Skip to main content

Full text of "Local-Scale Diversity and Between-Year "Frozen Evolution" of Avian Influenza A Viruses in Nature."

See other formats

OPEN 3 ACCESS Freely available online 


Local-Scale Diversity and Between-Year "Frozen rg\ 
Evolution" of Avian Influenza A Viruses in Nature asss* 

Alexander Nagy 1,2 *, Lenka Cerm'kova 1 , Helena Jifincova 2 , Martina Havh'ckova 2 , Jitka Horni'ckova 1 

1 State Veterinary Institute Prague, National Reference Laboratory for Avian Influenza and Newcastle Disease, Laboratory of Molecular Methods, Prague, Czech Republic, 

2 National Institute of Public Health, Centre for Epidemiology and Microbiology, National Reference Laboratory for Influenza, Prague, Czech Republic 


Influenza A virus (IAV) in wild bird reservoir hosts is characterized by the perpetuation in a plethora of subtype and 
genotype constellations. Multiyear monitoring studies carried out during the last two decades worldwide have provided a 
large body of knowledge regarding the ecology of IAV in wild birds. Nevertheless, other issues of avian IAV evolution have 
not been fully elucidated, such as the complexity and dynamics of genetic interactions between the co-circulating IAV 
genomes taking place at a local-scale level or the phenomenon of frozen evolution. We investigated the IAV diversity in a 
mallard population residing in a single pond in the Czech Republic. Despite the relative small number of samples collected, 
remarkable heterogeneity was revealed with four different IAV subtype combinations, H6N2, H6N9, H1 1 N2, and H1 1 N9, and 
six genomic constellations in co-circulation. Moreover, the H6, H1 1, and N2 segments belonged to two distinguishable sub- 
lineages. A reconstruction of the pattern of genetic reassortment revealed direct parent-progeny relationships between the 
H6N2, H11N9 and H6N9 viruses. Interestingly the IAV, with the H6N9 subtype, was re-detected a year later in a genetically 
unchanged form in the close proximity of the original sampling locality. The almost absolute nucleotide sequence identity 
of all the respective genomic segments between the two H6N9 viruses indicates frozen evolution as a result of prolonged 
conservation in the environment. The persistence of the H6N9 IAV in various abiotic and biotic environmental components 
was also discussed. 

Citation: Nagy A, Cerm'kova L, Jifincova H, Havlickova M, Horni'ckova J (2014) Local-Scale Diversity and Between-Year "Frozen Evolution" of Avian Influenza A 
Viruses in Nature. PLoS ONE 9(7): el 03053. doi:1 0.1 371 /journal.pone.01 03053 

Editor: Martin Beer, Friedrich-Loeffler-lnstitut, Germany 

Received January 20, 2014; Accepted June 25, 2014; Published July 30, 2014 

Copyright: © 2014 Nagy et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits 
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. 

Funding: The work was supported in part by research grant IGA NT 12493-3/201 1, Ministry of Health of the Czech Republic. No additional external funding was 
received for this study. The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. 

Competing Interests: The authors have declared that no competing interests exist. 

* Email: 


Influenza A virus (IAV) is a member of the genus Orlhomyx- 
oviridae with a genome composed of eight distinct negative-sense 
RNA segments. The segmented genome and the lack of 
proofreading activity of the virus RNA polymerase provide a 
basis for extreme genetic diversity. 

Monitoring studies carried out during the last two decades 
provided a large body of knowledge regarding the ecology of IAV 
in wild birds. Multiyear studies utilizing data from various bird 
populations sampled in different areas worldwide allowed us to 
identify the reservoir bird species and revealed the main ecological 
characteristics like the prevalence, subtype diversity, seasonality, 
environmental persistence, transmission routes, geographic distri- 
bution, intercontinental exchange, etc. [1-6]. Comprehensive 
evolutional analyses of the IAV genome have suggested extensive 
and ubiquitous reassortment [7,8] and rapid evolutionary dynam- 
ics in the avian reservoir [9]. 

Nevertheless, there are still aspects of IAV evolution and 
ecology in wild birds which have remained to be elucidated. For 
example, contrary to the IAV prevalence and subtype variety 
reported in the monitoring studies, only a few projects were 
focused on revealing the subtype diversity, genomic complexity, 
and dynamics of genetic interactions between the viruses occurring 
at the local-scale level, i.e. in the wild bird population residing in a 
single locality or in a single pond [10-12]. 

Another sparsely reported and not fully understood issue is the 
phenomenon of "frozen evolution" of the influenza virus in 
nature. The hypothesis of frozen evolution or frozen replication 
(both terms are used in the literature) is used to explain occasional 
evidence of anachronistic influenza virus genomes or genomic 
segments [13-16]. Anachronistic sequences exhibit unusually high 
or absolute sequence identity at the nucleotide level despite 
relative distance in reported time of detection. Therefore, it 
appears that they are being "frozen in time". This is the main 
difference to the hypothesis of evolutionary stasis [1,17]. Accord- 
ing to this hypothesis the IAV virus proteins are under strong 
purifying selection in avian reservoir as a result of adaptive 
optimum. Nevertheless, continuous circulation in the wild bird 
population results to continuous accumulation of nucleotide 
changes. However, from of these changes the synonymous 
mutations are selected predominantly. Hence, purifying selection 
results to circulation of phenotypically equivalent virus proteins [8] 
which appears to be at apparent stasis, despite the continuous 
accumulation of mutations at the nucleic acid level. Taking 
together, the two hypotheses relates to two different levels. Frozen 
evolution relates to the nucleotide (genotype) while the evolution- 
ary stasis to the amino acid (phenotype) conservation respectively. 

So, the frozen evolution is a hypothesis explaining significandy 
lover nucleotide mutation rate than expected [9] . The mechanism 
of frozen evolution therefore requires some kind of environmental 



July 2014 | Volume 9 | Issue 7 | e103053 

Local-Scale Diversity and Frozen Evolution of Influenza A Viruses 

persistence without the ability to replicate, accumulate nucleotide 
changes, and evolve. However, others consider anachronistic 
sequences as laboratory artifacts [18,19]. Therefore, the existence 
and operation of IAV frozen evolution in nature is unclear. 

The objective of our study was to help to elucidate these aspects 
of IAV evolution by investigating the genetic diversity and 
evolutionary relationships between the viruses detected one year 
apart from two neighboring localities. In 2009, unusually high 
IAV subtype diversity was observed in a sample pool from a 
mallard flock inhabiting a pond in the South Bohemian Region in 
the Czech Republic. A year after, the viruses collected in the 
vicinity of the first locality exhibited exceptionally high sequence 
similarity to that of the previously identified IAV strains. South 
Bohemia is known for its countless ponds and lakes which have 
been established since the 12 th century. The ponds are often 
interconnected with streams into cascades resulting to a dense 
network of pond and lake systems through the landscape the role 
of which in the ecology and perpetuation of the IAV has not been 
fully investigated yet. 

We analyzed the subtype diversity of AIV in each of the study 
localities. Further, the eight genomic segments of the co-circulating 
virus strains were examined to determine the nucleic and amino 
acid sequence diversity, genotype constellations, and patterns of 
genetic reassortment. Finally, we focused on the estimation of the 
evolutionary relationships between the viruses from the two 

The phylogenetic analysis and genotyping revealed remarkable 
genomic variability, with the identification of direct parent- 
progeny relationships between the newly emerging IAV genotypes. 
In addition, mutual comparison of IAV from the two localities 
indicated between-year frozen evolution. 

Materials and Methods 

Virus detection and isolation 

Cloacal and tracheal swabs (147C Virus Transport-Single 
Swab, Copan Innovation, Italy) were collected post mortem from 
hunting harvested mallards (Anas platyrhynchos) during the 
National avian IAV surveillance in the Czech Republic in 2009 
and 2010. As the National surveillance program no specific 
permissions were required to access the sampling localities and for 
sampling activities. The sample collection was coordinated with 
the Regional Veterinary Administration of the Czech Republic 
and the given hunting organizations. Our study did not involve 
endangered or protected species. All the specimens were collected 
post mortem after the hunting harvest. The GPS coordinates of 
the sampling localities were provided in Figure S 1 in File S 1 . The 
birds were not shot for the purpose of our study. No specific 
hunting permission was required to hunt the animals used in this 

The swabs were re-suspended in PBS buffer and the suspensions 
were then divided into aliquots and used either for molecular 
detection or virus isolation. Total nucleic acid was extracted using 
the MagNA Pure Compact and MagNA Pure LC extractors 
(Roche), employing the Total Nucleic Acid Extraction Kit (Roche) 
with an input volume of 200 or 400 ul and an elution volume of 
50 ju.1. The extracts were screened for IAV by the RT-qPCR 
method of Nagy et al. [20]. Subsequendy, the IAV-positive 
specimens were tested for all nine NA subtypes (OneStep RT- 
PCR kit, Qiagen) [21] and for the most common HA subtypes: 
H5, H7, H9 (QuantiTect Probe RT-PCR kit, Qiagen) [22,23], 
and H3, H4, H6, and HI 1 (OneStep RT-PCR kit, Qiagen; the 
primer sets are available on request). In addition, a universal HA 
typing approach was applied [24]. The results of conventional RT- 

PCR reactions were confirmed by sequencing and BLAST 
analysis [25] conducted by the National Center for Biotechnology 
Information (NCBI). 

Virus isolation was performed according to the methodology in 
reference [22]. 

Virus separation 

Allantoic fluid derived from the 2 nd passage of the co-infected 
sample P/18K (hemagglutination test titer of 128) positive for both 
the H6 and H 1 1 in RT-PCR was serially diluted in a range of 
10 1 to 10 4 in distilled water. Then, 100 ul of each dilution series 
were mixed with 100 |J.l of H6 or Hll antibodies and incubated 
for 30 min at room temperature. Subsequendy, 200 ul of allantoic 
fluid-antibody mixture for each dilution was inoculated into the 
allantoic sac of two specific pathogen free (SPF) embryonated 
chicken eggs and incubated at 37°C. After embryo death, the 
allantoic fluid was recovered and tested by the hemagglutination 
test [22]. Finally, the H6, Hll, N2, and N9 subtypes were 
determined via RT-PCR assays according to the above described 

Sequencing analysis 

Partial or whole genome amplification (OneStep RT-PCR kit, 
Qiagen) was performed with various combinations of the 
previously described primers [21,24,26-28] and primers from 
our primer library selected from the conserved and semiconserved 
regions of each genomic segments of the IAV (available on 
request). If needed, the second PCR round was carried out using 
the same primer combination. The amplification products of 
expected size were purified or cut from the agarose gel and 
purified by the High Pure PCR Product Purification Kit (Roche) 
and sequenced using the BigDye Terminator Cycle-Sequencing 
Ready Reaction Kit version 3.1 (Life Technologies). Besides the 
primers used for the amplification, sequencing primer sets 
(available on request) have also been employed to ensure full- 
amplicon or full-coding sequence read and increase the position 
coverage. Sequence analysis was performed on a 3130 genetic 
analyzer (Life Technologies). The particular sequence positions 
were covered 3-times on average. 

Phylogenetic analysis and genotyping 

The sequences were assembled and edited and the sequence 
quality was evaluated by the SeqScape software (Life Technolo- 
gies). BLAST analysis [25] was then performed for all segments of 
each individual isolate across the NCBI database. The sequences 
were aligned with the MAFFT program (Multiple Alignment using 
Fast Fourier Transformation) [29]. Subsequently, alignment 
trimming, and sequence identity matrix and sequence difference 
count matrix calculation at the nucleic and amino acid levels were 
performed using the BioEdit program [30]. Maximum 
likelihood (ML) trees were calculated using the MEGA software 
version 6.0 [31]. For each genomic segment the best nucleotide 
substitution model was inferred on the basis of the lowest Bayesian 
Information Criterion and Akaike information criterion scores. 
According to these selection procedures the following models were 
implemented: Hasegawa-Kishino-Yano + discrete Gamma distri- 
bution with 5 rate categories (HKY+G for PB2, PB1, PA, H6, and 
HI 1 sequences), Kimura 2-parameter +5G (K2+G for NP and MP 
sequences), and Tamura 3-parameter (T92 for N2, N9, and NS 
sequences). The robustness of nodes was evaluated by performing 
1000 bootstrap replicates. Trees were drawn by the TreeExplorer 
tool in the MEGA 6.0 program. For phylogenetic analysis of the 
Hll, N2, and N9 segments, which were represented only by a few 



July 2014 | Volume 9 | Issue 7 | e103053 

Local-Scale Diversity and Frozen Evolution of Influenza A Viruses 

amplicons, the data was supplemented with sequences from 
BLAST hits as well as additional IAV sequences of interest. 

The results of phylogenetic analyses were summarized by 
employing the digital genotyping approach [32]. 

GenBank submission 

The sequences were deposited in GenBank with the accession 
numbers listed in Table 1. 


Background information 

Cloacal and tracheal swabs from mallards (Anas platyrhynchos) 
inhabiting a pond in the South Bohemian Region (further referred 
to as locality P, Figure SI in File SI) and harvested by hunters on 
30 September 2009 were investigated for avian IAV during the 
National avian IAV surveillance program in the Czech Republic. 
Of 25 cloacal (K) and 25 tracheal (T) swabs, 10 (nine cloacal and 
one tracheal) were RT-qPCR positive. Subsequent analysis 
revealed the presence of three different IAV subtypes: H6N2 
(14K, 17K, 23K), H11N9 (25K), and H11N2 (12K). One sample 
(18K) showed H6, HI 1, and N9 positivity. Three swabs (3K, 4K, 
and 18T) were partially subtyped as N9, H6, and H6, respectively, 
and for one specimen (9K) the HA and NA subtypes were not 
determined. No H6N9 subtype was detected, even by means of 
repeated RT-PCR. Virus isolation was successful from two cloacal 
swabs (nos. 17 and 18). From swab no. 17K the H6N2 virus was 
retrieved (P/ 1 7K) whereas the allantoic fluid derived from swab 
no. 18K exhibited successful co-isolation of both H6 and Hll 
subtypes (P/18K). 

One year later, on 1 st November 2010, avian IAV of H6N9 
subtype was detected in mallards residing in a pond (further 
referred to as locality H, (Figure S2 in File SI)) located 
approximately at 25 km from the previous sampling area. Of 20 
cloacal and 20 tracheal swabs, 6 were RT-qPCR positive. Two 
tracheal swabs (nos. IT and 4T) exhibited H6N9 subtype positivity 
and the remaining four were not subtyped. Virus isolation was not 
successful. Preliminary analysis of the H6 and N9 sequences 
revealed unusually high sequence identity with IAV strains from 
the locality P. 

The ponds P and H belong to the artificially established and 
densely distributed pond systems of South Bohemia and are not 
interconnected with streams. These two ponds were managed by 
different hunting associations rearing mallards for hunting 
purposes. In both areas, the mallard population undergoes an 
annual de- and re-population cycle as follows: during spring, 3 
week-old ducklings are bought and raised indoors for two-four 
weeks. At the age of five or six weeks, the birds receive a vaccine 
shot against Clostridium botulinum and are released to the 
respective ponds. They are kept outdoors until the hunting 
harvesting takes place in autumn. Next year the entire re- and de- 
population cycle repeats. The hunting organizations were not in 
mutual contact and the ducklings for the two ponds were bought 
independently from different suppliers. The bank of pond P was 
lined with wooden huts attracting wild mallards to nest (Figure Sib 
in File SI). The mallard population on pond P in 2009 was 500 
birds and that on pond H in 2010 was 400 birds. 

Sequence and phylogenetic analysis 

The presented unique spatial, temporal, and IAV subtype 
setting prompted us to investigate the sequence and genomic 
relationships within and between the locality P and H viruses in 
more detail. To this end, partial or entire coding genome 
sequencing was performed and the phylogenetic relationships 

and nucleotide and amino acid sequence differences between the 
respective genome segments were inferred. Since our IAV pool 
contained two different HA and NA subtypes, phylogenetic trees 
were constructed for all of them (Figures la- j). The sequencing 
status of each specimen included in the analysis is summarized in 
Table 1. 

Overall, the phylogenetic analysis of the locality P viruses 
revealed remarkable sequence diversity. Depending on the 
genome segment, up to five distinct sub-lineages with significant 
bootstrap support were identified. The number of sub-lineages 
decreased from five (PB1 and PA) to four (PB2, NP, and MP) and 
three (NS); (Figures la-c, f, i, j). In addition, the NS segment was 
represented by both of the two alleles. Interestingly, remarkable 
sequence diversity was also observed within the H6, N2, and HI 1 
trees with two recognizable sub-clusters designated as 1 and 2 
(Figures Id, e, and g). This indicated deeper complexity and 
apparent co-circulation of two distinct H6 and H 1 1 genotypes 
within the same sampling locality. The N9 tree did not show 
discrete clustering of the sequences investigated (Figure lh). 

The nucleic acid sequence alignment of the regions used for 
phylogeny estimation showed diversity with peaks in a range from 
5.1 to 7.5% between the segments, namely: PB2 (78/1038; 7.5%), 
PB1 (60/825; 7.3%), PA (33/649; 5.1%), NP (55/797; 6.9%), MP 
(41/804; 5.1%). Regarding the segments encoding for the surface 
antigens, the major differences were identified within the H6 (61/ 
909; 6.8%), N2 (21/322; 6.6%) and HI 1 (12/596; 2%) sequences. 
The N9 amplicons were identical at the nucleotide sequence level 
(Tables S3a-j in File SI). 

The high nucleotide variation contrasted with the high identity 
at the amino acid sequence level. The differences spanned below 6 
residues within the investigated regions regardless of the segment 
considered. The only exception was the NS segment with 18 
amino acid differences in the NS1 and 21 between the NS2 
protein fragments, respectively, which corresponds to the known 
dual allelic structure in avian viruses. 

Contrary to the high genetic diversity within locality P, the 
phylogenetic analysis of the H6N9 strains (locality H) did not 
reveal any discrete sequence clustering and the two representative 
genomes were 100% identical at the nucleotide sequence level 
(Tables S3a-d, f, h-j in File SI). 

Finally, we established the relationships between the IAV strains 
from localities P and H both in terms of phylogenetic analysis and 
sequence identity. The results of phylogenetic analyses revealed 
that the PB2, PB1, and H6 segments of the H/H6N9 strains were 
closely related to those of the P/H6N2 sub-lineage 2 viruses while 
the PA, NP, N9, MP, and NS segments clustered within the P/ 
HI 1N9 sub-lineage 2. This was further supported by almost 100% 
nucleotide sequence identity between the H/H6N9 and corre- 
sponding P/H6N2 and P/Hl 1N9 segments (Tables S3a-d, f, h-j 
in File SI). 

Digital genotyping 

The summarization of the results of phylogenetic analysis within 
the segment identity matrix (SIM) revealed at least five distinct 
IAV genotypes which co-circulated in locality P (Figure 2b): H6N2 
genotype 1 (columns 2, 9, and 1 1), H6N2 genotype 2 (columns 5 
and 6), H11N9 (column 10), H11N2 (column 4), and one with 
unknown subtype (column 3). Except the H11N2 and H6N2 
genotype 2 viruses, which shared identical N2, MP and NS 
sequences (Figure 2c), no additional reassortment was observed. 
On the other hand, the two H6N9 strains IT and 4T, representing 
locality H, had identical genome constellations (Figure 2a, 
columns 12 and 13). 



July 2014 | Volume 9 | Issue 7 | e103053 

Local-Scale Diversity and Frozen Evolution of Influenza A Viruses 


Q\ CJ\ Q\ (J\ (J\ 

in m lo lo 

O* O* O* O* 
tji d d O* 



u u a u 

< ^ 

■— rsi i— rs 


£ E E 

ai ai ai 
ai ai ai 

4_, 4^ Q2 

E £ 




o\ r- 




ro m m fr) 



i- \o 
X X 

yQ vo vo 

c £ O 
O ^ 



CJi (Ji Qi 

cr> o o o o 

5 ^ e £ s> ^ ^ 

^ ^: ^ cn <r r-. co 

m *r o* i— ■— t— i— 

<N <N <N (N (N rN ("N 

o o o o o o o 

CT\ CTi Os 0"i 0"> CT> Os 

to lo lo io m io to 


_ro _ro 
~ro ~ro 

£ £ 

5 * 

£ £ 

5 4 

£ £ E 

5 5 5 

5> i 

Os O* Os O* o* 
u-i lo Lo lo m 

Isl N Nl Nl M 

u u u u u 

E £ £ £ E 

gi -Q 

ro S" 5 -O 

cr ? ro 

OJ r > 

-P u ro 1/1 

£ oju "ro 

i ^ * « 

- 2 o 

E ° 

I s 

£ *-> ° 
J5 aj «— 



July 2014 | Volume 9 | Issue 7 | e103053 

Local-Scale Diversity and Frozen Evolution of Influenza A Viruses 

2 o- 













Figure 1 . Phylogenetic trees of the locality P and H avian influenza viruses. Trees were generated with maximum-likelihood method in the 
MEGA 6.0 software on the basis of nucleotides 1251-2288 (1038) of PB2, 1465-2289 (825) of PB1, 783-1401 (649) of PA, 816-1724 (909) of H6, 679- 
1 274 (596) of H1 1 , 748-1 544 (797) of NP, 568-889 (322) of N2, 1 050-1431 (382) of N9, 203-1 006 (804) of MP, and 540-870 (331 ) of NS. The nucleotide 
substitution models implemented were listed in the Materials and methods section. Bootstrap values (1000 re-samplings) in percentages are 
indicated at each node. The locality P H6N9 strain (P/18KJH6N9) was highlighted in bold and the locality H H6N9 strains (H/1Tand H4/T) are bold and 
underlined. Each particular tree was supplemented with a nucleotide sequence identity matrix table (Tables S3a-j in File S1). The sub-clades of 
interest were highlighted with a segment specific color which is corresponding to Figure 2 and the abbreviations used with Table 1 respectively. 

Finally, the mutual comparison of the SIM columns derived 
from locality P and H IAV genomes clearly showed that the H/ 
H6N9 genome was assembled from the P/H6N2 genotype 2 and 
P/Hl 1N9 viruses at a 3:5 ratio (Figure 2d). Indeed, the PB2, PB1, 
and H6 segments of the H/H6N9 genome were acquired from the 
P/H6N2 genotype 2 viruses while the PA, NP, N9, MP, and NS 
segments from the P/Hl 1N9 strain. 

The results of phylogenetic analysis and digital genotyping 
indicated that the H/H6N9 strains detected in 2010 were 
apparently direct progenies of the previously co-circulating P/ 
H6N2 genotype 2 and P/H11N9 viruses. 

Investigation of co-infection 

The close genetic relationships uncovered between H6N9 
viruses from localities P and H drew our attention back to the 
H6/H11 co-infected cloacal swab specimen P/18K and to the 
IAV strains retrieved by virus isolation on ECE from that 

First of all, we focused on the allantoic fluid and estimated the 
IAV subtype composition by the H6, Hll, N2, and N9 specific 
RT-PCR tests. Surprisingly, the results revealed positivity for all 
segments but N2, which indicated that the allantoic fluid of P/18K 
evidendy represented a mixed population of the H6N9 and 
H11N9 subtypes. Subsequendy, these subtypes were successfully 
separated using H6 and Hll subtype-specific monoclonal 
antibodies and confirmed by RT-PCR tests and sequencing. 

In the next step, we inferred the genomic constellation of the 
antibody separated HI 1N9 and H6N9 IA viruses. Summarization 
of the sequencing results of the entire coding genome clearly 
demonstrated that the entire genome of the P/18K_H1 1N9 virus 
was identical to that of another P/H11N9 strain, P/25K 
(Figure 2e). This H11N9 strain was also identified as a putative 
five-segment donor of the H/H6N9 strains. Finally, genotyping of 
the antibody separated P/18K_H6N9 virus revealed identity 
between the P/18K_H6N9 and H/H6N9 strains, again supported 
by de facto 100% similarity at the nucleotide sequence level 
(Fig 2f). 



July 2014 | Volume 9 | Issue 7 | e103053 

Local-Scale Diversity and Frozen Evolution of Influenza A Viruses 

Figure 2. Segment identity matrix (SIM). The SIM was generated by plotting the influenza A virus (IAV) genomes against each other, with the 
relationships between the segments derived from the phylogenetic trees (Figure 1 a— j) highlighted with colored pixels. The virus nomenclature 
corresponds to that in Table 1 . The deduced genome constellations in the SIM were represented by columns 1 -1 3 and the pixels within the columns 
were aligned according to the conventional listing of the IAV genome segments (from left to right: PB2, PB1, PA, HA, NP, NA, MP, and NS). The color 
scheme for the segments is given at the bottom of the figure and corresponds to the tables S3 in File SI. Empty pixels mean unknown or 
undetermined. Figures: 2a, the entire SIM; 2b overview of the genomic diversity of locality P IAV. For information regarding the Figures 2c-f please 
refer to the text. 


Taking together all these data, since only one parental virus (P/ 
H6N2 genotype 2, Table 1) was successfully isolated, the antibody 
separation approach led to the retrieval of the second parental 
H11N9 strain as well as the progeny H6N9 strains. 

Finally, having the H6N9/H11N9 mixed allantoic fluid 
analyzed, the subtype composition of the original cloacal swab 
specimen was inspected employing the N2 specific RT-PCR test. 
The result was essentially the same like in the allantoic fluid case, 
i.e. the absence of the N2 segment. This indicated that the 
parental subtype H6N2 was apparently not present in this cloacal 
swab either. Thus, the original status of the co-infected cloacal 
specimen P/8K was HI 1N9 and H6N9 positive. 


Direct relationships between IAV from localities P and H were 
first indicated during the analysis of the H6N9/2010 IAV from 
locality H. The sequences obtained were first compared to our 
data containing genome sequences from various avian IAV 
detected in the Czech Republic between 2007 and 2011. This 

preliminary analysis showed absolute sequence identity to the 
viruses detected one year before in nearby locality P. 

Detailed analysis of the specimens collected from locality P 
revealed co-circulation of four subtype combinations, H6N2, 
H6N9, H11N2, and H11N9, and six genomic constellations four 
of which were entirely different. For two subtypes, H6N9 and 
H11N2, the reassortment pattern was indicated. Among the H6, 
H 1 1 , and N2 segments, two sub-clades could have been clearly 
recognized. One specimen showed co-infection with two sub- 
types, H11N9 and H6N9. 

Co-circulation of four entirely different IAV genotypes could 
theoretically led to double, triple, or even quadruple co-infections 
with a potential to generate 2 8 or even as many as 4 8 genomic 
constellations. This suggests, considering the high compatibility of 
the genomic segments [8] and the immunological naivety of the 
mallards, an IAV genotype explosion in locality P. Nevertheless, 
such extreme genomic diversity was not observed. In addition, the 
co-infection prevalence was lower than estimated previously 
[33,34]. From this point of view, it is reasonable to suppose that 
the IAV genomic diversity in locality P might have been greater 



July 2014 | Volume 9 | Issue 7 | e103053 

Local-Scale Diversity and Frozen Evolution of Influenza A Viruses 

than observed and that the disproportions presumably resulted 
from multiple factors like the small number of samples collected 
(representing only 5% of the birds reared in pond P) or the 
sampling bias relative to the culmination of the infection, as well as 
from additional environmental variables. Despite these limitations, 
our data clearly indicated that the IAV genetic diversity at a local- 
scale level can be unexpectedly complex which adds new evidence 
to the recent study of Wille and colleagues [11]. In addition we 
demonstrated that thorough sequence analysis and genotyping 
could reveal the most intimate genetic links and infer the very 
recent reassortment events between the co-circulating IAV strains. 
Therefore, obtaining deeper insight into the diversity and 
dynamics of IAV at the local-scale would require long-term 
monitoring efforts targeted on the same locality with using the 
advantage of sentinel birds [10,12] preferably in combination with 
parallel wild bird sampling [1 1] and followed by a detailed 
genotype analysis of all detected IAV strains. 

The analysis of the co-infected specimen from locality P showed 
the presence of two HA subtypes, H6 and HI 1, and a single NA 
subtype, N9. This constellation was observed both in the screened 
allantoic fluid and the original swab material. We used specific H6 
and Hll antibodies to separate these two subtypes. A similar 
strategy was applied previously to investigate IAV co-infections in 
wild ducks [33] although with a different experimental protocol. 
To this end, the allantoic fluid from the 2 nd passage was used as a 
starting material because the primary culture and the first passage 
were overgrown with bacterial contamination. This approach led 
to successful separation of the H11N9 and H6N9 viruses. 
Subsequent analyses revealed almost absolute sequence identity 
between the respective segments of the antibody separated and, 
let's say, native counterparts (P/Hl 1N9, P/H6N2, and H/H6N9). 
This sufficiently proved that, contrary to the previous observations 
of Lindsay and colleagues [35], in our specific case the two ECE 
passages did not alter the genomic status of the co-infecting viruses 
in terms of in vitro reassortment and ruled out artificial generation 
of the P/H6N9 subtype during the virus isolation efforts. So, the 
co-infected specimen evidently contained both the H6N9 and 
H11N9 viruses. 

Two conclusions can be drawn regarding the emergence of the 
H6N9 virus: i) the H6N9 subtype was evidendy present, if not 
originated, in locality P, ii) the H6N9 virus persisted in the same 
area as was suggested by its re-detection roughly one year apart in 
the nearby locality H. 

Genotyping of the antibody separated P/H6N9 virus showed 
that it was a 3:5 reassortant of the P/H6N2 genotype 2 and P/ 
H11N9 viruses with almost 100% identity of the respective 
genome segments at the nucleotide sequence level. At first sight, it 
indicates that the P/H6N9 virus represents a possible progeny of 
the P/H6N2 genotype 2 and P/H11N9 viruses. However, 
considering the close co-circulation of these viruses, it was not 
possible to determine which one is the parent and which one is the 
progeny. Furthermore, it is not clear whether the P/H6N9 virus 
emerged within the co-infected mallards or originated from 
elsewhere and subsequently co-infected the same bird along with 
the related (parental) P/Hl 1N9 strain. Nevertheless, the genotype 
constellations favor the suggested parent-progeny scenario. 

Despite a roughly one year interval between the P/H6N9 and 
H/H6N9 detection, both of the viruses retained identical subtype 
and genotype constellation. Unexpectedly, the entire genomes 
exhibited almost absolute nucleotide sequence identity. Consider- 
ing the rapid evolutionary dynamics of avian IAV [9], the one- 
year interval between the two H6N9 strains should mean at least 
13 nucleotide differences. Such or higher discrepancies between 
the isolation dates and unexpectedly high genetic conservation 

were previously attributed to laboratory artifacts [18,19]. Vertical 
audit of our entire virus isolation, amplification, sequencing, and 
sequence assembly procedure unequivocally excluded contamina- 
tion or data misinterpretation. In addition, the P and H/H6N9 
viruses were sequenced and analyzed one year apart and, in the 
meantime, various additional and unrelated avian IAV were 
isolated and sequenced by using the same primer sets and 

So, which mechanism would account for the exceptionally high 
sequence conservation of the H6N9 virus? To address this 
question, we performed epizootological investigations which 
included visiting the sample collection sites as well as communi- 
cation with the hunting association representatives and field 
veterinarians who assisted in specimen collection. The investiga- 
tion excluded any mutual contacts, cooperation, trade or 
involvement of other man-associated routes to allow artificial 
transmission, one-year preservation, and re-appearance of the 
H6N9 virus. Therefore, it is reasonable to hypothesize that the 
observed conservation of the H6N9 virus resulted from its 
environmental persistence and frozen evolution. We suppose that 
some of the H6N9 infected birds in locality P might have been 
frightened by hunters and escaped to nearby pond H where they 
disseminated the virus into the environment. Then, the H6N9 
strain persisted in the environment and infected the new and 
immunologically naive mallard flock re-populating pond H next- 

The frozen evolution is long considered as one of the 
mechanisms of influenza virus perpetuation in nature [13]. 
Nevertheless, its significance in the IAV ecology is not fully 
understood. In addition, there is no consistent view on this 
phenomenon in the literature [13—16,18,19]. In a recent study 
Shoham and colleagues [36] have demonstrated the ability of 
productive year-to-year preservation of avian IAV in arctic and 
sub-arctic ice which indicates that the frozen evolution evidendy 
might operate in nature. However, we observed this phenomenon 
in the temperate zone. Similarly, Globig and colleagues [10] 
reported avian H3N2 strains with unusually similar HA and NA 
sequences detected roughly three months apart in sentinel 
mallards kept at a pond in Southern Germany which is in a 
500 km distance from our sampling localities. This finding also 
supports our observations that besides extensive genetic variation, 
the frozen evolution and re-appearance of identical or unusually 
similar IAV strains may apparently act as an additional 
mechanism of virus perpetuation in wild aquatic birds. Again, 
additional and more complex surveillance efforts are required to 
fully elucidate this phenomenon. 

It has been thought that IAV do not prevail in the form of latent 
infection in birds. In addition, pond H underwent annual re and 
de-population cycles. Hence, the genetic conservation of the 
H6N9 virus suggests some kind of environmental persistence. This 
raised another important question: On which environmental 
matrix could the virus persisted? The H6N9 virus had to survive 
through winter, spring, and especially summer, which is relatively 
hot in the temperate zone, to re-appear during the autumn. 
Although recent data suggests that unfavorable environmental 
conditions during summer do not prevent circulation of avian IAV 
in the environment [37] apparendy none of the abiotic reservoirs 
of avian IAV studied so far like feces, intact pond water, or pond 
sediments [38-41] provide protection for a sufficiently long time 
period to ensure year-to-year preservation of the virus in our 
climatic zone. Moreover, the environmental persistence and 
subsequent productive re-infection requires preservation in a 
sufficiently concentrated state to prevent progressive dilution. This 
further argues against the majority of the abiotic components. 



July 2014 | Volume 9 | Issue 7 | e103053 

Local-Scale Diversity and Frozen Evolution of Influenza A Viruses 

Similarly, the significance of the biotic environmental components 
as long term reservoirs is also negligible [42-47] . 

Nevertheless, a mechanism has been characterized to date 
which can fit our assumptions. It has been demonstrated that 
feathers covered with preen oil could efficiently capture and 
concentrate the avian IAV from water [48]. Subsequendy, the 
virus particles adsorbed on bird's bodies may mediate infection 
through self-preening or allo-preening activities. Although feather 
swabs collected from experimentally preened birds were positive 
by virus isolation roughly for one month [49] it is not known how 
long the virus can survive in preened feathers. Can the 
hydrophobic preen oil on detached feathers provide a sufficiendy 
protecting environment for between-year persistence? 

Although our conclusions probably raise more questions than 
they answer the results of the presented study suggest that the IAV 
subtype and genotype diversity between the IAV at the local-scale 
level can be admirably complex. Evaluation of the most intimate 
genetic links between such viruses can reveal rarely observed 
phenomena like direct parent-progeny relationships between the 
co-circulating strains or frozen evolution. Further and more 
detailed studies are required to fully elucidate whether between 
year persistence and frozen evolution observed here is an isolated 
unique event or represents a more regular but yet unrecognized 
phenomenon in the evolution of the influenza virus in aquatic 

Supporting Information 

File SI Figure SI. Locality P. The samples in 2009 were 
collected from mallards inhabiting the pond called Putim (GPS 
coordinates 49°16'42.629"N, 14°8'9.881"E). The green flag on the 
map S 1 c represents the position of the S 1 a view. Figure S 1 b shows 
wooden huts distributed along the shore. The map was generated 
by the Figure S2: Locality H. The samples in 


1. Webster RG, Bean WJ, Gorman OT, Chambers TM, Kawaoka Y (1992) 
Evolution and eeology of inlluenza A viruses. Mierobiol Rev 56: 152—179. 

2. Olscn B, Munstcr VJ, Wallensten A, Waldenstrom J, Osterhaus AD, et al. (2006) 
Global patterns of inlluenza a virus in wild birds. Seienee 312: 384-388. 

3. Munster VJ, Baas C, Lexmond P, Waldenstrom J, Wallensten A, et al. (2007) 
Spatial, temporal, and speeies variation in prevalenee of inlluenza A viruses in 
wild migratory birds. PLoS Pathog 3: e6 1 . 

4. Wallensten A, Munster VJ, Latorre-Margalef N, Brytting M, Elmberg J, et al. 
(2007) Surveillanee of inlluenza A virus in migratory waterfowl in northern 
Europe. Emerg Infect Dis 13: 404-411. 

5. Krauss S, Obcrt CA, Franks J, Walker D, Jones K, et al. (2007) Influenza in 
migratory birds and evidence of limited intercontinental virus exchange. PLoS 
Pathog 3: el 67. 

6. Wilcox BR, Knutsen GA, Berdeen J, Goekjian V, Poulson R, et al. (2011) 
Influenza-A viruses in ducks in northwestern Minnesota: line scale spatial and 
temporal variation in prevalence and subtype diversity. PLoS One 6: c24010. 

7. Hatchcttc TF, Walker D, Johnson C, Baker A, Pryor SP, ct al. (2004) Influenza 
A viruses in feral Canadian ducks: extensive rcassortment in nature. J Gen Virol 
85: 2327-2337. 

8. Dugan VG, Chen R, Spiro DJ, Sengamalay N, Zaborsky J, et al. (2008) The 
evolutionary genetics and emergence of avian influenza viruses in wild birds. 
PLoS Pathog 4: el 000076. 

9. Chen R, Holmes EC (2006) Avian inlluenza virus exhibits rapid evolutionary 
dynamics. Mol Biol Evol 23: 2336-2341. 

10. Globig A, Fcreidouni SR, Harder TC, Grand C, Beer M, ct al. (2013) 
Consecutive natural inlluenza a virus infections in sentinel mallards in the 
evident absence of subtype-specific hemagglutination inhibiting antibodies. 
Transbound Emerg Dis 60: 395-402. 

11. Willc M, Tolf C, Avril A, Latorre-Margalef N, Wallcrstrom S, ct al. (2013) 
Frequency and patterns of rcassortment in natural influenza A virus infection in 
a reservoir host. Virology 443: 150-160. 

12. Tolf C, Latorre-Margalef N, Wille M, Bengtsson D, Gunnarsson G, ct al. (2013) 
Individual variation in inlluenza A virus infection histories and long-term 
immune responses in Mallards. PLoS One 8: e61201. 

13. Hayashida H, Toh H, Kikuno R, Miyata T (1985) Evolution of influenza virus 
genes. Mol Biol Evol 2: 289-303. 

2010 were collected from mallards resided the pond called 
Kahoun which is situated near the Hajany village (GPS 
coordinates 49°26'59.547"N, 13°49'52.568"E). The green flag 
on the map S2b represents the position of the panoramic view S2a. 
The map was generated by the Table S3: 
Nucleic acid sequence identity matrices of the locality 
P and H avian influenza virus segments. The matrices were 
constructed using the BioEdit program on the basis of nucleotides 
1251-2288 (1038) of PB2, 1465-2289 (825) of PB1, 783-1401 
(649) of PA, 816-1724 (909) of H6, 679-1274 (596) of Hll, 748- 
1544 (797) of NP, 568-889 (322) of N2, 1050-1431 (382) of N9, 
203-1006 (804) of MP, and 540-870 (331) of NS. The tables were 
highlighted with a segment specific color which is corresponding to 
Figures 1 and 2 and the abbreviations with Table 1 respectively. 


We gratefully acknowledge the excellent technical assistance of Mrs Eliska 
Vrzakova and Mrs Valeria Cermakova and the editorial assistance of Dr 
Eva Kodytkova. Special thanks go to Dr. Olga Janouchova and Dr Karel 
Krametbauer from the Regional Veterinary Administration of the Czech 
Republic, Ceske Budejovice and to the representatives of the hunting 
organization Haj and Mefiny Kocelovice for providing valuable field 
information. Last but not least we thank the State Veterinary Adminis- 
tration of the Czech Republic and all contributors of the Influenza Virus 
Resource database. 

Author Contributions 

Conceived and designed the experiments: AN LC HJ MH JH. Performed 
the experiments: AN LC. Analyzed the data: AN LC HJ MH JH. 
Contributed reagents/materials/analysis tools: AN LC HJ MHJH. Wrote 
the paper: AN. 

14. Endo A, Pecoraro R, Sugita S, Nerome K (1992) Evolutionary pattern of the H 
3 haemagglutinin of equine inlluenza viruses: multiple evolutionary lineages and 
frozen replication. Arch Virol 123: 73-87. 

15. Bountouri M, Fragkiadaki E, Ntafis V, Kancllos T, Xylouri E (2011) 
Phylogenctie and molecular characterization of equine H3N8 inlluenza viruses 
from Greece (2003 and 2007): evidence for rcassortment between evolutionary 
lineages. Virol J 8: 350. 

16. Chambers TM (2013) Equine/Canine/Feline/Seal influenza. 207. In: Webster 
RG, Monto AS, Bracialc TJ, Lamb RA. Textbook of Inlluenza, 2 nrl Edition, p. 

17. Webster RG (1998) Influenza: an emerging disease. Emerg Infect Dis 4: 436— 

18. Krasnitz M, Lcvine AJ, Rabadan R (2008) Anomalies in the influenza virus 
genome database: new biology or laboratory errors? J Virol 82: 8947-8950. 

19. Worobey M (2008) Phylogenctie evidence against evolutionary stasis and natural 
abiotic reservoirs of influenza A virus. J Virol 82: 3769-3774. 

20. Nagy A, Vostinakova V, Pirehanova Z, Ccrnikova L, Dirbakova Z, et al. (2010) 
Development and evaluation of a one-step real-time RT-PCR assay for universal 
detection of influenza A viruses from avian and mammal species. Arch Virol 
155: 665-673. 

21. Fcreidouni SR, Starick E, Grund C, Globig A, Mcttcnlcitcr TC, ct al. (2009) 
Rapid molecular subtyping by reverse transcription polymerase chain reaction of 
the neuraminidase gene of avian influenza A viruses. Vet Microbiol 135: 253- 

22. Anonymous (2006) Comission Decision 2006/437/EC approving a diagnostic 
manual for avian influenza as provided in Council Directive 2005/94/EC. 
OfTJ Eur Commun L237: 1-27. 

23. Monnc I, Ormelli S, Salviato A, De Battisti C, Bcttini F, ct al. (2008) 
Development and validation of a one-step real-time PGR assay for simultaneous 
detection of subtype H5, H7, and H9 avian influenza viruses. J Clin Microbiol 
46: 1769-1773. ' 

24. Phipps LP, Essen SC, Brown IH (2004) Genetic subtyping of inlluenza A viruses 
using RT-PCR with a single set of primers based on conserved sequences within 
the HA2 coding region. J Virol Methods 122: 1 19-122. 

25. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ (1990) Basic local 
alignment search tool. J Mol Biol 215: 403-410. 



July 2014 | Volume 9 | Issue 7 | e103053 

Local-Scale Diversity and Frozen Evolution of Influenza A Viruses 

26. Hoffmann E, StcchJ, Guan Y, Webster RG, Perez DR (2001) Universal primer 
set for the full-length amplifieation of all influenza A viruses. Arch Virol 146: 

27. Qiu BF, Liu WJ, Peng DX, Hu SL, Tang YH, et al. (2009) A reverse 
transcription-PCR for subtyping of the neuraminidase of avian influenza viruses. 
J Virol Methods 155: 193-198. 

28. Li OT, Barr I, Leung CY, Chen H, Guan Y, et al. (2007) Reliable universal RT- 
PCR assays for studying influenza polymerase subunit gene sequences from all 
16 haemagglutinin subtypes. J Virol Methods 142: 218-222. 

29. Katoh K, Misawa K, Kuma K, Miyata T (2002) MAFFT: a novel method for 
rapid multiple sequence alignment based on fast Fourier transform. Nucleic 
Acids Res 30: 3059-3066. 

30. Hall TA (1999) BioEdit: a user friendly biological sequence alignment editor and 
analysis program for Windows 95/98/NT. Nucl Acid Symp Scr 41: 95-98. 

31. Tamura K, Steelier G, Peterson D, Filipski A, Kumar S (2013) MEGA6: 
Molecular Evolutionary Genetics Analysis version 6.0. Mol Biol Evol 30: 2725— 

32. Nagy A, Gernikova L, Krivda V, Hornickova J (2012) Digital genotyping of 
avian influenza viruses of H7 subtype detected in central Europe in 2007—201 1. 
Virus Res 165: 126-133. 

33. Sharp GB, Kawaoka Y, Jones DJ, Bean WJ, Pryor SP, ct al. (1997) Coinfection 
of wild ducks by influenza A viruses: distribution patterns and biological 
significance. J Virol 71: 6128-6135. 

34. Wang R, Soli L, Dugan V, Runstadler J, Happ G, et al. (2008) Examining the 
hemagglutinin subtype diversity among wild duck-origin influenza A viruses 
using ethanol-fixed cloacal swabs and a novel RT-PGR method. Virology 375: 

35. Lindsay LL, Kelly TR, Plancarte M, Schobel S, Lin X, et al. (2013) Avian 
influenza: mixed infections and missing viruses. Viruses 5: 1964-1977. 

36. Shoham D, Jahangir A, Ruenphet S, Takehara K (2012) Persistence of avian 
influenza viruses in various artificially frozen environmental water types. 
Influenza Res Treat 2012: 912326. 

37. Henaux V, Samuel MD, Dusek RJ, Fleskes JP, Ip HS (2012) Presence of avian 
influenza viruses in waterfowl and wetlands during summer 2010 in California: 
arc resident birds a potential reservoir? PLoS One 7: e31471. 

38. Keelcr SP, Lcbarbcnehon C, Stallknecht DE (2013) Strain-related variation in 
the persistence of influenza A virus in three types of water: distilled water, filtered 
surface water, and intact surface water. Virol J 10: 13. 

39. Nazir J, Haumachcr R, Ike AC, Marschang RE (2011) Persistence of avian 
influenza viruses in lake sediment, duck feces, and duck meat. Appl Environ 
Microbiol 77: 4981-4985. 

40. Lang AS, Kelly A, Runstadler JA (2008) Prevalence and diversity of avian 
influenza viruses in environmental reservoirs. J Gen Virol 89: 509—519. 

41. Lebarbenchon C, Srcevatsan S, Lefevre T, Yang M, Ramakrishnan MA, et al. 
(2012) Reassortant influenza A viruses in wild duck populations: effects on viral 
shedding and persistence in water. Proc Biol Sci 279: 3967-3975. 

42. Stumpf P, Failing K, Papp T, Nazir J, Bohm R, et al. (2010) Accumulation of a 
low pathogenic avian influenza virus in zebra mussels (Dreisscna polymorpha). 
Avian Dis 54: 1183-1190. 

43. Faust C, Stallknecht D, Swayne D, Brown J (2009) Filter-feeding bivalves can 
remove avian influenza viruses from water and reduce infectivity. Proc Biol Sci 
276: 3727-3735. 

44. Huyvacrt KP, Carlson JS, Bentler KT, Cobble KR, Noltc DL, et al. (2012) 
Freshwater clams as bioconcentrators of avian influenza virus in water. Vector 
Borne Zoonotic Dis 12: 904-906. 

45. Oesterle PT (201 1) The role of freshwater snails in the transmission of influenza 
A viruses. Dissertation, Colorado State University. 

46. Horm VS, Gutierrez RA, Nicholls JM, Buchy P (2012) Highly pathogenic 
influenza A(H5N1) virus survival in complex artificial aquatic biotopes. PLoS 
One 7: c34160. 

47. Meixell BW, Borchardt MA, Spencer SK (2013) Accumulation and inactivation 
of avian influenza virus by the filter- fee ding invertebrate Daphnia magna. Appl 
Environ Microbiol 79: 7249-7255. 

48. Dclogu M, De Marco MA, Di Trani L, Raffini E, Cord C, et al. (2010) Can 
preening contribute to influenza A virus infection in wild waterbirds? PLoS One 
5: el 1315. 

49. Dclogu M, De Marco MA, Cotti C, Di Trani L, Raflini E, et al. (2012) Human 
and animal integrated influenza surveillance: a novel sampling approach for an 
additional transmission way in the aquatic bird reservoir. Italian Journal of 
Public Health 9: 29-36. 



July 2014 | Volume 9 | Issue 7 | e103053