DDX belongs to the DEAD box family of
DDX3 belongs to the DEAD-box family of proteins, ATP-dependent RNA helicases characterized by the presence of a highly conserved helicase core domain . This catalytic core is composed of two RecA-like domains containing motifs involved in ATP binding/hydrolysis, RNA binding and helicase activity , . DEAD-box proteins also function as RNA clamps for the assembly of large macromolecular complexes . Interestingly, DEAD-box proteins interact with RNA through the sugar backbone and thus, binding is rather unspecific , . Despite the high conservation of the helicase core domain, the mode of RNA binding and the enzymatic activity, the 37 DEAD-box proteins described so far in humans are expected to perform specific, non-redundant, functions within the cell , . As such, it has been suggested that regions flanking the catalytic core, which are variable both in length and amino TC-I 15 receptor composition, could be involved in conferring substrate specificity through the direct recruitment of protein co-factors or RNA partners .
Materials and methods
Discussion DDX3 is a host factor essential for HIV-1 replication as evidenced by several reports including different genome-wide siRNA screenings , . During viral replication, DDX3 was shown to promote nuclear export and translation of the unspliced mRNA. To do so, the RNA helicase interacts with: i) the viral mRNA, ii) the karyopherin CRM1 and iii) translation initiation factors eIF4GI and PABPC1 , . However, whether these two activities of DDX3 during the post-transcriptional control of HIV-1 gene expression are interconnected or independent has never been evaluated. In this work we first show that the N-terminal domain of DDX3 (amino acids 1–182) regulates the functions of this RNA helicase during HIV-1 unspliced mRNA translation (Fig. 1, Fig. 2). Given the fact that N-ter inhibits unspliced mRNA translation in a CRM1-dependent but 5′-UTR-independent manner, our data suggest that N-ter could be involved in a step of gene expression occurring after nuclear export but before the 40S ribosomal subunit is recruited onto the viral 5′-UTR (Fig. 3). In agreement with this idea, we observed that the regulatory effect of the N-terminal domain of DDX3 was independent of Rev but rather relied on nuclear export by CRM1 (Fig. 4, Fig. 5). Consistent with this data, a recent docking study aimed to characterize the human DDX3–CRM1–NES–RanGTP multimeric complex proposed an interesting mode of binding between regions present in the helicase domain of DDX3 and a region surrounding to the NES-binding pocket of CRM1 . Such a binding was stimulated by the binding of RanGTP to CRM1 but was independent of the NES-containing cargo . Unfortunately, there is no structural data of the full-length DDX3 and thus, the abovementioned simulations were performed in the absence of the N-terminal domain that we showed as an important determinant for interaction with CRM1 (Fig. 5). Indeed, our data suggest that DDX3 is incorporated into the CRM1-nuclear export viral mRNP through the N-terminal domain to promote viral mRNA translation. As such, the isolated N-terminal domain of DDX3 interferes with the recruitment of the full-length protein into this complex. Of note, our previous and present data indicate that DDX3 promotes HIV-1 gene expression independently of the viral protein Rev (Fig. 4) . These observations suggest that DDX3 is recruited onto the exported viral mRNP to play its functions most probably once Rev has been released from the complex. Interestingly, it was recently reported that the formation of a Ded1–CRM1–RanGTP trimeric complex in yeast resulted in a reduced RNA-stimulated ATPase activity of Ded1 . This phenomenon was dependent on the NES present in Ded1 and occurred concomitantly with the increase in the Km for the RNA substrate suggesting that the NES-dependent interaction between Ded1 and CRM1–RanGTP may modulate substrate specificity of the RNA helicase . Although the interaction between human DDX3 and CRM1 seems not to rely on a NES , it would be interesting to determine whether such a modulation of the enzymatic activity of DDX3 also occurs in the presence of the RRE/Rev–CRM1–RanGTP complex. Thus, it is tempting to speculate that DDX3 (and Ded1) could be acting on the CRM1-dependent pathway analogously to Dbp5 in the NXF1-dependent nuclear export pathway .