A G(enomic)P(ositioning)S(ystem) for Plant RNAPII Transcription

Post-translational modifications (PTMs) of histone residues shape the landscape of gene expression by modulating the dynamic process of RNAPII transcription. The contribution of particular histone modifications to the definition of distinct RNAPII transcription stages remains poorly characterized in plants. Chromatin Immuno-precipitation combined with next-generation sequencing (ChIP-seq) resolves the genomic distribution of histone modifications. Here, we review histone PTM ChIP-seq data in Arabidopsis thaliana and find support for a Genomic Positioning System (GPS) that guides RNAPII transcription. We review the roles of histone PTM “readers”, “writers” and “erasers”, with a focus on the regulation of gene expression and biological functions in plants. The distinct functions of RNAPII transcription during the plant transcription cycle may in part rely on the characteristic histone PTMs profiles that distinguish transcription stages. Arabidopsis RNA-Binding protein FCA lysine-specific demethylase 1 homolog and tri- but not monomethylation on histone H3 lysine 36 active transcription of genes involved in time regulation and other

thaliana (Col-0). We provide comparable metagene profiles that visualize the interplay between histone PTMs and RNAPII transcription stages ( Figure 1A). These ChIP-seq data reflect the localization of histone PTMs in the whole plant under normal growth condition, regardless of cell-, tissue-or condition-specific histone patterns [16][17][18]. This review covers recent advances in understanding how plant gene expression is underpinned by a histone PTM-based Genomic Positioning System (GPS) that guides RNAPII through transcription stages. RNAPII usually stalls near the 5'-end of genes after initiation, a phenomenon known as promoter proximal RNAPII stalling. RNAPII then enters the productive elongation stage to complete nascent RNA production of the full transcript. When RNAPII passes poly-(A) site (PAS) sequences at 3'-end of genes, RNAPII stalls again to assist transcriptional termination.

Histone PTMs define transcription stages
RNAPII thus performs different functions in transcription stages that are coordinated with different co-transcriptional molecular events (e.g. capping, splicing and poly-adenylation). These considerations raise the question: what molecular system informs RNAPII of the current transcription stage during transcriptional progression? Chromatin Immuno-precipitation followed by next-generation sequencing (ChIP-seq) resolved the genomic distribution profiles of many histone post-translational modifications (PTMs) and variants. Interestingly, the deposition of different histone PTMs or variants is spatially associated with different stages of transcription ( Figure 1A). The profile of histone PTMs may thus be connected to the definition of transcription stages that define distinct RNAPII activities.

Histone PTMs and transcription initiation
Transcription initiation controls the recruitment of RNAPII to promoters, and regulates the polymerase flux into the gene bodies. Transcription initiation relies on the assembly of the preinitiation complex (PIC) including RNAPII and general transcription factors (GTFs) at promoters.
Tri-methylation on histone H3 lysine 4 (H3K4me3) characterizes a well-studied chromatin modification associated with transcription initiation ( Figure 1A). In Arabidopsis, H3K4me3 density across transcription units peaks at 5'-end of genes, and high levels of These results support an indirect molecular connection between transcription initiation and H3K4me3 in plants. Perhaps this connection depends on the genomic context, since the SWI/SNF (Switch/Sucrose Non-Fermentable) chromatin remodeler complex controls the activation and repression of sense gene transcription and anti-sense non-coding transcription through PIC formation correlating with H3K4me3 levels at both gene ends [46]. In summary, H3K4me3 may promote PIC formation and RNAPII initiation in plants, yet the precise molecular mechanisms remain to be elucidated.

Histone PTMs and early transcriptional elongation
We sub-divide transcription elongation by RNAPII into early elongation and productive elongation [47]. Promoters coincide with nucleosome-depleted region (NDR) with resulting low levels of histone PTMs. In contrast, the first (i.e. +1) nucleosomes fall within the early elongation zone during RNAPII transcription. These nucleosomes dominate the genomic distribution of histone PTMs. In metazoans, early elongation refers the stage of RNAPII between transcription initiation and productive elongation linked to well-defined RNAPII promoter proximal pausing sites regulated by pausing factors such as negative elongation factor NELF (Negative Elongation Factor) [48]. Even though NELF is conspicuously absent in plants, RNAPII tends to stall at the position of the first nucleosome in gene bodies [20]. In addition, the distribution of RNAPII in plant promoter proximal regions is wider compared to metazoans, perhaps indicating an extension of the RNAPII early elongation stage in plants compared to metazoans. Di-methylation on histone H3 lysine 4 (H3K4me2) and tri-methylation on histone H3 lysine 36 (H3K36me3) peak slightly downstream of histone PTMs for transcription initiation [37, 49-52], thus could be associated with RNAPII early elongation and RNAPII stalling ( Figure 1A)[20]. However, the mechanistic connections between chromatin during early RNAPII elongation and RNAPII stalling are yet to be firmly established.
It is plausible to imagine a cross talk between transcription initiation and productive elongation that occurs during early elongation to facilitate progression further into the gene. In Arabidopsis, . These data may reflect an orchestrated progression through RNAPII transcription stages from initiation to early elongation. In conclusion, H3K4me2 during early RNAPII elongation might represent a molecular hub that coordinates the transition from transcription initiation to elongation through the interaction with histone acetylation.
During early transcription elongation, H3K36me3 often correlates with H3K4me2 at positions just downstream of H3K4me3 ( Figure 1). Roles of H3K4me3 and H3K36me3 in transcription initiation and elongation characterize these histone PTMs as excellent predictors for gene expression in plants [58]. In Arabidopsis, H3K36me3 acts in concert with other histone PTMs for active transcription (e.g. H3K4me3 and histone acetylation) to promote gene expression [59][60][61][62].
Moreover, H3K36me3 is highly enriched at temperature-regulated alternatively spliced genes, and a reduction of H3K36me3 affects alternative splicing outcomes in Arabidopsis [63]. Likewise, retained introns in the Arabidopsis spliceosome mutant brra2 often exhibit low H3K36me3 profiles [64]. These data link chromatin features during RNAPII transcription to pre-mRNA processing. Arabidopsis MRG (MORF Related Gene) proteins read H3K36me3 as well as H3K4me3 and mediate transcription activation by directing H4ac deposition near 5'-end of target genes [65,66]. The genomic distributions of H3K36ac and H3K36me3 overlap downstream of TSSs, albeit with antagonizing effects even though both are associated with active transcription [25]. Combinatorial effects on gene expression of histone PTMs of the same residue as suggested for H3K36 may increase the resolution to differentiate stages of RNAPII transcription.
In summary, the H3K36me3 peak during early RNAPII elongation is linked to chromatin features ahead of the peak, and to pre-mRNA processing after the peak, supporting a role in bridging RNAPII initiation and elongation.
The Arabidopsis pTEF-b subunit CDKC;2 regulates the global level elongating RNAPII (RNAPII-Ser2 Phosphorylation) transcription [73]. In addition, Arabidopsis SPT5 can be phosphorylated by CDKC;2, interact with PAF1C subunit VIP5 (VERNALIZATION INDEPENDENCE 5) and further influence H3K4me3 deposition on target loci [74]. During the productive transcriptional elongation stage, RNAPII translocates along the DNA template to synthesize a growing nascent RNA chain. In eukaryotes, the activity of elongating RNAPII is modulated by various elongation factors, including histone modifiers and splicing regulators [75]. RNAPII encounters few nucleosome barriers during transcription initiation and early elongation, while many nucleosomes need to be navigated during the productive elongation stage. The chromatin landscape shaped by histone PTMs on these intragenic nucleosomes thus provides the opportunity to regulate RNAPII elongation. In plants, a variety of histone PTMs localize to this stage and display nuanced distribution patterns. Histone PTMs that peaked at early elongation stage (i.e. H3K4me2 and H3K36me3) decline gradually towards 3'-end of genes. Monoubiquitination of histone H2B (H2Bub) and mono-methylation on histone H3 lysine 4 (H3K4me1) prominently cover most of the gene body without a clear peak. Di-methylation on histone H3 lysine 36 (H3K36me2) is gradually enriched towards the 3'-end of genes where it peaks, representing a histone PTM characterizing late stages of productive transcriptional elongation in plants ( Figure 1A). Interestingly, plant DUBm co-purifies with RNAPII subunits, mediator, histone chaperons and RNA processing factors, while HUB1 also co-purifies with transcription elongation factors. These data suggest a strong association of H2Bub biology and productive transcriptional elongation in plants [72,87].
H3K4me1 shows a similar distribution profile over gene bodies to H2Bub, but with a trend to increase towards 3'-gene ends ( Figure 1A). Interestingly, H3K4me1 may negatively correlate with initiation, potentially due to the dynamic conversion to the higher-order methylation states H3K4me2/me3 [37]. In gene bodies, H3K4me1 is enriched at cryptic intragenic TSSs that are repressed by the activity of the histone chaperone FACT complex in Arabidopsis [88]. The repressive effect of H3K4me1 on intragenic initiation appears to be distinct from H3K36 methylation, although SDG8, a H3K36 methyltransferase, has been proposed to read H3K4me1 as well as deposits H3K36me2/me3 [89]. In metazoans, H3K4me1 is classically associated with enhancers, whereas in plants, H3K4me1 is largely associated with RNAPII elongation and antagonizes the repressing effect of di-methylation on histone H3 lysine 9 (H3K9me2) [90].
It is tempting to speculate that the dynamics of H3K4 methylation-state conversion could guide the progression of plant RNAPII transcription.
H3K36me2 represents an additional key histone PTM for productive transcriptional elongation

Genomic information for transcriptional termination
The final stage of the transcription cycle represents transcriptional termination. RNAPII promoter DNA sequence displays elevated elongation (H3K36me2) signatures, and consistently reduced initiation and early elongation-associated histone PTMs H3K4me3 and H3K36me3. The underlying molecular mechanism is consistent with "transcriptional interference", where the act of RNAPII transcription changes the chromatin state at gene promoters to repress functional transcriptional initiation ( Figure 1B) [120]. In Arabidopsis, Genome-wide data support the idea that H3K36me2, when localized to promoters, correlates with negative gene expression.
Perhaps through a similar mechanism to transcription repression associated with enrichment of

Histone-based GPS service guides plant development and environmental responses
The chromatin states are dynamically regulated by the trios of histone "reader-writer-eraser" enzymes. Mutants defective in a particular histone modifying enzyme often profoundly affect gene expression and are associated with defects in plant growth, development and environmental response. We summarize the recent advances in understanding how histone PTM enzymes regulate biological processes in plants (Table 1 and Figure 2  (HISTONE ACETYLTRANSFERASE OF THE GNAT FAMILY 2) [188] and HAG3 [189][190][191][192].
Histone readers also contribute to the regulatory functions of histone PTMs. For example, histone readers and effectors not only recognize particular histone PTMs, but also mediate the downstream regulation and contribute to the crosstalk between different histone PTMs. Trios of "reader-writer-eraser" enzymes collectively regulate gene expression through the dynamics of histone PTMs/variants. In general, loss-of-function mutant of a particular "writer", "eraser" or "reader" enzyme will directly or indirectly affect the local or global histone PTM/variant levels, which may further impairs the RNAPII functions associated with transcription stages, reflected by changes of RNAPII occupancy or expression level of target genes. Even though the chromatin state profiles in many of the mutants listed in Table 1 is incomplete, it may be interesting to interpret the resulting phenotypic defects through the GPS model presented in this review. This view may reveal defects resulting in the mis-specification of transcription stages and associated defects in pre-mRNA processing. Future progress in this area would help to appreciate the biological significance of diverse spatially resolved effects chromatin modifications may have on gene isoform expression.

Conclusion Remarks and Future Directions
The roles of histone PTMs and associated factors in plant RNAPII transcription cycles remain largely uncharacterized. In this review, we summarized the current information on how histone   Table S1).The Yaxis represents the normalized ChIP-seq signal of histone PTMs. The raw values to plot these data are provided in (Table S2). The X-axis indicates the relative position across a gene (grey),   and for further precipitation. DNA that binds to histones is released by reverse cross-linking.
Purified DNA can be used in microarray (ChIP-chip) or library construction and followed by next-