Downloaded from www.microbiologyresearch.org by IP: 54.226.235.197 On: Mon, 21 Nov 2016 20:52:26 Short Communication CCR5 use by human immunodeficiency virus type 1 is associated closely with the gp120 V3 loop N-linked glycosylation site Peter Clevestig, 1 Lotta Pramanik, 1 Thomas Leitner 2 and Anneka Ehrnst 1,3 Correspondence Anneka Ehrnst Anneka.Ehrnst@mtc.ki.se 1,3 Microbiology and Tumor Biology Center 1 and Department of Laboratory Medicine, Division of Clinical Virology, Karolinska University Hospital Huddinge 3 , Karolinska Institutet, Box 280, SE-171 77 Stockholm, Sweden 2 Theoretical Biology and Biophysics Group, Los Alamos National Laboratory, Los Alamos, NM, USA Received 8 September 2005 Accepted 10 November 2005 Human immunodeficiency virus type 1 (HIV-1) enters cells through the chemokine receptors CCR5 (R5 virus) and/or CXCR4 (X4 virus). Loss of N-linked glycans and increased net charge of the third variable loop (V3) of the gp120 envelope glycoprotein have been observed to be important steps towards CXCR4 use. All reported sequences using CCR5 or CXCR4 exclusively, or using both, were gathered from the Los Alamos HIV Database and analysed with regard to the V3 N-linked glycosylation motifs (sequons) and charge. The V3 loop glycan had a sensitivity of 0?98 and a 0?92 positive predictive value in the context of CCR5 use. The difference from X4 was remarkable (P<10 ”12 ). Especially, the sequon motif NNT within the V3 loop was conserved in 99?2 % of the major clades. The results suggest a close association between the V3 loop glycan and CCR5 use and may provide new insight into HIV-1 tropism and help to improve phenotype-prediction models. To enter cells, human immunodeficiency virus type 1 (HIV- 1) uses the coreceptors CCR5 and/or CXCR4, together with the T-cell differentiation antigen CD4. Virus using the CCR5 receptor (R5 virus; Berger et al., 1998) or of the non-syncytium-tropic phenotype is suggested to be more transmissible (Zhang et al., 1993), whilst virus using the CXCR4 receptor (X4 virus; Berger et al., 1998) has been associated with disease progression (Connor et al., 1997; Scarlatti et al., 1997). The third variable (V3) loop of the HIV-1 gp120 envelope glycoprotein was recognized early on to play an important role in governing the choice of target cells (Hwang et al., 1991; De Jong et al., 1992; Fouchier et al., 1992; Shioda et al., 1992). Few amino acid substitutions and an increasing net charge of the V3 loop were sufficient to confer a change in cellular tropism in vitro (De Jong et al., 1992; De Wolf et al., 1994). A decreased number of N-linked glycosylation sites (sequons) in gp120, especially within and around the V3 region, has been demonstrated during evolution from the R5 to the X4 phenotype (Pollakis et al., 2001; Polzer et al., 2001, 2002). Computer-based models have been developed to predict the biological phenotype of HIV sequences (Briggs et al., 2000; Resch et al., 2001; Jensen et al., 2003; Pillai et al., 2003), based on multiple linear regression (Briggs et al., 2000), specific amino acids within V3 and overall charge among subtype B viruses (Resch et al., 2001) and other subtypes (Pillai et al., 2003). Machine learning has been used to develop phenotype classifiers (Resch et al., 2001) and position-specific scoring matrices have also been used (Jensen et al., 2003). These methods relate to the assumption or possibility that the R5 and X4 phenotypes may be evaluated or classified, based on the properties of their specific amino acids, from the same scale or sum of scales. If the R5- and X4-specific properties relate to different characteristics, however, such programs may overlook essential differences in the R5 versus X4 phenotypes. In a previous study (Clevestig et al., 2005), a phylogenetic analysis was performed on a large set of individual HIV-1 sequence clones from integrated DNA or plasma RNA, connected to sequences of isolates with determined co- receptor use. We made a preliminary observation (Clevestig et al., 2005) that R5 isolates and R5-associated sequences seemed to be connected to the presence of the N-linked glycosylation motif within the V3 loop of the env gp120 gene. As we studied only four individuals, linked epidemi- ologically as two mother–child pairs, we sought to explore this finding in the most epidemiologically and genetically Tables showing details of sequences retrieved and sequence align- ments are available as supplementary material in JGV Online. 0008-1510 Printed in Great Britain 607 Journal of General Virology (2006), 87, 607–612 DOI 10.1099/vir.0.81510-0