【病毒外文文献】2003 Amino acid substitutions within the heptad repeat domain 1 of murine coronavirus spike protein restrict viral antig

Amino acid substitutions within the heptad repeat domain 1 of murine coronavirus spike protein restrict viral antigen spread in the central nervous system Jean C Tsai a Linda de Groot a Josefina D Pinon a Kathryn T Iacono a Joanna J Phillips a Su hun Seo a Ehud Lavi b and Susan R Weiss a a Department of Microbiology University of Pennsylvania School of Medicine Philadelphia PA 19104 6076 USA b Department of Pathology and Laboratory Medicine University of Pennsylvania School of Medicine Philadelphia PA 19104 6076 USA Received 18 December 2002 returned to author for revision 25 February 2003 accepted 7 March 2003 Abstract Targeted recombination was carried out to select mouse hepatitis viruses MHVs in a defined genetic background containing an MHV JHM spike gene encoding either three heptad repeat 1 HR1 substitutions Q1067H Q1094H and L1114R or L1114R alone The recombinant virus which expresses spike with the three substitutions was nonfusogenic at neutral pH Its replication was significantly inhibited by lysosomotropic agents and it was highly neuroattenuated in vivo In contrast the recombinant expressing spike with L1114R alone mediated cell to cell fusion at neutral pH and replicated efficiently despite the presence of lysosomotropic agents however it still caused only subclinical morbidity and no mortality in animals Thus both recombinant viruses were highly attenuated and expressed viral antigen which was restricted to the olfactory bulbs and was markedly absent from other regions of the brains at 5 days postinfection These data demonstrate that amino acid substitutions in particular L1114R within HR1 of the JHM spike reduced the ability of MHV to spread in the central nervous system Furthermore the requirements for low pH for fusion and viral entry are not prerequisites for the highly attenuated phenotype 2003 Elsevier Science USA All rights reserved Keywords Murine coronavirus Viral pathogenesis CNS infection MHV spike Viral fusion Introduction The spike glycoprotein found on the virion envelope of the coronavirus mouse hepatitis virus MHV is a major determinant of viral pathogenicity Phillips et al 1999 Entry into susceptible cells and cell to cell spread of MHV are mediated by the spike Spike is synthesized as an 180 to 220 kDa glycosylated precursor that is posttranslationally cleaved into two 90 kDa subunits S1 and S2 Frana et al 1985 Oligomeric spike is present on both viral and in fected cell membranes The receptor binding domain RBD of spike has been mapped to the amino terminal 330 amino acid residues of the S1 subunit Kubo et al 1994 which is though to form the globular head of the spike The RBD is responsible for the initial attachment of MHV to the cell surface receptors members of the CEACAM subclass of the CEA gene family Dveksler et al 1991 This initial bind ing event is thought to trigger a conformational change in spike that allows S2 the stalk portion of the peplomer to initiate fusion of the two membranes Gallagher and Buch meier 2001 Matsuyama and Taguchi 2002a Zelus et al 2003 A candidate fusion peptide domain has been identi fied within S2 Luo and Weiss 1998 Fig 1 however the actual fusion peptide for the MHV spike has not been definitively identified Within S2 there are two heptad re peat HR domains HR1 and HR2 which are structural features common to all coronavirus spikes as well as of Corresponding author Department of Microbiology 203A Johnson Pavilion University of Pennsylvania Philadelphia PA 19104 6076 Fax H110011 215 573 4858 E mail address weisssr mail med upenn edu S R Weiss R Available online at Virology 312 2003 369 380 0042 6822 03 see front matter 2003 Elsevier Science USA All rights reserved doi 10 1016 S0042 6822 03 00248 4 attachment fusion proteins of many other viruses DeGroot et al 1987 More recently a third heptad repeat region located to the N terminal side of HR1 was predicted within the MHV spike Gallagher and Buchmeier 2001 Krueger et al 2001 The HR domains are believed to form H9251 helical configurations which have been though to play a role in the oligomerization of the spike peplomer DeGroot et al 1987 Luo et al 1999 It has also been believed that these domains participate in the conformational change of the spike peplomer which allows for the insertion of the puta tive fusion peptide into the cell membrane and that interac tions between HR1 and RBD may play a role in fusion Krueger et al 2001 Matsuyama and Taguchi 2002b However recently it has been shown that the S1 subunit contains sequences responsible for oligomerization Lewicki and Gallagher 2003 suggesting that the confor mational changes that spike undergoes during the process of fusion may be quite different from previously postulated The entry route for MHV has not been clearly defined Previous studies have demonstrated that the wild type JHM strain previously referred to as MHV 4 see Materials and methods can utilize both endosomal and nonendosomal pathways for entry Nash and Buchmeier 1997 Wild type JHM fuses at neutral pH however persistent JHM infection of a neuronal cell line OBL21a gave rise to an acid dependent fusion variant OBLV60 OBLV60 is unable to induce fusion at neutral pH and enters cells via the endo somal route Gallagher et al 1991 Nash and Buchmeier 1997 Sequencing of the spike genes of JHM and OBLV60 demonstrated that there were eight amino acid differences between these two spike proteins Five of these substitutions were located in the S1 subunit of the spike gene However it was demonstrated that the change from neutral pH in duced fusion to low pH induced fusion was dependent on the amino acid substitutions Q1067H Q1094H and L1114R all within the first heptad repeat HR1 of the S2 subunit Gallagher et al 1991 see Fig 1 JHM is a highly neurovirulent MHV strain Dalziel et al 1986 Gallagher et al 1991 following infection of sus ceptible weanling mice with JHM there is widespread dis tribution of infection in the brain often resulting in fatal encephalitis In contrast the OBLV60 variant is neuroat tenuated infection with OBLV60 is limited mainly to the neurons of the mitral and glomerular layers of the olfactory bulbs and no fatal encephalitis is observed Pearce et al 1994 We wanted to determine whether the three amino acid substitutions in the HR1 of the OBLV60 spike were responsible for its low pH dependent route of entry into cells in vitro and or the attenuation and limited spread of the virus in vivo It was also important to determine whether these phenotypes could be dissociated from each other We were particularly interested in the substitution at residue L1114 as this residue has been substituted in viruses with altered phenotypes isolated in several studies Gallagher et al 1991 Wang et al 1992 Saeki et al 1997 We have previously used targeted RNA recombination to demonstrate the important role spike plays as a major de terminant of pathogenicity in the central nervous system CNS Phillips et al 1999 Subsequently we used recom binant viruses expressing chimeric spike proteins to dem onstrate that the RBD determines the specificity of receptor utilization Tsai et al 2003 Here we focus on the pheno type of a virus encoding the three HR1 amino acids that were previously shown to be sufficient to result in the low Fig 1 Schematic diagram of wild type and mutant JHM spike proteins The OBL V60 spike is shown with asterisks marking the positions of the eight amino acid substitutions within the spike from amino to carboxyl ends K107R N134T L141F D597N Q600K Q1067H Q1094H L1114R from Gallagher et al 1991 Also shown is the JHM spike with the locations of the cleavage site represented by the vertical line separating the S1 and S2 subunits as well as the heptad repeat HR and transmembrane TM domains The HR1 is enlarged to show the locations of the amino acid substitutions in the mutant proteins containing either three substitutions in HR1 HR1 or L1114R alone a candidate fusion peptide pep1 Luo and Weiss 1998 within HR1 is also indicated 370 J C Tsai et al Virology 312 2003 369 380 pH dependent requirement for cell to cell fusion Gallagher et al 1991 Thus we have selected recombinant viruses with a defined genetic background expressing the JHM spike protein encoding the amino acid substitutions in HR1 of OBLV60 Q1067H Q1094H and L1114R or L1114 alone The three substitutions are sufficient to confer a neuroattenuated phenotype to this recombinant virus which similar to OBLV60 displays virus spread restricted to the olfactory bulbs Interestingly the L1114R substitution alone is capable of causing in vivo attenuation and restricted spread while displaying a fusion phenotype close to wild type as well as the ability to enter cells at neutral pH Results Selection and in vitro phenotypes of recombinant viruses with amino acid substitutions in the HR1 of the spike protein We wanted to determine whether the three mutations in HR1 of the JHM spike Q1067H Q1094H and L1114R were sufficient to cause in vivo attenuation and restriction of antigen to the olfactory bulbs If so we then wanted to determine whether these in vivo phenotypes were dependent on the in vitro property of acid pH dependent fusion and entry Gallagher et al 1991 Nash and Buchmeier 1997 This was not possible to accomplish with OBLV60 due to the presence of five other substitutions within S1 Fig 1 and an unknown number of amino acid substitutions outside of the spike Thus we used targeted RNA recombination to select isogenic recombinant viruses expressing the JHM spike gene encoding either the three HR1 amino acid sub stitutions SJHM HR1 or the L1114R alone SJHM L1114R as well as a virus expressing wild type JHM spike SJHM RA59 Kuo et al 2000 Phillips et al 2001 See Fig 1 for a schematic of the spike proteins expressed by these viruses All these viruses had genes other than spike derived from MHV A59 Phillips et al 1999 and are described under Materials and methods SJHM RA59 replicates to relatively low titers in mouse L2 fibroblasts in vitro similar to wild type JHM Phillips et al 1999 2001 Compared to JHM or SJHM RA59 viruses OBLV60 and SJHM HR1 replicated to much higher titers while SJHM L1114R replicated to titers intermediate be tween those of SJHM RA59 and SJHM HR1 Fig 2 The replication characteristics of OBLV60 SJHM HR1 JHM and SJHM RA59 in OBL21a cells were similar to those in L2 cells data not shown An in vitro fusion assay was performed to compare the levels of cell to cell fusion induced by recombinant viruses SJHM RA59 SJHM HR1 and SJHM L1114R As shown in Fig 3 SJHM HR1 lacks the ability to induce cell to cell fusion at neutral pH This is consistent with previous data on the inability of OBLV60 to induce cell to cell fusion at neutral pH Gallagher et al 1991 and confirms the obser vation that the three amino acid substitutions in HR1 Q1067H Q1094H and L1114R present in both SJHM HR1 and OBLV60 are sufficient to confer a fusion negative phenotype at neutral pH SJHM L1114R expressing the L1114R substitution alone does induce significant levels of cell to cell fusion at neutral pH however fusion induced by this recombinant is less extensive than that induced by SJHM RA59 Fig 3 Thus the L1114R substitution is not sufficient to confer the nonfusogenic phenotype of SJHM HR1 Fig 2 Replication of wild type and mutant viruses in L2 cell culture L2 cells were infected in duplicate at an m o i of 2 PFU cell At the indicated times postinfection cells were lysed and virus titered by plaque assay Each point represents the logarithmic mean titer of duplicate samples A Cells were infected wild type JHM virus wtJHM OBLV60 or recom binant viruses SJHM RA59 expressing wild type JHM spike JHM or SJHM HR1 expressing the JHM spike with three HR1 substitutions HR1 B Cells were infected with the following recombinant viruses SJHM RA59 JHM SJHM HR1 HR1 or two independent isolates of SJHM L1114R expressing JHM spike with the L1114R substitution L1114R A L1114R B Fig 3 Virus induced cell to cell fusion of L2 cells L2 cell monolayers were infected in duplicate at an m o i of 1 PFU cell with recombinant viruses expressing either wild type or mutant spike proteins At 7 and 9 h postinfection the cells were fixed and visualized at H11003200 magnification using phase contrast microscopy Fusion expressed as the mean percentage of nuclei in syncytium is plotted with the error bar representing the standard deviation 371J C Tsai et al Virology 312 2003 369 380 To determine whether similar to OBLV60 the mutant recombinant viruses rely on the acidification of the endoso mal vesicles to trigger fusion we tested the effects of the lysosomotropic agent ammonium chloride on the replica tion of SJHM RA59 SJHM HR1 and SJHM L1114R in vitro Fig 4 Vesicular stomatitis virus VSV which is known to enter cells via the endosomal pathway and to require low pH for entry was used as a control In the presence of 20 mM ammonium chloride the replication of VSV was reduced about 10 000 fold at 12 and 24 h postin fection of L2 cells Fig 4 Similarly the replication of SJHM HR1 was reduced more than 1000 fold in the pres ence of 20 mM ammonium chloride Similar levels of in hibition were observed for OBLV60 Gallagher et al 1991 our data not shown In contrast only a minimal reduction in replication level was observed when either SJHM RA59 or SJHM L1114R infection of L2 cells was carried in the presence of ammonium chloride Statistical analysis dem onstrated however that SJHM L1114R was inhibited to a greater extent than SJHM RA59 Similar patterns of inhi bition were observed in infections carried out in the pres ence of 40 H9262M chloroquine data not shown Thus the L1114R substitution alone does not result in an absolute dependence on low pH for virus entry Recombinant viruses with HR1 mutations are neuroattenuated and restricted in spread in the brain despite efficient replication Wild type JHM or recombinant viruses encoding the JHM spike SJHM RA59 are highly neurovirulent and dis seminate throughout the murine brain often causing mor tality Lavi et al 1990 Barnett et al 1993 Phillips et al 1999 OBLV60 is a highly neuroattenuated mutant of JHM Pearce et al 1994 To determine whether the three amino acid substitutions in the HR1 of OBLV60 or L1114R alone are sufficient to confer a neuroattenuated phenotype mice were inoculated either intracranially ic or intranasally in with serial dilutions of SJHM RA59 SJHM HR1 and SJHM L1114R and observed for disease and death Inocu lation with SJHM HR1 up to 5000 PFU ic and up to 10 5 PFU in or with SJHM L1114R up to 10 4 PFU ic did not result in any mortality throughout the 21 to 28 days the mice were observed In contrast mice inoculated with doses as low as 1 PFU ic or 1000 PFU in of SJHM RA59 exhibited clinical signs of disease such as hunched posture ruffled fur and abnormal gait with occasional paralysis at 4 5 days postinfection Thus recombinant viruses expressing either all three HR1 substitutions or L1114R alone induced clinically silent infections and are basically avirulent Even L1114R alone is sufficient to highly attenuate MHV To begin to understand the mechanism of the attenuation of SJHM HR1 and SJHM L1114R we examined the ability of these viruses to replicate and express viral antigen in the brain We used the in route of inoculation because it is a more natural route and detection of viral antigen following in inoculation is more discreet than following ic inoculation Thus mice were inoculated in with 5000 PFU each of SJHM RA59 SJHM HR1 or SJHM L1114R and titers of infectious virus in the brain at various times postinfection were determined by plaque titration Patterns of infection with all of these viruses were similar in terms of kinetics and extent of replication viral replication in the brain peaked at 5 days postinfection Fig 5 SJHM HR1 dis played slightly higher viral titers at 5 and 7 days postinfec tion as compared to SJHM RA59 however the difference Fig 4 Effect of lysosomotropic agents on viral replication in L2 cells L2 cell monolayers were treated for 1 h with 20 mM ammonium chloride and then infected with either vesicular stomatitis virus VSV or recombinant MHVs expressing wild type or mutant spikes as indicated The presence of 20 mM ammonium chloride was continuous throughout the infection At 0 12 and 24 h postinfection the level of infectious titer was measured by plaque assay and expressed as the difference between logarithmic mean H9004 log 10 PFU ml titer of the treated and untreated samples four independent samples per virus The error bars represent the standard deviations Fig 5 Replication of virus in brains of infected mice Four week old C57B1 6 mice were inoculated intranasally with 5000 PFU of recombinant viruses expressing wild type or mutant JHM spikes as indicated At the indicated times mice three per time point were sacrificed and brains harvested and homogenized Viral titers were determined by plaque assay of brain homogenates and expressed as the logarithmic mean titer per gram of tissue homogenate The standard deviations of the mean are plotted as error bars The limit of detection of this assay was 2 7 log 10 PFU g 372 J C Tsai et al Virology 312 2003 369 380 was less than 1 0 log 10 unit SJHM L1114R replicated less efficiently than SJHM RA59 however the difference be tween the titer of infectious virus in the brains of mice infected with these two viruses at 5 days postinfection was not significant two tailed t test P H11021 0 06 and P H11021 0 11 for each of the two SJHM L1114R isolates Thus although there is a striking difference in virulence after in inoculation with these viruses this difference cannot be explained by the extent of viral replication in the brains of infected animals To further investigate the attenuated phenotype of these mutants we used indirect immunofluorescence to visualize viral antigen in the brains of SJHM HR1 and SJHM L1114R infected animals Brain sections were obtained from mice infected in with 5000 PFU of SJHM RA59 SJHM HR1 and SJHM L1114R at 5 days postinfection These sections were labeled with polyclonal antisera di rected against MHV structural proteins and then detected using a biotin avidin labeled secondary antibody as de scribed under Materials and methods The difference in the patterns of viral antigen distribution observed in the brains of animals infected with SJHM RA59 and the viruses ex pressing mutant spike was striking The distribution of SJHM RA59 viral antigen was widespread Viral antigen was detected in olfactory bulbs as well as in other regions of the brain from the basal forebrain to the brain stem Phillips et al 1999 Fig 6 shows sections from the basal forebrain and the olfactory bulbs However in mice in fected with SJHM HR1 or SJHM L1114R viral antigen was primarily detected in the mitral layer of the olfactory bulbs and in the adjacent glomerular and external plexiform layers see below Outside the olfactory bulbs there were only occasional antigen positive cells The majority of an tigen positive cells looked morphologically similar to neu rons Thus both mutants showed patterns of antigen distri bution similar to OBLV60 Pearce et al 1994 this Fig 6 Viral antigen distribution in the brains of infected mice Indirect immunofluorescent staining of viral antigen with rabbit anti MHV A59 serum was carried out on paraffin embedded sagittal sections from brains collected on day 5 postinfection from mice infected with SJHM RA59 expressing wild type JHM spike a b SJHM HR1 expressing JHM spike with three HR1 substitutions c d or SJHM L1114R expressing the JHM spike with the L1114R substitution alone e f Shown are regions of the olfactory bulbs a c e and the basal forebrain d f g The secondary antibody was TRITC conjugated swine anti rabbit antibody Bars 20 H9262m 373J C Tsai et al Virology 312 2003 369 380 indicates that the L1114R substitution alone as well as the three HR1 substitutions is sufficient to restrict antigen ex pression to the olfactory bulbs We qu