The School of Molecular and Cellular Biology at the University of Illinois at Urbana-Champaign

Background and significance
During the last two decades, the cell nucleus has emerged as a complex organelle highly organized into many distinct functional domains. A true challenge for the 21rst century is to understand how these various sub-nuclear domains orchestrate a variety of major functions such as the maintenance of the genome, the distribution of the genetic information, and the assembly of the translational machinery. My long term goal is to understand the cellular mechanisms that govern pre-mRNA processing.

In cultured somatic cells, efforts to monitor sites of transcription are hindered by the dense grouping of somatic nuclear structures and the low resolution of light microscopy, which is inadequate to resolve fine chromatin structures for a detailed analysis of RNA transcriptional events. To overcome these limitations, I employ the amphibian oocyte nucleus, which contains chromosomes and organelles that are an order of magnitude larger than their somatic counterparts and, thus, are readily observable with transmitted light (figure 1)

Figure 1. (A) Narrow region of a nuclear spread showing one of the 18 chromosomes present in the nucleus of a Xenopus laevis oocyte. Several organelles are also readily visible: nucleoli (brightly labeled with DAPI), a Cajal body (arrow) and several IGCs (small rounded structures). (B) A particularly extended loop is presented. It is visible by phase contrast because of the high density of its surrounding nascent RNP fibrils. Scale bars are 5 �m.

Recently, I made the first visual observations ever of active RNA polymerase II (RNAPII) transcriptional units by differential interference contrast (DIC) in a living nucleus (Patel et al., in preparation). Together with our newly developed splicing assay (Patel et al., J. Cell Biol. 2007, 178: 937-949), this methodology breakthrough places my laboratory in a unique position to investigate the major components involved in RNA transcription and processing in vivo. We have developed the tools to study the steady-state dynamics of these factors in association with nascent transcripts, with a level of detail not possible in any other system.

In the nucleus of a Xenopus laevis oocyte, chromatin is organized within 18 lampbrush chromosomes (LBCs) and ~2000 non-chromosomal nucleoli. LBCs are best described as highly extended diplotene bivalent chromosomes, and their characteristic resemblance to a test-tube brush comes from the fact that each homologue consists of a heterochromatin axis from which are projected numerous pairs of lateral loops (figure 1). These loops correspond to active transcriptional sites by RNAPII and consist of euchromatin fibers surrounded by nascent RNP fibrils. The major goal of my laboratory is to determine the mechanisms that control the dynamic formation and activities of these loops in order to understand the co-transcriptional regulatory events of pre-mRNA maturation.

Structural and molecular definition of the chromosomal loops. In the current view, synthesis and processing of pre-mRNAs are thought to be coupled on the transcriptional unit. One of the major pre-mRNA maturation events is the removal of introns by the spliceosome, which is a large complex composed of the five major small nuclear ribonucleoprotein particles (snRNPs) and over 100 other splicing factors. To study the intranuclear trafficking of snRNPs in vivo, we exploited the fact that microinjected fluorescent snRNAs rapidly assemble into functional snRNPs. While other groups have been using the same strategy previously, my laboratory is the first one to report their association with the LBC loops. Interestingly, these data translated into a new and unique assay to investigate with an unprecedented spatial resolution the mechanisms that regulate the association of snRNPs with active transcriptional units. We found that, surprisingly, the spliceosomal assembly does not direct the recruitment of snRNPs to nascent transcripts. This conclusion is based on 1) the fact that the U1 snRNP and the U4/U6-U5 tri snRNPs associate with the chromosomal loops in absence of U2 snRNA, and 2) that non-functional forms of U1 and U2 snRNAs still associate with the active transcriptional units. These results carry important physiological implications as they suggest some level of preassembly of the splicing machinery on the nascent transcripts, possibly as a staging event for the efficient formation of an active spliceosome.

Importantly, our new assay can also determine the inherent properties of a spliceosomal snRNP that contribute to its association with nascent transcripts in vivo. Surprisingly, we showed that U1 and U2 snRNPs need not be functional for their association with the chromosomal loops. In the case of the U1 snRNP, we already characterized the first stem loop region (SL1) of U1 snRNA as necessary and sufficient for its association with pre-mRNAs. Interestingly, SL1 appears to also regulate the kinetic exchanges of the U1 snRNP between the nucleoplasm and two other discrete nuclear domains implicated in RNA transcription and processing, the Cajal bodies (CBs) and the interchromatin granule clusters (IGCs). We are currently characterizing the regulatory elements for the other major snRNPs.

While the association of splicing factors with chromosomal loops was already known, we wanted to demonstrate the functionality of these interactions in order to establish a unique cytological splicing assay. To achieve this goal, we demonstrated that the protein Y14, a core component of the Exon Junction Complex (EJC), associates with chromosomal loops in a splicing-dependent way (figure 2). Since EJCs are deposited by the spliceosome to mark exon-exon junctions, it demonstrates that pre-mRNA splicing occurs directly on the chromosomal loops. Using the recruitment of Y14 to nascent transcripts as a reliable indicator of spliceosomal activity represents a powerful new in vivo splicing assay that distinguishes itself from most other splicing systems in several ways: 1) co-transcriptional splicing of nascent transcripts can be studied in situ, in relation with the complex intranuclear trafficking of snRNPs; 2) it does not rely on a pre-mRNA splicing reporter nor on a specific nuclear extract; 3) it can survey all the endogenous active RNAPII transcriptional units simultaneously in one cell.

The association of Y14 with chromosomal loops also shows that EJCs engage nascent transcripts before their release from sites of transcription. Following their deposition on transcripts, EJCs are involved in recruiting components of the mRNP export machinery. We recently found that the export factor Aly/REF is also present on chromosomal loops and its association with pre-mRNAs, like that of Y14, depends on splicing. We are currently investigating other export factors such as the protein Tap. Collectively, our data strongly support a model where the synthesis of a spliced, export-competent mRNP occurs directly on transcriptional units.

Figure 2.Phase contrast and corresponding fluorescent micrographs of nuclear spreads from oocytes expressing a HA-tagged Y14. In absence of splicing activity, HA-Y14 does not associate with chromosomal loops. However, it is still recruited to CBs (arrow), and weakly to nucleoli. In this experiment splicing was inhibited by depletion of the U2 snRNA. (Adapted from Patel et al. 2007, accepted in J. Cell Biol.)

Concurrently, we have started to analyze the loop distribution of several other processing factors, such as hnRNPs A1, A2, Q, L and G, as well as nuclear factor 7 (Beenders et al., 2007, Mol. Cell. Biol., 27: 2615-2624). While several of these proteins were initially described as �mRNA chaperones�, they were recently shown to have a direct role in the regulation of splicing. Interestingly, although we found them associated with most loops, there are subtle differences in their patterns. In addition, we recently showed that the recruitment of hnRNP G to nascent RNP fibrils does not require its RNA Recognition Motif, but rather a discrete domain in its carboxyl terminus. The same may be true for other hnRNPs and this result highlights the need to determine in vivo the recruitment mechanisms of these factors in order to better understand how RNA processing is regulated co-transcriptionally.

Finally, it is critical to consider the overall structure of LBCs in order to fully understand the dynamics of the chromosomal loops. We have concentrated our initial efforts on factors involved in chromatin condensation and we showed that some of the condensins such as XCAP-D2, but not all of them, are components of LBCs (Beenders et al. 2003, Chrom. Res., 11:549-564). Since loops correspond to chromatin region where the sister chromatids are not associated, we are also currently analyzing the chromosomal association dynamics of several subunits of the cohesin complex, which is involved in sister chromatid cohesion. Lastly, we demonstrated that the methyl-cytosine binding protein 2 (MeCP2) distributes within LBCs in a unique axial pattern that suggests a role in loop formation and/or maintenance. The interest in MeCP2 lies within its ability to recruit histone deacetylases to methylated DNA and/or its possible involvement in chromatin loop formation in somatic systems to regulate gene expression. One conclusion of our work is that there is a constant exchange of these factors between chromatin and the surrounding nucleoplasm, highlighting the dynamic behavior of LBCs components. This property was further confirmed for the cohesin subunit MCD1 by FRAP analysis in oil-isolated nuclei (see below). In addition, we demonstrated that the methyl-C binding domain (MBD) of MeCP2 is necessary and sufficient for its recruitment to chromatin in vivo. Over-expression of the MBD, which should behave as a dominant negative form by preventing the recruitment of HDACs to LBC axes, will be carried out in the oocyte to test possible effects on the loop physiology.

In summary, using our newly established assays, we obtained results that reveal new level of regulation in the pre-mRNA maturation pathway in vivo. In particular we obtained original sets of data on the mechanisms that regulate the interaction of snRNPs with nascent transcripts and other organelles such as CBs and IGCs. These results suggest a critical staging of the various RNA processing factors directly onto the nascent RNP particles for splicing. In addition, we demonstrated a co-transcriptional recruitment of Y14 and Aly/REF, which underscores a complex interplay between transcription, processing, and export on the chromosomal loops. Finally, we began to investigate the mechanisms that regulate at the chromatin level the formation and/or maintenance of the loops.

Future directions
1) How is the recruitment of snRNPs to nascent transcripts and other organelles regulated?
Initially we plan to determine the essential characteristics of each snRNP that contribute to its association with chromosomal loops. We have already established for U1 snRNA that SL1 is critical for its association with nascent transcripts. We are now investigating which SL1-interacting factors are implicated. Interestingly, we also showed that SL1 influences the intranuclear trafficking of U1 snRNP, in particular its association with CBs and IGCs. Similar strategies and analyses will be applied to the other snRNPs. In turn, since we can monitor whether splicing occurs on chromosomal loops, it will be possible to test if the trafficking of snRNPs through CBs and other organelles is important for spliceosomal assembly on nascent transcripts.

2) What are the dynamics of nascent transcripts-associated processing factors?
To answer that question, we needed to observe transcriptional units with their associated nuclear domains in a live nucleus. Other groups demonstrated that an oocyte nucleus, isolated in mineral oil maintains all its activities for several hours, allowing observation of nuclear organelles by light microscopy. However, LBCs remained elusive. Over the last three years, we optimized the oil-isolation procedure in order to obtain the very first live images of LBCs from X. laevis oocytes. We showed that these chromosomes are morphologically identical to those observed in nuclear spread preparations and, importantly, their lateral loops are readily observable by DIC (Figure 3), which represent the very first live visualization of RNAPII transcriptional units. One powerful application will be the study of the steady-sate dynamics of loops components.

Figure 3. These two light microscopy images offer the first live visualization of several active RNAPII transcriptional units in Xenopus oocyte nuclei. Arrows indicate particularly extended chromosomal loops. (Adapted from Patel et al., 2007 in preparation).

In particular we plan to determine the exchange kinetics of the majors snRNPs. One strategy will be to perform FRAP analyses on fluorescently labeled snRNAs and/or GFP-tagged snRNP specific proteins. This approach could be complicated, however, by the fact that high levels of nucleoplasmic fluorescent snRNAs might compromise the visualization of chromosomal loops. Thus, a second strategy will be to use caged fluorescent snRNAs instead. Similarly, snRNP specific proteins will be expressed in fusion with a photo-activable GFP. Kinetic measurements should provide invaluable information on whether snRNPs are recruited to nascent transcripts individually or within a pre-assembled particle.

Similar strategies will also be used to test whether processing factors are stably bound with nascent RNP fibrils until their release from the chromosomal loops or if there is a constant dynamic exchange between the transcripts and the nucleoplasm while transcription proceeds.

In the long term, we plan to also investigate not only the dynamics of transcription factors and other non-splicing factors, but also of chromatin factors such as histones and their modifying enzymes.