Genome-Wide Function, Evolutionary Characterization and Expression Analysis of Sugar Transporter Family Genes in Pear (Pyrus bretschneideri Rehd)

Abstract

The sugar transporter (ST) plays an important role in plant growth, development and fruit quality. In this study, a total of 75 ST genes were identified in the pear (Pyrus bretschneideri Rehd) genome based on systematic analysis. Furthermore, all ST genes identified were grouped into eight subfamilies according to conserved domains and phylogenetic analysis. Analysis of cis-regulatory element sequences of all ST genes identified the MYBCOREATCYCB1 promoter in sucrose transporter (SUT) and monosaccharide transporter (MST) genes of pear, while in grape it is exclusively found in SUT subfamily members, indicating divergent transcriptional regulation in different species. Gene duplication event analysis indicated that whole-genome duplication (WGD) and segmental duplication play key roles in ST gene amplification, followed by tandem duplication. Estimation of positive selection at codon sites of ST paralog pairs indicated that all plastidic glucose translocator (pGlcT) subfamily members have evolved under positive selection. In addition, the evolutionary history of ST gene duplications indicated that the ST genes have experienced significant expansion in the whole ST gene family after the second WGD, especially after apple and pear divergence. According to the global RNA sequencing results of pear fruit development, gene expression profiling showed the expression of 53 STs. Combined with quantitative real-time PCR (qRT-PCR) analysis, two polyol/monosaccharide transporter (PLT) and three tonoplast monosaccharide transporter (tMT) members were identified as candidate genes, which may play important roles in sugar accumulation during pear fruit development and ripening. Identification of highly expressed STs in fruit is important for finding novel genes contributing to enhanced levels of sugar content in pear fruit.

Introduction

In higher land plants, sugars (including sucrose, monosaccharide and polyols) play important roles in plant growth, development and fruit flavor. In addition, sucrose represents the major photosynthetically assimilated carbon transported via the phloem between the source and sink (Van Bel 2003). After release from the phloem, sucrose is directly transported into sink cells, or is hydrolyzed by an extracellular invertase to fructose and glucose (Sherson et al. 2003). To date, it has been established that not only the unloading and the loading of the conducting complex, but also the allocation of sugars into sink cells and sources are controlled by sugar transporters (STs) that mediate the transport of polyols (Noiraud et al. 2001a, Juchaux-Cachau et al. 2007), monosaccharides (Buttner 2007) or sucrose (Kühn 2003, Kühn and Grof 2010). In previous studies, several ST genes have been isolated from flora, fauna and other species, such as two hexose transporters in Juglans regia (Decourteix et al. 2008), seven hexose transporters in Vitis (Fillion et al. 1999, Vignault et al. 2005, Hayes et al. 2007) and a few polyol transporters in Malus domestica (Watari et al. 2004), Prunus cerasus (Gao et al. 2003) and Olea europea (Conde et al. 2007). Among all of the ST gene families, the SUT gene family is a rather small protein family with six genes in pear and nine genes in Arabidopsis; however, the MST family is more diverse. Complete Arabidopsis genome sequencing revealed 53 MST members grouped into seven subfamilies (Buttner 2007). Furthermore, 52 putative ST genes in tomato (Reuscher et al. 2014), 63 putative ST genes in grape (Afoufa-Bastien et al. 2010), and five SUT genes (Aoki et al. 2003) and 65 MST genes in rice (Johnson and Thomas 2007) were identified in previous studies, indicating that STs can be found across the plant kingdom. This result shows that these seven subfamilies are ancient in higher plants (Johnson et al. 2006). In addition, the novel SWEET family that can transport sugars was recently identified. Due to its seven transmembrane domains, the SWEET family has been reported to belong to a different superfamily (Chen et al. 2010). However, because of its relative novelty and these differences, the SWEET transporter family will not be included in this study.

In previous research, some transporter gene expression data in advanced plant and Arabidopsis microarray data (the BAR database: http://bbc.botany.utoronto.ca) analyses have indicated that the expression of STs could be regulated by environmental and developmental factors, such as in yeast (Rolland et al. 2002), as well as VvHT1 (Atanassova et al. 2003, Conde et al. 2006). These results also indicate that the expression of STs might be regulated distinctly at the transcriptional level, but usually involves converging signaling pathways, depending on either metabolic and hormonal signals or developmental and environmental effects. However, in silico analysis of promoters of different genes involved in sugar storage, transport, carbon metabolism and mobilization clearly demonstrate the lack of common sugar-specific cis-regulatory elements (Sheen et al. 1999, Delrot et al. 2000, Rolland et al. 2006). This agrees with the fact that different types of transcription factors (AP2, MYB, bZIP, WRKY, B3 and EIN3) are involved in sugar-linked regulation of gene expression and are required for sugar signaling in plants (Rolland et al. 2006).

Due to the fast development of sequencing techniques, more and more genomes in plants have been sequenced in the past few years, and repeated episodes of small-scale and large-scale gene duplication events have been shown to play important roles during the evolution of gene families. Large-scale gene duplication includes segmental duplications and whole-genome duplications (WGDs) (Van de Peer and Meyer 2005). In pear, evidence has indicated that two WGDs occurred during pear genome evolution, with an ancient WGD event approximately 140 million years (Myr) ago (Fawcett et al. 2009), and a recent WGD event 30–45 Myr ago (Velasco et al. 2010). Small-scale gene duplication events, such as tandem duplications, also play important roles during gene family expansion (Taylor and Raes 2005). The sum of other small-scale duplications and tandem duplications are estimated to contribute duplicates on a scale comparable with large segmental duplications in rice (Yu et al. 2005). The evidence has indicated that tandem and segmental duplications are important during gene family expansion (Sharoni et al. 2011, Zhao et al. 2014).

Pear is one of the most important commercial fruits, and is cultivated in all temperate zone countries of both hemispheres. Sugar content is an important factor affecting fruit quality to a large extent. Therefore, the increase in sugar content is directly related to improvement of fruit quality. Unfortunately, most previous studies focused on sugar content evaluation in different pear varieties (Hudina and Śtampar 2000, Ito et al. 2002, Chen et al. 2007), and only a few sugar-related genes were cloned and studied (Iida et al. 2004). Thus, identification of important genes related to sugar transportation and accumulation, and understanding their functional mechanism will help improve sugar quality in pear. STs play an important role for pear growth and fruit quality; however, there has been limited reporting of the identification of STs, although they have been identified from other ligneous or herbaceous species. Recently, the pear (Pyrus bretschneideri Rehd) genome was sequenced and assembled by the strategy of BAC (bacterial artificial chromosome) by BAC, combined with whole-genome shotgun data, i.e. a total of ×194 genome coverage sequencing. The result showed that the assembled sequence accounts for 97.1% (512 Mb) of the estimated genome size of pear and includes 2,103 scaffolds with N50 at 540.8 kb. The high quality of the assembled sequence and annotation was assessed and confirmed using Sanger-derived BAC sequences along with RNA sequencing (RNA-seq) of different tissues and public protein database alignment (Wu et al. 2013). The high quality of the pear genome is suitable for genome-wide identification and analysis of gene families. The present study is the first to report on the genome-wide identification of ST genes in pear, together with phylogenetic, structural and evolutionary analysis. In addition, RNA-seq databases of pear fruit were used to determine the expression pattern for all ST genes and select key genes affecting sugar content. This study will help to reveal the roles of these ST genes in pear fruit development and sugar quality conformation, as well as provide gene resources for future genetic improvement of pear. The results obtained will also provide a reference in sugar regulation and quality improvement for other related fruit species.

Results

Identification and construction of the phylogenetic tree of the ST gene family in pear

In the present study, a total of 75 open reading frames (ORFs) encoding putative ST proteins were identified in the pear (cultivar: ‘Dangshansuli’) genome using the HMMER profile and BLASTp search for further analysis. On the basic of previous research in Arabidopsis, we renamed the ST of pear as the STP subfamily, which stands for the sugar transporter protein, tMT subfamily stands for the tonoplast monosaccharide transporter subfamily, VGT stands for the vacuolar glucose transporter subfamily, SFP stands for the sugar facilitator protein subfamily, INT stands for the inositol transporter subfamily, pGlcT stands for the plastidic glucose translocator subfamily, PLT stands for the polyol/monosaccharide transporter subfamily and SUT stands for the sucrose transporter subfamily (Table 1). Phylogenetic analysis of the 75 identified nucleotide sequences (Fig. 1) reveals that STs could be classified into eight separate subfamilies with two large subfamilies (the PLT subfamily and STP subfamily) and six small subfamilies based on the Arabidopsis ST sequences (the phylogenetic tree which contained the STs of pear and Arabidopsis is not shown). Among them, 23 and 20 ST genes were annotated as PLT and STP (Table 1). Across the maximum likelihood (ML) tree, most bootstrap values were ≥80, and eight nodes of each subfamily clade had a good bootstrap value. In addition, a consensus Neighbor–Joining (NJ) tree of all 75 ST nucleotide sequences with 1,000 bootstrap replicates revealed a topology that was similar to the ML topology (Supplementary Fig. S1). Combining the NJ topology and ML topology analysis, the phylogenetic tree of the ST gene family in the present study is highly reliable.

Fig. 1

Phylogenetic analysis of pear ST nucleotide sequences. The evolutionary history is inferred using maximum parsimony and maximum likelihood (ML). An ML tree is created by the PHYML program using 100 bootstrap replicates, and the best fitting substitution models for all data are determined with the Akaike information criterion (AIC) using ModelTest 3.06. An unrooted ML tree of pear MST nucleotide sequences is shown. The seven classes are marked by different colors. Different shapes and colors represent different motifs. SH1–15 represents motifs 1–15.

Table 1

Structural and biochemical information of ST members in pear

Gene ID	Chromosome (Mbp)	Strand	Gene model	Genomic (bp)	cDNA (bp)	Protein (amino acids)	E-value (% identity)
PbSTP1	13 (15.1)	−	Pbr003615.1	2,280	1,503	501	4.20E-125
PbSTP2	16 (20.6)	+	Pbr008093.1	2,091	1,503	501	7.00E-127
PbSTP3	16 (20.6)	−	Pbr008083.1	1,996	1,683	561	5.40E-127
PbSTP4	16 (20.6)	−	Pbr008081.1	2,280	1,488	496	3.20E-127
PbSTP5	16 (20.6)	−	Pbr008096.1	1,958	1,461	487	3.50E-128
PbSTP6	16 (20.6)	+	Pbr008080.1	1,958	1,122	374	3.50E-128
PbSTP7	16 (20.6)	+	Pbr008082.1	3,451	1,368	456	2.00E-124
PbSTP8	13 (15.1)	+	Pbr003614.1	3,701	1,311	437	4.70E-125
PbSTP9	16 (20.6)	−	Pbr008092.1	5,589	1,497	499	2.50E-120
PbSTP10	16 (20.6)	+	Pbr008084.1	5,401	1,599	533	8.00E-120
PbSTP11	13 (15.1)	+	Pbr003612.1	5,884	1,734	578	6.80E-124
PbSTP12	2 (22.1)	−	Pbr007315.1	2,947	1,734	578	5.70E-120
PbSTP13	2 (22.1)	+	Pbr033836.1	2,370	1,734	578	2.30E-120
PbSTP14	16 (20.6)	−	Pbr008094.1	3,254	1,728	576	3.40E-121
PbSTP15	15 (43.6)	−	Pbr034294.1	2,936	1,692	564	6.10E-120
PbSTP16	Scaffold 489.0.1 (0.3)	+	Pbr029168.1	2,914	1,740	580	8.00E-120
PbSTP17	2 (22.1)	−	Pbr033859.1	2,370	1,602	534	3.00E-116
PbSTP18	16 (20.6)	+	Pbr008095.1	1,970	1,746	582	1.80E-105
PbSTP19	13 (15.1)	+	Pbr003616.1	3,300	1,314	438	1.70E-120
PbSTP20	15 (43.6)	−	Pbr015498.1	3,027	1,314	438	1.10E-117
PbtMT1	6 (23.1)	+	Pbr015095.1	4,646	1,611	537	1.00E-47
PbtMT2	5 (28.4)	+	Pbr023965.1	4,150	1,611	537	3.90E-52
PbtMT3	5 (28.4)	−	Pbr033292.1	4,710	1,614	538	9.50E-52
PbtMT4	10 (26.2)	+	Pbr032130.1	3,232	1,581	527	4.70E-52
PbtMT5	15 (43.6)	+	Pbr037349.1	2,895	1,560	520	1.50E-51
PbtMT6	15 (43.6)	+	Pbr037348.1	2,891	1,605	535	6.20E-47
PbVGT1	1 (10.7)	+	Pbr013451.1	3,919	1,578	526	1.10E-96
PbVGT2	7 (15.3)	+	Pbr039977.1	4,191	1,578	526	1.40E-96
PbVGT3	11 (30.3)	−	Pbr018950.1	3,574	1,578	526	1.50E-94
PbSFP1	13 (15.1)	−	Pbr014788.1	3,745	1,578	526	1.20E-89
PbSFP2	15 (43.6)	+	Pbr020127.1	4,139	1,581	527	1.30E-97
PbSFP3	12 (22.8)	−	Pbr017113.2	3,303	1,584	528	1.20E-47
PbSFP4	8 (17.1)	+	Pbr006144.1	3,746	1,578	526	8.70E-86
PbSFP5	15 (43.6)	+	Pbr005950.1	3,674	1,584	528	3.60E-72
PbINT1	Scaffold 466.0 (0.4)	−	Pbr028155.1	6,813	1,686	562	1.50E-123
PbINT2	15 (43.6)	+	Pbr002731.1	6,266	1,455	485	8.00E-118
PbINT3	5 (28.4)	−	Pbr000049.1	2,941	1,419	473	4.30E-93
PbINT4	10 (26.2)	+	Pbr038498.1	2,984	1,419	473	1.10E-91
PbINT5	10 (26.2)	−	Pbr017619.1	2,977	1,473	491	1.10E-91
PbINT6	5 (28.4)	−	Pbr000048.1	2,607	1,443	481	1.50E-91
PbpGlcT1	17 (25.3)	−	Pbr010989.1	3,544	1,437	479	3.10E-84
PbpGlcT2	17 (25.3)	+	Pbr039652.1	5,591	1,437	479	8.40E-100
PbpGlcT3	Scaffold 786.0 (0.2)	−	Pbr037980.1	5,659	1,317	439	6.10E-92
PbpGlcT4	9 (22.4)	−	Pbr032591.3	5,974	1,533	511	1.00E-72
PbpGlcT5	Scaffold 477.0 (0.2)	+	Pbr028640.1	4,397	1,527	509	2.90E-74
PbpGlcT6	Scaffold 477.0 (0.2)	−	Pbr028631.1	4,397	1,527	509	2.90E-74
PbPLT1	Scaffold 764.0 (0.2)	+	Pbr037512.1	3,119	1,527	509	4.50E-102
PbPLT2	Scaffold 277.0 (0.2)	+	Pbr018464.1	3,188	1,533	511	8.10E-102
PbPLT3	5 (28.4)	−	Pbr038549.1	2,648	1,533	511	4.60E-104
PbPLT4	5 (28.4)	−	Pbr038546.1	4,268	1,524	508	5.40E-102
PbPLT5	7 (15.3)	−	Pbr018903.1	2,390	1,530	510	3.30E-104
PbPLT6	Scaffold 277.0 (0.5)	+	Pbr018463.1	3,279	1,584	528	4.00E-107
PbPLT7	7 (15.3)	+	Pbr018910.1	2,497	1,584	528	2.00E-102
PbPLT8	7 (15.3)	−	Pbr018908.1	2,497	1,584	528	2.00E-102
PbPLT9	Scaffold 764.0 (0.2)	+	Pbr037515.1	2,177	1,554	518	1.10E-105
PbPLT10	5 (28.4)	−	Pbr038547.1	2,352	1,554	518	4.20E-105
PbPLT11	Scaffold 637.0 (0.3)	+	Pbr034137.1	2,652	1,602	534	5.10E-104
PbPLT12	3 (27.4)	−	Pbr022830.1	2,650	1,554	518	2.20E-103
PbPLT13	3 (27.4)	−	Pbr040466.1	2,609	1,554	518	1.80E-107
PbPLT14	7 (15.3)	−	Pbr018906.1	2,654	1,653	551	1.10E-102
PbPLT15	8 (17.1)	−	Pbr019072.1	2,835	1,464	488	1.40E-53
PbPLT16	Scaffold 637.0 (0.3)	+	Pbr034138.1	2,329	1,509	503	1.70E-100
PbPLT17	Scaffold 637.0 (0.3)	+	Pbr034135.1	2,342	1,566	522	2.10E-100
PbPLT18	8 (17.1)	−	Pbr019074.1	2,293	2,217	739	1.10E-99
PbPLT19	Scaffold 277.0 (0.5)	+	Pbr018465.1	3,429	2,208	736	1.40E-100
PbPLT20	Scaffold 764.0 (0.2)	+	Pbr037511.1	2,509	2,208	736	1.30E-103
PbPLT21	Scaffold 764.0 (0.2)	+	Pbr037514.1	2,844	2,208	736	7.60E-98
PbPLT22	5 (28.4)	−	Pbr038548.1	4,356	2,259	753	2.70E-99
PbPLT23	8 (17.1)	−	Pbr019075.1	2,354	2,172	724	4.80E-75
PbSUT1	Scaffold 1139.0 (0.1)	−	Pbr003266.1	3,314	1,602	534	0
PbSUT2	Scaffold 160.1.13 (0.1)	+	Pbr009635.1	3,167	1,524	508	0
PbSUT3	10 (26.2)	+	Pbr018232.1	3,495	1,524	508	0
PbSUT4	Scaffold 512.0 (0.3)	+	Pbr030158.1	3,286	1,500	500	0
PbSUT5	8 (17.1)	+	Pbr025968.1	7,419	1,551	517	1E-160
PbSUT6	13 (17.1)	+	Pbr039114.1	5,734	1,839	613	1E-140

Conserved motifs and exon–intron organization of ST genes

Because the sugar transporter domain is essential for catalytic activity of ST proteins, the Multiple EM for Motif Elicitation (MEME) motif website search program was used to identify the conserved motifs from 75 ST proteins in pear. In this study, three distinct motifs, motif 1, 2 and 3, which all belong to the ST domain (Fig. 2), were located on the functional domains of all 69 MST proteins, but could not be found in SUT proteins, suggesting that the three distinct motifs may be necessary for MSTs. Interestingly, when we compared the conserved motifs between the SUT gene family and MST gene family, we found that the conserved motifs in the SUT gene family and conserved motifs in the MST gene family were quite different (Fig. 1), even though they had the same functional domains (ST domain); this result might due to functional differences between the SUT gene family (transport sucrose) and the MST gene family (transport monosaccharide). Although all MST proteins had three distinct motifs, the number of motifs in each of the seven subfamilies was different. As shown in Fig. 1, most genes in the two large MST subfamilies had 13 motifs, and in the other subfamilies the number of motifs varied from seven to nine. However, we found that the eight genes of the PLT subfamily did not have motif 10 (a conserved domain SH10), suggesting that SH10 was not critical for PLT member function, despite the conserved sequences and the fact that it was located on the functional domain of the MST gene family. Based on the results of the structural analysis, it was shown that all members with similar structure clustered into the same subfamily. Interestingly, the two large subfamilies of the MST family, the STP and PLT subfamily, shared similar structures, except the two motifs SH9 and SH8 that appear in the PLT subfamily, and SH12 and SH14 that appear in the STP subfamily.

Fig. 2

Sequence logos for three motifs of MST domains using the MEME program. MEME motifs are displayed by stacks of letters at each site. The total height of the stack is the ‘information content’ of that site in the motif in bits. The height of each letter in a stack is the probability of the letter at that site multiplied by the total information content of the stack. The x-axis represents the width of the motif and the y-axis represents the bits of each letter.

To gain insight into the structure of the ST genes, the exon and intron boundaries, which are known to play crucial roles in the evolution of multiple gene families, were analyzed. The results showed that the exon numbers of 75 ST genes ranged from two to 18 (Supplementary Fig. S2). Different subfamilies contained different exon numbers; the fact that PbSFP1 and PbSFP2 genes have 18 exons, and PbPLT17, PbPLT16, PbPLT20, PbPLT21 and PbPLT22 have only two exons, indicates that both exon gain and loss have occurred during the evolution of the ST gene family, which might lead to functional diversity of closely related ST genes. However, it was found that within each subfamily, genes usually have a similar number of exons.

Search for cis-elements involved in the transcriptional regulation of MST genes

In our study, a 2 kb promoter region for each of 62 ST genes was identified. For the other 13 ST genes, the identified sequence was shorter than 2 kb, because of the presence of another gene located <2 kb upstream. Finally, a Plant Cis-acting Regulatory DNA Elements (PLACE) website analysis was applied in this research and it identified 254 different cis-elements that have been classified per ST gene members.

First, a total of 20 common cis-regulatory elements were identified in the 2 kb promoter region, which were highly conserved among the 62 ST analyzed sequences. For some shorter promoters, any of the 20 common cis-elements might be missing. Those common cis-regulatory elements are also responsive to distinct plant hormones, such as cytokinins, and several environmental factors, as well as CO₂, light, dehydration stress and abiotic and biotic stresses (Table 2). At least six of the 20 common cis-acting elements, i.e. GATABOX, EBOXBNNAPA, IBOXCORE, GTGANTG10, TATABOX5 and GT1CONSENSUS, are required for transcriptional regulation by light, consistent with the roles of STs in sugar allocation between sink and source organs. A common cis-regulatory element named MYBCOREATCYCB1 was strongly represented in all ST genes (Table 2), indicating the importance of the MYBCOREATCYCB1 sequence in the promoter of pear MST genes.

Table 2

Common putative cis-elements identified in the promoter sequences of sugar transporters

SITE	Cis-element name	Sequence	Response
S000378	GTGANTG10	GTGA	Pollen
S000454	ARR1AT	NGATT	Cytokinins
S000176	MYBCORE	CNGTTR	Leaf, shoot
S000449	CACTFTPPCA1	CACT	Tetranucleotide
S000039	GATABOX	GATA	Light, leaf, shoot
S000314	RAV1AAT	CAACA	Root, rosette leaves
S000144	EBOXBNNAPA	CANNTG	Light, ABA, seeds
S000198	GT1CONSENSUS	GRWAAW	Light, leaf, shoot
S000494	EECCRCAH1	GANTTNC	CO2-responsive
S000502	MYBCOREATCYCB1	GANNTG	ABA, abiotic stress, cell cycle
S000447	WRKY71OS	TGAC	Giberellin repressor, ABA
S000265	DOFCOREZM	AAAG	C-metabolism, leaf
S000493	CURECORECR	GTAC	Copper; oxygen; hypoxic
S000407	MYCCONSENSUSAT	CANNTG	ABA, leaf, seed, cold
S000199	IBOXCORE	GATAA	Light, leaf, shoot
S000395	INRNTPSADB	YTCANTYY	Light-responsive
S000028	CAATBOX1	CAAT	Seed
S000103	SEF4MOTIFGM7S	RTTTTTR	Seed, storage protein
S000203	TATABOX5	TTATTT	Light
S000415	ACGTATERD1	ACGT	Dehydration stress, dark

The PLACE website is used for promoter sequence analysis.

SITE, a unique accession number is assigned to a motif sequence. The cis-element name, the sequence of the promoter and the signaling pathway are presented.

Another approach targeted unique cis-element sequences present in the promoter region of unique transporter genes, which might indicate expression specificity (Table 3). Finally, 24 unique cis-elements were identified in 17 single promoters. Interestingly, three unique cis-elements (O2F2BE2S1, LBOXLERBCS and ABREZMRAB28) out of 24 specific motifs were present only in PbpGlcT2, and two unique cis-elements were present in each of PbpGlcT1, PbSTP8, PbSTP19, PbtMT4 and PbINT5. The 11 ST genes have only one unique cis-element. This result indicated that a limited number of gene-specific cis-elements were concentrated in the promoter regions of a few transporter genes.

Table 3

Unique cis-elements identified only in the promoter sequence of a single MST gene

SITE	Cis-element name	Sequence	Response	Gene
S000489	SORLREP4AT	CTCCTAATT	Light	PbPLT9
S000233	HDMOTIFPCPR2	CTAATTGTTTA	Pathogen defense	PbSFP4
S000496	WRECSAA01	AAWGTATCSA	Wound	PbSTP9
S000085	INTRONUPPER	MAGGTAAGT	Splice junction	PbSFP1
S000154	SPHCOREZMC1	TCCATGCAT	ABA, seed	PbSFP3
S000367	E2FBNTRNR	GCGGCAAA	Cell cycle	PbSTP19
S000163	O2F2BE2S1	GCCACCTCAT	Storage protein, seed	PbpGlcT2
S000397	TE2F2NTPCNA	ATTCCCGC	Cell cycle, tissue	PbINT5
S000022	ARE1	RGTGACNNNGC	Antioxidant	PbSFP5
S000207	GT2OSPHYA	GCGGTAATT	Leaf, leaf, shoot	PbpGlcT1
S000234	AUXRETGA1GMGH3	TGACGTAA	Auxin	PbINT5
S000446	CARGNCAT	CCWWWWWWWWGG	Gibberellin	PbtMT4
S000301	CONSERVED11NTZMATP1	ACGTATTAAAA	Essential for gene expression	PbSTP19
S000361	HBOXPVCHS15	CCTACCNNNNNNNCTNNNNA	Defense	PbtMT4
S000056	HSELIKENTACIDICPR1	CNNGAANNNTTCNNG	Heat shock, pathogen	PbVGT3
S000302	LREBOXIPCCHS1	AACCTAACCT	Light, leaf, shoot	PbPLT14
S000095	RBCSBOX3PS	ATCATTTTCACT	Leaf, shoot	PbSTP8
S000455	PE2FNTRNR1A	ATTCGCGC	Cell cycle	PbpGlcT1
S000183	SP8BFIBSP8AIB	ACTGTGTA	Root	PbSTP8
S000126	LBOXLERBCS	AAATTAACCAA	Light, leaf, shoot	PbpGlcT2
S000133	ABREZMRAB28	CCACGTGG	ABA, seed, shoot	PbpGlcT2
S000343	AGL3ATCONSENSUS	TTWCYAWWWWTRGWAA	Leaf, flower, shoot	PbVGT1
S000368	CACGCAATGMGH3	CACGCAAT	Auxin	PbSTP6

The PLACE website is used for promoter sequence analysis.

SITE, a unique accession number is assigned to a motif sequence. The cis-element name, the sequence of the promoter and the signaling pathway are presented. Gene indicates the corresponding gene in which the cis-elements are found

Chromosomal localization, gene duplication events and collinearity analyses

The genomic distribution of ST genes on the pear chromosome was investigated in this study. Of 75 pear MST genes, 57 were mapped onto 15 chromosomes, excluding chromosomes 4 and 14, representing unbalanced distribution (Fig. 3). The largest number of ST genes was mapped onto chromosome 16 with 10 ST genes, and only one MST gene was located on each of chromosomes 1, 3, 6, 11 and 12. Additionally, 18 out of 75 ST genes were mapped onto different scaffolds.

Fig. 3

Chromosomal distribution and gene duplications of the ST gene family. The scale on the circle is in Mega bases. The gene IDs on the chromosomes indicate the positions of centromeres; the numbers of each chromosome are shown inside the circle of each bar. The WGD or segmental duplication genes are connected by a black line. In addition, after each gene ID, *0, *1, *2, *3 and *4 indicate singleton duplication, dispersed duplication, proximal duplication, tandem duplication and WGD or segmental duplication, respectively.

During the evolution of a gene family, tandem duplication and WGD/segmental duplication play important roles in generating new members. Therefore, in order to clarify the potential mechanism of evolution of the MST gene family, both tandem duplication and segmental duplication events were investigated in this study. The result of tandem duplication and WGD/segmental duplication analysis indicated that 24 ST genes could be assigned to WGD/segmental duplication blocks and nine ST genes were assigned to tandem duplication (Table 4). In addition, for all cDNA sequences of those genes, the similarity ranged from 66.91% to 99.68%, and all of the segmental gene pairs were found to have counterparts on segmental duplication blocks (Fig. 3). Interestingly, there are two members of the PLT subfamily assigned to the tandem duplication blocks, and none to the segmental duplication blocks in the pear genome (Table 5). All of six tMT subfamily members and four out of all SUT subfamily members were assigned to the segmental duplication blocks. In addition, in order to verify the reliability of WGD/segmental duplication in our study, both end sequences of PbtMT2 and PbtMT4 were analyzed as an example. The result indicated that the genes located on the ends of PbtMT2 and PbtMT4 are a WGD/segmental duplication region (Fig. 4).

Fig. 4

Collinearity relationships of ST genes in pear; 100 kb on each side flanking the genes PbtMT2 and PbtMT4. Segmental duplication pairs are connected with bands. Each black line represents a chromosome segment, and the chromosome number is to the right of the line. ST genes are shown in red, other genes are shown in green, arrowheads represents genes and their transcriptional orientation, and the suffix of the gene name is next to the line.

Table 4

Analysis of the MST gene families in pear

Paralog pairs	Positive selection?	No. of codon sitesa	Likely age of duplication (Myr)	Duplicate type
PbpGlcT3–PbpGlcT2	Yes	6/2**	∼18.11
PbpGlcT4–PbpGlcT1	Yes	19/1**/3*	∼124.65	WGD/segmental
PbpGlcT5–PbpGlcT6	Yes	24/2**/1*	∼1.09
PbVGT1–PbVGT2	No		∼79.63	WGD/segmental
PbSFP2–PbSFP4	No		∼27.87	WGD/segmental
PbSFP4–PbSFP5	No		∼26.08
PbPLT9–PbPLT13	No		∼73.61
PbPLT19– PbPLT21	No		∼11.27
PbPLT20–PbPLT6	No		∼13.54
PbPLT1–PbPLT2	No		∼4.74
PbPLT11–PbPLT18	No		∼3.09
PbPLT16–PbPLT23	No		∼8.20
PbPLT7–PbPLT8	Yes	10/4*	∼0.88
PbtMT5–PbtMT1	No		∼312.31	WGD/segmental
PbtMT2–PbtMT4	No		∼77.58	WGD/segmental
PbtMT3–PbtMT2	No		∼9.13	WGD/segmental
PpINT1–PpINT2	No		∼9.25
PpINT4–PpINT3	No		∼237.88	WGD/segmental
PpINT4–PpINT5	No		∼1.31
PbSTP9–PbSTP10	No		∼3.62
PbSTP9–PbSTP4	No		∼543.67	WGD/segmental
PbSTP2–PbSTP3	No		∼2.42
PbSTP4–PbSTP18	No		∼25.33
PbSTP5–PbSTP6	Yes	4/2*	∼0.83
PbSTP7–PbSTP14	No		∼1.53
PbSTP15–PbSTP16	No		∼11.10
PbSTP17–PbSTP15	No		∼73.88	WGD/segmental
PbSTP12–PbSTP13	No		∼2.21	WGD/segmental
PbSTP17–PbSTP12	Yes	1	∼1.27
PbSUT1–PbSUT4	No		∼7.56	WGD/segmental
PbSUT2–PbSUT3	No		∼8.05	WGD/segmental

^a Values without an asterisk show codon sites with posterior probability (PP) >50%; a single asterisk and double asterisks indicate PP >95% and PP >99%, respectively.

Table 5

Comparative analysis of ST gene families in pear, Arabidopsis and rice

Subfamily	Number of genes
	Pear	Duplicates		Arabidopsis	Duplicates	Rice	Duplicates
		Tandem	Segmental		Tandem		Tandem
STP	20	5	6	14	2	29	14
VGT	3	0	2	3	0	2	0
PLT	23	2	0	6	2	15	10
INT	6	2	2	4	0	3	0
tMT	6	0	6	3	0	6	0
SFP	5	0	2	19	13	6	4
pGlcT	6	0	2	4	0	4	0
SUT	6	0	4	9	1	5	0
Total	75	9	24	62	18	70	28

Estimation of positive selection at codon sites and history duplications of the ST family

Our results showed that all Ka/Ks paralog pairs of ST genes were less than one, indicating that ST genes have evolved mainly under purifying selection (data not shown). Following on from this, ML estimation of the dN/dS substitution rate ratios for paralog pairs in which each sequence came from the same duplication event at nodes in the pear ST nucleotide phylogeny were calculated using the branch-site models method. The result showed that six ST paralog pairs present a large number of codon sites under positive selection (Table 4). We also estimated duplication ages of ST paralog pairs; a pair of them (PbpGlcT4–PbpGlcT1) is near to approximately 140 Myr old, and others ranged in age from approximately 0.83 to approximately 18.11 Myr old. An estimated nucleotide substitution rate of the whole ST family in the pear is 0.05 substitutions per site per Myr. In addition, the ages of twelve segmental ST duplication gene events between approximately 2.21 and approximately 543.67 Myr old were estimated; the paralog pairs at the terminal nodes ranged from approximately 0.83 to approximately 73.61 Myr old in pear, but most of them appear after the second WGD (Fig. 5). In addition, the divergence times of the SUT gene family in pear were also explored; as shown in Fig. 5, estimation of gene duplicate divergence times revealed that the SUT gene subfamily began to diversify at approximately 8.08 Myr ago, and the same results were also found in the PLT, STP and SFP subfamily, and according to the result of divergence time estimation, four of six SUT genes began to diversify after divergence of pear and apple.

Fig. 5

Calibrated phylogenetic tree with gene duplicate divergence time estimates for the ST family. The 75 gene sequences of ST from pear and the phylogenetic tree topology constructed by PHYML. Dates are estimated using the ML method. Calibration points of different subfamilies are shown on the external nodes. Red square symbols at the external nodes indicate segmental duplication gene pairs, and black square symbols at the external nodes indicate paralog gene pairs.

Expression of ST gene family in pear

To investigate the transcript pattern of ST family genes during fruit development, the expression patterns during six developmental stages of pear fruit, from the early to mature stage, were analyzed using the RNA-seq database available from our previous research (Wu et al. 2013). Finally, a hierarchical cluster with the logarithm of average values for the 53 ST family members was generated. As shown in Fig. 6, ST family genes can be divided into two major groups based on their expression profiles. Group A contained 15 MST genes, 13 of them exhibiting preferential expression in some stages or low expression in other stages, and two of them, PbSTP12 and PbPLT1, with low expression in all stages, indicating that those genes may not play important roles in sorbitol accumulation during the whole of pear fruit development. In addition, 38 MST genes belong to group B, which showed high expression in different stages; among them, PbPLT9 and PbPLT22 had the highest expression levels during fruit development. The genes in group B could be further divided into three subgroups, B1, B2 and B3. Subgroup B1 included 17 ST genes, which showed high expression during all developmental stages, but almost all lower than the expression of subgroup B3. Subgroup B2 consisted of eight ST genes that displayed higher expression before July 28, and with low expression during the later stage of fruit development. Subgroup B3 comprised 13 ST genes that displayed higher expression than the other two subgroups in almost all developmental stages, suggesting that these 13 ST genes may play a more important role than other genes in pear fruit development.

Fig. 6

Expression patterns for pear ST genes from the RNA sequence. Heat map showing expression patterns of pear MST family genes in six stages. The fruit samples of the ‘Dangshansuli’ cultivar at April 22 (15 days after full blooming, DAFB), May 13 (36 DAFB), June 27 (81 DAFB), July 28 (110 DAFB), August 30 (145 DAFB) and September 21 (167 DAFB) were collected in 2011, which include the key stages of pear fruit development from early fruit setting to the mature stage. The expression of all ST genes identified in this research is measured by RNA-seq analysis using six stages of pear fruit development. The genes are located on the right and the different stages are indicated at the top of each column. The color scale represents reads per kilobase per million normalized log²-transformed counts, where light red indicates a high level, light green indicates a low level and black indicates a medium level.

Verification of gene expression by qRT-PCR

On the basis of the RNA-seq database, combining the content of sucrose, glucose, fructose and sorbitol analysis, we found that the expression levels of five ST (PbtMT2, PbtMT3, PbtMT4, PbPLT9 and PbPLT22) genes were closely related to sugar accumulation levels during pear fruit development and ripening, and may play more important roles than other genes. In order to verify that these genes were associated with sugar content during pear fruit development, the expression levels of five genes was analyzed by quantitative real-time PCR (qRT-PCR). Finally, the results of qRT-PCR analysis indicated that the expression levels of five genes are closely connected to the change of sugar content during pear fruit development (Fig. 7), one of them differing from the RNA-seq data. On the basis of the RNA-seq data analysis, the expression pattern of the PbtMT4 gene is up-regulated during the whole of pear fruit development. However, according to the results of the qRT-PCR analysis, the expression pattern of the PbtMT4 gene is up-regulated from May 1 to July 29, and down-regulated from July 29 to September 4. With the exception of the PbtMT4 gene, the other four ST genes show a similar trend to the RNA-seq data, indicating that our RNA-seq data are reliable.

Fig. 7

Expression profiles of five ST genes during pear fruit development by qRT-PCR analysis. Five fruit stages of ‘Dangshansuli’ were sampled depending on the status of pear development in 2013, May 1 (31 DAFB), May 27 (57 DAFB), June 23 (84 DAFB), July 29 (120 DAFB) and September 4 (157 DAFB), for qPT-PCR analysis. The relative mRNA levels of individual ST genes are normalized with respect to the housekeeping gene, tubulin, at different stages. The value on the x-axis represents the five stages of pear fruit development. The y-axis represents the relative mRNA (fold). The results were expressed using May 1 as a reference for each gene (relative mRNA level 1). Values represent the experiment among three independent biological repetitions.

Discussion

Identification, and phylogenetic and structural analysis of the ST gene family in pear

The genome sequence and RNA-seq profiles of pear provide a large amount of useful data to explore the functional diversity of the ST gene family from multiple perspectives. In this study, the search for the ST gene family in the pear translated genome has identified 75 STs, and, among them, six ST genes belong to the SUT family, indicating that SUT is a small family within the ST gene family, similar to results found in other plants, such as, four putative SUT genes were identified in grape (Afoufa-Bastien et al. 2010), five in rice (Aoki et al. 2003) and three in tomato (Reuscher et al. 2014). In addition, 69 were putative MSTs, showing that the number of MST members in pear is larger than in Arabidopsis (53 genes) (Buttner 2007), grape (61 genes) (Afoufa-Bastien et al. 2010), rice (65 genes) (Johnson and Thomas 2007) and tomato (49 genes) (Reuscher et al. 2014). We also compared the different subfamily members and gene duplication events among pear, Arabidopsis and rice (Table 5). Interestingly, as in rice, STP and PLT form the largest subfamilies in pear, perhaps due to the repeated regions encompassed by STP and PLT genes. As expected, conserved domains and phylogenetic analysis performed with these MST proteins revealed seven distinct subfamilies (Fig. 1). The same result has been found in grape, rice, Arabidopsis and tomato (Johnson et al. 2006, Afoufa-Bastien et al. 2010, Reuscher et al. 2014), which indicated that the classification of pear MST families was reliable and reasonable. As shown in Fig. 1, different subgroups have similar conserved domains, indicating that under normal circumstances the same subfamily members had the same function due to similar conserved domains. In addition, the results of the determination of exon–intron organization of ST genes have shown that the numbers of exons in 75 ST genes ranged from two to 18 (Supplementary Fig. 2), similar to tomato, in which the exon numbers of all ST genes ranged from one to 18 (Reuscher et al. 2014).

Cis-elements involved in the transcriptional regulation of ST genes

Based on sequences of cis-elements, ST gene promoters contained highly repetitive regions and several common motifs. Among them, motifs such as DOFCOREZM (DNA-binding with one finger) may play an important role not only in terms of response specificity via a combinatory control, but also in the regulation of gene expression in terms of activity levels for the ST. A similar result has been identified for AtSUC2 (Schneidereit et al. 2008), the expression of which in the companion cell is regulated by the close co-operation of binding sites for a putative HD-Zip transcription factor and a DOFCOREZM protein. In previous studies, several transporter gene promoters showed an important concentration of sugar-responsive elements, indicating their transcriptional regulation via sugars. The first to be clearly demonstrated was the transcriptional regulation of VvHT1 by glucose (Atanassova et al. 2003, Conde et al. 2006), confirmed by the fact that the VvHT1 promoter has the largest number of sugar-responsive motifs. In the present study, 20 common cis-regulatory elements were conserved in the promoter regions of ST gene family members (Table 2), the same number as has been identified in the ST gene family of grape (Afoufa-Bastien et al. 2010). Finally, the MYBCOREATCYCB1 promoter, which is required for transcriptional regulation of cyclin B1 during G₁/S to G₂/M transition in the cell cycle (Tréhin et al. 1997), was identified in all ST gene family members. However, it is different from a previous finding that the MYBCOREATCYCB1 sequence was exclusively found in SUC/SUT promoters in grape (Afoufa-Bastien et al. 2010). The MYBCOREATCYCB1 promoter could be found in SUT promoters for sucrose-dependent induction of Cyclin D3 gene expression (Riou-Khamlichi et al. 2000); this result indicated a possible concomitant regulation of some SUT genes in the cell cycle. In addition, it was interesting to find that the MYBCOREATCYCB1 promoter could be found in MST gene family members, which would provide a new research direction for cis-elements in the MST gene family of pear. In addition, this also indicated that different species might have a different transcriptional regulation mechanism in the MST gene family.

The ST gene family arose mainly through WGD/segmental duplication, accompanied by tandem duplications

It had been reported that a primary driving force of new functions in the evolution of genomes and genetic systems is gene duplication (Moore and Purugganan 2003), which is one of the major evolutionary mechanisms leading to functional speciation and diversification (Lynch 2000). As previously reported, the pear genome had undergone two rounds of WGD events, which have a great impact on the amplification of members of a gene family. In the present study, we found 75 ST genes in the pear genome that could be classified into eight subfamilies (Fig. 1) and distributed on 15 chromosomes or some scaffolds (Fig. 3). In addition, the results of the pear tandem duplication and segmental duplication analysis showed that 24 ST genes could be assigned to WGD/segmental duplication blocks and nine ST genes were assigned to tandem duplication. This result indicated that some ST subfamilies have increased rapidly during the course of evolution, and segmental duplication is the main mechanism for expansion of this ST gene family, accompanied by tandem duplications. The same phenomenon has also been found in the WRKY transcription factor family of soybean genes (Yin et al. 2013), where a majority of WRKY genes arose through segmental duplication, accompanied by tandem duplications, but different from Arabidopsis and rice, which experienced more tandem duplicated genes than WGD/segmental duplicated genes in the MST gene family (Johnson and Thomas 2007). For example, PLT and STP subfamilies are greatly expanded, with tandem duplications in rice accounting for 10 and 14 of those subfamily members, respectively (Table 5). In addition, between the vascular and the non-vascular lineages, STP and PLT subfamilies vary significantly in size, suggesting that the expansion of STP and PLT subfamilies could be related to the evolution of vascular plants, and indicating the increased importance of the sugar transporters in vascular plants (Johnson et al. 2006).

Positive selection and history of duplication of ST family genes

For functional proteins, many amino acids are not free to vary under functional constraints and strong structural traits. Thus, in order to detect positive selection of only a few amino acid residues, one should determine the variation in selective pressure among sites (Yang et al. 2000). In previous studies, the extent of positive selection on many protein families has been shown via phylogeny-based analyses of codon substitution (Smith and Eyre-Walker 2002, Weinberger et al. 2010), and has determined that positive selection at some codons is an important driver of protein evolution (Yang and Bielawski 2000). In this study, some of the members of the ST family have positive selection sites (Table 4). The PbpGlcT2–PbpGlcT3, PbpGlcT4–PbpGlcT1 and PbpGlcT5–PbpGlcT6 gene pairs had more positive selection sites than other gene pairs. Interestingly, all six of these genes belong to the pGlcT subfamily, which indicates that the pGlcT subfamily has evolved under positive selection to survive during evolution. In addition, positive selection appears in most duplicate pairs younger than approximately 20 Myr old in pear, except the PbpGlcT4–PbpGlcT1 (∼124.65 Myr old) gene pair, which also indicates that positive selection appears in most duplicate pairs after divergence of pear and apple. However, positive selection in Arabidopsis is seen in any duplicate pairs older than approximately 34 Myr in the MST gene family (Johnson and Thomas 2007).

In a previous study, gene duplicate divergence time estimates revealed that protogenes of each MST subfamily type were present in organisms leading to the land plant lineage at least as far back as the middle Proterozoic (Johnson and Thomas 2007). The result of gene duplicate divergence time of ST members in pear estimated in this study showed that the ST family had comparatively few members at 140 Myr ago, and the expansion of ST genes into large subfamilies continued after the second WGD (Fig. 5, 40 Myr ago). This result is similar to the research in Arabidopsis and rice, where the expansion of large subfamilies continued through the Cenozoic (65–0 Myr) (Johnson and Thomas 2007). Additionally, it was reported that pear and apple diverged from each other at 21.5–5.4 Myr ago (Wu et al. 2013). In this study, most members of the ST gene family were distributed at 20–0 Myr in phylogenetic trees according to the divergence time estimation (Fig. 5), indicating that most ST family members mainly arose after divergence of pear and apple. The comparison among pear, rice and Arabidopsis indicated that the different rounds of genome-wide duplication events and polyploidy led to the ST gene family expansion at inconsistent times among different species (Vision et al. 2000, Bowers et al. 2003). So, we can conclude that the ST family experienced large expansions resulting from the WGDs or multiple segmental duplications and continued to expand through the second WGD, and that most of the ST family began to expand after divergence of pear and apple.

Expression of the ST gene family during pear fruit development

For the ST gene family, we were most interested in those which play important roles during pear fruit development and ripening. In this report, transcript data showed that a total of 53 MST genes were expressed during pear fruit development. This result indicated that these expressed genes are functionally active, with 38 of them being expressed in all six stages during fruit development and ripening (Fig. 6). For PLT subfamily genes, nine of them had one motif missing during evolution, but genes PbPLT20, PbPLT21, PbPLT19 and PbPLT10 are still expressed during development of pear fruit. These results indicated that the loss of this N-terminal domain does not affect gene function.

Sucrose is a major phloem-translocated component and photosynthetic product in most plants. However, some plants synthesize carbohydrates other than sucrose in source leaves and translocate them to sink organs, such as polyols (often named sugar alcohols). Polyols are highly soluble, low molecular weight non-reducing compounds, which means they are suitable as translocating compounds. In many species of Rosaceae, the major phloem component is sorbitol, such as in pear, apple, apricot, peach, cherry and prune. In the phloem of apples, sorbitol comprises about 80% of translocated carbohydrates, and in mature apricot leaves up to 65–75% of translocated carbon is from sorbitol (Kühn et al. 1999, Lalonde et al. 2003). Even though sorbitol is very important in Rosaceae, the mechanism for phloem loading of sorbitol is still unclear (Noiraud et al. 2001b). In a previous study, two PmPLT genes were isolated from Plantago major, which can transport sorbitol (Ramsperger-Gleixner et al. 2004); the results indicated that PmPLT1 and PmPLT2 proteins were localized specifically in companion cells of source leaf phloem and showed their importance in phloem loading of sorbitol. Six PLT genes were detected in Arabidopsis, and have been described as non-specific hexose, pentose and polyol transporters expressed in different tissues (Klepek et al. 2005, Reinders et al. 2005). In addition, two sorbitol transporters (PcSOT1 and PcSOT2) were identified in sour cherry with high expression levels in fruit (Gao et al. 2003), and three MdSOT genes were isolated from apple source leaves (Watari et al. 2004). All these findings indicate that the PLT subfamily might play an important role in long-distance transport of assimilative sorbitol from leaves to fruits in Rosaceae species. In our study, 13 genes of the PLT subfamily were expressed during fruit development (Fig. 6). Out of all of them, the expression levels of PbPLT9 and PbPLT22 were believed to correspond to changes in sorbitol levels in pear fruit, with up-regulation from the early stage to the middle stage, and a slow decrease from the middle stage to near ripening (Fig. 6). In addition, the expression levels of two genes (PbPLT9 and PbPLT22) have been verified by qRT-PCR analysis (Fig. 7), on the basis of which we found that the expression levels of two PLT genes have similar trends with RNA-seq analysis. Based on our previous research, it was found that the content of sorbitol in pear fruit increased from the early stage to the middle stage, and slowly decreased from the middle stage to near ripening (data not shown). Thus, PbPLT9 and PbPLT22 should be considered as important candidate genes for the manipulation of sorbitol transport and accumulation for pear fruit.

The tMT subfamily has also been characterized in rice and Arabidopsis, as AtTMT1 and AtTMT2 were characterized as fructose/H+ or glucose antiporters and localized to the vacuolar membrane (Wormit et al. 2006, Schulz et al. 2011). In this study, all of the tMT subfamily genes were detected as expressed over the whole course of fruit development. Among them, the expression levels of three genes (PbtMT2, PbtMT3 and PbtMT4) were strongly believed to correspond to fructose and glucose levels during fruit development (Fig. 6), and similar results have been found in qRT-PCR analysis (Fig. 7). Fructose and glucose in ‘Dangshansuli’ have low levels at early stages, increasing sharply from the middle stage to near ripening during fruit development, with glucose levels then slowly decreasing during pear fruit ripening. In addition, previous studies have shown that PbtMT4 can be detected as differently expressed proteins and considered as candidate genes that could improve fruit quality during pear fruit development and ripening (Li et al. 2015). So, on the basis of RNA-seq, proteome and qRT-PCR analysis, PbtMT2, PbtMT3 and PbtMT4 are highly reliable candidate genes, which can increase fructose and glucose content during pear fruit development.

Conclusion

A total of 75 ST genes were identified in a genome-wide survey of the pear genome, and can be classified into eight subfamilies, with two large subfamilies (the STP subfamily and PLT subfamily) and six small subfamilies, as supported by the organization of conserved domains and phylogeny. Gene duplication analysis indicated that during expansion of the ST gene family, many subfamilies have evolved, and WGD/segmental duplications have played a more important role during the expansion of the VGT, tMT, pGlcT and SFP subfamilies in pear. Large expansions of the ST family continued through the second WGD, especially after the divergence of pear and apple. In addition, the estimation of positive selection at codon sites showed some amino acid sites belonging to pGlcT members under positive selection. The analysis of promoter sequences indicated that different species have different transcriptional regulation in the MST gene family, such as the MYBCOREATCYCB1 sequence exclusively found in SUC/SUT promoters in grape, but present in all ST gene families in pear. Finally, expression analysis revealed that most ST genes are expressed during fruit development. Among them, two PLT members and three tMT members showed consistent trends with sugar accumulation in fruit. Our results help to clarify the biological function of MST genes in pear development and have a significant influence on our knowledge of woody plant STs.

Materials and Methods

Identification of ST protein in pear

A search for all ST genes in the pear genome was performed using HMMER software (Eddy 1998) for all ST subfamilies. First, an ST domain (PF00083) downloaded from Pfam (http://pfam.sanger.ac.uk/) was used to search the pear protein database by HMMER software. Then, 167 putative ST proteins were identified in the pear genome with an E-value <1E-5. Secondly, all ST proteins were downloaded from the Arabidopsis database (http://www.arabidopsis.org/). Each Arabidopsis ST protein sequence was used as a query sequence in a Base Local Alignment Search Tool (BLAST) search (Altschul et al. 1997), searching against the 167 putative ST protein database of pear to find its best match sequence. Finally, a total of 75 STs protein sequences were identified for further analysis.

Alignment and phylogenetic tree analysis of the ST gene family

Multiple alignments of nucleotide sequences were performed using the Muscle program and Gblocks software (Castresana 2000). For phylogenetic tree analysis, the 75 pear ST nucleotide sequences using maximum parsimony and ML were used. An ML tree was created by the PHYML program (Guindon et al. 2010) using 100 bootstrap replicates, and the best fitting substitution models for all data were determined with the Akaike information criterion (AIC) using ModelTest 3.06 (Posada and Crandall 1998). The model selected was GTR + I + G, gamma distribution with four categories, and an estimated shape parameter of 1.5159. Representations of the calculated trees were constructed using Figtree. In addition, an NJ phylogenetic tree was created by MEGA 6.0 (Tamura et al. 2013).

Conserved motifs, cis-elements and gene structure of ST genes

ST protein sequences were analyzed by the MEME program (http://meme.nbcr.net/meme/cgi-bin/meme.cgi) to confirm the conserved motifs. MEME was employed using the following parameters: maximum number of motifs, 600; number of repetitions, any; optimum width, 15–60; and maximum number of motifs, 15. The results were generated as a txt file. Finally, the iTOL (interactive tree of life) program (Letunic and Bork 2011) (http://itol.embl.de/other_trees.shtml) was used to integrate the phylogenetic and structural tree. Promoter sequences (∼2,000 bp) of ST family genes were obtained from the pear Annotation Project database in our previous study (Wu et al. 2013). The cis-elements of promoters were identified by PLACE Web Signal Scan-PLACE (http://www.dna.affrc.go.jp/PLACE/signalup.html). The gene structure display server 2.0 (GSDS, http://gsds.cbi.pku.edu.cn) was used to illustrate exon and intron organization for individual ST genes by comparison of the cDNAs with their corresponding genomic DNA sequences from the pear genome database website (http://peargenome.njau.edu.cn).

Chromosomal locations and gene duplications of all ST genes

The chromosome Map Tool was used to determine the location of ST genes on pear chromosomes. The duplication pattern for each ST gene was analyzed in this study. In brief, the 42,812 protein-coding genes from the pear genomic database were analyzed using an all-vs-all local BLAST search with E-value <1E-5. The BLAST search outputs were imported into MCScanX software (http://chibba.pgml.uga.edu/mcscan2/) and 42,812 protein-coding genes were classified into various types of duplications including tandem, WGD/segmental, dispersed and proximal under a default criterion. If the pairs of genes were on the two segmental loci and are collinear gene pairs, we considered the gene pairs as segmental duplication gene pairs.

Estimation of positive selection at codon sites

To explore whether positive Darwinian selection drove the evolution of the ST gene family, the non-synonymous/synonymous substitution rate ratios (dN/dS) of all paralog pairs were analyzed using the coding sequence (CDS) of ST gene paralogs. If the dN/dS ratio was >1, it indicates that the gene pairs were under positive selection, or, alternatively, a dN/dS ratio of <1 indicates purifying selection and a dN/dS ratio = 1 indicates neutral evolution. In addition, ML estimation of the dN/dS for paralog pairs in which each sequence came from the same duplication event at nodes in the pear MST nucleotide phylogeny (square symbols in Fig. 5) was calculated using the branch site models method. The branch site test2 of positive selection was used in this study, as described in a previous study (Johnson and Thomas 2007), comparing the null model A1 (model = 2, NS sites = 2 and fix omega = 1) with the alternative model A (model = 2, NS sites = 2 and fix_omega = 0) to find codon sites under probable positive selection. The test of positive selection, with significance cut-offs of 5.41 and 2.71 at the 1% and 5% levels, respectively, was used. Codon sites under probable positive selection and genes with positive selection at the 5% level were identified using the Bayes Empirical Bayes method (Yang et al. 2005).

Estimation of divergence times

The 75 gene sequences of ST from pear and the phylogenetic tree topology constructed by the PHYML (Guindon et al. 2010) program were used. To estimate molecular evolutionary rates and divergence times, a Bayesian method implemented in MCMCtree in PAML (Yang 2007) and the independent rates model was applied to estimate the prior of rates among internal nodes. Each subfamily calibration was included through the time prior: the pGlcT subfamily divergence time was set at 967 (±58) Myr ago, the SFP subfamily divergence time was set at 866 (±53) Myr ago, and the VGT subfamily, STP subfamily, also called the hexose transporter family in grape, INT subfamily, PLT subfamily and tMT subfamily divergence times were set at 478 (±43), 835(±41), 689(±54), 833(±61) and 311(±14) Myr ago, respectively. All seven nodes were constrained with maximum and minimum ages (Johnson and Thomas 2007).

Genome-wide expression analysis of the ST gene family

To investigate the expression of ST gene family members, pear fruit samples of the ‘Dangshansuli’ cultivar on April 22 (15 days after full blooming, DAFB), May 13 (36 DAFB), June 27 (81 DAFB), July 28 (110 DAFB), August 30 (145 DAFB) and September 21 (167 DAFB) were collected in 2011, which included the key stages of pear fruit development from early fruit setting to the mature stage. RNA-seq libraries of six fruit developmental stages were constructed using an Illumina standard mRNA-Seq Prep Kit (TruSeq RNA and DNA Sample Preparation Kits version 2). The RNA-seq data can be downloaded from our center website (http://peargenome.njau.edu.cn/). Expression values of each gene were log transformed, and the cluster analyses were performed using cluster software with the hierarchical cluster method of ‘complete linkage’ and Euclidean distances. Finally, the Treeview program was used to display the results of the cluster analysis.

RNA extraction and first-strand cDNA synthesis

In our research, five fruit stages were sampled depending on the status of pear development in 2013, May 1 (31 DAFB), May 27 (57 DAFB), June 23 (84 DAFB), July 29 (120 DAFB) and September 4 (157 DAFB) for qPT-PCR analysis. Total genomic RNA was extracted from pear fruit according to the CTAB (cetyltrimethyl ammonium bromide) method (Gasic et al. 2004), and then DNase I (Invitrogen) was used to remove genomic DNA contamination. Finally, about 2 µg of total RNA was used for first-strand cDNA synthesis using a ReverTra Ace-aFirst Strand cDNA Synthesis Kit (TOYOBO Biotech Co. Ltd.) according to the manufacturer’s protocol.

Real-time PCR analysis

The primers used for amplifying five ST genes are listed in Supplementary Table S1. In the present study, the LightCycler 480 SYBR GREEN I Master (Roche) was used according to the manufacturer’s protocol. Each reaction mixture contained 10 µl of LightCycler 480 SYBR GREEN I Master, 0.4 µl of each primer, 1 µl of diluted cDNA and 7.4 µl of nuclease-free water. The qRT-PCR was performed on the LightCycler 480 (Roche) and all reactions were run as duplicates in 96-well plates. Each cDNA was analyzed in triplicate, and then the average threshold cycle (Ct) was calculated per sample. The qRT-PCR conditions were as follows: pre-incubation at 95°C for 10 min and then 40 cycles of 94°C for 15 s, 60°C for 30 s, 72°C for 30 s, with a final, extension at 72°C for 3 min, and reading the plate for fluorescence data collection at 60°C. A melting curve was performed from 60 to 95°C in order to check the specificity to the amplified product. Finally, the average threshold cycle (Ct) was calculated per sample; Pyrus tubulin (accession No. AB239681) was used as the internal control, and the relative expression levels were calculated with the 2^–ΔΔCt method descripted by Livak and Schmittgen (2001).

Authors’ contributions

J.L. carried out the experiments and data analysis, and produced a draft of the manuscript. D.Z., L.L. and X.Q. participated in the collinearity analysis, data analysis and preparation of figures. B.B. and S.W. contributed to sample collection and data analysis. S.Z. contributed with consultation. J.W. managed and designed the research and experiments.

Funding

This work was supported by the National Science and Technology Support Funds [2013BAD02B01-2], National Natural Science Foundation of China [31171928]; the Ministry of Education Program for New Century Excellent Talents in University [NCET-13-0864].

Disclosures

The authors have no conflicts of interest declared.

Abbreviations

Abbreviations
 
• AIC

  Akaike information criterion

 
• BAC

  bacterial artificial chromosome

 
• BLAST

  base local alignment search tool

 
• CDS

  coding sequence

 
• DAFB

  days after flower blooming

 
• dN/dS

  non-synonymous/synonymous substitution rate ratios

 
• INT

  inositol transporter

 
• MEME

  Multiple EM for Motif Elicitation

 
• ML

  maximum likelihood

 
• MST

  monosaccharide transporter

 
• Myr

  Million years

 
• NJ

  Neighbor–Joining

 
• ORF

  open reading frame

 
• pGlcT

  plastidic glucose translocator

 
• PLT

  polyol/monosaccharide transporter

 
• qRT-PCR

  quantitative real-time PCR

 
• RNA-seq

  RNA sequencing

 
• SFP

  sugar facilitator transporter

 
• ST

  sugar transporter

 
• STP

  sugar transporter protein

 
• SUT

  sucrose transporter

 
• tMT

  tonoplast monosaccharide transporter

 
• VGT

  vacuolar glucose transporter

 
• WGD

  whole-genome duplication