In a recent issue of AIDS, the studies of Fontella et al.  and Bello et al.  argued that the South American HIV-1 subtype C epidemics originated from a single introduction of a virus population from an Eastern African country (Burundi or Kenya). This ancestral stock very likely entered the continent through the Southern region of Brazil, where the prevalence of HIV-1C is high. Lately, a letter by de Macedo Brigido  suggested that both studies ignored cultural and historical issues that could elucidate the geographic region from which South American HIV-1C originated. The author subsequently linked the entrance of HIV-1C in South America with the migration of Mozambican refugees to Brazil from 1975 to 1980, during the civil war period soon after the independence of Mozambique from Portugal.
The presence of HIV-1C in Mozambique was first detected by Engelbrecht et al. , who analyzed the V3 region of samples obtained in 1996. Recent surveys have shown that HIV-1C samples in Mozambique are genetically diversified and likely resulted from several independent introductions from neighboring countries over the course of the epidemics [5,6]. The hypothesis of a Mozambican origin of South American HIV-1C would require sequences from this country to have a close phylogenetic affinity with the South American clade.
The work of Parreira et al.  identified and characterized circulating HIV-1 subtypes in patients from the city of Beira, the second largest city in Mozambique and an important seaport. They analyzed samples from 23 different countries, and only sequences from Brazil, India and Israel were grouped in monophyletic clusters. Mozambican env sequences were widely dispersed across multiple clusters. More importantly, no Mozambican env sample was closely related to the Brazilian group. Lahuerta et al.  studied env and protease sequences from samples obtained between 1999 and 2004 from women living in a semirural area of southern Mozambique. Their phylogeny also revealed that Mozambican sequences did not form a country-specific cluster.
In order to further evaluate the Mozambican ancestry hypothesis, we have screened the Los Alamos National Laboratory HIV database for sequences from Mozambique and Portugal to append to the dataset of Fontella et al. . Portuguese sequences were included because this country possibly received Mozambicans during the civil war. Samples were selected following the same criterion used in our original dataset, that is, genetically related sequences were avoided and the time range of the sample was maximized. This strategy resulted in 10 Mozambican and four Portuguese sequences of HIV-1 pol gene. To broaden the investigation, we also added 23 new Brazilian sequences from Santa Catarina, the Brazilian state suggested by de Macedo Brigido  to be the likely place of entrance of HIV-1C in South America. A sequence from Burundi was also included for the sake of comparison with the study by Bello et al. .
We used the maximum likelihood algorithm implemented in PhyML 3.0 (www.atgc-montpellier.fr/phyml) to reconstruct a phylogenetic tree using the general time reversible + gamma + proportion of invariant sites (GTR + I + G) model. Statistical credibility of clades was assessed with the approximate likelihood-ratio test (aLRT) . The estimated phylogeny showed that Mozambican samples were spread all over the tree as previously reported for the env gene topologies (Fig. 1). Samples from India and from Portugal were clustered in two monophyletic groups, as well as the South American sequences, which were clustered with an aLRT support value of 90. Sequences from Santa Catarina were nested within the Brazilian cluster. As Fontella et al.  and Bello et al.  have shown, samples from Burundi and Kenya were closely related to the South American clade.
The hypothesis that the origin of the HIV-1C epidemic in Brazil is related to the immigration of Mozambican refugees has a very appealing historical background; it attempts to link the viral transmission dynamics to a cultural bond, namely, the Portuguese language. However, we could not find any molecular connection between South American and Mozambican HIV epidemics. Evidently, the identification of the exact geographical location of the ancestral population that originated South American HIV-1C is impractical. Hence, Fontella et al.  proposed an Eastern African origin instead of pointing out a specific country. Although Mozambique is also located in the Eastern Africa, the suggestion of an origin associated with cultural ties is not likely based on the molecular data available so far.
1. Fontella R, Soares MA, Schrago CG. On the origin of HIV-1 subtype C in South America. AIDS 2008; 22:2001–2011.
2. Bello G, Passaes CP, Guimarães ML, Lorete RS, Matos Almeida SE, Medeiros RM, et al
. Origin and evolutionary history of HIV-1 subtype C in Brazil. AIDS 2008; 22:1993–2000.
3. de Macedo Brigido LF. On the origin of South America HIV-1C epidemic. AIDS 2009; 23:543–544.
4. Engelbrecht S, Koulinska I, Smith TL, Barreto J, van Rensburg EJ. Variation in HIV type 1 V3 region env sequences from Mozambique. AIDS Res Hum Retroviruses 1998; 14:803–805.
5. Parreira R, Piedade J, Domingues A, Lobão D, Santos M, Venenno T, et al
. Genetic characterization of human immunodeficiency virus type 1 from Beira, Mozambique. Microbes Infect 2006; 8:2442–2451.
6. Lahuerta M, Aparicio E, Bardaji A, Marco S, Sacarlal J, Mandomando I, et al
. Rapid spread and genetic diversification of HIV type 1 subtype C in a rural area of Southern Mozambique. AIDS Res Hum Retroviruses 2008; 24:327–335.
7. Anisimova M, Gascuel O. Approximate likelihood-ratio test for branches: a fast, accurate, and powerful alternative. Syst Biol 2006; 55:539–552.