Construction and screening of a switchgrass fosmid library
Panicum virgatum (cv. Alamo) leaf tissue was used to construct a fosmid library (unpublished data). In order to screen the library for polyubiquitin genes, P. virgatum EST sequence data (from JGI, Walnut Cove, CA) were aligned to genomic DNA sequences from rice and maize (NCBI). From these alignments, primers were designed to amplify fragments of 700-760 bp in size. Sequence specific primers (5'-TBACYGGMAAGACBATHACY-3', 5'-TCCTTYTGRATGTTRTARTC-3') were then used to screen the library. Fosmids identified to contain polyubiquitin genes were grown in 50-ml cultures containing 50 μl of fosmid induction solution (Epicentre Biotechnologies) to increase copy number, as per manufacturer's protocol. Nuclear-free fosmid DNA was extracted using the Qiagen Large Construct Kit. Approximately 10 μg of fosmid DNA [66 ng μl-1] was sheared to 2-10 kb using the Standard Hydroshear Shearing Assembly (Genomic Solutions) for 20 cycles at a speed code of 16. Sheared fragments between 3 - 8 kb were excised and shotgun libraries were built as described . Ten clones were randomly picked from the sub-clone library and digested with EcoRI to determine the average insert size as quality control. A total of 384 sub-clones were sequenced from both directions using ABI PRISM BigDye Chemistry (Applied Biosystems, Foster, CA) and run on an ABI 3730. The sequences were assembled using Phred/Phrap and annotated in Apollo . The cumulative data represent an approximate 13-fold coverage of each fosmid. Fosmid Pv9G7B5 contained two polyubiquitin genes in tandem and in the same orientation, both of which showed ≥ 99.5% sequence identity to switchgrass ESTs from callus, early floral development, late floral development, root, and stem tissues. Contig Pv9G7B5 was used for further isolation of switchgrass ubiquitin promoters.
Predictions were made for the location of the ubiquitin promoters and genes within the Pv9G7B5 fosmid using FGENESH  and GENSCAN  and further confirmed using blastn in GenBank and aligned with homologous ubiquitin sequences from other plant species in AlignX (Invitrogen, Carlsbad, CA). Based on these results, primers were designed to produce amplicons of PvUbi1 (1991 bp) and PvUbi2 (1861 bp) upstream of the predicted transcription start site (Additional file 1, Table S1). Identification of putative regulatory cis-elements within the promoter regions of PvUbi1 and PvUbi2 was performed using the PlantCARE database (http://bioinformatics.psb.ugent.be/webtools/plantcare/html) .
The full-length cDNAs (including 5'UTRs, coding sequence and 3'UTRs) of the PvUbi1 and PvUbi2 genes were identified using the 5'RACE-PCR and 3'RACE-PCR, respectively, in the GeneRacer™ kit (Invitrogen, Carlsbad, CA, USA). Total RNA extractions from leaves of switchgrass cv. Alamo were performed using the TRI reagent (MRC, Cincinnati, OH). Approximately 3 μg of total RNA were used for reverse transcription to generate cDNA. To remove trace contamination of genomic DNA, RNA was treated with DNase I according to manufacturer's instructions (Promega, Madison, Wisconsin, USA). The resulting 5' and 3'UTRs of cDNA of both genes were amplified with the GeneRacer™ kit and cloned into pCR®8/GW/TOPO® vector (Invitrogen) for sequence confirmation and analysis. The primers are listed in Additional file 1, Table S1.
Quantitative reverse transcriptase PCR (qRT-PCR)
Levels of PvUbi1 and PvUbi2 mRNA abundance were measured using quantitative reverse transcriptase PCR (qRT-PCR) in a variety of switchgrass tissues. Flower, leaf, stem and root tissues of three-month-old greenhouse-grown switchgrass (cv. Alamo), and callus generated from inflorescences of a switchgrass genotype (Alamo 2)  were used for RNA extraction. Total RNA was isolated using Tri-Reagent (Molecular Research Center, Cincinnati, OH), and DNA contamination was removed with DNase treatment (Promega, Madison, WI) following the manufacturer's instructions. A switchgrass actin gene (PvAct) was used as an internal control. Specific primers to the corresponding genes were designed (Additional file 1, Table S1) that amplify a single product for each corresponding gene, as confirmed by the melting temperature of the amplicons and gel electrophoresis. Approximately 3 μg of the total RNA from three independent experiments were synthesized into first strand cDNA using the High Capacity cDNA Reverse Transcription kit (Applied Biosystems, Foster City, CA) and qRT-PCR was conducted in triplicate using Power SYBR Green PCR master mix (Applied Biosystems) according to the manufacturer's protocol. Relative quantification was performed using the standard curve method, and transcript accumulation of each gene was normalized to the quantity of expressed switchgrass actin gene. For quality assurance purposes, only qRT-PCR assays that resulted in standard curves with the following parameters were considered: 1) linear standard curve throughout the measured area, 2) standard curve slope between -3.5 and -3.2, and 3) R2 value above 0.99.
Expression vector construction
All promoters (ZmUbi1, OsAct1, CaMV 35S, 2x35S, PvUbi1, PvUbi1+3, PvUbi1+9, PvUbi2, PvUbi2+3, PvUbi2+9) were amplified with specific primer sets shown in Additional file 1 (Table S1) and cloned into pCR8/GW/TOPO (Invitrogen, Carlsbad, CA). The PvUbi1 and PvUbi2 promoter variants were derived from the Pv9G7B5 contig mentioned above, the ZmUbi1 promoter from pAHC25 , the OsAct1 promoter from pCOR113 , the CaMV 35S promoter from pBin-m-gfp5-ER , and the 2x35S promoter from pMDC32 . DNA was confirmed by restriction enzyme digests for orientation, and clones containing the proper orientation were sequence-verified at the University of Tennessee Molecular Biology Resource Facility. These amplified promoter regions were introduced from the pCR8/GW/TOPO backbone into the binary vectors pGWB533 and pGWB535  using the Gateway® LR Clonase® II enzyme mix (Invitrogen). The pGWB533 vector contains the Gateway® cassette upstream of the uidA coding region (GUS), resulting in promoter:GUS fusion constructs used for initial promoter analysis and tobacco transformations. For comparison of different promoters in switchgrass and rice, the pCR8/GW/ZmUbi1 vector and the pGWB535 vector (containing the Gateway® cassette upstream of the firefly luciferase coding region (LUC)) were LR recombined and the resulting ZmUbi1:LUC cassette was cloned along with the Gateway® cassette upstream of GUS (cloned from pGWB533) and termed pHLucGWgus. The reporter gene cassettes were assembled, sequenced and annotated using Geneious v5.0.3 software . Each unique promoter described above was LR recombined into the Gateway-compatible site of the pHLucGWgus vector upstream and in the correct frame for GUS protein synthesis and sequence verified.
Plant materials and tissue culture
Switchgrass cv. Alamo genotype ST1 was provided by Zeng-Yu Wang from the Noble Foundation . Plants were maintained in the greenhouse by pruning tillers that matured beyond the boot stage  in a 42% sand, and 58% Fafard 3B soil mix (Conrad Fafard, Inc., Agawama, MA) in 12-liter plastic pots. Growth conditions consisted of a 12-h light/12-h dark cycle under 400-watt halide lamps. The greenhouse temperatures ranged from 20-27°C. Plants were watered daily, and fertilized weekly with 0.45 kg of Peters® Professional All Purpose Plant Food (St. Louis, MO) per 379 liters of water. The last culm node of switchgrass produced immature inflorescences at the E2-R0 stages . Culm nodes were identified as previously described by Alexandrova et al. . The 6.5-cm explants were surface-sterilized with 70% EtOH for 1 minute with gentle agitation. Explants were then placed in 15% Clorox® v/v supplemented with 0.01% Tween-20 (Fisher Scientific, Pittsburgh, PA, USA) and gently agitated for 3 minutes. All tissues were then rinsed three times at 2-minute intervals. Sterilized internodes were cut in half longitudinally  and explants were placed cut-side down on solid Murashige and Skoog (MS) medium  supplemented with B5 vitamins, 5 μM BAP and 3% sucrose. Explants were incubated in a growth chamber at 25°C, with cool-white fluorescent lighting (66-95 μE m-2 s-1) 16-h light/8-h dark cycle for 14 days. After 14 days of culture, immature inflorescences were used to initiate embryogenic callus cultures. The inflorescences were dissected out and cut to obtain sections of rachis tissue measuring 1 cm in length. Inflorescence pieces were placed on solid N6E medium  and incubated at 27°C in the dark with subculturing at three-week intervals. After the second subculture, callus was separated from the inflorescences and arranged in a 5 × 5-grid pattern on plates. Friable embryogenic callus tissue was bulked for eight months with subcultures at three-week intervals and used in particle bombardment experiments.
Seeds of rice cv. Taipei 309 were provided by the USDA National Plant Germplasm System. Kernels from dehusked seeds were surface-sterilized in 70% EtOH for 2 minutes at 100 RPM. Kernels were then transferred to a 60% Clorox® v/v supplemented with 0.01% Tween-20, stirred for 30 minutes and rinsed three times with H2O for two minutes. Sterilized kernels were dried, arranged in a 5 × 5-grid on modified NB medium (MNB) as per Chen et al.  and incubated in the dark at 27°C. Prior to particle bombardment, rice callus was induced, selected, and maintained as previously described for 5 months with transfers at 3-week intervals . All switchgrass and rice media were solidified with 2.5 g l-1 Gelzan™ (Caisson Laboratories, North Logan, UT, USA) and brought to pH 5.8 prior to autoclaving. Cultures were sealed in Petri dishes with 3M Micropore™ tape (St. Paul, MN, USA).
DNA particle bombardment of switchgrass and rice callus
Transient expression assays of Taipei 309 and ST1 embryogenic callus cultures were conducted following biolistic transformation using the Bio-Rad PDS-1000 (Bio-Rad Laboratories, Hercules, CA). The PDS-1000 was used for plasmid delivery with 7,584 kPa (1,100 psi) rupture disks, a microcarrier flight distance of 6 cm and a vacuum of 97 kPa (27 in) Hg [86, 87], with all hardware and reagents produced by Bio-Rad. Microprojectile preparation essentially followed Trick et al.  with the DNA amount decreased from 625 ng to 300 ng per bombardment, and 10 mg of 0.6 μm diameter gold (Au) particles used instead of 12 mg of 1 μm particles. Each bombardment consisted of a 10 μl aliquot placed on the macrocarrier and allowed to dry completely. Switchgrass and rice callus cultures were incubated for 6 h prior to bombardment on N6 osmotic medium with 0.6 M osmoticum (http://www.agron.iastate.edu/ptf/protocol/Callus%20bb.pdf), or 0.6 M NB osmotic medium , respectively. Each vector was used to bombard six replicate plates with 50 callus pieces per plate. To test the functionality of the promoter vectors and the validity of the bombardment assay, ten rice calli were selected from each of the first two replications and histochemically stained for observation of GUS. The five rice calli with the highest level of expression were selected and photographed (Additional file 1, Figure S3).
Stable transformation of rice
Stable transformations of rice were performed as described above for transient expression assays with three exceptions: three-month-old rice callus cultures were used and 150 ng of the pHLucGWgus vectors (containing the PvUbi1, PvUbi1+3, PvUbi1+9, PvUbi2, PvUbi2+3, PvUbi2+9, ZmUbi1, and CaMV 35S promoters) were used per bombardment. Rice callus cultures were incubated for 6 h pre- and 18 h post-bombardment on 0.6 MNB osmotic medium . Rice callus cultures were selected on MNBH50 as described  to ensure independent events were recovered. Positive transgenic calli were regenerated as described by Broothaerts et al.  on RGH6 medium solidified with Phytagel (6 g l-1) without selection and resulting plantlets were rooted for four weeks on 1/2 MS medium supplemented with B5 vitamins and hygromycin B (50 mg l-1) solidified with 3 g l-1 Gelzan™ in Magenta® GA-7 Plant Culture vessels. Regeneration and rooting occurred under a 23-h light/1-hr dark photoperiod provided by cool-white fluorescent light (66-95 μE m-2s-1) at 26°C. Prior to being moved to the greenhouse, a root sample was harvested and GUS-stained for all transgenics, and an untransformed control was regenerated without selection. Plants were allowed to grow in the greenhouse for approximately two months prior to harvesting tissue for GUS staining of leaf and stem tissues.
MUG and LUC assays
Following bombardment, gene expression was analyzed using luciferase and MUG assays. Thirty-six hours post-bombardment, 25 calli per replicate were ground in 50 μl of 1× lysis buffer (1× LB) . For the first two replications, five calli were stained for GUS . Upon lysing the cells, 350 μl of additional 1× LB were added to each sample. The cell lysates were centrifuged at 13,000 g for five minutes at ambient conditions; the tubes were then rotated 180° and spun again. The soluble protein extracts produced from each sample were used for 4-methylumbelliferyl β-D-glucuronide (MUG) and luciferase assays [63, 64]. For MUG assays, 50-μl of protein extract were added to 50 μl of assay buffer (1 mM MUG in 1× LB). Reactions were incubated at 37°C for 24 hours, and subsequently terminated with 100 μl of stop buffer (0.2 M Na2CO3 in H2O). Samples were read in duplicate with the BioTek® Synergy 2 fluorometer (BioTek, Winooski, VT, USA) at an excitation wavelength of 360/40 nm and an emission wavelength of 460/40 nm. The fluorometer was calibrated with 4-methyl umbelliferone (MU) standards in stop buffer. MUG results were expressed as micromole MU released hour-1. Luciferase activity was quantified twice for each replicate using 25 μl of protein extract. For each sample reading, 25 μl of sample extract in 1× LB buffer were diluted in 75 μl of Glo-lysis buffer, mixed with 100 μl of ONE-Glo™ Luciferase Assay buffer (Promega Corporation, Madison, WI, USA) and allowed to incubate at room temperature for 5 minutes. Non-specific GUS and luciferase activity was corrected, and normalization of the MUG data was accomplished using luciferase activity as previously described . Each unique GUS cassette allowed the measurement of gene expression to be quantified. The strength of each promoter was reported relative to that of the CaMV 35S control, normalized to 1, to create a dimensionless value of promoter strength [19, 62].
Agrobacterium-mediated transformation of tobacco
The vectors to be tested were transformed into A. tumefaciens EHA105 as previously described . EHA105 cells were maintained in liquid YEP medium and all incubations were performed at 28°C. Tobacco cv. Xanthi seeds were surface-sterilized, transformed, and regenerated using 50 mg l-1 hygromycin for selection according to published methods .
Plant tissues were stained for GUS activity in microwell plates and placed at 37°C overnight as described . For tobacco, intact tissue stains were made homogenous by vacuum infiltrating in solution for 30 minutes. For optimal visualization of stained tissues, chlorophyll was removed by repeatedly washing the tissue with a solution containing a 3:1 ratio of EtOH and acetic acid, ultimately storing tissue samples in 70% EtOH for imaging. For rice tissues, GUS staining assays were completed using a modified GUS buffer  brought to pH 7 , and chlorophyll was removed from the tissues as described by Cervera .
Data for relative expression of promoters using the MUG and LUC assays were subjected to Levene's test  to check for homogeneity of variance using the software package JMP® (Version 8.0.2 SAS Institute Inc., Cary, NC). When p ≤ 0.05, the data were considered to have unequal variances and were subjected to a square root transformation prior to ANOVA. Data sets with equal variances were subjected to ANOVA. If a significant difference was detected (p ≤ 0.05) using ANOVA, the least significant difference test (LSD) was employed to analyze the data for significant differences between treatments within an experiment (p = 0.05).
GenBank accession numbers
The PvUbi1 and PvUbi2 genes containing the promoters, 5' UTR exons and introns, polyubiquitin ORFs, and the 3' UTR regions have been deposited in GenBank (accession numbers HM209467 and HM209468, respectively).