Identification of two new CRF_BF in Rio de Janeiro State, Brazil

Phylogenetic trees and recombinant analysis from 1504 HIV-1 nucleotide pol sequences obtained between 1998 and 2005 from HIV-1 positive patients residents in Rio de Janeiro State, Brazil, identified two distinct BF pol recombinant profiles, with three samples in profile I and four samples in profile II. These epidemiologically unlinked patients had their samples selected for near full-length DNA genome sequencing. Bootscanning and informative site analysis show that all samples included within each BF pol recombinant profile share the same recombination pattern throughout the viral genome, allowing the description of two new CRFs_BF, designated as CRF39_BF and CRF40_BF.

Several HIV-1 subtypes, sub-subtypes, circulating recombinant forms (CRFs) and a huge number of unique recombinant forms (URFs) have been extensively described worldwide ^[1]. In Brazil, HIV-1 subtype B is the most prevalent, followed in a lower extension by subtype F (and BF recombinant forms) in most parts of the country ^[2]. Recently two new CRFs have been described in São Paulo, Southeastern region, involving subtypes B and F (CRF28_BF, CRF29_BF) ^[3]. Herein we report the presence of two new CRF_ BF, denominated CRF_BF39 and CRF_BF40, in Rio de Janeiro State, Brazil.

For this study, a total of 1504 HIV-1 nucleotide pol sequences obtained between 1998 and 2005 from HIV-1 positive patients who were residents in Rio de Janeiro State, Brazil, were analyzed. Phylogenetic trees were constructed by the maximum likelihood method with PAUP*, version 4.0b10 ^[4]. Analysis of recombination was performed by bootscanning, using the SimPlot software, version 3.5.1 ^[5]. From this screening, two distinct BF pol recombinant profiles were identified, in epidemiologically unlinked patients, with three samples in profile I and four samples in profile II.

These samples were selected for full-length DNA genome sequencing (8080 bp) (the polymerase chain reaction protocol and primers used in the study are available from the authors on request). Recombination breakpoints were characterized by bootscaning analysis (sliding window of 400 nt moving in steps of 20 nt using the Kimura two-parameter model), and informative site analysis as described previously ^[6]. These analyses clearly show that all samples included within each BF pol recombinant profile share the same recombination pattern throughout the viral genome, allowing the description of two new CRFs_BF, designated CRF39_BF and CRF40_BF, circulating in Rio de Janeiro State, Brazil.

The schematic near full-length genome structures of these CRFs and the Neighbour-joining analyses used to infer phylogenetic relationships of each sub-genomic fragment are illustrated in Fig. 1 ^[7].

Fig. 1

The CRF39_BF, represented by sequences 03BRRJ103, 03BRRJ327 and 04BRRJ179, displays six breakpoints along the genome (Fig. 1a). This CRF was mostly composed by subtype B (56%), and contains three segments of sub-subtype F1, as follows: two segments in the pol region; the first one in RT (position 2836-3020 relative to HXB2 genome) and the other one in integrase (3649-3866). The third F1 segment includes part of tat, all of vpu, env and the major part of the nef gene (5978-9383), giving a total fragment of 3405 nt.

The CRF40_BF, represented by sequences 05BRRJ055, 04BRRJ115, 05BRRJ200 and 04BRSQ46, presents a complex mosaic structure with ten breakpoints along the genome (Fig. 1b). This CRF was mostly composed by sub-subtype F1 (53%) distributed in five segments as follows: the first fragment is localized in the gag/pol region comprising the end of the p24 and beginning of the PR region (1340-2458); the second one corresponds to a segment of 1235 nt in the RT region (2631-3866); the third one includes part of the integrase and the beginning of vif (4584-5037); the fourth one comprises the env gene (6564-8034); and the fifth fragment includes part of the nef gene (8978-9360). The remaining fragments correspond to subtype B.

It is important to note the complexity of the genomic structures identified for the newly-identified CRF39_BF and CRF40_BF, which contrasts with the CRF28_BF and CRF29_BF previously described in Brazil mostly composed by B subtype, with only one and two F1 subtype fragments, respectively ^[3]. Furthermore, none of the breakpoints identified in CRF28_BF and CRF29_BF coincides with those described in our samples. Extensive differences in the genomic composition of the newly-described CRFs were also detected in relation to the CRF12_BF described in other South American countries ^[8]. Taken together, these results suggest a continuous complex interaction between subtypes B and F leading to new recombinant forms without any evidence of a common ancestor between CRF39_BF and CRF40_BF and the other CRFs_BF. Recent studies have also described the occurrence of a second generation of recombinants ^[9,10], confirming the rapid evolution of HIV as well as its adaptative capacity with new genomic compositions.

It is clear that distinct CRFs_BF co-circulate in the Brazilian HIV epidemic. National efforts are necessary to increase the number of HIV-1 full-length sequences from different Brazilian regions in the database, aiming to track the spread, evolution and the real distribution/prevalence of HIV-1 subtypes, URFs and CRFs in Brazil.

Acknowledgements

The authors would like to thank Dr Gonzalo Bello (FIOCRUZ) for helpful discussions and for critical reviewing of the manuscript. This study was partially supported by FIOCRUZ and Brazilian Ministry of Health.