Protein engineering

Protein engineering

25/02 Introduction

It’s the modification/changing of structure/function of proteins, useful to study proteins or to produce proteins, or

for example to understand the folding of proteins, improve an enzyme activity/stability, to produce a drug

(increase the amount, reduce the toxicity, improve bioavailability, etcetera). It’s also possible to widening the

range of substrates that can be modified by a specific enzyme, change specific residues or perform changes on a

larger scale.

Protein engineering: techniques used to produce proteins with new or modified sequences and consequently with

new or modified function. It can be done by chemical reactions, genetic engineering, using synthetic peptides etc.

1973: first experiment of DNA recombinant technology has been performed and published. By DNA recombinant

technology any change in the expressed protein can be performed, working at the nucleotide sequence level. The

result can be helpful, but it can also be unexpected.

· Techniques to analyze 3D structure of a protein at high resolution

NMR spectroscopy: limitations on protein size, <20-30 kDa. It needs proteins dissolved in water: the

§ analysis are performed in solution, where proteins show their natural dynamics. This is not useful for

membrane proteins. It’s the only technique that allows the study of disordered structures, partly folded or

denatured states at an atomic level. For NMR proteins that can’t be dissolved in waters can’t be studied and

it’s limited to small molecules, since information are close to the real protein structure but larger proteins

can’t be resolved.

X-ray crystallography: go deep in the 3D structure and see where atoms are exactly organized in space. But

§ the crystal is needed and is not easy to obtain, especially for membrane proteins or with hydrophobic

regions or post-translationally modified. In X-ray crystallography there are challenges with the

engineering of membrane proteins and IDPs (unable to crystallize). Crystallization may take years (empiric

approach for the crystal generation), but it can be applied to any size molecule and it gives information at

atomic resolution.

NMR and X-ray techniques are complementary. NMR and X-ray need a lot of mg of proteins, which is not

easy to obtain with recombinant DNA techniques.

Cryo-electron microscopy: also for complexes of proteins (IV structure, multi-subunit proteins, large

§ molecules), near to the atomic level but not as precise as X-ray crystallography. Cryo-electron microscopy

allows the freezing of the molecule in native conditions and larger molecules can be used. Resolution is

near the atomic resolution and samples are fast to prepare. Also possible to obtain the structure of a

protein bound to a substrate with this technique.

Computational prediction of protein structure: try to model the protein structure knowing the structure of

§ related proteins. Also, build the structure by patches, different domains from different origins. Must have

information related to homologous proteins or at least homologous domains.

On the 3D structure, is possible to decide which domain has to be modified. Protein engineering can be done even

without the knowledge of the 3D structure, but start with random mutagenesis and see what happens.

· Recombinant DNA techniques to produce a protein, a peptide, and to modify them

- Cloning strategies: knowledge of the sequence, possibility to amplify by PCR.

- Expression systems: availability of different hosts (organisms) for protein production (bacterial and

eukaryotic systems) different expression vectors. Expression system must consider also the folding and

yield.

- Purification strategies: different chromatographic approaches. Purification for a maximum yield and

remove contaminants(use: drugs, therapeutics). Used especially affinity chromatography. Add a tag (do we

want to remove it?) for purification with chromatography.

- Mutagenesis (see advanced molecular biology): methods for side directed mutagenesis, methods for

random mutagenesis. One round of mutagenesis is not enough.

- Screening: methods to identify and select the “best” mutant. Screening is essential. Screening and selection

are different. It is essential to identify the best mutant. When we are doing site- directed mutagenesis, we

can test the changes in the activity of the modified protein. But if we do a random mutagenesis we change

randomly the primary sequence of a protein. Maybe many of the mutations are deleterious for the protein

and in this situation we can’t produce and analyze every single mutant, it is too long. We have to set up a

screening method allowing to remove quickly all bad mutants (unable to fold properly, lose the activity,

etcetera) and we have to create libraries of mutants by random mutagenesis. 1

While modifying a protein, is important to consider the role of the protein itself.

Many experiments are done with proteins to improve their affinity or change it towards a particular substrate. Ex.

filaments phages (display, a sort of ELISA) to produce library of mutants of proteins and select the best phage

exposing the protein of interest. In this way, is possible to select millions of phages, each corresponding to a

different mutant, in a quick manner.

To screen a lot of proteins in a quick time, also for toxic proteins, there are in vitro systems, both for screening and

production.

To study protein stability → microcalorimetry: changing T and follow the denaturation by looking at CD

spectroscopy and monitoring the intrinsic fluorescence signal (information about the folding).

· History

Protein engineering techniques were initially developed to answer questions of enzymologists, in particular to test

theories on the mechanism of action of enzymes. Mutagenesis of “famous residues” allowed to confirm their

function in a specific protein, however many times paradigms were dismantled and forced to the revision of

previously accepted mechanisms.

Mutagenesis allowed the description of phenomenon such as allostery, folding, stability and specificity.

Fluorescence and NMR analyses allowed to monitor the folding process, leading to the identification of nucleation

cores and folding stories.

Determination of the energetic contribution of electrostatic and apolar interactions during folding were essential

to understand protein folding and stability of interactions.

Protein engineering showed promising results for the creation of enzymes with increased activity in organic

solvents, increased thermal stability and different pH optima. However it was not always easy to obtain the

desired result. The exchange of the Asp189 was insufficient to convert trypsin in chymotrypsin. Up to 15 changes

were needed to adapt the substrate specificity of trypsin to that of chymotrypsin. Although local interactions in an

enzyme are dominant factors in shaping substrate specificity, they are not the rule.

Sometimes even conservative modifications can have consequences on the final protein.

27/02

· Applications of protein engineering

Biotech industry:

Production of protein drugs (insulin, human growth hormone, erythropoietin, antibody for therapy and

§ diagnosis) and systems for the delivery of drugs.

Bioremediation: use of enzymes and modified bacterial or algae systems to degrade pollutants or as

§ biosensors.

Enzymes used in detergents, in industry, agriculture and food preparation (amylase, protease, cellulase,

§ lipase).

Basic research: to clarify aspects of protein folding, protein stability, mechanisms of enzyme catalysis and to study

molecular recognition and protein-protein interactions.

· Modifications that can be introduced in proteins:

- Replacement of an aa with another of the natural 20.

- Insertion of deletion of a stretch of aa.

- Introduction of chemical modifications for drug-delivery or to mimic post-translational modifications

(glycosylation, derivatization with lipids).

- Introduction of binding sites for metal.

- Introduction of non-natural aa.

· Methodologies used to introduce mutations

• Mutagenesis at the level of the gene.

• Chemical modification of native proteins.

• Solid phase synthesis of peptides and semi-synthesis of proteins (strategies of condensation and of

"ligation").

2

- Production of protein drugs -

The design of protein therapeutics integrates:

• Traditional goals of protein design: change stability, solubility, binding affinity, binding specificity, folding

kinetics.

• Biological and clinical parameters: serum half-life, injection-site absorption, immunogenicity, toxicity,

protease susceptibility, in vivo regulation, therapeutic mechanism.

Through protein production, storage and administration, novel and optimized protein therapeutics are obtained.

· Parameters

Stability: in oxidative condition, like replace free Cys residues (Cys → Ser). Increase the overall stability of a protein

and reduce susceptibility to proteases (mutations are introduced at the level of flexible loops that are typically

sites of proteolysis).

Solubility: important both at the level of expression but also at the level of administration. Aggregation results in

decreased activity, decreased availability and increased immunogenicity.

Pharmacokinetics: fusion proteins, PEG, glycosylation, altered state of oligomerization.

Affinity and specificity: antibodies.

Immunogenicity: for non-human proteins (Ab). May depend on solubility and stability properties of the protein and

route of administration. Problems when Abs neutralize not only the administered protein but also the endogenous

protein.

· Protein-based drugs

Product quality test: a lot of things to consider for the protein production (same microorganism, same T, same

purification, etcetera).

There are also much higher number of critical points in the production of a protein.

At the start ABs were rodent MABs (mice, rats), but immunogenic → chimeric ABs (variable regions, both light and

heavy, from rodents whereas the other region was of human origins, but

there was still immunogenicity) → humanized or reshaped ABs → all human ABs. There are different AB formats.

Improved insulin for diabetics:

- Fast-acting monomeric insulin (in therapy): insulin interacts with the receptor as a monomer. In order to

accelerate its arrival in the blood, immediately after meal consumption, it is necessary to inhibit the

aggregation of the protein into hexamers. By mutagenesis, researchers have altered the contact surfaces

between the molecules of insulin in the hexamer: juxtaposed negative charges or destabilized β-sheet

intermolecular interactions.

- Long-lasting insulin: to maintain basal levels of the protein. The Lys29B is acylated with a fatty acid. With

this modification, insulin associates with albumin present in subcutaneous fluids and it is therefore slowly

released into blood.

· Experimental approaching of protein engineering

• Rational design.

• Non-rational design.

Generation of random libraries.

o Molecular evolution.

o

• De novo protein design. 3

28/02 Rational design in mutagenesis

Introduction of a mutation at one or more specific sites of a protein by site directed mutagenesis.

The site of mutation can be determined on the basis of:

• The tertiary structure of the protein, using computer graphics. Obtaining co-crystallization of the substrate.

• Comparison with sequences of other proteins if the protein of interest belongs to a large family of proteins.

With homology modelling, first identify the structure and align the sequence with the family, important

because it can tell which are the most important residues (most conserved). Building the model for the

query sequence: core modeling, side chain modeling, loop modeling. Finally there’s the model evaluation

and the refinement.

Key points:

- Structure of a protein is defined by the amino acid sequence.

- Homologous proteins present an highly conserved sequence, but distantly related sequences may fold into

similar structures.

- Three-dimensional structure of proteins from the same family is more conserved than their primary

sequence.

Work flow:

1. Protein of interest without crystal structure.

2. Database searching with the aa sequence and multiple sequence alignment, to identify homologous

proteins in PDB.

3. Yes: possible to identify the homology model and build the 3D protein model.

4. No: search for the secondary structure prediction, so the software tell us if a region can fold and in which

or which region can’t and recognize the needed fold…. It’s possible to do a sequence-structure alignment

and identify a structure fitting with the one of interest, so it’s possible to do the homology model. Usually

the alignment is not possible, so there’s the prediction of a 3D structure of the protein that can be similar

or completely wrong.

Limits of rational design:

Availability of the 3D structure of the protein and knowledge of its structure- function relationships (often

§ mutations can be deleterious for the structure or function of the protein).

Mutations distant from the active site of an enzyme can significantly influence the properties and activity

§ of the protein. These sites are difficult to predict.

In order to consider the rational design of a target enzyme, several pieces of information are needed:

• A cloned gene coding for the enzyme.

• The sequence of the gene.

• Information on the chemistry of the active site (which amino acids in the sequence are involved in activity).

• Either a crystal/NMR structure of the enzyme, or the structure of another protein displaying a high degree

of structural homology.

• Co-crystallization of enzyme and substrate would be very helpful.

These informations are needed to have a clear idea of which amino acids should mutate to which likely effect.

- Selection of the right aa -

· Classification

Small: Gly, Ala.

§ Nucleophilic: Ser, Thr, Cys.

§ Hydrophobic: Val, Leu, Ile, Met, Pro.

§ Aromatic: Phe, Tyr.

§ Acidic: Asp, Glu.

§ Amide: Asn, Gln.

§ Basic: His, Lys, Arg.

§

4

· The peptide bond

The allowed dihedral angles Φ and Ψ are determined by the steric hindrance

of the lateral chains of aa residues.

In α-helical and β-sheet regions, aa residues have the same Φ and Ψ values.

Different secondary structure:

- α-helix.

- β-sheet.

- β-turns.

Amino acid properties:

- Branched at C (Val, Ile): destabilizes α-helix, while ok in β-sheet side chain projects out of plan of main

β

chain.

- No branch at C and R group can from H bonds (Ser, Asp, Asn): destabilizes α-helix because the moiety

β

forms H bonds that compete with backbone atoms forming H bonds in α-helix. Glu is quite present in

helices since R group is far away from the backbone.

- No branch at C and R group can’t form H bonds: ok in α-helix, side chain projects out of the main chain.

β

Except for Gly (too flexible) and Pro (too rigid).

· Interactions

The stability and folding of proteins, the interactions protein-ligand/protein and in general the functions of a

protein are mediated by several interactions between amino acid residues.

These interactions/stabilization are characterized by a free energy (ΔG°) due to the formation of H- bonds, van der

Waals interactions, ionic-bonds (typically ΔG° values are of about 5-20 kcal/mole).

01/03

- Additivity/non-additivity of mutational effects -

After the selection of the aa to be changed, the choice of the aa to be introduced, the production and purification of

the mutated protein, the results must be analyzed, the effect of the mutagenesis.

How to analyze mutants? How to have an idea of what happened in the protein? Set up a particular assay assessing

the differences. It depends on the modifications introduced and on the final aim. Sometimes the mutant protein

can’t fold in the system selected for the production (impossible to produce it), other times the modification is

selected in order to effectively change the activity of the protein. Often the modifications introduced are small

modifications and simple to verify.

The protein function is linked to the interaction between the protein residues. The stability and folding of proteins,

the interactions protein-ligand and protein-protein and in general the function of a protein are mediated by

several interactions between aa. These interactions are characterized by a free energy (ΔG°) due to the formation

of H-bonds, van der Waals interactions, ionic bonds. Typically when the bond is formed some energy is released

and it’s about ΔG° = 5-20 kcal/mole.

What changes introducing a mutation? Probably there is one small change in the overall ΔG, around 0.5-5

kcal/mole. Measure this ΔΔG° comparing WT ΔG° and mutant ΔG° using calorimetry (or micro-calorimetry),

changing T following protein structure, or denaturation in a denaturing agent, set up a particular T and change the

amount of denaturing agent. This can be done for each mutation: ΔΔG° is specific for each mutation.

It’s possible to evaluate if two residues influence each other in the overall structure. Suppose to mutagenize two

residues at the same time: determine the ΔΔG° (A, B) of the double mutant, the compare the two ΔΔG° of the with

those of the single mutants (A and B). If ΔΔG°(A, B) ≈ ΔΔG° (A) + ΔΔG° (B) the two mutations are additive.

When amino acid substitutions are made at well-separated locations in a single protein, their effects are generally

additive. This is due to the fact that alterations in the structure of a protein determined by the majority of the

mutations are very localized. However this is not always occurring, because there can be even long range distance

effects due to direct or indirect interactions. The 2 aa can be very far and even if they are additive they can have

long range distance effects, so there’s non-additive mutations. 5

Suppose that the ΔΔG°(A, B) is substantially different from the simple sum ΔΔG°(A) + &De