From Protein Structure to Function with Bioinformatics.pdf

4 Membrane Protein Structure Prediction 97Table 4.2 Beta-barrel transmembrane protein superfamilies, taken from the OPM database(Lomize et al. 2006b)SourceSuperfamilyOuter membranes of Gram-negative bacteria OMPA-likeOligomeric beta-barrels of Gram-positive OMPT-likebacteriaAutotransporter (AT)Trimeric autotransporterOM phospholipaseNucleoside-specific channel-forming membraneporinFadL outer membrane protein (FadL)OmpG porinTrimeric porinsSugar porinsOmp85-TpsB transportersLigand-gated protein channelsOuter Membrane Factor (OMF)Leukocidin-like(Dayhoff et al. 1978) or BLOSUM (Henikoff and Henikoff 1992). Such matricesare derived from globular protein alignments, and as amino acid composition,hydrophobicity and conservation patterns differ between globular and TM proteins(Jones et al. 1994a), they are in principle unsuitable for TM protein alignment. Anumber of TM-specific substitution matrices have therefore been developed, whichtake into account such differences. For example, the JJT TM matrix (Jones et al.1994b) was based on the observation that polar residues in TM proteins are highlyconserved, while hydrophobic residues are more interchangeable. Other matricessuch as SLIM (Müller et al. 2001), were reported to have the highest accuracy fordetecting remote homologues in a manually curated GPCR dataset, while PHAT(Ng et al. 2000) has been shown to outperform JJT, especially on database searching.However, to date, no independent study has accessed these TM-specific substitutionmatrices on a common dataset.Few novel methods have been developed to improve actual TM protein alignment.STAM (Shafrir and Guy 2004) implemented higher penalties for insertion/deletions inTM segments compared to loop regions, with combinations of different substitutionmatrices to produce alignments resulting in more accurate homology models.PRALINE (Pirovano et al. 2008), which integrates state-of-the-art sequence predictiontechniques with membrane-specific substitution matrices, was shown to outperformstandard multiple alignment techniques such as ClustalW (Higgins et al. 1994)and MUSCLE (Edgar 2004) when tested on the TM alignment benchmark set withinBaliBASE (Bahr et al. 2001). Recent adjustment to BLAST and PSI-BLAST (Altschuland Koonin 1998) to reflect the composition of the query sequence should theoreticallyimprove results for TM protein searches (Altschul et al. 2005), though again thishas not been assessed. An advanced alignment method T-Coffee (Notredame et al.