S were combined on single AAV2 capsid to 256373-96-3 biological activity produce double- and triple-mutant and efficiency of each vector was evaluated. (a) EGFP expression analysis at 48 h post-infection at MOI of 16103 vg/cell. (b) Quantification of transduction efficiency of each of the threonine-mutant AAV2 vectors. *P,0.005, **P,0.001 vs. WT AAV2. doi:10.1371/journal.pone.0059142.gcific threonine (T) residues on AAV2 capsids would likewise be expected to undergo phosphorylation, in the present 
study we systematically mutagenized each of the 17 surface-exposed T residues, and identified several single-mutant vectors that could increase the transduction efficiency up to 4-fold. Combinations of multiple T mutations on a single capsid identified modifications which further augmented the transduction efficiency up to ,10fold, compared with that of the WT AAV2 vector in HEK293 cells. It is of interest to note that two independent groups have previously reported mutations of specific T residues on AAV2 capsids. For example, Lochrie et al. [35] targeted the T residues at positions 330, 454, 455, 491, 503, and 550 in a tour de force effort to identify surface regions which bind antibodies, and DiPrimio et al. [41] targeted the T residue at position 659 in an effort to identify regions 15900046 critical for capsid assembly and genome packag-ing. In both studies, the T residues were substituted with either alanine (A), serine (S), or lysine (K) residues, or by peptide substitution. However, no increase in the transduction efficiency of any of the mutant vectors was observed. In contrast, in our studies, we substituted the surface-exposed T residues with valine residues. This further corroborates our recent observation of the critical role played by specific amino acid type in modulating the biological activity of AAV vectors [12,42]. When the most efficient threonine-mutation (T491V) was combined with a previously reported DprE1-IN-2 tyrosine triple-mutation (Y444+500+730F) [14] to generate a Y-T quadruple-mutant (Y444+500+730F+T491V) vector, the transduction 
efficiency of this vector was ,2?-fold higher than the tyrosine triple-mutant vector in murine hepatocytes, both in vitro and in vivo. However, combining the most efficient S-mutation (S662V) [12] with theLimits of Optimization of Recombinant AAV2 VectorsFigure 3. Evaluation of EGFP expression in H2.35 cell transduced with capsid optimized AAV2 vectors. The most efficient tyrosine, serine and threonine mutations were combined on single AAV2 capsid to produce several optimized AAV mutants. Efficiency of each vector was estimated on immortalized murine hepatocytes. (a) EGFP expression analysis at 48 h post-infection at MOI of 16103 vg/cell. (b) Quantification of transduction efficiency of each of the optimized scAAV2 vectors. *P,0.005, **P,0.001 vs. WT AAV2. doi:10.1371/journal.pone.0059142.gtyrosine triple-mutation negatively affected the transduction efficiency of the Y-S quadruple mutant (Y444+500+730F+S662V) vector as well as the Y-S-T pentuplemutant (Y444+500+730F+S662V+T491V) vector. Although several other combinations showed greater transduction efficiency compared with the WT AAV2 vector, neither combination of similar (quadruple, pentuple or sextuple-tyrosine; and triple and quadruple-threonine mutants), nor combination of the best performing YST mutations reached the level of expression from the triple-tyrosine mutant vector (Table S1). In view of the large number of combinations of mutations tested in the current studies, we focus.S were combined on single AAV2 capsid to produce double- and triple-mutant and efficiency of each vector was evaluated. (a) EGFP expression analysis at 48 h post-infection at MOI of 16103 vg/cell. (b) Quantification of transduction efficiency of each of the threonine-mutant AAV2 vectors. *P,0.005, **P,0.001 vs. WT AAV2. doi:10.1371/journal.pone.0059142.gcific threonine (T) residues on AAV2 capsids would likewise be expected to undergo phosphorylation, in the present study we systematically mutagenized each of the 17 surface-exposed T residues, and identified several single-mutant vectors that could increase the transduction efficiency up to 4-fold. Combinations of multiple T mutations on a single capsid identified modifications which further augmented the transduction efficiency up to ,10fold, compared with that of the WT AAV2 vector in HEK293 cells. It is of interest to note that two independent groups have previously reported mutations of specific T residues on AAV2 capsids. For example, Lochrie et al. [35] targeted the T residues at positions 330, 454, 455, 491, 503, and 550 in a tour de force effort to identify surface regions which bind antibodies, and DiPrimio et al. [41] targeted the T residue at position 659 in an effort to identify regions 15900046 critical for capsid assembly and genome packag-ing. In both studies, the T residues were substituted with either alanine (A), serine (S), or lysine (K) residues, or by peptide substitution. However, no increase in the transduction efficiency of any of the mutant vectors was observed. In contrast, in our studies, we substituted the surface-exposed T residues with valine residues. This further corroborates our recent observation of the critical role played by specific amino acid type in modulating the biological activity of AAV vectors [12,42]. When the most efficient threonine-mutation (T491V) was combined with a previously reported tyrosine triple-mutation (Y444+500+730F) [14] to generate a Y-T quadruple-mutant (Y444+500+730F+T491V) vector, the transduction efficiency of this vector was ,2?-fold higher than the tyrosine triple-mutant vector in murine hepatocytes, both in vitro and in vivo. However, combining the most efficient S-mutation (S662V) [12] with theLimits of Optimization of Recombinant AAV2 VectorsFigure 3. Evaluation of EGFP expression in H2.35 cell transduced with capsid optimized AAV2 vectors. The most efficient tyrosine, serine and threonine mutations were combined on single AAV2 capsid to produce several optimized AAV mutants. Efficiency of each vector was estimated on immortalized murine hepatocytes. (a) EGFP expression analysis at 48 h post-infection at MOI of 16103 vg/cell. (b) Quantification of transduction efficiency of each of the optimized scAAV2 vectors. *P,0.005, **P,0.001 vs. WT AAV2. doi:10.1371/journal.pone.0059142.gtyrosine triple-mutation negatively affected the transduction efficiency of the Y-S quadruple mutant (Y444+500+730F+S662V) vector as well as the Y-S-T pentuplemutant (Y444+500+730F+S662V+T491V) vector. Although several other combinations showed greater transduction efficiency compared with the WT AAV2 vector, neither combination of similar (quadruple, pentuple or sextuple-tyrosine; and triple and quadruple-threonine mutants), nor combination of the best performing YST mutations reached the level of expression from the triple-tyrosine mutant vector (Table S1). In view of the large number of combinations of mutations tested in the current studies, we focus.
