Avian influenza: mixed infections and missing viruses

Abstract

A high prevalence and diversity of avian influenza (AI) viruses were detected in a population of wild mallards sampled during summer 2011 in California, providing an opportunity to compare results obtained before and after virus culture. We tested cloacal swab samples prior to culture by matrix real-time PCR, and by amplifying and sequencing a 640bp portion of the hemagglutinin (HA) gene. Each sample was also inoculated into embryonated chicken eggs, and full genome sequences were determined for cultured viruses. While low matrix Ct values were a good predictor of virus isolation from eggs, samples with high or undetectable Ct values also yielded isolates. Furthermore, a single passage in eggs altered the occurrence and detection of viral strains, and mixed infections (different HA subtypes) were detected less frequently after culture. There is no gold standard or perfect reference comparison for surveillance of unknown viruses, and true negatives are difficult to distinguish from false negatives. This study showed that sequencing samples prior to culture increases the detection of mixed infections and enhances the identification of viral strains and sequences that may have changed or even disappeared during culture.

Figure 1

Matrix real-time PCR and HA subtyping results for paired samples (P, V) from 32 birds before and after inoculation in ECE. HA subtypes in the original VTM samples were determined by 640PCR/Sanger sequencing, while subtypes in allantoic fluid (ALF) were determined by both 640PCR and M-RTPCR/NGS. NT = not tested. Neg = matrix Ct >45 or HA subtype not detected.

Figure 2

Multi-segment RT-PCR results for allantoic fluid (ALF) samples that were positive by matrix real-time PCR. Full genome sequences were determined by next generation sequencing (NGS) for samples even if they lacked detectable bands. ^a Agarose gel profiles were examined for the presence/absence of individual segment bands. If all segment bands were absent on the gel, the sample was not put through NGS. Absent bands are highlighted. P = band present, A = band absent, NT = not tested by NGS and reported as negative in Figure 1. ^b The PB1 and PB2 segments are the same size and therefore cannot be resolved on agarose gels. ^c Sample 2578V was found to have two NS sequences so was designated “H3N8 mix.”