FIBexDB: a new online transcriptome platform to analyze development of plant cellulosic fibers

Abstract

Plant fibers constitute a major part of renewable biomass and are a valuable resource for various industries, composite material chemistry, and bioconversion processes. Plant fibers are specialized long cells with thick cell walls that are a convenient model for investigating fundamental issues such as individual cell biogenesis, cell wall formation, and cell specialization. Plant fibers usually accumulate secondary cell walls, which contain lignin, cellulose, and hemicelluloses like xylan, and are laid over thin primary cell walls. Some types of fibers like those found in the phloem of fiber crops and bending parts of hardwood tree species deposit cellulose-enriched tertiary cell walls onto the secondary cell walls (Gorshkova et al., 2018). The tertiary cell wall, also called gelatinous layer (G-layer), plays an important mechanical role; in the fibers of tension wood it serves to pull upward an inclined trunk or branch (Gardiner et al., 2014); in annual fiber crops, fibers with tertiary cell walls help to keep long and narrow stems in a vertical position. The tertiary cell wall does not contain lignin or xylan – an important advantage for plant bioconversion because these polymers hamper the saccharification efficiency of cellulose in lignocellulose. The biogenesis of specialized fibers in trees and annual fiber crops has been studied separately, despite potentially common underlying mechanisms. Thus, a comprehensive comparison of fibers from different sources is required to identify the basic common mechanisms, if present, and to determine possible differences. To tackle this issue, we chose two model plants, flax (Linum usitatissimum L.) and poplar (Populus spp.), because : (1) they are widely used for studying fiber biogenesis, and (2) even though both plants form tertiary cell walls in their fibers, the origins and mechanisms of induction are different (Gorshkova et al., 2018); (3) the poplar and flax genomes were fully sequenced (Tuskan et al., 2006; Wang et al., 2012). Primary phloem fibers in flax form tertiary cell walls constitutively during normal plant development, while xylary fibers in poplar deposit tertiary cell walls only when induced by environmental factors. Each model system has its advantages: flax fiber bundles can be easily isolated, giving the unique possibility to analyze the distinct single cell type at the certain developmental stage. Some transcriptomic studies have been conducted in fiber-containing samples of other fiber crops (Guerriero et al., 2017; Xie et al., 2020); however, datasets for purified fibers are not available for any of these species. In poplar tension wood, the development of tertiary cell walls occurs in fibers as a part of joint reactions of several cell types. Analogically, tertiary cell walls are also induced in flax xylem fibers located at the pulling stem side after plant inclination (Ibragimova et al., 2017), providing a system very similar to tension wood and permitting both interspecies and intraspecies comparisons of tertiary cell wall deposition and regulation. The emergence of NGS (next generation sequencing) technology boosted transcriptomic studies of various biological subjects, including tertiary cell wall formation, leading to the accumulation of a vast array of big data sets, most of which have been analyzed individually and have not been properly compared. To utilize these data more efficiently, a unified open database needs to be created. There are some databases containing information about gene expression in poplar: AspWood (Sundell et al., 2017) and PopGenIE (Sjödin et al., 2009). However, they do not include gene expression data during tension wood formation. By contrast, there are no available transcriptomic databases for fiber crops including flax at all. We therefore combined and normalized transcriptomic data for flax and poplar and created a new database, ‘FIBexDB’ (fiber expression database; https://ssl.cres-t.org/fibex/), storing our published and newly obtained data and extracted data from publicly available open resources. FIBexDB can be used to identify key common participants in the fibers of both plants during tertiary cell wall formation and to analyze the peculiarities inherent to each species. FIBexDB contains two connected but independent databases, one for each species. There are three entrances, namely (1) flax, (2) poplar, and (3) combined top pages (Fig. 1a). For flax, we collected RNA-sequencing (RNA-Seq) data for 61 samples from nine different projects, comprising transcriptomes for seven tissue types (shoot apical meristem, phloem fibers at different stages of development, core parenchyma, xylem tissue, hypocotyls, roots, and leaves), fibers from different flax cultivars and subspecies, as well as parts of mutant stems (reduced fibers) (Supporting Information Fig. S1; Tables S1, S2). For poplar, we collected RNA-Seq data for 88 samples from seven projects, comprising transcriptomes of roots, leaves, normal wood, tension wood of wild-type and transgenic poplars, and samples after treatments related to tension wood formation (Fig. S2; Tables S3, S4). To bridge the genes from the two species, information regarding homologous relationships between flax, poplar and Arabidopsis is stored. There are two major ways to use FIBexDB: (1) browsing expression patterns of single or multiple genes and finding coexpressed genes (gene search; Fig. 1b); and (2) finding upregulated/downregulated genes in particular situations/tissues (differentially expressed gene (DEG) finder; Fig. 1c). Gene expression can be searched by gene identifier (ID) like ‘Lus10028848’ for flax or ‘Potri.002G080700’ for poplar or by gene symbol. This gene search can also accept Arabidopsis gene symbols or locus identifiers, like ‘BGAL12’ or ‘AT4G26140’, and shows corresponding homologous genes in flax and/or poplar (Fig. 1b). The page for each flax and poplar gene (individual gene view) shows fundamental information such as coding sequence, gene symbol if present and homologous Arabidopsis and poplar or flax genes, as well as paralogous genes in each species (Fig. 1b), which are helpful to examine whether the homologous genes show similar expression patterns. The expression pattern in each tissue or condition is shown as a line graph (Fig. 1b). In addition, coexpressed genes calculated from the dataset employed in this database are also presented (Fig. 1b). The relationships between these genes are described as a coexpression network by Cytoscape (Shannon et al., 2003) using coexpression data and homology data for amino acid sequences to connect each gene (Fig. 1b). In this network view, genes with less extent of coexpression are omitted to make the number of genes in the network less than 50. Each gene in the network is labeled by its symbol or the gene symbol of the corresponding Arabidopsis gene if the flax/poplar gene does not have a symbol. If the corresponding Arabidopsis gene encodes a transcription factor (TF), it is shown as a rectangle in the network and the name of the TF family is indicated in the list of coexpressed genes. For further clarity, genes in the network are shown in several different colors representing subnetworks classified by the igraph software (Csardi & Nepusz, 2006). For example, the LusBGAL12 gene (Lus10028848) encoding β-galactosidase, which plays a significant role in tertiary cell wall development in flax fibers (Roach et al., 2011), shows specific expression in phloem fibers. However, in poplar, the closest gene Potri.006G144500 does not show high expression in tension wood. Instead, another β-galactosidase gene (Potri.002G080700, PtrBGAL16) shows specific upregulation in samples enriched with tension wood. Both Lus10028848 and Potri.002G080700 are coexpressed with genes encoding RG-I rhamnosyltransferase from the GT106 family and rhamnogalacturonan lyase family proteins involved in RG-I synthesis and modification; this indicates LusBGAL12 and PtrBGAL16 have similar functions in fibers of different origin. The sets of coexpressed TFs for LusBGAL12 and PtrBGAL16 (maximum 300 genes for each) include 19 and 9 genes, respectively, but they do not share common Arabidopsis closest genes (counterparts). The FIBexDB allows users to perform this kind of analysis easily. FIBexDB can be also used to list genes upregulated/downregulated in a particular tissue or condition via DEG finder (Fig. 1c). The DEGs in 87 and 153 comparisons and their combinations were calculated by DEseq2 package (Love et al., 2014) of R statistical software and can be extracted from flax and poplar, respectively, by simply clicking graphical icons in schematic pictures of flax and poplar (Fig. 1c). The list of extracted DEGs has information for the closest Arabidopsis gene. A unique feature of the DEG finder in FIBexDB is that users can set multiple search conditions simultaneously, even from both flax and poplar. In this case, Arabidopsis genes whose counterparts in flax and poplar meet each search condition are extracted and shown with the counterparts in flax and poplar. In this function, ‘counterparts’ of a particular Arabidopsis gene means flax and poplar genes that best-hit to the particular Arabidopsis gene in a Blast search. This unique feature allows users to find common important genes for tertiary cell wall formation in both species. Furthermore, the extracted genes can be classified by hierarchical or k-means clustering (de Hoon et al., 2004) by a few clicks (Fig. 1c). FIBexDB is largely aimed at analyzing flax and poplar gene expression profiles in tissues and conditions where tertiary cell walls are massively produced; this focus is not envisaged in any other database. With the collected set of samples and the system developed for data analysis, FIBexDB can provide new insights into the mechanisms of complex and important processes, such as cellulose formation, RG-I biosynthesis, and transcriptional regulatory network during tertiary cell wall formation and allows us to compare the development of distinct cell wall types namely primary, secondary, and tertiary cell walls. As tertiary cell walls develop in the fibers of many plant species under various physiological situations, the database can be further developed by adding relevant datasets to identify the important common molecular players as well as ones in specific samples. We welcome any collaboration in the plant fiber field aimed at improving and updating FIBexDB with new data for species containing fibers that form cellulose-enriched cell walls. The authors thank Prof. M. Deyholos (University of British Columbia, Canada) who provided seeds of the rdf mutant and corresponding background cultivar that were used to analyze transcriptomes in a specific sample set from an unpublished study. The authors thank T. Yanagida, K. Hosono, Y. Naraki (Space-Time Inc., Sapporo, Japan) for the design and user interface of the database. The authors thank Robbie Lewis, MSc, from Edanz Group (https://en-author-services.edanz.com/ac) for editing a draft of this manuscript. This work was supported by grants from the Russian Science Foundation (grant no. 19-14-00361; RNA-Seq of flax samples from gravistimulated plants, grant no. 20-44-07005, RNA-Seq data analysis) and Strategic International Collaborative Research project (grant no. JPJ0088379) promoted by the Ministry of Agriculture, Forestry and Fisheries, Tokyo, Japan. The authors also acknowledge financial support from the government assignment for the FRC Kazan Scientific Center of RAS (N Mokshina, OG, TG data normalization and discussion). No conflict of interest is declared. N Mitsuda developed the entire database system and HO and SS provided ideas for the user interface. OG, HT, and N Mokshina assisted with data analysis for data normalization. N Mokshina, N Mitsuda and TG designed and supervised the study. N Mokshina, N Mitsuda, and TG wrote the manuscript. All authors edited and/or made comments on the manuscript. Fig. S1 Schematic diagram of flax sample collection for RNA-Seq analysis. Information about flax samples is given in Tables S1 and S2. Fig. S2 Schematic diagram of poplar sample collection for RNA-Seq based on published data. Information about poplar samples is given in Tables S3 and S4 Table S1 List of experiments for flax transcriptomic data. Table S2 List of samples for flax transcriptomic data. Table S3 List of experiments for poplar transcriptomic data. Table S4 List of samples for poplar transcriptomic data. Please note: Wiley Blackwell are not responsible for the content or functionality of any Supporting Information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Topics & Concepts