Phage display
• Make a combinatorial library of genes of interest• Put genes into a vector so that each gene product is expressed on the surface of a bacteriophage
– Function encoded by each gene is on surface of phage– Gene is inside phage
• Take collection of phage and select those with desired properties (e.g., binding to something)• Similar methods: yeast display, bacterial display, ribosome display
See animation of phage display here: http://www.dyax.com/discovery/phagedisplay.html
Figure A-15
In vitro selection to produce human monoclonal antibodies or increase affinity of existing monoclonal antibody
Generate library of heavy and light chain variable regions using
spleen DNA. Or introduce random
mutations into variable regions
genes of a specific antibody.
Clone into a phage so that each phage
expresses one VH-VL surface
fusion protein.
Multiply phage display library in
bacteria, bind phage to surface
coated with antigen. Wash away unbound
phage.
Repeat procedure (multiply
recovered phage, bind to antigen,
wash away unbound phage)
for several cycles. Recover specific
high-affinity antigen binding VH-VL regions.
Phage display to select for sequence-specific DNA binding proteins
• Zinc finger proteins are modular.
• Each finger contains two anti-parallel -strands, an -helix, and a Zn atom.
• The -helix from each finger inserts into the major groove of DNA.
A General Strategy for Selecting High Affinity Zinc Finger Proteins for Diverse DNA Target Sites
Greisman et al., 1997, Science 275: 657
First use 3-finger/DNA crystal structure to determine important protein/DNA contacts
A General Strategy for Selecting High Affinity Zinc Finger Proteins for Diverse DNA Target Sites
Greisman et al., 1997, Science 275: 657
Originally bound to this sequence
RNA aptamers -- antibody-like properties
Aptamers have been made against small molecules, peptides, proteins, organelles, viruses, cells
Let the immune system make enzymes for you
Catalyst must bind more tightly to transition state than to products or reactants.
Catalytic Antibodies
• Raise antibodies against a transition state analog
• Screen hybridomas for antibodies that catalyze desired reaction
Wedemayer et al. (1997) Science 276, 1665-1669.
Catalytic antibody that hydrolyzes cocaineZhu et al., 2006, Structure 14: 205-216
Compound that wasinjected to raise
antibodies
Transition state analog that was
crystallized with Fab
(Nonpsychoactive products)
/ barrel enzymes evolved from a common ancestor
But this takes a long time…
10% of enzymes are barrels
In vitro evolution of enzymes
• Enzymes evolved for functions inside a living organism, not for biotechnology
– Might need long-term stability– Might need activity in non-aqueous solutions
• Produce new enzymes using recombinant DNA technology, but don’t know how by rational design
• Use directed evolution
Some considerations…
• Don’t start with random sequences because there are too many (20N)
• Instead start with lightly mutagenized gene (e.g., error-prone PCR) or high level of random mutations to small part of gene
Most mutations are destabilizing, so simply increasing protein stability can increase
mutational robustness
∆Gf
Marginally stable parent protein
∆Gf
stable unstable
Stabilized parent protein
These previously unacceptable mutations are now acceptable.
(critical threshold stability)
“Protein stability promotes evolvability.” JD Bloom, ST Labthavikul, CR Otey, and FH Arnold. Proc Natl Acad Sci, 103:5869-5874 (2006).
Evaluating stability
Two state unfolding transitionN <--> D
Monitor property of folded protein as function of increasing temperature
Transition midpoint (Tm)
Shows that class I MHC molecules require bound peptide for thermal stability
Sequence Space Structure Space
multiple sequences 1 structure
Protein Library Design
Mayo lab, Caltech
Sequence Space Structure Space
1 structure
Protein Fold Prediction versus Protein Design
1 sequence
Mayo lab, Caltech
Computational Protein Design:Rationale by the Numbers
Residues Sequences Mass 18 1023 Baseball 37 1048 Earth 42 1054 Sun 59 1077 Universe
Combinatorial Explosion
1 proteinp residues20 amino acid types20
p sequences
Mayo lab, Caltech
Optimization of Rotamers by Iterative Techniques (ORBIT)
Apply to protein fold stabilization, enzyme design
ComputationalProtein Design
RotamerLibraries
Methods
Applications
ProteinBackbones
CombinatorialOptimizationAlgorithms
Atom-BasedForcefields
NegativeDesign
Mayo lab, Caltech
De Novo Protein Design: Fully automated sequence selection
Dahiyat & Mayo, 1997, Science 278: 82-87
Comparison of original and designedprotein structuresTarget fold: Zif268 (a zinc finger)
Designed protein
Zif268 Zn finger
Stability Based Design: Protein G
• Bacterial protein involved in host immune system evasion
• 56 amino acid domain
• Objective was to stabilize protein while preserving structure and function
• Design focused on 26 core and boundary positions
• Combinatorial complexity, 106 amino acid sequences
Mayo lab, Caltech
O
N
H
O2NO
N
H
O2N
O-
C
O2N
N
O O O O HO O
catalytic antibody: kcat/kuncat = 106
Thorn et al., Nature 1995Debler et al., PNAS 2005
5-nitrobenzsoxazole
O O
ON
N
H
O
O
General Base Asp/Glu
*
*
Phe/Trp
HO
H-bond donorSer/Thr/Tyr
H N
N
H-bond donorHie/Hid/Gln/Asn
active site templateab initio t.s. model
Hu et al., JACS 2004
Kemp Elimination: a model system for enzyme design
Mayo lab, Caltech