Mouse hepatitis virus (MHV) isolates JHM.WU and JHM.SD promote severe central nervous system disease. However, while JHM.WU replicates robustly and induces hepatitis, JHM.SD fails to replicate or induce pathology in the liver. These two JHM variants encode homologous proteins with few polymorphisms, and little is known about which viral proteins(s) is responsible for the liver tropism of JHM.WU. We constructed reverse genetic systems for JHM.SD and JHM.WU and, utilizing these full-length cDNA clones, constructed chimeric viruses and mapped the virulence factors involved in liver tropism. Exchanging the spike proteins of the two viruses neither increased replication of JHM.SD in the liver nor attenuated JHM.WU. By further mapping, we found that polymorphisms in JHM.WU structural protein M and nonstructural replicase proteins nsp1 and nsp13 are essential for liver pathogenesis. M protein and nsp13, the helicase, of JHM.WU are required for efficient replication in vitro and in the liver in vivo. The JHM.SD nsp1 protein contains a K194R substitution of Lys194, a residue conserved among all other MHV strains. The K194R polymorphism has no effect on in vitro replication but influences hepatotropism, and introduction of R194K into JHM.SD promotes replication in the liver. Conversely, a K194R substitution in nsp1 of JHM.WU or A59, another hepatotropic strain, significantly attenuates replication of each strain in the liver and increases IFN-β expression in macrophages in culture. Our data indicate that both structural and nonstructural proteins contribute to MHV liver pathogenesis and support previous reports that nsp1 is a Betacoronavirus virulence factor.

IMPORTANCE The Betacoronavirus genus includes human pathogens, some of which cause severe respiratory disease. The spread of severe acute respiratory syndrome coronavirus (SARS-CoV) and Middle East respiratory syndrome coronavirus (MERS-CoV) into human populations demonstrates the zoonotic potential of emerging coronaviruses, and there are currently no vaccines or effective antivirals for human coronaviruses. Thus, it is important to understand the virus-host interaction that regulates coronavirus pathogenesis. Murine coronavirus infection of mice provides a useful model for the study of coronavirus-host interactions, including the determinants of tropism and virulence. We found that very small changes in coronavirus proteins can profoundly affect tropism and virulence. Furthermore, the hepatotropism of MHV-JHM depends not on the spike protein and viral entry but rather on a combination of the structural protein M and nonstructural replicase-associated proteins nsp1 and nsp13, which are conserved among betacoronaviruses. Understanding virulence determinants will aid in the design of vaccines and antiviral strategies.