Dynamic Deposition Of Histone Variant H3.3 Accompanies ... - PLOS

Dynamic H3.3 Replacement during Developmental Transition

We investigated the dynamics of H3.3 enrichment during the major developmental transition in vegetative plant life that leads to leaf formation. Leaf development is initiated from primordia that continuously arise from the shoot apical meristem (SAM). The SAM and the primordia comprise dividing cells [44]. Once a primordium enlarges through cell division, leaf patterning takes place while cells still divide. Subsequent cell differentiation coincides largely with the arrest of cell division. Thus, we compared H3 variant enrichment in meristem and leaf primordia (dividing tissue) and mature leaves (non-dividing tissue). Using data from cyclebase.org [45] and the transcriptomes obtained from each sample, we verified that dividing tissues expressed cell cycle regulated genes, including the five genes encoding H3.1 variants at levels higher than non-dividing tissues (Table 1 and Table S2). In animals, incorporation of H3.1 and H3.3 into chromatin depends on distinct assembly factors. While ASF1A and ASF1B are apparently required for deposition of both H3 types, the CAF-1 complex participates in H3.1 incorporation, while H3.3 incorporation depends on HIRA and DAXX [8], [13], [18], [46]. Except for DAXX, homologues of the H3 chaperones have been identified in the Arabidopsis genome (Table 1). Amongst these homologues, only the expression of the H3.1-specific CAF1 homolog FAS2 was strongly dependent on the cell cycle (Table 1). Together, the expression profiles of the H3 variants and their chaperones suggest that in Arabidopsis, as is the case in animals, H3.1 incorporation occurs primarily in dividing cells while H3.3 incorporation is largely independent of the cell cycle.

Download:

• PPTPowerPoint slide
• PNGlarger image
• TIFForiginal image

Table 1. Expression of histone H3 and potential histone H3 chaperone genes in dividing and non-dividing tissue.

https://doi.org/10.1371/journal.pgen.1002658.t001

To investigate H3.3 and H3.1 dynamics during the developmental transition from primordia to differentiated leaves, we selected two subsets of genes, according to their higher expression levels (at least five-fold) in either dividing or non-dividing tissue. Having gene sets that were preferentially expressed in either of the two tissue types, we could examine the changes in H3 variant levels that accompanied repression (Figure 3A) and induction (Figure 3B) of transcription during the developmental transition from dividing tissues (plain lines) to non-dividing tissues (dashed lines). Transcriptional repression was accompanied by a strong decrease of H3.3 levels at the 3′ end (Figure 3A). Conversely, activation of gene transcription at the developmental transition was reflected in an increase of H3.3 signal at the 3′ end (Figure 3B). H3.1 levels on the other hand were not affected at genes undergoing repression (Figure 3A) or activation (Figure 3B). Similarly, different groups of genes (cell cycle regulated genes, genes expressed in only one tissue, and control genes with similar expression) also supported that H3.3 enrichment changed dynamically according to the expression modulation (Figure S7). Moreover, when considering all the genes, we observed a positive correlation between expression change and the change in H3.3 enrichment that is modest, but highly significant (Spearman rank correlation of 0.28, p-value<1e-275) (Figure 3C). This was not the case for H3.1, H3 and IgG (Figure 3C).

Download:

• PPTPowerPoint slide
• PNGlarger image
• TIFForiginal image

Figure 3. Dynamics of H3.3 and H3.1 enrichment during development.

(A–B) Average profile of H3.3 (green) and H3.1 (orange) enrichment over genes that are five fold up-regulated in dividing tissue (plain line) compared to non-dividing tissue (dashed line) (194 genes) (A) and vice versa (88 genes) (B). The average H3.3 enrichment varies during development and follows expression changes (A–B). In contrast, the average H3.1 enrichment does not vary significantly during the developmental transition (A–B). (C) Scatterplots of the modulation of H3.3 (green), H3.1 (orange), H3 (blue) and IgG (dashed grey) enrichment rank versus gene expression rank in dividing compared to non-dividing tissue. A sliding window of 500 genes was applied on both the gene expression difference and the differential enrichment. Only the modulation of H3.3 shows a positive correlation with changes of levels of expression.

https://doi.org/10.1371/journal.pgen.1002658.g003

We conclude that the repression of gene expression during leaf differentiation is linked with a decrease in the H3.3 level, but not H3.1 level, suggesting that H3.3 may contribute to developmental transitions. Differentiation also requires the induction of gene expression, which correlates with gain of H3.3 enrichment at the 3′ end of some genes. H3.1 enrichment on the other hand, is not significantly affected by developmental transitions and appears to be a relatively stable chromatin component. This property would support a role of H3.1 in propagation of epigenetic patterns of histone modification through division, in agreement with the preference of H3.1 over H3.3 enrichment at heterochromatic regions, which need to be maintained in a transcriptional silent state.