|
|
Figure 1. The multiple aspects of gene expression.Gene expression can be regulated at different levels. Two families of proteins control most of the steps that link the genome to the proteome. Transcription factors are responsible for the first step and regulate RNA transcription with all the events after this point (RNA splicing, transport, stability, localization and translation) being regulated by RNA binding proteins.
The importance of post-transcriptional gene expression and the RNA cis-regulatory code.
Higher eukaryotes require coordinated regulation of multiple genes to accomplish complex phenotypic functions. This necessitates the regulated production of genes for generalized housekeeping functions as well as those for specialized functions. Most studies of gene expression have traditionally focused on transcription and are based on the idea that regulatory networks result from transcription factors interacting with promoters and enhancer elements in a combinatorial manner. In eukaryotic organisms, transcription is uncoupled from translation and physically separated by the nuclear membrane, thus eukaryotic gene regulatory networks require an additional level of coordination that links transcriptional and post-transcriptional processes. While there is no doubt that transcription contributes significantly to eukaryotic gene expression, there is often a poor correlation between mRNA levels and protein production. The discordance between mRNA and protein levels occurs in part because eukaryotic mRNAs are subject to post-transcriptional processing and regulation, which is predominantly mediated by RNA-binding proteins (RBPs) (Figure 1). Therefore, RBPs, especially those involved with mRNA interactions, are to post-transcriptional gene regulation as transcription factors are to transcription. RBPs provide a means to couple transcription to translation in eukaryotic systems (Figure 1) and are involved in all of the essential steps of post-transcriptional gene regulation including splicing, nuclear export, stability and ultimately translation. Cell-specific RBPs are involved in a variety of processes that are critical for appropriate protein expression (e.g., alternative splicing of messenger RNAs and translational control). RBP perturbation has been implicated in numerous clinical disorders and structural/functional studies of these proteins has begun to yield important insights into how they help to shape the protein expression programs unique to many cellular processes.
Not surprisingly, the number of RBPs has increased dramatically commensurate with the evolution of prokaryotes to eukaryotes in parallel with the development of the nuclear membrane and has continued to expand with higher eukaryotic evolution. Messenger RNA has traditionally been viewed as a passive molecule as it moves from transcription to translation. However, it is now clear that RBPs play a major role in regulating multiple mRNAs in order to orchestrate the concerted production of complex gene expression networks. On this basis, post-transcriptional and transcriptional gene expression networks appear to be very analogous. Although promoters residing on dispersed genes participate in the coordinated production of gene products, post-transcriptional regulation must also be maintained for coordinated protein production. In addition to eukaryotic genes being coordinately regulated by transcription factors that bind combinatorially to multiple promoter elements, we and others have demonstrated the coordinated post-transcriptional regulation of multiple mRNAs by messenger RNA-binding proteins (mRBPs). This demonstrates that, like transcription factors networking promoter elements, distinct subsets of mRNAs are regulated together by RBPs at the post-transcriptional level. It is therefore likely that gene expression in eukaryotes depends as much on the actions of mRBPs as it does on the actions of transcription factors.
|
Cis-mediated regulation though RNA binding proteins.
The number of identified RBPs has exploded in recent years. Their biological importance now is unambiguous, and much work focuses on how they act, contact one another, and are controlled. Much of this regulation is mediated via the non-coding regions of expressed mRNA, which includes the 5' and 3'-untranslated regions (UTRs) as well as the intronic regions.
As the complexity of translational regulators and their interactions has increased, principles have begun to emerge. A critical next key step will be identification of the full complement of RBP complexes associated with particular target mRNAs under defined circumstances. At this juncture, three important features of regulation by RBPs are clear: they can repress or activate, their control is combinatorial, and they form regulatory networks that are dynamic and controlled.
|
Figure 2. Two modes of regulation.Initiation factors and the 40S subunit are indicated. Two modes of regulation by 3'UTR-bound repressors commonly act in one of two modes, indicated by arrow. Some cause deadenylation; others interfere with the interaction between eIF4E and eIF4G. Activators relieve the repression by restoring the eIR4E-eIF4G interaction or increasing poly(A) length.
Repression and activation.
In most cases to date, the 3'UTR binding protein decreases expression, either by repressing translation or accelerating mRNA decay. These events almost invariably are correlated with removal of the poly(A) tail. Thus the action of 3'UTR-bound protein repressors closely mirrors that of microRNAs, which enhance deadenylation, translational repression, and decay. Two modes of repression have been identified. They both interfere with the effective circularization of the mRNA by the tripartite eIF4E-eIF4G-PAB complex, required for efficient translation (Figure 2). These RBPs can recruit a deadenylation complex and removes the poly(A) tail (Goldstrohm, 2006 ; Lykke-Andersen, 2005; Semotok, 2005) that elicits a repressive effect and RBPs can recruit protein cofactors that compete with the cap-binding protein, thereby competing with eIF4G, and hence blocking translation (Mendez, 2001; Richter, 2005 ; Thompson, 2006 ).
RBPs can also increase expression. Typically, activation consists of a relief of repression, and extension of the poly(A) tail. For example, phosphorylation of CPEB, an RRM protein, switches regulation from repression to activation by changing the protein contacts it forms while bound to the 3'UTR of its targets (Mendez, 2001; Richter, 2005); phosphorylation of mammalian hnRNP K relieves repression of 15-lipoxygenase mRNA by preventing the protein from binding the 3'UTR in partnership with hnRNP E proteins (Ostareck-Lederer, 2002). Thus regulation RBPs is dynamic.
Many biological processes are well-known for being regulated at the post-transcriptional level: embryogenesis, stem cell proliferation, spermatogenesis, sex determination, neurogenesis, erythropoiesisis. In most cases, the genes involved in the listed processes have regulatory elements placed on UTRs that serve to modulate their levels of expression. There are several examples connecting regulatory elements on UTRs to health related issues. For instance, Iron regulatory proteins (IRP) control several mRNAs (ferritin, transferring, mitochondrial aconitase, etc) that have a regulatory element named iron responsive element (IRE). IRPs bind to them in situations of iron deprivation and inhibit their translation. Mutations that affect the IRE can lead to human disease such as hereditary hyperferritinemia-cataract. Another good example is the amyloid-b precursor protein (APP) implicated in Alzheimer's and Down syndrome. Translation of APP mRNA is up-regulated by interleukin-1 through 5' UTR sequences (reviewed in Penalva, 2006).
|
Figure 3.Profiles of mRNAs associated with protein complexes identified using cDNA array analysis. RNA from precipitated mRNA-Protein complexes or total cell lysate was used to make probes for Mouse Arrays containing 600 double-spotted cDNA segments (CLONTECH). The RBPs ELAV/HuR, and PABP were targeted and compared to pre-bleed sera and total cellular RNA. An example of the specificity among mRNA profiles is demonstrated by the relative abundance of beta-actin mRNA as compared to ribosomal protein S29 mRNA (arrows a and b, respectively).
RNA-binding proteins and ribonomics.
There are an estimated 2-3 thousand RBPs encoded in the human genome making this family of proteins as abundant as the transcription factor family. For only a small percentage of RBPs, have the cis-regulatory element/binding site (CREBS for short) been well characterized. Most of these elements are located in the 5’ or 3’ untranslated regions (UTRs) of the mRNA. Our research focused on targeting RBPs to better understand post-transcriptional gene-expression networks. Previously, we developed genome-scale, antibody based technologies for immunoprecipitating endogenous RBP-mRNA complexes from cellular extracts and identifying the associated messages using microarray technology . This new approach to post-transcriptional functional genomics (termed Ribonomic profiling) greatly facilitated the high-throughput; quantitative identification of multiple mRNA targets of many RBPs and facilitated the analysis of the structural and functional relationship of the encoded proteins from these co-regulated mRNAs . Ribonomic profiling provided great insight into the infrastructure of coordinated eukaryotic post-transcriptional gene expression, which otherwise would not have been observed using total mRNA expression, which results from the integral involvement of RBPs in the mechanisms responsible for compartmentalizing and regulating these RNA subsets.
My lab is recognized as a leader in the genomic-scale, unbiased identification of RBP target mRNAs {Tenenbaum, 2000; Tenenbaum, 2002; Tenenbaum, 2003; Brown, 2001; Eystathioy, 2002; Keene, 2002 ; Intine, 2003 ; Penalva, 2004) having helped pioneer the technology and the field of Ribonomics. Figure 3 shows an example of array data (low-density) generated by profiling two RBPs: HuR, an ELAV RBP family member involved in the regulation of early response gene mRNAs, and the poly-A binding protein (PABP), which binds to the poly-A tail present in the vast majority of eukaryotic mRNAs. RNA profiles are clearly distinct and demonstrate the specificity for subsets of mRNAs obtained using ribonomic profiling with RBPs as compared to global or total RNA.
|
Figure 4.Comparison of total RNA with mRNA profiles associated with HuR before and after heat-shock treatment. Total RNA was compared to the RNA associated with the RBP HuR following heat-shock treatment array analysis. HeLa cells were heat-shocked at 450C for 60 min. and then returned to 370C. Post heat-shock samples were collected at 30, 60 and 120 min after return to 370C. Images of CLONTECH Stress Arrays containing 240 genes were analyzed using ATLASIMAGE 2.01 software. Differences in the relative amount of HuR associated mRNAs compared to the relative abundance in the total RNA was considered significant if 4-fold or greater. Red-bars; increased relative abundance, Blue-bars; decreased relative abundance and Green-bars; less-than a 4-fold change.
Figure 5.Comparison of the mRNA profiles associated with Hu proteins before and after treatment with RA. Panel A; Differences in the relative amount of HuR associated mRNAs compared to the relative abundance in the total RNA was considered significant if 4-fold or greater. Red-bars; increased relative abundance, Blue-bars; decreased relative abundance and Green-bars; less-than a 4-fold change. Panel B; Examples of HuR associated mRNAs showing differences in relative abundance as compared to total RNA before and after RA treatment. (****) = detected only in RA-treated cells.
Additionally, we have found that the repertoire of RNP associated RNA is dynamic, and can change under various cellular perturbations such as heat-shock (Figure 4) and Retinoic Acid (RA) treatment (Figure 5). These changes are frequently not detectable in the total RNA and likely result from mRNA associations being regulated in concert with the phenotypic changes occurring in response to heat shock (Gallouzi, 2000; Brennan, 2000; Gallouzi, 2001) or RA treatment (Tenenbaum, 2000 ). Specific transcripts of interest can also be studied and their RBP association verified individually.
For example, in Figure 6B IGF-2, Integrin-beta, Cyclin D2 and HSP84 abundance was examined by realigning and enlargement of their respective spots and by independent RT-PCR validation. The disparity between changes in the array profiles of total RNA and RBP associated RNA is readily visible and likely results from changes in the cellular environment in response to a biological inducer like RA. The insights afforded by targeting RBPs to study the coordinated regulation of functionally related subsets of mRNAs directly results from the integral involvement of RBPs in the mechanisms responsible for compartmentalizing and regulating these RNA subsets. Data generated using ribonomic profiling has led to several basic observations about the mRNP infrastructure; 1) mRNA binding proteins are associated with unique subpopulations of messages 2) the composition of these mRNA subsets can vary with changing cellular conditions, and 3) the same mRNA species can be found in multiple mRNP complexes. Based on these findings, we proposed a model of post-transcriptional gene expression in which mRNA-binding proteins regulate mRNAs as dynamic subpopulations.
|
Figure 6.Combinatorial organization of mRNAs in by RBPs. (Upper panel A). Representation of a polycistronic transcript from a prokaryotic operon containing open reading frames coding for four genes (ORFs 1-4). (Lower panel B). Four monocistronic mRNAs that contain different combinations of USER codes are represented (colored boxes a-d). mRNAs containing a specific USER code can be bound by a specific mRNA binding protein (colored shapes A-D). This model illustrates the potential for genetic information to be shuffled combinatorially as mRNA subsets. In this example, CREBS (also called USER codes) are placed in the 5' and 3' untranslated regions (UTRs) but could also reside in the coding regions. An important feature of the model is that a given RBP may regulate its own mRNA, thus affecting the regulation of its larger subset of mRNAs. The Venn diagrams depict example subpopulations of mRNAs that can be organized in various combinations by targeting the various CREBS.
The Post-Transcriptional Operon Model.
Data generated using ribonomic profiling has led to several basic observations about the mRNP infrastructure; 1) mRNA binding proteins are associated with unique subpopulations of messages 2) the composition of these mRNA subsets can vary with changing cellular conditions, and 3) the same mRNA species can be found in multiple mRNP complexes. Based on these findings, we proposed a model of post-transcriptional gene expression in which mRNA-binding proteins regulate mRNAs as dynamic subpopulations (Figure 6.). This model predicts that functionally related genes are regulated as subsets by specific mRNA-binding proteins that recognize sequence elements codified by cis-sequences in common among these otherwise independent mRNA species. Identifying these cis-regulatory elements/ binding-sites or CREBS are the focus of this proposal. Our data and others suggest that an mRNP infrastructure exists within mammalian cells consisting of mRNAs that are networked by multi-targeted RNA-binding proteins. Subpopulations of monocistronic mRNAs are coordinately regulated as groups and their protein products participate in shared biological processes or pathways, thereby providing a post-transcriptional analog to the classical polycistronic operon. Further, we proposed that mRNAs are utilized in multiple combinations in order to garner genetic complexity from a limited number of genes . Strong evidence supporting this model has recently been demonstrated . Polycistronic mRNAs present in prokaryotic operons appear to have been replaced in eukaryotes with monocistronic transcripts containing multiple RNA regulatory elements.
This model predicts that functionally related genes are regulated as subsets by specific RBPs that recognize sequence elements codified by cis-sequences in common among these otherwise independent mRNA species.
The comprehensive identification of these mRNA based cis-regulatory elements/Binding sites (CREBS) is the basis of my current NIH/NHGRI funded U01 ENCODE pilot project, which uses RIP-Chip (or ribonomics) methods in to immunoprecipitate endogenously formed RBP-RNA complexes and identify the associated RNA using genomic-scale analysis. Using tiling-array based RIP-Chip assays combined with RNase digestion methods we are foot-printing cis-regulatory elements/RBP-binding sites (CREBS) targeted by RBPs. The objectives of this Pilot project are to combine this method with three additional technologies to comprehensively catalog CREBS present in expressed ENCODE RNA. This is being accomplished by (1) characterizing the genome-wide associations of 6 representative RBPs with expressed mRNA using traditional RIP-Chip profiling and whole-genome expression arrays; (2) identifying specific CREBS contained in the untranslated regions (UTRs) of ENCODE expressed mRNAs using tiling-array based RIP-Chip foot-printing technology; (3) verifying and further increasing the resolution of identified ENCODE specific CREBS using bioinformatics combined with quantitative Real-Time PCR (qRT-PCR) and (4) biologically validating the function and RBP-binding activity of ENCODE identified CREBS using expression reporter assays. An additional objective of our current pilot project is to develop RIP-Seq using a "deep-sequencing" or 2nd-generation sequencing based platform. This will allow us to perform and compare RIP-Seq and RIP-Chip foot-printing data with respect to quality, quantity and cost.
|
The development of a high-throughput nanoscale immunoassay platform.
Immunoassays, and more specifically microarray-based immunoassays, are gaining in importance as a platform for the high- throughput detection of proteins. This technology has advanced greatly, with the development of new signal generation and detection techniques, surface chemistries and various assay formats. The applicability of current antibody based assays for profiling complex biological samples is still restricted at the moment and if simple strategies are used, such as protein labeling with dyes or haptens, detection is very limited. Some of the known limiting factors affecting immunoassays include low stability of antibody molecules, strong background signal due to inevitable protein adsorption on many kinds of surface supports, insufficient sensitivity of detection and most pronounced, the limitation of the mass-transport constraints on the reaction kinetics of many typical antibody-based assays.
To address these issues, we are developing novel nanotechnology wafer that employs a multiplexed, high-throughput assay that enables the simultaneous survey or capture of multiple proteins/nucleic acids of interest in a single reaction. This technology focuses on the fabrication of multiporous structures from a variety of materials including silicon, aluminum, SU8, etc. that serve as a host matrix to help in the attachment, filtering and analysis of selected chemicals of interest. The ability to create a high surface area in a minimal volume makes this process ideal for biological processing applications.
|
The development of genomic data mining informatics.
We have developed algorithms that use the genome as GPS coordinate system enabling all genetically derived material to be compared. This approach results in the discovery of new drug targets and follows on targets to existing drugs, improves the efficiency of drug screening, and facilitates drug repositioning. Use of these algorithms has led to the discovery of novel drug targets and a mechanism used by viruses that may serve as a platform for anti-viral therapies.
The Tenenbaum lab informatics portfolio is built upon the fundamental concept that the genome is a set of biologically relevant linear coordinates to which all genomics data can be converted. New findings point to a complex network in which genes, along with regulatory elements and other types of DNA sequences that do not code for proteins, interact in overlapping ways not yet fully understood. By taking a comprehensive look at the entire genome, my lab can efficiently analyzes regulatory elements as well as genes and all biological activity that is related to the genome. The first step in our approach is to convert all data to a common format. Then, high throughput analysis is possible along with the creation of highly accurate statistically valid controls. In addition, one algorithm generates hypotheses and elucidates relationships that were not necessarily being tested. Analysis can be performed using a single data sample versus a control, a single data sample against a set of data, or one set of data against another set.
|
microRNA-RBP co-regulation.
Research into small non-coding RNAs and more specifically microRNAs, has brought to light a whole new layer of post-transcriptional gene-expression regulation. Several recent studies have suggested that the biological targets of microRNAs reside in the 3’UTRs of many mRNA and may be influenced by the RNA-binding protein activity. Recent studies and our lab findings suggest that the cis-regulatory code targeted by microRNAs might be the same as that read by mRNA-binding proteins (Bhattacharyya, 2006). We are exploring this approach as an ideal means for bioinformatically identifying RBP binding-sites/regulatory elements. For example, we asked whether any of the known microRNAs would target a well-characterized mRNA regulatory element such as the stem-loop structure, which is present in the 3’-UTR of most higher eukaryotic histone mRNAs. Unlike most mRNAs, histone messages are not poly-adenylated. Instead, this family of mRNAs utilizes a 26 base stem-loop structure in their 3’-UTR called the histone stem-loop (HSL), which is regulated by the histone stem-loop binding protein (SLBP). This interaction may function analogously to that of the poly-A binding protein and its target, the poly-A mRNA tail and facilitates efficient translation. We blasted the human histone mRNA sequences containing the HSL against the characterized human microRNA dataset using the miRanda program. We found a family of microRNAs that specifically would interact with the HSL sequence with binding activity centered exactly on the HSL regulatory element.
Even more remarkable was that the microRNA-mRNA interaction would bind the HSL in a manner that would melt the stem-loop portion making the microRNA-mRNA complex and the SLBP binding site (or HSL regulatory element) mutually exclusive. This finding suggests that microRNAs and RNA-binding proteins may indeed be targeting an overlapping cis-regulatory code and therefore, information on microRNA targeting may prove extremely informative on identifying CREBS.
|
The response of Prostate Cancer cells to the drug Dutasteride
Male growth hormones or Androgens play an important role in development and growth of prostate cancer. Therefore, decreasing the concentration of androgens in the body and inhibiting the androgen-receptor (AR) has been the main focus of prostate cancer research. When detected early, while the cancer is localized, radiation therapy or surgery can be used to destroy or remove the cancer cells. However when cancer has spread to other parts of the body, systemic hormone therapy is used to deprive the cancer cells of the androgens that they need to grow, thereby causing them to die.
In the prostate the main circulating androgen (testosterone) is converted to its more potent form; dihydrotestosterone (DHT). This is achieved by the action of two enzymes called Steroid 5-α-Reductase Type I and Type II (SRD5A1 and SRD5A2). The drug "Dutasteride" inhibits both these enzymes, thereby depriving prostate cancer cells of the androgens. It was originally developed as a treatment for prostate enlargement (BPH). Many clinicians including Dr. Perrotti at St. Peter's Medical Center observed that dutasteride treatment for 6-10 weeks prior to prostate surgery caused a reduction in tumor volume and a decrease in mean serum prostate serum antigen (PSA) levels. They also reported that a combination of dutasteride with other hormone reducing drugs can slow tumor growth and prolong survival.
The purpose of this grant is to study the biological effects of dutasteride on prostate cancer cells, with the ultimate goal of understanding whether dutasteride can be used for treatment of prostate cancer.
We have selected two prostate cancer cell lines LnCaP (which is slow growing and androgen dependent) and PC3 (which is more aggressive and androgen independent) as prototypes for prostate cancer. These two cell lines will be treated with dutasteride and finasteride, RNA from these cells extracted after treatment and gene expression analyzed at the 'global' level by using microarray technology.
|
|
|
