center for biological sequence analysis technical university of denmark - dtu department of systems...

CENTER FOR BIOLOGICAL SEQUENCE ANALYSIS

Technical University of Denmark - DTUTechnical University of Denmark - DTUDepartment of systems biologyDepartment of systems biology

B cell epitopes and predictions


ECCB/ISMB-2009 - Immunological Bioinformatics Tutorial

Outline

•What is a B-cell epitope?

•How can you predict B-cell epitopes?



What is a B-cell epitope?

Antibody Fabfragment

• B-cell epitopes:

• Accessible structural feature of a pathogen molecule.

• Antibodies are developed to bind the epitope specifically using the complementary determining regions (CDRs).



The binding interactions

• Salt bridges

• Hydrogen bonds

• Hydrophobic interactions

• Van der Waals forces

Binding strength



The binding interactionHowever spatial composition matters!!!



The binding interaction



B-cell epitope classification

Linear epitopesOne segment of the amino acid

chain

Discontinuous epitope (with linear determinant)

Discontinuous epitopeSeveral small segments brought into proximity by the protein fold

B-cell epitope: structural feature of a molecule or pathogen, accessible and recognizable by B-cell

receptors and antibodies



B-cell epitope annotation

•Linear epitopes:

• Chop sequence into small pieces and measure binding to antibody

•Discontinuous epitopes:

• Measure binding of whole protein to antibody

•The best annotation method : X-ray crystal structure of the antibody-epitope complex




•Linear epitopes:

• Chop sequence into small pieces and measure binding to antibody

•Discontinuous epitopes:

• Measure binding of whole protein to antibody

•The best annotation method : X-ray crystal structure of the antibody-epitope complex

10%

90%




•No epitope is purely linear

• Epitopes contains linear determinants of 5 or more residues

Longest linear stretch in epitope



B-cell epitope data bases

•Databases:

• IEDB, Los Alamos HIV database, Protein Data Bank,

AntiJen, BciPep

•Large amount of data available for linear epitopes

•Few data available for discontinuous



B cell epitope prediction



The immune system – prediction tools

•Cytotoxic T cell epitope: (AROC ~ 0.9)• Will a given peptide bind to a given MHC class I molecule

•Helper T cell Epitope (AROC ~ 0.85)• Will a part of a peptide bind to a given MHC II molecule

•B cell epitope (AROC ~ 0.74)• Will a given part of a protein bind to one of

the billions of different B Cell receptors



B-cell – prediction tools•Sequence based prediction tools– Predominantly predicts linear

epitopes•Structure based epitopes– Predicts Conformational epitopes



Watch out for

overfitting



Sequence-based methods for prediction of linear

epitopes

TSQDLSVFPLASCCKDNIASTSVTLGCLVTGYLPMSTTVTWDTGSLNKNVTTFPTTFHETYGLHSIVSQVTASGKWAKQRFTCSVAHAESTAINKTFSACALNFIPPTVKLFHSSCNPVGDTHTTIQLLCLISGYVPGDMEVIWLVDGQKATNIFPYTAPGTKEGNVTSTHSELNITQGEWVSQKTYTCQVTYQGFTFKDEARKCSESDPRGVTSYLSPPSPL

Input

TSQDLSVFPLASCCKDNIASTSVTLGCLVTGYLPMSTTVTWDTGSLNKNVTTFPTTFHETYGLHSIVSQVTASGKWAKQRFTCSVAHAESTAINKTFSACALNFIPPTVKLFHSSCNPVGDTHTTIQLLCLISGYVPGDMEVIWLVDGQKATNIFPYTAPGTKEGNVTSTHSELNITQGEWVSQKTYTCQVTYQGFTFKDEARKCSESDPRGVTSYLSPPSPL

Output



Sequence-based methods for prediction of linear

epitopes• Protein hydrophobicity – hydrophilicity algorithms

• Parker, Fauchere, Janin, Kyte and Doolittle, Manavalan

• Sweet and Eisenberg, Goldman, Engelman and Steitz (GES), von Heijne

• Protein flexibility prediction algorithm

• Karplus and Schulz

• Protein secondary structure prediction algorithms

• PsiPred (D. Jones)



Propensity scales: The principle

• The Parker hydrophilicity scale

• Derived from experimental data

D 2.46E 1.86N 1.64S 1.50Q 1.37G 1.28K 1.26T 1.15R 0.87P 0.30H 0.30C 0.11A 0.03Y -0.78V -1.27M -1.41I -2.45 F -2.78L -2.87W -3.00

Hydrophilicity



• ….LISTFVDEKRPGSDIVEDLILKDENKTTVI….

(-2.78 + -1.27 + 2.46 +1.86 + 1.26 + 0.87 + 0.3)/7 = 0.39

Prediction scores:

0.39 0.1 0.6 0.9 1.0 1.2 2.6 1.0 0.9 0.5 -0.5

Epitope

Propensity scales: The principle



Turn prediction and B-cell epitopes

• Pellequer found that 50% of the epitopes in a data set of 11 proteins were located in turns

Turn propensity scales for each position in the turn were used for epitope prediction.

Pellequer et al.,Immunology letters, 1993

1

2

3

4



Blythe and Flower 2005

•Extensive evaluation of propensity scales for epitope prediction

• Conclusion:

– Basically all the classical scales perform close to random!

– Other methods must be used for epitope prediction



BepiPred

• Parker hydrophilicity scale

• PSSM

• PSSM based on linear epitopes extracted from the AntiJen database

• Combination of the Parker prediction scores and PSSM leads to prediction score

• Tested on the Pellequer dataset and epitopes in the HIV Los Alamos database



PSSM

• ….LISTFVDEKRPGSDIVEDLILKDENKTTVI….

2.46+1.86+1.26+0.87+0.3 = 6.75 Prediction value

A R N D C Q E G H I L K M F P S T W Y V S I1 0.10 0.06 0.01 0.02 0.01 0.02 2.46 0.30 0.01 0.07 0.11 0.06 0.04 0.08 0.01 0.11 0.03 0.01 0.05 0.08 3.96 0.372 0.07 0.00 0.00 0.01 0.01 0.00 0.01 0.01 0.00 0.08 0.59 1.86 0.07 0.01 0.00 0.01 0.06 0.00 0.01 0.08 2.16 2.163 0.08 1.26 0.05 0.10 0.02 0.02 0.01 0.12 0.02 0.03 0.12 0.01 0.03 0.05 0.06 0.06 0.04 0.04 0.04 0.07 4.06 0.264 0.07 0.04 0.02 0.11 0.01 0.04 0.08 0.15 0.01 0.10 0.04 0.03 0.01 0.02 0.87 0.07 0.04 0.02 0.00 0.05 3.87 0.455 0.04 0.04 0.04 0.04 0.01 0.04 0.05 0.30 0.04 0.02 0.08 0.04 0.01 0.06 0.10 0.02 0.06 0.02 0.05 0.09 4.04 0.28



ROC evaluation

Evaluation on HIV Los Alamos data

set



BepiPred performance•Pellequer data set:

– Levitt AROC = 0.66

– Parker AROC = 0.65

– BepiPred AROC = 0.68

•HIV Los Alamos data set

– Levitt AROC = 0.57

– Parker AROC = 0.59

– BepiPred AROC = 0.60



BepiPred

•BepiPred conclusion:

• On both of the evaluation data sets, Bepipred was shown to perform better

• Still the AROC value is low compared to T-cell epitope prediction tools!

• Bepipred is available as a webserver:

• www.cbs.dtu.dk/services/BepiPred



Prediction of linear epitopes

Con•only ~10% of epitopes can be classified as “linear” •weakly immunogenic in most cases•most epitope peptides do not provide antigen-neutralizing immunity•in many cases represent hypervariable regions

•Pro• easily predicted computationally

• easily identified experimentally

• immunodominant epitopes in many cases

• do not need 3D structural information

• easy to produce and check binding activity

experimentally



Sequence based prediction methods

• Linear methods for prediction of B cell epitopes have low performances

• The problem is analogous to the problems of representing the surface of the earth on a two-dimensional map

• Reduction of the dimensions leads to distortions of scales, directions, distances

• The world of B-cell epitopes is 3 dimensional and therefore more sophisticated methods must be developed



So what is more sophisticated?

• Use of the three dimensional structure of the pathogen protein

• Analyze the structure to find surface exposed regions

• Additional use of information about conformational changes, glycosylation and trans-membrane helices



Sources of three-dimensional structures

•Experimental determination • X-ray crystallography • NMR spectroscopy

• Both methods are time consuming and not easily done in a larger scale

•Structure prediction• Homology modeling• Fold recognition

• Less time consuming, but there is a possibility of incorrect predictions, specially in loop regions



Protein structure prediction methods

• Homology/comparative modeling • >25% sequence identity (seq 2 seq alignment)

• Fold-recognition• <25% sequence identity (Psi-blast search/ PSSM

2 seq alignment)

• Ab initio structure prediction• 0% sequence identity



What does antibodies recognize in a protein?



The binding interaction



The binding interactionHowever spatial composition matters!!!



What does antibodies recognize in a protein?

probe

Antibody Fabfragment

Protrusion index

Surface Exposed

Hydrophobic region



Structure base

prediction methods



DiscoTope

•Prediction of residues in discontinuous B cell epitopes using protein 3D structures

Pernille Haste Andersen, Morten Nielsen and Ole

Lund, Protein Science 2006

17.04.2008Præsentationens navn41 DTU Systembiologi, Danmarks Tekniske Universitet

Predicting B-cell epitopes


The DiscoTope method


•Some amino acids are preferred and disliked in the epitope


Epitope amino acid preferences

Kringelum et al, 2011, manuscript in preparation


•Some amino acids are preferred and disliked in the epitope•Epitopes reside on the surface of the protein


Kringelum et al, 2011, manuscript in preparation


•Predictions score for each residue are calculated by summing the epitope likelihood (propensity) of surrounding residues and subtracting the neighbor count


Andersen et al., 2006




Performance: Aroc = 0.700


On a dataset of 75 antigen-antibody complexes divided in 25 proteins


•Uneven spatial distribution of amino acid in epitopes

However position matters

Kringelum et al, 2012, manuscript in review

17/04/2008Presentation name49 DTU Sytems Biology, Technical University of Denmark

Propensity score function

Amino acid log-odds score

Weight factor


PS(THR256)



1)Identify Naighbor residues within 21.6Å (Cα-Cα)

PS(THR256)


Radius = 21.6Å


PS(THR256)



Radius = 21.6Å


PS(THR256)



2)Calculate summed propensity score: ls(THR) = -0.23

β256= 0.8*(1-0.0/21.6)+0.2 = 1.0

ls(THR)*β256 = -0.23*1.0 = -0.23

d256= 0.0Å


PS(THR256)



2)Calculate summed propensity score:

d255= 3.8Å

ls(THR) = -0.23

β256= 0.8*(1-0.0/21.6)+0.2 = 1.0

ls(THR)*β256 = -0.23*1.0 = -0.23

ls(ASP) = 2.5

β255=0.8*(1-3.8/21.6)+0.2 = 0.86

ls(THR)*β255 = 2.5*0.86 = 2.15


PS(THR256)




β256= 0.8*(1-0.0/21.6)+0.2 = 1.0

ls(THR)*β256 = -0.23*1.0 = -0.23

ls(ASP) = 2.5

β255=0.8*(1-3.8/21.6)+0.2 = 0.86

ls(ASP)*β255 = 2.5*0.86 = 2.15

ls(THR) = -0.23

β254=0.8*(1-6.1/21.6)+0.2 = 0.77

ls(THR)*β254 = -0.23*0.77 = -0.18 d254= 6.1Å


PS(THR256)




β256= 0.8*(1-0.0/21.6)+0.2 = 1.0

ls(THR)*β256 = -0.23*1.0 = -0.23

ls(ASP) = 2.5

β255=0.8*(1-3.8/21.6)+0.2 = 0.86

ls(ASP)*β255 = 2.5*0.86 = 2.15

ls(THR) = -0.23

β254=0.8*(1-6.1/21.6)+0.2 = 0.77

ls(THR)*β254 = -0.23*0.77 = -0.18

…........

………….

………….

ls(PRO) = 1.2

β254=0.8*(1-20.6/21.6)+0.2 = 0.24

ls(PRO)*β254 = 1.2*0.24 = 0.29

d247= 20.6Å


PS(THR256)




β256= 0.8*(1-0.0/21.6)+0.2 = 1.0

ls(THR)*β256 = -0.23*1.0 = -0.23

ls(ASP) = 2.5

β255=0.8*(1-3.8/21.6)+0.2 = 0.86

ls(ASP)*β255 = 2.5*0.86 = 2.15

ls(THR) = -0.23

β254=0.8*(1-6.1/21.6)+0.2 = 0.77

ls(THR)*β254 = -0.23*0.77 = -0.18

…........

………….

………….

ls(PRO) = 1.2

β254=0.8*(1-20.6/21.6)+0.2 = 0.24

ls(PRO)*β254 = 1.2*0.24 = 0.29

Summation of scores

PS(THR256) = 0.39


HS(THR256)

Half-sphere exposure

1)Create the upper half-sphere, which is the half-sphere where the residue side-chain is pointing


HS(THR256)

Half-sphere exposure

1)Create the upper half-sphere, which is the half-sphere where the residue side-chain is pointing2)Count neighbor residues within the half-sphere (nr of Cα-atoms)

3)As high counts means highly buried the counts are multiplied by -1

= 5

HS(THR256) = -5


1)Calculate Propensity score2)Calculate half-sphere exposure3)The final score is a weighted sum of the Propensity score and half-sphere exposure PS(THR256) = 0.39 HS(THR256) = -5

α = 0.115

DS = (1-α)*PS + α*HS

DS(THR256) = 0.885*0.39 + 0.115*(-5)

DS(THR256) = -0.23

DS(THR256)

The DiscoTope 2.0 Score


Performance and limitations

Average AUC = 0.741



Glycosylated proteins



Gp120 Hemagglutinine

•Glycosylation effects predictions



Small fragments (<120 residues) of larger biological units


• Inclusion of biological units enhance performance


Potassium Channel


• External Benchmark Dataset• 52 antigen:antibody structures• 33 homology groups

• Performance: 0.731 AUC



Conclusions• DiscoTope V2.0 outperforms similar methods

• High performance on 15/25• Medium performance on 7/25• Fail on 3/25

• Inclusion of surface measures does only slightly enhance predictions

• Use the entire biological unit, when possible• Small fragments (< 120 residues) have lower performance

• Glycosylation might course the prediction to fail• Check for clash between predicted epitopes and glycosylation sites



Rational vaccine design

>PATHOGEN PROTEINKVFGRCELAAAMKRHGLDNYRGYSLGN

WVCAAKFESNF

Rational Vaccine Design



Rational B-cell epitope design

• Protein target choice

• Structural analysis of antigenKnown structure or homology modelPrecise domain structurePhysical annotation (flexibility, electrostatics, hydrophobicity)Functional annotation (sequence variations, active sites, binding sites, glycosylation sites, etc.)Known 3D structure

Model




• Structural annotation

• Epitope prediction and ranking


Surface accessibilityProtrusion indexConserved sequenceGlycosylation status





• Structural annotation

• Epitope prediction and ranking

• Optimal Epitope presentation

Fold minimization, orDesign of structural mimics Choice of carrier (conjugates, DNA plasmids, virus like particles)Multiple chain protein engineering



Conclusions

• Rational vaccines can be designed to induce strong and epitope-specific B-cell responses

• Selection of protective B-cell epitopes involves structural, functional and immunogenic analysis of the pathogenic proteins

• When you can: Use protein structure for prediction

• Structural modeling tools are helpful in prediction of epitopes, design of epitope mimics and optimal epitope presentation

center for biological sequence analysis technical university of denmark - dtu department of systems...

Documents