Mol Genet Genomics DOI 10.1007/s00438-014-0977-3

ORIGINAL PAPER

Identification and characterization of the CONSTANS‑like gene family in the short‑day plant Chrysanthemum lavandulifolium Jianxin Fu · Liwen Yang · Silan Dai 

Received: 10 May 2014 / Accepted: 30 October 2014 © Springer-Verlag Berlin Heidelberg 2014

Abstract The CONSTANS (CO) and CONSTANS-like (COL) genes play key roles in the photoperiodic flowering pathways, and studying their functions can elucidate the molecular mechanisms underlying flowering control in photoperiod-regulated plants. We identified eleven COL genes (ClCOL1–ClCOL11) in Chrysanthemum lavandulifolium and divided them into three groups that are conserved among the flowering plants based on phylogenetic analysis. Most of the ClCOL genes are primarily expressed in the leaf and shoot apices, except for ClCOL6–ClCOL9, which belong to Group II. The expression levels of ClCOL4–ClCOL5 and ClCOL7–ClCOL8 are up-regulated under inductive short-day (SD) conditions, whereas ClCOL6 is downregulated under inductive SD conditions. The ClCOL genes exhibit four different diurnal rhythm expressions (Type I–Type IV). The Type I genes (ClCOL4–ClCOL5) are highly transcribed under light. The Type II genes (ClCOL1–ClCOL2, ClCOL10) display increased expression in darkness and are rapidly suppressed under light. Transcripts of ClCOL6–ClCOL9 and ClCOL11, belonging to Type III, are abundant in the late light period or

Communicated by S. Hohmann. Electronic supplementary material  The online version of this article (doi:10.1007/s00438-014-0977-3) contains supplementary material, which is available to authorized users. J. Fu · L. Yang · S. Dai (*)  Beijing Key Laboratory of Ornamental Plants Germplasm Innovation and Molecular Breeding, National Engineering Research Center for Floriculture and College of Landscape Architecture, Beijing Forestry University, Beijing 100083, China e-mail: [email protected]

at the beginning of the dark period. ClCOL3 belongs to Type IV, with high expression in the early light period and dark period. The peak expression levels of ClCOL4– ClCOL6 are decreased and postponed in the non-inductive night break (NB) and under long-day (LD) conditions, indicating that those genes may play an essential role in the flowering regulation of C. lavandulifolium. The overexpression of ClCOL5 promotes the flowering of Arabidopsis grown under LD conditions, suggesting that ClCOL5 may function as a flowering enhancer in C. lavandulifolium. This study will be useful not only for the study of the C. lavandulifolium photoperiod-dependent flowering process but also for the genetic manipulation of flowering time-related genes to change the flowering time in the chrysanthemum. Keywords  CONSTANS-like · Diurnal expression · Chrysanthemum lavandulifolium · Flowering time · Photoperiod Abbreviations AP1  APETALA1 CCA1 CIRCADIAN CLOCK ASSOCIATED 1 CCT CO, CO-like and TOC1 CDFs CYCLING DOF FACTORs CO  CONSTANS COL  CONSTANS-like COP1 CONSTITUTIVE PHOTOMORPHOGENESIS 1 CRY CRYPTOCHROME CTAB Cetyltrimethylammonium bromide DBB Double B-box FKF1 FLAVIN-BINDING, KELCH REPEAT F-BOX 1 FT  FLOWERING LOCUS T GI GIGANTEA Hd1  Heading date 1

13



Hd3a  Heading date 3a HOS1 HIGH EXPRESSION OF OSMOTICALLY RESPONSIVE GENES 1 LD Long-day LFY  LEAFY LHY LATE ELONGATED HYPOCOTYL MTP Metal tolerance protein NB Night break PHYB PHYTOCHROME B qRT-PCR Quantitative real-time RT-PCR RACE Rapid amplification of cDNA ends SAND SAND family protein SD Short-day SOC1  SUPPRESSOR OF OVEREXPRESSION OF CONSTANS 1 SPA1 SUPPRESSOR OF PHYA 1 STH Salt Tolerance Homologs STO Salt Tolerance-like protein SUR2  SUPERROOT 2 Tm Melting temperatures TUB2  Β-tubulin WT Wild-type ZT Zeitgeber time

Introduction The transition from vegetative to reproductive growth is an important developmental switch in angiosperms. To modulate the timing of flower transition, plants integrate signals from the environment and from endogenous regulatory pathways. Day length is one of the important environmental signals that can regulate the floral transition (Garner and Allard 1920). The ability to recognize and measure day length is the basis of the response to changes in day length. The regulation of flowering by the photoperiod has been extensively studied in two model plants: Arabidopsis, a long-day (LD) plant, and rice, a short-day (SD) plant (Izawa et al. 2003; Albani and Coupland 2010; Brambilla and Fornara 2013; Song et al. 2013). Despite different responses to day length in these two plants, the specific molecular components involved in the photoperiod pathway are strikingly conserved, although some appear to differ in their mode of regulation. These studies indicate that the CONSTANS (CO) gene in Arabidopsis (Putterill et al. 1995) and Heading date 1 (Hd1, CO homolog) in rice (Yano et al. 2000) play an important role in regulating photoperiod sensitivity. In Arabidopsis, the CO protein activates the expression of the FLOWERING LOCUS T (FT) gene only under the LD condition (Valverde et al. 2004), whereas in rice, the Hd1 protein can activate the expression of Heading date 3a (Hd3a, FT homolog) under SD conditions and suppress its expression under LD conditions (Hayama et al. 2003).

13

Mol Genet Genomics

The proteins of FT and Hd3a are long-sought florigens, and their expression leads to the induction of floral meristem identity genes, including LEAFY (LFY) and APETALA1 (AP1), and consequent flowering. The CO protein is a zinc finger transcriptional regulator containing two B-Boxes and a CO, CO-like and TOC1 (CCT) domain (Robson et al. 2001). The transcriptional regulation of CO mRNA and posttranscriptional regulation of the CO protein by the circadian clock and light ensure that the activation of FT transcription occurs only in the LD afternoon (Kobayashi and Weigel 2007). The peak expression levels of CO mRNA occur in LD afternoons but after dusk under SD conditions (Putterill et al. 1995; Suarez-Lopez et al. 2001). Multiple morning-phase Dof transcription factors, which are homologs of CYCLING DOF FACTORs (CDFs), can bind to the CO promoter and are involved in the repression of CO transcription in the morning (Fornara et al. 2009). At the same time, the core clock components, CIRCADIAN CLOCK ASSOCIATED 1 (CCA1) and LATE ELONGATED HYPOCOTYL (LHY), repress the transcription of FLAVIN-BINDING, KELCH REPEAT, F-BOX 1 (FKF1) and GIGANTEA (GI), both of which negatively regulate CDFs (Sawa et al. 2007). FKF1 and GI interact in a blue light-dependent manner in LD afternoons, thus inducing the degradation of CDFs. With the action of FKF1 and GI, CO mRNA can accumulate only in LD afternoons (Sawa et al. 2007). The molecular mechanism of CO protein regulation is also important for inducing FT expression in LD afternoons. The PHYTOCHROME B (PHYB) signal and the E3 ubiquitin ligase HIGH EXPRESSION OF OSMOTICALLY RESPONSIVE GENES1 (HOS1) are involved in the degradation of CO protein during the morning (Valverde et al. 2004; Lazaro et al. 2012), although it is not known whether the PHYB signal regulates CO protein expression through HOS1. The light signals are perceived by PHY, and CRYPTOCHROME (CRY) binds to the SUPPRESSOR OF PHYA-105 1 (SPA1) and CONSTITUTIVE PHOTOMORPHOGENIC 1 (COP1) complex in a blue light-dependent manner to suppress the activity of COP1 and the SPA1 complex and thus stabilize the CO protein in LD afternoons (Zuo et al. 2011). In the dark, although CO mRNA accumulates at high levels, the CO protein is immediately degraded by the complex of COP1 and SPA1 (Laubinger et al. 2006; Jang et al. 2008; Liu et al. 2008; Song et al. 2012). Therefore, only the peak expression of CO mRNA, which occurs in LD afternoons after the degradation of the CDFs by the GI-FKF1 complex, leads to CO protein accumulation. The CO/CO-like (COL) gene family includes different family members among diverse species, such as seventeen members in Arabidopsis (Robson et al. 2001), sixteen in rice (Griffiths et al. 2003), thirteen in sugar beet (Chia et al. 2008), nine in barley (Griffiths et al. 2003),

Mol Genet Genomics

four in Brassica napus (Robert et al. 1998), three in wheat (Nemoto et al. 2003), and two in the woody species Picea abies (Holefors et al. 2009). This gene family has been divided into three groups based on variations within the B-box region. The first group (AtCO and AtCOL1– AtCOL5) has two B-box domains and a CCT domain. The second group (AtCOL9–AtCOL15) has one B-box domain, one diverged zinc finger domain in the second B-box and a CCT domain, and the third group (AtCOL6–AtCOL8 and AtCOL16) has one B-box gene and a CCT domain (Robson et al. 2001; Griffiths et al. 2003). The first group has been further subdivided based on the number and sequence characteristics of the conserved domains in their middle region (M1–4 domains) into subgroup Ia, comprising AtCO and AtCOL1–AtCOL2 in Arabidopsis and subgroup Ic, comprising AtCOL3–AtCOL5 in Arabidopsis (Griffiths et al. 2003). The CO/COL genes were broadly expressed throughout the life cycle of the plant from the early stages of embryonic development through the reproductive phase, with expression in the cotyledons, leaves, shoot apices and inflorescence as well as in developing seeds (Ledger et al. 2001; Almada et al. 2009; Takase et al. 2011; Zhang et al. 2011). Therefore, not all of these genes in the CO/ COL family regulate flowering transition or are photoperiod related. The functions of most gene family members in Arabidopsis have been well characterized. AtCOL1– AtCOL2 genes have little effect on the flowering time, but they can regulate the period of circadian rhythms in a ratedependent manner (Ledger et al. 2001). AtCOL3 is a positive regulator of photomorphogenesis and is involved in the suppression of flowering, shoot branching, root development and anthocyanin accumulation (Datta et al. 2006). AtCOL5 transcription is under circadian regulation, but it may not have a role in regulating flowering in wild-type (WT) plants or may act redundantly with other flowering regulators (Hassidim et al. 2009). AtCOL7 can also promote the mRNA expression of SUPERROOT 2 (SUR2), which encodes a suppressor of auxin biosynthesis, to regulate shoot branching and shade avoidance responses in Arabidopsis (Wang et al. 2013; Zhang et al. 2014). AtCOL8 can be expressed in the seeds, leaves, flowers, and siliques. Transgenic Arabidopsis plants do not display any altered circadian rhythm under constant light conditions, but do display a late-flowering phenotype under LD conditions (Takase et al. 2011). AtCOL9 is involved in the regulation of flowering time through reducing the expression of CO and FT (Cheng and Wang 2005). In other plants, CO or COL genes can also be involved in tuber formation in potatoes (González-Schain et al. 2012), seasonal growth cessation in aspen trees (Bohlenius et al. 2006), and fruit ripening and stress responses in bananas (Chen et al. 2012). These studies indicate that different CO or COL genes have diverse functions in plant growth and development.

Chrysanthemum (Chrysanthemum morifolium), one of the most important ornamental plants is an obligate SD plant. Its flowering time can be precisely controlled by changing the day length to meet the demand for marketable flowers throughout the year. The discovery of genes responsive to the photoperiod could elucidate the molecular mechanism underlying the transfer of day length signal into the flowering time signal. However, there have been few reports regarding the expression and characterization of the CO/COL genes in chrysanthemums. Previously, only one COL gene, CmCOL1, has been characterized under different light quality conditions (Higuchi et al. 2012). However, the allopolyploid and self-incompatibility nature of chrysanthemums make it difficult to investigate its floral mechanism at the molecular level. A wild diploid chrysanthemum, C. lavandulifolium (2n = 18), also a SD plant that is widely distributed in the northeast regions of China (Ding et al. 2012), can be used as an ideal model plant for molecular biological studies on chrysanthemum due to its simple genetic background (Zhang and Dai 2009). In our previous study, an Illumina/Solexa library of C. lavandulifolium was constructed, and thirty-eight putative ClCOL genes containing the conserved zinc finger domain were identified (Wang et al. 2014). The expression levels of these thirty-eight putative ClCOL genes between the seedling stage and the visible bud stage were analyzed by an upgraded version of the digital genes expression method (Wang et al. 2014). Due to limitations of the transcriptome sequencing technology, these thirty-eight putative ClCOL genes only have an average length of 397 bp. In addition, in that report, the criterion that defined genes containing the conserved zinc finger domain as ClCOL was not accurate, as CO or COL proteins have been identified, containing one or two B-box domains near the amino (N)-terminus and a CCT domain near the carboxyl terminus. Some proteins exist, containing the conserved zinc finger domain but are not CO or COL proteins, such as the Salt Tolerance Homologs (STH), Salt Tolerance-like protein (STO) or Double B-box (DBB) proteins (Khanna et al. 2009); however, a recent study reported that the zinc finger protein can also regulate flowering time and abiotic stress tolerance by modulating gibberellin biosynthesis in the chrysanthemum (Yang et al. 2014). In this study, we attempted to isolate the full length of these ClCOL genes by 3′ and 5′ rapid amplification of the cDNA ends (RACE) method based on the transcriptome library and isolated eleven ClCOL genes, renaming them as ClCOL1–ClCOL11. The expression patterns of these gene family members in different tissues or different photoperiods were analyzed. Additionally, the function of an important ClCOL gene, ClCOL5, was further characterized. This study will be useful not only for further study of the C. lavandulifolium photoperiod-dependent flowering process but also for the genetic manipulation of

13



Mol Genet Genomics

Table 1  Primers and amplicon characteristics used for the expression patterns of ClCOL genes in C. lavandulifolium Regression coefficient (R2)

Amplicon length (bp)

96.6

0.998

101

60

104.6

0.996

190

60

108.4

1.000

93

58.1

94.4

0.998

181

58.1

98.2

0.995

253

60

98.4

0.993

241

60

98.9

0.998

213

60

99.9

0.995

118

60

95.7

0.994

242

60

109.2

0.997

110

60

95.7

0.984

206

Primer name

Primer sequence (5′–3′)

Tm (°C)

ClCOL1-F

GGTTATGCCTATTTCAGGGTCT

66

ClCOL1-R

ACGCTGCTTCATCTTCTTCTTC

ClCOL2-F

ACAGTGGATTTTTATCAGGAACGGA

ClCOL2-R

TCCATAAACTCCAAATACTCATCCTC

ClCOL3-F

CTTGAAGTAGGAGTTGTCCCAGACC

ClCOL3-R

TGGATAAACCTCTGTTGATGATTGG

ClCOL4-F

AATCGCATAGCCTTAGTGTATCC

ClCOL4-R

TTTCTGTTCTTCCTTTTCTCCTT

ClCOL5-F

CCTGCTCAAACCGATGTCAAAAT

ClCOL5-R

CGTGTAGAGCGAATGGATGAGAT

ClCOL6-F

CGATGGACTCTATGATGACTTTACC

ClCOL6-R

ATAACAAGGGTTTGGTTCGGTTT

ClCOL7-F

CAAGAGCGGATGTAAGGAAGC

ClCOL7-R

AAACCCTGAGACAACAATACAAGTG

ClCOL8-F

AGGACATAACACTGAAGGAAACTGC

ClCOL8-R

GTTTTCACCAACTGTGTGATGTGATT

ClCOL9-F

TTGAAGCAGGAGATTCCGTTGA

ClCOL9-R

GAAGCATCCTCAACAAGGTCAGT

ClCOL10-F

GTATCCTTTAGTGCCCGTGTTTG

ClCOL10-R

ATGTTTCTTCCTGTTTATTTTCCACT

ClCOL11-F

TGGACCCTAATGATTGTTGGCT

ClCOL11-R

TTCGGCATTGAGTTTCCTTACC

flowering time-related genes to change the flower time in the chrysanthemum.

Experimental procedures Plant material, growth conditions and treatment Seeds of C. lavandulifolium were germinated in plugs filled with vermiculite media under LD conditions (16 h light/8 h dark, 16 hL/8 hD, 108.4 μmol/m2/s, supplied by 28 W fluorescent lamps). After the seedlings had six true leaves, they were transferred to 9 cm flowerpots with turf and vermiculite (V:V  = 1:1) media under LD conditions. The room was kept at 22 ± 1 °C with ~60 % relative humidity. According to our previous results, plants that had fourteen true leaves were sensitive to inductive SD treatment (12 h light/12 h dark, 12 hL/12 hD), and therefore they were transferred to SD conditions to induce flowering (Fu et al. 2014). For analysis of the tissue-specific expression patterns of ClCOL genes in C. lavandulifolium, leaves, shoot apices, petioles, stems and roots were sampled when the floral buds of the plants were visible to the naked eye (approximately 15 days after SD induction). For analysis of the expression patterns of ClCOL genes in inductive SD or non-inductive LD conditions, plants that

13

PCR efficiency (%)

Table 2  Characterization of ClCOL gene family members in C. lavandulifolium Gene name

Unigene ID

Length of amino acid (aa)

NCBI E-value

ClCOL1

89927, 85822, 89447, 92955

376

4e−113

ClCOL2

24094

387

1e−111

ClCOL3

12883

318

7e−76

ClCOL4

98844, 100525

347

4e−69

ClCOL5

43518, 88705

375

2e−87

ClCOL6

94889, 1012

347

1e−127

ClCOL7

98860, 89454

367

5e−132

ClCOL8

40091, 92497, 98025

362

9e−88

ClCOL9

31846

434

2e−68

ClCOL10

97729, 39563

425

6e−87

ClCOL11

105777, 89975

412

5e−85

were grown under LD conditions were either transferred to SD conditions or maintained under LD conditions. The third and fourth true leaves below the shoot apices were sampled from either the plants grown in LD conditions or from the plants grown in SD conditions for different days. To analyze the diurnal expression patterns of the ClCOL genes in different photoperiods, the third and fourth true

Mol Genet Genomics

Fig. 1  Phylogenetic tree of ClCOL proteins in C. lavandulifolium with COL proteins from other plant species. The phylogenetic tree was generated using the neighbor-joining algorithm. A non-COL protein, Salt tolerance protein (STO, NP_172094), from Arabidopsis was chosen as the outgroup. The accession numbers for each protein from other plant species are as follows: Arabidopsis AtCO (NP_197088.1), AtCOL1 (NP_197089), AtCOL2 (NP_186887), AtCOL3 (NP_180052), AtCOL4 (NP_197875), AtCOL5 (AAM45054), AtCOL6 (AAM10103), AtCOL7 (NP_177528), AtCOL8 (NP_175339), AtCOL9 (NP_187422), AtCOL10 (AED95643), AtCOL11 (NP_193260), AtCOL12 (NP_188826), AtCOL13 (NP_182310), AtCOL14 (NP_973589), AtCOL15 (NP_174126), AtCOL16 (NP_173915); Vitis vinifera VvCOL5 (XP_002277953), VvCOL9 (XP_002264506), VvCOL13 (XP_002268490), VvCOL16 (XP_00228 2578); Solanum lycopersicum SlCOL (NP_001234448), SlCOL4 (XP_004253053), SlCOL6 (XP_004235805), SlCOL10 (XP_004243599),

SlCOL13 (XP_004247458), SlCOL14 (XP_004239301), SlCOL16 (XP_004236986); Solanum tuberosum StCOL4 (XP_006342453), StCOL5 (XP_006343545), StCOL10 (XP_006357679), StCOL13 (XP_0063 59430), StCOL15 (XP_006363266), StCOL16 (XP_006341612); Litchi chinensis LcCOL (AGS32266), Medicago truncatula MtCOL (XP_0 03610187); Glycine max GmCOL5 (NP_001240941); Prunus mume PmCOL2 (XP_008230340), PmCOL4 (XP_008220621), PmCOL5 (XP_0 08223533), PmCOL9(XP_008229772), PmCOL13 (XP_008234997), PmCOL15 (XP_008222351); Brassica nigra BnCOL1 (AAN09808); Helianthus annuus HaCOL1 (ADO61000), HaCOL2 (ADO60998); Populus deltoides PdCOL1 (AAS00054); Malus domestica MdCOL2 (XP_0083 41828), MdCOL6 (XP_008360050), MdCOL15 (NP_001280792); Chrysanthemum morifolium CmCOL (AEN71141); and Chrysanthemum seticuspe CsCOL (BAM67031)

13



13

Mol Genet Genomics

Mol Genet Genomics

◂ Fig. 2  Expression patterns of ClCOL genes in different tissues by

qRT-PCR analysis. Leaves (R), shoot apices (Sa), petioles (P), stems (S) and roots (R) were harvested from plants that had visible floral buds for analysis of the relative expression levels of the ClCOL genes. All of the expression levels of ClCOL genes were normalized against ClSAND expression. The error bars show the standard error of the results obtained from three biological replicates

leaves below the shoot apices were sampled at 2-h intervals over 24 h under LD, SD and 14 h light/10 h dark (14 hL/10 hD) conditions after a 7 days entrainment. The NB treatment consisted of a 2-h white light exposure in the middle of the dark period of the SD condition. The third and fourth true leaves below the shoot apices were sampled at 2-h intervals over 24 h under the NB condition. Three biological replicates were sampled for each time point and photoperiod condition. Gene isolation and phylogenetic analysis Total RNA was extracted from mixed samples of leaves and shoot apices, using the modified cetyltrimethylammonium bromide (CTAB) method (Chang et al. 1993) and pretreated with RNase-free DNase I (Promega, Madison, WI, USA) at 37 °C for 30 min to eliminate any genomic DNA contamination. The first strand of cDNA was synthesized based on 1 μg of total RNA using the M-MLV reverse transcription system (Promega, Madison, WI, USA) according to the manufacturer’s protocol. The 3′ RACE method (Invitrogen, Carlsbad, CA) was used to amplify the 3′ ends of these unigenes. PCR amplification was performed using the Abridged Universal Amplification Primer (AUAP, 5′-GGC CAC GCG TCG ACT AGT AC-3′) and gene-specific primers (Supplementary Table 1). For the ClCOL genes that do not contain an ATG start site in the unigene sequences, 5′-specific primers (Supplementary Table 1) were utilized according to the 5′ RACE System for Rapid Amplification of cDNA Ends Kits (Invitrogen, USA). The full length of cDNA for each gene was generated through PCR, with 1 cycle at 94 °C for 5 min, followed by 32 cycles of 30 s at 94 °C, 30 s at 60 °C, 2 min at 72 °C, and a final elongation for 7 min at 72 °C. The details of the primers used in amplification are described in Supplementary Table 1. All of the amplified products were cloned into the pMD18-T vector (Takara, Japan) and then sequenced by Sunbiotech Co., Ltd. (Beijing, China). Multiple sequence alignments were performed using the default parameters for ClustalW (Thompson et al. 1994). The phylogenetic tree was constructed using Molecular Evolutionary Genetics Analysis version 4.0 (MEGA 4.0, http://www.megasoftware.net) and the neighbor-joining method with the following parameters: Poisson correction, complete deletion, and bootstrap (1,000 replicates).

Expression analysis by quantitative real‑time PCR The myriad of transcripts were investigated using quantitative real-time PCR (qRT-PCR). The extraction of total RNA and the synthesis of the cDNA were performed as described above. The cDNA was diluted to 5 % (v/v) of its original concentration. The oligo-nucleotide primers for each ClCOL gene (Table 1) were designed using Primer Premier 5 software with melting temperatures (Tm) of 58.1– 66.0 °C, primer lengths of 21–26 bp, and amplicon lengths of 93–253 bp. PCR reactions were performed using a Mini Opticon Real-time PCR System (Bio-Rad, USA) with SYBR Premix Ex Taq (TaKaRa, Japan). The reactions were prepared in a total volume of 20 μl containing 2 μl of template, 0.4 μl of each primer (10 μM), 10 μl of 2× SYBR Premix Ex Taq and 7.2 μl of ddH2O. The PCR program was conducted with an initial step of 30 s at 95 °C, followed by 40 cycles at 95 °C for 5 s, the optimal annealing temperature for 30 s, and 72 °C for 30 s. The program was completed with a melting curve analysis using a temperature ramp from 60 to 95 °C. Negative controls were included in every reaction, and each qRT-PCR reaction was performed with three technical replicates. The ClSAND (SAND family protein, GenBank accession number: KF752605) gene was used as the endogenous control when analyzing the expression patterns of the ClCOL genes in different tissues, whereas the ClMTP (Metal Tolerance Protein, GenBank accession number: KF752604) gene was used as the endogenous control when analyzing the expression patterns of the ClCOL genes in leaves under different photoperiod treatments (Fu et al. 2013). The data were analyzed using CFX manager software and normalized to the reference gene after correcting for variations in amplification efficiency, as recommended in the CFX manager package 2.6. The fold change in gene expression between the gene of interest and a calibrator was determined using the 2−∆Ct method. Vector construction and Arabidopsis transformation ClCOL5 full-length cDNA was obtained by PCR amplification using gene-specific primers: 5′-GCG GAT CCA TGA GGA TGG TTG AAA C-3′ (BamH I site is underlined) and 5′-ACG CGT CGA CTT AAA GCG CCC G-3′ (Sal I site is underlined). The cDNA was inserted into a modified binary pBI121 vector (Wang et al. 2009). To generate transgenic plants, the vector was transformed into Agrobacterium tumefaciens strain EHA105, and Arabidopsis (ecotype Columbia) plants were transformed using the floral dip method (Clough and Bent 1998). The transformed progeny was selected by culturing on MS medium agar plates containing 50 mg/L of kanamycin. Kanamycin-resistant seedlings were transferred to soil after 2 weeks and grown

13



13

Mol Genet Genomics

Mol Genet Genomics

◂ Fig. 3  Expression patterns of ClCOL genes in leaves after treat-

ment for different numbers of days under inductive or non-inductive conditions. Plants that were grown under long-day (LD, 16 h light/8 h dark) conditions were transferred to short-day (SD, 12 h light/12 h dark) conditions or maintained under LD conditions. Leaves were sampled at ZT4 (zeitgeber time, 4 h after light on) in plants grown under LD or SD conditions for a different number days. The abbreviations indicate plants before SD induction (S1), SD induction for 1 day (S2), SD induction for 7 days (S3), SD induction for 14 days (S4) and SD induction for 38 days (S5). The closed circles indicate plants under SD conditions, and the open circles indicate plants under LD conditions. All of the gene expression levels were normalized against ClMTP. The error bars reflect the standard error of the results obtained from three biological replicates

at 23 °C under LD conditions. Homozygous T2 transgenic generation plants were identified on the basis of screening their progeny on the kanamycin medium and were then used for further assays. Gene expression analysis and phenotypic observation of transgenic Arabidopsis Rosette leaves from kanamycin-resistant and WT Arabidopsis plants were collected for RT-PCR validation. PCR analysis was performed using a CaMV 35S promoter-specific forward primer 5′-CGC AAG ACC CTT CCT CTA TAT A-3′ and ClCOL5-specific reverse primer 5′-ACG CGT CGA CTT AAA GCG CCC G-3′. The PCR program was conducted with an initial step of 30 s at 94 °C for 5 min, followed by 35 cycles of 30 s at 94 °C, 30 s at 55 °C, 2 min at 72 °C, and a final elongation for 7 min at 72 °C. The expression levels of AtFT in WT and transgenic Arabidopsis were evaluated using gene-specific primers (AB027504, 5′-TGG TGG AGA AGA CCT CAG GAA CT-3′ and 5′-TCA TTG CCA AAG GTT GTT CCA G-3′) in a qRT-PCR reaction. The TUB2 (β-tubulin) gene of Arabidopsis (GenBank accession number: AY054693) was used as an internal control, and its specific primers were 5′-ATC CGT GAA GAG TAC CCA GAT-3′ and 5′-AAG AAC CAT GCA CTC ATC AGC-3′. The qRTPCR conditions were followed as above described. For analysis of the phenotype of the WT and T2 transgenic line, randomly selected plants (>20 per line) were transferred to the soil and grown at 23 °C under LD conditions.

Results Isolation and characterization of ClCOL gene family members in C. lavandulifolium We blasted the sequences obtained using the 3′ RACE method in the NCBI BlastP database and found that some of the unigenes that are defined as ClCOL genes are STO genes because the predicted amino acid sequences lack the CCT

domain at the C terminus such as Unigene 77001 (data not shown). A number of the unigenes are the same gene, such as Unigene 98025, Unigene 92497 and Unigene 40091, but they were defined as different ClCOL genes in our previous study (Wang et al. 2014). Similarly, Unigene 1012 and Unigene 94889 are the same genes, but they were defined as different ClCOL genes in our previous study (Wang et al. 2014). Thus, we renamed these eleven complete cDNAs derived from the ClCOL genes amplified using the 3′ and 5′ RACE method in this study as ClCOL1–ClCOL11 (Table 2). The average length of predicted amino acids of these eleven ClCOL genes is 377 amino acid residues, with the shortest being the ClCOL3 protein (318 amino acid residues) and the longest being the ClCOL9 protein (434 amino acid residues). Using the BLASTP program in the NCBI database, the eleven ClCOL proteins were searched against the Arabidopsis protein database. All of the comparative E-values are below 5e−24, indicating that ClCOL proteins and AtCOL proteins are highly similar (Table 2). Based on phylogenetic analysis, the eleven ClCOL genes that were isolated from C. lavandulifolium were clearly divided into three groups: ClCOL1–ClCOL2 belong to Group Ia, ClCOL3–ClCOL5 belong to Group Ic, ClCOL6–ClCOL9 belong to Group II, and ClCOL10–ClCOL11 belong to Group III (Fig. 1). Furthermore, ClCOL1 is closely related to HaCOL1, whereas ClCOL2 is closely related to CsCOL. These two COL proteins are from the Compositae plants (Fig. 1). Tissue‑specific expression patterns of ClCOL genes in C. lavandulifolium To investigate the tissue-specific expression of the ClCOL genes in C. lavandulifolium, qRT-PCR was performed in different tissues, including the leaves, shoot apices, petioles, stems and roots from C. lavandulifolium plants at the stage when the floral buds are just visible. ClCOL1– ClCOL5 and ClCOL10 are mainly expressed in the leaves and shoot apices, whereas they are expressed at lower levels in other tissues such as the roots, stems and petioles (Fig. 2). The highest expression levels of ClCOL6 are only found in the stems and petioles (Fig. 2). ClCOL8 is highly expressed in leaves, shoot apices and stems (Fig. 2). ClCOL7 and ClCOL9 are constitutively expressed in all of the tested tissues and do not display tissue specificity (Fig. 2). The highest expression of ClCOL11 is found solely in the leaves (Fig. 2). Expression patterns of ClCOL genes under inductive or non‑inductive conditions In our previous study, the results indicated that the plants that had fourteen true leaves were sensitive to inductive SD treatment (12 h light/12 h dark, 12 hL/12 hD) and

13



13

Mol Genet Genomics

Mol Genet Genomics Fig. 4  Diurnal rhythm expression patterns of ClCOL genes in leaves under different photoperiod treatments. Leaves were sampled at 2-h intervals over 24 h in 12 h light/12 h dark (12 hL/12 hD), 14 h light/10 h dark (14 hL/10 hD) and 16 h light/8 h dark (16 hL/8 hD) conditions after a 7 days entrainment. Closed circles, open circles and open triangles indicate plants in 12 hL/12 hD, 14 hL/10 hD and 16 hL/8 hD conditions, respectively. All of the gene expression levels were normalized against ClMTP expression. The error bars represent the standard error of the results obtained from three biological replicates

proceeded to flower (Fu et al. 2014). For analysis of the expression patterns of the ClCOL genes in non-inductive LD or inductive SD conditions, qRT-PCR reactions were performed using RNA samples extracted from leaves sampled from the plants grown under LD conditions or from the plants grown under SD conditions treated for a different number of days. As shown in Fig. 3, the family members of the ClCOL genes have different expression patterns after SD treatment for a varying number of days. The expression levels of ClCOL1–ClCOL3 and ClCOL9–ClCOL11 display no significant differences between inductive SD and noninductive LD conditions. The expression levels of ClCOL4– ClCOL5 are dramatically up-regulated after 1 day of SD induction and are maintained at a high expression level until 16 days of SD induction, and then, they decrease. The plants of C. lavandulifolium need approximately 15 days to form floral buds and 38 days to form flowers in SD conditions (Fu et al. 2014). These results indicate that the expression levels of ClCOL4–ClCOL5 decrease when plants have formed floral buds. The expression levels of ClCOL7– ClCOL8 increase gradually after SD induction and attain the highest level until 16 days of induction, but under LD conditions, their expression levels are maintained at relatively low levels (Fig. 3). The expression levels of ClCOL6 decrease gradually after SD induction, but under LD conditions, its expression level is maintained at a relatively high level (Fig. 3). Therefore, ClCOL4–ClCOL5 and ClCOL7– ClCOL8 might function as flower inducers and ClCOL6 might function as a flower repressor in the photoperiodic flowering response of C. lavandulifolium. Diurnal oscillation of ClCOL genes expression in different photoperiods CO/COL genes in Arabidopsis and COL orthologs from other plant species display diurnal or circadian patterns of expression (Holefors et al. 2009; Campoli et al. 2012; González-Schain et al. 2012; Kikuchi et al. 2012). Thus, we studied the diurnal oscillation of ClCOL genes in different photoperiods based on the detection of mRNA abundance in fully opened leaves using qRT-PCR. The diurnal oscillation of ClCOL genes exhibits four patterns of light regulation (Fig. 4). Genes with a Type I expression pattern,

including ClCOL4–ClCOL5, are abundant in the light phase and maintain a relatively low expression level in the dark phase (Fig. 4). Type II genes, including ClCOL1– ClCOL2 and ClCOL10, display increased expression in the dark phase but decrease expression rapidly in the light period (Fig. 4). Type III genes, including ClCOL6– ClCOL9 and ClCOL11, are abundant in the late phase of the light period or at the beginning of the dark period, but are barely detectable in the early phase of the light period (Fig. 4). ClCOL3, a Type IV gene, is expressed highly in the early phase of the light period and in the dark period (Fig.  4). The results of a previous study indicate that 2 h of white light exposure in the middle of the dark period in SD can inhibit the flowering of C. lavandulifolium (Fu et al. 2014). The diurnal oscillations of these eleven ClCOL genes were analyzed under NB conditions. The peak expression levels of ClCOL4–ClCOL6 are postponed and decreased under NB and the non-inductive LD conditions (Figs. 4, 5). Overexpression of ClCOL5 leads to the promotion of flowering in Arabidopsis According to our results, both ClCOL4 and ClCOL5 are mainly expressed in the leaves and shoot apices (Fig. 2) and are up-regulated under inductive SD conditions (Fig.  3). The diurnal oscillations of ClCOL4 and ClCOL5 are the same both in the inductive or non-inductive photoperiods (Fig. 4, 5). Thus, we selected one gene, ClCOL5, which belongs to Group Ic and has the closest relationship to the AtCOL5 gene in Arabidopsis (Fig. 1) to analyze its biological function in C. lavandulifolium. ClCOL5, under the control of the CaMV 35S promoter, was introduced into WT Arabidopsis. A total of ten independent lines of 35S::ClCOL5 plants in the T1 generation were obtained, and we randomly selected two independent lines for further analysis. The RT-PCR reactions with the CaMV 35S promoter-specific forward primer and ClCOL5-specific reverse primer were performed using samples from the transgenic plants to confirm the presence of the transferred gene (Fig.  6a). The time from sowing to bolting or flowering was greatly shortened in these two T2 transgenic generation lines compared to WT Arabidopsis (Fig. 6b; Table 3). The WT Arabidopsis required 33.55 days from sowing to bolting and 40.52 days from sowing to flowering, while transgenic plants only required 26.00–26.15 days from sowing to bolting and 33.09–33.52 days from sowing to flowering (Table 3). Additionally, the rosette leaf numbers at the bolting or flowering stage were greatly reduced in the transgenic plants compared to WT Arabidopsis (Table 3). Next, we analyzed the expression level of AtFT in the transgenic lines. The expression levels of AtFT in the transgenic lines were significantly up-regulated in the 35S::ClCOL5

13



13

Mol Genet Genomics

Mol Genet Genomics Fig. 5  Diurnal rhythm expression patterns of ClCOL genes in leaves under night break (NB) conditions. The NB treatment consisted of a 2-h white light exposure in the middle of the dark period during the short-day (SD) conditions. Closed circles, open circles and closed triangles indicate plants in 12 h light/12 h dark (12 hL/12 hD, SD), 16 h light/8 h dark (16 hL/8 hD, LD) and NB conditions, respectively. All of the gene expression levels were normalized against ClMTP expression. The error bars represent the standard error of the results obtained from three biological replicates

transgenic plants (Fig. 6c), consistent with the phenotypic results.

Discussion The chrysanthemum is not only a typical obligate SD ornamental plant with high economic importance but also an excellent model to elucidate SD inductive flowering mechanisms. Using C. lavandulifolium, a wild diploid chrysanthemum, as a model to study the molecular mechanism of flowering will lead to a greater understanding of the mechanism in the chrysanthemum. In this study, we identified eleven ClCOL genes based on the results of our previous transcriptome database (Wang et al. 2014). The tissue-specific expression patterns and diurnal rhythm under different photoperiods of these ClCOL genes were analyzed. Most of the ClCOL genes are primarily expressed in the leaves and shoot apices, except for ClCOL6–ClCOL9, which belong to Group II. There are four diurnal rhythm expression patterns of these ClCOL genes. The expression levels of ClCOL4– ClCOL5 and ClCOL7–ClCOL8 are up-regulated under inductive SD conditions, whereas ClCOL6 is down-regulated under inductive SD conditions. The peak expression levels of ClCOL4–ClCOL6 are decreased and postponed under the non-inductive NB conditions and LD conditions, indicating that they have a close relation with the flowering of C. lavandulifolium. Overexpression of ClCOL5 can promote flowering in LD grown Arabidopsis, therefore it may function as a potential inducer in C. lavandulifolium. Previous studies indicated that the COL genes in Group I might be involved in the regulation of flowering time (Martin et al. 2004; Zhang et al. 2011; Campoli et al. 2012). To further examine the relationship among these putative ClCOL homologs and COL genes from other plant species, the COL amino acid sequences were downloaded for phylogenetic analysis (Fig. 1). These eleven ClCOL proteins can be divided into three groups: ClCOL1–ClCOL2 belong to Group Ia, ClCOL3–ClCOL5 belong to Group Ic, ClCOL6–ClCOL9 belong to Group II, and ClCOL10– ClCOL11 belong to Group III (Fig. 1). In Arabidopsis, the FT gene is activated by the CO protein, which belongs to Group Ia (Valverde et al. 2004). At the same time, AtCOL5, which belongs to Group Ic, has been demonstrated to

induce flowering and may act redundantly with other flowering regulators (Hassidim et al. 2009). In our studies, the ClCOL1–ClCOL5 genes belonging to Group I are mainly expressed in the leaves and shoot apices (Fig. 2). The transcript levels of ClCOL1–ClCOL3 are similar in the LD and the inductive SD conditions, which indicate that ClCOL1–ClCOL3 may have no relation to the photoperiod flowering in C. lavandulifolium (Fig. 3), but further studies are warranted to demonstrate the post-transcript levels of these genes because they are the key factor to activate the downstream FT gene to promote the flowering in Arabidopsis (Song et al. 2012). Meanwhile, the transcript levels of ClCOL4–ClCOL5 are dramatically up-regulated in the inductive SD condition compared to the LD condition (Fig. 3). Thus, these genes may function as flower promoters during the floral transition in the photoperiodic responses of C. lavandulifolium. We also investigated the diurnal rhythm of the ClCOL genes in different photoperiods. The diurnal oscillations of ClCOL1 and ClCOL2 are similar under both SD and LD conditions, with peak expression at dawn, but the expression of these two genes decreases rapidly in the light period before their accumulation at the end of dark period (Fig.  3). Thus, ClCOL1 and ClCOL2 are strongly induced by darkness and inhibited by light. The diurnal rhythms of ClCOL1 and ClCOL2 are similar to those of AtCOL1 and AtCOL2 in LD plant Arabidopsis (Ledger et al. 2001), BvCOL1 in LD plant sugar beet (Chia et al. 2008), VvCO and VvCOL1 in LD plant grape (Almada et al. 2009), StCO in SD plant potato (González-Schain et al. 2012), CrCOL1 and CrCOL2 in SD plant goosefoot (Drabesova et al. 2014) and PtCO2 in SD plant poplar (Bohlenius et al. 2006), in which the abundance of mRNA is restricted to the dark period under both photoperiod conditions. However, the functions of these COL genes are more diversified. The AtCOL1 and AtCOL2 genes in Arabidopsis have little effect on the flowering time, but they can regulate the period of circadian rhythms in a rate-dependent manner (Ledger et al. 2001). The late-flowering phenotype of Arabidopsis co-2 mutants is rescued by the over-expression of BvCOL1 in sugar beet, thereby suggesting that function of BvCOL1 is equivalent to AtCO, and it has been shown that BvCOL1 interacts appropriately with the endogenous downstream genes FT and SUPPRESSOR OF OVEREXPRESSION OF CONSTANS1 (SOC1) in the transgenic plants (Chia et al. 2008). The expression patterns of VvCO and VvCOL1 indicate that these two genes are involved in flower induction and dormancy in grapevine buds (Almada et al. 2009). Potato StCO and poplar PtCO2 function in photoperiodic tuberization (GonzálezSchain et al. 2012) and seasonal growth cessation in trees (Bohlenius et al. 2006). In our studies, the expression patterns of ClCOL1 and ClCOL2 displayed no difference in the LD and inductive SD conditions; therefore they may not be involved in photoperiod flowering. However, their peak

13



Mol Genet Genomics

Fig. 6  ClCOL5 function analysis in Arabidopsis transgenic plants. a Examination of 35S::ClCOL5 expression in two representative T2 generation transgenic plants. b Representative images of indicated plants grown in LD conditions for 25 days. The wild-type control flowers later than the two 35S::ClCOL5 transgenic plants. c The expression levels of AtFT in transgenic Arabidopsis. Transcript levels were determined using qRT-PCR analyses in 10-day-old WT and transgenic seedlings. The expression results were normalized against TUB2 expression. Each column represents an average of three biological replicates, and the bars indicate the standard errors

levels occur at the end of the dark period, indicating that they may be involved in measuring the length of the dark period, and thus their functions need to be further studied to elucidate the molecular mechanism of SD-inductive flowering in C. lavandulifolium. The diurnal oscillations of ClCOL4 and ClCOL5 are similar with abundance in the early light phase and maintenance of relatively low expression levels in the dark phase (Figs. 4, 5); however, the peak expression levels of ClCOL4 and ClCOL5 are postponed and decrease in non-inductive LD or NB conditions compared with the inductive SD condition. The diurnal oscillations of ClCOL4–ClCOL5 are the same as HvCO9 in barley, which is located in a grass species-specific CO-like subfamily subgroup and does not have an Arabidopsis counterpart (Griffiths et al. 2003; Hemming et al. 2008; Greenup et al. 2009). In our previous study, the diurnal expression patterns of ClFT in different photoperiods were analyzed, and the peak expression level of ClFT occurs 2 h after light exposure under 12 hL/12 hD conditions and at 4 h after light exposure under 14 hL/10 hD conditions; however, the peak expression levels of ClFT disappear under 16 hL/8 hD or NB conditions (Fu et al. 2014). In this study, the peak expression levels of ClCOL4–ClCOL5 coincided with the peak expression level of ClFT. Thus, we postulate that ClCOL4 and ClCOL5 may function upstream of ClFT and activate the expression of ClFT to inducing flowering in C. lavandulifolium. Because the expression patterns of ClCOL4 and ClCOL5 are the same in different photoperiods, we selected one of these two genes, ClCOL5, which has the closest relationship to AtCOL5 according to the phylogenetic analysis, to analyze its function. In the T2 generation of the transgenic plants, the bolting and flowering times are greatly shortened compared with WT Arabidopsis (Fig. 6b; Table 3). At the same time, the rosette leaf numbers at the bolting or flowering period were greatly reduced (Table 3). We also observed that the expression levels of AtFT are greatly up-regulated in transgenic Arabidopsis (Fig. 6c). Therefore, ClCOL5 may function as a floral inducer in C. lavandulifolium. In Arabidopsis, overexpression of the AtCOL5 gene causes early flowering with a concomitant

Table 3  Traits affected in the T2 generation of ClCOL5 transgenic plants in Arabidopsis Genotype

Bolting timea (days)

Leaf number at bolting

Flowering timeb (days)

Leaf number at flowering

n

WT 35S::ClCOL5#1

33.55 ± 0.95a 26.00 ± 1.86b

10.14 ± 0.83a 8.10 ± 1.02b

40.52 ± 1.58a 33.09 ± 1.20b

10.04 ± 0.84a 8.26 ± 0.86b

32 35

35S::ClCOL5#2

26.15 ± 1.95b

8.95 ± 0.85b

33.52 ± 1.27b

8.39 ± 0.66b

33

WT indicates wild-type Arabidopsis. Values followed by different letters in the same column represent significant differences at P