Adeno-associated virus does not replicate on its own. In nature, AAV is classified as a dependoparvovirus — a virus that requires co-infection by a helper virus, typically adenovirus or herpesvirus, to complete its replication cycle. This biological dependency is, paradoxically, one of the features that made AAV attractive as a gene therapy vector: a virus that cannot replicate without external help poses a fundamentally lower safety risk than one that can.
In clinical manufacturing, that biological dependency is reproduced in a controlled way through the three-plasmid system — a production architecture that supplies all the genetic components AAV needs to be manufactured, while keeping them deliberately separated to prevent production of replication-competent virus. Understanding how the three-plasmid system works, what each component provides, and why all three are required is essential to understanding AAV manufacturing at any level of depth.
Why Three Plasmids?
The wild-type AAV genome contains three genes: Rep, which encodes the replication proteins; Cap, which encodes the capsid structural proteins; and aap, which encodes the assembly activating protein embedded within the Cap reading frame. In the wild-type virus, these genes are flanked by inverted terminal repeats (ITRs) — short palindromic sequences that act as the origin of replication and the packaging signal.
To manufacture a recombinant AAV vector for therapeutic use, the therapeutic transgene replaces the Rep and Cap coding sequences in the genome, with only the ITRs retained to signal the replication and packaging machinery. But if the Rep and Cap genes are no longer in the vector genome, they must be supplied from somewhere else — otherwise the packaging machinery does not exist and no viral particles are produced.
The three-plasmid system solves this by dividing the required genetic information across three distinct DNA constructs, each introduced into producer cells simultaneously during transient transfection. Each plasmid is necessary. None is sufficient on its own. Together, they direct the producer cells to manufacture complete, functional, therapeutic viral vectors.
The Transfer Plasmid: The Therapeutic Payload
The transfer plasmid — also called the cis plasmid or the vector genome plasmid — carries the therapeutic gene that will ultimately be delivered to patient cells. It is the only component of the three-plasmid system that becomes part of the final drug product: its sequence, flanked by the two ITRs, is what gets packaged into each viral capsid.
The design of the transfer plasmid directly determines several product attributes. The promoter upstream of the transgene controls where and how much the therapeutic gene is expressed after delivery. The therapeutic coding sequence itself determines the protein or RNA that is produced. The polyadenylation signal at the 3′ end determines transcript stability. Regulatory elements such as introns, enhancers, and woodchuck hepatitis virus post-transcriptional regulatory elements (WPREs) can be incorporated to increase expression levels.
The ITRs on either side of the transgene cassette are the only cis-acting sequences required for genome replication and packaging. They are the molecular flag that the Rep proteins recognize when selecting which DNA to package into the assembling capsid. Anything flanked by ITRs can in principle be packaged; anything outside the ITRs is not packaged. This specificity is why the Rep and Cap sequences — which must be present in the producer cell but must not appear in the therapeutic vector — are placed on separate plasmids.
One critical constraint of the transfer plasmid is payload capacity. The AAV capsid can accommodate a genome of approximately 4.7 kilobases. Sequences larger than this cannot be efficiently packaged, leading to truncated or fragmented genomes and reduced vector potency. This size limitation is one of the defining strategic constraints in AAV gene therapy program design.
The Rep-Cap Plasmid: The Packaging Machinery
The Rep-Cap plasmid supplies the two sets of proteins that carry out genome replication and capsid assembly. Rep encodes four proteins — Rep78, Rep68, Rep52, and Rep40 — that collectively handle genome amplification and packaging. Cap encodes three structural proteins — VP1, VP2, and VP3 — that assemble into the icosahedral capsid shell, along with the assembly activating protein (AAP) encoded in an overlapping reading frame within Cap.
Rep proteins recognize the ITRs on the transfer plasmid, initiate replication of the vector genome, and use helicase activity to thread the single-stranded DNA genome into assembling capsid shells. Cap proteins self-assemble into the 60-subunit icosahedral structure in a 1:1:10 molar ratio of VP1:VP2:VP3, forming the protein shell that protects the genome and mediates cell entry and tissue targeting.
Because Rep proteins are cytotoxic at high concentrations — they interfere with host cell replication machinery — the Rep-Cap plasmid must be designed so that Rep expression is sufficient for productive packaging without killing the producer cells during the 72 to 96-hour post-transfection production window. This is one of the more delicate optimization challenges in AAV process development, and it is why promoter selection, codon optimization, and plasmid ratio during transfection are all active variables in manufacturing process design.
Critically, the Cap sequence on the Rep-Cap plasmid determines the serotype of the vector being produced — and therefore its tissue tropism. Swapping in a different Cap sequence produces a vector with a different tissue targeting profile. The Rep machinery is largely serotype-independent: the same Rep proteins can package genomes into capsids made of VP proteins from different serotypes, though with some efficiency differences across serotype combinations.
The Helper Plasmid: The Adenoviral Functions
In nature, AAV replication requires co-infection with a helper virus — historically adenovirus — that supplies transcriptional activators and other functions that AAV cannot provide for itself. In recombinant AAV manufacturing, these helper functions are provided in trans by the third plasmid: the helper plasmid.
The helper plasmid carries a subset of adenoviral genes: E2A, E4, and VA RNA. Each serves a distinct function. E2A encodes a DNA-binding protein involved in viral DNA replication. E4 encodes a set of regulatory proteins that promote viral mRNA accumulation and processing. VA RNAs are non-coding RNA species that suppress the innate antiviral response of the producer cell, allowing efficient transgene expression.
Notably, the helper plasmid does not carry the E1A or E1B adenoviral genes — the genes that would be required for adenoviral replication. HEK293 cells, the most common AAV producer cell line, constitutively express E1A and E1B because those genes were incorporated into the HEK293 genome during the original cell line creation. By using HEK293 cells, manufacturers obtain the E1A/E1B functions from the host genome rather than from an additional plasmid, while ensuring that no complete adenovirus can be assembled in the production process.
The functional consequence is that the helper plasmid provides just enough adenoviral helper activity to support AAV replication and packaging, without enabling production of wild-type adenovirus or replication-competent AAV. Residual adenoviral proteins present in the harvested vector must be removed during downstream purification, and their absence is verified as a standard lot release test.
Why All Three Are Necessary
Each plasmid supplies a non-redundant function. The transfer plasmid provides the genome to be packaged. The Rep-Cap plasmid provides the machinery to replicate and package that genome. The helper plasmid provides the transcriptional environment in which that machinery can operate efficiently.
Remove the transfer plasmid, and there is nothing to package — the Rep and Cap proteins are produced and assemble into empty capsids, but no therapeutic genome is incorporated. Remove the Rep-Cap plasmid, and there is no replication or packaging machinery — the therapeutic genome exists in the cell but cannot be amplified or loaded into capsids. Remove the helper plasmid, and Rep and Cap expression is insufficient — without adenoviral helper functions, AAV replication proceeds inefficiently or not at all.
The three-plasmid system is not merely a manufacturing convention. It is a deliberate biosafety architecture that distributes the genetic requirements for viral production across components that cannot individually — or even in pairs — produce a complete, infectious, replication-competent virus. That design principle is as important to the regulatory acceptability of AAV manufacturing as it is to its scientific function.
Understanding the role of each plasmid is the prerequisite for understanding why changes to any one of them — sequence modifications, promoter changes, plasmid ratios — must be evaluated against the full quality profile of the final vector product.