Virulence and Genetic Compatibility of Polymerase Reassortant Viruses Derived from the Pandemic (H1N1) 2009 Influenza Virus and Circulating Influenza A Viruses^,

Abstract

Gene mutations and reassortment are key mechanisms by which influenza A virus acquires virulence factors. To evaluate the role of the viral polymerase replication machinery in producing virulent pandemic (H1N1) 2009 influenza viruses, we generated various polymerase point mutants (PB2, 627K/701N; PB1, expression of PB1-F2 protein; and PA, 97I) and reassortant viruses with various sources of influenza viruses by reverse genetics. Although the point mutations produced no significant change in pathogenicity, reassortment between the pandemic A/California/04/09 (CA04, H1N1) and current human and animal influenza viruses produced variants possessing a broad spectrum of pathogenicity in the mouse model. Although most polymerase reassortants had attenuated pathogenicity (including those containing seasonal human H3N2 and high-pathogenicity H5N1 virus segments) compared to that of the parental CA04 (H1N1) virus, some recombinants had significantly enhanced virulence. Unexpectedly, one of the five highly virulent reassortants contained a A/Swine/Korea/JNS06/04(H3N2)-like PB2 gene with no known virulence factors; the other four had mammalian-passaged avian-like genes encoding PB2 featuring 627K, PA featuring 97I, or both. Overall, the reassorted polymerase complexes were only moderately compatible for virus rescue, probably because of disrupted molecular interactions involving viral or host proteins. Although we observed close cooperation between PB2 and PB1 from similar virus origins, we found that PA appears to be crucial in maintaining viral gene functions in the context of the CA04 (H1N1) virus. These observations provide helpful insights into the pathogenic potential of reassortant influenza viruses composed of the pandemic (H1N1) 2009 influenza virus and prevailing human or animal influenza viruses that could emerge in the future.

INTRODUCTION

The zoonotic transfer in April 2009 and global spread among humans of a novel influenza A (H1N1) virus, a reassortant between North American and European swine influenza virus lineages (17), prompted the first influenza pandemic of the 21st century (9, 12). Studies using mammalian models, such as mice, pigs, ferrets, and nonhuman primates, have shown that this virus is more pathogenic than seasonal influenza viruses (26, 33, 39). Although the pandemic (H1N1) 2009 influenza virus is highly transmissible, having spread to more than 214 countries, overall mortality rates remain relatively low (66) compared to those of previous influenza pandemics. This low mortality has been partially credited to the virus' lack of known virulence factors (17, 43).

The pathogenesis of influenza viruses is a polygenic trait wherein molecular determinants may differ among animal species (6, 43). Molecular factors that define the virulence and host range of influenza viruses frequently have been linked to the viral surface glycoproteins, particularly hemagglutinin, and their interaction with sialic acid receptors on the host's epithelial cells (21, 61). Nonstructural protein 1 (NS1) appears to contribute to the efficient replication of human influenza A viruses (i.e., 1918 pandemic virus) and is a virulence determinant of highly pathogenic avian influenza (HPAI) H5N1 viruses by restricting the induction of the host interferon response (11, 16, 60, 62). The influenza polymerase genes also play significant roles in virus adaptation and pathogenesis. Of the identified molecular markers of virulence on the viral polymerases, the glutamic acid (E)-to-lysine (K) alteration at position 627 in PB2 is perhaps most commonly observed, particularly among HPAI H5N1 and H7 viruses (21, 28, 40, 58). It also was reportedly responsible for viral replication at lower temperatures (33°C), allowing enhanced growth in the upper respiratory tract of mammals and efficient transmission altogether (22, 35, 57, 63). Some avian influenza viruses lacking the E627K substitution spontaneously acquire it during passage in mice, thereby increasing virus replication and pathogenicity (32, 34, 56).

In the absence of 627K, the aspartic acid (D)-to-asparagine (N) alteration at residue 701 of the PB2 protein has been shown to similarly enhance growth and transmission in guinea pigs (57) and to expand the host range of avian H5N1 viruses to mice and humans (13, 32, 57). The expression of the PB1-F2 protein, which is associated with mitochondrial targeting, membrane disruption, and subsequent apoptosis, was suggested to contribute to virulence (8, 18, 67). This alternatively spliced PB1 protein also contributes to the degree of the virulence of influenza virus (10, 67) and secondary bacterial infections (36). We also have previously reported that a threonine (T)-to-isoleucine (I) substitution in PA at position 97 resulting from passage in mice increased the virulence of a low-pathogenic avian H5N2 influenza virus (56). Surprisingly, these markers in PB2, PB1, and NS1 do not significantly affect the replication, pathogenesis, or transmission of the pandemic virus when individually incorporated (19, 20, 23, 68), suggesting that although these mutations may occur in nature, the generation of more virulent viruses is due to other mutations or genetic factors.

The continued spread of the 2009 pandemic influenza virus among human populations worldwide also caused the infection of various animals, most importantly turkeys and swine (46), presenting opportunities for genetic reassortment with cocirculating animal influenza viruses. Therefore, in this study, aside from examining the effect of adding specific molecular virulence markers to the background of the A/California/04/09 [CA04 (H1N1)] virus, we also examined whether reassortment with current human and animal influenza viruses could affect pathogenicity. We focused on the influenza virus' polymerase replication machinery, because viral proteins involved in transcribing and replicating the viral genome harbor determinants for host range and virulence (15, 42, 52). However, the replication and transcription activity of the viral polymerase complex also are restricting factors for reassortment (30). We used reverse genetics to generate a panel of mutant and reassortant viruses that subsequently were assayed for pathogenicity in a number of assays. Although most of the generated viruses, including those reassortant viruses containing the polymerase segments of HPAI H5N1 and seasonal human H3N2 viruses, had attenuated pathogenicity compared to that of the wild-type (WT) CA04 (H1N1) virus, some reassortants had significantly enhanced virulence in the mouse model. Our results provide insights into the pathogenic potential of reassortant viruses that could emerge in the future.

MATERIALS AND METHODS

Cells and viruses.

Madin-Darby canine kidney (MDCK) cells were grown in minimum essential medium with Eagle salts (EMEM; Lonza) containing 5% fetal bovine serum (FBS), and 293T human embryonic kidney (HEK) cells were grown in Dulbecco's modified Eagle's medium containing 10% FBS (Gibco). The human pandemic H1N1 virus, CA04 (H1N1), was obtained from St. Jude Children's Research Hospital. The human, avian, and swine influenza viruses, including the mammalian-passaged variants (Table 1), used in this study were from our repository of human and animal influenza viruses collected from the following sources: nasopharyngeal suctions of patients at the Chungbuk National University Hospital; lung tissue samples of swine from local pig farms; and fecal samples of birds from live bird markets, poultry farms, or the wild (2, 3, 29, 48, 55, 56). After the sequences of the rescued viruses were confirmed, virus stocks were prepared in 10-day-old embryonated chicken eggs and titrated in MDCK cells. Viral growth kinetics were determined in MDCK cells infected at a multiplicity of infection (MOI) of 10⁻⁴. Supernatants were harvested at 12, 24, 36, 48, and 60 h postinfection (pi) for virus titration in MDCK cells. Phenotypic characterization by plaque assay was done in 6-well plates of MDCK cells infected with serial 10-fold dilutions of reassortant viruses. Diameters of plaques formed were measured at 60 h pi.

Table 1.

Viruses used in this study^a

Virus strain	Subtype	Host	Abbreviation	PB2 alteration at:		PB1-F2 protein expression	PA alteration at aa 97
				aa 627	aa 701
A/California/04/09	H1N1	Human	CA04 (H1N1)	E	D	No	T
A/Puerto Rico/8/34	H1N1	Human	PR8 (H1N1)	K	D	Yes	T
A/Sw/Korea/CAN01/04	H1N1	Swine	Sw/CAN01 (H1N1)	E	D	Yes	T
A/Cheongju/H407/08	H3N2	Human	H407 (H3N2)	K	D	Yes	T
A/Sw/Korea/JNS06/04	H3N2	Swine	Sw/JNS06 (H3N2)	E	D	Yes	T
A/Dk/Korea/LPM91/06	H3N2	Avian	Av/LPM91 (H3N2)	E	D	Yes	T
A/EM/Korea/W150/06	H5N1	Avian	Av/W150 (H5N1)b	K	D	Yes	T
A/Ab/Korea/W81/05	H5N2	Avian	Av/W81 (H5N2)	E	D	Yes	T
A/Ab/Korea/ma81/07	H5N2	ma	Av/ma81 (H5N2)	E	D	Yes	I
A/Ab/Korea/ma81K/07	H5N2	ma	Av/ma81K (H5N2)c	K	D	Yes	I
A/Ab/Korea/W44/05	H7N3	Avian	Av/W44 (H7N3)	E	D	Yes	T
A/Ab/Korea/ma44/07	H7N3	ma	Av/ma44 (H7N3)	K	D	Yes	I
A/Md/Korea/6L/07	H7N6	Avian	Av/6L (H7N6)	E	D	Yes	T
A/Ck/Korea/04116/04	H9N2	Avian	Av/04116 (H9N2)	E	D	Yes	T
A/Ck/Korea/04164/04	H9N8	Avian	Av/04164 (H9N8)	E	D	Yes	T

^aThese human, swine, and avian influenza viruses were used to generate polymerase reassortant viruses with the pandemic (H1N1) 2009 virus (CA04). Accession numbers for the viral polymerase genes (PB2, PB1, and PA) used are the following: Sw/CAN01 (H1N1), EU798918, EU798898, and EU798878; H407 (H3N2), FJ009502, FJ009455, and FJ009455; Sw/JNS06 (H3N2), EU301177, EU301400, and EU301368; Av/LPM91 (H3N2), EU301207, EU301430, and EU301398; Av/W150 (H5N1), EU233746, EU233745, and EU233744; Av/W81 (H5N2), GU361390, GU361308, and GU361350; Av/ma81 (H5N2), HQ913050 (note that A/Ab/Korea/ma81K/07 differs from A/Ab/Korea/ma81/07 only by the presence of the E627K mutation in PB2; therefore, PB1 and PA for these two viruses are the same); Av/W44 (H7N3), HQ913044, HQ913051, and HQ913057; Av/ma44 (H7N3), HQ913048, HQ913055, and HQ913061; Av/6L (H7N6), HQ913047, HQ913054, and HQ913060; Av/04116 (H9N2), HQ913045, HQ913052, and HQ913058; Av/04164 (H9N8), HQ913046, HQ913053, and HQ913059. Abbreviations: aa, amino acid; ma, mouse adapted; Ab, aquatic bird; Ck, chicken; Dk, duck; EM, environment; Md, mallard; Sw, swine.

^bHighly pathogenic H5 avian influenza virus (29).

^cThis virus contains PB2_627K as an additional adaptive mutation.

Plasmids and generation of recombinant viruses.

All eight gene segments of the CA04 (H1N1) virus and the polymerase genes (PB2, PB1, and PA) of the 14 other influenza viruses used (Table 1) were synthesized and cloned into the pHW2000 plasmid vector to generate reverse genetic (Rg) viruses as described previously (56). All point mutant and recombinant viruses (Tables 2 and 3; also see Table S1 in the supplemental material) were rescued in six-well plates of cocultured 293T HEK and MDCK cell mixtures (3:1 ratio) that were transfected with the eight viral plasmids as indicated, each containing 1 μg of the respective gene segment. Transfection was performed by using TransIT-LT1 transfection reagent (Mirus Bio) according to the manufacturer's instructions. The transfection medium was removed after 6 h and replaced with Opti-MEM I (Gibco) containing 0.3% bovine serum albumin and 0.01% FBS. After 30 h, 1 ml of Opti-MEM I containing 0.2 μg/ml of l-1-tosylamido-2-phenylethyl chloromethyl ketone (TPCK)-treated trypsin (Sigma-Aldrich) was added to the transfected cells. The supernatant was harvested after 48 h and injected into 11-day-old embryonated chicken eggs for virus propagation. All rescued viruses were resequenced to ensure the absence of unwanted mutations and the presence of the desired gene combinations (Cosmo GeneTech, Seoul, Republic of Korea). The infectivity of the stock viruses was determined by calculating the 50% tissue culture infective dose (TCID₅₀) by the method of Reed and Muench (50). Briefly, viruses were serially diluted 10-fold (pipette tips were changed at every dilution point) in EMEM without FBS. MDCK cells in 96-well plates prepared 24 h previously were washed twice with 1× phosphate-buffered saline (PBS), followed by the addition of the virus dilutions (in replicate wells). After 1 h of virus adsorption, cells were washed twice with PBS and appropriate culture medium containing 1 μg/ml of TPCK-trypsin (Sigma-Aldrich) was added. Inoculated cells were incubated at 37°C and monitored daily for cytopathic effects within 60 h. Endpoint titers also were confirmed by performing a hemagglutination test using 0.5% turkey erythrocytes. All viruses were phenotypically characterized by performing plaque assays in MDCK cells as described previously (56). The handling of viruses and virus rescue were performed in an enhanced biosafety level 3 containment laboratory as approved by the Korean Centers for Disease Control and Prevention.

Table 2.

In vivo characterization of CA04 backbone polymerase mutant viruses^a

Virus name	PB1-F2 protein expression	PB2 alteration at:		MLD50b (log10 TCID50)	Survivalc (%)
		aa 627	aa 701
Rg CA04	No	E	D	5.8	80
CA04 PB2627K	No	K	D	5.8	80
CA04 PB2701N	No	E	N	5.8	80
CA04 PB1-F2	Yes	E	D	≥6.0	100
CA04 PB2627K-701N	No	K	N	5.8	80
CA04 PB1-F2 PB2627K	Yes	K	D	5.7	70
CA04 PB1-F2 PB2701N	Yes	E	N	5.9	90
CA04 PB1-F2 PB2627K-701N	Yes	K	N	5.9	90

^aAbbreviations: aa, amino acid; MLD₅₀, 50% mouse lethal dose; TCID₅₀, 50% tissue culture infectious dose as determined in Madin-Darby canine kidney cells.

^bDetermined by inoculating groups of 10 mice with 10^5.5 TCID₅₀ of the respective mutant viruses.

^cSurvival was monitored daily for 13 days pi.

Table 3.

Replicative efficiency of polymerase reassortant viruses in vitro^d

Reassortant no. and replication groupa	Virus name	Polymerase segment replaced in:			Titerb (TCID50/ml)	Mean (SD) plaque sizec (mm)
		PA	PB1	PB2
Highly replicative
r1	Rg CA04	CA04	CA04	CA04	1.6 × 107	0.92 (0.21)
r6	LPM91 PA + CA04	L91	CA04	CA04	1.0 × 107	0.99 (0.3)
r8	W81 PA + CA04	W81	CA04	CA04	1.6 × 107	0.83 (0.23)
r9	ma81 PA + CA04	ma81	CA04	CA04	1.0 × 107	1.41 (0.47)
r21	W81 PB1 + CA04	CA04	W81	CA04	1.6 × 107	1.05 (0.35)
r22	ma81 PB1 + CA04	CA04	ma81	CA04	1.6 × 107	0.89 (0.3)
r24	ma44 PB1 + CA04	CA04	ma44	CA04	1.0 × 107	0.84 (0.17)
r25	6L PB1 + CA04	CA04	6L	CA04	1.3 × 107	0.92 (0.31)
r26	04116 PB1 + CA04	CA04	04116	CA04	1.6 × 107	1.27 (0.33)
r28	PR8 PB2 + CA04	CA04	CA04	PR8	2.5 × 107	1.31 (0.36)
r29	CAN01 PB2 + CA04	CA04	CA04	CAN01	1.0 × 107	1.12 (0.36)
r48	W81 PA-PB1 + CA04	W81	W81	CA04	5.0 × 107	1.07 (0.38)
r49	ma81 PA-PB1 + CA04	ma81	ma81	CA04	1.6 × 107	0.93 (0.28)
r56	CAN01 PB1-PB2 + CA04	CA04	CAN01	CAN01	1.3 × 108	1.44 (0.26)
r61	W81 PB1-PB2 + CA04	CA04	W81	W81	1.0 × 107	1.21 (0.39)
r62	ma81 PB1-PB2 + CA04	CA04	ma81	ma81	3.2 × 108	1.93 (0.4)
r63	ma81K PB1-PB2 + CA04	CA04	ma81	ma81K	3.2 × 108	1.3 (0.5)
r70	CAN01 PA-PB2 + CA04	CAN01	CA04	CAN01	2.0 × 107	0.92 (0.24)
r77	ma81K PA-PB2 + CA04	ma81	CA04	ma81K	1.6 × 107	1.01 (0.19)
Moderately replicative
r5	JNS06 PA + CA04	JNS06	CA04	CA04	1.6 × 106	0.84 (0.17)
r10	W44 PA + CA04	W44	CA04	CA04	1.6 × 106	0.75 (0.24)
r13	04116 PA + CA04	04116	CA04	CA04	1.6 × 106	0.87 (0.27)
r14	04164 PA + CA04	04164	CA04	CA04	1.3 × 106	0.67 (0.19)
r15	PR8 PB1 + CA04	CA04	PR8	CA04	5.0 × 106	0.72 (0.19)
r16	CAN01 PB1 + CA04	CA04	CAN01	CA04	3.2 × 106	0.98 (0.31)
r18	JNS06 PB1 + CA04	CA04	JNS06	CA04	7.9 × 106	1.06 (0.37)
r23	W44 PB1 + CA04	CA04	W44	CA04	5.0 × 105	0.82 (0.32)
r30	H407 PB2 + CA04	CA04	CA04	H407	5.0 × 106	1.15 (0.18)
r31	JNS06 PB2 + CA04	CA04	CA04	JNS06	1.6 × 106	0.96 (0.28)
r32	LPM91 PB2 + CA04	CA04	CA04	L91	3.2 × 106	0.53 (0.08)
r33	W150 PB2 + CA04	CA04	CA04	W150	1.6 × 106	0.73 (0.06)
r34	W81 PB2 + CA04	CA04	CA04	W81	1.3 × 106	0.7 (0.19)
r35	ma81 PB2 + CA04	CA04	CA04	ma81	1.6 × 106	0.83 (0.19)
r36	ma81K PB2 + CA04	CA04	CA04	ma81K	1.6 × 106	1.01 (0.29)
r37	W44 PB2 + CA04	CA04	CA04	W44	5.0 × 105	0.8 (0.15)
r38	ma44 PB2 + CA04	CA04	CA04	ma44	1.6 × 106	0.84 (0.23)
r39	6L PB2 + CA04	CA04	CA04	6L	5.0 × 105	0.69 (0.15)
r50	W44 PA-PB1 + CA04	W44	W44	CA04	1.6 × 106	0.88 (0.22)
r52	6L PA-PB1 + CA04	6L	6L	CA04	5.0 × 106	1.15 (0.41)
r53	04116 PA-PB1 + CA04	04116	04116	CA04	1.3 × 106	1.13 (0.35)
r55	PR8 PB1-PB2 + CA04	CA04	PR8	PR8	7.9 × 106	1.07 (0.38)
r58	JNS06 PB1-PB2 + CA04	CA04	JNS06	JNS06	2.0 × 106	1.04 (0.23)
r59	LPM91 PB1-PB2 + CA04	CA04	L91	L91	2.0 × 106	0.62 (0.23)
r66	6L PB1-PB2 + CA04	CA04	6L	6L	1.6 × 106	0.65 (0.13)
r69	PR8 PA-PB2 + CA04	PR8	CA04	PR8	2.0 × 106	0.96 (0.5)
r72	JNS06 PA-PB2 + CA04	JNS06	CA04	JNS06	1.0 × 106	0.6 (0.15)
r79	ma44 PA-PB2 + CA04	ma44	CA04	ma44	5.0 × 106	0.95 (0.21)
r86	JNS06 PA-PB1-PB2 + CA04	JNS06	JNS06	JNS06	3.2 × 106	1.05 (0.27)
r87	LPM91 PA-PB1-PB2 + CA04	L91	L91	L91	1.3 × 106	0.64 (0.18)
r91	ma81K PA-PB1-PB2 + CA04	ma81	ma81	ma81K	5.0 × 106	1.08 (0.28)
Low replicative
r41	04164 PB2 + CA04	CA04	CA04	04164	3.2 × 104	0.5 (0.18)
r42	PR8 PA-PB1 + CA04	PR8	PR8	CA04	2.0 × 104	0.52 (0.14)
r45	JNS06 PA-PB1 + CA04	JNS06	JNS06	CA04	1.6 × 105	0.65 (0.1)
r64	W44 PB1-PB2 + CA04	CA04	W44	W44	3.2 × 105	0.68 (0.25)
r65	ma44 PB1-PB2 + CA04	CA04	ma44	ma44	2.0 × 105	0.84 (0.34)
r73	LPM91 PA-PB2 + CA04	L91	CA04	L91	1.3 × 104	0.43 (0.1)
r76	ma81 PA-PB2 + CA04	ma81	CA04	ma81	3.2 × 104	0.46 (0.2)
r90	ma81 PA-PB1-PB2 + CA04	ma81	ma81	ma81	1.3 × 105	0.45 (0.09)
r94	6L PA-PB1-PB2 + CA04	6L	6L	6L	2.0 × 104	0.62 (0.16)

^aThe “r” in the virus number denotes reverse genetics.

^bFinal titer after growth in MDCK cells for 48 h.

^cIn MDCK cells 60 h after inoculation.

^dVirus abbreviations: A/California/04/09, CA04; A/Puerto Rico/8/34, PR8; A/Sw/Korea/CAN01/04, CAN01; A/Cheongju/H407/08, H407; A/Sw/Korea/JNS06/04, JNS06; A/Dk/Korea/LPM91/06, LPM91; A/Em/Korea/W150/06, W150; A/Ab/Korea/W81/05, W81; A/Ab/Korea/ma81/07, ma81; A/Ab/Korea/ma81K/07, ma81K; A/Ab/Korea/W44/05, W44; A/Ab/Korea/ma44/07, ma44; A/Md/Korea/6L/07, 6L; A/Ck/Korea/04116/04, 04116; A/Ck/Korea/04164/04, 04164. Boldface entries denote the viral origin of the substituted gene segments for each reassortant.

Site-directed mutagenesis.

The characteristic mutations in PB2 (i.e., E627K and D701N) were generated by site-directed mutagenesis using the GeneTailor site-directed mutagenesis system (Invitrogen) according to the manufacturer's instructions. The stop codons at positions 12, 58, and 88 that are required for PB1-F2 protein expression were mutated to serine, tryptophan, and tryptophan, respectively. A marker in PA (T97I) previously reported to enhance virulence in mice (56) also was incorporated. All of these mutations (individually or simultaneously expressed) were generated in the CA04 (H1N1) virus background. To ensure that no unwanted mutations were introduced, the entire genomes of the rescued viruses were resequenced.

In vivo pathogenicity experiments.

The pathogenicity of the 58 viable (i.e., rescued) polymerase reassortant viruses were tested in 5-week-old BALB/c mice (Samtako, Republic of Korea). For each virus, groups of 10 mice were lightly anesthetized and intranasally (i.n.) inoculated with 30 μl of the virus at its full titer achieved after propagation in eggs. Mice were observed for 13 days postinfection. For viruses that killed more than 50% of the mice, the 50% lethal doses in mice (MLD₅₀s) were determined in additional groups of 10 mice with the cutoff being the loss of 20% of body weight, at which point animals were euthanized. The research protocol for the use of mice in this study were conducted in strict accordance and adherence to relevant policies regarding animal handling as mandated under the Guidelines for Animal Use and Care of the Korea Center for Disease Control (K-CDC) and was approved by the Medical Research Institute (approval number CBNU-IRB-2010-GM01) and Laboratory Animal Research Center (LARC) (approval number CBNUA-074-0904-01), a member of the IACUC of Chungbuk National University. Animal care and use in an enhanced biosafety level 3 containment laboratory was approved by the Animal Experiment Committee of Bioleaders Corp. (permit number BLS-ABSL-10-003).

Luciferase minigenome reporter assays.

A luciferase reporter plasmid (pHW72-Luc) that is driven by the polI transcription unit containing the noncoding region of the influenza A virus M gene was constructed as described previously (56). Luciferase activity was measured according to the protocol of Salomon et al. (52) to compare the activity of the recombinant viral polymerase complexes. 293T cells were prepared in 24-well plates 24 h before use and were transfected with 0.1 μg each of pHW72-Luc, pHW2000-PB2, pHW2000-PB1, pHW2000-PA, pHW2000-NP, and pCMV-β-galactosidase plasmids by using TransIT-LT1 transfection reagent as directed. After 4 h, the transfection medium was replaced with fresh DMEM (Gibco) containing 5% FBS. After 24 h, cells were washed with PBS and lysed for 30 min with 100 μl of lysis buffer (Promega). Cell lysates then were harvested, and luciferase activity was assayed in triplicate by using the Promega luciferase assay system. The results were normalized to the β-galactosidase activity level of the cells.

Phylogenetic alignment.

All of the viral genes of the plasmid used in this study were sequenced at Cosmo Genetech (Seoul, Republic of Korea) by using an ABI 373 XL DNA sequencer (Applied Biosystems, Foster City, CA). Sequences were edited by using the Lasergene sequence analysis software package (DNASTAR 5.0; Madison, WI). The Clustal X program (1) was used to align the sequences and build the phylogenetic trees by use of the neighbor-joining method. Full-length polymerase gene sequences of the viruses in this study and sequences available from GenBank were included in the phylogenetic trees, and they were viewed by using NJ Plot (49). The number of bootstrap replications was set to 100, and major tree branches were labeled with bootstrap values for reference. Branch lengths are proportional to sequence divergence and can be measured relative to the scale bar included in the figures. The GenBank accession numbers for the sequences of viral polymerase genes used in this study are provided in Table 1.

Statistical analyses.

The polymerase activities of the different combinations of viral polymerase complexes in cell culture were compared by using two-tailed paired t tests (Prism 5.0; GraphPad). TCID₅₀ and MLD₅₀ values were calculated by using the method of Reed and Muench (50).

RESULTS

Polymerase activity and pathogenicity of CA04 polymerase mutants.

Due to the low fidelity of the influenza viral RNA polymerase, the expression of molecular markers associated with high pathogenicity can be acquired through the accumulation of host-specific mutations during sustained transmission (43). To investigate the effect of known virulence markers on the prototype pandemic (H1N1) 2009 influenza virus, we first inserted the full-length cDNA of CA04 (H1N1) into the bidirectional expression plasmid pHW2000 (24) and generated mutant viruses by reverse genetics. Virulence markers from PB2 (627K and 701N), PB1 (expression of the PB1-F2 protein), and PA (97I) were incorporated, individually or simultaneously, by using site-directed mutagenesis (Table 2). Mutants expressing either of the PB2 virulence markers (627K or 701N) or the PB1-F2 protein had higher levels of polymerase activity (170, 140, and 80% higher, respectively) than the WT CA04 (H1N1) polymerase complex did (P < 0.05) (Fig. 1A, columns A to D). In contrast, polymerase complexes with the PA virulence marker had approximately 40% less activity than the WT polymerase (Fig. 1A, column E). The stepwise expression of the PB1-F2 protein with either of the PB2 virulence markers increased activity levels only slightly beyond those of the single PB2 mutants (Fig. 1A, columns G and H versus columns B and C), although these double mutants had substantially higher activity levels than the single PB1-F2 mutants did (Fig. 1A, column D). However, the activity levels of PB2–PB1-F2 mutants did not exceed those of the PB2 double mutant (CA04 PB2_627K-701N; column F) or the triple mutant virus (CA04 PB1-F2 PB2_627K-701N; column I), which had the highest polymerase activity of all of the viral polymerase complexes tested.

Fig. 1.

In vitro characterization of reverse genetic-generated polymerase mutant pandemic (H1N1) 2009 influenza viruses expressing different virulence markers. (A) Polymerase activity [relative to that of WT CA04 (H1N1) virus] of the reconstituted viral polymerase complexes (PB2, PB1, and PA genes) of the indicated polymerase mutants. Polymerase activity was measured as luciferase activity in a minigenome reporter gene assay; values are the means (standard deviations) of at least three assays. *, P < 0.05; **, P < 0.005 (two-tailed paired t test). (B) Growth kinetics of polymerase mutant viruses in MDCK cells inoculated at an MOI of 10⁻⁴. Viral titers were determined at the indicated time points.

We next generated recombinant WT and polymerase complex mutant viruses and observed their growth characteristics in MDCK cells; all of the viruses had replication kinetics similar to that of the WT virus (Fig. 1B). We then inoculated groups of BALB/c mice to ascertain the pathogenicity of the viruses in vivo. Intranasal inoculation with 10^5.5 TCID₅₀ of either of the mutant viruses or the WT virus yielded similar MLD₅₀ values (Table 2), indicating that replacing the molecular determinants (individually or simultaneously) in PB2 and PB1 could not significantly enhance the virulence of the 2009 pandemic influenza virus.

Activity of reassortant viral polymerase complexes in vitro.

Aside from undergoing adaptive mutations, influenza viruses also can evolve through the reassortment of the segmented viral genome. Therefore, we determined the effect of gene reassortment on the CA04 (H1N1) polymerase's activity by replacing each of the polymerase segments with corresponding subunits from human, swine, or avian isolates, including some mouse-adapted variants (Table 1). The polymerase genes of the mouse-adapted avian H5 and H7 viruses used in this study, A/Ab/Korea/ma81/07 (H5N2) [Av/ma81 (H5N2)] and A/Ab/Korea/ma44/07 (H7N3) [Av/ma44 (H7N3)], bear characteristic mutations in reference to their WT virus counterparts as a result of the serial lung-to-lung passage in mice (56). The well-characterized laboratory strain of human influenza virus A/Puerto Rico/8/34 (H1N1) [PR8 (H1N1)] also was included. The polymerase activity of the reassorted polymerase complexes was compared to that of WT CA04 (H1N1) viral polymerase complex (Fig. 2). Due to their influence on virulence, host range, and adaptation (11, 42), only viral PA, PB1, and PB2 were altered in these experiments; the CA04 (H1N1) nucleoprotein (NP) remained the same in all reassortants. The “r” designation on the polymerase reassortant complexes denotes reverse genetics, whereas the number corresponds to similar reassortant viruses created as shown in Table 3. Virus names reflect the viral source, including the subtype and the segment replaced with the remaining segments from CA04 (H1N1) virus.

Fig. 2.

Polymerase activity of reassortant viral polymerase complexes. The replication and transcription activity of reconstituted viral polymerase complexes in which the PB2, PB1, and/or PA genes of CA04 (H1N1) were replaced with cognate polymerase subunits from human, avian, and swine influenza viruses. (A) Single-gene reassortants; (B) double-gene reassortants; (C) triple-gene reassortants. Luciferase-based minigenome reporter assays were used to measure polymerase activity in the background of a WT CA04 (H1N1) NP plasmid. Activity values shown are the means from at least three assays. The first row at the bottom lists the recombinants by number (“r” denotes reverse genetics). Shaded reassortants were efficiently rescued. Asterisks indicate P values (*, P < 0.05; **, P < 0.005) that represent a statistically significant difference from the polymerase activity of the reconstituted viral polymerase of the homogeneous WT CA04 (H1N1) complex (100%), the baseline for comparison (as determined by two-tailed paired t tests). Error bars indicate standard deviations of the means. Virus abbreviations: A/California/04/09, CA04; A/Puerto Rico/8/34, PR8; A/Sw/Korea/CAN01/04, CAN01; A/Cheongju/H407/08, H407; A/Sw/Korea/JNS06/04, JNS06; A/Dk/Korea/LPM91/06, L91; A/Em/Korea/W150/06, W150; A/Ab/Korea/W81/05, W81; A/Ab/Korea/ma81/07, ma81; A/Ab/Korea/ma81K/07, ma81K; A/Ab/Korea/W44/05, W44; A/Ab/Korea/ma44/07, ma44; A/Md/Korea/6L/07, 6L; A/Ck/Korea/04116/04, 04116; A/Ck/Korea/04164/04, 04164.

With the exception of the r9 recombinant containing Av/ma81 (H5N2) PA, substitutions in the PA gene impaired the activity of all polymerase complexes relative to that of WT CA04 (H1N1) (r1), indicating incompatibility between the CA04 (H1N1) PB2-PB1 NP and the PA segments selected for single-gene reassortment (Fig. 2A, columns r1 to r14). Surprisingly, most of the mammalian- or avian-derived PB1 genes dramatically increased the polymerase activity of WT CA04 (H1N1); the exceptions were the A/Sw/Korea/CAN01/04 [Sw/CAN01 (H1N1)] and A/Dk/Korea/LPM91/06 [Av/LPM91 (H3N2)] isolates (column r15 to r27). Replacing the WT CA04 (H1N1) PB2 gene with PR8 (H1N1) PB2 (r28), Sw/CAN01 (H1N1) PB2 (r29), A/Cheongju/H407/08 [H407 (H3N2)] PB2 (r30), A/Sw/Korea/JNS06/04 [Sw/JNS06 (H3N2)] PB2 (r31), or mouse-adapted Av/ma81K (H5N2) PB2 (r36) or Av/ma44 (H7N3) PB2 (r38) segments elevated polymerase activity, whereas the other PB2 substitutions reduced it. In two-gene reassortants, the substitution of PB1-PB2 genes enhanced viral polymerase complex activity more than the substitution of PA-PB1 or PA-PB2 (Fig. 2B). This effect was most appreciable with the genes from the seasonal human H407 (H3N2) PB1-PB2 (r57) and HPAI A/EM/Korea/W150/06 [Av/W150 (H5N1)] PB1-PB2 (r60) viruses, which robustly augmented the polymerase activity of the pandemic virus. However, when the PA, PB1, and PB2 genes were replaced (i.e., three-gene reassortment), only the mammalian PR8 (H1N1) PA-PB1-PB2 (r83), H407 (H3N2) PA-PB1-PB2 (r85), and Sw/JNS06 (H3N2) PA-PB1-PB2 (r86) strains and mouse-adapted avian Av/ma81 (H5N2) PA-PB1-PB2 (r90), Av/ma81K (H5N2) PA-PB1-PB2 (r91), and Av/ma44 (H7N3) PA-PB1-PB2 (r93) strains, with the exception of Sw/CAN01 (H1N1) PA-PB1-PB2 (r84), enhanced polymerase activity (Fig. 2C).

Rescue and in vitro phenotyping of reassortant viruses.

As our mutant virus studies had shown an imperfect correlation between in vitro polymerase assays and viral virulence, we next attempted to create reassortant viruses with the combinations of polymerase genes discussed above, with the remaining segments in the CA04 (H1N1) virus background. In general, the efficiency of virus rescue moderately correlated with the polymerase activity assays. In particular, 28 of 36 (77.8%) polymerase combinations with robust in vitro polymerase activities were successfully rescued, whereas only 23 of 50 (46%) reassortants that exhibited low or impaired replication and transcription activities were rescued (Table 3 and Fig. 2). Interestingly, with respect to the human seasonal H407 (H3N2) and HPAI Av/W150 (H5N1) viruses, only their PB2 single-gene reassortants could be rescued.

However, generally single-gene PB1 or PB2 reassortants were more efficiently rescued than single-gene PA reassortants (Fig. 2A and Table 3). For example, only 7 PA reassor-tants were viable, whereas 9 PB1 and 12 PB2 reassortants were rescued (Table 3). PB1-PB2 double reassortants also were rescued at a high rate, which is consistent with their high polymerase activity in vitro. Taken together, these results suggest that PA is a limiting factor for reassortment in the CA04 (H1N1) virus background. Completely replacing the three polymerase genes with those of Sw/JNS06 (H3N2) PA-PB1-PB2 (r86), Av/LPM91 (H3N2) PA-PB1-PB2 (r87), Av/ma81 (H5N2) PA-PB1-PB2 (r90), Av/ma81K (H5N2) PA-PB1-PB2 (r91), or A/Md/Korea/6L/07 [Av/6L (H7N6)] PA-PB1-PB2 (r94) only allowed virus rescue (Fig. 2C, Table 3). Of the 40 possible single-gene reassortants, 29 (72.5%) were rescued, whereas only 24 of 41 two-gene (58.5%) and 5 of 14 three-gene (35.7%) reassortants were successfully rescued.

Of 95 possible polymerase reassortments in the CA04 (H1N1) virus background, only 58 (61%) of these could be rescued by reverse genetic techniques, suggesting compatibility limitations in the reassortment process. On the basis of their full titers, the reassortant viruses were phenotypically categorized into four viability groups (Table 3). The high replication efficiency group contains 18 reassortants (31% of rescued viruses) that grew efficiently in MDCK cells at titers similar to or greater than that of the parental Rg CA04 (H1N1) virus (≥1.0 × 10⁷ TCID₅₀/ml) (Table 3), and most formed plaques of greater than 1 mm in diameter. Twelve reassortants in the group had substantially higher polymerase activities than the pandemic virus in the minigenome assays (Fig. 2). The moderate replication efficiency group contains 31 reassortants (53.5% of rescued viruses), had titers ranging from 5.0 × 10⁵ to 1.0 × 10⁷ TCID₅₀/ml (Table 3), and formed plaques with mean sizes of 0.8 to 0.9 mm. More than half of these recombinants (17 viruses) had polymerase activities that were lower than or similar to that of the CA04 (H1N1) virus (Fig. 2).

The low replication efficiency group consists of nine reassortants (i.e., 15.5% of rescued viruses) that grew poorly in MDCK cells, yielding virus titers below 5.0 × 10⁵ TCID₅₀/ml (Table 3). Plaques formed were 0.4 to 0.6 mm in diameter (Table 3). The nonviable group consists of the 37 potential recombinants that were not successfully rescued in the plasmid-driven reverse genetics system despite numerous attempts (see Table S1 in the supplemental material). The polymerase activity of most of these viruses was very low compared to that of the WT CA04 (H1N1) viral polymerase complex, indicating severely impaired replication and transcription (Fig. 2). Ultimately, these categories guided our selection of viruses that were studied further in vivo.

Pathogenicity of reassortant viruses in vivo.

The plasmid-derived Rg CA04 (H1N1) virus was not highly pathogenic to mice (MLD₅₀ = 5.8 log₁₀ TCID₅₀) (Table 2). It replicated in mouse lungs, yielding maximum titers of 5.3 log₁₀TCID₅₀/g at 3 days after i.n. inoculation, and it caused a mean weight loss of approximately 20%. For comparison, in vivo pathogenicity testing of the 58 viruses was performed. On the basis of their pathogenicity in mice compared to that of the Rg CA04 virus, the polymerase reassortant viruses were further segregated into one of three virulence groups: higher virulence, similar virulence, or lower virulence.

Five reassortants [Av/ma81 (H5N2) PA (r9), Sw/JNS06 (H3N2) PB2 (r31), Av/ma81K (H5N2) PB2 (r36), Av/ma44 (H7N3) PB2 (r38), and Av/ma81K (H5N2) PA-PB2 (r77)] had higher virulence than Rg CA04 (H1N1), having MLD₅₀s of less than 5.3 log₁₀TCID₅₀ (Table 4). Interestingly, the possession of PB2 of Sw/JNS06 (H3N2) (r31 virus), which lacks any known virulence factors, was enough to increase the pathogenicity of CA04 (MLD₅₀ = 4.5 log₁₀ TCID₅₀), while the other four reassortants had mammalian-passaged avian PB2, PA, or both, which bear mutations associated with the adaptation (i.e., PB2_627K and PA_97I). All viruses in this group showed polymerase activity significantly higher (P < 0.005) than that of the WT CA04 (H1N1) polymerase complex (Fig. 2 and Table 4).

Table 4.

Pathogenicity in mice of polymerase reassortant viruses in the background of CA04 virus^c

Reassortant no. and pathogenicity group	Reassorted segments	Polymerase segment replaced in:			MLD50a (TCID50)	% Polymerase activityb (SD)
		PA	PB1	PB2
More pathogenic than Rg CA04
r9	ma81 PA + CA04	ma81	CA04	CA04	3.3	432 (39)
r31	JNS06 PB2 + CA04	CA04	CA04	JNS06	4.5	295 (9)
r36	ma81K PB2 + CA04	CA04	CA04	ma81K	4.5	283 (35)
r38	ma44 PB2 + CA04	CA04	CA04	ma44	4.5	260 (25)
r77	ma81K PA-PB2 + CA04	ma81	CA04	ma81K	3.3	321 (22)
As pathogenic as Rg CA04
r1	Rg CA04	CA04	CA04	CA04	5.8	100 (8)
r24	ma44 PB1 + CA04	CA04	ma44	CA04	5.6	687 (58)
r26	04116 PB1 + CA04	CA04	04116	CA04	5.8	245 (12)
r28	PR8 PB2 + CA04	CA04	CA04	PR8	6	730 (19)
r49	ma81 PA-PB1 + CA04	ma81	ma81	CA04	5.4	427 (5)
r62	ma81 PB1-PB2 + CA04	CA04	ma81	ma81	5.5	88 (12)
r63	ma81K PB1-PB2 + CA04	CA04	ma81	ma81K	5.5	183 (14)
Less pathogenic than Rg CA04
r5	JNS06 PA + CA04	JNS06	CA04	CA04	≥5.0	15 (1)
r6	LPM91 PA + CA04	L91	CA04	CA04	≥5.8	31 (1)
r8	W81 PA + CA04	W81	CA04	CA04	≥6.0	32 (1)
r10	W44 PA + CA04	W44	CA04	CA04	≥5.0	22 (10)
r13	04116 PA + CA04	04116	CA04	CA04	≥5.0	84 (8)
r14	04164 PA + CA04	04164	CA04	CA04	≥4.9	40 (8)
r15	PR8 PB1 + CA04	CA04	PR8	CA04	≥5.5	165 (13)
r16	CAN01 PB1 + CA04	CA04	CAN01	CA04	≥5.3	65 (4)
r18	JNS06 PB1 + CA04	CA04	JNS06	CA04	≥5.7	299 (22)
r21	W81 PB1 + CA04	CA04	W81	CA04	≥6.0	274 (7)
r22	ma81 PB1 + CA04	CA04	ma81	CA04	≥6.0	590 (25)
r23	W44 PB1 + CA04	CA04	W44	CA04	≥4.5	632 (27)
r25	6L PB1 + CA04	CA04	6L	CA04	≥5.9	418 (34)
r29	CAN01 PB2 + CA04	CA04	CA04	CAN01	≥5.8	217 (38)
r30	H407 PB2 + CA04	CA04	CA04	H407	≥5.5	746 (62)
r32	LPM91 PB2 + CA04	CA04	CA04	L91	≥5.3	24 (5)
r33	W150 PB2 + CA04	CA04	CA04	W150	≥5.0	105 (3)
r34	W81 PB2 + CA04	CA04	CA04	W81	≥4.9	38 (4)
r35	ma81 PB2 + CA04	CA04	CA04	ma81	≥5.0	46 (6)
r37	W44 PB2 + CA04	CA04	CA04	W44	≥4.5	26 (6)
r39	6L PB2 + CA04	CA04	CA04	6L	≥4.5	26 (3)
r41	04164 PB2 + CA04	CA04	CA04	04164	≥2.8	26 (4)
r42	PR8 PA-PB1 + CA04	PR8	PR8	CA04	≥2.6	95 (4)
r45	TX98 PA-PB1 + CA04	TX98	TX98	CA04	≥3.5	34 (3)
r48	W81 PA-PB1 + CA04	W81	W81	CA04	≥6.5	153 (17)
r50	W44 PA-PB1 + CA04	W44	W44	CA04	≥5.0	73 (12)
r52	6L PA-PB1 + CA04	6L	6L	CA04	≥5.5	220 (12)
r53	04116 PA-PB1 + CA04	04116	04116	CA04	≥4.9	26 (2)
r55	PR8 PB1-PB2 + CA04	CA04	PR8	PR8	≥5.7	188 (17)
r56	CAN01 PB1-PB2 + CA04	CA04	CAN01	CAN01	6.7	62 (3)
r58	JNS06 PB1-PB2 + CA04	CA04	JNS06	JNS06	≥5.1	208 (4)
r59	LPM91 PB1-PB2 + CA04	CA04	L91	L91	≥5.1	15 (2)
r61	W81 PB1-PB2 + CA04	CA04	W81	W81	≥5.8	112 (4)
r64	W44 PB1-PB2 + CA04	CA04	W44	W44	≥3.8	37 (3)
r65	ma44 PB1-PB2 + CA04	CA04	ma44	ma44	≥3.6	186 (13)
r66	6L PB1-PB2 + CA04	CA04	6L	6L	≥5.0	49 (5)
r69	PR8 PA-PB2 + CA04	PR8	CA04	PR8	≥5.1	263 (15)
r70	CAN01 PA-PB2 + CA04	CAN01	CA04	CAN01	≥6.1	43 (4)
r72	JNS06 PA-PB2 + CA04	JNS06	CA04	JNS06	≥4.8	58 (5)
r73	LPM91 PA-PB2 + CA04	L91	CA04	L91	≥2.4	18 (1)
r76	ma81 PA-PB2 + CA04	ma81	CA04	ma81	≥2.8	111 (3)
r79	ma44 PA-PB2 + CA04	ma44	CA04	ma44	≥5.5	103 (3)
r86	JNS06 PA-PB1-PB2 + CA04	JNS06	JNS06	JNS06	≥5.1	248 (3)
r87	LPM91 PA-PB1-PB2 + CA04	L91	L91	L91	≥4.9	55 (20)
r90	ma81 PA-PB1-PB2 + CA04	ma81	ma81	ma81	≥3.4	485 (15)
r91	ma81K PA-PB1-PB2 + CA04	ma81	ma81	ma81K	≥5.5	686 (9)
r94	6L PA-PB1-PB2 + CA04	6L	6L	6L	≥3.6	94 (15)

^aDetermined by inoculating groups of 10 mice with serial dilutions of their full titers.

^bPercentage of Rg CA04 polymerase activity.

^cThe “r” in the virus number denotes reverse genetics. Virus abbreviations: A/California/04/09, CA04; A/Puerto Rico/8/34, PR8; A/Sw/Korea/CAN01/04, CAN01; A/Cheongju/H407/08, H407; A/Sw/Korea/JNS06/04, JNS06; A/Dk/Korea/LPM91/06, LPM91; A/Em/Korea/W150/06, W150; A/Ab/Korea/W81/05, W81; A/Ab/Korea/ma81/07, ma81; A/Ab/Korea/ma81K/07, ma81K; A/Ab/Korea/W44/05, W44; A/Ab/Korea/ma44/07, ma44; A/Md/Korea/6L/07, 6L; A/Ck/Korea/04116/04, 04116; A/Ck/Korea/04164/04, 04164. Boldface entries denote the viral origin of the substituted gene segments for each reassortant.

Six polymerase reassortants [Av/ma44 (H7N3) PB1 (r24), Av/04116 (H9N2) PB1 (r26), PR8 (H1N1) PB2 (r28), Av/ma81 (H5N2) PA-PB1 (r49), Av/ma81 (H5N2) PB1-PB2 (r62), and Av/ma81K (H5N2) PB1-PB2 (r63)] were as pathogenic as the pandemic virus, with MLD₅₀s between 5.4 and 6.0 log₁₀ TCID₅₀ (Table 4). With the exception of the Av/ma81 (H5N2) PB1-PB2 (r62) recombinant, all reassortants showed significantly higher polymerase activity than WT CA04 (H1N1) (P < 0.05 to P < 0.005), and two reassortants [Av/ma44 (H7N3) PB1 (r24) and PR8 (H1N1) PB2 (r28)] had exceptionally high values.

Most of the polymerase reassortants tested had attenuated pathogenicity in mouse models compared to that of the pandemic CA04 (H1N1) virus (47 low-virulence viruses). MLD₅₀s for this group were greater than 6.0 log₁₀ TCID₅₀ (Table 4). The HPAI Av/W150 (H5N1) PB2 (r33) and human seasonal H407 (H3N2) PB2 (r30) reassortants and all of the low-replicative viruses (Table 3) belonged to this group.

In general, clinical signs of disease (ruffled fur, huddling, hunched back, and weight loss) were most evident among mice inoculated with viruses in the high-virulence group (data not shown). However, the severity of illness was not used as a measure of virulence due to the different inoculation doses used. To clinically compare the pathogenicity of the high-virulence and similar-virulence groups, we inoculated additional groups of 10 mice with 10^4.5 TCID₅₀ of each reassortant or the Rg CA04 (H1N1) virus and monitored their survival for 13 days. Severe influenza-like disease signs were observed among mice inoculated with the high-virulence viruses [Av/ma81 (H5N2) PA (r9), Sw/JNS06 (H3N2) PB2 (r31), Av/ma81K (H5N2) PB2 (r36), Av/ma44 (H7N3) PB2 (r38), Av/ma81K (H5N2) PA-PB2 (r77)], and the mortality rate was 50 to 100% (Fig. 3). In contrast, there were no deaths among mice inoculated with polymerase reassortants that showed virulence similar to that of Rg CA04 (H1N1) (Fig. 3), although signs of substantial illness were observed.

Fig. 3.

Survival of mice inoculated with reassortant viruses that have pathogenicity equal to or greater than that of the CA04 (H1N1) virus. Groups of 10 mice were inoculated with 10^4.5 TCID₅₀ of each virus. Survival was monitored daily for 13 days. An asterisk indicates viruses that were more pathogenic than Rg CA04 (H1N1) in vitro. A dagger indicates viruses that were as pathogenic as Rg CA04 (H1N1) in vitro.

DISCUSSION

In contrast to the three previous influenza pandemics, which caused millions of deaths worldwide, the pandemic (H1N1) 2009 influenza virus caused mild disease with a low mortality rate; these attributes are speculated to be due to the absence of specific molecular markers that are linked to high pathogenicity as well as a degree of preexisting cross-reactive immunity in humans (43). However, its continued circulation among humans and susceptible animal host species raises concern about the possible generation of more-virulent variants. The main mechanisms that facilitate the acquisition of virulence factors among influenza viruses are mutations due to faulty viral RNA replication and genetic reassortment with cocirculating influenza viruses (47). Influenza viruses encode three polymerase subunits (PB2, PB1, and PA) that form the trimeric viral RNA-dependent RNA polymerase responsible for the transcription and replication of the viral RNA genome in the nuclei of infected cells (14). Although the trimeric polymerase complex associates with NP to form the viral ribonucleoprotein complex required for viral replication, the polymerase subunits have been found more frequently to contain important determinants of pathogenicity (13, 15, 32, 52, 58). Notably, each of the previous 20th-century influenza virus pandemic strains have included a novel PB1 in addition to the new HA and/or NA segment (54). Further, the compatibility of the polymerase subunit proteins may be a restricting factor for influenza virus reassortment (30). Therefore, we focused our study on the influenza virus polymerases, which are essential for virus replication in host cells and which accelerate virus evolution through an error-prone RNA polymerase (38).

Expressing the 627K or 701N mutation in PB2 or driving PB1-F2 protein expression enhanced reporter gene expression compared to that of the WT CA04 (H1N1) polymerase complex but did not increase replication kinetics in MDCK cells or its pathogenicity in mice, findings which are in agreement with those of others (19, 23, 68). We also have shown here that even the simultaneous expressions of these residues (627K and 701N) do not significantly enhance replication and pathogenesis in the mouse model.

We also generated CA04 (H1N1) polymerase reassortant viruses possessing cognate polymerase subunits from swine, avian, and human viruses (Fig. 4). The pandemic (H1N1) 2009 influenza virus polymerase gene complex was derived from triple-reassortant (human-avian-swine) swine viruses and therefore is not an authentic porcine influenza virus complex. Specifically, the PB2 and PA genes originated from a North American avian-like virus, and the PB1 gene originated from a human-like H3N2 virus (17). Replacing these viral genes with segments from the genotypically similar triple-reassortant swine viruses Sw/JNS06 (H3N2) and Sw/CAN01 (H1N1) (48) had little overall effect on virus pathogenicity in mice, as expected. However, the PB2 of Sw/JNS06 (H3N2), which lacks any known virulence determinants, was sufficient to significantly elevate the virulence of the CA04 (H1N1) virus (the MLD₅₀ increased from 5.8 to 4.5 log₁₀ TCID₅₀). This finding suggests that Sw/JNS06 (H3N2)-like PB2 genes have the potential to increase the virulence of the pandemic virus, and it reinforces the concept that previously uncharacterized mutations in the polymerase complex enhance virus replication in mammals. Indeed, there is evidence that a number of sites on the PB2 segment are able to substitute for the functions of 627K or 701N (5, 31, 37, 68).

Fig. 4.

Phylogenetic trees of the viral polymerase genes used in this study. Virulence-related properties are indicated. The nucleotide sequences were aligned, and the phylograms were generated by the neighbor-joining method using the tree-drawing program Clustal X (1) and were viewed by using NJ plot (49). The scale bar represents the number of substitutions per nucleotide. Branch labels indicate the stability of the branches for 100 bootstrap replicates. Only those with bootstrap values of ≥60% are shown. Boldface entries represent viruses used in this study. Shaded viruses are pandemic (H1N1) 2009 influenza virus strains. Ab, aquatic bird; AGT, American green-winged teal; Av, avian; Em, environment; Hu, human; Sw, swine. An asterisk indicates viruses that induced polymerase activity at a higher level than that of Rg CA04 (H1N1). A dagger indicates viruses that resulted in replicative ability higher than that of Rg CA04 (H1N1). A double dagger indicates viruses that conferred virulence higher than that of parental Rg CA04 (H1N1) virus.

The avian polymerases used in these experiments represented viruses of the H3, H5, H7, and H9 subtypes and did not appear to affect the virulence of the pandemic (H1N1) 2009 influenza virus. Interestingly, most reassortants with pathogenicity equal to or greater than that of the CA04 (H1N1) virus contained polymerase genes from mammalian-adapted viruses, such as Av/ma44 (H7N3), Av/ma81 (H5N2), or Av/ma81K (H5N2), rather than those from their wild-type counterparts [Av/W44 (H7N3) and Av/W81 (H5N2)]. We observed that some of the mammalian-adapted polymerase genes that encode the PB2_627K and PA_97I residues appear to enhance pathogenesis. Unlike the case in the polymerase mutant CA04 (H1N1) viruses, which did not significantly affect in vitro and in vivo properties, the polymerase reassortant viruses bearing these markers [Av/ma81 (H5N2) PA (r9), Av/ma81K (H5N2) PB2 (r36), Av/ma44 (H7N3) PB2 (r38), and Av/ma81K (H5N2) PA-PB2 (r77)] were more virulent than the parental Rg CA04 (H1N1) virus. Therefore, more-virulent strains possessing these virulence factors may still emerge in the future through the process of reassortment.

This finding also suggests that avian viruses that have adapted to mammalian hosts have an effect on the pathogenicity of the pandemic virus. In various experimental models, influenza virus strains frequently appear to acquire virulence-enhancing mutations in the polymerase genes during the process of adaptation (51). In some regions of Asia, avian-like viruses (particularly H9N2 and HPAI H5N1) have been isolated continuously from pigs (53, 59, 69) and frequently from humans (65). Surprisingly, we found that no combinations of the polymerase genes from the clade 2.2 Korean HPAI Av/W150 (H5N1) or the seasonal human H407 (H3N2) virus significantly enhanced the disease phenotype of the CA04 (H1N1) virus. Although Av/W150 (H5N1) and H407 (H3N2) PB2 genes functioned well in the CA04 (H1N1) virus, they did not confer enhanced virulence. In contrast, Octaviani et al. (45) recently showed a high level of genetic compatibility between CA04 (H1N1) and HPAI H5N1 influenza viruses, indicating that the generation of new or more-virulent pandemic viruses is quite possible. It should be noted that the virus used in that study was a contemporary human H5N1 isolate (A/Vietnam/HN31694/2009), which supports our hypothesis that differences in genetic background (i.e., virus source, passage history, presence of characteristic mutations, and clade) contribute to the genetic compatibility of the polymerase genes for reassortment and pathogenicity.

Our rescue and replication data demonstrated that the polymerase gene segments of the pandemic virus are moderately compatible with those of the different viruses tested. With some exceptions, we found a functional cooperation between the PB2 and PB1 proteins. However, it is worth noting that PA appears to play a crucial role in maintaining the genetic stability of the CA04 (H1N1) virus. The replacement of PB1 or PB2 had less effect on the rescue of viable reassortant viruses than the replacement of PA, which severely impaired viability (Fig. 2 and Table 3). Such a tendency also can be observed in the CA04 (H1N1) mutant viruses, where polymerase activity and rescue were significantly impaired if the CA04 (H1N1) PA gene was mutated (Fig. 1). Nevertheless, rescue efficiency decreased when more polymerase segments were replaced. These findings suggest that the permutations and reassortments either modified the activities of the viral polymerase complex at the genomic RNA or protein level (23, 30) or disrupted the overall molecular interactions among the viral proteins and host cellular factors (41). Similarly, the discordance between polymerase activity and virus rescue in some cases [i.e., Sw/JNS06 (H3N2) PA (r5), Av/LPM91 (H3N2) PB2 (r32), Av/LPM91 (H3N2) PA-PB2 (r73), Av/ma44 (H7N3) PA-PB1-PB2 (r93), etc.] could be attributed to the effect of the gene substitutions on the other CA04 viral genes aside from the polymerase subunits (specific gene constellation).

Genetic reassortment is a significant driving element that accelerates the evolution of influenza A viruses. The pandemic (H1N1) 2009 influenza virus may have acquired its specific configuration of segments through multiple reassortment events, thereby achieving a high level of fitness to spread and a high capacity to reassort (4, 64). Although the polymerase reassortants tested generally did not increase the pathogenicity of the pandemic virus in our experiments, we cannot rule out the possibility that more-virulent viruses will emerge in the future, especially if the other viral genes (HA, NP, NA, M, and NS) have mutated or reassorted. High pathogenicity is a polygenic trait, and the polymerase genes in most cases work in concert with the NS and HA genes to produce virulent viruses (25). Octaviani et al. (44) demonstrated that reassortant viruses containing a human seasonal H1N1 (A/Kawasaki/UT-SAI-k23/2008) virus HA and the CA04 NA and M genes replicated in vitro to higher titers than their WT parental viruses, although their virulence was not tested. It also is worth mentioning that the PB1 genes of the human pandemic viruses in 1957 (Asian flu, H2N2) and 1968 (Hong Kong flu, H3N2), along with other segments, came from avian virus sources (27). The acquisition of the avian PB1 gene was speculated to be a crucial factor in the emergence of reassortants with a competitive advantage over seasonal influenza viruses with regard to replication and virulence (7). It is possible that our virulence assessment was limited by the failure of the stock viruses to replicate to sufficient infectious titers. Although this limitation could have been addressed by the additional passage of the reassortant viruses, we wished to avoid the risk of acquiring unwanted genetic changes. In addition, it appears that some of the rescued viruses, if not most, originally do not have the capacity to reach sufficiently high infectious titers, as shown by the low virus yields and production of small plaques in TCID₅₀ and plaque assays, respectively. More detailed systematic studies are needed to resolve these issues. However, the use of a wide variety of influenza viruses in the present study provides additional insights into the potential characteristics of reassortants between the pandemic (H1N1) 2009 influenza virus and prevailing human or animal influenza viruses. Because most current influenza research is focused largely on the molecular mechanisms of viral pathogenesis, our findings may help drive further studies on understanding the nature of gene interactions that also contribute to virus virulence and biology. Our findings also highlight the pivotal role of the viral polymerases in the evolution of influenza A viruses with respect to reassortment and the production of new, more-virulent pandemic viruses.