Plastidial Phosphoglucomutase (pPGM) Overexpression Increases the Starch Content of Transgenic Sweet Potato Storage Roots

Sweet potato (Ipomoea batatas), an important root crop, has storage roots rich in starch that are edible and serve as a raw material in bioenergy production. Increasing the storage-root starch contents is a key sweet potato breeding goal. Phosphoglucomutase (PGM) is the catalytic enzyme for the interconversion of glucose-6-phosphate and glucose-1-phosphate, precursors in the plant starch synthetic pathway. Plant PGMs have plastidial and cytosolic isoforms, based on their subcellular localization. Here, IbpPGM, containing 22 exons and 21 introns, was cloned from the sweet potato line Xu 781. This gene was highly expressed in the storage roots and leaves, and its expression was induced by exogenous sucrose treatments. The mature IbpPGM protein was successfully expressed in Escherichia coli when a 73-aa chloroplastic transit peptide detected in the N-terminus was excised. The subcellular localization confirmed that IbpPGM was localized to the chloroplasts. The low-starch sweet potato cultivar Lizixiang IbpPGM-overexpression lines showed significantly increased starch, glucose, and fructose levels but a decreased sucrose level. Additionally, the expression levels of the starch synthetic pathway genes in the storage roots were up-regulated to different extents. Thus, IbpPGM significantly increased the starch content of the sweet potato storage roots, which makes it a candidate gene for the genetic engineering of the sweet potato.


Introduction
Starch is the main storage form of carbohydrates in plants. During the day, leaves synthesize transitory starch in the chloroplasts through photosynthesis. It is then degraded into sucrose at night and transported to non-photosynthetic organs as an energy source [1,2]. ADP-glucose (ADPG) is the transitory starch substrate in chloroplasts and is synthesized by the successive catalysis of plastidial glucose phosphate isomerase, phosphoglucomutase (PGM) and ADP-glucose pyrophosphorylase. The ADPG is then transformed into amylose or amylopectin by the starch synthases [3].
The interconversion between glucose-1-phosphate (G1P), the precursor of ADPG, and glucose-6-phosphate (G6P) is catalyzed by PGM. Two types of PGM exist in plants, Primers starting with "q" are used for the qRT-PCR. The underlined nucleotides are restriction enzyme sites.

Expression Analysis of IbpPGM in the Sweet Potato
The total RNA was isolated from five different tissues (storage root, fibrous root, stem, leaf, and petiole) of Xu 781 plants grown in the field for approximately 100 days, and the first-strand cDNA was synthesized using PrimeScriptTM RT Reagent Kit with gDNA Eraser (Perfect Real Time) (TaKaRa, Beijing, China). The qRT-PCR was conducted on a 7500 Real-Time PCR System (Applied Biosystems, Waltham, MA, USA) to determine the IbpPGM transcript levels using the gene-specific primers qPGM-F/R. The control, IbActin, was amplified using the primers Actin-F/R.
The Xu 781 plants that were grown in the field for approximately one month, were used to investigate the response of IbpPGM to exogenous sucrose. Briefly, the leaf-petioles (10 cm) of the Xu 781 plants were cultured in water in darkness for 1 d as a starvation treatment. Then, they were supplied with water or 175 mmol L −1 sucrose in darkness at 28 • C. The qRT-PCR was conducted to determine the transcript levels of IbpPGM in the cuttings harvested at different time points (0, 2, 4, 6, 12, 24 and 48 h) after treatment. Three cuttings were used as biological replicates per time point for each treatment.

Prokaryotic Expression of IbpPGM
To determine whether IbpPGM encodes a mature protein, the gene was expressed in E. coli. The full-length ORF and the signal peptide-cleaved ORF (∆ORF) of IbpPGM were amplified with primers pET-F/R and pET-∆F/R, respectively. The sequence-verified fragments were ligated independently into the expression vector pET-28a. Then, the recombinant vectors pET-28a-IbpPGM and pET-28a-∆IbpPGM, as well as the pET-28a native vector, were introduced independently into the competent E. coli strain Transetta (DE3) cells (Transgen, Beijing, China). Fresh Luria-Bertani medium was inoculated independently with the positive clones, and they were cultured at 28 • C until the OD600 values reached 0.8. The soluble cytoplasmic proteins were prepared from the isopropyl β-D-thiogalactopyranosideinduced Transetta (DE3) cells. Then, the expressed IbpPGM protein was subjected to an SDS-PAGE analysis.

Subcellular Localization
The ORF of IbpPGM was amplified using primers 83-F/R and inserted into the pMDC83 vector between PacI and AscI cleavage sites. The recombinant vector pMDC83-IbpPGM and the native vector were introduced independently into the A. tumefaciens strain EHA105, and the positive strains were injected independently into the N. benthamiana leaf epidermal cells for the transient expression. Following the co-cultivation at 28 • C for approximately 36 h, the agroinfiltrated tobacco leaves were visualized using a laser scanning confocal microscope.

Production of the Transgenic Plants
To construct the overexpression vector, the sequence-verified ORF of IbpPGM, which had been amplified using primers OPGM-F/R, was inserted into pBI121 between BamHI and SacI to replace the glucuronidase (gusA) gene. Subsequently, the expression cassette 35S-IbpPGM-NOS was excised from the pBI121-IbpPGM vector using HindIII and PstI and then ligated between the same cleavage sites in pCAMBIA3301, to generate the recombinant overexpression vector pC3301-121-IbpPGM. This plasmid was transfected into the A. tumefaciens strain EHA105. The plant transformation and regeneration were performed, as described by Wang et al. [32] using embryogenic suspension cultures of Lizixiang established, as described by Liu et al. [34]. The putatively transgenic sweet potato plants were identified by histochemical GUS assays and PCR. Then, the positive transgenic lines were subjected to the qRT-PCR using primers qPGM-F/R, and the three lines with the highest IbpPGM expression levels were selected for further phenotypic analyses.

Quantification of the Carbohydrate Contents
The starch contents in the storage roots of the transgenic and WT plants were analyzed in accordance with the method of Smith and Zeeman [35], and high performance liquid chromatography (HPLC) was used to determine the sucrose, glucose, and fructose contents in the storage roots with the following method. First, 30 mg of 80 • C dried storage roots was dissolved in 0.7 mL of 80% ethanol to extract the sugars. Then, the sample was thoroughly vortexed and incubated for 2 h at 70 • C. The aliquots of 0.7 mL of HPLC-grade water and 0.7 mL chloroform were added to the sample. Then, after shaking several times, the mixtures were centrifuged at 12,000 g for 10 min. Then, 0.7 mL of the aqueous supernatant was transferred into 1.5-mL Eppendorf tubes and resuspended in 0.7 mL of chloroform. Following the centrifugation at 12,000 g for 10 min, 0.5 mL of the supernatant was transferred to a glass tube for the HPLC analysis of each sugar component. The Agilent technologies HPLC column (ZORBAX Carbohydrate column; 4.6 × 150 mm, 5 µm) with a differential refraction detector was used. The mobile phase consisted of 75% acetonitrile with a flow rate of 0.8 mL min −1 and the temperature of the column was maintained at 35 • C. The sugars were identified, based on the retention time of the standards, and the sample concentrations were calculated from the external standard curve. The quantifications were carried out with three replicates for each plant line.

Expression Analysis of the Starch Biosynthetic Genes
The transcript levels of 12 key genes in the starch biosynthetic pathways of the transgenic and WT storage roots were investigated using qRT-PCR. The 12 Ipomoea batatas genes were AGP-sTL1 and 2 (encoding the two small subunits of IbAGPase), AGP-TLI (encoding the large subunit of IbAGPase), granule-bound starch synthase I (GBSSI), soluble starch synthase I (SSI), SSII, SSIII, SSIV, starch branching enzymes I and II (SBEI and SBEII), isoamylase1 (ISA1), and pullulanase (PUL). The primers used to amplify these genes are listed in Table 1.

Cloning and Sequence Analysis of IbpPGM
The RACE method was used to clone IbpPGM from the Xu 781 sweet potato line. The cloned 2182-bp full-length IbpPGM cDNA contained a 1917-bp ORF that generated a 638-aa protein with a molecular weight of 69.3 kDa. The genomic sequence of IbpPGM was 5583 bp and contained 22 exons and 21 introns. TargertP 2.0 and ChloroP 1.1 predicted that the N-terminus of the IbpPGM protein contains a 73-aa chloroplast transit peptide. The molecular weight of IbpPGM without the transit peptide was 61.7 kDa. A multiplesequence alignment of the PGM protein showed that IbpPGM shared a conserved domain similar with those from several other plants and from Saccharomyces cerevisiae ( Figure 1). Additionally, pPGM had a chloroplast transit peptide that was not found in cPGM. A GSDS analysis revealed that pPGM and cPGM evolved into two different branches, with members of each branch having identical exon-intron structures, and IbpPGM was most closely related to the pPGM from Solanum lycopersicum (80.82% homology) ( Figure 2).

Expression of IbpPGM in the Sweet Potato
The qRT-PCR revealed that IbpPGM was expressed in all five tested tissues of Xu 781 plants, with the highest expression occurring in the storage root, followed, successively, by the leaves, the fibrous roots, the stems, and the petioles (Figure 3a). The 175-mM sucrose treatments of the leaf-petiole cuttings in darkness significantly induced the IbpPGM expression, which strongly increased at 12 h after treatment and peaked after 48 h, reaching approximately 28 times the value at 0 h ( Figure 3b).

Expression of IbpPGM in the Sweet Potato
The qRT-PCR revealed that IbpPGM was expressed in all five tested tissues of Xu 781 plants, with the highest expression occurring in the storage root, followed, successively, by the leaves, the fibrous roots, the stems, and the petioles (Figure 3a). The 175-mM sucrose treatments of the leaf-petiole cuttings in darkness significantly induced the IbpPGM expression, which strongly increased at 12 h after treatment and peaked after 48 h, reaching approximately 28 times the value at 0 h ( Figure 3b).

Expression of IbpPGM in E. coli
To investigate whether IbpPGM encodes the mature protein, we constructed a recombinant vector harboring the full-length ORF of IbpPGM, as well as one carrying the ORF with the transit peptide removed (ΔIbpPGM), to eliminate the effect of the chloroplast transit peptide on the prokaryotic expression. These two vectors were then expressed independently in E. coli (Figure 4). The full-length ORF of IbpPGM failed to express the corresponding 69.3-kDa protein, whereas ΔIbpPGM expressed a protein of approximately 61.7 kDa. This suggested that the chloroplast transit peptide inhibits the expression of IbpPGM in E. coli.

Expression of IbpPGM in E. coli
To investigate whether IbpPGM encodes the mature protein, we constructed a recombinant vector harboring the full-length ORF of IbpPGM, as well as one carrying the ORF with the transit peptide removed (∆IbpPGM), to eliminate the effect of the chloroplast transit peptide on the prokaryotic expression. These two vectors were then expressed independently in E. coli (Figure 4). The full-length ORF of IbpPGM failed to express the corresponding 69.3-kDa protein, whereas ∆IbpPGM expressed a protein of approximately 61.7 kDa. This suggested that the chloroplast transit peptide inhibits the expression of IbpPGM in E. coli. binant vector harboring the full-length ORF of IbpPGM, as well as one carrying the ORF with the transit peptide removed (ΔIbpPGM), to eliminate the effect of the chloroplast transit peptide on the prokaryotic expression. These two vectors were then expressed independently in E. coli (Figure 4). The full-length ORF of IbpPGM failed to express the corresponding 69.3-kDa protein, whereas ΔIbpPGM expressed a protein of approximately 61.7 kDa. This suggested that the chloroplast transit peptide inhibits the expression of IbpPGM in E. coli.

Subcellular Localization of IbpPGM in N. benthamiana
Both TargertP 2.0 and ChloroP 1.1 predicted that IbpPGM contained a chloroplast transit peptide. The expression vector pMDC83-IbpPGM was, therefore, transiently expressed in the N. benthamiana epidermal cells and visualized using a laser scanning confocal microscope ( Figure 5). In the tobacco epidermal cells, IbpPGM was observed in scattered patches and co-localized with the autofluorescence of the chloroplasts. This result concurred with the online predictions that IbpPGM localizes to the chloroplast.

Subcellular Localization of IbpPGM in N. benthamiana
Both TargertP 2.0 and ChloroP 1.1 predicted that IbpPGM contained a chloroplast transit peptide. The expression vector pMDC83-IbpPGM was, therefore, transiently expressed in the N. benthamiana epidermal cells and visualized using a laser scanning confocal microscope ( Figure 5). In the tobacco epidermal cells, IbpPGM was observed in scattered patches and co-localized with the autofluorescence of the chloroplasts. This result concurred with the online predictions that IbpPGM localizes to the chloroplast.

Overexpression of IbpPGM in the Sweet Potato
To functionally characterize IbpPGM, the recombinant vector pC3301-121-IbpPGM was introduced into the sweet potato cultivar Lizixiang. A total of 97 putative transgenic lines were obtained, and 10 lines were shown to be transgenic using GUS assays and PCR verification. The qRT-PCR indicated that the IbpPGM expression levels in these 10 transgenic lines were 1.3-15.3 times of the level in WT, with lines OX17, OX53, and OX85 having the three highest expression levels ( Figure 6). Then, the three OX lines and the wildtype plants were further propagated in vitro and we observed that the leaf area of the invitro plantlets of OX was much larger than that of WT and the growth rate of OX was also faster. However, when they were grown in the field, no differences were identified between WT and OX in leaf size and shape, vine growth vigor, or morphology of the storage roots ( Supplementary Figures S1 and S2). The overexpression of IbpPGM may have an influence on improving the photosynthesis of OX and under the lab tissue culture condition, this advantage of OX was magnified by the limited culture conditions, in terms of

Overexpression of IbpPGM in the Sweet Potato
To functionally characterize IbpPGM, the recombinant vector pC3301-121-IbpPGM was introduced into the sweet potato cultivar Lizixiang. A total of 97 putative transgenic lines were obtained, and 10 lines were shown to be transgenic using GUS assays and PCR verification. The qRT-PCR indicated that the IbpPGM expression levels in these 10 transgenic lines were 1.3-15.3 times of the level in WT, with lines OX17, OX53, and OX85 having the three highest expression levels ( Figure 6). Then, the three OX lines and the wild-type plants were further propagated in vitro and we observed that the leaf area of the in-vitro plantlets of OX was much larger than that of WT and the growth rate of OX was also faster. However, when they were grown in the field, no differences were identified between WT and OX in leaf size and shape, vine growth vigor, or morphology of the storage roots ( Supplementary Figures S1 and S2). The overexpression of IbpPGM may have an influence on improving the photosynthesis of OX and under the lab tissue culture condition, this advantage of OX was magnified by the limited culture conditions, in terms of light, temperature, humidity, and CO 2 concentration. When transplanted in the field, the photosynthetic ability and growth vigor of all plant lines were fully stimulated by the sufficient environmental factors and the phenotypic differences disappeared. lines were obtained, and 10 lines were shown to be transgenic using GUS assays and PCR verification. The qRT-PCR indicated that the IbpPGM expression levels in these 10 transgenic lines were 1.3-15.3 times of the level in WT, with lines OX17, OX53, and OX85 having the three highest expression levels ( Figure 6). Then, the three OX lines and the wildtype plants were further propagated in vitro and we observed that the leaf area of the invitro plantlets of OX was much larger than that of WT and the growth rate of OX was also faster. However, when they were grown in the field, no differences were identified between WT and OX in leaf size and shape, vine growth vigor, or morphology of the storage roots ( Supplementary Figures S1 and S2). The overexpression of IbpPGM may have an influence on improving the photosynthesis of OX and under the lab tissue culture condition, this advantage of OX was magnified by the limited culture conditions, in terms of light, temperature, humidity, and CO2 concentration. When transplanted in the field, the photosynthetic ability and growth vigor of all plant lines were fully stimulated by the sufficient environmental factors and the phenotypic differences disappeared.

Starch and Sugar Contents in the Transgenic Sweet Potato
The quantified levels of starch and sugar contents in the transgenic and WT lines are shown in Table 2. The overexpression of IbpPGM in the sweet potato significantly increased the starch contents in the storage roots. Moreover, the sucrose contents in the transgenic storage roots decreased significantly, whereas the glucose and fructose contents increased.

Expression Profiles of the Starch Biosynthetic Genes
The expression levels of the starch biosynthesis related genes in the transgenic lines were evaluated by the qRT-PCR (Figure 7). All 12 detected genes were located downstream of IbpPGM in the starch biosynthesis pathway and showed increased expression levels, to different extents, in the transgenic lines, compared with WT. Among the 12 genes, IbAGP-sTL1, IbAGP-sTL2, and IbAGP-TLI were responsible for the ADPG synthesis. IbGBSSI was involved in the amylose elongation, and the other eight genes were mainly involved in the amylopectin synthesis. Thus, the overexpression of IbpPGM promoted the accumulation of the starch biosynthesis precursors and resulted in the up-regulation of related downstream genes. stream of IbpPGM in the starch biosynthesis pathway and showed increased expression levels, to different extents, in the transgenic lines, compared with WT. Among the 12 genes, IbAGP-sTL1, IbAGP-sTL2, and IbAGP-TLI were responsible for the ADPG synthesis. IbGBSSI was involved in the amylose elongation, and the other eight genes were mainly involved in the amylopectin synthesis. Thus, the overexpression of IbpPGM promoted the accumulation of the starch biosynthesis precursors and resulted in the up-regulation of related downstream genes.

Discussion
The sweet potato is an important food crop with starchy storage roots that are raw materials for food, feed, and industrial uses. Traditional hybrid breeding results in a slow crop improvement rate and is not trait specific. Consequently, genetic engineering has become an effective way to increase the starch contents in the sweet potato storage roots, which is a primary goal of sweet potato breeding. The main functional genes in the sweet potato starch synthesis have been characterized. The overexpression of IbAATP, which is responsible for transporting ATP in the cytoplasm to plastids as an energy supply for starch synthesis, and IbSSI, which is involved in elongating the short chains of amylopectin, significantly increases the starch contents in the sweet potato storage roots [32,33]. Additionally, the overexpression of IbSnRK1 in the sweet potato improves the starch content, as well as the starch quality, in the storage roots [36].
In the plant starch synthetic pathway, PGM is responsible for the synthesis of the upstream precursor G1P. Functional studies of PGM have mainly focused on how the starch contents in the plants changed when this gene is mutated or subject to RNA interference [10,12,15,24]. However, few studies have investigated how the starch contents in plants, including the sweet potato, are altered when PGM is overexpressed. In this study, the coding and genomic sequences of IbpPGM were isolated from the Xu 781 sweet potato line using homologous cloning. A GSDS analysis showed that IbpPGM shared an identical gene structure (22 exons and 21 introns) with its counterparts from other species, such as Brassica napus and A. thaliana, whereas cPGMs shared a similar gene structure of 18 exons and 17 introns. This demonstrated that the functional differentiation of the PGM genes into plastidial and cytosolic forms was accompanied by gene structural changes. Additionally, the deletion of the chloroplast transit peptides from the cPGM proteins, compared with the pPGM proteins, also indicated that the gene function is closely related to the gene structure.
The expression level of IbpPGM was high in storage roots and leaves, which was consistent with its characterized role in starch synthesis. IbpPGM exists in amyloplasts in storage roots, whereas in leaves, it is mainly found in chloroplasts. Sucrose is an important signaling molecule in starch synthesis, and it induces the expression of starch synthesis-related genes through the abscisic acid pathway. When the sucrose contents in leaves (or in-vitro sucrose feeding) exceeds the need for respiration, starch synthesis is induced [37,38]. In sweet potato, the expression levels of IbGBSSI, IbAGP-sTL1, and IbSSI are all significantly induced by exogenous sucrose treatments [26,33,39]. In this study, the leaf-petiole cuttings were soaked in water in darkness as the starvation pretreatment to consume the endogenous carbohydrates and minimize respiration. The 0 h−6 h sucrose treatment was the recovery period for respiration, during which the IbpPGM expression was induced at a low level. When the continuous sucrose feeding exceeded the respiratory demand, the IbpPGM expression began to increase rapidly, reaching 28 times the 0 h level by the end of the treatment. However, the continuous feeding of water after the starvation pre-treatment did not induce the IbpPGM expression. In agronomic practices, spraying sucrose-based polymers on leaves may improve the yield and quality of some fruits, which may result from induced alterations in the carbohydrate metabolism in leaves and then fruits. Thus, determining whether a leaf spray of sugar-based growth regulators to increase the yield and quality of the sweet potato, would be of interest in the future.
To determine whether IbpPGM encodes a mature protein, two recombinant vectors for the prokaryotic expression were constructed. One carried the complete coding sequence of IbpPGM, and the other carried the same sequence minus the chloroplast transit peptide (∆IbpPGM). The former could not be expressed in E. coli, whereas the latter expressed a mature protein. Thus, the transit peptide of IbpPGM may form a complex secondary structure at the mRNA level that affects the initial translation in a prokaryotic system. The fluorescence of the GFP protein fused with the complete coding sequence of IbpPGM mainly targeted the chloroplasts, which indicated that the transit peptide was successfully translated in the eukaryotic system and directed the IbpPGM protein to the chloroplasts.
To characterize the function of IbpPGM, it was overexpressed in the sweet potato lowstarch cultivar Lizixiang, by infecting the embryogenic suspension cells, and 10 positive transgenic lines were obtained. The three lines having the highest IbpPGM expression levels exhibited significant 4.5%, 5.5%, and 12.0% increases in the storage-root starch content, compared with the WT line. Owing to the overexpression of IbpPGM, the precursor G1P accumulated, which up-regulated the downstream genes involved in amylose and amylopectin synthesis, leading to further increases in the starch contents (Figure 7). In heterotrophic organs, the catalytic substrate of pPGM for G1P is G6P, which is mainly derived from the degradation of sucrose and is transported into the amyloplasts from the cytoplasm [40]. In this study, the overexpression of IbpPGM consumed more G6P for the G1P synthesis, which accelerated the degradation of sucrose in the cytoplasm, leading to further increases in the glucose and fructose contents.

Conclusions
In the present study, the IbpPGM gene was cloned from the Xu 781 sweet potato line. IbpPGM was mainly expressed in the storage roots and leaves. Its expression was strongly induced by the exogenous sucrose treatments. The IbpPGM protein was subcellularly localized to the chloroplasts and was successfully expressed in E. coli when its 73-aa chloroplastic transit peptide was excised. The overexpression of IbpPGM significantly increased the starch contents of the transgenic sweet potato storage root and altered its soluble sugar levels. Meanwhile, the expression levels of starch biosynthetic genes in transgenic sweet potato storage roots showed increased expression levels to different extents. These results indicated that IbpPGM has the great potential as an important candidate gene for increasing the starch content of the sweet potato through genetic engineering.