|Home | About | Journals | Submit | Contact Us | Français|
Retroviral replication proceeds through the formation of a provirus, an integrated DNA copy of the viral RNA genome. The linear cDNA product of reverse transcription is the integration substrate and two different integrase activities, 3′ processing and DNA strand transfer, are required for provirus formation. Integrase nicks the cDNA ends adjacent to phylogenetically-conserved CA dinucleotides during 3′ processing. After nuclear entry and locating a suitable chromatin acceptor site, integrase joins the recessed 3′-OHs to the 5′-phosphates of a double-stranded staggered cut in the DNA target. Integrase functions in the context of a large nucleoprotein complex, called the preintegration complex (PIC), and PICs are analyzed to determine levels of integrase 3′ processing and DNA strand transfer activities that occur during acute virus infection. Denatured cDNA end regions are monitored by indirect end-labeling to measure the extent of 3′ processing. Native PICs can efficiently integrate their viral cDNA into exogenously added target DNA in vitro, and Southern blotting or nested PCR assays are used to quantify the resultant DNA strand transfer activity. This study details HIV-1 infection, PIC extraction, partial purification, and quantitative analyses of integrase 3′ processing and DNA strand transfer activities.
The vast majority of animal viruses replicate efficiently while maintaining their genomes separate from those of their hosts. Retroviruses, by contrast, are the only animal viruses that must intertwine their genomes with their hosts. The integrated form of the virus, the provirus, is required for efficient gene expression and ensures for equal segregation of the genetic material to both daughter cells upon division.
The retroviral genome is RNA. Retroviruses must therefore convert their genomes into double-stranded cDNA as a prelude to integration. The viral-encoded integrase enzyme then catalyzes two distinct endonucleolytic reactions en route to provirus formation. In the first reaction, which is called 3′ processing, integrase nicks the linear ends of the reverse transcript adjacent to the conserved sequence 5′-CA-3′, which in the majority of cases hydrolyzes a dinucleotide from each 3′ end [1–5] (Fig. 1). The second reaction, DNA strand transfer, links the viral and host DNAs together. Integrase uses the recessed CAOH ends to cleave the phosphodiester backbone of chromosomal DNA in a staggered fashion, which at the same time joins the viral 3′ ends to the resulting 5′-phosphates . The DNA recombination intermediate, with free viral DNA 5′ ends abutting single-stranded gaps [1, 2, 5], is repaired by host cell enzymes to yield the provirus flanked by the duplicated sequence of the staggered cut made during DNA strand transfer (Fig. 1).
Integrase proteins purified from a variety of recombinant sources display 3′ processing and DNA strand transfer activities in vitro. These simplified biochemical assays were central for dissecting critical aspects of enzyme function, as well as establishing high throughput screens for the identification of integrase inhibitors (refer to chapters by Grandgenett, Merkel, and Grobler within this Methods issue). During infection, integrase functions in the context of the preintegration complex (PIC), a relatively large nucleoprotein structure that is derived from the core of the incoming virion [7–9]. Following extraction, levels of integrase 3′ processing activity can be quantified by digesting the purified viral cDNA with restriction endonucleases to yield approximate 100 bp end regions, and analyzing the resulting structures after denaturation for dinucleotide loss from the U5 plus- and U3 minus-strands [1–5, 9] (Fig. 2). Native PICs moreover support DNA strand transfer activity in vitro [1, 2, 5, 10–12]. These seminal studies revealed several unknown mechanistic aspects of retroviral integration, including that (i) it occurred in the context of a higher-order nucleoprotein complex, (ii) proceeded in the absence of an added high-energy cofactor like ATP, and (iii) the immediate precursor was linear cDNA and not circular DNA forms that also arise during infection. PIC analyses remain central to investigations of integrase catalytic function in the context of the infected cell. This paper details biochemical analyses of HIV-1 PICs.
The original PIC assay utilized a genetic approach to quantify the level of in vitro integration . Extracts were prepared from cells infected with Moloney murine leukemia virus (MLV) modified to contain the bacterial SupF gene, and purified bacteriophage lambda gtWES DNA was added as a surrogate to cell chromosomes. Following in vitro integration, purified DNA was reacted with phage lambda packaging extract, and the resulting phage were plated on two different indicator cells. Bacteria that naturally complemented the nonsense codons in lambda gtWES yielded overall phage titer, whereas phage that had acquired SupF via MLV integration gained the ability to plate on suppressor-minus cells. The fraction of suppressor-minus colonies thus defined the frequency of MLV DNA integration. Importantly, the resulting proviruses lost 2 bp from each end of viral DNA and were flanked by a 4 bp duplication of lambda DNA sequence at the site of integration. The advent of this in vitro assay that recapitulated the known genetics of MLV integration in one fell swoop redefined the retroviral integration field.
Though quantitative, the Brown et al.  assay was laborious and moreover scored integration indirectly through phage lambda plating efficiencies. Deciphering viral cDNA end structures either before or after integration were therefore not approachable using the genetic screen. Direct physical assays were needed, and Fujiwara and Mizuuchi spearheaded the use of indirect end-labeling to analyze viral DNA end structures as well as to monitor the formation of the in vitro DNA recombination intermediate . Native PICs were reacted with linearized bacteriophage X174 target DNA, and purified DNAs were separated by agarose gel electrophoresis. Blotting with a virus-specific probe yielded two linear DNA species: the faster migrating cDNA substrate represented unreacted PICs, whereas the more slowly migrating band identified the MLV DNA recombination intermediate (Fig. 3). PIC activity could therefore be defined as the fraction of viral cDNA converted into integration reaction product. Monitoring integrase 3′ processing and DNA strand transfer activities by indirect end-labeling was quickly adapted to other viral systems such as HIV-1 [4, 11, 12] and Rous sarcoma virus .
Indirect end-labeling remains the industry-standard for quantifying integrase activities in the context of viral infection. Considering the utilization of defined target DNA molecules in vitro, DNA strand transfer reaction products can also be discerned via PCR amplification, and a number of assay designs have been described [13–18]. The main advantage of PCR over Southern blotting is that less input PIC material is in general required for a positive readout, an important consideration for high throughput approaches. One drawback of PCR is that the detection of the DNA recombination product is indirect. In other words, as compared to Southern blotting, the complete DNA recombination intermediate is not visualized. Most PCR assays also amplify just one of the two viral-target DNA junctions. This is expected in the vast majority of cases to represent the concerted integration of both viral DNA ends into opposing target DNA strands with defined spacing (Fig. 1, DNA recombination intermediate), though it should be noted that under certain conditions, for example when one of the viral DNA ends is mutated, the mutant end can insert independent of integrase function .
PICs are extracted from acutely-infected cells during the peak of reverse transcription, which can vary among different retroviruses (cells infected with HIV-1 are generally lysed 5–8 h post-infection, whereas MLV infections proceed overnight).
Relatively high multiplicities of infection are essential for PIC analyses. Additional parameters of cell culture, some of which are undefined, can greatly affect the level of in vitro DNA strand transfer activity. In our experience it is essential to keep cell lines in optimal working order. Maintain the concentrations of suspension cell cultures at 0.1–1.0 × 106/ml unless stated otherwise. Do not let adherent 293T cells reach confluency, and limit the amount of time these cells are exposed to trypsin during subculture. Grow all flasks broad side-down in a 5% CO2 humidified incubator to maximize gas exchange. Test different lots of fetal bovine serum (FBS) for those that yield optimal virus production from transiently-transfected 293T cells, and secure as many bottles of the optimal lot as feasible.
SupT1  and C8166-45  T-cells support efficient HIV-1 reverse transcription and are therefore preferred target cells . Infections are initiated by incubating uninfected cells with chronically-infected virus-producer cells or cell-free virus derived via transient transfection . Identify individual lots of cell lines that yield efficient levels of reverse transcription and in vitro PIC integration activity, and carefully expand these to make additional freezer vials. We routinely thaw fresh cells and expand these to sufficient numbers for each experiment.
MOLTIIIB cells, which are chronically infected with the IIIB strain of HIV-1 , as well as uninfected SupT1 and C8166-45 cells, are grown in RPMI 1640 media supplemented to contain 10% v/v heat-inactivated FBS, 100 IU of penicillin G sodium, and 0.1 mg of streptomycin sulfate per ml. MOLTIIIB cells are treated in one of two ways the day before infection. First, incubate the cells to be used in the co-culture at 4 × 105/ml with 10 ng/ml phorbol 12-myristate 13-acetate (PMA; Sigma-Aldrich catalog number P1585) to stimulate virus production [12, 22]. Plate a second lot of cells at 106/ml (without PMA) to prepare virus-conditioned media. Twenty-four hours later, co-culture 4 × 107 SupT1 cells with 4 × 106 PMA-stimulated MOLTIIIB cells in a final volume of 20 ml of virus-conditioned media for 5 h. The majority of cells should fuse to form giant multi-nucleated syncytia during this time.
High-titer virus is made by transfecting 293T cells with pNL4-3  or pNL43/XmaI, a somewhat smaller derivative lacking ~1.1 kb of flanking human DNA . Grow 293T cells in Dulbecco’s modified Eagles medium (DMEM) containing 10% FBS, 100 IU of penicillin G sodium, and 0.1 mg of streptomycin sulfate per ml (DMEM-FBS). A number of transfection techniques yield virus of sufficient titer. Calcium-phosphate co-precipitation is desirable due to its relatively low cost. It is important to treat resulting virus-containing supernatants with DNase to degrade the bulk of the plasmid DNA that remains after transfection. This step is more efficient with lipid-based as compared to calcium-phosphate transfection reagents. We therefore generate virus by transfection with Fugene 6 as recommended by the manufacturer (Roche Applied Science catalog number 11814443001) for experiments that examine reverse transcription and/or integration by PCR. For indirect end-labeling studies, virus generated by the following calcium-phosphate co-precipitation method suffices.
Cells lysed under non-denaturing conditions are separated into cytoplasmic and nuclear fractions.
HIV-1 PICs can be partially purified by gel filtration chromatography using fast performance liquid handling devices [9, 11, 12] or centrifugation [22, 27]. We prefer Sepharose CL-4B (Sigma-Aldrich catalog number CL4B200) spin columns assembled in Bio-Rad Econo-Pac columns (Fig. 4).
Purified viral DNA cleaved with restriction endonucleases is denatured, fractioned through sequencing gels, and transferred to a solid support to visualize the end structures via strand-specific riboprobes [1–5, 9, 26, 28].
Assay 0.5 ml of crude or Sepharose CL-4B-purified PIC preparation for determination of DNA strand transfer activity.
Our assay quantifies the extent of U5 att site integration into circular pTZ18U/PL plasmid DNA (Fig. 5). An initial exponential PCR amplifies HIV-plasmid DNA junctions, whereas the second real-time or kinetic PCR quantifies the resulting levels of HIV-1 DNA sequences [16, 18]. The revamped method here incorporates the AE3014 first round chimera primer comprised of bacteriophage λ and HIV-1 R sequences (Fig. 5). This design, adopted from Brussel and Sonigo  who used it to measure integration into human cellular DNA, increases overall signal-to-noise by incorporating λ DNA-specific primer AE3013 in the nested quantitative (Q) PCR. See Table 1 for primer and TaqMan probe sequences.
Many assays exist to measure levels of HIV-1 integrase catalytic function, and most of these utilize purified, recombinant protein. It however remains essential to quantify the levels of integrase activities in the context of virus infection. For example, indirect end-labeling analyses helped to pinpoint the functionality of an important pre-clinical class of HIV-1 integrase inhibitor to the DNA strand transfer step of integration . Raltegravir has more recently been approved for clinical use , and PIC analyses are predicted to remain central to investigations into mechanisms of drug resistance. Indirect end-labeling and nested PCR analyses of complexes isolated from lens epithelium-derived growth factor (LEDGF) knockout cells were moreover crucial to ascertain that this critical lentiviral integrase binding protein functioned downstream of PIC formation to affect the frequency and distribution of HIV-1 integration . PIC analyses will therefore remain a critical research tool as studies of host cell factors that potentially affect integrase function and HIV-1 integration expand [35–38].
We thank Michael Miller for technical advice and generous gifts of reagents. This work was supported by NIH grants AI039394 and AI052014 to A.E.
1Sterilize biochemical reagents by filtration through 0.22 μm filters.
2Make buffers K+/+ and K+/− fresh for each experiment. Digitonin stock solution (5% w/v in dimethyl sulfoxide) is stored at 4°C.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.