Three loblolly pine CesA genes expressed in developing xylem are orthologous to secondary cell wall CesA genes of angiosperms


  • C. Joseph Nairn,

    Corresponding author
    1. Warnell School of Forest Resources, University of Georgia, Athens, GA 30602, USA
      Author for correspondence: C. Joseph Nairn Tel: +1 706 542 1885 Fax: +1 706 542 8356 Email:
    Search for more papers by this author
  • Tamara Haselkorn

    1. Warnell School of Forest Resources, University of Georgia, Athens, GA 30602, USA
    Search for more papers by this author

Author for correspondence: C. Joseph Nairn Tel: +1 706 542 1885 Fax: +1 706 542 8356 Email:


  • • Specific plant cellulose synthases (CesA), encoded by a multigene family, are necessary for secondary wall synthesis in vascular tissues and are critical to wood production. We obtained full-length clones for the three CesAs that are highly expressed in developing xylem and examined their phylogenetic relationships and expression patterns in loblolly pine tissues.
  • • Full-length CesA clones were isolated from cDNA of developing loblolly pine (Pinus taeda) xylem and phylogenetic inferences made from plant CesA protein sequences. Expression of the three genes was examined by Northern blot analysis and semiquantitative RT-PCR.
  • • Each of three PtCesA genes is orthologous to one of the three angiosperm secondary cell wall CesAs. The PtCesAs are coexpressed in tissues of loblolly pine with tissues undergoing secondary cell wall biosynthesis showing the highest levels of expression. Phylogenetic and expression analyses suggest that functional roles for these loblolly pine CesAs are analogous to those of orthologs in angiosperm taxa.
  • • Based upon evidence from this and other studies, we suggest division of seed plant CesA genes into six major paralogous groups, each containing orthologs from various taxa. Available evidence suggests that paralogous CesA genes and their distinct functional roles evolved before the divergence of gymnosperm and angiosperm lineages.


Cellulose is a crystalline β-1,4-glucan synthesized from a UDP-glucose (UDP-Glc) substrate and is the most abundant biopolymer made by plants (Zhong et al., 2003). The biomass of cellulose makes it a global carbon sink and the properties of cellulose and its deposition are central to plant cell morphogenesis during primary cell wall biosynthesis (Arioli et al., 1998). Cellulose is also synthesized in large quantities during secondary cell wall biosynthesis, particularly in plant vascular tissues, and is the major component of the secondary cell walls, which are the primary component of wood and wood fiber.

The first plant cellulose synthases were identified in cotton (Gossypium hirsutum) fiber cDNA libraries using motifs homologous to the four UDP-Glc binding motifs in cellulose synthases from bacteria (Pear et al., 1996). Cellulose synthase (CesA) genes are represented in the genome of higher plants as a multigene family (Richmond, 2000) that is part of a larger superfamily of putative processive glycosyltransferases (Richmond & Somerville, 2000). The superfamily includes CesA genes and several classes of cellulose synthase-like (Csl) genes. The Csl genes have been postulated to be involved in the synthesis of other plant noncellulosic polysaccharides. Cloning and transgenic expression of a mannan synthase gene that is a member of the Csl gene family supports this hypothesis (Dhugga et al., 2004).

Cellulose synthesis complexes are visible as hexagonal rosette structures on the plasma membrane surface by electron microscopy (Delmer, 1999). The rosette complexes are transported to the plasma membrane following assembly in the Golgi (Haigler & Brown, 1986). Proteins encoded by the CesA genes are components of the rosette structures, as demonstrated by immunolocalization (Kimura et al., 1999).

Distinct functions for at least six of the 10 identified Arabidopsis (Arabidopsis thaliana) CesA genes have been proposed. Experimental evidence suggests that two different groups, of at least three genes each, are necessary for primary and secondary cell wall biosynthesis, respectively. The genes AtCesA1, AtCesA3, and AtCesA6 have nonredundant functions in primary cell wall biosynthesis as demonstrated in studies of plants with mutations in each of these genes (Arioli et al., 1998; Fagard et al., 2000; Scheible et al., 2001; Ellis et al., 2002). AtCesA1, AtCesA3 and AtCesA6 mutants have altered cell morphology including reduced cell elongation. Three other Arabidopsis CesA genes, AtCesA4 (IRX5), AtCesA7 (IRX3, FRA5) and AtCesA8 (IRX1, FRA6), are required for cellulose synthesis in the secondary cell walls of vascular tissues (Taylor et al., 1999; Taylor et al., 2000; Taylor et al., 2003; Zhong et al., 2003). Mutations in any one of these genes lead to severe defects in secondary cell walls of vascular tissue including a significant reduction in cellulose content. The encoded proteins are functionally nonredundant and interact to form multimers. Immunolocalization using antibodies to individual catalytic subunits and localization of CesA:GFP fusions in Arabidopsis roots indicate that all three subunits are required for assembly and subsequent targeting to the plasma membrane (Gardiner et al., 2003; Taylor et al., 2003).

Putative orthologs of the secondary cell wall CesA genes of Arabidopsis have been identified in the genomes of a number of seed plant taxa including rice (Oryza sativa), a monocot, and Populus (Populus tremuloides), a woody dicot species. Studies of CesA mutants of rice indicate that OsCesA4, OsCesA7 and OsCesA9, are orthologous to the Arabidopsis genes AtCesA8, AtCesA4 and AtCesA7, respectively, and have analogous, nonredundant functions in this monocot species (Tanaka et al., 2003). Characterization of expressed CesA genes in barley (Hordeum vulgare) included discovery of three genes that are orthologs of the three secondary cell wall CesA genes in Arabidopsis and rice. These three CesA genes are coexpressed in tissues predominantly undergoing secondary cell wall synthesis (Burton et al., 2004). CesA orthologs in Populus are highly expressed in developing xylem suggesting that these genes have analogous functions in secondary cell wall biosynthesis of this tree species (Wu et al., 2000; Samuga & Joshi, 2002; Joshi, 2003).

Gymnosperms are among the most important tree species, yet the role of CesA genes in gymnosperms has not been studied in any depth. In this regard, we report initial investigations of CesA genes encoding cellulose synthase catalytic subunits expressed at high levels in developing wood tissues of loblolly pine (Pinus taeda L.). An understanding of CesA genes and other genes associated with cellulose synthesis in loblolly pine should provide insight into wood production in pines and other conifers that are a major source of timber, pulp, paper, and many other fiber-derived products. Information derived from these types of investigations may be utilized for genetic manipulation of cellulose synthesis and deposition as well as tree improvement programs for sustainable wood production in future generations (Holland et al., 2000; Joshi, 2003).

Materials and Methods

Sequence analysis

Loblolly pine EST sequences representing putative CesA genes were downloaded from the Stanford Cell Wall web site ( Predicted peptides for the Arabidopsis CesA sequences AtCesA1-AtCesA10 were used to query the predicted peptide sequences of loblolly pine EST sequences in GenBank. ESTs with predicted amino acid sequences that shared ≥ 67% similarity with the Arabidopsis CesA proteins were used in further studies. Contig assembly was performed using the program GCG (Wisconsin Genetics Computer Group) and included the contigs PtCesA1 through PtCesA15 and the PtCesA singleton sequence from the Stanford Cell Wall web site. Gene nomenclature as described on the Stanford web site was adopted for loblolly pine cellulose synthase genes.

Putative orthologs of the Arabidopsis actin gene ACT2 (An et al., 1996) and 26S proteasome AAA-ATPase subunit RPT3 (AF123392) were identified in the loblolly pine EST database. Encoded peptide sequences for a loblolly pine actin, PtAct2, and a proteasome subunit, PtRpt3, shared 95.5% and 93.9%, respectively, with the putative Arabidopsis orthologs. (Table S1, available online as supplementary material). Consequently, PtAct2 and PtRpt3 were adopted as controls for gene expression analysis.

Plant material

Developing wood of a ≥ 40-yr-old loblolly pine (Pinus taeda L) was collected as previously described from crown (juvenile wood) and trunk (mature wood) (Lorenz & Dean, 2002). Roots, green tips (above the first internode), and young expanded needles were harvested from glasshouse grown plants. Unexpanded needles and emerging lateral shoots (candles) were harvested from field grown plants in early spring. The distal 3 cm of each lateral shoot was cut into three 1-cm sections. The sections representing the first centimeter from the tip were pooled, those representing the second centimeter from the tip were pooled, and those representing the third centimeter were pooled. Bark with attached phloem and cambial cells was peeled from glasshouse plants with a trunk diameter of c. 2.5 cm.

RNA isolation

Total RNA was isolated as previously described (Chang et al., 1993). Poly(A)+ mRNA was purified from total RNA using oligo-dT cellulose spin columns as per the manufacturer's protocol (Ambion Inc., Austin, TX, USA) for 5′ RACE and cDNA library construction. Poly(A) + mRNA was purified using oligo-dT magnetic beads following the manufacturer's protocol (Dynal Biotech LLC., Brown Deer, WI, USA) for Northern blot analysis and semiquantitative RT-PCR. RNA samples were quantitated by spectrophotometry at 260 and 280 nm.

RT-PCR and 5′ RACE

Primers were designed using multiple alignments of PtCesA consensus sequences derived from contigs composed of PtCesA EST sequences (Table S2, available online as supplementary material). RT-PCR was carried out using total RNA from developing xylem as the template. Partial sequences for PtCesA1 and PtCesA2 were used to design gene specific primers for 5′ RLM-RACE (Ambion). The sequences of the 5′ RACE product and the 3′ UTR of PtCesA1 were used to design 5′ forward primer PtC1–45F and 3′ reverse primer PtC1–13R for amplification of a full-length cDNA.

Library construction, screening and DNA sequencing

Loblolly pine cDNA libraries were constructed using the ZAP Express XR Library Construction Kit (Stratagene, LaJolla, CA, USA) as per the manufacturer's protocol except that Superscript III (Invitrogen Corp., Carlsbad, CA, USA) was used for reverse transcription and the reaction was carried out at 50°C for 50 min followed by 15 min at 55°C. Selection of cDNA was performed using cDNA Size Fractionation Columns (Invitrogen). Fraction 5, containing the largest cDNA molecules, and fractions 6–9, containing the remaining molecules over 1 kb were ligated and packaged in separate reactions. Probes were generated using insert DNA from partial clones of PtCesA1, PtCesA2 or PtCesA3 labeled with the Phototope kit (New England Biolabs Inc., Beverly, MA, USA). Libraries were transferred to nylon membranes, hybridized at high stringency (Sambrook & Russell, 2001) and chemiluminescent detection carried out with the Phototope Star Detection kit (New England Biolabs). Full-length clones for PtCesA1 (AY789650), PtCesA2 (AY789651), and PtCesA3 (AY789652) were sequenced by primer walking. Sequencing was accomplished by the dideoxy chain termination method (Sanger et al., 1977) using BigDye Terminator v3.1 and analyzed using an ABI 3700 automated sequencer (Applied Biosystems, Foster City, CA, USA).

Phylogenetic analysis

Full-length peptide sequences were obtained by conversion of the PtCesA1, PtCesA2 and PtCesA3 DNA sequences using GCG software (Genetics Computer Group, 1994). Full-length sequences for CesA genes from other plant taxa were obtained from the public database (Table 1). Multiple alignments of CesA peptide sequences were generated using Pileup (Genetics Computer Group, 1994) and CLUSTAL W (Thompson et al., 1994). Alignments generated with CLUSTAL W were used to infer phylogenies and perform bootstrap analysis of 1000 pseudoreplicates.

Table 1.  Designations for CesA sequences and their corresponding accession numbers.
Arabidopsis thaliana
AtCesA1AF027172Populus tremuloides 
AtCesA4 (IRX5)AF458083PtrCesA3AF527387
AtCesA7 (IRX3)AF088917PtrCesA6AY196961
AtCesA8 (IRX1)AF267742PtrCesA7AY162180
AtCesA9AC007019Solanum tuberosum 
Gossypium hirsutum Triticum aestivum 
GhCesA3AF150630Zinnia elegans 
Hordeum vulgare ZeCesA1AF323039
HvCesA1AY483150Zea mays 
Mesotaenium caldariorum ZmCesA4AF200528
Nicotiana alata ZmCesA6AF200530
Oryza sativa ZmCesA8AF200532
OsCesA9AK121170Ceratopteris richardii 
Pinus radiata CrCesAABE642569
Populus×canescens Physcomitrella patens 
Tortula ruralis 

Northern blot analysis

Poly(A) + enriched mRNA from tissues of loblolly pine was subjected to denaturing agarose gel electrophoresis using glyoxal and dimethylsulfoxide and transferred to nylon membranes (Sambrook & Russell, 2001). Triplicate gels were run and blotted using 500 ng of mRNA per lane. Gene specific primers were used to amplify c. 300 bp fragments from the first hypervariable region, HVR I, of PtCesA1, PtCesA2 and PtCesA3 (Table S2). Hybridization and washes were carried out at 68°C. Probe labeling and chemiluminescent detection were carried out using the Phototope system (New England Biolabs). Gene specificity of PtCesA probes was confirmed by hybridization to dot blots of plasmid clones containing full-length cDNA inserts for each of the PtCesA genes examined.

A 1003-bp fragment of the PtAct2 cDNA was amplified from loblolly pine cDNA using primers PtAct2–1F and PtAct2–3R (Table S2). An 840 bp fragment of PtRpt3 was amplified using primers PtRpt3–1F and PtRpt3–4R. Following hybridization with the PtCesA probes, blots were stripped and re-hybridized with PtAct2 and PtRpt3 probes to confirm equal lane loading of mRNA.

Semi-quantitative PCR

Semi-quantitative RT-PCR was carried out using poly(A) + enriched mRNA from tissues of loblolly pine. Gene specific primers for HVR I regions of PtCesA1 (PtC1-HVR1–1F and PtC1-HVR1–1R), PtCesA2 (PtC2-HVR1–1F and PtC2-HVR1–1R) and PtCesA3 (PtC3-HVR1–1F and PtC3-HVR1–1R) were used amplify c. 300 bp fragments from each of the genes (Table S2). Gene specific primers were used to amplify a c. 260 bp region of the PtAct2 (PtAct2–2F and PtAct2–3R) gene and a c. 290 bp fragment of the PtRpt3 (PtRpt3–5F and PtRpt3–4R) as internal controls.


Isolation of full-length PtCesA cDNAs

Contigs were assembled for loblolly pine CesA sequences using EST sequences listed on the Stanford cell wall web site (, and EST sequences from the public database. The resulting contigs did not represent full-length cDNA molecules. Sequences for PtCesA1, PtCesA2 and PtCesA3 were more highly represented in the EST databases than those for the remaining PtCesA genes examined. Preliminary data from semiquantitative PCR using gene specific primers also suggested that higher transcript levels were present in developing wood for PtCesA1, PtCesA2 and PtCesA3 than the remaining PtCesA genes examined. Consequently, PtCesA1, PtCesA2 and PtCesA3 were selected for further study.

The 5′ region of the PtCesA1 cDNA was cloned using a 5′ RACE procedure. Primers were designed for the 5′- and 3′ untranslated regions (UTR) of PtCesA1 and used to amplify and clone a full-length cDNA representing the PtCesA1 gene. Extension of the PtCesA2 using 5′ RACE did not produce full-length cDNA molecules and a number of the clones were truncated at approximately the same position, possibly due to RNA secondary structure. Full-length cDNA clones of PtCesA2 and PtCesA3 were obtained from loblolly pine cDNA libraries that were synthesized from RNA of developing wood. Full-length cDNA clones contained 5′- and 3′ UTRs.

DNA and protein sequences of PtCesA1, PtCesA2, and PtCesA3 were used to query sequences in GenBank using BLASTN and BLASTP, respectively ( Putative CesA sequences from a number of plant taxa were identified. Each of the three PtCesA peptide sequences contained conserved domains for the cellulose synthase protein family, PFAM03552, and glycosyltransferases, COG1215 (Marchler-Bauer et al., 2003; Tatusov et al., 2003; Bateman et al., 2004).

The full-length cDNA sequences for PtCesA1, PtCesA2 and PtCesA3 were compared with assembled contigs, including those for PtCesA1-PtCesA12, PtCesA14-PtCesA15 and PtCesA singletons from the Stanford Cell Wall web site ( Comparative sequences analysis confirmed that regions of the full-length PtCesA1, PtCesA2 and PtCesA3 cDNA clones correspond to the unigene sets of the same name listed on the Stanford cell wall web site. Analysis also indicated that the unigene set PtCesA11 represented a region of the PtCesA1 full-length cDNA, unigene set PtCesA12 represented a region of the PtCesA2 full-length cDNA and PtCesA14 represented a region of the PtCesA3 full-length cDNA sequence. In addition, PtCesA singleton sequences BQ701506 and BQ701242 represent regions within the full-length cDNAs of PtCesA1 and PtCesA2, respectively.

PtCesA 1–3 are orthologous to secondary cell wall CesA genes from angiosperms

Alignments of full-length plant CesA protein sequences were constructed using Pileup (Genetics Computer Group, 1994) and CLUSTAL W (Thompson et al., 1994). Multiple sequence alignments created by the two programs were similar and dendrograms of the alignments from the two programs were congruent with respect to branching order.

A phylogenetic tree was inferred from the multiple sequence alignment of full-length CesA peptide sequences using CLUSTAL W (Fig. 1). Six major clades, each containing CesA sequences from monocot and eudicot taxa, are supported by the inferred phylogeny and have high bootstrap support of 100%. Each of three of these clades contains one of the loblolly pine sequences, as well as sequences from monocot and eudicot taxa. These three clades each contain one of three putative secondary cell wall CesA sequences from Arabidopsis and rice (Taylor et al., 1999; Taylor et al., 2000; Tanaka et al., 2003; Taylor et al., 2003; Zhong et al., 2003). Each of the remaining three clades contains one of the Arabidopsis CesA sequences that have been implicated in cellulose synthesis of the primary cell wall (Arioli et al., 1998; Fagard et al., 2000; Scheible et al., 2001; Ellis et al., 2002) and representatives from additional monocot and eudicot taxa.

Figure 1.

Phylogeny of plant CesA genes inferred from a multiple alignment of full-length CesA protein sequences. Protein sequences were obtained from the cellulose synthase catalytic subunit genes of Arabidopsis thaliana, AtCesA; rice (Oryza sativa), OsCesA; aspen (Populus tremuloides), PtrCesA; loblolly pine (Pinus taeda), PtCesA; Zinnia elegans, ZeCesA; cotton (Gossypium hirsutum), GhCesA; hybrid poplar (Populus × canescens), PtcCesA; radiata pine (Pinus radiata), PrCesA; maize (Zea mays), ZmCesA; barley (Hordeum vulgare), HvCesA; potato (Solanum tuberosum), StCesA; wheat (Triticum aestivum), TaCesA; Persian tobacco (Nicotiana alata), NaCesA; and a green alga (Mesotaenium caldariorum), McCesA. Numbers indicate bootstrap support for each group from 1000 pseudoreplicates.

Sequence conservation of CesA proteins

Plant CesA protein sequences are highly conserved overall. The CesA protein sequence from the alga Mesotaenium shares 67–73% similarity with sequences from angiosperm and gymnosperm taxa. Sequence similarities between CesA proteins from gymnosperm and angiosperm taxa range from 68% to 84% and sequence similarities of over 99% are found between CesA proteins from taxa within the angiosperms and within the gymnosperms.

Available full-length plant CesA sequences contain the conserved D residues and the motif QVLRW is conserved in plant sequences including the green alga Mesotaenium caldariorum (Delmer, 1999; Roberts et al., 2002). The QVLRW motif is also present in partial CesA sequences from two bryophyte species (Tortula ruralis and Physcomitrella patens), and one pteridophyte species (Ceratopteris richardii). However, a partial sequence from T. ruralis (CN204913) and one from C. richardii (BE642569), that were identified as putative CesA sequences, contain variants of this motif.

The addition of loblolly pine full-length CesA sequences to those available from angiosperm taxa permitted the examination of sequence conservation across a broader taxonomic range than previous studies. The HVR II region of the alignments revealed conserved motifs in orthologous sequences that were not present in paralogous sequences examined (Fig. S1a, available online as supplementary material). Previous comparisons of CesA protein sequences revealed conserved motifs in the HVR II region of some putative orthologs and lead to the suggestion that the region be termed a ‘class-specific region’ (CSR) (Vergara & Carpita, 2001). Putative class specific (CS) motifs from secondary cell wall CesA proteins identified in this study were used to query the available barley CesA sequences (Burton et al., 2004) and the maize CesA sequences (Holland et al., 2000). Three barley sequences were identified, each containing a putative CS motif, and these represented the Group II CesA sequences that have been implicated in cellulose synthesis of the secondary cell wall. Putative CS motifs were not found in maize sequences ZmCesA1ZmCesA9 (Holland et al., 2000), but a a full-length cDNA sequence, ZmCesA10, was identified in GenBank which contained two motifs found in PtCesA2 and orthologous genes. The ZmCesA10 sequence grouped with the clade of CesA orthologs containing PtCesA2.

Comparative sequence analysis also revealed extensive conservation in the N-terminal region of CesA sequences within the clade that contained PtCesA3 and orthologous sequences (Fig. S1b, available online as supplementary material). Seed plant CesA sequences for the remaining two subgroups were not conserved at the primary sequence level in the N-terminal region.

Gene expression analysis

Gene specific probes from the HVR I regions of the three PtCesA genes were used to probe one of three triplicate blots (Fig. 2a). Identical patterns of expression were observed for PtCesA1, PtCesA2 and PtCesA3 in the loblolly pine tissues examined. Transcript levels were relatively high in mature and juvenile wood, apical shoot tips, lateral shoots and expanded young needles. Transcript levels in root tissue were relatively low and expression in unexpanded needles appeared to be near the limits of detection. No expression was detected in the samples containing bark, phloem and cambium.

Figure 2.

Transcript levels of PtCesA1, PtCesA2 and PtCesA3 in tissues of loblolly pine (Pinus taeda). RNA samples were isolated from juvenile wood (JW) and mature wood (MW), roots (R), shoot tips (ST), unexpanded needles (UN), expanded needles (EN), 1st cm of the distal tip of lateral shoots (LS1), 3rd cm from the distal tip of lateral shoots (LS3), bark with phloem and cambium (BPC). (a) Northern blot analysis of transcript levels. Triplicate gels containing 500 ng per lane of mRNA were hybridized at high stringency with gene specific probes from PtCesA1, PtCesA2 and PtCesA3. Blots were stripped and re-hybridized with probes for an actin gene, PtAct2, and a proteasome subunit gene PtRpt3, from loblolly pine. (b) Semiquantitative RT-PCR examining gene transcript levels in tissues of loblolly pine. Gene specific primers were used to amplify c. 300 bp fragments of the HVR I region from the three PtCesA genes. Primers for an actin gene, PtAct2, and a proteasome subunit gene, PtRpt3, were used to amplify regions of the respective genes from loblolly pine as internal controls.

Semiquantitative RT-PCR was also used to examine expression of the three CesA genes in loblolly pine tissues (Fig. 2b). Expression of each of the three genes is relatively high in juvenile and mature wood, lateral shoots and young expanded needles. Root tissues show intermediate levels of gene expression and unexpanded needles have relatively low levels of expression for each of the three genes examined.

Northern blot analysis suggests that PtCesA3 transcript levels are lower than those for PtCesA1 and PtCesA2. However, semiquantitative RT-PCR suggests that transcripts for the three PtCesA genes are present at similar levels. Different band intensities on Northern blots could have resulted from variation in probe specific activity or film exposure.


Full-length clones for three cellulose synthase genes, PtCesA1, PtCesA2 and PtCesA3, that are highly expressed in developing xylem of loblolly pine were isolated and characterized. Full-length cDNA molecules are necessary for functional studies, robust phylogenetic analyses, annotation of sequences, and for estimation of transcript levels when EST sequences are used for electronic Northerns to examine gene expression.

Fourteen CesA unigene sets (PtCesA1-PtCesA12 and PtCesA14-PtCesA15) have been assembled from loblolly pine EST sequences and a number of singleton sequences have been identified as well ( Each of the full-length PtCesA sequences corresponds to more than one of the unigene sets and singleton sequences. Based on this analysis, the number of putative CesA genes identified in loblolly pine is reduced from 14 to 10. This is similar to the Arabidopsis genome, which contains 10 identified members of the CesA gene family, and other diploid angiosperm species, for which relatively large numbers of EST sequences are available.

Evolution of the plant CesA multigene family

A phylogeny inferred from alignments of plant full-length CesA protein sequences indicates that each of the three genes, PtCesA1, PtCesA2 and PtCesA3, is orthologous to one of three groups of CesA genes from angiosperm taxa that are functionally nonredundant and are necessary for cellulose synthesis in the secondary cell wall of vascular tissues (Fig. 1). The phylogeny supports six major clades of paralogous sequences, each containing orthologous CesA sequences from a variety of plant taxa. Each of the six clades had high bootstrap support of 100% from 1000 pseudoreplicates.

Three of the clades contain putative secondary cell wall CesA genes and each of these three clades contains one loblolly pine, one rice, one Populus and one Arabidopsis CesA sequence as well as individual CesA sequences from additional plant taxa. Each of the three remaining clades includes one of the three Arabidopsis genes, AtCesA1, AtCesA3 and AtCesA6, that have been implicated in cellulose synthesis of the primary cell wall.

Phylogenetic inferences, which include the loblolly pine sequences presented herein, indicate that the evolution of distinct secondary cell wall CesA paralogs within the CesA multigene family preceded the divergence of the gymnosperm and angiosperm lineages, which occurred c. 300 million yr ago (Bowe et al., 2000). The phylogeny is congruent with those from previous studies of angiosperm CesA genes, which support evolution of distinct paralogous genes before the divergence of the moncot and eudicot lineages (Tanaka et al., 2003; Burton et al., 2004).

The N-terminal region of the PtCesA3 ortholog group is highly conserved in contrast to other groups of orthologs encoding secondary cell wall CesA genes and represents the most divergent of the three groups of secondary cell wall CesA proteins (Fig. S1b). The role of specific catalytic subunits and their stoichiometry within the cellulose synthesis complex have yet to be determined (Delmer, 1999; Doblin et al., 2002). Extensive sequence conservation across a broad range of plant taxa suggests that the N-terminus of this particular protein subunit may have a specific or specialized function within the cellulose synthase complex.

Gene expression of putative secondary cell wall CesA genes of loblolly pine

Expression patterns of the three loblolly pine CesA genes are consistent with functional roles that are analogous to those of their orthologous genes in angiosperm taxa. PtCesA1, PtCesA2 and PtCesA3 are coexpressed at relatively high levels in tissues undergoing secondary cell wall synthesis (Fig. 2). Differences in relative transcript levels for the three PtCesA genes observed in Northern blot analysis were not confirmed by semiquantitative RT-PCR. However, neither of these methods are fully quantitative and analysis by real time quantitative RT-PCR will be needed to provide an accurate assessment of relative transcript levels for the three PtCesA genes. Transcript levels for the three orthologous genes in barley, examined by real time quantitative RT-PCR, differed by 2–8 fold between the three paralogous genes (Burton et al., 2004) and transcript levels for OsCesA9 appear lower than those for paralogous secondary cell wall CesA genes in Northern blot analysis (Tanaka et al., 2003). Variation in RNA stability or translational efficiencies are possible explanations for unequal transcript levels (Burton et al., 2004).

In Arabidopsis and rice, three distinct, functionally nonredundant CesA genes are necessary for cellulose biosynthesis in the secondary cell wall of primary vascular tissues and orthologous genes in barley are coexpressed in tissues predominantly undergoing secondary cell wall synthesis (Tanaka et al., 2003; Taylor et al., 2003; Burton et al., 2004). CesA orthologs in Populus are expressed during development of secondary xylem (Joshi, 2003). Orthologous genes in loblolly pine are also highly expressed in developing xylem. Inferred phylogenetic relationships, however, indicate that the CesA orthologs of woody species do not share an evolutionary history distinct from those of herbaceous taxa. These data support analogous functions for orthologous CesA genes during secondary cell wall biosynthesis in primary vascular tissues of herbaceous species and secondary vascular tissues of woody species.

Based on phylogenetic inferences from a limited number of taxa and gene expression analysis, we hypothesize that six paralogous CesA genes were present in the genome of the common ancestor of extant seed plants and that distinct, nonredundant functions evolved before the divergence of extant seed plant lineages. Evidence suggests that three of these paralogs are involved in primary cell wall synthesis and three are necessary for secondary cell wall synthesis. Additional CesA genes may have arisen by subsequent duplication and divergence within individual taxa or lineages. Examination of the CesA gene family in nonseed plants, particularly bryophytes, ferns and related taxa, will be important in understanding the evolution and functional diversification of the CesA gene family and whether the evolution of these paralogous groups and their functional specificity preceded or coincided with the evolution of plant vascular tissue.


We thank W. Lorenz and J. Barnes for assistance in obtaining plant materials and Z. Ye for valuable comments on the manuscript. Funding for the project was provided by the University of Georgia's Daniel B. Warnell School of Forest Resources and the Office of the Vice President for Research.

Supplementary material

The following material is available as supplementary material at

Fig. S1 Alignment of regions from putative secondary cell wall CesA protein sequences.

Table S1 Accession numbers for ESTs used to assemble unigene sets representing full-length cDNA coding regions for loblolly pine (Pinus taeda) genes that were used as internal controls in gene expression analysis.

Table S2 Oligonucleotide primer sequences used for amplification of PtCesA cDNA molecules and control genes PtAct2 and PtRpt3.