Genome-wide identification and comparative analysis of GST gene family in apple (Malus domestica) and their expressions under ALA treatment

Abstract

Anthocyanins, a subclass of flavonoids, are synthesized at the cytoplasmic surface of the endoplasmic reticulum (ER), which then accumulate in vacuoles. Plant glutathione S-transferase (GST) genes are involved in anthocyanin transportation. Here, a total of 52, 42, 50, and 29 GST genes were identified from apple, pear, peach, and strawberry, respectively, through a comprehensive genome-wide survey. Based on phylogenetic analyses, the GST proteins of the four crops could be divided into the classes Phi, Tau, DHAR, TCHQD, and Lambda. The structure and chromosomal distribution of apple GST genes were further analyzed. The GST gene family expansion in apple likely occurred through tandem duplications, and purifying selection played a pivotal role in the evolution of GST genes. Synteny analysis showed strong microsynteny between apple and Arabidopsis/strawberry, but no microsynteny was detected between apple/strawberry/Arabidopsis and rice. Aminolevulinic acid (ALA), a key precursor of tetrapyrrole compounds, can significantly improve anthocyanin accumulation in fruits, Using RNA-seq and qRT-PCR analysis, we found that ALA treatment led to the differential expression of GST genes in apples. MdGSTF12 was strongly induced by ALA, suggesting that MdGSTF12 may play a role in ALA-induced anthocyanin accumulation. These results provide a detailed overview of GST genes in four Rosaceae species and indicate that GSTs are involved in ALA-induced anthocyanin accumulation.

Electronic supplementary material

The online version of this article (10.1007/s13205-020-02299-x) contains supplementary material, which is available to authorized users.

Introduction

Glutathione S-transferases (GSTs) are a large, ancient and diverse group of multi-functional proteins that are found in almost all aerobic organisms, from bacteria to human.

Plant GSTs were first identified in maize and have been extensively investigated to date (Frear and Swanson 1970). In most plants, GSTs are cytosolic and exist as monomers, homo- or heterodimers of 23–30 kDa subunits, depending on the type of GST. Each subunit contains an N-terminal domain with a glutathione binding (G-site), and a C-terminal domain that determines second-substrate specificity (C-site) (Edwards et al. 2000). GSTs also have a noncatalytic ligand-binding site (l-site), where hydrophobic substrates and natural products can bind (Dixon et al. 2010). Plant GSTs are divided into 11 classes based on their active site residues, protein sequence similarity and gene organization as follows: Tau, Phi, Theta, Zeta, Lambda, dehydroascorbate reductase (DHAR), tetrachlorohydroquinone dehalogenase (TCHQD), elongation factor 1 (EF1G), microsomal prostaglandin e synthase type 2 (mPGES2), glutathione transferases with two thioredoxin (GST2N), and glutathionyl hydroquinone reductase (GHR) (Sheehan et al. 2001). Among these classes of GSTs, Tau, Phi, Lambda, and TCHQD are plant specific.

GSTs mainly catalyze the conjugation of reduced tripeptide (γ-Glu-Cys-Gly) glutathione (GSH) to various reactive electrophiles. In plants, GST proteins are essential for proper development and physiological function. They play an important role in stress response (Roxas et al. 2000; Csiszár et al. 2014), and are involved in insecticide and herbicide degradation (Hayes et al. 2005). GSTs are responsive to plant hormones, such as auxins, ethylene, salicylic acid, abscisic acid and jasmonic acid (Moons 2005; Marrs 1996). Moreover, GSTs can also regulate the transport and metabolism of secondary compounds such as anthocyanin, flavonoids, and porphyrins (Dixon et al. 2010; Dixon and Edwards 2010). In recent years, genome-wide analyses have revealed that there are 53 GST genes in Arabidopsis (Sappl et al. 2009), 79 in rice (Jain et al. 2010; Soranzo et al. 2004), 25 in soybean, 42 in maize (McGonigle et al. 2000), 90 in tomato (Islam et al. 2017), 85 in pepper (Islam et al. 2019), 32 in pumpkin (Kayum et al. 2018), 23 in sweet orange (Licciardello et al. 2014), 90 in potato (Islam et al. 2018), and 20 in Dracaena cambodiana (Zhu et al. 2016).To date, despite the high number of whole genome sequences available, the structure of the GST gene family in Rosaceae remains unclear.

Anthocyanins belong to a class of flavonoids that is widely present in higher plants. They are crucial in attracting seed dispersers (Vogt et al. 1994), protecting against UV radiation (Sarma and Sharma 1999), and play a role in response to biotic and abiotic stresses (Chalker-Scott 1999). For many species, anthocyanins are a marker of fruit ripening, and fruit with higher anthocyanin content have higher market value and greater health benefits (Boyer and Liu 2004). Anthocyanins are synthesized in the cytosol and transported to the vacuole for storage. Compared with the biosynthetic process, the understanding of the transportation mechanism of anthocyanins is limited. Recently, three distinct models have been proposed for anthocyanin transport to the vacuole: vesicle trafficking, membrane transporters, and a GST mediated pathway, and these processes act in an integrative manner (Zhao 2015). GSTs transport anthocyanin from the endoplasmic reticulum (ER) to the vacuole (Dixon et al. 2010; Zhu et al. 2016; Sun et al. 2012; Alfenito et al. 1998; Mueller et al. 2000), and this function was first demonstrated in the maize bronze-2 (bz2, GST-like protein) mutant, which has a brown to bronze pigmentation phenotype because of the disruption of anthocyanin uptake into vacuoles (Marrs et al. 1995). Subsequently, a number of genes with similar functions have been confirmed in different plant species, such as Arabidopsis (TT19) (Sun et al. 2012), petunia (AN9) (Alfenito et al. 1998; Mueller et al. 2000), grape (VviGST1 and VviGST4) (Conn et al. 2008), carnation (Fl3) (Larsen et al. 2003), Litchi (LcGST4) (Hu et al. 2016) and cyclamen (CkmGST3) (Kitamura et al. 2012). GSTs likely act as carrier proteins rather than catalyzing the anthocyanin-GSH conjugation (Dixon and Edwards 2010; Sun et al. 2012; Zhao and Dixon 2010).

Aminolevulinic acid (ALA) is an essential biosynthetic precursor of tetrapyrrole compounds, and is widely applied to plants in agricultural production to improve photosynthesis (Youssef and Awad 2008; Memon et al. 2009), plant growth (Nguyen et al. 2016), and resistance to stress (Korkmaz et al. 2010; Akram and Ashraf 2013; Nguyen et al. 2016). Treatment with ALA also leads to increased anthocyanin accumulation in several fruit crops, such as apple (Feng et al. 2016; Zheng et al. 2017), pear (Changcheng et al. 2012), peach (Guo et al. 2013) and litchi (Feng et al. 2015). However, the biological mechanisms underlying the increase in anthocyanin accumulation after ALA treatment is poorly understood. Because of the known role of GSTs in regulating anthocyanin accumulation (Marrs et al. 1995; Mueller et al. 2000; Pérez-Díaz et al. 2016; Sun et al. 2012), a genome-wide analysis of GST genes was performed in four Rosaceae species, with a focus on apple. Additional information on physiochemical properties, subcellular localization, chromosomal location, phylogenetic relationship, and gene duplication events for each member has also been further analyzed. Moreover, the expression of apple GSTs after ALA treatment was analyzed. Our results provide foundational work for studying ALA-induced anthocyanin accumulation in apple.

Materials and methods

Plant materials and chemical treatments

The calli of ‘Fuji’ apple (Malus × domestica Borkh.) was cultured on Murashige and Skoog (MS) medium containing 1 mg/L 6-BA and 1 mg/L 2,4-D, and sub-cultured for 20 day intervals at 23 °C in the dark (Akram and Ashraf 2013). To study the expression profile of apple GSTs under ALA treatment, the control calli were treated with deionized water, while the calli was soaked in 50 mg/L ALA solution for 3 h in a shaking incubator. The calli were then retransferred to new MS medium culture dishes, and the dishes were cultured under continuous light at 22 °C for 72 h.

Identification of GSTs in apple and other species

The protein sequence, nucleotide sequence, and gene annotation files were obtained from the following online databases. Chinese white pear (Pyrus bretschneideri) datasets were downloaded from Pear Genome Project database (V1, https://peargenome.njau.edu.cn). Multiple datasets of Arabidopsis thaliana (TAIR10) and Prunus persica (Prupe1.0) were downloaded from the EnsemblPlants database (https://plants.ensembl.org/). Multiple datasets of Fragaria vesca (V1.1) and Oryza sativa (v7.0) were downloaded from JGI database (https://genome.jgi.doe.gov/portal/). Apple (Malus domestica) datasets were downloaded from the GDR database (https://www.rosaceae.org/, V3.0).

To identify GSTs in the apple, pear, strawberry, and peach genomes, BlastP and HMM methods were used. Arabidopsis and rice GST protein sequences were used as query against the protein databases of apple, pear, strawberry, and peach. The GST domain files (GST_N, PF00043; GST_C, PF02798) were downloaded from Pfam (https://pfam.xfam.org), and used to search against the four protein databases. The two search results were combined, redundant genes were removed, and the resulting putative GSTs were searched in the SMART database (https://smart.embl-heidelberg.de) to confirm the presence of common GST protein domains.

Bioinformatics analysis

Length of the protein sequence, genomic position, isoelectric point (pI), subcellular localization, protein molecular weight (MW), and subcellular location were analyzed using ProtParam and CELLO (https://web.expasy.org/protparam/ and https://cello.life.nctu.edu.tw/). Domain analysis of the protein sequences, protein sequence alignment, phylogenetic analysis, gene distribution, and gene structure display were done through MEGA, MapGene2Chro (https://mg2c.iask.in/mg2c_v2.0/), and GSDS (https://gsds.cbi.pku.edu.cn/index.php), respectively. MCScanX was used to analyze duplicate pairs and types, and the results were visualized using Circos (Wang et al. 2012). The number of non-synonymous substitutions per non-synonymous site (Ka) and the number of synonymous substitutions per synonymous site (Ks) for duplicated MdGST homologues was calculate using ParaAT and KaKs_Calculator (Zhang et al. 2012).

Measurement of anthocyanin

The pigments of apple calli were extracted from two grams (fresh weight) of samples using 1% (v/v) HCl-methanol at room temperature in the dark for 24 h. Pigments were measured using the absorbance at 530 nm of the extract solution with a spectrophotometer. The nmol of cyanidin-3-galactoside in 1 g of fresh sample was used to present the total anthocyanin content with a molar extinction coefficient of 3.43 × 10⁴ (Ubi et al. 2006). Three replicates were performed to calculate the mean values.

RNA isolation and RT-qPCR

The calli were harvested at 24, 48, and 72 h after ALA treatment to analyze gene expression patterns. Total RNA was extracted using the RNA simply Total RNA Kit (Tiangen, and Beijing, China) following the manufacturer’s protocol. The Revert Aid™ First-Strand cDNA Synthesis Kit (Transgen Biotech, China) was used to synthesize first-strand cDNA. cDNA products were diluted 30-fold in double-distilled H₂O for RT-qPCR. RT-qPCR reactions were conducted using diluted cDNA, forward and reverse primers, 2 × SYBER GREEN Master Mix and sterile water, for a final volume of 20 μl. The housekeeping gene MdUBQ (CTCCGTGGTGGTTTTTAAGT, GGAGGCAGAAACAGTACCAT) was used as an internal control. Primers for quantitative RT-PCR are listed in Supplemental Table 1. PCR was conducted using the following cycling conditions: 95 °C for 10 min, then 35 cycles at 95 °C for 15 s, followed by 62 °C for 1 min, and stored at 4 °C. The 2^−ΔΔCt method was used to calculate the expression levels for each sample (Livak and Schmittgen 2001).

Results

Identification and characterization of the GST gene family

To identify GST genes in apple, Chinese white pear, peach and strawberry, BlastP and HMM searches were conducted using protein sequence and domains from Arabidopsis and rice GSTs. The rice GST genes were selected based on past studies; however, the obsolete entries in the TIGR database, OsGSTU3 (LOC_Os10g38501) and OsGSTU4 (LOC_Os10g38495), were removed (Jain et al. 2010). Moreover, the candidate GST protein sequences from apple, pear, peach, and strawberry were verified using SMART. The combined Blastp and HMMER search resulted in all candidate proteins with complete GST-N- and GST-C-terminal domains, suggesting that our search was effective in identifying GSTs in these species. Finally, 52, 42, 50, 29, 77 and 53 genes were identified as possible members of the apple, pear, peach, strawberry, rice and Arabidopsis GST superfamily, respectively, and were used for further analysis (Table 1, Supplemental Tables 2–6). Remarkably, there were significantly more GST genes in rice compared to the other species included in this study, followed by Arabidopsis. Peach and apple had a similar number of GST genes, whereas strawberry had the fewest GSTs in the four studied species. We did not find an association between the number of GSTs and genome size. For example, the peach genome contains more GSTs compared to strawberry, even though the strawberry genome (240 Mb) is larger than the peach genome (224.6 Mb).

Table 1.

Detail information of apple GSTs

Gene subfamilies	Gene name	Gene identifier	Size (AA)	Genomics position	Theoretical pI	Mw (kDa)	Predicted Subcellular localization
PHI	MdGSTF1	MD17G1133800	214	Chr17:11942168–11946015	5.55	24,814.63	Cytoplasmic
	MdGSTF2	MD03G1282900	213	Chr03:36322193–36324159	7.06	23,854.87	Cytoplasmic
	MdGSTF3	MD03G1282700	232	Chr03:36305324–36,312,277	5.99	26,088.18	Cytoplasmic
	MdGSTF4	MD09G1147400	122	Chr09:11549037–11551449	5.09	14,042.22	Cytoplasmic
	MdGSTF5	MD17G1134200	205	Chr17:11987456–11988923	5.97	23,718.27	Cytoplasmic
	MdGSTF6	MD11G1303600	217	Chr11:41895006–41896381	5.03	24,517.14	Cytoplasmic
	MdGSTF7	MD03G1283000	217	Chr03:36329385–36330432	5.28	24,738.29	Cytoplasmic/ER
	MdGSTF8	MD09G1147100	204	Chr09:11539783–11,541,140	5.98	23,565.13	Cytoplasmic
	MdGSTF9	MD09G1147300	225	Chr09:11,544,507–11546275	5.48	25,990.84	Cytoplasmic
	MdGSTF10	MD17G1133600	214	Chr17:11930350–11932418	5.73	24,716.38	Cytoplasmic
	MdGSTF11	MD08G1006600	218	Chr08:537193–538944	6.84	24,561.02	Cytoplasmic
	MdGSTF12	MD17G1272100	215	Chr17:33407349–33409236	5.67	24,490.29	Cytoplasmic
	MdGSTF13	MD06G1012200	217	Chr06:1531953–1534667	6.13	24,284.89	Cytoplasmic
	MdGSTF14	MD17G1134300	278	Chr17:11997619–11999641	6.37	32,125.13	Cytoplasmic
TAU	MdGSTU1	MD16G1011600	225	Chr16:894058–894918	6.11	25,831.67	Cytoplasmic
	MdGSTU2	MD10G1172200	226	Chr10:26459270–26460291	5.93	25,946.76	Cytoplasmic
	MdGSTU3	MD05G1184400	223	Chr05:31287827–31289215	5.92	25,308.05	Cytoplasmic
	MdGSTU4	MD10G1172300	223	Chr10:26469318–26470297	6.26	25,813.95	Cytoplasmic
	MdGSTU5	MD05G1184300	221	Chr05:31259973–31261200	5.81	25,603.52	Cytoplasmic
	MdGSTU6	MD10G1172100	219	Chr10:26448956–26449772	5.71	25,337.27	Cytoplasmic
	MdGSTU7	MD05G1184200	221	Chr05:31257878–31258927	5.47	25,196.00	Cytoplasmic
	MdGSTU8	MD04G1139500	222	Chr04:22765396–22774986	5.43	25,524.45	Cytoplasmic
	MdGSTU9	MD00G1136300	223	Chr00:29541189–29539434	5.49	25,283.17	Cytoplasmic
	MdGSTU10	MD14G1232200	216	Chr14:31238440–31239843	5.78	24,965.81	Cytoplasmic
	MdGSTU11	MD16G1081100	231	Chr16:5706295–5707923	7.68	25,368.41	Cytoplasmic
	MdGSTU12	MD13G1227600	225	Chr13:22267032–22268321	7.02	25,882.17	Cytoplasmic
	MdGSTU13	MD04G1139400	221	Chr04:22756386–22758195	5.31	25,410.2	Cytoplasmic
	MdGSTU14	MD04G1139600	221	Chr04:22779085–22780589	5.21	25,496.3	Cytoplasmic
	MdGSTU15	MD02G1236000	219	Chr02:28232819–28234618	5.32	25,084.04	Cytoplasmic
	MdGSTU16	MD02G1236100	219	Chr02:28241902–28242900	5.22	25,094.06	Cytoplasmic
	MdGSTU17	MD13G1081900	234	Chr13:5750247–5751759	5.67	25,621.66	Cytoplasmic
	MdGSTU18	MD13G1082000	231	Chr13:5752474–5755270	5.86	25,423.48	Cytoplasmic
	MdGSTU19	MD05G1209700	306	Chr05:34122009–34124732	5.78	35,439.18	Cytoplasmic
	MdGSTU20	MD10G1197100	219	Chr10:29507624–29509326	6.03	25,392.42	Cytoplasmic
	MdGSTU21	MD17G1271700	220	Chr17:33239918–33240984	5.41	24,980.12	Cytoplasmic
	MdGSTU22	MD10G1196900	221	Chr10:29492441–29499355	6.15	25,880.07	Cytoplasmic
	MdGSTU23	MD17G1286500	176	Chr17:34609141–34610079	5.03	19,876.07	Cytoplasmic
	MdGSTU24	MD10G1196300	219	Chr10:29448059–29449491	5.41	25,510.73	Cytoplasmic
	MdGSTU25	MD05G1210700	219	Chr05:34204834–34206719	5.89	25,484.56	Cytoplasmic
	MdGSTU26	MD05G1211000	219	Chr05:34224366–34225801	5.18	25,524.41	Cytoplasmic
	MdGSTU27	MD09G1279700	220	Chr09:35654801–35655805	5.4	25,031.99	Cytoplasmic
	MdGSTU28	MD04G1033800	225	Chr04:3807094–3807959	5.82	26,262.47	Cytoplasmic
	MdGSTU29	MD05G1210600	218	Chr05:34198838–34200200	5.89	25,397.54	Cytoplasmic
	MdGSTU30	MD05G1210100	219	Chr05:34154934–34156291	5.53	25,490.49	Cytoplasmic
	MdGSTU31	MD05G1210200	219	Chr05:34165150–34166225	5.59	25,307.26	Cytoplasmic
	MdGSTU32	MD05G1210400	220	Chr05:34178564–34182130	6.02	25,614.82	Cytoplasmic
	MdGSTU33	MD05G1210000	219	Chr05:34146212–34147596	5.38	25,357.42	Cytoplasmic
DHAR	MdDHAR1	MD15G1438400	420	Chr15:53802878–53806470	6.58	47,450.67	Cytoplasmic
	MdDHAR2	MD08G1244100	419	Chr08:30929758–30932924	6.58	47,568.86	Cytoplasmic
	MdDHAR3	MD15G1403200	422	Chr15:50419135–50422795	6.11	48,043.31	Cytoplasmic
TCHQD	MdTCHQD1	MD15G1133600	267	Chr15:9698962–9701045	9.63	31,372.40	Mitochondrial
Lambda	MdGSTL2	MD12G1129400	236	Chr12:20460252–20463131	5.85	27,398.18	Cytoplasmic

The length of protein sequence, isoelectric point (pI), subcellular localization genomics position, and the protein molecular weight (MW) were analyzed. In apple, the smallest GST protein was MdGSTF4 (MD09G1147400) with 122 amino acids (aa), and MdDHAR3 (MD15G1403200) was the largest with 422 aa. The range of molecular weights and pI of the predicted apple GST proteins ranged from 14.04 to 48.04 kDa, and from 5.03 to 9.63, respectively. Most GSTs were predicted to localize in the cytoplasm, except MdCHQD1 (MD15G1133600), which was localized in the mitochondria. The properties of predicted GSTs from other species are listed in Supplemental Tables 2–4.

Classification and phylogenetic analysis of GST family genes.

To examine phylogenetic relationships of GST genes, the putative GST protein sequences from apple, pear, peach, and strawberry were aligned with Arabidopsis (Fig. 1 and Supplemental Figs. 1–3). All studied Rosaceae GST members were named according their protein homology with Arabidopsis. Phylogenetic analysis revealed that apple and pear contained the same five GST classes: Phi, Tau, DHAR, TCHQD, and Lambda. Peach contained the classes Phi, Tau, TCHQD, Lambda, and Theta, whereas strawberry only had four classes: Phi, Tau, TCHQD and Lambda. The Rosaceae species included in this analysis contained four GST classes specific to plants: Phi, Tau, TCHQD, and Lambda. Interestingly, the Tau subfamily was the largest group and accounted for more than half of the total number of GSTs in the studied species, while subfamily of TCHQD, Lambda or Theta were relatively small. For example, among the 52 MdGSTs in apple, 33 proteins (MdGSTU1–MdGSTU33) were divided into Tau, whereas only one protein (MdTCHQD1) was classified as belonging to TCHQD. These results suggest that the GST Tau subfamily members expanded more rapidly in plants, and likely play varied roles. In contrast, compared with Arabidopsis and rice, some Rosaceae species lost GST subfamilies like Zeta and EF1G, suggesting that Zeta and EF1G subfamilies were lost during evolution or after diverged in the last common ancestor. In general, most GSTs belonging to the four species included in the study were distinct from Arabidopsis homologues in the phylogenetic analysis, implying that most gene duplication events happened after their divergence from Arabidopsis.

Fig. 1.

Unrooted phylogenetic tree representing relationships among GST protein of apple and Arabidopsis. The neighbor-joining method in MEGA 7.0 was used to construct the unrooted phylogenetic tree, with 1000 bootstrap replicates. Different subfamilies were marked as different colors

Structure analysis of GST transcripts

Intron gain or loss can alter gene structure, playing a vital role during the evolution of gene families (Xu et al. 2012). To characterize the structural diversity of the GST family, the four Rosaceae exon–intron organization of GSTs was analyzed. Exon–intron numbers and positions varied in the four Rosaceae fruits genome (Fig. 2 and Supplemental Figs. 4–6). Overall, a similar number of exons and introns was found in most genes that were clustered in the same group, suggesting that gene structure was conserved within the same group. The number of introns of GSTs varied from one to nine. Almost all Tau subfamily GSTs in the four Rosaceae species contained one intron, some had two introns like MdGSTU19 and PpGSTU10, suggesting that the Tau subfamily was conserved during evolution. Phi subfamily GSTs generally had two introns in apple, pear, and peach but not strawberry. It may indicated the functional diversity of the strawberry Tau subfamily during the evolutionary process. Notably, the DHAR and Lambda subfamily GSTs had more introns.

Fig. 2.

Exon–intron structures of apple GST genes. The gene structures were drawn using the GSDS program. Yellow indicates protein-coding sequences (CDSs); blue indicates upstream/downstream sequences; black line indicates introns

Chromosomal location analysis

The 52 non-redundant MdGSTs were randomly distributed across the 17 chromosomes of apple, while according to the available genome, MdGSTU9 was not located in any chromosome according to the available genome. This may be because of the high level of heterozygosity or the quality of the apple genome sequence (Fig. 3). Chromosome 05 contained 11 GSTs, followed by chromosome 17 and chromosome 10, which had seven and six GSTs, respectively. Both chromosome 04 and 09 contained four GSTs. Chromosomes 03, 13 and 15 had three GSTs. Chromosomes 02, 08, and 16 contained two GSTs each. In contrast, there was only one GST located on chromosome 06, 11, 12, and 14, whereas GSTs were not identified on chromosome 01 and 07.

Fig. 3.

Chromosomal locations of GST genes in the genomes of apple. Blue strips represent chromosomes and chromosome number is labeled on the top of each chromosome, Chr00 is the gene scaffold. The scale represents megabases (Mb)

Gene duplication events, microsynteny analysis, and selection pressure on duplicated GSTs

Genome duplication events have occurred frequently during the course of plant genome evolution (Rensing 2014). Tandem and segmental duplication events often underlie the expansion of gene families (Cannon et al. 2004). Gene duplication events in the apple genome were investigated to elucidate the mechanism behind the expansion of the apple GST family. A total of 17 duplicated pair of genes were identified across the genome, including six segmental duplication pairs and 11 tandem duplication events (Table 2). Among the six segmental GST gene pairs, four belonged to the Tau class, and DHAR and Phi class had one each. In the 11 tandem duplicated gene pairs, 10 belonged to the Tau class, which may have contributed to the high number of Tau class GSTs in the apple genome, and only one pair belonged to the Phi subfamily. These results suggest that tandem duplication events played a significant role in the expansion of the Tau class GSTs in apple, and this pattern is similar to the expansion of GSTs in other plant species (Jain et al. 2010; Dong et al. 2016). In general, both kinds of duplication events were important for the GST gene family expansion in apple. Furthermore, the approximate date of the duplication event was calculated, and we found that duplications of GST genes in apple occurred from 0.20 Mya to 49.15 Mya (Table 2).

Table 2.

Ka/Ks analysis and estimation of the absolute dates of the duplication events for the duplicated GST gene pairs from apple

Duplicate pair (Gene ID)	Duplicate pair (Gene name)	Duplicate type	Ka	Ks	Ka/Ks	Purifying selection	Duplicate dates (million years)
MD02G1236000–MD02G1236100	MdGSTU15–MdGSTU16	Tendem	0.021	0.006	3.390	No	0.203
MD04G1139400–MD04G1139500	MdGSTU13–MdGSTU8	Tendem	0.054	0.153	0.353	Yes	5.114
MD04G1139500–MD04G1139600	MdGSTU8–MdGSTU14	Tendem	0.080	0.150	0.532	Yes	4.998
MD05G1184200–MD05G1184300	MdGSTU7–MdGSTU5	Tendem	0.310	1.475	0.210	Yes	49.154
MD05G1184200–MD10G1172100	MdGSTU7–MdGSTU6	Segmental	0.087	0.125	0.696	Yes	4.170
MD05G1184300–MD05G1184400	MdGSTU5–MdGSTU3	Tendem	0.250	1.408	0.177	Yes	46.918
MD05G1210000–MD05G1210100	MdGSTU33–MdGSTU30	Tendem	0.040	0.090	0.450	Yes	2.984
MD05G1210000–MD10G1196300	MdGSTU33–MdGSTU24	Segmental	0.047	0.204	0.232	Yes	6.797
MD05G1210100–MD05G1210200	MdGSTU30–MdGSTU31	Tendem	0.038	0.047	0.808	Yes	1.575
MD05G1210600–MD05G1210700	MdGSTU29–MdGSTU25	Tendem	0.031	0.104	0.296	Yes	3.474
MD08G1244100–MD15G1438400	MdDHAR2–MdDHAR1	Segmental	0.052	0.206	0.253	Yes	6.866
MD09G1147400–MD17G1133800	MdGSTF4–MdGSTF1	Segmental	0.151	0.300	0.503	Yes	9.997
MD09G1279700–MD17G1286500	MdGSTU27–MdGSTU23	Segmental	0.084	0.185	0.453	Yes	6.158
MD10G1172200–MD10G1172300	MdGSTU2–MdGSTU4	Tendem	0.102	0.331	0.309	Yes	11.023
MD13G1081900–MD13G1082000	MdGSTU17–MdGSTU18	Tendem	0.136	0.708	0.192	Yes	23.612
MD13G1081900–MD16G1081100	MdGSTU17–MdGSTU11	Segmental	0.118	0.810	0.146	Yes	27.006
MD17G1134200–MD17G1134300	MdGSTF5–MdGSTF14	Tendem	0.107	0.208	0.514	Yes	6.922

The non-synonymous (Ka), synonymous (Ks) and Ka/Ks ratio were calculated to determine the selective forces that may have led to the duplication events. A Ka and Ks ratio below one represents purifying or negative selection, a Ka and Ks ratio equal to one represents neutral selection, and a Ka and Ks ratio larger than one represents positive selection. In apple, the Ka/Ks ratios for the 16 duplicated pairs were < 1, and most were approximately 0.15, indicating that they experienced strong purifying selection. The duplicated pair MdGSTU15–MdGSTU16 had a Ka/Ks ratio of 3.39, which indicated that these genes were under positive selection (Table 2). The results suggest that purifying selection was the primary influence on the apple GST family genes.

In addition, microsynteny analysis of one monocotyledon (rice) and three dicotyledons (apple, strawberry, rice and Arabidopsis) were performed to clarify the relationship of GSTs between monocots and eudicots. Interspecies microsynteny can be used to identify the location of orthologous genes. A total of 11 orthologous gene pairs between apple and strawberry, and 10 orthologous gene pairs between apple and Arabidopsis were found, whereas only one ortholog was identified between strawberry and Arabidopsis (Table 3 and Fig. 4). This indicates that many of the apple and strawberry GSTs and their Arabidopsis counterparts appeared to be derived from a common ancestor. Interestingly, no orthologous gene pairs were found between apple and rice, strawberry and rice, or Arabidopsis and rice, which may be due to the closer relationship of apple/strawberry to Arabidopsis compared to rice. Additionally, all the Ka/Ks ratios from the orthologous GST pairs were lower than 0.5, suggesting that these genes underwent intensive purifying selection.

Table 3.

Ka/Ks analysis between the duplicated apple, strawberry, Arabidopsis, and rice GSTs homologues

Duplicate pair	Gene name	Ka	Ks	Ka/Ks	Purifying selection
AT1G02940–mrna10551	AtGSTF5–FvGSTF7	0.4201	2.0486	0.2051	Yes
AT1G10360–MD16G1081100	AtGSTU18–MdGSTU11	0.267	3.800	0.070	Yes
AT1G10370–MD13G1081900	AtGSTU17–MdGSTU17	0.240	3.816	0.063	Yes
AT1G17180–MD10G1197100	AtGSTU25–MdGSTU20	0.181	3.578	0.050	Yes
AT1G59670–MD13G1081900	AtGSTU15–MdGSTU17	0.402	1.786	0.225	Yes
AT1G69930–MD16G1081100	AtGSTU11–MdGSTU11	0.372	1.925	0.193	Yes
AT1G77290–MD15G1133600	AtTCHQD1–MdTCHQD1	0.233	2.036	0.114	Yes
AT2G47730–MD03G1282700	AtGSTF8–MdGSTF3	0.259	3.768	0.069	Yes
AT3G09270–MD04G1139500	AtGSTU8–MdGSTU8	0.350	1.753	0.200	Yes
AT3G62760–MD11G1303600	AtGSTF13–MdGSTF6	0.363	1.564	0.232	Yes
AT5G02780–MD12G1129400	AtGSTL1–MdGSTL2	0.187	4.023	0.046	Yes
mrna10383–MD10G1196300	FvGSTU16–MdGSTU24	0.109	0.652	0.167	Yes
mrna10550–MD03G1282700	FvGSTF6–MdGSTF3	0.252	0.622	0.406	Yes
mrna11271–MD15G1133600	FvCHQD1–MdTCHQD1	0.152	0.692	0.219	Yes
mrna18167–MD12G1129400	FvGSTL1–MdGSTL2	0.099	0.670	0.148	Yes
mrna18731–MD04G1139400	FvGSTU9–MdGSTU13	0.283	0.988	0.287	Yes
mrna22483–MD04G1033800	FvGSTU13–MdGSTU28	0.155	0.903	0.171	Yes
mrna25077–MD15G1403200	FvCHQD2–MdDHAR3	0.057	0.833	0.068	Yes
mrna28763–MD17G1133600	FvGSTF9–MdGSTF10	0.068	0.735	0.093	Yes
mrna31672–MD17G1272100	FvGSTF12–MdGSTF12	0.134	1.001	0.134	Yes

Fig. 4.

Microsynteny analysis of GSTs among apple, strawberry, rice, and Arabidopsis chromosomes. Grey bars denote microsyntenic regions between apple, strawberry, rice and Arabidopsis GST genes. Yellow bars denote microsyntenic regions within apple. The chromosome number is indicated at the top of each chromosome. Scale bars marked on each chromosome indicates chromosome length (Mb)

Effects of ALA treatments on anthocyanin content and expression profiles of GSTs

ALA promotes anthocyanin accumulation in fruits. We applied 50 mg/L ALA on apple calli, and the ALA-treated calli had significantly higher anthocyanin content compared to the untreated control (Fig. 5). Control apple calli exposed to light for 72 h had 3.74 times more anthocyanin content, while ALA-treated calli had 13.56 times higher anthocyanin content. Moreover, based on RNA-seq transcriptomic data (unpublished), most structural and regulator genes involved in anthocyanin biosynthesis were up-regulated after ALA-treatment (Supplemental Fig. 7).

Fig. 5.

Effects of 5-ALA on coloration level and anthocyanin content in apple calli. A and B: Color (a) and anthocyanin content (b) in apple calli. Letters over bars indicate significant differences (P < 0.05; t test) between control and treatment. The blue bar indicates control treatment and red bar indicates ALA treatment

GSTs are flavonoid-binding proteins that transport anthocyanins in plants. However, little is known about the functions of GSTs in ALA-induced anthocyanin accumulation. Thus, we conducted RNA-seq transcriptomic analysis to investigate the expression pattern of GSTs in apple (Fig. 6). ALA treatment led to the divergent expression of 16 GSTs. Among these 16 genes, only MdGSTF12 was induced in all time points after ALA treatment. To verify the expression pattern of GSTs, a comprehensive qRT-PCR analysis of all GSTs was performed, except eight genes for which we could not design appropriate primers. The expression profiles of MdGSTs were classified into three sets. The first set was up-regulated after ALA treatment at three time points, and this group included the 12 members of the Tau class (Fig. 7a). The second set of genes were down-regulated at the three time points after ALA treatment, and included four, six, one and one members from Phi, Tau, TCHQD and Lambda classes, respectively (Fig. 7b). The last set included seven members of Phi, ten members of Tau, and three members of DHAR, and they displayed a variable response to ALA at different time points (Fig. 7c). These results imply that there is a complex GST expression network in response to ALA in apple. Interestingly, we noticed that MdGSTF12 was the GST that was strongly up-regulated in both RNA-seq and qRT-PCR experiments, indicating that MdGSTF12 may be important for ALA-regulated anthocyanin accumulation.

Fig. 6.

Heatmap of significantly expression level of GST genes of apple calli at three time point. And the expression values mapped to a color gradient from low (green) to high expression (red) are shown at the right of the figure. Log (RPKM) is the log of the gene expression value (RPKM) in each sample

Fig. 7.

RT-qPCR results of GSTs in apple callis at different times after ALA treatment. a–c Three groups of MdGSTs based on the expression profile. MdGST genes were up-regulated after ALA treatment (a). MdGST genes were down-regulated after ALA treatment (b). MdGST genes were varied after ALA treatment (c). The blue bar indicates control treatment and red bar indicates ALA treatment in both the panel. Data represent the mean ± SD (standard deviation) of three independent biological replicates

Discussion

Increasing evidence indicates that GSTs play vital roles in plant development, biotic and abiotic stress tolerance, and primary and secondary metabolism (Moons 2005; Yousuf et al. 2012; Dixon et al. 2010). In recent years, GSTs have been identified in many plant species, and the number and composition of GST family members are different in various plants. However, no genome-wide identification of the GST gene family has been reported in Rosaceae species, including apple, peach and strawberry. Here, 52, 42, 50 and 29 GST genes were identified in apple, pear, peach, and strawberry, respectively (Table 1 and Supplemental Tables 1–3). The strawberry had the fewest GSTs, and this may have caused by the loss of repetitive genes in strawberry compared to other plants, or GSTs may have expanded slower. Recently, 46 GSTs were reported in the pear genome (Wang et al. 2018). The difference between our results may be due to the different methods used.

Bioinformatic analysis of GST genes

It is generally believed that the function of a protein is related to its subcellular localization (Dönnes and Höglund 2004). In the present study, the bioinformatics analysis indicated that most of GSTs were located on cytoplasm (Table 1 and Supplemental Tables 2–4), and this was consistent with previous studies (Dong et al. 2016; Kayum et al. 2018a; Islam et al. 2017a). However, the GSTs that are important for flavonoid transport were also associated with membranes, such as the endoplasmic reticulum (ER) and the vacuole (Zhao 2015; Conn et al. 2008; Gomez et al. 2011; Kitamura et al. 2010). This discrepancy may be due to the program limitation and the complexity of protein localization (Xiong et al. 2016). Phylogenetic analyses revealed that GSTs in the four studied species were divided into Phi, Tau, DHAR, TCHQD, Lambda, and Theta classes (Fig. 1 and Supplemental Figs. 1–3). Similar to other plant species (Csiszár et al. 2014; Islam et al. 2018; Wang et al. 2018), the Tau family was most gene-rich. Theta and Zeta GSTs were the most ancestral classes, and were absent in the four Rosacea species, suggesting that they may have been lost in apple, pear, and strawberry during evolution. Usually, it is thought that the major driving force for generating novel genes and gene family expansion occur through whole-genome duplication (WGD), segmental duplication and tandem duplication events (Rensing 2014). Indeed, we found that tandem duplication was more frequent than segmental duplication in apple (Table 2). A similar pattern was also observed in rice (Jain et al. 2010), G. raimondii and G. arboreum (Dong et al. 2016), Capsella rubella (He et al. 2016), and poplar (Lan et al. 2009). Gene duplication not only expands genome content, but is also important for generating novel gene functions, which can allow for organisms to adapt to complex environments (Freeling 2009). The Ka/Ks ratios of all apple GST pairs were lower than 1, except MdGSTU15–MdGSTU16, suggesting that most gene pairs underwent strong purifying selection in apple. The mean age for duplication events in apple was 12.75 Mya, and the most recent tandem and segmental duplication event was estimated to have occurred 0.20 Mya and 4.17Mya, respectively. This is much later than the recent apple whole genome duplication, which happened between 30 and 45 Mya (Velasco et al. 2010). Both multiple duplication events and the ages of genome duplication events suggest that the evolutionary history of GSTs occurred through a complicated process.

The gain or loss of an exon or intron plays an important role in the diversification of multi-gene families. This event may arise from the rearrangement and fusions of different chromosome fragments (Xu et al. 2012). We found that most of the same subfamilies shared similar gene structures in apple. However, we also noticed that some GSTs had different intron–exon structure within the same subfamily (Fig. 2). The gene structure characteristics of GSTs were conserved in potato, sweet potato and populus (Islam et al. 2018; Lan et al. 2009; Ding et al. 2017; Vaish et al. 2018).

In this study, strong microsynteny was detected between apple and Arabidopsis/strawberry. In contrast, no microsynteny was detected between apple/strawberry/Arabidopsis and rice (Fig. 4), which may be due to the evolutionary distance of dicots and monocots. We conclude that monocot and dicot GSTs exhibited differences in their evolutionary history and there was likely a lineage specific expansion and diversification of GSTs (Chi et al. 2010).

ALA promoted anthocyanin accumulation through GST pathway

Exogenous application of ALA at a suitable concentration can be a powerful tool to increase the anthocyanin content of plants (Feng et al. 2016, 2015; Ye et al. 2017; Xie et al. 2013). The molecular mechanisms underlying ALA-induced anthocyanin accumulation remain poorly understood. Previous studies revealed that ALA up-regulates anthocyanin biosynthetic structural genes, such as PAL, CHS, LDOX and UFGT, and regulatory genes like MYB, bHLH, and WD40 (Zheng et al. 2017; Ye et al. 2017; Xie et al. 2013). Moreover, the transcription factor MdMADS1 is involved in ALA induced anthocyanin accumulation in apple calli (Feng et al. 2016). However, to the best of our knowledge, there have not been any investigations that reveal the function of transport proteins in response to ALA treatment. To gain insight into the expression profiles of GSTs in response to ALA treatment, transcriptomic data was analyzed and validated using qRT-PCR (Figs. 6, 7). Clear divergence in expression patterns was observed among apple GSTs in response to ALA treatments.

Based on qRT-PCR analysis, 12 Tau members responded rapidly to ALA treatment and for a longer time. Tau GSTs are involved in detoxification of xenobiotics, signaling, stress tolerance, and transporting anthocyanins (Marrs et al. 1995; Loyall et al. 2000; Thom et al. 2002; Jha et al. 2011). Heterologous expression of a rice Tau class GST (OsGSTU4) in Arabidopsis improved tolerance to salinity and oxidative stresses (Sharma et al. 2014). It is well known that ALA improves plant tolerance to various stress (Balestrasse et al. 2010; Naeem et al. 2010, 2011). Therefore, our results suggest that ALA may up-regulate Tau class GSTs, improving the stress resistance of apple. Further studies are need to determine the detailed function of the specific MdGST.

MdGSTF12 was the only GST that was strongly up-regulated by ALA. Recently, several transcriptomic studies demonstrated that MdGSTF12 may be involved in anthocyanin accumulation. MdGSTF12 has been shown to be the most suppressed gene in a yellow-skin somatic mutant apple line with corresponding reductions in anthocyanin content (El-Sharkawy et al. 2015). In the flavonoid biosynthetic processes of red-fleshed apples, MdGSTF12 was strongly up-regulated (Wang et al. 2015). Applying NAA and 2,4-D to apple calli at higher concentrations inhibits anthocyanin and flavonoid transport into the vacuole and down-regulates MdGSTF12 (Ji et al. 2015). Interestingly, MdGSTF12 falls within the same clade as genes that have been previously characterized as anthocyanin transporters, such as AtGSTF12/TT19 and Phi class GSTs, e.g., VviGST4 (Pérez-Díaz et al. 2016) and LcGST4 (Hu et al. 2016) (Supplemental Fig. 8). VviGST4 and LcGST4 are considered the ortholog of TT19 from Arabidopsis. In a complementation experiment, both VViGST4 and LcGST4 rescued the anthocyanin-less phenotype of the Arabidopsis tt19 mutant (Hu et al. 2016; Pérez-Díaz et al. 2016). In Litchi, Hu et al. (2016) found that the transcription factor of anthocyanin biosynthesis, LcMYB1, could regulate the expression of LcGST4. Taken together, these findings highlight the important role of MdGSTF12 in anthocyanin accumulation, and further functional analyses are required to shed light into how MdGSTF12 regulates anthocyanin accumulation.

Although the three other species were not treated with ALA, existing studies have found that overexpression of the GST gene can cause anthocyanin accumulation. Reduced anthocyanin content in the petioles (rap) mutant, a white-fruited variety of wild strawberry, was caused by a premature stop codon of GST. Transient knock-down of RAP leads to reduced fruit coloration in strawberry (Luo et al. 2018). During the development of peach fruits, ppa11307 and ppa00812, which encode GST, play an important role in anthocyanin transport in peach flesh (Cao et al. 2018).

Conclusions

We identified 52, 42, 50 and 29 GSTs from apple, pear, peach and strawberry, respectively. A detailed bioinformatics analysis on the phylogenetic relationships, gene structure, chromosomal location, and gene duplication of GSTs was performed. The expressions of MdGSTs was affected by ALA treatment, indicating that GSTs may be involved in ALA-induced anthocyanin accumulation. This is the first systematic isolation and analysis of apple GSTs, and provides insight into the functional analysis of MdGSTF12.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

Conceived and designed the experiments: XF, LFSG, and JLW. Analyzed the data: XF and JZ. ALA treatment in apple calli observation: XF and JZ. RT-qPCR measuring: XF. Wrote and revise the paper: XF, YYA, and JLW.

Funding

This work was supported by the National Natural Science Foundation of China (31772253 and 31772283), and Key R&D projects in Jiangsu Province (BE2018389).

Compliance with ethical standards

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Ethics approval and consent to participate

Not applicable.