Evolution of Bcl-2 homology motifs: homology versus homoplasy

Abstract

Bcl-2 family proteins regulate apoptosis in animals. This protein family includes several homologous proteins and a collection of other proteins lacking sequence similarity except for a Bcl-2 homology (BH)3 motif. Thus, membership in the Bcl-2 family requires only one of the four BH motifs. On this basis, a growing number of diverse BH3-only proteins are being reported. Although compelling cell biological and biophysical evidence validates many BH3-only proteins, claims of significant BH3 sequence similarity are often unfounded. Computational and phylogenetic analyses suggest that only some BH3 motifs arose by divergent evolution from a common ancestor (homology), whereas others arose by convergent evolution or random coincidence (homoplasy), challenging current assumptions about which proteins constitute the extended Bcl-2 family.

Figure 1

Discovery timeline for the conglomerate Bcl-2 family. First publication year for proteins with documented or predicted structural similarity with Bcl-2 (above line) and for proteins reported to contain only the Bcl-2 homology (BH)3 motif (below line). (Note that the Bcl-2 protein coding region was published 2 years after the gene was identified in 1984.) For official and alternative protein names, see Figure 2; antiapoptotic Bcl-2-related proteins (green); Bax-like proapoptotic members (orange); Bid-like divergent Bcl-2 homologs (red); canonical BH3-only proteins (blue); other reported BH3-containing proteins (black); viral proteins (yellow). Techniques employed are indicated with superscripts: S, direct sequencing; H, hybridization; P, PCR; E, expressed sequence tag (EST) screen; C, coimmunoprecipitation; I, protein interactive cloning; Y, yeast two-hybrid; D, differential screening; G, genetic screen; F, functional screen; 3D, structural determination. Experimental identification: blue filled circles. Bioinformatics-based prediction of Bcl-2 family membership: pink filled circles. Visual inspection: beige filled circles.

Figure 2

Evidence for Bcl-2 homology (BH) motif compositions in Bcl-2 family proteins and candidate BH3-containing proteins. Schematic representation of protein size and BH motif composition for Bcl-2 homologous proteins (including Bcl-2-like, Bid-like, and Bax-like subgroups), canonical BH3-only proteins, and a large category of other reported BH3-containing proteins. Note that the various depicted BH3 motifs are supported by heterogeneous experimental evidence (Box 2), while BH1, BH2, and BH4 motifs lack experimental validations in most cases. A three-level color coding system indicates BH type (see color key) and predictability: (i) BH motifs predicted by InterProScan (see color key, meaning BH is predicted); (ii) BH motifs that appear in the NCBI entry for this protein (light shades, meaning BH can be uncertain due to automatic annotation); and (iii) BH motifs not predicted or annotated as above, but nonetheless reported with experimental evidence (colored circles). Transmembrane (TM) segments predicted by one (light gray) or both (dark gray) TMHMM Server v. 2.0 and Phobius web server , . Total amino acid (aa) number is indicated for proteins not drawn to scale. Abbreviations for non-human proteins: Ce, Caenorhabditis elegans; Dm, Drosophila melanogaster; NDV, Newcastle disease virus; Mm, Mus musculus; HCV, hepatitis C virus; Pl, Photorhabdus luminescens; Sp, Schizosaccharomyces pombe; SARS-CoV, human severe acute respiratory syndrome (SARS) coronavirus; Sc, Saccharomyces cerevisiae. Number of BH motif hits (right): the number of times a BH motif is predicted to occur within each protein based on the following searches (>1 denotes ambiguity). Predefined amino acid sequence signatures from InterPro (IPR) , from published patterns talis qualis (tq), and from patterns inferred from these publications additum (ad) were used to search the indicated protein and the number of exact matches to the following sequences is reported (shaded cells; unshaded cells had no hits): IPR003093 for BH4; IPR020731 for BH4 is from the PROSITE pattern PS01260 ([DS]-[NT]-R-[AE]-[LI]-V-X-[KD]-[FY]-[LIV]-[GHS]-Y-K-L-[SR]-Q-[RK]-G-[HY]-X-[CW]); 6-mer BH4 tq ([ILVCAGMFYWHKT]-[ILVCAGMFYWHKT]-X2-[FYWH]-[ILVCAGMFYWHKT]) ; 6-mer BH3 ad ([ILVCAGMFYWHKT]-[ILVCAGMFYWHKTE]-X2-[FYWHL]-[ILVCAGMFYWHKT]); IPR020728 BH3 is from PS01259 ([LIVAT]-X3-L-[KARQ]-X-[IVAL]-G-D-[DESG]-[LIMFV]-[DENSHQ]-[LVSHRQ]-[NSR]); 7-mer BH3 (L-X3-G-D-E); 9-mer BH3 ([LM]-[ERAKQL]-[RAQICEY]-[MILAFSE]-[ASG]-D-[DEQKGR]-[LFMIVD]-[NDVHSEQA]) ; 12-mer BH3 tq ([ILVCAGMFYWHKT]-X3-[ILVCAGMFYWHKT]-[KR]-X2-[AGS]-[DE]-X-[ILVCAGMFYWHKT]) ; 12-mer BH3 ad ([ILVCAGMFYWHKTE]-X3-[ILVCAGMFYWHKT]-[KRAQL]-X2-[AGS]-[DE]-X-[ILVCAGMFYWHKT]); 13-mer BH3 tq ([ILVCAGMFYWHKT]-[AGS]-X2-[ILVCAGMFYWHKT]-X2-[ILVCAGMFYWHKT]-[AGS]-D-[ED]-[ILVCAGMFYWHKT]-[NHDY]) ; 13-mer BH3 ad ([ILVCAGMFYWHKTE]-[AGSC]-X2-[ILVCAGMFYWHKT]-X2-[ILVCAGMFYWHKT]-[AGS]-D-[EDKQL]-[ILVCAGMFYWHKT]-[NHDYE]); IPR020717 BH1 from PS01080: [LVMENQ]-[FTLS]-X-[GSDECQ]-[GLPCKH]-X1/2-[NST]-[YW]-G-[RK]-[LIV]-[LIVC]-[GAT]-[LIVMF](2)-X-F-[GSAEC]-[GSARY]); and IPR020726 BH2 from PS01258 (W-[LIM]-X3-[GR]-G-[WQ]-[DENSAV]-X-[FLGA]-[LIVFTC]) (http://www.ebi.ac.uk/interpro/).

Figure 3

Phylogenetic tree of Bcl-2 homologous proteins. Neighbor-joining tree calculated using Poisson-corrected amino acid distances on 74 sites in the core Bcl-2 domain encompassing the Bcl-2 homology (BH)3 motif (or homologous site) through BH2. The tree was rooted with CED-9 sequences from Caenorhabditis. Numbers indicate bootstrap percentages after 1000 replications; values below 50% are not reported. Branch lengths are proportional to distances between sequences. Sequence names contain abbreviations of genus and species (see supplementary information 2 online); primarily antiapoptotic Bcl-2-related proteins (green), primarily proapoptotic Bax-like members (orange), and Bid-like divergent homologs (brown).

Figure 4

Bcl-2 homology (BH)4 and BH3 motif searches identify improbable numbers of hits throughout major phylogenetic groups. InterPro signatures and patterns derived from published alignments (Figure 2) were used to scan UniProt KnowledgeBase with reduced redundancy (no more than 60% identity, UNI60; release 2011_12) for sequences containing BH3 or BH4 motifs. Pie charts show the percent of proteins in each of the phylogenetic groups that were identified as hits by the indicated signature search. The total number of hits detected is shown within the pie chart.