Genomic organization and functional characterization of the chemokine receptor CXCR4, a major entry co-receptor for human immunodeficiency virus type 1.

CXCR4 is both a chemokine receptor and entry co-receptor for T-cell line-adapted human immunodeficiency virus type 1. The genomic organization and promoter function for the entire transcription unit of CXCR4 were determined. The gene contains 2 exons of 103 and 1563 base pairs (bp) interrupted by a 2132-bp intron precisely between codons 5 and 6 of the coding sequences. A transcription start site was identified 88 bp upstream of the initiation codon, and a polyadenylate addition site was identified 22 bp 3' to a polyadenylation signal. Transient expression assays defined a minimal promoter at positions -114 to +43 relative to the transcription start site. This region contains a TATA box, a nuclear respiratory factor-1 (NRF-1) site, and two GC boxes. Specific factor binding to the NRF-1 site and GC boxes were demonstrated by gel mobility shifts and DNase I footprinting. Site-directed mutagenesis showed that the NRF-1 site is crucial for promoter activity providing the first evidence for the regulation of a signal transduction gene by NRF-1. Sequences between -691 and -191 repress CXCR4 promoter activity. Further study of these regulatory elements will be important to understanding how CXCR4 functions as both a chemokine receptor and human immunodeficiency virus type 1 entry co-receptor.

numbers  resulting in a single, strongly hybridizing clone, CXCR4 -7a. Purified bacteriophage DNA from CXCR4 -7a was digested with the restriction endonuclease NotI, the 14-kilobase insert was subcloned into the pBluescriptIIKS-vector (Stratagene, Inc.) to generate pCXCR4 -7a, and a 5160-bp portion of the insert containing the CXCR4 transcription unit was completely sequenced on both strands using an ABI model 373A DNA sequencer and dye-terminator reactions (Applied Biosystems, Inc., Foster City, CA). Contigs were compiled using Sequencher version 3.0 software (Gene Codes Corp., Ann Arbor, MI). The final sequence, given in Fig. 1, was deposited in GenBank (accession number AF005058). Sequence analysis was performed using the Neural Network approach for prediction of splice donor/acceptors (25) and promoters (26) using interactive software. 3 Transcription factor binding site analysis was performed using the TESS and MatInspector interactive software packages. 4 Genomic Southern Analysis-High molecular weight DNA was extracted from the PBMCs of two HIV-1-seronegative human donors using the IsoQuick method (ORCA Research, Inc., Boswell, WA). Aliquots (10 g) of these genomic DNAs and 1-g amounts of purified, NotI-digested insert from pCXCR4 -7a were separately digested to completion with restriction endonucleases BamHI and PvuII. The products were electrophoresed through a 0.8% agarose gel, transferred to a nylon membrane (Hybond-N, Amersham Corp.), probed with the 32 P-random prime labeled pcDNA-fusin insert sequences, washed, and imaged using a model 850 PhosphorImager (Molecular Dynamics, Inc., Sunnyvale, CA).
5Ј RNA End-mapping-Total cellular RNA was purified from the human neoplastic T-cell lines HUT 78 and PM-1 cells using the RNAzol B technique (Tel-Test, Inc., Friendswood, TX). RNA (50-g amounts) was dissolved in 5 l of water and hybridized with 100 fmol of the 5Ј 32 P-end-labeled oligonucleotide probe PE1, 5Ј-CCCTCGGCGTCACTT-TGCTACCTGCTGC (sequence positions 991-964 according to Fig. 1), dissolved in 1 l of water at 58°C for 20 min in 50 mM Tris-HCl, pH 8.3, 50 mM KCl, 10 mM MgCl 2 , 10 dithiothreitol, 1 mM each deoxyribonucleotide triphosphate (dNTP), and 0.5 mM spermidine. Parallel reactions were performed with [ 32 P]PE1 in the absence of RNA. The reactions were then allowed to cool for 10 min at room temperature prior to incubation with 10 units of avian myeloblastosis virus reverse transcriptase (Promega, Inc., Madison, WI) in the presence of 50 mM Tris-HCl, pH 8.3, 50 mM KCl, 10 mM MgCl 2 , 10 mM dithiothreitol, 0.5 mM spermidine, 1.0 mM dNTP, and 2.8 mM sodium pyrophosphate at 42°C for 30 min. The reactions were electrophoresed through a denaturing 7% polyacrylamide sequencing gel along with a sequence ladder generated with unlabeled PE1 and pCXCR4 -7a template using [ 35 S]dATP and Sequenase version 2.0 (U. S. Biochemical Corp.), dried down, and visualized by autoradiography.
3Ј RNA End-mapping-A CXCR4-specific oligonucleotide 3ЈRACE1 5Ј-CCCAGCTGTTTATGCATAGA (sequence position 4537-4556) was used to perform 3Ј-rapid amplification of cDNA ends (3Ј-RACE) using 2 g of RNase-free, DNase I-treated total cellular RNA from HUT 78 cells with a kit purchased from Boehringer Mannheim. Control reactions performed in the absence of RNA produced no amplified material. The specific 240-bp 3Ј-RACE product was identified by hybridization with primer 3ЈRACE2, 5Ј-CAGTTTTCAGGAGTGGGTTG (sequence position 4666 -4685), agarose gel purified using the QIAEX II method (Qiagen, Inc., Chatsworth, CA), ligated into the vector pCR2.1 (Invitrogen, Carlsbad, CA), and the nucleotide sequence of multiple molecular clones was determined.
Promoter Mapping Experiments-A nested series of 5Ј deletions anchored to a common 3Ј sequence and a nested series of 3Ј deletions anchored to a common 5Ј sequence were generated from sequences representing the putative CXCR4 promoter elements and directionally cloned 5Ј to the chloramphenicol acetyltransferase (CAT) gene in the vector pKSCAT, a homologue of pSKCAT (27). The inclusive sequence positions of each deletion clone are 5Ј⌬1 (261-1037), 5Ј⌬3 (759 -1037), 5Ј⌬5 (836 -1037), 5Ј⌬7 (885-1037), 5Ј⌬9 (909 -1037), 3Ј⌬1 (759 -992), 3Ј⌬2 (759 -929), 3Ј⌬3 (759 -905). A modification of 5Ј⌬7, NRF-1⌬, was generated in which the nucleotides GCG at sequence positions 896 -898 were changed to TTT. All constructs were confirmed by DNA sequencing. Complete panels of these constructs were transfected in triplicate into A3.01, HUT 78, and Sup-T1 cells by electroporation using a method described previously (28) with the following modifications. For each electroporation, 17 g of CAT construction was electroporated into 1.5 ϫ 10 7 cells with 3 g of the ␤-galactosidase reporter construction pCMV␤-gal (28), 25 g of lysate was used for each CAT assay, and between 1 and 15% of the lysate was used to determine ␤-galactosidase activity. Positive control transfections were performed with the HIV-1 long terminal repeat reporter plasmid pU3R-III (29) and negative control transfections were performed with parental pKSCAT. Assays were performed within the linear range of the assay (1-25%). Raw percent conversions were corrected for background by subtraction of pKSCAT activity, normalized to ␤-galactosidase activity to control for transfection efficiency, and expressed as a relative percentage of the pU3R-IIIpositive control to normalize between experiments.
DNase I Footprinting Assays-A plasmid containing the 5Ј-flanking region of CXCR4 (5Ј⌬3) was digested with the restriction endonucleases KpnI and XmaIII to obtain a fragment (sequence position numbers 758 -933) with unique 3Ј-and 5Ј-overhanging termini, respectively. This restriction fragment was gel-purified and labeled using [␣-32 P]dCTP and the Klenow fragment of DNA polymerase holoenzyme (New England Biolabs, Inc.). Footprinting was performed using 16 g of A3.01 nuclear extracts prepared as described previously (30) and DNase I (Promega, Inc.) per the protocol specified in the technical bulletin of the manufacturer. The same unlabeled restriction fragment was used to generate a Maxam-Gilbert GϩA chemical sequencing ladder as described previously (31). Bovine serum albumin (16 g, New England Biolabs, Inc.) was used in place of nuclear extracts in the control reactions. The reaction products were deproteinized, electrophoresed through a 6% denaturing polyacrylamide gel, dried down, and visualized by autoradiography.
Electrophoretic Mobility Shift Assays-For NRF-1 studies, a 26-bp portion of the CXCR4 promoter (sequence positions 885-910) containing the consensus NRF-1 site was synthesized as two complementary oligonucleotides. A mutant form of these oligonucleotides was also synthesized (Fig. 5A). For standard binding experiments, 0.5-ng amounts of these 32 P-end-labeled probes were incubated with 8 g of A3.01 nuclear extracts at 10°C for 20 min in a reaction buffer containing 50 g/ml poly(dI-dC) (Pharmacia Biotech Inc.), 10 mM Tris-HCl, pH 7.5, 50 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, 5% (v/v) glycerol in a total reaction volume of 20 l. For supershift experiments, 1 l of NRF-1 antiserum (kindly provided by Dr. R. Scarpulla, Northwestern University) was subsequently added and incubated at room temperature for an additional 30 min. As a nonspecific control, 1 l of Sp4 antibody (Santa Cruz Biotechnologies, Inc.) was used for both NRF-1 and Sp1 EMSAs. For competition experiments, 100-fold molar excess (50 ng) of unlabeled wild-type or mutant oligonucleotide was added and allowed to incubate with the nuclear extract at room temperature for 20 min prior to the addition of labeled oligonucleotide. For GC box studies, a 46-bp portion of the CXCR4 promoter (sequence positions 842-887) containing GC boxes I and II ( Fig. 1) was synthesized as two complementary oligonucleotides. An oligonucleotide with mutations in both of these GC boxes was also synthesized (Fig. 5B). The reaction conditions differed from the NRF-1 experiments in that the poly(dI-dC) concentration was increased to 250 g/ml, 1 footprinting unit of purified Sp1 protein (Promega, Inc.) was used in place of nuclear extracts, 1 g of bovine serum albumin was added to each binding reaction, and 1 l of Sp1 antiserum (Santa Cruz Biotechnologies, Inc.) was used for supershifts. The DNA-Sp1 binding reactions were carried out at room temperature for 10 min. Binding complexes were resolved on native polyacrylamide gels and visualized by autoradiography essentially as described previously (32).

Isolation and Structural Organization of the Human CXCR4
Gene-Screening of a human PBMC genomic DNA library with a probe consisting of the entire CXCR4 coding sequence yielded three bacteriophage clones of which a single clone, CXCR4 -7a, hybridized with oligodeoxynucleotide probes representing the extreme ends of the published coding sequences (24). As this raised the possibility that the entire transcription unit was contained on the 14-kilobase pair insert of this clone, it was chosen for nucleotide sequencing. Comparison of the 5160-base pair genomic sequence obtained ( Fig. 1) with the cDNA sequence suggested that this was so. The 67-base pair 5Ј-untranslated region of the cDNA sequence was contiguously identified in the genomic sequence between positions 971 and 1037 with the initiating methionine codon beginning at position 1038. Following 15 base pairs of coding sequence corresponding to amino acid residues M-E-G-I-S, a consensus splice donor was encountered followed by a 2132-base pair intron (sequence position 1053-3184) precisely after the third base of codon 5 at position 1052. A consensus splice acceptor was identified at the 3Ј terminus of the intron followed by 1044 base pairs of coding sequences (positions 3185-4228) whose open reading frame started with the first base of codon 6 (isoleucine) and ended with a TAA codon. Taken together, these two exons comprised all 352 amino acids predicted from the cDNA sequence (24). The putative polyadenylation signal, AATAAA, was identified 498 bp 3Ј to the TAA codon starting at position 4726 of the genomic sequence, and the remaining 21 bp of the 3Ј-untranslated region predicted from the cDNA sequence was identified downstream contiguous with the published cDNA.
Identification of the CXCR4 Transcription Initiation and Termination Sites-To better define the 5Ј and 3Ј boundaries of the transcription unit, RNA end-mapping studies were performed. Primer extension studies were performed with a 32 Pend-labeled primer (PE1) from the 5Ј-untranslated region of CXCR4 using either no RNA template (lane (Ϫ)) or total cellular RNA from HUT 78 cells (lane (ϩ)) (Fig. 2). PE1 was also used to generate a dideoxy sequencing ladder from clone 5Ј⌬3 FIG. 1. Sequence and genomic organization of the human CXCR4 gene. Numbering for the CXCR4 gene is according to GenBank accession number AF005058. Regions matching transcriptional control consensus sequences are underlined and labeled. The consensus polyadenylate addition signal sequence AATAAA is also underlined. Deduced amino acid sequences are listed below the coding region using the single-letter code centered on the second base of each codon. Exon sequences are denoted by uppercase letters and intron, and flanking sequences are denoted by lowercase letters.
( Fig. 2, lanes GATC), which were co-electrophoresed with the primer extension reactions. The mock primer extension lane revealed multiple, nonspecific extension products. However, a family of three specific extension products was identified centered on the dominant, central C residue (Fig. 2, arrow) of the nucleotide triad 5Ј-ACT corresponding to the sense triad of 5Ј-AGT at positions 949 -951. Thus, the G at position 950 was designated the transcription start site/5Ј transcription unit border and assigned the identifier ϩ1. Similar 5Ј ends were identified using total cellular RNA from A3.01 and PM-1 cells (data not shown). The CXCR4 transcription start site is 21 bp upstream of the 5Ј cDNA terminus identified previously (24) and 32 bp downstream of a canonical TATA box, TATAA, at sequence positions 919 -923 (Fig. 1). The 3Ј end of the CXCR4 transcription unit was mapped using 3Ј-RACE to sequence position 4747 (data not shown) 22 bp downstream of a canonical polyadenylation addition signal, AATAAA, at positions 4726 -4731 in agreement with the published 3Ј cDNA terminus (24).
Based on these data, the CXCR4 gene comprised 2 exons of 103 and 1563 base pairs (Fig. 1, uppercase characters) interrupted by a single 2132-base pair intron (Fig. 1, lowercase  characters). Exon 1 contained 88 bp of 5Ј-untranslated sequences followed by 15 bp of coding sequences ending precisely after codon 5. Exon 2 contained 1044 bp of coding sequences from codon 6 to the termination codon TAA followed by 519 bp of 3Ј-untranslated sequences.
Potential recombination in the library was ruled out by digesting both CXCR4 -7a and PBMC genomic DNAs with restriction endonucleases BamHI and PvuII and performing Southern blotting using full length CXCR4 cDNA as a probe. The CXCR4 -7a sequence predicted a BamHI fragment of 3602 base pairs and a PvuII fragment of 2269 base pairs, which were identified by hybridization in both the clone and genomic DNA digests (data not shown).
Positional Mapping of the CXCR4 Promoter-The CXCR4 5Ј-flanking region sequences were searched for known promoter elements. The sequences immediately upstream of exon 1 and the ϩ1 site were identified as promoter and transcription start sites, respectively, by the Neural Net Promoter prediction algorithm (26). Sequences 5Ј to the transcription start site included a TATA box, a potential NRF-1 site, and four potential antisense GC boxes, of which the most distal pair, GC box III and IV, overlap (Fig. 1).
A nested series of 5Ј deletions were generated starting from position Ϫ689 relative to the transcription start site with the 3Ј anchor at position ϩ88, and multiple replicates were transfected into A3.01 cells (Fig. 3). Although moderate CAT activity was obtained with clone 5Ј⌬1, a 5-fold gain in signal intensity was consistently noted with clone 5Ј⌬3 which lacked sequences between Ϫ689 and Ϫ191 indicative of repressor sequences within this domain. Deletion of sequences to position Ϫ114 (clone 5Ј⌬5), which included the distal GC boxes III and IV at positions Ϫ138 to Ϫ126, had no effect on CAT activity. Further deletion of sequences to position Ϫ65 (clone 5Ј⌬7), which excised the proximal GC boxes I and II at positions Ϫ88 to Ϫ66, modestly decreased CAT activity to 40% of that seen with clones 5Ј⌬3 and 5Ј⌬5, whereas deletion of sequences to position Ϫ41 (clone 5Ј⌬9), which excised the NRF-1 site, essentially abolished CXCR4 promoter activity.
A nested series of 3Ј deletions were generated with the 5Ј anchor at position Ϫ191 (Fig. 3). Relative to clone 5Ј⌬5 (positions Ϫ191 to ϩ88), deletion of sequences to ϩ43 in clone 3Ј⌬1 had little effect on CAT activity, whereas deletion of sequences surrounding the transcription start site in clone 3Ј⌬2 (positions Ϫ191 to Ϫ21) and deletion of the TATA box in clone 3Ј⌬3 (positions Ϫ191 to Ϫ45) progressively reduced CAT activity. Taken together, the CXCR4 minimal promoter domain was mapped to positions Ϫ114 to ϩ43 (sequence positions 836 -992), which included GC boxes I and II, the NRF-1 site, the TATA box, and the transcription start site. Given that removal of the NRF-1 site in clone 5Ј⌬9 effectively ablated the CXCR4 promoter, a homologue to clone 5Ј⌬7 was generated (clone NRF-1⌬) by site-directed mutagenesis in which the core NRF-1 binding site nucleotides GCG (33) at positions 896 -898 (Fig. 1) were changed to TTT. This clone, similar to clone 5Ј⌬9, showed essentially no CAT activity, demonstrating the critical role of the NRF-1 binding site sequences in the CXCR4 promoter. Results are expressed as mean ϩ standard deviation relative to the activity of the control construct pU3R-III containing the HIV-1 promoter. The construct containing a mutated NRF-1 site is labeled NRF-1 ⌬. Differences in CAT activity between constructs were analyzed with the Wilcoxon matched-pairs tests. Constructs 5Ј⌬3 and 5Ј⌬5 produced significantly more CAT activity than 5Ј⌬7 and 5Ј⌬9 (p Ͻ 0.05). Activity of 5Ј⌬7 was also significantly greater than that of 5Ј⌬9 (p Ͻ 0.05). The activity of 5Ј⌬1 was significantly less than that of 5Ј⌬3 (p Ͻ 0.005). The activity of NRF-1⌬ was significantly lower than that of 5Ј⌬7 (p Ͻ 0.04). The 3Ј deletion constructs were all significantly different in their activity relative to each other (p Ͻ 0.01). Similar results were obtained by transfection of all of the constructs shown in Fig. 3

into both Sup-T1 and HUT 78 cells.
Identification of Potential cis-Acting Elements by Footprint Analysis-A restriction fragment obtained by digesting a plasmid containing sequence positions 758 -933 with KpnI and XmaIII was used for DNase I footprinting analysis with A3.01 cell nuclear extracts (Fig. 4). Compared with the control lane using bovine serum albumin (lane 2), there was a general decrease in band intensity from positions Ϫ88 to Ϫ49 relative to the transcription start site to include the two proximal GC boxes and the NRF-1 site which was heavily protected (lane 1). Taken together with the reporter gene analyses, the importance of these promoter elements was further substantiated.
Transcription Factors NRF-1 and Sp1 Bind to the CXCR4 Promoter-To determine if specific transcription factors corresponding to the NRF-1 site and GC boxes bound to their respective binding sequences in the CXCR4 promoter, electrophoretic mobility shift (EMSA) experiments were performed.
Oligonucleotides representing CXCR4 promoter sequences containing either the wild type or mutant NRF-1 binding sites were used as EMSA probes with A3.01 cell nuclear extracts (Fig. 5A). No complex formation was seen in the absence of extract for both wild-type and mutant probes (Fig. 5A, lanes 1  and 7, respectively). Wild-type probe generated one specific (C1) and two nonspecific (NS) complexes with A3.01 extract (lane 2), which was readily competed by unlabeled wild-type probe (lanes 3) but not an excess of mutant probe (lane 4). Addition of NRF-1 antiserum (lane 5) but not nonspecific Sp4 antibody (lane 6) resulted in the supershift of C1 to C2. Only nonspecific complexes were observed with the mutant probe (lane 8). These data strongly suggest that NRF-1 specifically binds to the NRF-1 binding site in the CXCR4 promoter.
Oligonucleotides representing CXCR4 promoter sequences containing either wild type or mutant GC box sites were also used as EMSA probes with purified Sp1 protein (Fig. 5B). Purified protein was used in lieu of nuclear extracts, as Sp1 has been previously shown to be expressed in T-cells (34,35). A mutant oligonucleotide, containing mutations in both GC boxes, was also used (Fig. 5B). No complex formation was seen in the absence of Sp1 for both the wild type and mutant probes (Fig. 5B, lanes 1 and 6, respectively). Wild-type probe generated a strong specific (C1) complex with Sp1 protein (lane 2), which was readily competed by an excess of unlabeled wildtype probe (lane 3). Addition of Sp1 antiserum resulted in the supershift of C1 to complex C2 (lane 4). Supershift was not seen with the nonspecific Sp4 antibody. Only a very weak C1 complex was seen with the mutant probe (lane 7). Taken together, these data strongly suggest that Sp1 specifically binds to the two proximal GC boxes in the CXCR4 promoter. DISCUSSION We have characterized the entire transcription unit of the human CXCR4 gene. The sequence of the gene has a very high GC content; the coding sequences are 50.3% GC, and the intron and 0.95 kilobase pairs 5Ј to the transcription start site are 56.9 and 53.1% GC, respectively, which is higher than 99% of human genes (36). The biologic significance of this observation is uncertain. However, the practical implication of this finding is that large portions of the gene are quite difficult to amplify by the polymerase chain reaction, which will make attempts to perform high through-put studies to look for genotypic polymorphisms in cohorts of subjects more difficult.
Polymorphisms that result in mutant chemokine receptor proteins have been identified for CCR5 (37,38) and CCR2B (39). CCR5 polymorphisms correlate with delayed progression to AIDS (21,40) and decreased susceptibility to HIV-1 infection (21,38,40), whereas the implications for CCR2B polymorphisms are less clear (19,39). The finding of substantial polymorphisms in the coding sequence for CXCR4 may be less likely than for the other chemokine receptors. CXCR4 mRNA is widely expressed in a variety of hematopoietic and non-hematopoietic tissues, including brain, heart, kidney, lung, and liver (41,42), which is a markedly larger distribution than the other identified chemokine receptors. There is so far only one identified ligand, SDF-1, which produces a lethal mutation when knocked out in mice (43), and both CXCR4 and SDF-1 genes show high interspecies conservation (9,41). All of these findings argue for the criticality of a functional CXCR4 gene product, and against the likelihood of finding significant polymorphisms in human cohorts.
The validity of our proposed genomic organization is supported by multiple lines of evidence. Primer extension analysis was performed in multiple cell lines and yielded a strong single nucleotide band with weaker bands for the adjacent nucleotides on either side, but no other significant start sites. This start site was predicted before the performance of these experiments by the Neural Network algorithm (26) and is further supported by the presence of nearby upstream elements including TATA and GC box elements. The intron-exon boundaries contain consensus splice donor and acceptor sequences, were predicted by Neural Network (25) and maintain the open reading frame as predicted by the cDNA sequences (24). The polyadenylate addition site was quite precise when sequenced in multiple clones, occurs immediately downstream of a consensus polyadenylation signal, and agrees with previously published work (24). The size of mRNA species predicted by sum of the two exons and a 100 -200-base pair poly(A) tail agrees with that observed by Northern analysis (6,24).
Analysis of the CXCR4 5Ј-flanking sequence showed the presence of several potential binding sites for known transcription factors. The importance of the two most proximal Sp1 sites, and to a much greater degree, the NRF-1 site in the basal promotion of the CXCR4 gene was shown by multiple, complementary experiments, including DNase I footprinting, transient transfection with promoter/CAT constructs, site-directed mutagenesis of the NRF-1 site, and mobility shift assays. The transient transfection experiments, deleting from both 5Ј and 3Ј ends of the 5Ј-flanking region, localized the critical basal promoter sequences between positions Ϫ114 to ϩ43 relative to the transcription start site (sequence positions 836 and 992). Deletion or mutation of the sequences containing the NRF-1 site at positions Ϫ61 to Ϫ49 (sequence positions 889 -901) essentially abolished CXCR4 promoter activity.
NRF-1 is a nuclear encoded gene product that has been shown to be important for the transcriptional regulation of multiple mitochondrial genes involved in organelle biogenesis and cellular respiration (33). Potential NRF-1 binding sites have also been identified in several genes important for cell maintenance, growth, and proliferation such as the genes for ornithine decarboxylase, 5-aminolevulinate synthase, bcl-2, and DNA polymerase-␣, and signal transduction genes such as those for cyclophilin, calmodulin, and murine GM-CSF (33). However, the functional significance of these observations is unclear. Following submission of this manuscript, Moriuchi and colleagues (45) described the recovery of the proximal portion of the CXCR4 promoter reported here and also demonstrated the importance of NRF-1 for CXCR4 transcriptional regulation. These two reports represent the first demonstration of the role of NRF-1 in the expression of a signal transduction gene. Given the evidence for the importance of CXCR4 in cellular development and its presence on immune effector cells, it is easy to postulate that NRF-1 serves to coordinate an increase in a cell's metabolic capacity in response to inflammatory or proliferative signals, preparing the cell to migrate or divide. This increased level of cellular activation has been previously shown to be important for high level HIV-1 replication in target cells (46). Transient transfections done in several cell lines consistently showed a degradation of promoter activity when sequences between Ϫ668 and Ϫ191 were present in the construct, suggesting the presence of negative regulatory elements in this portion of the sequence. Precise mapping of this region to delineate these elements is now in progress. An inducible repressor element could conceivably be a target for the development of therapeutic agents; if CXCR4 transcription could be blocked, entry of SI strains of HIV-1 into cells might be inhibited.
Identification of two new co-receptors that also mediate HIV-1 and SIV entry, such as Bonzo (47,48) and BOB (47), as well as reports of others (49), underscores the growing complexity of HIV-1 viral entry into target cells. Elucidation of the regulatory mechanisms that govern the interplay of these co- receptors will be important to a fuller understanding of HIV disease.