SND2, a NAC transcription factor gene, regulates genes involved in secondary cell wall development in Arabidopsis fibres and increases fibre cell area in Eucalyptus

Abstract

Background: NAC domain transcription factors initiate secondary cell wall biosynthesis in Arabidopsis fibres and vessels by activating numerous transcriptional regulators and biosynthetic genes. NAC family member SND2 is an indirect target of a principal regulator of fibre secondary cell wall formation, SND1. A previous study showed that overexpression of SND2 produced a fibre cell-specific increase in secondary cell wall thickness in Arabidopsis stems, and that the protein was able to transactivate the cellulose synthase8 (CesA8) promoter. However, the full repertoire of genes regulated by SND2 is unknown, and the effect of its overexpression on cell wall chemistry remains unexplored.

Results: We overexpressed SND2 in Arabidopsis and analyzed homozygous lines with regards to stem chemistry, biomass and fibre secondary cell wall thickness. A line showing upregulation of CesA8 was selected for transcriptome-wide gene expression profiling. We found evidence for upregulation of biosynthetic genes associated with cellulose, xylan, mannan and lignin polymerization in this line, in agreement with significant co-expression of these genes with native SND2 transcripts according to public microarray repositories. Only minor alterations in cell wall chemistry were detected. Transcription factor MYB103, in addition to SND1, was upregulated in SND2-overexpressing plants, and we detected upregulation of genes encoding components of a signal transduction machinery recently proposed to initiate secondary cell wall formation. Several homozygous T4 and hemizygous T1 transgenic lines with pronounced SND2 overexpression levels revealed a negative impact on fibre wall deposition, which may be indirectly attributable to excessive overexpression rather than co-suppression. Conversely, overexpression of SND2 in Eucalyptus stems led to increased fibre cross-sectional cell area.

Conclusions: This study supports a function for SND2 in the regulation of cellulose and hemicellulose biosynthetic genes in addition of those involved in lignin polymerization and signalling. SND2 seems to occupy a subordinate but central tier in the secondary cell wall transcriptional network. Our results reveal phenotypic differences in the effect of SND2 overexpression between woody and herbaceous stems and emphasize the importance of expression thresholds in transcription factor studies.

Figure 1

Absolute transcript abundance of SND2∩Ko genes represented on ATH1 22k arrays in Arabidopsis tissues and organs. Genevestigator V3 [39] was used for microarray data mining, and the anatomical cluster analysis tool was used to visualize and cluster the genes according to their tissue-specific expression patterns. Tissues/organs are staggered hierarchically, and the number of arrays on which the data are based is indicated in parentheses. Absolute transcript values are expressed as a percentage of their expression potential (E.P.), where E.P. is the mean of the top 1% of hybridization signals for a given probe set across all arrays. Cluster (a), highlighted in red, is comprised of 31 genes, including SND2 (*), which displayed preferential expression in tissues and organs where SND2 is expressed. Cluster (b) encompasses of 13 genes which displayed preferential expression in inflorescence stems and nodes, rosette stems, and in some cases the stamen, seedling hypocotyl and/or vasculature (stele) of roots.

Figure 2

RT-qPCR analysis of selected genes differentially expressed in inflorescence stems of eight-week-old SND2-OV(A) and wild type plants. SND2-OV(A) plants were grown alongside the wild type in three biological replicate pairs, with primary stems from six plants pooled per sample. SND2-OV(A) transcript levels were normalized to the wild type in each replicate (assigned a value of 1, for each gene), hence error bars indicate the standard error of the deviation from wild type across biological replicates. Significance was evaluated by a one-tailed paired t-test, in accordance with the expected direction of response for each gene; *P < 0.05.

Figure 3

(i) SCW thickness in IFs of eight-week-old wild type and T4 homozygous SND2-OV lines A, B and C. Measurements are based on scanning electron micrographs. Error bars indicate the standard error of the mean of three biological replicates (21-42 fibres were measured per line). *Significantly different from wild type according to homoscedastic two-tailed Student's t-test (P < 0.02). Transmission micrographs of representative IF regions of wild type and SND2-OV line C stems are shown in (ii) and (iii) respectively (scale bars = 20 μm).

Figure 4

Effect of SND2 overexpression on IF wall thickness in T1 generation stems. (A) Mean SCW thickness in IFs of eight-week-old wild type and T1 generation SND2-OV stems. Representative light microscopy images are shown in Additional file 1, Figure S5. Error bars indicate the standard error of the mean of eight wild type and seven T1 plants (26-48 fibres were measured per plant). *Significantly different from wild type based on homoscedastic two-tailed Student's t-test (P < 0.02). (B) Corresponding transcript abundance of total SND2 transcript in lower stems of six wild type and six SND2-OV T1 plants used for SCW measurements, as measured by RT-qPCR. The primer pair quantifies endogenous and transgenic SND2 transcript. Total SND2 transcript is ~435-fold relative to the wild type, represented here on a log₁₀ scale. Calibrated Normalized Relative Quantity (CNRQ) values were obtained by normalization against three control genes. Error bars indicate the standard error of the mean of six plants.

Figure 5

Proposed model of SND2-mediated SCW regulation in IFs. Solid lines indicate known direct protein-DNA interactions. Dashed lines indicate direct or indirect protein-DNA interactions. Master regulator SND1 is activated by a signal transduction pathway proposed by Oikawa et al. [56] (a). SND1 directly activates transcription of MYB103 and SND3 (b), and indirectly activates SND2 through an unknown intermediate (c; [21]). SND2 activates cellulose-synthesizing CesAs, either directly (d) or through the activation of MYB103 (e), which is known to activate SCW cellulose gene, CesA8 [21]. SND2 regulates hemicellulosic genes (f; Table 1), independently to a similar role played by direct SND1 targets MYB46, MYB83 or C3H14 [76-78]. SND2 plays a role in lignification through activation of lignin polymerization genes LAC4 and LAC17 (g; Table 1), but it does not regulate monolignol biosynthetic genes as is the case for MYB58, MYB63 and MYB85 (h) [21,60]. SND2 activates transcription of GPI-anchored FLA11/FLA12, CTL2 and other components of the signal transduction pathway (i), which leads to upregulation of SND1 (a).