To get insight into potential compartmentalization of HIV-1 replication at different locations within the gut, we focused on analyzing the Nef and RT encoding regions of HIV-1. These regions were amplified by nested PCR from DNA isolated from different gut tissues (esophagus, stomach, duodenum and colorectum) and peripheral blood lymphocytes (PBL). The samples were obtained from a previously described cohort of HIV seropositive homosexual men followed at the Southern Alberta Clinic (SAC), Calgary, Alberta, from 1993 to 1996 627. This cohort was recruited prior to the introduction of HAART (Highly Active Antiretroviral therapy) at the SAC in late 1997. Eight patients at various clinical stages of HIV infection/AIDS were selected and gut tissue samples from one visit were analyzed (Table 1). Although cDNA was synthesized and viral sequences could be detected with our real-time RT-PCR analysis using small amplicons (discussed in sections below), the sequences spanning the entire viral regions of interest were the most readily and consistently amplified from DNA. Therefore, we focused on these proviral sequences for the current study. Analysis of these sequences also exploits the "banking-effect" of proviral DNA in the chromosomal DNA of different cell populations offering some insight into "the history" of the infection 28 and seeding of the gut tissues. We chose to analyze the HIV-1 Nef and RT protein encoding sequences as these proteins have various effects on viral replication and the RT region has been shown to evolve by tissue compartment 8293031. Both proteins are targeted by the cellular immune response 323334, and therefore suitable targets to determine how HIV-1 evolves in different parts of the gut. The Nef protein is highly variable 32, but is relatively more conserved than the highly diverse envelope protein 35, which could make it easier to detect phylogenetic relationships within the patient population. Finally, the Nef protein has been implicated as an important pathogenic determinant of HIV-1 363738394041424344454647, and its analysis could shed some light on the evolution of pathogenic HIV-1 strains in the gut.

HIV-1 viral sequences were amplified from PBL and biopsy tissue DNA using our nested PCR protocol. For seven patients RT and Nef encoding sequences were obtained from 3 or more tissues (gut tissues and/or PBL), while for one patient (#8) only sequences from two tissues could be obtained (stomach and esophagus). Neighbour-Joining trees revealed clustering of the nef sequences by individual patient (bootstrap values of >90) (Fig. 1). The clones of the Nef encoding sequences also clustered by gut tissue from which they were obtained (i.e. esophagus, stomach, duodenum and colorectum). No mixed clustering with sequences of different tissues was observed, indicating that within a patient distinct HIV-1 quasispecies were found within different parts of the gut. Analysis of the clustering pattern for the patients for which we were able to obtain viral sequences from PBL revealed clustering of these sequences with each other (bootstrap value >90) (Fig. 1). This indicated that the viral quasispecies found in the periphery were different from those found in the gut. No obvious phylogenetic relationship of PBL sequences with sequences of a particular gut tissue (i.e. esophagus, stomach, duodenum or colon) was observed among the different patients. In contrast to the Nef encoding region, similar tight clustering for the RT encoding region was not found for any of the patients (Fig. 2). However, for various patients a large number of the RT encoding sequences from the esophagus and stomach clustered together. The latter could suggest that there may be a selection for a particular RT encoding sequence in these tissues among patients.

To corroborate our observations, the sequences from all clones for the Nef and the RT encoding region were used to derive consensus sequences for each tissue compartment for each patient. Bootstrap analysis of these consensus sequences revealed clustering of nef sequences by patient, and for two patients by upper (esophagus and stomach) and lower (duodenum and colorectum) gut tissue compartment (Fig. 3A) (bootstrap value >90). This suggested a clustering of nef sequences by upper or lower GI-tract in select patients. Similar to our previous results, clustering was again not observed for the RT encoding region (Fig. 3B). Analysis of the consensus Nef protein sequences obtained (Fig. 4) did not reveal any particular signature sequences for specific gut tissues. Taken together, these observations indicated that viral replication in the gut was compartmentalized, resulting in different HIV-1 quasispecies populating different parts of the gut.

To determine to what extent viral diversity differed among the different gut tissues, the mean total (d), and nonsynonymous (d_N) pair-wise distances were calculated for the Nef and RT encoding sequences obtained from the esophagus, stomach, duodenum and colorectum tissues of all patients (Fig. 5). Significantly lower d and d_N values (i.e. codon/amino acid changing substitutions) were observed for the RT encoding region for both the esophagus and the duodenum, compared to the stomach and colorectum (p < 0.001 and p < 0.05, respectively). In contrast, for the Nef encoding region, a significantly higher d value was observed in both duodenum and colorectum (p < 0.05). Further analysis of the Nef encoding region revealed a higher d_N value in the colorectum (p < 0.05), suggestive of a more diverse Nef protein in the colorectum. Similar to the analysis of the consensus Nef protein sequences (Fig. 4), analysis of all the inferred protein sequences for both the Nef and RT protein obtained from the different patients did not reveal any signature sequences for any of the tissue compartments, nor were there any obvious differences in domains important for protein function 36373839484950(data not shown).

The differences in viral diversity suggested that HIV-1 evolved to varying degrees in the different gut tissues. A previous report observed differences in viral loads between blood and colorectum 51, indicative of differences in viral growth between these compartments. To extend the observations obtained with our phylogenetic analysis, we used real-time RT-PCR to determine the relative levels of viral genomic RNA in our patients in esophagus, stomach, duodenum and colorectum (Fig. 6A). The fold increase in transcript levels was expressed relative to the levels observed in the esophagus, since in all patients we found the lowest level of transcript in this tissue. Significantly higher levels of viral genomic RNA in the colorectum compared to esophagus were observed in the 8 patients analyzed in this study (p < 0.05), suggesting HIV-1 replication differs in different parts of the gut. Finally, as our results indicated a more diverse Nef protein in the colorectum, we used real-time RT-PCR to analyze the expression of all HIV-1 RNA transcripts (genomic and mRNAs) containing the Nef protein open reading frame (Fig. 6B), as well as nef gene specific mRNA transcripts 52(Fig. 6C) in the different gut tissues. Indeed, an increased expression of viral mRNA and nef gene specific mRNA transcripts (p < 0.05) was observed among all patients in the colorectum.

Patients were enrolled from a previously described cohort of HIV seropositive homosexual men followed at the Southern Alberta Clinic (SAC), Calgary, Alberta, from 1993 to 1996 627. This study was reviewed and approved by the Office of Medical Bioethics of the University of Calgary and all patients signed an informed consent at enrollment. Patients were prospectively followed and laboratory testing included plasma viral load and CD4 counts, for each patient during each visit. In addition upper and lower endoscopies were performed in order to harvest tissue for further testing. This cohort was recruited prior to the introduction of HAART (Highly Active Antiretroviral therapy) at the SAC in late 1997. Antiviral therapies during the study consisted of no treatment, monotreatments with AZT, DDI, DDC, D4T, and 3TC or combinations thereof.

Gastrointestinal tract biopsies were taken during the study period at the level of the esophagus, stomach, duodenum and colorectum. The biopsies taken via endoscopic procedure were shipped cryopreserved and stored at -70°C, within 1 hour of collection 6. Plasma or serum was collected and stored at -70°C and peripheral blood lymphocytes (PBL) were isolated and stored in liquid nitrogen627. Eight patients at various clinical stages of HIV infection/AIDS were selected. All patients had a CD4 count of lower than 400 at the time of biopsy (Table 1). Nucleic acid sequence analyses were performed on the gut biopsies taken at one visit for this subset of patients.

Chromosomal DNA was isolated from the tissue biopsies from each patient using Trizol Reagent (Invitrogen, Burlington, ON). The HIV-1 proviral DNA regions encoding the Nef (negative factor) and RT (reverse transcriptase) protein were amplified using a nested PCR and 0.2 μg of chromosomal DNA as template. The nested PCR protocol for both the first and second round PCR, consisted of an initial denaturation step of 5 min at 94°C, 45 cycles of 1 min at 95°C, 1 min at the annealing temperature of the primer set used, 2 min at 72°C, followed by a final extension step of 10 min at 72°C. For the nef region, the primers used for the first round PCR were NEF8678F 5'-GTA GCT GAG GGG ACA GAT AG-3' and NEF9540R 5'-AGG CTC AGA TCT GGT GTA AC-3' (T_m = 50°C), and for the second round NEF8748F 5'-GAA GAA TAA GAC AGG GCT-3' and NEF9425R 5'-AGT CCC CAG CGG AAA GTC CC-3' (T_m = 40°C) 62. For the RT region, the first round primers consisted of RT2470 5'-GTA CAG TAT TAG GAC CTA CAC CTG-3' and RT3261 5'-ATC AGG ATG GAG TTC ATA ACC CAT CCA-3' (T_m = 55°C), and for the second round consisted of RT2604 5'-CCA AAA GTT AAA CAA TGG CCA TTG ACA-3' and RT3251 5'-AGT TCA TAA CCC ATC CAA AG-3' (T_m = 55°C). All viral sequences were amplified using pfu polymerase to reduce introduction of mutations during amplification. In addition, dilution experiments of the template were performed to prevent selective amplification of the most dominant viral sequences at the expense of less frequent viral sequences, and thus skewing sampling of the viral quasispecies. For these experiments, 2 to 10 fold dilutions of the template DNA were made. The dilutions at which still an abundant PCR product could be obtained were used for our sequence analysis.

The resulting Nef and RT encoding PCR fragments were isolated from gel using the Qiaquick gel extraction kit (Qiagen Inc., Mississauga, ON) and directly sequenced to obtain overall consensus sequences for the different tissue compartments. In addition, the gel purified PCR fragment were cloned by filling in the incomplete fragment ends with Klenow, phosphorylation using T4 polynucleotide kinase, and insertion into the EcoRV site of pSL1180. All enzymes were obtained from Invitrogen (Burlington, ON) and New England Biolabs (Pickering, ON) and used according to the manufacturer's specifications. Where possible up to five to ten clones per patient/sample were analyzed to determine the composition of the viral quasispecies for the Nef and RT encoding region. DNA sequences were determined by automated sequencing on an ABI sequencer (Applied Biosystems, Streetsville, ON) and a Li-Cor 4300 DNA Analysis sequencing system (Li-cor Biosciences, Lincoln, NE) according to the manufacturers' protocols and reagents. All sequences have been submitted to Genbank (accession numbers EF656787 to EF657121)

The inferred amino acid sequence for each cloned nucleic acid fragment was obtained for each sample. These protein sequences were screened for the integrity of the coding sequences (i.e. stop codons, deletions and/or insertions), as well as common amino acid motifs between the different patients and samples. In addition, changes in amino acid residues or motifs that have been shown to be important for function of the Nef and RT protein were analyzed 36373839484950. The nucleic acid sequences were subjected to an extensive phylogenetic analysis, using the MEGA version 3.1 software package 65. Neighbor-Joining trees were constructed using the Kimura-2-parameter model with 5000 replicates for the bootstrap analysis. Bootstrap values of >70 were considered significant. Reference sequences of prototypic HIV isolates were included to rule out possible contamination with laboratory strains. The MEGA software was also used to calculate mean total (d), nonsynonymous (d_N, i.e. codon changing substitutions) and synonymous (d_S, i.e. non-codon changing substitutions) distances for the nucleic acid sequences from all patient and tissue compartments as well as for individual tissue samples. All distances and phylogenetic analyses were performed through pair-wise comparison and pair-wise gap stripping of each sequence.

Total RNA was isolated from the gut tissues using Trizol (Life Technologies, Burlington ON) and used as template for the synthesis of cDNA using a previous described protocol 6266. Real-time quantitative PCR using this cDNA and the iCycler IQ system (Bio-Rad, Mississauga, Canada) was performed as follows. The cDNA was diluted 1/1 with water, and 5 μl was used per PCR reaction. Each 25 μl reaction contained 5 μl of cDNA, 11.5 μl of Supermix (Bio-Rad), 6.5 μl of SYBR-Green (1/50,000 dilution; Bio-Rad), 1 μl of fluorescein (1/10,000; Bio-Rad), and 1 μl of primer mix. The amplification protocol consisted of an initial denaturation step of 3 min at 95°C, followed by 45 cycles of 95°C for 30 s, 30 s at the annealing temperature (T_m) of the primers, and 72°C for 30 s, and a final elongation step of 72°C for 1 min. The amplicons were subjected to a melt curve analysis to ensure proper amplification, in which the temperature was raised from 65 to 99°C in 1°C increments, and data acquired for 8 s at each temperature increment. The viral genome, nef containing RNA, and nef mRNA levels were normalized against GAPDH (glyceraldehyde-3-phosphate dehydrogenase) and expressed as relative fold change compared to the RNA levels in the esophagus. All real-time RT-PCR quantifications were performed in duplicate and repeated with two different batches of cDNA. The primers used were: nef Forward (NEF8748F) 5'-GAA GAA TAA GAC AGG GCT-3', Reverse (NEF8879R) 5'-TCC CAC CCC ATC TGC TGC TGG-3' (Tm = 50°C, 2.5 μM each primer) 62, HIV-1 gag Forward 5'-TGC TAT GTC AGT TCC CCT TGG TTC TCT-3', and Reverse 5'-AGT TGG AGG ACA TCA AGC AGC CAT GCA AAT-3' (Tm = 60°C, 2.5 μM each primer)67, nef mRNA Forward 5'-GGC TAA CTA GGG AAC CCA CTG G-3', Reverse 5'-CAG GGA AGT AGC CTT GTG TGT GG-3' (Tm = 50°C, 10 μM each primer) 52, GAPDH Forward 5'-GAA GGT GAA GGT CGG AGT C-3', Reverse 5'-GAA GAT GGT GAT GGG ATT TC-3' (Tm = 50–60°C, 2.5 μM each primer).

All statistical analyses were performed using Graphpad InStat Version 3.01. (GraphPad Software, San Diego, CA) and p values < 0.05 were considered significant.