Structure of full-length cobalamin-dependent methionine synthase and cofactor loading captured in crystallo

Cobalamin-dependent methionine synthase (MS) is a key enzyme in methionine and folate one-carbon metabolism. MS is a large multi-domain protein capable of binding and activating three substrates: homocysteine, folate, and S-adenosylmethionine for methylation. Achieving three chemically distinct methylations necessitates significant domain rearrangements to facilitate substrate access to the cobalamin cofactor at the right time. The distinct conformations required for each reaction have eluded structural characterization as its inherently dynamic nature renders structural studies difficult. Here, we use a thermophilic MS homolog (tMS) as a functional MS model. Its exceptional stability enabled characterization of MS in the absence of cobalamin, marking the only studies of a cobalamin-binding protein in its apoenzyme state. More importantly, we report the high-resolution full-length MS structure, ending a multi-decade quest. We also capture cobalamin loading in crystallo, providing structural insights into holoenzyme formation. Our work paves the way for unraveling how MS orchestrates large-scale domain rearrangements crucial for achieving challenging chemistries.


Reviewers' Comments:
Reviewer #1: Remarks to the Author: The cobalamin dependent methionine synthase (MetH) is a complex multimodular protein that catalyzes the key step in the biosynthesis of methionine.The cobalamin cofactor is employed as a way-station for the methyl group that is derived from methyl tetrahydrofolate (MeTHF), on its way to Hcy to form Met. The cofactor cycles between the +1 and +3 oxidation states, as it undergoes methylation by MeTHF and demethylation by Hcy, respectively.The cofactor is periodically oxidized to the catalytically inactive +2 oxidation state.To return the protein to the active form, MetH from many sources also have an additional domain that binds S-adenosyl-L-methionine (SAM), which is used to reductively methylate the cofactor.MetH serves an important role in mammalian physiology.Mutations that are impair with various aspects of MetH activity are associated with various disease states.
The domains that bind various players in the activity of MetH have long been thought to be arranged as "beads on a string" with hinge regions that connect each of them.Pioneering studies by the labs of Rowena Matthews and the late Martha Ludwig established many of the basic structural and functional outlines of catalysis by this enzyme.Notably, individual domains were shown to be able to function independently, and when combined, recapitulate the overall activity of the protein, in trans.Structural studies on these various domains, starting with the first structure of a B12 binding domain solved in the early 1990's, have established the structural features of each of the active sites.However, the mechanism by which these domains interact with one another and the rules that dictate how each domain interacts with the cobalamin cofactor, in turn, have remained elusive.This paper by Koutmos, Yamada, and coworkers established some of the key structural details that have been missing in the MetH field.Notably, the authors show the X-ray crystal structure of the full apo-MetH, for the first time showing how the domains are arranged relative to one another.Notably, the SAM-binding domain is situated above the B12-binding Rossman fold, ready to accept and methylate B12.Using a smaller version of the protein they also show that the protein can be reconstituted with B12, and provide a snapshot of this initially reconstructed structure.These are very important insights.
MetH plays an important role in physiology and so even after decades of work on the protein, much remains to be understood, and 2023 is likely to be considered a watershed year in terms of structural details of MetH.This is the second paper this year on this topic, the first being a cryoEM/SAXS study form the Ando lab, which appeared in PNAS last month.Both studies employ a thermophilic homolog of the protein.While the overall impact of this work is somewhat dulled by the Ando work, this paper has several key strengths.First, the Ando study of the full length protein was at a lower resolution, and the SAM-binding domain was not well positioned in the structure.Second, the Ando study did not provide any insights into the cofactor delivery.Finally, the biochemical tools to express and purify MetH and its various domains that are established in this paper are likely to be transformative in additional studies.Therefore, the impact of the present study is significant.
Overall, the study is straightforward and presented in a logical manner.However, there are several areas where I think the manuscript can be improved.In general, I would recommend a careful reading of the entire manuscript and checking for consistency.I felt that different sections were likely written by different folks and there were some consistency issues.I apologize for not including page or line numbers.The printout on which I made all my changes cut them off.I am including enough context for each comment so that they can be located readily.
• Abstract, line 2 -"…and S-adenosyl-L-methionine" • Introduction, line 3 -cobalt • Introduction, 2nd paragraph -"…Co(I) state of the cofactor…" • Introduction and throughout the manuscript -please be consistent with the abbreviations.Use MeTHF for methyltetrahydrofolate, Met for methioniney, Hcy for homocysteine, etc.Also take a look at the figures and make sure that the abbreviations for the pieces of the protein (Cap, etc.) are consistent throughout.
• The following statement in the introduction is too cryptic and will be confusing to the reader: "…controlled by association and dissociation of different axial ligands to achieve each reaction…".It will be helpful here to expand this statement a little bit to catch all the nuances that are buried.
• The comment in the introduction "…along with the dearth of biochemical data, has raised further…".I think this is misleading.Yes, there isn't a log known about the cofactor loading, but quite a bit is known about chaperones, about the domains and their boundaries, etc. that this comment really does disservice to a lot of great biochemistry that has come out of Michigan in on this topic.
• "including homologs from…Escherichia coli..." • "…including around the catalytic His761.That I s found…".Please expand this a little bit and provide some context.This residue modulates reactivity of the cobalamin and is not strictly a catalytic residue.
• "…HB8 genome encodes a five-module…" • "Despite relatively high levels of sequence conservation…".What do you mean about relatively high?Be quantitative.
• The sections dealing with the biochemical data in Figure 2 are very confusing.Are these assays looking at reactions in trans, the way Goulding et al did them?What is really going on in the reactions?Why does the assay in Fig 2d plateau -is it because it is running out of substrate(s) or because the enzyme is getting inactivated?Are the various forms shown in Fig. 2c active in the expected reactions?Are they active in trans?
• …all factors that allow for robust biochemical and structural studies.tMS can be reconstituted with non-native cobalamins…".Please be specific here.I think that you are referring to the CN-Cbl reconstituted protein here as an example of non-native cobalamin.However, as the authors are aware there are a lot of other types of cobalamins that could have been placed here and I would bet that certain ones will not work.Perhaps spell out what you mean so as not to make a broad claim that is not supported by the data presented.
•. "..interacting with the ribotyl tail…".Technically, it is ribityl if the ribose was actually opened, as it is in flavin.In this case, it is the "ribosyl tail…" • In the Materials, check for consistency.I saw E. coli not italicized.The methyl tetrahydrofolate is already defined earlier -use the abbreviation.In the heading "Thermus thermoophilus MS" tMS is already defined -use it.
•Supporting Figure 3  • Supplementary Figure 5 -modify the title to "Acid-catalyzed conversion of tetrahydrofolate to methenyltetrahydrofolate". Should probably include a reference to the procedure here and in the methods.
• General comment on the PDB validation files -I am not a structural biologist but the validation files are clearly showing parameters that are in the "red zones" because of significant numbers of outliers.I appreciate the fact that MS is a large protein and refinement is probably a real pain.But, I think that the authors should either comment on these, or make an attempt to improve the statistics either by additional round of refinement, or perhaps processing at a lower resolution.
Reviewer #2: Remarks to the Author: This manuscript reports the isolation, characterization, and structure of cobalamin-dependent methionine synthase (MS) from Thermus thermophilus.MS performs three distinct methylation reactions on the substrates homocysteine, folate, and S-adenosylmethionine, and deficiencies in the human enzyme lead to pathologies and birth defects.Beyond its reactions, MS also undergoes periodic reactivation in which an inactive Co(II) species is reduced and methylated to reform the active cofactor.MS is a complex enzyme with five different domains, and the full length protein has been difficult to work with.By using the T. thermophilus homolog, the authors were able to crystallize and determine the structure of full-length MS in the apo form (without cobalamin), representing the first full-length structure and the first without cofactor.In addition, the structure of a three domain construct soaked with cyanocobalamin was determined.The structures reveal the overall arrangement of the domains, their interfaces with one another, and varying positions of key residue His791 and Tyr1132.The positions of these residues in the soaked structure suggest that the structure represents an intermediate on the way to reactivation.The work represents a step forward for the field and the T. thermophilus system is promising for future studies.1.The following statement in the introduction is too cryptic and will be confusing to the reader: "…controlled by association and dissociation of different axial ligands to achieve each reaction…".It will be helpful here to expand this statement a little bit to catch all the nuances that are buried.
We appreciate the reviewer's comments and therefore we have expanded the discussion of the role of the axial ligands of cobalamin to include specific examples for both the upper and lower axial ligands observed in MS.
2. The comment in the introduction "…along with the dearth of biochemical data, has raised further…".I think this is misleading.Yes, there isn't a log known about the cofactor loading, but quite a bit is known about chaperones, about the domains and their boundaries, etc. that this comment really does disservice to a lot of great biochemistry that has come out of Michigan in on this topic.
We have edited this sentence to emphasize the pioneering studies conducted in the Ludwig and Matthews' labs, including the addition of a more detailed background of the structural data obtained from said studies.We agree with the reviewer's comment and added that His761 can also serve as the lower-ligand in MS.We have however changed the text to provide succinct context with additional citations that discuss the multiple roles His761 plays.However, its potential role in acting as a key signaling residue is expanded on in greater detail in the discussion.An explicit mention that the full-length enzyme was used for all biochemical data shown has been included, and the Figure 2 legend has been revised to explicitly mention that the full-length enzyme was used in these assays.
The temperature-dependent assay in Fig. 2d does not show a plateau at the three lower temperatures (25 ˚C, 37 ˚C, and 50 ˚C), given that tMS displays optimal activity at 70 ˚C; at this temperature, a plateau is observed, likely due to substrate consumption.If oxidative inactivation was the root cause, a plateau would be expected for all temperatures tested, but this is not observed.
While we appreciate the reviewer's question regarding the truncated domains shown in Fig. 2c, and whether they are active in trans, this is beyond the scope of this paper.Though the excised domains are indeed active in trans, the main intention in 6. …all factors that allow for robust biochemical and structural studies.tMS can be reconstituted with non-native cobalamins…".Please be specific here.I think that you are referring to the CN-Cbl reconstituted protein here as an example of non-native cobalamin.However, as the authors are aware there are a lot of other types of cobalamins that could have been placed here and I would bet that certain ones will not work.Perhaps spell out what you mean so as not to make a broad claim that is not supported by the data presented.
While the reviewer makes a valid point on the wording of the sentence, and explicitly mentioning CN-Cbl only would be correct in the context of the paper, we were intentionally vague as to what non-native cobalamins could be loaded given that unpublished data has shown that tMS can bind other non-native cobalamins that have not previously been reported.As the reviewer noted,  127-138 (1997)).However, providing an extensive list of cobalamins that were found to bind to tMS is beyond the scope of this paper, though previously reported non-native cobalamins were added to narrow our initial statement.
7. General comment on the PDB validation files -I am not a structural biologist but the validation files are clearly showing parameters that are in the "red zones" because of significant numbers of outliers.I appreciate the fact that MS is a large protein and refinement is probably a real pain.But, I think that the authors should either comment on these, or make an attempt to improve the statistics either by additional round of refinement, or perhaps processing at a lower resolution.
We appreciate the reviewer's comments regarding the PDB validation files.The RSRZ outliers observed in the apo-Cap:Cob:Act structure are due to partial crystal twinning, the lower resolution data, and side-chain outliers: the residues and their sidechains are present at or near -Which PDB structures?Provide PDB accession numbers and references • Supporting Figure 4 -This figure is completely unnessary and inappropriate.This manuscript does not do anything with the human or E. coli proteins.So why is half of the figure showing those?In the second half, there are Xs where there should be numbers.With and without cofactor should be above the arrow not a distinct state.Etc."Obtain holo-MS with cobalamin bound" is redundant -holoMS is the cobalamin bound version.I recommend omitting this figure altogether.
65: the Cob domain appears to be orange, not red lines 89-90: it would be helpful to add a sentence summarizing what excised domain structures are available besides the study in ref. 38.This information is in SI Fig. 6, but would help here to introduce the state of the field.line 204: please explain what the distances in panel 4e are indicating.lines231-233: His761 appears to be on a loop and not on helix 1, which appears to maintain the same position in all the structures.What evidence indicates that the helix is more flexible when His761 is flipped out?Are the B-factors different?Are less residues involved in helical hydrogen bonds?line 261: " Compared to the apo-Cap:Cob:Act structure, the His761 imidazole side chain is ~6 Å from Co" -this comparison is not clear as there is no Co in the apo structure Why was the structure of the holo form not obtained without soaking?How do the authors know that the soaked structure accurately reflects what happens in solution?
3. "…including around the catalytic His761.That is found…".Please expand this a little bit and provide some context.This residue modulates reactivity of the cobalamin and is not strictly a catalytic residue.
4. "Despite relatively high levels of sequence conservation…".What do you mean about relatively high?Be quantitative.Specific values for the sequence identity of tMS relative to hMS and eMS have been added (33% and 34% identity respectively).5.The sections dealing with the biochemical data in Figure2are very confusing.Are these assays looking at reactions in trans, the way Goulding et al did them?What is really going on in the reactions?Why does the assay in Fig.2dplateau -is it because it is running out of substrate(s) or because the enzyme is getting inactivated?Are the various forms shown in Fig.2cactive in the expected reactions?Are they active in trans?
Fig. 2c is to show that tMS provides a Department of Chemistry, 930 N. University Ann Arbor MI 48109-1055 T: 734 647-1621 F: 734 647-4865 Office Rm: 3821 Please recycle robust biochemical model in which all excised domain constructs could be successfully expressed and purified.The figure legend has been edited to emphasize this point.
previous work has shown that eMS could bind n-propyl, thiocyanato-, azido-, chloro-, bromo-, and even CN-Cbl formed in situ (Hoover, D. M. et al.Interaction of Escherichia coli Cobalamin-Dependent Methionine Synthase and Its Physiological Partner Flavodoxin: Binding of Flavodoxin Leads to Axial Ligand Dissociation from the Cobalamin Cofactor.Biochemistry 36,