Conservation of Orientation and Sequence in Protein DomainDomain Interactions Stephen J. Littler and

1
Conservation of Orientation and Sequence in
Protein Domain-Domain InteractionsStephen J.
Littler and Simon J. HubbardJ.Mol.Biol.(2005)
345, 1265-1279

• CRB Evening Journal Club
• March 14, 2005

2
Challenge for Structural Genomics

• Identify the complete, prototypical set of
  structural domains
• Characterise how they interact with one another
  and how they orientate themselves within a single
  polypeptide chain to form a functional unit.
• Over 2000 novel domain-domain combinations
  estimated where structure still needs to be
  elucidated.

3
Previous Research Efforts

• Mostly focused on protein chains.
• Protein networks
• Work relating to domain interaction has been
  centered on chemical nature or packing at
  interfaces.
• Better understanding of domain-domain interaction
  is required to make prediction and complement
  structural genomics efforts.

4
Current Knowledge of Domain Interactions

• Correlated mutation patterns are prevalent
  between interacting domain partners.
• 2 and 3 domain combination exist in over 1/3 of
  structurally assigned multi-domain proteins.
• Low sequence identity homologous domain-pairs may
  not interact in the same way.

5
Survey of InteractionsStructures and Genomes

• Majority of domains limit combinations to very
  few partners.
• Like-like combinations
• Only one other domain superfamily
• N and C terminal orientation is invariable in
  majority of domain-domain combinations.
• Intra-chain domain interaction essential to
  function.

6
Accomplishments of this Paper

• Conducted comprehensive analysis of nature of
  intra-chain domain interactions using SCOP
  superfamilies.
• Analyze residue conservation in terms of sequence
  and volume.
• Examine overall orientation of domain partners
  found in intra-chain domain pairs.
• Reconsider the assignment of interactive domains
  within genome sequences.

7
Domain Assignment to Structure

• Define Interacting Domains
• PDB structures (PDB-SELECT _at_ 25 and 90)
• According to SCOP superfamilies
• Assignment with SUPERFAMILY hidden Markov model
  library.
• Interactions considered in pairwise fashion
  within single polypepetide chain.

8
Analyze Interface Residue Usage

• Assign residues into 5 accessibility classes
• Accessible
• Core buried
• Interface-A interdomain, within chain
• Interface-B interdomain, between chains
• Interface-C unassigned to domain

9
Analyze Interface Residue Usage

• Assess level of available surface used by
  Superfamily.
• Align structures to representative structure to
  create an interface map.
• Percentage of residues used at interfaces with a
  score, domain combination residue usage (DCRU).
  (one or more members)
• Allows for comparison between different domain
  superfamily in terms of total surface usage
  across the superfamily.

10
DCRU Analysis
Ranges 6.2 to 73.2 Mean 27.3
11
DCRU Conclusions

• DCRU values are varied.
• Little difference between multiple-partner and
  single-partner superfamilies.
• Single-partner superfamilies tend to reuse the
  same area of surface for domain combinations with
  the same,single partner.
• Example Chaining proteins.

12
Conservation of Domain Orientation

• Conservation of domain partner orientation is
  measured with respect to representative
  superfamily domain.
• Superposing member structures to the
  representative, complete with associated partner
  domains.
• Two measures of angular conservation between
  center of mass of interacting domains.
• Conservation of Interface Orientation(CIO)
• Conservation of Interface Angle (CIA)

13
Superfamily partner Orientation

• Proteins aligned with CE to representative domain
  interaction.
• Creates satellites around representative
  structure.
• Calculate Center of Mass (COM) for each domain.
• Only structurally aligned residues are used.
• Calculate vector between the representative
  domains COM and the partner domains COM.

14
Conservation of Interface Orientation

• Quantify orientation as a magnitude of the vector
  sum of all satellite unit vectors.
• A low CIO value indicates that angles between
  satellite vectors are small and domain
  orientation is conserved
• CIA is the mean angular difference in degrees
  between all pairwise comparisons of interaction
  vectors.

15
CIO and CIA for Single Partners
16
CIO and CIA for Multiple Partners
17
Categorization by Interaction Types

• a) Single partner/Single region (33)
• CIA lt7 and CIO 0.003
• b) Single partner/Multiple Region (13)
• CIA lt96.2 and CIO 0.97
• c) Multi-partner/Single Region (11) CIA lt30.2
  and CIO 0.07
• d) Multi-partner/Multiple Region (22)
• (79 Domain Superfamilies)

18
Conservation of Sequence and Volume

• Quantify the conservation of residues at
  inter-domain boundaries throughout evolution in
  sequence and in structure.

19
(No Transcript)
20
Distance Tendencies between domains

• Apic et. al defined domain combination with
  inter-domain distance cutoff of 30 residues.
• Only 482 of the 2545 domain combination had
  protein structures.
• Evaluate the utility of this cut off with respect
  to structural context.

21
Prediction Accuracy of Various Interdomain
Interaction Definitions
22
Take Home Message

• Comprehensive survey of domain-domain interaction
  in protein structure and genome sequence.
• Different promiscuity of interaction. Modes of
  interaction with respect to surface sites and
  binding angle was investigated.
• Sequence conservation at binding interfaces.
• Adopt an approach that utilizes individual
  probabilities based on domains to predict
  domain-domain interactions.

23
Structurally Underrepresented Domain-Domain
Interactions