Disregarding a poorly conserved insertion in CRM1 , all three modules are almost exactly the same size at 85-90bp long, further supporting the existence of a higher-order structure. Although the identity of this structure cannot be determined from the present data, its functional significance appears to be reflected in CRM1, where the ATG repeats/TGAT cores sequence appears to be reevolving in the ~36bp inserted sequence that displaced the original conserved block . Research in the Drosophila model suggests that the organizing proteins may in fact be PcG and/or TRX proteins, which are thought to be recruited to regulatory modules by a platform of multiple DNA binding proteins. In the Drosophila example however, the “platform” was spread over several hundred base pairs. In plants, a more comparable example might be the PcG binding site demonstrated in LEAFY COTYLEDONS2, where PcG proteins recognized an RLE motif that contained several cis-motifs in a region 50bp long. Interestingly, analysis of FERTILIZATION INDEPENDENT target sites found in A. thaliana found that they were enriched in four cis-motifs, and at least one of each can be identified in the three CRM’s revealed by the present study. The possibility of chromatin regulation also immediately suggests a plausible mechanism by which the 3’ enhancer region might repress CLV3 transcription.
The three CRM’s may serve as a nucleation site for a chromatin silencing mechanism,blueberry plants in pots allowing the silenced chromatin to spread in both directions until it blocks the 5’ promoter of AtCLV3, and presumably the promoter of the neighboring gene At2g27240 as well. However, this model is only weakly supported by the plant literature, as only as a single tenuous chain of evidence supports such an interaction: The TAAT core-motifs in CRM1 are bound by WUS protein, which in turn recruits TOPLESS, SAP18, and ultimately the histone deacetylase HDA19/SIN3-LIKE. This evidence is at least consistent with the repressive portion of WUS transcriptional activities. Although the biochemical details regarding how WUS activates transcription are not yet known, another example from the Drosophila model suggests that such bifunctional activity might be an emergent property of chromatin regulation. Transcription factors that recruit PcG proteins to transcriptional start sites were found to prefer H3K4me3 chromatin modifications. If interpreted correctly, this suggests that transcriptionally active promoters directly recruit their own repressor complex. When a similar model is extrapolated to plants, it is tempting to speculate that the reverse situation might also true: WUS as a repressive transcription factor, may recruit TRX proteins to silenced chromatin, thus activating CLV3 expression. The spread of chromatin silencing is also known to involve insulator motifs that limit the spread of such silencing, but the asymmetric structure of the three CRM’s makes it tempting to speculate that they have polarized activity.
Interestingly, such directional specificity has been observed in fission yeast centromeres, where strand specific repression depended on which Sin3 homolog was used to assemble a histone deacetylase complex. However, no comparable examples are known from plants. Another possible mechanism by which the 3’ enhancer region might affect AtCLV3 transcription is through chromatin looping. This typically involves 8-70kb stretches of DNA [98], all of which are considerably larger than the 1.5kb that separates the AtCLV3 promoter from the three cisregulatory modules. Considering that this small region only supports 5-8 nucleosomes, such short-distance looping might be difficult to achieve before transcriptional activation due to stearic interference. The possibility of looping with a distant enhancer element is also unlikely, as the previously identified 1.5kb +1.2 kb regulatory regions were sufficient to reproduce the AtCLV3 expression pattern. The presence of a potential miR414 site in the coding sequence of CLV3 is intriguing, as it may also offer another level of control. If this microRNA were to be expressed in the RM, its presence would be sufficient to explain the weak expression of AtCLV3 in L2-L4 tissues. This interpretation is consistent with the finding that miR414 is up regulated by cytokinin responses, and strong cytokinin responses are known to occur in the Rib Meristem. However, the putative target site in the 3rd exon is poorly conserved among the five orthologs , and others have suggested that the miR414 gene product itself does not fold properly. Still, it may be premature to dismiss miR414 as a pseudo-gene, as several additional target sites were also found in a naturally occurring transposon , just past the CLV3 regulatory region.
Between the three CRM’s identified in this study, it is possible that they can recruit up to 20 different transcription factors simultaneously. Currently, only WUS proteins have clearly been demonstrated to be part of this group, though a few other candidates can be inferred based on known protein interactions. The recognition that SAP18 binds to the EAR-domain of ERF3 for example, clearly suggests that WUS can interact with SAP18 through its own EAR-like domain, in addition to the previously established WUS-TPL interaction. Although TPL did not bind to HDA19, the observation that TOPLESS RELATED1 and HDA19 coimmuno precipitated suggests that they are at least part of the same protein complex. Thus it would be interesting to identify the proteins that co-immuno precipitate with WUS, as these may include the adjacent transcription factors and the higher order protein complexes. Among transcription factors, one likely candidate might be HAIRY APICAL MERISTEM1 At2g45160, a GRAS domain transcription factor. Originally identified in Petunia hybrida, the GRAS domain HAM1 is known to cooperate with the WUS ortholog TERMINATOR, and was later shown to physically interact with WUS in A. thaliana. This pattern is consistent with the structure of the cis-regulatory modules, particularly if HAM1 should bind to one of the cis-motifs on either side of the conserved +970 WUS binding site. It is also possible that STM might be another co-factor, as both WUS and STM were required to ectopically express AtCLV3 in leaf tissue. The WUS-CLV3 feedback loop has long been predicted to be an essential part of meristem structure within A. thaliana, yet evidence from the present study suggests that CLV3 orthologs are rather poorly conserved outside of the Brassicaceae. The lack of conservation may be related to the size of the CLE gene family, where current evidence suggests that most plant species have twenty or more paralogs. Many of these are co-expressed in the same tissues, and at least some are functionally interchangeable. However, it is also difficult to reconcile the WUS-CLV3 feedback loop with the number of evolutionary clades in each gene family, which would be expected to closely correspond if they represent a conserved feedback loop. Instead, the WOX gene family is organized into 3 recognizable clades, whereas CLE genes are divided into 13 distinct groups.
Their functions are also diametrically opposed, as WOX genes tend to be expressed in or near stem cells, while CLE genes are typically expressed in tissues that display terminal differentiation, such as trichomes, vasculature, stamens, the placenta, and abscission zones. If WUS is an activator of CLV3, it is also difficult to explain why CLV3 expression occurs as much as 24 hours after the appearance of WUS, a phenomenon that has been repeatedly observed plant embryos, and callus tissue studies. Although WUS and CLV3 do have reciprocal phenotypes in mutant backgrounds, and when ectopically expressed, it is surprising that the importance of the hypothesized feedback loop has not left a stronger evolutionary imprint. Instead, there are hints that the two genes may actually operate in different, but related pathways. One such pathways appears to involve an auxin-CLE connection, which is supported by the similarity of auxin responsive tissues and CLE gene family expression patterns in vasculature tissue, leaf tips, guard cells, and trichomes. This is consistent with the proposed CLV3 regulation by an auxin response element,draining plant pots and is futher supported by the synergistic interaction between auxin and exogenous CLE oligopeptides found in developing Zinnia elegans tracheids. Another pathway appears to involve a WUS-cytokinin connection, as WUS has been found to directly regulate cytokinin signaling by repressing A-type ARRs, and potentially has a role in activating cytokinin biosynthesis in A. thaliana and rice . In turn, these two mechanisms might be linked by the mutually exclusive pattern of auxin and cytokinin responses, which seem to be involved in pattern formation in different parts of the plant. Together these observations suggest that WUS and CLV3 might simply respond to the patterns produced by these hormones, providing sharper boundaries between zones and imparting tissue-specific cell identities. In the present study, the fortuitous finding that several significant CLV3 regulatory regions lie entirely within a naturally occurring transposon immediately suggests a novel hypothesis that could unite many different observations. A transposition event that introduced the cis-regulatory modules to AtCLV3 could easily explain the difficulty of identifying CLV3 orthologs outside of the Brassicaceae, as it implies that it occurred independently in other lineages, where similar transpositions may have involved other CLE paralogs. Repression of the transposon via siRNA pathways might also trigger chromatin silencing, leading to the repression of nearby genes, while replicative transposition might explain why the 13-bp TAATnnWnnTGAT motif seems to be widespread in the A. thaliana genome. On a more macroscopic level, the sudden introduction of the cis-regulatory modules might also immediately reduce the size of the SAM, as ectopic activation of CLV3 in the CZ would partially stimulate terminal differentiation by the CLE pathway, and thus indirectly suppress WUS expression. Such a mechanism would produce smaller plants overall, which is consistent with the size of A. thaliana and related species.In order to maintain the dynamic structure of the SAM, a feedback loop between WUSCHEL and CLAVATA3 has been proposed to be an integral part of meristem maintenance. Research over the past decade has successfully clarified many aspects of this model by identifying some of the intermediate steps between CLV3 transcription, perception and signal transduction pathways, though it is not yet known how this controls WUS transcriptional repression.
In contrast, studies of WUS regulated genes have identified several hundred candidates, and have shown that WUS binding to CLV3 regulatory sequences is necessary for CLV3 expression. However, there is an increasing amount of evidence to suggest that this feedback loop is at least partially influenced by hormone signaling pathways. One of the most striking examples of this occurs in rice, where cytokinin biosynthesis mutants produce a flower phenotype that is almost identical to the wus-1 mutant phenotype in A. thaliana. A more direct route of cross-talk was found through microarray experiments, which found that WUS repressed ARABIDOPSIS RESPONSE REGULATOR7 and ARR15,both of which are negative regulators of the cytokinin response pathway. This interaction is fully consistent with the strong pattern of cytokinin responses that occurs in the RM, and suggests that this pattern might be a result of WUS repression of a repressor, leading to activation. In addition, exogenous cytokinin treatments can increase WUS transcript levels, and WUS transcripts are increased when cytokinin catabolism is reduced in ckx3/ckx5 double mutant. Similar positive correlations have also been found in callus tissues and in microarray studies. More directly, WUS has evenbeen found to activate ARR1 transcription, a positive regulator of cytokinin responses, which in turn might explain why both WUS and ARR15 are simultaneously up regulated in SAM regeneration studies. However, this interpretation is somewhat inconsistent with the expression pattern of ARR7 and ARR15, which have been found to strongly overlap with the RM in numerous studies. How this is possible in a tissue that also expresses their direct negative regulator indicates that this system is not well understood. Auxins are involved in the WUS-CLV3 feedback loop, as this hormone has repeatedly been found to reduce WUS transcript levels. There also appears to be a tight correlation between the auxin transporter PIN1 and WUS induction during somatic embyogenesis, while mutation alleles of the auxin-sensitive POPCORN gene are known to disrupt WUS expression patterns. This relationship is perhaps most strongly supported by studies in root meristems, where the closely related WOX5 gene is known to participate in a complex feedback loop involving auxin biosynthesis with YUCCA6, auxin signal transduction with IAA17, auxin efflux with PIN1, and auxin influx with LAX3 carriers. In SAM tissues, the close juxtaposition of WUS and the YUC4 biosynthesis domain in the overlying CZ is also at least reminiscent of the activation of YUC1 by pWOX5:WOX-GR in the root meristemss.