CRF22_01A1 is involved in the emergence of new HIV-1 recombinants in Cameroon

Abstract

Cameroon is a West African country where high genetic diversity of HIV-1 has been reported. The predominant CRF02_AG is involved in the emergence of more complex intersubtype recombinants. In this study, we sequenced the full-length genome of a novel unique recombinant form (URF) of HIV-1, 02CAMLT04 isolated in blood donors in urban Cameroon. Phylogenetic tree and bootscan analysis showed that 02CAMLT04 was complex and appeared to be a secondary recombinant derived from CRF02_AG and CRF22_01A1. The genomic composition of 02CAMLT04 strain showed that it is composed of three segments; twenty four percent of the genome is classified as CRF02_AG, spanning most of the envelope gene. The remaining seventy six percent of the genome is classified as CRF22_01A1. In addition, the sequence analysis of 13 full-length sequences from HIV-1 positive specimens received from Cameroon between 2002 and 2010 indicated that five specimens are pure CRF22_01A1 viruses, and six others have homology with CRF22_01A1 sequences in either gag, pol or env region where as 6% of strains contain portions of CRF22_01A1. Further study demonstrated that CRF22_01A1 is a primary prevalence strain co-circulating in Cameroon and is involved in complex intersubtype recombination events with subtypes (D or F), subsubtypes (A1 or F2) and CRFs (CRF01_AE or CRF02_AG). Our studies show that novel recombinants between CRF22_01A1 and other clades and recombinant forms may be emerging in Cameroon that could contribute to the future global diversity of HIV-1 in this region and world wide.

INTRODUCTION

The genetic diversity of HIV-1 is very broad in Cameroon where all group M clades and several circulating recombinant forms CRFs (in particular CRF01_AE, CRF02_AG, CRF06_cpx, CRF09_cpx, CRF11_cpx, CRF13_cpx, CRF22_01A1, CRF36_cpx and CRF37_cpx), in addition to groups O, N and P viruses, have been identified^1–10. Among the broad genetic diversity of HIV-1 strains, CRF02_AG represents over 65% of HIV infections with an additional 26% classified as unique recombinant forms (URFs)^{8, 10}. Co-circulation of different subtypes and CRFs in Cameroon results in the continuing emergence of new intersubtype recombinants or even M/O recombinants¹¹. Therefore, Cameroon is an ideal area to investigate the genetic diversity of HIV-1 and its impact on the global pandemic. In 2002, we initiated a molecular and serologic epidemiology survey in blood donors in Douala and Yaoundé, in Cameroon to investigate HIV-1 genetic diversity, study virologic and immunologic characteristics of the viruses, evaluate the performance of US FDA licensed HIV-1 assays in the presence of numerous HIV-1 variants, and identify samples to serve as candidate reference reagents for diagnostics. We found that several group M subtypes (A, B, C, D, F₂ and G) and CRFs (CRF02_AG, CRF06_cpx, CRF11_cpx, CRF13_cpx, CRF19_cpx and CRF22_01A1) were circulating in this population. A new URF strain, 02CAMLT04, had CRF22_01A1^gag-CRF02_AG^env genotype¹², and a novel strain, 02CAMLT72 isolated in Douala, was designated as pure CRF22_01A1¹³.

CRF22_01A1 strain was initially identified in Cameroon in 2001⁴ and circulating in this country for many years. Brennan’s studies reported that the percentage of CRF22_01A1 in concordant specimens was 6.6% (1996–1999), 4.7% (2000–2002), and 6.6% (2003–2004) over the 9-year period¹⁴, however, molecular, epidemiological and evolutionary study of CRF22_01A1 phylogenetic association with different HIV-1 subtypes was not described. Here, we report 14 full-length genomic sequences of HIV-1 specimens isolated in Cameroon from 2002–2003 and 2006–2010. Of these 14 specimens, 5 were pure CRF22_01A1, 6 were determined to be recombinants of CRF22_01A1 with CRF02_AG. These new data support involvement of CRF22_01A1 in the emerging diversity of HIV-1 in Cameroon.

METHODS

Specimens

For HIV characterization studies, 441 blood samples from HIV-1-seropositive individuals were collected at Douala (LT), Bamenda (LPH), Buea (BDHS), Limbe (LB), Yaounde (ARC) and a few villages in Cameroon (NYU) in 2002–2010. A total of 208 viruses were isolated and genotyped based on partial sequence analysis, 63% of them were CRF02_AG, 26% were URFs, and 6% were CRF22_01A1 containing recombinants^12,15. In this study, 13 viruses containing CRF22_01A1 genome and 3 URF strains were selected for full-length sequence analysis, however nucleic acid was available for only 14 of them which were completely sequenced and characterized.

RNA Extraction, RT-PCR and Nearly Full-Length Genomic Sequencing

Viral RNA was extracted from the plasma samples using QIAamp Viral RNA Mini kit (Qiagen, Valencia, CA). Then, reverse transcription, polymerase chain reaction (PCR) and sequencing were performed as previously described^{6, 7}. The nearly full-length genomic sequences reported in this study have been deposited in the GenBank database (accession numbers: EU743964, JN864047-JN864059).

Phylogenetic and Recombination Sequence Analysis

Nucleotide sequences were aligned with HIV-1 reference strains of different subtypes, subsubtypes and CRFs from Los Alamos HIV database (A1_SE8538, A1_PS1044, A1_92UG037, A1_Q23 17, A1SE7535, A1_92RW008; A2_97CDKTB48, A2_94CY017; B_HXB2 LAI IIIB, B_671_00T36; C_BR025, C_ETH2220; D_ELI, D_94UG114; F1_VI850, F1_93BR020_1; F2_MP255, F2_02CM.0016BBY; G_DRCBL, G_92NG083; H_VI997, H_VI991; J_SE7887, J_SE7022; K_EQTB11C, K_MP535; 01_90CF4071, 01_90CF11697, 01_93TH051, 01_93TH253, 01_CM240; 02_pBD6_15, 02_IBNG)¹⁶ using Clustal W method¹⁷ as implemented in Vector NTI. The phylogenetic reconstructions were also performed with MEGA 4 software package using the neighbor-joining method¹⁸. To analyze the recombinant structure of the new viruses, the SimPlot 3.5.1 software was used to determine the percentage of similarity between selected pairs of sequences and to calculate bootscan plots¹⁹. The regions that did not cluster with any of the known subtypes were submitted to BLAST analysis (http://hiv-web.lanl.gov/BASIC-BLAST) to find the closely related sequences of other HIV-1 strains. The viral nucleotide positions correspond to the reference strain HXB2 (GenBank Accession Number K03455) used throughout this manuscript.

RESULTS

Near full-length genome sequencing was completed for 14 viruses of interest. One virus, NYU488, was collected from Bapile located in the East Province of Cameroon, and thirteen viruses were collected from the Southwest Province of Cameroon (Figure 1A). Phylogenetic analysis of these sequences confirmed their relatedness, as they all clustered together in the CRF22_01A1, CRF02_AG or subtype G radiation with a high bootstrap value (BV). As shown in Figure 2, six strains (02CAMLT04, LPH27MF, LB005, LB011, LB013 and LB054) were deeply clustered in the CRF22_01A1 radiation (BV, 100%) revealing a close relationship, 3 strains (LB045, NYU488 and ARC087) clustered between CRF22_01A1 and CRF01_AE suggesting they are CRF22_01A1 recombinants. 4 strains (ARC007, LB052, BDSH129 and LT66) clustered with CRF02_AG (BV, 93%) and LT31 clustered with subtype G (BV, 100%). These sequences were further analyzed by bootscanning to determine the phylogenetic mosaic structure. The 02CAMLT04 sequence was plotted against references restricted to the subsubtype A1, F2, subtype B, C, H, K, CRF22_01A1 and CRF02_AG. The genome of 02CAMLT04 was divided into three segments at the two breakpoints of nt 6951 and nt 9002, and the neighbor-joining trees were further built for each segment (Figure 3). The first segment extended from the gag to the beginning of the env gene (nt 789–6951), and clustered with CRF22_01A1 (BV, 100%). The second segment representing most of the env gene (nt 6952–9002) clustered with CRF02_AG radiation (BV, 100%) and the rest of the segments covering the nef gene (nt 9003–9408) clustered with CRF22_01A1. A significant portion (76%) of the 02CAMLT04 genome was in perfect alignment with CRF22_01A1 reference sequences except that most of the env gene clustered with CRF02_AG. These results demonstrated that the 02CAMLT04 virus was generated by the recombination of CRF22_01A1 and CRF02_AG. To our knowledge, this is the first report of a novel URF of CRF22_01A1 recombinant with CRF02_AG in full-length sequence.

Figure 1. CRF22_01A1 distribution in Cameroon (A) and Africa (B).

Approximate locations of the rural villages or cities from which the specimens were obtained are named. Circle indicates the specimens in Cameroon; light blue circle indicates CRF22_01A1 reference strains described in previous study¹³, black circle indicates new CRF22_01A1 containing strains characterized in this study. Specimens isolated in Africa and Saudi Arabia are indicated as red symbol.

Figure 2. Phylogentic analysis of the nearly full-length genome sequence.

Analysis was performed using the neighbor-joining methods with Kimura’s two-parameter method and bootstrap analysis (1000 replicates). The reference subsubtypes and CRFs were used to construct the tree, the five strains 02CAMLT72, 01CM.1867LE, 01CM.0001BBY, 02CM.3097MN and 02CM.1917LE were used as CRF22_01A1 references, some references have been omitted for clarity. Bootstrap values above 70% are shown, numbers at the nodes of the tree represent maximum parsimony bootstrap values. The scale bar represents 1% genetic distance. The new identified HIV-1 strains are indicated as red “●”, and blast searched HIV-1 strains are indicated as blue “Δ”.

Figure 3. Recombination in full-length of 02CAMLT04 sequence.

Bootscan plot were performed using SimPolt 3.5.1 software configured with 500 bootstrap replicates, 500 bp window, and a step size of 50 bp. The x-axis shows all 8771 aligned nucleotides of the sequence analyzed, and the y-axis shows the bootstrap value. The 02CAMLT04 was plotted against subtypes/CRF02_AG and a group of CRF22_01A1 references including 02CAMLT72, 02CM.1867LE, 01CM.0001BBY, 02CM.3097MN and 02CM.1917LE as indicated at the top panel. Breakpoints are indicated above the plot. Regions of the sequence alignment were extracted according to the indicated breakpoints. The reference subsubtypes and CRFs were fetched from Los Alamos HIV Sequence Database and initially used to construct the trees, some references have been omitted for clarity. Each segment was analyzed separately by neighbor-joining with the bootstrap above 70%. Representative trees are illustrated for regions a through e, Loci of genome segments are based on the HXB2 numbering engine. Scale bars represent 1–2% distance.

Bootscan analysis of these viruses revealed the presence of CRF22_01A1 homology, the overall structures of these CRF22_01A1 containing viruses are diagrammed in Figure 4. Five sequences (LPH27MF, LB005, LB011, LB013 and LB054) clustered phylogenetically with CRF22_01A1 references throughout the entire genome demonstrating that these viruses share the same genomic mosaic and can be designated as pure CRF22_01A1. Two viruses (ARC087 and NYU488) shared identical recombinant breakpoints and were largely CRF22_01A1 (80% of whole genome spanning mostly pol to nef genes) and 20% for CRF02_AG. LB045 virus was divided into 4 segments with a complex CRF02/CRF22/CRF02/CRF22 recombinant, about 65% of the LB045 genome was CRF22_01A1. BDSH129 virus was almost entirely CRF02_AG (84%) with a small piece of CRF22_01A1 (16%) spanning the gag gene and the beginning of the pol gene. LB052, a triple CRF02/CRF22/CRF02 recombinant, was also predominantly CRF02_AG (76%) with a small portion of CRF22_01A1 fragment spanning the most of env gene from the beginning. ARC007 virus was a subtype B/CRF02 recombinant strain. LT66 and LT31 viruses were a complex F2/CRF02/F2/CRF02/F2 and CRF01/G/CRF02/G recombinant, respectively (data not shown).

Figure 4. Diagram of the genomic structure of CRF22_01A1 containing recombinants.

HBX2 genomic regions are indicated at the top of the plot as reference structure. Breakpoint locations are based on the HXB2 number engine. Each of the resulting genomes was analyzed by SimPlot and Genotyping separately; recombinant structures of each strain were determined and listed. Red bar indicates the CRF22_01A1 regions, open bar indicates subtype D, F, subsubtype A1, F2, CRF01_AE (01) and CRF02_AG (02) as indicated.

To further understand the involvement of CRF22_01A1 in genetic recombination, the blast search of HIV sequence database against 6162 nt of 02CAMLT04 segment (nt 789–6951) yielded four high search score HIV-1 strains 01CM.0130NY, 01CM.1152NG, 02CM.3163MN and 01CM.4008HAN which were originally identified in Cameroon and classified as 01AF2U, A1U, AF2 and 01DU, respectively²⁰. Phylogenetic analysis showed that 01CM.0130NY virus clustered with high confidence to CRF22_01A1 radiation (BV, 100%, Figure 2), 02CM.3163MN and 01CN.1152NG clustered with other CRF22_01A1 containing recombinant(BV, 85%). 01CM.4008HAN was a unique strain clustered between CRF01_AE and subsubtype A1. Further bootscanning analysis showed that they were recombinants of CRF22_01A1 with HIV-1 subtypes F, D, subsubtype A1, F2 or CRFs, all of these strains had a unique mosaic pattern combining CRF22_01A1 (Figure 4). The detailed phylogenetic composition of CRF22_01A1 containing recombinants described above is listed in table (Supplemental Content).

DISCUSSION

We identified five pure CRF22_01A1 and six CRF22_01A1 containing URFs strains in Cameroon. We also reclassified four CRF22_01A1 containing recombinants from the HIV Sequence GenBank (Figure 4). Nine unique recombination patterns were identified; breakpoints were found throughout the genome, the most definitive breakpoints were at the beginning of the pol, accessory or beginning of the env region. CRF22_01A1 fragments may be identified in any region across the HIV whole genome. Table 1 summarizes the subtype/CRF assignments of recombinants containing CRF22_01A1 fragments evaluated in samples collected from 2000–2010. Since the identification of CRF22_01A1 in Cameroon, it may be emerging as a dominant HIV-1 variant in this population that could act as a parental subtype for recombination events. The emergence of more complex intersubtype recombinants containing fragments of CRF22_01A1, such as CRF36_cpx, have been reported in Cameroon⁷. Reanalysis of the CRF36 genome including CRF22_01A1 reference sequences showed that the CRF36_cpx contains the fragment of CRF22_01A1, CRF01_AE and CRF02_AG. Brennan et al found that 22.4% of HIV-1 URFs were CRF22_01A1 in the blood donor samples collected in 1996–2004 in Cameroon¹⁴. Our recent study of samples collected from 2006 to 2008 in Cameroon also identified three CRF02^gag/CRF22^pol/CRF22^env strains^12,13. These studies highlight the complexity of recombinants between CRF22_01A1 and CRF02_AG in Cameroon (Table 1). CRF02_AG appeared to be the dominant virus in the population and accounted for 60–68% of HIV-1 infections with an additional 26% classified as URFs^{2, 21–23}. This high degree of recombination could be due to preferential founder effects, and the subsequent introduction of the other genetic form, such as CRF22_01A1 or CRF02_AG, frequently resulting in recombinants containing fragments of CRF22_01A1 or CRF02_AG.

Table 1.

Subtype/CRF assignments of recombinants containing CRF22_01A1 fragment.

Isolates ID	collected year	country /village	patient information				Genotype			GenBank No
			age	sex	mar. status	sexual partner	gag	pol	env
			02CAMLT04	2002	Cam/Douala	ND	ND	ND	ND		CRF22	CRF22	CRF02	EU743964
NYU488	2006	Cam/Bapile	68	F	widow	heterosexual	CRF02	CRF22	CRF22	JN864047
ARC087	2006	Cam/Yaounde	38	F	married	heterosexual	CRF02	CRF22	CRF22	JN864048
LPH27MF	2006	Cam/Bamenda	37	F	married	multiple	CRF22	CRF22	CRF22	JN864049
BDSH129	2008	Cam/Buea	32	F	married	heterosexual	CRF22	CRF02	CRF02	JN864052
LB005	2010	Cam/Limbe	32	F	single	multiple	CRF22	CRF22	CRF22	JN864050
LB011	2010	Cam/Limbe	23	F	married	multiple	CRF22	CRF22	CRF22	JN864051
LB013	2010	Cam/Limbe	56	M	married	multiple	CRF22	CRF22	CRF22	JN864058
LB045	2010	Cam/Limbe	36	F	single	heterosexual	CRF02	CRF22	CRF22	JN864053
LB052	2010	Cam/Limbe	33	F	married	heterosexual	CRF02	CRF02	CRF22	JN864054
LB054	2010	Cam/Limbe	66	M	ND	heterosexual	CRF22	CRF22	CRF22	JN864059
NYU1126	2000	Cam/Mboy	-	-	-	-	CRF01	CRF02	CRF22	EF087995
NYU830	2000	Cam/Gadji	-	-	-	-	CRF01	CRF02	CRF22	EF087994
01CM.0001BBY	2001	Cam/Yaounde	-	-	-	-	CRF22	CRF22	CRF22	AY371159
CM53122	2001	Cam/Bertoua	-	-	-	-	CRF22	CRF22	CRF22/01	AY037284+ AY037285
01CM.0130NY	2001	Cam/Nyabessang	-	-	-	-	CRF22	CRF22	CRF22/F2	AY371167
01CM.1152NG	2001	Cam/Nyabessang	-	-	-	-	A1	CRF22	CRF22	AY371163
01CM.4008HAN	2001	Cam/Ndikinim eki	-	-	-	-	D	CRF22	D	AY371163
CE03	2001	Cam/Yaounde	-	-	-	-	CRF22	-	11_cpx	DQ056985; DQ056864
A1102	2001	Cam/Douala	-	-	-	-	CRF22	CRF02	CRF02	EU619435; EU618346; EU618919
02CAMLT72	2002	Cam/Douala	-	-	-	-	CRF22	CRF22	CRF22	EU743963
02CM.3097MN	2002	Cam/Manyamen	-	-	-	-	CRF22	CRF22	CRF22	GQ229529
02CM.1917LE	2002	Cam/Lomie	-	-	-	-	CRF22	CRF22	CRF22	GQ229530
02CM.1867LE	2002	Cam/Lomie	-	-	-	-	CRF22	CRF22	CRF22	AY371165
02CM.3163MN	2002	Cam/Manyamen	-	-	-	-	F2	CRF22	CRF22/F2	AY371160
4455-7	2003	Cam/Yaounde	-	-	-	-	CRF22	CRF02	CRF02	EU619276; EU618760; EU618187
06SP42-303556	2005-7	E. Guinea	-	-	-	-	CRF22	CRF22	-	EU342808; EU342854
08GQ346	2008	E. Guinea	-	-	-	-	CRF22	CRF22		FN557311
J11469	2008	S. Arabia	-	-	-	-	CRF22	CRF22	CRF22	DG375282; DG375234; DG375289
01771M8	2008-9	C. African	-	-	-	-	-	CRF22	-	HM117941
36-0041F1	2008-9	C. African	-	-	-	-	-	CRF22	-	HM117942
8-CP001029	2008-9	C. African	-	-	-	-	-	CRF22	-	HM117944
38-0147M1	2008-9	C. African	-	-	-	-	-	CRF22	-	HM117951
14-CP001245	2008-9	C. African	-	-	-	-	-	CRF22	-	HM117957

Cam: Cameroon; E. Guinea: Equatorial Guinea; C. African: Certral African Republic; S. Arabia: Saudi Arabia; ND: Not Documented; A dash (-) indicates the data was not obtained. The bolded font indicates the samples characterized in this study.

One interesting finding in this study is that all of the recombinants of HIV-1 characterized displayed a fairly similar breakpoint at the accessory gene and env region which spans from vif to gp120 gene. For example, a breakpoint of 02CAMLT04 genome in nucleotide position nt 6911 in gp120 was also observed around similar position in 02CAMLT72 (nt 7015), 01CM.1152NG (nt 6963), LB052 (nt 6099), LB045 (nt 6261), 01CM.4008HAN (nt 6088) and 01CM.0130NY (nt 6668). The segment from one third of the env to the 5’ end of the nef region (nt 7016–8895) in 02CAMLT72, showed a CRF22_01A1 virus derived from subsubtype A1, but was replaced by CRF02_AG in a similar region of 02CAMLT04 virus, suggesting that this region may be a hotspot for recombination between CRF22_01A1 and other subtypes/CRFs. The exact prevalence of the CRF22_01A1 intersubtype recombinants is still unclear. However, according to the current study, given that the CRF22_01A1-containing recombinants were possibly introduced to the population as early as 2000, CRF22_01A1 could be more broadly distributed in Cameroon even outside this country (Figure 1B). Although the CRF22_01A1 strain was prevalent over the 10-year period, the numbers of reported CRF22_01A1 cases were limited since reference strains of CRF22_01A1 were not well classified until now. The complexity of CRF22_01A1 strains in Cameroon seemed to increase during the 2002–2010 period. The CRF22_01A1 strain has also been identified in Saudi Arabia^{24, 25}, Equatorial Guinea^{26, 27}, Central African patients²⁸ and USA²⁹ indicating that the CRF22 strain was spreading to different geographic regions.

It has been well known that HIV-1 CRFs behave like a pure subtype of HIV-1 and are able to recombine with other subtypes or CRFs during HIV-1 viral reverse transcription when co-infection or super-infection occurs. We document here that CRF22_01A1 is circulating in Cameroon and is involved in complex intersubtype recombination events with subtypes (D or F), subsubtypes (A1 or F2) and CRFs (CRF01_AE or CRF02_AG). CRF22_01A1 recombinants seem to co-circulate at higher proportions than its pure prototype. Our findings further demonstrate the dynamic evolution of emerging variants in Cameroon which could potentially impact the phlyogenetic nature of the epidemic in this region in the future.