The ubiquitous and ancient ER membrane protein complex (EMC): tether or not?

The recently discovered endoplasmic reticulum (ER) membrane protein complex (EMC) has been implicated in ER-associated degradation (ERAD), lipid transport and tethering between the ER and mitochondrial outer membranes, and assembly of multipass ER-membrane proteins. The EMC has been studied in both animals and fungi but its presence outside the Opisthokont clade (animals + fungi + related protists) has not been demonstrated. Here, using homology-searching algorithms, I show that the EMC is truly an ancient and conserved protein complex, present in every major eukaryotic lineage. Very few organisms have completely lost the EMC, and most, even over 2 billion years of eukaryote evolution, have retained a majority of the complex members. I identify Sop4 and YDR056C in Saccharomyces cerevisiae as Emc7 and Emc10, respectively, subunits previously thought to be specific to animals. This study demonstrates that the EMC was present in the last eukaryote common ancestor (LECA) and is an extremely important component of eukaryotic cells even though its primary function remains elusive.


Amendments from Version 1
In response to reviewer comments, the results and discussion have been expanded to include a short section on Emc8 and Emc9 and a supplementary phylogenetic tree ( Figure S1). The analysis demonstrates that Emc8 and Emc9 are vertebratespecific paralogues resulting from a duplication of an ancestral protein in the lineage leading to vertebrates (Phylogenetic methods appear in the figure legend to Figure S1 -Associated references have also been added). These results suggest that vertebrate Emc8 and Emc9 should be renamed Emc8a and Emc8b, respectively. Figure 1 has been modified accordingly.

Introduction
Recent studies suggest that the EMC (Endoplasmic Reticulum Membrane Complex) is a multifunctional, multi-subunit protein complex. In Homo sapiens, the EMC comprises ten subunits, Emc1-10, whereas in Saccharomyces cerevisiae the complex comprises only Emc1-6 (Jonikas et al., 2009). The EMC has been implicated in several cellular processes. It has been implicated in ERAD (ER-associated degradation) (Christianson et al., 2012;Jonikas et al., 2009;Richard et al., 2013) but the molecular mechanism for how EMC triggers ERAD has remained elusive. Emc6 contains a Rab5 interacting domain and has been shown to interact with Rab5A in humans during autophagosome formation (Li et al., 2013). It has also been shown that the EMC is an ER-mitochondria tether in S. cerevisiae that interacts with the outer membrane protein Tom5 of the TOM (Translocase of the Mitochondrial Outer Membrane) complex (Lahiri et al., 2014). Most recently, the EMC has been implicated in the proper assembly of multi-pass transmembrane (TM) proteins (Satoh et al., 2015). These recent findings suggest that EMC involvement in ERAD may be due to secondary effects, as cells devoid of EMC components may result in either disruption of ER-mitochondria tethering, or the misfolding of multipass membrane proteins. Thus, the primary function of the EMC is still open for debate.
The ER-mitochondria encounter structure (ERMES), also involved in ER-mitochondria tethering, is a multifunctional protein complex implicated in both lipid transfer and mitochondrial outer membrane protein assembly (AhYoung et al., 2015;Kornmann et al., 2009;Meisinger et al., 2006, Meisinger et al., 2007Wideman et al., 2013;Wideman et al., 2010). However, ERMES as an ER-mitochondria tether is limited to a subset of eukaryote taxa (Wideman et al., 2013), suggesting that a universal ER-mitochondria tethering complex might exist. Lahiri et al. (2014) state in their title that the EMC is a conserved protein complex. However, by stating that a protein is conserved, cell biologists and biochemists often simply mean that the protein is present in S. cerevisiae (fungi) and animals. Since the clade comprising animals and fungi only accounts for one fifth of the diversity of eukaryotes (Adl et al., 2012), more work is necessary in order to support the claim made by Lahiri et al. Thus, I was prompted to investigate the taxonomic distribution of the EMC in order to (1) determine if it really is a conserved protein complex and (2) if it could possibly represent the pan-eukaryotic ER-mitochondria tether. Table S1 for accession numbers) from H. sapiens and S. cerevisiae were used as queries in BLAST (Altschul et al., 1997) and pHMMer (Finn et al., 2011) searches into the predicted proteomes of 70 organisms spanning the diversity of eukaryotes. Retrieved sequences were considered orthologous if they retrieved the original H. sapiens or S. cerevisiae EMC sequences as top hits when used as reciprocal BLAST or pHMMer queries into H. sapiens or S. cerevisiae predicted proteomes and did not retrieve any other closely related sequences (except in the case of Emc8 and Emc9, see below). In cases in which EMC components could not be identified in this manner, transcriptomes and genomes were searched using bioinformatically validated sequences from the previous step that were retrieved from closely related species. Genomes were downloaded from public repositories and genome project websites. See Table S1 for retrieved sequences.

Results and discussion
The EMC is an ancient and highly conserved protein complex Using homology-searching algorithms EMC candidate proteins were identified in the vast majority of sequenced genomes representing the complete diversity of eukaryotes ( Figure 1). Emc8 and Emc9 are homologues but only a single homologue could be detected in most genomes. A subset of opisthokont Emc8/9 sequences were subjected to a phylogenetic analysis demonstrating that vertebrate Emc8 and Emc9 are the result of a vertebratespecific duplication of Emc8 ( Figure S1; see legend for methods). Based on this knowledge, I suggest that the vertebrate Emc8 and Emc9 be renamed Emc8a and Emc8b, respectively.
A complete EMC (Emc1-8, 10) was found in at least one representative from each major lineage including animals, fungi, excavates, amoebozoa, green algae, plants, stramenopiles, alveolates, rhizaria, and haptophytes ( Figure 1). The relative sequence conservation of EMC components across diverse taxa suggests that the EMC has an ancient and critical role in cellular function.
Yeast Sop4 and YDR056C are Emc7 and Emc10, respectively Although previous reports suggest S. cerevisiae EMC comprises only six subunits, I identified Sop4 and YDR056C as orthologues of Emc7 and Emc10, respectively. Supporting this, Jonikas et al. (2009), the original discoverers of the EMC, show by co-immunoprecipitation analyses that Sop4 and YDR056C are interacting partners of FLAG-tagged Emc3. This experiment not only confirms my bioinformatic classification but also puts into perspective a previous study on Sop4's role in membrane protein quality control (Luo et al., 2002). Furthermore, tracing the evolutionary history of the EMC in fungi demonstrates that Emc8 was lost only in Ascomycetes and a few basally diverging fungi whereas most fungi retain Emc8 (as well as Emc7 and 10).

The EMC has been independently lost in several lineages
Although the EMC was identified in representative taxa from every major eukaryote supergroup, I was unable to identify even a single EMC member in the genomes of the microsporidians Nosema ceranae and Encephalitozoon cuniculi, the metamonad Giardia  (Field et al., 2013). Asterisks indicate presence of orthologue in a different member of the genus but absent in the indicated species (see Table S1).
Abbreviations intestinalis, the stramenopile Blastocystis hominis, the alveolate Theileria parva, and the red alga Cyanidioschyzon merolae ( Figure 1 and Figure 2). Trichomonas vaginalis, another metamonad retains only a rather divergent Emc2, that passed the test for orthology, but only weakly, suggesting that this protein is under relaxed selection, perhaps repurposed, or in the process of being lost. All other genomes from the remaining 65 species investigated contained clear representatives of EMC homologues (Figure 1).
These disparate organisms that lack the EMC prompted the question: What cellular or biochemical features tie these diverse organisms together? The microsporidians, metamonads and B. hominis all contain reduced anaerobic mitochondria-related organelles (MROs) and also lack the EMC. However, the amoebozoan Entamoeba histolytica retains Emc1-4, 7 and 10, the apicomplexan Cryptosporidium parvum retains Emc1-4, and 8, and the fungus Piromyces sp. retains Emc1-4, 6, 7, and 10, but all three organisms also contain extremely reduced MROs. T. parva and C. merolae contain relatively normal mitochondria but completely lack the EMC. Thus, it seems that further insight into the cell biology of these organisms is required to understand why only these few species from unrelated lineages have lost the EMC. At this point, of the proposed functions of the EMC, its involvement in multipass membrane protein assembly is the best candidate for generalization to other eukaryotes. It explains the connection to ERAD as a secondary effect of misassembled multipass proteins and explains why an organism with extremely reduced mitochondria (E. histolytica) might retain the EMC. Finally, although EMC involvement as an ER-mitochondria tether is attractive, the distribution of the only known MOM-localized interactor of EMC (Tom5) has not been identified in organisms other than animals and fungi (Maćasev et al., 2004). Thus, until an ancient interaction partner is identified, the role of EMC as an ancient tether remains speculative.

Conclusions
Since the vast majority of species from each major branch of eukaryotes retain the EMC it can be inferred that it was present in the last eukaryote common ancestor (LECA). Since the sequences of most of the identified EMC homologues are very similar, it can be inferred that its function has likely been retained in most  eukaryote lineages. Thus, the EMC is a generalizable eukaryotic feature as is its function-whatever it might be.

Data availability
All sequence data are freely available in online databases (NCBI, JGI, or independent genome sequencing project websites).

Competing interests
No competing interests were disclosed.

Grant information JGW is the recipient of the European Molecular Biology Organization (EMBO) Long-term Fellowship (ALTF 761-2014).
I confirm that the funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Acknowledgements I thank Joel Dacks for server access, computational time, and inspiration.
Supplementary Figure S1. Phylogenetic analysis of opisthokont Emc8/9 proteins. Proteins were aligned using MUSCLE (Edgar, 2004) and manually adjusted as needed using Mesquite (

Supplementary materials
Supplementary The article entitled "The ubiquitous and ancient ER membrane protein complex (EMC): tether or not?" authored by Jeremy Wideman presents a solid bioinformatics argument that the EMC protein complex is highly conserved amongst eukaryotes. The EMC was originally identified in budding yeast as a 6 subunit complex (Emc1-6) with a role in protein folding in the endoplasmic reticulum ( ). It was Jonikas ., 2009 et al later expanded to 10 subunits (Emc1-10) in mammals based on proteomic work studying ER-associated degradation ( ). This current bioinformatics study makes an nice contribution by Christianson 2011 et al., showing that the whole complex (all ten subunits, with a few exceptions) is widely present in eukaryotes (including invertebrates, fungi and plants). Although overlooked in the original submission, the issue raised by one reviewer about Emc8 and Emc9 being paralogs has now been resolved by the author.
My main issue with the article is the conclusion drawn by the author that the tethering function of the EMC (discovered by ) is likely not its conserved function. This is arrived at by comparing EMC Lahiri ., 2014 et al distribution among species with the presence of mitochondria/MROs. The article title also implies that the findings in this report call into question the role for the EMC in ER-mitochondrial tethering and PS transport (as its conserved function). I feel these are overstatements given the current analysis presented in the paper. Lahiri , show interactions between the EMC and Tom5, although these are not et al. demonstrated to be direct, and also state in their discussion that the EMC likely interacts with multiple subunits of the TOM complex (and cite unpublished data to the effect). Hence, judging a role for the EMC in tethering solely based on the presence Tom5 in species seems hardly sufficient to make such an argument. Lahiri , demonstrate a role for tethering by the EMC in PS transport to mitochondria, the et al. location of the phosphatidylserine decarboxylase (PSD) that converts the PS into PE. Hence, this transport step is required for PE synthesis by the PSD. Perhaps the author should investigate the coincidence of the EMC and mitochondrial-localized PSD enzymes in the species for which he uses to build arguments against a role for the EMC in tethering. A quick search revealed that two species mentioned in the paper, and , which completely lack the EMC, contain PSD enzymes C. merolae T. parva that lack mitochondrial-targeting signals (28% and 50% probability, respectively; compared to 95% for S. PSD); hence, the EMC would not be needed in these organisms for PE synthesis. The third cerevisiae EMC-lacking species mentioned, , has a mitochondrial-targeted PSD (99.9% probability) and a B. hominis second non-mitochondrial PSD (0.01% probability), indicating that there is not an absolute requirement for ER-mitochondrial PS transport and hence, for the EMC in PE synthesis in this species.
If the author feels that an analysis of the co-occurrence of the EMC and mitochondrial-localized PSD enzymes (and/or TOM complex -all subunits) is beyond the scope of this paper, I feel the paper should be revised, including the title, to de-emphasize the argument that tethering is not a conserved function of F1000Research enzymes (and/or TOM complex -all subunits) is beyond the scope of this paper, I feel the paper should be revised, including the title, to de-emphasize the argument that tethering is not a conserved function of the EMC. An additional minor point, the author should name Emc7 and Emc10 on the SGD website for the benefit of the yeast bioinformatic community.
I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
No competing interests were disclosed. It was gratifying to see that the author has further expanded his study based on my observations. The new finding of Wideman that Emc8 and Emc9 are vertebrate-specific paralogues explains the gain of EMC components in vertebrates. We'll, however, have to wait for future research to know whether the duplication of Emc8 in these higher eukaryotes has any functional relevance. I feel that the author has rightfully addressed my major concern and thus approve the indexation of this revised manuscript.

I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
No competing interests were disclosed. Competing Interests:

F1000Research
As you suspected, some of your other comments are out of the scope of the paper, but I would like to comment on them here for anyone that is interested.
First, regarding your comment "it would be interesting to correlate the presence of various TOM components in these 'amitochondriates' with the various EMC components": yes this would be interesting, however, I believe best included in a larger study on the evolution of protein import pathways. Tom5, the only known MOM interactor of EMC is found only in Opisthokonts (Macasev et al. 2004), although this has not been investigated in detail for quite some time. The protein is so short (~50aa) that it is easily missed in bioinformatic analyses; even if the protein is more widespread, it may be that it will only be identified biochemically. Additionally, most amitochondriates have extremely divergent Tom complexes (e.g. , microsporidians, Entamoeba ), and it is unlikely that even if a small protein like Tom5 is present in these organisms that it Giardia will be detectable by phylogenetic analysis.
Second, prokaryotes do not seem to have any close homologues (based on a preliminary BLAST into NCBI prokaryote database) but some weak homology can be detected. Further investigation is beyond the scope of this project.
Third, the likelihood of HGT of EMC components is quite low in this case given the high frequency of retention of EMC across all eukaryotes. Also, for many of the proteins it is unlikely that phylogenies would resolve HGTs as many of the proteins are very short and support values would be low.
No competing interests were disclosed. The article titled "The ubiquitous and ancient ER membrane protein complex (EMC): tether or not?" authored by Jeremy G. Wideman is a comprehensive study to determine if the EMC proteins are truly conserved. Using homology searching algorithms the author has shown that except for a few branches the EMC proteins are present in the vast majority of the eukaryotes and reasoned in favor of the presence of EMC proteins in the last eukaryote common ancestor (LECA). In addition the author has also identified two genes in to be orthologues of Emc7 and Emc10 which supports the finding of Jonikas S. cerevisiae et . where these two proteins were co-immunoprecipitated along with the other EMC proteins. al In light of the increasing scientific attention on the EMC proteins over past few years and their multifaceted roles in cell physiology I find this article to be quite relevant in delivering a thorough understanding of this protein complex from the evolutionary perspective. Thus this study by Wideman will be helpful in further understanding of the biology of the EMC proteins. On its scientific merit I consider this article to be substantially important for getting published with F1000Research.
However there is one major concern, which I'd like to be addressed before endorsing the acceptance of this article. The author has described Emc9 to be present only among the vertebrates. However the