By serologic screening of sentinel mice housed in several NIAID animal facilities, we found evidence initially for norovirus infection, and in addition, we were able to detect the virus by RT-PCR in representative mice 
. This study focused on the further characterization of MNV strains isolated from mice housed in three of the animal facilities (). While all facilities housed mice from the same commercial vendors, each facility also housed mice received from separate non-commercial sources. Ten MNV strains (designated as the NIH strains) were isolated following passage in RAW264.7 cells, and their full-length genomes were sequenced. The genomes were similar in length (7382 or 7383 nt) and organized into three major open reading frames (ORFs): ORF1 (nt 6–5069) encoding the nonstructural polyprotein, ORF2 (nt 5056–6681) encoding the major capsid protein, VP1, and ORF3 (nt 6681–7307) encoding the minor capsid protein, VP2 (). In addition, the genomes contained a predicted conserved fourth ORF (ORF4) spanning nt 5069–5710 that would encode a putative protein of 23.6 kDa 
Murine norovirus strains isolated from NIH mice.
Sequence analysis of NIH MNV strains.
Comparison of the nucleotide sequences showed a close genetic relationship among the NIH MNV strains, with a range of 87.2–99.7% overall nucleotide identity. The nucleotide identities with the prototype strain, MNV-1, ranged from 87.4–89.1%, and with all other full-length MNV sequences available in GenBank, the identities ranged from 87.4–99%. Comparison of the deduced amino acid sequences among the NIH strains showed that the proteinase was the most highly conserved protein (approximately 97% similarity), while the VP1 (major capsid protein) showed the most variability (as low as 89.6% similarity). Analysis of the MNV-1 VP1 capsid protein showed that the shell (S) region of the capsid protein (amino acid residues 49 to 218) was 100% conserved among the NIH MNV strains while the protruding (P) region (amino acids 229 to 537) was most variable, consistent with data reported for other highly-related murine noroviruses 
. The P2 domain (amino acids 278–415) within the P region is exposed predominantly on the surface of the virion and has been proposed as the major site of immune pressure and receptor binding for calicivirus virions 
. Consistent with this, passage of the virulent MNV-1 strain in cell culture yielded virus in which lysine was changed to glutamic acid at position 296 in the P2 domain. This single amino acid substitution resulted in loss of virulence of the MNV-1 strain in STAT1-deficient mice (17). Of interest, a glutamic acid residue at position 296 was conserved in the cell-culture adapted NIH MNV strains in this study, as well as the NIH-2409 virus prior to adaptation (data not shown) 
. A leucine residue at position 386 involved in the formation of a neutralization epitope in MNV-1 
was conserved in the P2 domain of the NIH MNV strains (data not shown).
The replication strategy for murine norovirus involves the proteolytic processing of the ORF1 polyprotein into several nonstructural proteins by the virus-encoded cysteine proteinase (Pro). The proteinase mediates five cleavages at dipeptide sites 341
, and 1177
to produce six proteins: NS1-2 (Nterm), NS3 (NTPase), NS4 (p18), NS5 (VPg), NS6 (Pro) and NS7 (Pol) 
. Alignment of the deduced ORF1 polyprotein sequences of the NIH MNV strains with that of MNV-1 showed that these dipeptide cleavage sites were conserved in location, but with some minor variation in the P1′ positions at amino acid residues 706 and 871 ().
A phylogenetic analysis of the complete genome sequences of the NIH MNV strains was performed to compare their relatedness with that of other murine noroviruses (). The NIH strains clustered with Genogroup V murine noroviruses, but did not share complete identity with other strains reported thus far. Eight of the strains formed a distinct monophyletic group (NIH- 4421, 2747, 2409, 2411, 2410, 2750, 4428, and 4431, with nucleotide variation not exceeding 1%). Strain NIH-A114 did not cluster with other viruses and formed a distinct branch. Strain NIH-D220 clustered most closely with a phylogenetic group that included several Charles River (CR) and Washington University (WU) strains. The three genetic groupings of the NIH strains corresponded to their isolation from three distinct animal facilities in the institute at different times ().
Phylogenetic analysis of NIH MNV viral genomes in comparison with representative viruses previously published.
A previous study had identified amino acid substitutions in the NS4 and VP1 proteins that emerged during cell-culture passage of the MNV-1 strain. In this study, we followed the evolution of MNV NIH-2409 during cell culture passage to examine whether similar substitutions occurred. Strain NIH-2409 was isolated from a RAG2−/− mouse in a facility (B14) that has operated for more than 20 years, and that accepts vendor animals and transfer animals from within NIH as well as imported animals. The mesenteric lymph node tissue was homogenized and added to a monolayer of RAW264.7 cells and incubated for 7 days. There was no visible cytopathic effect (CPE) in RAW264.7 cells following the initial 7-day infection period. However, an immunofluorescence assay showed the presence of occasional cells with a perinuclear, localized area of positive reactivity with the MNV-1-specific antibodies (as indicated by arrows in ) that was not present in the mock-infected RAW264.7 cells (). The passage 1 (P1) material was added to a new monolayer and the visible CPE increased with each subsequent cell passage. After three passages, the material was subjected to plaque purification, and small plaques were observed (). Following two additional rounds of plaque purification (selecting large plaques), the plaque morphology resembled MNV-1 ().
Serial passage of MNV NIH-2409 in RAW264.7 cells.
The “consensus sequence” of the genome of the virus in the original MLN specimen (P0) was determined by the direct sequencing of RT-PCR products that spanned the genome. Only a few nucleotide changes were observed between the first two serial passages (P1 and P2) and after three rounds of plaque purification (3xPP), consistent with previous studies of MNV-1 
(). At nucleotide position 2166, a mixed population was detected that would result in either a serine or alanine at amino acid residue 721 in the NS4 region of the ORF1 polyprotein. Following three rounds of plaque purification of P3, an alanine became the predominant sequence at position 721. Of note, no adaptive amino acid mutations were observed in the VP1 (capsid) protein for this strain.
Sequence analysis of viral genomes during passage of MNV NIH-2409 in cell culture.