The structure of the teleost Immunoglobulin M core provides insights on polymeric antibody evolution, assembly, and function

Polymeric (p) immunoglobulins (Igs) serve broad functions during vertebrate immune responses. Typically, pIgs contain between two and six Ig monomers, each with two antigen binding fragments and one fragment crystallization (Fc). In addition, many pIgs assemble with a joining-chain (JC); however, the number of monomers and potential to include JC vary with species and heavy chain class. Here, we report the cryo-electron microscopy structure of IgM from a teleost (t) species, which does not encode JC. The structure reveals four tIgM Fcs linked through eight C-terminal tailpieces (Tps), which adopt a single β-sandwich-like domain (Tp assembly) located between two Fcs. Specifically, two of eight heavy chains fold uniquely, resulting in a structure distinct from mammalian IgM, which typically contains five IgM monomers, one JC and a centrally-located Tp assembly. Together with mutational analysis, structural data indicate that pIgs have evolved a range of assembly mechanisms and structures, each likely to support unique antibody effector functions.


Reviewer #2 (Remarks to the Author):
In this manuscript, the authors report the cryo-electron microscopy structure of the immunoglobulin M from a teleost species (tIgM).The structural characteristics of tIgM were analyzed in detail.The tIgM structure is terameric and has no JC domain, which is quite different from that of human IgM.Together with comparative and mutational analyses, the tIgM structure has provided the first glimpse of both a teleost pIg and a pIg lacking a JC, revealing distinct modes of assembly compared to mammalian pIgs.Overall, this manuscript is well written, the experiments are well designed, and the quality of the data is high.The results would provide new insights into evolution and diversity of vertebrate pIg structure and function.But this manuscript still has some issues that need to be revised as listed below.
What does the lane L in Supplementary Figure 1 panel b mean?It should be the meaning of maker?
2. Line 107: Despite C2 biological symmetry…, what is the C2 mean?And "an axis running between between Fcμ2…" this sentence repeats a word "between".
3. Supplementary figures in the article are not cited in order.For example: Supplementary Figure 4 (line 227) should be cited first, then Supplementary Figure 5. 4. The residues Lys417 mentioned on line 149 is not shown in Figure 3-b or -c? 5. Lines 183-184: the cited figures are not only Supplementary Figure 5b, but also Supplementary Figure 5d and 5e; Line 190: Supplementary Figure 5c should be changed to 5f.Figures 5a,5b,5c,5d (Line 217,221,225,and 232).7. Line 260: Supplementary Figure 5c should be changed to 5f.

Figures should be cited in the order of
8. We know that the constant domain CH2 and Fabs of human IgM are flexible and poorly resolved in cryo-EM maps.During the construction of tIgM, whether do the authors try to construct the full-length tIgM (adding CH2 domain or Fabs)? 9. Line 510: it would be more appropriate to change the title to "Automic model building, refinement, and validation".

Reviewer #3 (Remarks to the Author):
Oligomerization of some immunoglobulins is thought to play a role in a wide range of functions.Despite their prevalence in vertebrates little is known about oligomerization processes, their evolution or the structure of those that lack a joining chain.
In this manuscript Lyu and co-workers present the structure of polymeric IgM from trout, which does not encode a joining chain.The structure shows that despite the fact that the eight heavy chains which form a tetramer of dimeric-IgM molecules have identical protein sequences, they adopt different folds.This results in a structure which is very distinct from the previously solved mammalian IgM oligomer structure.The authors then undertaken mutational analysis and analyze fusion proteins to provide data to support discussion on evolution and assembly.
The manuscript is generally of a very high level and contains exciting and interesting new data.The work supports the conclusions and the methodology is both sound and provide in sufficient detail.
However, there are some points which should be addressed.Specifically: 1) The authors refer to "electrostatics interactions" in several places e.g.lines 124, 149, when there are no charged residues involved.
2) It may be useful to have additional information e.g.RMSD as a supplementary table on the three distinct conformations of the six B-type domains.Linked to this in lines 217-219 it would be useful to know with which chains the RMSD was made and whether 1.27Å vs 1.18Å is really significantly different once the differences between the "three distinct conformations" of the B-type domains is considered?
3) Supplementary figure 1b provides supporting evidence that an inter-chain disulfide bond is formed, but this is not cited in the section relating to this possible disulfide.Adding a similar N/R SDS-PAGE gel for the C445S monomer would also support the arguments made.
4) The authors are in places a little loose with referring to their own figures.For example, supplementary figure 1 should be cited on line 92 not line 94; on lines 24-25 neither Fig 2a of supplementary figures 3a/3b show an interface "rich in electrostatic interactions"; 5d is an inappropriate citation on line 138; it should be 5h and not 5a on line 253 and 5f not 5c on line 260 -also 5c-d should possibly be cited around here in the text.
5) It may be useful for it to be reiterated to the reader around line 273 that the "domains differed, structurally but not in sequence, in the FG loop, G strand and pre-Tp loop regions".
Linked to this the section lines 282-313 could be potentially misleading to the casual reader as it has the implication that the sequences from tFcu A-type and B-type structures are different.I would suggest that the section (and the corresponding part of the discussion) be slightly reworded to make it clear this is not the case.This may also be supported by an alignment from say the start of strand F to the end of the protein for the wild-type and chimeric proteins used in Figure 6 (this could be a supplementary figure).This information is already included in Figure 5d, but the alignment is too small to read and requires the reader to work out where the boundaries are in the aligned sequences between the different chimeras.

Minor comments/suggestions:
Line 49, should it be "basolateral" rather than "basal lateral"?line 66, insert "a" to make "which has a distinct".Line 179, insert "bonds" to make "potential disulfide bonds between".Line 193 opening sentence would benefit from a reference.Lines 545-547, it may be useful to the reader to add common names of species as used in supplementary figure 4 to the names currently used in the text e.g."Oncorhynchus mykiss (trout)".
While readers could construct them themselves it may be useful to have all of the protein sequences in the supplementary material as well as an alignment of the native and chimeric proteins used in figure 6.
The font size of the proteins sequences in Figure 5d and supplementary figure 4 would greatly benefit from being a larger, readable size.
Supplementary figure 1, the legend talks about an N-terminal signal sequence, but this is absent from the figure.
Supplementary figure 1, the legend talks about "incomplete reduction" for the reduced sample but it does not say anything about the high mw bands in the N lane nor that the gel strongly implies that the non-reduced SDS-treated sample is a disulfide linked dimer.I would suggest that the later get added and that either the extra bands in both lanes get mentioned or that the extra bands in both lanes do not get mentioned (not the current mix). 1

Responses to reviewer #1 comments:
Reviewer comment: This work is quite interesting, revealing structural patterns of multimeric immunoglobulins (Ig) throughout evolution.These data in the bony fish are certainly useful as a jumping off point for other studies of isotypes in jawed vertebrates.The data are convincing and interesting.There is a major omission in this work, however.The description of evolution of multimeric Igs on pp 3, 4, 12, 13, 19, 20, 22 are incomplete; this detracts from the work in a comparative sense, and readers will get the wrong impression of what is known in the field.
Response: We thank the reviewer for the positive feedback and construct comments, which we have incorporated into the revised manuscript.Specifically, we have endeavored to broaden discussion and reference to the evolution of multimeric Igs.