Types of Molecular Biology Vectors: Cloning to Expression

A vector is a DNA vehicle that carries a foreign insert into a host cell so the insert can be replicated, selected, or expressed. Most routine cloning still uses plasmids, but the same job family includes shuttle systems, viral particles, and large-insert backbones.

Cloning teams choose a vector by matching insert size, host range, selectable marker, and whether the construct must only be stored or also expressed. The comparison below starts with the major classes, then unpacks origin, marker, MCS, and promoter decisions before wet-lab work.

What a vector is in molecular biology

Labs often say plasmid when they mean the circular backbone in the freezer. Vector names the job: move an insert into a host and keep that insert selectable. Many vectors are plasmids. Packaged viral genomes and some linear or large-insert systems still do the carrying job without being a routine cloning plasmid sitting in E. coli at high copy.

The useful question is therefore not a second definition. It is which job the construct must finish. Storage and subcloning need a stable, easy-to-prep backbone and a multiple cloning site. Expression needs a promoter, ribosome-binding or Kozak context, terminator, and a host that can run that cassette. Transfer between species needs dual replication and dual selection. Delivery into hard-to-transfect cells may need a viral packaging system. Cloning a 150 kb locus needs a low-copy large-insert host, not a 3 kb pUC derivative.

Four backbone parts show up in almost every decision, even when the class name changes. The origin of replication (ori) sets host range and copy number. The selectable marker decides how you keep the construct and whether that marker is legal in the downstream cell type. The multiple cloning site (MCS) or equivalent insertion strategy decides how the insert goes in. The promoter (or its absence) decides whether the insert is a silent passenger or an expressed cassette. Mis-matching any one of those four is a more common failure than picking the wrong marketing class name.

Overview table of major vector classes

Scan the classes first, then use the H3 notes for the failure modes that the table cannot hold.

Cloning plasmids

Cloning plasmids are optimized for making more DNA, cutting it, and putting pieces together. Classic features are a bacterial ori, a marker such as ampicillin or kanamycin resistance, and an MCS inside a reporter (often lacZα) so empty versus recombinant colonies can be distinguished. They may include a primer-binding region for routine sequencing. They usually lack a complete expression cassette for the host you care about, and that absence is a feature when the insert is toxic or when you only need a holding construct.

Copy number is the hidden variable. A pUC-type ori at high copy is convenient for minipreps and painful for inserts that titrate a repressor or encode a leaky toxic product. Switching to a moderate- or low-copy ori is a cloning decision, not an expression-philosophy debate. Restriction uniqueness in the MCS still matters for scar-sensitive work, even if the lab has moved most assemblies to Gibson or Golden Gate. The junctions are still sequence features that must be true on the map.

Expression plasmids

Expression plasmids add the parts that turn an insert into a product: promoter, 5' untranslated context, coding-sequence frame, optional tags, and a terminator. The same insert that was harmless in a cloning backbone can become a metabolic burden once a strong promoter is present. That is why many workflows clone first, then move the ORF into an expression backbone, instead of forcing one plasmid to be both a construction chassis and a production cassette.

Host match is non-negotiable. A T7 promoter is not a mammalian promoter. A CMV cassette will not rescue an E. coli expression failure. Ribosome-binding sites, Kozak sequences, codon usage, and secretion signals travel with the host decision. Induction method (IPTG, analog, serum withdrawal, doxycycline, or constitutive) belongs on the same line as the promoter name. If those fields are missing from the map, the wet-lab "expression plasmid" is only a cloning plasmid with extra arrows.

Shuttle vectors

Shuttle vectors carry two replication systems and usually two markers so the construct can live in both a convenient prep host and an assay host. A common pair is E. coli plus Saccharomyces cerevisiae: a ColE1-type ori and ampicillin marker for bacteria, plus a yeast ori (or integration homology) and a yeast auxotrophic or drug marker. Other pairs exist for bacteria-to-bacteria transfer, Agrobacterium workflows, or bacteria-plus-mammalian maintenance, depending on the lab's organisms.

The failure mode is silent loss of one ori or one marker during assembly. A colony that grows on bacterial ampicillin has not proven that the yeast marker is intact. Each host's selection must be verified after construction, and the map should show both origins as annotated features, not as a footnote that says "shuttle." Prep in the easy host, then confirm the construct still transforms and selects in the second host before you invest in the phenotype assay.

Viral vectors

Viral vectors use a viral packaging pathway to deliver DNA or RNA into target cells. Lentiviral, retroviral, AAV, and adenoviral systems differ in cargo limit, integration behavior, tropism, and biosafety classification. In the cloning lab they still begin as plasmids: transfer plasmids, packaging plasmids, and envelope plasmids. The vector that matters to the cell is the packaged particle, not the bacterial prep.

Use a viral system when plasmid transfection cannot reach the cell type, when you need longer expression after a single delivery, or when the experiment requires a defined tropism. Do not use one as a default "stronger plasmid." Packaging limits are hard. Oversized genomes produce low titer or incomplete particles. Institutional biosafety review is part of the design, not an after-the-fact permit. Titer method, multiplicity of infection, and empty-particle controls belong in the experiment record beside the map.

Large-insert systems: cosmids, BACs, and YACs

When the insert is tens to hundreds of kilobases, standard plasmids become unstable or simply too small. Cosmids use lambda cos sites and can carry intermediate inserts with phage packaging. Bacterial artificial chromosomes (BACs) use a low-copy F-plasmid ori and are a standard way to hold large genomic fragments in E. coli. Yeast artificial chromosomes (YACs) can hold still larger inserts in yeast, at the cost of more rearrangement risk and more awkward DNA handling.

These systems are still vectors: they carry and maintain an insert. They are poor first choices for a 1 kb ORF. Low copy means low yield. Verification needs restriction fingerprinting, long-read sequencing, or tiled PCR, not a single diagnostic digest that worked for a 6 kb plasmid. If the scientific object is a whole locus, an operon cluster, or a large repeat-containing region, start with the size ceiling, not with the promoter you wish you had.

How to match a vector to the experimental job

Match the job before you match a catalog name. Write the host, the insert size, whether the insert must be expressed, whether you must move between species, and whether delivery requires a particle. Those five answers usually collapse the table to one class. Only then pick ori, marker, MCS chemistry, and promoter.

Decision parts, in the order that prevents wasted oligos:

1. Host and biosafety. The organism that must maintain or express the construct sets legal markers and legal promoters. A bla marker that is fine in cloning E. coli may be unwanted in a cell line that already carries beta-lactamase reporting.
2. Insert size and stability. If the fragment exceeds ordinary plasmid comfort, stop considering high-copy cloning backbones and move to BAC, cosmid, or a reduced-copy plasmid.
3. Replication origin and copy number. High copy for prep and mutagenesis; lower copy for toxic inserts and large fragments. Dual ori only when you truly need two living hosts.
4. Marker. One selectable marker per host you will grow. Confirm it does not collide with existing resistances in the strain or cell line.
5. Promoter and MCS. Empty promoter for storage; host-matched promoter for expression. Choose assembly chemistry (restriction MCS, Gibson overlaps, Golden Gate overhangs) that your map can actually simulate.

Cloning versus expression is the most common mix-up. A cloning plasmid can hold an ORF. It does not thereby become an expression plasmid. An expression plasmid can be cloned into, but using it as the construction chassis raises leaky-expression risk. Shuttle versus two separate plasmids is the next mix-up: if you only ever prep in E. coli and transfect mammalian cells without replicating in them, you may need an expression plasmid, not a true shuttle.

Viral versus plasmid is a delivery question. If chemical or electroporation transfection already reaches the cells and you need only transient expression, a plasmid is the smaller system. If you cannot transfect, or you need integration or durable expression that the plasmid does not provide, evaluate a viral particle and its cargo limit. Public backbone collections, including the Zettalab Plasmid Library, are starting points for names and maps. They are not a substitute for checking insert size, marker legality, and license terms on the specific record you intend to use.

Annotating vector features before wet-lab work

A class name on a tube is not a map. Before restriction digests, assembly, or transfection, annotate ori, every marker, the MCS or assembly junctions, promoters, terminators, tags, primer sites, and any packaging signals. Then translate the ORF in the intended frame. Most failed "vector choices" discovered at the bench were map errors: a promoter pointing the wrong way, a scar that destroyed a Kozak sequence, or a leftover att site that shifted numbering.

Annotation should be specific enough that another person can regenerate the same feature table. "AmpR" without coordinates is weaker than a feature with start, end, strand, and the evidence (sequence file versus vendor datasheet). "CMV" without the exact promoter fragment leaves enhancer boundaries unknown. Viral transfer plasmids need packaging signals marked so an accidental deletion is visible before packaging week.

In silico assembly belongs in the same pass. Simulate the junctions, confirm unique overhangs or homology arms, and export a post-assembly map that still lists ori and markers. Sequence-design workspaces such as ZettaGene can hold that annotated map beside later alignment of colony PCR or Sanger reads. A standalone editor can do the same job if the exported file and the experiment record share one construct ID. If the software is removed, the GenBank or equivalent file should still explain why that backbone was legal for the host.

Selection criteria for the software layer are separate from vector biology. If the team is choosing a construction tool rather than a backbone class, use a dedicated software evaluation, such as Zettalab's plasmid construction software guide. The biological decision remains: ori, marker, MCS, promoter, and class constraints. Documentation of that decision can follow the lab's workflow guides so the map, the oligo order, and the transfection note stay joined.

FAQ

What types of vectors are used in molecular biology and cloning?

The types that cover most bench work are cloning plasmids, expression plasmids, shuttle vectors, viral vectors, and large-insert systems such as cosmids, BACs, and YACs. Cloning plasmids store and rearrange DNA in a convenient bacterial host. Expression plasmids add a host-matched cassette so the insert can be transcribed. Shuttle vectors keep dual origins and markers so one construct lives in two species. Viral vectors package cargo for delivery into cells that plasmids transfect poorly. Large-insert systems exist because ordinary plasmids cannot stably hold very long loci. Choose by job and size first, then by ori, marker, and promoter, not by a familiar catalog nickname.

What is the difference between cloning vectors and expression vectors?

A cloning backbone is built for propagation, restriction or assembly, and storage. An expression backbone is built so a host can run a transcription cassette. The same circular DNA can contain both kinds of parts, but the jobs remain different. If you only need more DNA or a holding construct, a strong promoter is a risk, not a benefit, especially with toxic ORFs. If you need protein or RNA output, a cloning-only plasmid that lacks the correct promoter, 5' context, and terminator will not start working because the insert is present. Move the ORF after the junctions are proven, or start on an expression backbone and keep copy number and leakiness under control.

When should a lab use a viral vector instead of a plasmid?

Use a viral particle when plasmid transfection cannot reach the target cells, when you need a tropism the plasmid mix cannot provide, or when the experiment requires integration or longer expression after one delivery. Stay with a plasmid when transfection already works and you need only transient expression or a simple reporter. Viral systems add packaging plasmids, cargo limits, titer assays, and biosafety review. They are not a louder version of the same backbone. Measure cargo length against the system's published packaging comfort, and record multiplicity of infection beside the map. If those extras are not required by the biology, the bacterial plasmid remains the smaller, easier system.

How do you choose an origin of replication and a selectable marker?

Choose the ori for host range and copy number. High-copy bacterial origins help minipreps and site-directed mutagenesis. Lower-copy origins help toxic inserts and larger fragments. A second ori appears only when the construct must replicate in a second species. Choose the marker for the host you will actually plate or culture, and check existing resistances in that strain or cell line. Ampicillin in cloning E. coli does not prove a yeast auxotrophic marker is intact, and a mammalian drug marker is useless if you never transfect those cells. Write both choices on the map with coordinates. Do not infer them from the vendor's class name.

Do labs need different software for each vector class?

No. The classes differ in biology. The software task is the same: annotate features, simulate assembly, and keep the map joined to later verification. A workspace that handles plasmids can usually represent a BAC or a viral transfer plasmid if insert length and feature types are not artificially capped. Evaluate tools by map fidelity, assembly simulation, and record linkage, not by a separate product SKU per class. A sequence editor and a public backbone catalog are helpers, not different biology. Neither replaces checking ori, marker, promoter, and license on the specific file you clone from.

Conclusion

Vector class names are shortcuts for jobs: store, express, shuttle, deliver, or hold a large locus. The parts that make the shortcut true are ori, marker, MCS or assembly junctions, and promoter. Plasmids are the common physical form, not the whole category. Annotate those parts on a map before wet-lab work, and verify the construct in every host you claim it can enter. To browse public backbone records while you keep design files in the same research workspace, open the Zettalab plasmid collection and confirm size, marker, and license on the entry you intend to use.