Download Structural Bioinformatics - LCQB

Structural
Bioinformatics
Elodie Laine
Master BIM-BMC Semestre 3, 2016-2017
Laboratoire de Biologie Computationnelle et Quantitative (LCQB)
e-documents: http://www.lcqb.upmc.fr/laine/STRUCT
e-mail: [email protected]
Introduction
Elodie Laine – 20.09.2016 What is a protein?
1-­‐Dimensional text WPLSSSVPSQKTYQGSYGFRLGFLH 2-­‐Dimensional series of strand and helices 3-­‐Dimensional set of points/shape z y x Elodie Laine – 20.09.2016 What is a protein?
tumor-­‐
supressor P53 DNA binding domain Elodie Laine – 20.09.2016 What is a protein?
tumor-­‐
supressor P53 DNA binding domain Elodie Laine – 20.09.2016 What is a protein made of?
one amino-acid aRginine
lysine (K)
aspartate (D)
glutamate (E)
asparagiNe
glutamine (Q)
Cysteine
Methionine
Histidine
Serine
Threonine
Valine
Leucine
Isoleucine
phenylalanine (F)
tYrosine
tryptophan (W)
Glycine
Alanine
Proline
20 amino acids Pep?dic bond Elodie Laine – 20.09.2016 Protein structure
1st level of organisaMon : primary structure …QNCQLRPSGWQCRPTRGDCDLPEFCPGDSSQCPDVSLGDG… ~10’s to ~1000’s of amino acid residues Covalent bonds 1 protein = 1 polypep?dic chain Elodie Laine – 20.09.2016 Protein structure
2nd level of organisaMon : secondary structure β-­‐sheet α-­‐helix Backbone-­‐backbone weak chemical bonds Other elements: 310 helix > turns > loops > random coil Elodie Laine – 20.09.2016 Protein structure
3rd level of organisaMon : terMary structure A protein sequence adopts a parMcular fold in soluMon, which corresponds to a free energy minimum Types of non-­‐covalent interac:ons: •  salt bridges •  hydrogen bonds •  hydrophobic contacts •  pi-­‐pi stacking… Elodie Laine – 20.09.2016 Protein structure
4th level of organisaMon : quaternary structure Arrangements of domains within a protein or of proteins within a macro-­‐
molecular assembly Elodie Laine – 20.09.2016 Protein structure representations
sticks
surface
spheres
cartoon
Elodie Laine – 20.09.2016 Protein structure representations
sticks
protein kinase ~ 300 amino acid residues ~ 5000 atoms (15 000 dof) Each atom is colored according to its element (N: blue, O: red, C: grey). The sMcks represent the covalent bonds (~1.5 Angstroms) between atoms. The atoms are at the extremiMes or intersecMon of the sMcks. Elodie Laine – 20.09.2016 Protein structure representations
spheres
The sphere represents the volume taken by the atom. The radius of the sphere depends on the type of atom. It is called the van der Waals radius. Elodie Laine – 20.09.2016 Protein structure representations
The molecular surface delimits the volume that is not penetrated by water molecules in soluMon. It is obtained by rolling a probe (sphere of 1.4 Angstroms) on the protein. surface
1 Angstrom = 10-­‐10 m = 0.1 nm Elodie Laine – 20.09.2016 Biochemical units
Hydrophobic side chain
Elodie Laine – 20.09.2016 Biochemical units
Polar side chain
Charged side chain
Elodie Laine – 20.09.2016 Biochemical units
Special
Cys reduc:on oxida:on Cys bond Chirality
disulfide bridge The translation machinery of
protein has evolved to utilize
only one of the chiral forms of
amino acids : the L-form
Elodie Laine – 20.09.2016 Structural units
Planar peptide units
Backbone torsion angles
The conformation of the whole main chain of
the polypeptide is completely determined when
the rotation angles φ (N-Cα) and ψ = (CαC’) are defined with high accuracy.
Elodie Laine – 20.09.2016 Structural units
Ramachandran diagram
+180 Most combinations of φ and ψ
for an amino acid are not allowed
because of steric collisions between
the side chain and main chain
-­‐180 -­‐180 Ramachandran et al. (1963) J. Mol. Biol.
Ramachandran et al. (1968) Adv. Protein. Chem.
+180 Elodie Laine – 20.09.2016 Structural units
Ramachandran diagram
Glycine can adopt a much wider
range of conformations than the
other residues and thus plays a
structurally very important role
Ramachandran et al. (1963) J. Mol. Biol.
Ramachandran et al. (1968) Adv. Protein. Chem.
Elodie Laine – 20.09.2016 Sequence-structure-function paradigm
Dynamics Tumour-supressor protein p53’s
disordered segments help it interact
with hundreds of partners.
Elodie Laine – 20.09.2016 Structured & disordered protein building blocks
Elodie Laine – 20.09.2016 Multiple isoforms from a single gene
Does one sequence code for one protein?
Alternative splicing produces several
isoforms from the same gene, by combining
subsets of exons in different ways
Elodie Laine – 20.09.2016 Protein functions
sensor of light
stores iron ions inside cells
supports organs and tissues
hormones
pumps drugs and poisons out of cells
recognizes foreign objects
breaks down
food in the
stomach
copies the information
held in a DNA strand
rotary motor powered by
electrochemical energy
forms structural girders
Elodie Laine – 20.09.2016 Protein functions
The enzyme aconitase is a
key player in the central
pathway of energy
production. It converts
citrate into isocitrate.
Moonlighting
proteins lead
double lives,
performing two
entirely different
functions
The iron regulatory
protein 1 interacts with
messenger RNA to
control the levels of iron
inside cells.
Elodie Laine – 20.09.2016 Protein folding prediction problem
The geometric configuration of a protein’s native state determines its macroscopic
properties, behaviour and function.
The number of possible conformations for a given protein is astronomical.
ex: 100aa, 3 conf/aa => 5 1047 conf
1 fold/10-13 sec => 1027 years (age of Universe: 1010 years)
And yet protein do fold spontaneously in a matter of milliseconds.
How can it be ?
Levinthal’s paradox
Hydrophobicpolar (HP) 2Dlattice model
Elodie Laine – 20.09.2016 Protein dynamics spatio-temporal scales
Na?ve state forma?on Elodie Laine – 20.09.2016 Protein structures are about...
v Chemistry
Ø  atoms, bonds, chirality, pH...
v Physics
Ø  forces, molecular mechanics...
v Biology
Ø  functions, processes, cellular environment...
v Mathematics
Ø  combinatorics (of interactions, of domains)...
v Informatics
Ø  automated algorithms, data analysis...
Elodie Laine – 20.09.2016 Algorithms in structural bioinformatics: what for ?
Ø  To predict protein structures
•  experimental data analysis and
3-dimensional model building
•  secondary or tertiary structure prediction based on the sequence
known protein sequences
increasing gap
protein sequences with function
known protein structures
Elodie Laine – 20.09.2016 Algorithms in structural bioinformatics: what for ?
Ø  To
predict protein structures
Ø To
compare protein structures
•  classification of proteins (divergent/convergent evolution)
•  identification of active sites, functional motifs or binding sites
Phylogenetic tree of
38 CATH Architecture
domain structures
Elodie Laine – 20.09.2016 Algorithms in structural bioinformatics: what for ?
Ø  To
predict protein structures
Ø To
compare protein structures
Ø To
simulate protein motions
•  atomic-level description of the mechanisms underlying protein activity
•  characterization of intermediate conformations that can be targeted by drugs
Elodie Laine – 20.09.2016 Algorithms in structural bioinformatics: what for ?
Ø  To
predict protein structures
Ø To
compare protein structures
Ø To
simulate protein motions
Ø To
characterize protein interactions
•  protein interaction sites identification and complex structures prediction
•  discimination between true partners in the cell and non-interactors
Elodie Laine – 20.09.2016 Algorithms in structural bioinformatics: what for ?
Ø  To
predict protein structures
Ø To
compare protein structures
Ø To
simulate protein motions
Ø To
characterize protein interactions
Ø To
discover and design drugs
•  putative druggable pockets identification
•  binding mode and relative affinity prediction
Elodie Laine – 20.09.2016 Conclusion
•  Proteins are composed of amino acid residues linked by a peptidic
bond. There exist 20 types of amino acids with different physicochemical properties.
•  A protein is a polypeptidic chain. Four levels of organization
determine the 3D atomic coordinates of a protein structure.
•  Proteins fulfill various biological functions: structural, enzymatic, of
transport… Some proteins can have multiple functions.
•  Proteins are dynamic and flexible objects which adapt their shape in
response to environmental conditions.
•  Algorithm in structural bioinformatics can help predict and classify
protein structures, describe their motions and interactions, design new
drugs.
Elodie Laine – 20.09.2016