Rpn11-mediated ubiquitin processing in an ancestral archaeal ubiquitination system

While protein ubiquitination was long believed to be a truly eukaryotic feature, recently sequenced genomes revealed complete ubiquitin (Ub) modification operons in archaea. Here, we present the structural and mechanistic characterization of an archaeal Rpn11 deubiquitinase from Caldiarchaeum subterraneum, CsRpn11, and its role in the processing of CsUb precursor and ubiquitinated proteins. CsRpn11 activity is affected by the catalytic metal ion type, small molecule inhibitors, sequence characteristics at the cleavage site, and the folding state of CsUb-conjugated proteins. Comparison of CsRpn11 and CsRpn11–CsUb crystal structures reveals a crucial conformational switch in the CsRpn11 Ins-1 site, which positions CsUb for catalysis. The presence of this transition in a primordial soluble Rpn11 thus predates the evolution of eukaryotic Rpn11 immobilized in the proteasomal lid. Complementing phylogenetic studies, which designate CsRpn11 and CsUb as close homologs of the respective eukaryotic proteins, our results provide experimental support for an archaeal origin of protein ubiquitination.

3. Page6 "is found distant from classical Ub and more closely related to prokaryotic SAMPs like MoaD and ThiS." Comment: Incorrect. No archaeal SAMPs like bacterial ThiS and archaeal SAMP1 is bi-functional. First, no archaeal SAMPs like bacterial ThiS of thiamine biosynthesis. Archaea lack steps of bacterial ThiF-ThiS-ThiG-ThiH pathway of synthesizing the thiazole ring, instead archaea apply eukaryote-like Thi4 mechanism to mobilize sulfur to form the thiazole ring (Hwang et al., 2017). Second, archaeal SAMP1 could be isopepetide-linked to hundreds of protein substrates, like eukaryotic ubiquitin, however, bacterial MoaD doesn't have this feature. Archaeal SAMP1 is bifunctional, on one hand like eukaryotic ubiquitin for diverse protein substrates modification 4. Page7 "CsUb is an extremely thermostable monomeric protein (Tm > 95°C in circular dichroism heat-denaturation studies, not shown)." Comment: "not shown" is not ok. Other places in the manuscript also have this "data not shown" problems.
5. Page7 "In vivo, the protein is made as a precursor of 87 amino acids (Ub-pre), from which the last 9 residues are removed upon maturation. " Comment: No data support. As you mentioned C. subterraneum could not be cultured so far, did you mean you co-express CsUb-pre (Csub_C1474) and CsRpn11 (Csub_C1473) in E. coli and find last 9 residues of CsUb-pre is removed by Csub_C1473 in E. coli as the in vivo assay? If so, show the data.
6. Page7 "CsUb also superimposes best with human Ub (DALI Z-Score 12.2), but is more distinct from so far characterized prokaryotic Ub-like proteins, such as H. volcanii MoaD/SAMP1 (Z = 6.9) or ThiS/SAMP2 (Z = 5.4)." Comment: Don't agree. H. volcanii SAMP2 is not ThiS-like and reasons were mentioned above. SAMP2 is eukaryotic ubiquitin-like, based on both functional and structural similarities, for example the following reasons: First, SAMP2 could be isopepetide-linked to hundreds of protein substrates, like eukaryotic ubiquitin (Humbard et al., 2010). Second, SAMP2 isopeptide modified protein substrate (e.g., TBP2) could be degraded in eukaryotic-like ubiquitin-proteasome pathway, in which archaeal SAMP2 (resemble eukaryotic ubiquitin), JAMM2 (resemble eukaryotic Rpn11), AAA ATPase PAN2, Cdc48c (resemble eukaryotic Rpt) and 20S proteasomes (resemble eukaryotic 20S proteasomes) were participated (Fu et al., 2016). Third, SAMP2 could apply a structure conformation (when in complex with JAMM1) extremely similar to eukaryotic ubiquitin and is recognized by archaeal JAMM1 in a similar way as ubiquitin is recognized by eukaryotic JAMM (Cao et al., 2017). How does the structure of Ub compare to CsUb vs. PfSAMP2 in complex with JAMM1? Did you compare the similarity of Ub vs. CsUb to Ub vs. PfSAMP2 in complex with JAMM1? If they don't show significant difference, then the statement "CsUb also superimposes best with human Ub but is more distinct from SAMP1 or SAMP2" is incorrect. 7. Page8 "The physiological relevance of these dimers is unclear, considering that for all determined JAMM structures the dimerization interfaces differ ( Figure S2 8. Page8 "In structure superimpositions, CsRpn11 displays the highest similarity to S. cerevisiae Rpn11 (DALI Z-Score 21.1), but is less similar to so far determined structures of other prokaryotic JAMMs, such as P. furiosus JAMM-120 (Z = 15.1)." Comment: How about "CsRpn11 (Csub_C1473)" similarities to eukaryotic CSN5, BRCC36 and AMSH/AMSH-LP? It is necessary to show. For example, if Csub_C1473 is also very similar to AMSH or even more similar to AMSH, is that indicate Csub_C1473 is AMSH-like rather than Rpn11-like. I do have some concerns, especially when consider: First, Csub_C1473 is active independently without requirement to form JAMM+/JAMM-heterodimer or larger complex, which is more like AMSH rather than Rpn11. Second, Asgard archaea do have AMSH-related ESCRT homologs (Zaremba-Niedzwiedzka et al., 2017). 9. Page 11 "Our results show that the Ub binding mode of AMSH, Rpn11 and CSN5 was already established in an archaeal ancestor" Comment: Inappropriate. Ub-like SAMP2 binding mode of archaeal JAMM1 is extremely similar to Ub binding mode of eukaryotic JAMM and has been shown in previous study (Cao et al., 2017).
10. Page12 "In fact, homologs of eukaryotic DUBs have not been found yet in archaea." Comment: Incorrect. Did you mean non-JAMM metalloprotease type DUBs (for example, cysteine protease type DUBs)? if so, write it clearly. 11. Page12 "Likewise, neither linearly linked human Ub (HsUb2), nor Lys48-or Lys63-isopeptide linked HsUb2 were suitable substrates ( Figure 4E)." Comment: This evidence appears opposing your statement that Csub_C1474 is CsUb-pre and Csub_C1473 is CsRpn11. It supports Csub_C1474 is another type/group of archaeal Ub-like proteins and Csub_C1473 is Csub_C1474-specific protease. In contrast, archaeal JAMM1 is more like Rpn11 in its ability to cleave true Ub-Ub dimers (Lys48 or Lys63 linkage) (Cao et al., 2017), compared to the archaeal Csub_C1473 ("CsRpn11").
12. Page 12-13 "At 45°C, truncated CsRpn11 processed CsUb-pre with a KM of 24.2 μM and a kcat of 3.5/min and bound the reaction product with a KD of 14.6 μM. With a KM of 31.6 μM and a kcat of 4.0/min, the full-length version displayed similar reaction parameters, indicating that the CsRpn11 C-terminal helices are indeed not critically required for catalysis. " Comment: Why perform at 45 °C? Is it optimized temperature for CsRpn11 activity or close to physiological temperature that C. subterraneum grow/live? If neither of them, it is hard to know whether "C-terminal helices" impact catalysis or not. For example, if 45°C is far away from optimized temperature of CsRpn11 activity, the difference of kinetics behavior of CsRpn11 (full length) and CsRpn11 (without C-terminal helices) may not be significant, but if tested at optimized temperature, the much higher difference may occur.
13. Page14 "Though to our knowledge metal ion specificity has not been tested for other JAMM proteases," Comment: Incorrect. Metal ion specificity was tested in HvJAMM1 before (Hepowit et al., 2012).
17. Page 17 "While Ub conjugation was long thought to be restricted to eukaryotes and to have evolved from distant prokaryotic homologs, such as the ThiS or MoaD conjugation system." Comment: Incorrect. ThiS or MoaD systems are not conjugation systems, but SAMP and ubiquitin systems are. ThiS and MoaD system are systems of thiamine biosynthesis and MoCo biosynthesis, respectively.
18. Page18 "bioinformatic discovery of complete Ub modification systems in C. subterraneum, and more recently in Asgard archaea, suggests that the fundamental function of this system, protein tagging, was already established in archaea. The first true archaeal Ub modification systems, however, co-existed side by side with the universal archaeal SAMP/JAMM system," Comment: This statement indicates "archaeal CsUb-pre (Csub_C1474)/CsRpn11 (Csub_C1473)" and "archaeal SAMP/JAMM system" are two different systems, which I don't agree. To me, the results of this study indicate Csub_C1474 is another type of ubiquitin-like protein in archaea (for example SAMP4), besides SAMP1, SAMP2 and SAMP3. Csub_C1473 is another type (JAMM-3 type) of JAMM proteases in archaea, besides JAMM1 and JAMM2. This statement indicates "CsUb-pre (Csub_C1474)/CsRpn11 (Csub_C1473)" system is first true archaeal Ub modification systems, which I don't agree. Asgard archaea and C. subterraneum have eukaryotic E3-like component (RING-finger protein) is a main point that the authors think "CsUbpre (Csub_C1474)" system is first true archaeal Ub modification systems. However, SAMP system also has eukaryotic E3-like component, for 19. Page19 "one can ask how an archaeal system could possibly operate with just a few components, with for example only two small E3 RING finger proteins in C. subterraneum, compared to hundreds of sophisticated E3 enzymes in eukaryotes" Comment: Misleading, need to modify this statement. First, RING finger proteins in eukaryotes are also limited (e.g., Rbx1, Rbx2), instead of hundreds. Substrate receptors of E3 in eukaryotes are "hundreds" or even more. A lot of substrate receptors of E3 in acrhaea including C. subterraneum may also exist. Second, the claim "one can ask how an archaeal system could possibly operate with just a few components" is based on limited in vitro evidence, since the organism C. subterraneum can not be cultivated so far. 20. Figure 1 and Figure 1 legends Comment: Need some modification based on reasons mentioned in previous comments. For one example, "Archaeal MoaD" and "Archaeal ThiS" are inappropriate and misleading.

Figure 4
Comment: The Mass results appear wrong. Figure 4A show CsUb-pre 12.462kDa and CsUb 11.344 kDa. However, based on my calculation Csub_C1474 (CsUb-pre) of 87 amino acid MW: 9.11 kDa; CsUb of 78 amino acid MW: 8.15 kDa. If you have some specially treatment to "Csub_C1474 (CsUb-pre)", please write clearly.
In addition, no standard (e.g., CsUb or CsUb-pre with known amount) was loaded on the same gel, how could you quantify gel bands of CsUb or CsUb-pre for kinetics calculation? 22. Figure 6 Comment: Control needed (e.g, "EDTA-treated Rpn11 with substrate") to make sure the "CsUb" band is due to "Csub_C1473 (CsRpn11)" cleavage of substrate. As I concern: during incubation, a little portion of substrate may be reduced and CsUb band occur.
Reviewer #2 (Remarks to the Author): The authors provide a biochemical and structural analysis of an archeal JAMM-type deubiquitinating enzyme (csRpn11), which is remarkably similar to its eukaryotic Rpn11 counterpart. The different csRPN1 structures described in this manuscript -with and without bound archeal ubiquitin -are representatives of the different catalytic steps of the DUB reaction. This manuscript is very interesting for the evolution of the ubiquitin pathway.
The work described is solid and well-executed; the representation and discussion of the results are clear and leave little to be desired. However, there are a few minor points that could be improved on.
-page 7, lower part: it is not correct to claim that eukaryotic ubiquitin has 'equivalent Lys residues' compared to csUb. As the authors correctly state later, the lysines are found at different positions, which is hardly reconcilable with their 'equivalence'.
-page 8, bottom line: what is meant by the term 'circular JAMM protease core' ? I can't see anything circular here.
-page 9ff, in connection to figure 3: The authors describe several structural features (helices, insert regions), which are not indicated in the structure. In general, I think that structure 3B should be rendered bigger and should contain labels for the insert regions and other features mentioned in the text.
-page 11, top, explains the contribution of particular residues by refering to the yeast ScRpn11 sequence. However, the alignment figure 3A shows only archeal and human Rpn11 sequences, making it very hard to appreciate discussions of "Gly-66 in ScRpn11". The authors should either use the residue numbering of human Rpn11 or alternatively include the yeast sequence in figure  3A.
-page 13, the authors write "..the linear substrate Ub-TAMRA.." I don't see why Ub-TAMRA should be considered a "linear substrate". The manuscript does not describe the source of the the Ub-TAMRA used by the authors, but the commercially available Ub-TAMRA has the Ub-cterminus bound to the epsilon-amino group of a lysine residue (which in turn is coupled to TAMRA). Thus, Ub-TAMRA emulates an isopeptide linkage rather than the peptide linkage found in linear chains. The authors should clarify this.
- Figure 4e: I wonder why in this gel, the linear HsUb2 runs at a much higher MW than the K48and K63-linked diUb species. In our hands, all three human diUb species run (almost) at the same height. Since there is no MW ladder provided, it is not clear how big the apparent MW difference really is. Also, I did not find a source for the used diUb species in the materials section. This should be rectified.
We would like to thank both reviewers for their comments and constructive criticism. All points that were raised are addressed in the revised manuscript and explained in detail below.
subterraneum. New insights into biochemical, structural and substrate recognition features of Csub_C1473 are shown. One main point that this study claim is Csub_C1473 is archaeal Rpn11, however, I disagree. Reasons as follows: First of all, the key feature/biological role of eukaryotic Rpn11 is its association with AAA ATPase and 20s core particle for ubiquitinmediated protein degradation. This feature discriminates Rpn11 from other eukaryotic JAMM proteases, such as CSN5, AMSH, BRCC36. If without this feature, it is hard to think Csub_C1473 is Rpn11-like, rather than AMSH-like or CSN5-like or BRCC36-like?Second, some of the evidence provided in this study doesn't support or even oppose Csub_C1473 is Rpn11-like. For example, Csub_C1473 is active independently and behaves as a monomer in solution, which makes Csub_C1473 resemble AMSH rather than Rpn11 (require association with protein partners for activity). For another example, Csub_C1473 ("CsRpn11") process Csub_C1474 ("CsUb-pre") into mature form, but authors also mention other eukaryotic DUBs rather than Rpn11 do this Ub-pre processing job. In addition, Csub_C1473 has substrate discrimination feature and does not cleave eukaryotic Ub2 (M1, K48 and K63 linkage), this appears indicating Csub_C1474/Csub_C1473 system separates from Ub/Rpn11 system.To me, results in this study indicate Csub_C1473 is a novel JAMM-type protease in Archaea (for example, archaeal JAMM-3 type in Fig.1) and Csub_C1474 is another type of Ub-like protein in Archaea (for example, another type of SAMP). You can only draw conclusions based on what the results show, and further claim that Csub_C1473 is archaeal Rpn11 and Csub_C1474 is archaeal ubiquitin is not right. Treating "Csub_C1473/Csub_C1474" system as first true archaeal Ub modification system is inappropriate, especially when we consider archaeal SAMP system contains components with both sequence and function similarities to eukaryotic Ub, E1, E3, JAMM metalloprotease, AAA ATPase and 20S core particle ( As is the case for the eukaryotic ubiquitination system, also the Caldiarchaeum ubiquitination system did not arise in a single step de novo. Obviously, some of its components' features can be traced to other members of their respective protein families, e.g. SAMPs and JAMM proteases ( Fig. 1). In that sense, these protein groups are certainly of great interest in an evolutionary context too. In addition, they are of course pivotal for archaeal physiology, as is evident from the numerous findings by Maupin-Furlow and colleagues, cited in our paper and by this reviewer.
1. Page6 "Remarkably, while prokaryotic JAMMs which cluster distant from Rpn11 are functionally associated with just a SAMP-or Ub-like protein and an E1-like homolog," Comment: Inaccurate. Not just a SAMP-or Ub-like protein, at least three Ub-like proteins reported (e.g., SAMP1, SAMP2 and SAMP3) and related with archaeal JAMMs.
With 'a SAMP-or Ub-like protein' we did not mean any number, but were referring to the type of protein as such. To clarify this, we have modified the sentence. We have altered the sentence on p6 and now compare the phylogenetic features of Urm1 to prokaryotic 'homologs' like MoaD and ThiS, which should be a more suitable term.

Page7 "CsUb is an extremely thermostable monomeric protein (Tm > 95°C in circular
dichroism heat-denaturation studies, not shown)."Comment: "not shown" is not ok.
Other places in the manuscript also have this "data not shown" problems. Supplementary Figure 1. There is no other 'data not shown' in the revised manuscript. 5. Page7 "In vivo, the protein is made as a precursor of 87 amino acids (Ub-pre), from which the last 9 residues are removed upon maturation. "Comment: No data support.

The CD melting curve for CsUb is now shown in
As you mentioned C. subterraneum could not be cultured so far, did you mean you coexpress CsUb-pre (Csub_C1474) and CsRpn11 (Csub_C1473) in E. coli and find last 9 residues of CsUb-pre is removed by Csub_C1473 in E. coli as the in vivo assay? If so, show the data.
The altered sentence on p7 does not make any reference to in vivo studies anymore. 6. Page7 "CsUb also superimposes best with human Ub (DALI Z-Score 12.2), but is more distinct from so far characterized prokaryotic Ub-like proteins, such as H. best with human Ub but is more distinct from SAMP1 or SAMP2" is incorrect.   We have changed the sentence accordingly (p11). The sentence has been modified (p12).

Our statement is based on the DALI Z-scores, in which a higher Z-score implicates
11. Page12 "Likewise, neither linearly linked human Ub (HsUb2), nor Lys48-or Lys63isopeptide linked HsUb2 were suitable substrates ( Figure 4E)."Comment: This evidence appears opposing your statement that Csub_C1474 is CsUb-pre and Csub_C1473 is CsRpn11. It supports Csub_C1474 is another type/group of archaeal Ub-like proteins and Csub_C1473 is Csub_C1474-specific protease. In contrast, archaeal JAMM1 is more like Rpn11 in its ability to cleave true Ub-Ub dimers (Lys48 or Lys63 linkage) (Cao et al., 2017), compared to the archaeal Csub_C1473 ("CsRpn11"). Fig. 3), it is conceivable that also non-native substrates are being cleaved. for CsRpn11 activity or close to physiological temperature that C. subterraneum grow/live? If neither of them, it is hard to know whether "C-terminal helices" impact catalysis or not. For example, if 45°C is far away from optimized temperature of CsRpn11 activity, the difference of kinetics behavior of CsRpn11 (full length) and

If the features of a substrate sequence happen to allow an interaction with the protease (see interaction sites in
CsRpn11 (without C-terminal helices) may not be significant, but if tested at optimized temperature, the much higher difference may occur.
CsRpn11 performs very well at 45°C, but can be assayed up to 60°C, which would be in the physiological growth range of Caldiarchaeum. The reaction kinetics of CsRpn11 and CsRpn11 ∆149-202 at 60°C are now shown in Supplementary Fig. 9. As expected, both proteins are similarly active, confirming that the C-terminal helices, which are absent in other JAMM proteases, do not influence CsRpn11 activity, neither at 45°C nor at 60°C.
13. Page14 "Though to our knowledge metal ion specificity has not been tested for other JAMM proteases," Comment: Incorrect. Metal ion specificity was tested in HvJAMM1 before (Hepowit et al., 2012).

We have missed this, presumably because only Zn 2+ was reported to result in an active
JAMM protease. The statement on p14 has been modified accordingly.
The inhibitors thiolutin and 8-TQ were the topic of recent high-profile publications, as cited in our paper, with the authors' attention focusing on Rpn11. As only eukaryotic proteins were analysed so far, we were curious to see possible effects on prokaryotic proteases and tested CsRpn11, as it is the subject of this paper. It was not our intention to use inhibitor binding as a way to establish phylogenetic relationships. We agree with the reviewer that this is not feasible, and testing various proteins with various inhibitors would not provide useful information in that regard.
The binding of thiolutin and 8-TQ to other archaeal JAMM proteases is of course a distinct possibility. We followed the reviewer's suggestion and exposed CsJAMM to the inhibitors. Indeed, both thiolutin and 8-TQ affect its proteolytic activity, albeit the effects on CsJAMM are significantly weaker than the ones seen on CsRpn11 (Supplementary Fig. 8B, Fig. 5C). We have changed the sentence accordingly.

While this is in line with
18. Page18 "bioinformatic discovery of complete Ub modification systems in C.
subterraneum, and more recently in Asgard archaea, suggests that the fundamental function of this system, protein tagging, was already established in archaea. The first true archaeal Ub modification systems, however, co-existed side by side with the universal archaeal SAMP/JAMM system,"Comment: This statement indicates "archaeal CsUb-pre (Csub_C1474)/CsRpn11 (Csub_C1473)" and "archaeal SAMP/JAMM system" are two different systems, which I don't agree. To me, the results of this study indicate Csub_C1474 is another type of ubiquitin-like protein in archaea (for example SAMP4), besides SAMP1, SAMP2 and SAMP3. Csub_C1473 is another type (JAMM-3 type) of JAMM proteases in archaea, besides JAMM1 and JAMM2.This statement indicates "CsUb-pre (Csub_C1474)/CsRpn11 (Csub_C1473)" system is first true archaeal Ub modification systems, which I don't agree. Asgard archaea and C. subterraneum have eukaryotic E3-like component (RING-finger protein) is a main point that the authors think "CsUb-pre (Csub_C1474)" system is first true archaeal Ub modification systems. However, SAMP system also has eukaryotic E3-like component, for example, archaeal  19. Page19 "one can ask how an archaeal system could possibly operate with just a few components, with for example only two small E3 RING finger proteins in C.
subterraneum, compared to hundreds of sophisticated E3 enzymes in eukaryotes" Comment: Misleading, need to modify this statement.First, RING finger proteins in eukaryotes are also limited (e.g., Rbx1, Rbx2), instead of hundreds. Substrate receptors of E3 in eukaryotes are "hundreds" or even more. A lot of substrate receptors of E3 in acrhaea including C. subterraneum may also exist.
Second, the claim "one can ask how an archaeal system could possibly operate with just a few components" is based on limited in vitro evidence, since the organism C. subterraneum can not be cultivated so far.   Figure 4A show CsUb-pre 12.462kDa and CsUb 11.344 kDa. However, based on my calculation Csub_C1474 (CsUbpre) of 87 amino acid MW: 9.11 kDa; CsUb of 78 amino acid MW: 8.15 kDa. If you have some specially treatment to "Csub_C1474 (CsUb-pre)", please write clearly.In addition, no standard (e.g., CsUb or CsUb-pre with known amount) was loaded on the same gel, how could you quantify gel bands of CsUb or CsUb-pre for kinetics calculation?
As described in Methods (p21), CsUb-pre was produced with an N-terminal His6-tag and a TEV protease cleavage site. These additional sequences account for the extra mass. In our assays, we observed no difference between tagged and untagged CsUb-pre. The figure legend has been modified for clarity.
For the kinetics calculation, it goes without saying that a standard was loaded on the same gel, which can now be seen in Supplementary Fig. 7A and B, where full gels are shown. The description in Methods has been modified to make this point clearer (p26). Figure 6Comment: Control needed (e.g, "EDTA-treated Rpn11 with substrate") to make sure the "CsUb" band is due to "Csub_C1473 (CsRpn11)" cleavage of substrate. As I concern: during incubation, a little portion of substrate may be reduced and CsUb band occur.

22.
3. page 9ff, in connection to figure 3: The authors describe several structural features (helices, insert regions), which are not indicated in the structure. In general, I think that structure 3B should be rendered bigger and should contain labels for the insert regions and other features mentioned in the text.
We very much appreciate this suggestion, and agree that the now enlarged Figure 3B, together with the labels, provides a much clearer picture of the proteins' features.
4. page 11, top, explains the contribution of particular residues by refering to the yeast ScRpn11 sequence. However, the alignment figure 3A shows only archeal and human Rpn11 sequences, making it very hard to appreciate discussions of "Gly-66 in ScRpn11".
The authors should either use the residue numbering of human Rpn11 or alternatively include the yeast sequence in figure 3A. 5. page 13, the authors write "..the linear substrate Ub-TAMRA.." I don't see why Ub-TAMRA should be considered a "linear substrate". The manuscript does not describe the source of the the Ub-TAMRA used by the authors, but the commercially available Ub-TAMRA has the Ub-cterminus bound to the epsilon-amino group of a lysine residue (which in turn is coupled to TAMRA). Thus, Ub-TAMRA emulates an isopeptide linkage rather than the peptide linkage found in linear chains. The authors should clarify this.
We have corrected the sentence describing Ub-TAMRA accordingly (p13).