Why the Insulin Batch Was Inactive
A small fermentation unit tries to make human insulin in Escherichia coli. A technician homogenises leftover human pancreatic tissue from a surgery, extracts high-molecular-weight DNA, digests it with BamHI, ligates the fragments into the BamHI site of pBR322 and transforms E. coli HB101. Transformants are selected on ampicillin and then replica-plated onto tetracycline; about 40% of the ampicillin-resistant colonies fail to grow on tetracycline. The best colony is confirmed to carry an insulin sequence by colony hybridisation with an insulin cDNA probe, grown in a 20-litre fermenter, induced and lysed. No insulin is detected by radioimmunoassay. The manager wants to repeat the run with more glucose and a longer induction period; the consultant says the design, not the fermentation, is wrong.
1. Why is no insulin produced even though the human sequence is present? Give at least three independent reasons and rank them.
Read the worked answer
Reason 1 - introns (absolute block). The human insulin gene contains introns. E. coli has no spliceosome and cannot remove them, so the transcript cannot be translated into the correct proinsulin. A cDNA copy made from mature beta-cell mRNA is needed instead.
Reason 2 - promoter recognition (absolute block). pBR322 is a cloning vector, not an expression vector. Even a perfect coding sequence placed there sits behind a human promoter that E. coli RNA polymerase does not recognise. A bacterial promoter (lac, trp, T7) with a ribosome-binding site is required.
Reason 3 - post-translational processing (major). Human insulin is made as proinsulin and cleaved by specific proteases (PC1/PC3 and carboxypeptidase E) to remove the C-peptide, and the A and B chains are then joined by two disulfide bonds. E. coli performs neither the cleavage nor reliable disulfide formation, so even a correct cDNA may yield an inactive single chain or misfolded aggregates.
Reason 4 - tissue source (practical). Beta cells are a tiny fraction of pancreatic tissue, so the insulin gene is a very rare sequence in a genomic library and the clone obtained may be incomplete.
Ranking. The promoter and the introns are absolute blocks that no change in fermentation conditions can overcome; processing defects come next. Adding glucose and extending induction cannot fix any of them.
2. What cloning strategy would actually produce active insulin, and how would you confirm success?
Read the worked answer
Start from mRNA, not genomic DNA. Isolate mRNA from insulin-producing beta cells, use reverse transcriptase to make a cDNA library, and screen it with an insulin probe. cDNA carries no introns, so E. coli can translate it.
Use an expression vector. Insert the cDNA downstream of a strong, regulatable bacterial promoter with a good ribosome-binding site. Because the A and B chains must be joined by disulfide bonds, a common design is to express the A and B chains separately as fusion proteins with a short leader, cleave the leader chemically (for example with cyanogen bromide at a methionine linker), and then oxidise the two chains together under controlled conditions so the correct disulfide bonds form. Alternatively express proinsulin in a host that can secrete and process it, such as yeast, or in a mammalian cell line.
Fermentation and recovery. Induce at the right cell density, harvest, lyse, purify by chromatography, and fold.
Confirmation. Three levels are needed: presence of the DNA insert (restriction mapping and sequencing, not just hybridisation), presence of the protein (SDS-PAGE and Western blot with anti-insulin antibody), and biological activity (radioimmunoassay plus a receptor or cell-based assay). A band on a gel or a positive hybridisation alone does not prove active hormone, which is exactly the trap this unit fell into.
3. Interpret the tetracycline replica-plating result. Does 40% mean that 40% of the colonies are recombinants making insulin?
Read the worked answer
What the test detects. The BamHI site of pBR322 lies inside the tetracycline-resistance gene, so a fragment inserted there disrupts the gene: the recombinant is ampicillin-resistant but tetracycline-sensitive, whereas the empty vector is resistant to both. Growth on ampicillin selects for plasmid-bearing cells; replica plating onto tetracycline then distinguishes recombinants (no growth) from non-recombinants (growth). This is insertional inactivation, and 40% is a plausible recombinant frequency.
Why 40% is not 40% insulin producers. First, tetracycline sensitivity can also result from deletions, rearrangements or instability of the plasmid, not only from a genuine insert. Second, many different BamHI fragments were ligated, so a recombinant colony may carry an unrelated human fragment; hybridisation identified the insulin sequence only in the chosen colony. Third, the fragment can insert in either orientation, and only one orientation places the coding strand correctly relative to the promoter - and pBR322 has no usable promoter for this purpose anyway. Fourth, the figure is a proportion of ampicillin-resistant colonies, not of all transformants.
What is needed. Restriction mapping and sequencing of the junction to confirm insert, size and orientation, followed by an expression test. The replica-plating result is a screening step, not evidence of a product.