A bacterial artificial chromosome (BAC) is an engineered DNA molecule that replicates as a single-copy plasmid inside E. coli, carrying inserts of roughly 100–300 kilobases. It is built from the F-plasmid's strict copy-control machinery, which keeps the construct stable and limits rearrangement even with very large inserts. That combination of capacity and fidelity is why BACs underpin genome libraries, physical maps, and transgenic models. This article explains the replication mechanism, the real size ceiling, how BACs compare with YACs and cosmids, common failure modes, and the concrete steps labs take to keep clones intact.
The F-Plasmid Replication Mechanism That Makes BACs Work
Every BAC descends from the fertility factor (F-plasmid) of E. coli. The F-plasmid is a low-copy, self-transmissible episome that normally carries genes for conjugation. BAC vectors strip out the transfer genes and keep only what is needed for stable maintenance. What remains is a compact replicon with two functional zones: oriS, the origin of vegetative replication, and the par region, which actively partitions copies to daughter cells.
Replication is controlled by repE, which encodes the initiator protein, and by incC, an incompatibility locus that counts copies. The system holds the plasmid at one to two copies per cell. In plain terms: the BAC behaves like a single, carefully counted chromosome rather than a runaway plasmid. That low copy number is the central design feature, because it minimizes the opportunity for homologous recombination between repeated sequences inside a large insert.
The parA, parB, and parC genes form the partitioning cassette. They push plasmid copies to opposite cell poles before division. A BAC that loses its par region drops clones at measurable frequency, which is why most modern vectors retain it. The origin of transfer (oriT) is usually kept as well, not because conjugation is wanted, but because it allows mobilization when a helper plasmid is supplied. Selectable markers (typically chloramphenicol resistance via cat) and a multicloning site complete a standard 7–8 kb backbone.
The stability gain is quantitative, not marginal. A high-copy pUC-based plasmid carrying a 150 kb insert typically deletes or rearranges within a few passages. A single-copy BAC of the same insert shows rearrangement rates low enough that clone archives remain usable after years of frozen storage. That is the entire engineering rationale in one sentence.
Why BACs Beat Cosmids, YACs, and Plasmid Vectors
The BAC occupies a specific niche defined by insert size, copy number, and host. Cosmids top out around 45 kb. Yeast artificial chromosomes (YACs) reach into the megabase range but show chimerism and rearrangement at rates that made large-scale projects painful. Ordinary plasmids are convenient but cannot hold a eukaryotic gene with its full regulatory landscape. BACs land in the 100–300 kb band with bacterial handling speed.
| Vector type | Typical insert size | Copy number in host | Main strength | Main weakness |
|---|---|---|---|---|
| Plasmid | 0.1–10 kb | 500–700 | Fast, simple cloning | Tiny capacity |
| Cosmid | 35–45 kb | 10–20 | Efficient packaging | Repeated-sequence instability |
| BAC | 100–300 kb | 1–2 | Stability with large inserts | Slow DNA prep yield |
| YAC | 0.2–2 Mb | 1 | Largest inserts | Chimerism and rearrangement |
| PAC (P1-based) | 80–100 kb | 1–2 | Alternative single-copy host | Smaller ceiling than BAC |
The PAC (P1 artificial chromosome) deserves a mention because it is often confused with the BAC. PACs use the P1 bacteriophage replicon and offer a similar single-copy profile at somewhat smaller inserts. Many genome projects ran BAC and PAC libraries in parallel and chose per-region based on what cloned cleanly.
The trade-off is yield. One to two copies per cell means a liter of culture produces far less BAC DNA than the same volume of a high-copy plasmid prep. Labs routinely grow 500 mL to 1 L of culture and use large-column purification. That is not a defect; it is the price of the stability that makes a 200 kb clone worth having.
Insert Size Limits and the Real Numbers Behind Them
The practical BAC ceiling is usually quoted as 300 kb, but the true figure depends on which number you measure. The largest inserts reliably recovered from standard libraries cluster at 150–250 kb. Above roughly 300 kb, transformation efficiency into E. coli falls sharply, and clones that do transform often carry deletions. A few specialized BACs have reached 350–500 kb, but these are exceptions that require modified hosts and careful verification.
Library construction reflects this. The human and mouse genome BAC libraries built for the Human Genome Project used average inserts near 150–200 kb. A library targeting sixfold coverage of a 3 billion base pair genome therefore needs roughly 90,000–120,000 clones at 200 kb each. That arithmetic sets the plate count and freezer footprint for any large sequencing or mapping effort.
Electroporation efficiency is the bottleneck. Transformations of ligated BAC DNA into standard E. coli strains yield on the order of 10³ to 10⁴ colonies per microgram, compared with 10⁶ or higher for small plasmids. Adding a 100 kb insert cuts that further. This is why BAC cloning is a construction project rather than a routine bench task. The 2017 FDA guidance on genome editing and gene therapy products repeatedly cites the need to verify large-insert vectors for rearrangements, reflecting how common partial deletions are in practice.
Building a BAC: Timeline From Design to Verified Clone
BAC construction and verification follow a predictable sequence with several quality gates. The timeline below reflects a typical 6–10 week cycle for a straightforward library or a single large-insert clone, and shorter if a pre-made library is available.
| Stage | Typical duration | Key checkpoint |
|---|---|---|
| Vector preparation and insert digestion | 3–5 days | Verified backbone and insert ends |
| Ligation and electroporation | 2–3 days | Colony count and insert screening |
| Clone screening by PCR or colony hybridization | 3–7 days | On-target insert confirmation |
| Restriction fingerprinting | 2–4 days | No internal deletions |
| End sequencing and assembly check | 5–10 days | Correct junctions, no chimerism |
| Large-scale prep and final QC | 3–5 days | Yield and purity for downstream use |
| Optional transgenesis in a model organism | 2–12 weeks | Germline transmission and copy number |
The two checkpoints that matter most are fingerprinting and end sequencing. Fingerprinting with rare-cutting enzymes detects internal deletions that PCR across the insert would miss. End sequencing catches chimeras formed during ligation. Skipping either is a common reason a clone fails months later during functional work.
Where Bacterial Artificial Chromosomes Are Used
BACs carry full-length eukaryotic genes with their native promoters, enhancers, and introns. That is the property that drives most current use. A BAC spanning 150–200 kb can recapitulate endogenous expression patterns that a 5 kb cDNA construct cannot. This is why transgenic mouse lines carrying human BACs exist for hundreds of disease loci, and why BAC transgenics are a workhorse of neurobiology.
Genomic libraries remain a core application. The human and mouse BAC libraries built in the 1990s and 2000s still supply clones for functional studies and for filling gaps in genome assemblies. Physical mapping, positional cloning, and fluorescence in situ hybridization (FISH) probes often trace back to BAC clones. Diagnostic FISH probes for cancer cytogenetics, including those for BCR-ABL1 and MYC rearrangements, are frequently BAC-derived.
In synthetic biology and gene therapy, BACs serve as stable carriers for large gene clusters. They are used to shuttle polyketide synthase clusters, immunoglobulin loci, and similar multi-gene cassettes into model systems. BACs are also candidate vector backbones for agroinfiltration in plant biotechnology because of their ability to hold large T-DNA inserts. These uses are established in research settings, while clinical use of BAC-based vectors remains limited and is regulated case by case, per FDA guidance on gene therapy products.
Misconceptions and Edge Cases
A frequent mistaken belief is that BACs are always stable because they are low-copy. Low copy number reduces recombination probability, but it does not eliminate it. Inserts containing long direct repeats, such as immunoglobulin or olfactory receptor clusters, still recombine in E. coli. The correct claim is that BACs are more stable than high-copy vectors, not that they are inert.
Another edge case is the "empty vector" problem. BACs are large enough that a backbone without an insert can outgrow the insert-bearing clone during library amplification. The result is a library that looks normal by colony count but has lost a fraction of its inserts. Pooled growth rather than individual clone handling is the main cause. Labs that plan to keep libraries for years should amplify minimally and store multiple independent copies.
A third misconception concerns size and function. A large insert does not guarantee that the gene is intact. Rearrangements during ligation or host propagation can truncate a gene while leaving the total insert length unchanged. That is why insert size checks alone are insufficient and why end sequencing and internal fingerprinting remain standard.
One counterintuitive detail: the strict copy control that makes BACs stable also makes them slower to isolate. Labs accustomed to minipreps from high-copy plasmids are often surprised by the volume of culture needed to recover enough BAC DNA for sequencing or transgenesis. That constraint is intrinsic to the design, not a technique error.
Practical Actions for Anyone Working With BACs
The following steps reflect standard practice in genomically demanding labs and reduce the most common failure modes.
- Verify new clones with at least two independent methods: restriction fingerprinting and end sequencing. Do not rely on PCR across the insert alone.
- Use a recombination-deficient host strain (such as recA mutants) for propagation, and confirm the strain before starting any library.
- Grow 500 mL to 1 L of culture for a single BAC prep; do not scale down to miniprep volumes and expect adequate yield.
- Amplify libraries only when needed, and store the original glycerol stocks without pooling.
- Record the vector backbone, insert size, and antibiotic marker for every clone; archived annotations are frequently lost and expensive to reconstruct.
- Red flags that should stop a project and trigger re-verification: unexpected fragment sizes on fingerprinting, loss of the par region on sequencing, and any clone that shows growth in the absence of the selecting antibiotic.
The final practical point is documentation. BACs are the kind of resource that works quietly for a decade and then fails when the original construct notes are missing. A well-annotated clone bank is worth as much as the clones themselves.