Coding region, the neglected post-transcriptional code

Abstract

The control of mammalian mRNA turnover and translation has been linked almost exclusively to specific cis-elements within the 5′- and 3′-untranslated regions (UTRs) of the mature mRNA. However, instances of regulated turnover and translation via cis-elements within the coding region (CR) of mRNAs are accumulating. Here, we describe the regulation of post-transcriptional fate through trans-binding factors (RNA-binding proteins and microRNAs) that function via CR sequences. We discuss how the CR enriches the post-transcriptional control of gene expression, and predict that new high-throughput technologies will enable a more mainstream study of CR-governed gene regulation.

Introduction

In mammalian cells, a tight regulation of gene expression on multiple levels ensures proper responses to acute damaging agents, proliferative stimuli and developmental cues. Protein expression patterns are strongly influenced by a wide array of post-transcriptional processes affecting every aspect of mRNA metabolism, from its synthesis to its degradation.^1–3 Among these processes, changes in the stability and translation of mature mRNA potently affect protein output. Two main types of mRNA-binding regulators have been implicated in the control of mRNA stability and translation: RNA-binding proteins (RBPs) and noncoding RNAs, particularly microRNAs.

RBPs that regulate turnover and translation (sometimes named TTR-RBPs) associate with target mRNAs via different RNA-binding domains and regulate their stability and translation. Mammalian TTR-RBPs constitute a large family of proteins that includes human antigen (Hu) proteins [HuR (HuA), HuB, HuC, HuD], AU-binding factor 1 (AUF1), T-cell intracellular antigen 1 (TIA-1) and TIA-1-related (TIAR) proteins, tristetraprolin (TTP), polypyrimidine tract-binding protein (PTB), CUG triplet repeat RNA-binding protein (CUGBP), fragile X mental retardation protein (FMRP), the coding region determinant-binding protein (CRD-BP) and heterogeneous nuclear ribonucleoproteins (hnRNP) A1, A2, C1/C2 (reviewed in ref. 4 and 5). The vast majority of these RBPs have been shown to affect mRNA turnover and translation by interacting with the 3′-untranslated region (UTR) of target mRNAs, and in some cases with the 5′UTR.^3,6,7

MicroRNAs are ∼22 nt-long noncoding RNAs, which are loaded onto RNA-induced silencing complexes (RISCs) that contain the Argonaute (Ago) RBPs as core components. In mammalian cells, they typically repress the translation of target mRNAs and/or destabilize mRNA by forming incomplete Watson-Crick base-paring.^8,9 Like RBPs, almost all microRNAs have been reported to function by interacting with the 3′UTR of target mRNAs, but they occasionally interact with the 5′UTR.^10–12 Through their influence on the production of proteins encoded by target mRNAs, both RBPs and RISC-loaded microRNAs have been implicated in important biological processes, including cell differentiation, cell cycle progression, carcinogenesis and the response to immune and stress stimuli.

As mentioned above, the UTRs provide fertile ground for regulation by both RBPs and RISC-bound microRNAs, sometimes binding individually, but often binding combinatorially. Thus, RBPs can compete or cooperate with other RBPs to elicit changes in mRNA abundance and/or translation;^13,14 likewise, microRNAs can synergize with other microRNAs that interact with a shared target transcript.¹¹ Recent examples are also emerging of functional interconnections (cooperative or competitive) between RBPs and RISC-microRNAs on shared UTRs.^15,16

A small but growing number of studies indicate that RBPs and microRNAs elicit similar post-transcriptional gene regulation by interacting with coding region (CR) elements. Here, we review prominent examples of post-transcriptional gene control via the CR. With novel deep-sequencing methodologies for transcriptome analysis, we anticipate that much more regulation through the CR will soon come into view.

The UTR Bias

As the CR is the template for amino acid information and protein biosynthesis, this mRNA region was widely considered to be the exclusive domain of ribosomes. Accordingly, the past two decades' efforts to identify cis-elements of post-transcriptional control focused on the 5′UTR and the 3′UTR. In the 5′UTR, several regulatory motifs have been described, for example the polypyrimidine tract, the internal ribosome entry site (IRES), the iron response element (IRE), and the 5′ terminal oligopyrimidine (5′TOP) tract (reviewed in ref. 6); several sites of interaction with microRNAs have also been reported in the 5′UTR.^10–12 In the 3′UTR, translation and stability are governed by sequences such as the cytoplasmic polyadenylation elements (CPEs), U- and AU-rich elements (AREs), GU-rich elements (GREs) and other elements rich in CUGs, CUs, etc.^17–19 Additionally, the vast majority of mammalian microRNAs described have been shown to function through binding to 3′UTR sequences.⁸

While the CR has received less attention, much of the cellular mRNA is in fact excluded from polysomes (e.g., nuclear mRNAs and cytoplasmic mRNAs in transit or storage) or is occupied sparsely by ribosomes^20,21 (Fig. 1). Therefore, the non-translating mRNA pool and mRNA segments are available for interaction by RBPs and noncoding RNAs. The resurgence in interest in CR regulatory elements is largely driven by the recent discovery of CR-directed microRNAs, although several RBPs that function through target CR sequences have also been identified. Below, we discuss several reports of CR-directed trans-binding factors (Table 1).

Figure 1.

CR-bound RBPs and microRNAs in the mammalian cell. RBPs and microRNAs may be found associated with mRNAs during transport, while mRNAs are stored in specialized cytoplasmic locales (e.g., stress granules or processing bodies), or in polysomes with low ribosome occupancy. Mature mRNAs may also associate with RBPs and microRNAs in the nucleus.

Table 1.

mRNAs regulated through the CR

Factors	Target mRNAs	Regulators	Function	References
CR-binding RBPs	APP	hnRNPC FMRP	Translation ↑ Translation ↓	31 30, 31
	β-TrCP	CRD-BP (1)	Stability à	28
	c-Myc	CRD-BP (1)	Stability ↑	22–27
	c-fos	Unr	Stability ↓	33
	PAI-2	Unknown 50-kDa RBP	Stability ↓	35
	MnSOD	Unknown RBP	Stability ↓	34
	VEGF	Unknown RBP	?? (2)	36
CR-binding miRNAs	p16	miR-24	↓ (3)	37
	HuR	miR-519	↓	43
	Dicer	Let-7	↓	38
	Nanog	miR-296, miR-470	↓	40
	Oct4	miR-470	↓	40
	Sox2	miR-134	↓	40
	Hoxa9	miR-126	↓	41
	Dnmt3b	miR-148	↓	42
	ZNFs	miR-181a	↓	44

Examples of RBPs and microRNAs that interact with the CR. Mammalian mRNAs regulated via changes in stability and translation through CR sequences. Columns list factors interacting with the CR (RBPs, microRNAs), the target mRNAs regulated via CR elements, the RBPs/microRNAs involved, the effect on expression the mRNA, and relevant references. (1) CRD-BP is also named IMP-1 and IGF2BP1; (2) reduction is likely due to lowered mRNA stability; (3) all microRNAs reduced expression of targets by diminished mRNA stability and/or translation.

CR-interacting RBPs

CRD-BP.

Despite the overall scarcity of CR cis-elements described in mammalian mRNAs, the first such sequence was actually described almost two decades ago by the Ross laboratory.²² A 182-nt segment within the c-myc CR, the coding region determinant (CRD), strongly increased the half-life of the c-myc mRNA. The CRD-BP was later identified as a protein with four K homology (KH) domains which enhanced c-myc mRNA stability by protecting the transcript from endonucleolytic attack; other names of CRD-BP include IGF-II mRNA-binding protein (IMP-1, IMP-2, IMP-3) and zip code-binding protein (ZBP1).^23–27 To stabilize the c-myc mRNA, CRD-BP requires additional factors, including hnRNP I, Syncrip, YBX1R and DHX9.²⁷ Recently, CRD-BP was also reported to bind to the CR of the β-TrCP mRNA and stabilized it, thereby increasing the abundance of the encoded protein, β-TrCP. In colon cancer cells responding to β-catenin signaling, CRD-BP promoted the expression of both c-myc and β-TrCP, leading to the activation of NFκB and the inhibition of apoptosis.²⁸

FMRP.

The fragile X mental retardation protein (FMRP) has important developmental roles and is encoded by FMR1, a gene whose mutation is linked to the fragile X mental retardation syndrome.²⁹ FMRP binds the CR of the amyloid precursor protein (APP) mRNA and represses APP mRNA translation in response to the metabotropic glutamate receptor agonist DHPG (dihydroxyphenylglycol).³⁰ The inhibition of APP expression appears to be due, at least in part, to the FMRP-mediated recruitment of APP mRNA to processing bodies (PBs), specialized sites of mRNA degradation and translational repression.³¹

hnRNP C.

We recently reported that hnRNP C bound to the APP mRNA CR, causing an enhancement in APP translation.³¹ This function was explained, at least in part, by the competition between hnRNP C and FMRP for binding to the same RNA segment within the APP CR. According to the proposed model,³¹ these two RBPs regulate APP mRNA translation competitively: FMRP suppresses APP translation by recruiting APP mRNA into PBs, while hnRNP C enhances APP translation by competing with FMRP binding and hence blocking the localization of APP mRNA in PBs.

Unr.

This cold-shock domain (CSD)-containing RBP, previously implicated in IRES function,³² was shown to bind to the major CR determinant of instability (mCRD) within the c-fos mRNA. By interacting with the poly(A)-binding protein (PABP), Unr promoted the translation-coupled decay of c-fos mRNA.³³

Unknown CR-binding RBPs.

The manganese superoxide dismutase (MnSOD) mRNA has a CR determinant that confers stability to the transcript.³⁴ Similarly, the plasminogen activator inhibitor type 2 (PAI-2) mRNA has stability elements within the CR that prevent its degradation by interacting with an as-yet-unknown 50- to 52-kDa RBP.³⁵ The vascular endothelial growth factor (VEGF) mRNA also has a regulatory region within the CR which affects mRNA stability, but the RBP(s) involved remains to be identified.³⁶

CR-interacting microRNAs

The past few years have also seen a rising number of examples of microRNAs that target sequences within the mRNA CR. For example, the CR of p16 mRNA, which encodes an inhibitor of cyclin-dependent kinases, was shown to be the target of miR-24, which primarily repressed its translation.³⁷ Likewise, let-7 interacted with the CR of the Dicer mRNA and suppressed expression of Dicer, a critical RNase implicated in microRNA biogenesis.^38,39 Expression of differentiation factors Nanog, Oct4 and Sox2 was similarly repressed through the association of microRNAs miR-296, miR-470 and miR-134 with the corresponding CRs.⁴⁰ MicroRNAs targeting the CRs of mRNAs encoding the homeobox protein Hoxa9, the zinc finger protein ZNF, the methyltransferase DNMT3B and the RBP HuR (miR-126, miR-181a, miR-148 and miR-519, respectively) similarly repressed the biosynthesis of the encoded proteins.^41–44

Progress in the identification of CR-directed microRNAs has been hampered by the exclusion of CR sequences from the databases accessed by microRNA prediction algorithms (e.g., TargetScan, PicTar, miRBase, etc.). However, recent transcriptome-wide sequencing analyses have revealed that CRs likely harbor numerous microRNA sites. For example, HITS-CLIP analysis by the Darnell laboratory showed that 25% of associations between Ago protein and mRNAs were in the CR,⁴⁵ although the putative microRNA-mRNA interactions at these CR sites await individual study.

Conclusion

For two decades, there were only anecdotal examples of CR cis-elements affecting the mRNA's post-transcriptional fate. With the recent discovery that numerous microRNAs function through CR sequences, there has been a surge in interest in regulatory elements present within the CR. The increased access to high-throughput, transcriptome-wide sequencing technologies, as well as the inclusion of CR sequences in microRNA-mRNA prediction algorithms will be especially important to propel progress.

A greater appreciation of CR-acting microRNAs will consequently rise the recognition of CR-acting RBPs. Indeed, RBPs and RISC-microRNAs share an affinity for mature mRNAs, both rely on the same cellular machineries to repress or increase gene expression (PBs, ribosomes, ribonucleases, etc.,), and in several instances, they have important functional interactions, sometimes cooperative or competitive.^15,16

It remains to be determined whether CR-directed microRNAs/RBPs interact with mRNAs engaged in transport, with mRNAs present in specialized cytoplasmic domains, or with translating mRNAs sparsely occupied with ribosomes. The increased availability of technologies that identify exact RBP/microRNA binding sites and ribosome occupancy, will allow these important questions to be addressed in depth. With renewed appreciation that CRs encode key information on mRNA turnover and translation, the stage is ready to incorporate CR-interacting factors into the increasingly rich constellation of regulators of mammalian gene expression programs.

Abbreviations

CR

coding region

RBP

RNA-binding protein

UTR

untranslated region