In this work a new Rev-interacting protein has been identified using the yeast two- hybrid system. The protein was termed Risp (Rev-interacting shuttle protein), because it shuttles between the nuclear and the cytoplasmic compartments.
The Risp gene is widely expressed in human cells and conserved among various species, most probably as part of a larger gene. High identity (99%) with the C- terminal part of a large brain cDNA clone for KIAA0592 protein (Nagase et al., 1998) has been found, whereas no high sequence homology has been associated with any protein with known function. However, a weak and partial homology appeared with several RNA-/DNA-binding and shuttle proteins. This might indicate that Risp protein - or the larger protein containing it - could be a member of a new family of nucleocytoplasmic shuttle proteins with RNA-/DNA-binding function.
In HeLa cells Risp-GFP localizes in both nuclear and cytoplasmic compartments, but clearly accumulates in the cytoplasm, indicating the presence of a strong nuclear export signal (NES). The identification of Risp NES was confirmed by localization analysis of different transfected segments of Risp and by cytoplasmic accumulation of nuclear microinjected BSA-fusion protein containing a peptide from the C-terminal part of Risp. The inhibition of the Crm1-mediated nuclear export, induced by the presence of leptomycin B, showed that the nuclear export of Risp is Crm1-dependent.
As it is rich in leucine and isoleucine residues, the Risp NES might be related with other leucine-rich NESs. To determine if these hydrophobic amino acids are also crucial for the nuclear export of Risp, the export activity of Risp NES mutant peptides and Risp mutant protein with alterations in one or more of the leucine and isoleucine positions has to be analyzed.
Risp has been shown to share a specific motif with other Rev-interacting proteins. This consists of 12 amino acids [P (AELNR) (KHILST) (KCRST) T N (PD) F (GLQ) (LST) (LIN) (EAGQS)] termed RIP motif. This is the first indication for the presence of a specific and common motif shared between Rev-interacting factors, including Rev itself. A region containing a sequence matching to the consensus motif was identified in every Rev-interacting protein. Nevertheless, a particular group of cellular proteins, mostly involved in the Rev/RRE nuclear export, showed to be highly specific versus the motif. No correlation was found between the motif and the Rev-binding ability, in fact the region of Risp harboring the RIP motif was neither essential nor sufficient for the Rev-binding in the yeast two hybrid system.
Therefore, the role and the relevance of the consensus motif present in Risp and in the other Rev-interacting factors should be in detail investigated in future studies. For example mutational analysis and deletions of the domain in the Risp protein will be established to analyze whether the function of Rev could be modified.
Although, it is not yet clear which is the minimal sequence part required for the Risp- Rev interaction, we can assume that the Risp-binding region in Rev should be located in a region overlapping the basic domain deputed to RNA binding/NLS (34- 50aa) and the second multimerization region (52-60aa). In the basic domain already other Rev-interacting proteins such as B23, p32, and importin β are supposed to bind Rev, whereas other proteins involved directly and indirectly to the Rev nuclear export, such as Crm1, eIF-5A, hRab/rip, are known to bind Rev in its activation domain (Table 2.3 and Fig. 5.1).
Fig. 5.1 Rev and cellular interacting factors. It is known that Rev is composed of several domains harboring distinct binding sites for cellular proteins. Here an updated scheme of the primary Rev structure is proposed. The leucine-rich activation domain in the C-terminus (aa 73-84) contains a nucleocytoplasmic shuttling signal (NSS) mediating the Rev nuclear export and nuclear import. In the nucleus, Rev NSS binds directly together with RanGTP to the export receptor Crm1 to form a stable complex Rev/Crm1/RanGTP able to dock and translocate the nuclear pore complex. Several nucleoporins (hRip/Rab, yRip1p1/Nup42, Nup98, Nup153, Nup214, NLP-1) and eIF-5A, identified as Rev NES-interacting proteins, are involved to mediate the nuclear export of Rev/Crm1/RanGTP complex. The interacting import protein (?) mediating the Rev import via the NSS domain is not yet discovered. The arginine-rich stretch located toward the N-terminus (aa 34-50) serves as a RNA-binding domain as well as a nuclear/nucleolar localization signal (NLS/NOS). This is, in turn, closely flanked on both sides by residues that mediate Rev multimerization (aa 12-29 and 52-60). Importin β and p32, have been proposed to bind directly with the Rev basic domain. In the cytoplasm, the import receptor importin β recognizes the REV NLS and thus mediates the Rev nuclear entry. The interaction of Rev with the protein p32, normally associated with the cellular splicing factor SF2/ASF, may participate in the removal of splicing factors from intron-containing RNAs and permit a nuclear accumulation of RNAs substrate for Rev-mediated nuclear export. A binding between the nucleolar phosphoprotein B23 and Rev has been also observed. However, the accumulation of both proteins in the nucleolus may reflect the ability of the interaction, while no B23 role in supporting Rev function has been demonstrated. Finally, in this study, we showed the presence of a new Rev-interacting protein, Risp, that binds Rev in its RNA-binding and the II multimerization domain.
Activation domain
Nucleocytoplasmic shuttling signal (NSS)
RNA-binding domain & Nuclear/nucleolar localization (NLS/NOS) Oligomerization I Oligomerization II eIF-5A B23 N S S NLS/NOS RanGTP Risp Nucleoporins
Preliminary experiments suggested that Risp, as a Rev-interacting protein, is able to inhibit Rev-trans-activation, while Risp does not interfere with Tat in a Tat-trans- activation assay. In fact, the over-expression of Risp-GFP was able to reduce the production of the Rev-dependent structural viral protein p24gag up to 70% and fairly specifically the SLIIB-CAT expression in a Tat-Rev/SLIIB-binding assay. Therefore, it is possible that the binding of Risp to Rev in its RNA-binding and multimerization domain may compete with the binding of Rev to its SLIIB RNA target and to other Rev molecules inhibiting thereby Rev function.
Since Risp most probably is only a part of a larger protein, it is also important to isolate its full-length cDNA and repeat the Rev-trans-activation studies. Currently, Chris Bickel, in the group of Dr. Brack-Werner, is trying to enlarge to the 5’ end of the full-length cDNA. It will be of high interest to determine whether the original protein containing Risp is able to interfere with the Rev function and thus to limit HIV-1 replication.
Moreover, since both Risp and Rev are using the same export receptor Crm1, an additional role of Risp could also be proposed at the level of the nucleocytoplasmic translocation of Rev. The direct binding of Risp to Crm1 could compete with the Rev-Crm1binding and therefore reduce Rev nuclear export. On the other hand, Risp bound to Rev in its RNA-binding and multimerization domain may accelerate the nuclear export of Rev, providing also its NES, and therefore facilitating the association with Crm1and RanGTP to form an active export complex in the absence of HIV-1 RNA.
Of course, most of the implications reported here are still speculative, requiring future experimental work to obtain a better understanding of the function(s) of Risp and its interactions with Rev.