Control systems for membrane fusion in the ancestral eukaryote; evolution of tethering complexes and SM proteins

Abstract

Background: In membrane trafficking, the mechanisms ensuring vesicle fusion specificity remain to be fully elucidated. Early models proposed that specificity was encoded entirely by SNARE proteins; more recent models include contributions from Rab proteins, Syntaxin-binding (SM) proteins and tethering factors. Most information on membrane trafficking derives from an evolutionarily narrow sampling of model organisms. However, considering factors from a wider diversity of eukaryotes can provide both functional information on core systems and insight into the evolutionary history of the trafficking machinery. For example, the major Qa/syntaxin SNARE families are present in most eukaryotic genomes and likely each evolved via gene duplication from a single ancestral syntaxin before the existing eukaryotic groups diversified. This pattern is also likely for Rabs and various other components of the membrane trafficking machinery.

Results: We performed comparative genomic and phylogenetic analyses, when relevant, on the SM proteins and components of the tethering complexes, both thought to contribute to vesicle fusion specificity. Despite evidence suggestive of secondary losses amongst many lineages, the tethering complexes are well represented across the eukaryotes, suggesting an origin predating the radiation of eukaryotic lineages. Further, whilst we detect distant sequence relations between GARP, COG, exocyst and DSL1 components, these similarities most likely reflect convergent evolution of similar secondary structural elements. No similarity is found between the TRAPP and HOPS complexes and the other tethering factors. Overall, our data favour independent origins for the various tethering complexes. The taxa examined possess at least one homologue of each of the four SM protein families; since the four monophyletic families each encompass a wide diversity of eukaryotes, the SM protein families very likely evolved before the last common eukaryotic ancestor (LCEA).

Conclusion: These data further support a highly complex LCEA and indicate that the basic architecture of the trafficking system is remarkably conserved and ancient, with the SM proteins and tethering factors having originated very early in eukaryotic evolution. However, the independent origin of the tethering complexes suggests a novel pattern for increasing complexity in the membrane trafficking system, in addition to the pattern of paralogous machinery elaboration seen thus far.

Figure 1

Location and functions of multi-subunit tethering complexes and SM proteins in an idealised eukaryotic cell. A generalised endomembrane system with only the major trafficking routes is shown, including the multiple recycling pathways that intersect with the Golgi complex and various endosomal subcompartments. The locations of tethering complexes discussed in this article are shown in red and of SM proteins in green. Note that the precise location is not always clear, for example the HOPS complex is known to function in late endocytic steps, but the functional and physical subdivision of the late endosomal population is not precise. Further, some factors may function in more than one locale. The figure is based on Figure 1 from Morgan et al., [64] with the permission of the authors.

Figure 2

Composition, evolutionary conservation and general structure of tethering complexes. The individual subunits of each of the complexes are shown as ovals, arranged according to interaction data from many sources (including pull down, yeast two hybrid, direct structural visualisation and genetic data). For clarity, not all interactions are shown. For example, there is evidence to support interactions between COG 1 and COGs 3 and 4 and additional interactions within the exocyst complex. Individual subunits are colour-coded for evolutionary conservation; red designates a subunit that has wide taxon distribution, whilst subunits in green are absent from multiple taxa. Interactions with factors outside of the tethering complex are indicated in black – the line indicates the approximate interaction interface, if known. See Figure 4 and additional file 3 and the text for full details.

Figure 3

Sequence relationships between tethering factor subunits. Panel A, intercomplex relationships, panel B; intracomplex relationships. Sequence connections between subunits are colour-coded for clarity, and are based on PSI-BLAST hits, but the choice of colour is arbitrary. In panel A, the individual connections have been gathered together into grey ribbons between the complexes for clarity. Note that PSI-BLAST analyses retrieve multiple subunits of GARP, COG and exocyst in a reciprocal manner, while only two such associations are identified between Dsl1 and exocyst. HOPS and TRAPP are not interconnected with the other tethering complexes.

Figure 4

Distribution of tethering complex subunits across representative eukaryotic taxa. Data are based on BLAST results together with alignments – typically the S. cerevisiae or H. sapiens sequences were used as queries, as described fully in methods. Large taxon groupings are colour coded, and a key defining the factors represented by each sector is given at bottom. Filled sectors = an identification based on a clear reverse BLAST result and/or additional evidence through analysis of the sequence by Clustal [58]. Open sectors = not found. Individual BLAST results are provided in additional file 3. Note Dsl3/Sec39p has been omitted from this figure as it is only found in S. cerevisiae.

Figure 5

Phylogenetic analysis of SM proteins from eukaryotes in the major sampled supergroups. This analysis shows that the four SM protein families robustly form separate clades as shown by the bold support values and denoted by the vertical bars to the right of each clade. Support values are given in the order of Bayesian posterior probabilities/ML bootstraps/ML corrected distance bootstraps. That each family encompasses all of the sampled eukaryotic diversity is illustrated by the colour-coded taxon names, which follow the scheme of Figure 4. Support values for all nodes supported by 0.95 posterior probability and 50% bootstrap support or better, are illustrated symbolically. This analysis demonstrates that the SM protein families are ubiquitously found in the sampled eukaryotes and their evolution likely pre-dates the last common eukaryotic ancestor.