Down-regulation of glycosyltransferase 8D genes in

Tree Physiology 31, 226–236 doi:10.1093/treephys/tpr008

Research paper

Down-regulation of glycosyltransferase 8D genes in Populus trichocarpa caused reduced mechanical strength and xylan content in wood Quanzi Li1,4, Douyong Min2, Jack Peng-Yu Wang1, Ilona Peszlen2, Laszlo Horvath2, Balazs Horvath2,3, Yufuko Nishimura1, Hasan Jameel2, Hou-Min Chang2 and Vincent L. Chiang1,2 Biotechnology Group, Department of Forestry and Environmental Resources, North Carolina State University, Campus Box 7247, 2500 Partners II Bldg, 840 Main Campus Drive, Raleigh, NC 27695, USA; 2Department of Forest Biomaterials, North Carolina State University, Raleigh, NC 27695, USA; 3Present address: CaptiveAire, Inc., Youngsville, NC 27596, USA; 4Corresponding author ([email protected]) Received November 23, 2010; accepted January 29, 2011; handling Editor Chunyang Li

Members of glycosyltransferase protein families GT8, GT43 and GT47 are implicated in the biosynthesis of xylan in the secondary cell walls of Arabidopsis. The Arabidopsis mutant irx8 has a 60% reduction in xylan. However, over-expression of an ortholog of Arabidopsis IRX8, poplar PoGT8D, in Arabidopsis irx8 mutant could not restore xylan synthesis. The functions of tree GT8D genes remain unclear. We identified two GT8 gene homologs, PtrGT8D1 and PtrGT8D2, in Populus trichocarpa. They are the only two GT8D members and are abundantly and specifically expressed in the differentiating xylem of P. trichocarpa. PtrGT8D1 transcript abundance was >7 times that of PtrGT8D2. To elucidate the genetic function of GT8D in P. trichocarpa, the expression of PtrGT8D1 and PtrGT8D2 was simultaneously knocked down through RNAi. Four transgenic lines had 85–94% reduction in transcripts of PtrGT8D1 and PtrGT8D2, resulting in 29–36% reduction in stem wood xylan content. Xylan reduction had essentially no effect on cellulose quantity but caused an 11–25% increase in lignin. These transgenics exhibit a brittle wood phenotype, accompanied by increased vessel diameter and thinner fiber cell walls in stem xylem. Stem modulus of elasticity and modulus of rupture were reduced by 17–29% and 16–23%, respectively, and were positively correlated with xylan content but negatively correlated with lignin quantity. These results suggest that PtrGT8Ds play key roles in xylan biosynthesis in wood. Xylan may be a more important factor than lignin affecting the stiffness and fracture strength of wood. Keywords: glycosyltransferase, Populus trichocarpa, wood formation, xylan

Introduction Wood is important for tree growth because it is a water-conductive and supportive vascular tissue. Wood has been widely used as a raw material for construction, pulp and paper production, and fuel. Wood is also regarded as an important renewable source of energy (Plomion et al. 2001, Ragauskas et al. 2006). Modification of wood composition and properties could reduce the recalcitrance to saccharification for conversion of lignocellulosic biomass to ethanol (Chen and Dixon 2007, Lee et al. 2009a).

Wood formation occurs with the development of xylem and the differentiation of the secondary cell wall, which represents most of the biomass produced by plants. The secondary cell wall is typically composed of three major components: ~50% cellulose, ~30% hemicelluloses and ~20% lignin. In the wood of angiosperms including poplar, xylan is a predominant hemicellulose and accounts for ~20% of the total dry weight of wood. Xylan is a linear polymer with a backbone composed entirely of (1 → 4)-linked β-d -xylose units and is partially substituted by 4-O-methyl-α-d-glucuronic acid (4-O-Me-GlcA)

© The Author 2011. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]

Downloaded from treephys.oxfordjournals.org at D H Hill Library - Acquis Dept S on May 19, 2011

1Forest

Glycosyltransferase 8D genes in wood formation 227 heterologous system, such as yeast or insect cells (Zhong et al. 2005, Persson et al, 2007). In Populus alba × tremula (aspen), a few GT genes have been studied for their involvement in xylan biosynthesis during wood formation. PoGT47C, PoGT8D and PoGT43B show high sequence similarity to FRA8, IRX8 and IRX9, respectively. Transgenic PoGT47C expression in the fra8 mutant of Arabidopsis restored the stem strength and thickness of the secondary wall of fibers and vessels to the wild-type level, indicating that PoGT47C is a functional ortholog of FRA8 (Zhou et al. 2006). The genetic function of PoGT47C in xylan bio synthesis was further revealed in PoGT47C RNAi-silencing transgenic P. alba × tremula plants (Lee et al. 2009a, 2009b). Over-expression of PoGT43B in the irx9 mutant rescued the defects in plant size and secondary wall thickness and partially restored the xylan content, indicating that PoGT43B is a functional ortholog of IRX9 involved in GX biosynthesis. Complementation experiments in parvus Arabidopsis mutant further suggested that three glycosyltransferases, GT8E, GT8F/ PdGATL1.1 and PdGATL1.2, which belong to the E and F groups of the GT8 family, respectively, are functional orthologs of PARVUS (Kong et al. 2009, Lee et al. 2009a, 2009b,). The functions of the poplar GT8D gene family are, however, not clear because over-expression of PoGT8D in irx8 did not affect the mutant phenotypes (Zhou et al. 2007). Here we report the characterization of the two members of the Populus trichocarpa GT8D family, PtrGT8D1 and PtrGT8D2, which have a 97 and 89% nucleotide sequence identity to P. alba × tremula PoGT8D, respectively. The genetic function of the GT8D family was tested in transgenic P. trichocarpa where the expression of both PtrGT8D1 and PtrGT8D2 was suppressed using RNAi. Transgenics with strong down-regulated PtrGT8D1 and PtrGT8D2 expression exhibited a drastic reduction in xylan content, and an increase in lignin quantity in stem wood. These alterations were accompanied by a reduction in wall thickness in stem fiber cells and a more brittle stem phenotype. The dwarf and irregular xylem phenotypes observed in irx8 were not detected.

Materials and methods Plant materials The P. trichocarpa Nisqually-1 plants were grown in 1/2 MiracleGro Soil (Scotts Miracle-Gro products, Maysville, OH, USA) and 1/2 Metro-Mix 200 (Sun Gro, Bellevue, WA, USA) in the greenhouse (17–26 °C, 16 h light/8 h dark cycle with supplemental light of ~300 µE m−2 s−1). All wild-type and transgenic plants used for analysis were 5 months old. After the bark was peeled from the stem, the differentiating xylem tissues were scraped with a single-edge razor from the surface of the peeled log, and phloem tissues were scraped from the inner surface of the bark. All the collected tissues were frozen in liquid nitrogen immediately.

Tree Physiology Online at http://www.treephys.oxfordjournals.org


through α-(1 → 2)-glycosidic linkages (Timell 1967, Shimizu et al. 1976, Johansson and Samuelson 1977, Andersson et al. 1983). A portion of the backbone is acetylated at either C-2 or C-3 of the xylose residues. Golgi-targeted glycosyltransferases are involved in the biosynthesis of xylan and other non-cellulosic polysaccharides (Keegstra and Raikhel 2001, Scheible and Pauly 2004, Lerouxel et al. 2006, Ye et al. 2006). The characterization of Arabidopsis mutants with collapsed xylem has identified genes encoding glycosyltransferases that may be involved in xylan biosynthesis, Arabidopsis Fragile Fiber 8/IRREGULAR XYLEM 7 (FRA8/IRX7), IRX8, IRX9, IRX14, IRX14-L, PARVUS, IRX10 and IRX10-Like (IRX10-L), IRX9, and IRX9-L (Zhong et al. 2005, Bauer et al. 2006, Lee et al. 2007a, 2007b, Pena et al. 2007, Persson et al. 2007, Brown et al. 2009, Wu et al. 2009). Five mutants, fra8, irx8, irx9, irx14 and parvus-3, showed severe dwarf phenotypes, with an ~40–60% xylan content reduction compared with wild type. Although the other two mutants, irx10 and irx10-l, displayed either a weak or no mutant phenotype, a severe phenotype was observed in the irx10 and irx10-l double mutant. IRX10 and IRX10-L are functionally redundant in elongation of the xylan backbone (Brown et al. 2009, Wu et al. 2009). FRA8 is a putative glucuronyltransferase involved in the transfer of GlcA residues onto xylan (Zhong et al. 2005). Further analysis of the fra8 and f8h mutants suggests their roles in the biosynthesis of the reducing end and in the backbone of xylan (Wu et al. 2010). IRX10, IRX10-L, F8H and FRA8 all belong to the GT47 family of glycosyltransferases. IRX8 (GT8D) and PARVUS belong to the GT8 family, which is a large family consisting of seven groups (A–G). PARVUS and IRX8 belong to the C group and D group, respectively. Structural analysis of xylan by NMR showed that the glycosyl sequence was absent in irx7, irx8 and parvus-3 mutants, indicating the involvement of FRA, GT8D and PARVUS in the synthesis of the glycosyl sequence at the GX reducing end (Brown et al. 2007, Pena et al. 2007). PARVUS may be involved in the synthesis of a xylan primer (Lee et al. 2007b). Both IRX9 and IRX14 are members of the GT43 family. Compared with wild type, xylan chain length was significantly reduced in irx9 and irx14 mutants, indicating that, besides IRX10 and IRX10-L, IRX9 and IRX14 are also involved in xylan chain elongation, perhaps through a complex (Brown et al. 2007, Pena et al. 2007). Two other GT43 members, IRX9-L/ I9H and IRX14-L/I14H, have redundant functions with IRX9 and IRX14, respectively, and all four GT43 members form two functionally non-redundant groups essential for GX backbone elongation (Lee et al. 2010, Wu et al. 2010). The xylan s ynthesis-related function of IRX9 and IRX14 was supported by reduced xylan synthase activities in irx9 and irx14 mutants (Brown et al. 2007, Lee et al. 2007a). The biochemical functions of FRA8 and IRX8 have not been demonstrated due to the difficulty of producing their recombinant proteins in a

228 Li et al.

Amplification of gene coding sequences

Southern blotting An 800 bp fragment of PtrGT8D2 was excised with NheI and SpeI from pMT/V5-PtrGT8D2. The fragment was purified with a QIAquick Gel Extraction Kit (Qiagen), and used as a template for preparing the 32P radioisotope-labeled hybridization probe using the DECA Primer II Random Primed DNA labeling kit (Ambion, Austin, TX, USA). Genomic DNA was isolated by a CTAB method (Aitchitt et al. 1993). Twenty micrograms of DNA were digested with EcoRI, HindIII and EcoRV. After electrophoresis and denaturation, the DNA was transferred to a Hybond N+ nylon membrane (Amersham, Piscataway, NJ, USA). Prehybridization was performed at 50 °C for 30 min and hybridization at 50 °C for 2 h in Express Hyb™ hybridization solution (Clontech, Mountain View, CA, USA). The membrane was washed in 2× SSC (20× SSC is 3 M NaCl, 0.3 M sodium citrate, pH 7.0), 0.1% sodium dodecyl sulfate (SDS) at room temperature for 30 min, and in 0.5× SSC, 0.1% SDS at 55 °C for 10 min. Autoradiography was performed at −70 °C for 60 h.

Northern hybridization Total RNA from roots, leaves, phloem (including part of the bark), xylem, stem internodes 1–2 and stem internodes 7–8

Tree Physiology Volume 31, 2011

Quantitative reverse transcription PCR Total of young roots, mature roots, leaves, apex, stems of 1–3 internodes, xylem and phloem was isolated using the CTAB method (Chang et al. 1993). The RNA, treated with RNase-free DNase I and purified with the RNeasy Plant RNA Isolation Kit was reverse transcribed into cDNA with TaqMan reagents. Quantitative reverse transcription polymerase chain reaction (qRT–PCR) was carried out using the Applied Biosystems 7900 HT Sequence Detection System in 25 µl reactions containing 1 ng of RNA, 200 nM each primer and 12.5 µl of SYBR Green PCR Master Mix. The PCR program was 95 °C 10 min, 45 cycles at 95 °C 15 s and 60 °C 1 min, followed by a thermal denaturing cycle that is used to derive the dissociation curve of the PCR product to verify the amplification specificity. The plasmid DNA containing the gene was used as a standard for establishing a quantitative correlation between the copy number of the target transcript molecules and the Ct values (Suzuki et al. 2006). Each sample was run in triplicate. The primers for expression analysis of PtrGT8D1 were 8D1rF (5′-GAATTTATGGACGAAGTC AAGAACAC-3′) and 8D1rR (5′-GCTGCTTCGGTATGCTACTTGA TGCT-3′). The primers for PtrGT8D2 were 8D2-rF (5′-TCGCAA AGGGGAAAGTCCCAGTT-3′) and 8D2-rR (5′-GTATTTGGGACC AAGTGTTTGCAGC-3′). The real-time PCR products were purified with a QIAquick PCR Purification Kit, and 3 µl were used for sequencing to confirm the product.

Promoter-driven glucuronidase expression in tobacco A 2.6 kb DNA fragment containing the 5′ upstream region of PtrGT8D1 was amplified with the primers 5′-GAAGCTTGAACAAA ACAAACTACTTGGATCAAC-3′ and 5′-TGGATCCGGCTATGTCATT ATTTGGCAATCACAC-3′, and a 2.9 kb fragment of the 5′ upstream region of PtrGT8D2 was amplified with 5′-GAAGCTTTG GGTGGAATGGTTCTAGGATGCCAG-3′ and 5′-TGGATCCATTTCAT CATTTGGCAGTCGCACTCA-3′. The promoter fragment was cloned into pBI121, replacing the CAMV 35S promoter, resulting in binary vectors pBI121-PtrGT8D1P and pBI121-PtrGT8D2P, which were introduced into Agrobacterium tumefaciens strain C58. Transformations with Nicotiana tabacum leaf disks were conducted as described by Horsch et al. (1985). Stem sections from 8–12 internodes were cut by hand and immediately put in glucuronidase (GUS) staining solution (50 mM NaHPO4, pH 7.2,


Total RNA from differentiating xylem of P. trichocarpa was isolated with the RNeasy Plant RNA Isolation Kit (Qiagen, Valencia, CA, USA). Two micrograms of total RNA were reverse transcribed to cDNA (Omniscript RT Kit; Qiagen) in 20 µl and then diluted to 200 µl after transcription. The amplification of the PtrGT8D1 and PtrGT8D2 coding sequences was conducted in 50 µl reactions containing 1× polymerase chain reaction (PCR) buffer, 0.4 mM forward and reverse primers, 4 µl of cDNA template, 2 mM dNTP and 2.5 U of PfuUltra highfidelity DNA polymerase (Stratagene, Santa Clara, CA, USA). The PCR reactions were carried out for 32 cycles of 94 °C for 45 s, 56 °C for 45 s, 68 °C for 1 min, followed by 68 °C for 20 min. The primer sequences for PtrGT8D1 amplification are 5′-CGTTATGGAGCAGCTTCATATATCGCCGAGTTTG-3′ and 5′-AG ATGGCCTAATATGACAGCC-3′, and the primer sequences for PtrGT8D2 amplification are 5′-CGTTATGGAGCAGCTTCATAT ATCACCAAGTTTGAG-3′ and 5′-TGTCCTAATATGACAGCCCGTA ATG-3′. The PCR products were purified with the QIAquick Gel Extraction Kit (Qiagen), and ligated into pMT/V5-His-TOPO vectors (Invitrogen, Carlsbad, CA, USA). Sequencing was conducted using an Applied Biosystems 3730XI DNA analyzer. Sequencing primers for PtrGT8D1 are MT forward primer (5′-CATCTCAGTGCAACTAAA-3′), BGH reverse primer (5′TAGAAGGCACAGTCGAGG-3′) and PtrGT8Ds-5s (5′-GTACAG GCATGTAGCATC-3′), and primers for PtrGT8D2 are MT forward primer, BGH reverse primer and primer 8D2-rR (5′-GTATTT GGGACCAAGTGTTTGCAGC-3′).

was isolated using a CTAB method (Chang et al. 1993). Twenty micrograms of total RNA were fractionated by electrophoresis on 1.2% formaldehyde agarose gels. Then, RNA was transferred to a nylon membrane (Fritsch et al. 1989). Prehybridization was performed in Express Hyb™ hybridization solution at 42 °C for 2 h. The probe used is the same as used for the Southern blots. Hybridization was conducted at 42 °C for 48 h. Filters were washed in buffer I (2× SSC, 0.2% SDS) at room temperature for 20 min and in buffer II (0.2× SSC, 0.2% SDS) at 68 °C for 20 min.

Glycosyltransferase 8D genes in wood formation 229 2 mM 5-bromo-4-chloro-3-indolyl-β-glucuronic acid, 0.5 mM K3Fe[CN]6 and 0.5 mM K4Fe[CN]6). After vacuum infiltration for 30 min, the materials were incubated at 37 °C until blue staining appeared, and were observed under a Zeiss (Stemi DV4) microscope.

RNA silencing of PtrGT8D1 and PtrGT8D2

Mechanical properties Two or three ~20 cm long stem sections were cut from the bottom portion of each tree sample. The stem sections were labeled and kept in plastic bags, which contained some water to prevent drying. One segment with a length-to-width ratio of 22 was cut from each stem section. A three-point bending test was conducted to measure the modulus of elasticity (MOE) and the modulus of rupture (MOR) in green condition by an MTS Alliance RF/300 universal mechanical tester following Horvath et al. (2010b).

Light and scanning electron microscopy For quantitative wood anatomy, four transverse microtome sections of 20 µm from the bottom of each stem were cut and stained with safranin to enhance contrast, washed with deionized water, dried on a warm plate at 75 ± 5 °C and mounted on glass slides with Permount. From each slide, five

Cell wall composition analysis The stem from the 15th internode to the bottom was used for the chemical composition analysis using a modified Klason lignin method. Briefly, air-dried wood was ground. Wood particles of 40–60 mesh were extracted with benzene/ethanol (2:1, v/v) for 10 h. In all, 0.1 g of sawdust was hydrolyzed with 1.5 ml of 72% (w/w) H2SO4 at room temperature for 1.5 h. The mixture was diluted with 56 ml of water and autoclaved at 121 °C for 1 h. The mixture was filtered through a fine coarseness crucible and acid-insoluble lignin was determined gravimetrically. The filtrate was used to determine acid-soluble lignin content by UV-VIS absorption at wavelength 205 nm (HP8453E UV-VIS spectrophotometer). The extinction coefficient used was 110 AU l/g cm (Dence 1992). The concentration of sugars (arabinose, rhaminose, galactose, glucose, xylose and mannose) in the filtrate was quantified by ion exchange chromatography (IC) . The IC system (Dionex IC-3000; Dionex) was equipped with an ion exchange PA1 (Dionex) column, a pulsed amperometric detector with a gold electrode, and a Spectra AS 300 autosampler. Prior to injection, samples were filtered through 0.2 µm nylon filters (Millipore), and a volume of 25 µl was loaded. The column was pre-equilibrated with 250 mM NaOH and eluted with Milli-Q water at a flow rate of 1.1 ml min−1 with a gradient of water for 45 min, 200 mM NaOH for 15 min and water for 15 min. Fucose was used as an internal standard.

Results We used the sequences of Arabidopsis IRX8 and P. alba × tremula PoGT8D to blast the P. trichocarpa genome and identified two homologous gene models (JGI ID) 834622 and 569069. The two gene models had 97 and 89% nucleotide sequence identity, respectively, with PoGT8D. Southern blot analysis confirmed that there are two GT8D genes in the P. trichocarpa genome (data not shown). We designated these two genes as PtrGT8D1 and PtrGT8D2, and cloned their coding sequences.

Tree Physiology Online at http://www.treephys.oxfordjournals.org


An RNA-silencing construct with an inverted repeat was prepared following Miki et al. (2005) (Figure 3a). A 680 bp GUS linker was amplified with a pair of primers, 5′-TGACCTCGAGGTCGAC GATATCGTCGTCATGAAGATGCGGAC-3′ and 5′-CTAGACTAGTCC CGGGGGTACC ATCCACGCCGTATTCGGTG-3′, and cloned into the pCR2.1 Vector at XhoI/SacI, resulting in pCR2.1-GL. The inverted repeat sequence consists of chimeric sequences from PtrGT8D1 and PtrGT8D2. A 200 bp fragment was amplified from PtrGT8D1 with the primers 8D1s-F (5′-GTCTAGAGGATCCGCTC TCAAGCTCAGAGAAATG-3′) and 8D1-R (5′-CAAGCGCAGGGAC GAGTTCCG-3′). A 330 bp fragment was amplified from PtrGT8D2 with the primers 8D2-F (5′-CGGAACTCGTCCCTGCGCTTG-3′) and 8D2s-R (5′-GCTCGAGTTTCCACGATTGCAGACGAGC-3′). With the two fragments as templates, the sense strand of the chimeric fragment was generated by overlapping PCR with the primers 8D1s-F and 8D2s-R (Warrens et al. 1996). The antisense chimeric fragment was amplified with the primers 8D1as-F (5′- GGAGCTCCATATGGCTCTCA AGCTCAGAGA A ATG -3′) and 8D2as-R (5′-GACTAGTTTTCCACGATTGCAGACGAGC-3′). The sense and antisense fragments were inserted in pCR2.1-GL at XbaI/XhoI and SpeI/SacI, respectively, to produce pCR2.1sense-GL-antisense. The sense-GL-antisense fragment was excised with XbaI/SacI and cloned into pBI121-PtrGT8D1P behind PtrGT8D1 promoter by replacing the GUS gene. The construct was introduced into A. tumefaciens strain C58 by the freeze thaw method (Holsters et al. 1978) and transferred into P. trichocarpa following the established method (Song et al. 2006).

digital images were selected randomly along the radius at ×200 magnification for the assessment of vessel lumen area fraction (%), vessel lumen diameter (μm), cell wall area fraction (%) and ray area fraction (%). The anatomy was measured using an image analyzer system, which consists of a light microscope (Nikon E200), a 3CCD color video camera (Sony DXC-390) and Image-Pro Plus 4.5 software (Horvath et al. 2010a). Cell wall thickness of fibers was studied on scanning electron microscope images taken at ×3000 magnification using an AIF Hitachi S-3200N SEM (Analytical Instrumentation Facility, NCSU). From each specimen, five areas were randomly selected and images were taken to measure the double cell wall thickness of fibers (μm) using Revolution 1.6.0 software.

230 Li et al. Sequence analysis showed that the PtrGT8D1 cDNA sequence is identical to that of the annotated 834622, and the PtrGT8D2 cDNA sequence has one base different from that of the annotated 569069. PtrGT8D1 and PtrGT8D2 share 91% nucleotide sequence identity and 92% protein sequence identity.

PtrGT8Ds are highly expressed in the differentiating stem xylem of P. trichocarpa

PtrGT8D1 and PtrGT8D2 promoter::GUS fusion in tobacco To drive tissue-specific down-regulation of PtrGT8D expression, we amplified a 2.6 kb promoter fragment of PtrGT8D1 and a 2.9 kb promoter fragment of PtrGT8D2 from P. trichocarpa genomic DNA. The promoter fragments were fused with the β-GUS reporter gene and transformed into tobacco plants to test their activity. Five independent transgenic tobacco plants for each promoter::GUS construct were analyzed. In PtrGT8D1 promoter::GUS transgenic tobacco, strong GUS staining was observed in the stem differentiating xylem, and weak staining was observed in the phloem (Figure 1c). In the leaves, GUS expression was present in a few cells within the vascular bundle (Figure 1d). The PtrGT8D2 promoter::GUS fusion displayed identical GUS expression in tobacco plants to PtrGT8D1 promoter::GUS (data not shown).

RNAi down-regulation of PtrGT8D1 and PtrGT8D2 expression in P. trichocarpa To examine the functions of GT8D genes during wood formation, we knocked down the expression of both PtrGT8D1 and PtrGT8D2 simultaneously. To do this, we prepared a single

Figure 1. Expression analyses of PtrGT8D1 and PtrGT8D2 in wild-type P. trichocarpa. (a) Northern blotting of PtrGT8Ds. Twenty micrograms of total RNAs were loaded per lane. The ethidium bromide-stained total RNA was included as a loading control. The radioisotope (32P)-labeled 800 bp DNA fragments of the 3′ portion of the PtrGT8D2 coding region sequence were used as probes. (b) Quantitative RT–PCR analysis of PtrGT8D1 and PtrGT8D2 expression in P. trichocarpa. Error bars represent standard errors of triplicate assays. (c) Glucuronidase staining in the stems of transgenic tobacco plants containing PtrGT8D1 promoter::GUS transgene. (d) Glucuronidase staining in the leaf petiole of transgenic tobacco plants containing PtrGT8D1 promoter::GUS transgene. PtrGT8D2 promoter::GUS fusions displayed identical expression patterns in tobacco plants to PtrGT8D1 promoter::GUS.

Tree Physiology Volume 31, 2011


Northern blotting and quantitatively real-time PCR analysis of PtrGT8D expression We used RNA gel blots to estimate the transcript abundance of PtrGT8D1 and PtrGT8D2 in P. trichocarpa. Their transcripts were detected in roots, xylem and stem internodes 7–8 and were not detected in leaves, stem internodes 1–4 or stem phloem (Figure 1a). We further used qRT–PCR to distinguish the expression of PtrGT8D1 and PtrGT8D2. Quantitative RT–PCR results showed that, consistent with the results of northern blots, PtrGT8D1 was abundantly expressed in differentiating xylem, and PtrGT8D2 had the highest specificity of expression in xylem among all the tissues analyzed. The absolute transcript level of PtrGT8D1 was ~7 times higher than that of PtrGT8D2 in xylem. Quantitative RT–PCR also detected relatively low transcript levels of PtrGT8D1 in roots, leaves and stem phloem, and did not detect PtrGT8D1 transcripts in apexes and stem (1–3 internode). Low transcript levels of PtrGT8D2 were detected in old roots and stem phloem (Figure 1b) and no expression was detected in

young roots, apexes or leaves. Sequencing of real-time PCR products verified that they are authentic PCR products of PtrGT8D1 and PtrGT8D2, and confirmed that the two pairs of primers could distinguish PtrGT8D1 and PtrGT8D2 transcripts.

Glycosyltransferase 8D genes in wood formation 231

RNAi construct containing a PtrGT8D1-specific sequence and a PtrGT8D2-specific sequence as RNAi-silencing trigger sequences (Figure 2a). The construct was transferred into P. trichocarpa (Song et al. 2006). We obtained five putative transgenic lines. Polymerase chain reaction analyses of the DNA isolated from the five lines confirmed the genome integration of the PtrGT8D sense:Gus linker:antisense fragment in four lines (data not shown). These four transgenic lines are named PtrGT8D-TL2, PtrGT8D-TL3, PtrGT8D-TL5 and PtrGT8D-TL7. Quantitative RT–PCR was used to analyze PtrGT8D1 and PtrGT8D2 expression reduction in transgenic plants. The transcript levels of both PtrGT8D1 and PtrGT8D2 in these four lines were reduced (Figure 2b and c). The transgenic plant with the most reduced expression level of PtrGT8Ds was PtrGT8D-TL2, where the PtrGT8D1 expression level was reduced to 1.9% of that in the wild type and the PtrGT8D2 expression level was reduced to 45% of that in the wild type. PtrGT8D1 expression in PtrGT8D-TL5 and PtrGT8D-TL7 is 5–6% of that in the wild type, and PtrGT8D2 expression is 14–19% of that in the wild type. The transcript levels of PtrGT8D1 and PtrGT8D2 expression in PtrGT8D-TL3 were 14 and 22%, respectively, of those in the wild type. The sums of the expression of PtrGT8D1 and PtrGT8D2 in PtrGT8D-TL3, -TL5, -TL2 and -TL7 were 14.7, 6.0, 5.7 and 6.6%, respectively, of those in the wild type. The reason that PtrGT8D1 has a stronger reduction than PtrGT8D2 is maybe because the trigger sequences of PtrGT8D1 are more efficient in RNA silencing than those of PtrGT8D2, or because the gene with the higher expression level can be silenced more efficiently.

Chemical composition analysis To investigate the roles of PtrGT8Ds in the biosynthesis of polysaccharides during wood formation, we analyzed the cell wall composition of transgenic wood. The most obvious composition change is the reduction in xylan content, ranging from 64 to 71% of wild type (Table 1). Compared with the wild type, galactan and mannan content had a slight increase. The quantities of cellulose and other hemicelluloses remained essentially unchanged. The lignin content of all the low-xylan transgenics was increased (Table 1). t-Test analysis showed that the P values for the change in xylan, lignin, galactan and mannan content were all