Bacterial Artificial Chromosomes: Structure, Cloning, and Genomic Uses

Bacterial artificial chromosomes, universally called BACs, are circular DNA molecules engineered to carry large chunks of foreign DNA inside the common lab bacterium E. coli. Built on the backbone of a naturally occurring fertility plasmid, a single BAC can hold inserts ranging from roughly 100 to over 300 kilobases, which is far more than a standard plasmid can manage and enough to capture entire gene clusters or sprawling regulatory regions in one piece. That capacity made BACs indispensable for the Human Genome Project and continues to earn them roles in virology, drug discovery, transgenic animal research, and early-stage gene therapy.

What a BAC Is Made Of

The BAC vector descends from the F plasmid (also called the fertility factor), a naturally low-copy DNA element that E. coli maintains at one or two copies per cell. Researchers stripped the F plasmid down to its essential replication and partitioning genes and then added a cloning site where foreign DNA can be inserted, along with selectable markers like antibiotic-resistance genes so that only bacteria carrying the BAC will grow on selective plates. The low copy number is not a bug; it is a deliberate feature. Having just one or two copies per cell means the host’s own recombination machinery has very little opportunity to rearrange the insert, which keeps the cloned DNA faithful to its original sequence.

Some newer BAC designs add a second, switchable origin of replication so researchers can temporarily boost the copy number when they need a lot of DNA. One approach uses a tightly controlled promoter system that responds to the sugar L-arabinose. Under normal growth conditions, the BAC stays at single copy. When arabinose is added to the culture medium, the second origin fires and the copy number climbs, producing enough DNA for downstream experiments. Adding glucose to the medium can push the system back toward a true single-copy state, improving long-term stability to a level comparable to DNA integrated into the host chromosome itself.¹

Building a BAC Library Step by Step

A BAC library is a collection of thousands of BAC clones, each carrying a different overlapping fragment of an organism’s genome. Together, the clones represent the entire genome many times over, so that any given region can be found in at least a few independent clones. Building one is a multi-day process that hinges on keeping the source DNA as intact as possible.

The first challenge is extracting very large DNA without breaking it. Genomic DNA at the sizes needed for BAC cloning is fragile, so cells or purified nuclei are typically embedded in low-melting-point agarose plugs before any enzymes touch them. All the lysis, protein digestion, and washing steps happen while the DNA is immobilized in that gel matrix, which shields the long molecules from the shearing forces of pipetting and vortexing.² The embedded DNA is then partially cut with a restriction enzyme, most commonly MboI or a controlled combination of EcoRI endonuclease and EcoRI methylase, to generate a random spread of overlapping fragments.³

Because the partial digestion produces fragments spanning a wide range of sizes, the next step is to pick out the ones in the sweet spot, usually somewhere between 100 and 300 kilobases. Pulsed-field gel electrophoresis separates these enormous molecules by size in a way that conventional gel electrophoresis cannot. Researchers run marker lanes alongside the digest, stain only those outer lanes to see where the desired size range migrated, then cut the corresponding unstained zone from the middle of the gel and recover the DNA by electroelution.⁴

The size-selected fragments are then ligated into the BAC vector, and the resulting constructs are introduced into E. coli by electroporation, a brief high-voltage pulse that opens transient pores in the bacterial membrane. Getting large DNA molecules inside a bacterium is far harder than getting small ones in. Molecules around 240 kilobases transform roughly 30-fold less efficiently, on a molar basis, than molecules of 80 kilobases, and the optimal voltage and pulse conditions differ from those used for small plasmids.⁵ The choice of E. coli strain matters enormously as well; strains that perform equally well with small DNA can differ by as much as 30-fold in their ability to take up large molecules. Careful optimization of growth temperature, washing buffers, and cell concentration has pushed transformation efficiencies for 120-kilobase BACs up to hundreds of millions of transformants per microgram of DNA.⁶

Why BACs Replaced Yeast Artificial Chromosomes

Before BACs became the workhorse, yeast artificial chromosomes (YACs) were the standard tool for cloning large DNA. YACs can carry even bigger inserts, sometimes exceeding a megabase, but they come with serious drawbacks. Chimeric clones, where two unrelated pieces of genomic DNA get joined into a single YAC during the cloning process, are common. Library construction is tedious, and purifying a clean YAC insert away from the host yeast’s own chromosomes is difficult. BACs solved most of these problems in one stroke by moving the work into E. coli, an organism that grows faster, is simpler to manipulate, and produces DNA that is easy to purify.

Early tests of BAC libraries in plants illustrated the advantages clearly. A sorghum BAC library of over 13,000 clones had a combined average insert size of 157 kilobases, with some inserts reaching 315 kilobases. Those large inserts remained stable after more than 100 generations of growth in liquid culture, and fluorescence hybridization to sorghum chromosomes detected no chimeric clones.⁷ The stability stems from the single-copy replication strategy: with only one BAC molecule in the cell, there is no partner molecule for recombination to act on, which is a major advantage over YACs, where recombination-driven rearrangements are a well-known headache.⁸

Other cloning systems sit below BACs in insert capacity. Cosmids and fosmids typically carry inserts around 40 kilobases, while P1-derived artificial chromosomes (PACs) hold roughly 85 kilobases on average.⁹ BACs fill the gap between these smaller-insert systems and the unwieldy YACs, combining large capacity with the convenience and reliability of E. coli cloning.

BACs and the Human Genome Project

BAC libraries were central to the public Human Genome Project. The international consortium built a physical map of the genome by fingerprinting hundreds of thousands of BAC clones and assembling them into overlapping contigs that tiled across each chromosome. That map then guided the selection of individual BACs for sequencing, and the finished sequences were stitched together using the map as a scaffold.¹⁰

Three BAC libraries, known as RPCI-11, RPCI-13, and Caltech D1/D2, contributed the vast majority of clones to the fingerprint map and the sequencing effort. A validated set of nearly 33,000 BAC clones was later assembled to span the genome, covering at least 98 percent of the fingerprint map and 99 percent of the assembled sequence, with an effective resolving power of about 79 kilobases.¹¹ That clone set became a permanent resource: researchers could order specific BACs from repositories and use them for regional studies, gene hunting, or clinical diagnostics without having to rebuild anything from scratch.

The same strategy was applied to other organisms. Physical maps built from BAC libraries have supported genome projects for soybean, flax, maize, rice, cattle, and many more species.¹² In crop plants especially, BAC-based maps remain valuable for positional cloning of genes tied to agronomic traits and for anchoring new sequence assemblies.¹³ High-throughput pooling strategies allow labs to screen massive BAC libraries quickly, identifying positive clones without manually deconvoluting complex pools.¹⁴

Rewriting Genes Inside a BAC

One of the most powerful things about BACs is that they can be surgically modified inside E. coli using a technique called recombineering. The basic idea is to borrow recombination enzymes from bacteriophage Lambda (the “Lambda Red” system) and use them to swap, delete, or insert DNA sequences at precise locations within the BAC insert. Because BACs carry such large pieces of a genome, this lets researchers make point mutations, delete a regulatory element, or knock in a reporter gene without having to reconstruct the surrounding region from smaller fragments.

The approach has been used to build transgene constructs for animal studies. A BAC carrying a gene of interest can be modified so that a fluorescent reporter replaces the gene’s coding sequence, or so that a specific mutation is introduced to model a disease. The finished construct is then injected into fertilized eggs to make transgenic mice or other animals.¹⁵ Because the BAC carries not just the gene but also its distant regulatory sequences, BAC transgenes tend to produce expression patterns that mirror the gene’s natural behavior more faithfully than small-plasmid transgenes do. They are more likely to drive copy-number-dependent expression that does not depend heavily on where the transgene lands in the host chromosome.¹⁶

Cloning Entire Viral Genomes

Large DNA viruses, particularly herpesviruses, have genomes in the 100- to 250-kilobase range, which fits neatly within a BAC’s capacity. Researchers have exploited this by cloning entire herpesvirus genomes into BACs, turning each viral genome into a stable E. coli plasmid that can be manipulated with all the tools of bacterial genetics. When the modified BAC DNA is transfected back into mammalian cells, the cells produce infectious virus, allowing researchers to study the effect of each genetic change on viral behavior.

This strategy has been applied to bovine herpesvirus 1, where a BAC clone of the full genome produced virus with replication kinetics indistinguishable from the parent strain.¹⁷ It has also been used for Kaposi’s sarcoma-associated herpesvirus (KSHV), where a full-length BAC clone was rescued in E. coli, modified, and then used to generate infectious virions after transfer into human cells.¹⁸ More recently, direct cloning methods have been developed that bypass some of the older multi-step approaches, allowing researchers to capture a herpesvirus genome into a BAC and observe viable virus emerging from transfected cells within just a few days.¹⁹

The same recombineering tools described earlier can be applied to viral BACs. For coronaviruses like MERS-CoV, Lambda Red recombination on a BAC carrying the viral genome allows researchers to introduce point mutations, insertions, or deletions and then recover recombinant virus to test hypotheses about pathogenesis and drug targets.²⁰ This kind of reverse genetics would be far more cumbersome without a stable, easily modified clone of the complete genome.

Mining Microbes for New Drugs

Soil bacteria in the genus Streptomyces are the source of a huge fraction of the antibiotics and other bioactive compounds used in medicine, but many of their biosynthetic gene clusters are “cryptic,” meaning they stay silent under standard laboratory growth conditions. BAC libraries offer a way around this problem. Researchers can build a BAC library from a Streptomyces genome, transfer individual BAC clones into a well-characterized surrogate host strain, and then screen the resulting colonies for antimicrobial activity or other bioactive properties.

One group demonstrated this by building a BAC library from Streptomyces rochei, expressing it in S. lividans, and screening for antimicrobial compounds. The screen recovered the known gene clusters for streptothricin and borrelidin but also turned up two novel linear lipopeptides and an entirely unknown antibiotic cluster that had never been detected in the original organism.²¹ Because many biosynthetic pathways span 50 to 100 kilobases or more, capturing them in a single BAC clone is critical. A specialized shuttle vector called pSBAC was designed to move between E. coli and Streptomyces, and it was validated by cloning and expressing the approximately 90-kilobase meridamycin gene cluster, one of the largest polyketide biosynthetic pathways known at the time.²²

BACs in Gene Therapy Research

Most gene therapy vectors, whether viral or non-viral, have strict size limits. Adeno-associated virus vectors top out at about 4.7 kilobases of cargo, and lentiviral vectors hold roughly 8 to 10 kilobases. Many human disease genes, with all their introns, regulatory elements, and distant enhancers, far exceed those limits. BACs offer one potential solution: deliver the whole genomic locus, regulatory architecture and all, in a single construct.

This idea has been tested in proof-of-concept experiments. Researchers delivered a 135-kilobase BAC carrying the entire human low-density lipoprotein receptor (LDLR) gene locus to mouse livers using a hydrodynamic injection technique. The BAC DNA was retained in hepatocytes, and human LDLR protein was still detectable four months after delivery.²³ In cell culture, a peptide-lipid delivery system was used to transfect BAC-based constructs up to 242 kilobases into human cells. About 60 percent of cells took up smaller constructs, while constructs as large as 242 kilobases still reached over 10 percent of cells. When a 143-kilobase BAC carrying the NBS1 gene was delivered to cells from patients with Nijmegen breakage syndrome, at least 57 percent of the cells that showed reporter expression also produced functional nibrin protein.²⁴

These are still laboratory-stage experiments, not clinical therapies. Delivering large DNA molecules efficiently and safely into specific tissues in a living person remains a formidable engineering problem. But the results demonstrate a principle that no small-vector system can match: a BAC can carry a gene in its full genomic context, regulatory elements included, and produce physiologically relevant expression.

Practical Headaches of Working with BACs

For all their advantages, BACs are not easy to handle. The DNA molecules are enormous by molecular-biology standards, and they break easily. Vigorous shaking, repeated pipetting, or even routine vortexing can shear a 200-kilobase BAC insert into useless fragments. Labs working with BACs use wide-bore pipette tips, avoid agitation wherever possible, and sometimes add polyamines or salt to the purification buffer to compact the DNA strands and reduce shearing. Standard commercial plasmid-prep kits are not designed for molecules this large; specialized kits or manual alkaline-lysis protocols are needed.²⁵

While BACs are generally stable, they are not immune to rearrangement. Inserts containing long tandem repeats or highly repetitive sequences can sometimes undergo deletions or rearrangements even at single copy, because the host cell’s repair machinery occasionally acts on the repeats. This is a concern for genomes rich in satellite DNA or transposable elements.⁸ Researchers typically screen newly constructed libraries for rearrangements by comparing restriction-digest fingerprints of individual clones to expected patterns, and they avoid using clones that show discrepancies.

Yield is another practical limitation. Because BACs replicate at one or two copies per cell, a standard overnight culture produces far less DNA per milliliter than a high-copy plasmid prep. The conditionally amplifiable BAC systems described earlier help, but they add complexity. And the electroporation step during library construction remains inefficient for the very largest inserts, which means that building a deep-coverage library from an organism with a large genome requires a lot of ligation reactions and a lot of electroporations.

BACs in an Era of Long-Read Sequencing

Modern long-read sequencing platforms can now generate reads of tens or even hundreds of kilobases from a single molecule of native genomic DNA, which has led some to ask whether BAC libraries are obsolete. For de novo genome sequencing of organisms with good reference genomes, BACs are indeed less central than they once were. A long-read sequencer can span many of the repetitive regions that used to require a BAC-based physical map to resolve.

But BACs have not disappeared. They remain the tool of choice when a researcher needs a stable, renewable physical clone of a specific genomic region, for example to build a transgenic construct with all the surrounding regulatory sequence, to capture and manipulate a viral genome, or to express a large biosynthetic gene cluster in a heterologous host. A sequencing read is data; a BAC clone is a physical reagent that can be propagated indefinitely, shared between labs, modified by recombineering, and injected into cells. Those are fundamentally different things, and no sequencing technology replaces the need for the reagent itself. BAC repositories continue to distribute clones worldwide, and new BAC libraries are still being built for organisms where functional genomics, positional cloning, or comparative mapping requires physical clones tied to a genetic map.