The ER transmembrane complex (EMC) can functionally replace the Oxa1 insertase in mitochondria

TLDR: Observations indicate that protein insertion into the ER is functionally conserved to the insertion mechanism in bacteria and mitochondria and adheres to similar topological principles.

Abstract: Two multisubunit protein complexes for membrane protein insertion were recently identified in the endoplasmic reticulum (ER): The guided entry of tail anchor proteins (GET) complex and ER membrane complex (EMC). The structures of both of their hydrophobic core subunits, that are required for the insertion reaction, revealed an overall similarity to the YidC/Oxa1/Alb3 family members found in bacteria, mitochondria and chloroplasts. This suggests that these membrane insertion machineries all share a common ancestry. To test whether these ER proteins can functionally replace Oxa1 in yeast mitochondria, we generated strains that express mitochondria-targeted Get2-Get1 and Emc6-Emc3 fusion proteins in Oxa1 deletion mutants. Interestingly, the Emc6-Emc3 fusion was able to complement an{Delta} oxa1 mutant and restored its respiratory competence. The Emc6-Emc3 fusion promoted the insertion of the mitochondrially encoded protein Cox2 as well as of nuclear encoded inner membrane proteins though was not able to facilitate the assembly of the Atp9 ring. Our observations indicate that protein insertion into the ER is functionally conserved to the insertion mechanism in bacteria and mitochondria and adheres to similar topological principles.

Figures

Fig 2. Expression of a mitochondria-targeted Emc6-Emc3 fusion protein (mito-EMC) suppresses the growth defect of a oxa1 mutant. (A) Schematic representation of the fusion proteins used in this study. The segments shown in purple are derived from Oxa1 in order to ensure import and insertion of the proteins into the inner membrane. (B) Simple growth test on glycerol plates on which cells can only grow if they generate energy from respiration. (C) Cells were grown in galactose medium, from which tenfold serial dilutions were dropped onto plates with the respective carbon sources. An empty vector (ev) control is shown for comparison. (D) Growth curves of oxa1 cells with the indicated expression plasmids. Shown are mean values of three technical replicates (n=3).

Fig 4. While mito-EMC mediates the insertion of Cox2, it fails to take over the Oxa1facilitated assembly of the Atp9 oligomer. (A) oxa1 mutants expressing Oxa1, mito-GET or mito-EMC were grown in galactose medium to log phase. Cytosolic translation was inhibited by cycloheximide. 35S-methionine was added to radiolabel mitochondrial protein synthesis. Cells were harvested and lysed. Proteins were precipitated by trichloroacetic acid and visualized by SDS-PAGE and autoradiography. (B) Mitochondria were isolated from the indicated strains and incubated in the presence of 35Smethionine in translation buffer for 30 min at 30°C. Mitochondria were washed and newly synthesized proteins visualized by autoradiography. (C, D) Mitochondrial translation products were radiolabeled for 30 min before mitochondria were lysed with Triton X-100. The extract was either directly loaded (T, equivalent to 10% total) or after immunoprecipitation with antibodies raised against the C-terminus of Oxa1 (Oxa1C), subunits of the ATPase or with preimmune serum (PIS). Arrowheads depict immunoprecipitated ATPase subunits. (E) The levels of oligomycin-sensitive ATPase were measured in isolated mitochondria and quantified. Shown are mean values of three replicates. (F) While mito-EMC is able to mediate the membrane insertion of Cox2, it does not efficiently promote the assembly of Atp9 into its oligomeric form. Thus, mito-EMC seems only to be able to take over some of the Oxa1mediated functions in mitochondria.

Fig 1. The members of the Oxa1 superfamily are structurally similar, however, their phylogenetic relationship is unclear. (A) All shown proteins serve as membrane insertases. Whereas the representatives of bacteria, mitochondria and chloroplasts are monomeric proteins with five or six transmembrane domains, the insertases of the ER form oligomeric complexes of two (Get1/Get2) or multiple (EMC) subunits. Recent studies on the structure of the GET and EMC complexes proposed that Emc3 and Get1 are structurally related to YidC, Oxa1 and Alb3. The transmembrane domains that share structural similarity are indicated in darker color. (B) Blast searches identified 460 homologs of DUF106, Emc3, Get1, Alb3/Alb4, Oxa1 and YidC proteins from different species (see Fig. S1 and Supplementary Data Files 1-3). This tree supports the relatedness of these protein groups and is consistence with the hypothesis that Emc3 is related to a DUF106-like ancestor protein. While YidC/Oxa1/Alb3 proteins form a closely related group, the branches to Get1, Emc3 and DUF106 are much longer and, thus, their sequences are likely more deviated.

Fig 3. Protein insertion by mito-EMC, like that by Oxa1, depends on negative charges in the transferred sequence. (A) Schematic representation of the import and Oxa1-mediated membrane insertion of nuclear encoded proteins. The scissors indicate the proteolytic removal of the N-terminal mitochondrial targeting sequence by the matrix processing peptidase. Protein insertion depends on Oxa1 and the membrane potential (). (B, C) Radiolabeled Oxa1 and Su9(1-112)-DHFR were incubated with mitochondria isolated from the indicated mutants for 30 min at 25°C. Samples were divided into three fractions and treated with or without proteinase K (PK) under isosmotic or hypoosmotic (Swelling) conditions. Upon swelling, protease treatment generates a fragment (frag., white arrowhead) from the membrane-embedded protein, but not from the translocation intermediate that still resides in the matrix (black arrowhead). (D) Insertion efficiencies were quantified from three biological replicates and quantified. Mean values and standard deviations are shown.

Full text

The ER transmembrane complex (EMC) can functionally replace the Oxa1 insertase in
mitochondria
Büsra Güngör
1
, Tamara Flohr
1
, Sriram G. Garg
2
, Johannes M. Herrmann
1
*
1, Cell Biology, University of Kaiserslautern, Erwin-Schrödinger-Strasse 13, 67663
Kaiserslautern, Germany
2, Molecular Evolution, Heinrich-Heine-University of Düsseldorf, Universitätsstrasse 1,
40204 Düsseldorf, Germany
* to whom correspondence should be sent
Cell Biology, University of Kaiserslautern, Erwin-Schrödinger-Strasse 13, 67663
Kaiserslautern, Germany, +49 631 2052406, hannes.herrmann@biologie.uni-kl.de
Roles:
BG, TF: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software,
Visualization, Writing – review & editing
SG: Conceptualization, Data curation, Software, Writing – review & editing
JMH: Conceptualization, Investigation, Methodology, Resources Validation, Funding acquisition, Writing –
original draft
Short title: The EMC core can replace Oxa1 in mitochondria
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 2, 2021. ; https://doi.org/10.1101/2021.08.02.454725doi: bioRxiv preprint

Abstract
Two multisubunit protein complexes for membrane protein insertion were recently identified
in the endoplasmic reticulum (ER): The guided entry of tail anchor proteins (GET) complex
and ER membrane complex (EMC). The structures of both of their hydrophobic core subunits,
that are required for the insertion reaction, revealed an overall similarity to the
YidC/Oxa1/Alb3 family members found in bacteria, mitochondria and chloroplasts. This
suggests that these membrane insertion machineries all share a common ancestry. To test
whether these ER proteins can functionally replace Oxa1 in yeast mitochondria, we generated
strains that express mitochondria-targeted Get2-Get1 and Emc6-Emc3 fusion proteins in
Oxa1 deletion mutants. Interestingly, the Emc6-Emc3 fusion was able to complement an

oxa1 mutant and restored its respiratory competence. The Emc6-Emc3 fusion promoted the
insertion of the mitochondrially encoded protein Cox2 as well as of nuclear encoded inner
membrane proteins though was not able to facilitate the assembly of the Atp9 ring. Our
observations indicate that protein insertion into the ER is functionally conserved to the
insertion mechanism in bacteria and mitochondria and adheres to similar topological
principles.
Keywords
EMC complex / Evolution / Membrane Insertion / Mitochondria / Oxa1 / Topogenesis
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 2, 2021. ; https://doi.org/10.1101/2021.08.02.454725doi: bioRxiv preprint

Introduction
Membranes of bacteria and eukaryotic cells contain different protein translocases. These pore-
like structures transport unfolded polypeptides across membranes and, in case of membrane
proteins, laterally integrate them into the lipid bilayer [1]. Examples are the SecY/Sec61
complexes of the bacterial inner membrane and the endoplasmic reticulum (ER)[2, 3], the
beta barrel-structured outer membrane translocases of bacteria, mitochondria and chloroplasts
[4, 5], and the translocases of the mitochondrial inner membrane (TIM23 and TIM22
complexes) [6, 7]. These translocases belong to distinct non-related protein families and
developed independently during evolution.
Protein translocation can also be mediated by a second group of translocation machineries
which do not form defined pores but rather facilitate protein translocation by local distortion
and compression of lipid bilayers [8]. Such a mechanism was recently proposed for the ER-
associated degradation (ERAD) pathway machinery [9].
Locally distorted and compressed lipid bilayers are also used by insertases, which integrate
hydrophobic proteins into membranes. Substrates of these insertases include membrane
proteins that lack large hydrophilic regions on the trans-side of the membrane (such as in the
case of tail-anchored proteins) or multispanning membrane proteins whose more complex
topogenesis relies on the cooperation of insertases with a canonical protein translocase.
The mitochondrial protein Oxa1 was discovered in the early 90s as the first representative of
these insertases [10, 11] and served as the founding member of the YidC/Oxa1/Alb3 family.
These closely related and functionally exchangeable proteins [12-15] mediate membrane
insertion of proteins into the inner membranes of bacteria and mitochondria as well as in the
thylakoid membrane of chloroplasts [16-23]. Several YidC structures were published recently
which suggest that these monomeric proteins serve as enzymes that accelerate the
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 2, 2021. ; https://doi.org/10.1101/2021.08.02.454725doi: bioRxiv preprint

spontaneous (though often inefficient, slow and non-productive) partitioning of hydrophobic
segments into the lipid bilayer [24-27].
Two recently identified protein complexes serve as insertases in the ER membrane: The Get1-
Get2 (in vertebrates WRB-CAML) complex which facilitates the insertion of tail-anchored
proteins [28] and the ER membrane complex (EMC), a multimeric insertase consisting of
eight (yeast) or nine (animals) subunits which acts independently of as well as in conjunction
with the Sec61 translocon in the topogenesis of multispanning ER proteins [29-33]. The
structures of both complexes were recently solved [34-36], revealing a striking similarity of
the reaction centers formed by Get1-Get2 and Emc6-Emc3 with the architecture of
YidC/Oxa1/Alb3 proteins despite very limited sequence similarity. On the basis of their
structural architecture, it was proposed that all these insertases are members of one related
group of proteins, which was named the Oxa1 superfamily [37].
In this study, we report that mitochondrial Oxa1 protein can be functionally replaced by the
core components of the EMC complex, at least in respect to its role in the membrane insertion
of proteins. Unlike Oxa1, EMC is however unable to facilitate the assembly of the F
o
sector of
the ATPase, presumably because it is not recognized by mitochondrion-specific assembly
factors [38, 39]. Our study shows that the EMC complex of the ER and the Oxa1 insertase of
mitochondria fulfill analogous molecular functions, consistent with their proposed structural
similarity.
Results
The core components of the various membrane insertases display similar topology
despite limited sequence identity
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 2, 2021. ; https://doi.org/10.1101/2021.08.02.454725doi: bioRxiv preprint

The YidC, Alb3 and Oxa1 insertases of bacteria, chloroplasts and mitochondria are
characterized by five conserved transmembrane domains (Fig. 1A). The DUF106 protein
family of archaea was proposed to be a distant relative, although it only has three
transmembrane domains which show similarity to the transmembrane domains 1, 2 and 5 of
Oxa1. It was suggested that the DUF106 family gave rise to Emc3 and Get1 on the basis of
very similar overall structural organization [34, 35].
To assess a potential phylogenetic relationship among these proteins, we screened for
potential related proteins of Oxa1, Alb3, YidC, DUF106, Emc3 and Get1 and identified 460
unique homologs across eukaryotes and prokaryotes (see supplemental data set 1-3).
Phylogenetic trees were calculated based on trimmed alignments (Fig. 1B). These trees
supported the common origin of Oxa1, Alb3 and YidC very well and also indicated good
bootstrap support for their relationship with members of the DUF106, Get1 and Emc3
families (Fig S1, supplemental data set 1-3). Even though the similarity of individual proteins
is low, the inclusion of the large number of sequences allowed the construction of a well-
supported tree that supports the relatedness of these different groups of insertases. Even if
analogy based on convergent evolution cannot be formally excluded, homology based on
common ancestry appears more likely.
A mitochondria-targeted EMC core restores respiration of

oxa1 cells
The recently solved structures of the EMC and GET complexes [34-36] suggested that their
core centers, consisted of Emc6/Emc3 and Get2/Get1 respectively, resembling the structural
organization of YidC. This inspired us to clone the respective regions of Emc6/Emc3 and
Get2/Get1 into fusion proteins which also contained the matrix-targeting sequence (MTS) of
Oxa1 to ensure mitochondrial import of these proteins, the C-terminal ribosome binding
domain of Oxa1 necessary for its interaction with the mitochondrial translation machinery
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 2, 2021. ; https://doi.org/10.1101/2021.08.02.454725doi: bioRxiv preprint

Frequently asked questions

Q1. What contributions have the authors mentioned in the paper "The er transmembrane complex (emc) can functionally replace the oxa1 insertase in mitochondria" ?

In this paper, the authors showed that protein insertion into the ER is functionally conserved to the insertion mechanism in bacteria and mitochondria and adheres to similar topological principles. 

Q2. What is the name of the protein translocases?

Substrates of these insertases include membraneproteins that lack large hydrophilic regions on the trans-side of the membrane (such as in thecase of tail-anchored proteins) or multispanning membrane proteins whose more complextopogenesis relies on the cooperation of insertases with a canonical protein translocase. 

Q3. What is the role of the mito-EMC?

The authors observed that mito-EMC is able to mediate the insertion of the mitochondrial translationproducts which range from proteins of rather simple topology (Atp8 and Cox2) to multi-passproteins with several transmembrane domains (Cox1, Cox3, cytochrome b and Atp6). 

Q4. What is the role of the Sec61 translocon in the insertion of multispanning proteins?

Studies on reconstituted liposomes showed that while the EMC complex isable to integrate proteins of rather simple topology on its own [32], the assistance of theSec61 translocon is often necessary for accurate topogenesis of multispanning proteins [31]. 

Q5. What was the reaction used to remove the precursors?

The remaining precursors outside of the mitochondria were removed by protease treatment (PK) for 30 min. 2 mM PMSF was added to stop protein degradation. 

Q6. What is the role of the mito-EMC insertase?

(4) In addition to its role as an insertase for membrane proteins, Oxa1 was proposed tofacilitate the assembly of oligomeric complexes [54, 55]. 

Q7. How many ml per g of cell pellets were resuspended?

Pellets were resuspended in 2 ml per g wet weight in MP1 (100 mM Tris, 10 mM DTT), incubated for 10 min at 30°C and centrifuged again (5 min, 3000 g). 

Q8. How long did mitochondria remain incubated at 30°C?

Radiolabeling was stopped by addition of an excess of cold methionine, and mitochondria were further incubated (chase) at 30°C for 0, 10 or 30 min. 

Q9. What was the effect of mito-EMC on the insertion of mitochondria?

Radiolabeled precursor proteins of different model substrates wereincubated with these mitochondria to allow their import and intramitochondrial sorting. 

Q10. What grants were used to fund this project?

This project was funded by grants from the Deutsche Forschungsgemeinschaft (DIP MitoBalance and HE2803/9-1) and the Landesschwerpunkt BioComp. 

Q11. What is the name of the pore-like structures?

These pore-like structures transport unfolded polypeptides across membranes and, in case of membraneproteins, laterally integrate them into the lipid bilayer [1]. 

Q12. How many ml per g of zymolyase washed?

Pellets were washed with 1.2 M sorbitol and centrifuged (5 min, 3000 g) before resuspending in 6.7 ml per g wet weight in MP2 (20 mM KPi buffer pH 7.4, 1.2 M sorbitol, 3 mg per g wet weight zymolyase from Seikagaku Biobusiness) and incubated at 30°C for 60 min. 

Q13. What is the effect of mito-EMC on insertion efficiency?

Previousstudies have shown that the Oxa1-mediated insertion efficiency is dependent on the negativecharge in the inserted region and that positively charged regions are not exported by Oxa1[46]. 

Q14. What are the examples of membrane proteins?

Examples are the SecY/Sec61complexes of the bacterial inner membrane and the endoplasmic reticulum (ER)[2, 3], thebeta barrel-structured outer membrane translocases of bacteria, mitochondria and chloroplasts[4, 5], and the translocases of the mitochondrial inner membrane (TIM23 and TIM22complexes) [6, 7]. 

Q15. What is the mechanism of insertion in the mito-EMC core?

Thus,our in vivo reconstitution indicates that, in principle, the EMC core is able to mediate theinsertion of a broad range of membrane proteins in the absence of a Sec61-mediated insertionactivity. 

Q16. What is the name of the mitochondrial protein Oxa1?

The mitochondrial protein Oxa1 was discovered in the early 90s as the first representative ofthese insertases [10, 11] and served as the founding member of the YidC/Oxa1/Alb3 family. 

Q17. What is the effect of mito-EMC on mitochondria?

When mitochondria were lysed after the labelingreaction and Oxa1 and mito-EMC were isolated by immunoprecipitation with antibodiesagainst the C-terminus of Oxa1, subunits of the ATPase were only co-isolated with Oxa1 butnot with mito-EMC (Fig. 4C). 

Q18. What is the phylogenetic relationship between the DUF106 family and the Oxa?

The DUF106 proteinfamily of archaea was proposed to be a distant relative, although it only has threetransmembrane domains which show similarity to the transmembrane domains 1, 2 and 5 ofOxa1.