Professional Documents
Culture Documents
www.elsevier.com/locate/meegid
Abstract
The human immune deficiency virus (HIV) exhibits strikingly tremendous amount of genetic variability. Such feature is critically
important for the virus to adapt to environmental changes by escaping the host immune system and by escaping candidate vaccine. Therefore,
understanding of such diversity is fundamental for the design of successful drugs or vaccine, which is urgently needed to bring the HIV/AIDS
epidemic under control. In this study, we investigated the magnitude of diversity of the HIV-1 near full-length genomes from
patients previously assigned as infected with non-recombinant HIV-1 subtypes B and F1 variants based on small portion of viral genome.
HIV-1 proviral DNA was extracted from 14 samples previously classified in our laboratory as six subtypes B and eight subtypes F on the basis
of small amplicon sequencing. Reamplifications of DNA from these samples were carried out by an overlapping PCR followed by
direct sequencing. The data were phylogenetically inferred. Sequence analysis revealed that two out of six partially identified subtype B and
six out of eight partially identified subtype F were in fact BF recombinants throughout their full genomes. Two pairs BF recombinants had
identical genomic recombination structure and distinct from the Argentinean CRF 12_BF strains, probably represents a novel circulating
recombinant forms in Brazil. Our data provided new genetic material of some of the HIV-1 subtypes currently circulating in the country
and points to the widespread of BF recombinants which are expected to change the epidemic nature by approaching the level of subtype B
in Brazil.
# 2006 Elsevier B.V. All rights reserved.
Table 1
Characteristics of patient samples included in this study
Sample Year of collection Sex Age Mode of transmission Subtype identification
Partial genome subtype Complete genome
region
01BR042 2001 M 38 Heterosexual F1 rt a BF1
01BR047 2001 M 38 Heterosexual F1 Protb, rt BF1
01BR087 2001 M 49 Bisexual F1 prot, rt and envc F1
01BR125 2001 F 39 Heterosexual F1 prot, rt and env F1
01BR226 2001 F 43 Heterosexual F1 prot, rt BF1
01BR323 2001 F 44 Heterosexual F1 prot, rt and env BF1
02BR002 2002 F 29 Heterosexual B prot, rt and env B
02BR005 2002 F 29 Heterosexual B prot, rt BF1
02BR006 2002 F 30 Heterosexual B prot, rt BF1
02BR008 2002 M 37 Homosexual B prot, rt B
02BR011 2002 M 27 Bisexual B prot, rt B
02BR013 2002 F 35 Heterosexual B prot, rt B
02BR033 2002 M 36 Bisexual F1 prot, rt BF1
02BR034 2002 M 31 Bisexual F1 prot, rt BF1
a
Reverse transcriptase.
b
Protease.
c
Envelope (V3–V5).
370 S. Sanabani et al. / Infection, Genetics and Evolution 6 (2006) 368–377
and 72 8C for 3 min; and a final extension step at 72 8C for weight matrix. Aligned sequences were manually edited and
10 min. The first round PCR products (2 ml aliquots) was then trimmed to the minimal shared length in the BioEdit Sequence
amplified with the nested primers under the same conditions. Alignment Editor Program (Hall, 1999).
Efficient PCR amplification was optimized in advance by The gap-stripped aligned sequences were screened for the
manipulating the concentration of MgCl2 and annealing presence of recombination patterns by the similarity plot and
temperatures using an Eppendorf gradient PCR cycler. The bootscan methods as implemented in the SIMPLOT program v
expected sizes of the amplified products were verified using 3.2 beta (Lole et al., 1999; Salminen et al., 1995). These
ethidium-bromide staining after agarose gel electrophoresis. methods identify putative positions of recombination by
Both DNA complementary strands were sequenced directly visually plotting percentage of permuted trees along the length
from purified PCR products in an overlapping fragment of 400 of each sequence. Sequences with suggestive recombination
nucleotides by using variety of internal sequence specific against reference panel by similarity plot were subsequently
primers, fluorescent-dye terminators, and Taq polymerase on an analyzed by bootscan. For both methods, nucleotide distances
automated sequencer (ABI 3100, Applied Biosystems Inc., were calculated in a sliding window of 420 bp moving in steps
Foster City, CA). The data from the sequenced fragments were of 20 bp by using the F84 model of evolution (maximum
edited, assembled into contiguous sequences, and a consensus likelihood) and transition/transversion ratio of 2.0. Bootstrap
of the both strands was formed by the Sequencher program values of 70% were considered definitive for assignment of
(Gene Code Corp., Ann Arbor, MI). All the sequences were parenthood. Exploratory trees of DNA segments between
checked for contamination by BLAST search against HIV-1 breakpoints were constructed.
sequence database and among themselves (Korber et al., 1995). Phylogenetic relationships were established by maximum
Full genome sequences were aligned with reference likelihood and neighbor joining methods with PAUP*4.0b10
sequences representing subtypes A–D, F–H, J and K obtained (Rogers and Swofford, 1998). Model-test (Posada and Crandall,
from the Los Alamos database (http://hiv-web.lanl.gov) using 1998) was used to determine the optimal nucleotide substitution
the CLUSTAL X program (Thompson et al., 1997) with the model for each sequence. The substitution model GTR + I + G
‘‘slow-accurate’’ default alignment parameters and IUB DNA was chosen as the best fitting evolutionary model for likelihood
Fig. 1. Phylogenetic relationship generated from nearly full-length HIV-1 sequences of six non-recombinant Brazilian isolates from the current study (Marked with
black circles and rhombus). Nucleotide sequences were compared with reference sequences of subtype A–D, F–H, J and K (http://hiv-web.lanl.gov). The tree was
constructed by maximum likelihood method under the GTR + I + G substitution model. Values at branches indicate bootstrap support based on 1000 iterations under
NJ method using ML distances and NNI swapping. The scale bar represents 0.1 nucleotide substitution per site.
S. Sanabani et al. / Infection, Genetics and Evolution 6 (2006) 368–377 371
analysis based on the Akaike information criterion parameters. Maximum likelihood trees were constructed of isolates that
After the starting tree was obtained by neighbor joining, did not show evidence of recombination by similarity plot
heuristic searches for likelihood was performed by using the analysis. Fig. 1 depicts the phylogenetic tree of HIV-1 of 44
NNI branch-swapping algorithm. Phylogenies were boot- sequences, including six isolates from the current study and 38
strapped with 1000 (neighbor-joining) replicates. A Kimura reference strains (GenBank and Los Alamos database)
2P model in MEGA version 2.1 was used to calculate the representing subtypes A–D, F–H, J and K. The computed
genetic distances between and within isolates. overall distance among these six isolates was 5.5–15.7% based
The nucleotide sequences from the 14 HIV-1 isolates have on 8374 nucleotides. Four isolates clustered with subtype B
been submitted to GenBank (accession no. DQ358799- reference sequences (100%; bootstrap), showing nucleotides
DQ358812). distances from 6.7% to 8.2%. Isolates 01BR087 and 01BR125
fell into subtype F1 reference group in 100% of the bootstrap
3. Results trees with distances of 6.0–8.7%.
Isolates were then grouped according to their references and
Reanalysis of HIV-1 subtypes were determined for 14 compared on the base of their genetic distances. The maximum
samples by near full-length genome sequencing. No evidence genetic variability among subtypes B and F1 viruses, including
of sample contamination was observed during the BLAST the F1 Brazilian reference strains were 8.4% and 6.0%,
search. The majority of open reading frames were intact and respectively. The mean variation between both subtypes ranged
opened in all isolates. All samples were previously character- from 14.3% to 15.6%. To gain detailed insight about clustering
ized by partial sequencing and showed six sequences to cluster pattern of our isolates and to define their genetic relation with a
with subtype B, eight with subtype F as shown in Table 1. Full corresponding reference sequences from various countries,
genome similarity plot of all isolates revealed that four of the other phylogenetic trees were constructed independently for
previously identified subtypes B were pure and two were BF each subtype.
recombinants (data not shown). Of the eight isolates previously Evolutionary related subtype B isolates and one sequence
identified as subtype F only two were pure strains throughout from Argentina formed a separate cluster supported by a weak
their genomes and the other six were BF recombinants. bootstrap value of 53% (Fig. 2A). Subtype B sequences from
Fig. 2. Phylogenetic relationship generated from non-recombinant B and F1 subtypes in a nearly full-length tree. (A) Rooted tree showing the geographic cluster of
the four Brazilian subtype B sequences and representative of subtype B. The SIV CPZGAB was used to root the tree. (B) Rooted tree showing the geographic cluster
of the two Brazilian subtype F1 isolates and representative of clade F. The HIV HXB2 was used as an outgroup. References sequences used in A and B, labelled
according to their sources, were obtained from the database (http://hiv-web.lanl.gov). The trees were constructed by maximum likelihood method under the
GTR + I + G substitution model. Values at branches indicate bootstrap support based on 1000 replications under NJ method using ML distances and NNI swapping.
The scale bar represents 0.1 nucleotide substitution per site.
372 S. Sanabani et al. / Infection, Genetics and Evolution 6 (2006) 368–377
Argentina and Colombia showed the highest inter-isolate between our isolates and the strains from Finland, Belgium and
distances with maximum values of 10.8%. Maximum like- France were 7.2%, 7.5% and 8.9%, respectively.
lihood tree of the near full length sequences of subtype F1 The stability of the cluster of our subtypes B and F1 isolates
identified in this study showed strong clustering (100% were verified by maximum likelihood trees, independently
bootstrap values) with the Brazilian subtype F1 reference made from the HIV-1 gag, pol and env regions using the same
strains. Genetic distances were calculated despite the existence set of reference strains. The results showed that the four subtype
of few number of full-length subtype F1 sequences in the B defined isolates did not form a single cluster in any regions as
databases. The results showed that the mean genetic distances they did in the full genome analysis (data not shown). Unlike
Fig. 3. Phylogenetic relationship of subtype F1 full-length gag (A), pol (B) and env (C) reading frames. The trees were constructed by maximum likelihood method
under the GTR + I + G substitution model. Values at branches indicate bootstrap support based on 1000 replications under NJ method using ML distances and NNI
swapping. Only bootstraps 70% are shown. Trees were rooted using HIV HXB2. The two Brazilian isolates are identified by circle. Branch lengths are drawn to
scale.
S. Sanabani et al. / Infection, Genetics and Evolution 6 (2006) 368–377 373
clade B, the ability of subtype F1 isolates to form a distinct against a panel of reference sequences of known subtypes by
cluster remained unaffected in all three regions and showed bootscan analysis as shown in Fig. 4. Two isolates, 02BR005
results consistent with those obtained by the full-length tree and 02BR006 were subtype B throughout the majority of their
(Fig. 3). genomes. Isolate 02BR005 had small fragment of subtype F
Data from similarity plot analysis have suggested that at the 30 end of gp 120 and beginning of the 50 end of gp 41
eight isolates, previously assigned either as subtype B or F1, while 02BR006 showed two small fragments of subtype B
were recombinant forms between subtypes B and F1. To fully (Fig. 5). All recombinant isolates, except 02BR005 and
explore this mosaic structure and to accurately identify the 01BR042, shared a common crossover site located between
positions of crossover sites, all sequences were scanned positions 506–551 nt downstream the gag ATG initiation
Fig. 4. Bootscan analysis of the near full length genomes of the eight BF recombinant viruses compared to representatives of HIV-1 subtype reference sequences.
Distances were computed after stripping gaps using F84 (ML) model of substitution. Sliding window of 420 bp moving in steps of 20 nt and 100 replicates of the
dataset were used.
374 S. Sanabani et al. / Infection, Genetics and Evolution 6 (2006) 368–377
Fig. 5. Schematic representation of the near full length genomic structure and breakpoints profile of eight BF1 recombinant isolates identified in this study. The region
of subtypes B and F1 are indicated at the bottom. Positions of breakpoints are marked with grey arrowhead and numbered according to the HXB2 sequence.
codon which corresponds to the N terminal end of p24gag As depicted in Figs. 4 and 5, the two isolates were comprised
protein. The same isolates, except 02BR033 and 02BR006, of subtype F1 throughout the majority of their genome with
shared other crossover points at positions 330 relative to the three dispersed stretches of subtype B. The exploratory tree
nef start codon. Strains 01BR047 and 01BR226 showed analysis of the seven fragments, delimited by crossover sites,
substantial homology in their mosaic patterns and shared four in isolates 01BR323 and 02BR034 showed them to cluster
crossover points at nearly identical positions as shown in closer to each other with statistical values of more than 70%
Figs. 4 and 5. Maximum likelihood trees of the individual in most of the fragments (Fig. 7). Weak clustering of both
fragments between breakpoints of the later isolates showed isolates was observed in the second fragment (592–1022;
them to cluster with significant bootstrap support to the alignment positions), probably due to few informative sites in
relevant subtypes and positioned closer to each other in some this region.
of the fragments but with low bootstrap values (<70%) In attempt to find sequences with recombinant structure
(Fig. 6). Based on the similarities of the apparent similar to any of our isolates, multiple sequence alignments
recombination pattern and phylogenetic results of subge- including all BF1 published isolates and CRF12_BF were built.
nomic regions, we believed that isolates 01BR047 and The alignment was subjected to bootscan analysis to define and
01BR226 are descendents of the same recombinant strain. compare the breakpoints for each sequence. Using this
Other striking similarity in the recombination profile was approach, we have not found any sequences to share identical
observed between strain 01BR323 and 02BR034, which recombination structure or breakpoints to any of our
exhibited three out of four breakpoints at identical positions. recombinant strains.
S. Sanabani et al. / Infection, Genetics and Evolution 6 (2006) 368–377 375
Fig. 6. Exploratory tree analysis based on fragments between breakpoints of isolates 01BR047 and 01BR226. ML trees were constructed from regions between
recombination breakpoints as indicated by bootscan plot. The best-fitted evolutionary model was determined for each fragment. Values at the respective nodes
indicate bootstrap support based on 1000 replications under NJ method using ML distances and NNI swapping. The O.BE.87.ANT70 isolate was used to root the tree
but not shown for simplicity. The scale bar represents 0.1 nucleotide substitution per site.
Fig. 7. Exploratory tree analysis based on fragments between breakpoints of isolates 01BR323 and 02BR034. ML trees were constructed from regions between
recombination breakpoints as indicated by bootscan plot. The best-fitted evolutionary model was determined for each fragment. Values at the respective nodes
indicate bootstrap support based on 1000 replications under NJ method using ML distances and NNI swapping. The O.BE.87.ANT70 isolate was used to root the tree
but not shown for simplicity. The scale bar represents 0.1 nucleotide substitution per site.
circulating in Brazil, but in low proportion. This may be genomes. Analysis of individual fragments from these isolates
attributed to the replacement of genuine subtype F1 by showed them to cluster to each other in different branches
emerging BF1 recombinants as a result of unknown selective regardless of the non-significant bootstrap support, suggesting
advantage. that they are derived from common recombinant progenitor.
All sequenced recombinant genomes were BF1 mosaics. The other pair includes isolates 01BR323 and 02BR034, which
Neither of them resembled CRF12_BF suggesting that they exhibited almost identical recombination profile and strong
arose either from unrelated ancestors or independent recombi- monophyletic clustering of all but one distinct regions
nation events and indicating that the Brazilian epidemic may suggesting that they might have arisen from a common
harbor its own recombinant forms. ancestor. Moreover, the background history of infection of all
Although only eight recombinant forms were fully four patients did not show any direct epidemiological link.
sequenced, we were able to identify a full range of BF1 Although genomes in both pairs had a very similar genome
mosaics from genomes composed mostly by subtype B to structure, some coincidental breakpoints did not fall exactly at
others in which subtype F1 genetic material was predominant. the same place. We believe that this is a result of punctual
This large sequence diversity present in a limited sample size mutations removing subtype specific signature nucleotides and
suggests that HIV-1 intersubtype recombination is happening thus causing a dislocation in breakpoint positions in the
very frequently in Brazil probably due to HIV-1 transmission bootscan analysis. Other factor that affects the precise
networks operating in the country. However, we were still able localization of breakpoints is the post-recombination evolution
to identify shared breakpoints in these structures, which may in the recombinant strains.
either reflect them as hot spots for recombination or that viruses In this report, we provided evidences to indicate that the two
shared indeed in some extent a common ancestry. pairs of 01BR047/02BR226 and 01BR034/02BR034 are
Analysis of recombinant strains showed particular mosaic representative of two novel CRFs, however, because they did
pattern in two pairs of samples. The first pair consists of not meet one of the formal criteria required for assigning them
isolates 01BR047 and 01BR226, which demonstrated sig- as new CRFs, their formal classification will have to wait
nificant similarity to one another on their virtual complete sequencing of another virus with similar genomic structure.
S. Sanabani et al. / Infection, Genetics and Evolution 6 (2006) 368–377 377
We were able to fully sequence 14 HIV-1 genomes in Brazil. Hall, T.A., 1999. BioEdit: a user-friendly biological sequence alignment editor
and analysis program for Windows 95/98/NT. Nucleic Acids Symp. Ser. 41,
Since there is a lack of full-length genome data in the HIV-1
95–98.
epidemic in our country, we believe that these results will Hierholzer, J., Montano, S., Hoelscher, M., Negrete, M., Hierholzer, M., Avila,
contribute for a better understanding of this epidemic. M.M., Carrillo, M.G., Russi, J.C., Vinoles, J., Alava, A., Acosta, M.E.,
Furthermore, only 6 out of 14 were non-recombinant viruses Gianella, A., Andrade, R., Sanchez, J.L., Carrion, G., Russell, K., Robb, M.,
indicating a high recombination process that continue to Birx, D., McCutchan, F., Carr, J.K., 2002. Molecular epidemiology of HIV
generate BF1 recombinants. Therefore, we assumed that BF1 type 1 in Ecuador, Peru, Bolivia, Uruguay, and Argentina. AIDS Res. Hum.
Retrovirus. 18, 1339–1350.
recombinant in Brazil is the second most common clade after Korber, B.T., Learn, G., Mullins, J.I., Hahn, B.H., Wolinsky, S., 1995. Protect-
subtype B and expected to change the epidemic nature in this ing HIV databases. Nature 378, 242–244.
country. It is noteworthy that only two subtype F genomes were Lole, K.S., Bollinger, R.C., Paranjape, R.S., Gadkari, D., Kulkarni, S.S., Novak,
non-recombinant indicating that subtype F is less represented in N.G., Ingersoll, R., Sheppard, H.W., Ray, S.C., 1999. Full-length human
immunodeficiency virus type 1 genomes from subtype C-infected serocon-
Brazil than previously thought.
verters in India, with evidence of intersubtype recombination. J. Virol. 73,
152–160.
Acknowledgement Posada, D., Crandall, K.A., 1998. Model-test: testing the model of DNA
substitution. Bioinformatics 14, 817–818.
This work was supported by grant 0110719-5 from the Robertson, D.L., Sharp, P.M., McCutchan, F.E., Hahn, B.H., 1995. Recombina-
Fundação de Amparo a Pesquisa do Estado de São Paulo tion in HIV-1. Nature 374, 124–126.
Robertson, D.L., Anderson, J.P., Bradac, J.A., Carr, J.K., Foley, B., Funkhouser,
(FAPESP). R.K., Gao, F., Hahn, B.H., Kalish, M.L., Kuiken, C., Learn, G.H., Leitner,
T., McCutchan, F., Osmanov, S., Peeters, M., Pieniazek, D., Salminen, M.,
References Sharp, P.M., Wolinsky, S., Korber, B., 2000. HIV-1 nomenclature proposal.
Science 288, 55–56.
Bongertz, V., Bou-Habib, D.C., Brigido, L.F., Caseiro, M., Chequer, P.J., Couto- Rogers, J.S., Swofford, D.L., 1998. A fast method for approximating maximum
Fernandez, J.C., Ferreira, P.C., Galvao-Castro, B., Greco, D., Guimaraes, likelihoods of phylogenetic trees from nucleotide sequences. Syst. Biol. 47,
M.L., Linhares de Carvalho, M.I., Morgado, M.G., Oliveira, C.A., Osma- 77–89.
nov, S., Ramos, C.A., Rossini, M., Sabino, E., Tanuri, A., Ueda, M., 2000. Sa Filho, D.J., Sanabani, S., Diaz, R.S., Munerato, P., Brunstein, A., Fusuma, E.,
HIV-1 diversity in Brazil: genetic, biologic, and immunologic characteriza- Sabino, E.C., Janini, L.M., 2005. Analysis of full-length human immuno-
tion of HIV-1 strains in three potential HIV vaccine evaluation sites. deficiency virus type 1 genome reveals a variable spectrum of subtypes B
Brazilian network for HIV isolation and characterization. J. Acquir. Immune and F recombinants in Sao Paulo, Brazil. AIDS Res. Hum. Retrovirus. 21,
Defic. Syndr. 23, 184–193. 145–151.
Brindeiro, R., Vanderborght, B., Caride, E., Correa, L., Oravec, R.M., Berro, O., Sabino, E.C., Shpaer, E.G., Morgado, M.G., Korber, B.T., Diaz, R.S., Bongertz,
Stuyver, L., Tanuri, A., 1999. Sequence diversity of the reverse transcriptase V., Cavalcante, S., Galvao-Castro, B., Mullins, J.I., Mayer, A., 1994.
of human immunodeficiency virus type 1 from untreated Brazilian indivi- Identification of human immunodeficiency virus type 1 envelope genes
duals. Antimicrob. Agents Chemother. 43, 1674–1680. recombinant between subtypes B and F in two epidemiologically linked
Caride, E., Brindeiro, R., Hertogs, K., Larder, B., Dehertogh, P., Machado, E., individuals from Brazil. J. Virol. 68, 6340–6346.
de Sa, C.A., Eyer-Silva, W.A., Sion, F.S., Passioni, L.F., Menezes, J.A., Salminen, M.O., Carr, J.K., Burke, D.S., McCutchan, F.E., 1995. Identification
Calazans, A.R., Tanuri, A., 2000. Drug-resistant reverse transcriptase of breakpoints in intergenotypic recombinants of HIV type 1 by boot-
genotyping and phenotyping of B and non-B subtypes (F and A) of human scanning. AIDS Res. Hum. Retrovirus. 11, 1423–1425.
immunodeficiency virus type I found in Brazilian patients failing HAART. Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgins, D.G.,
Virology 275, 107–115. 1997. The CLUSTAL_X windows interface: flexible strategies for multiple
Carr, J.K., Salminen, M.O., Albert, J., Sanders-Buell, E., Gotte, D., Birx, D.L., sequence alignment aided by quality analysis tools. Nucleic Acids Res. 25,
McCutchan, F.E., 1998. Full genome sequences of human immunodefi- 4876–4882.
ciency virus type 1 subtypes G and A/G intersubtype recombinants. Thomson, M.M., Delgado, E., Herrero, I., Villahermosa, M.L., Vazquez-de
Virology 247, 22–31. Parga, E., Cuevas, M.T., Carmona, R., Medrano, L., Perez-Alvarez, L.,
Carr, J.K., Avila, M., Gomez Carrillo, M., Salomon, H., Hierholzer, J., Cuevas, L., Najera, R., 2002. Diversity of mosaic structures and common
Watanaveeradej, V., Pando, M.A., Negrete, M., Russell, K.L., Sanchez, ancestry of human immunodeficiency virus type 1 BF intersubtype recom-
J., Birx, D.L., Andrade, R., Vinoles, J., McCutchan, F.E., 2001. Diverse BF binant viruses from Argentina revealed by analysis of near full-length
recombinants have spread widely since the introduction of HIV-1 into South genome sequences. J. Gen. Virol. 83, 107–119.
America. AIDS 15, F41–F47. Thomson, M.M., Sierra, M., Tanuri, A., May, S., Casado, G., Manjon, N.,
Cornelissen, M., Kampinga, G., Zorgdrager, F., Goudsmit, J., 1996. Human Najera, R., 2004. Analysis of near full-length genome sequences of HIV
immunodeficiency virus type 1 subtypes defined by env show high fre- type 1 BF intersubtype recombinant viruses from Brazil reveals their
quency of recombinant gag genes. The UNAIDS network for HIV isolation independent origins and their lack of relationship to CRF12_BF. AIDS
and characterization. J. Virol. 70, 8209–8212. Res. Hum. Retrovirus. 20, 1126–1133.