Retroviruses and Reverse Transcription

A retrovirus is an enveloped virus whose genome consists of two identical copies of positive-sense single-stranded RNA, roughly 7–12 kilobases each, packaged together as a diploid genome inside the virion. This places retroviruses in Group VI of the The Baltimore Classification System, the group defined precisely by its dependence on reverse transcription. Unlike other positive-strand RNA viruses (Group IV), whose genomic RNA is translated directly by host ribosomes, retroviral RNA is never translated as genomic RNA in the cytoplasm of a newly infected cell. Instead, it is first copied into DNA, and that DNA — not the RNA — becomes the template for everything downstream. The enzyme that performs this copying, reverse transcriptase (RT), is a multifunctional polymerase carried inside the virion itself, so the virus does not have to wait for host machinery to catalyze this unusual reaction. RT's product is a double-stranded DNA copy of the genome called the provirus, flanked by two identical long terminal repeat (LTR) sequences that did not exist in that form in the original RNA — they are synthesized during reverse transcription itself, in a process that duplicates terminal sequences and generates the promoter/enhancer elements needed for host RNA polymerase II to later transcribe the provirus. The provirus is then inserted into the host cell's chromosomal DNA by a second viral enzyme, integrase, at essentially random (though not perfectly random) genomic locations. Once integrated, the provirus behaves exactly like a cellular gene: it is replicated along with host DNA at every cell division, transcribed by host polymerase II, and its transcripts are spliced, exported, and translated using entirely host machinery. This is the defining feature that distinguishes retroviruses from every other RNA virus family — the genome becomes a permanent, heritable part of the host genome, not a transient cytoplasmic replication intermediate.

Step 1 — Entry and uncoating. The retroviral envelope glycoprotein binds a host cell receptor (for HIV, CD4 plus a coreceptor such as CCR5), triggering membrane fusion and release of the viral core into the cytoplasm. The core contains two RNA genome copies, RT, integrase, and a host-derived tRNA primer already annealed to a specific site on the genome called the primer binding site (PBS), typically located near the 5′ end of the RNA.

Step 2 — Minus-strand DNA synthesis begins. RT's RNA-dependent DNA polymerase (RDDP) activity extends the tRNA primer, synthesizing a short DNA segment complementary to the 5′ end of the genomic RNA (the R and U5 regions). This produces "minus-strand strong-stop DNA," typically 100–150 nucleotides long.

Step 3 — First strand transfer. Because the RNA template used to make this short DNA is degraded by RT's RNase H activity, and because the R sequence is repeated at both ends of the genome, the nascent DNA can dissociate and re-anneal to the R sequence at the 3′ end of the same or the other genomic RNA copy. This "jump" is called strand transfer and is essential — it relocates the growing DNA strand to a template position from which it can copy the rest of the genome.

Step 4 — Full minus-strand synthesis. RT elongates the DNA continuously along the genomic RNA, while RNase H concurrently degrades the RNA template behind the advancing polymerase, except for two purine-rich RNA fragments resistant to RNase H cleavage — the polypurine tracts (PPTs).

Step 5 — Plus-strand synthesis and second strand transfer. The PPT RNA fragments serve as primers for plus-strand DNA synthesis, using the newly made minus-strand DNA as template. This copies through the tRNA primer sequence, generating a plus-strand strong-stop DNA that must again jump (second strand transfer) to the other end of the minus strand, guided by complementary PBS sequences, to complete a full-length double-stranded DNA.

Step 6 — LTR generation. Because the process duplicates the U3 and U5 regions at both ends of the DNA (U3 originally only at the 3′ end of the RNA, U5 only at the 5′ end), the final double-stranded DNA product carries two identical LTRs, each with the structure U3-R-U5, at both ends. This is the molecular reason the LTRs exist at all: they are generated de novo by the mechanics of strand-transfer, not copied directly from a single continuous template.

Step 7 — Nuclear entry. The blunt-ended, double-stranded proviral DNA, still associated with integrase and other viral proteins in a pre-integration complex (PIC), must cross the nuclear envelope. For lentiviruses like HIV, the PIC can cross an intact nuclear pore, permitting infection of non-dividing cells — a key difference from simple retroviruses like murine leukemia virus, which generally require nuclear envelope breakdown during mitosis.

Step 8 — Integration. Integrase first removes two nucleotides from each 3′ end of the viral DNA (3′ processing), exposing recessed CA-3′-OH ends. Integrase then catalyzes strand transfer, using these 3′-OH groups to attack phosphodiester bonds on opposite strands of host chromosomal DNA, staggered by a small number of base pairs (5 bp for HIV). This concerted cleavage-ligation reaction covalently joins viral DNA to host DNA at both ends simultaneously. The staggered cut leaves short single-stranded gaps that host DNA repair enzymes fill in, which duplicates a short stretch of host sequence flanking the provirus — the hallmark target-site duplication seen at every retroviral integration site.

Step 9 — Provirus expression. The integrated provirus is transcribed by host RNA polymerase II from the promoter in the 5′ LTR. Full-length transcripts serve three roles: genomic RNA for progeny virions, unspliced mRNA for Gag and Gag-Pol polyproteins, and singly/multiply spliced mRNA for Env and accessory proteins. Gag polyprotein drives assembly at the plasma membrane; the viral protease (itself encoded within Pol and initially part of the Gag-Pol polyprotein) cleaves Gag and Gag-Pol into mature structural proteins and enzymes during or after budding, a maturation step required for infectivity.

Genome organization. Almost all retroviruses share the basic gene order 5′-gag-pol-env-3′, flanked by LTRs:

Complex retroviruses such as HIV additionally encode regulatory and accessory genes (tat, rev, nef, vif, vpr, vpu in HIV-1) overlapping the structural genes in different reading frames, which control latency, immune evasion, and replication efficiency — foreshadowing the greater regulatory complexity covered when discussing HIV and Its Life Cycle. RT itself deserves emphasis as an enzyme with three distinct catalytic activities in one polypeptide (or heterodimer, in HIV's p66/p51 RT): RNA-dependent DNA polymerase, DNA-dependent DNA polymerase, and RNase H, an unusual multifunctionality that makes it both mechanistically elegant and, because it lacks 3′→5′ proofreading exonuclease activity, highly error-prone — introducing roughly 1 mutation per 10⁴–10⁵ nucleotides copied per replication cycle, a major driver of retroviral genetic diversity and drug resistance.

The problem with the dogma. By the early 1960s, the central dogma of molecular biology — articulated by Francis Crick — held that sequence information flows DNA→RNA→protein, and explicitly that RNA cannot be copied back into DNA. Yet Howard Temin, studying Rous sarcoma virus (RSV, an RNA tumor virus) in chicken cells, observed something the dogma could not explain: RSV infection and the resulting cell transformation were blocked by actinomycin D, a drug that inhibits DNA-dependent RNA synthesis, and by inhibitors of DNA synthesis applied early in infection. This implied that a DNA intermediate was required for an RNA virus to replicate and transform cells — Temin's "DNA provirus hypothesis," first proposed in 1964, which was met with deep skepticism.

The decisive experiment (1970). Temin, working with Satoshi Mizutani, and independently David Baltimore, working with Moloney murine leukemia virus, both demonstrated directly that purified retrovirus particles contain an enzyme that synthesizes DNA from an RNA template. The key experimental design: 1. Purify intact virions of RSV (Temin/Mizutani) or Moloney MLV (Baltimore) away from any host cell contamination. 2. Permeabilize the virion envelope with a mild detergent (e.g., Nonidet P-40) to allow substrate access without destroying internal enzyme activity. 3. Supply the permeabilized virions with radiolabeled deoxynucleotide triphosphates (dNTPs) and Mg²⁺. 4. Measure incorporation of radiolabel into an acid-insoluble (polymeric, macromolecular) product. 5. Show the product is DNA, by its sensitivity to DNase but resistance to RNase, and confirm it is complementary to the viral RNA genome by nucleic acid hybridization. Both groups found robust, template-dependent DNA synthesis occurring inside detergent-disrupted virions, and the product hybridized specifically to the viral genomic RNA — direct proof that an RNA-templated DNA polymerase existed inside the virus particle itself, not contributed by a contaminating host enzyme. Both papers were published back-to-back in Nature in June 1970. This was airtight because the enzyme was demonstrated to be an intrinsic component of a purified, cell-free virion, ruling out artifacts from host-cell DNA polymerases, and the RNA-dependence of the DNA product was proven biochemically, not just inferred from drug-sensitivity experiments.

Consequence. The discovery earned Temin and Baltimore (with Renato Dulbecco, for related tumor virus work) the 1975 Nobel Prize in Physiology or Medicine. It resolved the paradox of RNA tumor viruses causing permanent, heritable cellular transformation — because their genome becomes DNA and integrates — and it forced a formal amendment of the central dogma to allow RNA→DNA information flow as a real, if specialized, biological pathway. This enzyme was later purified, characterized structurally, and became the essential reagent for cDNA cloning and modern molecular biology.

A second pillar of evidence: integration itself. Subsequent work in the 1970s–80s (including studies by Robin Weiss, Stephen Goff, and others) demonstrated that proviral DNA is covalently linked to host chromosomal DNA by restriction enzyme mapping and Southern blotting: cutting host genomic DNA at sites flanking a known integrated provirus produced DNA fragments of a size consistent with covalent, colinear attachment to unique flanking host sequences that differed at every independent infection event — proving random, chromosome-embedded integration rather than episomal persistence. The discovery of target-site duplications flanking every provirus, always a fixed small number of base pairs for a given retrovirus, provided the structural signature predicted by the staggered-cut integration mechanism later confirmed biochemically with purified integrase protein and model DNA substrates in vitro (Craigie, Bushman, and colleagues in the late 1980s and 1990s), reconstituting 3′ processing and strand transfer with purified components.

Example 1 — Calculating expected mutation load in a single replication cycle. HIV-1 has a genome of about 9,700 nucleotides (~9.7 kb). Reverse transcriptase's in vivo error rate is commonly cited around 3 × 10⁻⁵ mutations per nucleotide copied per replication cycle (published estimates for HIV-1 RT range roughly 10⁻⁴–10⁻⁵ depending on assay). Using 3 × 10⁻⁵ per base: Expected mutations per genome per cycle = 9,700 × (3 × 10⁻⁵) ≈ 0.29 mutations per genome per replication cycle. Because roughly 10¹⁰ new virions are produced per day in an untreated HIV-infected patient, and each genome differs on average by close to one mutation from its immediate parent, essentially every possible single point mutation across the genome is generated multiple times per day. This is why HIV populations exist as a quasispecies — a cloud of related but genetically distinct variants — and why resistance mutations to a single antiretroviral drug can pre-exist in an infected person even before treatment begins, explaining the clinical rule that HIV must always be treated with combination therapy (typically three drugs from at least two classes), since resistance to any single drug is nearly guaranteed to arise by chance alone, but simultaneous resistance mutations to three independent drug targets are vastly less probable in the same genome.

Example 2 — Reading an integration-site Southern blot to distinguish clonal expansion from ongoing infection. Suppose a researcher studies a patient's peripheral blood T cells using ligation-mediated PCR to map retroviral integration sites, followed by deep sequencing, in a gene therapy trial using a retroviral vector to correct a genetic immunodeficiency (a real methodology used in SCID-X1 gene therapy trials).

• If integration sites are highly diverse across thousands of independent cells (each provirus at a different chromosomal position with different flanking sequence), this indicates a broad, polyclonal population of independently transduced cells — normal, expected outcome.
• If, instead, a large fraction of cells share the identical integration site (same chromosome, same flanking target-site duplication sequence), this indicates that a single transduced cell has since clonally expanded, since integration is a one-time, essentially random event and cannot recur at the exact same base pair independently in unrelated cells.
• In the real-world SCID-X1 trials of the early 2000s, exactly this pattern was found in several patients who developed leukemia: dominant clones carried integrations near the LMO2 proto-oncogene, and insertional activation of LMO2 by the strong retroviral LTR promoter/enhancer was implicated in driving oncogenic clonal expansion.

This worked example shows how the same molecular fact — random, semi-permanent integration with host-sequence duplication — is simultaneously the basis of retroviral vector gene therapy and the mechanism of its most serious complication, insertional oncogenesis.

Howard Martin Temin (1934–1994) trained under Renato Dulbecco at Caltech, studying Rous sarcoma virus in cell culture using the quantitative focus-assay methods Dulbecco had pioneered for animal viruses. Working at the University of Wisconsin–Madison from 1960 onward, Temin noticed that RSV-transformed cells retained their transformed, cancerous phenotype even after the original infecting virus was long gone and diluted away through many cell divisions — behavior far more consistent with a stably inherited genetic change than with a transient cytoplasmic RNA replication cycle. His insistence on the DNA provirus hypothesis, in the face of considerable professional skepticism through the mid-to-late 1960s, exemplifies how a rigorously reasoned but unfashionable hypothesis can be vindicated by decisive biochemistry once the right experiment is designed. David Baltimore, working at MIT on unrelated RNA tumor viruses using similar detergent-disruption biochemistry, reached the same conclusion independently and simultaneously, and the two published together in the same 1970 issue of Nature — a striking case of convergent discovery validating a genuinely correct but previously heterodox idea.

Antiretroviral therapy. Reverse transcriptase and integrase are two of the principal drug targets in modern HIV treatment. Nucleoside/nucleotide RT inhibitors (NRTIs, e.g., tenofovir, emtricitabine) are chain-terminating dNTP analogs; non-nucleoside RT inhibitors (NNRTIs) bind an allosteric pocket in RT; integrase strand-transfer inhibitors (INSTIs, e.g., dolutegravir, bictegravir) block the strand-transfer chemistry directly, and are now first-line therapy in most treatment guidelines because of their high barrier to resistance — a topic elaborated further under Antiviral Strategies and Drug Targets.

Gene therapy vectors. Retroviral and lentiviral vectors, engineered to delete most viral genes but retain the cis-acting LTR and packaging sequences, are used to stably integrate therapeutic genes into patient cells ex vivo, as in CAR-T cell cancer immunotherapy and treatments for inherited immunodeficiencies and blood disorders; modern self-inactivating (SIN) vector designs specifically reduce the LTR enhancer activity implicated in the earlier insertional oncogenesis cases.

Molecular biology and diagnostics. Reverse transcriptase, purified from Moloney MLV or avian myeloblastosis virus, is an indispensable laboratory reagent, converting RNA into complementary DNA (cDNA) for RT-PCR diagnostics (including SARS-CoV-2 testing), RNA sequencing library preparation, and cloning of eukaryotic genes from mRNA.

Genome evolution and human biology. Endogenous retroviruses, fixed in the human germline over evolutionary time, constitute a measurable fraction of the genome and illuminate both ancient infection history and cases of molecular domestication, such as syncytin genes in placental development — a striking example of viral genes being repurposed for essential host physiology, connecting this topic to broader questions in RNA Virus Replication Strategies and genome evolution.

Cancer biology. Insertional mutagenesis by integrating retroviruses, first understood through studies of avian leukosis virus activating the MYC oncogene and later directly observed in human gene therapy trials, remains a central paradigm for how integration site choice and LTR enhancer strength can convert a therapeutic vector into an oncogenic event, informing the design of safer, self-inactivating vectors used today.

Ask the tutor: How does HIV's use of accessory genes and nuclear pore transport let it infect non-dividing cells, unlike simple retroviruses?