Noroviruses belong to the virus family Caliciviridae, a group of small icosahedral, nonenveloped viruses that possess a polyadenylated single-stranded positive-sense RNA genome (Jiang et al., 1990). Caliciviruses are diverse in their associated disease syndromes, host specificity, and genetic characteristics. However, a striking property shared by nearly all caliciviruses is their global and ubiquitous distribution in their respective hosts. The human noroviruses are highly successful in this characteristic, with evidence for multiple infections throughout life in the majority of individuals (Jing et al., 2000; Smit et al., 1997; O’Ryan et al., 1998).

Human noroviruses are genetically diverse, and early studies noted the presence of two major genetic groups, now called genogroups (G) (Ando et al., 1994; Wang et al., 1994). The norovirus RNA genome (approx. 7.5–7.6 kb in size) is organized into three major open reading frames (ORFs 1–3). The length of the RNA genome can vary among strains, and the nucleotide boundaries of the ORFs are shown for representative GI (Norwalk virus) and GII (Lordsdale virus) human norovirus strains in Fig. 3.5.1A.

The RdRp of the Caliciviridae, like that of other positive strand RNA viruses, does not have an editing function, resulting in a comparably higher error rate per genome per replication cycle than DNA viruses (Duffy et al., 2008). The subsequent nucleotide mutations can lead to fixed substitutions at the viral population level, a key component in adaptation. Genetic drift may best be illustrated by the evolutionary patterns of the predominant GII.4 noroviruses that show the emergence of a new pandemic variant every few years (Fig. 3.5.3). The genetic changes of an emerging variant are often clustered in epitopes in the surface-exposed P2 region of the capsid, suggesting a role for herd immunity and selective antibody pressure (antigenic drift) in its evolution (reviewed in Debbink et al., 2012). In immunocompromised individuals, the population of RNA molecules is even more diverse, presumably because of the absence of selective pressure to drive escape from antibodies and other immune responses (Bull et al., 2012).

Deep sequencing of clinical specimens has shown that norovirus strains exist as a mixed population of genetically closely related RNA variants in the host, with a predominant RNA sequence that can comprise >60% of the total population (Bull et al., 2012). However, transmission of this mixed population into a new host may create a fitness bottleneck, and the emergence of a minor population can give rise to new population diversity (Bull et al., 2012). The highly infectious nature of norovirus and its efficient spread by person-to-person transmission (Koopmans, 2008) creates ample opportunity for the emergence of new variants via this mechanism. A recent study of the norovirus genomic RNA population in an immunocompromised host showed evidence for the presence of novel GII.4 variants with the potential for transmission into a susceptible population (Vega et al., 2014a). The impact of transmission bottlenecks in norovirus epidemiology will require additional viral metagenomics studies. (See Chapter 5.1.)

During calicivirus replication, two major positive-sense RNA species are generated, with one corresponding to the full-length genome (approx. 7.6 kb in size) and the other to a 3′-end coterminal subgenomic region (approx. 2.6 kb in size) (Neill and Mengeling, 1988). The abundant subgenomic RNA serves as a template for translation of the VP1 and VP2 capsid proteins. There is evidence that during coinfection of a cell with two caliciviruses, the two genomic RNAs replicate in proximity so that recombination events can readily occur (Abente et al., 2013), likely due to template switching by the RdRp (Bull et al., 2005). A recent in vitro study found that after coinfection of cells with two parental feline calicivirus strains, recombination was detected as early as 6 h after infection at a rate of 6.8 × 10(−6) single direction recombinant genomes generated per parental virus genome (Symes et al., 2015). Although the recombination rate is not known for the human noroviruses, epidemiological studies suggest that recombination is not a rare event (Lu et al., 2015). A hotspot for recombination in the noroviruses exists at the junction between ORF1 and ORF2 as a breakpoint (BP) for template switching, yielding new combinations of VP1 and polymerase that may confer adaptive advantages to the new strain (Bull et al., 2005; Mans et al., 2014; Shen et al., 2012) (Fig. 3.5.3). For example, the “GII.Pb” polymerase has been detected globally for nearly a decade and has been associated with at least five different GII capsid proteins (GII.1, GII.2, GII.3, GII.4, and GII.13) (Lim et al., 2013; Reuter et al., 2005). The GII.4 noroviruses may have additional recombination BPs within ORF2 and between ORF2 and 3, leading to new combinations of VP1 and VP2 (Kamel et al., 2009; Eden et al., 2013). Evidence for additional BPs in ORF1 was found in the murine norovirus (MNV) VPg, RdRp, and protease genes following coinfection of mice with two distinct MNV strains (Zhang et al., 2015). A common feature of the MNV BP sites was high sequence identity between the two parental strains, strongly suggesting that any highly conserved region may, in theory, be a site for template switching and recombination.

Norovirus genotypes vary among geographical locations, seasons, and years. Outbreaks occur more often in the winter months of temperate climates (Ahmed et al., 2013). A link has been proposed between higher particle stability and low absolute humidity (below 0.007 kg/kg air), which could explain, in part, the increase of norovirus transmission and illness during cold months when these conditions predominate (Colas de la Noue et al., 2014). The marked diversity and unpredictable pattern of norovirus genotypes is reflected in epidemiologic surveys conducted across the globe. The diversity is especially striking in environmental studies, reflecting the ubiquitous distribution of these viruses (Fernandez et al., 2012; van den Berg et al., 2005). Two consistently observed features in most clinical surveys are the predominance of GII compared to GI viruses, and the higher prevalence of the GII.4 genotype (Matthews et al., 2012; Vega et al., 2014b). In contrast, studies of environmental samples, waterborne outbreaks, or shellfish may find a predominance of GI strains (Matthews et al., 2012; Lowther et al., 2012). The molecular basis for this variation in genotypic distribution is not yet known, but the emergence of a pathogenic strain likely involves properties of the virus (eg, physicochemical features, replicative fitness) and the susceptible host (eg, immune status, HBGA profile).

Serologic surveys have shown evidence for the acquisition of both GI and GII norovirus-specific antibodies early in life in most populations, although regional differences in rates of antibody acquisition and specificity have been reported (Jiang et al., 2000; Menon et al., 2013; Gray et al., 1993). Susceptibility to norovirus disease spans all age groups, and this lifetime susceptibility enables the occurrence of sharp gastroenteritis outbreaks in a wide range of settings involving otherwise healthy individuals. Severe, life-threatening norovirus illness occurs most often in the young and old, suggesting that both, immunological immaturity in the young and immunosenescence in the elderly, hamper virus clearance. This is consistent with the link between immunocompromised patients and chronic norovirus infection, in which impaired immune function affects virus clearance (Vega et al., 2014a, reviewed in Bok and Green, 2012).

Noroviruses are transmitted by the fecal–oral route, and shed in the feces. It is estimated that approximately 14% of norovirus outbreaks are foodborne (Verhoef et al., 2015). The virus is apparently most infectious in the early phase of the acute illness, when the symptoms of vomiting and/or diarrhea are most severe and norovirus shedding is at its peak (Zelner et al., 2013). The infectious dose has been estimated in volunteer studies. In one such study, aggregation of Norwalk virus particles was detected and the ID₅₀ dose of disaggregated virus was estimated to be approximately between 18 and 1015 genome equivalents (Teunis et al., 2008). In a second dose-response study of Norwalk virus, a higher ID₅₀ dose of 1320–2800 genome equivalents was reported, with dose depending on secretor and blood group status (Atmar et al., 2014). Norovirus is shed to similar levels (as measured by RT-qPCR) in symptomatic and asymptomatic individuals, making it problematic to use virus genome copy numbers as a marker of disease (Teunis et al., 2014).