Background

Switchgrass (Panicum virgatum L.) is a C4 warm season perennial forage grass and a leading lignocellulosic feedstock for renewable bioenergy production in the United States [1]. It is well adapted to eastern North America with a wide range of genomic variation, high biomass yield, efficient use of water and nutrients, and resilience to biotic and abiotic stresses [2]. Genetic engineering of switchgrass has been performed to increase biomass [3,4,5,6,7,8], modify flowering [9, 10], improve plant architecture [11], reduce cell wall recalcitrance (i.e., resistance of cell wall to deconstruction and conversion into biofuels) [11,12,13,14,15,16,17,18,19], and improve water and nutrition use efficiency [20].

So far, the number of promoters used for switchgrass transformation and genetic engineering has been very limited. The list consists of constitutive promoters such as the cauliflower mosaic virus (CaMV) 35S promoter [3, 20,21,22], the rice actin 1 (OsAct1) promoter [23], and the ubiquitin promoters from maize (Ubi-1) [5,6,7,8, 10, 11, 14, 16,17,18,19, 24,25,26], rice (rubi2 and rubi3) [27, 28], and switchgrass (PvUbi1 and PvUbi2) [29]. Constitutive promoters allow high levels of constant gene expression in all tissues at all developmental stages, and permit first-order analysis of phenotypes conferred by a transgene. Such continuous high level of expression of some transgenes may cause side effects to the host plants, such as homology-dependent gene silencing [30, 31], unintended impacts on growth and development [32,33,34], and abnormal morphology [32, 35,36,37,38,39,40]. A good example is when switchgrass MYB4 (PvMYB4) gene was overexpressed in transgenic switchgrass under the control of ZmUbi1 promoter [12, 17]. The best PvMYB4 overexpression line during a two-year field experiment produced 32% more biofuel and 63% more biomass than the non-transgenic switchgrass, which represents a doubling of biofuel production per hectare and the highest gain among all of the reported field-grown genetically modified feedstocks [12]. However, since PvMYB4 encodes an R2–R3-type transcription factor that acts as a negative regulator of many lignin biosynthetic genes, it was observed that negative growth effects and yield penalties were associated with the high expression levels of the transgenic PvMYB4 in several of the field-grown switchgrass overexpression lines [12]. The best PvMYB4 overexpression lines in terms of low lignin content and high biofuel and biomass production had low-to-moderate ectopic expression levels. The high-level ectopic expression lines exhibited reduced tiller height, plant width (i.e., the diameter at the mid-section of each whole plant; spread of tillers), and tiller numbers [12, 17], and did not survive the first winter in the field [12]. Moreover, the high-level ectopic expression lines also suffered from weak, undeveloped, or diminished (so-called mushy) root systems, indicating a disruptive effect of PvMYB4 overexpression on the root system [12]. Thus, there is an urgent need for green tissue-specific promoters to limit PvMYB4 overexpression to the aboveground tissues, where the harvestable biomass is produced each growing season.

Multiple green tissue-specific promoters have been well characterized in some monocot species such as maize and rice. These include the promoters of maize phosphoenolpyruvate carboxylase (PEPC) [41,42,43], pyruvate orthophosphate dikinase (PPDK) [44] and the small subunit of ribulose-1, 5-bisphosphate carboxylase/oxygenase (rbcS) [43], rice light-harvesting complex II chlorophyll-a/b binding gene (Lhcb; also known as Cab) [45,46,47], rbcS [48], Leaf Panicle 2 (LP2) [49], D54O [50], and DX1 [51]. These genes are tightly light-inducible in green tissues, and their proteins are involved in photosynthesis that converts light energy into sugar. For example, Lhcb encodes a protein that functions in photosystems I and II by binding to chlorophyll. PEPC in C4 plants encodes a cytosolic enzyme that catalyzes the conversion of phosphoenolpyruvate and bicarbonate to four carbon acid oxaloacetate and inorganic phosphate, even though its primary function in C3 plants is anaplerotic by replenishing the tricarboxylic acid cycle with intermediates [52]. PsbR encodes the subunit R of the photosystem II 10 kDa polypeptide.

In the present study, we identified and functionally characterized three switchgrass green tissue-specific promoters (i.e., PvLhcbp, PvPEPCp, and PvPsbRp; p stands for promoter) in endogenous plant tissues as well as in transgenic rice. We characterized the functions of truncations of each promoter. We also fused the green tissue promoters with PvMYB4 to study the effects of targeted green tissue expression of this transcriptional repressor in transgenic switchgrass on plant growth and sugar release.

Results

Sequence analysis and expression patterns of three green tissue-specific genes in switchgrass

Using the rice green tissue-specific genes OsLhcb, OsPEPC, and OsPsbR as the query sequences, the BlastP search for their homologous sequences in the switchgrass genome returned 11, 8, and 4 candidate sequences, respectively, with high amino acid sequence similarities (Additional file 1: Figs. S1–S3). The cDNA sequence analysis indicated that all of these candidate sequences had the same exon/intron structures as their rice homologs (Additional file 1: Figs. S4–S6) except 6 out of the 11 switchgrass Lhcb (PvLhcb) sequences, which contained a single intron at variable positions while their rice homologs were intronless (Additional file 1: Fig. S4).

The 11, 8, and 4 switchgrass candidate sequences, which, respectively, corresponded to OsLhcb, OsPEPC, and OsPsbR, had the highest sequence identity to 8, 6, and 3 Noble Foundation Gene Atlas unitranscript entries, respectively (Additional file 1: Figs. S7–S9). In silico expression analysis revealed that these unitranscript entries showed highly variable expression levels in different switchgrass tissues (Additional file 1: Figs. S7–S9). The entries of PvLhcb and the switchgrass PsbR (PvPsbR) were mainly expressed in the aboveground tissues with AP13CTG19188 (i.e., the entry of Pavirv00047797m and Pavirv00024895m) and AP13CTG07332 (i.e., the entry of Pavirv00009702m) being the strongest expressed entries of both genes, respectively (Additional file 1: Figs. S7, S9). Both entries were highly expressed in leaves, internodes, nodes, inflorescences that were taller than 200 mm in length, and whole flowers. They were moderately expressed in inflorescences of rachis, primary and secondary branch meristem of 0.5–3.0 mm in length at the initiation stage, inflorescences of the glume and floret developmental stages of 10–20 mm in length, and inflorescences of 50–150 mm in length. They were weakly expressed in the whole crown of E4 stage plant, and minimally expressed in roots (Fig. 1; Additional file 1: Figs. S7, S9). Unexpectedly, all of the entries, including KanlCTG00012, which is the entry of Pavirv00033161m, of the switchgrass PEPC (PvPEPC) had extremely low expression levels in different tissues (Additional file 1: Fig. S8). Based on the sequence similarities, gene structure, and in silico expression patterns, Pavirv00047797m, Pavirv00033161m, and Pavirv00009702m were selected as the potential switchgrass homologs of the rice green tissue-specific genes OsLhcb, OsPEPC, and OsPsbR for further analysis and named PvLhcb, PvPEPC, and PvPsbR hereafter, respectively (Fig. 1). Those genes were the targets for promoter characterization and manipulation.

Fig. 1
figure 1

In silico expression profiles (fold expression) of the unitranscripts of the three switchgrass green tissue-specific genes in different tissues. The cDNA sequences of the three switchgrass green tissue-specific genes as well as the PvUbi1 gene, i.e., Pavirv00047797m, Pavirv00033161m, Pavirv00009702m, and Pavirv00038913m, were used to blast the Noble Foundation switchgrass gene expression atlas PviUT V1.2 (http://switchgrassgenomics.noble.org/). AP13CTG19188, KanlCTG00012, AP13CTG07332, and KanlCTG00705 were the unitranscripts of PvLhcb, PvPEPC, PvPsbR, and PvUbi1, respectively. E4-root, whole root system of E4 stage plant; E4-crown, whole crown of E4 stage plant; E4-node, pooled nodes of E4 the tiller; E4-LFB, pooled leaf blade from E4 tiller; E4-LSH, pooled leaf sheath of the E4 tiller; E4i3m, middle 1/5 fragment of internode 3; E4i3mVB, vascular bundle isolated from 1/5 fragment of internode 3; E4i4t, top 1/5 fragment of internode 4; E4i4m, middle 1/5 fragment of internode 4; E4i4b, bottom 1/5 fragment of internode 4; inflo-meristem, inflorescence of rachis, primary and secondary branch meristem initiation stages 0.5–3.0 mm; inflo-floret, inflorescence of glume and floret development stages 10–20 mm; Inflo-REL, inflorescence 50–150 mm; Inflo-PEM, inflorescence > 200 mm

Full size image

Real-time RT-PCR analysis demonstrated that the expression of PvLhcb, PvPEPC, and PvPsbR were highest in leaf blade and leaf sheath, followed by inflorescence, internode, and node, but marginally expressed in root and seed of switchgrass plants grown in the greenhouse at 28 °C under 16-h day/8-h night photoperiods (390 μE/m2/s) at the R1 growth stage (Fig. 2). The relative expression levels of each gene in root and seed were comparable to each other. However, the relative expression levels of PvPEPC in most of the aboveground tissues were much higher than that of the PvLhcb and PvPsbR genes. When compared to the PvUbi1 expression levels, PvPEPC had 104.4-, 43.7-, 12.7-, and 12.0-fold higher expression in the leaf blade, leaf sheath, inflorescence, and internode, respectively (Fig. 2). PvLhcb and PvPsbR in the four aboveground tissues had 3.8–8.8 and 0.9–8.6 times higher expression than that of PvUbi1, respectively (Fig. 2). Moreover, exposure to light with an intensity of 390 μE/m2/s enhanced the expression of PvLhcb, PvPEPC, and PvPsbR in the switchgrass shoot to 38.6, 560.2, and 76.2 times higher than PvUbi1, respectively (Fig. 3). In contrast, the three genes had minimal expression levels when grown in the dark in comparison to the PvUbi1 (Fig. 3). As a result, we concluded these three switchgrass genes are green tissue-specific and highly light-inducible, and their promoters were used for further functional analysis.

Fig. 2
figure 2

Endogenous relative expression levels of the three switchgrass green tissue-specific genes in different tissues measured by real-time RT-PCR. RNA was extracted from switchgrass cv. ‘Alamo’ grown in the greenhouse at 28 °C under 16-h day/8-h night photoperiods (390 μE/m2/s) at the R1 growth stage. Relative quantification was performed using the standard curve method with PvUbi1 as the internal control gene. Bars represent the mean values of three independent replicates ± standard errors (vertical bars)

Full size image
Fig. 3
figure 3

Light induction of the relative expression of the three switchgrass green tissue-specific genes in the shoots of 3-week-old switchgrass cv. ‘Alamo’ seedlings measured by real-time RT-PCR. RNA was extracted from the shoots of switchgrass seedlings grown in the greenhouse at 28 °C under 16-h day/8-h night photoperiods (390 μE/m2/s). Relative quantification was performed using the standard curve method with PvUbi1 as the internal control gene. Bars represent the mean values of three independent replicates ± standard errors (vertical bars). Bars with asterisk are significantly different from the dark treatments at p ≤ 0.05 as calculated by t test. Dark, 3-week-old seedlings growing under 24-h dark conditions at 28 °C; light, 3-week-old seedlings growing under 16/8-h light/dark cycles at 28 °C

Full size image

GUS expression driven by each green tissue-specific promoter in stable transgenic rice

The promoter sequences of the PvLhcb, PvPEPC, and PvPsbR genes (i.e., PvLhcbp, PvPEPCp, and PvPsbRp) obtained from the switchgrass genome were 764, 1878, and 2009 bp in length, respectively (Additional file 1: Figs. S10–S12). The start codon and the translation termination codon of PvLhcb were located 764 and 1153 bp, respectively, downstream from the beginning of the 5′-end of the contig165429 (Additional file 1: Fig. S10). The start codon and the translation termination codon of PvPEPC were located 1977 and 4863 bp, respectively, downstream of the start codon of its nearest upstream gene (Pavir.J09521.1; unknown function) that was reversely orientated on the contig12474 (Additional file 1: Fig. S11). The start codon and the translation termination codon of PvPsbR were located 4350 and 4746 bp, respectively, downstream of the start codon of its nearest upstream gene (Pavir.Fa02068.1; a pentatricopeptide repeat-containing protein) that was reversely orientated on Chr06a (Additional file 1: Fig. S12).

The GUS activities under the control of each of the three green tissue-specific promoters were examined in T0 transgenic rice at two different developmental stages, i.e., the seedling stage and the heading stage. Histochemical GUS analysis of transgenic rice plants at the seedling stage revealed ectopic constitutive GUS expression in leaves, stems, and roots of the three positive controls (Fig. 4). The transgenic rice expressing GUS under the control of PvLhcbp, PvPEPCp, or PvPsbRp had strong GUS expression in leaves and stems, which was comparable to that in the three positive controls (Fig. 4). However, GUS expression driven by each green tissue-specific promoter was only marginally detectable in roots (Fig. 4). Thus, we concluded that PvLhcbp, PvPEPCp, or PvPsbRp have green tissue-specific activity in transgenic rice.

Fig. 4
figure 4

Histochemical GUS analysis of leaf, stem, and root of T0 stable transgenic rice containing each of the three green tissue-specific promoters at the seedling stage. Stable transgenic rice containing each of the promoters of PvUbi2, ZmUbi1, and CaMV 35S were used as the positive controls. The histochemical GUS assay was conducted on at least 10 transgenic lines with the representatives being shown

Full size image

A more sensitive fluorometric GUS assay was then used to measure GUS activities in leaf blade, leaf sheath, stems, and panicles of T0 transgenic rice plants at the heading stage. The root was not included here in the fluorometric GUS analysis due to a slight expression level of each of the three green tissue-specific genes in the roots of non-transgenic switchgrass (Fig. 2) and GUS-reporter transgenic rice (Fig. 4). Transgenic rice containing the 35S::GUS-positive control showed strong GUS expression in leaf blade, leaf sheath, and panicles with a relatively moderate GUS expression in stems (Fig. 5). Transgenic rice containing GUS driven by each of the three promoters showed moderate GUS expression in leaf blades and leaf sheath of about 15.5–38.8% of the 35S activities in both tissues (Fig. 5). All of the three promoters exhibited relatively low GUS activities in stems. In panicles, the GUS activities driven by PvLhcbp and PvPsbRp were relatively moderate, whereas that driven by PvPEPCp was low in panicles (Fig. 5).

Fig. 5
figure 5

Quantitative fluorometric GUS analysis of leaf blade, leaf sheath, stem, and panicles of T0 stable transgenic rice containing each of the three green tissue-specific promoters at the heading stage. The rice plants were grown in the greenhouse at 25–29 °C under 12-h photoperiods (390 μE/m2/s). Stable transgenic rice containing the promoter of CaMV 35S was used as positive control. The fluorometric GUS assay was conducted on at least 10 transgenic lines at the heading stage with the representatives being shown

Full size image

GUS expression driven by 5′-end serial deletions of each green tissue-specific promoter in stable transgenic rice

The 5′-end serial promoter deletions (Fig. 6) were generated to assess the functionality of various portions, including the core promoter regions, of PvLhcbp, PvPEPCp, and PvPsbRp. Histochemical and fluorometric GUS analysis of each of the T0 transgenic rice plants at the seedling stage revealed that each promoter deletion conveyed a strong GUS expression in leaves and stems (Figs. 7, 8, 9, Additional file 1: Figs. S13, S14) and a marginal GUS expression in roots as did its promoter (Figs. 7, 8, 9). All of the shortest promoter fragments (199 to 275 bp in length) conveyed strong GUS expression in the leaves and stems of transgenic rice. The only exceptions came from the 464- and 231-bp-long promoter deletions of PvLhcbp (i.e., PvLhcbp-1 and -2) and the 1210-bp-long deletion of PvPsbRp (i.e., PvPsbRp-2), which conveyed a moderate-to-strong (PvLhcbp-1 and -2) and low (PvPsbRp-2) GUS expression in roots (Figs. 7, 9).

Fig. 6
figure 6

Scheme of the promoter::GUS fusion constructs used in stable rice transformation experiments. a The PvLhcb promoter and its serial deletions driving GUS expression. b The PvPEPC promoter and its serial deletions driving GUS expression. c The PvPsbR promoter and its serial deletions driving GUS expression. Lines represent promoters with promoter length (base pairs; bp) being indicated above each line. NosT, Nos terminator

Full size image
Fig. 7
figure 7

Histochemical GUS analysis of leaf, stem, and root of T0 stable transgenic rice containing each of the serial promoter deletions of PvLhcbp. The histochemical GUS assay was conducted on at least 10 transgenic lines of each construct at the seedling stage with one representative line of each construct being shown

Full size image
Fig. 8
figure 8

Histochemical GUS analysis of leaf, stem, and root of T0 stable transgenic rice containing each of the serial promoter deletions of PvPEPCp. The histochemical GUS assay was conducted on at least 10 transgenic lines of each construct at the seedling stage with one representative line of each construct being shown

Full size image
Fig. 9
figure 9

Histochemical GUS analysis of leaf, stem, and root of T0 stable transgenic rice containing each of the serial promoter deletions of PvPsbRp. The histochemical GUS assay was conducted on at least 10 transgenic lines of each construct at the seedling stage with one representative line of each construct being shown

Full size image

Green tissue-specific expression of the PvMYB4 gene in stable transgenic switchgrass

In order to examine the application of these green tissue-specific promoters, the three promoters were used to individually drive PvMYB4 expression in stable transgenic switchgrass. Real-time RT-PCR analysis of 5 randomly selected T0 lines of each promoter showed up to 164-fold overexpression in shoots, but only fourfold overexpression in roots, compared to the expression of the endogenous PvMYB4 gene in non-transgenic controls (Fig. 10). In contrast to the compromised aboveground growth and the weak and diminished root systems of the transgenic switchgrass ectopically overexpressing the PvMYB4 gene at high levels under the control of the constitutive ZmUbi1 promoter [12], green tissue-specific expression of the PvMYB4 gene in transgenic switchgrass produced normal root systems (Fig. 11). No phenotypic difference was observed in the aboveground tissues between the T0 transgenic switchgrass plants expressing PvMYB4 under the control of each of the three promoters and the non-transgenic switchgrass (Fig. 11). Thus, aboveground tissue-specific overexpression of PvMYB4 is an effective strategy to target cell wall phenotype while producing wild-type-like root systems in transgenic switchgrass.

Fig. 10
figure 10

Relative expression of the PvMYB4 gene in the shoot and root of T0 stable transgenic switchgrass under the control of each of three green tissue-specific promoters measured by real-time RT-PCR. RNA was extracted from at least 10 transgenic lines of each construct at the E5 growth stage with 5 representative lines of each construct being shown. Relative quantification was performed using the standard curve method with PvUbi1 as the internal control gene. Bars represent the mean values of three independent replicates ± standard errors (vertical bars). N.T., non-transgenic switchgrass

Full size image
Fig. 11
figure 11

Representative images of the shoots and root systems of pot-grown transgenic switchgrass overexpressing the PvMYB4 gene under the control of each of the three green tissue-specific promoters. The PvMYB4 expression levels under the control of the ZmUbi1 promoter in transgenic switchgrass were low in line 6, moderate in line 8, and high in line 1 [12], and used as the controls. Plants were grown in 4-l pots and images were taken at the R1 growth stage. N.T., non-transgenic switchgrass

Full size image

Total lignin content and S/G lignin monomer ratios were then examined from whole tillers of each of the five randomly selected T0 lines of each promoter using pyrolysis molecular beam mass spectrometry (py-MBMS) of cell wall residues (CWRs). In comparison to the non-transgenic switchgrass, green tissue-specific expression of PvMYB4 resulted in a significant reduction in total lignin content by 1.9–7.4% in one line of the promoters PvLhcbp (line 6) and PvPsbRp (line 24), and in two lines of the promoter PvPEPCp (lines 5 and 10) (Fig. 12a). Green tissue-specific expression of PvMYB4 also resulted in a significant reduction in the S/G lignin monomer ratios by 2.9–9.9% in a total of 10 lines of the three promoters, i.e., PvLhcbp (4 lines), PvPEPCp (3 lines), and PvPsbRp (3 lines) (Fig. 12b), including the same four lines with significantly less total lignin content (Fig. 12a). The three lines of both PvPEPCp and PvPsbRp had comparable reduction in levels of both total lignin content and the S/G lignin monomer ratios, which were greater than the reduction levels in both traits of the four lines of PvLhcbp.

Fig. 12
figure 12

Measurements of total lignin content, S/G lignin monomer ratios, and total cell wall sugar release from the whole tillers of T0 stable transgenic switchgrass under the control of each of three green tissue-specific promoters. The whole tillers were harvested at the E5 growth stage with 5 representative lines of each construct being shown. Total lignin content and S/G lignin monomer ratios were measured using pyrolysis molecular beam mass spectrometry (py-MBMS) of cell wall residues (CWR). Sugar release efficiencies were determined by hot water pretreatment of CWR, followed by enzymatic hydrolysis. Bars represent the mean values of three independent replicates ± standard errors (vertical bars). Bars with asterisk are significantly different from controls at p ≤ 0.05 as calculated by t test. N.T., non-transgenic switchgrass

Full size image

Sugar release efficiencies were determined by hot water pretreatment of cell wall residues, followed by enzymatic hydrolysis. All 10 lines including the three different promoter constructs that endowed significant reduction in S/G lignin monomer ratios exhibited significantly higher (9.3–16.2%) release of total sugars compared with the non-transgenic switchgrass (Fig. 12c). Contrasting results in this regard were from 3 lines: PvPEPCp (line 10) and PvPsbRp (lines 3 and 24), which did not show significant increase in total sugar release than the non-transgenic switchgrass. In addition, lines 6 and 13 of PvPEPCp, which did not produce significant reduction in S/G lignin monomer ratios, had significantly increased total sugar release (Fig. 12b, c). Interestingly, the increase in the total sugar release was comparable among all of the above-mentioned 9 lines using the three promoters (Fig. 12c). These 9 lines also exhibited significant increase in xylose release even though only 5 out of the 9 lines had significant increases in glucose release (Additional file 1: Table S1).

Discussion

The green tissue-specific promoters described here confer high levels of light-inducible gene expression in aboveground tissues, providing an effective means to avoid the side effects caused by transgene expression from constitutive promoters. Here we identified and functionally characterized three switchgrass promoters that were highly green tissue-specific and light-regulated. Their expression patterns were very similar to those of their rice homologs [53,54,55,56,57]. To our best knowledge, these are the first characterized green tissue-specific promoters from switchgrass.

The expression levels of PvPEPC detected by real-time RT-PCR were much higher than those of the other two promoters (PvPsbR and PvLhcb), whereas PvPEPC expression level detected by in silico expression analysis was much lower than the other two promoters; PvPEPC expression levels detected by the GUS reporter in transgenic rice were comparable to that of the other two promoters. This discrepancy might be a result of inaccuracy of in silico expression analysis. It might also be because PvPEPCp provided only a portion of its native promoter strength, partially due to the absence of its 5′UTR. Since PvPEPCp comprised the region between the start codons of its own and its nearest upstream gene (Pavir.J09521.1) that was reversely orientated, the PvPEPC promoter could potentially be a bidirectional promoter. Further experiments are needed to examine the bidirectional functionality of PvPEPCp. In addition, the promoter strength of the three green tissue-specific genes in the aboveground tissues of transgenic switchgrass at the heading stage was dramatically lower than that of the 35S promoter, indicating a developmental variation in strength of the three promoters.

Minimal core promoter regions can be used for parts’ selection for synthetic promoter engineering [58,59,60]. Most of the 5′-end serial deletions of each promoter provided the same promoter activities as their promoters in transgenic rice, and even the minimal promoter regions tested (199–275 bp in length) were sufficient for providing strong green tissue-specific expression in monocots. These minimal promoters might serve as desirably compact core promoters in monocots, and could be used together with green tissue-specific motifs [50, 51, 61] for synthetic promoter engineering in cereal crops. The findings that the two deletions of PvLhcbp (i.e., PvLhcbp-1 and -2) and one deletion of PvPsbRp (i.e., PvPsbRp-2) conveyed some GUS expression in roots (Figs. 7, 9) indicate that the promoter regions upstream of PvLhcbp-1 and PvPsbRp-2 might contain sequences inhibiting root expression of both promoters, whereas PvLhcbp-1 and -2, and the first 590-bp-long region of PvPEPCp2 might contain sequences conferring root-preferential expression.

The application of these green tissue-specific promoters in transgenic switchgrass demonstrated that all three promoters restricted PvMYB4 expression primarily to the aboveground tissues, resulting in maintenance of normal growth of root systems. Surprisingly, we found that the green tissue-specific expression of PvMYB4 affected the reduction in total lignin content less than the reduction in S/G lignin monomer ratios. This is contrary to what was found for constitutive PvMYB4 overexpression by the ZmUbi1 promoter in transgenic switchgrass, which resulted in a greater reduction in total lignin content than in S/G ratio [12, 17]. This discrepancy could reflect the strong activity of the ZmUbi1 promoter (and hence the PvMYB4 gene expression) throughout plant development in cell types that do not normally produce secondary cell walls, compared with the green tissue promoters that are only highly active in maturing aboveground tissues that typically form secondary cell walls.

Moreover, the green tissue-specific expression of PvMYB4 resulted in a smaller increase in total sugar release than observed following constitutive overexpression of PvMYB4 by the ZmUbi1 promoter [12]. The green tissue expression of PvMYB4 also led to more xylose release than glucose release, contrary to constitutive PvMYB4 overexpression by the ZmUbi1 promoter [12]. It is worthwhile to point out that we did not find a linear correlation between the PvMYB4 expression levels and lignin content, S/G ratio, or sugar release. This observation could be attributed to non-linear impacts of changes in the PvMYB4 transcription factor on the network of downstream genes in the lignin pathway.

The identification and functional characterization of these three green tissue-specific promoters in switchgrass provide novel tools for monocot genetic engineering and synthetic promoter development when aboveground tissues are targeted. Green tissue-specific expression of PvMYB4 or other transgenes provides a highly effective strategy for crop trait improvement while maintaining unmodified root systems.

Conclusions

We identified and functionally characterized three green tissue-specific promoters from switchgrass in transgenic rice and switchgrass plants; these promoters could be useful for genetic engineering of monocots when aboveground specificity is preferred. Green tissue-specific expression of PvMYB4 is an effective strategy for improvement of transgenic biofuel feedstocks.

Methods

Switchgrass sequence analysis and in silico expression analysis

The amino acid sequences of the rice Lhcb, PEPC, and PsbR genes were used as the query sequences to BlastP against the switchgrass genomic DNA database (V0.0) on the Phytozome website (www.phytozome.net) in order to obtain their homologous sequences in the switchgrass genome. The International Rice Genome Sequencing Project (IRGSP) gene IDs of the three rice Lhcb genes OsLhcb1-1, OsLhcb1-2, and OsLhcb2-1 were Os09g17740 [54, 55, 57], Os1g41710 [54], and Os03g39610 [55], respectively. The IRGSP gene IDs of the five plant-type rice PEPC genes Osppc1, 2a, 2b, 3, and 4 were Os02g0244700, Os08g0366000, Os09g0315700, Os01g0758300, and Os01g0208700, respectively [56]. The IRGSP gene IDs of the three rice PsbR genes OsPsbR1, 2, and 3 were Os07g05360, Os07g05365, and Os08g10020, respectively [53]. The deduced amino acid sequences of the switchgrass homologs of these rice genes were aligned together with their respective query protein sequences using ClustalX 2.0 [62] (Additional file 1: Fig. S1). The promoter sequences (just upstream the start codon), genomic DNA sequences, and cDNA (including 5′ and 3′UTR) sequences of each switchgrass homolog were obtained from Phytozome. The unitranscript entry of each switchgrass homolog was obtained from the Noble Foundation switchgrass gene expression atlas PviUT V1.2 (http://switchgrassgenomics.noble.org; [63]) using each switchgrass cDNA sequence as the query. The switchgrass gene expression profiles were then obtained from the Noble Foundation switchgrass gene expression atlas (https://switchgrassgenomics.noble.org/download_seq.php; [63]).

Characterization of target gene expression in non-transgenic switchgrass

Switchgrass cv. ‘Alamo’ was used for tissue-specific RNA and genomic DNA extraction. Switchgrass plants were grown in the greenhouse at 28 °C under 16-h day/8-h night photoperiods (390 μE/m2/s). The dark/light experiments were conducted by germinating cv. ‘Alamo’ seeds at 28 °C and growing seedlings under 16-h day/8-h night photoperiods (390 μE/m2/s) or without light for three weeks until RNA extraction.

Switchgrass genomic DNA extraction, PCR amplification, and vector construction

Switchgrass genomic DNA was extracted from the leaf blade at the R1 growth stage [64] using a CTAB method [65]. The genomic DNA was used as the template for PCR amplification of the promoter regions of each green tissue-specific gene using sequence-specific primers (Table 1). The PCR product of each gene was gel purified and cloned into the pCR™8/GW/TOPO® vector (Invitrogen™, Carlsbad, CA, USA) for sequencing confirmation. Each promoter was subcloned upstream of the uidA (GUS) gene in the pMDC162 vector with the help of the Gateway® LR Clonase™ II enzyme (Invitrogen™, Carlsbad, CA, USA) for sequence confirmation and then stable rice transformation (see below).

Table 1 Primer sequences used in this study
Full size table

Similarly, each 5′-end serial deletion of each green tissue-specific promoter was PCR amplified and cloned into pCR™8/GW/TOPO® for sequencing confirmation, and then subcloned into pMDC162 for stable rice transformation. The switchgrass Ubiquitin 2 (i.e., PvUbi2), maize Ubiquitin 1 (i.e., ZmUbi1), and CaMV 35S promoters were also cloned into pMDC162, which was used as the positive control vector.

Switchgrass RNA extraction and real-time RT-PCR

Total RNA was extracted from leaf blade, leaf sheath, inflorescence, internode, node, root, and seeds of switchgrass (cv. Alamo) or shoots and roots of transgenic switchgrass expressing PvMYB4 under the control of each green tissue-specific promoter at the E5 stage using Tri-Reagent (Molecular Research Center, Cincinnati, OH, USA). RNA quality was analyzed with an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). One µg of RNA from each tissue was treated with DNase I (Invitrogen™, Carlsbad, CA, USA) and then used for reverse transcription with the High-Capacity cDNA Reverse Transcription kit (Applied Biosystems, Foster City, CA, USA).

The relative expression of each endogenous green tissue-specific gene in each switchgrass tissue or the PvMYB4 transgene in the shoots or roots of transgenic switchgrass was quantified by real-time quantitative reverse transcription PCR (real-time RT-PCR) using sequence-specific primers (Table 1). A switchgrass Ubiquitin gene (i.e., PvUbi1) was used as the internal control (Table 1) [29]. The real-time RT-PCR reactions were conducted using the Power SYBR Green PCR master mix (Applied Biosystems, Foster City, CA, USA) on a 7900HT Fast Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). The standard curve method was used for relative expression analysis normalized by PvUbi1 [29].

Rice transformation

Seeds of japonica rice (O. sativa L.) cv. TP309 were provided by the USDA National Plant Germplasm System. All of the promoters and serial deletions driving GUS expression were introduced into cv. ‘TP309 by Agrobacterium-mediated transformation [66]. Transgenic rice plants were grown in growth chambers at 27 °C under 12-h photoperiods for 2 weeks before being transferred to a greenhouse and grown at 25–29 °C under 12-h photoperiods.

Histochemical GUS analysis in transgenic rice

Histochemical GUS assay was performed according to the published protocol [67]. The leaf, stem, and root of each individual plant were incubated in GUS staining solution (200 mM potassium phosphate at pH 7.0, 0.1% Triton X-100, 1 mg/ml X-Gluc, 10% DMSO) at 37 °C for 6–10 h. After staining, the samples were bleached with 70% (v/v) ethanol, and images were taken under a dissecting microscope (Fisher Scientific™ Stereomaster™ Track Pole, Pittsburgh, PA, USA) using a digital camera (Infinity X-32, Lumenera Corporation, Ottawa, ON, USA). For each construct, at least ten independent transgenic lines were subjected to histochemical GUS assays.

Fluorometric GUS assay in transgenic rice

Quantitative fluorometric assay for GUS activities was conducted according to the published protocol [67]. Total protein concentration was quantified by the Bradford assay [68]. A fixed excitation (365 nm)–emission (460 nm) wavelength fluorometer was used to determine the relative fluorescence units of each sample for three times at intervals of 10 min. Calibration was performed by reading 1000 units for 100 pmol of 4-methylumbelliferone (4-MU) using a 100 nM MU solution. Results were expressed as pmol 4-MU produced (mg protein)/min.

Switchgrass transformation

Each of the three green tissue-specific promoters was individually cloned into the 5′-end of the PvMYB4 gene [17] with the octopine synthase terminator (OCST) being the terminator. Each cassette (individual promoter-PvMYB4-OCST) was subcloned into pMDC99 for sequencing confirmation and stable switchgrass cv. ‘Alamo’ transformation [31].

Analysis of lignin content and composition and cell wall sugar release in transgenic switchgrass

Lignin content, composition, and sugar release from cell wall residues were measured as previously described [69]. Specifically, 300 mg shoot samples were pyrolyzed at 500 °C in 80-µl stainless steel cups using an Extrel single-quadrupole molecular beam mass spectrometer. Lignin content was determined from the relative intensities of the peaks corresponding to the lignin monomers, while S/G ratio was calculated by dividing the sum of the intensities of the syringyl peaks by the sum of the intensities of the guaiacyl peaks.

Sugar release was determined by high-throughput pretreatment and enzymatic hydrolysis [12, 70]. Amylases were used to remove soluble extractives and starch from the biomass samples, followed by ethanol extraction in a Soxhlet extractor [71]. The resulting cell wall residues were loaded into custom-made 96-well metal plates in triplicate. Samples containing ~ 1.7% solids in water (w/w) were pretreated with condensing steam at 180 °C for 17.5 min. Then, enzymatic hydrolysis was conducted by incubation at 50 °C for 70 h with Ctec2 enzyme cocktail (Novozymes North America, Franklinton, NC) at 70 mg protein/g biomass. Glucose and xylose released into the liquid were quantified by colorimetric assays (Megazyme Intl., Bray, Ireland).

Statistical analysis

Analysis of variance (ANOVA) was performed for statistical analyses (p < 0.05) (SAS 9.2 for Windows; SAS Institute, Cary, NC).