A birth-to-death view of mRNA from the RNA recognition motif perspective

Authors

  • Terri Goss Kinzy,

    1. UMDNJ Robert Wood Johnson Medical School Graduate School of Biomedical Sciences and Rutgers, The State University of New Jersey Joint Program in Molecular Biosciences, NJ
    2. Department of Molecular Genetics, Microbiology and Immunology, UMDNJ Robert Wood Johnson Medical School, NJ
    Search for more papers by this author
  • Lauren A. De Stefano,

    1. UMDNJ Robert Wood Johnson Medical School Graduate School of Biomedical Sciences and Rutgers, The State University of New Jersey Joint Program in Molecular Biosciences, NJ
    Search for more papers by this author
  • Anthony M. Esposito,

    1. UMDNJ Robert Wood Johnson Medical School Graduate School of Biomedical Sciences and Rutgers, The State University of New Jersey Joint Program in Molecular Biosciences, NJ
    Search for more papers by this author
  • Jennifer M. Hurley,

    1. UMDNJ Robert Wood Johnson Medical School Graduate School of Biomedical Sciences and Rutgers, The State University of New Jersey Joint Program in Molecular Biosciences, NJ
    Search for more papers by this author
  • Rohini Roy,

    1. UMDNJ Robert Wood Johnson Medical School Graduate School of Biomedical Sciences and Rutgers, The State University of New Jersey Joint Program in Molecular Biosciences, NJ
    Search for more papers by this author
  • Anibal J. Valentin-Acevedo,

    1. UMDNJ Robert Wood Johnson Medical School Graduate School of Biomedical Sciences and Rutgers, The State University of New Jersey Joint Program in Molecular Biosciences, NJ
    Search for more papers by this author
  • Kai-Hsiung Chang,

    1. UMDNJ Robert Wood Johnson Medical School Graduate School of Biomedical Sciences and Rutgers, The State University of New Jersey Joint Program in Molecular Biosciences, NJ
    Search for more papers by this author
  • Jonathan Davila,

    1. UMDNJ Robert Wood Johnson Medical School Graduate School of Biomedical Sciences and Rutgers, The State University of New Jersey Joint Program in Molecular Biosciences, NJ
    Search for more papers by this author
  • Jennifer M. Defren,

    1. UMDNJ Robert Wood Johnson Medical School Graduate School of Biomedical Sciences and Rutgers, The State University of New Jersey Joint Program in Molecular Biosciences, NJ
    Search for more papers by this author
  • Jesse Donovan,

    1. UMDNJ Robert Wood Johnson Medical School Graduate School of Biomedical Sciences and Rutgers, The State University of New Jersey Joint Program in Molecular Biosciences, NJ
    Search for more papers by this author
  • Patricia Irizarry-Barreto,

    1. UMDNJ Robert Wood Johnson Medical School Graduate School of Biomedical Sciences and Rutgers, The State University of New Jersey Joint Program in Molecular Biosciences, NJ
    Search for more papers by this author
  • Anibal Soto,

    1. UMDNJ Robert Wood Johnson Medical School Graduate School of Biomedical Sciences and Rutgers, The State University of New Jersey Joint Program in Molecular Biosciences, NJ
    Search for more papers by this author
  • Riza M. Ysla,

    1. UMDNJ Robert Wood Johnson Medical School Graduate School of Biomedical Sciences and Rutgers, The State University of New Jersey Joint Program in Molecular Biosciences, NJ
    Search for more papers by this author
  • H. Liesel Copeland,

    1. UMDNJ Robert Wood Johnson Medical School Graduate School of Biomedical Sciences and Rutgers, The State University of New Jersey Joint Program in Molecular Biosciences, NJ
    Search for more papers by this author
  • Paul R. Copeland

    Corresponding author
    1. UMDNJ Robert Wood Johnson Medical School Graduate School of Biomedical Sciences and Rutgers, The State University of New Jersey Joint Program in Molecular Biosciences, NJ
    2. Department of Molecular Genetics, Microbiology and Immunology, UMDNJ Robert Wood Johnson Medical School, NJ
    • UMDNJ Robert Wood Johnson Medical School Graduate School of Biomedical Sciences and Rutgers, The State University of New Jersey Joint Program in Molecular Biosciences, NJ
    Search for more papers by this author

  • This work was supported by NSF MCB 516688, NIH GM068077 and GM077073.

Abstract

RNA binding proteins are a large and varied group of factors that are the driving force behind post-transcriptional gene regulation. By analogy with transcription factors, RNA binding proteins bind to various regions of the mRNAs that they regulate, usually upstream or downstream from the coding region, and modulate one of the five major processes in mRNA metabolism: splicing, polyadenylation, export, translation and decay. The most abundant RNA binding protein domain is called the RNA Recognition Motif (RRM)1. It is probably safe to say that an RRM-containing protein is making some contact with an mRNA throughout its existence. The transcriptional counterpart would likely be the histones, yet the multitude of specific functions that are results of RRM based interactions belies the universality of the motif. This complex and diverse application of a single protein motif was used as the basis to develop an advanced graduate level seminar course in RNA:protein interactions. The course, utilizing a learner-centered empowerment model, was developed to dissect each step in RNA metabolism from the perspective of an RRM containing protein. This provided a framework to discuss the development of specificity for the RRM for each required process.

A wide range of proteins are able to associate with mRNA via highly conserved RNA-binding domains and play a pivotal role in mRNA metabolism and hence, in the post-transcriptional regulation of gene expression. The RNA recognition motif (RRM), alternatively is one of the most abundant eukaryotic protein domains that mediate interaction between mRNAs and proteins. The pFAM database currently lists over 12,000 sequences containing RRM motifs derived from over 400 species[1]. In humans there are ∼500 RRM-containing proteins annotated, suggesting that ∼2% of the genome is dedicated to RRM production. The observation that precursor mRNAs and nuclear mRNAs are almost always found complexed with proteins led to the initial detection of RRMs 20 years ago[2]. Since then, RRM containing proteins have been extensively studied in conjunction with their function in the development of RNA and gene expression[3–5]. The RRM comprises a consensus RNA-binding sequence about 75–85 amino acids long, located at N-termini of proteins, the structure of which is a β1α1β2β3α2β4 fold[6]. RNA-binding protein motifs are functionally conserved between eukaryotes and prokaryotes and often bind other proteins in addition to RNA[7]. Here we review the function of several RRM-containing proteins that are crucial to the process of RNA metabolism from transcription through decay highlighting specific proteins involved at each stage. Figure 1 illustrates the life-cycle of an mRNA making note of the RRM-containing proteins discussed below.

Figure 1.

An illustration of the RRM-containing proteins (italics) that function throughout the lifecycle of an mRNA.

mRNA SPLICING

The spliceosome is an intricate organization of protein/RNA complexes whose primary purpose is to remove introns from pre-mRNA. Accurate splicing of eukaryotic mRNA precursors (pre-mRNAs) requires the combined effort of five specific small nuclear RNAs (snRNAs) and an estimated 300 proteins[8]. The spliceosome recognizes the 5′ and 3′ splice sites, causes these sites to bend around the branch point to bring the edges of the exons together and stimulates cleavage and subsequent ligation of the free ends to form the final exon[9]. The ability of the spliceosome to complete this task depends on RNA binding proteins that interact with pre-mRNA and the five snRNAs that network to form the spliceosome. Although many of these proteins interface through protein–protein interactions, a significant percentage of these proteins bind directly to the pre-mRNA and snRNAs themselves. They do this through multiple RNA binding domains, one of which includes RRM. The characterization of the spliceosome and its component parts has yet to be completed. Though only a few of the identified proteins directly interact with snRNA, many of the RRM-containing proteins are essential for a functional spliceosome. This activity is physiologically essential as dysfunctional spliceosomes are linked to diseases such as dementia and Parkinson's.

The first step of spliceosome formation involves many of the major RNA binding proteins in the splicing process. To recognize the 5′ end of the splice site, the U1 snRNP binds to the GU site directly downstream of the 5′ site[9]. This is facilitated by one of the best-studied RRM-containing proteins, U1A, which contains two RRMs (Fig. 2). The first RRM has been shown to bind to the second hairpin loop of the U1 snRNA[10]. Binding of U1A aids in the directionality U1 snRNP binding to the GU recognition site of the mRNA.

Figure 2.

Alignment of the RRM motifs discussed in this review including the level of conservation, quality of conservation, overall consensus and predicted secondary structure, as indicated. Alignments were made using the MUSCLE algorithm and presented in JalView format.

U1A is one of the proteins that has been used to define the RRM motif. It is highly similar to another protein involved in splicing, U2B. Though the two proteins are 94% similar and 75% identical, they bind to different targets with high specificity[11]. These similarities and differences have helped to explain some of the basics of specificity in RNA binding proteins. This minimal difference between the two proteins can be reduced to about nine amino acids, which when mutated are enough to change the specificity of the protein, suggesting that RRM specificity resides likely at the single amino acid level[12].

Another spliceosomal protein, Prp8, plays many roles in splicing and contains a single RRM. Deletions of the RRM have demonstrated a suppression of cold sensitive defects that are caused by an inability of U4 to unwind from U6. This suggests that the RRM in Prp8 may function to help U4 unwind from U6 and allow U6 to attach to the U2 snRNP and the pre-mRNA[3]. Thus it is clear from the very beginnings of RNA processing and metabolism that the myriad RRM proteins participating in splicing illustrate both the commonality and specificity of this motif.

mRNA POLYADENYLATION

The next step in the maturation of the pre-mRNA is a highly regulated event that includes cleavage and polyadenylation at the 3′-end[13]. Polyadenylation is a complex process that requires the interaction between pre-mRNA and several multi-subunit proteins that contain RRMs. Alternative polyadenylation of the same mRNA can produce distinct transcripts and hence plays an important role in increasing diversity in gene expression. Major developmental and differentiation decisions are determined by alternative processing of the poly(A) site[14].

Occurring during and immediately after transcription, polyadenylation is a multi-step process that requires the initial recognition of a highly conserved core signal (AAUAAA) by cleavage and polyadenylation specificity factor (CPSF). Further downstream is a GU-rich downstream sequence element (DSE) that is recognized by the heterotrimeric cleavage and stimulation factor (CstF) through its 64 kDa subunit CstF-64[15]. CPSF and CstF are both RRM-containing proteins that exhibit cooperative binding on the pre-mRNA. The CPSF/CstF/pre-mRNA polyadenylation complex recruits other proteins including many cleavage factors (CFI and CFII) and polyadenylate polymerase (PAP). Subsequently, the mRNA is cleaved downstream of the poly(A) sequence by an as-yet unidentified component of the cleavage complex possessing endoribonuclease activity. PAP synthesizes the poly(A) tail and poly(A)-binding protein (PABP) binds to and stabilizes it. PABP is yet another RRM-containing protein that is involved in polyadenylation (see below and[16]).

The significance of RRM-containing proteins that comprise the polyadenylation complex becomes clear as the role of CstF in polyadenylation is explored. CstF consists of three subunits named according to their apparent molecular weight: 77, 64 and 50 kDa[17]. Mammalian CstF-64 has an RRM at its N-terminal end that binds to GU-rich DSEs in pre-mRNA molecules. As determined by NMR, the RNA-binding domain of CstF-64 adopts the typical RRM-fold, β1α1β2β3α2β4, (Fig. 2) and the β-sheet was confirmed as the RNA recognition surface of CstF-64. In addition CstF-64 also has a long C-terminal α-helix (helix C) that lies perpendicular to the β-sheet thereby occluding the RNA recognition surface. Although CstF-64 plays an important role by recognizing DSEs, these sequences are not highly conserved. Binding affinity for UU repeats within the GU-rich DSE was observed to be higher and required unfolding of helix C. It is speculated that in order for CstF-64 to bind mRNA, helix C unfolds to expose the high affinity UU dinucleotide binding pocket of CstF-64 and simultaneously triggers an active conformation of the CSPF/CstF/pre-mRNA complex[18].

RRM-containing proteins involved in other cellular processes also participate in the regulation of polyadenylation. U1A, for example, has an important role in spliceosome formation (see above) and participates in an auto-regulatory mechanism controlling its own polyadenylation. The N-terminal RRM of U1A appears to have a higher affinity for RNA than the C-terminal RRM[19]. It has been shown that two molecules of U1A protein, via the N-terminal RRM, recognize a region of the U1A mRNA known as the polyadenylation-inhibitory element RNA (PIE-RNA;[20]). This results in the inhibition of polyadenylation of the pre-mRNA due to inhibition of poly A polymerase. Overall, the striking feature of the RRM proteins that participate in polyadenylation is the similarity in biochemical action, usually the provision of an RNP platform for complex assembly, yielding a wide array of functions.

CYTOPLASMIC POLYADENYLATION

Polyadenylation is not limited to the nucleus. Some mRNAs that remain dormant and require spatial and/or temporal translation, as found during early development and in dendrites, are polyadenylated in the cytoplasm. Two cis-acting factors in the 3′-UTR are required for the addition of a poly(A) tail in the cytoplasm: the cytoplasmic polyadenylation element (CPE; UUUUAU), and the hexanucleotide polyadenylation signal, (AAUAAA;[21]. The CPE binding protein (CPEB), binds the CPE and concomitantly represses translation of CPE-containing mRNAs though an interaction with maskin, a translation initiation factor 4E (eIF-4E) binding protein that blocks recruitment of eIF-4G[22]. Cytoplasmic polyadenylation is activated upon phosphorylation of CPEB, an event that recruits cytoplasmic CPSF[23]. This ultimately leads to poly(A) synthesis on the 3′-UTR by Gld2[24], a poly(A) polymerase. Cytoplasmic polyadenylation is coupled to translation intitiation, reportedly by PABP associating with the poly(A) tail and eIF-4G, which in turn disrupts the maskin/eIF-4E complex[25]. Recently, CPE containing mRNAs in Xenopus oocytes were demonstrated to be polyadenylated in the nucleus and de-adenylated in the cytoplasm in a CPE dependent manner[26]. This study also reported that the poly(A)-specific ribonuclease was found in complex with the cytoplasmic polyadenylation machinery and dissociates upon phosphorylation of CPEB.

CPEB was first identified in Xenopus oocytes as a 62 kDa dual RRM-containing protein (Fig. 2) required for cytoplasmic polyadenylation during oocyte maturation. Hake and colleagues[27] found that CPEB possesses a zinc finger-like motif. Both RRMs as well as the zinc finger-like motif are required for binding CPE-containing mRNAs. Importantly, this work suggested that CPEB and CPEs interact in a 1:1 ratio as some CPE-mRNAs possess more than one CPE.

Interestingly, CPEB was shown to interact with amyloid precursor-like protein 1 (APLP1) in a yeast two-hybrid screen using a mouse brain cDNA library. Phosphorylation of SPEB and subsequent polyadenylation of CPE-containing mRNAs was enhanced in Xenopus oocytes in the presence of APLP1[28]. APLP1 was also found to enhance CPE-dependent, and therefore cytoplasmic polyadenylation dependent, translation of a luciferase reporter mRNA in cultured neurons. Together these data present a model for localized cytoplasmic polyadenylation that has possible implications for CPEB's purported role in synaptic plasticity[29,30]. Additionally, there is evidence that CPEB may be involved in transport or localization of some mRNAs[31,32]. Interestingly, nuclear and cytoplasmic polyadenylation utilize entirely different sets of proteins to achieve the same end result, and the fact that the RRM is utilized as a platform for complex assembly in both cases provides an excellent example of the plasticity of RRM utilization.

RNA EXPORT

In eukaryotic cells the division between the cytoplasm and the nucleus helps ensure that only properly processed RNAs reach the cytoplasmic translational machinery. The RNA nuclear export mechanism provides a quality control step in regulating RNA biogenesis and maturation. Previously identified export pathways require specific RNA-binding proteins for RNA shuttling. Only a limited number of transcripts in the transcriptome bind to these proteins in a sequence dependent manner. If most of the RNA-protein binding occurs in a sequence independent manner, how does the cell determine which export pathway will transport a nascent RNA to the cytoplasm? One proposed parameter specifying export pathways is RNA length. For example, U snRNAs or small RNAs less than 300 nucleotides long are exported via a CRM1 dependent pathway[33]. CRM1 is a member of the importin-beta family, whose members are among the first RNA export factors identified[34].

In contrast, the bulk of cellular mRNAs (over 300 nts) are exported in a CRM1-independent manner via TAP/NXF1 and its binding partner p15[35]. TAP is the human homolog of Mex67p, previously identified to function in nuclear export in yeast. TAP has two nuclear localization sequences and one nuclear export sequence that are vital for export activity[36]. It also has UBA and NTF2-like domains that bind to FG repeats on nucleoporins, both of which are needed for efficient mRNA export[37]. TAP was first identified as a cellular cofactor necessary for the export of constitutive transport element (CTE)-containing RNAs found in several retroviruses. These retroviruses do not encode their own RNA export factor, thus TAP binds to the CTE and makes use of the host nuclear export machinery[38,39]. The CTEs fold into a helical RNA structure that presents two identical RNA loops which interact directly with TAP through a surface loop within its RRM domain and a concave surface of the adjacent leucine-rich repeat domain[40].

Endogenous cellular mRNAs interact with TAP through various adaptor proteins, a significant set of which are the REF/Aly family of RRM-containing proteins. Depletion of the REF family protein Yra1p in yeast or its three known homologs in C. elegans results in accumulation of poly(A)-containing mRNA within the nucleus[41,42]. REF/Aly, along with THO, Tex1p and UAP56/Sub2p are the major components of the transcription-export (TREX) complex[43]. The TREX complex serves to couple the transcriptional processing and the export of mRNA. GST-pull-down experiments indicate that UAP56/Sub2p directly and specifically interacts with REF/Aly and thus mediates the recruitment of REF/Aly to spliced mRNAs[44].

The REF binding proteins contain a central RRM domain and two highly conserved regions at the N and C termini that are unique to this family of proteins[45]. The REF binding proteins are an evolutionarily conserved family, with members present in yeast, Drosophila, C. elegans, Xenopus, plants, and humans. The RRM present in REF/Aly has a β1α1β2β3α2β4 (Fig. 2) fold, which is quite similar to most RRMs[45]. However, this structure is unique because there is a substitution of a highly conserved Phe/Tyr residue, present in most RRMs, to Asp[45]. This substitution is conserved in Aly and its homologues and probably accounts for its unique function. The REF/Aly family proteins interact directly with and help recruit TAP/Mex67p to cellular mRNAs[42]. The REF proteins then likely act as a scaffold between TAP and the RNA in the resulting RNP complex. The UBA and NF2 domains at the C-terminus of TAP bind to the nucleoporins within the nuclear pore complex and the RNA-protein complex exits the nucleus. The export pathway taken by any given transcript must be determined by the mRNA binding proteins associated to the transcript, and the RRM plays a central role in coordinating these events.

mRNA LOCALIZATION

When cells need to rapidly respond to stimuli, the specific localization of an mRNA can modulate its translation in a spatial and temporal manner inside the cell. This mechanism frequently causes an asymmetric distribution of a protein. mRNA localization has been observed to play a crucial role during development in Drosophila and Xenopus as well as in yeast and somatic cells such as neurons and fibroblasts[46–49]. Localization of transcripts is widely believed to be coupled to translational repression, thereby controlling where, and depending upon translation activation signal, when, proteins are expressed.

β-Actin is an abundant protein that is found at the leading edge of cells that are under constant actin polymerization and remodeling[50–52]. The polarized expression of β-actin to the leading edge of cells is accomplished by targeting the β-actin mRNA to a specific location resulting in compartmentalized synthesis. The targeting of β-actin mRNA is mediated by a 54 nucleotide region in the 3′UTR composed of several A/C rich elements (ACACCC), known as zipcode sequences[53]. Analyses of the zipcode element have shown that several regions are conserved among β-actin from various species, but are absent in other actin isoforms, suggesting a specific role in the regulation of β-actin mRNA localization and expression[54]. It has been shown that the localization of β-actin mRNA occurs in an energy dependent manner, yet independent of new protein synthesis. In addition, the β-actin mRNA transport is blocked by the disruption of the actin cytoskeleton[53].

Work by Ross et al[53], isolated and identified a 68 kDa protein that recognizes and binds to the zipcode sequence in β-actin mRNA which forms an RNA-protein complex that promotes the specific localization of the mRNA in chicken embryo fibroblasts (CEFs). The protein, named Zipcode Binding Protein 1 (ZBP1), is composed of two RRMs (Fig. 2) in the N-terminal region and four hnRNP K homology (KH) domains in the C-terminal end. Recently, Farina et al[55] dissected the region of ZBP1 that is important in the regulation and localization of β-actin mRNA in CEFs. Surprisingly, they demonstrated that the two KH domains most proximal to the C-terminus of the protein (and not the two RRMs) are sufficient for binding the zipcode sequence in β-actin mRNA, for granule formation, and association with the cytoskeleton. It was also demonstrated that the N-terminal region is required for the proper localization of ZBP1-containing granules[55]. The binding of ZBP1 occurs co-transcriptionally with its target mRNA, forming an RNA-protein complex that is subsequently exported from the nucleus as a ribonucleoprotein particle[29]. When the complex reaches the cytoplasm, ZBP1 represses the translation of β-actin mRNA by preventing the formation of the 80S ribosomal complex, but not the 48S complex, in a manner similar to the repression of 15-lipoxygenase translation by hnRNP-K[56]. The ZBP1/β-actin complex is transported along the cytoskeleton to the leading lamellopodia, where ZBP1 is phosphorylated by a non-receptor tyrosine kinase (Src kinase) This phosphorylation causes a decrease in ZBP1 binding affinity and release of β-actin mRNA thus allowing the ribosome to assemble and synthesize β-actin protein[57]. By this mechanism, therefore, the β-actin mRNA is protected by ZBP1 from its birth in the nucleus, preventing its premature translation or degradation in the cytoplasm, until it reaches its final destination where the protein is finally translated and then assembled into the actin cytoskeleton.

TRANSLATION

Protein synthesis is implicated in the control of cell growth, proliferation, and differentiation[58]. Although a great deal of this process has been elucidated, much remains to be understood. In eukaryotes, protein synthesis can be divided into three general steps: initiation, elongation and termination. Each step requires specific RNA-protein interactions including those involving an RRM. One of the best characterized RRM-containing proteins involved in translation is the poly(A) binding protein (PABP). PABP is a 71 kDa phylogenetically conserved protein involved in translation initiation and mRNA stabilization[58]. During translation, this protein binds the poly(A) tail of an mRNA molecule with high specificity and physically interacts with eIF4G, promoting mRNA circularization and enhancing ribosome recruitment and translation[59]. These facts support the “closed loop” model for translation initiation[60]. Additionally, PABP can bind to the 5′ cap structure of an mRNA, negatively regulating decapping activity of Dcp2[61].

Biochemical analysis of PABP revealed that it contains four RRMs and a C-terminal proline-rich region (Fig. 2). RRM1 and RRM2 are located at the N-terminal region of PABP and are involved in poly(A) binding[58]. RRM3 and RRM4 are involved in AU-rich region binding. The carboxyl terminal third of PABP contains an unstructured region, as well as a α-helical peptide-binding domain[58]. The RRMs in PABP have been determined to be involved in the ability of this protein to bind different sequences of RNA and also in forming protein–protein interactions that can dramatically affect its function. The activity of PABP and its ability to promote protein translation can be post-translationally regulated by protein–protein interactions and this regulation also involves RRM binding. For example, Paip1, another RRM containing protein, has been shown to bind to PABP via RRM1 and RRM2 in a 1:1 stoichiometry to promote translation[62,63]. On the other hand Paip2 has been shown to bind to PABP via RRM3 and RRM4 and decrease its affinity for poly(A)[62]. These factors thus provide yet another example of the basic architecture of RRMs being used for, in this case, opposing functions within RNA metabolism.

mRNA DECAY

The control of cytoplasmic mRNA turnover plays a major role in targeting the precise timing and expression of many gene products in eukaryotes. In mammals, A + U-rich elements (AREs) direct the rapid turnover of many labile mRNAs, including several oncoproteins, cytokines and G-protein coupled receptors[64]. AREs generally consist of one or more overlapping AUUUA pentamers within or near a U-rich region. However, ARE-directed turnover mechanisms are largely heterogeneous in extent of decay as well as in decay characteristics[65]. This type of heterogeneity arises from the diverse RNA- protein and protein–protein interactions that regulate turnover.

Several ARE-binding proteins have already been identified and cloned. AUF1 (also referred to as hnRNPD) was first isolated from the erythroleukemic K562 cell line[66]. AUF1 exists as a family of four protein isoforms denoted by their molecular weights: p37, p40, p42 and p45 resulting from alternative splicing of a common pre-mRNA, and all associate with RNA substrates via a tandem pair of RRMs[67]. Each AUF1 isoform contains a common N-terminus, both RRMs, and a glutamine-rich domain. The p40, p42 and p45 isoforms are determined by the presence or absence of a 19-amino acid insert or a 49-amino acid insert, respectively, while p37 lacks both inserts[68]. AUF1 mediates ARE-directed decay as part of a multi-subunit complex with itself and with additional cellular factors such as: translation initiation factor eIF4G, PABP, and the heat shock proteins Hsc70 and Hsp70.

Although the binding of AUF1 to AREs results in destabilization of the mRNA, there are also proteins such as the Hu family of proteins which bind to AREs and confer stability. The Hu proteins have been shown to interact with AREs through two of their three conserved RRM domains[69], and the stabilizing effects of this interaction have been well characterized in neuronal development where the HuD protein has been shown to be essential for proper differentiation of neuronal precursors[70]. The solution of the crystal structure for HuD bound to the c-fos mRNA elucidated the structural basis for recognition of the ARE sequence by the RRMs of HuD[71]. The two RRMs which interact with the ARE show typical RRM topology, each organized into β1α1β2β3α2β4 (Fig. 2) forming two interacting surfaces which cover approximately 49% of the ARE surface. This extensive interaction may act to shield the ARE and prevent destabilizing proteins from binding. Only one pair of residues was found to interact between the two RRMs and it is unclear if this interaction between RRMs is essential for proper ARE binding. On the basis of this structure of HuD bound to c-fos mRNA as well as HuD bound to TNF-α mRNA, the consensus binding sequence was predicted to be N-U/C-U-N-N-U/C-U-U/C. Superposition of this RRM over several other RRM crystal structures revealed two conserved binding pockets, the U4/U10 and U3/U9 which may be responsible for binding site specificity. These studies provide structural insight into how the specificity of RRM binding to RNAs may be achieved. Though we have focused on RRM containing proteins which bind to AREs in order to regulate mRNA stability, other RRM containing proteins such as PABP[72] or the yeast protein Rbp1p[73] regulate the stability of mRNA by binding to other sequence elements further demonstrating the flexibility of the RRM to accommodate different recognition sequences for different mRNAs. These, as well as other proteins and sequence elements which have yet to be characterized, show the importance of the RRM containing proteins in contributing to the specificity that is essential for proper regulation of mRNA turnover.

SUMMARY

In conclusion, the RNA binding platform is one of the most abundant eukaryotic protein domains and the list of RRM-bearing proteins that are involved in post-transcriptional events is extensive. RRM containing proteins maintain a limited number of critical motifs and structural similarity, yet RRMs influence an astonishing breadth of targets and functions. The determinant of specificity of each RRM is something that is not yet understood and requires further investigation. Sequence specificity, typified by the comparison between U1A and U2B[12], structural specificity, demonstrated by ARE recognition via the RRMs of HuD[71] and mRNA length recognition, as in the case of CRM1[33], have all been suggested to play a role in the binding of the RRMs and other protein-RNA binding motifs to RNA. What actually influences the RRM-mRNA interactions is something that is still not understood and warrants additional investigation.

USING MOTIF-BASED COURSE DESIGN AS A TOOL IN GRADUATE EDUCATION

Graduate students are more likely to advance if they learn to explore the connections between processes, moving beyond simple understanding of fact and function towards application of knowledge for promoting testable questions. As adults they may learn best when they see the connections between the course content and their own research or professional goals, when they assume responsibility for their learning, and when they participate in discussion that promotes reflection and transference of content to broader areas[74,75]. Furthermore adult learning theories suggest that adults do better with strategies that incorporate peer learning, highlight the connections between their own knowledge and experiences with new concepts, and provide a challenging yet collegial environment[75]. A new graduate course structure was developed to promote these goals by two methods: thematically demonstrating the multitude of functions possible through a single protein domain and requiring students to generalize and integrate these connections through not only presentation but also writing. To be effective writers students need guidance and a more systematic approach which utilizes groups, peer learning such as with peer review, and supports both the shared discourse community and exemplification of normal practice[76]. Drawing on this theoretical background this paper presents a scientific review as well as an applied pedagogical approach to graduate level science education.

The diversity of RNA binding motifs and the multitude of biological functions they perform make the organization of a graduate course on this topic challenging. A one credit graduate seminar course entitled RNA:Protein Interactions has been offered by the Joint Graduate Program in Molecular Genetics, Microbiology and Immunology at UMDNJ Robert Wood Johnson Medical School and Rutgers, The State University of New Jersey for several years. Initial course organization focused on covering the topic based on the different types of RNA binding motifs.

However, a subsequent course was designed to instead focus on the RRM as the predominant RNA binding motif and utilize the diversity of biological functions as a platform to dissect how a single motif is modified to perform different functions. To encourage students to assume responsibility for their learning and incorporate writing skills, which students could programmatically see as professionally important, a change in both theoretical and instructional approaches was adopted. Over 12 weeks the students were assigned in pairs to cover a review on the biological process and then a paper on one RRM containing protein that functions in that pathway. At the end of the semester, each pair of students wrote a one page review, integrating the pathway and the RRM example. By working in pairs, students were able to develop editing skills. These one page reviews were merged into a single review edited by all the students and the course directors.

At the instructional level, this unique approach enabled students to gain a deeper understanding since they were more actively involved in explaining the course content. In addition to acquiring skills in reading the primary scientific literature and preparing and giving presentations, students developed the ability to write a section of a scientific paper and go through the revision and submission process.

STUDENT EVALUATIONS

The student evaluations for the first offering of the course were one motivating factor for the new course deign. First, anonymous structured feedback on that course indicated that 73% of the students responding (response rate 92%) disagreed or strongly disagreed that the work load was too high. Thus, the rigor of the course was increased. This resulted in a complete shift in the second offering to where 100% of the students responding (response rate 67%) agreed or strongly agreed the work load was too high. Second, student comments suggest the change was beneficial. Changes were noted from the first offering where some students suggested “assign review article for each paper, for better introduction” and requested “more good papers” to the second offering where the pairing of a review with a primary literature paper and students picking their own papers were positively commented on. Notable were two students who remarked on learning: “The format of one review presentation of a theme and then the research paper presentation is good and you can get a better idea of the subject”; “I liked the fact that this term we …[were given] responsibility that is not found in other seminars”. The end result of aiming to publish the review paper was likely a contributing factor to the concern on workload but also was positively received “the idea of publishing a review paper at the end of the semester from the work done in class I think is an excellent way to motivate us”. Overall, the students in the first offering rated the course overall 82% in the top 2 units of a five point scale (excellent or above average) while in the second offering 100% rated the course in the top 2 units (excellent or good).

  • 1

    The abbreviations used are: RRM, RNA Recognition Motif; snRNA, small nuclear RNA; CPSF, cleavage and polyadenylation specificity factor; DSE, downstream sequence element; CEFs, chicken embryo fibroblasts; PAP, polyadenylate polymerase; PABP, poly(A) binding protein; CPEB, CPE binding protein; APLP1, amyloid precursor-like protein 1; CTE, constitutive transport element.

Ancillary