Folding, Assembly, Flexible Systems Maxim Petoukhov EMBL, Hamburg - PowerPoint PPT Presentation

Folding, Assembly, Flexible Systems Maxim Petoukhov EMBL, Hamburg Outstation

Outline Outline • Introduction • Combined rigid body and ab initio modelling • Unfolded proteins • Modular proteins with disordered linkers • Transient complexes and weak oligomers • Examples • Conclusions

Re: Atomic Structure Based Modelling Re: Atomic Structure Based Modelling • Conventional rigid body modelling allows quaternary structure analysis of macromolecular complexes, oligomers and modular proteins • The use of complementary data significantly reduces the ambiguity – The resulting models are still of low resolution

Re: Quantitative Analysis of Mixtures ∑ = ( ) ( ) I s v I s k k k OLIGOMER: volume fractions of individual components Prerequisite: models (profiles) of components are known

Small angle scattering: resolution lg I(s) 2.00 1.00 0.67 0.50 0.33 Resolution, nm 8 Resolution is defined by the magnitude of the Atomic largest scattering vector structure in the diffraction pattern d=2 π /s max 7 Shape 6 5 F old 0 5 10 15 s, nm -1

Concept of Dummy Residues Concept of Dummy Residues • Proteins are (folded) polypeptide chains composed of amino acids • At a resolution of ~1 nm each amino acid can be represented as one entity (dummy residue) • For simplicity DRs are – Identical – Centered at the C α positions = < … … > D.I. Svergun, M.V. Petoukhov, & M.H.J. Koch (2001) Biophys. J. 80, 2946-53

Building native-like folds of polypeptides • Using DR-type models and protein-specific penalty functions Secondary Primary Excluded sequence structure volume Number of neighbours 6 5 4 3 2 1 0 0.2 0.4 0.6 0.8 1.0 Shell radius, nm Bond angles & Knowledge-based Neighbors dihedrals distribution potentials distribution

(Mis)-Folding from SAXS? • “Native-like” folds of lysozyme

Modelling of of multidomain multidomain proteins proteins Modelling MRGSHHHHHH GSGVPSRVIH IRKLPIDVTE GEVISLGLPF GKVTNLLMLK GKNQAFIEMN TEEAANTMVN YYTSVTPVLR GQPIYIQFSN HKELKTDSSP NQARAQAALQ AVNSVQSGNL ALAASAAAVD AGMAMAGQSP VLRIIVENLF YPVTLDVLHQ IFSKFGTVLK IITFTKNNQF QALLQYADPV SAQHAKLSLD GQNIYNACCT LRIDFSKLTS LNVKYNNDKS RDYTRPDLPS GDSQPSLDQT MAAAFGLSVP NVHGALAPLA IPSAAAAAAA AGRIAIPGLA GAGNSVLLVS NLNPERVTPQ SLFILFGVYG DVQRVKILFN KKENALVQMA DGNQAQLAMS HLNGHKLHGK PIRITLSKHQ NVQLPREGQE DQGLTKDYGN SPLHRFKKPG

Modelling of of multidomain multidomain proteins proteins Modelling • A combined approach is proposed to built the models of multidomain proteins with large and flexible interdomain linkers • The latter are represented as DR chains which are attached to the appropriate terminals in rigid domains. • A single modification of a model is a rotation about one or two randomly selected DR(s).

Modelling of of multidomain multidomain proteins proteins Modelling

Building native- -like folds of linkers like folds of linkers Building native Dihedral angles, degrees 150 Absence of 100 steric clashes 50 0 i+K -50 r -100 -150 20 40 60 80 100 120 140 160 Bond angles, degrees Bond angles & i dihedrals distribution Neighbors distribution along the sequence Loop compactness = 3 Rg 3 n may also be required id l

Simultaneous fitting of multiple data sets from Simultaneous fitting of multiple data sets from deletion mutants deletion mutants lg I, relative 11 10 9 8 0.5 1.0 1.5 2.0 s, nm -1

BUNCH: Modelling Modelling of of multidomain multidomain proteins proteins BUNCH: • Search of the optimal positions and orientations of rigid domains and probable conformations of DR linkers, those fit the SAXS data. • Proper bond and dihedral angles in the DR chains are required together with the absence of overlaps. • The scattering pattern is calculated from partial amplitudes of domains and form-factors of DR comprising the loops using spherical harmonics. ∞ l ( ) ( ) ( ) 2 ∑ ∑ ∑ ∑ = π + 2 ( ) ( ) k i 2 | | I s A s D s lm lm = = − 0 l m l k i • Multiple scattering curves fitting from deletion mutants Petoukhov M.V., Svergun, D.I. (2005). Biophys. J. 89 , 1237-1250

Structure and RNA interactions of polypyrimidine tract binding protein PTB is an important regulator of alternative splicing, which allows the production of multiple mRNA transcripts from a single pre-mRNA species. PTB contains four domains (RNA recognition motifs, RRMs), whose structure is solved by NMR. K134 K137 N H133 L136 D F98 2 N 3 H62 Q96 1 K92 4 K64 K94 C K65 R122 R185 K266 N L255 B K259 3 2 C 5 K238 1 4 K271 A F216 R254 K218 Multiple scattering curves from I187 NMR: high resolution structures deletion mutants fitted simultaneously Q223 of RRM1 and RRM2 Collaboration: S.Curry (London)

Structure and RNA interactions of polypyrimidine tract binding protein Further restraints (e.g. from NMR) are required Overlap of the typical ab initio to resolve the orientational ambiguity and rigid body models Petoukhov, M. V., Monie, T. P., Allain, F. H., Matthews, S., Curry, S., and Svergun, D. I. (2006). Structure 14 , 1021-1027.

CORAL: Crossing SASREF & BUNCH • Bunch: • Sasref: – Single polypeptide – Does not account for chain missing portions – Impossible to fix more than one domain

Random Loop Library for Combined Modelling RANLOGS database CORAL

Hybrid Modelling in Coral RANLOGS database CORAL Novel feature: consorted movements

Kratky Plots to Detect Disorder Unfolded Patterns of globular and flexible proteins Folded Multi-domains with flexible linkers 20

Biomolecules are Dynamic Entities ► Local Protein Fluctuations: Backbone and side-chains Protein vibrations, loop motions and breathing to facilitate interactions and catalysis ► Concerted domain motions: Linkers as Hinges Specific and limited domain jumps between relative positions often linked with catalysis in on/off mechanisms ► Highly flexible regions or domains An astronomical number of conformations are available. Flexible multi-domain proteins (MD) and Intrinsically Disordered Proteins (IDPs) Linked with signaling and regulation

Flexible Proteins: Importance ► Many biological functions such as transcription, regulation, cell cycle control, require extensive flexibility ► Is more common in higher organisms that have to perform more and more controlled functions. Disorder is correlated with complexity ► High selectivity, moderate affinity, and promiscuity are properties often linked to flexibility ► In these systems partially structured conformations or transient long range interactions can be crucial for biological activity. Structural studies are important

Indications (not Proofs!) of Flexibility ► Smooth Scattering profiles and featureless Kratky Plots ► Large R g and D max ► Absence of correlation peaks in the p(r) function ► Low correlation densities in ab initio reconstructions ► Isolated domains in rigid body modelling ► Prediction of disorder using bioinformatics tools http://www.idpbynmr.eu/home/science/research-tools.html

Flexibility as mix of different conformations D sin sr ∫ = π ( ) 4 ( ) I s p r dr sr 0 For monodisperse systems the scattering is proportional to that of a single particle averaged over all orientations ∑ = ( ) ( ) I s v I s k k k v k = volume fraction I k (s) = scattering intensity from the k -th component 24

Detection of Flexibility PolyUbiquitin Molecules Flexible Rigid 2,3,4 and 5 Ubiquitin (72 AA) domains connected by 20 AA linker (RanCH) Flexible Multidomain Proteins present less features than rigid counterparts Bernadó Eur. Biophys. J. 2009, 39, 769

Detection of Flexibility SAXS curves Analysis of the overall size descriptors (R g , p(r) , Kratky) Rigid Scenario Modelling: ab initio (DAMMIN/DAMMIF) Go for flexibility! and Rigid body (BUNCH/CORAL) Flexible Analysis of the differences Scenario

Ensemble Optimization Method curve structures

Ensemble Optimization Method

Ensemble methods in SAXS ... ... N 1 ∑ = ( ) ( ) I s I n s N = 1 n 5 ... R g 1 R g 2 R g 3 R g 4 R g ρ (R g )

C r o s s i n N g Mutations Experimental Elitism Curve

R1 R2 R3 Crossing N … Mutations G Generations Experimental Elitism Curve R1 R2 R3 R g Distribution … Ensemble Optimization Method (EOM) Bernadó, Mylonas, Petoukhov, Blackledge, and Svergun. J Am Chem Soc 2007, 129 :5656-64.

Modelling: Native vs. Random φ C α ψ Quasi C α -C α Ramachandran plot C α Bond angles vs. Dihedral angles G. Kleywegt , Validation of protein models from C α coordinates alone , JMB , 1997, 273, 371-376 Theoretical distribution of the bond and dihedral angles for random chains R g R 0 Persistence Length R g = R 0 ·N ν ν Solvent ‘quality’ Several experimental and theoretical studies establish ν ≈ 0.588 as an indication of the ‘random coil’ in chemically denatured (Urea or GuHCl) proteins. Kohn et al. PNAS , 2004, 101, 12491 N 32

Missing loops (i.e. flat electron density map) … Nter.pdb Cter.pdb curve.dat Kratky Plot vs. apoferritin MRIGMV……..GGVQSHVLQ…..VLRDAGHEVS…….PHVKLPDYVS seq.seq missing loop 30 AA pool 33