Construction of Alignments for Subfamilies of EF-Hand Calcium-Binding Proteins

EF-Hand CaBP Data
Library Sequence Information

Construction of Multiple Sequence Alignments for EF-Hand Calcium-Binding Proteins

All alignments were generated with CLUSTALW, using the online interface provided by the Network Protein Sequence Analysis website. All settings were left at their defaults (weight matrix = blosum, gap opening penalty = 10.0, gap extension penalty = 0.05, hydrophilic gaps = on, hydrophilic residues = GPSNDQERK, residue specific gap penalties = on).

These alignments were constructed in order to determine the amino acids at each position of the EF-hand consensus sequence (defined for this purpose to be the binding loop and nine residues on either side). Therefore, the most care was taken to ensure that the alignments were reasonable within these regions, and these are the most reliable regions of the alignments. Furthermore, since the purpose of the alignments was to allow comparisons of residues at the consensus positions, the number of gaps within the EF-hands were minimized. Gaps were allowed in the binding loops where necessary, and in the flanking helices only when other members of the subfamily did not require the gap. This means that the alignment between subfamilies should not be taken as one that accurately reflects the sequence similarity between the subfamilies.

For some of the subfamily alignments, it was necessary to adjust the CLUSTALW output by hand in order to align the EF-hands correctly. The necessary corrections for each subfamily are listed below. if a subfamily does not appear on this list, its alignment is the direct result of the CLUSTALW run.

the calpain subfamily: the inclusion of the ALG-2 sequence in the CLUSTALW run caused problems in the alignment of the first EF-hand, as judged by the reported locations of the ancestral binding loops in the various proteins. To circumvent this problem, an alignment of the calpain subfamily without the ALG-2 sequence was generated using CLUSTALW (NOALG2). An alignment of sorcin, granclacin, and ALG-2 was also generated from CLUSTALW (NOCALPS). The NOCALPS alignment was then used to paste the ALG-2 sequence into the NOALG2 alignment. Some minor changes were also made to the fourth binding loop in the sorcins and grancalcin based on the NOCALP alignment (these changes involved moving gaps in the sequence by one or two residues). The alignment of the SCHJA sorcin was also adjusted, based on its closer similarity to ALG-2 (by visual inspection). These changes were slightly more significant, but still restricted to a small region (the region just after binding loop 1), and were supported by the NOCALP alignment.
SPARC/BM-40: the inclusion of a fragmentary SPARC sequence from pig lead to obviously incorrect alignments-- regions in the bovine sequence failed to align with identical and nearly identical regions in the other sequences. Therefore, the alignment was regenerated without any fragmentary SPARC sequences.
spectrin alpha chains: it was necessary to delete much of the non-EF-hand sequences in order to get the putative EF-hand sequences to align. Even so, there is some uncertainty about the alignment of the EF-hands. The SWISS-PROT files for SPCN_HUMAN and SPCA_HUMAN report different regions as the first EF-hand loop, which do not align in the alignment. It seems that the SWISS-PROT file for SPCN_HUMAN is in error, based on the reported NMR structure of chicken alpha spectrin, which seems to match the sequence in SPCN_CHICK. In this structure, the first EF-hand corresponds to the sequence identified in the SPCA_HUMAN SWISS-PROT file.