CCR5{Delta}32 59537-G/A Promoter Polymorphism Is Associated with Low Translational Efficiency and the Loss of CCR5{Delta}32 Protective Effects

We have recently demonstrated that the CCR5 ${Delta}$ 32 protein interactswith CCR5 and CXCR4 and down-modulates their cell surface expression.We have also reported the absence of detectable expression ofthe truncated CCR5 ${Delta}$ 32 protein in four out of six human immunodeficiencyvirus-infected (HIV⁺) CCR5^–/– individuals. To explainthe defect in protein expression in these samples, we clonedand sequenced the promoter regions of the six HIV⁺ individuals.We have identified several polymorphisms in the CCR5 ${Delta}$ 32 promoterregion, but these polymorphisms were not associated with significantdifferences in mRNA levels. Coupled in vitro transcription/translationand polyribosome analysis demonstrated a strong associationbetween a variant genotype designated CCR5 ${Delta}$ 32 59537-A/A and alow translation efficiency. Protein analysis indicated thatthe peripheral blood mononuclear cells from two of the HIV⁺CCR5^–/– individuals carrying the CCR5 ${Delta}$ 32 59537-A/Avariant expressed trace amounts of CCR5 ${Delta}$ 32 protein compared tothe individuals carrying the CCR5 ${Delta}$ 32 59537-G/G genotype. Theresults imply that the absence of CCR5 ${Delta}$ 32 protein in two HIV⁺individuals is due to a genetic defect in the translation ofthe protein. Together, these results highlight the importanceof the CCR5 ${Delta}$ 32 protein as an HIV suppressive factor and providefurther insight into the mechanism of the protective effectof the CCR5 ${Delta}$ 32 mutation.

INTRODUCTION

Top

INTRODUCTION

The CC chemokine receptor 5 (CCR5) is the coreceptor utilizedby human immunodeficiency virus type 1 (HIV-1) to infect humanCD4⁺ macrophages and T lymphocytes (4, 10, 12-14). A naturallyoccurring 32-base pair deletion in the human CCR5 gene is highlyassociated with resistance to HIV-1 infection (8, 11, 13, 14,38). CCR5 ${Delta}$ 32 encodes a truncated protein that is not detectedon the cell surface and therefore is not functional as a coreceptor(8, 11, 13, 14, 38). The majority of individuals carrying thetwo alleles of the CCR5 ${Delta}$ 32 mutation (CCR5^–/–) arehighly protected against HIV-1 infection (reviewed in reference32). Individuals who are heterozygous for the mutant allele(CCR5^+/–) are not protected against infection, but onceinfected, their progression to AIDS is delayed (8, 11, 13, 27,36, 38), indicating that partial resistance can occur in thepresence of a single copy of CCR5 ${Delta}$ 32.

In rare cases, CCR5 ${Delta}$ 32 homozygosity was associated with HIV-1infection (5, 7, 15, 22, 28, 33, 34, 36), but in these cases,the mechanism of infection has not been defined. In most cases,the exclusive use of CXCR4 by virus isolates or the presenceof the env gene sequences typical of X4 viruses was observed.Isolation of the dual-tropic (R5X4) HIV-1 from infected individualshomozygous for the CCR5 ${Delta}$ 32 allele has also been reported (15,16, 31).

Our previous work suggested that the HIV resistance in CCR5 ${Delta}$ 32homozygotes may result from both the genetic loss of CCR5 fromthe cell surface as well as the active down-regulation of CXCR4expression by the mutant CCR5 ${Delta}$ 32 protein. We and others havedemonstrated that the CCR5 ${Delta}$ 32 protein may form heterodimers withthe wild-type CCR5 and CXCR4 proteins, which are retained inthe endoplasmic reticulum and result in reduced cell surfaceexpression of the coreceptors (2, 6, 9, 37). We have previouslydemonstrated the absence of detectable CCR5 ${Delta}$ 32 protein in fourinfected CCR5^–/– individuals (1, 2). Here, we haveexamined the effect of polymorphisms in the promoter regionand the 5' untranslated region (UTR) on CCR5 ${Delta}$ 32 expression.

To explain the absence or low abundance of the CCR5 ${Delta}$ 32 proteinin infected CCR5^–/– individuals, we cloned the promoterregions of the HIV-infected (HIV⁺) CCR5 ${Delta}$ 32 homozygous individuals.We identified a 59537-G/A polymorphism in exon 3 (E3) that isa part of the CCR5 ${Delta}$ 32 promoter region and the 5' UTR. Promoteranalysis of luciferase reporter constructs did not reveal significanteffects on mRNA transcription; however, the 59537-A/A genotypewas strongly associated with low translational efficiency ofthe CCR5 ${Delta}$ 32 protein. The results provide important insight intothe mechanism of resistance to HIV-1 infection and pathogenesis.

MATERIALS AND METHODS

Top

MATERIALS AND METHODS

Cell lines and CCR5^–/– PBMC samples. Human embryonic kidney (HEK293) and HeLa cells were purchasedfrom ATCC and cultured in Dulbecco's modified Eagle's mediumcontaining 10% fetal bovine serum. The peripheral blood mononuclearcell (PBMC) samples (six HIV⁺ and some HIV-negative [HIV^–]CCR5^–/– cells) have been described previously (1).The clinical profile of patient VH has been described previously(7, 31). VM and VH are two CCR5^–/– individuals;VM is uninfected, while VH is infected, and the mode of infectionappears to be via homosexual intercourse. The three HIV⁺ SEROCOsamples are all from CCR5^–/– individuals who arehomosexuals. The SEROCO sample 2 has been reported previously(36). The Shep sample 164 has been documented previously (34)and examined in our previous studies (1, 2). The MACS infectedCCR5^–/– sample, ID 20705, has been reported previously(15). The other HIV^– CCR5^–/– samples and thesamples from individuals homozygous for the wild-type CCR5 allele(CCR5^+/+) were obtained from the MACS cohort. PBMCs were stimulatedwith phytohemagglutinin (PHA) and cultured in RPMI 1640 mediumsupplemented with 10% fetal bovine serum, penicillin (100 units/ml),streptomycin (100 µg/ml), and recombinant interleukin2 (IL-2; 100 units/ml).

Reverse transcription (RT)-PCR analysis of mRNA transcription in CCR5^–/– PBMCs. Total RNA of transfected cells was isolated from PBMCs usingan RNeasy mini-kit (Qiagen, MD) reagent. The RNA (1 µg)was reverse transcribed using oligo(dT)₂₀ primer and SuperScriptIII reverse transcriptase (Invitrogen, Carlsbad, CA). The resultingfirst-strand DNA (2 µl) was amplified to a final volumeof 25 µl containing 10 pmol of each primer and 1 unitof Taq polymerase (Promega, Madison, WI). The primers used toamplify the E2- ${Delta}$ 32 transcript (183 bp) were forward primer (FP)5'-ATTCTGTGTAGTGGGATG-3' and reverse primer (RP) 5'-TGAAGGAGGGTGGAGTTAA-3'.The primers used to amplify the E3- ${Delta}$ 32 transcript (54 bp) wereFP 5'-CTCATCTGGCCAGAAGAG-3' and RP 5'-CGGGGAGAGTTTCTTGTA-3'.Amplification conditions were denaturation at 94°C for 30s, annealing at 50°C for 30 s, and extension at 72°Cfor 60 s for 35 cycles. The amplified products (5 µl)were subjected to electrophoresis on 2% agarose and stainedwith ethidium bromide.

PCR cloning of the CCR5 ${Delta}$ 32 promoter regions. Genomic DNA was isolated by using a DNeasy tissue kit (Qiagen,MD) reagent. The primers were designed according to GenBanksequence U95626. The upstream promoter (P_U) was amplified usingthe FP 5'-AAGCTAGCAGGAAATGGAAGCTTGGGCA-3' and the RP 5'-CGAAGCTTCGTGACCTTGGCTCTAGAAT-3'.The downstream promoter (P_D) was amplified using the FP 5'-GCAAGCTTCGCCTTCTTAGAGATCACAA-3'and the RP 5'-CGAAGCTTCCAAACTGTGACCCTTTCC-3'. The amplifiedP_U fragment maps to sequences between 57236 and 58761, whilethe amplified P_D fragment represents sequences from 58491 to59728. Briefly, DNA samples and primers were first denaturedat 94°C for 2 min and then subjected to 35 cycles of PCRamplification (95°C for 15 s, 55°C for 30 s, and 68°Cfor 1 min per kb). PCRs were performed using Platinum Pfx DNApolymerase (Invitrogen). PCR products were purified by usinga Quick Gel Extraction kit (Invitrogen) through an agarose geland cloned into a pGEM T Easy vector (Promega) after A tailing.The resulting positive promoter clones were sequenced to confirmtheir identities. The P_D and P_U fragments cloned into the pGEMT-Easy vector were digested with HindIII and then ligated tomake P_U+P_D, which was joined at an overlapping HindIII siteat a position from 58525 to 58530 (Fig. 1C).

View larger version (30K):
[in this window]
[in a new window]

FIG. 1. (A) Schematic diagram of the CCR5 ${Delta}$ 32 gene promoter. The promoter regions and potential transcripts resulting from CCR5 ${Delta}$ 32 are as originally described for CCR5 by Mummidi et al. (30). E1, E2, E3, and E4 represent the four exons, P_U refers to the upstream promoter, and P_D refers to the downstream promoter, as described by Mummidi et al. (30). The position of the 59537 G/A-polymorphism in the E3 region is indicated. This polymorphism corresponds to –1951, according to the numbering system that designates the first nucleotide of the translation start site of CCR5 ${Delta}$ 32 as position 1 and the nucleotide immediately upstream of this as position –1. (B) RT-PCR analysis of endogenous CCR5 ${Delta}$ 32 mRNA isolated from PBMCs with the indicated CCR5 genotypes. Labeling on the right indicates the identity of the primers used. Equivalent amounts (5 µg) of poly(A)⁺ mRNA were used in this analysis. The primers to amplify β-actin were used to control for gel loading. The amplified products were analyzed with a 1% agarose gel.

The P_U, P_D, and P_U+P_D promoter region fragments in the pGEMT-Easy vector were digested with NotI, filled in with Klenowfragment, ligated into the pRF vector (Promega) that had beendigested with NheI/XbaI to remove the T7 promoter and the Renillaluciferase gene, and blunt ended. Promoter activity assays weredone using a dual-luciferase reporter assay kit according tothe manufacturer's instructions (Promega). Briefly, 0.1 millionHeLa cells per well were seeded into a 24-well plate for overnightgrowth. Cotransfection of 0.5 µg DNA of the P_U, P_D, orP_U+P_D promoter construct (Fig. 1C) along with 0.2 µg ofthe pRL-CMV vector (Renilla luciferase under cytomegalovirus[CMV] promoter control) was accomplished by using Lipofectamine2000 (Invitrogen). The empty pGL3B vector was used as a blankcontrol. Transfected cells were incubated for 72 h and thenwashed twice in cold phosphate-buffered saline (PBS) bufferand lysed with 100 µl passive lysis buffer (Promega) byshaking for 15 min at room temperature. A 20-µl sampleof cell lysate was taken to measure luciferase activity by usingan automated AutoLumat LB 953 Berthold luminometer.

PCR cloning of CCR5 ${Delta}$ 32 5' UTR fragments and construction of 5' UTR reporter genes. The CCR5 ${Delta}$ 32 5' UTR E1 was amplified using the FP 5'-CTTCAGATAGATTAT-3'and the RP 5'-CGAGTACTATGCCAGATACGTAGG-3'. E2/3 was amplifiedusing the FP 5'-CGAGTACTATTCTGTGTAGTGGGATG-3' and the RP 5'-CGAGTACTCGGGGAGAGTTTCTTGTA-3'.E3 was amplified using the FP 5'-CGAGTACTCTCATCTGGCCAGAAGAG-3'and the RP 5'-CGAGTACTCGGGGAGAGTTTCTTGTA-3'. The E1, E2/3, andE3 5' UTR fragments were subcloned into the pGEM T Easy vector,digested with Scal, and ligated to make E1/2/3 or E1/3 withan overlapping Scal site. For constructing the T7 promoter luciferasereporter system, the pRF vector (Promega) was digested withEcoRV/XbaI to remove the Renilla luciferase gene and then self-ligatedto make a new pT7FL vector. The CCR5 ${Delta}$ 32 5' UTR E3, E2/3, E1/3,and E1/2/3 fragments in the pGEM T Easy vector were digestedwith SpeI/EcoRI and subcloned into the same SpeI/EcoRI digestionsite of the pT7FL vector. The structure of the different 5'UTR fragments was designed as reported by Mummidi et al. (30).

To study the in vivo translation efficiency of the constructs,0.5 µg of each expression vector was transfected intoHEK293 cells, using Lipofectamine 2000 reagent (Invitrogen),together with 0.2 µg of P Tracer-LacZ (Invitrogen), whichcontains the LacZ gene under the control of the T7 promoter.After 4 h of transfection, the expression was activated by infectionwith the vaccinia virus vTF7-3. On the following day, the cellswere washed twice in cold PBS buffer and then lysed with passivelysis buffer. The dual-luciferase enzymatic activity assay wascarried out according to the manufacturer's protocol (Promega)and measured by luminometer. The β-galactosidase activitywas measured according to standard procedure (3). The translationefficiency was calculated by dividing the luciferase units bythe β-galactosidase units.

For in vitro-coupled transcription/translation, we utilizedthe linked in vitro SP6/T7 transcription/translation nonradioactivekit (catalog no. 1 814 346; Roche). The system is based on invitro transcription of DNA sequences inserted downstream ofthe T7 promoter into capped RNA, using linearized or supercoiledDNA as the template. Without any further purification, the transcribedmRNA is subsequently translated in rabbit reticulocyte lysates.All the components required for the in vitro transcription reaction,as well as for the nonradioactive in vitro translation reaction,are provided in the kit, premixed and ready to use. Briefly,1 µg of linear DNA of each construct (E1/2/3, E3G, orE3A) was incubated in T7 transcription buffer at 30°C for15 min and frozen at –70°C, after a 10-µl aliquotwas taken for Northern blotting analysis. The samples were thenthawed, and a 10-µl aliquot from each construct was mixedwith 40 µl of the translation mix. The mixtures were incubatedfor 1 h at 30°C, and then 5 µl of the mixtures wasdiluted in 45 µl of luciferase reaction buffer, and theluciferase produced was measured using a luminometer.

Construction of a CMV-5' UTR E3G or E3A luciferase reporter system and polyribosome analysis. For the CMV promoter luciferase reporter system, we used thevector pFL-CMV as described in the previous section. E3 containingthe G or A polymorphism fragment in the pGEM TEasy vector wasdigested with SpeI/EcoRI and subcloned into the same SpeI/EcoRIdigestion site as that of the pFL-CMV vector. To analyze thetranslational efficiency of the E3 region, HEK293 cells werecotransfected with 0.5 µg of the recombinant E3-pFL-CMVvector together with 0.2 µg of the pRL-CMV vector (whichcontains the Renilla luciferase gene under the control of theCMV promoter). The Renilla luciferase served as an internalcontrol for transfection efficiency normalization, consideringthe pGL3B vector (Promega) only as a blank control. The transfectedcells were lysed at 3, 6, 12, and 48 h and diluted in 45 µlluciferase reaction buffer, and the luciferase produced wasmeasured using a luminometer.

For polyribosome analysis, approximately 5 x 10⁶ cells (HEK293cells) were transfected with 10 µg of pFL-CMV-5' UTR Gor pFL-CMV-5' UTR and incubated at 37°C for 48 h. The cellswere treated with 0.35 mM cycloheximide (Sigma, St. Louis, MO)for 10 min, washed with PBS, and lysed with lysis buffer (20mM HEPES-KOH [pH 7.4], 10 mM KCl, 5 mM MgCL2, 0.3% NP-40, 0.5mM dithiothreitol, 200 U/ml RNasin [Roche, Indianapolis, IN],and 15 µl/ml protease inhibitor mixture [Sigma], with50 mM EDTA [pH 8.0]). The cell lysates were scraped into a 1.5-mlmicrocentrifuge tube and passed five times or more through a27-gauge needle. Nuclei were pelleted by centrifugation at 12,000rpm for 20 min at 4°C, and the supernatant was layered ontop of a 20% to 60% linear sucrose gradient in gradient buffer(10 mM HEPES-KOH [pH 7.4], 150 mM KCl, 0.5 mM MgCL2, 1 mM dithiothreitol,50 U/ml RNasin). The gradient was made with Gradient Maker (catalogno. GM-100; CBS Scientific, Del Mar, CA). Gradients were centrifugedat 35,000 rpm for 2 h in a Beckman SW41 rotor and recoveredin 18 equal fractions. The optical density was measured at 254nm. Fractions were treated with DNase (GIBCO-BRL, Rockville,MD) and then extracted twice with phenol, and the RNA was precipitatedwith ethanol, washed twice with 70% ethanol, and stored at –80°C.

For Northern blotting analyses, the entire RNA sample from eachfraction was used. Samples were separated on 1.2% agarose gelscontaining 15% formaldehyde, transferred to Zeta-Probe membranes(Bio-Rad Laboratories, Hercules, CA), and incubated with a luciferase-specificDNA probe. The luciferase probe was prepared by using a randomprimer DNA-labeling system (Gibson-BRL) with a gel-purifiedLuc PCR fragment and [³²P]dCTP (GE Healthcare, Piscataway, NJ).The membranes were cross-linked using a UV cross-linker (Stratagene,La Jolla, CA) and hybridized to the ³²P-labeled probe. Afterundergoing hybridization and washing, the membrane was exposedto X rays at –70°C and developed.

Metabolic cell labeling and immunoprecipitation. The CCR5^–/– PBMCs were activated with PHA plus IL-2for 3 days and metabolically labeled with 200 µCi of [³⁵S]methionineand 200 µCi of [³⁵S]cysteine (Amersham Corp., Pittsburgh,PA) for 3 h, as described previously (2). The cell lysates wereimmunoprecipitated by using a previously described polyclonalantibody raised against the last 31 frame-shift amino acidsin the protein, the CCR5 ${Delta}$ 32-specific antibody (2). The resultingimmune complexes were reacted with protein A-Sepharose beads(GE Healthcare, Piscataway, NJ), washed three times, and analyzedby sodium dodecyl sulfate-polyacrylamide gel electrophoresis(SDS-PAGE). The gel was then dried and exposed to X-ray film.

RESULTS

Top

RESULTS

RT-PCR analysis of the CCR5 ${Delta}$ 32 transcripts in PBMCs from six HIV⁺ patients. To analyze CCR5 ${Delta}$ 32 RNA transcription, we first compared endogenousmRNA expression by RT-PCR analysis using poly(A) mRNA isolatedfrom the different CCR5^–/– PBMCs. Previous studiesdemonstrated multiple transcripts from the CCR5 promoter (30).Figure 1A shows the CCR5 promoter organization and the predictedmRNA transcripts as originally described (30). Mummidi et al.identified two CCR5 promoter regions: the P_U, a weak promoterwhich resides proximal to E1, and the P_D, a stronger promoterwhich is located upstream of E3 (30). Both CCR5 promoter regionshave been demonstrated to be active in CD4⁺ T lymphocytes (23).Figure 1A shows a schematic structure of the predicted CCR5 ${Delta}$ 32mRNA transcripts as previously reported (30) and the locationof the 59537-G/A polymorphism. This polymorphism correspondsto –1951, according to the numbering system that designatesthe first nucleotide of the translation start site of CCR5 ${Delta}$ 32as position 1 and the nucleotide immediately upstream of thissite as position –1.

RT-PCR analysis using primers that amplify the large transcript(CCR5A in reference 30) indicated the absence of this mRNA speciesin four HIV⁺ CCR5^–/– samples (Fig. 1B, E2 to E4).However, RT-PCR analysis using primers that map to the E3 region(Fig. 1B, E3 to E4) indicated no significant differences inthe abundance of this RNA species among the tested CCR5^–/–samples (Fig. 1B, E3 to E4). The quality of samples loaded wasverified by amplification of β-actin mRNA from the sameRNA quantity (Fig. 1B, bottom gel). The absence of a detectablelarge transcript in the HIV⁺ CCR5^–/– PBMCs and thefact that it is unknown which transcribed mRNA species are translatedin vivo prompted us to examine the promoter regions for potentialpolymorphisms.

Polymorphisms in CCR5 ${Delta}$ 32 promoter regions and their effect on reporter gene expression. We sequenced both promoter regions in the CCR5^–/–samples and identified a number of sequence polymorphisms inall CCR5^–/– PBMC samples, as indicated in Table1. To determine whether the promoter polymorphism had any effecton the transcriptional activity of CCR5 ${Delta}$ 32 promoter DNA, fragmentscorresponding to the P_U and P_D regions cloned from the CCR5^–/–samples were cloned upstream of the firefly luciferase reportergene (Fig. 2A).

View this table:
[in this window]
[in a new window]

TABLE 1. Sequence polymorphisms of the CCR5 promoter region^a

View larger version (44K):
[in this window]
[in a new window]

FIG. 2. Transcriptional activities of the CCR5 ${Delta}$ 32 promoter variants. (A) Schematic diagram illustrating the different promoter constructs used in subsequent experiments to measure the transcriptional activities of the promoter variants described in this study. The promoter regions identified by the indicated GenBank nucleotide numbers were cloned upstream of the firefly luciferase, using the simian virus 40 polyadenylation signal (SV40 PA) for the 3' end processing and the termination signal. The 59537 polymorphism corresponds to the reported –1951 position relative to that of the translation initiation codon. (B) The CCR5 Pr-Luciferase plasmids were individually cotransfected with a plasmid containing the Renilla luciferase into HEK293 cells and incubated for 48 h at 37°C. Transcriptional activity of the cloned promoters was measured by calculating the ratio of firefly:Renilla luciferase production. (C) Northern blotting analysis of the luciferase mRNA produced in transfected HEK293 cells. Blots were probed with a fragment corresponding to the firefly luciferase. The broken line represents the background levels measured in empty vector-transfected HEK293 cells with no luciferase production. The blots were also probed with a DNA fragment corresponding to β-actin to control for gel loading. Rel., relative levels.

In order to determine whether these polymorphisms impact CCR5promoter activity, we utilized a reporter assay system thathas been used previously to analyze the transcriptional activityof CCR5 promoter (17, 30). We generated promoter constructsthat contained either P_U, P_D, or both promoter regions joinedin one fragment upstream of the firefly luciferase (Fig. 2A).The different promoter plasmid constructs were cotransfectedinto HEK293 cells along with plasmid DNA encoding the Renillaluciferase under the CMV promoter. The results demonstrate nosignificant differences in CCR5 ${Delta}$ 32 promoter activities associatedwith these polymorphisms (Fig. 2B). Northern blotting analysisof the mRNA produced in HEK293 cells transfected with the differentpromoter constructs confirmed the relative abundance of theexpressed RNA species driven from each promoter construct (Fig.2C). These results demonstrate that the observed promoter polymorphismshad no significant impact on CCR5 ${Delta}$ 32 transcriptional activity.

Translational analysis of the 5' UTRs of CCR5 ${Delta}$ 32. Since the CCR5 ${Delta}$ 32 59537-G/A polymorphism is located in E3 ofthe CCR5 gene, it is reasonable to assume an effect on translationalefficiency. Since it is located in E3, this polymorphism (Fig.3A) is present in the 5' UTR of all transcripts that are transcribedfrom the CCR5 promoter (30). It has been demonstrated previouslythat complex splicing and multiple transcriptional start sitesresult in several CCR5 transcripts that differ in their 5' UTR(30). To analyze the translational efficiency of the different5' UTRs, the sequences of the different 5' UTRs were constructedas previously described (30) and were cloned upstream of thefirefly luciferase under the control of the T7 promoter (Fig.3A). Production of the mRNA was activated by infection witha vaccinia virus recombinant encoding T7 RNA polymerase. The5' UTR constructs were individually cotransfected with the pT7-LacZvector (containing LacZ under T7 promoter control) into HEK293cells, and the translation efficiency of the UTR was measuredas the ratio of luciferase:β-galactosidase. The resultsindicate that the shortest 5' UTR (E3) was the most efficientat translation and produced the highest luciferase:β-galactosidaseratio (Fig. 3B). Low translation efficiency was observed consistentlywith the longer 5' UTR that had E2 and E3 (E2/3) and consistentlywith the background levels obtained with the longest UTR (i.e.,E1/2/3) (Fig. 3B). To determine the translational efficiencyof the E3 5' UTR containing the A polymorphism (E3A), we comparedits efficiency with that of its counterpart that has a G atthat position (E3G). We consistently found that E3G has a significantlyhigher translational efficiency than E3A (Fig. 3B).

View larger version (44K):
[in this window]
[in a new window]

FIG. 3. Translational analysis of the 5' UTRs of CCR5 ${Delta}$ 32. (A) Schematic diagrams of the maps of the pGL3-based plasmid construct containing the different 5' UTRs. The location of the 59537-G/A polymorphism in the CCR5 ${Delta}$ 32 promoter is indicated. The 5' UTR regions of CCR5 ${Delta}$ 32 (as originally described by Mummidi et al. [30]) were individually cloned under the T7 promoter in the pGL3 vector (Promega) and upstream of the firefly luciferase gene. The plasmids containing the different T7-UTR-luciferase constructs were cotransfected along with pT7-lacZ into HeLa cells, and expression of the constructs was activated by infection with vTF7-3 encoding the T7 RNA polymerase. (B) Expression efficiency of the constructs transcribing the luciferase gene linked with the different 5' UTRs. Transfection efficiency was determined by cotransfecting pT7-lacZ that contains β-galactosidase under T7 promoter (translation efficiency was calculated as the luciferase/β-gal ratio). The experiment was reproduced at least three times. (C) In vitro transcription/translation of the indicated 5' UTR constructs. The indicated 5'UTR plasmid construct (1.0 µg) was incubated in rabbit reticulocyte lysate (Promega) supplemented with T7 polymerase. Results are shown for the luciferase produced in one experiment performed in duplicate and reproduced three times. (D and E) Northern blotting analysis of the firefly luciferase mRNA produced in transfected HeLa cells (D) and as a result of in vitro transcription (E). Rel., relative levels.

To verify the results obtained with the T7 polymerase system,we used an in vitro-coupled transcription/translation system(Promega). The translational efficiency of the E3G and E3A UTRconstructs was examined by measuring the luciferase activityproduced with rabbit reticulocyte lysate supplemented with T7RNA polymerase. The results of these experiments consistentlyproduced translational efficiency of E3G that was higher thanE3A (Fig. 3C). The construct containing the E1/2/3 UTR had avery low translational efficiency in these experiments (Fig.3C). Northern blotting analysis confirmed that the observedresults were not due to higher mRNA produced from the E3G-basedconstructs (Fig. 3D) or the E3G RNA produced in the in vitro-coupledtranscription/translation system (Fig. 3E). Taken together,these results suggest that the mRNA containing the E3 UTR isthe RNA species most efficiently translated and that the G/Apolymorphism in E3 significantly affects translational efficiency.

The 59537-G 5' UTR but not the 59537-A 5' UTR augments ribosomal association. To determine the in vivo effect of the G/A polymorphism, wetransfected 293 cells with the constructs containing the polymorphic5' UTR under the control of the CMV promoter (Fig. 4A). Thedifferent 5' UTR constructs were individually transfected alongwith the plasmid DNA encoding Renilla luciferase under the CMVpromoter. Transfected cells were lysed after incubation for3, 6, 12, and 48 h, and luciferase activity was measured. Theresults indicate that the E3A 5' UTR construct produced significantlylower luciferase activity than the E3G 5' UTR, even after 48h of incubation (Fig. 4B).

View larger version (43K):
[in this window]
[in a new window]

FIG. 4. Expression efficiency and distribution of transcripts containing the CCR5 ${Delta}$ 32 5' UTR regulatory elements in polysome gradients. (A) The structural map of the reporter constructs containing either E3G or E3A 5' UTR upstream of the firefly luciferase reporter gene is shown. (B) The protein expression efficiency of the two E3G and E3A 5' UTR constructs in transfected HEK293 cells is shown. To normalize for the transfection efficiency, the cells were transfected with another plasmid containing Renilla luciferase under the control of the CMV promoter. The cells were lysed and analyzed for luciferase expression at the indicated times posttransfection. (C and D) Polysome analysis of luciferase mRNA extracted from transfected HEK293 cells that were transiently transfected with 2.5 µg of E3G or E3A 5' UTR reporter constructs and incubated for 48 h. Cell lysis and sucrose gradient sedimentation analyses of polysomes were performed as described in Materials and Methods. Fractions (0.5 ml each) were collected, and their optical density was measured at a wavelength of 254 nm (C). The fractions were labeled from top to bottom to indicate their positions in the gradient. RNA was extracted from each fraction and subjected to Northern blotting analysis (D). The blots were hybridized with a probe specific to the firefly luciferase gene. The two blots at the bottom show the polysomal profiles as a result of EDTA treatment. The addition of 50 nM EDTA results in the dissociation of ribosomes from mRNA and a shift of luciferase mRNA to the top of the gradient.

To directly examine the effect of the 5' UTR polymorphism ontranslational efficiency in vivo, we isolated poly(A) mRNA fromcells transfected with the chimeric 5' UTR constructs. Transcriptionallyactive ribosomes (polysomes) were separated from the nonactive(free ribosomal), using a sucrose gradient. Cytoplasmic extractswere prepared from the transfected cells and analyzed. Eighteenfractions were collected, and absorbance was measured (Fig.4C). The RNA extracted from the gradient fractions was analyzedby Northern blotting hybridization using a probe correspondingto the luciferase-coding region (Fig. 4D). To analyze the cytoplasmicdistribution of luciferase RNA, all 18 fractions of the ribosomalprofile were subjected to Northern blotting analysis. The resultsdemonstrate that the chimeric RNA with the E3G 5' UTR was completelyassociated with actively translating polysomes, while the chimericRNA carrying E3A 5' UTR was partially associated with polysomes,indicating a recruitment defect. Figure 4C shows the absorbanceprofiles of the 18 fractions.

Figure 4D shows that the chimeric mRNA with the 5' UTR (E3G)is almost completely associated with actively translating polysomes,while the mRNA with the polymorphism (E3A) is only partiallyassociated with the polysome fractions. Polysome disruptionby EDTA treatment results in a dramatic shift of the sedimentationpattern from the heavy polysomes to the lighter and free fractionsof the gradient (Fig. 4D, bottom two blots). Treatment withEDTA resulted in the disappearance of the mRNA from the polysomefraction, confirming the specificity of this assay. These resultssuggest a translational control mechanism in two HIV⁺ CCR5^–/–individuals with a 59537-A polymorphism in the 5' UTR.

Endogenous CCR5 ${Delta}$ 32 protein expression in the HIV⁺ PBMC samples correlate with the in vitro translation data. We have previously reported that the CCR5 ${Delta}$ 32 protein was notdetected by Western blotting in four of the six HIV⁺ samples(1). Since the CCR5 ${Delta}$ 32 protein was not detected in the two samples(sample 1 and sample 2) that have a 59537-A/A genotype, we examinedthe endogenous mRNA expression by RT-PCR and compared that tothe three uninfected CCR5^–/– samples with a 59537-G/Ggenotype. The results demonstrated abundant mRNA expressionin the two samples with the 59537-A/A genotype that was similarin intensity to that obtained with the samples with the 59537-G/Ggenotype (Fig. 5A). We then examined the CCR5 ${Delta}$ 32 protein expressionin PBMC samples radiolabeled with [³⁵S]methionine and [³⁵S]cysteine.The radiolabeled proteins were immunoprecipitated with antibodiesthat specifically detect the CCR5 ${Delta}$ 32 protein. The immunoprecipitatedproteins were analyzed with SDS-PAGE and detected by autoradiography.The results show high CCR5 ${Delta}$ 32 protein expression in the CCR5^–/–PBMC samples with a 59537-G/G genotype and relatively low levelsin the samples with a 59537-A/A genotype (Fig. 5B). Immunoprecipitatingequivalent amounts of cell lysates with antibodies specificto β-actin revealed equivalent gel loading. These resultsstrongly support our findings of the translational efficiencyof the 5' UTR constructs and indicate that the expression ofthe CCR5 ${Delta}$ 32 protein is critical for the resistance phenotype.

View larger version (58K):
[in this window]
[in a new window]

FIG. 5. Endogenous expression of CCR5 ${Delta}$ 32 mRNA and protein in PBMCs isolated from CCR5^–/– individuals with 59537G/A polymorphism. (A) RT-PCR analysis of endogenous mRNA expression. Poly(A)⁺ mRNA isolated from PHA plus IL-2-activated PBMCs was used as a template to amplify E3 to E4 transcripts in the indicated samples. The amplified products were analyzed with a 1% agarose gel. Amplification of β-actin was performed as a control. (B) The PHA plus IL-2-activated PBMC samples were radiolabeled with [³⁵S]methionine and [³⁵S]cysteine for 6 h. Lysates of the radiolabeled cells were prepared, and equivalent amounts were immunoprecipitated with either anti-CCR5 ${Delta}$ 32 antibodies or antibodies to β-actin. The positions of the CCR5 ${Delta}$ 32 and β-actin proteins are indicated. Uninfected (UN)^–/– refers to PBMCs from HIV^– CCR5^–/– individuals. Infected (IN)^–/– refers to PBMCs from HIV⁺ CCR5^–/– patients.

DISCUSSION

Top

DISCUSSION

Polymorphisms in CCR5 have been implicated in modulating AIDSdisease progression and resistance to HIV-1 infection (32).Several studies have suggested that polymorphisms in the CCR5promoter region may result in different levels of mRNA transcriptionthat might explain some of the observed variability in the progressionto AIDS (19, 25, 26, 29). The literature reported differentdata concerning CCR5 promoter polymorphisms. While some studiesreported an association between the CCR5 promoter polymorphismsand the transcription efficiency (26), others have found thatpromoter polymorphisms had no significant effects on the CCR5promoter activity in vitro (24). Here, we report a number ofpromoter polymorphisms in a number of HIV⁺ CCR5^–/–individuals, none of which had a significant impact on the transcriptionalactivity of the promoter in vitro. We report for the first timethat one of these polymorphisms that are located in E3 of theCCR5 ${Delta}$ 32 promoter (and is a part of the 5' UTR) had a significanteffect on the translation efficiency of the CCR5 ${Delta}$ 32 protein.Our results do not absolutely prove that the 59537 variant issolely responsible for HIV infection of these CCR5 ${Delta}$ 32 individuals.Other unknown host factors might also be involved in the lossof the CCR5 ${Delta}$ 32 protective effect in vivo.

The 59537-A polymorphism in the 5' UTR was associated with weaktranslational activity of the CCR5 ${Delta}$ 32 protein in two differentunrelated HIV⁺ CCR5^–/– individuals. This findingis supported by immunoprecipitation experiments showing verylow levels of the endogenous CCR5 ${Delta}$ 32 protein in the two samplesof HIV⁺ PBMCs. The two samples with the 59537-A genotype hadsignificantly lower levels of the CCR5 ${Delta}$ 32 protein than the threeother CCR5^–/– PBMC samples with the 59537-G genotype.Mutations that cause disease through increased or decreasedefficiency of mRNA translation have been extensively documented,defining translational pathophysiology as a novel mechanismof human disease (reviewed in references 20 and 21). For example,it has been demonstrated that a G-to-C transversion in the 5'UTR in the second exon of the BRCA1 gene is able to affect mRNAtranslation in vitro and in vivo (35), possibly affecting tumordevelopment and/or progression.

The low translational efficiency of the 59537-A 5' UTR may explainthe low abundance of the CCR5 ${Delta}$ 32 protein in two infected CCR5^–/–individuals. The reporter assays implied that 59537-G in the5' UTR is associated with the higher efficiency of CCR5 ${Delta}$ 32 proteinexpression that correlated with resistance to HIV-1 infection.We have previously reported that the protective effect of theCCR5 ${Delta}$ 32 protein in the two HIV⁺ samples with a 59537-A genotypecould be rescued by the expression of recombinant CCR5 ${Delta}$ 32 protein(1), implying that the absence of the protective effect is stronglyassociated with the low levels of CCR5 ${Delta}$ 32 protein expression.The low expression levels of the CCR5 ${Delta}$ 32 protein in the two HIV⁺CCR5^–/– subjects could not be due to a change theCCR5 ${Delta}$ 32 protein amino acid sequence, since no polymorphisms intheir open reading frames were detected (data not shown).

Previous studies described an association between acceleratedAIDS progression and homozygosity for the CCR5P1/P1 genotype(25). Martin et al. described a polymorphic region in the CCR5promoter, from E1 to E3 (GenBank no. 58934 to 59537). Analysisof that region revealed 10 polymorphic nucleotide residues thatspecify 10 CCR5 promoter region haplotype alleles designatedCCR5P1 through CCR5P10. The CCR5P1/P1 genotype contains a Gat position 59537 (59537-G/G genotype) and has been found tobe more frequent than the 59537-A/A genotype (25). We foundthat the presence of a G in the 5' UTR of a reporter gene resultsin more efficient protein translation. Martin et al. reportedthat the 59537-G/G genotype in CCR5^+/+ individuals is associatedwith accelerated disease progression (25). It is possible thatthe increased CCR5 expression in vivo might result in a largernumber of cells that initially get infected with HIV-1. It isalso possible that other factors may be involved in the accelerateddisease progression of these individuals and that the 59537-G/Ggenotype is only one of the factors involved.

It is not clear why four of the HIV⁺ CCR5^–/– samplesdid not express the E2-to-E4-related transcript. This transcriptis similar to the previously described large CCR5A transcript(30). The E2-to-E4-related transcript could also represent oneof the truncated mRNA species that has the E2 and E3 regionsfused, as previously reported (30). Analysis of the effect ofthe promoter polymorphisms in the reporter gene activation assaydid not reveal significant differences for the observed polymorphicpromoter regions. The protein translation experiments consistentlydemonstrated that the CCR5 ${Delta}$ 32 transcripts with the shortest 5'UTR (E3 fused to E4) results in the most efficiently translatedprotein. We were unable to detect a significant translationalefficiency for the longest 5' UTR region that has the threeexons fused (E1/2/3), suggesting that it is not a major translationproduct in vivo.

The observation that the 59537-G/A variant was not associatedwith a transcriptional defect does not exclude a possible rolefor other polymorphisms in promoter activity. Previous studiessuggested that a 59029-G/A polymorphism in the CCR5 promoterwas associated with a differential regulation of CCR5 transcription(26). Other studies demonstrated that some individuals who practicehigh-risk sexual behavior may resist infection because theyposses one CCR5 ${Delta}$ 32 allele plus a 59029-A/G genotype combination,which is apparently associated with relatively weak CCR5 expression(18). This newfound complexity in the vagaries of CCR5 ${Delta}$ 32 expressionand function might increase the clarity of the genomic analysisof risk factors for HIV infection and disease progression.

ACKNOWLEDGMENTS

We thank Hassan Naif for providing VH and VM samples, HaynesSheppard and Susan Buschbinder for the 164 sample, and the MACSfor the HIV⁺ and HIV^– CCR5^–/– samples. Wealso thank Ken Cornetta and Jeremy Sanford for critical comments.

This study was supported by NIH grant A152019-01 to G.A. Q.J.was supported by a scholarship from the Chinese ScholarshipCouncil, Beijing, China.

Human PBMC samples were obtained under the Indiana Universityexempt IRB study no. 4.

FOOTNOTES

* Corresponding author. Mailing address: Department of Microbiology and Immunology, Indiana University School of Medicine, 635 Barnhill Drive, Rm. 420, Indianapolis, IN 46202. Phone: (317) 278-3698. Fax: (317) 274-7592. E-mail: galkhati{at}iupui.edu

Published ahead of print on 19 December 2007.

Present address: Department of Neurology, Nanjing Medical University,Jiangsu Province, China 210029.

REFERENCES

Top