1. Arne Elofsson (arne@bioinfo.se) On the evolution of membrane proteins Homology detection Blast, PSIBLAST are they good ? SHRIMP (Bernsel, submitted) Evolution of TM-helices Internal duplications >50% of +6TM proteins

2. Arne Elofsson (arne@bioinfo.se) Problems with homology detection of TM proteins Methods optimized on globular proteins High scoring false positives due to similarity between unrelated TM regions Hedman et al, 1998 (Pmembr): Fold recognition methods improve this fact Globular detections improved by profile- profile (HMM-HMM) comparisons

3. Arne Elofsson (arne@bioinfo.se) Profile-profile methods improve detection of membrane proteins

4. Arne Elofsson (arne@bioinfo.se) Classification of GPCRs at 1% FPR

5. Arne Elofsson (arne@bioinfo.se) How does TM proteins evolve Mutation rates any differences in respect to the different compartments. Few “multi-domain” membrane proteins Many complexes Internal symmetry

6. Arne Elofsson (arne@bioinfo.se) Slower evolution in the membrane

7. Arne Elofsson (arne@bioinfo.se) Internal duplications on sequence level Searched for internal duplication events in 87 prokaryotic genomes Out of 38,124 membrane proteins, 377 duplication events could be detected (~1%) Internal duplications has been hard to detect on sequence level αααααα XABC J. Mol. Biol. 2004, 339:1-15

8. Arne Elofsson (arne@bioinfo.se) Lactose Permease

9. Arne Elofsson (arne@bioinfo.se) 1xqf (NH 3 -channel) Antiparallel duplication

10. Arne Elofsson (arne@bioinfo.se) 1xqf Antiparallel duplication

11. Arne Elofsson (arne@bioinfo.se) 12-TM chain from Cytochrome C-oxidase(Triplication)

12. Arne Elofsson (arne@bioinfo.se) Internal symmetry in α-helical inner membrane proteins 2-fold symmetry axis Parallel to membrane Odd number of TMH Domains anti-parallel to each other In plane with membrane Even number of TMH Domains parallel to each other N middle C C N Parallel: LacY GlpT Sav1866 ABC transporter AcrB EmrD ADP/ATP carrier Antiparallel: LeuT SecY BtuCD AQP1 GlpF AmtB ClC H + /Cl - exchange transporter NhaA

13. Arne Elofsson (arne@bioinfo.se) Many (most) proteins have internal duplications

14. Arne Elofsson (arne@bioinfo.se) Alpha-helical inner membrane proteins Support for the idea that α-helical membrane proteins increase in size and complexity by gene duplication DuplicationGene fusion 4TMH2*4TMH8TMH gene

15. Arne Elofsson (arne@bioinfo.se) Duplications in TM proteins

16. Arne Elofsson (arne@bioinfo.se) Both parallel and antiparallel duplications exist

17. Arne Elofsson (arne@bioinfo.se) But in order to have antiparallel domains, they must be inserted with opposing topology! In order for an antiparallel symmetry to be explained by a duplication event, each subunit of the homodimer must be inserted with opposite topology That implies that a sequence can have dual topology, be inserted either way in the membrane Have been constructed in labs (Cell 62, 1135-1141 & Cell 77, 401- 412) inside outside membrane N C N C

18. Arne Elofsson (arne@bioinfo.se) Hypothesis for antiparallel internal symmetry Nat. Struct. Mol. Biol. 2006, 13: 94-96

19. Arne Elofsson (arne@bioinfo.se) Dual topology candidates from E. coli paper

20. Arne Elofsson (arne@bioinfo.se) Detection of internal duplications Structural alignments Finds most of the manually detected examples. Sequence alignments Sequence alignments do not find most. Using SHRIMP about 80% can be found Also finds a few extra. Some problems defining exact duplication unit Genome scanning Ongoing.

21. Arne Elofsson (arne@bioinfo.se) Acknowledgements Erik Granseth Anni Kauko Jenny Falk Andreas Bernsel Kristoffer Illergård

22. Arne Elofsson (arne@bioinfo.se) Final comments Thanks to EMBRACE Please fill out the questionaire..