Detection and validation of QTLs for flowering time in morning glory

Abstract

Japanese morning glory (Ipomoea nil), a short day plant, has been used for studying flowering times. Here, quantitative trait loci (QTL) analysis for days from sowing to flowering (DTF) of F₂ between I. nil var. Tokyo Kokei Standard (TKS) and I. hederacea line var. Q65, an early flowering variety, revealed seven QTLs: QTL Ipomoea Flowering 1–7 (qIF1–7). The position of qIF3, which had the most significant effect among the seven QTLs, corresponds with that of I. nil (or I. hederacea) CONSTANS (InCO/IhCO) in the linkage map. There is a single-base InDel in the coding sequence of InCO/IhCO. The single-base deletion (SBD) causes a frame-shift mutation and loss of function in TKS allele (inco-1). I. nil accessions bearing inco-1 tend to flower early, similarly to rice varieties bearing the loss of function allele of CO ortholog, hd1. The TKS allele of qIF3 reduces DTF and corresponds with the inferred effect of inco-1. Based on the distribution of inco-1, a hypothesis was proposed that the SBD in inco-1 might have played an important role in the expansion of Japanese morning glories, originally native to the tropical regions of the Americas, into temperate Asia.

Introduction

Japanese morning glory (Ipomoea nil), a short-day (SD) plant that is highly sensitive to the photoperiod, has been used as a model plant to study flowering (Hayama et al. 2007, Higuchi et al. 2011, Takeno 2016). Genetic pathways that promote flowering in response to seasonal cues were first reported in the model species Arabidopsis thaliana (Andrés and Coupland 2012), a long-day (LD) plant in which CO expression is regulated by light. The circadian clock plays a central role in regulating photoperiodic flowering (Andrés and Coupland 2012). CO encodes a zinc finger transcription factor, which binds to the promoter of FLOWERING LOCUS T (FT), a florigen gene, and which positively affects FT expression (Cao et al. 2014, Putterill et al. 1995).

For SD plants, intensive genetic studies of flowering have been conducted in rice (Hori et al. 2016). Several genes key to controlling the heading date of rice have been identified using quantitative trait locus (QTL) analysis and have been cloned. Heading date 1 (Hd1), the ortholog of Arabidopsis CO, is a major gene controlling the heading date in rice (Yano et al. 2000). In fact, Hd1 regulates expression of the Heading date 3a (Hd3a) rice florigen gene, delays heading under LD, and promotes heading under SD (Hayama et al. 2003, Izawa et al. 2002, Kojima et al. 2002, Tamaki et al. 2007, Yano et al. 2000).

From the study of I. nil flowering, the homologous genes of Arabidopsis flowering genes have been isolated (Hayama et al. 2007, Higuchi et al. 2011, Liu et al. 2001, Zheng et al. 2009). The I. nil CO ortholog, INIL11g18779, was originally named PnCO because of the former scientific name of Japanese morning glory: Pharbitis nil. As presented herein, the prefix “Pn” of genes derived from I. nil is replaced with “In”. Therefore, PnCO will be referred to as Ipomoea nil CONSTANS (InCO) herein. InCO, which is highly expressed in SD condition, has multiple splicing variants: a transcript containing a long intron (InCO (li)), a transcript containing a short intron (InCO (si)), and a transcript containing no intron (InCO (ni)) (Liu et al. 2001). Only the InCO (ni) mRNA encodes complete CONATANS-like protein with a CO, CO-like TOC1 (CCT) domain. It can restore an Arabidopsis co mutant. By contrast, InCO (li) and InCO (si) encodes truncated protein lacking the terminal CCT domain (Liu et al. 2001). Two I. nil homologous genes of FT, PnFT1 and PnFT2, expressed only in inductive SD conditions and overexpression of PnFT1, promote flowering in Arabidopsis and I. nil (Hayama et al. 2007). Transgenic plants constitutively expressing I. nil ortholog of GIGANTEA, PnGI, form fewer flowers than non-transgenic plants do. However, the expression level of InCO does not differ dramatically in transgenic plants compared to non-transgenic plants (Higuchi et al. 2011). Consequently, although the homologous genes of flowering related genes in Arabidopsis have been reported in I. nil, their function in photoperiodic flowering remains unclear.

The whole genome sequence of I. nil ‘Tokyo Kokei Standard’ (TKS) has been sequenced and useful genome sequences are now available (Hoshino et al. 2016a). EST-SSR marker of I. nil has also been reported: approximately half of them are polymorphic between TKS and Ivy-leaved morning glory, I. hederacea ‘Q65’ (Q65) (Ly et al. 2012). In addition, single nucleotide polymorphisms (SNP) become easily detected by comparing mRNA sequences between two accessions using the next generation sequencing (NGS) (Edwards et al. 2012). Although interspecific hybridizations between Q65 and Japanese strains of I. nil is not easy and reported that two pods and three seeds were obtained by pollination using 22 flowers (Yoneda and Takenaka 1981), F₁ hybrids are fertile and F₂ population is applicable to construct linkage maps. Consequently, these techniques and plant materials enable us to map QTLs for days from sowing to flowering (DTF) in morning glory.

Actually, I. nil has varieties of various flowering times (Yoneda and Takenaka 1981). Consequently, natural variations in I. nil are attractive resources for mapping QTLs affecting flowering time in dicot SD plants. Flowering in late July in Ibaraki, Japan, TKS is a standard variety of Japanese morning glory. Q65 was probably transported from the United States to Japan along with grains. It blooms approximately two weeks earlier than TKS. This report describes QTL mapping of DTF in the F₂ population between Q65 and TKS and a frame-shift mutation found in the TKS allele of InCO, which is the candidate gene of the most prominent QTL of DTF.

Materials and Methods

Plant materials

I. nil accessions TKS (Q1065), Q31, Q33, Q61, Q62, Q63, and Q1187 and an I. hederacea accession, Q65, and the F₂ population derived from the interspecific cross Q65 × TKS were provided by the Morning glory stock center of Kyushu University with support by the National Bio-Resource Project of the MEXT, Japan. An I. nil accession, PI227365 was provided by Agricultural Research Service, United States Department of Agriculture. I. nil varieties, Pekin-tendan (PKT) and Yakuyou-shirohana (YYS) were derived from the genetic stock of Ibaraki University (Hoshino et al. 2016b). An I. nil variety, Violet, was purchased from Marutane Seed Co., Kyoto, Japan. After dividing 192 individuals from the F₂ population, each of the 96 individuals was cultivated in 2011 and 2012. These plants were cultivated under natural conditions at the College of Agriculture, Ibaraki University (36°2ʹ14ʺN, 140°12ʹ47ʺE), Ami, Ibaraki, Japan.

SSR marker development from I. nil and I. batatas

Using 10% acrylamide gel electrophoresis, 326 SSR markers from I. nil (Ly et al. 2012) and 1250 EST-SSR markers from I. batatas (Kazusa DNA Lab, Japan) were analyzed. Polymorphic markers between TKS and Q65 were selected for genotyping the F₂ population (Supplemental Table 1). DNA was extracted from fresh leaves of each plant using a DNeasy Plant Mini kit (Qiagen Inc.). Each 6 μl of reaction mixture contained 1xNH₄ reaction buffer (Bioline; Meridian), 0.2 mM dNTP, 3 mM MgCl₂, 0.2 μM each primer, 0.15 unit Taq polymerase (Bioline; Meridian) and 1.0 ng of template DNA. Polymerase chain reactions (PCR) were run using a modified ‘touchdown PCR’ program (Sato et al. 2005). EST-SSR markers derived from I. nil (Ly et al. 2012) and I. batatas (Supplemental Table 1) showing polymorphisms between Q65 and TKS in polyacrylamide gel electrophoresis were used for linkage-maps construction.

Development and genotyping of SNP markers

After mRNA was isolated from 0.1 g buds of Q65 using an RNeasy Plant mini kit (Qiagen Inc.), it was sequenced using a Genetic Analyzer II (Illumina Inc.). Using software (Maq; Li et al. 2008), the sequence reads were mapped to a reference sequence, the non-redundant I. nil ESTs derived from TKS buds cDNA (Hoshino et al. 2016a, Ly et al. 2012), and were used to detect SNPs between Q65 and TKS. The detected SNPs were then used to develop SNP markers.

The RNA-seq data were archived at the DNA Data Bank of Japan under accession number DRA010196. These SNPs were genotyped by high-resolution melting (HRM) analysis (Liew et al. 2004) or melting temperature (T_m)-shift primer method (Fukuoka et al. 2008, Wang et al. 2005). Primers for HRM analysis were designed with Primer 3 (Rozen and Skaletsky 2000) (Supplemental Table 2). In designing of primers for HRM, Primer 3 parameters of PCR-amplicon length, optimal melting temperature, and primer length were set respectively as 80–110 bp, 60°C, and 20 bases. PCR amplification and DNA melt curve analysis were performed in a total volume of 10 μl containing 10 mM Tris-HCl (pH 8.3), 65 mM KCl, 1.5 mM MgCl₂, 0.2 μM each primer, 0.2 mM dNTP, 1.25% glycerol, 0.4× Eva Green (Biotium Inc.), 0.25 U Taq DNA polymerase (Ampliqon AS) and 1.0 ng of template DNA (Fukuoka et al. 2008). The PCR for HRM analysis was performed in High Resolution Melt/DNA binding Dye/DNA/PCR with HRM Curve mode of the Eco real-time PCR system (Illumina Inc.). The PCR conditions were 94°C for 3 min, 40 cycles of 96°C for 20 s, 58°C for 1 min, and 72°C for 30 s followed by melt curve analysis using the default settings of the system.

Primers for T_m-shift primer method were designed using ‘tms_primer_designer.pl’ (Fukuoka et al. 2008) (Supplemental Table 3). The PCR for Tm-shift primer method was performed in Quantification/DNA binding Dye/Standard curve mode of an Eco real-time PCR system (Illumina Inc.). The PCR conditions were 94°C for 3 min, 40 cycles of 95°C for 20 s, 58°C for 30 s, and 72°C for 30 s with subsequent melt curve analysis using the default settings of the system.

Linkage map construction and QTL analysis

From genotyping data of the F₂ population, we constructed linkage maps using QTL IciMapping software ver. 4.2 (Meng et al. 2015) and MapChart software (Voorrips 2002). DTF was calculated from the date of sowing to the date on which the first flower of an individual plant opened. Using each constructed linkage map and DTF-trait data, we detected QTLs for DTF with R/qtl package (Broman et al. 2003). We conducted interval mapping and composite interval mapping, respectively using the function, “scanone” and “cim”. Multiple QTL models were constructed by “scantwo”, “makeqtl”, “fitqtl”, “refineqtl” and “stepwiseqtl”. To establish alpha = 0.05 significant thresholds, the data were permutated 1000 times in each experiment. The estimated percent variances explained for the QTL and effects of QTL were calculated using the function “fitqtl”.

PCR amplification and DNA sequencing of the Q65 ortholog of InCO (IhCO)

Four primer sets were used for PCR amplification of IhCO and its 5ʹ flanking region (Supplemental Table 4). PCR amplifications were performed using KOD FX Neo DNA polymerase (Toyobo Co. Ltd., Osaka, JAPAN). The nucleotide sequences of the PCR products were analyzed with a capillary sequencer (ABI3130; Life Technologies Inc.) using the primers presented in Supplemental Table 4. The DNA sequence data derived from IhCO were archived at the DNA Data Bank of Japan under accession nos. LC716650 and LC768960. The alignments of nucleotide sequences and putative amino-acid sequences were found using Clustal W ver. 2.1 (Larkin et al. 2007). Splicing site prediction was performed with NetGene2-2.42 server (Hebsgaard et al. 1996, https://services.healthtech.dtu.dk/services/NetGene2-2.42/).

Quantitative reverse transcription PCR (qRT-PCR) analysis

TKS and Q65 plants were cultivated at 25°C under continuous light (195 μmol m^–2s^–1) until five days after sowing and were grown under SD (10 h light/14 h dark) or LD (14 h light/10 h dark) conditions for 3 days. Then, cotyledons were harvested after treating 8 h or 14 h of dark period from the last light period. Total RNA was isolated from cotyledons using an SV Total RNA Isolation System (Promega Corp.). cDNA was synthesized using Prime Script RT Master mix (Takara Bio Inc.). Real-time PCR was conducted using a Light Cycler 96 system (Roche) with TB Green^® Premix Ex Taq™ II (Takara Bio Inc.). Three primer sets, ni (forward primer (Fwd) 5ʹ-AGGGACTGCAGCAGCATAAC-3ʹ and reverse primer (Rev) 5ʹ-TGGAGACCATATCCCGTGTT-3ʹ), si (Fwd 5ʹ-CCATCAGTCACACAGTCTCCA-3ʹ and Rev 5ʹ-TGAAGCTGTGGAGGCATCTG-3ʹ) and li (Fwd 5ʹ-TCCAATAAAACCCAACGTCA-3ʹ and Rev 5ʹ-GGGGAAGTTGCATACCTTGA-3ʹ), were designed and used for qRT-PCR analyses of three InCO/IhCO splicing variants. The primer set for InACTIN4 (Fwd 5ʹ-GAATACTTGTATGCCACGAGCA-3ʹ and Rev 5ʹ-GGATTGCCAAGGCAGAGTAT-3ʹ) was used as the internal reference gene. Real-time PCR conditions were 94°C for 3 min, 40 cycles of 95°C for 10 s, 58°C for 30 s, and 72°C for 15 s, followed by melt curve analysis using default settings of the system. Relative expression levels for each sample were calculated based on the comparative Ct method. Welch’s two-tailed t-test was used to compare the expression levels between TKS and Q65.

Results

DTF in F₂ populations

In 2011 and 2012, seeds of the F₂ population were sown respectively on May 20 and June 7. The DTF values of individuals were measured (Fig. 1): those for Q65 and TKS were, respectively, 54 and 62 days in 2011 and 38 and 52 days in 2012. The ranges of DTF in the F₂ populations of 2011 and 2012 were 46–80 days and 39–76 days, respectively; more than half of individuals flower later than both parents in 2012 (Fig. 1).

Fig. 1.

Distribution of days from sowing to flowering among F₂ population derived from a cross: Q65 × TKS. A and B respectively show the F₂ population in 2011 and 2012. The black arrow and white arrow respectively indicate flowering days of Q65 and TKS.

Marker development and linkage-map construction

The F₂ populations cultivated in 2011 and 2012 were genotyped by DNA markers for constructing genetic linkage maps. The linkage map of the 2011 population contained 250 loci including 53 Tm-shift-SNP markers, 37 HRM-SNP markers, 143 EST-SSR markers of I. nil, 16 EST-SSR markers of I. batatas and one phenotypic marker, covered a total length of 1799 cM and included 15 linkage groups (Supplemental Fig. 1, Supplemental Tables 1–3). The linkage map of the 2012 population contained 192 loci including 43 Tm-shift-SNP markers, 34 HRM-SNP markers, 97 EST-SSR markers of I. nil, 17 EST-SSR markers of I. batatas and one phenotypic marker, covered a total length of 1664 cM and consisted of 15 linkage groups (Supplemental Fig. 2, Supplemental Tables 1–3). These constructed linkage groups were anchored to chromosomes using pseudo-chromosomes of Asagao_1.1 (Hoshino et al. 2016a) (Supplemental Figs. 1, 2).

QTL analysis for DTF

In the F₂ population cultivated in 2011, four QTLs for DTF, designated as qIF1, qIF2, qIF3, and qIF4 were mapped in the vicinity of the DNA marker of IES0160 on chromosome 5, Contig11987 on chromosome 9, Contig4567.156 on chromosome 11, and Contig05575 on chromosome 14 based on a multiple QTL model, respectively (Table 1, Supplemental Figs. 1, 3, 4). In this QTL model, the phenotype y is modeled as y = qIF1 + qIF2 + qIF3 + qIF4 + qIF1:qIF4 (A colon between two QTLs denotes interaction between the QTLs). The percentage of phenotypic variance explained (PVE) by these QTLs was 14.5%–23.1% (Table 1); the total PVE was 73.7% based on the multiple QTL model.

Table 1.Multiple QTL mapping of DTF in the cross between Q65 and TKS

QTL name	Year	Nearest marker	Chr.	Position (cM)	99% CIa (cM)	LOD	%varb	ac SEd	de SE	P (F)
qIF1	2011	IES0160	5	37	29–42	10.0	16.2	–5.4 ± 0.9	0.5 ± 1.2	3.2E-07
qIF2	2011	Contig11987	9	10	2–17	10.7	17.6	6.0 ± 1.4	–3.1 ± 1.4	5.7E-10
qIF3	2011	Contig4567.156	11	28	25–34	13.1	23.1	6.9 ± 0.9	4.2 ± 1.2	4.4E-12
	2012	Contig4567.156	11	50	37–64	5.6	13.0	4.2 ± 0.8	0.9 ± 1.5	1.4E-05
qIF4	2011	Contig05575	14	16	10–20	9.1	14.5	–4.6 ± 1.0	1.8 ± 1.2	1.1E-04
qIF1:qIF4f	2011	–	–	–	–	6.2	9.2	–	–	4.8E-05
qIF5	2012	rJMFF041I11	4	43	0–72	2.9	6.4	–3.3 ± 0.9	0.4 ± 1.2	2.9E-03
qIF6	2012	Contig683.0110	10	88	83–91	10.4	27.3	–4.0 ± 0.8	–1.9 ± 1.2	1.6E-07
qIF7	2012	JMFN020H15	14	42	34–54	10.2	26.5	–1.1 ± 1.0	–3.5 ± 1.3	2.4E-07
qIF6:qIF7	2012	–	–	–	–	8.9	22.4	–	–	3.3E-07

^a 99% CI, 99% Bayesian credible interval of the QTL; ^b %var, percentage of variance explained by the QTL; ^c a, additive effect; ^d SE, standard error; ^e d, dominance effect; ^f Colons show interactions between two QTLs.

In the F₂ population cultivated in 2012, a multiple QTL model was obtained consisting of three QTLs other than qIF3. The three QTLs, qIF5, qIF6, and qIF7 were mapped in the vicinity of the DNA marker of rJMFF041I11 on chromosome 4, Contig683.0110 on chromosome 10, and JMFN020H15 on chromosome 14, respectively (Table 1, Supplemental Figs. 2, 3, 5). In this QTL model, the phenotype y is modeled as y = qIF3 + qIF5 + qIF6 + qIF7 + qIF6:qIF7. The PVE by these QTLs was 6.3%–27.3% (Table 1); the total PVE was 58.4% based on the multiple QTL model.

qIF3 showed the highest PVE among these QTLs (23.1%) in the 2011 analysis and demonstrated stability by being detected as a significant QTL again in 2012. However, the LOD value of qIF3 in 2012 was reduced by approximately one-third compared to that in 2011. The qIF1 and qIF2 also demonstrated the readily apparent effects (Table 1, Supplemental Fig. 4) and detected in the simple interval mapping (SIM) or the composite interval mapping (CIM) (Supplemental Fig. 3). However, no QTLs except qIF3 were detected neither in SIM and CIM in 2012 analysis (Supplemental Fig. 3). In 2012, seeds were sown two weeks later than in 2011. The shorter period of long-day condition in 2012 population might reduce power for detection of QTL related to photoperiodism.

qIF1 and qIF4 in 2011 and qIF6 and qIF7 in 2012 were detected as interacting QTLs (Supplemental Fig. 6A–6C). In individuals with a homozygous Q65 allele of qIF7, the plants that are homozygous for the TKS allele of qIF6 exhibited a delayed flowering compared to plants homozygous and heterozygous for the Q65 allele of qIF6 (Supplemental Fig. 6C). A significant interaction (LOD = 7.12 > 6.39 (P = 0.05)) was also detected between the chromosomal region at 116 cM of chromosome 14 (14@116) and at 123 cM of chromosome 5 (5@123) (Supplemental Fig. 6D). The effects of TKS and Q65 allele of 5@123 reversed depending on 14@116 genotypes.

Polymorphism between the TKS and Q65 alleles of InCO/IhCO

The strongest effect QTL for DTF, qIF3 was co-located with InCO/IhCO in the linkage maps. Also, Contig4567.156, the DNA marker the nearest to the LOD peak of qIF3, was developed based on polymorphism in the EST encoding InCO (Table 1, Supplemental Figs. 1, 2, Supplemental Table 2). Additionally, when we searched the confidence interval of qIF3 using the Japanese morning glory genome database, no homologs of genes related to flowering time in other plant taxa, except for InCO, were identified. Consequently, InCO/IhCO might be a candidate of qIF3. Therefore, we compared the DNA sequence of InCO/IhCO between TKS and Q65.

First, using PCR amplification and agarose gel electrophoresis, the 5ʹ flanking regions of InCO/IhCO were compared among Q65 and four I. nil accessions: PKT, TKS, Q63, and YYS (Fig. 2). The size of PCR products amplified from Q65 was 2.3 kbp and approximately 300 bp larger than that of other accessions (Fig. 2), which is attributable to the structural difference of the 5ʹ flanking region of InCO/IhCO between Q65 and other accessions including TKS. Then, the PCR products of Q65 were sequenced and aligned with the DNA sequence of TKS (Hoshino et al. 2016a) (Supplemental Fig. 7). Short interspersed element (SINE)-like 168 bp insertion sequence was found in the reverse strand of approximately 1.9 kb upstream of the putative transcription start site (Supplemental Fig. 7). There were a conserved putative RNA polymerase III promoter containing A-box and B-box (Galli et al. 1981) and target site duplications (Fig. 3). A few thousand copies of this SINE-like sequence were detected by homology search against the Asagao 1.2 genome (Hoshino et al. 2016a) using the BLAST program. DNA sequences downstream of the SINE-like sequence were highly polymorphic between TKS and Q65. There was a long TA repeat (more than 200 bp) in the Q65 sequence (Supplemental Fig. 7).

Fig. 2.

Agarose gel electrophoresis of PCR products amplified from the 5ʹ flanking regions of InCO/IhCO in five morning glory accessions: PKT, Pekin tendan; TKS, Tokyo Kokei standard; Q63; YYS, Yakuyo shirohana; and Q65.

Fig. 3.

A SINE-like sequence found in the 5ʹ flanking region of IhCO. DNA sequences surrounded by square frame show the SINE-like sequence. TSD, target site duplication. TSDs are shown by underlining. The locations of putative boxA and boxB are shown by a gray background.

In addition to the polymorphisms in the 5ʹ flanking region, the coding sequence (CDS) of Q65 had 10 single-base substitutions (SBS) and 3 insertions against TKS sequence. Of the 10 SBSs, 7 are synonymous substitutions, which represent functional constraints of InCO/IhCO. Of the three insertions, the two insertions were 9 and 12 nucleotides, causing an additional 3 and 4 amino acids, respectively, but one was a single-base insertion causing a frame-shift of the translational reading frame (Fig. 4A). The Q65 allele has a single cytosine base present between positions 7617445 and 7617446 on Chr11 of TKS in the genome sequence (Asagao_1.2) (Fig. 4A). If the splice sites are identical to InCO (ni), the transcript reported for the ‘Violet’ allele (Liu et al. 2001), then the Q65 mRNA would encode an incomplete protein without a CCT domain due to the insertion. Therefore, the splice-donor site of Q65 allele was assessed using NetGene2-2.42 server (Hebsgaard et al. 1996) (Fig. 4B). As a result, the splice-donor site of InCO (ni) was not detected confidently and 26 nucleotides (nt) downstream from the splice-donor site of InCO (ni) was the most confident splice-donor site (Fig. 4B). The transcript spliced at 26 nt downstream was correspondent with the transcript reported by Liu et al. (2001) as InCO (si). InCO (si) encodes an incomplete protein, which results from premature stop codon in ‘Violet’ allele (Liu et al. 2001), but in the Q65 allele, it encodes complete 433 aa protein with the CCT domain (Supplemental Fig. 8).

Fig. 4.

InCO/IhCO-splice-site prediction at Q65 allele. (A) The genomic DNA sequence around an intron of Q65 allele. Numbers at the right side of DNA sequences show nucleotide numbers from the putative transcription start site. The boxed cytosine is inserted into the TKS allele. A white-reverse triangle shows a splice donor site reported by Liu et al. (2001) as no-intron mRNA (InCO (ni)). A black reverse triangle shows a splice donor site reported by Liu et al. (2001) as short-intron mRNA (InCO (si)) and also as predicted by NetGene2 v. 2.4 as the most confident splice-donor site. The DNA sequence with gray background denotes the intron of InCO (si). “v” above the DNA sequence signifies the SNP site position. ‘A/G’ shows the adenine of TKS allele and the guanine of Q65 allele at the SNP site. Primer pairs used for expression analysis are shown as arrows. (B) A splice-donor-site prediction by Net Gene2 v. 2.4. Pos. denotes the nucleotide position from the putative transcription start site.

Distribution of the single-base InDel in the CDS of InCO among accessions derived from natural populations

To determine the ancestral form of the single-base InDel in the CDS of InCO, the single-base InDel sites were investigated in eight I. nil accessions derived from natural populations in addition to TKS and Q65 (Table 2). Accessions Q33 and Q1187, derived from the South American continent, have cytosine insertion at the single-base InDel site as Q65 (Table 2). Although the number of accessions surveyed was limited, accessions that have the same sequence as TKS in the single-base InDel are limited to those from the Asian countries (Table 2). Since I. nil originates from the tropical regions of the Americas (Austin et al. 2001), the Q65-type sequence of the single-base InDel is likely to represent the ancestral form. Moreover, the fact that InCO orthologs of Ipomoea species I. triloba and I. trifida also contain the Q65-type sequence in the single-base InDel also support the idea that the Q65-type sequence of the single-base InDel is ancestral (Fig. 5). Thus, TKS type allele (inco-1) might emerged from the ancestral allele of InCO (InCO-2) by the single-base deletion (SBD) in the CDS (Fig. 6).

Table 2.DTF and the single-base insertion/deletion in the CDS of InCO/IhCO among morning glory varieties

Variety	Origin	Sequencea	DTFb (± S.E.)
Q65	–	CCG	44 ± 1.3	n = 3
Q61	China	C-G	53 ± 2.9	n = 5
Violet	Japan	C-G	60 ± 4.3	n = 3
Q62	Nepal	C-G	65 ± 0.7	n = 5
TKS	Japan	C-G	66 ± 0.9	n = 3
PI227365	Iran	C-G	95 –	n = 1
Q31	China	CCG	103 ± 3.6	n = 5
Q1187	Paraguay	CCG	109 ± 2.9	n = 5
Q63	Guinea	CCG	133 ± 4.1	n = 3
Q33	Brazil	CCG	138 ± 1.9	n = 4

^a DNA sequence of the single base insertion/deletion site in the InCO/IhCO CDS.

^b Days from sowing to flowering at Ami, Ibaraki, Japan. Seeds were sown on May 20, 2022.

Fig. 5.

DNA sequence alignment of InCO orthologous genes around the single-base InDel causing frameshift in InCO. DNA sequence of I. triloba (XR_004098766.1) and of I. trifida (CP025648.1) were aligned with Clustal W (Ver. 1.83, 2003). Gray background denotes the single-base-InDel site.

Fig. 6.

InCO/IhCO alleles and comparison of the expression levels of three transcripts of InCO/IhCO in Q65 and TKS. (A) Map of the InCO/IhCO alleles and the PCR primer pairs. Narrow black boxes show untranscribed regions. Wider boxes show exons. A narrow white box shows an intron. White triangles and vertical broken lines show cryptic splice sites producing InCO (ni). S1, S2, and S3 of inco-1 respectively show stop codon sites of InCO (li), InCO (si), and InCO (ni). Gray parts of the boxes show regions homologous to IhCO protein in a.a. Light-gray parts of inco-1 show that InCO (ni) protein is homologous to IhCO protein. The diagonal stripe pattern shows the region where the InCO (ni) protein is homologous with the IhCO protein, but not with the InCO (si) protein. ni, si, and li indicated by arrows and broken lines are primer pairs used for qRT-PCR. These ni, si, and li were used respectively for amplifying splicing variants, InCO (ni), InCO (si), and InCO (li). (B, C) Relative expression analysis of three InCO splicing variants using qRT-PCR. *, **, and *** respectively denote significant difference at P < 0.05, P < 0.01, and P < 0.001 by Welch’s two-tailed t-test).

DTFs of the accessions presented in Table 2 were investigated in 2022. They were from 44 days of Q65 to 138 days of Q33. The accessions without the SBD in the CDS of InCO/IhCO tended to have longer DTFs than those with the SBD, except for Q65 (Table 2).

Comparison of InCO/IhCO expression between TKS and Q65

The expression levels of three transcripts, InCO (ni), InCO (si), and InCO (li) were compared between TKS and Q65 in SD and LD conditions using qRT-PCR (Fig. 6B, 6C). The relative expression levels of the Q65 allele tended to be higher than TKS in all three transcripts across all conditions. Especially, the difference between TKS and Q65 was the most pronounced in InCO (si) (Fig. 6B). The expression level of InCO (si) was sevenfold higher in Q65 than in TKS after 14 h of darkness in the SD condition.

In all conditions, among the transcripts of three types, InCO (li) exhibited the highest expression level, as reported earlier in the relevant literature (Liu et al. 2001), followed by InCO (si). The lowest expression level was observed for InCO (ni) (Fig. 6B, 6C). Although quantities differed among the three transcripts, the expression patterns responding to the day length and the dark period were mutually similar. These tendencies were observed for both TKS and Q65.

When comparing expression levels of the transcripts encoding the InCO protein with an intact CCT domain at 14 h after dark in the SD condition, the expression level of InCO (si) in Q65 was 13 times more than those of InCO (ni) in TKS.

Discussion

InCO/IhCO, a candidate gene for the largest effect QTL, qIF3, for DTF

In this study, seven QTLs for DTF, qIF1–7, were detected in the F₂ population of Q65 and TKS (Table 1). InCO/IhCO is located near the LOD peak of the largest effect QTL, qIF3. InCO is known to the ortholog of CO (Liu et al. 2001), which is a central regulator of photoperiodism in Arabidopsis (Andrés and Coupland 2012). If InCO were qIF3, then InCO/IhCO would be expected to have polymorphisms causing functional differences between TKS and Q65. In fact, SBD causes a frame-shift in the InCO/IhCO CDS (Fig. 4, Supplemental Fig. 8). The expression analysis and splicing site prediction demonstrated that the transcript variant InCO (si) is expected to produce a protein with an intact CCT domain in the Q65 allele (Fig. 6, Supplemental Fig. 8). Actually, InCO (ni) was earlier thought to be a transcript without introns and encoding the functional InCO protein (Liu et al. 2001), but it appears to be a minor transcript (Figs. 4, 6). Although the splicing-variant InCO (ni) of TKS allele encodes functional CONSTANS-like protein (Liu et al. 2001), the expression level of TKS InCO (ni) is much less than that of Q65 InCO (si) (Fig. 6B). Reportedly, the expression levels of CONSTANS affect the flowering time in Arabidopsis (Rosas et al. 2014). Therefore, the TKS allele of InCO might be defective because of the frame shift caused by the SBD. For that reason, we designate the TKS/Violet type defective allele as inco-1 and the putative ancestral allele without the SBD as InCO-2.

In addition to differences in the expression levels, the amino acid sequence encoded by TKS InCO (ni) differs from that encoded by Q65 InCO (si) (Supplemental Fig. 8). In TKS InCO (ni), three nonsynonymous substitutions are present, and due to alternative splicing, the region corresponding to 26 amino acids in the Q65 InCO (si) protein is replaced by a distinct 17-amino-acid sequence (Supplemental Fig. 8). Although both proteins retain key motifs such as the N-terminal B-box-type zinc fingers and the C-terminal CCT domain (Supplemental Fig. 8), their predicted three-dimensional structures differ substantially (Jumper et al. 2021). The Violet InCO (ni), which shares the same sequence as TKS InCO (ni), has been reported to complement the co mutant of Arabidopsis and promote flowering (Liu et al. 2001). However, I. nil is a short-day plant, and InCO is predicted to function analogously to Hd1 in rice, acting as a flowering repressor under long-day conditions (Yano et al. 2000). Therefore, whether the protein encoded by TKS InCO (ni) is fully functional remains unclear, and further investigation using I. nil will be necessary to clarify the functions of transcript variants in InCO/IhCO.

The amino acid sequence encoded by InCO (si) of the African accession Q63 is identical to that of Q65 InCO (si), except for the asparagine–proline and asparagine repeat regions (data not shown). Considering that Q63 and Q65 belong to different species and are genetically distant (Ly et al. 2012), the InCO/IhCO amino acid sequence is likely subject to strong functional constraints. In contrast, the presence of three nonsynonymous substitutions between the Q65 and TKS alleles imply that the TKS allele of InCO might have experienced relaxed functional constraints due to loss of function.

The allelic difference of InCO/IhCO corresponds with the effects of qIF3 on DTF

If InCO/IhCO were qIF3, the phenotypic effects expected from the genotypes of InCO/IhCO on the DTF are assumed to correspond to the phenotypic effect of qIF3. The Q65 allele of qIF3 has the effect of increasing DTF, whereas the TKS allele of qIF3 has the effect of decreasing DTF (Table 1, Supplemental Fig. 4). Therefore, if InCO/IhCO were qIF3, then it can be inferred that InCO/IhCO allele without the SBD would increase DTF; also, inco-1 allele with SBD would decrease DTF. This inference corresponds with results showing that the accessions bearing inco-1 tend to flower earlier than those bearing the InCO/IhCO allele without the SBD, except Q65 (Table 2). Among the strains examined for this study, Q65 flowers the earliest, but it has the IhCO allele without the SBD (Table 2). The fact that Q65 flowers the earliest among accessions examined for this study does not contradict the notion that the InCO/IhCO allele without the SBD has an effect of delaying flowering. If InCO/IhCO were qIF3, then the Q65 allele, that is the IhCO allele without the SBD, would also exert a delaying effect on flowering because the Q65 allele of qIF3 has the effect of increasing DTF (Table 1, Supplemental Fig. 4). Consequently, the allelic differences in the structure and the predicted effect for DTF of InCO/IhCO can well explain the allelic differences of qIF3.

In rice, dysfunctional alleles of Hd1 (hd1), the rice ortholog of CONSTANS, are known to reduce photoperiod sensitivity. Moreover, they have the effect of promoting heading in field conditions (Ebana et al. 2011, Fujino et al. 2019, Hori et al. 2016, Mo et al. 2021, Yano et al. 2000). The early flowering tendency of rice cultivars with hd1 corresponds with that of I. nil varieties with inco-1 (Table 2). The photosensitivity of crops with ancestors originating from tropical regions, which are classified as short-day plants, tend to be reduced or lost during domestication and expansion to high-latitude areas (Lin et al. 2021). The I. nil varieties with inco-1 were found exclusively in Asia but not in the Americas, the origin of I. nil, and Africa. Consequently, inco-1 might have emerged in Asia, and might have contributed to the spread of I. nil across temperate Asia by reducing its photoperiod sensitivity.

Genetic mechanism for the early flowering of Q65

The fact that Q65 was the earliest flowering among the strains investigated for this study (Table 2) despite having the ancestral allele of InCO/IhCO implies that Q65 possesses a mechanism for early flowering that is distinct from that of Asian I. nil strains. Genes other than IhCO might be responsible for causing Q65 to flower early.

Seven QTLs for DTF were detected in this study. Among the seven QTLs, the Q65 allele of qIF1, qIF4, qIF5 and qIF6 have the effect of reducing DTF and enhancing flowering (Table 1, Supplemental Figs. 4, 5). In addition to these QTLs, two interactions also reduced DTF when both QTLs were homozygous for the Q65 allele (Supplemental Fig. 6C, 6D). However, the effects of these QTLs are not stable enough to be detected in both 2011 and 2012, and were not high enough to fully explain the early flowering of Q65 (Table 1, Supplemental Figs. 4, 5). Therefore, additional genetic studies using suitable experimental lines might be necessary to elucidate the causal genetic mechanism of the early flowering in Q65.

In rice, the combination of the loss of function of two genes, Ghd7 and Osppr37, results in extremely early flowering. Given this genetic background, Hd1 has been reported to promote earlier heading (Fujino et al. 2019). Similarly, in Q65, the loss of function of multiple genes other than IhCO can be presumed to contribute to early flowering.

Polymorphism in the 5ʹ-upstream regulatory region of InCO/IhCO

Insertion of a SINE-like retrotransposon, which does not exist in TKS allele of InCO/IhCO, was observed in 5ʹ-flanking region of Q65 allele (Fig. 3, Supplemental Fig. 7). SINEs are known to cause changes in the chromatin structure and to affect gene expression (Elbarbary et al. 2016). The expression levels of Q65 allele of InCO/IhCO are greater than those of TKS (Fig. 6). Therefore, structural differences and the SINE-like sequence insertion upstream of Q65 allele (Supplemental Fig. 7) might affect the expression level of InCO/IhCO.

An intron-containing InCO/IhCO transcript variant, InCO (li)

In both Q65 and TKS, InCO (li) exhibits the highest expression level among the three splicing variants, although it includes an intron sequence and encodes truncated protein without CCT domain (Liu et al. 2001) (Fig. 6). In Arabidopsis, an intron-containing CO transcript variant, COβ, also encodes a truncated protein without CCT domain caused by premature stop codon. The COβ protein interacts with the functional form of CO protein, COα, and reduces its stability (Gil et al. 2017). Therefore, the InCO (li) protein might also interact with the functional InCO/IhCO protein and might affect InCO/IhCO functions. However the expression level of COβ is less than COα (Gil et al. 2017). Therefore, the highest expression level of InCO (li) among splicing variants is a unique and interesting feature of InCO/IhCO.

Conclusion

This study suggests that the TKS allele of InCO, a candidate of the most significant QTL for DTF, qIF3, is likely dysfunctional due to the SBD. In tropical-origin SD plants like rice, it is known that a loss-of-function mutation in the Hd1 enhances flowering. Similarly, the loss of function in InCO might promote flowering in I. nil. Accessions carrying the same SBD with the TKS allele tend to flower earlier, and within the accessions analyzed, those carrying the SBD were suggested to originate from temperate regions of Asia. Based on these results, we hypothesize that the SBD in InCO is promoting earlier flowering in mid-latitudes and enabling I. nil, which originated in the tropical regions of the Americas, to adapt to temperate Asia.

Author Contribution Statement

Conceptualization: HK and TK; resources provision: EN and AH; mapping population development: EN; DNA marker development: TI, HK, HF, SS, and SI; investigation: HK, TI, KE, and TK; writing–original draft preparation: HK and TK; writing–review and editing: TK, EH, and AH; supervision: AH, EN, MO, and NW; project administration: TK; funding acquisition: TK and MO. All authors have read and agreed to the published version of the manuscript.

 Acknowledgments

This work was partially supported by JSPS KAKENHI Grant Number 18K05569. This work was also partially supported by the Cooperative Research Grant of the Plant Transgenic Design Initiative (PtraD) at the Gene Research Center, University of Tsukuba.

Literature Cited

•  Andrés,  F. and  G.  Coupland (2012) The genetic basis of flowering responses to seasonal cues. Nat Rev Genet 13: 627–639.
•  Austin,  D.F.,  K.  Kitajima,  Y.  Yoneda and  L.  Qian (2001) A putative tropical American plant, Ipomoea nil (Convolvulaceae), in pre-Columbian Japanese art. Econ Bot 55: 515–527.
•  Broman,  K.W.,  H.  Wu,  Ś.  Sen and  G.A.  Churchill (2003) R/qtl: QTL mapping in experimental crosses. Bioinformatics 19: 889–890.
•  Cao,  S.,  R.W.  Kumimoto,  N.  Gnesutta,  A.M.  Calogero,  R.  Mantovani and  B.F.  Holt (2014) A distal CCAAT/NUCLEAR FACTOR Y complex promotes chromatin looping at the FLOWERING LOCUS T promoter and regulates the timing of flowering in Arabidopsis. Plant Cell 26: 1009–1017.
•  Ebana,  K.,  T.  Shibaya,  J.  Wu,  K.  Matsubara,  H.  Kanamori,  H.  Yamane,  U.  Yamanouchi,  T.  Mizubayashi,  I.  Kono,  A.  Shomura et al. (2011) Uncovering of major genetic factors generating naturally occurring variation in heading date among Asian rice cultivars. Theor Appl Genet 122: 1199–1210.
•  Edwards,  C.E.,  T.L.  Parchman and  C.W.  Weekley (2012) Assembly, gene annotation and marker development using 454 floral transcriptome sequences in Ziziphus celata (Rhamnaceae), a highly endangered, Florida endemic plant. DNA Res 19: 1–9.
•  Elbarbary,  R.A.,  B.A.  Lucas and  L.E.  Maquat (2016) Retrotransposons as regulators of gene expression. Science 351: aac7247.
•  Fujino,  K.,  U.  Yamanouchi,  Y.  Nonoue,  M.  Obara and  M.  Yano (2019) Switching genetic effects of the flowering time gene Hd1 in LD conditions by Ghd7 and OsPRR37 in rice. Breed Sci 69: 127–132.
•  Fukuoka,  H.,  K.  Miyatake,  S.  Negoro,  T.  Nunome,  A.  Ohyama and  H.  Yamaguchi (2008) Development of a routine procedure for single nucleotide polymorphism marker design based on the T_m-shift genotyping method. Breed Sci 58: 461–464.
•  Galli,  G.,  H.  Hofstetter and  M.L.  Birnstiel (1981) Two conserved sequence blocks within eukaryotic tRNA genes are major promoter elements. Nature 294: 626–631.
•  Gil,  K.E.,  M.J.  Park,  H.J.  Lee,  Y.J.  Park,  S.H.  Han,  Y.J.  Kwon,  P.J.  Seo,  J.H.  Jung and  C.M.  Park (2017) Alternative splicing provides a proactive mechanism for the diurnal CONSTANS dynamics in Arabidopsis photoperiodic flowering. Plant J 89: 128–140.
•  Hayama,  R.,  S.  Yokoi,  S.  Tamaki,  M.  Yano and  K.  Shimamoto (2003) Adaptation of photoperiodic control pathways produces short-day flowering in rice. Nature 422: 719–722.
•  Hayama,  R.,  B.  Agashe,  E.  Luley,  R.  King and  G.  Coupland (2007) A circadian rhythm set by dusk determines the expression of FT homologs and the short-day photoperiodic flowering response in Pharbitis. Plant Cell 19: 2988–3000.
•  Hebsgaard,  S.M.,  P.G.  Korning,  N.  Tolstrup,  J.  Engelbrecht,  P.  Rouze and  S.  Brunak (1996) Splice site prediction in Arabidopsis thaliana pre-mRNA by combining local and global sequence information. Nucleic Acids Res 24: 3439–3452.
•  Higuchi,  Y.,  K.  Sage-Ono,  R.  Sasaki,  N.  Ohtsuki,  A.  Hoshino,  S.  Iida,  H.  Kamada and  M.  Ono (2011) Constitutive expression of the GIGANTEA ortholog affects circadian rhythms and suppresses one-shot induction of flowering in Pharbitis nil, a typical short-day plant. Plant Cell Physiol 52: 638–650.
•  Hori,  K.,  K.  Matsubara and  M.  Yano (2016) Genetic control of flowering time in rice: integration of Mendelian genetics and genomics. Theor Appl Genet 129: 2241–2252.
•  Hoshino,  A.,  V.  Jayakumar,  E.  Nitasaka,  A.  Toyoda,  H.  Noguchi,  T.  Itoh,  T.  Shin-I,  Y.  Minakuchi,  Y.  Koda,  A.  Nagano et al. (2016a) Genome sequence and analysis of the Japanese morning glory Ipomoea nil. Nat Commun 7: 13295.
•  Hoshino,  A.,  Y.  Yoneada and  T.  Kuboyama (2016b) A Stowaway transposon disrupts the InWDR1 gene controlling flower and seed coloration in a medicinal cultivar of the Japanese morning glory. Genes Genet Syst 91: 37–40.
•  Izawa,  T.,  T.  Oikawa,  N.  Sugiyama,  T.  Tanisaka,  M.  Yano and  K.  Shimamoto (2002) Phytochrome mediates the external light signal to repress FT orthologs in photoperiodic flowering of rice. Genes Dev 16: 2006–2020.
•  Jumper,  J.,  R.  Evans,  A.  Pritzel,  T.  Green,  M.  Figurnov,  O.  Ronneberger,  K.  Tunyasuvunakool,  R.  Bates,  A.  Židek,  A.  Potapenko et al. (2021) Highly accurate protein structure prediction with AlphaFold. Nature 596: 583–589.
•  Kojima,  S.,  Y.  Takahashi,  Y.  Kobayashi,  L.  Monna,  T.  Sasaki,  T.  Araki and  M.  Yano (2002) Hd3a, a rice ortholog of Arabidopsis FT gene, promotes transition to flowering downstream of Hd1 under short-day condition. Plant Cell Physiol 43: 1096–1105.
•  Larkin,  M.A.,  G.  Blackshields,  N.P.  Brown,  R.  Chenna,  P.A.  McGettigan,  H.  McWilliam,  F.  Valentin,  I.M.  Wallace,  A.  Wilm,  R.  Lopez et al. (2007) Clustal W and Clustal X ver. 2.0. Bioinformatics 23: 2947–2948.
•  Li,  H.,  J.  Ruan and  R.  Durbin (2008) Mapping short DNA sequencing reads and calling variants using mapping quality scores. Genome Res 18: 1851–1858.
•  Liew,  M.,  R.  Pryor,  R.  Palais,  C.  Meadows,  M.  Erali,  E.  Lyon and  C.  Wittwer (2004) Genotyping of single-nucleotide polymorphisms by high-resolution melting of small amplicons. Clin Chem 50: 1156–1164.
•  Lin,  X.,  C.  Fang,  B.  Liu and  F.  Kong (2021) Natural variation and artificial selection of photoperiodic flowering genes and their applications in crop adaptation. aBIOTECH 2: 156–169.
•  Liu,  J.,  J.  Yu,  L.  McIntosh,  H.  Kende and  J.A.D.  Zeevaart (2001) Isolation of a CONSTANS ortholog from Pharbitis nil and its role in flowering. Plant Physiol 125: 1821–1830.
•  Ly,  T.,  H.  Fukuoka,  A.  Otaka,  A.  Hoshino,  S.  Iida,  E.  Nitasaka,  N.  Watanabe and  T.  Kuboyama (2012) Development of EST-SSR markers of Ipomoea nil. Breed Sci 62: 99–104.
•  Meng,  L.,  H.  Li,  L.  Zhang and  J.  Wang (2015) QTL IciMapping: Integrated software for genetic linkage map construction and quantitative trait locus mapping in biparental populations. Crop J 3: 269–283.
•  Mo,  Y.,  C.M.  Lee,  H.M.  Park,  S.K.  Ha,  M.J.  Kim,  J.  Kwak,  H.S.  Lee,  J.H.  Lee and  J.U.  Jeung (2021) Hd1 allele types and their associations with major agronomic traits in Korean rice cultivars. Plants (Basel) 10: 2408.
•  Putterill,  J.,  F.  Robson,  K.  Lee,  R.  Simon and  G.  Coupland (1995) The CONSTANS gene of Arabidopsis promotes flowering and encodes a protein showing similarities to zinc finger transcription factors. Cell 80: 847–857.
•  Rosas,  U.,  Y.  Mei,  Q.  Xie,  J.A.  Banta,  R.W.  Zhou,  G.  Seufferheld,  S.  Gerard,  L.  Chou,  N.  Bhambhra,  J.D.  Parks et al. (2014) Variation in Arabidopsis flowering time associated with cis-regulatory variation in CONSTANS. Nat Commun 5: 3651.
•  Rozen,  S. and  H.  Skaletsky (2000) Primer3 on the WWW for general users and for biologist programmers. In: Misener, S. and S.A. Krawetz (eds.) Bioinformatics methods and protocols: Methods in molecular biology. Humana Press, Totowa, NJ, pp. 365–386.
•  Sato,  S.,  S.  Isobe,  E.  Asamizu,  N.  Ohmido,  R.  Kataoka,  Y.  Nakamura,  T.  Kaneko,  N.  Sakurai,  K.  Okumura,  I.  Klimenko et al. (2005) Comprehensive structural analysis of the genome of red clover (Trifolium pratense L.). DNA Res 12: 301–364.
•  Takeno,  K. (2016) Stress-induced flowering: The third category of flowering response. J Exp Bot 67: 4925–4934.
•  Tamaki,  S.,  S.  Matsuo,  H.L.  Wong,  S.  Yokoi and  K.  Shimamoto (2007) Hd3a protein is a mobile flowering signal in rice. Science 316: 1033–1036.
•  Voorrips,  R.E. (2002) MapChart: Software for the graphical presentation of linkage maps and QTLs. J Hered 93: 77–78.
•  Wang,  J.,  K.  Chuang,  M.  Ahluwalia,  S.  Patel,  N.  Umblas,  D.  Mirel,  R.  Higuchi and  S.  Germer (2005) High-throughput SNP genotyping by single-tube PCR with T_m-shift primers. Biotechniques 39: 885–893.
•  Yano,  M.,  Y.  Katayose,  M.  Ashikari,  U.  Yamanouchi,  L.  Monna,  T.  Fuse,  T.  Baba,  K.  Yamamoto,  Y.  Umehara,  Y.  Nagamura et al. (2000) Hd1, a major photoperiod sensitivity quantitative trait locus in rice, is closely related to the Arabidopsis flowering time gene CONSTANS. Plant Cell 12: 2473–2483.
•  Yoneda,  Y. and  Y.  Takenaka (1981) Genshoku asagao kensaku zukan (Natural-Color Illustrated monograph of Japanese morning glory), 2nd edn. Hokuryukan, Tokyo, pp. 50–58 (in Japanese).
•  Zheng,  C.C.,  D.  Potter and  S.D.  O’Neill (2009) Phytochrome gene expression and phylogenetic analysis in the short-day plant Phabitis nil (Convolvulaceae): differential regulation by light and an endogenous clock. Am J Bot 96: 1319–1336.