Lab partners: two viruses work together only in cell culture

From the Bloom Lab, Basic Sciences Division

Viruses such as influenza mutate so rapidly that they exist as genetically diverse “quasispecies”, even within a single patient. This property has important implications for understanding viral infections, and explains why we need a new flu shot each year. The Bloom Laboratory in the Basic Sciences Division studies the evolutionary dynamics of viral populations using deep sequencing approaches. In previous work, they showed that different influenza viruses can work together during infection—two viruses differing at a single site on a surface protein grew better together than individually. “However, it remained unclear whether the viruses also cooperate in actual human infections, or only in the lab”, says Dr. Jesse Bloom. Since clarifying this point is key to understanding what evolutionary forces select for viral cooperation, Dr. Bloom and the lead author of the original paper, graduate student Katherine Xue, initiated a follow-up study. Their results, which were recently published in mSphere, revealed that the cooperative interaction they observed was a consequence of laboratory selection.

Influenza binds to and enters host cells via sialic acid receptors on the host cell surface. In order for new viral particles to be efficiently released after viral replication, influenza cleaves sialic acid receptors using one of its surface proteins, neuraminidase (NA). Interestingly, “a particular mutation in NA had been observed by labs all over the world that were growing clinical flu samples in cell culture”, says Katherine Xue. This variant, known as G151, carries an aspartic acid (D) to glycine (G) mutation at position 151 in NA that renders the protein unable to cleave sialic acid receptors. As a result, G151 grows much worse than D151 in cell culture. To the researchers’ surprise, however, a mixed population of D151 and G151 viruses grew significantly better than D151 alone. A possible explanation for this phenomenon is that G151’s ability to bind sialic acid receptors without cleaving them makes it especially good at host cell entry, while D151’s cleavage activity promotes host cell exit and can compensate for G151’s deficiency. 

Patient samples are routinely passaged in the laboratory prior to sequencing. This process may introduce mutations that were not present during the original infection. Image provided by Katherine Xue

In the original study, it was unclear whether D151 and G151 viruses actually coexist outside the lab, in real-world influenza infections. Other research groups had only detected G151 in samples that had been “passaged”, meaning that they had undergone multiple rounds of replication in the laboratory setting after being isolated from patients (see Figure). Because the passaging process may allow mutations to arise that were not present during the actual infection, the authors needed access to original samples to definitively establish whether G151 exists in natural viral populations. “In collaboration with public health officials at UW Medicine and the Washington State Department of Health, we were able to obtain the original nasal swab samples for several viral isolates that had the cooperating variant after being grown in the lab”, explains Ms. Xue.

To determine whether G151 viruses were present prior to passaging, the authors performed deep sequencing on the nasal swab samples, which would detect G151 even if it was present in extremely low numbers compared to D151 or other variants. “We did not find it!” recalls Dr. Bloom. The G151 variant was not present at levels higher than would be expected from library preparation or sequencing errors in any of the samples tested.

The finding that cooperation between D151 and G151 viruses is unlikely to be significant in natural infections serves as a reminder that the laboratory environment does not always represent real-world conditions. In this case, there are many possible explanations as to why cooperation is selected for only in the lab. For example, laboratory infections are often performed with higher viral densities than might occur during natural infections, and D151/G151 cooperation is only expected to occur when the same cell is infected by both variants. In addition, “bottlenecks during flu transmission in nature may simply be too narrow for cooperating groups of viruses to survive,” suggests Dr. Bloom. Moving forward in the viral evolution field, it will be important to sequence clinical samples as soon as they are isolated to minimize the introduction of artificial selection.

Xue K, Greninger AL, Perez-Osorio A and Bloom JD. 2018. Cooperating H3N2 Influenza Virus Variants Are Not Detectable in Primary Clinical Samples. mSphere. 3(1).

This research was supported by the Fannie and John Hertz Foundation, National Science Foundation, National Institutes of Health, Howard Hughes Medical Institute and the Simons Foundation.