Regulated nucleocytoplasmic trafficking of viral gene products: A therapeutic target?

Abstract

The study of viral proteins and host cell factors that interact with them has represented an invaluable contribution to understanding of the physiology as well as associated pathology of key eukaryotic cell processes such as cell cycle regulation, signal transduction and transformation. Similarly, knowledge of nucleocytoplasmic transport is based largely on pioneering studies performed on viral proteins that enabled the first sequences responsible for the facilitated transport through the nuclear pore to be identified. The study of viral proteins has also enabled the discovery of several nucleocytoplasmic regulatory mechanisms, the best characterized being through phosphorylation. Recent delineation of the mechanisms whereby phosphorylation regulates nuclear import and export of key viral gene products encoded by important human pathogens such as human cytomegalovirus dengue virus and respiratory syncytial virus has implications for the development of antiviral therapeutics. In particular, the development of specific and effective kinase inhibitors makes the idea of blocking viral infection by inhibiting the phosphorylation-dependent regulation of viral gene product nuclear transport a real possibility. Additionally, examination of a chicken anemia virus (CAV) protein able to target selectively into the nucleus of tumor but not normal cells, as specifically regulated by phosphorylation, opens the exciting possibility of cancer cell-specific nuclear targeting. The study of nucleoplasmic transport may thus enable the development not only of new antiviral approaches, but also contribute to anti-cancer strategies.

Introduction

Viruses remain the least understood of human pathogens, the patent lack of knowledge with respect to pathogenic mechanisms and host–parasite interactions being highlighted by the general lack of effective drugs against viruses and limited success in developing vaccines. Progress in understanding how viral gene products traffic in eukaryotic host cells, and how important regulation thereof may be in controlling infection, provides new hope that this information may be exploited in the future in the development of antiviral therapeutics.

The present review focuses on recent findings for a subset of gene products from quite diverse viruses with respect to the importance of regulated trafficking into and out of the nucleus, and the role of specific phosphorylation therein. The precisely scheduled trafficking and regulation thereof for these viral gene products appears to be critically important to the viral infectious cycle, making strategies to hinder these pathways of great interest for the future in terms of developing novel pharmaceutical approaches.

Eukaryotic cells are highly compartmentalized biological systems; the genetic information resides in the cell nucleus, separated from the apparatus responsible for protein synthesis, which resides in the cytoplasm. Separation of the two compartments is effected by a double membrane structure, the nuclear envelope (NE), passage through which may only occur through the NE-embedded ca. 120-MDa multiprotein-constituted nuclear pore complexes (NPCs) [1]. Molecules < ca. 50 kDa can freely diffuse through the aqueous channel delimitated by the NPC, whereas the transport of larger molecules occurs through an active, signal-dependent process [2]. Examples of such molecules include cellular mRNAs that need to progress after transcription from the nucleus to the cytoplasm in order to be translated, and proteins such as histones and transcription factors, synthesized in the cytoplasm that need to be transported into the nucleus in order to perform their functions [3].

Active transport of molecules through the NPC is mediated by members of the importin (IMP) family of cellular transporters that recognize specific targeting signals on cargo molecules. Nuclear localization sequences (NLSs) and nuclear export sequences (NESs) represent the signals for transport into and out of the nucleus, respectively (Fig. 1) [2]. The study of viral proteins has been proven invaluable in delineating the molecular details of nucleocytoplasmic transport. Over 20 years ago, Kalderon et al. demonstrated that simian virus 40 (SV40) large tumor antigen (T-ag), a multifunctional protein normally localizing in the host cell nucleus, mislocalized to the cytoplasm upon mutation of a short basic sequence [4], which, when present in frame to the sequence of otherwise cytoplasmic proteins such as pyruvate kinase, was sufficient to relocalize them to the cell nucleus [5]. Importantly, mutation of this sequence does not alter T-ag's biochemical properties, such as the ability to bind to double-stranded (ds) DNA or the tumor suppressor p53. The 7-amino-acid (aa) highly basic sequence (PKKKRKV¹³², single letter aa code) was dubbed a “nuclear location signal” [6] or NLS. Since then a number of different types of NLS have been identified which can be classified on the basis of their sequence, and their ability to be recognized by specific IMPs. Transport can be mediated, either by IMPβ1 or one of its several homologues which are able to recognize the NLS-carrying cargo directly, or IMPβ1 together with the adapter molecule IMPα, where the latter recognizes the NLS. Mammalian cells possess a number of distinct forms of IMPα and IMPβ which can differ widely in NLS-binding specificity [7], [8], [9], [10], [11], [12] and expression pattern in different tissues [13], [14], [15], [16].

The best understood NLSs are classically highly basic sequences, generally lysine-rich, conferring interaction with the IMPα/IMPβ1 heterodimer. These NLSs either resemble the T-ag-NLS in being a single cluster of basic residues (see above) or are bipartite in nature, with two clusters of basic residues separated by a spacer region of 10–13 aa [17]; an example is the NLS from the histone assembly factor nucleoplasmin (KR-11aas-KKKK¹⁷⁰). IMPα has two functional domains, a short basic N-terminal domain, the IMPβ-binding (IBB) domain which is responsible for binding to IMPβ1, and a large region of 10 repeat sequences, that contains 2 NLS-binding regions. In the absence of IMPβ1, the IBB domain binds to the major NLS-binding pocket of the repeat region of IMPα, preventing NLS binding. Binding of IMPβ to IMPα displaces the IBB from the NLS-binding site on IMPα, allowing IMPα to interact with NLSs [18] (Fig. 1(2)). Once in the nucleus, binding of the monomeric guanine–nucleotide binding protein Ran in activated GTP-bound form to IMPβ mediates a conformational change, resulting in the release of IMPβ from the IBB domain, which then competes with the NLS for the major binding site on IMPα, effecting cargo release into the nucleus (Fig. 1(3)).

Certain basic often arginine-rich NLSs are recognized specifically by IMPβ1 directly; an example is the RQARRNRRRRWR⁴⁶ sequence from the human immunodeficiency virus (HIV-1) Rev protein [19]. Other types of NLS are recognized by homologues of IMPβ1; the Gly-/aromatic aa-rich M9 NLS sequence (GNYNNQSSNFGPMKGGNFGGRSSGPYGGGGQYFAKPRNQGGY³⁰⁵ [20]) from the heterogeneous ribonucleoprotein particle (hnRNP) A1 protein, for example, is recognized by IMPβ2 (transportin) [21]. In the case of proteins transported by an IMPβ alone (Fig. 1(1)), cargo release in the nucleus is dependent on Ran-GTP binding to the IMPβ to dissociate the NLS-carrying protein–IMPβ complex, in analogous fashion to pathways dependent on the IMPα/IMPβ1 heterodimer (see Fig. 1(3)).

Nuclear protein export is analogous to IMP-dependent nuclear protein import, mediated by the “exportin” (EXP) members of the IMPβ superfamily. Over 10 years ago, several groups identified a strongly hydrophobic 11 aa stretch (the nuclear export signal—NES) within the HIV-1 Rev protein (LQLPPLERLTL⁸³) that was responsible for the export from the nucleus to the cytoplasm, as well as being able to mediate nuclear export of heterologous proteins [22]. Rev's NES is recognized by EXP1, or chromosome region maintenance protein 1 (CRM1) [23]. CRM1 is the best characterized EXP, mediating nuclear export of many proteins, as well as certain mRNAs and spliceosomal U small nuclear (sn)RNAs; its activity can be specifically inhibited by the drug leptomycin B (LMB) [24]. Other EXPs include the cellular apoptosis susceptibility factor CAS (Cse1p in yeast), the EXP for IMPα  responsible for recycling it to the cytoplasm subsequent to a round of nuclear import [25]. Unlike IMPs, EXPs can only bind their cargoes when complexed to Ran-GTP (see Fig. 1(4)); subsequent to passage through the NPC to the cytoplasm, hydrolysis by Ran of GTP to GDP results in dissociation of the complex (Fig. 1(5)).

The IMPα NLS-binding domain is determined by the repetition of 10 modular motifs named Armadillo (ARM) repeats. Each ARM repeat is formed by ca. 40 aa organized in 3 α-helices (H1, H2 and H3), in turn arranged in a superhelical structure. The ARM repeats 2–4 and 7–8, which contain strategically positioned asparagine and tryptophan residues, form the major and minor binding pockets for conventional basic NLSs [26], [27], [28]. Although able to bind to both binding pockets, monopartite NLSs such as that of T-ag bind preferentially to the major binding site (ARM repeats 2-4), whereas the two basic clusters of bipartite NLSs are able to bind simultaneously to both [28], [29]. Both pockets retain binding sites for six aas (P₁–P₆ and P1₁–P1₆ for the major and the minor binding sites, respectively). Interactions between the aas of the NLS and the binding pockets of IMPα1 include backbone interactions of the side chains of the aas of the NLS with conserved asparagine side chains on IMPα, binding of aliphatic portions of the lysine side chains of the NLS to the shallow grooves formed by the H3 helices on the surface of IMPα, and salt bridges between the extended Lys side chains with conserved negatively charged residues that surround the hydrophobic grooves on the surface of IMPα [28]. The grooves formed by IMPα in positions P₂, P₄ and P₅ present a negative charge that can also form electrostatic interactions with positive charges provided by the side chains of basic residues of the NLS, with the groove in P₂ being too small to accommodate an arginine residue. Mutation of K¹²⁸ (+ 2 position of the NLS) of the T-ag NLS [5] binding to the P₂ site completely abolishes nuclear targeting activity, thermodynamic studies confirming that the Lys-P₂ interaction makes the highest energetic contribution to IMPα/NLS binding. The lowest contribution is conferred by residues in the + 1 and + 6 positions of the NLS, whereby substitution by several different aa does not impair NLS functionality [30]. However, a negative charge in position + 1 of the NLS might have a detrimental effect on the NLS–IMPα interaction, due to electrostatic repulsion with the negative charges present on the IMPα binding site (see Fig. 2(3)). Accordingly, the current core consensus sequence for IMPα binding can be considered X-K-(K/R)-X-(K/R)-X, where X is any aa [28], [30].

Eukaryotic cells can tightly regulate the nuclear import rate of certain cargoes by modulating a variety of factors such as the levels and distribution of IMPs/EXPs [13], [16], NPC number and composition [3], [13], [16], [31], and the affinity of the cargo for the respective IMPs/EXPs, which in most cases is regulated by the phosphorylation of the cargo in vicinity of the NLS/NES [3], [13], [16], [31]. As there is a direct correlation between the affinity of IMPs for the cargoes and the level/efficiency of nuclear accumulation in both mammalian and yeast cells [32], [33], it is clear that modulating the affinity of an NLS/NES for an IMP/EXP can alter the kinetics of nuclear transport markedly.