C5Orf15 Protein Structure
6.1 Changing a protein on purpose #
6.1.1 Directed evolution #
The cycle is diversify, express, select or screen, and repeat. Frances Arnold showed it could improve enzymes for unnatural conditions. She shared the 2018 Nobel prize with Smith and Winter, for phage display of peptides and antibodies.
Error-prone PCR and Stemmer’s DNA shuffling make the variation. Saturation mutagenesis explores every amino acid at a few positions. Selection ties the property to growth. Screening is required when it does not, which is most industrial properties. You get what you select for: a screen on an artificial substrate can worsen the real one. Phage display keeps genes whose products bind a surface. It does not, unless specially arranged, select serum stability or catalysis. Affinity can rise. So can stickiness.
6.1.2 Rational design #
Rational design changes a residue a structure nominates. Replacing a bleach-sensitive methionine in subtilisin is the early industrial example. A wrong hypothesis often yields only a dead enzyme, which does not say why. Structures are snapshots and sometimes the wrong ligand. The usual compromise is a structure-chosen library, then a new structure of the winner. Design returns an explanation that may be false. Evolution returns a winner that may be solving another problem.
Site-directed mutagenesis is the laboratory step behind a rational substitution: one specified codon, the method recognised alongside PCR in 1993. It does not choose the codon. Activity, stability, and the original substrate still have to be measured, because a more stable enzyme is often a slower one.
6.2 Fusions and antibodies #
6.2.1 Fusion proteins beyond the purification tag #
A fusion can add GFP and a localisation report, an Fc domain and a longer half-life, or a secretion signal. Etanercept is a tumour-necrosis-factor receptor fused to Fc; Immunology owns the cytokine. The fusion is a new protein. The partner can hide a sorting signal, and the linker can be too short or cleaved too soon.
A His-tag is the small case of the same idea. It is chosen because six histidines are less likely than GFP or an Fc domain to dominate the fold, and they can still break a protein or sit on an inactive fragment. The measurement that counts is activity of the molecule you made.
6.2.2 Monoclonal antibodies and humanisation #
Köhler and Milstein’s 1975 hybridoma immortalises one specificity: the antibody of one B cell and the growth of a myeloma. A mouse antibody is immunogenic in humans and is cleared, often with an immune reaction.
A chimaeric antibody keeps mouse variable domains and human constant domains. A humanised antibody keeps only the mouse complementarity-determining regions; framework residues are often put back because grafting drops affinity. Fully human antibodies come from phage display or from mice with human immunoglobulin loci. Adalimumab was the first approved fully human phage-display antibody. Fc sequence and glycosylation set effector function and half-life. That biology is Immunology; the CHO host is Chapter 3.
6.3 Pathways and sensors #
6.3.1 Metabolic engineering and artemisinic acid #
Flux control is shared: coefficients along a pathway sum to one, so overexpressing a supposed rate-limiting enzyme often does little. A competitor drains the precursor, an intermediate is toxic, or a cytochrome P450 lacks its reductase. Balancing beats maximising every gene.
Artemisinin comes from Artemisia annua, and botanical supply has been uneven. Yeast makes farnesyl pyrophosphate for sterols. Ro, Keasling, and colleagues reported in 2006 a heterologous route from that precursor to artemisinic acid, using amorphadiene synthase, CYP71AV1, and a reductase. Later strains reached grams per litre. The sterol pathway competes, and intermediates can be toxic, so supply and competition had to be engineered too. The acid is not artemisinin; chemistry finishes it. Factory constraints and a competing plant crop remained. A pathway paper is not a supply chain.
6.3.2 Biosensors #
A biosensor turns recognition — transcription factor, riboswitch, enzyme, or antibody — into fluorescence, current, or growth. Used to screen a pathway, it is only as honest as its specificity: if it also sees an upstream intermediate, selection will accumulate that intermediate. A glucose oxidase electrode is the old clinical example of the same idea. These notes do not build either device.
6.4 Key terms, summary, and self-check #
Key terms. Directed evolution; DNA shuffling; phage display; rational design; fusion protein; hybridoma; chimaeric antibody; humanisation; flux control; heterologous pathway; artemisinic acid; biosensor.
Summary. Directed evolution gets what the assay rewards. Rational design is fast when the structure is right and opaque when it is wrong. Antibodies are humanised, or made human, so they can be dosed; fusions add a second function. Metabolic control is shared. Yeast made artemisinic acid, a real heterologous product, and scale-up, chemical conversion, and agricultural competition stayed separate. A biosensor cannot rescue a non-specific screen.
Self-check.
1. Why can a large library still yield the wrong enzyme?
2. A designed substitution kills activity. What can you not conclude?
3. What mouse sequence remains in a chimaeric antibody, and in a humanised one?
4. Why was adding the plant genes not enough for artemisinic acid, and why did titre not finish the supply problem?
5. A sensor fires on product and on an upstream intermediate. What will selection accumulate?
—
—
—