Improving virus production through quasispecies genomic selection and molecular breeding

Virus production still is a challenging issue in antigen manufacture, particularly with slow-growing viruses. Deep-sequencing of genomic regions indicative of efficient replication may be used to identify high-fitness minority individuals suppressed by the ensemble of mutants in a virus quasispecies. Molecular breeding of quasispecies containing colonizer individuals, under regimes allowing more than one replicative cycle, is a strategy to select the fittest competitors among the colonizers. A slow-growing cell culture-adapted hepatitis A virus strain was employed as a model for this strategy. Using genomic selection in two regions predictive of efficient translation, the internal ribosome entry site and the VP1-coding region, high-fitness minority colonizer individuals were identified in a population adapted to conditions of artificially-induced cellular transcription shut-off. Molecular breeding of this population with a second one, also adapted to transcription shut-off and showing an overall colonizer phenotype, allowed the selection of a fast-growing population of great biotechnological potential.


Construction of vectors G1RL0 and G1RCMsKp
For the construction of the G1RL0 vector, five mutations were sequentially introduced into the G1RC 1 vector. Two of them (C140T and A194G) were introduced at once, following a method previously described 2 , while the other three (A394G, C473T, C647A) were individually introduced, using the QuikChange Site-Directed Mutagenesis kit (Agilent Technologies). All introduced mutations were confirmed by sequencing using vector-derived external primers. The list of primers is provided in Supplementary Table 7.
For the construction of the G1RCMsKp vector, the same site-directed mutagenesis procedure was used to introduce restriction sites for the MscI and KpnI enzymes between the IRES and the FLuc gene. All introduced restriction sites were confirmed by sequencing using vector-derived external primers.
Primers used are shown in Supplementary Table 7.

Recognition by H7C27 and K34C8 monoclonal antibodies.
Virus stocks used in the analysis of recognition by antibodies were treated with 1% NP-40 for 30 min at 37ºC. Viruses recovered in the supernatants were subjected to three 30-s sonication cycles of at 60 W.
H7C27 mAb recognizes the glycophorin A binding site 3, 4 while K34C8 mAb is directed against the immunodominant site 3 . While H7C27 epitope is present in the protomers, procapsids and capsids, the epitope recognized by the K34C8 mAb is present only in procapsids and capsids 5 . For the recognition with each individual mAb an indirect sandwich ELISA was performed 6 , in which particles were captured by a convalescent-phase serum and detected with H7C27 or K34C8 mAbs. All mAbs were used at the highest dilution (1/10000) yielding recognition and were detected using a labeled anti-mouse IgG. An average of 1.5 x 10 6 TCID50 per well was used.
Recognition of virus stocks by a polyclonal convalescent serum was also tested. A direct sandwich ELISA was performed in which particles were captured and detected with the same convalescent-phase serum. In the detection step the antibodies from the convalescent-phase serum were labeled and used at the highest dilution (1/1400) yielding recognition.
A cut-off level was established corresponding to the mean ± 3 SD of the unspecific recognition of mock-infected FRhK-4 cell extracts. Three different stocks of each population were tested twice.

Physical Stability.
Treatments at 61ºC for 5 min, pH 2 for 1h at 37ºC and 1% biliary salts for 4h at 37ºC were performed and the resistance of each population evaluated as previously described 7 . To quantify virus decay, a control test of non-treated viruses kept for the same length of time at 37°C was run in parallel. The Log10 reduction of the virus titer after each treatment (Nt) compared with the infectious titer of the parallel non-treated virus control (N0) was figured. Three different stocks for each population were tested and all samples were titrated twice.

Sucrose-Iodixanol gradients
Infected cell culture supernatant fluids were centrifuged at 1,500 x g for 10 minutes at 4°C to pellet any cellular debris and further clarified by centrifuging twice at 10,000 X g for 30 minutes at 4°C. Viruses were further concentrated by ultracentrifugation at 100,000 X g for 1 hour at 4°C. Finally, the resulting pellet was resuspended in 1 ml of phosphate-buffered saline and briefly sonicated, loaded onto a pre-formed 6-50% iodixanol-(OptiPrep, Axis-Shield) sucrose step gradient, and then centrifuged at 205,000 X g for 2 hours and 45 minutes at 4°C using an SW41 Ti rotor in a Beckman Coulter Optima L-90K Ultracentrifuge.
Later, approximately 20 fractions of 0.5 ml each were collected from the gradient and the density of each fraction was determined using a refractometer.
RNA from gradient fractions was extracted using the Nucleospin RNA Virus Extraction Kit (Macherey-Nagel) and HAV genome copy numbers were determined by real-time qRT-PCR as previously described 8

Statistical analysis.
Statistical differences between the different virus populations regarding virus production/cell, plaque diameter, monoclonal and polyclonal antibodies recognition and physical stability were assessed using the Student's t-test (twosided), after verifying the normality of data with the Kolmogorov-Smirnov test.