The MDCK-II cell line (collection of Institute of Virology, Marburg) and human epithelial colon carcinoma cell line (CACO-2) (European Collection of Cell Cultures; ECACC) were cultivated in DMEM containing 10% (v/v) bovine fetal calf serum (FCS) (Gibco BRL). Cell monolayers were infected with influenza virus mutants and incubated with DMEM without serum at 37°C for different periods. To prepare virus samples, virus-containing culture fluid was clarified at 6000 x g for 20 min and then pelleted through 20 ml of 20% (w/v) sucrose prepared in PBS at 30,000 rpm (SW 50.1 rotor) for 2.5 hrs. Virus pellets were analyzed by protein gel electrophoresis.

Viral gene segments PB1, PA, PB2, HA, NA, NP, M and NS were cloned into the plasmid pHH21 (Wagner et al, 2005). For rescue of recombinant FPV, 1μg of each of these plasmids was transfected into 10⁶ 293T cells along with expression plasmids encoding the subunits of the influenza virus polymerase pcDNA-PB1 (1μg), pcDNA-PB2 (1μg), pcDNA-PA (0.2μg), and pcDNA-NP (1μg) using Lipofectamine 2000 (Invitrogen) (Wagner et al, 2005). Four hours post-transfection MDCK cells (2x10⁵) were added and the obtained mixed 293T-MDCK cell culture was co-cultivated in DMEM containing 0.5% (v/v) FCS. Four to seven days post- transfection the supernatant containing 10³-10⁶ HAU/ml of virus (passage 0) was removed, aliquoted and kept at −80°C. To obtain caspase mutants, site-directed mutagenesis of pHH21-FPV-NP and pHH21-FPV-M plasmids were constructed using the Quick-change mutagenesis kit (Stratagene, The Netherlands) using the following caspase site-specific primers:

M2nn₂₃/fo: 5′cagcagaatgctgttaacgttaacgatggtcattttg

Eleven-day old chickens (Lohmann Brown line PK-13) were infected with different virus variants and housed under isolation conditions in separate rooms to exclude cross-infection; all chickens received a standard grower diet throughout the experiment. Parenteral routes of virus inoculation were found to be suitable for testing of pathogenicity of avian influenza viruses (Brugh and Beard 1986; Scholtissek and Rott 1984). Accordingly, an intramuscular (i.m.) route of infection was chosen to provide a more accurate pathogenicity comparison of different inoculation doses of the generated recombinant viruses. Birds were infected intramuscularly with virus mutants (0.15ml per leg) and survival and disease signs, such as reduction in movement, loss of body weight, ruffled feathers, body temperature, etc., were monitored daily. Titers of anti-virus antibodies in blood samples were determined by hemagglutination inhibition (HAI) test: 30μl of two-fold serially diluted chicken sera were incubated with 30μl of reference A/FPV/Ro/34 virus (8 HA units) at 22°C for 1hr followed by incubation with 50μl of 10% (v/v) chicken red blood cells at 4°C for 45min. The HAI titer was defined as the highest serum dilution inhibiting virus hemagglutination. All painful manipulations with chickens, such as bleeding, i.m. inoculations, etc., were made under light ether narcosis.

Polypeptides were electrophoresed in 12% (w/v) polyacrylamide gels using Tris-glycine-SDS buffer followed by autoradiography as previously described (Zhirnov et al, 1999). For electrophoresis, cellular polypeptides were dissociated in 2% (w/v) SDS, 0.05 M dithiotreitol, 0.02 M Tris-HCl (pH 6.8) for 10min at 85°C.

After SDS-PAGE the polypeptides were transferred from the gel onto Protran-nitrocellulose membranes (pore-size 0.45-μm) (Schleicher and Schuell) by semidry electro-blotting and processed with virus-specific antibodies as previously described (Zhirnov and Klenk, 2003). Then membranes were exposed to horseradish peroxidase (HRP)-conjugated secondary anti-species antibodies (Dako), followed by ECL visualization on Kodak film.

MDCK cells grown in 24-well plates were incubated with 0.4 ml/well of ten fold virus dilutions in DMEM. After 60min incubation at 37°C, the virus inoculum was removed, and cells were covered with 1.0ml of 1% (w/v) agarose (culture quality; Sigma) in DMEM. 48 hrs after infection, cells were fixed with 4% (v/v) paraformaldehyde and stained with anti-influenza virus antibody followed by visualization of virus foci with TMB insoluble substrate “True Blue” (KPL) (Zhirnov and Klenk, 2003).

As a first step we inspected proteins NP and M2 of highly-pathogenic avian influenza virus A/FPV/Ro/34 (H7N1) to identify caspase cleavage motifs. For this purpose a bioinformatics tool GrabCas predicting potential caspase cleavage sequences was applied (Backes et al, 2005). The NP was found to possess two caspase cleavage motifs: the N-terminal ETG₁₆/G and EEYD₅₇₆/N at the C-terminus, as well the M2 protein carried the motifs DSSD₂₃/G in the N-terminal exomembrane domain and VDVDD₈₇/G in the C-terminal cytoplasmic tail. Additional inspection of other viral proteins revealed numerous caspase cleavage motifs in the proteins PB1, PA, PB2, NS1, NA and HA (summarized in Table 1). The proteins NEP, M1, and the hypothetical NSP (Zhirnov et al, 2007) appeared to lack any caspase cleavage motifs. Wide distribution of caspase motifs in different influenza virus proteins implied that these sites could play an important role in functioning of the viral proteins and virus biology.

These observations prompted us to study the roles of the NP and M2 caspase motifs in virus replication and pathogenicity. These sites were altered by site-directed mutagenesis and virus mutants were generated using a reverse genetics approach. Two types of caspase motif alterations were generated. The fist type was generated to alter the avian-like non-cleavable motif ETG₁₆/G within the NP protein to a human-like cleavable ETD₁₆/G (NPdg) motif. The second type was developed to alter cleavable caspase motifs EEYD₅₇₆/N (NP), DSSD₂₃/G (M2), and VDVDD₈₇/G (M2) into non-cleavable ones. For such alterations the residues E and D at positions -1, -3, and/or -4 (underlined), which are well known to be the most important for the recognition and cleavage by specific caspases (Thornberry et al, 1997), were either substituted for an N residue or deleted in the different mutants (summarized in Table 2). These complex mutations were introduced into caspase motifs to impede the appearance of wild type back revertants. For example, the mutants M2nn₈₇ and M2nn₈₇del possessed a similar attenuated phenotype, however, the first one was less stable and reverted to the WT genotype after 1-2 passages in chicken (data not shown). To impede such reversion, the deletion mutant M2nn₈₇del was generated and studied here in detail.

Next, replicative potential of the generated virus caspase mutants in cultured cells was compared with wild type virus generated by reverse genetics (WTRG) using a focus formation test in MDCK cells. It was found that mutations in the caspase motifs, as a rule, were not lethal and most of mutants the efficiently replicated in MDCK cells (Table 2) and produced normal large focuses in canine MDCK (Figure 1) and human CACO-2 cells (data not shown). This observation indicated that mutations in caspase motifs were not restrictive for virus replication in cells of different origin. Only mutations in the caspase motif located at the C-terminus of the NP appeared to be lethal and restricted virus replication. Interestingly, the mutants NPgd and M2nn₈₇del formed large-sized focuses (500 cells and more per focus), similar to the wild type WTRG virus, whereas M2nn₂₃ produced predominantly small-sized ones. Examination under the light microscope, showed a lot of single virus-positive cells and small foci (15-50 cells per one focus) in cell cultures infected with the last mutant, implying a low ability of this mutant to assembly virions and/or infect neighbouring cells (not shown). Next, virus titers produced in MDCK cells under one-cycle growth were evaluated (Figure 2A). Consistent with the focus assay, mutants NPgd and M2nn₈₇del, like WTRG virus, produced high titers of progeny virus while M2nn₂₃ mutant synthesized markedly lower yields. Collectively, the above data indicated that two mutants (NPgd and M2nn₈₇del) possessed high replicative ability, similar to WTRG virus, but that another one (M2nn₂₃) displayed low replication in cultured cells.

The major viral proteins NP, M1, NS1, and M2 were analyzed as markers of virus synthesis in infected cells. Levels of these proteins synthesized by WTRG and mutant viruses in infected cells and in virions were compared. All proteins were detected in infected cells and their amounts were similar to those in WTRG-infected cells (Figure 2B) and WTRG virions (Figure 2C). The polypeptide profile of virus synthesized in NPgd-infected cells was also similar to WTR virus (not shown). The mutated viral protein NPgd was found to be cleaved in infected cells and its apoptotic form aNP was clearly detected (Figure 2B; lane NPgd) supporting a critical role of D16 in this cleavage.

Conversely, the mutated protein M2nn₈₇del was not cleaved in infected cells as late as 30hpi, when WTRG M2 protein was clearly cleaved to form aM2 (Figure 2D). Interestingly, aNP and aM2 appeared in infected cells sequentially at 15 and 30 h.p.i., respectively, suggesting that NP was less resistant to caspase cleavage in infected cells than M2. Thus, sufficient levels of mutant NP and M2 proteins in infected cells and virions indicated that mutations in caspase sites did not affect significantly intracellular synthesis and maturation of the mutated proteins. It was noted that synthesis of M1 was markedly increased in cells infected with M2nn₈₇del mutant (Figure 2B, bottom). In this mutant M2-specific caspase mutations did not overlap the M1 open reading frame but modified a 3′ non-translated region of the M1 mRNA that could elevate its intracellular stability and translation efficacy.

Influenza FPV virus is well known to replicate in all organs of infected chickens and develop a pantropic infection disseminating throughout the body by viremia (Feldman et al, 2000). Thus, viremia was chosen as an indicator of the infectious process and virus spreading in the chicken. Accordingly, titers of mutant viruses in blood were estimated by focus assay on days 3 and 5 after intramuscular infection of chickens. Virus was regularly detected on days 3 and 5 in the blood of WTRG-infected birds in high concentrations, 10⁵-10⁶ pfu/ml of blood. Markedly lower virus titers, such as 10²-10⁴ pfu/ml, appeared in blood of NPgd-infected chickens on day 3 pi In chickens infected with M2nn₈₇del and M2nn₂₃ mutants viremia was not detected by the focus assay used, indicating that virus titers in these birds were less than the threshold level of 10² pfu/ml of blood. On day 5 pi viremia was not revealed in all birds infected with virus mutants. Thus, in chickens infected with caspase mutants, virus titers in blood were markedly lower than in WTRG-infected ones. These data suggest that wild type WTRG virus effectively replicated and disseminated in chickens, whereas caspase mutant viruses were markedly attenuated in the chicken.

Pathogenic properties of the obtained FPV caspase mutants were assayed. Groups of 11-day-old chickens were infected either with WTRG, NPgd, M2nn₂₃ or M2nn₈₇del virus mutants. It was found that the WTRG variant displayed the highest (100%) virulence for chickens (Figure 3A). Markedly lower lethality (~40%) was observed in the NPgd-infected group. M2nn₂₃ and M2nn₈₇del did not kill chickens either at a low dose of 300 pfu (Figure 3A) or even at the massive infection dose of 10⁵ pfu (data not shown).

To evaluate this further the pathogenic potential of these caspase mutants, the dynamics of disease symptoms caused by the different virus mutants was studied. Chickens were infected with a low dose (~100 pfu/animal) of each virus mutant and signs of illness were monitored. The severity of disease symptoms was found to correlate with the lethal ability of the virus mutant. Chickens infected with WTRG virus displayed high fever (42-43°C), rapid loss of body weight and death within 4-5 days post-infection. Both M2nn₂₃ and M2nn₈₇del mutant viruses were almost apathogenic and did not cause visible symptoms or loss of body weight (Figure 3B). The NPgd mutant induced disease of an intermediate severity. During 4-7 days post-infection, chickens infected with this mutant displayed signs of disease, such as fever (41-42°C), ruffled feathers, lowered activity, delay of body weight gain, and then infected birds rapidly recovered (Figure 3B). These data supported the concept that caspase mutant viruses succumbed to host defense mechanisms and displayed attenuated virulence in chickens.

In order to investigate the antibody responses induced by caspase mutant viruses, 11-day-old chickens were given a primary boost with 50-100 pfu of either the NPgd, M2nn₂₃ or M2nn₈₇del mutant virus. On day 20 after the booster infection chickens were challenged with a high, lethal dose (10³ LD₅₀ /chicken) of field FPV virus. Sera from these birds were collected on day 10 after challenge infection and analyzed in the HAI test. First, chickens vaccinated with NPgd or M2nn₈₇del caspase mutants did not display any signs of disease after challenge infection; in the M2nn₂₃-boosted group chickens had minor transient symptoms of disease during days 3-6 after challenge and fully survived after challenge; placebo non-boosted birds rapidly died during 3-5 days after FPV challenge. Second, chickens initially infected with caspase mutant viruses displayed high HAI antibody titers that reached values of 2⁹-2¹² on day 10 after challenge infection. High antibody titers were retained in chickens for at least 30 days (the period of observation). There was a clear correlation between the booster virus dosage and the antibody titer developed in birds; the higher the booster dose of virus inoculated, the higher the post-challenge antibody titers appearing the in blood of birds (Table 3). Thus, these data show that non-lethal booster infection with caspase mutant viruses induced high antibody titers effectively protecting chickens against a lethal dose of field virus.