A Cellular Cofactor for the Constitutive Transport Element of Type D Retrovirus

See allHide authors and affiliations

Science  30 May 1997:
Vol. 276, Issue 5317, pp. 1412-1415
DOI: 10.1126/science.276.5317.1412


A human nuclear protein that specifically interacts with the constitutive transport element (CTE) of simian retrovirus was identified as adenosine 5′-triphosphate–dependent RNA helicase A. This protein could bind to functional CTE but not to inactive CTE mutants. The interaction of helicase A with CTE was distinct from previously described helicase activity of this protein. Helicase A shuttled from the nucleus to the cytoplasm in the presence of a transcription inhibitor or in cells transiently overexpressing CTE-containing RNA. In vivo colocalization of helicase A and CTE was observed in experiments that combined in situ hybridization and immunostaining. These results suggest that helicase A plays a role in the nuclear export of CTE-containing RNA.

Normal cellular mRNAs are exported from the nucleus as fully spliced RNA. However, retroviruses need to export partially spliced or unspliced RNA to the cytoplasm, both as templates for protein synthesis and as genomic RNA to be packaged in progeny virions. The complex retroviruses, including the human pathogenic retroviruses human T cell leukemia virus and human immunodeficiency virus (HIV), mediate this process through trans-acting proteins (Rex and Rev, respectively) that bind to their cognate RNA targets (RxRE and RRE, respectively). The simple retroviruses do not encode such trans-acting proteins but rather use cis-acting sequences that presumably interact directly with cellular nuclear export proteins. One example is the CTE of type D retroviruses, which is able to functionally replace Rev or RRE in subgenomic constructs and infectious HIV clones (1, 2). Here we report the identification and characterization of a candidate for the cellular cofactor of CTE.

Wild-type and a nonfunctional mutant CTE (ΔCTE) were biotinylated and used in RNA selection experiments (3, 4). A 140-kD protein was reproducibly selected by wild-type but not by mutant CTE (Fig.1A, lanes 1 and 2). Use of a panel of CTE deletion mutants revealed a complete correlation of CTE function [as determined in a chloramphenicol acetyltransferase (CAT) reporter assay (4)] and the ability to select the 140-kD protein (Table1). We separated the 140-kD protein from other cellular proteins by SDS–polyacrylamide gel electrophoresis (SDS-PAGE) and excised it from a polyvinylidene difluoride (PVDF) membrane after blotting. We subjected tryptic peptides to microsequencing as described (5). Three independently sequenced internal peptides matched with three different parts of one protein in a BLAST homology search: human adenosine triphosphate (ATP)–dependent RNA helicase A, a DEAD-box helicase that belongs to the DEAH-box subfamily (6). We confirmed the identity of the 140-kD protein by immunoblotting with specific antibodies against human helicase A (Fig.1A, lane 3). Direct interaction between CTE and helicase A was further shown by gel-shift experiments with purified protein. Complexes were formed between helicase A protein and CTE in the sense but not in the antisense orientation (Fig. 1B, lanes 1 to 4). There were two apparent complexes, suggesting that the protein might have oligomerized. Helicase A associates with and translocates along single-stranded tails of double-stranded RNA templates in vitro in the presence of ATP (7). We carried out RNA-protein interaction assays under the conditions used by Lee and Hurwitz (7). CTE remained bound to helicase A in the presence of ATP, suggesting that this interaction is not associated with the helicase activity of the protein (Fig. 1B, lanes 5 to 7).

Figure 1

In vitro interaction between CTE and human helicase A (23). (A) A 140-kD nuclear protein was specifically selected by wild-type (lane 1) but not mutant (lane 2) CTE. The protein was specifically recognized by a polyclonal antibody to human ATP-dependent RNA helicase A (lane 3). (B) Gel-shift assay of CTE with purified helicase A protein. Helicase A formed complexes with CTE but not with CTE antisense probe (lanes 1 to 4). Hydrolysis of ATP did not result in dissociation of the protein (lanes 5 to 7). The ATP chase experiments were done as described by Lee and Hurwitz (7).

Table 1

Specificity of interaction between CTE and helicase A. SRV, simian retrovirus; MPMV, Mason-Pfizer monkey virus.

View this table:

The exact mechanism of CTE-mediated nuclear export of unspliced mRNA is not clear, but because it can substitute for Rev or RRE function, the two pathways are likely to use a similar underlying mechanism. Rev shuttles between the nucleus and cytoplasm and contains a nuclear export signal in its activation domain (8, 9). Rev also appears in the cytoplasm with the overexpression of RRE-containing viral RNA (10). However, helicase A has not been observed previously to be a shuttling protein. We confirmed that the normal localization of helicase A is predominantly nuclear (Fig. 2A, a). In cells transiently transfected with pDM138CTE, a reporter construct in which CTE promotes CAT gene expression by exporting unspliced RNA containing the CAT sequences in an intron, the distribution of helicase A changed markedly, allowing detection of the protein at all stages of nuclear-cytoplasmic trafficking (Fig. 2A, b to e). In contrast, cells transfected with pDM128, which contains RRE instead of CTE in the intron (Fig. 2A, f); pDM138, which lacks CTE; and pDM138ΔCTE, which contains a nonfunctional CTE, all retained helicase A predominantly in the nuclei (11). The cytoplasmic enrichment of helicase A in the presence of CTE is not likely due to a block of nuclear import of newly synthesized protein because treatment with the translation inhibitor cyclohexamide had no effect on cytoplasmic accumulation. Several other constructs that express CTE in different contexts (Fig. 2B) also induced cytoplasmic accumulation of helicase A (12), suggesting that CTE acts as a constitutive, position-independent signal that triggers nuclear export of helicase A.

Figure 2

Helicase A shuttles between the cytoplasm and the nucleus in the presence of CTE-containing RNA (24). (A) (a) Helicase A concentrated in the nucleus in normal cells. (b to e) In cells transfected with pDM138CTE, helicase A was detected at all stages of nuclear-cytoplasmic trafficking. (f) Transfection of pDM128 did not change the normal nuclear distribution of the protein. (B) Three constructs were used in this study that expressed CTE in different contexts. pDMCTE is the same as pDM138CTE except that the CAT gene is deleted, leaving CTE alone in the intron; pDCTE is a construct in which the DNA fragment for CTE was cloned into pCDNA3; and CTE is now expressed as exon RNA.

Because not all the cells were transfected in the transient assay system used, we investigated whether cytoplasmic distribution of helicase A occurs in the same cells expressing CTE. Double labeling of the RNA and protein in CTE-transfected cells was carried out by in situ hybridization and immunostaining, and their subcellular localization was visualized by confocal microscopy. All cells that were negative for CTE staining also had no significant staining of helicase A in the cytoplasm, suggesting that the observed cytoplasmic accumulation of helicase A was a direct result of CTE overexpression. On the other hand, in cells expressing high levels of CTE in the cytoplasm, helicase A had relocated into the cytoplasm and associated with CTE (Fig.3). The in vivo colocalization of CTE and helicase A was observed at various times after transfection and with different detection modules used to visualize RNA and protein in the cells (13). There were, however, some cells that expressed CTE in the cytoplasm while helicase A was still enriched in the nuclei, suggesting a dynamic process of RNA and protein shuttling in the transfected cells.

Figure 3

Helicase A colocalizes with CTE in vivo. HeLa cells were fixed 20 hours after transfection. CTE and helicase A were detected by combining in situ hybridization (CTE) and immunostaining (helicase A) techniques. (A) Transmitted light image. (B) Immunodetection of helicase A. (C) Detection of CTE-containing RNA by in situ hybridization with digoxigenin-labeled antisense RNA probe. Both positively transfected and untransfected cells are shown. (D) Double labeling reveals colocalization of helicase A and CTE in the transfected cells.

Rev and heterogeneous ribonucleoprotein (hnRNP) A1 are two well-characterized RNA-binding proteins involved in RNA transport. They accumulate in the cytoplasm with inhibition of transcription (8,14), suggesting that ongoing transcription is required for the rapid shuttling of these proteins back to the nuclei. To determine whether helicase A has intrinsic shuttling capability, we treated HeLa cells for 3 hours with various transcriptional inhibitors and stained the cells with antibodies to helicase A. Redistribution of helicase A to the cytoplasm was observed when the cells were treated with actinomycin D (5 μg/ml) and kept at 37°C (Fig. 4A, lower right). Nuclear localization of the protein was retained in untreated cells or in treated cells kept at 4°C after addition of actinomycin D (Fig. 4A, upper), suggesting that nuclear export of the protein is energy-dependent. Treatment with actinomycin D did not make the nuclei generally leaky because the distribution of another nuclear protein, splicing factor SC35, was not altered (Fig. 4B). Again, treatment of the cells with cyclohexamide (50 μg/ml) had no effect on the cytoplasmic distribution of helicase A in the presence of actinomycin D, suggesting that the cytoplasmic staining was a result of export of preexisting protein. 5,6-Dichlorobenzimidazole riboside (DRB; 100 μM), which inhibited RNA polymerase II (Pol II) and promoted cytoplasmic accumulation of Rev and hnRNP A1 in the cytoplasm (8,13, 15), did not alter nuclear localization of helicase A. In addition, actinomycin D (0.04 μg/ml) alone, which is sufficient to inhibit RNA Pol I transcription (16), or in combination with 100 μM DRB, also had no effect (17) (Fig. 4A, lower left). Thus, these data in combination with the observation that actinomycin D (5 μg/ml) inhibits transcription by all three RNA polymerases suggest that inhibition of Pol III transcription is necessary for cytoplasmic accumulation of helicase A.

Figure 4

Inhibition of transcription accumulates helicase A in the cytoplasm. Cyclohexamide (50 μg/ml) was added to all cell cultures 3.5 hours before the cells were fixed. Helicase A was detected by indirect immunofluorescence labeling (20). (A) Inhibitors of RNA synthesis by various RNA polymerases had different effects on the subcellular distribution of helicase A. Inhibition of Pol I and Pol II transcription [actinomycin D (0.04 μg/ml) and 100 μM DRB] did not cause significant change of nuclear localization of the protein, whereas actinomycin D (5 μg/ml) accumulated helicase A in the cytoplasm at 37°C. (B) Actinomycin D (5 μg/ml) treatment did not alter nuclear localization of splicing factor SC35.

In summary, we have identified RNA helicase A as an inherent shuttling protein that interacts with CTE in vitro and associates with CTE in its trafficking from the nucleus to the cytoplasm in vivo. The fact that there was perfect correlation between CTE function and helicase A binding for a number of CTE mutants strongly suggests that helicase A plays a role in CTE function (Table 1). The minimal functional CTE has been mapped to a 173-nucleotide fragment and is composed of two essential loops that are predicted to be the binding sites of the cellular cofactors (18). Two of our nonfunctional mutants (M21 and M14) contain deletions that destroy formation of loop A without affecting loop B (Table 1), whereas ΔCTE contains mutations that are predicted by computer analysis to unravel the stem structure between the loops and thus affect formation of both loops. The failure of helicase A to bind to all three mutants suggests that loop A of CTE is required for the interaction and, as in the case for CTE function, loop B alone is not sufficient for binding to helicase A even though loops A and B have identical sequences (19). We have previously identified two additional nuclear proteins that bind to wild-type CTE but not to ΔCTE (4). Therefore, other proteins may be involved in CTE-mediated gene expression.

Several DEAH-box RNA helicases participate in mRNA transport. A 144-kD human helicase–like protein, human RNA helicase 1, facilitated export of cellular mRNA by releasing the RNA from the spliceosome after splicing was completed (20). A yeast gene capable of encoding a 120-kD putative RNA helicase protein rescued the mutant phenotype of a conditional yeast mutant defective in export of polyadenylate RNA (21). Given the RNA-binding and shuttling ability of helicase A, it is conceivable that it participates in certain cellular RNA export pathways. Simian retroviruses have likely tapped into this pathway and use helicase A as a cofactor in nuclear export of CTE-containing RNA by means of a specific RNA-protein interaction. It will be interesting to determine whether helicase A also plays a role in the replication cycle of other retroviruses, including complex retroviruses such as HIV.


View Abstract

Stay Connected to Science

Navigate This Article