The objectives of this project, and related progress in the past year, are reviewed below. (1) Development of a reverse genetics system that can be used to modify the antigenicity and virulence of rotaviruses (RVs) and to develop new vaccine candidates. The RV genome consists of 11 double-stranded RNA segments. The positive-sense RNA of each segment typically contains short 5'and 3'untranslated regions (UTRs) that flank a single open reading frame (ORF). Occasionally, individual RV genome segments undergo rearrangement due to partial head-to-tail duplication of a segment sequence. In most cases, rearrangements initiate after the stop codon, leaving the ORF and the encoded protein unaffected. During virus replication, naturally occurring segments with sequence rearrangement are preferentially packaged over wild-type segments. The mechanism for the improved-replication phenotype is unclear, but may be connected to duplication of RNA sequences involved in packaging of the RV genome. As an extension of these observations, we used reverse genetics to engineer duplications into the segment encoding NSP2. The duplications were 25, 50, 100 and 200 bases in length and were inserted into the 3'UTR of the NSP2 segment such that ORF was unaffected and retained the capacity to encode NSP2. Recombinant RVs containing the duplicated sequences grew to high titer and were genetically stable. They were, however, unable to compete with wild-type virus in mixed infections, suggesting that the location and /or length of the engineered sequence duplication did not provide for preferential packaging of the segment. These results indicate that sequence duplication alone (or increased segment length) is not the only factor responsible for enhancing the packaging activity of an RNA segment. Indeed, our work raises the possibility that many rearrangements that arise in vitro and in vivo go undetected and are eliminated from the virus population. (2) Elucidation of mechanisms that RV proteins use to antagonize the interferon signaling pathway. The RV nonstructural protein NSP1 inhibits the expression of type I interferon (IFN), thereby promoting viral spread. NSP1 is predicted to function as an E3 ubiquitin ligase by inducing polyubiquitination and subsequent proteasome-mediated degradation of the IFN-regulatory factors IRF3, IRF5, and IRF7. All IRF proteins share an N-terminal DNA-binding domain, while IRF3, IRF5, and IRF7 contain a C-proximal IRF-association domain (IAD), which mediates dimerization. IRFs are maintained in an inactive state in the cytoplasm through an autoinhibitory domain that interacts with the IAD, burying key residues required for IRF dimerization and preventing IRF nuclear accumulation. C-terminal phosphorylation induces charge repulsions, causing the autoinhibitory domain to unmask the IAD, allowing formation of transcriptionally active IRF dimers. To identify the region of IRF proteins targeted for degradation by NSP1, we generated a series of IRF3 and IRF7 truncation mutants and co-expressed each with NSP1. We found that the IAD was necessary and sufficient for degradation by NSP1. Disruption of conserved residues required for dimerization of IRF3 did not interfere with NSP1-mediated degradation of the protein. Similarly, constitutively-expressed dimeric forms of IRF3 were targeted for degradation by NSP1. These results indicate that NSP1 recognizes the dimerization domain of both the monomeric and dimeric form of IRF proteins for degradation. The ability to target IRF proteins in both states suggests multiple steps of the IFN signaling pathway are targeted by NSP1, allowing for more efficient inhibition of IFN expression. (3) Analysis of the diversity and evolution of the RV genome: (a) Characterization of asymptomatic and symptomatic G10P11 RV Infections in Vellore, India, suggest a possible role of Aichi virus in diarrheal disease. Human G10P11 RVs causing symptomatic and asymptomatic infections have been described in neonates living in Vellore. Towards the goal of identifying possible genetic determinants responsible for the different phenotypes, the genomes of G10P11 viruses in stool samples from neonates/infants with asymptomatic infections (n=20) and with symptomatic infections (n=19) were recovered and analyzed using a combination of Sanger, 454, and Illumina sequencing methods. The results of RV sequence alignments and phylogenetic analysis showed no differences in the nucleotide or amino acid sequences of G10P11 RVs associated with asymptomatic and symptomatic infections. Surprisingly, deep sequencing revealed the presence of Aichi viral RNA in 2 of the 20 asymptomatic infections and 6 of the 19 symptomatic infections. These results indicate that genetic determinants are not responsible for differences in symptomatic and asymptomatic G10P11 RV infections. Our results raise the possibility that other infectious agents, such as Aichi virus, may be responsible for some diarrheal disease previously associated with symptomatic strains of G10P11 RV. (b) Increased incidence of G2P4 RV infections as Vanderbilt University Medical Center (VUMC) during 2010-2011 reveal distinct genotype 2 alleles. As part of the CDCs New Vaccine Surveillance Network, stool specimens were collected from infants and young children presenting with acute gastroenteritis at VUMC during the season preceding the introduction of rotavirus vaccines in the US and prospectively since. Most laboratory-confirmed rotavirus infections at VUMC were caused by G1P8 rotaviruses during the 2005-06, 2006-07, and 2007-08 seasons and G3P8 strains during the 2008-09 season. Relative to the preceding years, an increase in rotavirus infections were noted for the 2010-11 season, for which G2P4 viruses were largely responsible (12 G2P4, 13 G3P8, 1 G9P8, 1 G12P8). To better understand the diversity and relationships of the G2P4 viruses, their genomes were sequenced. All the G2P4 viruses contained complete genotype 2 (DS1-like) constellations (I2-R2-C2-M2-A2-N2-T2-E2-H2), indicating a lack of reassortment with more commonly circulating genotype 1 (Wa-like) viruses (I1-R1-C1-M1-A1-N1-T1-E1-H1). Phylogenetic analysis showed the genes of the G2P4 viruses could be resolved into at least two subgenotype alleles and most of the viruses could be resolved into either of two clades based on the conserved nature of their allele constellations. Our findings suggest, like G1P8 and G3P8 viruses, a limited amount of genetic exchange occurs between the G2P4 virus clades co-circulating in the same geographical location during the same year. (c) Vaccine-derived NSP2 segment in RV from vaccinated children with gastroenteritis in Nicaragua. The efficacy of RV vaccines can be low in developing countries. For example, a vaccine efficacy against severe diarrhea of only 58% was noted in a 2007-2009 Nicaraguan study with RotaTeq. To understand the basis of vaccine failure, the genomes of RVs isolated from vaccinated Nicaraguan children with gastroenteritis were sequenced. The results revealed that all the viruses had typical genotypes (11 G1P8, 1 G3P8) and that nine of the G1P8 viruses and the single G3P8 virus had genome constellations common to human RVs. However, two of the G1P8 viruses had atypical constellations, G1P8I1R1C1M1A1N2T1E1H1, due to the presence of a genotype-2 NSP2 (N2) gene. The sequence of the N2 NSP2 gene was identical to the RotaTeq N2 NSP2 gene, indicating that the two atypical viruses originated by reassortment of human G1P8 RVs with RotaTeq viruses. These data suggest that the high level of vaccine failure in Nicaragua is neither due to antigenic drift nor to the emergence of new antigenically-distinct virus strains. Our data also suggest that the widespread use of RotaTeq has led to the introduction of vaccine genes into circulating human RVs.
Showing the most recent 10 out of 22 publications
