Analysis of the CCR5 gene coding region diversity in five South American populations reveals two new non-synonymous alleles in Amerindians and high CCR5*D32 frequency in Euro-Brazilians

Abstract

The CC chemokine receptor 5 (CCR5) molecule is an important co-receptor for HIV. The effect of the CCR5*D32 allele in susceptibility to HIV infection and AIDS disease is well known. Other alleles than CCR5*D32 have not been analysed before, neither in Amerindians nor in the majority of the populations all over the world. We investigated the distribution of the CCR5 coding region alleles in South Brazil and noticed a high CCR5*D32 frequency in the Euro-Brazilian population of the Paraná State (9.3%), which is the highest thus far reported for Latin America. The D32 frequency is even higher among the Euro-Brazilian Mennonites (14.2%). This allele is uncommon in Afro-Brazilians (2.0%), rare in the Guarani Amerindians (0.4%) and absent in the Kaingang Amerindians and the Oriental-Brazilians. R223Q is common in the Oriental-Brazilians (7.7%) and R60S in the Afro-Brazilians (5.0%). A29S and L55Q present an impaired response to β-chemokines and occurred in Afro- and Euro-Brazilians with cumulative frequencies of 4.4% and 2.7%, respectively. Two new non-synonymous alleles were found in Amerindians: C323F (g.3729G > T) in Guarani (1.4%) and Y68C (g.2964A > G) in Kaingang (10.3%). The functional characteristics of these alleles should be defined and considered in epidemiological investigations about HIV-1 infection and AIDS incidence in Amerindian populations.

Introduction

The human immunodeficiency virus type 1 (HIV-1) epidemic shows great variation among the different Brazilian regions. A progressive reduction in the number of deaths from acquired immunodeficiency syndrome (AIDS) was observed after the introduction of potent antiretroviral therapy in 1996, but the deceleration of the AIDS epidemic was not homogenous throughout all the Brazilian regions (Brito et al., 2005). The Southeast region has experienced the lowest increase in the AIDS epidemic from 1990 to 1996, contrasting with a steep rise in the North and South regions (Szwarcwald et al., 2000). Since 1996, the incidence rates of AIDS in Brazil as a whole and in the State of São Paulo in particular show a trend towards stability, whereas in the Brazilian Northeast the incidence rates of the disease continue to grow (Brito et al., 2005). The different spreading of the disease is due to multiple variables, including biological, behavioural, demographic and economic/political factors that influence the rate of contact between infected and susceptible individuals, as well as the individual's infectiousness and susceptibility. Among these factors are genetic variants of host genes that facilitate or hamper viral entry into the cells and modulate immune responses against the infection.

The chemokine (C-C motif) receptor 5 gene (CCR5) comprises three exons. The polypeptide of 352 amino acid residues is encoded by exon 3 (formerly named exon 4) (Mummidi et al., 1997). CCR5 transduces the signals of several different chemokines in phagocytes and T lymphocytes and serves as an essential co-receptor for the entry of R5-tropic HIV-1 into those cells (Blanpain et al., 2001). This is the viral form that most frequently infects people in Brazil (Ferraro et al., 2001). Therefore, CCR5 alleles that code for proteins poorly or not expressed at the cell surface are strong candidates for protection against the infection and for the delay of AIDS onset. This is the case of the truncated CCR5*D32 allele, and probably also of the Fs299 and R60S alleles (Dean et al., 1996; Shioda et al., 2001; Tamasauskas et al., 2001). CCR5*D32 was also favourably associated with autoimmune diseases such as multiple sclerosis, rheumatoid arthritis and type 1 diabetes mellitus, but increases the risk for abdominal aortic aneurysm and sarcoidosis (for a review, see Navratilova, 2006).

The interaction between the CCR5 receptor and its ligands can block HIV-1 entry and thus retard disease progression. The A29S and L55Q alleles encode products with a reduced affinity for (C-C motif) chemokines and might be associated with a shorter time interval from HIV infection to AIDS onset (Howard et al., 1999).

During AIDS, the acquisition of mutations in the HIV-1 gp120 envelope glycoprotein gene leads to the switch from primary R5 (CCR5-using) to highly cytopathic X4 (CXCR4-using) HIV-1 variants. According to the somatic hypermutation hypothesis, this switch takes place in the germinal center B cells, due to aberrant somatic hypermutation of the gp-120-coding region of the HIV-1 env gene (Suslov, 2004). This process seems to be more effective in CCR5*D32 heterozygotes, which were found at a 2.5 times higher risk of harbouring X4 HIV-1 variants before the onset of highly active antiretroviral therapy. The presence of X4 variants in the patients seems not to compromise the therapy outcome (Brumme et al., 2005), whereas the presence of a CCR5*D32 allele was found associated with a better response (Accetturi et al., 2000; Guerin et al., 2000).

In order to better understand the diversity of the CCR5 gene and to supply data for studies on the functional effect and epidemiological consequences of the CCR5 variants, we investigated the distribution of CCR5*D32 and other known exon 3 coding region CCR5 alleles in five populations of South Brazil. These alleles and their known functional characteristics are listed in Table 1. We also sequenced part of the coding region of the gene, in order to search for new variants.

Table 1.

Functional characteristics (in vitro) of previously known CCR5 alleles*.

DNA level	g.2846G > T	g.2925T > A	g.2941G > T	g.3315_3346del32	g.3429G > A	g.3654delC	g.3757C > T	g.3765C > T	g.3777A > T
SNP (NCBI) database	rs1800939	rs1799863	rs1800940	-	rs1800452	rs34962689	-	rs1800944	rs1800945
Protein level	p.Ala29Ser	p.Leu55Gln	p.Arg60Ser	p.Ser79_Thr88del	p.Arg223Gln	p.Phe299fs	silent	p.Ala335Val	p.Tyr339Phe
Common nomenclature	A29S	L55Q	R60S	D32	R223Q	Fs299	P332P	A335V	Y339F
Gene transcription							?
Membrane expression			X	X		X
HIV-1 infection			X	X		X
Response to chemokines	x	x		X		X

*(Blanpain et al., 2000; Howard et al., 1999; Zhao et al., 2005) x: impaired/ diminished; X: blocked or absent; ?: unknown.

The allele nomenclature at the DNA and protein levels follows guidelines of the Human Genome Variation Society (http://www.genomic.unimelb.edu.au/mdi/mutnomen/). For practical purposes, we used the common nomenclature adopted by most authors throughout this article (Ansari-Lari et al., 1997; Carrington et al., 1997; Zhao et al., 2005).

“Major allele” DNA and protein reference sequences were AF031237 and NP_000570, respectively. Numbers for alleles nominated at the DNA and protein levels obey the nucleotide and amino acid residue numbers in the reference sequences.

Materials and Methods

Samples

One hundred and seventy two Afro-Brazilians, 172 Euro-Brazilians, 18 Oriental-Brazilians, 115 Guarani (89 of which belong to the M'byá sub-group) and 160 Kaingang were investigated. All individuals were randomly selected and live in the State of Paraná, in South Brazil, with the exception of 26 Guarani belonging to the Kaiowá and Ñandeva subgroups which live in the State of Mato Grosso do Sul in Central-Western Brazil. For some CCR5 alleles, the number of individuals analysed was lower. The classification of individuals as Euro-Brazilians and Afro-Brazilians was based on morphological features. The Euro-Brazilians included 53 unrelated German-speaking individuals whose ancestors came from or joined Mennonite settlements in South Brazil. No HLA genotyping data was available for this subsample. Based on HLA allelic frequencies previously determined for all other population samples, an average European component of 34% and an average Amerindian component of 6% were estimated for the Afro-Brazilians. For the non-Mennonite Euro-Brazilians, the African and Amerindian components are approximately 9% and 5%, respectively (Braun-Prado et al., 2000; Probst CM, MSc Dissertation, Universidade Federal do Paraná, Curitiba, 2000; Probst et al., 2000). The average admixture values of the Guarani and Kaingang with the immigrants from Europe and Africa were estimated to be 4% and 7%, respectively (Petzl-Erler et al., 1993; Probst CM, MSc Dissertation, Universidade Federal do Paraná, Curitiba, 2000; Tsuneto et al., 2003). The gene flow between these two Amerindian groups is also low, being approximately 1.4% in Guarani and 0.5% in Kaingang (Petzl-Erler et al., 1993).

Typing method

DNA was extracted from peripheral blood cells using the standard phenol/chloroform/isoamyl alcohol or salting-out techniques. The coding region of exon 3 of the CCR5 gene was amplified by PCR as described previously (Boldt and Petzl-Erler, 2002). The product was applied on nylon membranes in the form of dot-blots and allowed to hybridize with sequence-specific oligonucleotide probes (SSOP, Table 2), according to the protocol of the XII International Histocompatibility Workshop (Fernandez-Viña and Bignon, 1997). Part of the coding region of exon 3 was additionally sequenced using the CCR5rev internal primer in 13 Guarani and 29 Kaingang, one Euro-Brazilian and five Oriental-Brazilian samples. These samples and 59 additional Guarani and 55 additional Kaingang samples were also sequenced using the CCR5for internal primer. One Guarani M'bya individual was genotyped only by sequencing. Sequencing reactions were performed with BigDye Terminator version 1.1 chemistry (Applied Biosystems, Foster City, CA). The sequences of the primers and probes are listed in Table 2.

Table 2.

CCR5 PCR primers and sequence-specific probes.

	Sequence 5'→ 3'	Variant
PCR primer CCR5m	TATGCACAGGGTGGAACAAG	——————————-
PCR primer CCR5jn	CACAACTCTGACTGGGTCAC	——————————-
Seq. primer CCR5for	AATGAGAAGAAGAGGCACAGGGCT	——————————-
Probe CCR5 9- CCR5 9+	AAGCAAATCGCAGCC AAGCAAATCTCAGCC	+ A29S
Probe CCR5 1- CCR5 1+	CTCATCCTGATAAAC CTCATCCAGATAAAC	+ L55Q
Probe CCR5 10- CCR5 10+	GCAAAAGGCTGAAGA GCAAAAGTCTGAAGA	+ R60S
Probe CCR5 2- CCR5 2+	CAGTATCAATTCTGG CCATACATTAAAGATAG	+ D32
Probe CCR5 3+ CCR5 3-	CTCTGTTTCGGTGTC CTCTGCTTCAGTGTC	+ R223Q
Probe CCR5 14- CCR5 14+	CATCTATGCCTTTGT TCATCTATGCTTTGT	+ Fs299
Probe CCR5 6- CCR5 6+	AGGCTCCCGAGCGAG AGGCTCCTGAGCGAG	+ P332P
Probe CCR5 7- CCR5 7+	GAGCGAGCAAGCTCA GAGCGAGTAAGCTCA	+ A335V
Probe CCR5 8- CCR5 8+	TCAGTTTACACCCGA TCAGTTTTCACCCGA	+ Y339F

PCR: polymerase chain reaction

Seq.: sequencing; +: major allele; in bold: variant nucleotides.

Statistical analysis

Genotype and allele frequencies were obtained by direct counting with the aid of the Convert program version 1.1 (Program distributed by the author, CM Probst). The Hardy-Weinberg equilibrium and population homogeneity hypotheses were tested using the approach of Guo and Thompson and the Raymond and Rousset test, respectively, in the ARLEQUIN software package version 3.1 (http://cmpg.unibe.ch/software/arlequin3) (Excoffier et al., 2005). p = 0.05 was adopted as the significance limit.

Results

The CCR5 genotype distributions met the Hardy-Weinberg equilibrium expectations in all populations. The frequency of the most common CCR5 allele varied from 88% to 100% (Table 3). Alleles Fs299 and P332P were not observed in the population samples studied. The other alleles were seen in at least one population, at frequencies varying from about 0.5% to 5% for most of them, except D32 and R223Q.

Table 3.

CCR5 allelic frequencies and standard deviations in various populations.

Population	A29S	L55Q	R60S	D32	R223Q	Fs299	P332P	A335V	Y339F
Afro-American1,2 n = 50	0.015 ± 0.015 (32)	0.008 ± 0.003 (332)	0.013 ± 0.013 (38)	0.019 ± 0.003 (1015)	0	0	0.01 ± 0.01	0.025 ± 0.007 (242)	0.026 ± 0.015 (58)
Afro-Brazilian n = 37	0.017 ± 0.017 (29)	0.027 ± 0.019	0.05 ± 0.05 (10)	0.02 ± 0.008 (172)	0	0	0	0.015 ± 0.007 (172)	0 (11)
Euro-American1,2 n = 50	nt	0.041 ± 0.008 (354)	nt	0.1 ± 0.004 (2605)	0.016 ± 0.016 (32)	0	0	0.006 ± 0.006 (87)	0 (121)
Euro-Brazilian n = 172	0.007 ± 0.007 (69)	0.02 ± 0.008	0 (156)	0.093 ± 0.016	0	0	0	0.003 ± 0.003	0.006 ± 0.004
Mennonites n = 53	0 (30)	0.038 ± 0.019	0 (39)	0.142 ± 0.034	0	0	0	0	0
Non-Mennonites n = 119	0.013 ± 0.013 (39)	0.013 ± 0.007	0 (117)	0.071 ± 0.017	0	0	0	0.004 ± 0.004	0.008 ± 0.006
Chinese3 n = 785	0	0	0	0	0.047 ± 0.005	0.005 ± 0.002	0	0	0
Oriental-American1,2 n = 100	nt	0	nt	0	0.04 ± 0.014	0.04 ± 0.014	0	0	0
Oriental-Brazilian n = 13	nt	0	0 (11)	0 (16)	0.077 ± 0.053	0 (12)	0 (12)	0 (18)	0 (12)
Guarani n = 27	0	0 (34)	0	0.004 ± 0.004 (115)	0 (34)	0	0	0.013 ± 0.008 (115)	0
Kaingang n = 31	0	0	0	0 (160)	0	0	0	0 (160)	0
Hispanic1 n = 50	nt	0.01 ± 0.01	nt	0.03 ± 0.012	0.02 ± 0.014	0	0	0	0

n: number of individuals; nt: not tested; in parenthesis: number of individuals if different from “n”.

¹(Ansari-Lari et al., 1997); ²(Carrington et al., 1997); ³(Zhao et al., 2005); shadowed in italics: this work.

Three D32 homozygotes were found among the Euro-Brazilians. The D32 heterozygote frequencies were 4.1% (7/172) in Afro-, 15.1% (26/172) in Euro-Brazilians, and 0.9% (1/115) in the Guarani Amerindians. We did not find the CCR5*D32 allele in Oriental-Brazilians nor in the Kaingang Amerindians. This allele was more frequent in the Euro-Brazilian sample (9.3%) than in any other sample previously investigated in Latin America (Table 4). The frequency of the D32 allele rose to 14.2% in a subsample of 53 German-speaking Euro-Brazilians, whose ancestors came from or joined Mennonite settlements in the past. Two of the three homozygotes and 15 of the 26 heterozygotes seen in the Euro-Brazilian sample belonged to this group. Nevertheless, there was no statistically significant difference between the frequency distribution of the CCR5 genotypes of the Mennonite and the non-Mennonite Euro-Brazilians investigated (p = 0.08, exact test of population differentiation).

Table 4.

D32 allelic frequencies and standard deviations in Latin American populations.

Population	n	D32 freq.	Region	Reference
Admixed Mexican	212	0.014 ± 0.006	————, MX	(Zuniga et al., 2003)
Amerindian Mayo	70	0
Amerindian Teenek	61	0
Amerindian Mazatecan	61	0.016 ± 0.011
Afro-Jamaican	242	0.01 ± 0.005	————, JM	(Hisada et al., 2002)
Colombian	150	0.027 ± 0.009	Medellin, CO	(Diaz et al., 2000)
Amerindian	172	0.009 ± 0.005	Arequipa, PE	(Calzada et al., 2001)
Amerindian Tikuna	191	0	Northwest Amazonas, BR	(Leboute et al., 1999)
Amerindian Baniwa	46	0
Amerindian Kashinawa	29	0	Southwest Amazonas, BR	(Leboute et al., 1999)
Amerindian Kanamari	34	0
Amerindian Tiriyó	180	0	North Amazonas, BR	(Grimaldi et al., 2002)
Amerindian Waiampi	221	0
Six Amerindian groups	89	0	North Pará, BR	(De Pinho Lott Carvalhaes et al., 2004)
Brazilian	394	0.03 ± 0.006
Afro-Brazilian	67	0.008 ± 0.008
Oriental-Brazilian	111	0
Brazilian	104	0.02 ± 0.01	Recife, Pernambuco, BR	(de Souza et al., 2006)
Admixed Brazilian	549	0.026 ± 0.005	Northeast Bahia, BR	(Grimaldi et al., 2002)
Afro-Brazilian	54	0.019 ± 0.013	Rio de Janeiro, BR	(Chies and Hutz, 2003)
Brazilian	115	0.056 ± 0.015	São Paulo, BR	(Munerato et al., 2003)
Brazilian	100	0.035 ± 0.013	Ribeirão Preto, São Paulo, BR	(Passos Jr and Picanço, 1998)
Euro-Brazilian	102	0.044 ± 0.014	Paraná, Santa Catarina and Rio Grande do Sul, BR	(Chies and Hutz, 2003)
Brazilian	127	0.055 ± 0.014		(Kaimen-Maciel et al., 2007)
Eight Amerindian groups	241	0.013 ± 0.005		(Hunemeier et al., 2005)
Afro-Brazilian	172	0.02 ± 0.008	Paraná, BR
Euro-Brazilian	172	0.093 ± 0.016
Oriental-Brazilian	16	0		This work.
Guarani	114	0.004 ± 0.004
Kaingang	160	0
Brazilian	100	0.05 ± 0.015	Londrina, Paraná, BR	(Brajão de Oliveira et al., 2007)
Euro-Brazilian	99	0.066 ± 0.018	Santa Catarina, BR	(Grimaldi et al., 2002)
Euro-Brazilian	59	0.068 ± 0.023	Alegrete, Rio Grande do Sul, BR	(Vargas et al., 2006)
Afro-Brazilian	13	0.038 ± 0.038
Admixed Brazilian	31	0.064 ± 0.032
Afro-Brazilians	58	0.009 ± 0.009	Rio Grande do Sul, BR	(Chies and Hutz, 2003)
Amerindian Chiriguano	42	0.012 ± 0.012	Northwest Argentina, AR	(Mangano et al., 2001)
Argentinean	751	0.03 ± 0.004	————, AR	(Gonzalez et al., 2001)
Chilean	62	0.024 ± 0.014	————, CL	(Desgranges et al., 2001)

n: number of individuals; freq.: frequency; ISO 3166-1 codes indicate countries.

Allele R223Q was observed in Oriental-Brazilians but not in the other population samples (Table 3). It occurred in the heterozygotic state in two of 13 Oriental-Brazilians (heterozygote frequency of 15.4%).

Sequencing analysis of the coding region of exon 3 revealed a new allele in the Guarani (g.3729G > T), causing the substitution of cysteine by phenylalanine at amino acid residue 323 (p.Cys323Phe) in the C-terminal intracellular segment of the protein. The p.Cys323Phe allele occurred in two heterozygotes out of the 72 Guarani individuals whose DNA was sequenced, which allowed estimating an allelic frequency of 1.4% in the Guarani population. The DNA carrying this variant was reamplified and resequenced to confirm the presence of this new allele. In the Kaingang, sequencing revealed another new allele (g.2964A > G) causing the substitution of tyrosine by cysteine at the conserved residue 68 (p.Tyr68Cys) in the second transmembrane part of the protein. This allele occurred in five heterozygotes out of 29 sequenced individuals, which allowed estimating a frequency of 10.3% in the Kaingang population. We also confirmed the presence of the R223Q allele in one heterozygote out of the 5 Oriental-Brazilians whose exon 3 was sequenced.

Discussion

This is the first study investigating the A29S and R60S alleles in European-derived populations. Also, alleles other than D32 have not been analysed before in Amerindians. Based on the CCR5 allelic frequencies in the Chinese, North- and South-American populations (Table 3), it is possible to infer that A29S,R60S,A335V and Y339F most likely originated in Africa; L55Q and D32 in Europe; R223Q and Fs299 in Asia. The P332P allele, found only once in one heterozygote Afro-American (Ansari-Lari et al., 1997), was not found in our population samples nor in screenings of about 700 Afro-Americans, 700 Euro-Americans and 785 Chinese (Ansari-Lari et al., 1997; Carrington et al., 1997; Zhao et al., 2005). The presence of the D32 allele in the Guarani seems to be the result of gene flow from Neo-Brazilians, as suggested for Mura and Kaingang in another study (Hunemeier et al., 2005).

The high D32 frequency in Euro-Brazilians is similar to the frequencies found in Central Europe (Stephens et al., 1998). It is compatible with the greater European component in the Euro-Brazilian population of the Paraná State, in comparison to other, previously analysed Brazilian populations of predominantly European ancestry (Probst et al., 2000). The D32 frequency in the Mennonite subsample is two times higher than in the non-Mennonite Euro-Brazilian subsample and equals the high D32 frequencies in North Europe (Stephens et al., 1998; Yudin et al., 1998). The frequency of L55Q, another allele with likely European origin, is three times higher in Mennonite compared to non-Mennonite Euro-Brazilians. The Mennonites have Friesian origin (North of Germany and the Netherlands) and exist as a religious Anabaptist group since the second half of the XVI century. The majority of individuals in this subsample are direct descendants from 200 Mennonite families that left their villages in the Ukraine and in Siberia and arrived in South Brazil in 1930 (Pauls Jr., 1976). Thus, a founder or bottleneck effect associated to random genetic drift most probably caused the rise in the D32 and L55Q allelic frequencies in this population.

The R223Q allele is the most frequent variant in the Chinese population. It is equally distributed in HIV-1 infected and non-infected Chinese groups and has similar HIV-1 co-receptor activity as the major CCR5 allele (Zhao et al., 2005). Other populations have thus far not been investigated. We also found this allele among the Oriental-Brazilians.

The cysteine residue we found mutated to phenylalanine at codon 323 (p.Cys323Phe) in two heterozygote Guarani individuals is not conserved in CCR2, the homologous C-C chemokine-receptor protein with the highest sequence similarity to CCR5 (75%). The substitution of the same residue by alanine was found to decrease the expression of the CCR5 protein on the cellular membrane by preventing receptor palmitoylation (Blanpain et al., 2001). A change in the secondary structure and function may also be expected from the replacement of this residue by phenylalanine. In the Kaingang, sequencing revealed another new allele (g.2964A > G) causing the substitution of tyrosine by cysteine at the otherwise conserved residue 68 (p.Tyr68Cys) in the second transmembrane part of the protein. This allele seems to be very common in the Kaingang population and restricted to it. Possible protective effects of both alleles regarding HIV-1 infection and progression to AIDS have to be established in appropriate cohorts attending Amerindian(-derived) populations.

In summary, we studied the distribution of the CCR5 coding region alleles in various Brazilian populations and noticed a high D32 frequency in the Euro-Brazilian population of the Paraná State in South Brazil. The D32 frequency is even higher among the Mennonites and is the highest thus far reported for Latin America. We also identified two new coding CCR5 mutations in the Amerindian populations, whose functional characteristics should be defined and considered in epidemiological investigations about HIV-1 infection and AIDS incidence in Amerindian populations.