CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

B cell epitopes and predictions

Pernille Haste Andersen,

Ph.d. student

Immunological Bioinformatics

CBS, DTU

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

B cells and antibodies

Antibodies are produced by B lymphocytes (B cells)

Antibodies circulate in the blood

They are referred to as “the first line of defense” against infection

Antibodies play a central role in immunity by attaching to pathogens and recruiting effector systems that kill the invader

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

What is a B cell epitope?

Antibody Fabfragment

B cell epitope

B cell epitopes

Accessible and recognizable structural feature of a pathogen molecule (antigen)

Antibodies are developed to bind the epitope with high affinity by using the complementarity determining regions (CDRs)

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

Motivations for prediction of B cell epitopes

Prediction of B cell epitopes can potentially guide experimental epitope mapping

Predictions of antigenicity in proteins can be used for selecting subunits in rational vaccine design

Predictions of B cell epitopes may also be valuable for interpretation of results from experiments based on antibody affinity binding such as ELISA, RIA

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

Computational Rational Vaccine Design

>PATHOGEN PROTEINKVFGRCELAAAMKRHGLDNYRGYSLGNWVCAAKFESNF

Rational Vaccine Design

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

B cell epitopes, linear or discontinuous?

Classified into linear (~10%) and discontinuous epitopes (~90%)

Databases: AntiJen, IEDB, BciPep, Los Alamos HIV database, Protein Data Bank

Large amount of data available for linear epitopes

Few data available for discontinuous epitopes

In general, B cell epitope prediction methods have relatively low performances

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

Discontinuous B cell epitopes

•

SLDEKNSVSVDLPGEMKVLVSKEKNKDGKYDLIATVDKLELKGTSDKNNGSGVLEGVKADKCKVKLTISDDLGQTTLEVFKEDGKTLVSKKVTSKDKSSTEEKFNEKGEVSEKIITRADGTRLEYTGIKSDGSGKAKEVLKG

• ..\Discotope\1OSP_epitope\1OSP_epitope.psw

An example: The epitope of the outer surface protein A from Borrelia Burgdorferi (1OSP)

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

The binding interactions

Salt bridges Hydrogen bonds Hydrophobic interactions Van der Waals forces

Binding strength

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

B-cell epitopes are dynamic

Many of the charged groups and hydrogen bonding partners are present on highly flexible amino acid side chains.

Most crystal structures of epitopes and antibodies in free and complexed forms have shown conformational rearrangements upon binding.

“Induced fit” model of interactions.

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

B-cell epitope classification

Linear epitopes

One segment of the amino acid chain

Discontinuous epitope (with linear determinant)

Discontinuous epitope

Several small segments brought into proximity by the protein fold

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

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

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

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

B-cell epitope data bases

• Databases: AntiJen, IEDB, BciPep,

Los Alamos HIV database, Protein

Data Bank

• Large amount of data available for

linear epitopes

• Few data available for discontinuous

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

B cell epitope prediction

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

Sequence-based methods for prediction of linear epitopes

Protein hydrophobicity – hydrophilicity algorithms Parker, Fauchere, Janin, Kyte and Doolittle, ManavalanSweet and Eisenberg, Goldman, Engelman and Steitz (GES), von Heijne

Protein flexibility prediction algorithm Karplus and Schulz

Protein secondary structure prediction algorithms GOR II method (Garnier and Robson), Chou and Fasman, Pellequer

Protein “antigenicity” prediction :Hopp and Woods, Welling

TSQDLSVFPLASCCKDNIASTSVTLGCLVTGYLPMSTTVTWDTGSLNKNVTTFPTTFHETYGLHSIVSQVTASGKWAKQRFTCSVAHAESTAINKTFSACALNFIPPTVKLFHSSCNPVGDTHTTIQLLCLISGYVPGDMEVIWLVDGQKATNIFPYTAPGTKEGNVTSTHSELNITQGEWVSQKTYTCQVTYQGFTFKDEARKCSESDPRGVTSYLSPPSPL

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

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

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

Propensity scales: The principle

….LISTFVDEKRPGSDIVEDLILKDENKTTVI….

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

Prediction scores:

0.38 0.1 0.6 0.9 1.0 1.2 2.6 1.0 0.9 0.5 -0.5

Epitope

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

Evaluation of performance

• A Receiver Operator Curve (ROC) is useful for finding a good threshold and rank methods

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

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

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

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

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

BepiPred: CBS in-house tool

• Parker hydrophilicity scale • Hidden Markov model• Markov model based on linear epitopes

extracted from the AntiJen database• Combination of the Parker prediction scores

and Markov model leads to prediction score• Tested on the Pellequer dataset and

epitopes in the HIV Los Alamos database

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

ROC evaluation

Evaluation on HIV Los Alamos data set

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

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

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

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

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

How can we get information about the three-dimensional structure?

Structural

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

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

Protein structure prediction methods

• Homology/comparative modeling

>25% sequence identity (seq 2 seq alignment)• Fold-recognition/threading

<25% sequence identity (Psi-blast search/ seq 2 str alignment)

• Ab initio structure prediction

0% sequence identity

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

A data set of 3D discontinuous epitopes

A data set of 75 discontinuous epitopes was compiled from structures of antibodies/protein antigen complexes in the PDB

The data set has been used for developing a method for predictions of discontinuous B cell epitopes

Since about 30 of the PDB entries represented Lysozyme, I have used homology grouping (25 groups of non-homologous antigens) and 5 fold cross-validation for training of the method

Performance was measured using ROC curves on a per antigen basis, and by weighted averaging of AUC values

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

Epitope log-odds ratios Frequencies of amino acids in epitopes

compared to frequencies of non-epitopes

Several discrepancies compared to the Parker hydrophilicity scale which is often used for epitope prediction

Both methods are used for predictions using a sequential average of scores

Predictive performance of B cell epitopes:Parker 0.614 AUCEpitope log–odds 0.634 AUC

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

3D information: Contact numbers

• Surface exposure and• structural protrusion can• be measured by residue• contact numbers

The predictive performance:

Parker 0.614 AUCEpitope log–odds 0.634 AUCContact numbers 0.647 AUC

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

DiscoTope : Prediction of Discontinuous epitopes using 3D structures

• A combination of:– Sequentially averaged epitope log-odds values of

residues in spatial proximity– Contact numbers

-0.145

+0.346+1.136

+0.691+0.346+1.136+1.180+1.164

Contact number : K 10

DiscoTope prediction value

Sum of log-odds values

.LIST..FVDEKRPGSDIVED……ALILKDENKTTVI.

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

DiscoTope : Prediction of Discontinuous epiTopes

Improved prediction of residues in discontinuous B

cell epitopes in the data set

The predictive performance on B cell epitopes:

Parker 0.614 AUC

Epitope log–odds 0.634 AUC

Contact numbers 0.647 AUC

DiscoTope 0.711 AUC

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

Evaluation example AMA1

• Apical membrane antigen 1 from Plasmodium falciparum (not used for training/testing)

• Two epitopes were identified using phage-display, point-mutation (black side chains) and sequence variance analysis (side chains of polyvalent residues in yellow)

• Most residues identified as epitopes were successfully predicted by DiscoTope(green backbone)

DiscoTope is available as web server: http://www.cbs.dtu.dk/services/DiscoTope/

..\Discotope\1Z40_epitope\1Z40_movie.mov

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

Future improvements

Add epitope predictions for protein-protein complexes

Visualization of epitopes integrated in web server

Testing a score for sequence variability fx based on entropy of positions in the antigens

Combination with glycosylation site predictions

Combination with predictions of trans-membrane regions

Assembling predicted residues into whole epitopes

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

Presentation of the web server

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

Presentation of the web server output

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

Use the CEP server

• Conformational epitope server

http://202.41.70.74:8080/cgi-bin/cep.pl

• Uses protein structure as input• Finds stretches in sequences which

are surface exposed

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

Use the DiscoTope server

• CBS server for prediction of discontinuous epitopes

• Uses protein structure as input• Combines propensity scale values

of amino acids in discontinuous epitopes with surface exposure

• www.cbs.dtu.dk/services/DiscoTope

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

Rational vaccine design

>PATHOGEN PROTEINKVFGRCELAAAMKRHGLDNYRGYSLGNWVCAAKFESNF

Rational Vaccine Design

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

Rational B-cell epitope design

• Protein target choice

• Structural analysis of antigen Known structure or homology model Precise domain structure Physical annotation (flexibility,

electrostatics, hydrophobicity) Functional annotation (sequence

variations, active sites, binding sites, glycosylation sites, etc.)

Known 3D structure

Model

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

Rational B-cell epitope design

Surface accessibility Protrusion index Conserved sequence Glycosylation status

• Protein target choice• Structural annotation• Epitope prediction and ranking

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

Rational B-cell epitope design• Protein target choice• Structural annotation• Epitope prediction and ranking• Optimal Epitope presentation

Fold minimization, or Design of structural mimics Choice of carrier (conjugates, DNA

plasmids, virus like particles) Multiple chain protein engineering

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

Multi-epitope protein design

Rational optimization of epitope-VLP chimeric proteins:

Design a library of possible linkers (<10 aa)

Perform global energy optimization in VLP (virus-like particle) context

Rank according to estimated energy strain

B-cellepitope

T-cellepitope

CEN

TER

FO

R B

IOLO

GIC

AL S

EQ

UEN

CE A

NA

LY

SIS

TEC

HN

ICA

L U

NIV

ER

SIT

Y O

F D

EN

MA

RK

DTU

Conclusions

• 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