Efficient Generation of Recombinant Influenza A Viruses Employing a New Approach to Overcome the Genetic Instability of HA Segments

Influenza A viruses (IAVs) are the most relevant and continual source of severe infectious respiratory complications in humans and different animal species, especially poultry. Therefore, an efficient vaccination that elicits protective and neutralizing antibodies against the viral hemagglutinin (HA) and neuraminidase (NA) is an important strategy to counter annual epidemics or occasional pandemics. With the help of plasmid-based reverse genetics technology, it is possible that IAV vaccine strains (IVVS) are rapidly generated. However, the genetic instability of some cloned HA-cDNAs after transformation into competent bacteria represents a major obstacle. Herein, we report efficient cloning strategies of different genetically volatile HA segments (H5- and H9-subtypes) employing either a newly constructed vector for reverse genetics (pMKPccdB) or by the use of the Escherichia coli strain HB101. Both approaches represent improved and generalizable strategies to establish functional reverse genetics systems preventing genetic changes to the cloned (HA) segments of IAV facilitating more efficient rescue of recombinant IAV for basic research and vaccine development.


Introduction
Influenza viruses, belonging to the family Orthomyxoviridae, are divided according to the antigenic difference between their nucleoprotein (NP) and matrix 1 (M1) proteins into three genera; A, B or C [1]. Out of these three genera, influenza A viruses (IAV) are the most likely to cause serious infections in different species including humans [2]. IAV is composed of eight negative sense genomic RNA segments within a lipid-bilayer envelope. According to the showed comparable replication efficiency. We thereby demonstrate that these two cloning strategies provide a powerful tool to prevent genetic instability of cloned viral segments and simplify the generation of complete reverse genetics systems and vaccine strains for influenza A viruses.

Cells, viruses and viral cDNA
Madin-Darby canine-and human embryonic kidney cells, MDCK-II and 293T, were maintained in Dulbecco's Modified Eagle Medium (DMEM) (Gibco, Invitrogen, USA) supplemented with 100 I.U./ml penicillin, 100 μg/ml streptomycin and 10% fetal bovine serum (FBS). All cells were incubated at 37°C in the presence of 5% CO 2 . Virus stock of the influenza virus A/chicken/Saudi Arabia/CP7/1998 (H9N2/SA) was obtained by inoculation of 100 μl (100 FFU) through allantoic route into 10-day-old embryonated chicken eggs. After 48 h incubation at 37°C in an egg incubator (Heka, Germany), eggs were chilled at 4°C overnight. The allantoic fluid was then harvested and centrifuged at 2500 rpm for 5 min to remove cell debris. The cDNA for A/chicken/Egypt/VSVRI/2009 (H5N1/VSVRI) was obtained from the Egyptian Veterinary Serum and Vaccine Research Institute (VSVRI).

Full-length RT-PCR amplification of HA and NA segments
The viral RNA of the H9N2/SA virus was extracted using "RNeasy Protect Kit" (Invitrogen, USA) according to the manufacturer's instructions. The HA and NA segments of H9N2/SA were then amplified in separate reactions using "SuperScript III One-Step RT-PCR Platinum Taq High Fidelity" (Invitrogen, USA). Briefly, 5μl of extracted vRNA was mixed with 25μl "2× Reaction Mix", 2μl (0.4 μM each) of the Blunt-Ba-HA-F and Blunt-Ba-HA-R primers for the HA segment (Table 1) or universal primers [38] with BsmBI restriction sites for the NA segment (Table 1) of H9N2/SA and 1 μl of enzyme mix. The total volume was adjusted to 50 μl using nuclease-free water (Ambion, USA) and subjected to cDNA synthesis at 55°C for 30 min, followed by pre-denaturation (94°C for 2 min), PCR amplification (35 cycles: 94°C/15 sec for denaturation, 58°C/30 sec for annealing and 68°C/3 min for extension) and final extension (1 cycle: 68°C/7 min). For H5N1/VSVRI, the cDNA generated using the Uni-12 primer [38] was used as a template. The mutibasic cleavage site (PQGEGRRKKR/GLF) in the H5/VSVRI was converted to a monobasic cleavage site (PQGETR/GLF) by amplifying the HA1 and HA2 separately using specific primers (Table 1). However, the NA segment of H5N1/VSVRI was amplified also using the universal primers [38] with BsmBI restriction sites (Table 1). Briefly, 2 μl cDNA were mixed with 5 μl 10× PCR Buffer, 1 μl 10mM dNTP mixture (0.2 mM each), 1.5 μl of 50 mM MgCl 2 (1.5 mM), 2 μl of each Primer (40 pmoles) and 0.2 μl of Platinum Taq DNA Polymerase (10 U/μl) (Invitrogen, USA). The total volume was adjusted to 50 μl using nuclease-free water (Ambion, USA), followed by pre-denaturation at 94°C for 2 min, PCR amplification (35 cycles: 94°C/30 sec for denaturation, 58°C/30 sec for annealing and 68°C/3 min for extension), then final extension at 68°C/7 min. The PCR products were purified using "GeneJET Gel Extraction Kit" (Thermo Scientific, USA). The obtained RT-PCR fragments were next digested with BsaI and BsmBI for according to manufacturer protocol to be used further cloning.

Generation of chimeric HA plasmids
In an attempt to circumvent the observed instability of the H9N2/SA-HA sequence, which we suspected to be located in the HA2-part, plasmids encoding chimeric HAs were generated by exchanging the HA1 sequence of the pHW2000 plasmid (pHW-HA-Bei) stably encoding the A/Chicken/Beijing/2/1997 (H9N2/Bei) HA with the corresponding sequence from the pSMART-LC-Kan plasmid (pSMART-HA-SA) possessing the HA-cDNA of A/chicken/Saudi Arabia/CP7/1998 (H9N2/SA) and vice versa. For this the HA1 sequences of pHW-HA-Bei and pSMART-HA-SA were excised using "FastDigest" AlwNI and EcoRI (Thermo Scientific, USA) for 15 min at 37°C and the HA1 sequence of the H9N2/Bei was cloned into pSMART-HA-SA, while the HA1 sequence of H9N2/SA was cloned into pHW-HA-Bei. To assure that plasmids to express the chimeric HAs have the same vector backbone the chimeric HA-cDNA in pSMART-HA-Bei/SA containing the HA1 of the H9N2/Bei-HA and HA2 of the H9N2/SA-HA was subcloned into pHW2000 after excision using BsaI-HF for 15 min at 37°C and ligation into BsmBI pre-cut pHW2000 using T4 ligase (Thermo Scientific, USA) at 22°C for 1 h. The resulting two pHW2000 plasmids (pHW-HA-Bei/SA, pHW-HA-SA/Bei) with the chimeric HAs (S1 Fig.) were transformed into competent XL-1 blue E. coli and analyzed for correct positive transformants.

Construction of pMKPccdB vector
To generate pMKPccdB vector with BsaI, BsmBI or AarI cloning/restriction sites, Pol-I/ccdB/Pol-II cassettes containing the three different cloning/restriction sites were transferred from the respective pMPccdB vector versions [39] using "FastDigest" BsrBI enzyme (Thermo Scientific, USA) at 37°C for 15 min. The gel-purified cassettes were ligated into the EcoRV-blunt ends of pSMART-LC-Kan vector using T4 DNA ligase (Thermo Scientific, USA) according to the manufacture instructions. The ligation reactions were then transformed into "One Shot ccdB Survival 2 T1R" competent E. coli cells (Invitrogen, USA) according to the manufacture instructions and cultured on LB agar plates. 16-18 h later, five individual colonies were picked up and cultured for 18 h at 37°C in 5 ml of LB media containing 30 μg/ml of Kanamycin (SERVA, Germany) followed by plasmid DNA extraction using mi-Plasmid Miniprep Kit (Metabion, Germany). The bacterial culture containing the correct plasmid DNA construct, confirmed by enzymatic digestion and sequencing, was subjected to large scale culture (250 ml LB media) and plasmid DNA extraction/purification using NucleoBond Xtra Maxi (Macherry-Nagel, Germany).

Site directed mutagenesis
To silently mutate all BsaIand BsmBI-recognition sites in the backbone of pMKPccdB, "Quickchange II XL Site-Directed Mutagenesis Kit" (Agilent Technologies, USA) was used as previously described [40]. The mutated forward and reverse primers ( Table 2) were designed using PrimerX software package (http://www.bioinformatics.org/primerx/). Briefly, 1 μl of template plasmid DNA (2 ng) was mixed with 0.5 μl of each forward and reverse primer (2.5 pmoles/μl), 0.25 μl of 40 mM dNTP mix (10 mM each), 1.25μl "10× Pfu Ultra Buffer" and 0.25 μl of "Pfu Ultra Hotstart polymerase enzyme". The total volume was finally adjusted to 12.5 μl with nuclease-free water (Ambion, USA). The reaction was kept at 95°C for 5 min followed by 18 cycles of 50 sec at 95°C for denaturation, 50 sec at 60°C for primer annealing and 5 min at 68°C for extension. The final extension was held at 68°C for 7 min. After the PCR program was finished, 0.25 μl of DpnI (20 U/μl) was added directly to the reaction and incubated at 37°C for 60 min to digest the methylated and hemi-methylated parental strands. An aliquot of 2 μl of the final reaction was transformed into 20 μl "One Shot ccdB Survival 2 T1R" competent E. coli cells". The correct clones were confirmed by restriction analysis and sequencing of the extracted plasmid DNA.
Cloning of HA-and NA-cDNAs from H9N2/SA and H5N1/VSVRI An A-overhang was added to the purified HA-and NA-RT-PCR product of H9N2/SA and H5N1/VSVRI, which were then cloned into the T-overhangs of the classical cloning/sequencing vector pCR2.1 (Invitrogen, USA) as previously described [39]. The pCR2.1 constructs carrying the cloned NA-segments, proved by enzymatic digestion and subsequent sequencing, were digested by "FastDigest" BsmBI enzyme (Thermo Scientific, USA) at 37°C for 15 min, and then the NA-cDNAs were ligated into BsmBI-linearized pHW2000 using T4 ligase (Thermo Scientific, USA). The ligation reactions were transformed into DH5-alpha (New England Biolabs, USA) according to manufacture instructions. Since the cloned HA-segments of H9N2/SA showed instability in the genetic backbone of pCR2.1 sequencing vector, the full-length RT-PCR products of HA segments amplified with primers containing EcoRV and BsaI sites were digested with EcoRV enzyme (Thermo Scientific, USA) for 1 h at 37°C to produce blunt-ends to allow the ligation into the EcoRV-linearized pSMART-LC-Kan (Lucigen, USA). The positive and correct vector/insert constructs were digested with BsaI-HF (NEB, Germany) for 15 min at 37°C. The digested product was then purified using "QIAquick PCR Purification Kit" (Qiagen, Germany). Meanwhile, the pHH21, pHW2000, pMPccdB and pMKPccdB were linearized with the "FastDigest" BsmBI (Thermo Scientific, USA) at 37°C for 15 min. The linearized vectors were then purified using "GeneJET Gel Extraction Kit" (Thermo Scientific, USA) and ligated to BsaI-sticky ends of the amplified cDNAs using T4 DNA ligase (Thermo Scientific, USA) according to the manufacture instructions. The ligation products were transformed into Top10 (Invitrogen, USA), DH5-alpha (New England Biolabs, USA), DH10B (Invitrogen, USA), XL1-Blue (Stratagene, USA), Stbl2 Table 2. Primers to mutate BsmBI and BsaI sites in the genetic backbone of pMKPccdB.

Western blot analysis
Confluent 293T cells seeded in a 6-well plate, were transfected in triplicates with 2 μg plasmid DNA/well of pHW2000 and pMKPccdB plasmids encoding the HA segment of H5N1/ VSVRI (pHW-H5 HA/VSVRI and pMKP-H5 HA/VSVRI, respectively). Trans-IT2020" (Mirus, USA) was us as a transfection reagent as previously described above in "Chloramphenicol Acetyl Transferase (CAT) assay". After 24 h, cells were lysed as previously described [43]. Cell lysates were separated on precast gradient NuPAGE Novex 4-12% Bis-Tris Protein Gels (Invitrogen, Germany) and transferred onto Immobilon-FL polyvinylidene fluoride (PVDF) membranes (Merck Millipore, Germany) AT 150 Volts/cm 2 . Following protein transfer, PVDF membranes were blocked using blocking buffer (1× TBS (20 mM Tris-HCl, pH 7.6, 140 mM NaCl) containing 5% non-fat dry milk) for 1 h at room temperature (RT). The membranes were washed 1 time using washing buffer (1× TBS-Tween: TBS containing 0.05% Tween20). Afterwards, detection of viral HA protein and cellular beta actin was achieved using primary mouse anti-influenza A virus (H5N1/HA1) monoclonal antibody (abcam, England) and rabbit anti-beta Actin monoclonal antibody (abcam, England) diluted in blocking buffer at a dilution of 1:5000 and 1:10000, respectively. One hour later, the membranes were washed for 3 times with washing buffer and incubated with secondary goat antimouse IRDye 800 and goat anti-rabbit IRDye 680 (1:10000 in blocking buffer containing SDS (1:1000 dilution of a 10% stock solution)), and added to the membranes in the dark for 1 h. After 3 times washing with washing buffer and one time with 1× TBS, the proteins were visualized using an "Odyssey Infrared Imaging System" and the according application software package (LI-COR, Germany). To quantify the amount of the protein, basic "Quantity One" software was used (Bio-Rad, Germany).

Transfection and rescue of recombinant viruses
To generate the different prototype IVVS, 1 μg of each pHW2000-or pMKPccdB plasmid encoding HA and NA segments of H9N2/SA or H5N1/VSVRI together with pHW2000 plasmids encoding the remaining segments of influenza virus A/Puerto Rico/8/34 (PR8, H1N1) were cotransfected into a co-culture of 293T/MDCK-II cells (ratio 3:1, 10 cm 2 ) as previously described with minor modifications [14,39,44]. Briefly, the transfection mixture consisting of 180 μl "Opti-MEM" (Gibco, Invitrogen, USA) and 8 μg of plasmid DNA (1 μg of each plasmid) along with 16 μl "Trans-IT2020" (Mirus, USA) were incubated for 45 min at RT. The transfection mixture was then diluted to 1 ml using Opti-MEM and transferred to 80-90% confluent cell monolayer to allow transfection. The cells were then incubated for 8 h at 37°C in the presence of 5% CO2. The transfection medium was replaced with 1 ml of infection medium composed of "Opti-MEM" containing 100 I.U./ml penicillin, 100 μg/ml streptomycin and 0.2% BSA (PAA, Germany) and the cell cultures were incubated for another 12 h. Afterwards, an additional 1 ml of infection medium containing 2 μg/ml TPCK-treated trypsin was added (Sigma-Aldrich, USA). The cell culture supernatant was harvested 48 h after addition of the TPCKtreated trypsin and cell debris was removed by centrifugation at 2500 rpm for 5min at 4°C. An aliquot of 500 μl of each supernatant was used for the inoculation of fresh MDCK-II cells (25 cm 2 ) and the cultures were then incubated for 72 h in the presence of TPCK-treated trypsin (1 μg/ml). To detect and characterize the recovered viruses, standard HA assay as well as RT-PCR using subtype-specific HA primers was performed (primers not shown). The titration of the recombinant viruses was performed using foci assay as described previously [45].

Virus growth curves
To investigate the growth kinetic of the prototype IVVS expressing HA and NA of H5N1/ VSVRI and H9N2/SA that were either generated using pMKPccdB or pHW2000 reverse genetic systems, confluent monolayers of MDCK-II cells ( . Aliquots collected from the different viruses were then titrated using standard hemagglutinin assay and focus assay [45,46].

Results
Instability of cloned HA-cDNA of H9N2/SA and H5N1/VSVRI To generate the IVVS, the cDNA of HA and NA segments of the candidate strain have to be precisely and correctly cloned into a suitable reverse genetic (RG) system. Herein, the NA-cDNAs of H9N2/SA and H5N1/VSVRI were cloned into pCR2.1 for sequencing and further sub-cloning into the standard RG vector pHW2000. Both HA-cDNAs of H9N2/SA-and H5N1/VSRVI cloned in the backbone of pCR2.1showed genetic alterations. Upon transformation of DH5-alpha at 37°C none of the transformants contained the correct vector/insert construct as demonstrated by the enzymatic digestion (data not shown). Furthermore, repeated trials to transform pCR2.1 constructs with the respective HA-cDNAs into different E. coli strains (Top10, DH10B, XL1-Blue, Sure, Stbl2 and Stbl4) failed to obtain the correct plasmid constructs at both the standard temperature of 37°C and at lowered temperatures of 30 and 25°C and different antibiotic concentrations (25, 50 and 100μg/ml ampicillin). However, all transformants analyzed contained the correct vector/insert DNA when the same cDNAs were cloned into the cloning vector pSMART-LC-Kan. Therefore, the insert DNA of these positive clones carrying the HA-cDNA of H9N2/SA and H5N1/VSVRI in pSMART-LC-Kan, confirmed by enzymatic digestion and sequencing, were subsequently sub-cloned into the RG vectors pHW2000 [44], pMPccdB [39] and pHH21 [14] and transformed into different competent bacterial strains (Top10, DH5-alpha, DH10B, XL1-Blue, Sure, Stbl2, Stbl3 and Stbl4), propagated at different temperatures (25, 30°C and 37°C) and antibiotic concentrations (25,50 and 100 μg/ml ampicillin). To our surprise, both sub-cloned HA-cDNAs were unstable again under these conditions in the different vectors and bacterial strains. The enzymatic digestion of different constructs encoding both cDNAs did not match the expected pattern (data not shown). Although, a few clones of pHW2000, pMPccdB and pHH21 carrying the HA-cDNA of H9N2/SA were demonstrated positive by enzymatic digestion, the sequencing results showed that they were not functional due to an insertion or a deletion of one nucleotide in the HA2 region that would lead to frame-shifting and hence inactive protein expression (Fig. 1). Chimeric H9-HA encoding-RG vectors In comparison with the genetically stable HA segment of A/Chicken/Beijing/2/1997, H9N2 (H9N2/Bei), the HA1 region of H9N2/SA-HA shows a sequence homology of 88.5%, while the homology of the HA2 regions is 98.2%. Since the nucleotide deletion/insertions appeared in the HA2 (Fig. 1) after transformation of different RG vectors encoding the H9N2/SA HA-cDNA, we thought to generate a chimeric HA gene comprised of the H9N2/SA-HA1 region (containing the important antigenic sites for the generation of neutralizing antibodies [47]) ligated to the H9N2/Bei-HA2 region. By this approach we aimed to obtain a stable plasmid to create a recombinant vaccine virus expressing the desired epitopes [47]. To this point, two chimeric HAs were constructed-pHW-HA-SA/Bei and pHW-HA-Bei/SA (control), cloned into pHW2000 and transformed into XL-1 blue. The first chimeric HA segment was composed of the HA1 region of H9N2/SA and the HA2 region as well as the 3 0 and 5 0 non-coding regions (NCR) of H9N2/Bei (S1A Fig.). Under investigation the sequence of this chimeric HA also showed insertion/deletion instability in the HA2 region. In contrast, the sequence of the other chimeric HA possessing the HA1 of H9N2/Bei and the HA2 region as well as the 3 0 and 5 0 -NCR of H9N2/SA (S1B Fig.) was genetically stable. Based on this finding, we favor that the instability of the cloned HA of H9N2/SA-cDNA is probably related to the sequence of the HA1 region, but its impact appears in the sequence of the HA2 region, preventing the use of the chimeric HA constructs.

Efficient cloning and high genetic stability of HA segments upon transformation of vector-DNA into HB101 competent cells
Since different E. coli genotypes with recA1 (DH5-alpha, DH10B, Top10 and Stbl4), recA-deficient (XL-1 blue), recB-deficient (Sure competent cells) or recA13/HB101-derived (Stbl3) failed to maintain the insert stability, we sought to transform the ligation reaction of pHW2000, pMPccdB or pHH21 with the HA-cDNA of H9N2/SA into recA13/HB101 competent cells. Notably, the HB101 could faithfully maintain insert stability. Following transformation, two phenotypes of the transformants were detected on the agar plates, large and small colonies (Fig. 2: A and C). Only transformants of the small-sized colonies showed the expected restriction digestion pattern (Fig. 2: B and D) as well as the correct nucleotide sequence even under commonly used temperature condition and antibiotic concentration (37°C, 100 μg/ml ampicillin). Nevertheless, unlike the transformants of large-sized colonies containing incorrect plasmids, the transformants of small-sized colonies containing stable plasmids only grew poorly, probably to restrain the expression of an aberrant toxic protein product. However, not only toxic gene products, but also other mechanisms or bacterial pathways affecting recombination and/ or removal of toxic sequences could account for the genetic instability. To confirm the advantageous ability of HB101 to stabilize the vector/insert DNA, also the previously unstable HA-cDNA of H5N1/VSVRI cloned into pHW2000 vector was transformed into HB101 competent cells. Similarly, the small-sized HB101 transformants harbored the correct vector/insert constructs. In a trial to ensure that this reduction in colony size is related to the cloned HA-cDNA of H9N2/SA and H5N1/VSVRI, the HB101 cells were also transformed with stable pHW-HA-Beijing and pHW-HA-KanI (pHW2000 plasmids encoding the HA segment of A/Chicken/Beijing/2/1997(H9N2) or A/Thailand/1(KAN-1)/2004(H5N1)). The pHW-HA-Beijing and pHW-HA-KanI transformants showed homogenous common large-sized colonies (S2 Fig.). This demonstrates that stability of genetic volatile inserts is probably maintained in HB101 by reduced bacterial growth.

Low copy number pMKPccdB vector stabilizes HA-cDNAs of H9N2/SA and H5N1/VSVRI
Based on the fact that the H9-HA-cDNA was stable in the backbone of pSMART-LC-Kan and on the previous study of Zhou et al., regarding the cloning of an unstable PB2 segment of A/ Turkey/Ontario/7732/1966 (H5N9) into pHH21 [36], we replaced the backbone of pMPccdB [39] by pSMART-LC-Kan to construct an improved pMKPccdB (Fig. 3). The improved vector also contains the ccdB gene as a negative selection marker, which helps to ameliorate the  screening process [39,48]. Furthermore, all BsmBIand BsaI-sites in the backbone of pMKPccdB vector were masked by silent mutations and the final construct was generated in three versions with AarI, BsmBI or BsaI cloning sites. Although the H9N2/SA and H5N1/VSVRI HA-cDNAs cloned into the established pHH21, pHW2000, pMPccdB vectors were not stable in a wide range of E. coli strains except for the recA13/HB101, the same cDNAs showed high stability when cloned into the improved pMKPccdB vector and transformed into commonly used DH5-alpha, DH10B and XL-1 blue competent cells. This stability was illustrated by enzymatic digestion and verified by correct nucleotide sequencing (data not shown). To prove that the pMKPccdB is fully functional in regard to the rescue of recombinant IAV by RG, recombinant virus was generated using pMKPccdB plasmids encoding the entire eight viral segments of H9N2/Bei (data not shown).
To further test whether the backbone of the vector, would affect the Pol-II transcription efficiency of pMKPccdB we cloned the PB1 segment of influenza virus A/Hamburg/04/09 (Hamburg/H1N1) into pMKPccdB (pMKP-PB1-Hamburg) or pHW2000 (pHW-PB1-Hamburg, control). These were transfected into 293T cells together with pHW2000 plasmids containing the cloned PB2-, PA-and NP-segment of Hamburg/H1N1 expressing the according mRNAs and pPol-I-CAT-RT to generate a vRNA-like Pol-I-transcript encoding the chloramphenicol acetyl transferase (CAT) reporter gene in order to reconstitute viral ribonucleoprotein complexes (vRNP) expressing CAT [15]. When analyzing the CAT activities of the vRNPs reconstituted with either pMKP-PB1-Hamburg or pHW-PB1-Hamburg we found that they were comparable (Fig. 4A).
Furthermore, the HA protein expression level of pHW2000 and pMKPccdB encoding the HA of H5N1/VSVRI (pHW-H5 HA/VSVRI and pMKP-H5 HA/VSVRI, respectively) was determined in eukaryotic 293T cells as a direct indicator of the efficiency of the Pol-II directed mRNA transcription. Confirming our results of the CAT activity derived from the plasmid based in vitro RNP reconstitution, the HA protein production level of pMKPccdB was comparable to that of pHW2000 plasmid (Fig. 4B). These findings clarify that the pSMART-LC-Kan backbone of pMKPccdB does not impair the function of the Pol-I/ccdB/Pol-II cassette derived from the pMPccdB vector in eukaryotic cells.

Recombinant IVVS (6+2) viruses rescue and growth curve kinetic
After we had demonstrated that we could generate genetically stable RG plasmids for the H9N2/SA-HA and the H5N1/VSVRI-HA, either with the help of the E. coli strain HB101 or with the new pMKPccdB vector in common E. coli strains, we next aimed to rescue 6+2 reassortant IVVS with the according HA and NA of either H9N2/SA or H5N1/VSVRI and RG plasmids carrying the remaining segments of PR8.
doi:10.1371/journal.pone.0116917.g003 Figure 4. Effect of pSMART-LC-Kan cassette on the transcription efficiency of pMKPccdB versus pHW2000 in eukaryotic cells. A) CAT activity in 293T cells: cells were transfected with pPOLI-CAT-RT plasmid and pHW2000 plasmids encoding PB2, PA and NP of A/Hamburg/04/09 (Hamburg/H1N1) together with pHW2000 or pMKPccdB encoding PB1 segment of the Hamburg/H1N1. 24 h post transfection, cell extracts were prepared and assayed for CAT activity with [ 14 C] chloramphenicol and thin-layer chromatography as previously described [15]. B) Western blotting analysis of expressed HA protein from H5N1/VSVRI: 293T cell monolayers were transfected in triplicates with pHW2000 or pMKPccdB plasmid DNA encoding the HA segment of H5N1/VSVRI (pHW-H5 HA/VSVRI and pMKP-H5 HA/VSVRI, respectively). 24 h post transfection, cells were lysed, subjected to SDS-PAGE, transferred onto PVDF membrane and detected by HA-specific antibodies. The amount of HA protein expressed by pMKP-H5 HA/ VSVRI plasmid was similar to that expressed by pHW-H5 HA/VSVRI. Cellular beta actin was used for H5N1/LP compared to PR8-pHW-H5N1/LP as determined by standardized HA assay (Fig. 5C) and focus forming assay (Fig. 5D).

Discussion
The RG approach has facilitated and accelerated the basic research on IAV as well as vaccine development in the past two decades [18,52,53]. To generate recombinant IVVS, the RT-PCR copies of the eight viral segments are cloned precisely into an RG system to produce vRNAs of the eight viral genomic segments and at least the mRNAs of the viral polymerase subunits (PB1, PB2, PA) and the nucleoprotein (NP). To obtain these vRNAs, the RT-PCR copies of the corresponding viral segments are cloned between an RNA polymerase I (Pol-I) promoter and normalization. Error bars show standard deviation (SD) of the mean (N = 3). (ns) indicates a non-significant difference (P value<0.05).
doi:10.1371/journal.pone.0116917.g004 terminator or the HDV ribozyme [13,14]. In order to express the necessary subunits of the viral polymerase and the NP that together with the vRNAs form the biological active RNPs, as well as the other viral proteins, the viral segments are also cloned between an RNA polymerase II (Pol-II) promoter and polyadenylation sequence [14,15,37]. The correct plasmids producing the vRNAs and mRNAs of IAV are then transfected together into suitable cells [54], to recover the recombinant IAV [14,44]. However, the occasional genetic instability of plasmids encoding the different viral genome segments (specially of HA and NA of candidate circulating or emerging IAV) in E. coli hinders a reliable and rapid construction of complete RG systems and the generation of IVVS. These are important pre-requisites to combat these strains and to assess their replication, virulence, pathogenicity, host range, and transmissibility. In this study, we demonstrate two different solutions to solve the stability problem of the cloned HA-segments from H5-and H9-subtypes transformed into different E. coli competent cells.
The plasmid vectors (for RG systems) are generally made from DNA sequences, derived from different sources such as bacteria and/or viruses and/or mammals. Occasionally, the foreign plasmid DNA can restrain or stop the growth of the bacterial host, undergoes structural rearrangement or can rapidly be lost during bacterial cell divisions [55]. Additionally, it is also well known that E. coli competent cells can introduce specific mutations and/or deletions in the foreign DNA to avoid the extensive aberrant expression of possibly toxic proteins and to maintain it in an inactive form. Therefore, different recA mutations were generated to minimize the occurrence of such insertions, deletions and plasmid DNA rearrangement caused by the recA protein of the bacteria [32].
Recently, several studies have demonstrated that vaccination with chimeric HA constructs can improve the antigen yield against IAVs and also elicit efficient polyclonal humoral responses to the HA ecto-and stalk domain protecting against heterologous and heterosubtypic challenges [56][57][58]. In our case, we thought that this approach might be helpful to overcome the problem of HA instability as we detected the bacterial artifacts in the stalk domain (HA2), but not in the globular head domain (HA1) of the H9N2/SA-HA. Nevertheless, based on our findings with chimeric HA constructs, that the origin of the observed instability was related to the sequence of the HA1 region, but its impact appeared in stalk domain (HA2), this strategy of chimeric HAs did not provide a solution to avoid the genetic instability of the cloned HAs.
To find an E. coli strain that would stably maintain the recombinant plasmids (e.g. pHH21, pMPccdB and pHW2000) containing the HA-cDNAs of H9N2/SA and H5N1/VSVRI, we tested various recA alleles of standard competent bacteria; recA1 (DH5-alpha, DH10B, Top10 and Stbl4), recA-deficient (XL-1 blue), recB-deficient (Sure competent cells) and recA13 (HB101 and Stbl3). Unlike recA1, recAand recB-deficient competent cells, the HA-RT-PCR products of H9N2/SA and H5N1/VSVRI were stably maintained in recA13/HB101 competent cells. The HB101 strain is a hybrid of E. coli K12 and B-bacteria, containing the recA13 mutation that minimizes recombination and supports insert stability. In addition, it carries the hsdS20(r B − m B −) restriction and methylation minus genotype, derived from E. coli B-bacterium, to prevent the cleavage of cloned DNA by endogenous restriction enzymes and allows efficient cloning of the methylated DNA [59]. In contrast, the various competent cells, which failed to stabilize the plasmids encoding HA-segments of H9N2/SA and H5N1/VSVRI are derivatives of the wild type E. coli K12 strain. Therefore, we think that the ability of HB101 to stably maintain the inserts is related to a combination of the recA13 mutation and the genomic sequence derived from E. coli B-strain. Despite the fact that the Stbl3 E. coli strain also accommodates the recA13 mutation and is derived from HB101, it did not stabilize the cloned HA-segments (data not shown). This strain was engineered to produce high yield of plasmid DNA and to enhance the cloning of methylated genomic DNA. It was also modified to optimally stabilize cloned products into lentiviral vectors [55]. Presumably, these modifications of the Stbl3 strain lead to a loss of the advantageous property of the parental HB101 strain. Importantly, highly unstable HA segments like the H9N2/SA-HA were stable in HB101 in small-scale and largescale cultivations at 30°C. Nevertheless, in general the HB101 strain produces only low yields of plasmid DNA, which is also unprotected from restriction enzymes due to the restrictionmodification minus phenotype (r − , m − ).
Recently, the low-copy RG vector, pGJ3C3 was developed by substitution of the high copy number origin of replication (ori) from pUC in the standard unidirectional RG vector pHH21 with the low copy number ori of pSMART-LC-Kan (pBR322) [36]. Nevertheless, when we used the pHW2000, which also contains the low copy number ori (pBR322), the unstable HA constructs could not be stabilized in recA1 and recA-deficient bacteria. To this point, we constructed the bidirectional pMKPccdB RG vector based our previously published pMPccdB vector [39] and pSMART-LC-Kan to overcome the instability of the previously mentioned HA-cDNAs transformed into recA1 and recA-deficient competent bacteria. The improved RG system, pMKPccdB, combines the main features of pSMART-LC-Kan and pMPccdB. Several distinct elements derived from pSMART-LC-Kan ensure the maintenance of unstable inserts. Firstly, the repressor of primer (ROP) protein helps to down regulate and minimize the plasmid replication [60]. Secondly, the kanamycin selection is more stringent than ampicillin selection. Thirdly, the presence of three terminators derived from pSMART-LC-Kan prevents unwanted protein expression based on accidental promoter sequences present in the cloned genes. Despite the fact that cytomegalovirus (CMV) immediate-early promoter (derived from pMPccdB) has the ability to initiate gene expression in E. coli [61], we assume that due to the low copy number pBR322 ori, the ROP gene and three termination sites of pMKPccdB unwanted gene expression is prevented in E. coli. Furthermore, the change of the vector backbone of pSMART-LC-Kan did not affect the viral protein expression by the resulting pMKPccdB in eukaryotic cells when compared with standard pHW2000 vector in RG systems.
Moreover, the pMKPccdB has an efficient and strong negative selection marker, the ccdB gene, allowing only E. coli clones to grow that carry the plasmid with the inserted DNA [39,62]. To further improve the vector for an efficient one-step cloning procedure [63,64], the sites of the commonly used type-II restriction enzymes BsaI and BsmBI, present in the pSMART-LC-Kan-derived backbone and in the ccdB gene (derived from of pMKPccdB) were deleted. Ultimately, the low copy number pMKPccdB RG vector containing the cloned cDNAs of the desired HA segments was successfully employed to reliably generate IVVS (6+2) for IAVs with indistinguishable replication characteristics compared to recombinant viruses engineered by the standard pHW2000 RG vector and HB101 strain. We should mention that our attempt to combine both strategies (transformation of HB101 with pMKPccdB) only resulted in a poor plasmid DNA yield and therefore was not superior to transformation of standard E. coli (e.g. XL1-blue) with pMKPccdB.
In summary, these improved strategies facilitate cloning and consequently the rescue of recombinant IAVs with unstable HA segments in the attempt to develop recombinant IVVS and will further allow basic research to focus on scientific questions regarding the biological characteristics of RG-derived recombinant IAVs.
Supporting Information S1 Fig. HA chimeras. Schematic diagram of HA chimeras constructed to circumvent the observed instability of the H9N2/SA-HA sequence, which we suspected to be located in the HA2part, by exchanging the HA1 sequence of the pHW2000 plasmid (pHW-HA-Bei) stably encoding the A/Chicken/Beijing/2/1997 (H9N2/Bei) HA with the corresponding sequence from the pSMART-LC-Kan plasmid (pSMART-HA-SA) possessing the HA-cDNA of A/chicken/Saudi Arabia/CP7/1998 (H9N2/SA) and vice versa. A) The chimeric HA segment of H9N2/Bei with the HA1 region of H9N2/SA cloned into pHW2000 (pHW-HA SA/Bei) showed genetic instability in the E. coli strain XL-1 blue. B) The chimeric HA of H9N2/SA with the HA1 region of H9N2/Bei cloned into pHW2000 (pHW-HA-Bei/SA) showed genetic stability in the E. coli strain XL-1 blue. (TIF) RT-PCR and PCR using specific primer pairs for HA and NA segments and by B) standard HA assay. As a template for RT-PCR and direct PCR, the purified viral RNA was extracted from the supernatants harvested 72 h post transfection. Unlike the PCR reactions, the RT-PCR reactions resulted in the expected HA and NA amplification products for both viruses, indicating that the products did not result from residual transfected plasmid DNA. This confirmed that the positive HA titer was related to the rescued viruses. (TIF)