Subcellular localization of mRNAs is a key mechanism for the synthesis of proteins close to their site of function. The mRNA encoding MT-1 (metallothionein-1) is localized in the perinuclear cytoplasm, where it is associated with cytoskeletal-bound polysomes. This localization relies on sequences present in the 3′-UTR (3′-untranslated region). The present study aims to characterize the cis-acting localization element(s) within the 3′-UTR. Using transfected cells expressing tagged MT-1 differing in their 3′-UTRs (deleted or mutated), the section(s) of this region required for directing MT-1 transcripts to the perinuclear cytoplasm has been investigated. Different 3′-UTRs were also used in UV cross-linking experiments that highlighted two distinct regions (nt 26–30 and 66–76) necessary for the binding of a protein of approx. 50 kDa, presumably involved in the mRNA targeting. The poor sequence homology between the MT-1 3′-UTR of various species, together with the bipartite nature of the required cis-element, indicates the involvement of a particular structure in the localization signal. The secondary structure of the MT-1 3′-UTR was investigated using enzymic and chemical probing. Current structural analysis of mutant 3′-UTRs will allow the critical structural features of the MT-1 mRNA perinuclear localization signal to be defined.
There is considerable evidence that certain mRNAs are targeted to specific cytoplasmic regions in a wide range of cells from yeast to mammals. It is generally proposed that such messenger localization provides a mechanism to restrict synthesis of the encoded proteins close to where they function. In Drosophila and Xenopus, different maternal mRNAs, such as bicoid and Vg1, are directed to either pole of the oocyte/embryo . This determines gradient formation of the encoded morphogens that are critical for the embryo polarity and future developmental fate. Localization of RNA transcripts to specific subcellular sites also occurs in differentiated somatic cells. For example, in oligodendrocytes, the myelin basic protein mRNA is transported along the processes before being associated with the myelin compartment . In less specialized cells, some mRNAs are specifically found at the periphery, such as the β-actin mRNA at the leading edge of spreading fibroblasts, whereas others, including those coding for the transcription factors MYC and FOS, as well as for MT-1 (metallothionein-1), are present in the perinuclear cytoplasm where they are associated with the cytoskeleton [3,4].
Mechanisms for the sorting of mRNAs generally involve their transport and/or anchoring by means of an association with cytoskeletal components. In particular, perinuclear MT-1, c-myc and c-fos mRNAs are found enriched in cytoskeletal-bound polysome fractions . As observed in most cases of mRNA localization, this specific association depends on cis-acting element(s) present in the 3′-UTR (3′-untranslated region). Removal of this region causes loss of mRNA and subsequent protein localization. It has been shown that the 3′-UTR-mediated association of MT-1 mRNA with the cytoskeleton around the nucleus is essential for the nuclear location of the MT-1 protein during the G1/S phase transition of the cell cycle . This suggests a physiological role for perinuclear mRNA localization, which was further supported by the influence of the messenger targeting on MT function as a protective factor against DNA damage and apoptosis in stressed CHO (Chinese-hamster ovary) cells, previously transfected with the full or 3′-UTR-lacking MT gene .
To date, little is known about the nature of the signal(s) that direct(s) mRNA localization. In some instances, signal recognition by the localization machinery seems to involve structural features of the mRNA (e.g. bicoid), whereas in others the primary sequence appears as the critical determinant (Vg1 and β-actin) [3,7,8]. We investigated the nature of the PNLS (perinuclear localization signal) present in the rat MT-1 3′-UTR. Our approach uses a deletion/mutagenesis analysis of this region combined with localization assays (immunocytochemistry and in situ hybridization), determination of its structure by enzymic/chemical probing and phylogenetic sequence comparison.
As a first step towards the characterization of the MT PNLS, we performed a sequence alignment of MT-1 3′-UTRs from various mammalian species to determine the existence of potential conserved motifs (Figure 1). More closely related species (rat, hamster and mouse) present two regions of high homology (75 and 84% homology in regions 26–45 and 71–101 respectively). This degree of conservation decreases substantially when two additional species (human and pig) are used for this analysis (∼50% homology for both regions). Interestingly, the rat sequence possesses a repetition of a CACC motif between positions 67 and 76. It has been shown previously that a tandem repeat of a ACACCC sequence present in the zipcode of the chicken β-actin mRNA is essential for the transcript localization, whereas CAC-containing motifs are required for the targeting of Vg1 and VegT mRNAs to the vegetal pole of Xenopus oocytes [3,9]. However, the CACC repeat present in the rat MT-1 3′-UTR is not conserved in the other species. Only one such motif appears in the mouse and human 3′-UTRs, whereas none is present in the hamster sequence and two CAC motifs occur in the pig 3′-UTR close to the same position. It is noteworthy that a comparable situation has been described for the β-actin mRNA. Although only one ACACCC motif is present in the human β-actin mRNA, this messenger was still found to be targeted to the periphery of chicken fibroblasts . Therefore the MT-1 PNLS does not seem to rely on the repeat sequence by itself but, instead, probably requires particular structural features of the 3′-UTR possibly involving one or two CAC motif(s).
Comparison of the 3′-UTRs of various
Gene constructs in which the rat MT-1 coding region is linked to different sequences from the rat MT-1 3′-UTR (deleted or mutated, particularly nt 26–45 and the CACC repeat) were thus prepared by cloning into the pcDNA4/Hismax-TOPO vector based on PCR or using site-directed mutagenesis. These constructs were transfected into CHO cells and stable transfectants selected by growth in G418-containing medium. The effects of the different deletions/mutations were then screened by studying MT localization by immunocytochemistry. In parallel, the constructs are being used to generate RNA transcripts for protein-binding assays (UV cross-linking) and enzymic/chemical cleavage. Radiolabelled in vitro transcribed RNAs, in the presence or absence of an excess of unlabelled sequence-specific RNA competitors, are incubated with protein extracts from CHO cells. After cross-linking of the ribonucleoprotein complexe(s) and removal of the non-covalently bound RNA by digestion with the RNase A, the protein(s) are resolved on a 12.5% SDS/polyacrylamide gel and visualized by autoradiography. For RNA probing, in vitro transcribed RNAs are 5′-end-labelled with [γ-32P]ATP and submitted to cleavage by specific RNases (V1, T1, T2, A) or lead. Cleavage products are separated on a 10% sequencing gel, visualized by autoradiography and analysed for the determination of single- and double-stranded cuts. So far, our results suggest that the CACC repeat is important for MT-1 mRNA localization but also that it requires an appropriate structural context to function efficiently. In particular, we showed that this repeat sequence by itself confers some localization properties when inserted in the 3′-UTR of a heterologous mRNA (D. Nury, M. Levadoux-Martin, H. Chabanon and J.E. Hesketh, unpublished work). Currently, the critical features of the localization signal are under investigation.
Lipids, Rafts and Traffic: A Focus Topic at BioScience2004, held at SECC Glasgow, U.K., 18–22 July 2004. Edited by G. Banting (Bristol, U.K.), N. Bulleid (Manchester, U.K.), C. Connolly (Dundee, U.K.), S. High (Manchester, U.K.) and K. Okkenhaug (Babraham Institute, Cambridge, U.K.)