Host And Viral Determinants For MxB Restriction Of HIV-1 Infection

Host organisms have developed a wide array of innate immunity proteins to prevent or mitigate viral infection. Expression of these proteins is often induced by interferon signaling, as their potent activities can come at the cost of host cell or organism viability. The study of retroviruses in particular has uncovered many examples of innate antiviral proteins, including apolipoprotein B mRNA editing enzyme, catalytic polypeptide-like (APOBEC) 3G [1], tripartite motif (TRIM) 5α [2] and TrimCyp [3], SAM domain and HD domain-containing protein (SAMHD) 1 [4],[5], and bone marrow stromal cell antigen (BST) 2/Tetherin [6],[7]. Recently, Myxovirus resistance protein 2 (MxB) was discovered to exhibit potent antiviral activity against HIV-1 infection [8]-[11].

In humans, MxB is one of two members of a family of dynamin-like large GTPases. Mx family proteins are found in almost all vertebrates, demonstrating their evolutionary importance for host organisms [12]. The X-ray crystal structure of human MxA showed that these proteins can be divided into three structurally folded domains: a globular GTPase domain, a largely C-terminal alpha helical stalk domain, and a series of alpha helices found in sequences adjacent to these domains which fold in the protein tertiary structure to form the bundle signaling element (BSE) [13]. Mx proteins also possess relatively unstructured N-terminal regions of varying length. These proteins readily multimerize into higher-order structures, with MxA known to form large ring-like assemblies [14],[15].

Human MxA inhibits infection by a large number of negative-stranded RNA viruses, though it also counteracts other viral families [12]. The ability of MxA to inhibit a diverse set of viruses while specifically targeting different proteins across these families is atypical amongst innate immune proteins [16], suggesting an antiviral mechanism distinct from those of previously discovered restriction factors. Despite the breadth of known antiviral activities of MxA, little is known about the antiviral potential of MxB. Until recently, MxB was not thought to possess antiviral activity, which suggested that its purpose was solely to function during cellular nucleo-cytoplasmic transport [17]. In contrast to MxA, which is cytoplasmic, MxB localizes to the nuclear rim [18], which may play into its distinct pattern of antiviral activity.

A large-scale screen assessing the activities of interferon stimulated gene products against a panel of viruses first uncovered an antiviral activity of human MxB against HIV-1 [8]. More recently, a series of papers found MxB to be a key component of the interferon-mediated response against the early steps of HIV-1 infection [9]-[11]. The reverse transcription complex wherein HIV-1 RNA is reverse transcribed into double stranded linear DNA carries a fraction of the virion capsid (CA) protein [19],[20], and CA mutations can accordingly confer resistance to restriction by MxB [9]-[11]. The viral integrase (IN) protein processes the long terminal repeat (LTR) ends of the viral DNA to yield the integration-competent preintegration complex (PIC), which subsequently transports the viral DNA into the nucleus for IN-mediated integration [21]. Although the recent studies agreed that MxB expression potently inhibited the early phase of HIV-1 replication including integration, they differed in terms of where in the lifecycle infection was blocked. For example, the measured effect on 2-LTR-containing DNA circles, which is utilized as a marker for PIC nuclear import [22], ranged from completely [9] or partially affected [11] to completely unaffected [10]. Additionally, initial results from these studies suggested MxB antiviral activity was independent of its GTPase activity [9],[11], yet dependent on the inclusion of an N-terminal sequence harboring a nuclear localization signal (NLS) [11]. Here we clarify that MxB restricts PIC nuclear import as well as HIV-1 integration. Concordantly, our results confirm the critical nature of the N-terminal region NLS, which can be functionally exchanged by the heterologous basic-type NLS from simian virus (SV) 40 large T antigen.