Cloning a glycosylated human gene into pET-28a

The internal NdeI site is the central problem. Digestion of the PCR product at both the designed ends and the internal site cuts the insert into two fragments, so intact full-length insert is rare and ligation efficiency collapses. That explains the low colony count. The 9 colonies with no T7 product are not true recombinants; they are most likely escapers or plate artefacts, or colonies whose plasmid is too large or rearranged for the PCR, and they should be re-streaked and retested rather than counted as clones.

Redesign:

• Choose enzymes absent from the insert, verified by restriction mapping of the actual sequence, for example NcoI and XhoI, or use Gibson assembly or SLIC, which need no internal-site compatibility.
• Gel-purify the correct insert band and phosphatase-treat the vector to reduce self-ligation.
• Use a 3:1 molar insert-to-vector ratio, include a no-insert control, and check the vector digest is complete.
• Confirm every clone by restriction digest and full-length sequencing of both strands.

Reasoning: adding antibiotic cannot fix a cloning design error; it only selects harder for the few surviving plasmids.

Taq polymerase lacks proofreading and has an error rate of roughly 1 in 10^4 to 10^5 bases per cycle. Over 30 cycles on 1.8 kb, the expected number of errors per molecule is about 1800 x 30 x 10^-4, which is around 5 errors per molecule in the worst case, and lower but still substantial in the best case. Most errors are substitutions, but a single-base deletion that shifts the reading frame is exactly the kind of low-probability event that appears when many molecules are cloned and only one is sequenced.

The correct response is not simply to repeat with Taq:

• Re-amplify with a high-fidelity proofreading polymerase, use 18 to 22 cycles, and increase template quality.
• Sequence at least three independent clones in both directions.
• If the same deletion recurs, check whether it sits in a primer or a secondary-structure hotspot; redesign the primer if so.
• Subclone the verified insert into the expression vector after sequencing, not before.

Reasoning: the construct must be sequence-verified before any functional conclusion is drawn, because a frameshift produces a truncated, inactive protein.

E. coli is not acceptable for the intended assay. It cannot perform N-linked glycosylation, so the product will be aglycosylated. For a human glycoprotein this often means misfolding, loss of activity, altered pharmacokinetics and potential immunogenicity. Faster expression is not a justification when the required post-translational modification is missing.

Acceptable alternatives are CHO or HEK293 cells for authentic human-type glycosylation, or Pichia pastoris if high-mannose glycans are acceptable. The choice must be justified by the assay's requirements, not convenience.

Biosafety and regulatory points:

• E. coli K-12 is a low-risk host, but the recombinant still requires containment, autoclaving of waste and a strain record under institutional biosafety rules.
• The kanamycin-resistance marker should be recorded and, where possible, avoided in strains intended for larger scale.
• Mammalian expression brings additional requirements: mycoplasma and adventitious-agent testing, a master cell bank, viral clearance validation, and documented segregation.
• If the protein is for human use, GMP and medicine-regulation requirements apply; if for research only, institutional biosafety committee approval is still required.

The judgement is that host choice and containment are scientific and regulatory decisions, not just scheduling ones.