Role of Loc1p in assembly and reorganization of nuclear ASH1 messenger ribonucleoprotein particles in yeast
Edited by James E. Dahlberg, University of Wisconsin Medical School, Madison, WI, and approved November 19, 2013 (received for review August 12, 2013)
Significance
Cytoplasmic mRNA localization is preceded by the formation of nuclear pre-mRNPs. To date the requirement of the nuclear mRNP assembly is not well understood. We used ASH1 mRNA localization from budding yeast to understand the mechanisms of nuclear priming for cytoplasmic mRNA transport. We found that the nuclear factor Loc1p is required to stably and specifically tether the bona fide ASH1 RNA-binding protein She2p to RNA in the nucleus, offering an explanation for the requirement of Loc1p for ASH1 mRNA localization. Because Loc1p is not part of the cytoplasmic ASH1 mRNP, it must be removed before mRNA transport. We show that the cytoplasmic transport factor She3p displaces Loc1p from the ASH1 mRNP, allowing for the maturation of the localizing ASH1 mRNP.
Abstract
Directional transport of mRNA is a universal feature in eukaryotes, requiring the assembly of motor-dependent RNA-transport particles. The cytoplasmic transport of mRNAs is preceded by the nuclear assembly of pre-messenger ribonucleoprotein particles (mRNPs). In budding yeast, the asymmetric synthesis of HO 1 (ASH1) pre-mRNP originates already cotranscriptionally and passes through the nucleolus before its nuclear export. The nucleolar localization of ASH1 mRNA protein 1 (Loc1p) is required for efficient ASH1 mRNA localization. Immunoprecipitation experiments have revealed that Loc1p forms cocomplexes with other components of the ASH1 transport complex. However, it remains unclear how Loc1p is recruited into this mRNP and why Loc1p is important for ASH1 mRNA localization. Here we demonstrate that Loc1p undergoes a direct and specific interaction with the ASH1 mRNA-binding Swi5p-dependent HO expression protein 2 (She2p). This cocomplex shows higher affinity and specificity for RNA bearing localization elements than the individual proteins. It also stabilizes the otherwise transient binding of She2p to ASH1 mRNA, suggesting that cooperative mRNA binding of Loc1p with She2p is the required nuclear function of Loc1p for ASH1 mRNA localization. After nuclear export, myosin-bound She3p joins the ASH1 mRNP to form a highly specific cocomplex with She2p and ASH1 mRNA. Because Loc1p is found only in the nucleus, it must be removed from the complex directly before or after export. In vitro and in vivo experiments indicate that the synergistic interaction of She2p and She3p displaces Loc1p from the ASH1 complex, allowing free Loc1p to rapidly reenter the nucle(ol)us. Together these findings suggest an ordered process of nuclear assembly and reorganization for the maturation of localizing ASH1 mRNPs.
Sign up for PNAS alerts.
Get alerts for new articles, or get an alert when an article is cited.
Messenger RNA localization is a universal feature of eukaryotes (1–3). By complementing transcriptional control (4), it fulfills a variety of functions, including the establishment of cell polarity and specialization of subcellular regions. In recent years, the directional transport of asymmetric synthesis of HO 1 (ASH1) mRNA in budding yeast has emerged as a particularly well-suited model to study mechanistic principles of RNA localization. Here, comparably few proteins participate in the directional transport of ASH1 mRNA and about 30 other transcripts (5, 6).
Chromatin-immunoprecipitaton experiments revealed that the dedicated RNA-binding Swi5p-dependent HO expression protein 2 (She2p) binds already cotranscriptionally to nascent ASH1 mRNA (7, 8). Two additional RNA-binding proteins, pumilio-homology domain family protein 6 (Puf6p) and heterogeneous nuclear RNP K-like protein 1 (Khd1p), are also present in the nucleus, bind to ASH1 mRNA, and act in the cytoplasm as translational repressors during ASH1 transport (9–12). A fourth nuclear factor, termed localization of ASH1 mRNA protein 1 (Loc1p), has been implicated in the assembly of nuclear pre-messenger ribonucleoprotein particles (mRNPs). Like Puf6p, Loc1p is a nuclear protein (13) with an enrichment in the nucleolus (14, 15) and participates in the assembly of the large ribosomal subunit (16–18).
The composition of the nuclear ASH1 mRNP was previously analyzed by coimmunoprecipitation experiments. By using She2p as bait, Puf6p, Loc1p, and ASH1 mRNA were copurified (11, 19). The translational inhibitor Khd1p, which interacts with the 5′ region of ASH1 mRNA (12), was absent from such purifications (11, 20). These data suggest the formation of a nuclear mRNP consisting of ASH1 mRNA, She2p, Loc1p, and Puf6p.
Genomic deletion of LOC1, similar to PUF6 (9, 11), leads to less efficient ASH1 mRNA localization (13). Although Loc1p does not shuttle between nucleus and cytoplasm (13), loc1∆ cells show an up-regulation of cytoplasmic ASH1 mRNA translation (13, 21), indicating that this protein is required for translational repression during localization. ASH1 mRNA contains four cis-acting localization elements, which are also termed zip-code elements. These elements are specifically recognized by the transport complex. In the 5′–3′ order of the ASH1 mRNA the zip-code elements are termed E1, E2A, E2B, and E3. Puf6p- (11) and Loc1p-dependent translational repression (21) seems to be mediated only by the E3 zip-code element. Thus, Loc1p might be required to remodel the nuclear precomplex for the assembly of translational repressors at the E3 element.
In a point-mutated She2p that is unable to enter the nucleus, no Puf6p and very little Loc1p are found associated with ASH1 mRNA (19). Thus, She2p appears to play an important role for the assembly and/or recruitment of the nuclear precomplex including Puf6p and Loc1p. Vice versa, preventing nuclear export of She2p with a temperature-sensitive mex67-5 mutant results in accumulation of She2p and ASH1 mRNA in the nucleolus (14). Together, these observations indicate that the nucle(ol)ar fate is an essential step for the assembly of mature, translationally silent ASH1 mRNA-transport complexes.
ASH1 mRNA is exported into the cytoplasm along with She2p and Puf6p, whereas Loc1p remains located in the nucleolus (13, 14, 22). Once in the cytoplasm, the ASH1 precomplex interacts with another subcomplex consisting of the type V myosin motor myosin 4 (Myo4p) and the RNA-binding protein She3p. The nuclear and cytoplasmic precomplexes form the mature transport complex via the direct interaction between She2p and She3p and their synergistic binding to the zip-code elements of the ASH1 mRNA (8, 23–28).
Although Puf6p is associated with the localizing ASH1 mRNP, in vitro binding studies suggested that Puf6p and She2p interact only indirectly via their joint binding to the ASH1 mRNA (8). For Loc1p and She2p or for Loc1p and Puf6p no such information is available. It also remains unclear whether these factors depend on each other for RNA binding, whether they compete for ASH1 mRNA interaction, and whether they all participate in the same mRNP or form different subcomplexes. Furthermore, no mechanism has been described that explains how Loc1p is removed from the complex prior to or during nuclear export.
Here we report the characterization of the nuclear ASH1 mRNP assembly. We found that Loc1p and She2p directly and specifically interact with each other. This cocomplex of She2p and Loc1p shows a significantly better affinity, stability, and specificity toward ASH1 mRNA. Because the RNA binding of She2p alone is very transient, this cooperative interaction of Loc1p and She2p with ASH1 mRNA appears to be essential for stably tethering She2p to nuclear ASH1 mRNA. We do not observe a similar recruitment of Puf6p to the mRNP by Loc1p, She2p, or She3p. Thus, the importance of both factors for Puf6p binding to ASH1 mRNA is likely more indirect. Because Loc1p is absent in the cytoplasm, it has to be removed from the mRNP during complex maturation. We found that She3p outcompetes Loc1p from its cocomplex with She2p and RNA, which results in the synergistic cocomplex of She2p, She3p, and ASH1 mRNA as well as free Loc1p. This outcompetition offers a mechanistic explanation for the removal of Loc1p from the mRNP. Consistently, genomic deletion of SHE3 in vivo perturbs the nucleolar localization of Loc1p.
Results
Loc1p Binds Directly and Specifically to She2p.
Loc1p was shown to copurify with She2p from yeast extracts (19). To test whether this interaction is direct, we performed pull-down experiments with recombinant maltose binding protein (MBP)-tagged Loc1p and She2p. To exclude indirect, RNA-mediated interactions, proteins were purified under stringent conditions and successful removal of nucleic acid contaminations was verified before binding experiments (Materials and Methods; for Loc1p, see Fig. S1A). She2p was retained on amylose resin in the presence of Loc1p but not with MBP alone or in the absence of an MBP-tagged protein (Fig. 1A). Thus, the pull-down experiments suggest a direct and quantitative interaction between Loc1p and She2p.
Fig. 1.
Loc1p is involved not only in mRNA localization but also in ribosome biogenesis. It has been proposed that the effect of Loc1p on ASH1 mRNA localization might be only an indirect effect of impaired ribosome function (21). On the other hand, Loc1p has been found in complex with ASH1 mRNA, She2p, and Puf6p (13, 19), suggesting a direct and functional link. We performed additional in vitro pull-down experiments in the absence of RNA and found that Loc1p interacts selectively with She2p, but not with Puf6p, His-tagged She3p (She3p-H6), or GST (Fig. 1 A and B). Thus, Loc1p binding to She2p is selective, which rather indicates a direct role of Loc1p in ASH1 mRNA localization.
Loc1p and She2p Form a Joint Cocomplex on the E3 Zip-Code Element of the ASH1 mRNA.
Yeast three-hybrid experiments were used to identify Loc1p as an ASH1 E1 and E3 zip-code–binding protein (13). Binding to the E3 zip-code element was further confirmed in electrophoretic mobility shift assays (EMSAs), using either yeast extracts or bacterial extracts with recombinant Loc1p. In EMSAs with purified Loc1p we show that this protein indeed binds directly to the E3 zip-code element (Fig. 1C). The appearance of a higher molecular weight band further indicates that a second binding event occurs at a higher protein concentration.
Next, we asked whether the protein interaction between Loc1p and She2p is altered by binding to zip-code RNA. For better purification of large amounts of zip-code RNA, we used the tRNA-scaffold technique (29). We fused the 118-bases-long ASH1 E3 element to a tRNA (ASH1 E3-118 tRNA), expressed it in bacteria, and purified it. Using amylose resin and MBP-Loc1p, we pulled down She2p as well as ASH1-E3-118 tRNA (Fig. 1D), suggesting the formation of a ternary complex of these components.
Cooperative Binding of She2p and Loc1p to ASH1 mRNA.
The pull-down of She2p with RNA and Loc1p indicates a joint complex, but fails to yield direct information about their RNA binding within the complex. Furthermore, pull-down experiments are performed at micromolar concentrations that greatly exceed the nanomolar concentrations of these proteins in the cell. To directly assess RNA binding under more physiological conditions, we performed UV cross-linking experiments with radioactively labeled ASH1 E3 RNA. Because UV cross-linking efficiency occurs preferentially with certain amino acids, such experiments yield semiquantitative information on affinities. As demonstrated before (8), at a concentration of 250 nM, She2p alone showed a well-detectable cross-linking product (Fig. 2A, lane 3). Also Loc1p alone could be cross-linked to the RNA (Fig. 2A, lanes 4–6). However, this cross-linking gave a clearly detectable signal only at 1 µM protein concentration. The combination of Loc1p and She2p led to a strong increase of the Loc1p–RNA cross-link (Fig. 2A, lanes 7–9). This observation indicates that She2p is required for efficient recruitment of Loc1p to ASH1 mRNA. Because the She2p cross-link also showed an increased intensity in the presence of Loc1p, these results suggest cooperativity in RNA binding.
Fig. 2.
To validate these observations with a more quantitative assay, we performed EMSAs with constant, low concentrations of MBP-Loc1p and increasing concentrations of She2p (Fig. 2B). As previously shown (8, 30), in our EMSAs She2p alone does not form a protein–RNA complex stable enough to yield a significant mobility shift even at 1 µM She2p concentration (Fig. 2B). In contrast, Loc1p alone gives rise to a distinct band shift at a concentration of 125 nM. When both proteins are combined, a much stronger band shift was observed (Fig. 2B). Supershift experiments with increasing antibody concentrations further confirmed that She2p was indeed present in this shifted band (Fig. S2A, Left). Mobility shift assays with gradient gels that better separate larger complexes show the additional slow-migrating band expected for the ternary complex of She2p, Loc1p, and RNA (Fig. S2A, Right). These findings, together with the UV cross-linking experiments, indicate a cooperative recruitment of She2p and Loc1p onto the ASH1 mRNA and a stabilization of She2p on the RNA by Loc1p.
Cocomplex of Loc1p and She2p Forms Only with Zip-Code RNA.
Next, we wanted to assess the specificity of the ternary complex formed by She2p, Loc1p, and RNA for the ASH1 E3 zip-code element. First, we performed pull-down experiments with glutathione resin and GST-tagged She2p as bait. In these experiments with ASH1 E3-51 RNA and tRNA competitor but in the absence of Loc1p, we observed copurification of ASH1 E3-51 RNA but not of the tRNA competitor (Fig. 3A). Addition of Loc1p resulted in the pull-down of ASH1 E3-51 RNA and Loc1p, but not of tRNA competitor. Interestingly, in the experiment with She2p and Loc1p, a higher molecular weight band was observed in the urea PAGE (Fig. 3A, asterisk). Because this band disappeared after phenol-chloroform extraction (Fig. 3A, Right), the ternary complex of She2p, Loc1p, and ASH1 E3-51 RNA was likely not completely disrupted by the presence of urea in the gel. Such a high molecular weight band was not observed in absence of Loc1p (Fig. 3A, Left).
Fig. 3.
When we repeated these pull-down experiments with the HIV TAR-57 tRNA as an unspecific, zip-code–lacking control instead of an ASH1 zip-code RNA, we failed to observe copurification of Loc1p or RNA (Fig. 3B). Because She2p and Loc1p undergo a direct protein–protein interaction in the absence of RNA (Fig. 1A), this result suggests different binding modes of this complex for specific and unspecific RNAs. The disruption of the protein–protein interaction between She2p and Loc1p in the presence of HIV TAR-57 RNA (Fig. 3B) suggests competitive binding events, instead of the previously observed cooperative binding of She2p and Loc1p to zip-code RNA (Figs. 1D and 2 A and B). Consistent with this interpretation is the observation that in an amylose-resin pull-down, MBP-Loc1p copurified She2p only in the presence of zip-code–containing RNAs (Fig. S2B).
To confirm this difference in specific and unspecific RNA-binding modes by a different approach, we performed again UV cross-linking experiments. Radioactively labeled ASH1 E3-51 tRNA efficiently cross-linked with She2p, Loc1p, and as a positive control She3p (Fig. 3C). Also in the cocomplex, Loc1p and She2p both efficiently cross-linked with the zip-code–containing RNA, as observed before (Fig. 2A). In contrast, the control HIV TAR-57 tRNA failed to show major cross-linking bands with She2p, Loc1p, or their cocomplex. Only She3p, which served as a control for unspecific binding (8), showed a clear cross-link (Fig. 3C). Also in competition experiments we observed a preference for ASH1 E3 zip-code element (Fig. S2C). In summary, these experiments suggest a clear preference in binding for the ASH1 E3-51 RNA over the control RNA.
Cocomplex of She2p and Loc1p Shows Selectively Increased Affinity and Stability for Zip-Code RNA.
For a quantitative assessment of the observed difference in RNA binding, we performed double-filter–binding experiments (31). In these experiments, a nitrocellulose membrane retains protein-bound radioactively labeled RNA, whereas a second nylon membrane retains free, unbound RNA. We used the radioactivity retained on the first membrane to determine equilibrium dissociation constants (Kd) and radioactivity on the second, nylon membrane to estimate the fraction of unbound RNA. Because filter-binding assays involve washing steps after initial binding, the latter can be used to assess complex stabilities.
Binding of the She2p-Loc1p cocomplex to ASH1 E3-51 RNA showed an improved affinity (Kd = 18 nM) compared with that of the individual proteins (Kd for She2p = 82 nM and for Loc1p = 52 nM; Fig. 3 D and E and Fig. S3). The cocomplex of She2p and Loc1p had an over 40-fold lower affinity for the HIV TAR-57 tRNA (Kd = 769 nM). Interestingly, the She2p-Loc1p cocomplex had a 5-fold lower Kd for the control RNA compared with Loc1p alone (769 nM vs. 152 nM; Fig. 3 D and E and Fig. S3). The opposite effect was observed in binding experiments with the ASH1 RNA (18 nM vs. 52 nM; Fig. 3 D and E and Fig. S3), further supporting the notion that the modes of binding for zip-code–containing and unspecific RNAs might be different.
Also the amounts of protein-free RNA retained on the nylon membrane indicated a difference in binding. The cocomplex of Loc1p and She2p as well as Loc1p alone retained almost no free RNA on the nylon membrane, indicating that the vast majority of RNA was stably associated with the protein-bound fraction at the nitrocellulose membrane (Fig. 3D). In contrast, She2p alone was unable to retain a comparable amount of RNA on the first membrane (Fig. 3D). These results indicate that the interaction of She2p with RNA is stabilized by Loc1p. This stabilization effect was quantified by comparing signal intensities on the nylon membrane with the stable ternary complex of Loc1p, She2p, and RNA defined as 0% retention. Whereas we did not observe a difference in complex stability between Loc1p alone and the ternary complex, She2p showed a larger fraction of free RNA bound to the nylon membrane (Fig. 3F). The latter confirms the previously described (8, 30) low complex stability of the She2p–RNA interaction.
In Pull-Down Experiments Loc1p and She2p but Not Loc1p and Puf6p Form a Cocomplex with RNA.
Like Loc1p, Puf6p coimmunoprecipitates with She2p and ASH1 mRNA from yeast extracts (19) and localizes to the nucleolus (14, 15). However, in pull-down experiments Loc1p and Puf6p fail to interact (Fig. 1B). To study Puf6p interactions in the presence of RNA we used the E3-118 zip-code element, which contains two conserved Puf6p binding sites (see Table S3) (11). Initially we tested whether Puf6p can be pulled down by MBP-Loc1p or H6-Loc1p in the presence of ASH1 E3-118 RNA. As before, we ensured that both proteins reproducibly showed monodisperse behavior in size-exclusion chromatography and normal RNA-binding capacity (Fig. 1C and Fig. S1 B and C). Puf6p was not retained by MBP-Loc1p on amylose beads (Fig. 4A), although Puf6p alone efficiently binds to E3 RNA (Fig. S1C) (8). Even by using pull-down conditions more closely resembling a physiological situation (250 nM protein) and different Loc1p and Puf6p tags, we were not able to detect an interaction between Loc1p and Puf6p (Fig. S4 A and B). Amylose pull-down experiments with MBP-Loc1p, She2p, Puf6p, and ASH1 E3 mRNA showed similar results. All factors were retained on amylose resin, except for Puf6p (Fig. 4B). These data indicate that the proteins Loc1p and She2p form a stable cocomplex with ASH1 mRNA, in which Puf6p does not participate.
Fig. 4.
UV Cross-Linking Reveals That Puf6p Is Not Recruited by Other Factors to the ASH1 E3 Zip-Code RNA.
To directly assess RNA-binding events at near-physiologic concentrations, we performed UV cross-linking experiments with radioactively labeled ASH1 E3-118 RNA. Puf6p alone showed only a faint RNA cross-link (Fig. 4C, lane 4; for a longer exposure of this UV cross-link, see Fig. S4D). Also the presence of other proteins, such as She2p (Fig. S4C, lanes 7 and 8), She2p and She3p (Fig. 4C, lanes 6 and 7, and Fig. S4D), or She2p and Loc1p (Fig. 4C, lanes 8 and 9, and Fig. S4D) did not enhance the intensity of the Puf6p RNA cross-link. Vice versa, Puf6p did not alter the cross-linking of Loc1p, She2p, and She3p or combinations thereof. This indicates that Puf6p is not recruited by any of these factors to the ASH1 E3 zip-code RNA.
The Cocomplex of She2p and She3p Displaces Loc1p from the RNA.
She2p and ASH1 mRNA, but not Loc1p mRNA, is exported into the cytoplasm (13, 22). Therefore, a mechanism must be in place that allows Loc1p to be removed from the mRNA before or during its export into the cytoplasm. Because Loc1p contains multiple potential nuclear-localization signals (Fig. S5), one mechanism could be that a strong nuclear reimport of Loc1p is sufficient to strip off this protein from the RNA. Alternatively, Loc1p could be actively displaced from the RNA by another component of the transport complex that joins the mRNP during or after nuclear export. A potential candidate for such an active displacement is the myosin adapter and RNA-binding protein She3p. In the cytoplasm, it forms a high-affinity cocomplex with She2p and ASH1 mRNA. Also Loc1p binds to both She2p (Fig. 1A) and ASH1 mRNA (Figs. 1 C and D and 2). Thus, for an active displacement, She3p would need to outcompete Loc1p from both binding partners.
We first tested whether the addition of She3p to Loc1p, She2p, and ASH1 E3 element changes the association of Loc1p with the RNA. To detect RNA-binding events at near-physiologic concentrations, we used again UV cross-linking of radioactively labeled ASH1 E3 zip-code RNA. Whereas we recapitulated the Loc1p recruitment to the ASH1 RNA by She2p (Fig. 5A, compare lanes 3–5 with lanes 6 and 7), the combined presence of She2p and She3p almost completely abolished RNA cross-linking by Loc1p (Fig. 5A, lanes 8 and 9). This finding indicates that at nanomolar, near-physiological concentrations the She2p-She3p complex displaces Loc1p from zip-code–containing RNA.
Fig. 5.
Formation of the Ternary Complex of She2p, She3p, and ASH1 RNA Disrupts the Interaction between She2p and Loc1p.
This UV cross-linking experiment does not rule out the possibility that Loc1p remains directly bound to She2p, even when it is displaced from the RNA. Hence, we tested in pull-down experiments whether She3p alters the interaction of She2p and Loc1p in the absence and presence of RNA. Indeed, we found that MBP-Loc1p fails to interact with She2p and She3p in the absence of the E3 element (Fig. 5B). We also performed these experiments with the ASH1 E3 element and an unspecific tRNA competitor. Also under these experimental conditions Loc1p was not able to pull down She2p or She3p (Fig. 5C). We conclude that She3p also outcompetes the protein interaction of Loc1p with She2p.
Mutational Studies with She2p Indicate Overlapping Binding Sites of She3p and Loc1p.
Because we observed an outcompetition of Loc1p by She3p, these proteins most likely share overlapping binding sites on She2p. To test this assumption we performed pull-down experiments with mutant versions of She2p. It was previously shown that the She2p mutation of leucine 130 into tyrosine [She2p (L130Y)] impairs its oligomeric state and its in vivo interaction with She3p (30). A recent study further showed that the deletion of the very C terminus of She2p [She2p (∆C)] has no strong effect on in vitro binding to She3p, whereas the deletion of its protruding helix [She2p (∆hE)] abolishes the She3p interaction (8). We first recapitulated these results (Fig. S6A) and then tested these mutant versions of She2p for Loc1p binding. We found that Loc1p shows reduced binding to She2p (L130Y), whereas She2p (∆C) is bound like wild-type She2p (Fig. S6B). The observation that mutations in She2p affect Loc1p and She3p binding is consistent with the assumption that both proteins share an overlapping interaction surface on She2p.
Deletion of Cytoplasmic She3p Alters Nuclear Distribution of Loc1p.
The observed displacement of Loc1p from the complex with She2p and ASH1 E3 RNA by She3p (Fig. 5) suggests that this mechanism might also contribute in vivo to the removal of Loc1p from the ASH1 mRNP. She3p has been previously shown to be exclusively cytoplasmic (8). First, we wanted to elucidate whether Loc1p is found at or close to the nuclear periphery, where an interaction with She3p could potentially occur. We performed double staining of GFP-tagged Loc1p and mCherry-tagged nuclear pore marker Nup120. As previously shown by Du et al. (14), Loc1p is clearly found in the nucleolus (Fig. 6A). The overlay with Nup120-mCherry shows that this nucleolar localization of Loc1p also reaches parts of the nuclear periphery.
Fig. 6.
If She3p is indeed necessary for removing Loc1p from ASH1 mRNA during nuclear export, the Loc1p distribution should be altered in a she3∆ strain. Because the majority of Loc1p is likely involved in ribosome biogenesis inside the nucleolus, alterations in Loc1p distribution are expected to be relatively mild. We therefore overexpressed ASH1 mRNA from a 2µ plasmid and focused on cells in anaphase when ASH1 expression peaks (32). In these cells with wild-type She3p expression, GFP-Loc1p distribution is clearly limited to the nuclear crescent that is typical for the yeast nucleolus (Fig. 6A) (14). In contrast, cells that are deficient for She3p show a dispersed GFP-Loc1p staining in the nucleoplasm (Fig. 6B). This observation indicates that cytoplasmic She3p indeed affects Loc1p distribution in the nucle(ol)us. Occasionally, we also observed a weak cytoplasmic staining of GFP-Loc1p in she3∆ cells. However, this is an infrequent event and weaker than the nucleoplasmic staining in mutant cells.
Discussion
To date little is known about the molecular mechanisms involved in the early nuclear stages of localizing mRNPs and how nucleolar transitions are organized. Nucleolar trespassing of pre-mRNPs before cytoplasmic mRNA localization is not limited to the SHE complex in yeast. For instance, mammalian Staufen2 also passes through the nucleolus before its nuclear export (33). Here we present a study in which we systematically assessed the interplay between transport factors that associate with ASH1 mRNA already in the nucleus. This allows us to provide a mechanistic understanding of these early events.
It was recently shown that the interaction of Loc1p and Puf6p with ASH1 mRNA is impaired when She2p is prevented from entering the nucleus (19). This suggests interdependence between these protein–RNA interactions. Furthermore, Loc1p remains in the nucleus but is nevertheless important for cytoplasmic mRNA localization. Together, these findings indicate that the nuclear assembly of the pre-mRNP is important for cytoplasmic mRNA localization and translational repression.
A puzzling observation of previous studies was that the in vitro interaction between She2p and ASH1 mRNA is very transient (8, 30). However, because She2p is cotranscriptionally recruited to ASH1 mRNA (7, 8) and exported in an RNA-dependent manner to the cytoplasm (22), the formation of a stable complex in the nucleus is expected to occur. In the cytoplasm the interaction of ASH1 mRNA and She2p is stabilized by the formation of a highly specific cocomplex with cytoplasmic She3p (8). An unresolved question is therefore whether and how She2p binding is modulated to form a stable complex with ASH1 mRNA in the nucleus and whether such a nuclear complex would be specific for ASH1 mRNA.
We found that She2p specifically binds to Loc1p even in the absence of RNA. We further observed a Loc1p-dependent stabilization of She2p binding to RNA. Loc1p alone shows only a modest preference for ASH1 E3 RNA over HIV TAR-57 RNA (about 3-fold stronger Kd; Fig. 3E). However, Loc1p in cocomplex with She2p shows an about 40-fold better Kd for ASH1 E3 RNA over HIV TAR-57 RNA, clearly indicating a specific selection of zip-code RNA.
Our pull-down experiments further indicated that zip-code–containing RNA is bound in a different mode by She2p and Loc1p than a control RNA (Fig. 3 A and B and Fig. S2B). In fact, ternary complexes formed only with zip-code RNA. Together with UV cross-linking and filter-binding assays, these findings suggest that Loc1p is required to stably and specifically associate She2p with nuclear ASH1 mRNA. The cooperative, mutual recruitment and stabilization of Loc1p and She2p on the RNA are in agreement with the previous observation that in a nuclear localization-deficient She2p strain almost no ASH1 mRNA was immunoprecipitated with TAP-tagged Loc1p (19).
An interesting question arising from these insights is where this stable ternary complex forms in the nucleus. For She2p a cotranscriptional recruitment to the ASH1 mRNA has already been described (Fig. 7, Upper, A), whereas the exact location of Loc1p joining the early mRNP is unknown. One possible scenario is that Loc1p encounters She2p at or in the nucleolus, where Loc1p is enriched. Loc1p has a theoretical isoelectric point of 10.3, which is very similar to that of the vast majority of ribosomal proteins. Although nucleolar localization signals are difficult to predict (34), the considerable number of putative signals in Loc1p (Fig. S5) indicates that this protein is very likely localizing directly to the nucleolus without the need of a nucleolar hub protein. In contrast, She2p does not possess detectable nucleolar localization signals. Therefore, the interaction of Loc1p might not only stabilize She2p on the ASH1 mRNA, but also piggyback She2p into the nucleolus (Fig. 7, Upper, A and B). Such indirect nucleolar recruitment mechanisms have been reported already for other proteins without intrinsic nucleolar localization signal, such as Nucleolin and Nucleophosmin (35).
Fig. 7.
Because She2p associates with ASH1 mRNA already cotranscriptionally, it is also tempting to speculate that Loc1p binds to She2p already after or even before its association with the ASH1 transcription complex in the nucleoplasm (Fig. 7, Upper, A). The recruitment of Loc1p could help to stably tether She2p to the ASH1 mRNA already at the onset of mRNP biogenesis. This complex would remain associated until joining of She3p and formation of the She2p-She3p complex in the cytoplasm displaces Loc1p from the mRNP (see below).
In a mutant yeast strain with cytoplasmically trapped She2p, the nuclear interaction between Puf6p and ASH1 mRNA is strongly impaired (19). In our in vitro assays we failed to observe a direct interaction of Puf6p with She2p or Loc1p or a recruitment of Puf6p by these factors to the ASH1 mRNA. Hence, there must be another explanation for this dependence of Puf6p on She2p. Like for puf6Δ, the LOC1 deletion strain exhibits defects in ASH1 translational repression (13, 14). Furthermore, Puf6p shows a strong nucleolar enrichment (15) similar to that of Loc1p. Therefore, Loc1p-dependent piggybacking of She2p and ASH1 mRNA into the nucleolus would be an obvious mechanism to bring the mRNP in the vicinity of Puf6p and to allow its loading into the complex (Fig. 7, Upper, C). Such a scenario would be consistent not only with our in vitro observations but also with the dependence of Puf6p binding to ASH1 mRNA on She2p in vivo (19). Interestingly, Loc1p- as well as Puf6p-dependent translational repression is mediated by the E3 zip-code element (11, 21), indicating that this element fulfills a special function in ASH1 mRNA.
After nucleolar trespassing, the ASH1 mRNP is exported into the cytoplasm and Loc1p has to be removed from the ASH1 mRNP. We observed that the cocomplex of She2p and She3p is able to displace Loc1p from the ASH1 mRNP. She3p is strictly cytoplasmic and constitutively tethered to the cytoskeleton (8). Furthermore, with a molecular weight of 47 kDa She3p is too large to freely diffuse through the nuclear pore (36). Also the nuclear protein Loc1p was reported not to shuttle between nucleus and cytoplasm, as judged by shuttling and heterokaryon assays (13). Because these techniques are not very sensitive, minor subfractions might have remained undetected. Based on the observation that the deletion of SHE3 does affect the nuclear localization and occasionally leads to cytoplasmic accumulation of Loc1p, it is conceivable that an outcompetition of Loc1p by the She2p-She3p cocomplex indeed occurs at or close to the nuclear pore (Fig. 7, Upper, D).
From the data presented here, we conclude that a stable and specific ternary complex is formed within the nucleus consisting of She2p, Loc1p, and ASH1 mRNA. This pre-mRNP trespasses the nucleolus in a Loc1p-dependent manner. Because we do not observe a direct interaction of Puf6p with the pre-mRNP constituents or recruitment to the ASH1 mRNA, it seems likely that Puf6p joins the mRNP independently, probably within the nucleolus. After nucleolar trespassing, the mRNP is exported into the cytoplasm, where the mature transport complex together with She3p and Myo4p is formed. Our data implicate that She3p outcompetes Loc1p for the interaction with She2p and ASH1 mRNA during or shortly after nuclear export. The reported assembly line of the nuclear mRNP ensures that all required factors associate and dissociate at the right time to escort the ASH1 mRNA from its site of transcription all of the way to its localization at the bud tip.
Materials and Methods
Protein Expression and Purification.
Full-length Loc1p was expressed with an N-terminal MBP-tag or H6-tag in Escherichia coli strain BL21 (DE3). After induction with 0.25 mM isopropyl β-D-1-thiogalactopyranoside in the logarithmic growth phase cells were cultured for 16 h at 18 °C and harvested. Cells were sonicated at 4 °C in lysis buffer (20 mM Hepes, pH 7.4, 500 mM KCl, 0.5 mM EDTA, complete protease inhibitor or 20 mM K-Phosphate, pH 7.5, 500 mM KCl, 50 mM Imidazole, and complete protease inhibitor). After centrifugation the clarified supernatant was applied to an amylose/Ni Sepharose resin column and washed with high-salt buffer (1 M KCl) to remove nucleic acids and subsequently with low-salt buffer (150 mM KCl). The protein was eluted with 20 mM maltose or 750 mM imidazole in low-salt buffer and further purified using ion exchange chromatography (SP FF; GE Healthcare), heparin affinity chromatography, and size exclusion chromatography (e.g., Superdex S200; GE Healthcare).
Full-length She2p and Puf6p were expressed and purified as GST-fusion proteins in E. coli strain BL21 (DE3) and purified as previously described (8, 37).
His6-tagged She3p was coexpressed with She2p in insect cells (High Five) using the bac-to-bac system (Invitrogen) and purified with affinity, ion exchange, and size exclusion chromatography as previously described (8). Plasmids and primers used in this study are listed in Tables S1 and S2.
To remove RNA contaminations from each of the above-mentioned proteins, a high-salt washing step was carried out during the affinity chromatography and validated via measurement of the OD254 to OD280 ratio. Proteins with an OD254 to OD280 ratio below 0.6 were specified as RNA/DNA free.
RNA Preparation.
RNA was produced by in vitro transcription (E3-118), tRNA scaffold expression, and purification (E3-51 tRNA, E3-118 tRNA, and TAR-57 tRNA) or total chemical synthesis (E3-51; Dharmacon) (38). For in vitro transcription the 118-nt DNA fragment of the ASH1 gene (plasmid pRS405-ASH1 in Table S1) was PCR amplified using primers with the T7 promotor sequence at the 5′ end. This fragment was used as a template for in vitro transcription (MEGAshortscript Kit; Ambion). The transcribed RNA was purified as described in the manufacturer’s instructions, using DNase digestion, phenolic extraction, and ethanol precipitation. The integrity of the RNA was confirmed by agarose gel electrophoresis.
tRNAMet fusion constructs (E3-51 tRNA, E3-118 tRNA, and TAR-57 tRNA) were overexpressed in E. coli JM 101 cells, using 2x typtone-yeast media. The RNA was phenol-chloroform extracted, ethanol precipitated, and further purified by ion exchange chromatography (DAEA and MonoQ columns), as described in ref. 29. Pure fractions were identified by 8% (vol/vol) Tris/borate/EDTA (TBE) Urea PAGE, pooled, and ethanol precipitated.
For UV cross-linking and EMSA experiments in vitro-transcribed E3-118 zip-code RNA, synthesized E3-51, or recombinant E3-51 tRNA, E3-118 tRNA, or HIV TAR-57 tRNA constructs were radioactively labeled using [γ32P]ATP and T4 polynucleotide kinase. Free nucleotides were separated from the RNA, using either Sepharose spin columns (NucAway; Ambion, E3-51 element) or phenol-chloroform extraction (E3-118, E3-51 tRNA, E3-118 tRNA, and HIV TAR-57 tRNA). For interaction studies with Puf6p the E3-118 or E3-118 tRNA zip-code RNA was used, because it includes the corresponding Puf6p binding element (11). For interaction studies without Puf6p the E3-51 zip-code element was used because of better recovery rates after isotope labeling. The sequences of RNAs used in this study are listed in Table S3.
EMSA.
In a total volume of 20 µL 5 nM radioactively labeled RNA, 100 µg/mL tRNA competitor, and the indicated protein concentration were mixed in HNMD buffer (20 mM Hepes, pH 7.8, 200 mM NaCl, 2 mM MgCl2, 2 mM DTT) with 4% glycerol. After 25 min incubation at 25 °C the RNA protein complexes were resolved by native TBE PAGE (6% polyacrylamide, 1× TBE running buffer). Gels were incubated for 15 min in fixing solution [10% (vol/vol) acetic acid, 30% (vol/vol) methanol] and vacuum dried. Gels were analyzed with radiograph films. For supershift EMSAs monoclonal She2p antibody (30) was included in the reaction. DTT was omitted from the HNMD reaction buffer for supershift EMSAs. For detection of cocomplexes 4–20% TBE gradient gels were used.
UV Cross-Linking Experiments.
In a total volume of 20 µL 5 nM radiolabeled RNA, 100 µg/mL tRNA, and the indicated protein concentrations were mixed in HNMD buffer. After 25 min incubation at 25 °C the samples were subjected to 2.25 J UV radiation over 15 min on ice. The distance between the samples and the UV lamp was ∼8 cm. After adding SDS loading dye, samples were incubated for 5 min at 90 °C and separated on 10% SDS/PAGE. Gels were fixed, dried, and analyzed with radiograph films.
In Vitro Pull-Down Experiments.
In a total volume of 100 µL 7.5 or 10 µM protein, 4 µM E3-51/E3-118 tRNA, and 5 mg/mL yeast tRNA competitor (unless stated otherwise) as indicated were mixed in appropriate pull-down buffer (20 mM Hepes, pH 7.8, 150 mM or 200 mM NaCl, 2 mM MgCl2, 2 mM DTT or 20 mM K phosphate, pH 7.5, 150 mM or 200 mM KCl, 2 mM MgCl2, and 30 mM imidazole). Samples were centrifuged for 10 min and supernatant was incubated for 30 min at 4 °C on a rotating wheel with 50 µL amylose resin (MBP-Loc1p), GST-Sepharose (GST-Puf6p), or Ni-NTA (She3p-H6). Binding reactions were washed four times with 200 µL and once with 50 µL pull-down buffer. Bound proteins were eluted with maltose (amylose resin), glutathione (GST-Sepharose), or imidazole (Ni-NTA). On SDS/PAGE 10% of the input, 20% of the last wash step, and 20% of the elution were analyzed by Coomassie blue staining. RNA was analyzed on a 1.5% agarose gel or an urea PAGE with GelRed DNA stain, according to ref. 38.
Filter-Binding Assays.
In a total volume of 80 µL the indicated protein concentrations and 0.5 nM of radiolabeled RNA (ASH1 E3-51 tRNA, HIV TAR-57 tRNA) were incubated for 25 min in filter-binding buffer (10 mM Hepes, pH 7.4, 100 mM NaCl, 1 mM DTT, and 5 mM MgCl2) supplemented with 30 µg/mL yeast tRNA competitor. Samples were applied to nitrocellulose and nylon membranes, using a Dot Blot Aparatus (BioRad), and washed twice with filter-binding buffer. Membranes were air dried and analyzed by phospho-imaging. Quantification of signal intensities was carried out using the Dot Blot Analyzer macro within the ImageJ program and further analyzed with the program Origin 8.6.
Generation of Yeast Strains.
Strain RJY4358 (LOC1-GFP::HIS3MX6) was generated by one-step tagging of W303a according to ref. 39. To construct strain RJY4359 (LOC1-GFP::HIS3MX6 she3∆::URA3), RJY4358 was transformed with a DNA construct containing the URA3 gene flanked by sequences of the 5′ and 3′ regions of SHE3, which was obtained by PCR amplification from the she3∆::URA3 locus of strain RJY90 (40). Before imaging, both strains were transformed with plasmid pRJ88 (YEplac181-ASH1), which allows overexpression of ASH1 mRNA during mitosis, and plasmid pRS314-Nup120-mCherry, which allows detection of nuclear pores. Genotypes of used yeast strains as well as plasmids for transformation are listed in Tables S1 and S4.
In Vivo Imaging.
For imaging, a single yeast colony from a fresh plate grown at 30 °C was inoculated in 1 mL of synthetic complete (SC) medium with 2% glucose but lacking leucine and tryptophane. Cells were grown at 30 °C for 4–4.5 h to enrich for mitotic cells. Cells were collected by short spin and resuspended in 100 µL of SC medium. One microliter of cell suspension was placed on a multiwell slide coated with agarose before microscopic observation. For each mitotic cell pair, Z-stacks containing 50 images at 200-nm distance were acquired for GFP, mCherry, and DIC. Due to the weaker signal of Nup120-mCherry, mCherry images were deconvoluted before mounting. Image processing and mounting were performed with AxioVision software version 4.8 (Zeiss). GFP and mCherry overlay pictures were generated using Adobe Photoshop CS3. Essentially the mCherry Layer was duplicated and inserted in the GFP file. By using exclusion in the layer option an overlay was generated and one representative cell was magnified.
Acknowledgments
We thank Sigrun Jaklin and Vera Roman for their support. We thank Ed C. Hurt for generously providing the plasmid pRS 314-Nup120-mCherry. This work was supported by the Deutsche Forschungsgemeinschaft (A.N., R.-P.J., and D.N.), the Helmholtz Association (Grant NG-VH142 to D.N.), the Bayerisch-Französisches Hochschulzentrum (A.N. and D.N.), and the Boehringer-Ingelheim Fonds (M.M.).
Supporting Information
Supporting Information (PDF)
Supporting Information
- Download
- 1015.91 KB
References
1
CE Holt, SL Bullock, Subcellular mRNA localization in animal cells and why it matters. Science 326, 1212–1216 (2009).
2
D St Johnston, Moving messages: The intracellular localization of mRNAs. Nat Rev Mol Cell Biol 6, 363–375 (2005).
3
RP Jansen, D Niessing, Assembly of mRNA-protein complexes for directional mRNA transport in eukaryotes—an overview. Curr Protein Pept Sci 13, 284–293 (2012).
4
D Niessing, et al., A cascade of transcriptional control leading to axis determination in Drosophila. J Cell Physiol 173, 162–167 (1997).
5
RG Heym, D Niessing, Principles of mRNA transport in yeast. Cell Mol Life Sci 69, 1843–1853 (2012).
6
N Paquin, P Chartrand, Local regulation of mRNA translation: New insights from the bud. Trends Cell Biol 18, 105–111 (2008).
7
Z Shen, A St-Denis, P Chartrand, Cotranscriptional recruitment of She2p by RNA pol II elongation factor Spt4-Spt5/DSIF promotes mRNA localization to the yeast bud. Genes Dev 24, 1914–1926 (2010).
8
M Müller, et al., A cytoplasmic complex mediates specific mRNA recognition and localization in yeast. PLoS Biol 9, e1000611 (2011).
9
Y Deng, RH Singer, W Gu, Translation of ASH1 mRNA is repressed by Puf6p-Fun12p/eIF5B interaction and released by CK2 phosphorylation. Genes Dev 22, 1037–1050 (2008).
10
N Paquin, et al., Local activation of yeast ASH1 mRNA translation through phosphorylation of Khd1p by the casein kinase Yck1p. Mol Cell 26, 795–809 (2007).
11
W Gu, Y Deng, D Zenklusen, RH Singer, A new yeast PUF family protein, Puf6p, represses ASH1 mRNA translation and is required for its localization. Genes Dev 18, 1452–1465 (2004).
12
K Irie, et al., The Khd1 protein, which has three KH RNA-binding motifs, is required for proper localization of ASH1 mRNA in yeast. EMBO J 21, 1158–1167 (2002).
13
RM Long, et al., An exclusively nuclear RNA-binding protein affects asymmetric localization of ASH1 mRNA and Ash1p in yeast. J Cell Biol 153, 307–318 (2001).
14
TG Du, et al., Nuclear transit of the RNA-binding protein She2 is required for translational control of localized ASH1 mRNA. EMBO Rep 9, 781–787 (2008).
15
WK Huh, et al., Global analysis of protein localization in budding yeast. Nature 425, 686–691 (2003).
16
CR Urbinati, GB Gonsalvez, JP Aris, RM Long, Loc1p is required for efficient assembly and nuclear export of the 60S ribosomal subunit. Mol Genet Genomics 276, 369–377 (2006).
17
Y Ho, et al., Systematic identification of protein complexes in Saccharomyces cerevisiae by mass spectrometry. Nature 415, 180–183 (2002).
18
ML De Marchis, A Giorgi, ME Schininà, I Bozzoni, A Fatica, Rrp15p, a novel component of pre-ribosomal particles required for 60S ribosome subunit maturation. RNA 11, 495–502 (2005).
19
Z Shen, N Paquin, A Forget, P Chartrand, Nuclear shuttling of She2p couples ASH1 mRNA localization to its translational repression by recruiting Loc1p and Puf6p. Mol Biol Cell 20, 2265–2275 (2009).
20
M Oeffinger, et al., Comprehensive analysis of diverse ribonucleoprotein complexes. Nat Methods 4, 951–956 (2007).
21
S Komili, NG Farny, FP Roth, PA Silver, Functional specificity among ribosomal proteins regulates gene expression. Cell 131, 557–571 (2007).
22
C Kruse, et al., Ribonucleoprotein-dependent localization of the yeast class V myosin Myo4p. J Cell Biol 159, 971–982 (2002).
23
F Böhl, C Kruse, A Frank, D Ferring, RP Jansen, She2p, a novel RNA-binding protein tethers ASH1 mRNA to the Myo4p myosin motor via She3p. EMBO J 19, 5514–5524 (2000).
24
A Heuck, et al., Monomeric myosin V uses two binding regions for the assembly of stable translocation complexes. Proc Natl Acad Sci USA 104, 19778–19783 (2007).
25
A Heuck, et al., The structure of the Myo4p globular tail and its function in ASH1 mRNA localization. J Cell Biol 189, 497–510 (2010).
26
AR Hodges, EB Krementsova, KM Trybus, She3p binds to the rod of yeast myosin V and prevents it from dimerizing, forming a single-headed motor complex. J Biol Chem 283, 6906–6914 (2008).
27
RM Long, W Gu, E Lorimer, RH Singer, P Chartrand, She2p is a novel RNA-binding protein that recruits the Myo4p-She3p complex to ASH1 mRNA. EMBO J 19, 6592–6601 (2000).
28
PA Takizawa, RD Vale, The myosin motor, Myo4p, binds Ash1 mRNA via the adapter protein, She3p. Proc Natl Acad Sci USA 97, 5273–5278 (2000).
29
L Ponchon, G Beauvais, S Nonin-Lecomte, F Dardel, A generic protocol for the expression and purification of recombinant RNA in Escherichia coli using a tRNA scaffold. Nat Protoc 4, 947–959 (2009).
30
M Müller, et al., Formation of She2p tetramers is required for mRNA binding, mRNP assembly, and localization. RNA 15, 2002–2012 (2009).
31
I Wong, TM Lohman, A double-filter method for nitrocellulose-filter binding: Application to protein-nucleic acid interactions. Proc Natl Acad Sci USA 90, 5428–5432 (1993).
32
N Bobola, RP Jansen, TH Shin, K Nasmyth, Asymmetric accumulation of Ash1p in postanaphase nuclei depends on a myosin and restricts yeast mating-type switching to mother cells. Cell 84, 699–709 (1996).
33
P Macchi, et al., The brain-specific double-stranded RNA-binding protein Staufen2: Nucleolar accumulation and isoform-specific exportin-5-dependent export. J Biol Chem 279, 31440–31444 (2004).
34
V Sirri, S Urcuqui-Inchima, P Roussel, D Hernandez-Verdun, Nucleolus: The fascinating nuclear body. Histochem Cell Biol 129, 13–31 (2008).
35
E Emmott, JA Hiscox, Nucleolar targeting: The hub of the matter. EMBO Rep 10, 231–238 (2009).
36
D Görlich, U Kutay, Transport between the cell nucleus and the cytoplasm. Annu Rev Cell Dev Biol 15, 607–660 (1999).
37
D Niessing, S Hüttelmaier, D Zenklusen, RH Singer, SK Burley, She2p is a novel RNA binding protein with a basic helical hairpin motif. Cell 119, 491–502 (2004).
38
FT Edelmann, A Niedner, D Niessing, Production of pure and functional RNA for in vitro reconstitution experiments. Methods, 10.1016/j.ymeth.2013.08.034. (2013).
39
C Janke, et al., A versatile toolbox for PCR-based tagging of yeast genes: New fluorescent proteins, more markers and promoter substitution cassettes. Yeast 21, 947–962 (2004).
40
RP Jansen, C Dowzer, C Michaelis, M Galova, K Nasmyth, Mother cell-specific HO expression in budding yeast depends on the unconventional myosin myo4p and other cytoplasmic proteins. Cell 84, 687–697 (1996).
Information & Authors
Information
Published in
Classifications
Submission history
Published online: December 9, 2013
Published in issue: December 24, 2013
Keywords
Acknowledgments
We thank Sigrun Jaklin and Vera Roman for their support. We thank Ed C. Hurt for generously providing the plasmid pRS 314-Nup120-mCherry. This work was supported by the Deutsche Forschungsgemeinschaft (A.N., R.-P.J., and D.N.), the Helmholtz Association (Grant NG-VH142 to D.N.), the Bayerisch-Französisches Hochschulzentrum (A.N. and D.N.), and the Boehringer-Ingelheim Fonds (M.M.).
Notes
This article is a PNAS Direct Submission.
Authors
Competing Interests
The authors declare no conflict of interest.
Metrics & Citations
Metrics
Altmetrics
Citations
Cite this article
Role of Loc1p in assembly and reorganization of nuclear ASH1 messenger ribonucleoprotein particles in yeast, Proc. Natl. Acad. Sci. U.S.A.
110 (52) E5049-E5058,
https://doi.org/10.1073/pnas.1315289111
(2013).
Copied!
Copying failed.
Export the article citation data by selecting a format from the list below and clicking Export.
Cited by
Loading...
View Options
View options
PDF format
Download this article as a PDF file
DOWNLOAD PDFLogin options
Check if you have access through your login credentials or your institution to get full access on this article.
Personal login Institutional LoginRecommend to a librarian
Recommend PNAS to a LibrarianPurchase options
Purchase this article to access the full text.