Plant glycoside hydrolase family 9 (GH9) comprises typical endo-β-1,4-glucanase (EGases, EC184.108.40.206). Although GH9A (KORRIGAN) family genes have been reported to be involved in cellulose biosynthesis in plants, much remains unknown about other GH9 subclasses. In this study, we observed a global gene co-expression profiling and conducted a correlation analysis between OsGH9 and OsCESA among 66 tissues covering most periods of life cycles in 2 rice varieties. Our results showed that OsGH9A3 and B5 possessed an extremely high co-expression with OsCESA1, 3, and 8 typical for cellulose biosynthesis in rice. Using two distinct rice non-GH9 mutants and wild type, we performed integrative analysis of gene expression level by qRT-PCR, cellulase activities in situ and in vitro, and lignocellulose crystallinity index (CrI) in four internodes of stem tissues. For the first time, OsGH9B1, 3, and 16 were characterized with the potential role in lignocellulose crystallinity alteration in rice, whereas OsGH9A3 and B5 were suggested for cellulose biosynthesis. In addition, phylogenetic analysis and gene co-expression comparison revealed GH9 function similarity in Arabidopsis and rice. Hence, the data can provide insights into GH9 function in plants and offer the potential strategy for genetic manipulation of plant cell wall using the five aforementioned novel OsGH9 genes.
Citation: Xie G, Yang B, Xu Z, Li F, Guo K, Zhang M, et al. (2013) Global Identification of Multiple OsGH9 Family Members and Their Involvement in Cellulose Crystallinity Modification in Rice. PLoS ONE 8(1): e50171. https://doi.org/10.1371/journal.pone.0050171
Editor: Joshua L. Heazlewood, Lawrence Berkeley National Laboratory, United States of America
Received: June 1, 2012; Accepted: October 22, 2012; Published: January 4, 2013
Copyright: © 2013 Xie et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: The project was supported in part by grants from the National 973 pre-project (2010CB134401), the National Transgenic Project (2009ZX08009-119B), the National 111 Project (B08032), and HZAU Changjiang Scholar Promoting Project (52204-07022). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Cellulose is the major wall polysaccharide in plants and has a wide application for biofuel, paper, and other chemical products , . Due to their crystalline property, cellulose microfibrils are highly recalcitrant to biomass saccharification . Hence, understanding cellulose biosynthesis and crystallization is essential.
Cellulose is a fibrous polymer of glucose units linked by β-1, 4-glucosidic bonds. It can self-associate into non-crystalline and crystalline microfibrils in a plant cell wall, providing mechanical strength and flexibility during plant growth and development . Over the past years, the crystallinity index (CrI) has been applied to account for lignocellulose crystallinity by characteristic X-ray diffraction (XRD) patterns and solid-state 13C nuclear magnetic resonance (NMR) spectra , . In higher plants, cellulose is synthesized at the plasma membrane by a symmetrical rosette of six global protein complexes, with each complex containing several structurally similar cellulose synthase (CESA) subunits . AtCESA1, 3, and 6 in Arabidopsis and OsCESA1, 3, and 8 in rice have been identified for primary cell wall formation, whereas AtCESA4, 7, and 8 and OsCESA4, 7, and 9 are responsible for cellulose biosynthesis in the secondary cell walls, respectively –. Furthermore, AtCESA3ixr1-2 and AtCESA6ixr2-1 mutants show increased biomass saccharification efficiency and reduced cellulose crystallinity index (CrI) , . CESA1A903V and CESA3T942I mutants also display reduced cellulose microfibril crystallinity . In addition, other genes, such as COBRA and KORRIGAN, have been reported to contribute to cellulose biosynthesis , .
Endo-β-1, 4-glucanases (EGases, EC220.127.116.11) have been found in prokaryotic and eukaryotic organisms. They are crucial for cell wall degradation and remodeling because they can cleave the internal β-1,4-glycosidic bond between two glucose moieties in the center of a polysaccharide chain –. Plant EGase enzymes belong to subgroup E2 of glycoside hydrolase family 9 (GH9) with three subclasses (A, B, C) , . GH9A comprises membrane-anchored proteins , , GH9B comprises secreted proteins with only one catalytic domain (CD), and GH9C possesses a CD and a distinct C-terminal extended cellulose-binding domain (CBD) that binds to crystalline cellulose as the bacterial cellulase did . Although GH9 family proteins are usually able to hydrolyze artificial soluble cellulose derivatives, such as carboxymethyl cellulose (CMC) or hydroxyethyl cellulose (HEC) , biochemical analyses have revealed their specificity for different substrates in vitro –.
A number of KORRIGAN (kor) mutants of GH9A family genes show reduced crystalline cellulose in plants , , . For instance, ectopic overexpression of PttKOR1 in Arabidopsis kor1-1 mutant leads to a higher cellulose crystallinity , whereas RNA interference (RNAi) of PaxgKOR could reduce cellulose level and increase cellulose crystallinity . In rice, silencing of OsGH9A3 results in reduction of cell elongation and cellulose content; it also causes an increase of pectin content in leaves . In addition, KORRIGAN protein can either cleave sterol-cellodextrin substrate , or remove glucan chains incorrectly assembled in the growing microfibrils . Although both GH9B and GH9C have been reported with activities for cello-oligosaccharide release and xyloglucan cleavage in plants , –, little is known about their functions in cellulose biosynthesis and crystallization in rice , .
Rice, an important food crop worldwide, is a model for gene function analysis in monocotyledonous plants. The completion of the rice genome sequencing may likely identify the potential function of the entire GH9 family genes in rice, based on bioinformatics analysis and related biological characterization. In this study, we initially observed a gene co-expression profiling between OsGH9 and OsCESA, and then performed a comparative analysis of the multiple OsGH9 genes' functions in cellulose synthesis and assembly, especially in their involvement in modification of lignocellulose crystallinity in rice.
Identification of OsGH9 family in rice
Searching local rice genome database yielded a total of 25 OsGH9 family genes located in 9 chromosomes in rice, except 1 gene (LOC_Os01g64140) with a partial DNA sequence (Table S1). Based on the nomenclature of GH9 family in Arabidopsis , GH9 family proteins can be divided into three subclasses (A, B, and C) (Figure 1A). GH9A subclass (OsGH9A1, 2 and 3) contained the charged and hydrophilic amino acid-rich N-termini, transmembrane domain (TM), and proline-rich C-termini. The GH9A proteins had large tails with 71 to 73 amino acid residues (Table S1). In addition, all GH9A proteins contained polarized targeting signals (LL and YXXΦ) and six predicted glycosylation sites (data not shown). Particularly, OsGH9B18 showed the sequence homology and motif similar to OsGH9B13 (Figure 1), despite that its orthologue in Arabidopsis has been grouped as AtGH9A4 (Figure S4). By comparison, all GH9B and GH9C proteins did not contain any TM, but had a catalytic domain (CD) and an N-terminal cleavable signal peptide (SP) without significant sequence conservation in the extracellular secreted proteins, as shown by PSORT prediction tool (data not shown). GH9B proteins displayed diversity in the transmembrane helix and pI value, and GH9C proteins showed a putative cellulose-binding domain (CBD) as the bacterial cellulase did (Table S1).
(A) Phylogenetic tree of OsGH9 family proteins: Clustal X program and tree construction using MEGA3.1 used for multiple alignment analysis of 25 rice GH9 family proteins. (B) Exon-intron comparison of OsGH9 family genes: GSDS http://gsds.cbi.pku.edu.cn/ was used for analysis of the exon-intron structures. (C) Motif distributions of OsGH9 family proteins. The 25 motifs of OsGH9 family proteins annotated in Table S8 using the interProScan search program http://www.ebi.ac.uk/Tools/InterProScan/, and then identified using the MEME program (version 4.0) http://meme.sdsc.edu/meme/cgi-bin/meme.cgi.
Furthermore, there were at least 4 exons in GH9A and GH9C genes, and 1 to 7 exons in GH9B genes (Figure 1B). MAST analysis (http://meme.nbcr.net/) showed that GH9A1, 2, 3 and GH9C1, 2, 3, 4 contained identical and conserved motifs, whereas those in GH9B were relatively variable (Figure 1C), suggesting a diverse structural organization in GH9B family in rice.
Co-expression profiling between OsGH9 and OsCESA family
As KORRIGAN genes have been identified to show co-expression with cellulose synthase gene CESA in Arabidopsis , we initially observed the co-expression profiling between entire OsGH9 and OsCESAs family genes among 66 tissues (Table S9) from the most periods of life cycles of 2 rice varieties (ZS97 and MH63) using cDNA chip's CREP data (http://crep.ncpgr.cn) . As a result, 25 OsGH9 family genes can be classified into 3 expression clusters (I, II, III) with 6 subunits (Ia, Ib, Ic, IIa, IIb, IIc) (Figure 2). In general, OsGH9B8, 9, 10, and 11 genes in Cluster Ia were highly co-expressed with OsCESA4, 7, and 9 typical for secondary cell wall biosynthesis , whereas OsGH9A3 and OsGH9B5 genes in Cluster Ib showed strong co-expression with OsCESA1, 3, 8, 5, and 6 for primary cell wall formation . In particular, Clusters Ic, IIa, and IIb presented a tissue-specific expression in the radicle, panicle, and calli, respectively, and Clusters IIc and III showed a weak co-expression pattern. Correlation analysis further confirmed that OsGH9A3 and OsGH9B5 had significant co-expression with OsCESA1, 3, and 8 (p<0.01), suggesting that OsGH9A3 and OsGH9B5 may have a role in cellulose biosynthesis (Tables 1, S2, and S3). Notably, a significant correlation was also found among OsGH9B1, 2, 3, and 16 (Tables 2, S3), confirming their tight co-expression observed in Cluster IIa.
The cDNA chip data of 66 tissues obtained from two rice var. ZS97 and MH63 at CREP database http://crep.ncpgr.cn/cgi/home.pI website, and the transcription profiling of OsGH9 and OsCESA family genes performed by the hierarchical cluster method.
Furthermore, 18 GH9 family gene expressions were detected by qRT-PCR (Figure S1). Both OsGH9A3, B5, and OsCESA1, 3, 8 were highly expressed in radicle, plumule, and internodes tissues. Although OsGH9B1, 3, and 16 genes were determined with high expression levels, OsGH9B2 gene was undetectable by qRT-PCR in most tissues. In addition, OsGH9A2 and OsGH9C1 showed the specifically high expression in the radicle tissues. OsGH9B6 was highly expressed in stamen tissues, and other OsGH9 family genes were expressed in several tissues in rice.
Analysis of two rice mutants
After large-scale screening and identification of rice mutants in morphological phenotypes and cell wall characteristics including cell wall components, compositions and cell wall degradability, we selected two distinct T-DNA (Osfc4 and Osfc11) mutants with genetic backgrounds of Nipponbare (NPB, wild type) in order to identify GH9 family's potential function on cellulose biosynthesis and crystallization (Figure 3). Both homozygous mutants exhibited a normal growth phenotype during most of their life cycles. However, both were detected with fragile culms compared with the wild type (Figure 3A). Carbohydrate analysis indicated that the two mutants had much lower cellulose and relatively higher hemicellulose levels than wild type in their mature stem tissues (Figure 3B). The stem tissues from 1st to 4th internodes at booting stage of rice presented a consistent increasing course of cellulose biosynthesis, so we observed a typical alteration of cellulose content among mutants and wild type (Figures 3C and 3D). However, the two mutants and the wild type showed a large difference in cellulose production in 1st and 4th internodes, providing the experimental materials excellent for functional analysis of GH9 family genes . In addition, both mutants were genetically identified to be the non-GH9 genes involved in cell wall biosynthesis and modification (data not shown), which allowed the functional analysis of entire OsGH9 family genes in rice.
(A) Phenotypes of two rice mutants (Osfc4, Osfc11) and wild type (NPB): upper panel, rice growth at mature stage; middle panel, manual bending of mature culms; down panel, stem dissection under light microscope. (B) Cell wall composition of mutants and wild type in the mature stems: Cellulose, hemicelluloses and lignin (% DW, dry weight) contents expressed as means ± SD (n = 3). (C) Diagram of four internodes at booting stage. (D) Cellulose contents (% DW, dry weight) of four internodes as shown in (C): Data as means ± SD (n = 3).
Correlation among cellulase activity, lignocellulose crystallinity, and OsGH9 RNA transcripts
Using two rice mutants and wild type, we initially detected cellulase activity in situ and in vitro in the stem internode tissues at booting stage. As shown in Figure 4A, the cellulase activity in situ could be observed in the cell walls of vascular bundles of internodes in the stem tissues. The cellulase activity in vitro with substrate resorufin cellobioside was quantitatively determined with a dynamic alteration during stem internode growth and development (Figures 4B and S2). As a result, both mutants and wild type displayed much higher cellulase activities in young internode (1st and 2nd) tissues than in the old ones (3rd and 4th), indicating cellulase predominant activity in primary cell wall biosynthesis. Meanwhile, we detected lignocellulose crystallinity in all tissues with a consistent increase CrI value from 1st to 4th internodes in both mutants and wild type (Figure 4C). Furthermore, a correlation analysis was conducted between cellulase activity and lignocellulose CrI with R2 value at 0.44 (Figure 4D). Despite the R2 value being less than 0.5, the correlation coefficient value significantly reached −0.70 (p<0.05), suggesting that cellulase may modify cellulose crystallinity.
(A) Observation of cellulase activity in situ of the 3rd internode tissues: The detection of cellulase activity in situ, and the bar indicated 50 µm. (B) Assay of in vitro cellulase specific activity (U. mg−1 protein) in four internodes tissues: Total proteins of four different internodes were used for in vitro cellulase assay by Markergene fluorescent Cellulase Assay Kit. (C) Lignocellulose crystallinity index (CrI) of four internodes tissues. (D) Correlation (n = 12) between cellulase specific activity and lignocellulose CrI among 12 internode samples of four stems in mutants (fc4 and fc11) and wild type (NPB). (E) Correlation coefficients (n = 12) between OsGH9B1, 3, and 16 transcript levels and cellulase specific activity or lignocellulose CrI.
We then analyzed the representative gene expression levels of 3 clusters (Ia, Ib, and IIa) in 4 internode tissues of mutants and wild type by qRT-PCR analysis (Figure S1, Table S4) to test the GH9 family genes' involved roles in the alteration of lignocellulose crystallinity. GH9B1 and GH9B16 in Cluster IIa showed extremely high correlation coefficient values (0.813 to 0.902, at p<0.01, respectively), either positively with cellulase specific activity or negatively with lignocellulose CrI (Figure 4E and Table S5). Although GH9B3 in Cluster IIa did not display a significant correlation with cellulase activity, its coefficient value related to lignocellulose CrI reached −0.716 at p<0.01. Hence, our findings suggest that OsGH9B1, B3, and B16 may have enzymatic activities for lignocellulose crystallinity modification in rice in terms of stem internode growth and development. By contrast, other GH9 genes in Clusters Ia and Ib did not show any significant correlation with lignocellulose CrI or cellulase activity.
Furthermore, OsGH9A3 and OsGH9B5 exhibited significantly positive co-expression (p<0.01) with OsCESA1, 3, and 8 among the four stem internodes of mutants and wild type, but other OsGH9B genes did not show any positive co-expression (Tables S4, S6). Notably, OsGH9A1 was highly co-expressed with OsCESA3, 4, 8 (p<0.05), whereas OsGH9B1 and OsGH9B16 were negative with OsCESA7 (p<0.01). It was confirmed that OsGH9A3 and B5, other than OsGH9B1, 3, 16, may play a role in cellulose biosynthesis, although additional evidences need to provide. It also suggests that OsGH9A1 may have an effect on cellulose biosynthesis at least in the rice stem internode growth and development.
Comparison with AtGH9 family in Arabidopsis
According to the catalytic domain (CD) analysis, 25 putative AtGH9 family genes with 3 subclasses (A, B, and C) can be found in Arabidopsis  Based on the co-expression profiling among the 63 tissues in Arabidopsis, 25 AtGH9 genes can also be classified into 3 clusters (Figure S3, Table S10). By comparison, AtGH9A1 and B7 in Cluster Ic, like OsGH9A3 and B5 in Cluster Ib, were highly co-expressed with AtCESA1, 3, and 6 typical for primary cell wall biosynthesis in Arabidopsis (Table 3). Furthermore, AtGH9B1 and B2 showed high co-expression in the flower and carpel tissues, similar to OsGH9B1, B2, B3, and B16 expression in panicle tissue. In addition, phylogenetic analysis was consistent with the co-expression patterns in rice and Arabidopsis (Figure S4). Due to the close relationship in the phylogenetic tree, AtGH9A1 and B7 were suggested to have a role in cellulose biosynthesis, whereas AtGH9B1 and B2, like OsGH9B1, 3, 16 in rice, may have enzymatic activities for lignocellulose crystallinity modification in Arabidopsis.
Large-scale co-expression has been performed to identify the gene functions in plant cell wall formation across plant species . In Arabidopsis, KORRIGAN mutants have been characterized as GH9A family genes involved in cellulose biosynthesis , –. The regression analysis of the 408 publicly available microarray data sets could even reveal the AtKOR1 (AtGH9A1) co-expression with AtCESA1, 3, and 6 typical for primary cell wall biosynthesis . In aspen tree, RT-PCR, in situ hybridization, and tissue-print assays demonstrated the co-expression of PtrKOR with PtrCESA1, 2, and 3 genes associated with the secondary cell wall synthesis in xylem cells . Although several GH9A (KOR) genes have been identified in Arabidopsis, rice, and other plants , , little remains known about other OsGH9 subclass functions. In this study, global co-expression profiling and correlation analysis indicated that two subclasses of OsGH9 family genes (OsGH9A3 and OsGH9B5) were highly co-expressed with OsCESA1, 3, and 8 genes typical for cellulose synthesis in rice. Similarly, two subclasses of AtGH9 family genes (AtGH9A1 and AtGH9B7) were also identified for cellulose biosynthesis in Arabidopsis. It suggests that genetic silencing of OsGH9A3 or OsGH9B5 or both genes in the transgenic plants may clarify their potential role, like KORRIGAN, in cellulose biosynthesis in rice and other plants. More importantly, co-expression patterns of the genes could somewhat suggest the potential interaction or coordination of their proteins. For instance, based on the high co-expression pattern, OsGH9B1, 3, and 16 were initially suggested to be involved in lignocellulose crystallinity modification, other than in cellulose biosynthesis, which was sequentially confirmed by integrative analysis among gene expression level, cellulase activity, and lignocellulose CrI. Obviously, gene co-expression profiling and correlation analysis are powerful tools in the identification of the entire GH9 family genes' functions in rice and other plants.
The in situ hybridization of endo-β-1, 4-glucanase in differential tissues has been reported in pine , Brassica napus , and aspen trees . A novel, real-time fluorogenic assay with resorufin-β-cellobioside as a substrate has been used for observing glycoside hydrolase activity in planta. Recently, the resorufin glycosides have been detected with high sensitivity in muro cellulase enzyme activity assay due to the significant resorufin ionization at typical apoplastic pH values . By this means, KOR1-overexpressing Arabidopsis plants have been found to show increased cellulase activity in stem tissues compared with wild-type plants . In this study, the cellulase activity in situ was observed specifically in the cell wall of the four internode stems in rice, and the cellulase activities in vitro were quantified and compared between mutants and wild type. Notably, both wild type and mutants displayed a consistent increase at cellulose levels and CrI values from 1st to 4th internode, but showed a constant decrease at cellulase activity in vitro except at 2nd internode of wild type. It suggests that the four internodes would be model materials accounting for cellulase effect on cellulose biosynthesis and lignocellulose crystalline feature in rice and other plants. Despite two mutants and wild type showed a difference at each internode, we could conduct a correlation analysis using all four internodes of wild type and mutants. The correlation calculations among OsGH9 expression level, cellulase activity, and lignocellulose crystallinity, could indicate that OsGH9B1, 3, and 16 have specific cellulase activities for lignocellulose crystallinity modification, other than for cellulose biosynthesis. By contrast, OsGH9A3 and B5 did not show any significant correlation either with cellulase activity or lignocellulose CrI. Hence, we could interpret that the changed lignocellulose crystallinity of AtGH9A1 (Kor) may be a consequence of cellulose biosynthesis, because recent report has indicated that CesA mutant could result in the lignocellulose crystallinity alteration .
Mutant selection and reverse genetic analysis have been broadly applied to identify the target gene function in plants. However, both approaches have their typical limitations and disadvantages, especially if the target gene is lethal to plant growth, or genetically redundant, or functionally dependent on isoform coordination , , . Alternatively, we found that analysis of non-GH9 rice mutants is a useful approach for identification of the multiple OsGH9 family genes. Because the two rice mutants used in this study were genetically identified as the non-GH9 mutants (data not shown), we could conduct a correlation analysis among all detectable OsGH9 gene expressions, cellulase activity, and lignocellulose crystallinity in the wild type and two mutants, suggesting that OsGH9B1, 3, and 16 have a coordinate function on lignocellulose crystallinity modification. To our knowledge, the GH9B subclass gene function in plants has not yet been discovered. Although the GH9B function could be investigated by reverse genetic analysis, the extremely high co-expression and coordination among OsGH9B1, 3, and 16 suggest that silencing of the individual OsGH9B isoform gene may not result in any significant alteration of lignocellulose crystallinity. In this case, co-silencing of OsGH9B1, 3, and 16 in a mutant may be essential for functional analysis in the future. Hence, non-GH9 mutants are advantageous for function analysis of the entire OsGH9 family genes.
A total of 25 GH9 family genes with 3 subclasses (A, B, and C) were identified in both rice and Arabidopsis. GH9A (KORRIGAN) proteins, such as OsGH9A3 and AtGH9A1, containing the transmembrane domain, could be co-localized with CESA complex , indicating their involvement in cellulose biosynthesis. However, GH9B subclass, such as OsGH9B5 and AtGH9B7, comprised the secreted proteins, suggesting that GH9B should be distinguished from GH9A for cellulose biosynthesis. OsGH9B1/2/3/16 and AtGH9B1/2 were proposed to have the enzymatic activity for lignocellulose crystallinity alteration, other than for cellulose biosynthesis, indicating that these proteins may have specific activities for post-modification of cellulose microfibers in the cell wall. In addition, GH9C subclass comprised the secreted proteins, but had a C-terminal CBD to crystalline cellulose, suggesting a specific role in the turnover of crystalline cellulose , . Although GH9C function was not identified in this study, OsGH9C1 and AtGH9C1/3 were highly and specifically expressed in the radicle tissues, providing potential for further investigation in the future.
Global gene co-expression profiling and correlation analysis based on microarray data from 66 tissues of 2 rice varieties indicated the OsGHA3 and B5 potential role in cellulose biosynthesis. Integrative analysis of OsGH9 gene expression level, cellulase specific activities in situ and in vitro, and lignocellulose crystallinity (CrI) in distinct two rice mutants and wild type revealed that OsGH9B1, 3, and 16 may have enzymatic activities for lignocellulose crystallinity modification. The results can provide new insights into OsGH9 function in plants and offer a strategy for genetic manipulation of OsGH9 genes toward bioenergy crop breeding in rice.
Materials and Methods
Plant materials and growth condition
A japonica rice var. Nipponbare (NPB) and T-DNA insertion homozygous mutant lines (Osfc4, Osfc11) were grown in the field of Huazhong Agricultural University, Wuhan, China. When the tip of panicle just protruded out of the flag leaf at booting stage, four different internodes of stems were collected to analyze the OsGH9 family gene expression, in situ and in vitro cellulase activities and cell wall component determination. When the panicles protruded out 2 cm above the top of the flag leaf, stems were sampled for the analysis of qRT-PCR and cell wall component. All tissues were obtained from 3–6 plants and pooled for each biological replicate in the biological triplicate.
Phylogenetic and structural analysis
The Hidden Markov Model (HMM) profile of the glycoside hydrolase family 9 domain (PF00759.1) was downloaded from PFam http://pfam.sanger.ac.uk/. We employed a name search and the protein family ID PF00759.1 for the identification of OsGH9 genes from the rice genome. Information about the chromosomal localization, coding sequence (CDS), amino acid (AA) and full length cDNA accessions was obtained from TIGR http://www.tigr.org and KOME http://cdna01.dna.affrc.go.jp/cDNA. The multiple alignment analysis with counterparts in Arabidopsis was performed using the Clustal X program (version 1.83) , the unrooted phylogenetic trees were constructed with the MEGA3.1 program  and the neighbor joining method with 1,000 bootstrap replicates.
Motif of protein sequences were analyzed and identified using the MEME program (version 4.0) http://meme.sdsc.edu/meme/cgi-bin/meme.cgi with the following parameters: number of repetitions, any; maximum number of motifs, 25; optimum motif width set to >6 and <200. The motifs were annotated using the InterProScan http://www.ebi.ac.uk/Tools/InterProScan/search_program. The exon-intron structure analysis was performed using GSDS http://gsds.cbi.pku.edu.cn/ . The protein transmembrane helices were predicted by the TMHMM Server V2.0 http://www.cbs.dtu.dk/services/TMHMM/ .
Co-expression profiling analysis
The transcriptional profile data of GH9 family and CESA family genes in 33 tissues (Table S9) of Zhenshan 97 (ZS97) and Minghui 63 (MH63), Z1–Z33 and M1–M33 (Figure 2) was respectively obtained from the CREP database http://crep.ncpgr.cn. Massively parallel signature sequencing (MPSS) data website: http://mpss.udel.edu/rice/mpss_index.php  was used to determine the transcriptional profiles of the genes with conflicting probe set signals. The expression values were log-transformed, and cluster analysis was performed using a software cluster with Euclidean distances and the hierarchical cluster method of complete linkage clustering as described in , . Correlation coefficients of these gene expressions were also calculated to determine whether they are significantly different at 0.01 or 0.05 percent levels, respectively
The gene expression profiling of AtGH9 and AtCESA families in the 66 tissues of Arabidopsis (Table S10) was based on the Gene Expression Omnibus database http://www.ncbi.nlm.nih.gov/geo/ using the GSE series accession numbers GSE5629, GSE5630, GSE5631, GSE5632, GSE5633 and GSE5634. The raw data were processed with the Affymetrix Microarray Analysis Suite (MAS Version 5, Affymetrix) . Subsequent analysis of the gene expression data was performed in the statistical computing language R http://www.r-project.org using packages available from the Bioconductor project http://www.bioconductor.org as described in .
Total RNA was isolated from samples using RNAprep pure Plant Kit (DP432, TIANGEN BIOTECH), and 5 µg total RNA was reverse transcribed with an oligo(dT)18 primer in a 50 µl reaction using an M-MLV Reverse Transcriptase (Promega, USA) according to the manufacturer's instructions. The qRT-PCR was performed in a 20 µl reaction system (cDNA template 2.0 µl, 2×SYBR Green1 Mix10 µl, primer-F 0.5 µl, primer-R 0.5 µl, MilliQ 7.0 µl) with SYBR Green qPCR kit (ZOMANBIO, China) on Two Color Real-time PCR Detection System (MyiQ2, BIO-RAD) using the following program: 2 min at 95°C followed by 40 cycles of 15 sec at 95°C, 15 sec at 60°C, 25 sec at 72°C. Ubiquitin gene (AK059011) was used as an internal standard in the qRT-PCR. The gene expression unit was subjective to the percentage of the target gene expression value relative to the internal standard (Ubiquitin gene). All quantitative PCR experiments were performed in biological triplicate. All the gene-specific primers used were listed in Table S7.
In situ and in vitro cellulase activity assay
In situ cellulase assay was performed as described previously by Takahashi et al  with minor modifications. The 3rd internodes of rice stems were hand-sectioned (≈100 µm) and placed in 0.1 M MES buffer (pH 6.0). Tissue sections were transferred to 0.1 mM Resorufin Cellobioside (Markergene™ Fluorescent Cellulase Assay Kit, Product M1245, USA) in 0.1 M MES buffer (pH 6.0) for 30 min, and then the fluorescence intensity was detected by time-lapse confocal laser scanning microscopy (Carl Zeiss, LSM 510 META) equipped with a green fluorescent filter (excitation 567 nm, emission 595 nm). The same scanning parameters were used for all samples.
In vitro cellulase activity was also detected by Markergene™ Fluorescent Cellulase Assay Kit. Four internodes of rice stems with 0.1 g each were ground to powder in liquid nitrogen, and suspended in 300 µl reaction buffer (100 mM sodium acetate buffer, pH 6.0) at 4°C for 5 min. The supernatant was collected after centrifuge twice at 18,000 g at 4°C for 15 min (Eppendorf Centrifuge 5417 R). 50 µl of supernatant was incubated with 50 µl of 0.5 mM Resorufin Cellobioside (substrate reagent) and reacted for 5 min in a black flat-bottomed 96-well microtiter plate (Greiner Microlon).
Fluorescence intensity of resorufin released was measured with excitation (550 nm) and emission (595 nm), at 35°C for 45 cycles with a cycle time of 1 min using Multimode microplate reader (TECAN Infinite M200) according to the method described by Takahashi J et al . Fluorescence values of blank (50 µl substrate reagent was added to 50 µl reaction buffer) were subtracted at each time point. A standard curve of Resorufin ranging from 0 to 50 µM was prepared to determine concentration of Resorufin in rice internode tissues extracts reacted with Resorufin Cellobioside. The protein concentration of samples was measured by Bradford method in triplicate. All the reactions were performed with biological triplicates.
Plant cell wall fractionation and wall polysaccharide analysis
The plant tissues, including internodes or stems of different stages, were first heated at 105°C for 20 min, dried to constant weight at 60°C for about 7 d and kept dry until use. The extraction and fractionation of cell wall polysaccharides were performed as previously described by Peng et al with minor modification . The crude cell wall material was extracted with 0.5% (w/v) ammonium oxalate and heated for 1 h in a boiling water bath. The remaining pellet was suspended in 4 M KOH containing 1.0 mg.ml−1 sodium borohydride for 1 h at 25°C, and the combined supernatant was neutralized, dialyzed and lyophilized for total hemicelluloses analysis. The KOH non-extractable residue was further extracted with acetic-nitric acids for 1 h at 100°C and the remaining pellet was used for cellulose determination. All samples were carried out in biological triplicate.
Colorimetric assay of total hexoses and pentoses: UV–VIS Spectrometer (V-1100D, Shanghai MAPADA Instruments Co., Ltd. Shanghai, China) was used for the absorbance reading. Hexoses were detected using the anthrone/H2SO4 method, and pentoses were detected using the orcinol/HCl method. For cellulose determinations, the cellulose sample was dissolved in 67% (v/v) H2SO4 (1.0 ml) with shaking at 25°C for 1 h, and then 10.0 µl aliquot was used for the anthrone/H2SO4 method. Total hemicelluloses level was subject to the sum total of hexoses and pentoses. Considering that high pentoses level can affect the absorbance reading at 620 nm for hexoses content by anthrone/H2SO4 method, the deduction from pentoses reading at 660 nm was carried out for final hexoses calculation. A series of xylose concentrations were analyzed for plotting the standard curve referred for the deduction, which was verified by GC-MS analysis. All experiments were carried out in biological triplicate.
Total lignin content was determined by two-step acid hydrolysis method. The solution was filtered with membrane filter (0.22 µm). 20 µl solution was injected into a HPLC (Waters 1525 HPLC) column Kromat Universil C18 (4.6 mm×250 mm, 5 µm) operating at 28°C with CH3OH∶H2O∶HAc (25∶74∶1, v/v/v) carrier liquid at the flow rate of 1 ml.min−1. All experiments were carried out in biological triplicate.
Detection of the crystallinity index
Detection of crystallinity index of lignocellulose (CrI) was described by Xu et al  with a minor modification. The internodes of rice stem tissues at booting stages were cut into small pieces through 40 mesh sieve. The fine raw biomass powder of the plant tissue was laid on the glass sample holder (35×50×5 mm), analyzed under plateau conditions: Ni-filtered Cu Kα radiation (λ = 0.154056 nm) generated at 40 kV and 18 mA, and scanned at speed (0.0197°/s ) from 10° to 45°. CrI was calculated as 100×(I200−Iam)/I200, where I200 is the intensity of the 200 peak (θ = 22.5°), Iam is the intensity at the minimum between the 200 and 110 peaks (θ = 18.5°). Standard error was detected at ± 0.05∼0.15 using five samples in triplicate.
The SPSS 17.0 was used for statistical analysis. Correlation coefficients were generated by performing Spearman rank correlation analysis for all pairs of measured traits across the whole population. This analysis used average values calculated from all original determinations for a given traits pair.
Representative OsGH9 and OsCESA gene expression pattern by qRT-PCR in six tissues in rice variety (NPB). Eighteen representative OsGH9 family genes and six OsCESA genes (OsCESA4, 7 and 9, OsCESA1, 3 and 8) detected by qRT-PCR in six tissues of japonica rice variety Nipponbare (NPB).
Time course of in vitro cellulase specific activities in four internodes of mutants (fc4 and fc12) and wild type (NPB) at booting stage. Total proteins used for in vitro cellulase activity using the Resorufin Cellobioside as substrate in the time-course of 45 cycles with a cycle time of 1 min each at 35°C.
Co-expression profiling among AtGH9 and AtCESA family genes in Arabidopsis. The expression profiling of AtGH9 and AtCESA family genes based on Arabidopsis microarray data GSE5629, GSE5630, GSE5631, GSE5632, GSE5633 and GSE5634, and performed by the hierarchical cluster analysis.
Phylogenetic comparison of GH9 families in rice and Arabidopsis. The sequences of GH9 family proteins obtained from rice (http://rice.plantbiology.msu.edu/) and Arabidopsis (http://www.arabidopsis.org/) were aligned with Clustal X program and then constructed a phylogenetic tree using MEGA3.1 software.
Information of 25 OsGH9 genes in rice.
Correlation coefficients between OsGH9 and OsCESA expression levels in 66 tissues of ZS97 and MH63 (n = 66).
Correlation coefficients among OsGH9 expression levels in 66 tissues of ZS97 and MH63 (n = 66).
Transcript levels (%) changes of OsGH9 and OsCESA in four internodes of mutants (fc4 and fc11) and wild type (NPB) at booting stages.
Correlation coefficients between OsGH9 genes expression level and cellulase specific activity or lignocellulose CrI in 12 internodes of mutants (fc4 and fc11) and wild type (NPB) at booting stages (n = 12).
Correlation coefficients between OsGH9 and OsCESA expression levels in rice at booting stage (n = 12).
Primer pairs of genes for qRT-PCR.
List of 25 putative motifs in the OsGH9 family proteins.
Tissue samples from 33 developmental stages of two rice varieties.
Conceived and designed the experiments: LP. Analyzed the data: GX BY. Wrote the paper: GX BY. Revised and finalized the manuscript: LP. Performed the bioinformatics data analysis: LW KG. Completed major experiments: GX BY. Completed the microscopic observation: ZX. Participated in rice mutant selection: FL KG MZ LW WZ YW.
- 1. Delmer DP (1999) CELLULOSE BIOSYNTHESIS: Exciting Times for A Difficult Field of Study. Annu Rev Plant Physiol Plant Mol Biol 50: 245–276.
- 2. Ragauskas AJ, Williams CK, Davison BH, Britovsek G, Cairney J, et al. (2006) The path forward for biofuels and biomaterials. Science 311: 484–489.
- 3. Himmel ME, Ding SY, Johnson DK, Adney WS, Nimlos MR, et al. (2007) Biomass recalcitrance: engineering plants and enzymes for biofuels production. Science 315: 804–807.
- 4. Kaida R, Kaku T, Baba K, Oyadomari M, Watanabe T, et al. (2009) Loosening xyloglucan accelerates the enzymatic degradation of cellulose in wood. Mol Plant 2: 904–909.
- 5. Park S, Baker JO, Himmel ME, Parilla PA, Johnson DK (2010) Cellulose crystallinity index: measurement techniques and their impact on interpreting cellulase performance. Biotechnol Biofuels 3: 10.
- 6. Segal LCJ, Martin AE, Conrad CM (1959) An empirical method for estimating the degree of crystallinity of native cellulose using the X-ray diffractometer. Textile Research J 786–794.
- 7. Pear JR, Kawagoe Y, Schreckengost WE, Delmer DP, Stalker DM (1996) Higher plants contain homologs of the bacterial celA genes encoding the catalytic subunit of cellulose synthase. Proc Natl Acad Sci U S A 93: 12637–12642.
- 8. Taylor NG, Howells RM, Huttly AK, Vickers K, Turner SR (2003) Interactions among three distinct CesA proteins essential for cellulose synthesis. Proc Natl Acad Sci U S A 100: 1450–1455.
- 9. Persson S, Paredez A, Carroll A, Palsdottir H, Doblin M, et al. (2007) Genetic evidence for three unique components in primary cell-wall cellulose synthase complexes in Arabidopsis. Proc Natl Acad Sci U S A 104: 15566–15571.
- 10. Tanaka K, Murata K, Yamazaki M, Onosato K, Miyao A, et al. (2003) Three distinct rice cellulose synthase catalytic subunit genes required for cellulose synthesis in the secondary wall. Plant Physiol 133: 73–83.
- 11. Wang L, Guo K, Li Y, Tu Y, Hu H, et al. (2010) Expression profiling and integrative analysis of the CESA/CSL superfamily in rice. BMC Plant Biol 10: 282.
- 12. Scheible WR, Eshed R, Richmond T, Delmer D, Somerville C (2001) Modifications of cellulose synthase confer resistance to isoxaben and thiazolidinone herbicides in Arabidopsis Ixr1 mutants. Proc Natl Acad Sci U S A 98: 10079–10084.
- 13. Desprez T, Vernhettes S, Fagard M, Refregier G, Desnos T, et al. (2002) Resistance against herbicide isoxaben and cellulose deficiency caused by distinct mutations in same cellulose synthase isoform CESA6. Plant Physiol 128: 482–490.
- 14. Harris DM, Corbin K, Wang T, Gutierrez R, Bertolo AL, et al. (2012) Cellulose microfibril crystallinity is reduced by mutating C-terminal transmembrane region residues CESA1A903V and CESA3T942I of cellulose synthase. Proc Natl Acad Sci U S A 109: 4098–4103.
- 15. Nicol F, His I, Jauneau A, Vernhettes S, Canut H, et al. (1998) A plasma membrane-bound putative endo-1,4-beta-D-glucanase is required for normal wall assembly and cell elongation in Arabidopsis. EMBO J 17: 5563–5576.
- 16. Schindelman G, Morikami A, Jung J, Baskin TI, Carpita NC, et al. (2001) COBRA encodes a putative GPI-anchored protein, which is polarly localized and necessary for oriented cell expansion in Arabidopsis. Genes Dev 15: 1115–1127.
- 17. Libertini E, Li Y, McQueen-Mason SJ (2004) Phylogenetic analysis of the plant endo-beta-1,4-glucanase gene family. J Mol Evol 58: 506–515.
- 18. Davison A, Blaxter M (2005) Ancient origin of glycosyl hydrolase family 9 cellulase genes. Mol Biol Evol 22: 1273–1284.
- 19. Hayashi T, Yoshida K, Park YW, Konishi T, Baba K (2005) Cellulose metabolism in plants. Int Rev Cytol 247: 1–34.
- 20. Henrissat B (1991) A classification of glycosyl hydrolases based on amino acid sequence similarities. Biochem J 280 (Pt 2) 309–316.
- 21. Robert S, Bichet A, Grandjean O, Kierzkowski D, Satiat-Jeunemaitre B, et al. (2005) An Arabidopsis endo-1,4-beta-D-glucanase involved in cellulose synthesis undergoes regulated intracellular cycling. Plant Cell 17: 3378–3389.
- 22. Somerville C (2006) Cellulose synthesis in higher plants. Annu Rev Cell Dev Biol 22: 53–78.
- 23. Lopez-Casado G, Urbanowicz BR, Damasceno CM, Rose JK (2008) Plant glycosyl hydrolases and biofuels: a natural marriage. Curr Opin Plant Biol 11: 329–337.
- 24. Molhoj M, Johansen B, Ulvskov P, Borkhardt B (2001) Expression of a membrane-anchored endo-1,4-beta-glucanase from Brassica napus, orthologous to KOR from Arabidopsis thaliana, is inversely correlated to elongation in light-grown plants. Plant Mol Biol 45: 93–105.
- 25. Yoshida K, Komae K (2006) A rice family 9 glycoside hydrolase isozyme with broad substrate specificity for hemicelluloses in type II cell walls. Plant Cell Physiol 47: 1541–1554.
- 26. Urbanowicz BR, Bennett AB, Del Campillo E, Catala C, Hayashi T, et al. (2007) Structural organization and a standardized nomenclature for plant endo-1,4-beta-glucanases (cellulases) of glycosyl hydrolase family 9. Plant Physiol 144: 1693–1696.
- 27. Karczmarek AFS, Lichocka M, Sobczak M, Kurek W, et al. (2008) Expression of two functionally distinct plant endo-beta-1,4-glucanases is essential for the compatible interaction between potato cyst nematode and its hosts. Mol Plant Microbe Interact 21: 791–798.
- 28. Zuo J, Niu QW, Nishizawa N, Wu Y, Kost B, et al. (2000) KORRIGAN, an Arabidopsis endo-1,4-beta-glucanase, localizes to the cell plate by polarized targeting and is essential for cytokinesis. Plant Cell 12: 1137–1152.
- 29. Szyjanowicz PM, McKinnon I, Taylor NG, Gardiner J, Jarvis MC, et al. (2004) The irregular xylem 2 mutant is an allele of korrigan that affects the secondary cell wall of Arabidopsis thaliana. Plant J 37: 730–740.
- 30. Takahashi J, Rudsander UJ, Hedenstrom M, Banasiak A, Harholt J, et al. (2009) KORRIGAN1 and its aspen homolog PttCel9A1 decrease cellulose crystallinity in Arabidopsis stems. Plant Cell Physiol 50: 1099–1115.
- 31. Maloney VJ, Mansfield SD (2010) Characterization and varied expression of a membrane-bound endo-beta-1,4-glucanase in hybrid poplar. Plant Biotechnol J 8: 294–307.
- 32. Zhou HL, He SJ, Cao YR, Chen T, Du BX, et al. (2006) OsGLU1, a putative membrane-bound endo-1,4-beta-D-glucanase from rice, affects plant internode elongation. Plant Mol Biol 60: 137–151.
- 33. Peng L, Kawagoe Y, Hogan P, Delmer D (2002) Sitosterol-beta-glucoside as primer for cellulose synthesis in plants. Science 295: 147–150.
- 34. Lane DR, Wiedemeier A, Peng L, Hofte H, Vernhettes S, et al. (2001) Temperature-sensitive alleles of RSW2 link the KORRIGAN endo-1,4-beta-glucanase to cellulose synthesis and cytokinesis in Arabidopsis. Plant Physiol 126: 278–288.
- 35. Ohmiya YSM, Shiroishi M, Amano Y, Kanda T, et al. (2000) Evidence that endo-1,4-beta-glucanases act on cellulose in suspension-cultured poplar cells. Plant J 24: 147–158.
- 36. Park YW, Tominaga R, Sugiyama J, Furuta Y, Tanimoto E, et al. (2003) Enhancement of growth by expression of poplar cellulase in Arabidopsis thaliana. Plant J 33: 1099–1106.
- 37. Hartati S, Sudarmonowati E, Park YW, Kaku T, Kaida R, et al. (2008) Overexpression of poplar cellulase accelerates growth and disturbs the closing movements of leaves in sengon. Plant Physiol 147: 552–561.
- 38. Chundawat SP, Bellesia G, Uppugundla N, da Costa Sousa L, Gao D, et al. (2011) Restructuring the crystalline cellulose hydrogen bond network enhances its depolymerization rate. J Am Chem Soc 133: 11163–11174.
- 39. Molhoj M, Pagant S, Hofte H (2002) Towards understanding the role of membrane-bound endo-beta-1,4-glucanases in cellulose biosynthesis. Plant Cell Physiol 43: 1399–1406.
- 40. Xie G, Peng L (2011) Genetic engineering of energy crops: a strategy for biofuel production in China. J Integr Plant Biol 53: 143–150.
- 41. Persson S, Wei H, Milne J, Page GP, Somerville CR (2005) Identification of genes required for cellulose synthesis by regression analysis of public microarray data sets. Proc Natl Acad Sci U S A 102: 8633–8638.
- 42. Wang L, Xie W, Chen Y, Tang W, Yang J, et al. (2010) A dynamic gene expression atlas covering the entire life cycle of rice. Plant J 61: 752–766.
- 43. Bosch M, Mayer CD, Cookson A, Donnison IS (2011) Identification of genes involved in cell wall biogenesis in grasses by differential gene expression profiling of elongating and non-elongating maize internodes. J Exp Bot 62: 3545–3561.
- 44. Ruprecht C, Mutwil M, Saxe F, Eder M, Nikoloski Z, et al. (2011) Large-scale co-expression approach to dissect secondary cell wall formation across plant species. Front Plant Sci 2: 23.
- 45. Bhandari S, Fujino T, Thammanagowda S, Zhang D, Xu F, et al. (2006) Xylem-specific and tension stress-responsive coexpression of KORRIGAN endoglucanase and three secondary wall-associated cellulose synthase genes in aspen trees. Planta 224: 828–837.
- 46. Zhang JW, Xu L, Wu YR, Chen XA, Liu Y, et al. (2012) OsGLU3, a putative membrane-bound endo-1,4-beta-glucanase, is required for root cell elongation and division in rice (Oryza sativa L.). Mol Plant 5: 176–186.
- 47. Loopstra CA, Mouradov A, Vivian-Smith A, Glassick TV, Gale BV, et al. (1998) Two pine endo-beta-1,4-glucanases are associated with rapidly growing reproductive structures. Plant Physiol 116: 959–967.
- 48. Molhoj M, Ulvskov P, Dal Degan F (2001) Characterization of a functional soluble form of a Brassica napus membrane-anchored endo-1,4-beta-glucanase heterologously expressed in Pichia pastoris. Plant Physiol 127: 674–684.
- 49. Ibatullin FM, Banasiak A, Baumann MJ, Greffe L, Takahashi J, et al. (2009) A real-time fluorogenic assay for the visualization of glycoside hydrolase activity in planta. Plant Physiol 151: 1741–1750.
- 50. Burton RA, Gibeaut DM, Bacic A, Findlay K, Roberts K, et al. (2000) Virus-induced silencing of a plant cellulose synthase gene. Plant Cell 12: 691–706.
- 51. Crowell EF, Gonneau M, Stierhof YD, Hofte H, Vernhettes S (2010) Regulated trafficking of cellulose synthases. Curr Opin Plant Biol 13: 700–705.
- 52. Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG (1997) The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res 25: 4876–4882.
- 53. Kumar S, Tamura K, Nei M (2004) MEGA3: Integrated software for Molecular Evolutionary Genetics Analysis and sequence alignment. Brief Bioinform 5: 150–163.
- 54. Guo AY, Zhu QH, Chen X, Luo JC (2007) [GSDS: a gene structure display server]. Yi Chuan 29: 1023–1026.
- 55. Krogh A, Larsson B, von Heijne G, Sonnhammer EL (2001) Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes. J Mol Biol 305: 567–580.
- 56. Nobuta K, Venu RC, Lu C, Belo A, Vemaraju K, et al. (2007) An expression atlas of rice mRNAs and small RNAs. Nat Biotechnol 25: 473–477.
- 57. Team RDC (2009) R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing.
- 58. Peng L, Hocart CH, Redmond JW, Williamson RE (2000) Fractionation of carbohydrates in Arabidopsis root cell walls shows that three radial swelling loci are specifically involved in cellulose production. Planta 211: 406–414.
- 59. Xu N, Zhang W, Ren S, Liu F, Zhao C, et al. (2012) Hemicelluloses negatively affect lignocellulose crystallinity for high biomass digestibility under NaOH and H2SO4 pretreatments in Miscanthus. Biotechnol Biofuels 5: 58.