Journal of Genetics, Vol. 97, No. 2, June 2018, pp. 439–451 © Indian Academy of Scienceshttps://doi.org/10.1007/s12041-018-0929-5

RESEARCH ARTICLE

Evolutionary, interaction and expression analysis of floral meristem identitygenes in inflorescence induction of the second crop in two-crop-a-year grapeculture system

RONGRONGGUO1,2, BOWANG3, LINGLIN4, GUOCHENG4, SIHONGZHOU2, SHUYUXIE4, XIAOFANGSHI4, MUMING CAO4, YING ZHANG4 and XIANJIN BAI1∗

1Guangxi Academy of Agricultural Sciences, Nanning 530007, People’s Republic of China2Guangxi Crop Genetic Improvement and Biotechnology Laboratory, Nanning 530007, People’s Republic of China3Agricultural College, Guangxi University, Nanning 530004, People’s Republic of China4Grape and Wine Research Institute, Guangxi Academy of Agricultural Sciences, Nanning 530007, People’s Republic ofChina*For correspondence. E-mail: baixianjin@126.com.

Received 25 January 2017; revised 12 September 2017; accepted 21 September 2017; published online 19 May 2018

Abstract. The fruitfulness of grapevines (Vitis vinifera L.) is determined to a large extent by the differentiation of uncommittedmeristems, especially in the second-crop production of some varieties, where the intermediate of inflorescence and tendril accounts fora significant proportion in two-crop-a-year grape culture system. The differentiation of uncommitted lateral meristem was reportedto be regulated by a network, whose backbone was composed of several floral meristem identity genes. In the present study, thephylogenetics of grape floral meristem identity genes with their orthologues in other species, and their conserved domain andinteraction networks were analysed. In addition, the effects of chlormequat chloride and pinching treatments on the expressionprofiles of floral meristem identity genes and content of gibberellic acid (GAs) and zeatin riboside (ZR), as well as the ratio ofZR/GAs in buds that were used to produce the second crop, and the ratio of inflorescence induction of the second crop were studiedin ‘Summer Black’. The present results showed that floral meristem identity genes of grape and their orthologues in one or moreamongMalus domastic, Citrus sinensis, Theobroma cacao,Nicotiana tabacum, Solanum lycopersicum and Glycine hirsutum, probablyoriginated from a common ancestor. Interaction networks of six grape-floral meristem identity genes indicated that the inflorescenceinduction and floral development were regulated by one more complex network, and expression profiles of genes that involved in thisnetwork could be affected by each other. Expression profiles of eight floral meristem identity genes were affected by chlormequatchloride and pinching treatments, and higher expression levels of FT, TFL1A and TFL1B, as well as lower expression levels ofLFY from 3 days before full bloom to 11 days after full bloom were thought to play important roles in promoting the formation ofinflorescence primordial of the second crop, and higher expression levels of CAL A, SOC1 and TFL1A at 18 days after full bloom(DAF) could promote the development of inflorescence primordial. In addition, lower ratio of ZR/GAs at 3 days before full bloomand4 days after full bloom could promote the formation of uncommitted lateral meristems in chlormequat chloride and pinching-treatedplants, and higher ratio at 11 days after full bloom was the main reason for the formation of more inflorescences after chlormequatchloride treatment.

Keywords. grapevine; inflorescence induction; floral meristem identity; evolution; expression profiles;chlormequat chloride treatment.

Electronic supplementary material: The online version of this article (https://doi.org/10.1007/s12041-018-0929-5) contains supplemen-tary material, which is available to authorized users.

439

440 Rongrong Guo et al.

Introduction

The inflorescence induction involved the activity of agenetic network that acted in meristems to specify flo-ral identity, and the main output of this network wasthe initiation of a developmental patterning programmefor the generation of floral organs. The first characteristicof meristem identity genes was their capacity to integratethe environmental and endogenous signs that regulate theonset of flowering, and the second was the mutual regu-latory interactions established between them. Finally, thethird feature was the overlap between the meristem iden-tity and other developmental programmes that operatedsimultaneously to regulate different aspects of the con-struction of flowers (Blazquez et al. 2006).

In various angiosperms, some proteins were proved toform the backbone of the regulatory network controllingfloral meristem identity (FMI) (Blazquez et al. 2006; Ben-lloch et al. 2007). For example, LEAFY (LFY) encodeda transcription factor that was so far found only in theplant kingdom (Maizel et al. 2005), and served as a flowermeristem identity regulator activating the transcription ofother FMI genes (Parcy et al. 1998). APETALA1 (AP1),CAULIFLOWER (CAL), FRUITFULL (FUL) and SUP-PESSOR OF OVEREXPRESSION OF CONSTANS 1(SOC1) are members of MADS-box genes family in Ara-bidopsis thaliana.AP1 andCAL are expressed throughoutyoung floral meristems and in flower development, theirexpression became restricted to the first and secondwhorlsof the flower, and to floral pedicels. FUL was a close rel-ative of AP1 and CAL, but their expression patterns inthe inflorescence were almost complementary, and it wasable to partially replace the function of AP1 and CAL(Blazquez et al. 2006). SOC1 (AGL20) integrated signalsfrom the photoperiod, vernalization and autonomous flo-ral induction pathways, and the level of SOC1 activitywas critical to the control of flowering time (Lee et al.2000). FLOWERING LOCUS T (FT) and TERMINALFLOWER 1 (TFL1) are members of PEBPs gene fam-ily in A. thaliana (Kobayashi et al. 1999). FT interactedwith FD, a bZIP transcription factor that promoted flow-ering in A. thaliana through direct activation of AP1 inthe shoot apical meristem (SAM) (Abe et al. 2005; Wiggeet al. 2005). TFL1 controlled the length of the vegetativephase and delayed the flowering transition, and main-tained the identityof inflorescencemeristems, thus, playinga role in controlling the inflorescence architecture (Shan-non and Meeks-Wagner 1993; Bradley et al. 1997). InA. thaliana, genetic and molecular analyses have demon-strated the interaction between TFL1 and LFY in theestablishment of inflorescence architecture (Shannon andMeeks-Wagner 1991; Bradley et al. 1997).

Grapevine (Vitis vinifera L.) is a woody perennial vinewith a pattern of organ formation and development quitedistinct to those previously described for annual herba-ceous plants (Mullins et al. 1992). In adult grapevine

plants, the SAMproduced a characteristic sequence of leafprimordia and uncommitted meristems (UCM) (Tuckerand Hoefert 1968; Pratt 1974; Gerrath and Posluszny1988). The UCM gave rise to tendrils for a long period oftime before the plant initiates flowering, and upon flower-ing induction, the first twoor threeUCMs in the latent budon the primordial shoot had the potential to differentiateinto inflorescences, while the following UCM differenti-ated into tendrils (Pratt 1970; SrinivasanandMullins 1981;Posluszny andGerrath 1986; Gerrath and Posluszny 1988;Morrison 1991). When it differentiated, the UCM firstdivided to form two arms, if a tendril primordium wasformed, there was no further branching. If an inflores-cence primordium was formed, the UCM proliferated togivemultiple branch primordials, with each branch being aprotuberance of undifferentiatedmeristematic tissue (Car-mona et al. 2002). As both structures were formed fromthe UCM, tendrils and inflorescences were proposed to behomologous organs (Boss et al. 2003; Crane et al. 2012).

Grape bud fruitfulness, expressed as the number of clus-ters per bud, was thought to be influenced by geneticbackground, environmental conditions and horticulturalpractices (Crane et al. 2012). Based on the above, itappeared that control of UCM fate may be to a largeextent responsible for the degree of fruitfulness (Craneet al. 2012). It was reported that cytokinin could promotethedevelopmentof inflorescences rather than tendrils fromthe newly formed UCM (Srinivasan and Mullins 1978;Srinivasan andMullins 1980b). Gibberellic acid (GA) pro-motedUCM initiation, but inhibitedUCMdifferentiationinto inflorescences, favouring tendril development (Srini-vasan and Mullins 1980a). Unlike its orthologues in A.thaliana that played roles in delaying the flowering transi-tion,VvTFL1Awas involved in the control of inflorescencedevelopment, and its activity coulddelay acquisitionof flo-ral meristem identity, thus, extended the branching periodfor the UCM, resulting in the differentiation of inflores-cence primordia (Crane et al. 2012). The orthologues ofA. thaliana AP1, FUL-like, FT andLFY in grape,VvAP1,VvFUL-L, VvFT and VvLFY were reported to be associ-ated with tendril development (Calonje et al. 2004; Craneet al. 2012). Except for their common roles in tendrildevelopment, they had respective roles in flower or fruitdevelopment. For example, VvAP1 has a role in floweringtransition and flower development, and VvFUL-L has arole in floral transition and carpel and fruit development(Calonje et al.2004).OverexpressionofVvFT in transgenicA. thaliana could generate early floweringphenotypes, sug-gesting its role in flowering promotion (Carmona et al.2007). VvLFY not only participated in flower meristemspecification but also in themaintenance of indeterminacybefore the differentiation of derivatives of the apical meris-tem: flowers, leaves, or tendrils (Carmona et al. 2002). TheA. thaliana SOC1 orthologue,VvSOC1, whose expressionlevel was high in the axillary buds at the timewhen inflores-cence primordialswere initiated in them, suggesting its role

Function analysis of floral meristem identity genes in inflorescence induction of the second crop of grape 441

in the early stages of inflorescence development (Sreekan-tan and Thomas 2006).Although, the functions of cytokinin and GA, as well

as the function of floral integrators in UCM developmentwere studied in grape, the mechanism that controls inflo-rescence induction and the strategy by which the abovefactors manipulate this mechanism are still not clear.In temperate regions, grapevine requires two consecu-

tive growing seasons to flower (Calonje et al. 2004). Inrecent years, grape production was expanded to warmersubtropical areas, where the climatic conditions allow theproduction of two crops per year. In the two-crop-a-yeargrape culture system, the growing season for the first, orsummer, crop was usually from February to June, and thatof the second, or winter, crop was from August to Decem-ber (Xu et al. 2011). The fruitfulness of the second crop ofsome cultivars, like ‘SummerBlack’, ismuchworse in somesubtropical areas, and the chlormequat chloride (CCC),a plant growth inhibitor, and pinching were usually usedto promote the inflorescence induction, thus, enhancingthe fruitfulness (Naoki et al. 1978). Therefore, the worsefruitfulness of the second crop of some cultivars in the two-crop-a-year culture production provides a good materialto study the mechanism that controls inflorescence induc-tion.In the present study, the expression levels of eight FMI

genes in the developing buds that were used to produce thesecond crop of grape post CCC and pinching treatmentswere analysed, and the contents ofGAsand zeatin riboside(ZR) were measured in the same buds. Finally, the resultof inflorescence induction was calculated after the secondcrop starts development.

Materials and methods

Plant material and treatments

The grape variety ‘Summer Black’ (V. labrusca × V.vinifera) used in this study was grown in the rain sheltergreenhouse atGuangxi Academy ofAgricultural Sciences,Nanning, Guangxi, China (22◦82′N, 108◦37′E). Three-year-old ‘Summer Black’ grape plants were used for CCCand pinching treatments.Pinching treatment was conducted when the sixth leaf

(from the base of the cane) expanded fully, the shoot sec-tion above the sixth leaf was excised artificially. All thelateral shoots were erased except the top one, whose shootapical was excised until there were eight leaves on it. CCCtreatment was conducted by spraying leaves with 316µMCCC (Solarbio, Beijing, China, the purity is 98%), at 10and 1 d before full bloom. Control plants’ leaves weresprayed with purified water. The shoot apical of controland CCC treatment was excised when there were 14 leaveson the cane, and all lateral shoots were erased. Thirty budsat the 5th and 6th nodes from the base that were used to

Table 1. Dates of sampling and cor-responding growth stages.

Date DAF Growth stage

March 30th −9 H17April 1st −7 H18April 5th −3 19April 12th 4 J27April 19th 11 J29April 26th 18 K31May 10th 32 M35May 24th 46 M37

DAF, days after full bloom.

produce the second crop usually were sampled at −9, −7,−3, 4, 11, 18, 32 and 46 d after full bloom (DAF, corre-sponding to 10, 12, 16, 23, 30, 37, 51 and65dafter pinchingtreatment, and 1, 3, 7, 14, 21, 28, 42 and 56 d after the firstCCC treatment), and immediately frozen in liquid nitro-gen and stored at –80◦C until use. According to the studyby Coombe (1995), the relationship between dates of sam-pling and grape growth stages are shown in table 1. All thesamples were used to analyse the expression levels of eightfloral meristem identity genes, and measure the contentsof GAs and ZR. All the experiments were repeated thrice.

Sequence and phylogenetic analysis

Orthologous genes of the eight grape FMI genes in othercropswere searched inNCBI through sequence alignment.Multiple alignments of protein sequences fromV. vinifera,Theobroma cacao, Nicotiana tabacum, Solanum lycoper-sicum, A. thaliana, Glycine max, Malus domastic, Gossyp-ium hirsutum, Citrus sinensis were performed using theClustalW program. Phylogenetic trees were constructedwith the MEGA 5.0 software using the neighbour-joining(NJ) method, and the bootstrap test was replicated 1000times (Guo et al. 2013).

Prediction of Pfam domain and interaction network

The Pfam domain and interaction network of eight grapeFMI genes were predicted using SMART (http://smart.embl-heidelberg.de/smart/set_mode.cgi?NORMAL=1)(Letunic et al. 2012).

Expression analysis of grape FMI genes

Total RNA was extracted using the Spectrum plant totalRNA kit, according to the manufacturer’s instructions(Sigma, USA). A total of 500 ng of DNAse-treated totalRNA was used for first strand cDNA synthesis, using amixture of poly(dT) and random hexamer primers along

442 Rongrong Guo et al.

with PrimeScript RTase (TaKaRa Biotechnology, Dalian,China). The cDNA samples were then diluted six-fold andstored at –40◦C for the quantitative real-time PCR (qRT-PCR) analysis.The grape Actin1 gene (GenBank accession number

AY680701) was used as internal control. Gene-specificprimers were designed for the eight grape FMI genes(table 1 in electronic supplementary material at http://www.ias.ac.in/jgenet/). QRT-PCR analysis was conductedwith a Roche LightCycler480 real-time PCR instrument(Roche, Basel, Switzerland). Each reactionwas carried outin triplicate with a reaction volume of 20µL containing1 µL each primer (1.0µM), 2µL cDNA, 10µL SYBRGreen I Master (Rocheta et al. 2014) and 6µL sterile dis-tilled water. The PCR parameters were 95◦C for 5 min,followed by 40 cycles of 95◦C for 10 s, 60◦C for 20 s,72◦C for 20 s. Melt-curve analyses were performed usinga program with 95◦C for 5 s, 65◦C for 1 min, and then aconstant increase to 97◦C. Relative expression levels wereanalysed using the LightCycler480 software and the nor-malized expression method.

Measurement of endogenous hormones contents

Endogenous hormones, including GAs and ZR, wereextracted and purified as described by Bollmark et al.(1988) and He (1993). The contents of GAs and ZR weremeasured using themethod of enzyme-linked immunosor-bent assay (ELISA) as described previously (Yang et al.2001).

Measurement of inflorescence induction ratio of the secondcropping

The shoot was pruned at mid-August, and the cane abovethe 6th node from the base was excised. Buds at the 5thand6thnodes fromthebaseof canewerebrokendormancywith 2.5% (w/w) cyanamide, until there was no bleeding atthepruningwound.Numberof inflorescences, tendrils andthe intermediate of inflorescence and tendril as describedby Crane et al. (2012) was counted 20 d postcyanamidetreatment, and their per cent was calculated.

Statistical analysis

All experiments were repeated independently thrice foreachof the threebiological replicates.Results arepresentedas means and standard errors, and were calculated usingExcel (Microsoft,USA). Paired t-test was performedusingthe SPSS Statistics 17.0 software (IBM, Beijing, China) toassess significant differences.

Results and discussion

Phylogenetic analysis of FMI genes from various plant species

Three phylogenetic trees were constructed with eight FMIgenes from various plant species (figures 1–3). To pro-vide a reference for the evolutionary history of plant FMIgenes, orthologues of AP1, CAL A, FUL, SOC1, LFY,FT/TFL1 proteins from nine species, including grape,Theobroma cacao, Nicotiana tabacum, Solanum lycoper-sicum, A. thaliana, Glycine max, Malus domastic, Gossyp-ium hirsutum and Citrus sinensis were analysed.It was found that AP1, CAL A, FUL and SOC1

were members of MADS-box transcription factor fam-ily through sequence alignment, and the former threebelong to AP1/FUL sub-group as described previously(Wang et al. 2015). As shown in figure 1, AP1, CALA and FUL from the analysed nine species clusteredtogether, indicating the three genes originated from thesame ancestor. Moreover, VvSOC1 was more closelyrelated to TcSOC1 and CsSOC1, while VvAP1 was moreclosely related to MdAP1 (figure 1). Thus, it could beconcluded that SOC1 in grape, T. cacao and C. sinen-sis, as well as AP1 in grape and apple shared a commonancestral gene, and their functions could be similar. Itwas reported that the citrus SOC1-like gene, CsSL1 (cor-responding to CsSOC1 in the current study) was ableto shorten the time taken to flower in the Arabidopsiswile-type ecotypes Columbia and C24 (Tan and Swain2007) and MdMADS5 (corresponding to MdAP1 in thecurrent study) was involved in flower-bud formation,regulation of floral meristems and floral organ develop-ment of the apple (Kotoda et al. 2002; Mimida et al.2013). Therefore, the grape SOC1 may have function inpromoting early-flowering, and AP1 has roles in flower-bud formation, floral meristems regulation and floralorgan development. The flower meristem identity reg-ulator, LFY in grape, N. tabacum and S. lycopersicumclustered together (figure 2), implying the proteins inthe above three species come from a common ancestor.NFL1 (corresponding toNtLFY in the present study) wasreported to play a critical role in the allocation of meris-tematic cells that differentiated lateral structures such asleaves and branches, thereby determining the architec-ture of the wild-type tobacco shoot, and it also controlsthe cell proliferation and cell identity, thus, regulated flo-ral meristem development (Ahearn et al. 2001), and thetomato FALSIFLORA (FA, corresponding to SlLFY inthe present study) could regulate both floral meristemidentity and flowering time in tomato in a similar wayto LFY of Arabidopsis (Molinero-Rosales et al. 1999). Itwas reported that VvLFY participated in flower meristemspecification, and had roles in the maintenance of inde-terminacy before the differentiation of derivatives of theapicalmeristem: flowers, leaves, or tendrils (Carmona et al.2002), thus, its function was similar to that of NFL1 andSlFA.

Function analysis of floral meristem identity genes in inflorescence induction of the second crop of grape 443

Figure 1. Phylogenetic analysis of AP1, CAL A, FUL and SOC1 proteins from grape and other plants. Different coloursrepresent proteins from different plant species. The accession numbers for each sequence are as follows: V. vinifera AP1(AAT07447.1), CAL A (XP_002263017.1), FUL (XP_010660493.1), SOC1 (XP_010661478.1); T. cacao TcAP1 (XP_007045796.2),TcCAL A-1 (XP_017973950.1), TcCAL A-2 (XP_017974314.1), TcSOC1 (XP_017969659); N. tabacum NtAP1c (OIS95986),NtAGL8 (NP_001312134.1), NtCAL A (XP_016466964), NtSOC1 (NP_001312958); S. lycopersicum SlAP1 (NP_001234665.1),SlCAL A (XP_010316206.1), SlFUL2 (NP_001294867), SlSOC1 (XP_004248574); A. thaliana AtAP1 (NP_177074.1), AtCALA (NP_564243.1), AtFUL (NP_568929.1), AtSOC1 (NP_182090.1); G. max GmCAL A-1 (XP_003547792), GmCAL A-2(XP_006581051.1), GmSOC1 (NP_001236377); M. domastic MdAP1 (ACD69426), MdCAL A-1 (XP_008374663), MdCAL A-2(XP_008393257), MdSOC1 (NP_001280855); G. hirsutum GhCAL A-1 (NP_001313746), GhCAL A-2 (XP_016744306), GhCALA-3 (NP_001313669.1), GhMADS23 (NP_001314083.1), GhSOC1 (XP_016735195); C. sinensis CsAP1 (NP_001275828), CsCALA-1 (XP_00647715), CsCAL A-2 (XP_006482432), CsSOC1 (NP_001275772).

Figure 2. Phylogenetic analysis of LFY proteins from grapeand other plants. The accession numbers for each sequenceare as follows: V. vinifera VvLFY (AAM46141.1), G. hir-sutum GhLFY (NP_001314573.1), M. domestic MdLFY(NP_001280945.1), C. sinensis CsLFY (AAR01229.1),G. max GmLFY (NP_001304626.1), N. tabacum NtLFY(XP_016449328.1), S. lycopersicum SlLFY (NP_001234388.1),A. thaliana AtLFY (NP_200993.1).

FT, TFL1A and TFL1B belonged to the FT/TFL1gene sub-family (Carmona et al. 2007). A phylogeneticanalysis was also conducted using five grape FT/TFL1proteins togetherwithorthologues fromother eight species(figure 3). The present result showed that VvFT andVvTFL1Aweremore closely related toMdFT1,TcHD3A,GhHD3A and TcCEN2, GhCEN2, respectively, andVvTFL1B to MdTFL1-1 and MdTFL1-2, VvTFL1C toNtCEN1 and SlCEN1, VvMFT to CsMFT, TcMFT andGhMFT (figure 3), indicating FT in grape and its ortho-logues in M. domastic, T. cacao and G. hirsutum, TFL1Ain grape and its orthologues in T. cacao and G. hirsu-tum, TFL1B in grape and its orthologues in M. domastic,TFL1C in grape and its orthologues in N. tabacum andS. lycopersicum, and MFT in grape, C. Sinensis, T. cacaoand G. hirsutum originated from a common ancestor, andtheir functionsmay be similar. It was reported thatMdFT1

444 Rongrong Guo et al.

Figure 3. Phylogenetic analysis of FT-TFL1 proteins from grape and other plants. The accession numbers for each sequenceare the following: V. vinifera VvMFT (DQ871594), VvFT (DQ504308), VvTFL1A (DQ871591), VvTFL1B (DQ871592),VvTFL1C (DQ871593); T. cacao TcCEN2 (XP_007015849.1), TcHD3A (XP_007028083.1), TcMFT (XP_017978764.1), TcSP(XP_007027716.1), TcCEN1(XP_017973069.1);N. tabacumNtMFT (XP_016507970), NtCEN1 (AF145259), NtCEN2 (AF145260),NtCEN4 (AF145261), NtCEN5 (AF145262), NtTFL1 (XP_01644627.1); S. lycopersicum SlSP2G (AY186734), SlSP3D(AY186735), SlSP6A (AY186737), SlSP5G (AY186736), SlFLT (XP_004250075), SlCEN1 (XP_004234370), SlSP9D (AY186738),SlSP (U84140); A. thaliana AtMFT (AF147721), AtTSF (AF152907), AtFT (AF152096), AtBFT (NM_125597), AtTFL1(U77674), AtATC (AB024714); G. max GmTFL (ACJ61501.1), GmHD3A (NP_001240185.1), GmMFT (NP_00123689.1),GmBFT (NP_001236597.1); M. domastic MdMFT (XP_008374830), MdFT1 (AB161112), MdTFL1-1 (AB052994), MdTFL1-2(AB162046);G. hirsutum GhCEN2 (XP_016747639.1), GhHD3A (XP_016692018.1), GhMFT (XP_016708673.1), GhSP(XP_016676616.1), GhCEN1 (XP_016743136.1); C. sinensis CsMFT (XP_006467912), CsFT (AB027456), CsTFL1 (AY344245).

could function to induce early flowering by activating thedownstream flowering-related genes (Kotoda et al. 2010;Li et al. 2010; Tränkner et al. 2010), which was similarto that of VvFT (Carmona et al. 2007). The other genein the FT/TFL1 gene sub-family of apple, MdTFL1, wasreported to be involved in the maintenance of the vegeta-tive phase in apple and suppression of the expression leadsto early flowering (Kotoda andWada 2005; Szankowski etal. 2009), whose function was similar to that of VvTFL1(Crane et al. 2012).

Conserved domain and interaction networks of FMI genes ingrapevine

The conserved domain and interaction networks of thefour FMI genes that belonged to MADS-box gene fam-ily are shown in figure 4. One MADS domain and one

K-box domain were included in the four proteins (fig-ure 4a). Moreover, the interaction network of FUL wasthe same as that of AP1, proteins that interacted withthem were members of mitogen-activated protein kinase(MAPK) family or homologue of MAPK (figure 4, b & d;table 2 in electronic supplementary material). Several pro-teins that contained in the interaction network of CALA were also included in the AP1 and FUL, such asVIT_19s0014g00220.t01, VIT_06s0004g03620.t01, VIT_17s0000g02570.t01, VIT_15s0046g02010.t01, VIT_12s0142g00130.t01 and VIT_18s0001g13010.t01 (figure 4,b–d), implying interaction networks of the three proteinswere crossedwith eachother. Besides, the abovementionedsix proteins, three members of MADS-box family wereincluded in the interaction network of CAL A, separately,they wereMADS4, VIT_00s0313g00070.t01, correspond-ing to MADS6 and MADS52, respectively, in previous

Function analysis of floral meristem identity genes in inflorescence induction of the second crop of grape 445

Figure 4. The distribution of MADS and K-box domains of grape (a) AP1, CAL A, SOC1 and FUL and the interaction networkof (b) AP1, (c) CAL A and (d) FUL.

study (Wang et al. 2015), as well as MADS8, correspond-ing to SOC1 in the present work (figure 4c; table 3 inelectronic supplementary material) indicating that therewas interaction among the members of MADS-box fam-ily. In addition, one transcriptional corepressor SEUSS(VIT_18s0001g15320.t01) was also included in the inter-action network of CAL A (figure 4c; table 3 in electronicsupplementary material).There was one FLO_LFY domain in the grape LFY

domain, and one helix structure was seen at the C-terminal of this domain (figure 5a). Three MADS-boxproteins, MADS8 (SOC1 in the present work), AP3(MADS46 in the previous study (Wang et al. 2015)) andTM6, as well as one receptor protein kinase CLAVATA1(VIT_18s0001g14610.t01), one axial regulator YABBY1(VIT_15s0048g00550.t01), one protein UNUSUALFLO-RAL ORGANS (VIT_01s0011g01220.t01), one home-box protein knotted-1-like 1 (VIT_14s0060g01180.t01),

one protein FRIGIDA (VIT_19s0014g00030.t01), onechromatin structure-remodelling complex protein SYD(VIT_05s0020g02020.t01) and one ATP-dependent heli-case BRM (VIT_15s0046g02290.t01) were included in theinteraction network of LFY (figure 5b; table 4 in electronicsupplementarymaterial). It should be noted thatMADS8,termed as SOC1 in the present work, was included in theinteraction networks both of CAL A and LFY, indicatingthe the three proteins interact with each other.One PBPdomainwas found in TFL1AandTFL1Bpro-

teins, and proteins in the interaction networks were thesame (figure 5, c–f), implying that the functions of the twoproteins may be similar. Except for the LFY protein andone bZIP protein, FD (VIT_18s0001g14890.t01), otherproteins included in the interaction networks of TFL1Aand TFL1B were ribosomal proteins (figure 5, d&f; table5 in electronic supplementary material). The participationof LFY in the interaction networks of TFL1A andTFL1B

446 Rongrong Guo et al.

Figure 5. The distribution of conserved domain and three-dimensional structure of grape (a) LFY, (c) TFL1A and (e) TFL1B and(b, d, f) the interaction network of the three proteins.

Function analysis of floral meristem identity genes in inflorescence induction of the second crop of grape 447

Figure 6. Expression profiles of eight inflorescence and floral meristem identity genes in the developing buds postCCC and pinchingtreatments. Data represent mean values± SD from three independent experiments. *Statistical significance (*0.01< P < 0.05, **P <0.01, student’s t-test) between treatments and control.

indicated their interactionnetworkswere crossedwith eachother. It was worth noting that the FD protein was pre-dicted to interact with TFL1A and TFL1B in grape, whileits orthologues in A. thaliana and G. hirsutum, AtFD andGhFD, were reported to interact with AtFT and GhFT,respectively (Abe et al. 2005;Wigge et al. 2005; Zhang et al.2016).

From the above mentioned interaction networks of thesix proteins, it could be concluded that the eightFMIgenesin the present work were interacted with one more com-plex network, and their expression profiles were affectedby each other.

Effect of CCC and pinching treatments on expression profiles ofgrape FMI genes

CCC and pinching treatments were widely applied topromote the bud differentiation, thus, enhancing the fruit-fulness (Naoki et al. 1978). To determine whether grapeFMI genes were responsive to CCC and pinching treat-ments, the 5th and 6th buds from the base of the cane of‘Summer Black’ following CCC and pinching treatmentswere sampled at several differentiation stages, and abso-lute expression levels of the eight grape FMI genes in these

buds were determined using quantitative real-time PCR.To make the result more intuitionistic, the relative expres-sion level of each gene at –9 DAF of control was set as 1,and their relative expression levels at all stages post CCCand pinching treatments, as well as other stages of con-trol were in the ratio of their absolute expression levels tothat at –9 DAF of control. Then, the relative expressionlevel of each gene postCCC and pinching treatments werecompared with that of the control at the same stage, andif it was lower than that of the control, this gene was neg-atively regulated by the treatment. As shown in figure 6,the relative expression level ofVvAP1was almost 20 timeshigher than that of control, and picked at 11 DAF, thendecreased, and increased post 18 DAF, and then becamestable at 32 DAF postpinching treatment, while its relativeexpression level was higher than that of control just at 32and 46 DAF postCCC treatments. The relative expressionlevel of VvCAL A was upregulated at –3, 4, 18, 32 and46 DAF postCCC and pinching treatments, and down-regulated at 11 DAF postCCC treatment (figure 6). TheVvFT was upregulated at –3, 4 and 11 DAF postCCC andpinching treatments, and it was upregulated at –7 and 32DAF postpinching treatments, and downregulated at –9and 18 DAF and upregulated at 46 DAF postCCC treat-ment. The othermember ofAP1/FUL sub-group,VvFUL,

448 Rongrong Guo et al.

whose relative expression level was higher than that of thecontrol at –9, –3, 4 and 46 DAF postCCC and pinch-ing treatments, and it was downregulated at 11 and 18DAF postCCC treatments (figure 6). Further, it shouldbe noted that the variation trends of VvAP1 and VvFULwere similar, even though therewas considerable differencebetween their relative expression levels, and this would bedue to the same interaction (figures 4, b&d; 6). The rel-ative expression level of VvLFY was lower than that ofthe control at 11, 18 and 32 DAF postCCC and pinchingtreatments, and besides the three stages, its relative expres-sion level at –7, –3 and 46 DAF was lower than that ofthe control postpinching treatment. SOC1 was downreg-ulated at –9, –7, 11 and 46 DAF, and upregulated at 18DAF postCCC and pinching treatments. Meanwhile, itwas downregulated at 32 DAF postCCC treatment. Therelative expression level ofVvTFL1Awas higher than thatof the control at 4, 11 and 18 DAF postCCC and pinchingtreatments, and lower at –9 DAF. And it was downreg-ulated and upregulated at –7 and –3 DAF, respectively,postpinching treatment. The paralogous of VvTFL1A,VvTFL1B were upregulated at –3, 11, 32 and 46 DAFpostCCC and pinching treatments, and upregulated at –9 DAF and 4 DAF postCCC and pinching treatments,respectively, and downregulated at –9 and –7 DAF post-pinching treatments.It was reported that VvTFL1A might be involved in

the control of inflorescence development, and its highexpression level may be beneficial for the differentiationof inflorescence primordia (Crane et al. 2012). Moreover,VvAP1, VvFUL-L, VvFT and VvLFY were associatedwith tendril development (Calonje et al. 2004; Craneet al. 2012). The expression level of VvSOC1 was highwhen inflorescence primordial was being initiated in buds(Sreekantan andThomas 2006). In the present study, whenthe expression profiles of grape FMI genes were com-bined with the result of inflorescence induction postCCCand pinching treatments, it was found that higher relativeexpression levels of FT, TFL1A and TFL1B, as well aslower relative expression levels of LFY from –3 DAF to11 DAF, and higher relative expression levels of CAL A,SOC1 and TFL1A at 18 DAFmay promote the formationof inflorescence of the second crop (figures 6 and 8). There-fore, it was speculated that the most important period forinflorescence induction in the 5th and 6th buds from thebase of the cane of ‘SummerBlack’ was from–3DAF to 18DAF. The functions ofVvTFL1A,VvLFY andVvSOC1 ininflorescence inductionwere consistentwithprevious stud-ies in grape (Carmona et al. 2002; Sreekantan andThomas2006; Crane et al. 2012), and similar to their orthologues inN. tabacum,M. domestica, S. lycopersicum and C. sinensisthat were in the phylogenetic analysis (Molinero-Rosaleset al. 1999; Ahearn et al. 2001; Kotoda and Wada 2005;Tan and Swain 2007; Szankowski et al. 2009). The differ-ences of the relative expression profiles of other FMI genesin the present study and previous reports were caused by

differences of varieties, climatic conditions and cultivationmeasures.It seemed possible that high sequence similarity was

not necessarily correlated with similar expression levels,because proteins with very similar sequences, presumablyperforming similar biochemical functions, were needed indifferent periods during growth and development (Guoet al. 2013). In the present study, AP1, CAL A, as well asFUL that belonged to the AP1/FUL subgroup ofMADS-box gene family; and FT, TFL1A as well as TFL1B thatbelonged to the FT/TFL1 subfamily of PEBPs family haddifferent expression levels postCCC and pinching treat-ments (figure 6). Therefore, it could be speculated thatthese genes carried out similar biological functions in flo-ral meristem identity through different expression levels,and their expression levels were influenced by proteins intheir interaction networks.

Effect of CCC and pinching treatments on content of plantendogenous hormones

It was proposed that the requirement for a specific balanceof hormones for flower formation was readily applica-ble to woody perennials (Zeevaart 1976). In grapevine,inflorescence formation was regulated at two levels: for-mation ofUCMand differentiation ofUCM; andGAandcytokinin were the principal regulators of flowering (Vas-concelos et al. 2009). It was reported that cytokinin couldpromote the development of inflorescence rather than ten-dril from the newly formed UCM, when it was applied toshoot tips (Srinivasan and Mullins 1978 Srinivasan andMullins 1980b) and GA promoted UCM initiation, butinhibited UCM differentiation into inflorescence (Srini-vasan andMullins 1980a). In the present study, the contentof GAs in developing buds was higher at –9 and 32 DAF,and lower at –7, 4 to 18DAF than that of control postCCCtreatment, and higher at 4 DAF and lower at –9 and 11–46DAF postpinching treatments (figure 7a). The ZR contentwas decreased at –7, –3, 4, 18 and 46DAF and increased at32 DAF postCCC treatment, and decreased at –9, –3 and11–46 DAF postpinching treatments (figure 7b). The ratioof ZR/GAs was higher at –7, 11 and 32 DAF postCCCtreatment than that of the control, and higher at –7 DAFpostpinching treatment, and it was lower at –3 and 4 DAFfollowing CCC and pinching treatments (figure 7c).

When the expression profiles of grape FMI genes werecombinedwith the ratio of ZR/GAs and the result of inflo-rescence induction postCCC and pinching treatments, itcould be concluded that higher expression levels ofFT,TFL1A and TFL1B, as well as lower expression levelsof LFY, and lower ratio of ZR/GAs from –3 DAF to 4DAF might promote the formation of UCM, and higherexpression levels of FT,TFL1A andTFL1B, lower expres-sion levels of LFY, and higher ratio of ZR/GAs at 11DAF were benefitted for the differentiation of UCM into

Function analysis of floral meristem identity genes in inflorescence induction of the second crop of grape 449

Figure 7. Endogenous content of (a) GAs and (b) ZRs in the developing buds postCCC and pinching treatments and the ratio of (c)ZR/GAs. Data represent mean values ± SD from three independent experiments. *Statistical significance (*0.01 < P < 0.05, **P <0.01, student’s t-test) between treatments and control.

inflorescence primordial of the second crop (figures 6–8).Moreover, higher expression levels of CAL A, SOC1 andTFL1A at 18 DAF would promote the development ofinflorescence primordial (figure 6).

Effect of CCC and pinching treatments on inflorescence inductionof the second crop

It was reported that CCC and pinching treatments couldpromote the inflorescence induction (Naoki et al. 1978).Both these methods were used to enhance the fruitfulnessof the second crop in warmer subtropical areas, wherethe two-crop-a-year grape culture system was appliedwidespread. In this survey, effect of CCC and pinchingtreatments on inflorescence induction of the second cropwas tested, and the results showed that the ratio of inflo-rescence was 36 and 10% higher than that of the controlpostCCC and pinching treatments, respectively, while theratio of intermediate of inflorescence and tendril was 24%and 5% lower than that of the control postCCC and pinch-ing treatments, respectively, and the ratio of tendril was12% lower than that of the control postCCC treatment(figure 8). It should be noted that the main factor thatlimit the fruitfulness of the second crop was the formationof higher proportion of intermediate of inflorescence andtendril, and further study should be focussed on this.In conclusion, in the present research, the phylogenetics

of eight grape FMI genes with their orthologues in otherspecies, their conserved domain and interaction networkswere analysed. In addition, the effects of CCC and pinch-ing treatments on the expression profiles of the eight FMIgenes and content of GAs and ZR, as well as the ratiobetween them, and the inflorescence induction of the sec-ond crop were studied in ‘Summer Black’. Results impliedthat AP1 in grape,M. domastic, SOC1 in grape,C. sinensisandT. cacao, FT in grape, and its orthologues inM.domas-tic, T. cacao and G. hirsutum, TFL1A in grape, and itsorthologues in T. cacao and G. hirsutum, TFL1B in grape

Figure 8. The result of inflorescence induction of the secondcrop postCCC and pinching treatments. Data represent meanvalues ± SD from three independent experiments. *Statisticalsignificance (*0.01 < P < 0.05, **P < 0.01, student’s t-test)between treatments and control.

and its orthologues in M. domastic, LFY in grape and itsorthologues in N. tabacum and S. lycopersicum were orig-inated from a common ancestor. Interaction networks ofsix grapeFMI genes indicated that the inflorescence induc-tion and floral development were regulated by one morecomplex network. Expression profiles of eight FMI geneswere affected by CCC and pinching treatments, higherexpression levels ofFT, TFL1A and TFL1B, and lowerexpression levels of LFY from –3 DAF to 11 DAF mightpromote the formation of inflorescence primordial of thesecond crop, and higher expression levels ofCALA,SOC1and TFL1A at 18 DAF could promote the development ofinflorescence primordial. Lower ratio ofZR/GAsat –3 and4DAF could promote the formation of UCM in CCC andpinching treated plants, and higher ratio at 11 DAF wasthe main reason for the formation of more inflorescencefollowing CCC treatment.

450 Rongrong Guo et al.

Acknowledgements

This work was supported by postdoctorate foundation ofGuangxi Academy of Agricultural Sciences (GNKB2014024)and the National Natural Science Foundation of China(31460501 and 31401821).

References

AbeM., Kobayashi Y., Yamanoto S., DaimonY., Yamaguchi A.,Ikeda Y. et al. 2005 FD, a bZIP protein mediating signals fromthe floral pathway integrator FT at the shoot apex. Science309, 1052–1056.

Ahearn K. P., Johnson H. A., Weigel D. and Wagner D. R.2001NFL1, aNicotiana tabacum LEAFY -Like gene, controlsmeristem initiation and floral structure. Plant Cell Physiol. 41,1130–1139.

BenllochR.,BerbelA., Serrano-MislataA.andMaduenoF.2007Floral initiation and inflorescnece architecture: a comparativeview. Ann. Bot. 100, 659–676.

Blazquez M. A., Ferrandiz C., Madueno F. and Parcy F. 2006How floral meristems are built. Plant Mol. Biol. 60, 855–870.

Bollmark M., Kubat B. and Eliasson L. 1988 Variations inendogenous cytokinin content during adventitious root for-mation in pea cuttings. J. Plant Physiol. 132, 262–265.

Boss P. K., Buckeridge E. J., Poole A. and Thomas M. R. 2003New insights into grapevine flowering. Funct. Plant Biol. 30,593–606.

Bradley D., Ratcliffe O. J., Vincent C., Carpenter R. and Coen E.1997 Inflorescence commitment and architecture in Arabidop-sis. Science 275, 80–83.

Calonje M., Cubas P., Martinez-Zapater J. M. and CarmonaM.J. 2004 Floral meristem identity genes are expressed duringtendril development in grapevine. Plant Physiol. 135, 1491–1501.

Carmona M. J., Calonje M. and Martinez-Zapater J. M. 2007The FT/TFL1 gene family in grapevine. Plant Mol. Biol. 63,637–650.

CarmonaM. J., Cubas P. andMartinez-Zapater J.M. 2002VFL,the grapevine FLORICAULA/LEAFY ortholog, is expressedinmeristematic regions independently of their fate.Plant Phys-iol. 130, 68–77.

Coombe B. G. 1995 Adoption of a system for identifyinggrapevine growth stages. Aust. J. Grape Wine R. 1, 100–110.

Crane O., Halaly T., Pang X. Q., Lavee S., Perl A., Vankova R.et al. 2012 Cytokinin-induced VvTFL1A expression may beinvolved in the control of grapevine fruitfulness. Planta 235,181–192.

Gerrath J.M. andPoslusznyU. 1988Morphological andanatom-ical development in the Vitacese. I. Vegetative development inVitis riparia. Can. J. Bot. 66, 209–224.

Guo R. R., Xu X. Z., Carole B., Li X. Q., Gao M., ZhengY.et al. 2013 Genome-wide identification, evolutionary andexpression analysis of the aspartic protease gene superfamilyin grape. BMC Genomics 14, 554.

He Z. 1993 Guidance to experiment on chemical control in cropplants. In: Guidance to experiment on chemical control in cropplants (ed. Z. P.He), pp. 60–68. BeijingAgriculturalUniversityPublishers, Beijing.

Kobayashi Y., KayaH.,GotoK., IwabuchiM. andAraki T. 1999A pair of related genes with antagonistic roles in mediatingflowering signals. Science 286, 1960–1962.

Kotoda N. and Wada M. 2005 MdTFL1, a TFL1-like gene ofapple, retards the transition from the vegetative to reproductivephase in transgenic Arabidopsis. Plant Sci. 168, 95–104.

Kotoda N., Hayashi H., Suzuki M. Igarashi M., Hatsuyama Y.,Kidou S. et al. 2010 Molecular characterization of FLOW-ERING LOCUS T -like gene of apple (Malus × domesticaBorkh.). Plant Cell Physiol. 51, 561–575.

Kotoda N., Wada M., Kusaba S., Kano-Murakami Y., MasudaT. and Soejima J. 2002 Overexpression of MdMADS5, anAPETALA1-like gene of apple, causes early flowering in trans-genic Arabidopsis. Plant Sci. 162, 679–687.

Lee H., Suh S. S., Park E., Cho E., Ahn J. H., Kim S. G. etal. 2000 The AGAMPOUS-LIKE 20 MADS domain proteinintegrates floral inductive pathways inArabidopsis.Genes Dev.14, 2366–2376.

Letunic I., Doerks T. and Bork P. 2012 SMART7: recent updatesto the protein domain annotation resource.Nucleic Acids Res.40, D302–D305.

LiW.M., TaoY.,YaoY.X.,HaoY. J. andYouC.X. 2010Ectopicover-expression of two apple Flowering Locus T homologues,MdFT1 and MdFT2, reduces juvenile phase in Arabidopsis.Biol. Plantatum 54, 639–646.

Maizel A., Busch M. A., Tanahashi T., Perkovic J., Kato M.,Hasebe M. et al. 2005 The floral regulator LEAFY evolves bysubstitutions in the DNA binding domain. Science 308, 260–263.

Mimida N., Komori S., Suzuki A. andWadaM. 2013 Functionsof the apple TFL1/FT orthologs in phase transition. Sci. Hor-tic. 156, 106–112.

Molinero-Rosales N., Jamilena M., Zurita S., Gomez P., CapelJ. and Lozano R. 1999 FALSIFLORA, the tomato orthologueof FLORICAULA and LEAFY, controls flowering time andfloral meristem identity. Plant J. 20, 685–693.

Morrison J. C. 1991 Bud development in Vitis vinifera L. Bot.Gaz. 152, 304–315.

Mullins M. G., Bouquet A. and Williams L. E. 1992 Biology ofthe Grapevine. Cambridge University Press, Cambridge, UK.

Naoki U., Akira S. and Takashi T. 1978 Effects of CCC andpinching treatment on the inflorescence induction and devel-opment on lateral shoots in grapevine. J. Jpn. Soc. Hort. Sci.47, 151–157.

Parcy F., Nilsson O., Busch M. A., Lee I. and Weigel D. 1998 Agenetic framework for floral patterning. Nature 395, 561–566.

Posluszny U. and Gerrath J. M. 1986 The vegetative and floraldevelopment in the hybrid grape cultivar Ventura.Can. J. Bot.64, 1620–1631.

Pratt C. 1970 Reproductive anatomy in cultivated grapes- aReview. Am. J. Enol. Vitic. 22, 92–109.

Pratt C. 1974 Vegetative anatomy of cultivated grapes- a Review.Am. J. Enol. Vitic. 25, 131–150.

Rocheta M., Becker J. D., Coito J. L., Carvalho L. and AmâncioS. 2014 Heat and water stress induce unique transcriptionalsignatures of heat-shock proteins and transcription factors ingrapevine. Funct. Integr. Genomics 14, 135–148.

Shannon S. and Meeks-Wagner D. R. 1991 A mutation in theArabidopsis TFL1 gene affects inflorescence meristem devel-opment. Plant Cell 3, 877–892.

Shannon S. and Meeks-Wagner D. R. 1993 Genetic interactionsthat regulate inflorescence development in Arabidopsis.PlantCell 5, 639–655.

Sreekantan L. and Thomas M. R. 2006 VvFT and VvMADS8,the grapevine homologues of the floral integrators FT andSOC1, have unique expression patterns in grapevine and has-ten flowering in Arabidopsis. Funct. Plant Biol. 33, 1129–1139.

Srinivasan C. andMullinsM. G. 1981 Physiology of flowering inthe grapevine-A Review. Am. J. Enol. Vitic. 32, 47–63.

Srinivasan C. and Mullins M. G. 1978 Control of floweringin the grapevine (Vitis vinifera L.). Plant Physiol. 61, 127–130.

Function analysis of floral meristem identity genes in inflorescence induction of the second crop of grape 451

Srinivasan C. and Mullins M. G. 1980a Effects of temperatureand growth regulators on formation of anlagen, tendrils andinflorescences in Vitis vinifera L. Ann. Bot. 45, 439–446.

SrinivasanC. andMullinsM.G. 1980bFlowering inVitis: effectsof genotype on cytokinin-induced conversion of tendrils intoinflorescences. Vitis 19, 293–300.

Szankowshi I., Waidmann S., Omar E. D. S., FlachowskyH., Hättasch C. and Hanke M. V. 2009 RNAi-silencing ofMdTFL1 induces early flowering in apple. Acta Hortic. 839,633–636.

Tan F. C. and Swain S. M. 2007 Functional characterization ofAP3, SOC1 and WUS homologues from citrus (Citrus sinen-sis). Physiol. Plantarum 131, 481–495.

Tränkner C., Lehmann S., Hoenicka H., Hanke M.V., FladungM., Lenhardt D. et al. 2010 Over-expression of an FT -homologous gene of apple induces early flowering in annualand perennial plants. Planta 232, 1309–1324.

Tucker S. C. and Hoefert L. 1968 Ontogeny of the tendril inVitisvinifera. Am. J. Bot. 55, 1110–1119.

Vasconcelos M. C., Greven M., Winefield C. S., Trought M. C.T. and Raw V. 2009 The flowering process of Vitis vinifera: AReview. Am. J. Enol. Vitic. 60, 411–434.

Wang L., Yin X. J., Cheng C. X., Wang H., Guo R. R., Xu X. Z.et al. 2015 Evolutionary and expression analysis of a MADS-box gene superfamily involved in ovule development of seededand seedless grapevines. Mol. Genet. Genomics 290, 825–846.

Wigge P. A., Kim M. C., Jaeger K. E., Busch W., Schmid M.,Lohmann J. U. et al. 2005 Integration of spatial and tempo-ral information during floral induction inArabidopsis. Science309, 1056–1059.

XuC.M.,ZhangY.L.,ZhuL.,HuangY. andLuJ. 2011 Influenceof growing season on phenolic compounds and antioxidantproperties of grape berries from vines grown in subtropicalclimate. J. Agric. Food Chem. 59, 1078–1086.

Yang J. C., Zhang J. H.,Wang Z.Q., ZhuQ. S. andWangW. 2001Hormonal changes in thegrainsof rice subjected towater stressduring grain filling. Plant Physiol. 127, 315–323.

Zeevaart J. A.D. 1976. Physiology of flower formation.Ann. Rev.Plant Physiol. 27, 321–348.

Zhang X. H., Wang C. C., Pang C. Y., Wei H. L., Wang H. T.,SongM.Z. et al. 2016Characterizationand functional analysisof PEBP family genes in upland cotton (Gossypium hirsutumL.). PLoS ONE 11, e0161080.

Corresponding editor:Manoj Prasad