CoEvolution of Type A Influenza Virus and Its Hosts

1
Co-Evolution of Type A Influenza Virus and Its
Hosts
Stephen Huff1, Catherine Putonti1,2, Chen Feng2,
Yuriy Fofanov1,2 (1) Dept. of Biology and
Biochemistry, University of Houston, Houston, TX,
77204 (2) Dept. of Computer Science, University
of Houston, Houston, TX, 77204
Abstract Coevolution is not only a likely factor
in the emergence of infectious disease but also
the perpetual survival of many of the infectious
diseases within the population. Many viral and
bacterial pathogens have been found to mimic the
function of host proteins as a means of
manipulating the host for the pathogens benefit.
There has recently been evidence of a bias in
nucleotide usage within the genomic sequence of
several bacteriophages and eukaryotic viruses
corresponding to their host range. Furthermore,
the authors recently observed that the genomic
sequences of several viral pathogens that infect
humans appear to be closer to the human genome
than their near-neighbors that do not infect
humans. In order to assess if specificity to a
particular host can be observed at the genetic
level, we have measured the distance between
all available influenza type A segment sequences
and the human genomic sequence. Here the
distance is quantified as the average number of
mutations necessary to convert each subsequence
(length 16-20) present in each influenza segment
sequence to the nearest sequence present in the
human genomic sequence. As a result of these
computations, the differences in the distance
from the human genome between strains isolated
from different host organisms (including human,
avian, and swine, amongst others) was found to be
statistically significant.
Results
Distance values plotted for 3 segments of the
influenza virus separating those segments which
were isolated from human and those segments
isolated from avian hosts.
Distance values plotted for 3 segments of the
influenza virus by serotypes for those segments
which were isolated from avian hosts.
statistically significant.
Distance values plotted for 3 segments of the
influenza virus by year of isolation including
all serotypes and all segment 6 genomes isolated
from all hosts.
Distance values plotted for 3 segments of the
influenza virus by serotypes for those segments
which were isolated from human hosts.
Segment 2
In many cases, each peak corresponds to a
separate distribution of the D related to
separate years of isolation. For segments 2 and
4, the distance tend to move closer to the
human genome over time. Results for the
remaining segments appear to be ambiguous, with
only moderate trends apparent in the plots.
Curiously, segment 2 of the H3N2 serotype virus
produced an exceptionally large peak that
appeared within the last ten years shifted
further away from the human genome, though the
implications of this remain to be elucidated.
Each serotype appears to have a definitive range
of distances from the human. The multiple peaks
observed when all serotype sequences were
considered is actually a result of the multiple
individual serotypes. While all of these segment
sequences were isolated from human hosts,
different serotypes, e.g. H1N1 in segments 2 and
4 appear significantly closer to human than
others regardless of the number of samples
(sequences) available.
Segment 4
Segment 6
Conclusions Segment by segment genomic distance
(d) analysis of 20,294 complete influenza type A
genome sequences (including all serotypes and
hosts) revealed co-evolutionary trends within the
virus. In many cases, the viral genome appeared
to move closer to the human genome over time,
especially in the case of certain serotypes that
are prevalent disease agents in the human host,
including the H3N2, H1N1, H5N1 and H9N2
serotypes. Using the metric introduced here it
is possible to isolate regions of segments or
entire segments that have possibly chosen to
mimic the host genomic sequence at the nucleotide
level in an effort to either evade host detection
or to increase virulence for the particular host
organism.
Acknowledgements Partial support of this work.
Additional funding was provided by the Department
of Homeland Security Science and Technology
Directorate, award NBCHC070054, and the Texas
Learning Computation Center..