconds. The peel was then rinsed with deionized water and immediately photographed using bright-field optics on a microscope.

Tissue collection for qRT-PCR
Plants were grown in the greenhouse under LD conditions with a 14-hour day length. To cover much of the developmental range of candidate gene expression, coleoptiles were sampled from NTG (non-transgenic plants) and UBI pro :ZmCCT10 transgenic plants but not from the BSV pro :ZmCCT10 plants due to limited seed availability. Afterwards, the first true leaf was sampled, and then leaves were sampled (with a punch) twice a week beginning from leaf # 4. Due to expression of the maize florigen gene ZCN8 at the tip of the leaf [30], leaves were consistently sampled 3-5 cm from the tip. To capture diurnal expression patterns, leaves were sampled in the morning when lights were turned on in the greenhouse and 12 hours later in the evening. When the BSV pro :ZmCCT10 plants reached 24 leaves, it was technically challenging to sample such tall plants and thus leaves 24-31 were sampled at the final dissection of the plants in the afternoon. Tissues collected for qRT-PCR were immediately flash frozen after being sampled. Additional NTG and UBI pro :ZmCCT10 transgenic plants were grown in a Conviron CMP6050 growth chamber under 14-hour days (28˚C days and 26˚C nights) to sample meristems. Meristems were dissected at every V stage starting at V1 until the meristem transitioned to a tassel (~V4 for NTG and~V4 to V7 for UBI pro :ZmCCT10 transgenic plants depending on the CCT10 allele). Due to limited seed availability, BSV pro :ZmCCT10 plant meristems were sampled in the greenhouse under LD conditions with a 14-hour day length during the initial T1 experiment. Meristems were sampled at stages V6, V9, V11, V13, V17 and V19 from adult plants and from plantlets at stages V32 and V33. Images were taken of each meristem with a Leica MSV269 dissecting scope, then immediately flash frozen to be used for qRT-PCR.

RNA isolation and qRT-PCR
Total RNA was isolated using Qiagen RNeasy reagents (https://www.qiagen.com/us/shop/ Lab-Basics/Buffers-and-Reagents/) with the nucleic acid bound to columns of a 384 well binding plate purchased from the Pall Corporation (http://www.pall.com/main/oem-materialsand-devices/product.page?lid=gri78l6g). DNA was removed from the RNA samples using Roche DNAse I Recombinant (https://www.lifescience.roche.com/shop/en/us/products/ dnase-i-recombinant-rnase-free) and synthesis of cDNA was done using Applied Biosystems High Capacity cDNA Reverse Transcription kits (https://www.thermofisher.com/order/ catalog/product/4368813?ICID=search-product). Quantitative PCR was done using hydrolysis probe and SYBR based reactions. Primers and probes were designed using Applied Biosystems Primer Express software (https://www.thermofisher.com/order/catalog/product/4363993? ICID=search-product) using nucleotide sequences published in Genbank. Hydrolysis probebased PCR was performed using Bioline Sensi-fast mix (http://www.bioline.com/us/sensifastprobe-lo-rox-kit.html) while SYBR-based PCR was run using Applied Biosystems PowerUp SYBR Green Master Mix (https://www.thermofisher.com/order/catalog/product/A25741? ICID=search-product). All reactions were run on an Applied Biosystem Viia7 Real-Time PCR instrument using the manufacturer's conditions. Relative gene expression was calculated by normalizing against maize eukaryotic initiation factor 4-gamma gene (GenBank accession # EU967723). A list of primers and probes are shown in S5 Table. To identify genes which expressions were different in transgenic apices compared to NTG plants, T-test was performed for every developmental stage. Expression level were considered statistically significant with p<0.05 (S6 Table).

Over-expression of ZmCCT10 produces dramatic effects on multiple traits in T0 maize plants
To investigate how over expression of ZmCCT10 may impact flowering, we used a transgenic approach and chose to constitutively over-express the ZmCCT10 coding region. To explore how diverse allelic variation of the ZmCCT10 protein might differentially alter flowering, ZmCCT10 alleles from different maize inbred lines with distinct flowering characteristics were selected, including the day-length neutral early flowering Gaspé Flint, the temperate B73, and the SD-sensitive tropical CML436 and CML311 lines [30]. Two teosinte accessions, the SDsensitive wild progenitor of maize, were also used (accessions PI 441934 and PI 422162). To identify conserved protein domains, maize ZmCCT10 and sorghum SbGHD7, the closest maize CCT10 homolog which functions as a floral repressor under long days, were compared [47]. Their amino acid alignment showed the proteins were conserved with 61% amino acid identity (S1 Fig). To investigate how the level of ZmCCT10 expression could affect flowering, we used two constitutive promoters of different strengths to drive expression of six different alleles. The maize ubiquitin promoter (UBI pro ) is widely used in cereal transgenic studies as a constitutive promoter with a high level of expression [59]. However, the Banana Streak Virus promoter (BSV pro ) exceeds the level of expression driven by the UBI pro and directs expression in all tissue tested with exception of pollen [60]. Using both the UBI pro and BSV pro and the genomic and/or cDNA sequences of six ZmCCT10 alleles, we constructed a cohort of constructs that provided a wide range of allelic and expression combinations in which to study the effect of ZmCCT10 overexpression on plant phenotype (S1 Table).
All constructs were transformed into the day-neutral early flowering Gaspé Flint line ( Fig  1A). Between 9-10 single copy transgenic events were generated for each construct. Because T0 transgenic plants were produced in batches, representative T0 data were collected for each experimental batch grown in the greenhouse at the same time (S1 Table).
As expected for over-expression of a flowering repressor, the transgenic plants exhibited a late flowering phenotype which was manifested as taller plants with more leaves compared to the non-transgenic (NTG) parental Gaspé Flint line (Fig 1B and 1C). Delayed flowering, recorded as time to pollen shed (shedding) and silk exertion (silking), was also observed (S1 Table). Typical time for Gaspe Flint to shed and silk is about 30 days.
In addition to the expected late flowering phenotype, the transgenic plants also displayed unexpected features. The UBI pro :ZmCCT10 T0 plants produced atypically long axillary branches subtending the ears (Fig 1B). Moreover, BSV pro :ZmCCT10 transgenic plants exhibited a severe modification to the entire plant architecture (Fig 1C and 1D). The BSV pro :ZmCCT10 transgenic plants were extremely tall, up to 300 cm, on average, and produced up to 38 leaves (Fig 1C and 1D  In contrast, the non-transgenic Gaspé Flint line produced only a few brace roots at the base of the plant (Fig 1F). Unexpectedly, as plants matured, a vegetative, highly branched leafy structure emerged from the top of many of the BSV pro :ZmCCT10 plants in lieu of tassels ( Fig 1C). More than 50% of the BSV pro :ZmCCT10 transgenic plants produced these vegetative apical structuresphyllody-(S2E Fig) and were composed of reversion plantlets. This phenotype was not observed in any of the NTG or UBI pro :ZmCCT10 plants.

Pleiotropic effects of ZmCCT10 are independent of allelic variations but are dosage dependent
We further examined ZmCCT10 overexpression effects in the T1 families. We selected 12 constructs composed of six cDNA ZmCCT10 alleles driven by either the UBI pro or BSV pro (S2 Table). The T1 families of 20 plants from two events were planted for the UBI pro : ZmCCT10 constructs. Due to the reduced fertility of the BSV pro :ZmCCT10 transgenic plants, the available T1 seed was limited and, in a few cases, only 3-4 plants were planted (S2 Table). Consistent with T0 observations, both UBI pro :ZmCCT10 and BSV pro : ZmCCT10 constructs modified plant morphology in the T1 plants, with the more extreme manifestations being observed in the BSV pro :ZmCCT10 plants. A weak allelic effect was detected in the UBI pro :ZmCCT10 constructs where UBI pro :ZmCCT10 Gaspé and UBI pro : ZmCCT10 B73 showed a smaller effect on plant height and leaf number. This allelic effect was not observed in the BSV pro :ZmCCT10 plants and we surmise that this lack of phenotypic effect may have been overridden by the very high level of transgene expression (S2 Table). Thus, we considered the level of transgene expression as the major factor correlated with phenotype severity. For this reason, we analyzed data by grouping the UBI pro : ZmCCT10 and BSV pro :ZmCCT10 results regardless of their allele background.
The T1 families showed a strong association between the level of transgene expression and trait measured. This

Modification of vegetative traits in T1 maize transgenic plants
The T1 transgenic families exhibited modification of plant phenotypes consistent with the T0 generation. The average BSV pro :ZmCCT10 T1 plant height was 263 cm compared to 105 cm for the UBI pro :ZmCCT10 plants and~70 cm for the NTG siblings ( Table 1). The growth rate of the BSV pro :ZmCCT10 plants was only 2.8 cm/day compared to~4.2 cm/day of the UBI pro : ZmCCT10 and NTG plants (Table 1). Despite their slower growth rate, the BSV pro :ZmCCT10 transgenic plants produced more internodes because they remained in a vegetative stage longer, which also resulted in the increased plant height.
Internode length depends on position within the plant shoot. As observed for NTG and UBI pro :ZmCCT10 plants, the internodes below the ear were progressively longer closer to the ear and then became shorter above the ear (Fig 2A). Thus, internode length displayed an acropetal, bottom-up gradient below the ear and then switched to a top-down basipetal gradient above the ear, which is characteristic for organ growth before and after floral transition [4]. The BSV pro :ZmCCT10 internodes were on average 1.5-fold shorter and 2-fold wider than NTG. No significant differences were found between UBI pro :ZmCCT10 and NTG plants ( Fig  2A and S5 Fig). Diameters of the BSV pro :ZmCCT10 stalks measured at the 1 st internode averaged 31 mm compared to 15 mm for NTG and 20 mm for UBI pro :ZmCCT10 plants (S3 Table). This observation may explain the slower relative growth of the BSV pro :ZmCCT10 plants since they are allocating more assimilates to stalk-width growth.
The UBI pro :ZmCCT10 plants produced an average of 13 leaves compared to 9-10 leaves for NTG plants, whereas the BSV pro :ZmCCT10 plants produced 36  Leaf size showed an acropetal growth pattern below the ear and a basipetal pattern above the ear (Fig 2). In NTG and UBI pro :ZmCCT10 plants, leaves below the ear displayed a sharp increase in length and width closer to the ear resulting in an increased length/width ratio. Leaves above the ear became shorter and narrower and the length/width ratio sharply declined (Fig 2B, 2C and 2D). Typically, the biggest leaf on a plant is the ear leaf. The BSV pro :ZmCCT10 plants displayed the characteristic acropetal pattern of leaf size but the basipetal pattern was disrupted after the 10 th leaf. Leaf size remained relatively constant above node 10 reflecting the longer and wider leaf shape (Fig 2B, 2C and 2D). This indicated an increased leaf growth rate, longer duration of growth, or both in the BSV pro :ZmCCT10 plants compared to NTG or UBI pro :ZmCCT10 plants. Consistent with the T0 observations, brace roots were found as high as node 37 in the BSV pro :ZmCCT10 plants, which was above the node of primary ear formation (Table 1); whereas NTG and UBI pro :ZmCCT10 plants produced brace roots below the ear (NTG  Over-expression of the photoperiod regulator ZmCCT10 in maize internode 4-6 and UBI pro :ZmCCT10 internode 5-9). Because brace roots typically form at juvenile nodes [7] we decided to examine phase change in T1 families. To determine juvenile and adult growth phases, we first observed the presence of macro-hairs (an adult trait). Macrohairs were present in NTG, UBI pro :ZmCCT10 and BSV pro :ZmCCT10 plants starting at leaf #4 (S4 Table). To determine juvenile and adult traits at the cellular level, we examined epidermal peels stained with Toluidine Blue O (TBO) [7]. Based on observations of bulliform cells, epidermal hairs, cell wall invagination, and TBO staining, all epidermal peels were juvenile at leaves 2-3 and began transitioning at leaf 4 (S7 Fig). NTG leaves were fully adult by leaf 7 and above, whereas UBI pro :ZmCCT10 and BSV pro :ZmCCT10 leaves were fully adult by leaf 8. There were no significant differences between NTG and transgenic plants for the juvenile-toadult transition based on cell morphology and TBO staining of epidermal peels. Therefore, we conclude that juvenile-to-adult phase change was not affected in these transgenic plants. Transgene expression had a significant effect on tassel branching. NTG plants produced tassels with an average of 7-9 primary branches and no secondary branches ( Table 2 and  For the NTG and UBI pro :ZmCCT10 plants, 93-100% of the plants produced ears that silked 2 to 6 days later than they shed pollen. The unusually long shank subtending the ears observed in T0 UBI pro :ZmCCT10 plants was reproduced in the T1 families. The average shank length of the UBI pro :ZmCCT10 plants was 31 cm which was a 3-fold increase compared to the~12 cm long shanks produced by NTG (Table 2). There was one event (11 plants) of the BSV pro : ZmCCT10 PI422162 construct that produced silking ears. These plants were used to measure ear and shank length for the BSV pro :ZmCCT10 plants ( Table 2). The transgenic ears were~2 cm longer than NTG ears ( Table 2). No differences in shank length were found. After careful dissection of the mature T1 BSV pro :ZmCCT10 plants with no visible ears, we found initiated ears but most them had aborted at early stages of development and never exerted silks.

A high level of ZmCCT10 expression caused formation of apically-induced plantlets
A striking feature of the BSV pro :ZmCCT10 transgenic plants was the development of the phyllody phenotype upon plant maturity ( Fig 1D). More than 50% of the T0 plants developed this phenotype (S2E Fig), which was reproduced in 43% of the T1 BSV pro :ZmCCT10 plants ( Table 2). This feature was never observed in the NTG or UBI pro :ZmCCT10 plants suggesting a strong relationship between the phyllody phenotype and the higher level of ZmCCT10 transgene expression.
Dissection of the highly branched vegetative structure from the T0 plants revealed complex arrangements of vegetative structures resembling multiple plantlets, each with an individual shoot axis (Fig 3). The number of countable plantlets varied from 2 to 26 (S2C Fig). Plantlets were at different stages of development with larger, more mature plantlets positioned on the periphery of the tassel and younger, less developed plantlets closer to the center. Some apices still had a central growing point that could generate more plantlets (Fig 3B and 3C). Dissection of the individual plantlets revealed a complex mix of abnormal tassel-like and ear-like organs. In a few cases, ear-like organs were fertile and could be fertilized to produce kernels (Fig 3A,  Row 4). Dissection of the T1 BSV pro :ZmCCT10 phyllody tops revealed the presence of apices consisting of only plantlets or a combination of plantlets and tassel branches (S9 Fig). The most mature looking plantlets (S 10A) when planted into soil developed large root systems (S10B, S10C and S10D Fig). However, the shoot growth of these transplanted plantlets was stunted with secondary plantlets developing in place of reproductive organs (S10C and S10D Fig).
To capture morphological changes in the meristem during and after floral transition, we dissected and imaged shoot apices from NTG and T1 transgenic sib plants. NTG plants developed very fast, as is typical for the Gaspé Flint line. Transitioning of the vegetative SAM to an inflorescence meristem (IM) occurred between V1-V2 stages (Fig 4A and 4B) followed by branch meristem (BM) initiation ( Fig 4C) and developing tassel (DT) (Fig 4D) at V2-V3. By the V4 stage the immature tassel (IT) was fully formed and committed to maturation (Fig 4E). The apices of the UBI pro :ZmCCT10 transgenic plants were sampled every V-stage until an immature tassel had formed (Fig 4F-4J). The floral transition occurred at the V3-V5 stages, two stages later than NTG plants (Fig 4G). UBI pro :ZmCCT10 apices showed initiation of multiple BMs (Fig 4H) that resulted in highly branched tassels (Fig 4I and 4J and S8B Fig). Due to the limited number of BSV pro :ZmCCT10 T1 plants, we sampled apices at fewer V-stages and were only able to observe the vegetative SAM up to the V19 stage (Fig 4K). At this late stage, the SAM was swollen at the base (Fig 4K), that was not observed in NTG or UBI pro :ZmCCT10 plants (Fig 4A and 4F). During removal and dissection of the very tall, latest flowering BSV pro : ZmCCT10 plants at 130 days after sowing, we found two plants with an IM, and several plants with developing tassels (DT), which were characteristically highly branched (Fig 4L and 4M). Interestingly, two BSV pro :ZmCCT10 plants were found that had a phyllody top, consisting of a developing tassel with plantlets initiating at lower positions ( Fig 4N).
To investigate the morphology of the apical meristem in the phyllody plantlets, we dissected the plantlets from several T1 BSV pro :ZmCCT10 plants. The younger plantlets had a vegetative appearing SAM while others had developed a typical BSV pro :ZmCCT10 highly branched immature tassel (S11A, S11B, S11E and S11F Fig). The shoot meristem of older plantlets had a mixed morphology, showing a combination of an immature main tassel rachis with numerous shoot-like structures basally positioned (S11C, S11D, S11M and S11O Fig), which we expected might form secondary plantlets. Other abnormalities were observed in the plantlet apical meristems, including ear-like organs surrounded by leaf-like primordia (S11J, S11K and S11L Fig) and a dissected spikelet containing both anthers and a cluster of pistillate florets (S11P, S11Q and S11R Fig).

Expression of flowering regulators in leaves affected by ZmCCT10 overexpression
To gain insight into the gene network regulated by ZmCCT10 in leaves, we selected candidate genes based on the regulatory flowering network models proposed for maize [61] and rice [62,63]. Even though ZmCCT10 might regulate as many as 1117 genes in leaves [39], we selected 20 representative genes with proven or predicted functions in flowering pathways in leaves and assayed their expression by qRT-PCR (S5 Table).
The circadian clock plays a central role in regulation of photoperiodic flowering time in plants and so several clock genes were assayed. The maize genes GIGZ1A/1B and CONZ1 are homologs of Arabidopsis flowering genes GIGANTEA (GI) and CONSTANS (CO) [64]. As previously shown, the duplicated genes GIGZ1A and GIGZ1B have identical expression patterns [64]. For this reason, GIGZ1A/1B were tested using the same assay (S5 Table). GIGZ1A/ Over-expression of the photoperiod regulator ZmCCT10 in maize 1B were not expressed in the coleoptiles but from leaf # 1 and onwards GIGZ1A/1B exhibited a typical diurnal expression pattern with an evening peak (Fig 5A). Hence GIGZ1A/1B genes appeared to be not regulated by ZmCCT10 and likely functions upstream of ZmCCT10 in the photoperiod pathway, as is the case in rice [62]. CONZ1 gene expression was slightly reduced in the UBI pro :ZmCCT10 transgenic plants and was significantly repressed in the BSV pro : ZmCCT10 plants suggesting its regulation by ZmCCT10 (Fig 5B).
The APETALA2-like transcription factor GL15, that regulates the transition from juvenile to adult leaf identity in maize and is a target of miR172 [15], was highly expressed in coleoptiles and in juvenile leaf # 1. Thereafter, its expression decreased in leaf # 4 (Fig 5C), which is consistent with our observation that the onset of the transition from juvenile-to-adult phases occurred in leaf # 4 (S7 Fig). GL15 was diurnally expressed with an evening peak and its expression pattern was nearly identical in NTG and transgenic plants suggesting no regulation by ZmCCT10 (Fig 5C).
The AP2-like ZmRap2.7 gene is a repressor of flowering and is a target of miR172 [65]. The ZmRap2.7 expression pattern was similar to GL15 which was higher in the coleoptiles and in juvenile leaf # 1 with a diurnal evening peak (Fig 5D). No significant differences were observed in NTG and transgenic plants placing ZmRap2.7 outside of ZmCCT10 regulation.
ZmEhd1 is a homolog of the rice Ehd1 activator of flowering under short days [43]. According to the rice photoperiod model, Ehd1 is directly repressed by Ghd7 [62,63]. In agreement with the rice model, expression of ZmEhd1 was reduced in the UBI pro :ZmCCT10 transgenic plants and completely repressed in the BSVpro:ZmCCT10 plants positioning ZmEhd1 downstream of ZmCCT10 (Fig 5E).
There are 15 FT-like genes in maize (the ZCN genes) that encode florigen-like proteins but only seven of them are expressed in leaves at some stages of development [28]. The ZCN8 and ZCN7 genes were shown to have florigen activity [29,30,66]. We examined the function of five leaf-expressed ZCN genes in Arabidopsis and found that four genes, ZCN12/14/15/18, displayed florigenic activity in Arabidopsis while ZCN26 did not (S12 and S13 Figs). This finding suggested the potential for six ZCN proteins (including ZCN7/8) to function as florigen triggers in maize. Expression of all seven leaf-expressed ZCN genes was examined in the ZmCCT10 transgenic plants. Expression levels of ZCN8, ZCN7, ZCN12 and ZCN15 were reduced in the UBI pro :ZmCCT10 transgenic plants and completely repressed in the BSV pro : Over-expression of the photoperiod regulator ZmCCT10 in maize ZmCCT10 transgenic plants (Fig 5F, 5G, 5H and 5I). Thus ZCN8, ZCN7, ZCN12 and ZCN15 appeared to be negatively regulated by ZmCCT10. Three of the ZCN genes, ZCN14, ZCN18 and ZCN26, exhibited complex expression patterns (Fig 5J, 5K and 5L). ZCN14 expression was slightly up-regulated in the UBI pro :ZmCCT10 transgenic plants and its expression was continuously elevated in the BSV pro :ZmCCT10 transgenic plants reaching its maximum in leaves 20-21 around the time when floral transition took place in at least some BSV pro :ZmCCT10 plants (Fig 5J). Thus, ZCN14 is not downstream of ZmCCT10 and may potentially be influenced by long day-dependent mechanisms. ZCN18 expression was slightly reduced in the UBI pro : ZmCCT10 plants but was up-regulated in the BSV pro :ZmCCT10 plants where it showed diurnal expression with a morning peak (Fig 5K). Thus, expression of ZCN18 appears to be induced by high levels of ZmCCT10 expression. ZCN26 showed the opposite expression pattern of ZCN18. It was up-regulated in the UBI pro :ZmCCT10 transgenic plants and was slightly repressed in the BSV pro :ZmCCT10 plants in leaves #10-11 and completely repressed in older leaves (Fig 5L). Because ZNC26 does not have florigenic activity and displays this unusual pattern, we speculate that its function may be unrelated to flowering time.
We also examined the TFL1-like ZCN1/3 genes which are antagonists of FT-like genes and have been shown to delay flowering and modify plant architecture when over-expressed in maize plants [54]. Out of six TFL1-like genes (ZCN1-6), only ZCN1 and ZCN3, are expressed in leaves at early growth stages [54]. The duplicated ZCN1 and ZCN3 genes showed higher expression in the coleoptiles and much lower expression in leaves at all stages of development in both NTG and transgenic plants ( Fig 5M) and thus, are likely not regulated by ZmCCT10.
The MADS box genes, ZMM4 and ZMM15 are markers of the floral transition in maize. They are activated during this transition, first in the reproductive inflorescences and then in the leaves [34]. ZMM4 expression increased in leaves two stages later after the floral transition in NTG and UBI pro :ZmCCT10 plants (Fig 5N). In the BSV pro :ZmCCT10 plants, a slight expression increase was observed in leaves # 28-30 ( Fig 5N). ZMM15 expression increased in leaves # 6-7 of NTG plants and 2 leaf stages later in the UBI pro :ZmCCT10 transgenic plants (Fig 5O). In the BSV pro :ZmCCT10 transgenic plants, ZMM15 expression was first detected in leaf #22 and increased in leaves # 24-31 reaching a higher level than in NTG or UBI pro :ZmCCT10 plants (Fig 5O). This finding suggests that both ZMM4 and ZMM15 are regulated by ZmCCT10 but ZMM15 may be activated independently of the photoperiod pathway.

Expression of meristem identity genes in the shoot apices affected by ZmCCT10 overexpression
To gain insight into the gene network(s) regulated by ZmCCT10 in the shoot apex, genes were selected with known and predicted functions in the floral transition and inflorescence development (S5 Table). For qRT-PCR analysis, the apices were grouped into five categories according to their developmental stage by morphology, including vegetative SAM, reproductive inflorescence meristem (IM), branch meristem (BM) initiation, developing tassel (DT), and immature tassel (IT) (Fig 4). Due to the limited number of BSV pro :ZmCCT10 plants available for dissection, fewer apices from this genotype were sampled and those at the BM initiation stage were missed. Overall, we surveyed 30 genes with various meristem functions (S5 Table). Expression of 20 genes were different in transgenic apices at more than one developmental stage (S6 Table). As a general trend, the magnitude of transcriptional changes was higher in the BSV pro : ZmCCT10 apices compared to UBI pro :ZmCCT10 apices (Fig 6 and S14 Fig).
The flowering activators DLF1 [27] and ZmSOC1 [67] known also as ZmMADS1 [68] showed increased expression in transgenic apices (Fig 6A and 6B) suggesting their function independent of ZmCCT10. Expression of the flowering repressor ZCN2, a maize ortholog of Arabidopsis TFL1, showed similar expression trends in NTG and transgenic plants (Fig 6C). Expression of ZFL1 that controls inflorescence phyllotaxy in maize [69,70] was slightly lower in transgenic apices compared to NTG plants (S14A Fig, S6 Table) suggesting a partial regulation of ZFL1 by the photoperiod pathway.
The inflorescence meristem identity AP1-like and SEP-like MADS genes were selected based on their key functions in inflorescence meristem specification in maize [34, 71] and rice [71][72][73]. As expected, the AP1-like MADS box genes, ZAP1, ZMM28, ZMM4 and ZMM15 were expressed at very low levels at vegetative stages in all genotypes. Their expression increased in the IM in NTG plants with ZAP1 having the highest expression level (Fig 6D). In BSV pro :ZmCCT10 apices, the expression levels of ZAP1 and ZMM28 were significantly reduced (Fig 6D and 6E) which indicates they function downstream of ZmCCT10. However, the expression of ZMM4 and ZMM15 was not significantly different in transgenic and NTG apices (Fig 6F and 6G).
The SEP-like MADS box genes ZMM24 and ZMM31 displayed contrasting expression patterns (Fig 6H and 6I). Expression of ZMM24 was repressed in both UBI pro :ZmCCT10 and Over-expression of the photoperiod regulator ZmCCT10 in maize BSV pro :ZmCCT10. But ZMM31 expression was moderately increased in UBI pro :ZmCCT10 apices at certain stages and was strongly increased in BSV pro :ZmCCT10 apices. ZMM31 expression was also high in plantlet apices (Fig 6I). This observation suggests that expression of the AP1-like and SEP-like genes is not tightly co-regulated by ZmCCT10 even though AP1 and SEP MADS proteins seems to function together in specifying inflorescence meristem identity [73].
To address the morphological changes in the tassel, such as the increased number of branches, the RAMOSA (RA) genes that control branching were assayed [74,75]. In NTG plants, RA1, RA2, RA3 and REL2 showed normal developmental expression patterns, consistent with original observations [75,76]. Expression was not detected in the SAM, was detected in the IM, was increased in the BM and DT, and declined in the IT stages in NTG (Figs 6J, S14B, S14C and S14D). Although minor expression differences were observed for RA1, RA3 and REL2 in both UBI pro :ZmCCT10 and BSV pro :ZmCCT10 transgenic apices (S14B, S14C and S14D Fig), RA2 expression was strongly increased in the transgenic plants, with BSV pro : ZmCCT10 apices showing a 250-fold increase (Fig 6J). The RA2 gene is required for transcriptional activation of RA1 [77], thus one would expect RA1 expression to also be up-regulated in transgenic plants, but this was not observed.
Given the critical function of auxin in plant development and shoot architecture [78,79] we assayed the auxin biosynthetic genes VT2 [80] and SPI1 [81]. Surprisingly expression of these genes showed opposing patterns (Fig 6K and 6L). Expression of VT2 was significantly increased in UBI pro :ZmCCT10 and BSV pro :ZmCCT10 apices (Fig 6K) but expression of SPI1 was either not different than NTG or was reduced in UBI pro :ZmCCT10 and BSV pro :ZmCCT10 apices (Fig 6L). Elevated expression of VT2 was also observed in the roots of dissected apical plantlets (S10 Fig), but SPI1 expression was not detected in roots. VT2 encodes a tryptophan aminotransferase [80] and SPI1 encodes a YUCCA-like flavin monooxygenase [81]. Both genes belong to different tryptophan dependent auxin biosynthesis pathways [81]. This finding suggests that those pathways may be regulated independently of each other and that regulation of the SPI1 pathway might be influenced by photoperiod. To further investigate effects on auxin pathway genes, expression of the auxin efflux transporter ZmPIN1a (Zea mays pinfor-med1a) [82] and the auxin signaling genes BIF1, BIF4 and BA1 [83] were assayed. Expression of ZmPIN1a was not affected in transgenic plants (S14L Fig). Expression of BIF1 was increased for both transgenic constructs, whereas expression of BIF4 and BA1 were not statistically different (Figs 6M, S14M and S14N).
To study the contribution of genes regulated by miR172, which has an antagonistic function to miR156, we selected the spikelet meristem identity genes AP2-like transcription factor BD1 (BRANCHED SILKLESS1 [87], TS6/IDS1 (TASSELSEED6/INDETERMINATE SPIKELET1, and SID1 SISTER OF INDETERMINATE SPIKELET1) [88,89]. The overall expression patterns of these genes were not statistically different between NTG and transgenic apices (S14E, S14F and S14G Fig), although BD1 transcript accumulation was slightly decreased in UBI pro : ZmCCT10 apices (S14F Fig). We took advatage of the prolonged vegetative phase of transgenic plants to compare expression of the miR156 and miR172 targeted genes during aging of the SAM. The SAM of the UBIpro :ZmCCT10 plants transtioned to an IM on average 20 days after sowing whereas the SAM of the BSV pro :ZmCCT10 plants took up to 60 days. Transcript accumulation of the miR156-targeted genes, ZmSBP18, TSH4/SBP6, TGA1, UB2/ZmSBP30, UB3/ZmSBP8 and ZmSBP14, showed an age dependent increase in expression (S15A, S15B, S15C, S15D, S15E and S15F Fig). This finding suggests that the expression of miR156 decreased over the extended vegetative period of growth resulting in accumulation of targeted SBP transcripts. Interestingly TGA1 mRNA accumulation peaked around the floral transition in the apices of UBI pro : ZmCCT10 and BSV pro :ZmCCT10 transgenic plants (S15C Fig) suggesting a putative involvement in the transition from vegetative to reproductive development.
Transcript accumulation for the miR172-targeted AP2-like genes showed an inverse pattern to the SBP genes, with expression being higher in the younger stage SAM and declining in older SAMs (S15G, S15H and S15I Fig). This pattern suggests that expression of miR172 is low in young apices and increased during extended vegetative growth. This finding is consistent with observations that miR156-miR172 regulation of the juvenile-to-adult transition in leaves is not altered in transgenic plants and these miRNA species are not regulated by ZmCCT10.

Duration of the vegetative phase shapes maize plant architecture
Genetic evidence suggests that ZmCCT10 is a regulator of the photoperiod pathway in maize repressing flowering under long days in sensitive lines [39,49]. ZmCCT10 is a nuclear localized protein with the potential to repress transcription, as was demonstrated for the homologous rice GHD7 protein [42]. For these reasons, over-expression of the ZmCCT10 protein in transgenic day-neutral (DN) maize was expected to act as a dominant inhibitor of the photoperiod pathway. This transgenic model allowed us to decipher which components of the photoperiod pathway are still responsive to ZmCCT10 modulation in day neutral maize, to identify genes in the autonomous pathway that connect to the photoperiod pathway, and to study the impact of modulating these pathways on maize growth and development.
Over-expression of ZmCCT10 with promoters of different strength, UBI pro :ZmCCT10 and BSV pro :ZmCCT10 , revealed that the phenotypic effects of ZmCCT10 are quantitatively related to transcript abundance. Both UBI pro :ZmCCT10 and BSV pro :ZmCCT10 transgenic plants showed multiple changes to plant architecture, but the BSV pro :ZmCCT10 transgenic plants developed more extreme phenotypes, in general. Quantitative transgene action may be explained by the diurnal turnover of the ZmCCT10 protein. This idea is supported by the observation that the homologous GHD7 protein in rice is degraded at night, but high protein levels accumulate each day correlated with high levels of GHD7 transcription [42]. If this is the case for the ZmCCT10 protein, a higher transcriptional level may maintain a higher protein level.
Manipulation of expression of the ZmCCT10 repressor with promoters of different strength revealed a strong relationship between duration of vegetative growth and plant architecture. As expected during the prolonged vegetative phase, plants produced more phytomers, recorded as the number of nodes with an attached leaf. This is clearly demonstrated by comparison of NTG Gaspé Flint plants with the shortest vegetative phase (1-2 days) producing a maximum of 10 leaves at maturity, compared to the BSV pro :ZmCCT10 transgenic plants with the longest vegetative stage (up to 60 days) producing up to 60 leaves. Since NTG Gaspé Flint plants transitioned to reproductive development at the V1 stage, when they are still in the juvenile phase, this suggests that phase change might not be required for initiation of reproductive development in maize.
Morphological parameters of the individual phytomer are also modified in transgenic plants. Internode length is shorter and stalk diameter is wider compared to NTG plants. Leaf shape along the shoot is also modified (Fig 2) with leaves above the ear continuing growth resulting in long, wide leaves on the upper shoot of transgenic plants (Fig 1). Transgenic BSV pro :ZmCCT10 plants also developed adventitious brace roots at every internode up to the top of the plants (Fig 1). Another atypical phenotype produced only by UBI pro :ZmCCT10 transgenic plants was an unusually long axillary branch (shank) subtending the ears (Fig 1D,  Table 2). Extreme elongation of ear shanks was reported for the jasmonic acid-deficient double opr7, opr8 mutants [90] hinting to a possible perturbation of jasmonic acid biosynthesis in UBI pro :ZmCCT10 transgenic plants. All these modifications dramatically changed the overall architecture of the BSV pro :ZmCCT10 transgenic plants indicating that prolonged vegetative growth could disturb multiple developmental processes.
Delaying the transition to reproductive development also induced changes in reproductive traits. We observed increased tassel branching in both UBI pro :ZmCCT10 and BSV pro : ZmCCT10 transgenic plants (S8 Fig) relative to controls. In BSV pro :ZmCCT10 transgenic plants, all ears aborted, suggesting axillary inflorescence development was suppressed by increased apical dominance. Alternatively, constant strong ZmCCT10 transcription may repress transition of the axillary branch meristem from vegetative to reproductive growth. It is important to emphasize that all maize plants with delayed floral transition displayed similar modifications to plant architecture. Production of more and bigger leaves, thicker stalks, brace roots at higher nodes, and highly branched tassels, was seen in the late flowering mutants id1 [33] and dlf1 [27]; in transgenic lines with less florigenic ZCN8 gene activity [91], in TFL1-like over-expression lines [54,91], and in tropically adapted maize lines grown under long days [37,92]. Moreover, transformation of the tassel into a phyllody phenotype observed in the BSV pro :ZmCCT10 transgenic lines (Fig 1), was similar to the "ball of shoots" phenotype observed in the null id1-m1 mutant [33] and in some tropical lines grown under long day conditions [92]. Regardless of the genetic or environmental cues, prolonged vegetative growth leads to similar phenotypic modifications to plant architecture suggesting a common physiological mechanism. The nature of this mechanism remains unclear.

Suppression of the photoperiod pathway does not affect transition from juvenile to adult phase but does delay the transition from vegetative to reproductive development
We found no effect of transgene expression on juvenile-to-adult phase change in transgenic events using such leaf morphological features as bulliform cells, epidermal hairs, cell wall invagination, and TBO staining (S7 Fig). In agreement with phenotypic observations, expression of the GL15 gene, a regulator of juvenile-to-adult phase leaf traits and a target of miR172 [15], is not affected in transgenic plants (Fig 5C). mRNA levels of miR172 targeted gene, ZmRap2.7, a repressor of flowering [65] also did not change in transgenic plants (Fig 5D). This finding suggests that the "adult" miR172-mediated aging pathway is not regulated by ZmCCT10 and is not connected to the photoperiod pathway.
Higher ZmCCT10 expression did result in a significant increase in leaf number and later time to flower which is evidence of repression of floral transition. We investigated the morphology of shoot apices and found that ZmCCT10 expression delayed the transition of the vegetative SAM to an IM (Fig 4). The floral transition occurred two stages later in the UBI pro : ZmCCT10 transgenic plants (V3-V4) compared to NTG controls (V1) and up to 20 stages later in the BSV pro :ZmCCT10 transgenic plants. Suppression of the photoperiod pathway delayed transition from vegetative to reproductive development but had no effect on phase change.

Disruption of meristem identity gene expression may cause the perturbed inflorescence morphology in ZmCCT10 transgenic plants
We surveyed the expression of 30 meristem identity genes and found that for most of them, their expression was altered in BSV pro :ZmCCT10 apices. The most common expression change was increased expression in the IM samples (Fig 6I, 6J, 6K, 6N and 6P). Among them, the two most striking were the 250-fold increase in RA2 expression (Fig 6J) and the 12-fold increase in expression of the auxin biosynthesis gene VT2 (Fig 6K), suggesting that auxin levels might be elevated in transgenic apices that contribute to alteration of meristem development. The mechanisms by which over accumulation of RA2 or VT2 contributes to the highly branched, vegetative phyllody phenotype is not clear. In fact, RA2 normally functions to suppress the initiation of long branches in the tassel, since loss of function ra2 mutants have highly branched tassels. Perhaps, increased branching in the ZmCCT10 transgenics is mediated by a different branching pathway, like the UB2/UB3 pathway, and RA2 expression is activated in response to suppress the extra branching.
Expression of ZCN14/18 is independent of ZmCCT10, positioning them outside of the photoperiod pathway. ZCN18 is likely not involved in flowering time control regardless of its potential florigenic activity in Arabidopsis (S13 Fig) because its expression is independent of the floral transition and the ZCN18 protein does not interact with DLF1 in a Y2H assay [30].
To the contrary, ZCN14 may function as part of the autonomous pathway that enables flowering even under non-inductive conditions in photoperiod sensitive backgrounds. We base this hypothesis on two lines of evidence. First, ZCN14 is expressed in leaves, consistent with a role as a floral activator, but only under LDs in tropical germplasm [30], indicating it is LD responsive. Second, the ZCN14 protein interacts with DLF1 in a Y2H assay [30], and is phylogenetically quite closely related to the rice, barley, and wheat florigenic proteins [28]. Therefore, we suggest a ZCN14-DLF1 FAC may form under LD conditons, allowing photoperiod sensitive genotypes to flower, even under non-inducitve conditions. It is important to emphasize that DLF1 expression is not suppressed in transgenic apices (Fig 6A), indicating the DLF1 protein is not rate limiting in forming a FAC.

A conceptual network model for regulation of flowering time in maize
The genetic network model for flowering time in maize [61] is less elaborated compared to rice [93] due to a limited number of flowering time mutants and a lack of flowering time QTLs with a large effect [94]. An updated network model was recently advanced based on results from transcriptome profiling of temperate maize and obligate SD teosinte [95]. Another approach is transgenic manipulation of flowering time which provides additional information to populate the maize network. Over-expression of ZmCCT10 helped define components of the photoperiod dependent and independent pathways and allowed us to further refine the flowering network (Fig 7) published previously [61].
In our model, the photoperiod pathway is represented by the conserved module ZmCCT10 (homolog of Ghd7)-ZmEhd1-ZCN8/7 (FT-like florigen). The circadian clock is the primary regulator of the photoperiod pathway but over expression of ZmCCT10 does not affect expression of the circadian genes GIGZ1A/B (homolog of GIGANTIA), indicating the lack of feedback from ZmCCT10 to the circadian clock. The conz1 gene, the maize homolog of CONSTANS, is down regulated in ZmCCT10 transgenic plants but its connection to other genes in this pathway is not clear, so it was not included in this model.
ZCN8 appears to be a conversion point of the photoperiod, autonomous and aging pathways [31,40,95,96] and its regulation is complex. In rice, Ehd1 is one of many upstream activators of the florigenic genes Hd3a/RFT1 [93,97]. In maize, ZmEhd1 may play a similar activation function of ZCN8 and other florigenic genes but this gene order must be further validated. Recently, the MADS-box transcription factor ZmMADS1 (alias ZmSOC1, in the model) was shown to be a direct, positive regulator of ZCN8 reinforcing the idea of multiple inputs for ZCN8 regulation.
In ZmCCT10 transgenic plants, expression of the florigenic genes ZCN8/7/12/15 are repressed, positioning them downstream of ZmCCT10. In contrast, expression of ZCN14 increases over time in ZmCCT10 transgenic plants, suggesting it function outside the photoperiod pathway, perhaps in the autonomous pathway. Within the autonomous pathway, ID1 is a major flowering time regulator and the pathway itself may be regulated by sugar signaling through a recently proposed "sugar clock" [95].
In day neutral, and tropical lines under inductive SDs, florigen is supplied by the ZCN8/7 genes, which moves to the SAM [29,30] where it interacts with DLF1 to form the FAC. In the shoot apical meristem, DLF1 is expressed independently of the photoperiod pathway, allowing for a constant supply of this bZIP transcription factor to the FAC [24]. This complex activates expression of the meristem identity AP1-like MADS box genes ZAP1, ZMM28, ZMM4 and ZMM15.
The aging pathway in leaves and apical meristems are not repressed in ZmCCT10 transgenic plants which places them outside of the photoperiod pathway. However, during prolonged vegetative growth, we observed an age-dependent activation of SBP genes in the SAM. This finding suggests that some SBP proteins might have the ability to bypass the need for FAC function and activate AP1-like MADS box genes independent of the florigen-dependent pathway for flowering. A similar situation was shown to function in Arabidopsis where expression of the SPL3, SPL13 and SPL9 (squamosa promoter binding protein-like) genes could trigger expression of AP1, FUL1 and LEAFY and, thus, bypass the need for FT-FD function to induce the floral transition [98,99]. If a similar bypass mechanism functions in maize, it has yet to be determined.

Conclusion
Maize was domesticated from the tropical progenitor-grass teosinte, which requires short-day photoperiods to flower. Over time, as cultivation moved to higher latitudes, maize's requirement for short-day photoperiod induction was reduced. As a result, maize has been adapted to grow in a wide range of photoperiods, with tropical maize flowering faster in short days, and temperate maize being relatively day neutral. Genetic studies revealed that ZmCCT10, which functions as a repressor of flowering, controls this short-day requirement. We investigated the . Genes tested in this study are framed by a blue outline. In the leaves, three pathways are depicted: aging, photoperiod and autonomous. The photoperiod pathway is represented by the module ZmCCT10 (rice Ghd7)-ZmEhd1-ZCN8/7/12/15 (FT-like genes) which is conserved in other short-day monocots, rice and sorghum. ZmCCT10 is likely a repressor of the flowering activator ZmEhd1 gene which is a homolog of rice Ehd1 and sorghum SbEhd1. ZmCCT10 is not expressed under inductive SDs releasing ZmEhd1 from repression and activating the FT-like ZCN8/7/12/15 genes. ZmCCT10 is expressed under noninductive LDs leading to repression of the FT-like ZCN8/7/12/15genes which occurs directly or indirectly via repression of ZmEhd1. However, the FT-like ZCN14 gene is not repressed by ZmCCT10 and may be placed in the autonomous pathway. In the aging pathway, ZmRap2.7 represses flowering via suppression of ZmEhd1 or ZCN8/7 genes. In the autonomous pathway, ID1 is required for activation of ZCN8/7 and possibly ZmEhd1 or other upstream regulators of ZCN8/7. Once transported to the shoot apical meristem (SAM), the florigenic ZCN8/7proteins interact with DLF1 forming the florigen activation complex (FAC, yellow oval). The florigen-antagonistic protein ZCN2 may compete with ZCN8/7 in the FAC leading to delayed flowering. Expression of DFL1 and ZCN2 is independent of the photoperiod pathway. Formation of the FAC in the SAM induces transcription of the AP1-like MADS box meristem identity genes ZAP1, ZMM4 and ZMM15 marking the onset of inflorescence specification. The ZFL1/2 genes are positioned downstream of the AP1-like genes. role this gene plays in flowering and other developmental processes by generating transgenic plants overexpressing ZmCCT10. Our phenotypic analysis of transgenic events containing either a constitutive promoter with high strength, or a constitutive promoter with very-high strength, showed that ZmCCT10 over-expression produced the expected late flowering phenotype, with the very-high level expressing events showing a dramatically prolonged vegetative period of growth and severe morphological developmental defects. Transcript expression analyses indicated that many genes that promote flowering are repressed and thus are downstream of ZmCCT10. We also showed that specific genes in modules affecting other developmental transitions are linked to photoperiod dependent or independent regulation. Our analysis allowed us to update the maize flowering time network model.