Changes in cell wall polysaccharide composition, gene transcription and alternative splicing in germinating barley embryos

Journal of Plant Physiology 191 (2016) 127–139

Contents lists available at ScienceDirect

Journal of Plant Physiology journal homepage: www.elsevier.com/locate/jplph

Physiology

Changes in cell wall polysaccharide composition, gene transcription and alternative splicing in germinating barley embryos Qisen Zhang a , Xiaoqi Zhang b , Filomena Pettolino c , Gaofeng Zhou d , Chengdao Li a,b,d,∗ a

Australian Export Grains Innovation Centre, 3 Baron-Hay Court, South Perth, WA 6155, Australia Western Barley Genetics Alliance, Murdoch University, 90 South Street, Murdoch, WA 6150 Australia Agriculture Flagship, CSIRO, Black Mountain, ACT 2601, Australia d Department of Agriculture and Food Western Australia, 3 Baron-Hay Court, South Perth, WA 6155, Australia b c

a r t i c l e

i n f o

Article history: Received 20 September 2015 Received in revised form 17 December 2015 Accepted 17 December 2015 Available online 2 January 2016 Keywords: Alternative splicing Cell wall Differentially expressed gene Germination Polysaccharide RNA-sequencing

a b s t r a c t Barley (Hordeum vulgare L.) seed germination initiates many important biological processes such as DNA, membrane and mitochondrial repairs. However, little is known on cell wall modifications in germinating embryos. We have investigated cell wall polysaccharide composition change, gene transcription and alternative splicing events in four barley varieties at 24 h and 48 h germination. Cell wall components in germinating barley embryos changed rapidly, with increases in cellulose and (1,3)(1,4)-␤-d-glucan (20–100%) within 24 h, but decreases in heteroxylan and arabinan (3–50%). There were also significant changes in the levels of type I arabinogalactans and heteromannans. Alternative splicing played very important roles in cell wall modifications. At least 22 cell wall transcripts were detected to undergo either alternative 3 splicing, alternative 5 splicing or intron retention type of alternative splicing. These genes coded enzymes catalyzing synthesis and degradation of cellulose, heteroxylan, (1,3)(1,4)-␤-d-glucan and other cell wall polymers. Furthermore, transcriptional regulation also played very important roles in cell wall modifications. Transcript levels of primary wall cellulase synthase, heteroxylan synthesizing and nucleotide sugar inter-conversion genes were very high in germinating embryos. At least 50 cell wall genes changed transcript levels significantly. Expression patterns of many cell wall genes coincided with changes in polysaccharide composition. Our data showed that cell wall polysaccharide metabolism was very active in germinating barley embryos, which was regulated at both transcriptional and posttranscriptional levels. © 2015 Elsevier GmbH. All rights reserved.

1. Introduction The cell wall is unique to plants and is important for development, growth and survival in diverse environments. The plant

Abbreviations: AG, arabinogalactan; AIR, alcohol insoluble residues; AS, alternative splicing; A3S, alternative 3 splicing; A5S, alternation 5 splicing; IBSC, International Barley Sequencing Consortium; CesA, cellulose synthase A; CslF, cellulose synthase-like F; DEG, differentially expressed gene; EI, (1,3)(1,4)-␤-d-glucan endohydrolase isoenzyme I; GT, glycosyl transferase; HvUGAE, UDP-glucouronic acid epimerase; HvUGE, UDP-Glc epimerase; HvUXE, UDP-Xyl epimerase; HvUXS, UDP-Xyl synthase; HX, heteroxylan; IR, intron retention; IUM, initially unmapped; QC, quality control; RPKM, reads per kilo base-pair per million mapped reads; TF, transcription factor; XET, xyloglucan endotransglucosylase/hydrolase. ∗ Corresponding author at: Australian Export Grains Innovation Centre, 3 BaronHay Court, South Perth, WA 6155, Australia. E-mail addresses: [email protected] (Q. Zhang), [email protected] (X. Zhang), [email protected] (F. Pettolino), [email protected] (G. Zhou), [email protected] (C. Li). http://dx.doi.org/10.1016/j.jplph.2015.12.007 0176-1617/© 2015 Elsevier GmbH. All rights reserved.

cell wall is a complicated mixture of polysaccharides including cellulose, heteroxylan (HX), (1,3)(1,4)-␤-d-glucan, pectin, lignin and other polymers. Cell wall composition varies between different parts of the barley grain. Cell walls of barley endosperm contain high levels of (1,3)(1,4)-␤-d-glucans (75% v/v) and substantial amounts of HX (20% v/v), while cell walls in aleurone layers contain high levels of HX (71%,v/v) and moderate amounts of (1,3)(1,4)-␤d-glucans (26% v/v) (Fincher, 2010). Cellulose levels are extremely low in the cell walls of endosperm and aleurone layers (2%, v/v). Cell walls in barley grains also contain small amounts of other polysaccharides including glucomannans, arabinogalactans, arabinans and xyloglucans. Biosynthesis and modification of plant cell wall polysaccharides require a suite of enzymes. The number of genes coding for these enzymes and regulatory proteins is more than 1200 (Burton et al., 2010). HX backbones are synthesized by members of the glycosyltransferase family 43 (GT43) and glycosyltransferase family 47 (GT47) (Scheller and Ulvskov, 2010). The

128

Q. Zhang et al. / Journal of Plant Physiology 191 (2016) 127–139

glucuronosyl substitution of the backbone is catalyzed by members of the glycosyltransferase family 8 (GT8) (Mortimer et al., 2010), while the ␣-(1,3)-linked arabinosyl substitution for the HX backbone is catalyzed by members of the glycosyltransferase family 61 (GT61) (Anders et al., 2012). Members of the cellulose synthaselike family F (CslF) and cellulose synthase-like family H (CslH) (both GT2) are responsible for the biosynthesis of (1,3)(1,4)-␤-dglucans (Burton et al., 2006; Doblin et al., 2009). Furthermore, a suite of hydrolytic enzymes may regulate polysaccharide levels in germinating barley grains. These include (1,4)-␤-d-xylan endohydrolases (GH10), arabinoxylan arabinofuranohydrolase (AXAH, GH51), ␣-l-arabinofuranosidase (GH3) and ␤-d-xylosidase (GH3) for HX degradation (Fincher, 2010), and (1,3)(1,4)-␤-d-glucan endohydrolase (GH17), (1,4)-␤-d-glucan glucohydrolase (GH1) and ␤-d-glucan exohydrolase (GH3) for (1,3)(1,4)-␤-d-glucan degradation (Fincher, 2010). Several glycosyl hydrolases in GH5, GH9 and GH16 may also be involved in modifications to (1,3)(1,4)-␤-dglucans, as implicated by association mapping with two-row spring and winter barley (Houston et al., 2014). Synthesis of cellulose in barley is catalyzed by cellulose synthases (CesA, GT2) with at least seven CesA genes identified in barley (Burton et al., 2004). Xyloglucan endotransglucosylase/hydrolase (XET) is another important cell wall enzyme family. It catalyzes the cross-linking between different polysaccharides including cellulose, (1,3)(1,4)-␤-d-glucan and xyloglucans (Hrmova et al., 2009, 2007). Alternative splicing (AS) plays a key role in post-transcription regulations. It regulates many biological processes including hormone-mediated signal transduction and stress- or lightinduced responses (Eckardt, 2013; Reddy et al., 2013; Staiger and Brown, 2013; Syed et al., 2012; Wu et al., 2014). It determines tissue-specific differentiation patterns and controls plant development and adaptation to environmental conditions. Several types of AS events have been detected in plants including alternative 3 splicing (A3S), alternative 5 splicing (A5S) and intron retention (IR). Seed germination initiates a transition from metabolically inactive to active phases for many biological processes and eventually develops to a growing seedling. Cellular membrane and mitochondria repairs have been known to be important in the initiation of seed germination (Bewley, 1997). Transcription of cell wall genes is also very active in germinating barley seeds (An and Lin, 2011). However, little is known about cell wall polysaccharide biosynthetic activities and their regulations at transcriptional and post-transcriptional (e.g., AS) levels. In order to understand the cell wall biosynthetic activities during the seed germination, we have established the cell wall polysaccharide composition by monosaccharide linkage analysis at 24 and 48 h germination in four barley varieties. Monosaccharide linkage data showed that cell wall polysaccharide modification activities were very active during barley seed germination. We have also conducted RNA-sequencing experiments (RNA-seq) to investigate the involvement of AS events and differentially expressed genes (DEG) on cell wall biosynthesis. RNA-seq data showed that transcription of several cell wall genes was very active in germinating embryos and also coincided with changes in profiles of cellulose, HX, (1,3)(1,4)-␤-glucan and arabinan. Thus, we concluded that cell wall modification was one of the most important biological processes during the seed germination. Transcription and post-transcription regulations play key roles in the cell wall modifications.

2. Materials and methods 2.1. Plant materials The barley (Hordeum vulgare L.) varieties Bass, Baudin, Harrington and Stirling were used for this study. Seeds were sterilized with

1% (w/v) hypochrorite solution for 15 min, rinsed with 3–5 volumes of running water, transferred to a 14 cm Petri dish covered with two layers of filter paper and 10 mL water and incubated in the dark. Germinating embryos were separated from endosperms at 24 h and 48 h. Many small batches of barley seeds were used to induce germination and embryos were collected at 24 h and 48 h. The time needed for collecting embryos were not lasted for more than 20 min for each small batch. Different batch embryos were pooled and about 200–500 mg of embryo tissues were used for cell wall preparation and RNA purification. The embryos were stored at −80 ◦ C before use. 2.2. Cell wall preparation and cell wall polysaccharide analyses Three biological replications of embryos (collected at different days) were used for cell wall preparation. For preparation of alcohol insoluble residues (AIR), barley embryos were ground in liquid nitrogen to fine powder, extracted 5–6 times with 80% ethanol (v/v), and once each with acetone and methanol. AIR was de-starched by amylase according to Pettolino et al. (2012). Monosaccharide-linkage analysis was performed by methylation with methyl iodide in sodium hydroxide and DMSO (Ciucanu and Kerek, 1984) followed by hydrolysis, reduction and acetylation. Data were calculated as mol% of total AIR (Pettolino et al., 2012). Monosaccharide linkages (mol%) and relative polysaccharide proportions were deduced from the partially methylated alditol acetates that were separated and analyzed by GC–MS according to Pettolino et al. (2012). 2.3. Sequencing and bioinformatics analysis strategy Total RNA was extracted from embryos with phenolSDS reagents (http://onlinelibrary.wiley.com/doi/10.1002/ 0471142727.mb0403s09/pdf, accessed on 7 December 2015). RNA sequencing and bioinformatics analyses followed a flow chart in Supplemental Fig. S1. 2.3.1. Sequencing We assessed the quality of the RNA preparation by the Agilent 2100 Bioanalyzer system with a RNA integrated number value greater than 8, after genomic DNA contamination was removed by DNase. We enriched messenger RNA by magnetic beads with Oligo (dT) and fragmented them. The first strand cDNA was synthesized with the mRNA fragments as templates using reverse transcriptase and random hexamer primers. The second-strand cDNA was synthesized with first strand cDNA as template after adding reaction buffer, dNTPs, RNase H, DNA polymerase and MgCl2 , purified with QIAquick PCR Purification Kit (Qiagen) and end-repaired with T4 DNA polymerase and Klenow DNA polymerase. After adding an “A” base to the 3 end, cDNA was ligated to the sequencing adapters with a ‘T’ base overhang at the 3 end. DNA fragments with length from 250 bps to 500 bp were selected from electrophoresis gels and purified with QIAquick PCR Purification Kit before PCR amplification (15 cycles). The quality of cDNA libraries were assessed by an Agilent 2100 Bioanalyzer and ABI StepOnePlus Real-Time PCR system. RNA sequencing was done on a HiSeq2000 (Illumina) using a paired-end sequencing protocol (Pease and Sooknanan, 2012). 2.3.2. Raw data quality control and alignment to reference genome A quality control step was carried out on raw data by using FastQC software (version 0.11.2, http://www.bioinformatics. babraham.ac.uk/projects/fastqc/, accessed on 7 December 2015). Briefly, the percentage of A/G should roughly equal that of T/C and the rate of N should be less than 1%, while the proportion of low

Q. Zhang et al. / Journal of Plant Physiology 191 (2016) 127–139

quality (Qs < 20) bases should be less than 10% after base composition were determined. The dirty reads with sequencing adapters, N rate over 10% and proportion of low quality reads more than 50% were all discarded. The samples with low quality values on raw data were re-sequenced (Supplemental Fig. S1). After raw data QC, clean reads were aligned to the barley reference genome (International Barley Sequencing Consortium, IBSC) (Mayer et al., 2012) with SOAPaligner/Soap2 (version 2.21) software. 2.3.3. Alignment QC We conducted an alignment QC by using SOAPcoverage. Coverage, depth and mapped rates (including unique mapped reads, total mapped reads and unmapped reads) were evaluated. 2.3.4. Gene expression analysis We used RPKM (reads per kilobase-pair per million mapped reads) to represent mRNA levels, since it eliminated the effects of gene length and data amount. RPKM =

106 C NL/103

where, C is the number of reads mapped perfectly on a gene X, N is the total number of reads uniquely mapped on reference genes and L is the length of gene X. Finally, FDR (false discovery rate) and DEGs between 24 h and 48 h were calculated with an algorithm developed on the theoretical basis of the Poisson Distribution and Multiple Testing (Benjamini and Yekutieli, 2001). A stringent threshold of FDR ≤ 0.001 and the absolute value of log2 ratio ≥1 (for DEG) were used to represent a significant change. 2.3.5. Alternative splicing transcript identification Alternative splicing generates different variants of mRNA transcripts. They are translated into distinguishable proteins (Black, 2003; Stamm et al., 2005). We evaluated a few software for detecting splicing junctions, including TopHat, SpliceMap and MapSplice and selected SOAPsplice (version 1.10) for this work (Huang et al., 2011). SOAPsplice produced the highest call rate (detected true junction number/true junction in total) and the lowest false positive rate (detected false junction number/the number of junctions all detected). Briefly, SOAPsplice first mapped complete reads to the reference genome. Initially unmapped (IUM) reads were then mapped with the spliced alignment algorithm, in which, SOAPsplice divided the IUM reads into two segments, which are expected to be derived from different exons in the premature mRNA. SOAPsplice first found the longest 5 end segment of an IUM read that could be mapped to the reference, then aligned the remaining segment to the reference sequences. Important criteria were: 1) Each segment should be longer than 8 bp. 2) No mismatch and no gaps were allowed in the alignment of each segment. 3) Distance of two segments is expected to range between 50 bp and 50,000 bp. This range covers the majority of known intron size in Eukaryote (Trapnell et al., 2009). 4) The boundary of an intron should be in the form of “GT–AG”, “GC–AG” or “AT–AC”. When spliced alignment produced multiple hits, a splice junction candidate with “GT–AG” boundary was given the highest priority, followed by candidates with “GC–AG” and “AT–AC” boundaries (Burset et al., 2000). 5) When the segments had multiple hits to the reference, consideration was given to the cases where one segment had a unique hit while the other had multiple hits, or each segment had at most 3 hits. The closest pair of hit was reported for this read. SOAPsplice ignored the other cases.

129

6) Since the reads might span more than 2 exons, we applied an additional step to detect junctions with reads longer than 50 bp. For the reads between 50 bp–100 bp, SOAPsplice spliced the read into two equal size segments. For reads longer than 100 bp, SOAPsplice spliced the reads into multiple segments of 50 bp from the 5 end until the remaining segment was in length between 50 and 100 bp. SOAPsplice then spliced this remaining part into two equal segments. SOAPsplice considered each segment as a sub-read, and treated it with the above alignment step. 7) Finally, SOAPsplice checked and concatenated the separated alignment hits for sub-reads to build the alignment for the original read (Huang et al., 2011). 2.4. Annotation of cell wall genes For the current barley genome, annotation of cell wall genes was not completed. Thus, important cell wall gene families coding for the enzymes catalyzing the synthesis of major cell wall polysaccharide components were manually annotated. Firstly, GenBank accession numbers and gene sequences of cloned cell wall genes were obtained from the literature (Supplemental Table S1) and GenBank, respectively. Secondly, the gene sequences were used to blast the in-house WGS morex contig sequence databases downloaded from IBSC website ftp://ftpmips.helmholtz-muenchen.de/ plants/barley/public data/, accessed on 5 December 2015 (Mayer et al., 2012). The corresponding morex contigs and gene models were obtained from top hit with a low e-value (<1e-5). The manually annotated genes included CesA, CslF, CslH, GT43, GT47, GT75, GT61 families, XET and nucleotide sugar inter-conversion genes as listed in Supplemental Table S1. The other cell wall genes were according to IBSC annotations in barley genomic databases (Mayer et al., 2012). 2.4.1. Quantitative RT-PCR Three biological replications of embryos (collected from different days) were used for RNA extraction and quantitative RT-PCR. Four cell wall genes ((1,3)(1,4)-␤-d-glucan endohydrolase isoenzyme I (EI), MLOC 62746; XET, MLOC 18499; CslF6, MLOC 57200 and HvUXE2, MLOC 60662) and a Gibberellin receptor GID1L2 gene (MLOC 59369) were selected as representatives for the confirmation of gene expression by quantitative RT-PCR. The primer pairs used were listed in Supplemental Table S2. Fresh total RNA was prepared by using RNeasy Plant Mini kit from Qiagen (Qiagen Pty., Ltd., Victoria, Australia). Contamination of genomic DNA was removed by RNase-Free DNase Set according to manufacturer’s protocol (Qiagen Pty., Ltd., Australia). First strain cDNA was synthesized by using SensiFAST cDNA Synthesis Kit (Bioline Pty/Ltd., Australia). The reaction mixture contained 2 ␮g total RNA, 4 ␮l TransAMP buffer, 1 ␮l reverse transcriptase and water to make 20 ␮l. First strain cDNA synthesis was carried out at 25 ◦ C for 10 min, 42 ◦ C for 25 min and 48 ◦ C for 15 min. The reaction was stopped by incubating at 85 ◦ C for 5 min. Gene transcript levels were detected by using SeniFAST HRM Kit from Bioline (Bioline Pty/Ltd., Australia). The reaction mixture contained 5 ␮l 2 x SensiFAST HRM mix, 0.5 ␮M forward primer, 0.5 ␮M reverse primer, 1 ␮l 20x diluted cDNA and water to make 10 ␮l. The program used for the reaction was 1 cycle of 95 ◦ C for 2 min, 40 cycles of 95 ◦ C for 5 s, 60 ◦ C for 10 s and 72 ◦ C for 20 s. At the end of the amplification reaction, a melting curve was determined to ensure that primer pairs used for quantitative RTPCR generated only one single PCR product. The melting program was as following: ramping from 65 ◦ C to 95 ◦ C, rising 1◦ each step, waiting for 90 s of the pre-melt condition on the first step, waiting 5 s for each step afterward. Barley actin was used as reference gene for quantitative RT-PCR. The transcript levels were expressed

130

Q. Zhang et al. / Journal of Plant Physiology 191 (2016) 127–139

as the ratios of tested gene expression over actin gene expression (Supplemental Fig. 2). The RNA-seq data was deposited to the NCBI GEO database with accession number GSE66024. 3. Results 3.1. Cell wall polysaccharide composition changed rapidly in germinating barley embryo wall Barley cell wall monosaccharide linkages were analyzed by a methylation method and polymer contents were deduced according to Pettolino et al. (2012). There were rapid changes in monosaccharide linkage in germinating embryo wall between 24 h and 48 h (Supplemental Table S3). The deduced polysaccharide composition was also dynamic particularly with the major wall components including HX, cellulose, xyloglucans and (1,3)(1,4)-␤d-glucans (Table 1). 3.1.1. Heteroxylan Heteroxylan consists of a (1,4)-linked ␤-d-xylopyranosyl residue backbone, some of which is substituted with single ␣-larabinofurosyl residues at the C(O)3 position and at C(O)2 to a lesser extent. The degree of substitution varies considerably and depends on tissue type, physiological conditions and developmental stage. This polymer was a major component in germinating barley embryo walls, comprising 35–41% at 24 h (Table 1, mol%) which is less than that found in aleurone layers (71%, v/v) and more than that in endosperm cell walls (20%, v/v) (Fincher, 2010). At 48 h, HX content had decreased by 3–15% in four barley varieties (Table 1). The ratios of substituted xylose over unsubstituted xylose were 4.1–4.9 at 24 h, and decreased to 1.8–3.3 at 48 h (Table 1), indicating that arabinosyl residues were rapidly removed from HX backbones in the germinating embryos. The removal of arabinosyl substituents from HX backbone chains would alter the physicochemical properties of the polymer with a subsequent enhancement of wall strength. The unsubstituted regions of (1,4)-␤-xylan facilitate hydrogen bonding between neighboring HX chains and between HX and other cell wall polysaccharides, including cellulose, (1,3)(1,4)-␤-glucan and xyloglucan (Fincher, 2009). However, a balance of wall strength and wall loosening is very important in the extruding radicle and coleoptiles in the germinating embryos. 3.1.2. Cellulose Cellulose is a linear molecule with (1,4)-linked ␤-dglucopyranosyl residues. Twenty-four or 36 cellulose molecules aggregate to form microfibers to serve as structural frames in the cell wall. Cellulose was the second richest component in germinating barley embryo wall with about 20% (mol%) at 24 h. At 48 h, cellulose levels had increased by 20–50% in four barley varieties. Germinating barley embryos had much higher cellulose levels than that in the cell walls of the aleurone layer or the endosperm, where cellulose content was only 2% (Fincher, 2010). 3.1.3. Xyloglucan Xyloglucans consist of a (1,4)-linked ␤-d-glucosyl residue backbone, some of which is substituted with xylosyl side chains. It was the third richest component in germinating barley embryo wall ranging from 12 to 15% in four barley varieties (Table 1). At 48 h, xyloglucan levels had not significantly changed. 3.1.4. Arabinan Arabinan is a molecule with (1,5)-linked ␣-l-arabinofuranosyl residues. There was a significant amount of arabinan (about 5%) in

germinating barley embryo wall at 24 h. The levels of this polymer had decreased at 48 h by 12–50% (Table 1). 3.1.5. (1,3)(1,4)-ˇ-d-glucan Polymer of (1,3)(1,4)-␤-d-glucan is a linear molecule with (1,3)and (1,4)-linked ␤-d-glucosyl residues. Cell wall in germinating barley embryos contained 1.8–2.5% of (1,3)(1,4)-␤-d-glucan at 24 h. At 48 h, the levels of (1,3)(1,4)-␤-d-glucan increased by 30–100% (Table 1). Germinating barley embryo walls had much lower levels of (1,3)(1,4)-glucans than that in the cell walls in the aleurone layer (26%, w/w) or the starchy endosperm (75%, w/w) (Fincher, 2010). 3.1.6. Type I arabinogalactan (AG), type II AG and heteromannans Type I AG, type II AG and heteromannans are minor components of barley embryos at about 2 mol%. At 48 h, the levels of these three polymers had decreased substantially. About 13% of other cell-wall-related monosaccharide linkages did not fit into any of the above mentioned polysaccharides according to the predicted structure. These form the category listed as “other” (Table 1). 3.2. Determination of genes responsible for cell wall modifications After determining the active changes in cell wall polysaccharide composition, we investigated the regulation of cell wall modifications at transcriptional and post-transcriptional levels by conducting RNA-seq experiments. We pooled a couple hundred of embryos for RNA-seq experiments for four barley varieties (Bass, Harrington, Stirling and Baudin) at 24 h and 48 h germination and used stringent conditions for RNA-seq data QC and reliable software for bioinformatics analysis. Furthermore, we discussed only common gene transcription activities and alternatively spliced genes occurred in at least three out of four barley varieties. 3.2.1. RNA-seq read statistics and alignments to barley reference genome All four barley RNA libraries from Bass, Baudin, Harrington and Stirling generated similar numbers of reads (approximately 26 million) and accounted up for about 2.3 G base pairs for each library (Supplemental Table S4). The reads mapped to genomic and annotated gene sequences were 67% and 45%, respectively. Gene coverage estimated the percentage of a gene covered by reads. About 60% of the genes had 90–100% gene coverage, 12% of genes had 80–90% gene coverage and 15% had 50–80% gene coverage (Supplemental Table S5). We have made sure that that the RNA-seq data had high qualities before gene expression analysis. 3.2.2. Confirmation of gene expression by quantitative RT-PCR Four cell wall genes (EI, MLOC 62746; XET, MLOC 18499; CslF6, MLOC 57200 and HvUXE2, MLOC 60662) and a Gibberellin receptor GID1L2 gene (MLOC 59369) were selected as representatives for quantitative RT-PCR confirmation. The expression profiles detected by both RNA-seq and quantitative RT-PCR were very similar in four barley varieties. For example, the expression levels of Gibberellin receptor gene and HvUXE2 gene were lower than the expression levels of other three genes for both set of data. Decreases in transcript levels at 48 h were observed for most of these genes in the data collected from both RNA-seq and quantitative RT-PCR experiments (Supplemental Fig. S2). 3.2.3. The mRNA levels of genes responsible for cellulose synthesis There are at least seven CesA genes in barley (Supplemental Table S1) (Burton et al., 2004). Four genes (CesA1, CesA2, CesA3 and CesA6) are thought to be primary cell wall genes, while CesA4, CesA5/7 and CesA8 are secondary wall genes. Transcripts of three primary wall genes (CesA1, CesA2 and CesA6) were most abundant

Q. Zhang et al. / Journal of Plant Physiology 191 (2016) 127–139

131

Table 1 Polysaccharide composition of germinating barley embryos. Polysaccharides

Bass

Harrington

24 h

48 h

5.9 ± 0.3 3.0 ± 0.1 0.8 ± 0.0 41.3 ± 2.0 2.2 ± 0.2 12.3 ± 0.6 2.0 ± 0.3 20.2 ± 0.7 12.4 ± 1.4 4.9 ± 0.3

Arabinan Type I AG Type II AG Heteroxylan (1,3)(1,4)-glucan Xyloglucan Heteromannan Cellulose Other Sub/unSub xyl

24 h

4.3 ± 0.1 2.5 ± 0.2 1.0 ± 0.1 36.3 ± 2.0 3.1 ± 0.5 15.0 ± 0.1 1.7 ± 0.1 24.3 ± 0.7 11.6 ± 1.7 3.3 ± 0.3

Stirling 48 h

5.6 ± 0.3 2.8 ± 0.1 0.7 ± 0.2 35.8 ± 2.2 2.5 ± 0.4 14.2 ± 0.2 2.2 ± 0.1 21.5 ± 1.1 14.6 ± 2.0 4.1 ± 0.3

24 h

1.9 ± 0.1 1.6 ± 0.1 0.9 ± 0.1 31.0 ± 2.2 5.6 ± 0.4 13.9 ± 0.4 1.6 ± 0.1 34.5 ± 0.8 9.1 ± 1.0 1.8 ± 0.1

5.1 ± 0.1 2.7 ± 0.1 0.6 ± 0.0 41.4 ± 2.6 1.8 ± 0.1 12.6 ± 0.5 2.1 ± 0.2 20.0 ± 0.6 13.8 ± 1.7 4.4 ± 0.4

Baudin 48 h

24 h

4.3 ± 0.3 2.2 ± 0.1 0.9 ± 0.1 39.9 ± 1.5 2.3 ± 0.2 13.5 ± 0.7 1.8 ± 0.2 23.8 ± 0.7 11.4 ± 0.7 2.7 ± 0.1

5.4 ± 0.2 2.7 ± 0.3 0.8 ± 0.1 37.9 ± 1.7 1.9 ± 0.3 14.4 ± 1.2 2.2 ± 0.1 21.5 ± 0.9 13.3 ± 1.5 4.2 ± 0.4

48 h 2.6 ± 0.3 1.5 ± 0.1 1.3 ± 0.1 32.0 ± 2.1 4.4 ± 0.4 13.5 ± 1.2 1.5 ± 0.1 32.9 ± 1.1 10.3 ± 0.7 1.8 ± 0.1

Barley cell wall monosaccharide linkages were analyzed by a methylation method with polymer contents deduced according to Pettolino et al. (2012). Sub/unSub xyl: substituted xylosyl residues over unsubstituted xylosyl residues in HX backbone. Other: cell wall-related monosaccharide linkages which did not fit into any of the polysaccharides according to predicted structure; AG: arabinogalactan. Data were means of three replicates with indications of standard deviations.

350 300

a

MLOC_55153.1 (CesA1) MLOC_62778.2 (CesA2) MLOC_61930.2 (CesA3) MLOC_55081.3 (CesA6)

RPKM

250 200 150 100 50 0 20

MLOC_66568.3 (CesA4)

b

MLOC_43749.1 (CesA5/7)

RPKM

15

MLOC_68431.4 (CesA8)

10 5 0 40

c

Cellulose

mol %

30

Har: Harrington Stir: Stirling Bau: Baudin

20 10 0 Bass 24 h

Bass 48 h

Har 24 h

Har 48 h

Sr 24 h

Sr 48 h

Bau 24 h

Bau 48 h

Fig. 1. The mRNA levels of CesA genes. The mRNA levels were determined by RNA-seq from geminating embryos. The gene annotations were according to Supplemental Table S1. (a) Transcript levels of primary wall CesA genes, (b) transcript levels of secondary wall CesA genes, (c) cellulose contents in embryos with indications of standard deviations

at 24 h (>250 RPKM, Fig. 1a). The mRNA levels of CesA3 were 5–13fold lower than those of CesA1, CesA2 and CesA6. The mRNA levels of all three secondary wall CesA genes (CesA4, CesA5/7 and CesA8) were 10-fold lower than those of CesA1, CesA2 and CesA6 (Fig. 1b). The mRNA level of CesA5/7 increased significantly at 48 h in Bass, Stirling and Baudin (Table 2), but the other secondary CesA mRNA had small increases (Fig. 1b). The high transcript levels of primary wall CesA genes would result in high enzyme levels for continuous primary cell wall synthesis in the germinating embryos. Thus, a significant increase in cell wall cellulose (Fig. 1c) could mainly attribute to primary wall components produced by a high level of primary cell wall enzymes and also contained some secondary wall

components synthesized by the secondary wall enzymes such as CesA5/7. 3.2.4. The mRNA levels of the genes coding for glycosyltransferase families GT43, GT47, GT75 and GT61 proteins Families GT43 and GT47 proteins are involved in xylan backbone biosynthesis, while families GT61 and GT8 proteins catalyze the substitutions of xylan backbones with side chains (Anders et al., 2012; Scheller and Ulvskov, 2010). Family GT75 proteins have been implicated in HX biosynthesis as shown in co-immunoprecipitation experiments (Zeng et al., 2010). A GT43 (MLOC 8254) gene and a GT47 (MLOC 61533) gene were highly transcribed in

132

Q. Zhang et al. / Journal of Plant Physiology 191 (2016) 127–139

Table 2 Cell wall DEG. Cell wall genes whose expression levels changed significantly in at least three out of four barley varieties between 24 h and 48 h. Gene annotations were according to IBSC. GeneID

Bass

Har

Stir

Bau

Readable Description

1.20 1.31 0.91 −1.59 1.47 1.04 0.69 1.03 0.62 1.12 1.32 0.45 1.33 −2.01 −0.90 1.29 1.41 −0.02 −0.68 −0.74 1.12 −2.04 1.45 −0.65 −0.77 1.61 1.85 0.83 −1.22 2.08 0.62 1.75 −0.58 −0.03 −1.72 −0.90 0.73 1.08 −0.72 −0.74 −0.18 −0.44 1.48 0.68 1.53 −1.87 1.22 −0.06 1.86 1.33

2.19 2.42 2.21 −2.45 2.90 1.66 1.43 1.85 1.30 1.61 1.35 1.10 1.56 −3.02 −3.99 1.25 0.72 -1.92 −2.94 −3.36 1.26 −2.86 0.04 −1.52 −2.35 0.96 3.30 1.86 −1.58 5.24 1.84 3.43 −2.11 −1.12 −2.08 −3.78 1.66 3.61 −1.55 −1.78 −2.26 −1.65 1.21 1.57 1.24 −2.34 2.34 −1.44 3.32 2.86

␣-galactosidase ␣-glucosidase like protein ␣-glucosidase like protein ␤-glucosidase like protein ␤-d-glucosidase ␤-d-xylosidase ␤-galactosidase ␤-glucosidase Endo-1,4-␤-d-glucanase Cellulose synthase (CesA5/7) Endo-1,3;1,4-␤-d-glucanase Endo-1,4-␤-xylanase A Expansin Expansin Expansin 1 expansin B4 Expansin protein Expansin protein Expansin-B6 Expansin-B6 Glucan endo-1,3-␤-glucosidase 1 Glucan endo-1,3-␤-glucosidase 7 Glycosyltransferase GT61, Glycosyltransferase Glycosyltransferase GT43 Pectin lyase-like superfamily protein Pectin lyase-like superfamily protein Pectinesterase Pectinesterase inhibitor domain containing protein Similarity to alpha galactosidase UDP-glycosyltransferase UDP-glycosyltransferase UDP-glycosyltransferase UDP-Glycosyltransferase superfamily protein UDP-Glycosyltransferase superfamily protein UDP-Glycosyltransferase superfamily protein UDP-Glycosyltransferase superfamily protein UDP-Glycosyltransferase superfamily protein UDP-Glycosyltransferase superfamily protein UDP-Glycosyltransferase superfamily protein UDP-Glycosyltransferase superfamily protein Xyloglucan endotransglucosylase/hydrolase Xyloglucan endotransglucosylase/hydrolase 7 Xyloglucan endotransglucosylase/hydrolase protein Xyloglucan endotransglucosylase/hydrolase protein Xyloglucan endotransglucosylase/hydrolase protein Xyloglucan endotransglucosylase/hydrolase protein xyloglucan endotransglycosylase (PM5) Xylose isomerase

Log2 ratio MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC

15778.1 15619.3 68876.3 13058.1 56792.1 62475.1 34610.2 9865.1 69717.1 43749.1 58795.3 60943.2 54344.1 63724.2 43237.1 60217.2 64368.1 70189.1 76675.1 51411.1 8669.1 5621.1 55095.1 68728.2 70966.4 4722.2 54501.1 54500.1 37134.2 39585.1 12729.2 68938.2 66675.1 79717.1 65675.1 57690.1 64192.1 65245.1 60784.1 61721.1 72804.1 16287.1 56250.1 71749.1 64805.1 8320.2 21703.1 61972.1 14727.2 55107.1

2.45 3.03 2.12 −2.43 2.47 2.13 1.78 2.15 1.57 1.69 2.18 1.17 3.25 −2.05 −3.07 1.61 2.58 −2.23 −2.52 −3.57 2.03 −2.42 1.67 −1.33 −2.11 1.79 2.87 1.49 −1.51 3.53 1.80 3.12 −1.50 −1.30 −2.08 −3.34 2.05 3.05 −1.43 −1.19 −1.69 −1.65 1.78 1.90 2.06 −2.60 2.85 −1.75 3.59 3.00

1.87 1.77 1.77 −1.62 0.81 1.14 1.41 0.59 1.49 0.40 0.58 1.22 0.17 −0.23 −4.33 2.62 1.21 −2.84 −3.98 −5.73 1.08 −2.03 2.14 −1.47 −3.36 1.10 1.56 2.28 −0.71 4.54 1.54 1.84 −2.62 −1.94 −2.26 −4.86 1.85 1.10 −1.34 −1.82 −2.63 −2.41 −0.41 2.63 -0.30 −1.10 3.52 −1.71 2.51 1.79

geminating embryos (150–250 RPKM) (Fig. 2a). There were at least two GT75 family transcripts in barley embryos; the mRNA level of a GT75 gene (MLOC 6065) was extremely high (1000–2500 RPKM) (Fig. 2b). Two GT61 family transcripts were also detected in germinating embryos; the mRNA level of a GT61 gene (MLOC 6357) was moderate (30–100 RPKM), while mRNA level of the other (MLOC 68728) was high (over 100 RPKM) in Bass, Harrington and Baudin at 24 h (Fig. 2c). Two of the highly-expressed HX backbone genes (MLOC 8254, MLOC 61533) had a slight decrease in mRNA levels at 48 h, but the GT61 family gene (MLOC 68728) responsible for side chain modifications showed a substantial decrease in mRNA levels (Fig. 2c). The data showed that the highly expressed GT43 (MLOC 8254) and GT47 (MLOC 61533) could have contributed to the maintenance of the high levels of HX in germinating embryos (Table 1 and Fig. 2d). On the other hand, the significant decrease in the transcript level of the side chain gene (MLOC 68728) was

consistent with a decrease in the ratios of substituted xylose over unsubstituted xylose (Table 1).

3.2.5. The mRNA levels of the genes responsible for xylan degradation Transcripts of three ␤-d-xylosidase (Fig. 3a) and six endo-1,4-␤d-xylanase (Fig. 3b) were detected in germinating barley embryos. The transcripts of most of these genes were below 10 RPKM. However, transcripts coded a ␤-d-xylosidase (MLOC 62475) and an endo-1,4-␤-xylanase (MLOC 64936) were abundant (100–400 RPKM). The mRNA levels of the ␤-d-xylosidase gene had increased more than two-fold at 48 h in all four barley varieties (Table 2), while expression of the endo-1,4-␤-xylanase gene had decreased substantially in Harrington, Stirling and Baudin (Fig. 3b). Elevated mRNA levels of the ␤-d-xylosidase (MLOC 62475) gene could be

Q. Zhang et al. / Journal of Plant Physiology 191 (2016) 127–139

300

a

MLOC_4722.2 (GT43) MLOC_8254.1 (GT43) MLOC_64806.1 (GT47) MLOC_61533.1 (GT47) MLOC_76930.1 (GT47)

RPKM

200 100 0 3000

RPKM

133

b

MLOC_63185.1 (GT75) MLOC_6065.1 (GT75)

2000 1000 0

RPKM

150

c

MLOC_68728.2 (GT61) MLOC_6357.1 (GT61)

100 50 0 50

d

HX

Mol %

40

Arabinan

30

Har: Harrington Stir: Stirling Bau: Baudin

20 10 0 Bass 24 h

Bass 48 h

Har 24 h

Har 48 h

Sr 24 h

Sr 48 h

Bau 24 h

Bau 48 h

Fig. 2. The mRNA levels of families GT43, GT47, GT75 and GT61 genes. The mRNA levels were determined by RNA-seq from geminating embryos. The gene annotations were according to Supplemental Table S1. (a) GT43 and GT 47 family gene expression; (b) GT75 gene expression; (c) GT61 gene expression, (d) HX and arabinan levels with indications of standard deviations.

associated with reduced HX contents in germinating embryos (Table 1 and Fig. 2d). 3.2.6. The mRNA levels of the genes responsible for (1,3)(1,4)-ˇ-d-glucan synthesis and degradation The family GT2 members (CslF and CslH) catalyze the biosynthesis of (1,3)(1,4)-␤-d-glucan in barley (Burton et al., 2008; Burton et al., 2006; Doblin et al., 2009). The mRNA levels of CslF6 (MLOC 57200) were very high (over 200 RPKM) and showed a substantial decrease at 48 h in four barley varieties (Fig. 4a), while the mRNA levels of CslH and other CslF genes were very low. CslF6 could have played very important roles in the synthesis of (1,3)(1,4)-␤d-glucan in germinating embryos, since it was the only GT2 family gene which were highly transcribed. Degradation of (1,3)(1,4)-␤-d-glucan is catalyzed by the glycoside hydrolase family GH17, GH1 and GH3 proteins (Fincher, 2010; Fincher et al., 1986). The mRNA of (1,3)(1,4)-␤-d-glucan endohydrolase isoenzyme I gene (EI, GH17) (MLOC 62746) and a ␤glucosidase gene (MLOC 37740) was most abundant in germinating embryos (400–1500 RPKM) (Fig. 4b). The transcript levels of these genes increased substantially in Bass, Stirling and Baudin at 48 h. Transcripts of several other genes implicated in glucan hydrolysis were also very abundant including glucan endo-1,3-␤-glucosidase (MLOC 5621), glucan endo-1,3-␤-glucosidase (MLOC 10919), ␤-dglucosidase (MLOC 67725) genes, to name a few (Supplemental Table S6). However, these genes showed decreasing transcription profiles at 48 h.

3.2.7. The mRNA levels of expansin and XET genes Both XET and expansins play key roles in auxin- and gravityinduced cell elongations (Cosgrove, 2000; Zhang et al., 2011). They facilitate structural modifications to cell walls. The enzyme XET catalyzes the cross-linking between different polysaccharides including cellulose, (1,3)(1,4)-␤-d-glucan and xyloglucans (Hrmova et al., 2009, 2007). The mRNA of three XET genes (MLOC 18499, MLOC 61972 and MLOC 4539) were abundant (>100 RPKM). The transcript levels of these XET genes decreased substantially at 48 h (Fig. 5a). The mRNA levels of four expansin genes (MLOC 63724, MLOC 73204, MLOC 70189 and MLOC 12906) were most abundant at 24 h (>200 RPKM) (Fig. 5b). The transcript levels of these genes also decreased substantially at 48 h. 3.2.8. The mRNA levels of genes coding for nucleotide sugar inter-conversion enzymes Plant cell wall polysaccharide biosynthesis requires activated precursors including UDP-Glc, UDP-Gal, UDP-Xyl, UDP-Ara, UDPGluA and UDP-GalUA (Zhang et al., 2006, 2005, 2010). The mRNA levels of the genes coding for UDP-Glc dehydrogenase (MLOC 5287, MLOC 63077 and MLOC 70967), UDP-Glc epimerase (MLOC 72010) and UDP-GluA epimerase (MLOC 65544) were very high (>200 RPKM) at 24 h (Fig. 6). It is worth notice the three UDPGlc dehydrogenases would catalyze carbon entry from UDP-Glc to other UDP-sugars including UDP-GluA, UDP-GalUA, UDP-Xyl and UDP-Ara. The high transcript levels of these nucleotide sugar interconversion genes would result in a large amount of enzymes and

134

Q. Zhang et al. / Journal of Plant Physiology 191 (2016) 127–139

400

MLOC_60721.1 MLOC_54205.1 MLOC_62475.1

a

RPKM

300 200 100 0

RPKM

500 450 400

MLOC_75090.1 MLOC_15272.2 MLOC_73983.1 MLOC_60943.2 MLOC_64936.1 MLOC_61506.1

b

350 300 250 200 150 100 50 0

Har: Harrington Stir: Stirling Bau: Baudin

Bass 24h

Bass 48h

Har 24h

Har 48h

Sr 24h

Sr48h

Bau 24 h

Bau 48 h

Fig. 3. The mRNA levels of HX degradation genes. The mRNA levels were determined by RNA-seq from geminating embryos. The gene annotations were according to IBSC barley genomic database (Mayer et al., 2012). (a) Transcript levels of ␤-d-xylosidase genes, (b) transcript levels of endo-1,4-␤-d-xylanase genes.

generate a large activated-substrate pool for active cell wall modifications. 3.2.9. Co-expression of cell wall genes with transcription factor genes Fifty-one of transcription factor (TF) genes underwent significant changes in transcript levels between 24 h and 48 h (Supplemental Table S7). Thirty-one of them showed increases in transcript levels, while 20 TF genes decreased transcription activity. Co-expression analyses showed that transcription of many TF genes were either positively or negatively correlated with transcription of cell wall genes (Supplemental Table S8). Some of TF genes (MLOC 54950, MLOC 70777, MLOC 38567, MLOC 7325, MLOC 60672, MLOC 66841, MLOC10221 and MLOC 64636) were co-expressed with over 10 cell wall genes (Supplemental Table S8). It was noticed that three transcription factor genes (MLOC 7077, MLOC 75519 and MLOC 53580) showed similar co-expression patterns with cell wall genes, all of them were co-expressed with at least a glycosyltransferase gene (MLOC 68728), a cellulase gene (MLOC 69717) and a XET gene (MLOC 4539). 3.3. Numbers and types of AS events in germinating embryos Large numbers of AS events (2635–3901) and genes (2229–3201) were detected in the germinating barley embryo (Supplemental Table S9). About 1.2 AS events per AS gene were estimated (Supplemental Table S9). The current gene annotation in barley has 15754 intron-containing genes (Mayer et al., 2012). Thus, 14–20% of intron-containing genes underwent AS during in germinating barley embryos. Moreover, A3S was the major AS event in the germinating barley embryo, accounting for 39–46% of total AS events, while IR and A5S accounted for 29–36% and 18–23%

Table 3 Number of AS transcripts commonly occurring in at least three out of four barley varieties. AS types

24 h

48 h

Turnover

A3S IR A5S Total

299 (45) 229 (34) 141 (21) 669

228 (34) 216 (32) 108 (16) 552

155 (52) 105 (46) 80 (57)

Common AS genes occurring in at least three out of four barley varieties were identified at 24 h and 48 h. Number in the brackets indicate the% of each type of AS transcripts over total AS genes. The last column (turnover) is the number of AS transcripts appeared at 24 h, but disappeared at 48 h. The numbers in brackets indicate% of AS transcripts disappeared at 48 h.

of total AS events, respectively. Of other AS event types, 2–3% was exon skipping, while alternative first exon and alternative last exon were negligible, and the mutually-exclusive exon was not detectable. Most of AS transcripts detected at 24 h had disappeared at 48 h (Supplemental Table S9), indicating that AS was dynamic. We have also analyzed common AS transcripts detected in at least three out of four barley varieties at 24 h and 48 h, the number of AS genes were 669 and 552 at 24 h and 48 h, respectively (Table 3). The percentages of A3S, IR and A5S transcripts over total AS transcripts were 34–45%, 32–34% and 16–21%, respectively. About half of commonly detected AS transcripts identified at 24 h were not present at 48 h (Table S3). 3.3.1. Regulations by alternative splicing Cell wall modifications could be regulated by AS, since at least 22 of the genes responsible for cell wall polysaccharide metabolism were detected to undergo AS in germinating barley embryos (Table 4). These included genes coding for enzymes

Q. Zhang et al. / Journal of Plant Physiology 191 (2016) 127–139

400

135

MLOC_52692.1(CslF8) MLOC_53007.1(CslH) MLOC_59289.1(CslF3) MLOC_59327.1(CslF9) MLOC_57200.2 (CslF6)

a

RPKM

300 200 100 0 2000

MLOC_62746.1 (EI)

RPKM

MLOC_37740.3

b

1500 1000 500 0 7 6

(1,3)(1,4)-β-D-Glucan

c

5 Mol %

4 3

Har: Harrington Stir: Stirling Bau: Baudin

2 1 0 Bass 24 h

Bass 48 h

Har 24 h

Har 48 h

Sr 24 h

Sr 48 h

Bau 24 h

Bau 48 h

Fig. 4. The mRNA levels of genes implicated in (1,3)(1,4)-␤-d-glucan synthesis and degradations. The mRNA levels were determined by RNA-seq from geminating embryos. The gene annotations were according to Supplemental Table S1 and IBSC. (a) Transcript levels of CslF genes, (b) transcript levels of (1,3;1,4)-␤-glucanase (EI) gene and a ␤-glucosidase gene, (c) (1,3)(1,4)-␤-d-glucan levels with indications of standard deviations.

MLOC_18499.1 MLOC_76160.3 MLOC_71748.1

a 500

MLOC_61972.1 MLOC_60000.1 MLOC_5607.1

MLOC_4539.2 MLOC_61233.2 MLOC_56250.1

MLOC_8320.2 MLOC_80115.1

450

RPKM

400 350 300 250 200 150 100 50 0

RPKM

b 1000 900 800 700 600 500 400 300 200 100 0

MLOC_63724.2 MLOC_43237.1 MLOC_76675.1

Bass 24h

Bass 48h

MLOC_73204.3 MLOC_65431.2 MLOC_64368.1

Har 24h

MLOC_70189.1 MLOC_51411.1 MLOC_15691.1

Har 48h

Sr 24h

MLOC_12906.1 MLOC_81957.1 MLOC_16693.3

Sr48h

Har: Harrington Stir: Stirling Bau: Baudin

Baudin 24h

Baudin 48h

Fig. 5. The mRNA levels of XET and expansin genes. The mRNA levels were determined by RNA-seq from geminating embryos. The gene annotations were according to IBSC barley genomic database (Mayer et al., 2012). (a) Transcript levels of XET genes, (b) transcript levels of expansin genes.

136

Q. Zhang et al. / Journal of Plant Physiology 191 (2016) 127–139

MLOC_5287.1(HvUGD2) MLOC_63077.1(HvUGD1) MLOC_72010.2(HvUGE3)

1400

MLOC_70967.1(HvUGD3)

1200

MLOC_65544.2(UGAE1) MLOC_60662.1(HvUXE2)

1000 RPKM

MLOC_5738.2(HvUXE1) MLOC_70713.1(HvUGE1)

800

MLOC_67741.2(HvUXE3)

600

MLOC_71845.1(UGAE2)

400 200 0 Bass 24h

Bass 48h

Har 24h

Har 48h

Sr 24h

Sr 48h

Bau 24h

Bau 48h

Har: Harrington Stir: Stirling Bau: Baudin

Fig. 6. The mRNA levels of genes coding for nucleotide sugar inter-conversion enzymes. The mRNA levels were determined by RNA-seq from geminating embryos. The gene annotations were according to Supplemental Table S1.

Table 4 Alternative splicing of cell wall transcripts. Gene ID

MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC MLOC

67725 37740 7890 55153 34425 54898 12906 70189 5621 54331 10919 63574 75464 60662 72010 44472 57500 67741 60000 76160 18499 62746

Bass

Harrington

Stirling

Baudin

24 h

48 h

24 h

48 h

24 h

48 h

IR IR IR A3S A3SIR IR A3S A3S A3SIR A3S IR

IR IR IR A3S A3S

IR

IR

IR IR IR A3S A3S IR IR

IR IR IR

AS IR A3S A3S

A3S A3S A3S

IR IR A3SIR A3SIR IR A5S

A3SIR

A3S IR IR IR A3SIR A3S IR IR A5S IR

A5S

IR A3S IR IR A3S A3S

IR A3S A3S

A3S A3S IR

A3S A3S IR IR

IR IR A3S A3S IR IR IR A5S

A3S IR IR

A3S IR

IR

A5S

IR IR A3S IR A3S IR IR IR A5S A5S

IR IR IR A3S A3S IR IR IR A5S A5S IR

A5S

24 h IR IR A3S A3S IR IR A3S IR A3S A3S IR A3S A3S IR IR IR A3S IR A3S IR IR

Readable descriptions 48 h

IR A3S IR IR

A3S IR

IR IR IR

A5S A5S A5S

A5S IR A5S

B-d-glucosidase B-glucosidase B-glucosidase CesA1 Endo-1,3,1,4-␤-d-glucanase Endoglucanase Expansin protein Expansin protein GH17 Glucan endo-1,3-␤-glucosidase Glucan endo-1,3-␤-glucosidase 4 Glucan endo-1,3-␤-glucosidase 5 Glucan endo-1,3-␤-glucosidase 7 HvUXE2 UDP-glucuronic acid epimerase 3 UDP-GT UDP-GT HvUXE3 XET XET XET (1,3;1,4) ␤-glucanase (EI)

Barley embryos were used for RNA-seq experiments. Alternative splicing was detected by software SOAPsplice. Cell wall genes were according to the annotations in either Table 1 or IBSC. Only those genes with same AS types in at least three out of four barley varieties were listed. A3S, alternative 3 splicing; A5S, alternative 5 splicing; IR, intron retention; GT, Glycosyltransferase; XET, Xyloglucan endotransglucosylase/hydrolase.

catalyzing metabolism of major cell wall components such as cellulose, HX, (1,3)(1,4)-␤-d-glucan and xyloglucans. A CesA1 (MLOC 55153) transcript underwent A3S in all four barley varieties (Bass, Harrington, Stirling and Baudin). The product of this gene catalyzed cellulose synthesis (Burton et al., 2004) and the mRNA level of this gene was very high in germinating embryos (Fig. 1a). Two UDP-Xyl epimerase transcripts (MLOC 60662 and MLOC 67741) underwent IR type of AS. They coded enzymes catalyzing the interconversion of UDP-Xyl and UDP-Ara, which are the activated substrates for HX synthesis (Zhang et al., 2010). Three XET transcripts underwent either IR type of AS (MLOC 18499) or A5S (MLOC 60000 and MLOC 76160). The XET transcripts coded for enzymes modifying cell wall polysaccharide structure (Hrmova et al. 2009, 2007,7). Furthermore, other cell wall transcripts undergoing AS commonly occurred in at least three barley varieties coded

for EI (MLOC 62746), ␤-d-glucosidase (MLOC 67725, MLOC 37740 and MLOC 7890), expansin (MLOC 12906 and MLOC 70189) and glucan endo-1,3-␤ glucosidase (MLOC 54331, MLOC 10919 and MLOC 75464), to name a few (Table 4). Some of the transcripts underwent two types of AS such as genes coding for endo(1,3)(1,4)-␤-d-glucanase (MLOC 34425) and UDP-GluA epimerase (MLOC 72010, Table 4).

4. Discussion 4.1. An active cell wall modification is observed in germinating embryos Barley seed germination starts after the uptake of water. Many biochemical processes transition from quiescent to active

Q. Zhang et al. / Journal of Plant Physiology 191 (2016) 127–139

phases. Some of cellular organelles undergo rapid repair to become fully functional, including cellular membrane and mitochondria (Bewley, 1997). Here, we have demonstrated that cell wall composition and possibly structures undergo very active modifications in germinating embryos as deduced from rapid changes in cell wall polysaccharide components (Table 1) and the abundance of the transcripts coding for major cell wall enzymes and proteins (Figs. 1–6). Nevertheless, the monosaccharide linkage analysis showed that the major cell wall composition are HX and cellulose, accounting for 32–41% and 20–34%, respectively. Metabolic trend of these polymers is a decrease in HX contents accompanied with removal of branch chains from HX backbones as shown by a decrease in ratios of substituted over un-substituted xylose (Table 1) and an increase in cellulose contents. Germinating barley embryos consist of two major organs, the elongating roots and coleoptiles. They grow rapidly and require their cell wall expansible during cell elongation. Primary wall possesses the expandable properties with a low cellulose content (Carpita and Gibeaut, 1993). On the other hand, cell wall in these organs need to maintain their rigidity for supporting growth and keeping cell shapes. An increase in secondary wall may have started, therefore resulting in a constant modification of the cell wall in the elongating embryos.

4.2. Changes in transcript profiles are consistent with cell wall modification activities Two important biological processes may occur on germinating embryo wall. One is a breakdown of old cell wall such as wall in coleorhiza for emerging rootlets to recycle carbon sources for new plant growth. A suit of cell wall hydrolases are involved in this process. Thus, we have observed substantial increases in transcription activities of hydrolyzing genes coding for ␣-galactosidase, ␣-glucosidase, ␤-galactosidase, ␤-glucosidase, ␤-d-xylosidase, ␤d-glucanase, pectinlyase, and pectin esterase, to name a few (Table 2). The transcripts of two genes coding for glucan hydrolyzing enzyme EI (MLOC 62746) and ␤-glucosidase (MLOC 37740) were not only abundant, but also showed an increasing pattern at 48 h (Fig. 4b, Table 2). All these gene products may play important roles in the degradation of old cell wall and recycle of carbon sources in germinating embryos. The second important biological process is for cell elongation, which requires cell wall loosing in elongating rootlets and emerging coleoptiles (Cosgrove, 2000). The cell wall loosing agents include expansin, XET and endo1,4-␤-d-glucanase (Cosgrove, 2005). One of the genes coding for endo-1,4-␤-d-glucanase (MLOC 69717) increases transcript levels for more than two folds in three out of four barley varieties (Table 2), while seven expansin genes and seven XET genes were detected to have significant changes in transcript levels (Table 2). The changes in transcript levels of these genes may be related to the cell elongation in germinating embryos. Furthermore, substitutions of HX backbone with arabinosyl residues is catalyzed by GT61 family enzymes (Anders et al., 2012; Chiniquy et al., 2012). A decrease in transcription activities of GT61 family genes (MLOC 68728 and MLOC 6357, Fig. 2c) is as expected, since the ratios of substituted/unsubstituted HX decrease in germinating embryos (Table 1). A decrease in substitution would facilitate the aggregation of xylan backbones with each other and with other polymers; this would change cell wall physicochemical properties (Fincher, 2009). Moreover, the increase in transcription activities of CesA5/7 gene (MLOC 43749, Fig. 1b) implies an increase in secondary wall biosynthesis. Both the decrease in HX backbone substitutions and the increase in secondary wall biosynthesis would increase wall rigidity as required for root and coleoptile extension. Overall, a balance of cell wall loosing and cell wall rigidity is the most

137

important in germinating embryos. Therefore, changes in transcriptions of important genes are accordingly. 4.3. Regulations of cell wall modifications by transcription factors in germinating embryos Regulation of secondary wall metabolism by TFs has been intensively studied in Arabidopsis and other vascular plants. A cascade TF regulations exist with NAC domain TFs acting as master switches and MYB TFs servicing as second-level regulators (Taylor-Teeples et al., 2015; Zhong et al., 2010, 2008). Barley genome contains at least 52 NAC and 119 MYB domain TF genes (Supplemental Table S10). Two NAC (MLOC 37104 and MLOC 61270) and one MYB TF genes are highly transcribed. The expression level of a NAC TF gene (MLOC 61270) halved at 48, while the MYB and the other NAC TF genes do not show changes in transcript levels between 24 h and 48 h. Co-expression analysis for the three secondary wall genes (MLOC 66568, MLOC 43749 and MLOC 68431) shows that they may co-expressed only with a BHLH transcription factor gene (MLOC 37971) and a GATA transcription factor gene (MLOC 70809), but not with currently annotated NAC and MYB TFs (Supplemental Table S11). Little information is available on the regulation of primary wall metabolism by TFs. Our RNA-Seq data show that some of TFs may regulate cell elongation, since expressions of these TF genes are correlated with the expression of the cell wall loosing (XET and cellulase) agents (Supplemental Table S8). 4.4. Possible involvements of alternative splicing in cell wall modifications Alternative splicing occurs at a post-transcriptional regulatory stage and controls many developmental processes in planta (Eckardt, 2013; Reddy et al., 2013; Staiger and Brown, 2013). It generates more than one mRNA variant from precursor mRNA transcripts and regulates transcript levels by introducing premature termination codons leading to a nonsense-mediated decay (Kalyna et al., 2012; Liu et al., 2013). It also produces transcript isoforms with altered sequences which have different functionalities or substrate binding specificities (Syed et al., 2012). Roles of AS have been studied in many biological processes including nutrient uptake, flowering time, circadian clock, organ differentiation, biotic and abiotic stress responses in Arabidopsis, oil formation in soybean; and temperature and light responses in Physcomitrella patens (Chang et al., 2014; Eckardt, 2013; Li et al., 2013; Wu et al., 2014). However, little is known whether AS is involved in cell wall modifications. Here, we have detected at least 22 cell wall transcripts undergoing AS (Table 4). They coded for enzymes/proteins implicated in major cell wall polysaccharide modifications including HX, cellulose and (1,3)(1,4)-␤-d-glucans. Alternative splicing may regulate cell wall modifications by a nonsense-mediated decay pathway to control adequate transcript and subsequent enzyme levels. It can also increase cell wall transcript diversities and alter the activities or specificities of cell wall enzymes towards their substrates or counterparts in a protein complex. Structures of cell wall polysaccharides are very complicated. For example, substitutions of polymers can be very different with residues at different positions such as arabinofuanose at either C(O)3 position or at C(O)2 position in HX. The residual composition of backbones also vary enormously in cell wall polymers such as (1,3)(1,4)-␤-d-glucan, which consists of either 1,3 or 1,4 linkages. An increase in cell wall transcript and subsequent enzyme diversities thus would be very useful. However, no direct evidence is available for the AS cell wall enzyme variants binding to different polysaccharide sites. In human keratinocyte, two functional different receptors (KGFR, KGFR-2) were derived from the same gene by AS (Miki et al., 1992). Thus, it is possible that the end products of AS variants of a cell wall gene

138

Q. Zhang et al. / Journal of Plant Physiology 191 (2016) 127–139

react with different polymer substrates or same polymer at different sites. Nevertheless, detection of AS transcripts of cell wall genes are the first step for the study of regulations of cell wall modifications by AS. 5. Conclusion Monosaccharide linkage analysis showed that cell wall modification was very active in germinating barley embryos with increased cellulose and (1,3)(1,4)-␤-d-glucan and reduced HX and arabinan. We have also found that transcriptions of several cell wall genes coincide with the changes in cell wall polymers. Furthermore, we detected at least 22 alternatively spliced cell wall transcripts, which may also play important roles in cell wall modifications. Authors contribution Q.Z.: cell wall preparation, data analysis, write MS; C.L.: conceived and designed the research; X.Z.: RNA preparation for RNA-seq; F.P.; cell wall monosaccharide linkage analysis; G.Z.; support and provided useful comments on the manuscript. Acknowledgement This research was fund by Australian Grains Research and Development Corporation to C. Li. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jplph.2015.12. 007. References An, Y.Q., Lin, L., 2011. Transcriptional regulatory programs underlying barley germination and regulatory functions of Gibberellin and abscisic acid. BMC Plant Biol. 11, 105. Anders, N., Wilkinson, M.D., Lovegrove, A., Freeman, J., Tryfona, T., Pellny, T.K., et al., 2012. Glycosyl transferases in family 61 mediate arabinofuranosyl transfer onto xylan in grasses. PNAS 109, 989–993. Benjamini, Y., Yekutieli, D., 2001. The control of the false discovery rate in multiple testing under dependency. Ann. Stat. 29, 1165–1188. Bewley, J., 1997. Seed germination and dormancy. Plant Cell 9, 1055––1066. Black, D.L., 2003. Mechanisms of alternative pre-messenger RNA splicing. Ann. Rev. Biochem. 72, 291–336. Burset, M., Seledtsov, I.A., Solovyev, V.V., 2000. Analysis of canonical and non-canonical splice sites in mammalian genomes. Nucleic Acids Res. 28, 4364–4375. Burton, R., Shirley, N., King, B., Harvey, A., Fincher, G., 2004. The CesA gene family of barley, quantitative analysis of transcripts reveals two groups of co-expressed genes. Plant Physiol. 134, 224 2–36. Burton, R.A., Jobling, S.A., Harvey, A.J., Shirley, N.J., Mather, D.E., Bacic, A., et al., 2008. The genetics and transcriptional profiles of the cellulose synthase-like HvCslF gene family in barley. Plant Physiol. 146, 1821–1833. Burton, R.A., Ma, G., Baumann, U., Harvey, A.J., Shirley, N.J., Taylor, J., et al., 2010. A customized gene expression microarray reveals that the brittle stem phenotype fs2 of barley is attributable to a retroelement in the HvCesA4 cellulose synthase gene. Plant Physiol. 153, 1716–1728. Burton, R.A., Wilson, S.M., Hrmova, M., Harvey, A.J., Shirley, N.J., Medhurst, A., et al., 2006. Cellulose synthase-like CslF genes mediate the synthesis of cell wall (1,3;1,4)-s-d-glucans. Science 311, 1940–1942. Carpita, N.C., Gibeaut, D.M., 1993. Structural models of primary cell walls in flowering plants: consistency of molecular structure with the physical properties of the walls during growth. Plant J. 3, 1–30. Chang, C.Y., Lin, W.D., Tu, S.L., 2014. Genome-wide analysis of heat-sensitive alternative splicing in Physcomitrella patens. Plant Physiol. 165, 826–840. Chiniquy, D., Sharma, V., Schultink, A., Baidoo, E., Rautengarten, C., Cheng, K., et al., 2012. XAX1 from glycosyltransferase family 61 mediates xylosyltransfer to rice xylan. PNAS 109, 17117–17122. Ciucanu, I., Kerek, F., 1984. A simple and rapid method for the permethylation of carbohydrates. Carbohydr. Res. 131, 209–217. Cosgrove, D.J., 2000. Expansive growth of plant cell walls. Plant Physiol. Biochem. 38, 109–124.

Cosgrove, D.J., 2005. Growth of the plant cell wall. Nat. Rev. Mol. Cell Biol. 6, 850–861. Doblin, M.S., Pettolino, F.A., Wilson, S.M., Campbell, R., Burton, R.A., Fincher, G.B., et al., 2009. A barley cellulose synthase-like CSLH gene mediates (1,3;1,4)-␤-d-glucan synthesis in transgenic Arabidopsis. PNAS 106, 5996–6001. Eckardt, N.A., 2013. The plant cell reviews alternative splicing. Plant Cell 25, 3639. Fincher, G., 2009. Revolutionary times in our understanding of cell wall biosynthesis and remodeling in the grasses. Plant Physiol. 149, 27–37. Fincher, G.B., 2010. Biochemistry, physiology, and genetics of endosperm mobilization in germinated barley grain. In: Ullrich, S. (Ed.), Barley. Wiley-Blackwell, pp. 449–477. Fincher, G.B., Lock, P.A., Morgan, M.M., Lingelbach, K., Wettenhall, R.E.H., Mercer, J.F.B., et al., 1986. Primary structure of the (1 → 3,1 → 4)-␤-d-glucan 4-glucohydrolase from barley aleurone. PNAS 83, 2081–2085. Houston, K., Russell, J., Schreiber, M., Halpin, C., Oakey, H., Washington, J., et al., 2014. A genome wide association scan for (1,3;1,4)-beta-glucan content in the grain of contemporary 2-row spring and winter barleys. BMC Genomics 15, 907. Hrmova, M., Farkas, V., Harvey, A.J., Lahnstein, J., Wischmann, B., Kaewthai, N., et al., 2009. Substrate specificity and catalytic mechanism of a xyloglucan xyloglucosyl transferase HvXET6 from barley (Hordeum vulgare L.). Febs J. 276, 437–456. Hrmova, M., Farkas, V., Lahnstein, J., Fincher, G.B., 2007. A Barley xyloglucan xyloglucosyl transferase covalently links xyloglucan, cellulosic substrates, and (1,3;1,4)-beta-d-glucans. J. Biol. Chem. 282, 12951–12962. Huang, S., Zhang, J., Li, R., Zhang, W., He, Z., Lam, T.-W., et al., 2011. SOAPsplice: genome-wide detection of splice junctions from RNA-Seq data. Front. Genet. 2, 46. Kalyna, M., Simpson, C.G., Syed, N.H., Lewandowska, D., Marquez, Y., Kusenda, B., et al., 2012. Alternative splicing and nonsense-mediated decay modulate expression of important regulatory genes in Arabidopsis. Nucleic Acids Res. 40, 2454–2469. Li, W., Lin, W.-D., Ray, P., Lan, P., Schmidt, W., 2013. Genome-wide detection of condition-sensitive alternative splicing in Arabidopsis roots. Plant Physiol. 162, 1750–1763. Liu, J., Sun, N., Liu, M., Liu, J., Du, B., Wang, X., et al., 2013. An autoregulatory loop controlling Arabidopsis HsfA2 expression: role of heat shock-induced alternative splicing. Plant Physiol. 162, 512–521. Mayer, K., Waugh, R., Langridge, P., Close, T.J., Wise, R.P., Graner, A., et al., 2012. A physical, genetic and functional sequence assembly of the barley genome. Nature 491, 711–716. Miki, T., Bottaro, D.P., Fleming, T.P., Smith, C.L., Wilson, H.B., Chan, A.M.L., et al., 1992. Determination of ligand-binding specificity by alternative splicing: Two distinct growth factor receptors encoded by a single gene. PNAS 89, 246–250. Mortimer, J.C., Miles, G.P., Brown, D.M., Zhang, Z., Segura, M.P., Weimar, T., et al., 2010. Absence of branches from xylan in Arabidopsis gux mutants reveals potential for simplification of lignocellulosic biomass. PNAS 107, 17409–17414. Pease, J., Sooknanan, R., 2012. A rapid, directional RNA-seq library preparation workflow for Illumina sequencing. Nat. Methods i-ii. Pettolino, F.A., Walsh, C., Fincher, G.B., Bacic, A., 2012. Determining the polysaccharide composition of plant cell walls. Nat. Protoc. 7, 1590–1607. Reddy, A.S.N., Marquez, Y., Kalyna, M., Barta, A., 2013. Complexity of the alternative splicing landscape in plants. Plant Cell 25, 3657–3683. Scheller, H.V., Ulvskov, P., 2010. Hemicelluloses. Ann. Rev. Plant Biol. 61, 263–289. Staiger, D., Brown, J.W., 2013. Alternative splicing at the intersection of biological timing, development, and stress responses. Plant Cell 25, 3640–3656. Stamm, S., Ben-Ari, S., Rafalska, I., Tang, Y., Zhang, Z., Toiber, D., et al., 2005. Function of alternative splicing. Gene 344, 1–20. Syed, N.H., Kalyna, M., Marquez, Y., Barta, A., Brown, J.W.S., 2012. Alternative splicing in plants coming of age. Trends Plant Sci. 17, 616–623. Taylor-Teeples, M., Lin, L., de Lucas, M., Turco, G., Toal, T.W., Gaudinier, A., et al., 2015. An Arabidopsis gene regulatory network for secondary cell wall synthesis. Nature 517, 571–575. Trapnell, C., Pachter, L., Salzberg, S.L., 2009. TopHat: discovering splice junctions with RNA-Seq. Bioinformatics 25, 1105–1111. Wu, H.-P., Su, Y.-s., Chen, H.-C., Chen, Y.-R., Wu, C.-C., Lin, W.-D., et al., 2014. Genome-wide analysis of light-regulated alternative splicing mediated by photoreceptors in Physcomitrella patens. Genome Biol. 15, R10. Zeng, W., Jiang, N., Nadella, R., Killen, T.L., Nadella, V., Faik, A., 2010. A glucurono(arabino) xylan synthase complex from wheat contains members of the GT43, GT47, and GT75 families and functions cooperatively. Plant Physiol. 154, 78–97. Zhang, Q., Hrmova, M., Shirley Neil, J., Lahnstein, J., Fincher Geoffrey, B., 2006. Gene expression patterns and catalytic properties of UDP-d-glucose 4-epimerases from barley (Hordeum vulgare L.). Biochem. J. 394, 115–124. Zhang, Q., Pettolino, F.A., Dhugga, K.S., Rafalski, J.A., Tingey, S., Taylor, J., et al., 2011. Cell wall modifications in maize pulvini in response to gravitational stress. Plant Physiol. 156, 2155–2171. Zhang, Q., Shirley, N., Lahnstein, J., Fincher, G.B., 2005. Characterization and expression patterns of UDP-d-glucuronate decarboxylase genes in barley. Plant Physiol. 138, 131–141. Zhang, Q., Shirley, N.J., Burton, R.A., Lahnstein, J., Hrmova, M., Fincher, G.B., 2010. The genetics, transcriptional profiles, and catalytic properties of UDP-alpha-d-xylose 4-epimerases from barley. Plant Physiol. 153, 555–568.

Q. Zhang et al. / Journal of Plant Physiology 191 (2016) 127–139 Zhong, R., Lee, C., Ye, Z.H., 2010. Evolutionary conservation of the transcriptional network regulating secondary cell wall biosynthesis. Trends Plant Sci. 15, 625–632.

Zhong, R., Lee, C., Zhou, J., McCarthy, R., Ye, Z., 2008. A battery of transcription factors involved in the regulation of secondary cell wall biosynthesis in arabidopsis. Plant Cell 20, 2763––2782.

139