Studies on rotavirus NSP5 phosphorylation and its interaction with NSP2

Rotavirus NSP5 is a non-structural protein that localises in cytoplasmic viroplasms of infected cells. NSP5 interacts with NSP2 and undergoes a complex posttranslational hyper-phosphorylation , generating species with reduced PAGE mobility. This process has been suggested to be due in part to autophosphorylation. Here, we show that it rather works as an autoregulator of its own phosphorylation though the activation of cellular kinases. In this thesis, it is described the development of an in vitro phosphorylation assay using as a substrate an in vitro-translated NSP5 deletion mutant that was phosphorylated by extracts from MA 104 cells transfected with NSP5 mutants but not by extracts from mock-transfected cells. The phosphorylated products obtained showed shifts in mobility similar to what occurs in vivo. From these and other experiments, we concluded that NSP5 activates a cellular kinase(s) for its own phosphorylation. Three NSP5 regions were found to be essential for kinase(s) activation. Glutathione-S-transferase-NSP5 mutants were produced in E. coli and used to determine phosphoacceptor sites. These were mapped to four serines (153, 155, 163 and 165) within an acidic region with homology to casein kinase 2 (CK2) phosphorylation sites. CK2 was able to phosphorylate NSP5 in vitro. NSP5 and its mutants fused to enhanced green fluorescent protein were used in transfection experiments followed by virus infection and allowed the determination of the domains essential for viroplasm localisation in the context of the virus infection. A second hyper-phosphorylation assay was also developed. This is an in vivo assay in which , two constructs are co-transfected . One of them tagged with 11 aa served as substrate while the other was used to map the domains required to induce activation of the cellular kinase. We learn that the two activities can be separated , demonstrating that the hyper-phosphorylation is a process in trans, with one molecule activating and the other being phosphorylated. We described a motif a (from amino acids 63 to 67) in region 2 with the amino acidic sequence SDSAS. This motif, and in particular phosphorylated serine 67, is responsible to trigger the hyper-phosphorylation process. In fact, mutation of serine 67 to aspartic acid in the full-length NSP5 allowed hyper-phosphorylation of NSP5 in the absence of NSP2, suggesting that NSP2 could produce a conformational change in NSP5 to expose motif a (serine 67), thus allowing phosphorylation of serine 67. On the other hand, NSP5 substrate characteristics were mapped in region 4 (amino acids 131 to 179). The serines 153, 155, 163 and 165, that are CK2-like phosphorylation sites , are in part responsible for the hyperphosphorylation. Two other serines, 137 and 142, that are a putative sites for PKC phosphorylation are good candidates to be also substrates in NSP5 hyperphosphorylation. Moreover, the c-terminal tail (T) of NSP5 of 18 aa was found to be also necessary for activation of the cellular kinase. Although its role is not yet clear, it is possible that a dimersation in trans with another NSP5 molecule can explain the results. Alternatively, a direct interaction with a cellular kinase may be required to permit its activation Viroplasms are discrete structures formed in the cytoplasm of cells sustaining rotavirus replication that constitute the machinery of replication of the virus. In this thesis, it has been investigated the relative localisation of NSP5 and NSP2 within viroplasms as well as the dynamics of viroplasm formation in cells infected with rotavirus that also express NSP5 or NSP5 fused to EGFP. The results showed NSP2 localising more internally with respect to NSP5. The number of viroplasms was shown to first increase and then to decrease in post-infection time, while the area of each one increased , suggesting a fusion between them. The interaction between NSP2 and NSP5 was investigated using two different assays, namely a two-hybrid system and an in vivo binding assay. Both methods gave essentially the same results, indicating that in NSP5 the N-terminal region (33 aa) as well as C-terminal part (amino acids 131 to 198) are required for bind ing to NSP2. These two regions were able to confer to EGFP ability to local ise in viroplasm and to form VLS with NSP5.