KSRP PROMOTES THE MATURATION OF A GROUP OF miRNA PRECURSORS

Abstract

microRNAs (miRNAs) are small noncoding RNAs that down-regulate gene expression by reducing stability and/or translation of target mRNAs. In animals, miRNAs arise from sequential processing of hairpin primary transcripts by two RNAse III domain-containing enzymes, namely Drosha and Dicer, to generate a mature form of about 22nucleotides. In this chapter we discus our latest findings indicating that KSRP is an integral component of both Drosha and Dicer complexes. KSRP binds to the terminal loop sequence of a subset of miRNA precursors promoting their maturation. Our data indicate that the terminal loop is a pivotal structure where activators of miRNA processing as well as repressors of miRNA processing act in a coordinated way to convert cellular signals into changes in miRNA expression processing. This uncovers a new level of complexity of miRNA mechanisms for gene expression regulation.

INTRODUCTION

miRNA precursors are encoded within diverse functional regions of the genome with more than half of human miRNA genes encoded within introns, while others are transcribed as independent monocistronic or polycistronic units.¹ The majority of miRNAs are transcribed by RNA polymerase II and, subsequently, they are capped and polyadenylated.^1,2 It has been recently demonstrated that Drosha cleavage of intronic pri-miRNAs occurs cotranscriptionally and precedes splicing.³ Additional work from Bozzoni and coworkers further strengthened the connection between transcription and Drosha cleavage also for intragenic pri-miRNAs.⁴

The evolutionarily-conserved mechanism by which primary miRNAs (pri-miRNAs) are processed first to precursor miRNAs (pre-miRNAs) and then to mature miRNAs involves two consecutive endonucleolytic cleavages executed by multiprotein complexes containing the RNase III enzymes Drosha and Dicer, respectively.⁵ Drosha processes the pri-miRNA into a ~70 nt hairpin pre-miRNA.⁵ Through the interaction with Exportin-5 and Ran-GTP, the pre-miRNA is transported into the cytoplasm, where it undergoes a second round of processing catalyzed by Dicer.⁵ This cleavage event gives rise to a double-stranded ~22 nt product composed of the mature miRNA guide strand and the miRNA* (star) passenger strand.⁵ The mature miRNA is then loaded into the RNA-Induced Silencing Complex (RISC) while the passenger strand is degraded. In the context of the RISC, miRNAs posttranscriptionally regulate the expression of target genes.^5,6 Although major progress has been made in understanding the basic mechanism of miRNAs biogenesis, many questions remain unanswered. Specific chapters of this book, besides recently published reviews, extensively describe the majority of the known aspects of miRNA biogenesis regulation^1,2,7 Thus we will focus our discussion on some open questions regarding mechanistic and functional aspects of miRNA biogenesis controlled by the multifunctional single-strand RNA-binding protein KSRP (KH-type splicing regulatory protein).

CO-ACTIVATORS AND CO-REPRESSORS OF miRNA PRECURSOR MATURATION

Pioneering studies from different laboratories have envisaged the possibility that miRNA maturation is a finely regulated process that involves co-regulators and is suited to respond to changing cellular conditions.^8–13

Recent studies have revealed that different regulatory proteins participate in the specific control of individual miRNAs. The DEAD-box RNA helicases p68 (DDX5) and p72 (DDX17) were identified as components of the Drosha-mediated processings complex and, in the absence of both DDX5 and DDX17 the expression of approximately 35% of miRNAs (and pre-miRNAs) was reduced without concomitant changes in the levels of corresponding pri-miRNAs.⁹ This suggested a role for DDX5/DDX17 in promoting the Drosha-mediated processing of a subset of miRNAs. The precise mechanism of DDX5/DDX17-mediated processing is still unclear, but may involve re-arrangement of the RNA hairpin, which results in enhanced Drosha recruitment or stability. Alternatively, as DDX5/DDX17 are known to interact with a variety of proteins, they may serve as a scaffold for the recruitment of multiple factors to the Drosha complex. Indeed, a recent paper revealed that the tumor suppressor and transcription factor p53 interacts with the Drosha complex through the association with DDX5 and facilitates the maturation of a restricted population of pri-miRNAs in response to DNA damage in cancer cells.¹⁴ The role of DDX5/DDX17 as adaptors for signal- and cell-specific cofactors in miRNA processing is further supported by the positive regulation of Drosha-mediated processing mediated by the DDX5-interacting Smad proteins, the signal transducers of the TGF-beta family signaling cascade (see also chapter by Hata and Davis).⁸ Very recent data from the Kato laboratory, reviewed in the chapter by Fujiyama-Nakamura et al, suggested a mechanism whereby estradiol-bound estrogen receptor α blocks Drosha-mediated processing of a subset of miRNAs by binding to Drosha in a DDX5/DDX17-dependent manner and inducing the dissociation of the microprocessor complex from the pri-miRNA.¹⁵

Smad proteins as well as p53 and estrogen receptor α do not directly bind to RNA. Conversely, other Drosha cofactors are RNA-binding proteins. The heterogeneous nuclear ribonucleoprotein (hnRNP) A1, which interacts with RNA through its RNA recognition motifs and appears to be involved in many aspects of RNA life, originally proved to be required for the maturation of miR-18a, a member of the miR-17-92 cluster.¹⁰ We recently demonstrated that also the KH-type splicing regulatory protein (KSRP) is able to interact with select miRNA precursors.¹² KSRP is a multifunctional single-strand RNA-binding protein that affects many steps of RNA life including splicing, localization and degradation.^16–20 Recently, we demonstrated that KSRP binds to the terminal loop (TL) of a cohort of miRNA precursors and interacts with both Drosha and Dicer to promote miRNA maturation.^12,21 In Figure 1, the miRNAs whose biogenesis is regulated by KSRP are listed. Interestingly, also hnRNP A1 binds to the TL of a group of pri-miRNAs, which partly overlaps that interacting with KSRP.^10,22 Even more intriguingly, KSRP interacts with hnRNP A1 and regulates its expression by affecting its mRNA half-life.²³ A relevant difference between hnRNP A1- and KSRP-activated miRNA maturation is that, in contrast to hnRNP A1, KSRP functions not only in the nuclear maturation of pri-miRNAs to pre-miRNAs but also in the cytoplasmic maturation of pre-miRNAs into miRNAs, thus representing a link between nuclear and cytoplasmic events. Indeed, we also obtained evidence that KSRP interacts with Exportin-5 in an RNase A-sensitive way thus suggesting that KSRP is associated with the TL of target miRNA precursors during nucleo-cytoplasmic transit.

Figure 1.

A schematic model of the interplay between positive and negative co-regulators of miRNA precursors maturation, including a list of miRNAs whose maturation is controlled by KSRP.

KSRP-RNA recognition is a complex event and relies both on the sequence selectivity of the KH domains and on the actual availability of single stranded RNA sequences at the recognition site.^17,24 We have examined the sequence preference of the four KH domains of KSRP. Three of them (KH1, KH2 and KH4) display a moderate selectivity towards specific sequences, while KH3 can discriminate strongly in favour of short G-rich stretches.²⁴ If a G-rich stretch is available within the single stranded RNA target (e.g., in the let-7a precursors) KH3 will bind to it, defining KSRP’s binding frame. However, if the RNA target does not include a G-rich stretch, the four KH domains of KSRP will explore all available sequences in order to optimize the binding affinity of KSRP (e.g., in the miR-21 precursors). In either case, the structural setting of the single-stranded sequences will play a major role in KSRP binding, as KSRP-RNA interaction is a multi-domain event and steric hindrance may limit the space available to the protein domains within the RNA 3D structure. The size and conformation of the TL of miRNA precursors vary significantly and, together with the presence of a G-rich sequence, could represent a powerful selector for KSRP recognition.¹² In conclusion, we propose a general model for KSRP-RNA interactions based on the differential use of multiple domains that underscores the adaptability of the protein to a broad range of single-strand RNA sequences.

A number of manuscripts appeared in the last year proving that miRNA maturation can also be negatively regulated through the intervention of cofactors. We have already mentioned the estrogen receptor α, which exerts an indirect negative function on miRNA maturation acting through DDX5/DDX17.¹⁵ The RNA-binding protein Lin-28 is able to repress the maturation of let-7 family members and this effect is mediated by its interaction with the TL of let-7 precursors (see also chapter by Lehrbach and Miska).^13,25,26 In addition to inhibition of the Drosha-mediated processing step, Lin-28 also inhibits the Dicer-mediated processing of let-7 family members.^13,27,28 Indeed, similarly to KSRP, Lin-28 is a shuttling protein and it is abundant in the cytoplasm, suggesting that this may be the primary location of its interaction with let-7 precursors. Data from Wulczyn laboratory suggest that Lin-28 compete with Dicer for access to pre-let-7.²⁷ Additionally, Kim and coworkers recently reported that Lin-28 promotes the 3′-uridylation of pre-let-7, which inhibits Dicer-mediated processing and leads to degradation of pre-let-7 itself.^29–31

While Lin-28 repressor function seems to be restricted to let-7 family members, the double-stranded RNA-binding proteins NF90 and NF45 reduce the production of a broader spectrum of pre- and mature miRNAs. NF90 and NF45 interact with the stem of a group of pri-miRNAs to preclude their binding to DGCR8, an essential member of the Drosha complex and, in turn, their maturation into pre-miRNAs.³²

Figure 1 summarizes the proposed scenario of activators and repressors of miRNA maturation.

IMPACT OF KSRP AND OTHER CO-ACTIVATORS AND CO-REPRESSORS OF miRNA PRECURSOR MATURATION ON CELL PROLIFERATION, DIFFERENTIATION AND CANCER

As a single miRNA modulates the expression of many targets simultaneously, it is able to rapidly modify complex cellular functions that require the coordinated regulation of gene networks, in response to environmental cues. Therefore, regulation of miRNA biogenesis may serve as an important line of response to promote the modulation of gene expression programs in response to cellular stimuli. As KSRP as well as the majority of co-regulators mentioned in this review are able to influence maturation of discrete groups of of miRNAs, this could allow the coregulation of miRNAs implicated in certain cellular functions in response to certain cellular stimuli.

A good example of how the control of maturation of a group of miRNAs by a single cofactor can affect a cellular function is represented by the KSRP-directed maturation of myogenic miRNAs in response to differentiative stimuli in myoblasts.¹² Upon serum withdrawal, C2C12 myoblasts undergo differentiation into myotubes. This event is accompanied by enhanced expression of some “myogenic” miRNAs (including miR-1, miR-133a and miR-206) reported to have a causative role in the differentiation process.^33,34 KSRP knockdown impairs myogenic miRNA maturation, increases the expression of some of their targets and inhibits C2C12 myoblasts differentiation.¹² These data link the stimulation of pri-miRNA processing by KSRP to a mammalian differentiation program.

Some evidence suggests a functional interplay between positive and negative co-regulators of pri-miRNA maturation. Lin-28 and mature let-7g show reciprocal expression patterns during both embryonic development and embryonic stem cell differentiation thus supporting a role of Lin-28 in let-7g regulation during embryogenesis.^1,2 Even though Lin-28 and KSRP do not share common binding sites in the TL of let-7 family precursors,^12,25,26 our data suggest that when Lin-28 is expressed in undifferentiated embryonic carcinoma cells, KSRP cannot interact with pri-let-7g. When Lin28 is not expressed, as in differentiated P19 cells or in undifferentiated P19 upon specific Lin-28 knockdown, KSRP is able to promote let-7g maturation.¹² We propose that the TL is a pivotal structure where miRNA-processing-co-activators (KSRP and possibly additional RNA-binding proteins) as well as miRNA-processing co-repressors (as exemplified by Lin-28 for let-7 and possibly additional RNA-binding proteins) function in a coordinated way to convert proliferation and differentiation cues into changes of miRNA expression. In other words, the occurrence of a co-activator and a co-repressor for regulation of miRNA maturation extends the concept of opposing co-regulators, analogous to events now well established for DNA-binding transcription factors (Fig. 1).

The differential expression of miRNA processing co-regulators, as well as their posttranslational modification and sub-cellular localization, may serve to regulate miRNA expression in a tissue-or context-dependent manner. For example, phosphorylation of hnRNP A1 by MAPK p38 has been reported to promote cytoplasmic localization of hnRNP A1.^35,36 Similarly, KSRP can be phosphorylated by two distinct Ser/Thr kinase, MAPK p38 and Akt. While MAPK p38-mediated phosphorylation affects the binding efficiency KSRP of to some RNA substrates, Akt promotes nuclear accumulation of KSRP through interaction with 14-3-3.^37–39 Intriguingly, phosphorylation by Akt impairs the ability of KSRP to interact with some enzymes responsible for decay of labile mRNAs.³⁹ An important future challenge will be to systematically dissect pathways that modulate the function of KSRP and other RNA-binding proteins in the regulation of miRNA biogenesis.

Furthermore, it is now clear that the function of a single cofactor can vary depending on the cellular context in which both the cofactor and its target miRNAs are expressed. For example KSRP, which is ubiquitously expressed, affects, through miRNA maturation control, cellular fuctions as different as myoblast diffretiation and inflammatory response to microbic products (such as lipopolysaccharide, LPS) in macrophages. In this last case, the regulation of a single miRNA (miR-155) is responsible for the KSRP-mediated response to LPS.²¹

Finally, our studies showed that KSRP knockdown limits, in a let-7a-dependent way, cell proliferation by influencing the expression of let-7a targets such as MYC and NRAS.¹² A seminal study from Thomson and coworkers implicated the regulation of miRNA precursor processing in cell transformation and cancer.⁴⁰ Indeed, a notable global reduction of mature miRNAs has been observed in cancers.⁴¹ In addition, the importance of miRNA processing regulation for tumorigenesis has been experimentally proved by Drosha, Dgcr8 or Dicer knockdown.⁴² The demonstration that Lin-28 is overexpressed in primary human tumors and human cancer cell lines,^43–45 which is linked to the down-regulation of let-7 expression,¹³ supports the idea that altered control of miRNA biogenesis may critically impact on cancer pathogenesis, representing a stimulus for further intense investigations.

CONCLUSION

Altogether, we conclude that KSRP also serves as a previously unsuspected component of both Drosha and Dicer complexes and regulates the biogenesis of a subset of miRNAs. KSRP binds in a sequence-specific fashion to the TL of a subset of pri- and pre-miRNAs and functions as a co-activator for miRNA processing. The evidence that both co-activators (such as KSRP) and co-repressors (such as Lin-28) of miRNA maturation exist and their interplay is required to precisely regulate miRNA expression in specific cellular contexts provides the rational basis to identify additional co-regulators of miRNA processing, stimulating therefore future research in this area of gene expression regulation.