Role of capsid proteins in parvoviruses infection

The parvovirus genomes are approximately 5.0 kb in length and is enclosed within an icosahedron capsid (T = 1), which is 18–26 nm in diameter. At both ends of the genome, inverted terminal repeats (ITR) are formed by palindromic sequences, which are assembled into different shapes of a hairpin structure (depending on virus). In most of the parvoviruses, the viral DNA encodes two open reading frames (ORF). The ORF1 encodes nonstructural proteins (NS), and the ORF2 encodes two or three viral particle (VP) proteins which assemble the viral capsid; the VP proteins share a common termination codon. However, a few parvoviruses possess more than two ORFs, including, for example, members of the Bocaparvovirus, which have an extra ORF that encodes a nuclear phosphoprotein NP1 [12‐15]. The NS protein is a replicate protein that control genomic replication, is cytotoxic to host cells and is a cause of apoptosis [16, 17]. Successful examples of continuously producing viral particles in cell lines did not occur until a lac repressor-operator system was included, which first successfully overcame the cytotoxicity problem and allowed for stringent regulation of these proteins [18]. The NS1 protein has several replication-related regions, for example, a DNA-binding region, an ATP binding region, a helicase domain and a transactivation domain [16]. In the B19, AAV, minute virus of mice (MVM), canine parvovirus (CPV), porcine parvovirus (PPV), bovine parvovirus (BPV) and GPV, the VP1 contains the entire sequence of the VP2. Compared with the VP2, the VP1 has an extra length of ~140 amino acids at the N-terminal, with a phospholipase A₂ (PLA₂) domain and an NLS. The VP2 constitutes the primary component of the capsid protein and is highly conserved. The VP3 is a cleavage product that appears after the translation of VP2, and not all the parvoviruses have this protein. Although sequence diversity exists among the different species, these viruses have similar structures as determined by the three-dimensional structures, such as for B19 [19], AAV [20‐27], MVM [28], CPV [29, 30], BPV [9] and the Aleutian mink disease virus (ADV) [31].

The multifunctional capsid is responsible for the adsorption on and the entrance into the host cell, intracellular transport and localization, viral egress and induction of the immune response. Figure 1 shows the process of viral infection. In the following sections, the function of each capsid protein during the viral infection will be discussed.

In most parvoviruses, the VP1 N-terminal changes location from internal to being external of the capsid when exposed to some extreme chemical and physical conditions, such as heat or low pH. Moreover, the heat-induced exposure of the VP1 N-terminal is an irreversible process [32, 33]. The conformational changes of the capsid VP1 and the exposure of the VP1 N-terminal play a considerable role in viral infection. In previous studies with B19, a recombinant capsid VP1 N-terminal with a unique region (VP1u) was readily recognized by a specific monoclonal antibody, whereas the natural capsid was not recognized. Furthermore, after a heat or a low-pH treatment, the native VP1u was accessible to the antibody [34]. These results revealed a primarily conformational difference between the native and recombinant virion for VP1u and preliminarily showed that the B19 VP1u peptides were originally located inside the capsid. Additionally, the CPV and MVM virions exposed the VP1 N-terminal during endosomal trafficking, which also argued that the VP1 N-terminal was initially hidden in the viral capsid [33, 35]. Mutational experiments confirmed that the VP1 N-terminal of the MVM directly affected the infectivity of the virus [36]. Additionally, the direct effect the VP1 N-terminal on infectivity was observed with AAV-2. The microinjection of wild-type AAV-2 directly into the cytoplasm resulted in low infectivity, but the infectivity improved with the exposure of the VP1 N-terminal, which strongly suggested again that the VP1 N-termini are essential for the intracellular virion to initiate infection [37]. Additional studies revealed that B19 VP1u became external after primary cellular receptor attachment [38, 39]. Furthermore, the VP1u mediated internalization was a highly restricted process and only occurred in the permissive erythroid lineages, which suggested a viral tropism and pathogenesis [40].

The NLS is a nuclear localization signal that assists in the navigation of viruses to the nucleus, which leads to further nuclear translocation. A previous study identified a typical NLS on the VP1 N-terminal residues 4–13 in the CPV and demonstrated that the nuclear translocation was an ATP-dependent process [41]. An NLS was also identified in the MVM [36]. Moreover, the VP1 N-terminal region of the MVM is composed of four basic amino acid clusters, BC1, BC2, BC3 and BC4. Mutational and biochemical studies determined that BC1 and BC2 exhibited a great potential for nuclear transport [36]. The sequences of the BCs are highly conserved among parvoviruses; thus, we inferred that the BC elements, which contain the NLS, most likely participate in the nuclear transport in other representative parvoviruses, as well. A recent study showed that the PPV was distinctly different from the closely related parvoviruses because the PPV had more than one activated NLS, including a novel nuclear localization motif (NLM) [42]. One of the NLSs is typically located on the VP1 N-terminal and plays a role in the early steps of infection, whereas the other NLS is a novel NLM that targets the VP2 trimers to the nucleus late in the infection [42]. Additionally, in the genus Bocaparvovirus, the NLSs display some distinct properties in comparison with the other parvoviruses. Direct experiments suggest that the human bocavirus NP1 protein possesses a nonconventional NLS and is capable of transporting β-galactosidase fusion proteins to the nucleus [43], which indicates that the NP1 plays a role in nuclear translocation. Notably, in ADV, the VP1 binds the DNA, and the virus can not replicate efficiently in the nucleus because of the absence of the NLS [44].

In the majority of parvoviruses, the VP1 N-termini have a secretory PLA₂ (sPLA₂s) homology domain, which contains the catalytic site of a sPLA₂ and a conserved Ca²⁺ binding loop. However, a few of the viruses do not have this motif. The PLA₂ is a lipolytic enzyme that destroys the membrane and allows the viruses to escape from the lysosome. Additionally, the PLA₂ is located initially within the native capsid but is then turned to the outside after heat or pH treatments [34]. In addition to the conserved amino acids within the VP1u PLA₂ motif, some nonconserved amino acid residues around the VP1u also have an effect on PLA₂ activity. The B19 mutants of the PLA₂ motif showed a significant decrease in PLA₂ activity and viral infectivity, which suggested that the PLA₂ played an important role in the B19 life cycle [45]. Furthermore, in their study, Deng et al. demonstrated that the integrity of the cell membrane was destroyed by PLA₂ in a B19 infection. When the UT7-Epo cells were incubated with purified VP1u proteins, the cell morphology changed with time until finally the change was abrupt, resulting even in death, with an increase in the VP1u treatment. However, in comparison, the cells with mutant VP1u proteins or those in the control treatment did not change [46]. The identical effect was also validated for the AAV [47]. Additionally, the structural and conformational changes that occurred during the VP1 N-terminal externalization from the inner capsid to the outside were physically revealed [48]. In a comparison with humans, the question remains whether the same process occurs in animals. In the MVM, the phosphorylation of PLA₂ activity also influenced the stage of virus released from the endosome. After A9 cells were transfected with VP1u mutant infectious clones, the PLA₂ activity and the MVM infectivity were both abrogated [49]. Similarly in CPVs, when the capsid was incubated with PLA₂ inhibitors, the infectivity of the CPVs was significantly reduced and the endosome membrane permeability was changed during the period of CPV infection [35]. These results also demonstrated that PLA₂ activity is essential for effective infection.

Viral infection begins with the adhesion of the virus to a cell surface receptor [50]. In viruses, there is also an “anti-receptor” that is found in the parvovirus VP2 protein, which attaches to the cell receptor and begins internalization (Table 1). With the use of the cryo-electron microscopy technique at the resolution of 8-Å, a depression at the icosahedral threefold axes of the B19 VP2 capsid was found, which bound to the cellular receptor globoside [51]. In the AAV-2, the heparin sulphate proteoglycan is a dominant receptor [52], which shares a common receptor α5β1 with B19. When the AAV-2 was combined with heparin, at ~18 Å resolution under cryo-electron microscopy, the conformational changes of the capsid were revealed; the tip of the protrusions on the three-fold axes became flat, and the top of the channel located on top of the five-fold axes grew broader [53]. In contrast to B19, there is a distinct “spike” on the three-fold axes of CPV VP2, based on observations of the 3D structure. On the VP2 capsid surface, there is a putative receptor-binding region located in a depression with a canyon-like shape. Moreover, one of the binding regions is located between the two icosahedral three-fold axes, and the other binding region is on the five-fold axes [30]. The BPV capsid VP2 showed a common parvovirus feature, obvious protrusions encircling the threefold axes, which indicated a potential site for receptor recognition [9]. For ADV, the 3D structure of VP2 capsid have been determined to 22 Å resolution, a dimple at the twofold axes was indicated to be involed in the recognition of cellular receptor [31].

Notably, a new NLS of the B19 virus was located on the VP2 N-terminal, which could facilitate nuclear transport [54]. In the case of MVM, the VP2 N-terminal also acted as a nuclear export signal (NES) [55]. In further studies, when the temperature reached a specific level, the VP2 N-terminal region of the MVM transferred from the inner capsid through the five-fold axes to the outside of the virion [56]. After a series of conformational changes and the exposure of the NLS and NLM, the VP1 cooperated with the VP2 and formed a trimer with the assistance of VP2-chaperone activity to finally traffic through the nuclear pore complex [57]. Additionally, the VP2 N-terminal was phosphorylated during the late stage of the MVM life cycle, and therefore, the viruses spread to neighbouring cells efficiently prompted by the activated VP2. Miller et al. hypothesised that the N-terminal of the VP2 together with the NS2 had an influence on the viral egress from the nucleus [58]. In the case of ADV, a D534 residue of the VP2 enabled the ADV-G to replicate in mink, which induced a continuous immune response [59].

The VP3 generally occurs only when the viral genome has completed the capsid assembly and packaging. In the MVM, the VP3 is generated from the cleavage of VP2 at approximately 25 amino acids from the N-terminal. A trypsin digestion experiment demonstrated that this proteolytic reaction occurred only in the mature virion, with an intact genome. Although the VP1 has an identical proteolytic site, cleavage does not occur [60]. An identical proteolytic phenomenon also appeared in the ADV, but with some dissimilarity. Previous studies showed that during an ADV infection or when the viral capsid only was expressed an extra, an unknown 26 kDa protein was produced, which was recently identified as a split product of the capsid VP1 and VP2. Based on the results, the caspase family was activated during the expression of the ADV capsid proteins, and the caspase-7 response was to specifically cleavage the capsid at a distinct site [61].

The studies on VP1 and VP2 are sufficiently comprehensive, but research on VP3 is lacking, and further studies are required to reveal the discrete functions of this protein. For birds, the VP3 is a prominent protein in both GPV and Muscovy duck parvovirus (MDPV), in which the VP3 induced a distinct immune response [62]. Recombinant GPV capsid proteins were expressed and purified to apply to an in vivo assay, and all of the VLPs induced a strong immune response in the susceptible geese, whereas the VLPs-VP2 and VLPs-VP3 induced higher concentrations of neutralized antibodies than the VLPs-VP1 [11].

Notably, a 23 kDa protein encoded by the AAV ORF2 promoted the assembly of the VP3 capsid and was named as one of the assembly-activating proteins (AAP). When the AAP was expressed, some of the VP3 was transformed into the nucleus to form the capsid [63], which indicated that the AAP enabled the transport of VP proteins to the nucleoli and was required for the capsid package. The inference was that the VP3 was responsible for the capsid assembly and the virion stability. However, the role of VP3 in the parvovirus life cycle remains to be fully understood.

For the viral capsid, engineers conducted a variety of innovations. Other viruses, for example, a non-enveloped virus in the Picornavirus family that causes foot-and-mouth disease (FMD), have much in common with parvoviruses, and interference against the VP1 structural protein successfully produced a protective effect in both cells and suckling mice [64]. This result indicated that RNAi technology might provide a therapeutic measure to treat virus infection. In the parvoviruses, the dominant antigenic determinants are located on the capsid. Langeveld et al. used an immunofluorescent assay to analyse the B-cell epitopes on structural proteins and, based on this research, obtained the CPV polypeptide vaccine [65]. Moreover, the VLPs-VP2 of the CPV expressed in Escherichia coli and inoculated in mice showed subtle differences between the native viruses and the VLPs in neutralizing antibody titres, in addition to the immune response of the T-cells [66]. The CPV VLPs shared properties closely with the native virus and could be profitable candidates for therapeutic vaccines. Furthermore, researchers used the relation between CPV and its cellular receptor TfR (also exist in human cells) to explore a potential nano-container for tumor targeting. In their study, the accessible lysines on the viral capsid were derivatized with dye molecules. After incubation with TfR expression or those that lacked the TfR expression cell lines, Researchers found that the internalization of the labelled CPV-VLPs was observed only in the TfR expression cell lines [67], which indicated that the CPV-VLPs could act as a target delivery substance.

Adeno-associated viruses (AAVs) are nonpathogenic members of the Parvoviridae family. The last 10 years has witnessed a surge of studies on the use of AAVs as vectors. For example, a recombinant triple-tyrosine mutant AAV-2 vector significantly increased the efficiency of gene transduction, approximately three-fold compared with the wild type [68]. Thus, we could minimize the therapeutic dose and produce a preferred level of protection. More recently, Giridhara et al. found that AAV vectors activated the NF-κB pathway, which suggested an application in gene therapy [69]. Furthermore, Mirta et al. constructed a recombinant AAV capsid vector, which contained a fragment of a tumour-targeting sequence, and the modified mutant resulted in high transduction in tumour cells, but the transduction in 293 T cells was poorly induced. By contrast, the native AAV capsid was highly expressed in the 293 T cell lines, and this altered tropism suggested a strategy to develop an in vivo-targeted vaccine [70].