Poliovirus is one of the first recognized viruses to undergo recombination 
. Recombination has been found frequently in wild polioviruses, VDPVs, and vaccine-related polioviruses 
. Usually, the first crossover site is in the P2
nonstructural region and often within the 3D
coding region 
. Natural intertypic capsid recombination between Sabin strains is a relatively rare phenomenon, possibly due to structural constraints that maintain the integrity of the capsid shell. The integrity of the capsid region of poliovirus seems to be very important for propagation of the viruses themselves. Some in vitro
experiments have shown that if the crossover site is located in the capsid-coding region, the intertypic recombinant chimera will be nonviable 
; in addition, some other experiments have shown that, compared to its parental strains, the intertypic recombinant chimera with the crossover site in the capsid-coding region is unstable 
Some similar intertypic recombinant chimera have been reported previously 
, but to the best of our knowledge, 136 nucleotides insert is the longest donor sequence that has been reported, so the longest poliovirus type 2 insert in the VP1
coding region of poliovirus capsid recombinant has been reported herein.
Although the first crossover sites were different from each other among the 10 VP1
capsid recombinants, they all gained type 2 NAg3a amino acids in the VP1 structure protein in the Sabin 3 background. Although the replacement of NAg3a in these recombinant viruses could result in partial antigenic changes, as shown before 
, the overall type 3 antigenic structure was not significantly altered since all virus isolates were readily neutralized by type 3 polyclonal serum. This meant that NAg3a may have altered some of the type 3-specific antigenic properties but that Sabin 3 had not acquired any type 2-specific characterizations, possibly based on the fact that the inserted type 2 NAg3a is located on the surface of the virion and is implicated in receptor binding, allowing more freedom for aberrant folding 
The NAg3a amino acid of the Sabin 3 strain seems atypical; other type 3 wild poliovirus isolates that have circulated worldwide in recent years (after 1980) have sequences of NAg3a more like the Sabin 2 strain. But the type 3 wild poliovirus isolates from earlier year (before 1967) still have sequences of NAg3a more like the Sabin 3 strain. The possibility exist that recent type 3 wild poliovirus isolates may be a recombinant having NAg3a sequences derived from another strain during between 1967 and 1980. It seems likely, but is hard to prove, that perhaps the type 3/type 2 recombination events in the 3′ end of VP1 coding region may have a higher fitness.
The results of age estimation guided us towards the reconstruction of the potential histories of these clinical isolates. OPV strains had replicated in the human gut for 38–120 days (from the Ks estimate) and for 65–127 days (from the Kt estimate) after the initial OPV doses were given. Based on the fact that type 2 and type 3 vaccine viruses can replicate and be excreted by immunocompetent vaccinees for about 3 months after vaccination 
and considering the small dispersion of ages, the possibility that each particular recombination event occurred in the person from whose stool specimen the virus was isolated cannot be excluded. During the short time period, nucleotide substitutions and genetic exchanges were selected very quickly, probably as a response to different selective advantages, to increase the fitness of the Sabin strains for replication in the human gut by preventing the accumulation of harmful mutations.
The limitation with the method of the estimation of evolution used in this study is that this approach assumes a fixed substitution rate and multiplies it by the estimated number of nucleotide substitutions (total substitutions or synonymous substitutions), but in fact, the actual amount of evolution (number of substitutions) mainly depends on the virus effective population size inside the host and the frequency of replication. Additionally, the recombination events would lead to a change in the selection pressure, and most likely, there would be an increase in selection resulting in an increased evolutionary rate, meaning that the ages of the VP1 capsid recombinants may have been overestimated. This is most obvious in the case of CHN11144, whose estimated ages (65 days from the Ks estimate or 118 days from the Kt estimate) are greater than the time from administration of OPV to sampling (41 days).
All the children had one or more OPV before sampling, and the intervals between the dates of administration of last OPV and sampling are from 26 days to 1027 days, which indicated not all natural VP1 capsid recombinants directly derived from the OPV strains they received, most likely, some children were affected by them from environment. There was evidence that, like other vaccine-like polioviruses, VP1 capsid recombinants could survive and be detected from the environmental sewage 
Genetic attenuation usually decreases viral fitness, and all the VP1
capsid recombinants had nucleotide substitutions of U-to-C at nt472 in the 5′-UTR
region and a transition C-to-U at nt2493, which resulted in a Thr-to-Ile substitution of residue 6 of VP1, and these were direct reversions to the neurovirulent precursor of Sabin 3, Leon/USA/1937. These two nucleotide substitutions have also been frequently found in type 3 VDPV strains, including 2002 VDPV isolates from sewage in Estonia 
, 2006 iVDPV isolates from an Iranian child 
, and in the Sabin 3-related polioviruses 
, which may indicate that these “hot spots” faced intensive selective pressures as the OPV strain replicated in the human gut.
Five of the VP1 capsid recombinants (CHN5275, CHN6060, CHN6218, CHN6356 and CHN12121) did not lose the temperature-sensitive phenotype of parent P3/Sabin strain. This result differed to three previously reported capsid recombinants which had almost entirely lost the temperature sensitivity 
, but was similar to a recent report 
, and other five VP1 capsid recombinants in this study had almost entirely lost the temperature sensitive phenotype. It had been reported that the temperature sensitivity phenotype of type 3 poliovirus is mainly attributable to a difference in residue 91 of the VP3 capsid protein 
, but this amino acid substitution could not be found in the VP1 capsid recombinants in this study except strain CHN11185h (), and this might explain the preservation of temperature-sensitive phenotype of the viruses. Besides residue 91 in VP3 capsid region, 11 other amino acid substitutions have been suggested to be related to temperature sensitivity, four in VP1 region (residues 34, 54, 132 and 263), four in VP2 region (residues146, 200, 215 and 265), and three in VP3 region (residues 108, 175 and 178) 
. Among these, the substitution Ala to Thr or Ala to Val at position 54 in the VP1 capsid region were found in strains CHN5275, CHN6053 and CHN12092.
In conclusion, 10 natural type 3/type 2 intertypic VP1 capsid-recombinant polioviruses, in which the first crossover sites were found to be in the VP1 coding region, were isolated and characterized. In spite of the complete replacement of NAg3a by type 2-specific amino acids, the serotypes of the recombinants were not altered, and they were totally neutralized by polyclonal type 3 antisera but not at all by type 2 antisera. NAg3a of the Sabin 3 strain seems atypical; other wild-type poliovirus isolates that have circulated in recent years have sequences of NAg3a more like the Sabin 2 strain. It is possible that recent type 3 wild poliovirus isolates may be a recombinant having NAg3a sequences derived from another strain during between 1967 and 1980, and the type 3/type 2 recombination events in the 3′ end of the VP1 coding region may result in a higher fitness.