StMYB44 negatively regulates anthocyanin biosynthesis at high temperatures in tuber flesh of potato.

High temperatures are known to reduce anthocyanin accumulation in a number of diverse plant species. In potato (Solanum tuberosum L.), high temperature significantly reduces tuber anthocyanin pigment content. However, the mechanism of anthocyanin biosynthesis in potato tuber under heat stress remains unknown. Here we show that high temperature causes reduction of anthocyanin biosynthesis in both potato tuber skin and flesh, with white areas forming between the vasculature and periderm. Heat stress reduced the expression of the R2R3 MYB transcription factors (TFs) StAN1 and StbHLH1, members of the transcriptional complex responsible for coordinated regulation of the skin and flesh pigmentation, as well as anthocyanin biosynthetic pathway genes in white regions. However, the core phenylpropanoid pathway, lignin, and chlorogenic acid (CGA) pathway genes were up-regulated in white areas, suggesting that suppression of the anthocyanin branch may result in re-routing phenylpropanoid flux into the CGA or lignin biosynthesis branches. Two R2R3 MYB TFs, StMYB44-1 and StMYB44-2, were highly expressed in white regions under high temperature. In transient assays, StMYB44 represses anthocyanin accumulation in leaves of Nicotiana tabacum and N. benthamiana by directly suppressing the activity of the dihydroflavonol reductase (DFR) promoter. StMYB44-1 showed stronger repressive capacity than StMYB44-2, with both predicted proteins containing the repression-associated EAR motif with some variation. StMYB44-1 conferred repression without a requirement for a basic helix-loop-helix (bHLH) partner, suggesting a different repression mechanism from that of reported anthocyanin repressors. We propose that temperature-induced reduction of anthocyanin accumulation in potato flesh is caused by down-regulation of the activating anthocyanin regulatory complex, by enhancing the expression of flesh-specific StMYB44 and alteration of phenylpropanoid flux.

Introduction

Anthocyanins are a large group of flavonoid compounds, providing a wide range of colours in many plant species (Holton and Cornish, 1995; Petroni and Tonelli, 2011). Anthocyanins are not only essential for plant performance, but also have beneficial effects on human health due to antioxidant, antimutagenic, and anticarcinogenic capacities (He and Giusti, 2010; Pojer ; Khoo ).

The MBW transcriptional regulatory complex, consisting of R2R3 MYB, basic helix–loop–helix (bHLH), and WD40/WDR regulators (Ramsay and Glover, 2005; Hichri ), is central to the regulation of anthocyanin levels. Most R2R3 MYBs involved in the control of anthocyanin biosynthesis are activators that enhance the expression of biosynthetic pathway genes, such as AtMYB75 (or PAP1), AtMYB90 (or PAP2), AtMYB113, and AtMYB114 in Arabidopsis (Stracke ; Gonzalez ), MdMYBA, MdMYB1, MdMYB10, and MdMYB110a in apple (Espley ; Chagné ), SlMYB12 in tomato (Ballester ), and StAN1 in potato (Jung ; Liu ).

In addition to MYB activators, it has also been reported that MYB transcription factors (TFs) can repress flavonoid/anthocyanin accumulation. There are two distinct classes of MYB TFs which negatively regulate anthocyanin accumulation: R3 MYB and R2R3 MYB repressors, which have one or two repeats of the MYB domain region, respectively. The R2R3 MYB repressors belong to the subgroup 4 (SG4) type MYB TFs, and are characterized by two conserved domains in their C-terminus required for their repressive activity, the C1() and the C2/EAR motif (LxLxL or DLNxxP; Kranz ; Kagale and Rozwadowski, 2011). The EAR motif is the most dominant transcriptional repression motif identified in plants. These R2R3 MYB repressors include AtMYB3/4/6 in Arabidopsis (Jin ), MdMYB16/17/111 in apple (Lin-Wang ), FaMYB1 in strawberry (Aharoni ), and PhMYB4 and PhMYB17 in petunia (Albert ). The R3 MYB repressors have been reported in several plant species, including CAPRICE (CPC), TRIPTYCHON (TRY), and AtMYBL2 in Arabidopsis. While AtCPC and AtTRY do not contain any repression motif in the C-terminus, a repression motif (TLLLFR) was found in AtMYBL2 (Dubos ; Zhu ). R3 MYB repressors exhibit passive suppression by competing with MYB activators for binding sites of bHLHs. For example, PhMYBx, a petunia homolog of AtCPC, interacts with PhAN1 and PhJAF13 to inhibit anthocyanin accumulation (Koes ).

The accumulation and stability of anthocyanins are influenced by environmental conditions such as light, temperature, pH, water stress, wounding, and pathogen infection (Chalker-Scott, 1999; Lin-Wang ; Sivankalyani ). Temperature influences anthocyanin metabolism in a diverse range of plant species such as Arabidopsis (Kim ), grape (Mori ; Movahed ; Pastore ), apple (Lin-Wang ; Wang ), blood orange (Eugenio ), blueberry (Liu ), and petunia (Shvarts ). In Arabidopsis seedlings, high ambient temperature decreases both anthocyanin accumulation and expression of several anthocyanin biosynthetic pathway genes (Kim ), whereas low temperatures increase anthocyanin content (Leyva ). In apple, anthocyanin biosynthesis is reduced at high temperatures, concomitant with a reduction in the transcriptional levels of the regulation complex (Lin-Wang ). In tomato, the TFs SlAN2, SlAN1, and SlJAF13 mediate anthocyanin biosynthesis under low temperature, resulting in anthocyanin accumulation (Kiferle ; Qiu ). The temperature-specific effect on anthocyanin biosynthesis, particularly in relation to heat-induced repression, raises concerns in relation to the effects of climate warming and the potential reduction in dietary anthocyanin levels. Thus, elucidation of the influence of environmental temperature on the regulatory mechanism of anthocyanin biosynthesis in plants is necessary if pigmentation and consequent crop quality are to be maintained.

Potato (Solanum tuberosum), currently the most important non-cereal crop world-wide, is sensitive to temperatures (Xiao and Guo, 2010). Recently, there has been increased interest in understanding the regulatory mechanism of anthocyanin production in coloured potatoes as they are a rich source of anthocyanins, in particular acylated derivatives (Fossen and Andersen, 2000). Anthocyanin synthesis in the tuber periderm of potato is controlled by three loci—R, P, and D—which encode the biosynthetic enzymes dihydroflavonol 4-reductase (DFR), flavonoid 3',5'-hydroxylase (F3'5'H), and R2R3 MYB-StAN1, respectively (Jung , 2009; Y. Zhang ). StAN1 is involved in tuber skin and flesh pigmentation (Y. Zhang ; Payyavula ). Furthermore, there are amino acid variations in the C-terminus of StAN1 due to the presence of a repeated 10 amino acid motif, and one copy of this motif appeared optimal for activating anthocyanin accumulation (Liu ). In addition to StAN1, StMYBA1 and StMYB113 were also found to influence anthocyanin biosynthesis in tobacco by transient assays (Liu ). In tuber flesh, one locus (Pf), which is tightly linked to I, is associated with anthocyanin accumulation; however, Pf alone is insufficient for complete tuber pigmentation (De Jong, 1987). Tuber-specific overexpression of the MYB TF gene StMtf1 resulted in elevated amounts of phenylpropanoids, and tubers contained increased levels of anthocyanin (Rommens ). Expression analysis was used to investigate gene expression patterns associated with the accumulation of purple tuber anthocyanins and 27 differentially expressed genes (DEGs) including a single domain MYB TF, which is similar to the soybean MYB73, were identified as differentially expressed in white and purple sectors (Stushnoff ). In potato sprouts, 22 compounds and 119 transcripts including those involved in anthocyanin biosynthesis, hormones, TFs, and signaling-related genes were identified, and regulation of potato pigmentation was revealed by network analysis of the metabolome and transcriptome (Cho ). Besides the activators—StMYB12 (PGSC0003DMT400018841) and an R2R3 MYB (PGSC0003DMT400029235), which were targets of miR828, and MYB-36284 (PGSC0003DMT400036284), which was targeted by TAS4 D4(–)—were suggested to play a negative role in anthocyanin production in the potato tuber (Bonar ).

In this study, we explore the regulatory networks of anthocyanin biosynthesis for reduced purple flesh colouration of the potato cultivar ‘Heimeiren’ (HM) caused by high temperature. We show that heat stress results in a reduction in anthocyanin accumulation of potato flesh around vascular rings and down-regulation of several anthocyanin biosynthesis pathway genes and the transcriptional activator complex (StAN1, StbHLH1). Conversely, high temperature led to an up-regulation of the chlorogenic acid (CGA) and lignin biosynthesis pathway genes. Furthermore, we identify potato flesh-specific repressors StMYB44-1/-2, belonging to SG22 type R2R3 MYBs. When tested, these repressors, were able to reduce the activity of the TF StAN1 on the promoter of the key anthocyanin gene, DFR. Our findings provide new insights into the regulatory mechanism associated with anthocyanin biosynthesis in pigmented potato.

Materials and methods

Plant materials and growth conditions

The tetraploid purple Solanum tuberosum cultivar ‘Heimeiren’ (purple skin and purple flesh) (Fig. 1A) was cultivated in 20 cm diameter pots at 21/17 °C (day/night) under a 16 h cool, white fluorescent light at 200 μmol m−2 s−1 irradiance, and 8 h dark photoperiod. After 1 month of growth, three biological replicates composed of four plants each were transferred to 26/22 °C day/night and three biological replicates composed of four plants each were maintained at 21/17 °C (control) under a 16 h/8 h photoperiod at 200 μmol m−2 s−1 irradiance. Six fresh tubers for each treatment were peeled carefully using a scalpel, white and purple flesh tissues were cut each into small pieces after harvest, and the skin and flesh were then immediately frozen in liquid nitrogen and stored at −80 °C.

Fig. 1.

Biosynthesis of anthocyanins in potato tubers under two different temperature conditions. (a) HM grown at 26/22 °C (i) and 21/17 °C (ii). (b) Levels of total anthocyanin content of skin and flesh (white region and purple region) of HM. The data represent the means ±SE of three biological replicates. Statistical significance was determined by one-way ANOVA; significant differences between means (LSD, P<0.05) are indicated, where letters (a, b, c, etc.) above the bar differ. Tissues denoted by * showed no detectable levels of anthocyanin content. CS and CF represent the skin and flesh under control (21/17 °C); HS, HPF, and HWF represent skin, purple flesh region, and white flesh region under high temperature (26/22 °C).

Determination of anthocyanin content

The total anthocyanin content (TAC) was determined by the pH differential spectrophotometry method described by C. Zhang and Liu . Anthocyanins were extracted from 1 g samples in methanol (0.05% HCl) and absorbance of the anthocyanin extracts was measured at 510 nm and 700 nm in buffers at pH 1.0 and 4.5 by a spectrophotometer (UV-2550, Shimadzu, Japan). Absorbance (Abs) was calculated as Abs=(A510 nm–A700 nm)pH1.0–(A510 nm–A700 nm)pH4.5 with a molar extinction coefficient for cyanidin 3-glucoside of 26 900 (C. Zhang ; Yang et al., 2012). The TAC was calculated using the following equation and expressed as milligrams of cyanidin 3-glucoside equivalents per 100 g FW.

Where e is cyanidin 3-glucoside molar absorbance [26 900 ml (mmol cm)–1], L is the cell path length (1 cm), MW is the molecular weight of anthocyanin (449.2 g mol–1), D is a dilution factor, V is the final volume (ml), and G is the mass of FW (mg).

RNA extraction and qPCR

Total RNA of skin and flesh from HM under high temperature and control were extracted using the PureLink Plant RNA Reagent Kit (Invitrogen, USA) according to the manufacturer’s instructions. The RNA was quantified by using a Nanodrop ND-1000 spectrophotometer (Nanodrop Technologies, USA), and quality was assayed on a 1% agarose gel. Elimination of genomic DNA contamination and first-strand cDNA synthesis were carried out using oligo(dT) according to the manufacturer’s instructions (SuperScript III, Invitrogen, USA). For quantitative real-time PCR (qRT-PCR) analysis, SYBR® Premix Ex Taq™ II (Takara Bio, Inc., Japan) was used according to the manufacturer’s instructions. qPCR conditions were as follows: 30 s at 95 °C, followed by 40 cycles of 5 s at 95 °C, 30 s at 60 °C, followed by 65–95 °C melting curve detection. The qPCR efficiency of each gene was obtained by analyzing the standard curve of a cDNA serial dilution. StEF-1α (AB061263) (Tang ) was used for template normalization. Relative abundance was calculated with the comparative Ct (2−ΔΔ Ct) method. The primers are listed in Supplementary Table S1 at JXB online.

RNA-seq data analysis

RNA sequencing (RNA-seq) was conducted with three biological replicates; for each replicate, the samples of two or three tubers grown from four plants were used. Total RNA from white flesh and purple flesh of cv. HM at 26/22 °C was extracted as described as above. Enrichment of mRNA, fragment interruption, addition of adaptors, size selection, PCR amplification, and next-generation Illumina sequencing were performed by Sagene Biotech Co., Ltd (Guangzhou, China). High-quality clean reads were obtained by trimming the raw reads filtering out contaminants, adaptors, phred scores <20, and uncertain bases N. The cleaned data were aligned to PGSC_DM_v3.4 gene models downloaded from the Solanaceae Genomics Resource at Michigan State University (http://solanaceae.plantbiology.msu.edu/pgsc_download.shtml) by Bowtie2 (v2.2.9). The unmapped reads were assembled by de novo transcriptome assembly using Trinity v2.2.0 to generate unigenes. The numbers of mapped clean reads were counted (by fragment) and subjected to differential expression analysis using the edgeR package (http://www.r-project.org/). Genes with the absolute value of log2FC (log2fold change) not less than 1 and a false discovery rate (FDR) <0.05 were considered as significant DEGs.

DEGs were annotated against the non-redundant database (nr), SwissProt/UniProt Plant Proteins, Kyoto Encyclopedia of Genes and Genomes (KEGG), SwissProt/UniProt Plant Proteins, COG/KOG (Cluster of Orthologous Groups of proteins), and the potato protein database (ftp://ftp.jgi-psf.org/pub/compgen/phytozome/v9.0/Stuberosum) by BLASTX, and the cut-off E-value was 1.0e-5. DEGs were then subjected to enrichment analysis of Gene Ontology (GO) functions and KEGG pathways. The RNA-seq data set is available at the NCBI Sequence Read Archive (SRA) with the accession number: PRJNA528682.

Construction of expression vectors

Six pairs of primers, MYB44-1F/MYB44-1R, MYB44-2F/MYB44-2R, MYB15-1F/MYB15-1R, MYB15-2F/MYB15-2R, MYB15-3F/MYB15-3R, and MYB48F/MYB48R, were designed to amplify the whole coding sequences of the StMYB44-1, StMYB44-2, StMYB15-1, StMYB15-2, StMYB15-3, and StMYB48 genes using cDNA templates synthesized from white flesh of cv. HM. PCR amplification was performed using Platinum Taq DNA Polymerase High Fidelity (Invitrogen, USA). PCR products with EcoRI and HindIII restriction sites were cloned into the multiple cloning sites (MCS) of the binary vector pSAK277 by using an In-Fusion® HD Cloning Kit (Clontech Laboratories, Inc., Mountain View, CA, USA). Primer sequences used for expression vector construction are listed in Supplementary Table S1.

Transient assays of gene function

Transient assays, or dual luciferase assays, were performed in tobacco (Nicotiana benthamiana or N. tabacum) as previously reported (Hellens ; Espley ). The promoters of the potato DFR gene prom-3-StDFR in pGreenII 0800-LUC (Liu ), TFs StAN1-R1 in pSAK277, StbHLH1-2 in pHEX2 (Liu ), StMYB44-1, StMYB44-2, StMYB15-1, StMYB15-2, StMYB15-3, and StMYB48 in pSAK277, and the negative control β-glucuronidase (GUS) in pSAK277 driven by the Cauliflower mosaic virus (CaMV) 35S promoter were used in transient assays. These constructs were individually transformed into Agrobacterium tumefaciens strain GV3101 by the electroporation method. Agrobacterium cultures containing the reporter cassette and the cassettes containing StMYB44-1/StMYB44-2/StMYB15-1/StMYB15-2/StMYB15-3/StMYB48, StAN1-R1, and StbHLH1-2 were mixed with a ratio of 1:3:3:3. Two-week-old N. benthamiana plants with 3–4 leaves, grown under glasshouse conditions, were available for infiltration with Agrobacterium (OD600=0.8–1). Agrobacterium infiltration processes were as described by Hellens ; LUC and REN activities were measured and analyzed by using a VICTOR3 Multilabel Readers (Perkin Elmer, Boston, MA, USA).

In a separate colour assay, Agrobacterium cultivation and infiltration preparation were performed as described above for the transient dual luciferase assay. StMYBs alone or a mixture of equivalent doses of Agrobacterium cultures containing StAN1-R1 and StMYBs were injected into young leaves of 2-week-old seedlings of N. tabacum to test the function of StMYBs. Photos were taken 7 d after infiltration.

BiFC assays

Constructs for investigating in planta interactions were produced in the pSPYNE-35S and pSPYCE-35S vectors by using bimolecular fluorescence complementation (BiFC) assays. The ORFs of StMYB44-1 and StbHLH1-2 were cloned into pSPYNE-35S and pSPYCE-35S to generate StMYB44-1–YFPN and StbHLH1–YFPC. The primers used for plasmid construction are listed in Supplementary Table S1. These constructs were transformed into the Agrobacterium GV3101 strain by electroporation. Nicotiana benthamiana leaves were infiltrated with the mixed Agrobacterium strains. Fluorescence signal was detected 48 h after infiltration using a confocal laser-scanning microscope (FV10-ASW, Olympus, Tokyo, Japan), with excitation at 488 nm and detection with a 500–530 nm band-pass filter for yellow fluorescent protein (YFP).

Statistical analysis

For anthocyanin analysis, data are presented as means (±SE) of three biological replicates of pooled tubers. For dual luciferase promoter activation assays and qPCR analyses, data are presented as means (±SE) of four biological replicates. Statistical significance was determined by one-way ANOVA followed by the least significant difference (LSD) computed at P<0.05.

Accession numbers

Sequence data from this study can be found in the GenBank/European Molecular Biology Laboratory (EMBL) database under the following accession numbers: StMYB44-1 (MK410941; PGSC0003DMG400003316), StMYB44-2 (MK410942; PGSC0003DMG400007994), StMYB15-1 (MK410943; PGSC0003DMG400017223), StMYB15-2 (MK410944; Unigene 025307), StMYB15-3 (MK410945; PGSC0003DMG400013405), and StMYB48 (MK410946; PGSC0003DMG400015536).

Results

High temperature reduces anthocyanin concentration in skin and purple flesh of potato cv. HM

The concentration of anthocyanins in the purple skin and flesh of HM was 95.2±5.13 mg 100 g–1 FW and 49.60±4.06 mg 100 g–1 FW, respectively, at 21/17 °C (control). High temperature reduced the total anthocyanin content in skin and flesh of HM significantly; in particular, the flesh around the vascular ring became white and no anthocyanins were detected. The anthocyanin content of purple flesh was 25.83±2.82 mg 100 g–1 FW (Fig. 1), representing a 47.9% decrease.

RNA-seq identifies differentially expressed phenylpropanoid biosynthetic genes in purple and white flesh

Under higher day and night temperatures (at 26/22 °C), purple fleshed potato showed no anthocyanin accumulation around and outside the vascular ring (termed HWF), but inside the vascular ring there was still anthocyanin present (termed HPF; Fig. 1). In order to ascertain the mechanism underlying this response, RNA-seq analysis was conducted on purple flesh sectors and white flesh sectors of HM grown at 26/22 °C to identify DEGs.

A total of 56 genes, annotated as phenylpropanoid pathway genes, including lignin, chlorogenic acid (CGA), and anthocyanin biosynthetic genes, were found to be differentially expressed between HPF and HWF grown under higher temperature (Tables 1, 2; Fig. 2).

Table 1.

Comparative abundance of transcripts encoding flavonoid and core phenylpropanoid biosynthetic genes between HWF and HPF

Genes	Gene IDa	HWF/HPFb		Genes	Gene IDa	HWF/HPFb
		FPKM	Log2FC			FPKM	Log2FC
PAL-1	Unigene015288	3029/92	5.5	PAL-2	PGSC0003DMG400031457	1004/38	4.8
PAL-3	Unigene015289	940/4	8.4	PAL-4	Unigene023245	465/3	7.3
PAL-5	PGSC0003DMG401021549	358/157	1.3	PAL-6	PGSC0003DMG402021564	338/27	3.7
PAL-7	Unigene023250	332/2	7.7	PAL-8	Unigene023246	300/33	3.3
PAL-9	Unigene024861	293/6	5.8	PAL-10	Unigene030145	102/1	7.3
PAL-11	Unigene023253	83/2	5.0	PAL-12	Unigene024862	73/2	5.4
PAL-13	PGSC0003DMG400031365	46/1	5.4	PAL-14	Unigene023251	31/0	6.8
PAL-15	Unigene018958	28/0	6.0	PAL-16	Unigene031857	16/1	4.0
PAL-17	Unigene028547	11/1	4.5	PAL-18	Unigene023249	9/0	6.7
C4H-1	PGSC0003DMG402030469	160/26	2.7	C4H-2	PGSC0003DMG401030469	107/12	3.2
C4H-3	XLOC_073935	68/12	2.6	C4H-4	Unigene034727	13/1	4.1
C4H-5	Unigene045407	11/0	11.5	C4H-6	PGSC0003DMG400013684	3/0	6.3
4CL-1	Unigene009898	209/6	5.2	4CL-2	PGSC0003DMG400014223	173/2	6.6
4CL-3	XLOC_063106	73/0	7.8	4CL-4	XLOC_080128	44/2	4.5
4CL-5	PGSC0003DMG400028929	43/1	5.3	4CL-6	PGSC0003DMG400003155	22/1	4.2
4CL-7	Unigene020888	10/0	4.9	4CL-8	Unigene042883	9/0	5.2
CHS-1	Unigene016883	25/152	−2.5	CHS-2	PGSC0003DMG400019110	5/18	−1.9
CHS-3	PGSC0003DMG400029620	2/7	−1.5	CHS-4	Unigene000266	7/1	3.4
CHS-5	PGSC0003DMG400001635	2/0	3.8
CHI	Unigene019712	2/17	−3.0
F3'H-1	PGSC0003DMG400023119	3/40	−3.7	F3'H-2	PGSC0003DMG400018064	9/0	5.1
F3H	Unigene022050	1/7	−3.4
F3'5'H-1	PGSC0003DMG400000425	21/86	−2.0	F3'5'H-2	Unigene024395	14/50	−1.7
F3'5'H-3	Unigene024394	4/24	−2.3
DFR-1	Unigene024567	63/163	−1.2	DFR-2	Unigene016587	3/0	4.2
ANS-1	Unigene032308	33/71	−1.0	ANS-2	XLOC_074754	18/41	−1.1
ANS-3	Unigene032309	0/7	−4.0	ANS-4	PGSC0003DMG400022746	2/6	−1.5
ANS-5	PGSC0003DMG400033567	0/1	−5.2	ANS-6	Unigene020805	2/0	6.5
ANS-7	PGSC0003DMG400003880	1/0	5.2
UFGT-1	Unigene029267	9/67	−2.8	UFGT-2	PGSC0003DMG400024344	6/26	−2.0
UFGT-3	Unigene027807	5/19	−1.8

Gene ID beginning with ‘PGSC...’ is the gene ID from the Database of Potato Genome Sequencing Consortium (http://potatogenomics.plantbiology.msu.edu); ‘Unigene…’ is the gene ID from de novo assembly of white flesh and purple flesh under high temperature; while ‘XLOC...’ is novel gene identified by RNA-seq.

The genes that are differentially expressed (FDR <0.05, absolute value of the log2FC≥1, and FPKM>1 applied) in the HWF/HPF library.

Table 2.

Comparative abundance of transcripts encoding CGA and lignin biosynthetic genes between HWF and HPF

Genes	Gene IDa	HWF/HPFb		Genes	Gene IDa	HWF/HPFb
		FPKM	Log2FC			FPKM	Log2FC
C3'H-1	PGSC0003DMG400007178	10/2	2.2	C3'H-2	Unigene005411	5/1	3.1
C3H	PGSC0003DMG400003289	3/0	2.9
HQT	PGSC0003DMG400011189	92/19	2.5
HCT-1	Unigene045297	21/2	3.8	HCT-2	PGSC0003DMG400014152	20/3	3.1
HCT-3	Unigene039930	15/1	3.5	HCT-4	Unigene031173	6/0	7.6
CCR-1	Unigene023847	12/6	1.0	CCR-2	PGSC0003DMG400019825	4/1	2.0
CCoAOMT-1	PGSC0003DMG400006448	98/310	−1.6	CCoAOMT-2	Unigene013813	56/175	−1.6
CCoAOMT-3	XLOC_003173	53/1	5.5	CCoAOMT-4	PGSC0003DMG400006214	48/11	2.2
CCoAOMT-5	Unigene034816	47/0	7.6	CCoAOMT-6	Unigene021201	3/0	5.8
CCoAOMT-7	PGSC0003DMG400025882	2/0	3.5
COMT-1	Unigene009460	11/2	2.8	COMT-2	Unigene000724	10/3	1.8
COMT-3	PGSC0003DMG400011266	8/2	2.2	COMT-4	PGSC0003DMG400000560	6/1	3.0
F5H-1	Unigene000601	5/1	3.0	F5H-2	Unigene036710	3/0	4.0
F5H-3	PGSC0003DMG400003546	2/0	3.3
CAD-1	Unigene007358	92/8	3.7	CAD-2	PGSC0003DMG400009638	13/2	2.8
CAD-3	XLOC_057899	12/2	3.0	CAD-4	Unigene036134	8/4	1.1
CAD-5	Unigene014277	5/1	3.4	CAD-6	Unigene014275	3/0	9.6
CAD-7	Unigene023648	3/0	6.3
PER-1	Unigene007823	0/2	−2.7	PER-2	PGSC0003DMG401025083	149/6	4.8
PER-3	Unigene008746	96/11	3.2	PER-4	Unigene009319	92/22	2.1
PER-5	Unigene032251	67/1	6.3	PER-6	Unigene027540	48/1	6.4
PER-7	PGSC0003DMG400025492	45/17	1.5	PER-8	PGSC0003DMG400022341	38/2	4.4
PER-9	PGSC0003DMG400020252	35/0	8.4	PER-10	Unigene003204	32/0	10.0
PER-11	Unigene047127	31/0	6.3	PER-12	Unigene005680	30/10	1.7
PER-13	Unigene003205	30/0	7.7	PER-14	PGSC0003DMG400027614	29/0	11.6
PER-15	Unigene031419	18/1	4.9	PER-16	Unigene003206	13/0	5.9
PER-17	Unigene042889	13/3	2.1	PER-18	Unigene019365	12/1	4.0
PER-19	Unigene027122	11/2	2.4	PER-20	PGSC0003DMG400024967	11/1	3.8
PER-21	PGSC0003DMG400018031	10/0	8.7	PER-22	PGSC0003DMG400015106	10/0	10.4
PER-23	Unigene003208	9/0	10.4	PER-24	PGSC0003DMG400000511	8/1	3.1
PER-25	Unigene013527	7/2	1.8	PER-26	PGSC0003DMG400005273	7/3	1.1
PER-27	Unigene021138	7/0	5.0	PER-28	PGSC0003DMG400012589	6/1	3.5
PER-29	PGSC0003DMG402025083	6/3	1.2	PER-30	XLOC_083356	6/0	7.6
PER-31	Unigene007331	6/0	5.2	PER-32	PGSC0003DMG400014055	4/0	5.4
PER-33	Unigene031420	2/0	8.8

Gene ID beginning with ‘PGSC...’ is the gene ID from the Database of Potato Genome Sequencing Consortium (http://potatogenomics.plantbiology.msu.edu); ‘Unigene…’ is the gene ID from de novo assembly of white flesh and purple flesh under high temperature; while ‘XLOC...’ is novel gene identified by RNA-seq.

The genes that are differentially expressed (FDR <0.05, absolute value of the log2FC≥1, and FPKM>1 applied) in the HWF/HPF library.

Fig. 2.

Expression of DEGs involved in the phenylpropanoid pathway comparing HWF and HPF under higher temperature. The expression pattern represents the log2FC. The numbers on the bottom of each grid indicate different alleles or gene family members for each gene, shown in Tables 1 and 2. The gene names with both red and green indicate that some versions of the gene are up-regulated and the rest are down-regulated.

DEGs annotated as phenylalanine ammonia-lyase (PAL) (EC 4.3.1.24), trans-cinnamate 4-monooxygenase (C4H) (EC 1.14.13.11), and 4-coumarate-CoA ligase (4CL) (EC 6.2.1.12), which together provide the precursor products for use in the flavonoid pathway, were all more highly expressed in HWF with a 1.1–11.5 log2fold increase (Table 1; Fig. 2). A total of 18 StPAL genes were found to be differentially expressed in the HWF compared with the HPF library, of which 10 StPAL genes had an FPKM >100 with a 1.3–7.7 log2FC increase. The FPKM values of two StPAL genes were >1000, of which one StPAL (Unigene015288) was newly discovered with an FPKM value which reached up to 3029. StPAL genes were located on chromosomes 3, 9, and 10, except StPAL-1 (Unigene015288) and StPAL-10 (Unigene030145) which did not map. Six StC4H genes and eight St4CL genes were also found to be up-regulated in HWF. StC4H genes were located on chromosomes 5 and 6, and St4CL genes were located on chromosomes 3 and 6 (Fig. 3).

Fig. 3.

Chromosomal location of anthocyanin biosynthetic genes and differentially expressed MYB TFs. The pathway gene IDs with red or blue indicate that the versions of the gene are up-regulated or down-regulated in HWF and their names are shown in parentheses. The MYB TFs in black with upward red arrows or downward blue arrows are up-regulated or down-regulated, respectively, in HWF.

In contrast, almost all the flavonoid pathway genes (18 out of 24 genes), including chalcone synthase (CHS) (EC 2.3.1.74), chalcone isomerase (CHI) (EC 5.5.1.6), flavonoid 3'-monooxygenase (F3'H) (EC 1.14.13.21), flavanone 3β-hydroxylase (F3H) (EC:1.14.11.9), flavonoid 3',5'-hydroxylase (F3'5'H) (EC 1.14.13.88), dihydroflavonol 4-reductase (DFR) (EC 1.1.1.219), leucoanthocyanidin dioxygenase/anthocyanidin synthase (LDOX/ANS) (EC 1.14.11.19), and anthocyanidin 3-O-glucosyltransferase (UFGT) (EC:2.4.1.115), were all down-regulated in HWF. Although the expression of two StCHS genes, one StF3'H, one StDFR, and two StANS genes was higher in HWF than in HPF, the FPKM was <9 (Table 1; Fig. 2).

Three genes annotated as StCHS, the first committed enzyme in the flavonoid pathway, were down-regulated in HWF. One StF3H, whose product converts naringenin to dihydroflavonol, one StF3'H gene, and three StF3'5'H genes, which determine the type of dihydroflavonol formed, were also down-regulated in HWF. One gene for StDFR was down-regulated in HWF, as were five StLDOX/ANS gene models, which produce anthocyanidin. Three gene models for StUFGT, the glycosylation step to anthocyanin, were all represented in the down-regulated gene set by 1.8–2.8 log2FC in HWF (Table 1; Figs 2, 3).

Interestingly, a total of 64 DEGs involved in CGA and lignin biosynthesis, including cinnamoyl-CoA reductase (CCR) (EC 1.2.1.44), ferulate 5-hydroxylase (F5H) (EC 1.14.-.-), caffeate O-methyltransferase (COMT) (EC 2.1.1.68), cinnamyl alcohol dehydrogenase (CAD) (EC 1.1.1.195), 3-O-methyltransferase (CCoAOMT) (EC:2.1.1.104), P-coumaroyl quinate/shikimate 3'-hydroxylase (C3'H) (EC 1.14.13.36), 4-coumarate 3-hydroxylase (C3H) (EC:1.14.14.9) and hydroxycinnamoyl-CoA shikimate/quinate hydroxycinnamoyl transferase (HCT/HQT) (EC 2.3.1.133) were up-regulated in HWF (Table 2; Fig. 2).

RNA-seq identifies differentially expressed transcription factors

Comparative transcriptome analysis revealed that 43 MYB TFs (both R2R3 MYBs and R3 MYBs) were differentially expressed in the purple flesh (HPF) versus the white flesh (HWF) library (Supplementary Table S2). Phylogenetic analysis was performed on these differentiated MYB TFs and flavonoid-related MYB sequences from other dicot species. The result showed that the 43 differentially expressed MYB TFs could been classified into 10 of the 25 SGs of Arabidopsis (Dubos ) (Fig. 4; Supplementary Table S2). Of these, 11 MYB TFs were highly expressed in HPF, with 32 MYB TFs being more highly expressed in HWF (Fig. 4). Motifs and domains of these MYB TFs are shown in Supplementary Figs. S1, S2. Of interest was the gene PGSC0003DMG400003316, designated as StMYB44-1, homologous to AtMYB44 (AT5G67300), which showed a 3.26 log2fold increase in HWF with an FPKM value of up to 121, the highest among all the differentially expressed MYB TFs. PGSC0003DMG400007994 was a close BLAST match to AtMYB44, designated as StMYB44-2, and also showed higher expression in HWF. Both the AtMYB44-like MYB TFs were located on chromosome 4 (Fig. 3).

Fig. 4.

Phylogenetic relationships and expression profiles of differentially expressed MYB TFs in HWF and HPF under high temperature. (a) Phylogenetic tree of 43 MYB proteins. The unrooted Neighbor–Joining phylogenetic tree was constructed with MEGA6 by using full-length amino acid sequences with 1000 bootstrap replicates. (b) Expression profiles of differentially expressed MYBs in HWF and HPF libraries. Normalized FPKM values are represented as colours ranging from white (0: little or no expression) to dark blue (1: highest expression within the two libraries). Each column represents a library, and each row represents a gene. The number in the colour bar on the left indicates the range of the highest FPKM value within the two libraries for each gene.

In Arabidopsis, the R2R3 MYB TFs in SG4 have been shown to be active repressors of phenylpropanoid biosynthesis, resulting in less anthocyanin accumulation (Jin ). Three differentially expressed R2R3 MYB TFs (Unigene001595, PGSC0003DMG400030548, and Unigene000965) were in the same clade as SG4, located on chromosomes 1 and 5, and were up-regulated in HPF. This suggests that they are not related to the reduction in anthocyanin biosynthesis in white sectors. In addition, the expression of StAN1 (Unigene028994), the regulator of anthocyanin biosynthesis in potato tubers and leaves (Jung ; Payyavula ; D’Amelia ), was higher in HPF (FPKM=126) than in HWF (FPKM=69), with 0.79 log2fold increase.

The expression of 12 MYB TFs related to AtMYB15 (SG2) was higher in HWF, with three (PGSC0003DMG400017223, Unigene 025307, and PGSC0003DMG400013405) being highly expressed (FPKM value ranging from 63 to 111).

Forty-four bHLH TF genes were differentially expressed in the two tissues, including two novel transcripts and StbHLH35 (PGSC0003DMG400004087), which were in greater abundance in HWF, with an FPKM value >100. The expression of StbHLH1 (PGSC0003DMG400012891), which has been reported to be involved in anthocyanin biosynthesis in potato leaves and tuber (D’Amelia ; Liu ), was higher in HPF than in HWF, while the expression of StJAF13 showed no difference. Surprisingly, five WD40 TF family StAN11 genes, namely TRANSPARENT TESTA GLABRA 1 (PGSC0003DMG400026477, PGSC0003DMG400000561, Unigene034776, Unigene034778, and Unigene034779), were more highly expressed in HWF with an FPKM from 24 to 101, and a 2.6–5.3 log2fold increase (Supplementary Table S3).

A high correlation between qRT-PCR and RNA-seq data sets

To further validate our RNA-seq expression profiling data, 13 differentially expressed structural genes and TFs were selected for qPCR assays using qPCR with gene-specific primers (Supplementary Table S1). Selection was based on previous publications or blast match to Arabidopsis. qPCR analysis indicated that anthocyanin biosynthesis-related genes (StCHS, StF3H, StF3'5'H, StDFR, StANS, StUFGT, StAN1, and StbHLH1) were all up-regulated in HPF, while StPAL, StC4H, St4CL, StMYB44-1, and StMYB44-2 were up-regulated in HWF (Fig. 5). Although the relative expression of the selected genes varied between the RNA-seq data set and qPCR analysis, a high correlation (R2=0.8930) described by a simple linear regression equation, y=1.1362x+0.0639, suggests good consistency between the two analysis methods.

Fig. 5.

Transcriptomic and qPCR analysis of the genes in the phenylpropanoid biosynthetic pathway and transcriptional regulators. FPKM, FPKM obtained by mapping RNA-seq reads to reference genes and de novo assembly; qPCR, real-time PCR. qPCR data are presented as means (±SE) of four technical replicates. PAL, PGSC0003DMG400031365; 4CH, PGSC0003DMG402030469; 4CL, PGSC0003DMG400003155; CHS, PGSC0003DMG400029620; F3H, PGSC0003DMG400003563; F3'5'H, PGSC0003DMG400000425; DFR, PGSC0003DMG400003605; ANS, PGSC0003DMG400022746 and XLOC_074754; UFGT, Unigene029267; AN1, Unigene028994; bHLH1, PGSC0003DMG400012891; MYB44-1, PGSC0003DMG400003316; and MYB44-2, PGSC0003DMG400007994. PF and WF represent the purple flesh region and white flesh region under high temperature (26/22 °C).

Down-regulation of the anthocyanin pathway is via a reduction in the transcriptional activation complex

The effect of high temperature on the potato transcription activators was measured by qPCR. The results showed that the higher temperature significantly reduced the expression of transcription activators and anthocyanin pathway genes (Fig. 6). This suggests that reduced transcription activator abundance results in a reduced ability to trans-active the potato anthocyanin pathway genes, or that repressive TFs are reducing expression of the both transcriptional activator genes and biosynthetic genes.

Fig. 6.

Relative expression of the anthocyanin-related activators StAN1 and StbHLH1, and anthocyanin biosynthetic genes in skin and flesh under two different temperature conditions. Error bars represent the means ±SE of three biological replicates. Statistical significance was determined by one-way ANOVA; significant differences between means (LSD, P<0.05) are indicated, where letters (a, b, c, etc.) above the bar differ.

Transcriptome analysis revealed that two R2R3MYB TF genes, StMYB44-1 and StMYB44-2, were up-regulated in HWF under high temperature. qPCR analysis confirmed that StMYB44-1 and StMYB44-2 were highly expressed in white flesh, but not in purple flesh or skin of HM, under either high temperature or control (Fig. 7). The deduced protein of StMYB44-1 and StMYB44-2 shares 56.8% and 36.2% sequence identity with AtMYB44, respectively. As shown in Fig. 8, the conserved amino acid signature of the bHLH-interacting motif ([DE]Lx2[RK]x3Lx6Lx3R) was not found in StMYB44. In the C-terminal region of the protein, the SG22 group contains the conserved motifs 22.1 (TGLYMSPxSP) and 22.3 (GxFMxVVQEMIxxEVRSYM) (Stracke ), and further analysis showed that another conserved motif, 22.2 (D/EPP/MTxLxLSLP), is present between motifs 22.1 and 22.3 (Zhou ). Motif 22.2 is partially conserved with the C2 (EAR) motif found in subgroup 4 R2R3 MYB repressors.

Fig. 7.

Relative expression of StMYB44-1 and StMYB44-2 in skin and flesh under two different temperature conditions. Error bars represent the means ±SE of three biological replicates. Statistical significance was determined by one-way ANOVA; significant differences between means (LSD, P<0.05) are indicated, where letters (a, b, c, etc.) above the bar differ.

Fig. 8.

Amino acid sequence alignment of StMYB44-1, StMYB44-2, AtMYB44, and other known flavonoid-related repressors. Conserved residues and partial conservation are shown in black. The R2 and R3 domains are indicated with grey and grayish white boxes, respectively. The conserved motifs C1 and C2 in the C-terminus of SG4 are shown in red boxes, and the conserved motifs 22.1 (TGLYMSPxSP), 22.2 (D/EPP/MTxLxLSLP), and 22.3 (GxFMxVVQEMIxxEVRSYM) in the C-terminus of SG22 are shown in green boxes. The following GenBank or Arabidopsis TAIR accession numbers were used: Arabidopsis thaliana AtMYB4 (AT4G38620), AtMYB44 (AT5G67300), Fagaria ananassa FaMYB1 (AF401220), Prunus persica PpMYB18(KT159234), and Solanum lycopersicum SlTHM27(NP_001233975).

As both StMYB44s are highly expressed in white flesh of tubers under high temperature, we cloned the CDS of both genes and performed transient Agrobacterium infiltration assays in N. tabacum and N. benthamiana leaves (Fig. 9; Supplementary Fig. S3). Transient colour assays in tobacco leaves showed that infiltration of either StMYB44-1 or StMYB44-2 induced no anthocyanin production, while the MYB activators of anthocyanin in potato, StAN1-R0, StAN1-R1, and StAN1-R3 (Liu ), resulted in an intense red pigmentation in tobacco leaves after 4 d. In contrast, no pigmentation was observed when StAN1-R0/StAN1-R1/StAN1-R3 were co-infiltrated with StMYB44-1 (Fig. 9a), suggesting StMYB44-1 was able to suppress anthocyanin accumulation. When StAN1-R1 was co-infiltrated with StMYB44-2, anthocyanin accumulation was inhibited compared with StAN1-R1 alone, but repression was not as strong as that caused by StMYB44-1 (Fig. 9a).

Fig. 9.

Functional analysis of the potato StMYB44-1 and StMYB44-2 TFs. (a) Transient colour assay of the StMYB44-1 and StMYB44-2 activities in tobacco leaf. StAN1-R0, StAN1-R1, and StAN1-R3 were infiltrated alone on the left side of a tobacco leaf, and the combination of StAN1-R0, StAN1-R1, and StAN1-R3 with StMYB44-1 was co-infiltrated on the right side of the tobacco leaf (i). StAN1-R1 and combination of StAN1-R1 with StMYB44-2 were infiltrated on the leaf side of tobacco leaf, and empty vector (EV) and StMYB44-2 were infiltrated alone on the right side of the tobacco leaf (ii). Photos were taken 7 d after infiltration. (b) StMYB44-1 and StMYB44-2 prevent the activation of the promoter StDFR-LUC, by co-filtration with StAN-R0, R1, and R3 with StbHLH1 using the dual luciferase assay. The error bars stand for the SE of four biological replicates. Statistical significance was determined by one-way ANOVA; significant differences between means (LSD, P<0.05) are indicated, where letters (a, b, c, etc.) above the bar differ.

To assay the effect of StMYB44-1 on the anthocyanin pathway gene transcription level, the promoter of the potato DFR gene prom-3-StDFR (Liu ) fused to the luciferase reporter was used. Co-infiltration of StAN1-R0/StAN1-R1/StAN1-R3 with StbHLH1 was able to activate the promoters of prom-3-StDFR, whilst co-infiltration of StMYB44-1 or StMYB44-2 inhibited the activity of prom-3-StDFR. The repressive ability of StMYB44-1 was much stronger than that of StMYB44-2 (Fig. 9b). These results suggest that StMYB44-1 and StMYB44-2 are negative regulators of anthocyanin biosynthesis. It is possible that StMYB44-1 has a stronger repressive capacity due to the presence of the EAR motif, while StMYB44-2 has a weaker repressive ability and an altered EAR motif.

In addition, the function of three AtMYB15-like MYB TFs (PGSC0003DMG400017223, Unigene 025307, and PGSC0003DMG400013405) and one AtMYB48-like (PGSC0003DMG400015536), designated as StMYB15-1, StMYB15-2, StMYB15-3, and StMYB48, was investigated. These were highly expressed in white flesh under high temperature (Fig. 4). Transient colour assays in tobacco leaves showed that when StAN1-R1 was co-infiltrated with StMYB15-1, StMYB15-2, StMYB15-3, and StMYB48, respectively, no pigmentation change was observed. This suggests that these MYB TFs, induced under high temperature, are not related to anthocyanin biosynthesis (Supplementary Fig. S3).

StMYB44 is an StbHLH1-independent repressor

According to a previous study, the MYB StAN1 partners with StbHLH1 to regulate anthocyanin biosynthesis in potato tubers (Liu ). BiFC was performed to investigate whether StMYB44 also interacted with StbHLH to limit anthocyanin biosynthesis under high temperature by competing with StAN1. The StMYB44-1–YFPN and StbHLH1–YFPC recombination vectors were constructed and were transformed into N. benthamiana leaves by Agrobacterium-mediated transformation. YFP fluorescence signal was observed in N. benthamiana epidermal cells of the positive control, but not co-expressing StMYB44 and StbHLH1 (Supplementary Fig. S4). The result suggests that StMYB44 does not interact with StbHLH1 to suppress anthocyanin biosynthesis, although the mechanism needs to be further investigated.

Discussion

Temperature is a major environmental factor influencing anthocyanin metabolism in a diverse range of plant species, often resulting in consequences for fruit or vegetable quality and, therefore, financial returns to the grower or retailer. High temperature reduces anthocyanin accumulation in plants by inhibiting the expression of anthocyanin activators and anthocyanin biosynthetic genes and/or enhancing that of repressors (Rowan ; Lin-Wang ; Rehman ). In our study, a transcriptomic analysis by RNA-seq was conducted on purple flesh and white flesh in tubers in response to gene expression at high temperature.

Results showed that under a higher temperature regime, anthocyanin content was significantly reduced around the vascular ring in tubers. This regime resulted in a day and night temperature differential of 4 °C, demonstrating the sensitivity to temperature of anthocyanin production in these tubers. The response was strongest around the phloem, perhaps due to the influence of sugar signals. The expression of anthocyanin activators StAN1 and StbHLH1, and anthocyanin biosynthetic genes StCHS, StCHI, StF3H, StF3'H, StF3'5'H, StDFR, StANS, and StUFGT was reduced in white flesh. Anthocyanins, CGA, lignin, and flavonols share common steps in the phenylpropanoid pathway. CGA is synthesized predominantly through the HQT pathway, with HQT expression directly or indirectly mediating PAL expression (Payyavula ). In our study, the StHQT gene (PGSC0003DMG400011189) is highly expressed in HWF and the expression of 18 StPAL genes was higher in HWF than in HPF. Our study also showed that all the genes related to lignin biosynthesis were highly expressed in HWF. There were no flavonol-related DEGs between HPF and HWF. These results suggest that suppression of the anthocyanin branch was associated with a re-routing of flux into the CGA or lignin biosynthesis branches.

The expression of a MYB73-like gene (a single domain MYB TF) is strongly associated with purple pigmented sectors of tuber flesh in potato cultivars (Stushnoff ). In our study, we found that the nucleotide sequence of Unigene029963 shares 85.4% identity with StMYB73-like (bf_mxflxxxx_0055g04.t3m.scf), but it was not differentially expressed between the HWF and HPF library, with FPKM <8. It is also reported that StMYB12 (PGSC0003DMT400018841), an R2R3 MYB (PGSC0003DMT400029235), and MYB-36284 (PGSC0003DMT400036284) might play a negative role in anthocyanin production in potato tuber (Bonar ). However, in our data set, these genes showed almost no expression in the HWF and HPF libraries.

StMYB44s are negative regulators of anthocyanin biosynthesis

In plants, diverse developmental processes and physiological responses are mediated by MYB proteins. A distinct subfamily of these TFs is essential for activating phenylpropanoid biosynthesis (Sablowski ; Laurent ). The Arabidopsis R2R3 MYB proteins can be divided into 25 SGs according to the conserved R2R3 region and motifs in the C-terminus (Stracke ). Our study identified 43 differentially expressed MYBs between HWF and HPF under high temperature, which were characterized into seven SGs. The expression of 12 MYBs was higher in HPF and that of 21 MYBs was higher in HWF. We further investigated the function of six highly expressed MYBs by using transient assays, and the results showed that StMYB44s (StMYB44-1 and StMYB44-2) could suppress anthocyanin biosynthesis in N. tabacum and N. benthamiana leaves, with StMYB44-1 showing a comparatively stronger repressive ability. qPCR results demonstrated that under high temperature StMYB44-1 and StMYB44-2 were only highly expressed in white flesh compared with that in purple flesh either under high temperature or control, suggesting that StMYB44s are activated in flesh.

In Arabidopsis, AtMYB44, AtMYB70, AtMYB73, and AtMYB77 belong to subfamily SG22. AtMYB44 mediates responses to aphid attack, high salinity, and drought stress, and is regulated by phosphorylation (Jung ; Lü ; Seo ; Persak and Pitzschke, 2014). It is also reported that AtMYB44 positively regulates the enhanced elongation of primary roots induced by N-3-oxo-hexanoyl-homoserine lactone in Arabidopsis (Zhao ). In potato, StMYB44 negatively regulates Pi transport in potato by suppressing StPHO1 expression (Zhou ). To our knowledge, this is the first report of the clade acting as a repressor to suppress anthocyanin biosynthesis in tuber flesh activated by high temperature.

StMYB44s may confer repressive ability via the C-terminal EAR motif (LxLxL) without interacting with StbHLH1

The EAR motifs (LxLxL or DLNxxP) are the most predominant transcriptional repression motifs so far identified in plants (Kagale and Rozwadowski, 2011). Two distinct classes of MYB TFs, R3 MYB and R2R3 MYB repressors, have been found to negatively regulate anthocyanin accumulation. The studied R2R3 MYB repressors, which negatively regulate anthocyanin biosynthesis, belong to the SG4 type (Kranz ). MYB repressors are able to passively repress anthocyanin biosynthesis by coupling with bHLH proteins to compete with MYB activators in the MBW complex, thereby reducing its activation capability. In addition, the R2R3 MYB repressors can also actively suppress the transcription of downstream genes through their repression motif to switch the MBW complex from activation to repression. R3 MYB repressors exhibit passive suppression due to the lack of the repression motif, so compete with R2R3 MYB activators for binding to the bHLH partner. An exception is MYBL2, which has a TLLLFR repression motif which is essential for the repressive function (Matsui ; Albert ).

In the Solanaceae, MYB repressors such as PhMYB27 and PhMYBx have been reported in petunia (Albert ). PhMYB27, an R2R3 MYB repressor, functions as part of the MBW complex and suppresses anthocyanin biosynthesis via its C-terminal EAR motif by targeting both the anthocyanin pathway genes and PhAN1. PhMYBx inhibits the ability of the MBW complex to activate the DFR promoter by binding to PhAN1. It has also been reported that the R3 domain of these MYBs is crucial for the repressive activity. For example, in poplar, site-directed mutations in the bHLH-binding motif result in the loss of repressive activity of PtMYB182 (Yoshida ). In peach, the bHLH-binding motif in the R3 domain is associated with the repressive activity of PpMYB18 (Zhou ). In our study, the conserved amino acid signature of the bHLH-interacting motif ([DE]Lx2[RK]x3Lx6Lx3R) was not found in StMYB44, and BiFC assays also imply that StMYB44 does not interact with StbHLH1, suggesting that the repressive activity is bHLH independent.

It is possible that StMYB44-1 exerted its strong repressive ability through the presence of a complete EAR motif (LxLxLx) in the C-terminal region. The sequence of this motif is less conserved in StMYB44-2, with a substitution of the final L with P (LxLxPx) which could account for its weaker repression. This difference in repressive capacity was also evident in the activation assays. Moreover, the StMYB44s could directly repress the DFR promoter activity without interacting with StbHLH1, due to a lack of a bHLH-binding motif, suggesting a different repressive mechanism, which requires further investigation.

It has been reported that StMYB44 can physically interact with AtWRKY6 and StWRKY6 in vivo (Zhou ). Amongst the many functions of WRKYs, they have been shown to be involved in flavonoid biosynthesis. For example, overexpression of apple WRKY11 promotes the expression of anthocyanin biosynthetic genes and increases the accumulation of flavonoids and anthocyanin in apple calli (Wang ). In our data, 64 genes annotated as WRKYs were differentially expressed in the HWF compared with the HPF library; only three were highly expressed in HPF (FPKM values of 101–645). This possible reduction in the expression of WRKY TFs in the reduced anthocyanin tissues requires further investigation.

In sweet potato, the expression of IbWD40 positively correlates with anthocyanin accumulation in different cultivars, and anthocyanin accumulates after transformation of IbWD40 into Arabidopsis (Dong ). In our study, five WD40-like genes were highly expressed in HWF (a 2.6–6.7 log2fold increase with FPKM values from 24 to 100) and showed less expression in the purple flesh (Supplementary Table S3).

It has been suggested that heat stress reduces anthocyanin biosynthesis by enhancing degradation rather than impairing anthocyanin biosynthesis, such as in grape berries and in plum fruits. One mechanism would be an increased activity of peroxidase enzymes (class III peroxidase) driving degradation at high temperature (Movahed ; Niu ). In our work, 33 genes annotated as peroxidase were highly expressed in HWF by a 1.1–11.6 log2fold increase, with FPKM values from 2 to 149 (Table 2). The three peroxide genes with the highest FPKM values were annotated as peroxidase 3, peroxidase 12-like, and peroxidase N1-like. The function of these genes requires further investigation.

Our results confirm that high temperature in potato causes a reduction in the transcription of anthocyanin structural and regulatory genes as well as elevating the expression of a novel anthocyanin repressor, StMYB44. We propose a model (Fig. 10) where the enhanced transcription of StMYB44 is likely to down-regulate the expression of core genes of the anthocyanin pathway (such as DFR), as well as potentially TFs (such as transcriptional activation complex members StAN1 and StbHLH1) whose role it is to activate these genes. This may impact on the flux of the phenylpropanoid pathway in the tuber, with suppression of the anthocyanin branch associated with a re-routing of flux into the CGA or lignin biosynthesis branches.

Fig. 10.

Schematic showing the StMYB44s activated by high temperatures to suppress anthocyanin accumulation in potato tuber flesh. Dotted lines may indicate other levels of control. MBS indicates MYB-binding sites.

Supplementary data

Supplementary data are available at JXB online.

Fig. S1. Motif compositions and MYB DNA-binding domains of differentially expressed MYB TFs in HWF and HPF under high temperature.

Fig. S2. Sequence information of motifs in 43 differentially expressed MYB TFs.

Fig. S3. Transient colour assay of the StMYB15-1, StMYB15-2, StMYB15-3, and StMYB48 activities in tobacco leaf.

Fig. S4. Bimolecular fluorescence complementation assays showing no interaction between StMYB44-1 and StbHLH1 proteins in epidermal cells of .

Table S1. Sequences of primers used for gene cloning, vector construction, and qRT-PCR.

Table S2. Forty-three differentially expressed MYB TFs in the HPF versus the HWF library.

Table S3. Other differentially expressed TFs in the HPF versus the HWF library.

Click here for additional data file.

Click here for additional data file.

Click here for additional data file.

Click here for additional data file.