Binding of influenza viruses to sialic acids: reassortant viruses with A/NWS/33 hemagglutinin bind to alpha2,8-linked sialic acid

We have examined the specificity of binding of A/NWS/33 hemagglutinin (HA), exploring the effects of fucosylation, changing the Gal-GlcNAc linkage between the second and third sugars, and binding affinity for alpha2,8-linked sialic acid. The HA of A/NWS/33(HA)-Tokyo/67(NA) (NWS-Tok, H1N2) virus binds to 3'-linked sialyllactose with 10-fold higher affinity than 3' sialyllactosamine and 3-fold higher affinity than 6' sialyllactosamine. The P227H mutation in A/NWS/33(P227H)(HA)-A/Memphis/31/98(NA) (NWS-Mem/98, H1N2) results in sevenfold lower affinity for 3' sialyllactose, but binding to 6' sialyllactosamine is unchanged. The apparent switch from 3' to 6' specificity is solely due to a loss of Siaalpha2,3 binding. Fucosylation of the third sugar and changing the linkage between second and third sugars had little effect on binding by NWS-Tok, but marked effects on A/NWS/33(P227H)(HA)-tern/Australia/G70c/75(NA) (NWS-G70c, H1N9) and NWS-Mem/98. NWS-Tok, NWS-G70c, and NWS-Mem/98 bind to alpha2,8-bisialic acid with high affinity. NWS-Mem/98 can also bind to alpha2,8-trisialic acid, but with lower affinity. Together, these data show that alpha2,8-linked sialic acid, fucosylation of the third sugar, and linkage between the second and third sugars could play important roles in allowing efficient virus binding to its host cell. The finding that influenza viruses have the potential to bind to alpha2,8-linked sialic acid is a new influenza virus-receptor interaction pathway.     

Cell lines that support replication of a novel herpes simplex virus 1 UL31 deletion mutant can properly target UL34 protein to the nuclear rim in the absence of UL31

Previous results indicated that the herpes simplex virus 1 (HSV-1) U(L)31 gene is necessary and sufficient for localization of the U(L)34 protein exclusively to the nuclear membrane of infected Hep2 cells. In the current studies, a bacterial artificial chromosome containing the entire HSV-1 strain F genome was used to construct a recombinant viral genome in which a gene encoding kanamycin resistance was inserted in place of 262 codons of the 306 codon U(L)31 open reading frame. The deletion virus produced virus titers approximately 10- to 50-fold lower in rabbit skin cells, more than 2000-fold lower in Vero cells, and more than 1500-fold lower in CV1 cells, compared to a virus bearing a restored U(L)31 gene. The replication of the U(L)31 deletion virus was restored on U(L)31-complementing cell lines derived either from rabbit skin cells or CV1 cells. Confocal microscopy indicated that the majority of U(L)34 protein localized aberrantly in the cytoplasm and nucleoplasm of Vero cells and CV1 cells, whereas U(L)34 protein localized at the nuclear membrane in rabbit skin cells, and U(L)31 complementing CV1 cells infected with the U(L)31 deletion virus. We conclude that rabbit skin cells encode a function that allows proper localization of U(L)34 protein to the nuclear membrane. We speculate that this function partially complements that of U(L)31 and may explain why U(L)31 is less critical for replication in rabbit skin cells as opposed to Vero and CV1 cells.     

Molecular epidemiological and clinical aspects of hepatitis D virus in a unique triple hepatitis viruses (B,C, D) endemic community in Taiwan

The molecular epidemiological and clinical aspects of hepatitis D virus (HDV) in a unique HBV, HCV, and HDV triple virus endemic community in southern Taiwan were investigated. A total of 2,909 residents aged 45 or older were screened for hepatitis B surface antigen (HBsAg), anti-HCV antibody, and anti-HDV antibody (specifically for HBsAg-positive carriers). Factors that might be associated with HDV infection, viral nucleic acid detection, and genotyping of HBV, HCV, and HDV were investigated. The prevalence of HBsAg and anti-HCV were 12.6% (366/2,909) and 41.6% (1,227/2,909), respectively. For HBsAg carriers, 15.3% (56/366) were positive for anti-HDV assay. Living in a higher endemic district of HCV infection (odds ratio [OR] = 3.2; 95% confidence interval [CI] = 1.7-6.3), male gender (OR = 1.9; 95% CI = 1.1-3.6) and co-infection with HCV (OR = 1.8; 95% CI = 1.0-3.3) were significantly independent factors associated with HDV infection. The detection rate of HDV RNA among anti-HDV-positive patients was only 12.7% (7/55). The mean HBV titer of triple infection group was significantly lower than in the HBV/HDV co-infection group (2.23 vs 3.05 in log(10), copies/ml, P = 0.046). HCV RNA detection among the triple infection group showed 47.4% (9/19) viremia rate and viral loads of 579,121 IU/ml in median (16,803-1,551,190 IU/ml). The prevalent genotype of HBV was type B (23/25); HCV was 1b (7/9) and HDV was IIa/IIb (4/4). Only the presence of HCV RNA predicted the presence of elevated ALT significantly (OR = 25.0; 95% CI = 3.39-184.6). In conclusion, the geographical aggregation of HDV infection paralleled that of HCV infection in this community. HCV suppressed the replication of HBV among triple vital infection patients. HBV and HDV lapsed into a remission or nonreplicative phase in most cases, and HCV acted as a dominant factor in triple viral-infected individuals. Only the presence of HCV RNA was associated with elevated ALT values, but not HBV or HDV.     

Characterization of a 105-kDa plasma membrane associated glycoprotein that is involved in West Nile virus binding and infection

This study attempts to isolate and characterize West Nile virus-binding molecules on the plasma membrane of Vero and murine neuroblastoma cells that is responsible for virus entry. Pretreatment of Vero cells with proteases, glycosidases (endoglycosidase H, alpha-mannosidase), and sodium periodate strongly inhibited West Nile virus infection, whereas treatments with phospholipases and heparinases had no effect. The virus overlay protein blot detected a 105-kDa molecule on the plasma membrane extract of Vero and murine neuroblastoma cells that bind to WN virus. Treatment of the 105-kDa molecules with beta-mercaptoethanol resulted in the virus binding to a series of lower molecular weight bands ranging from 30 to 40 kDa. The disruption of disulfide-linked subunits did not affect virus binding. N-linked sugars with mannose residues on the 105-kDa membrane proteins were found to be important in virus binding. Specific antibodies against the 105-kDa glycoprotein were highly effective in blocking virus entry. These results strongly supported the possibility that the 105-kDa protease-sensitive glycoprotein with complex N-linked sugars could be the putative receptor for WN virus.     

Mismatched hemagglutinin and neuraminidase specificities in recent human H3N2 influenza viruses

The hemagglutinin (HA) of influenza viruses initiates infection by binding to sialic acid on the cell surface via alpha2,6 (human) or alpha2,3 (avian) linkage. The influenza neuraminidase (NA) can cleave both alpha2,3- and alpha2,6-linked sialic acids, but all influenza NAs have a marked preference for the non-human alpha2,3 linkage. Recent H3N2 influenza viruses have lost the ability to agglutinate chicken red blood cells. To determine if changes in HA specificity or affinity correlate with NA specificity or activity, we examined red cell binding and elution of a series of H3N2 viruses. We found that the NA activity of many influenza viruses does not release binding by their HA. In some egg-adapted strains, lack of elution correlates with low levels of viral NA activity, and these elute rapidly when bacterial NA is added. However, a Fujian-like virus, A/Oklahoma/323/03, does not elute by its own NA or with Vibrio cholerae sialidase, and it binds to red cells pre-treated with V. cholerae sialidase. It elutes after addition of the broad specificity Micromonospora viridifaciens sialidase. Human glycophorin inhibits A/Oklahoma/323/03 hemagglutination 6-fold better than fetuin. We conclude that specific forms of sialic acid are used as receptor by recent human H3N2 influenza viruses, perhaps involving branched alpha2,6 sialic acid or alpha2,8 sialic acid structures on O-linked carbohydrates. The virus itself has no O-linked glycans, so even though the NA is not able to cleave receptors on cells, the viruses will not self-aggregate. It will be important to monitor efficacy of neuraminidase inhibitors in case there are NA-resistant receptors in the human respiratory tract that allow the viruses to be less dependent on NA activity.     

Sensitivity of prostate tumors to wild type and M protein mutant vesicular stomatitis viruses

Because of its potent ability to induce apoptosis, vesicular stomatitis virus (VSV) is an attractive candidate as an oncolytic virus for tumor therapy. Previous studies have suggested that VSV selectively infects tumor cells due to defects in their antiviral responses making them more susceptible to VSV infection than normal cells. We tested this hypothesis in the prostate tumor system by comparing LNCaP and PC-3 prostate tumor cells to benign human prostatic epithelial cells from patient prostatectomy specimens. We compared the cell killing ability of a recombinant virus containing a wild-type (wt) M protein (rwt) and an isogenic M protein mutant virus (rM51R-M) that induces interferon (IFN) in infected cells and should display a greater selectivity for tumor cells. Our results showed that in single-cycle infection experiments, LNCaP cells were sensitive to killing by both wt and mutant viruses, while PC-3 cells were highly resistant to VSV-induced cell killing. LNCaP and benign prostate cells were similarly susceptible to both viruses, indicating that normal prostate cells are not inherently resistant to killing by VSV. In each of the cell lines, the rM51R-M virus induced similar levels of apoptosis to rwt virus, showing that the M protein does not play a significant role in apoptosis induction by VSV in these cells. In multiple-cycle infection experiments, LNCaP cells were more sensitive than benign prostatic epithelial cells to virus-induced cell killing by rM51R-M virus, but not rwt virus. Both viruses were equally effective at reducing LNCaP tumor volume in vivo following intratumoral and intravenous inoculation in nude mice, while PC-3 tumors were resistant to VSV treatment. None of the mice treated with rM51R-M virus died as a result of virus infection, while 50-71% of mice treated with rwt virus succumbed to virus infection. Similarly, when inoculated by the more sensitive intranasal route, the rM51R-M virus was less pathogenic than the rwt virus from which it was derived. These results indicate that M protein mutant viruses are superior candidates as oncolytic viruses for therapies of prostate tumors, but future strategies for use of VSV will require testing individual tumors for their susceptibility to virus infection.     

Processing of Norwalk virus nonstructural proteins by a 3C-like cysteine proteinase

Expression of Norwalk virus nonstructural polyprotein precursor in vitro resulted in rapid cotranslational cleavage at specific sites. The cleavage products were similar to those previously identified for Southampton virus, a highly related virus. We inactivated the virally encoded proteinase responsible for cleavage of the nonstructural polyprotein by mutation of the putative catalytic cysteine residue, which resulted in production of full-length polyprotein precursor. NV proteinase was expressed in Escherichia coli as a glutathione S-transferase fusion and purified by GST-affinity chromatography. Activity of the purified proteinase was demonstrated by incubation with the full-length precursor protein. trans cleavage of the nonstructural protein precursor resulted in cleavage products similar to those observed during cotranslational cleavage, however, at lesser efficiency. NV proteinase displayed sensitivities to cysteine and serine protease inhibitors similar to poliovirus 3C proteinase, suggesting that NV proteinase is a member of the viral cysteine proteinase family. We propose that the proteinase may play a regulatory role in viral replication.