Asp222 in the wild-type CA/04 HA is a characteristic residue predominantly found in human-adapted H1N1 viruses. This residue has been shown to play a key role in providing optimal contacts with the penultimate Gal sugar in glycans terminated by α2-6-linked sialic acid [25], [26]. Therefore a D222G mutation is likely to affect the molecular contacts of this key residue position with the glycan receptor. It was demonstrated previously that an analogous mutation in the prototypic 1918 pandemic (A/South Carolina/1/1918 or SC18) HA resulted in shifting its glycan receptor-binding from high specificity for α2-6 sialylated glycans to a mixed binding to both α2-6 (albeit at a lower affinity than wild-type HA) and α2-3 sialylated glycans [27], [28].

To investigate the effect of D222G mutation on glycan-receptor binding properties of CA/04 virus, we analyzed the binding of both wild-type and mutant virus in a dose-dependent fashion to representative α2-3 and α2-6 sialylated glycans on a glycan array platform (Figure 1). Our analysis showed that D222G mutation resulted in a modest reduction in the binding avidity to α2-6 sialylated glycans (6′SLN and, to a lesser extent, 6′SLN-LN) in comparison with the wild-type virus. The notable difference in glycan-binding properties was the substantial increase in binding of D222G mutant virus to α2-3 sialylated glycans (relative to its α2-6 binding) compared with that of wild-type virus. These results with CA/04 virus are consistent with previous observations demonstrating an increase in α2-3 binding of other 2009 H1N1 viruses carrying the D222G mutation [24].

A recent study demonstrated that the introduction of D222G in a 2009 H1N1 virus did not alter the pathogenicity or transmissibility of a 2009 H1N1 isolate A/Netherlands/602/09 in ferrets [22]. However, the use of a parental strain which already possesses a highly transmissible phenotype may limit the ability to identify additional molecular changes which could confer enhanced transmissibility [29]. In contrast, the 2009 H1N1 CA/04 virus, which exhibits reduced transmissibility in the ferret model compared with seasonal H1N1 viruses, and previously enabled us to identify an HA mutation which conferred efficient transmission by respiratory droplets in ferrets was used in this study to more stringently assess the impact of D222G on virus transmission [30], [31].

Ferrets inoculated with the D222G virus exhibited similar clinical signs as CA/04 wild-type virus and a reverse-genetics derived CA/04 virus, with ferrets exhibiting transient weight loss (mean maximum weight loss of 11.3% on day 6–7 p.i.) and fever (mean maximum fever 1.9°C over baseline day 2 p.i.) before returning to baseline levels [30], [31]. D222G virus-infected ferrets further exhibited a mild, transient lymphopenia (22% decrease in circulating lymphocytes on day 3 p.i.) which was comparable to ferrets infected with other 2009 H1N1 viruses [32]. D222G virus was detected in the ferret upper respiratory tract (6.7±0.2 log₁₀ PFU/ml nasal turbinates) and the lower respiratory tract (6.3±0.6 log₁₀ PFU/g lung tissue) at similar titers as 2009 H1N1 wild-type viruses day 3 p.i. [22]. Moreover, the D222G mutant virus was detected in the intestinal tract of 1/3 ferrets day 3 p.i. (2.29 log₁₀ PFU/ml), and in rectal swabs collected on days 1 and 3 p.i. in 2/3 ferrets (average titer 1.4±0.2 log₁₀ PFU/ml), comparable to other 2009 H1N1 isolates [31].

To evaluate the transmissibility of D222G virus by respiratory droplets, three ferrets were inoculated with 10⁶ PFU of virus. Approximately 24 hours p.i., inoculated-contact pairs were established by placing naïve ferrets in each of three adjacent modified cages with a perforated side wall, allowing air exchange between ferrets in the absence of direct or indirect contact. Nasal washes (NW) were collected on alternate days p.i. or post-contact (p.c.) for virus titration. Detection of infectious virus in NW of contact ferrets and seroconversion of contact ferrets were considered evidence of virus transmission. Consistent with the experimental transmission data obtained with wild type 2009 H1N1 viruses [31], CA/04 virus did not spread by respiratory droplet to every contact ferret; infectious virus was detected in the nasal washes of two of three contact ferrets (Figure 2A) [30]. D222G virus replicated efficiently in the upper respiratory tract of inoculated ferrets, reaching peak mean titers of 7.5±0.2 log₁₀ PFU/ml day 1 p.i., significantly higher than CA/04 virus at this time (p<0.03) [30]. Similar to CA/04 virus, D222G mutant virus was detected in NW of two of three contact ferrets, with virus titers >7 log₁₀ PFU/ml on day 3 p.c. (Figure 2B). While detectable virus was not observed in NW of the third contact ferret, all three contact ferrets seroconverted to homologous virus by day 20 p.i. (HI titer range 320–640). This pattern of D222G virus transmission by respiratory droplets was confirmed in a duplicate experiment that resulted in virus detection in NW in two of three contact ferrets and seroconversion of all contact ferrets (data not shown). Ferret lung and NW samples from inoculated and contact animals that were harvested were sequenced to confirm the mutation. All samples yielded the expected mutant sequence; reversion of the mutation (222G to 222D) was not observed.

To determine whether the D222G mutation conferred heightened virulence of a 2009 H1N1 virus in a mouse model, we inoculated mice i.n. with 10⁵ PFU of either CA/04 or D222G virus. Mice inoculated with D222G virus by the intranasal route exhibited greater weight loss than CA/04 virus-inoculated mice; one mouse from the D222G group was euthanatized on day 11 p.i. due to excessive weight loss (Table 1). However, neither virus exhibited >50% lethality in mice when inoculated at this dose. While both viruses replicated efficiently in the lungs of mice on days 3 and 6 p.i., D222G virus replicated to a significantly higher mean titer compared with CA/04 virus at day 6 p.i. (p<0.05). Virus was found infrequently and at low titer in the nose and intestines of mice from both groups on day 3 p.i. only (Table 1 and data not shown). Similar titers of virus in the lung and sporadic detection of virus in intestinal tissue with these viruses was observed with numerous other 2009 H1N1 viruses in this model [33].

A recent report indicated that the presence of D222G enhanced the ability of a 2009 H1N1 virus to cause ocular disease in mice, with virus detected in the eye of one mouse following intranasal inoculation [22]. We did not observe extrapulmonary spread of virus to the eyes of mice following intranasal inoculation with either CA/04 or D222G virus, and ocular disease was not observed during visual examination of mice during the course of infection (data not shown). However, to better assess the ability of the D222G mutation to cause ocular disease in mice, we inoculated mice by the ocular route with both viruses. The right eye of each mouse was lightly scarified with a 2-mm corneal trephine blade, and 10⁵ PFU of each virus in a 5 µl volume was deposited onto the corneal epithelial surface and massaged into the eye with the eyelid. Mice inoculated by the ocular route with either virus did not exhibit substantial morbidity, and virus was not detected in the eye, nose, intestine, or lung on day 3 or 6 p.i. (data not shown). These results indicate that CA/04 virus is not well suited to infect mice by the ocular route, and the presence of D222G does not confer an ocular tropism in this model.

Previous studies have demonstrated that select influenza viruses which do not demonstrate an ocular tropism in mice nonetheless are capable of high-titer replication in human ocular cells [34]. To rule out the possibility that the D222G mutation confers an ocular tropism not observed in our murine model, we compared the ability of CA/04 and D222G viruses to replicate in primary human corneal epithelial cells (HCEpiC), which predominantly express α2-3 linked sialic acids [34]. HCEpiC were infected at a MOI of 0.01 and supernatants were collected at indicated times p.i. to quantify infectious virus (Figure 3, open symbols). Both CA/04 and D222G viruses were detected at low levels (<3 log₁₀ PFU/ml) through 72 hrs p.i. in this cell type. In contrast, both viruses replicated efficiently in the human bronchial epithelial cell line Calu-3 (Figure 3, closed symbols). Similar to a previous study, we observed a slight delay in D222G virus replication in Calu-3 cells at 24 hrs p.i., with titers of D222G virus being 10-fold lower than CA/04 virus (p<0.02) [22]. However, at 48 and 72 hrs p.i., D222G virus replicated to significantly higher titer than CA/04 virus in the human respiratory cells (p<0.0005). Primary human lung blood microvascular endothelial cells (HMVEC-LBI) express a greater proportion of α2-3 linked sialic acids compared with Calu-3 cells (H. Zeng, personal communication), but did not support efficient replication of either CA/04 or D222G viruses. Taken together, these findings indicate that the D222G mutation confers enhanced replication in respiratory epithelial cells but does not alter the ocular tropism of 2009 H1N1 virus.

The eight reverse-genetics plasmids used for the rescue of recombinant influenza A/California/04/09 (CA/04) virus were constructed as described previously [30], [44], [45]. The mutation from Asp to Gly at HA position 222 was achieved by altering the position 222 codons from GAT(D) to GGT(G) using a Stratagene QuikChange II Site-Directed Mutagenesis kit (Agilent Technologies, Santa Clara, CA) per manufacturer's instructions on the CA/04 HA plasmid template.

CA/04 and 222G viruses were rescued as described previously [30]. Briefly, 293T cells were transfected with the eight CA/04 plasmids and co-cultured with MDCK cells 12 hrs post-transfection. Rescued viruses were isolated by plaque purification on MDCK cells. Coding sequences of plasmids and rescued viruses were confirmed by automated sequencing performed at the Influenza Sequence Activity, Influenza Division, Centers for Disease Control and Prevention. Experiments with CA/04 and 222G viruses were conducted under biosafety level 3 containment with enhancements, in accordance with guidelines of the WHO. (https://www.who.int/csr/resources/publications/swineflu/Laboratorybioriskmanagement.pdf).

A streptavidin plate array comprising representative biotinylated α2→3 and α2→6 sialylated glycans as described previously [28], [30] (Table S1) was used for the analysis. 3′SLN, 3′SLN-LN, 3′SLN-LN-LN are representative avian receptors. 6′SLN and 6′SLN-LN are representative human receptors. The biotinylated glycans were obtained from the Consortium of Functional Glycomics through their resource request program. Streptavidin-coated High Binding Capacity Pierce 384-well Microplates (Pierce, Rockford, IL) were loaded to the full capacity of each well by incubating the well with 50 µl of 2.4 µM of biotinylated glycans overnight at 4°C. Excess glycans were removed through extensive washing with PBS. The viruses were diluted based on hemagglutinating units (HAU, obtained using 0.5% turkey erythrocytes) with 1X PBS+1% BSA. 50 µl of the diluted virus was added to each of the glycan – coated wells and incubated overnight at 4°C. This was followed by three washes with 1X PBST (1X PBS+0.1% Tween-20) and three washes with 1X PBS. Each of the wells was blocked with 1X PBS+1% BSA for 2 h at 4°C. The wells were incubated with primary antibody (ferret anti – CA/04 antisera; 1∶500 diluted in 1X PBS+1% BSA) and incubated for 5 h at 4°C. This was followed by three washes with 1X PBST and three washes with 1X PBS. Finally the wells were incubated with the secondary antibody (goat anti – ferret HRP conjugated antibody from Abcam; 1∶500 diluted in 1X PBS+1% BSA). The wells were washed with 1X PBST and 1X PBS as before. The binding signals were determined based on the HRP activity using the Amplex Red Peroxidase Assay (Invitrogen) according to the manufacturer's instructions. The assays were done in triplicate and appropriate negative controls were included.

Male Fitch ferrets (Triple F Farms, Sayre, PA), 8–10 months of age and serologically negative by hemagglutination inhibition (HI) assay for currently circulating influenza viruses, were used in this study. Ferrets were housed for the duration of the experiment in cages within a Duo-Flow Bioclean environmental enclosure (Lab Products, Seaford, DE). The pathogenesis of each virus was assessed following intranasal (i.n.) inoculation of 10⁶ PFU of virus in a 1 ml volume, a dose reported to consistently infect ferrets with human or avian influenza viruses, was determined as previously described [46], [47]. Ferret hematologic and blood chemistry analyses were performed as previously described [32]. Transmission of virus by respiratory droplets was performed as previously described [38]. Nasal washes were collected from ferrets every other day for at least 9 days post-inoculation (p.i.) or contact (p.c.) and titrated by standard plaque assay using MDCK cells. Post-exposure serum was collected from virus-inoculated (15–19 days p.i.) and contact (18–21 days p.c.) ferrets and tested for H1-specific antibodies against homologous virus by HI assay using 0.5% turkey erythrocytes [48].

Female BALB/c mice (Charles River Laboratories, Wilmington, MA), 8–12 weeks of age, were inoculated i.n. with 10⁵ PFU of CA/04 or D222G virus in a 50 µl volume as previously described [33]. Ocular infection of mice with CA/04 and D222G viruses was performed as previously described using 10⁵ PFU in a 5 µl volume [34]. Five mice per group were monitored daily for 14 days p.i. for morbidity, as measured by body weight, and mortality. Any mouse which lost >25% of its pre-inoculation body weight was euthanatized. Three mice per group were euthanatized on days 3 and 6 p.i., and eye, nose, intestine, and lung tissues were collected to determine replication and systemic spread of virus. Tissues were homogenized in 1 ml of cold PBS, and clarified homogenates were titrated by standard plaque assay. Statistical significance for these and all experiments was determined using Student's t test.

The bronchial epithelial cell line Calu-3 (ATCC, Manassas, VA) was cultured on membrane inserts and infected apically as previously described [49]. Primary human corneal epithelial cells (HCEpiC) were obtained from ScienCell (San Diego, CA) at passage 1 and were prepared for virus infection as described previously [34]. Primary human lung blood microvascular endothelial cells (HMVEC-LBI) were obtained from Lonza (Walkersville, MD) and cultured with Endothelial Cell Basal Medium-2 (Lonza) (Zeng et al, in preparation). Virus was added to Calu-3, HCEpiC, or HMVEC-LBI monolayers at a multiplicity of infection (MOI) of 0.01 for one hour before washing. Cell type specific serum-free media was added to all wells with the addition of 300 µg/L N-p-tosyl-L-phenylalanine chloromethyl ketone-treated (TPCK) trypsin (Sigma-Aldrich, St. Louis, MO) in cultures using primary human cells. Aliquots of culture supernatant taken post-infection (p.i.) from triplicate cultures were immediately frozen at −80°C until titrated for the presence of infectious virus by standard plaque assay.

All animal research described in this study was specifically approved by CDC's Institutional Animal care and Use Committee (IACUC). The animal research was conducted under the guidance of CDC's IACUC and in an Association for Assessment and Accreditation of Laboratory Animal Care International-accredited facility.