
Almost every plasmid cloning project follows the same five stages, whether you are moving one gene with restriction enzymes or building a ten-part Golden Gate assembly: design, prepare, join, transform, and verify.
The method changes what happens in the middle. It does not change the order of the stages, and it does not change where the failures come from. A single base in the wrong place on Monday gives you a construct that grows happily in bacteria on Friday and does not express the protein you wanted.
This guide walks through the five stages and the points where things most often go wrong.
At a glance
- Workflow: Design in silico, prepare the DNA, assemble the construct, transform competent cells, verify clones.
- Key distinction: Colonies on a selective agar plate are candidates, not evidence of a correct construct.
- Final check: Sequence across the insert, the junctions, and anything that was amplified by PCR.
Step 1: In silico design
A cloning experiment starts with the sequence you intend to end up with. Before you order primers or thaw an enzyme, build the planned construct on screen and inspect it carefully.
Three things are worth checking every time:
- The reading frame, the start and stop codons, and any fusion junction in an expression construct.
- Your vector and insert for unintended internal restriction sites.
- The length, melting temperature and uniqueness of each overlap or overhang, if your method uses them.
Tools such as Photo51, SnapGene and Benchling (see our Photo51 vs. SnapGene vs. Benchling comparison) let you lay out the final plasmid, read the coding sequence and see where each junction falls, so that a frameshift or a duplicated overhang shows up now rather than after two rounds of screening. A design check is cheap and it removes a whole category of avoidable failure. It cannot tell you whether the construct will express, or whether the host will tolerate it.
Step 2: Insert and vector preparation
Once the design is settled, you need the physical pieces:
- Vector: Usually an existing circular plasmid, opened with restriction enzymes or amplified as a linear backbone by PCR.
- Insert: Amplified from genomic DNA, cDNA or another plasmid, or ordered as synthetic DNA. During PCR, a custom primer can add restriction sites or assembly overlaps to the ends in the same step.
- Cleanup: Check fragment size and purity. Depending on the next reaction, you may need to remove enzymes, template DNA, salts or unwanted fragments, with a column cleanup or a gel extraction.
Some assembly kits tolerate a small volume of unpurified PCR product, which saves a step when the band is clean. What they do not tolerate is carryover from extraction buffers. Guanidine salts and residual ethanol inhibit the enzymes in the next reaction, and this is a common reason a well-designed assembly gives nothing.
Step 3: The assembly reaction
This is where your insert and vector are physically joined, and the protocol depends entirely on the method you chose at the design stage:
- Traditional restriction cloning joins digested fragments with T4 DNA ligase.
- Gibson Assembly and NEBuilder HiFi use an exonuclease, a polymerase and a ligase to join fragments through matching overlaps.
- Golden Gate Assembly combines a Type IIS enzyme and a ligase in one tube for ordered, multi-part assemblies.
- Gateway cloning uses site-specific recombination instead of cutting and joining at all.
Working in moles, not microliters
Do not mix equal volumes of insert and vector. The reaction responds to the molar ratio of the fragments, while a concentration in ng/µL tells you only about mass.
A 3:1 insert-to-vector molar ratio is a common starting point for a single-insert ligation. Overlap and multi-fragment methods need their own ratios, and these change with the number of fragments, so take the figure from the protocol for your method rather than from habit.
To convert, use each fragment's length:
ng of insert = ng of vector × (insert length in bp ÷ vector length in bp) × molar ratioToo much insert favors unwanted products, such as several inserts joined together. Too little reduces the yield of the product you want.
Step 4: Transformation
You now have an assembled plasmid in a tube. To make copies of it, you need it inside a host, almost always competent E. coli.
- Competent cells: Use cells prepared for chemical transformation or for electroporation. Commercial ultra-competent cells are sold with reported efficiencies above 109 cfu/µg, and are worth the cost when your assembly is precious or your product is a nicked circle. Cells prepared in the lab are usually fine for routine subcloning.
- Heat shock: The most common method. A standard protocol mixes DNA with cells on ice, holds them at 42 °C for 30 to 45 seconds, then returns them immediately to ice. The temperature spike changes the membrane briefly enough for DNA to enter. Strains differ in how much they tolerate, so follow the protocol for the strain you have rather than a generic one, particularly for the time on ice and the length of the heat shock.
- Recovery: After the heat shock, add rich medium such as SOC or LB and incubate at 37 °C for about an hour before plating. Do not skip this. The cells are fragile, and they need time to repair their membranes and express the resistance gene before they meet the antibiotic.
Step 5: Screening and verification
Spread the recovered cells on agar containing the selective antibiotic. Only cells carrying your plasmid will grow.
This is the point where it is easiest to believe you are finished. You are not. Bacteria take up empty, re-ligated vector readily, and an empty vector carries the same resistance marker as your construct, so it grows just as well.
- Colony PCR: Pick three to ten colonies and screen them with primers that flank the insert site, to check the fragment size. Grow the promising ones in liquid medium for a miniprep.
- Whole-plasmid sequencing: Long-read nanopore services sequence the entire construct in one run, with no primers to design. Some commercial providers advertise roughly 10 to 15 euros or dollars per plasmid with next-day results, which can be cheaper than Sanger for a construct where you would otherwise order three or more reads. You can usually send an overnight culture or a miniprep directly.
- Sanger sequencing: Still the better choice for a single short check, and often more accurate per base on a short fragment, which makes it useful for confirming a difficult insert region.
The difference is coverage rather than accuracy. Sanger reads only where your primer binds, so it cannot show you what it was not aimed at. A deletion that removes the primer site is invisible, because the primer only reports the molecules that still match your intended sequence. Plasmid dimers, duplications and backbone rearrangements are missed for the same reason. Very GC-rich regions, such as the CAG promoter, and repeats longer than a single read are difficult or impossible to read by Sanger.
Prices and turnaround vary a great deal by provider, and an institutional core facility is not always cheaper or faster than a commercial service. Check both before you decide what routine verification looks like in your lab.
Sequence anything that was amplified by PCR, even with a high-fidelity polymerase. Proofreading reduces the error rate, but does not remove it.
Troubleshooting
Zero colonies on the plate
Symptom: You followed the protocol, the cells should be competent, and the plate is bare.
Cause: Competent cells lose efficiency quickly if freezer temperatures fluctuate or if they are thawed warm, and a dead batch looks exactly like a failed assembly.
Fix: Run a positive control transformation alongside the experiment, using about 10 pg of an intact, verified plasmid such as pUC19. If the control plate is also bare, the cells are the problem. Discard them and use a fresh batch before you touch anything in the assembly.
Every colony contains an empty vector
Symptom: You have plenty of colonies, and colony PCR shows a backbone with no insert.
Cause: Your cut vector closed back on itself instead of accepting the insert, or uncut template survived into the transformation.
Fix: Treat the digested backbone with an alkaline phosphatase such as rSAP to remove its 5' phosphates, so it can only close if a phosphorylated insert bridges the gap. Better still, cut with two enzymes that leave incompatible ends. If your vector or insert was amplified by PCR from a plasmid template, digest the product with DpnI to remove the methylated template.
Poor purity readings and a failed assembly
Symptom: Your gel-extracted fragments give a 260/230 ratio below about 1.5, and the assembly produces nothing.
Cause: Guanidine salt carried over from the gel-dissolving buffer. These salts inhibit the enzymes used in assembly reactions.
Fix: When you use a silica column for gel extraction, let the wash buffer sit on the column for five minutes before spinning, and add an extra ethanol wash before elution. Where your PCR gives a clean band of the right size, a column cleanup instead of a gel extraction avoids the problem entirely.