Pre-existing immunity against swine-origin H1N1
influenza viruses in the general human population
Jason A. Greenbauma,1, Maya F. Kotturia,1, Yohan Kima, Carla Oseroffa, Kerrie Vaughana, Nima Salimia, Randi Vitaa,
Julia Ponomarenkob, Richard H. Scheuermannc, Alessandro Settea, and Bjoern Petersa,2
aDivision of Vaccine Discovery, La Jolla Institute for Allergy and Immunology, La Jolla, CA 92037;bSan Diego Supercomputer Center, University of California
at San Diego, La Jolla, CA 92093; andcDepartment of Pathology and Division of Biomedical Informatics, University of Texas Southwestern Medical Center,
Dallas, TX 75390
Communicated by Howard M. Grey, La Jolla Institute for Allergy and Immunology, La Jolla, CA, October 8, 2009 (received for review August 31, 2009)
A major concern about the ongoing swine-origin H1N1 influenza
virus (S-OIV) outbreak is that the virus may be so different from
seasonal H1N1 that little immune protection exists in the human
population. In this study, we examined the molecular basis for
pre-existing immunity against S-OIV, namely the recognition of
viral immune epitopes by T cells or B cells/antibodies that have
been previously primed by circulating influenza strains. Using data
of B-cell epitopes present in recently circulating H1N1 strains are
conserved in the S-OIV, with only 17% (1/6) conserved in the
hemagglutinin (HA) and neuraminidase (NA) surface proteins. In
contrast, 69% (54/78) of the epitopes recognized by CD8?T cells
are completely invariant. We further demonstrate experimentally
that some memory T-cell immunity against S-OIV is present in the
adult population and that such memory is of similar magnitude as
protection from infection is antibody mediated, a new vaccine
based on the specific S-OIV HA and NA proteins is likely to be
required to prevent infection. However, T cells are known to blunt
disease severity. Therefore, the conservation of a large fraction of
T-cell epitopes suggests that the severity of an S-OIV infection, as
far as it is determined by susceptibility of the virus to immune
attack, would not differ much from that of seasonal flu. These
results are consistent with reports about disease incidence, sever-
ity, and mortality rates associated with human S-OIV.
databases ? epitopes ? meta-analysis ? pandemic
ganization (WHO) criteria for a pandemic (1). As this virus
contains a unique combination of gene segments from both
North American and Eurasian swine lineages, and is antigeni-
cally distinct from seasonal human influenza A (2), there is
concern that little protective immune memory exists in the
general human population. This concern was confirmed by
reports that neutralizing antibodies against S-OIV are found
nearly exclusively in persons born before 1957, presumably
because of their exposure to H1N1 influenza strains that did not
circulate after that time (3, 4). Together with reports of differ-
ences in pathogenicity of the virus in animal models compared
of a pandemic with major public health consequences. At the
same time, the incidence of clinically severe cases so far appears
to be similar to that experienced for seasonal flu: According to
CDC estimates, ?1 million people were infected with S-OIV
between April 15 and July 24, 2009, leading to 5,011 hospital-
izations and 302 deaths (http://www.cdc.gov/h1n1flu/surveil-
lanceqa.htm). This seeming contradiction highlights the need to
better understand the interaction of this pathogen with the
The focus of the present study was to examine the presence of
immune memory against S-OIV in the human population.
Adaptive immune responses against influenza (and other patho-
he spread of the ongoing human swine-origin H1N1 influ-
enza virus (S-OIV) outbreak meets the World Health Or-
gens) are triggered upon T-cell or B-cell receptors recognizing
viral immune epitopes. Recognition of epitopes by antibodies
and T cells that were induced after past influenza infections or
vaccinations are key components of immune memory and pro-
tection from infection. Importantly, a virus can carry substantial
sequence differences in some regions but still be recognized by
the immune system if the virus retains sequence identity in
regions including the immune epitopes. Therefore, we specifi-
cally examined whether there are immune epitopes in S-OIV
that are likely targets of pre-existing immunity (i.e., if epitopes
that were present in the H1N1 seasonal flu strains from between
1988 and 2008 are also present in the S-OIV strains).
We obtained sequence and source information on previously
defined epitopes from the Immune Epitope Database (IEDB)
(7, 8), which was developed for the purpose of cataloging and
making available epitope information in a single repository to
the scientific community. Curated epitope information from all
manuscripts published to date that characterize epitopes in
influenza is contained in the IEDB. Swine influenza sequence
information was obtained from the National Center for Bio-
technology Information (NCBI) (9) and the Global Initiative on
Sharing Avian Influenza Data (GISAID) (10) influenza se-
quence databases. The present report analyzes how well these
known influenza epitopes, and therefore the targets of immu-
nological memory, are conserved in S-OIV isolates and exper-
imentally validates the results.
A Significant Fraction of Epitope Sequences Are Conserved in S-OIV.
The overall goal of the analysis presented in this section was to
establish whether the epitopes described in the literature and
presumably associated with pre-existing immunity in the general
population, would be conserved in S-OIV sequences.
At the time of this writing, the epitope information available
in the IEDB related to influenza A encompassed information
from 594 references (journal articles and direct submissions),
describing 3,724 distinct molecular structures (linear and dis-
continuous peptides) derived from influenza A that were tested
experimentally for interaction with immune receptors. This
roughly doubled the amount of information on influenza
epitopes that was available in 2006 (7), emphasizing the impor-
tant contribution and greatly enhanced throughput of recent
influenza epitope mapping efforts, which were stepped up since
the emergence of H5N1 avian flu.
Author contributions: A.S. and B.P. designed research; J.A.G., M.F.K., and C.O. performed
research; J.A.G., M.F.K., Y.K., K.V., N.S., R.V., J.P., R.H.S., A.S., and B.P. analyzed data; and
J.A.G., M.F.K., J.P., R.H.S., A.S., and B.P. wrote the paper.
The authors declare no conflict of interest.
Freely available online through the PNAS open access option.
1J.A.G. and M.F.K. contributed equally to this work.
2To whom correspondence should be addressed. E-mail: email@example.com.
This article contains supporting information online at www.pnas.org/cgi/content/full/
December 1, 2009 ?
vol. 106 ?
no. 48 ?
We focused our analysis on influenza A epitope mapping
experiments that have the most relevance for human immunity
and considered only those epitopes that are present in the
sequences of recently circulating strains of H1N1 (i.e., isolates
from 1988 to 2008). We considered an epitope relevant if it
showed recognition by antibodies or by T cells in the context of
human MHC and if the epitopes were mapped in the context of
whole influenza organisms or proteins. As B-cell epitopes are
infrequently identified in humans (7), B-cell epitopes mapped in
the context of any host organism were included. We think that
this is justified, as studies have shown substantial overlap for
B-cell epitopes defined in different species, presumably because
the structural constraints associated with their recognition is
similar across species (11).
We distinguished between two primary categories of epitopes:
(i) those recognized by B cells/antibodies, and (ii) those recog-
nized by T cells. The latter were further categorized into those
recognized by CD8?T cells in the context of major histocom-
patibility complex (MHC) class I molecules and those recog-
nized by CD4?T cells in the context of MHC class II. If multiple
epitopes in the same category had sequences that were nested
within each other, a representative epitope was chosen that was
associated with the highest specific immune response. Full
details of the specific IEDB queries are given in Materials and
Methods. As shown in Table 1, a total of 26 B-cell epitopes were
found in recently circulating H1N1 strains, whereas 139 CD4?
and 78 CD8?T-cell epitopes were found.
Next, we asked how many of these experimentally defined
are totally conserved in the emerging S-OIV strains. Human
S-OIV sequences were retrieved from isolates originating in
Mexico, the United States, and elsewhere, as described in
Materials and Methods. As shown in Table 1, a substantial
number of epitopes present in recent seasonal H1N1 strains are
also found to be 100% conserved in S-OIV sequences. Notably,
whereas only ?31% of the B-cell epitopes are conserved in
S-OIV strains, up to 41% and 69% of the CD4?and CD8?T-cell
epitopes, respectively, are conserved. The complete list of
individual epitopes is provided in Table S1.
We also assessed whether seasonal influenza viruses of the
H3N2 subtype contained additional epitopes conserved in S-
OIV strains. A repeat of the analysis above showed that this was
not the case, as all epitopes that are conserved between the
H3N2 subtype and as S-OIV are also shared with seasonal H1N1
isolates. Overall, these data demonstrated that a sizeable frac-
tion of epitopes derived from H1N1 seasonal flu strains and
described in the literature is conserved in S-OIV sequences. This
raised the possibility that some level of immunity against S-OIV
sequences might exist in the general population.
Distribution of Conserved Epitopes in Different Influenza Proteins.
The proteins hemagglutinin (HA) and neuraminidase (NA),
which make up the majority of the viral surface, are more
variable than other influenza proteins. Therefore, it can be
expected that epitopes from those proteins are less likely to be
conserved across strains than others. As shown in Table 2, this
is indeed the case. For T-cell epitopes, only five of 43 (12%) HA
and NA epitopes are conserved, as opposed to 106 of 174 (61%)
epitopes derived from other proteins. Similarly, at the level of
B-cell epitopes, only a single HA/NA epitope among six is
conserved (17%), whereas seven of 20 of the remaining epitopes
in other proteins are conserved (35%). This shows that for both
B-cell and T-cell epitopes, fewer epitopes are conserved in the
virion surface antigens.
For B-cell responses, neutralizing epitopes are primarily lo-
cated on the virion surface proteins HA and NA. That means
that the majority of B-cell epitopes conserved in S-OIV are
unlikely to be neutralizing. For example, the NP protein contains
9 H1N1 B-cell epitopes, four of which are conserved in S-OIV
sequences. However, to be targets of directly neutralizing anti-
bodies, epitopes need to be exposed on the virion surface.
Therefore, although epitopes in the NP protein are often
targeted by serological immune responses, they are not protec-
tive as they are thought to be inaccessible from the surface (12).
A single HA B-cell epitope is conserved in S-OIV. To
determine its location in the protein, we mapped the epitope
onto the three-dimensional structure of a homologous protein of
the S-OIV HA trimer complex for the precursor HA (Fig. 1A
and B) and onto a homology model of the cleaved HA (Fig. 1C
and D), as the epitope is located at the N-terminal of the fusion
peptide. In the precursor, the epitope is solvent accessible, with
an accessible surface area (ASA) of 632 Å2, which is similar to
the surface area of interfaces in known structures of antibody–
peptide complexes. In the cleaved HA, the epitope is partially
buried in the cavity of HA, into which the fusion peptide inserts
Table 1. Number of influenza A H1N1 epitopes in the Immune
S-OIV, swine-origin H1N1 influenza virus.
Table 2. Distribution of epitopes among the influenza proteins
The total number of epitopes in the H1N1 seasonal flu strains from 1988–2008 (Total) as well as the number of epitopes conserved
in swine-origin H1N1 influenza virus (S-OIV) (Cons.) are listed.
www.pnas.org?cgi?doi?10.1073?pnas.0911580106 Greenbaum et al.
after precursor cleavage, with an ASA of 294 Å2. The location
of this epitope in the conserved HA2 glycopeptide is atypical, as
most neutralizing antibodies target HA1, and it is unclear how
frequent pre-existing immune responses are against this peptide.
At the same time, this epitope was mapped to a monoclonal
antibody that inhibited cell fusion in vitro and provided protec-
tion from challenge in vivo (13). Taken together, these data
suggest that although the one B-cell epitope that is conserved
might be partially exposed in some HA isoforms, the vast
majority of known B-cell epitopes are not conserved in S-OIV
and are unlikely to be relevant for protection.
Experimental Demonstration of Memory T-Cell Response Recognizing
H1N1 Influenza Sequences. The analysis presented above suggests
that a significant level of T-cell immunity might pre-exist in the
general population against epitope sequences that are found
totally conserved in S-OIV. To experimentally address this
corresponded to the CD4?T-cell epitopes totally conserved in
S-OIV sequences; the second group consisted of CD8?T-cell
epitopes conserved in S-OIV; and two additional groups of
peptides consisted of CD4?and CD8?T-cell epitopes not
conserved in S-OIV sequences. These four separate peptide
pools were tested with peripheral blood mononuclear cells
(PBMCs) from normal blood donors for induction of interferon
(IFN)–? secretion in direct ex vivo ELISPOT assays. PBMC
from anonymous blood donors from the San Diego region were
banked in the context of an independent study of influenza
responses, and collected at least 1 year before the onset of the
current S-OIV pandemic.
We found that, on average, responses significantly greater
than zero were detected against the pool of epitopes conserved
in S-OIV sequences, both in terms of CD4?and CD8?T-cell
responses (P ? 0.01 for all epitopes sets, one-tailed single sample
t test) (Fig. 2). The responses to the conserved S-OIV epitope
pools were similar in magnitude to the responses observed with
their nonconserved counterparts, and the observed difference
was not statistically significant (average of 40 spot-forming cells
[SFC]/106cells for conserved and 41 SFC for nonconserved
CD8? T-cell epitopes (P ? 0.94, two-tailed paired t test), and 51
respectively (P ? 0.14, two-tailed paired t test). Information on
age and sex was available only for half of the 20 donors (Table
S2); thus investigating correlations of responses with these
parameters was not possible.
For selected donors, we sought to further document that the
responding cells were of the memory phenotype. Accordingly,
intracellular cytokine and cell-surface staining assays were used to
characterize the IFN-? responses against CD4?and CD8?T-cell
epitopes conserved in S-OIV. Specifically, CD4?or CD8?T cells
producing IFN-? were examined for the expression of various
memory markers. We found that the responding CD4?IFN-??T
cells, stimulated with the epitope pool conserved in S-OIV, were
CD45RAmedCD62LloCCR7?(Fig. 3). Likewise, the CD8?IFN-??
T cells stimulated with the epitope pool conserved in S-OIV were
also CD45RAmedCD62LloCCR7?. Overall, these data show that
both CD4?and CD8?T-cell memory responses against S-OIV
responses against epitopes conserved in S-OIV is at least compa-
frontal (A) and orthogonal (B) views. Monomer chains are in white, magenta, and green. Epitopes in each monomer are identical and shown in blue. (C and D)
Quaternary structure of the cleaved HA modeled for the representative swine influenza HA sequence (ACP41934.1): frontal (C) and orthogonal (D) views. HA2
in each chain.
Location of the conserved epitope in the S-OIV HA protein structure. (A and B) Quaternary structure of the HA precursor (Protein Data Bank ID: 1HA0):
S-OIV. PBMC from normal individuals (n ? 20) were stimulated with pools of
either CD4?or CD8?T-cell epitopes from recent seasonal H1N1 influenza
strains that were either absolutely conserved or not conserved in S-OIV
sequences. Responses were measured through ex vivo IFN-? ELISPOT assays.
Error bars represent SEM.
Detection of pre-existing CD4?and CD8?T-cell immune responses to
Greenbaum et al.PNAS ?
December 1, 2009 ?
vol. 106 ?
no. 48 ?
rable to that of other epitopes in recently circulating influenza
Comparison of Literature-Defined Epitope Content for Seasonal In-
fluenza and S-OIV at the Individual Isolate Level.Next,weaskedhow
the level of pre-existing immunity for an average S-OIV isolate
compares to that of an average seasonal H1N1 influenza isolate.
For 559 seasonal H1N1 flu isolates from 2000 to 2008, we
determined the number of epitopes present in each isolate that
were also present in H1N1 seasonal flu isolates of the preceding
20 years. As before, such epitopes are considered likely targets
of pre-existing immune responses. Table 3 lists the median
number of epitopes on a per-virus basis as 16 B cell, 93 CD4?,
and 66 CD8?T-cell epitopes. These numbers were used as a
baseline against which we compared the number of epitopes
conserved in S-OIV isolates, again on a per-virus basis. Inter-
estingly, no variability in the epitope sequences within different
S-OIV isolates was detected. Thus, the number of conserved
epitopes in any S-OIV isolate is 8, 57, and 54 for B-cell, CD4?,
and CD8?T-cell epitopes, respectively. Thus, it would be
predicted that the levels of pre-existing immunity against S-OIV,
albeit reduced in comparison to the levels directed against other
H1N1 strains, would still be significant, particularly in the case
of T-cell epitopes, in general, and CD8?T-cell epitopes, in
To test this hypothesis, we compared the magnitude of re-
the peptide pools corresponding to CD4?and CD8?T-cell
epitopes conserved in S-OIV with peptide pools corresponding
to CD4?and CD8?T-cell epitopes found in 2008 seasonal
influenza. For this comparison, we used PBMC from blood
donations obtained in 2007. The results obtained are shown in
We found that significant CD4?and CD8?T-cell responses
were detected to epitopes conserved in both S-OIV and 2008
seasonal influenza (P ? 0.05 for all epitope sets, one-tailed
single-sample t test). As predicted, the response to the epitopes
in 2008 seasonal influenza was higher than the response to the
S-OIV epitopes, in terms of both CD4?and CD8?T cells (17
versus 28 SFC/106for CD4?T cells, and 44 versus 51 SFC/106for
CD8?T cells in response S-OIV and seasonal influenza
epitopes, respectively); however, these differences did not reach
statistical significance (P ? 0.25 for CD4?responses and P ?
Overall, the level of pre-existing immunity was higher for CD8?
T cells compared with CD4?T cells, as this likely reflects a
higher level of conservation of the S-OIV CD8?T-cell epitopes
with recently circulating 2008 seasonal influenza.
memory phenotype. PBMC from two representative normal donors were
the surface markers CD4 and CD8, respectively. Signal intensities of the gated
cells for memory phenotypic markers CD45RA, CD62L, and CCR7 are shown.
S-OIV–specific CD4?and CD8?T cells appear to have an effector
Table 3. Median number of epitopes per virus isolate found
conserved in strains from preceding 20 years
YearB-cell T-cell, CD8?
Abbreviation: S-OIV, swine-origin H1N1 influenza virus.
2008 seasonal influenza. PBMC from normal individuals (n ? 20) were stim-
ulated with pools of either CD4?or CD8?T-cell epitopes that were conserved
in S-OIV or seasonal influenza 2008 sequences. Responses were measured
through ex vivo IFN-? ELISPOT assays. Error bars represent SEM.
www.pnas.org?cgi?doi?10.1073?pnas.0911580106Greenbaum et al.
We provide here an analysis of immune epitopes found in the
emergent S-OIV strains. We found that, overall, 49% of the
epitopes reported in the literature and present in recently
circulating seasonal H1N1 are also found totally conserved in
S-OIV. Interestingly, the number of conserved epitopes varied
greatly as a function of the class of epitopes considered. Al-
though only 31% of the B-cell epitopes were conserved, 41% of
the CD4?and 69% of the CD8?T-cell epitopes were conserved.
It is known that crossreactive T-cell immune responses can exist
even between serologically distinct influenza A strains (14, 15).
Based on this observation and the data presented above, we
hypothesized that it is possible that immune memory responses
against S-OIV exist in the adult population, at the level of both
B and T cells. However, our analysis also suggested that T cells
would mediate such responses predominantly, and in particular
In terms of B-cell responses, there is an average of only 16
known B-cell epitopes with the potential to elicit a memory
immune response in a seasonal H1N1 influenza strain. As it is
known that pre-existing antibodies against previous seasonal
influenza strains do not provide broad protection from subse-
quent infections, a further drop from 16 to 8 B-cell epitopes in
S-OIV is likely to mean that little memory B-cell responses
against S-OIV exist in the human population. Moreover, only a
single conserved epitope was found between the HA and NA
surface proteins, which are the primary targets of neutralizing
antibodies. This suggests that few pre-existing neutralizing an-
tibodies will be present in the general human population, which
is in agreement with the results of experimental studies that
essentially found no neutralizing antibodies against S-OIV in the
general human population under the age of 60 years (3, 4).
The same experimental studies indicated that an increased
number of persons who contracted influenza before the 1957 flu
pandemic apparently have retained antibody responses that are
neutralizing against S-OIV (3, 4). The overlap between the
number of epitopes found in influenza sequences in 1957 and
previous years, versus S-OIV, was not significantly different
from the more recently circulating strains (10 B-cell, 52 CD4?,
59 CD8?). However, this is likely due to few influenza strains
from those years having been used to identify epitopes. To better
address this issue, we mapped S-OIV regions homologous to
identified epitopes and compared their sequence similarity in
influenza strains preceding 1957 or from the past 20 years.
Taking the average sequence identity, we found a significantly
higher similarity between the strains preceding 1957, with 126 of
all 200 epitopes and 58 of 78 epitopes from the HA and NA
protein having a higher similarity hit in strains preceding 1957
(P ? 3.78 E-6 according to a paired, one-tailed t test). This could
explain why the presence of neutralizing antibodies in the human
population depends on exposure to influenza strains preceding
In terms of T-cell responses, epitopes are more highly con-
served overall. In particular, 69% of the epitopes targeted by
CD8?T-cell responses in seasonal H1N1 influenza isolates are
also present in all S-OIV isolates. The comparatively lower
fraction of conserved CD4?epitopes (41%) likely reflects the
relative larger size of the CD4?T-cell epitope as compared with
CD8?T-cell epitopes, which decreases the likelihood that a
CD4?T-cell epitope is totally conserved. As predicted by the
bioinformatic analysis, we were able to experimentally detect
significant levels of pre-existing T-cell immunity to sequences
totally conserved in S-OIV. Further analysis proved that this
level of immunity is comparable to that observed in sequences
conserved in the seasonal H1N1 influenza isolates. Although
T-cell responses do not prevent infection, they do contribute to
the clearance of infected target cells, and such pre-existing
immunity may lead to a less severe course of disease (16–19).
It is tempting to speculate on the significance of our finding
in terms of the disease severity and death rate of the S-OIV as
compared with seasonal human H1N1 influenza virus. It was
initially feared that the current S-OIV would be much more
lethal than seasonal H1N1 influenza. In fact, a recent study
demonstrated that a S-OIV isolate (from a hospitalized patient
in the United States) caused more severe pathogenicity in
infected mice, ferrets, and nonhuman primates than a currently
circulating human H1N1 influenza virus (3). Not all studies in
animal models unequivocally suggest increased virulence, as
transmission via respitory droplets in ferrets was found to be less
efficient for S-OIV compared with seasonal H1N1 by some
investigators (6) but not others (5). On the other hand, S-OIV
isolates were consistently found to replicate better in the lung
tissue of animal models (3, 5, 6). At the same time, the potential
pathogenic nature of S-OIV is not supported by the data
currently available that suggests that a large number of suspected
infections in the United States and a disproportionately low
number of deaths are associated with S-OIV. We propose that
the divergence between disease severity observed in most animal
studies and that found in the human population could be due to
the contribution of pre-existing T-cell–mediated immunity to
lessening disease severity. This was not a component of the
animal studies, as naive animals without previous exposure to
seasonal H1N1 influenza were infected with the S-OIV isolate.
Clearly several mechanisms contribute to the overall infectiv-
ity and virulence of a given influenza isolate/strain. Different
strains may vary in their capacity to interfere with IFN pathways
based on the activity of the particular sequence of the NS1
protein, as demonstrated in the case of 1918 H1N1 (20). Also,
changes in the HA receptor-binding domain can play a role in
(21, 22). These mechanisms are not addressed by our analysis.
Our analysis does, however, demonstrate that as far as immune
other seasonal H1N1 influenza isolates, perhaps explaining the
relatively mild nature of the S-OIV strain.
A counterargument against the relevance of our findings is as
follows: If conserved T-cell epitopes are indeed relevant targets
of protective pre-existing immune responses, why are they not
the neutralizing B-cell epitopes in HA and NA? There are four
possible explanations. First, our analysis demonstrated that most
B-cells epitope are derived from the more variable external
proteins, whereas a significant fraction of T-cell epitopes are
derived from the more conserved internal proteins. T-cell
epitopes in the variable HA and NA proteins were just as likely
to be mutated as B-cell epitopes. This suggests that mutations in
fitness, so that the virus is less capable of avoiding immune
responses against them. Second, it can be argued that influenza
viruses have found alternative ways to avoid the effect of T-cell
responses, notably by inhibiting IFN-related viral responses in
the host (23). This would mean that the virus does not need to
evade T-cell responses by mutation, as it has found other ways
of escaping from the immune response. Third, antibodies pro-
vide sterilizing immunity from infection, which make it essential
for the virus to escape from this response to spread. T-cell
responses, on the other hand, are thought to lessen disease
severity. If the virus retains the ability to infect and spread from
host to host, the pressure on avoiding T-cell immune recognition
may be less pronounced. Fourth, T-cell responses are much more
host specific than antibody responses, as they require presenta-
tion by different types of MHC complexes. Because the S-OIV
has been circulating through swine in the recent past, there has
Greenbaum et al. PNAS ?
December 1, 2009 ?
vol. 106 ?
no. 48 ?
not been the selection pressure to eliminate HLA-specific pep-
tides from the viruses.
We would like to point out several caveats that need to be kept
in mind when examining the results of our study. Our analysis is
based on epitope sequences that have been reported in the
literature and, as such, does not include epitopes that have not
yet been identified. This is a particularly serious limitation in
terms of epitopes present in strains dating back to years in which
epitope identification technologies were not widely available or
used. A second caveat is that we consider an epitope to be
conserved between two strains only if the sequence is 100%
identical. Clearly, this is a very conservative assumption, as
cross-reactivities between influenza epitopes with individual
residue substitutions have been observed frequently (24, 25).
The qualitative findings of our analysis should not be altered
significantly when allowing for such cross-reactivities, however,
as the relative conservation of T-cell and B-cell epitopes will
likely remain unchanged.
In conclusion, we have conducted an analysis of epitopes in
S-OIV based on experimentally identified epitopes cataloged
in the IEDB and on viral sequences that were rapidly published
in sequence databases available to scientists worldwide. An
initial version of the epitope analysis in this report was published
online through the IEDB to allow rapid dissemination of knowl-
edge. Our analysis provides insights into the relative conserva-
tion of B-cell and T-cell epitopes and has allowed formulating
the hypothesis experimentally tested in the present report. We
hope that the analysis and datasets provided with it prove useful
to experimental and computational immunologists, and that
awareness is raised of the ability to track immune epitope
conservation across multiple viral strains. Finally, this experi-
ence has demonstrated how ‘‘real time’’ exchange of information
on the Internet can catalyze the scientific process in the eye of
an imminent public health threat.
Materials and Methods
influenza A were retrieved from the IEDB, as described in the SI Text.
Querying for Epitopes Conserved in Circulating H1N1. The epitopes in the
human-relevant set described above were searched against H3N2 and H1N1
Querying for Epitopes Conserved in S-OIV (H1N1 2009). The set of circulating
H1N1 epitopes were searched against all S-OIV sequences (H1N1, 2009) and
the conservation of epitopes was calculated in the same manner as above.
Querying for Influenza Sequences. Circulating H1N1 and H3N2 influenza
sequences were obtained from the NCBI Influenza Virus Resource (http://
www.ncbi.nlm.nih.gov/genomes/FLU/FLU.html). For details on the queries,
see the SI Text.
At the time this analysis was performed, the NCBI database did not contain
sequences from Mexican isolates. These were retrieved from the GISAID
database. Table S3 lists the isolates for which sequences were obtained. All
Mexican isolates were submitted by the Centers for Disease Control and
Prevention (Atlanta, GA, Rebecca Garten).
Homology Modeling, Image Generation, and ASA Calculation. The homology
model for Fig. 1 was generated using the SWISS-MODEL homology-modeling
server (26), as described in the SI Text.
PBMC Isolation. PBMC were isolated from heparinized blood by gradient
centrifugation with a Histopaque-1077 (Sigma-Aldrich, St. Louis, MO), sus-
pended in fetal bovine serum (FBS) containing 10% dimethyl sulfoxide, and
cryopreserved in liquid nitrogen.
27. In brief, 2 ? 105PBMC (CD8?assays) or enriched CD4?T cells were
incubated with pools of peptides (1 ?g/ml per peptide). CD4?T cells were
positively enriched by incubating PBMC with anti-CD8 microbeads (Miltenyi
Biotech, Auburn, CA) for 15 min at 4 °C. After washing, PBMC were resus-
pended in magnetic cell sorting (MACS) buffer and passed through a magne-
tized LD column (Miltenyi Biotech). Enriched CD4?T cells were collected by
described in ref. 27.
Intracellular IFN-? Staining. PBMC were stimulated with pools of peptides (1
for an additional 6 h with the peptide pools. After incubation, cells were
stained for cell surface antigens CD4, CD8, CD45RA, CD62L, and CCR7. Cells
were then fixed, permeabilized, and stained for intracellular IFN-? using a
Cytofix/Cytoperm kit according to manufacturer’s directions (BD Biosciences).
After washing, samples were resuspended in phosphate-buffered saline, and
data were acquired on an LSRII flow cytometer (BD Biosciences). The fre-
quency of CD4?and CD8?T cells responding to each peptide pool was
quantified by determining the total number of gated CD4?or CD8?and
IFN-??T cells using FlowJo software (Tree Star, San Carlos, CA).
ACKNOWLEDGMENTS. This work was supported by National Institutes of
Health (NIH) contracts HHSN26620040006C, N01-AI30039, and N01-AI40041.
1. Cohen J, Enserink M (2009) Swine flu. After delays, WHO agrees: The 2009 pandemic
has begun. Science 324:1496–1497.
2. Garten RJ, et al. (2009) Antigenic and genetic characteristics of swine-origin 2009
A(H1N1) influenza viruses circulating in humans. Science 325:197–201.
3. Itoh Y, et al. (2009) In vitro and in vivo characterization of new swine-origin H1N1
influenza viruses. Nature, in press.
4. Centers for Disease Control and Prevention (2009) Serum cross-reactive antibody
response to a novel influenza A (H1N1) virus after vaccination with seasonal influenza
vaccine. MMWR 58:521–524.
5. Munster VJ, et al. (2009) Pathogenesis and transmission of swine-origin 2009 A(H1N1)
influenza virus in ferrets. Science 325:481–483.
6. Maines TR, et al. (2009) Transmission and pathogenesis of swine-origin 2009 A(H1N1)
influenza viruses in ferrets and mice. Science 325:484–487.
7. Bui HH, Peters B, Assarsson E, Mbawuike I, Sette A (2007) Ab and T-cell epitopes of
influenza A virus, knowledge and opportunities. Proc Natl Acad Sci USA 104:246–251.
to blueprint. PLoS Biol 3:e91.
9. Bao Y, et al. (2008) The influenza virus resource at the National Center for Biotech-
nology Information. J Virol 82:596–601.
10. Bogner P, Capua I, Lipman DJ, Cox NJ, et al. (2006) A global initiative on sharing avian
flu data. Nature 442:981.
11. Nguyen HH, et al. (2007) Heterosubtypic immunity to influenza A virus infection
requires a properly diversified antibody repertoire. J Virol 81:9331–9338.
Dev Biol (Basel) 115:25–30.
13. Prabhu N, et al. (2009) Monoclonal antibodies against the fusion peptide of hemag-
glutinin protect mice from lethal influenza A virus H5N1 infection. J Virol 83:2553–
14. Liew FY, Russell SM, Appleyard G, Brand CM, Beale J (1984) Cross-protection in mice
infected with influenza A virus by the respiratory route is correlated with local IgA
antibody rather than serum antibody or cytotoxic T-cell reactivity. Eur J Immunol
response induced in trivalent inactivated vaccine recipients. J Virol 82:9283–9287.
16. McMichael AJ, Gotch FM, Noble GR, Beare PA (1983) Cytotoxic T-cell immunity to
influenza. N Engl J Med 309:13–17.
17. Webby RJ, et al. (2003) Protection and compensation in the influenza virus-specific
CD8? T-cell response. Proc Natl Acad Sci USA 100:7235–7240.
18. McElhaney JE, et al. (2006) T-cell responses are better correlates of vaccine protection
in the elderly. J Immunol 176:6333–6339.
histocompatibility complex-restricted T cells have delayed viral clearance and in-
creased mortality after influenza virus challenge. J Exp Med 175:1143–1145.
20. Basler CF, et al. (2001) Sequence of the 1918 pandemic influenza virus nonstructural
gene (NS) segment and characterization of recombinant viruses bearing the 1918 NS
genes. Proc Natl Acad Sci USA 98:2746–2751.
21. Stevens J, et al. (2006) Structure and receptor specificity of the hemagglutinin from an
H5N1 influenza virus. Science 312:404–410.
22. Yen HL, et al. (2009) Changes in H5N1 influenza virus hemagglutinin receptor binding
domain affect systemic spread. Proc Natl Acad Sci USA 106:286–291.
23. Garcia-Sastre A, et al. (1998) Influenza A virus lacking the NS1 gene replicates in
interferon-deficient systems. Virology 252:324–330.
24. Kreijtz JH, et al. (2008) Cross-recognition of avian H5N1 influenza virus by human
cytotoxic T-lymphocyte populations directed to human influenza A virus. J Virol
25. Lee LY, et al. (2008) Memory T cells established by seasonal human influenza A
infection cross-react with avian influenza A (H5N1) in healthy individuals. J Clin Invest
26. Schwede T, Kopp J, Guex N, Peitsch MC (2003) SWISS-MODEL: An automated protein
homology-modeling server. Nucleic Acids Res 31:3381–3385.
27. Oseroff C, et al. (2005) HLA class I-restricted responses to vaccinia recognize a broad
Sci USA 102:13980–13985.
www.pnas.org?cgi?doi?10.1073?pnas.0911580106 Greenbaum et al.