Lift-out cryo-FIBSEM and cryo-ET reveal the ultrastructural landscape of extracellular matrix

Zens et al. present a workflow to structurally characterize natively preserved extracellular matrix (ECM) using lift-out cryo-FIBSEM and cryo-ET. Employing cell-derived matrices to mimic authentic ECM, they reveal the intricate network of extracellular fibers in the context of matrix-secreting cells. Their findings expand the structural atlas of the ECM.

The relatively minor concerns and useful suggestions are numerous, but all appear to be readily addressable.Besides resolving the technical questions and wording/citation concerns, it would indeed be helpful and reassuring if possible to confirm that the filaments are indeed collagen fibers using second harmonic generation, though this approach is not absolutely essential.
We look forward to receiving a revised manuscript from you with a point-by-point response to each of the comments of these conscientious reviewers.We will re-review the manuscript at the Editor level to confirm that each of the points have been resolved within reason.We thank you for your interest in the Journal of Cell Biology and for this valuable contribution.
While you are revising your manuscript, please also attend to the following editorial points to help expedite the publication of your manuscript.Please direct any editorial questions to the journal office.

GENERAL GUIDELINES:
Text limits: Character count for a Report is < 20,000, not including spaces.Count includes title page, abstract, introduction, the joint Results & Discussion, and acknowledgments.Count does not include materials and methods, figure legends, references, tables, or supplemental legends.Figures: Reports may have up to 5 main text figures.To avoid delays in production, figures must be prepared according to the policies outlined in our Instructions to Authors, under Data Presentation, https://jcb.rupress.org/site/misc/ifora.xhtml.All figures in accepted manuscripts will be screened prior to publication.***IMPORTANT: It is JCB policy that if requested, original data images must be made available.Failure to provide original images upon request will result in unavoidable delays in publication.Please ensure that you have access to all original microscopy and blot data images before submitting your revision.***Supplemental information: There are strict limits on the allowable amount of supplemental data.Reports may have up to 3 supplemental figures.Up to 10 supplemental videos or flash animations are allowed.A summary of all supplemental material should appear at the end of the Materials and methods section.
Please note that JCB now requires authors to submit Source Data used to generate figures containing gels and Western blots with all revised manuscripts.This Source Data consists of fully uncropped and unprocessed images for each gel/blot displayed in the main and supplemental figures.If your paper will include cropped gel and/or blot images, please be sure to provide one Source Data file for each figure that contains gels and/or blots along with your revised manuscript files.File names for Source Data figures should be alphanumeric without any spaces or special characters (i.e., SourceDataF#, where F# refers to the associated main figure number or SourceDataFS# for those associated with Supplementary figures).The lanes of the gels/blots should be labeled as they are in the associated figure, the place where cropping was applied should be marked (with a box), and molecular weight/size standards should be labeled wherever possible.Source Data files will be made available to reviewers during evaluation of revised manuscripts and, if your paper is eventually published in JCB, the files will be directly linked to specific figures in the published article.
Source Data Figures should be provided as individual PDF files (one file per figure).Authors should endeavor to retain a minimum resolution of 300 dpi or pixels per inch.Please review our instructions for export from Photoshop, Illustrator, and PowerPoint here: https://rupress.org/jcb/pages/submission-guidelines#revised The typical timeframe for revisions is three to four months.While most universities and institutes have reopened labs and allowed researchers to begin working at nearly pre-pandemic levels, we at JCB realize that the lingering effects of the COVID-19 pandemic may still be impacting some aspects of your work, including the acquisition of equipment and reagents.Therefore, if you anticipate any difficulties in meeting this aforementioned revision time limit, please contact us and we can work with you to find an appropriate time frame for resubmission.Please note that papers are generally considered through only one revision cycle, so any revised manuscript will likely be either accepted or rejected.
When submitting the revision, please include a cover letter addressing the reviewers' comments point by point.Please also highlight all changes in the text of the manuscript.
We hope that the comments below will prove constructive as your work progresses.We would be happy to discuss them further once you've had a chance to consider the points raised in this letter.
Thank you for this interesting contribution to Journal of Cell Biology.You can contact us at the journal office with any questions at cellbio@rockefeller.edu.---------------------------------------------------------------------------Reviewer #1 (Comments to the Authors (Required)): I congratulate the authors on this heroic effort aiming to provide a first structural 3D atlas of the ECM at molecular resolution.The systematic effort to optimize specimen preparation of cell culture-based 3D-ECM platforms for compatibility with cryo-FIB lift-out and high quality cryo-ET provide a valuable starting point for studies focused on complex cell-ECM interactions that better recapitulate in vivo scenarios in a controlled 3D in vitro system.The quality of the data is outstanding and the accompanying annotations are stunning.The authors complement their structural characterization with systematic proteomics and light microscopy-based measurements.While the manuscript remains descriptive, the work in my view opens an exciting avenue for the study cell-matrix interfaces and illuminating critical gaps in our knowledge of ECM synthesis, secretion and assembly, and the effect of critical mutations.I recommend publication, and would have liked to see more quantification of the volume fractions of the different ECM structures identified.These might be useful in the future, especially if MS data is presented in parallel in terms of protein abundance, to possibly identify potential candidates for newly identified ECM structures (more below).

Minor comments:
The manuscript text is at times too wordy and therefore convoluted, especially in the introduction and in some of the results description.It may benefit from simplification and consolidation.
Abstract: "sample thinning by cryo-lift out focused ion beam milling".Lift out is an extraction procedure.Microfabrication and thinning are done by FIB.I recommend rephrasing for clarity to a broader audience."This allows us to visualize ECM for the first time in its hydrated, cellular context".The authors may want to reward "cellular" to "native" or similar.
Figure 1: I wonder whether the authors could comment on the origin of the observed directionality of both the extracellular and intracellular filaments.There is more information in the supplementary figures, but it is not explained or discussed.Does this evolve over time from cell seeding to the end point of the cryo-EM investigation?Is it directed by something in the EM grid substrate?Page 3: "Altogether, these results confirmed that CDMs harvested on or later than Day 14 represent bona fide ECM assemblies".It would be useful if the authors could point to previous publications on the expected morphology of "bona fide" ECM in support of their conclusion.
Page 4: with regard to the testing of the cryo-protectants, it would be useful if the authors could provide further information, and possibly references.For example, for Dextran, it is important to provide the molecular weight as this is an important factor for osmolarity and discuss evaluation of the effect of osmolarity on cell/extracellular native preservation in this context.
Page 4: how do the authors interpret the "empty areas" in the extracellular space?This is where a parallel presentation of the ECM proteomics could be useful.
Figure 2: how do the authors interpret the large vacuoles observed primarily in cells located towards the top of their lamellae?The authors may want to discuss this if appears to be a common feature.
Page 6: "The variation in both diameter and repeat pattern along the filament z-axis suggests extensibility and assembly variability."And "the amorphous matrix was observed only in close proximity to ECM fibers...This suggests a co-assembly between the components in these entities."These are unnecessary overinterpretations of the data.I recommend that the authors rephrase, especially when the structures cannot be identified.
Page 6, sections "The absence of other filament-like structures in CDMs": here it would be useful to explicitly spell out the candidates from proteomics measurements, or the previous reports the authors mention (without providing references!) and the expected structural features, potentially even from structure predictions.For example, Table 2 "ECM proteins in TIFF CDMs identified by MS" could perhaps be presented in a plot form that also reflects relative abundance (possibly only the fibril-forming proteins, while others are aggregated, for simplicity) in the main text and related to the related volume fractions occupied by the different ECM structures as observed by cryo-ET.
I recommend that the authors include in their discussion/conclusion some mention on the potential of "waffle" or "Serial lif-out" methodologies to gain higher throughput and more insights into similar model systems given the dimensions of their specimens.
Movies S4 and S4: some of the fuzzy amorphous density appears to be continuous with the Col-I fibrils.It might be useful for the authors to mention this if this appears to be a common feature in other data as it might indicate that these structures are precursors of Col-I assembly.Methods, cryo-FIB: unclear at which point the half-moon grid are mounted into autogrids.Please add.
Data availability: it would be extremely valuable for the community if the authors deposit the tomograms and segmentations associated with this work.I would even strongly recommend deposition of the full data, from raw movies to segmentations in EMPIAR.
Reviewer #2 (Comments to the Authors (Required)): In this article entitled "Unveiling the ultrastructural landscape of extracellular matrix via lift-out cryo-FIBSEM and cryo-ET", Zens et al. reporting a technical development based on cryoEM and FIBSEM technologies.They adapted and combined methodologies (references 24, 28, 31 and 36) to enable visualization of cell derived matrices in a hydrated context.This is of outmost relevance considering the deleterious effects of classic protocols including chemical fixation and dehydration steps.This article is well written and well-illustrated.It has an appreciable part of "trial and error" reporting of the scientific process, and results in establishment of a successful and promising protocol for the study and for future studies.This is the primary merit of the study and as such may become seminal in the field.Limitations of the manuscript are related to the relatively low number of conclusive points that are made by the authors.Indeed, the use of a single biological condition (TIFF derived CDM) is somewhat limiting the general value of the study.Having said that, owing to the laboarious nature of this approach, limiting efforts to a single sample time is understandable.

Major comments:
The authors should consider toning down this statement: "Another aggravating factor for the structural annotation of ECMs is the heterogeneity of tissue-derived ECM material.In contrast, cell derived matrices (CDMs) are a highly adaptable and versatile tool that is increasingly used to recapitulate the complexity of native tissue ECM (29-31)." I reality, CMD are not recapitulating the complexity of native tissue, which is the results of many cell types interacting together and jointly contributing to the matrix compositions and remodelling.Only in vivo observation can fully recapitulate in vivo complexity and CDM is rather a (valuable) reductionist in vitro model system to study matrix biology.
The authors are taking care not to overinterpret their findings and this is commendable.However, it would be beneficial for the reader to if some more concrete conclusions could be drawn.some conclusions out of this work.
Related to this topic, the "beaded filaments" and the absence of fibronectin visualization are suprising and need more explanation.Can the authors rule out the possibility that "amorphous ECM" is in fact not well-preserved during sample preparation?
Can the rare fibers that are observed in the amorphous ECM be a low percentage of preserved material?Can this be fibronectin filaments?Can the beaded filaments be fibronectin filaments?
One major limitation of the study is the lack of comparisons between for example condition or treatments that are expected to influence the CDM.For example, genetic manipulation of the TIFF or CDM made by another cell type.If these are beyond the scope of this study, the authors should discuss this point and mention this limitation.

Minor comments:
The authors abstract first paragraph is somewhat misleading, especially when mentioning Col VI assembly as remaining "enigmatic", considering that there is no conclusive statement made in the manuscript regarding Col VI.This study is an exquisite example of how new technology can facilitate novel information.The only areas of potential improvement are some missing citations to credit the original CDM work and methodology (and similar) as well as a suggestion to combine the Cryo-generated data with SHG (second harmonic generation of polarized light) using the same 3D sample.The latter is merely a suggestion that could significantly improve the interpretation of the data (regarding collagen fibers).

Specific MINOR suggestions:
1.In the introduction, when CDMs are first described as a valid in vivo mimicry approach, the authors may want to cite the original paper and methods depicting these cell/ECM functional units:  1C, the day 14 fibronectin staining looks as if the immunofluorescence followed a process whereby fixing of the cells was done first and permeabilization followed.The original methods have provided troubleshooting for this process (to avoid the apparent void in fibers) by conducting a simultaneous fix/permeabilization step followed by added fixing (published in this version of CDM production methods and quality control analyses: PMCID: PMC5058441).The authors may want to repeat this using this advice and potentially generate a higher quality image/figure .4. Regarding data in supplemental figure 2, CDMs have been shown to mature enough (e.g., incorporate enough fibrous collagen) in vitro to include collagen fibers that are detected using the second harmonic generation of polarized light (obtained with most multiphoton microscopes: PMCID: PMC7442735).Hence, the authors may want to include this orthogonal approach to support their interpretation of the data (to support that the 25-60 nm filaments are indeed collagen fibers as suggested) and demonstrate the high quality of their CDMs.
5. The statement suggesting that the growth of new CDM (fibers) preferentially takes place on the top layers is not well justified and should probably be omitted, while the density of fibers statement could prevail.

Response to Reviewers:
We thank the reviewers for their evalua3on of our manuscript and have responded to their ques3ons point-by-point below.
Reviewer #1 (Comments to the Authors (Required)): I congratulate the authors on this heroic effort aiming to provide a first structural 3D atlas of the ECM at molecular resolu3on.The systema3c effort to op3mize specimen prepara3on of cell culture-based 3D-ECM plaNorms for compa3bility with cryo-FIB liQ-out and high quality cryo-ET provide a valuable star3ng point for studies focused on complex cell-ECM interac3ons that beSer recapitulate in vivo scenarios in a controlled 3D in vitro system.The quality of the data is outstanding and the accompanying annota3ons are stunning.The authors complement their structural characteriza3on with systema3c proteomics and light microscopy-based measurements.While the manuscript remains descrip3ve, the work in my view opens an exci3ng avenue for the study cell-matrix interfaces and illumina3ng cri3cal gaps in our knowledge of ECM synthesis, secre3on and assembly, and the effect of cri3cal muta3ons.
We thank the reviewer for the posi3ve comments.
I recommend publica3on, and would have liked to see more quan3fica3on of the volume frac3ons of the different ECM structures iden3fied.These might be useful in the future, especially if MS data is presented in parallel in terms of protein abundance, to possibly iden3fy poten3al candidates for newly iden3fied ECM structures (more below).
We have performed the suggested quan3fica3ons and provide more details on the results and the limita3ons associated with such quan3fica3ons below.

Minor comments:
The manuscript text is at 3mes too wordy and therefore convoluted, especially in the introduc3on and in some of the results descrip3on.It may benefit from simplifica3on and consolida3on.
In the course of revision, the text has changed in several instances.We hope this has consolidated the text.
Abstract: "sample thinning by cryo-liQ out focused ion beam milling".LiQ out is an extrac3on procedure.Microfabrica3on and thinning are done by FIB.I recommend rephrasing for clarity to a broader audience."This allows us to visualize ECM for the first 3me in its hydrated, cellular context".The authors may want to reward "cellular" to "na3ve" or similar.
We have reworded the sentences accordingly (changes in bold): "We have developed a 3D-ECM pla2orm compa6ble with sample thinning by cryo-focused ion beam milling, the li6-out extrac8on procedure and cryo-electron tomography.""This allows us to visualize ECM for the first 6me in its hydrated, na8ve context." Figure 1: I wonder whether the authors could comment on the origin of the observed direc3onality of both the extracellular and intracellular filaments.There is more informa3on in the supplementary figures, but it is not explained or discussed.Does this evolve over 3me from cell seeding to the end point of the cryo-EM inves3ga3on?Is it directed by something in the EM grid substrate?Previous publica3ons have shown the direct influence of cells, specifically fibroblasts, on the orienta3on of ECM fibers.This is mediated by cytoskeleton-ECM interac3ons such fibrillar adhesions (Geiger et al, 2001;PMID: 11715046;Harris et al 1981, PMID: 7207616).Through these adhesions, cells have been reported to exert forces on ECM fibers such as fibronec3n and collagens, and to align them (Piotrowski-Daspit et al., 2017;PMID: 28793224).In agreement with this, throughout the 14+ days of CDM culturing, the fibroblasts seem to organize the ECM fibers as they are producing them (see Figure 1, where the alignment of cells and ECM is visible already early on).Accordingly, CDMs generated from TIFF cells and other epithelial cells have been shown to have highly aligned ECM fibers (Kaukonen et al., 2017, PMID: 29048422;Lansky et al., 2019, PMID: 32055794), matching our observa3ons.The EM substrate itself seemingly has no influence on this direc3onality, as the same organiza3on can be observed when growing cells on glass cover slips or cell culture dishes (for example as shown in Figure S1) or as reported in Kaukonen et al., 2017.We now also address this point more explicitly in the text: "Previous publica6ons have reported the direct influence of cells, specifically fibroblasts, on the orienta6on of ECM fibers, which are mediated by cytoskeleton-ECM interac6ons via adhesion complexes (Geiger et al., 2001;Harris et al., 1981).Cells align ECM fibers such as fibronec6n and collagens by exer6ng forces through cell-ECM adhesion interac6ons (Piotrowski-Daspit et al., 2017)." Page 3: "Altogether, these results confirmed that CDMs harvested on or later than Day 14 represent bona fide ECM assemblies".It would be useful if the authors could point to previous publica3ons on the expected morphology of "bona fide" ECM in support of their conclusion.
We have now rephrased this sentence to further clarify the point made and added cita3ons to support our statement.Specifically, we toned down our statement from "represen6ng bona fide ECM assemblies" to "mimicking ECM assemblies found in 6ssue".(Changes in bold) "Altogether, these results confirmed that CDMs harvested on or later than Day 14 mimic ECM assemblies found in 8ssue (Fitzpatrick and McDeviY, 2015;Ahlfors and Billiar, 2007) and should allow visualiza6on of ECM components in their na6ve environment of matrix-secre6ng cells." We are convinced that the composi3on of our CDMs is similar to ECM in 3ssue based on the MS data we provide, as well as our microscopy results which show higher-order assembly of ECM components such as collagen and fibronec3n.Previous publica3ons use CDMs to study complex cellular behaviours such as cell migra3on (Hakkinen et al., 2011, PMID: 20929283), and in regenera3ve medicine (Fitzpatrick et al., 2015, PMID: 25530850;Bello et al., 2012, PMID: 11721649).While CDMs cannot represent an exact copy of ECM in 3ssues, they are a physiologically close representa3ve system enabling us to study basic structures such as the ones described in this publica3on.
Page 4: with regard to the tes3ng of the cryo-protectants, it would be useful if the authors could provide further informa3on, and possibly references.For example, for Dextran, it is important to provide the molecular weight as this is an important factor for osmolarity and discuss evalua3on of the effect of osmolarity on cell/extracellular na3ve preserva3on in this context.
All cryo-protectants were either prepared in PBS, PB or cell medium.These buffers are commonly used in cell culture and 3ssue experiments by us and others, hence their physiological proper3es are evident.Each cryo-protectant listed in Table S3 was chosen based on its previously described use for high pressure freezing.References for studies and protocols employing them were already given in the main text, as for example below.
The cryo-protectant working best for us was Dextran (molecular weight of 40kDa).This Dextran has been used in previous studies performing high-pressure freezing such as Bharat et al, 2018 (PMID: 29681471).In order to describe this in more detail in the revised manuscript, we now provide this informa3on in the main text and methods (changes in bold)."A degassed cryoprotectant solu6on containing 10% (w/v) high molecular weight Dextran (40 kDa) in 0.1 M phosphate buffer (PB) showed the highest success and resulted in complete vitrifica6on in several samples with low addi6onal background (Figure 2

, Figure S4). High MW Dextran is a polymer that is non-penetra8ng and has been shown to have liLle osmo8c effect on 8ssues (Dahl and Staehelin, 1989; Al-Amoudi, 2004), resul8ng in its rou8ne use to facilitate vitrifica8on of biological samples by HPF (Dahl and Staehelin, 1989; Sader et al., 2009; Bharat et al., 2018; Mesman, 2013; Zhang et al., 2021)."
We also added informa3on to the legend of Table S3 as follows: "An overview table showing the tested cryoprotectant/buffer combina3ons, detailing the achieved vitrifica3on status and the introduced background for each combina3on.The buffers were chosen based on their physiological proper9es and their common use for biological specimens." Page 4: how do the authors interpret the "empty areas" in the extracellular space?This is where a parallel presenta3on of the ECM proteomics could be useful.
The reviewer makes a valuable point here.Indeed, we did not aim to imply that these areas are absolutely void of any molecules, but are devoid of large structures like collagen, amorphous density, or similar.There are likely soluble proteins and/or sugars present in these areas, which we simply cannot visualize sufficiently with the used methodology.However, the "empty areas" could also match with being buffer/water-rich areas, finng with a hydrogel.Our HPF carriers were specifically designed to not compress the specimen and to retain the CDM in an unaltered state.It would be temp3ng to speculate that in 3ssues the ECM could be more compressed than is the case for our CDMs.Hence, it is conceivable that in a 3ssue, these spaces might be smaller.However, future work will be required to unambiguously show this.
While our proteomics data tells us that we have many secreted soluble, smaller proteins in the extracellular regions of our CDMs, such as growth factors, or proteases, it cannot give us any indica3on about the posi3oning of these proteins.Hence, we cannot confidently define which ECM components might be located in these seemingly empty spaces (or if any are).
Figure 2: how do the authors interpret the large vacuoles observed primarily in cells located towards the top of their lamellae?The authors may want to discuss this if appears to be a common feature.
The reviewer is correct, the presence of vesicles located in cells primarily towards the top of the CDM is interes3ng.One poten3al interpreta3on of these large vesicles is that the top cells might be more ac3ve in the produc3on and secre3on of ECM components and the overall remodelling of the ECM.However, the content of these vesicles is unclear at this point.In addi3on, following the sugges3on from reviewer #3, that our ini3al statement linking the growth of new CDM to the top layers should probably be omiSed, we have refrained from further men3oning these vesicles in the manuscript.
Page 6: "The varia3on in both diameter and repeat paSern along the filament z-axis suggests extensibility and assembly variability."And "the amorphous matrix was observed only in close proximity to ECM fibers...This suggests a co-assembly between the components in these en33es."These are unnecessary overinterpreta3ons of the data.I recommend that the authors rephrase, especially when the structures cannot be iden3fied.
Mul3ple studies have suggested different ECM fibers to be extensible and variable in their repeat paSern and diameter, especially when responding to forces applied by cells.We would therefore like to keep this sentence in the manuscript, but have adapted it and added further references to underline this point: "The varia6on in both diameter and repeat paYern along the filament z-axis suggests extensibility and assembly variability, matching several studies on the proper8es of both fibronec8n and fibrillin (Glab and Wess, 2008;Klotzsch et al., 2009;SherraL et al., 2001;Dzamba and Peters, 1991)." Similarly, we s3ll would like to men3on the poten3al of a co-assembly of molecules in the amorphous matrix and ECM fibers, and have rephrased the sentence accordingly to tone it down: "Based on these observa8ons, it is temp8ng to speculate that the components in these en88es might co-assemble." Page 6, sec3ons "The absence of other filament-like structures in CDMs": here it would be useful to explicitly spell out the candidates from proteomics measurements, or the previous reports the authors men3on (without providing references!) and the expected structural features, poten3ally even from structure predic3ons.For example, Table 2 "ECM proteins in TIFF CDMs iden3fied by MS" could perhaps be presented in a plot form that also reflects rela3ve abundance (possibly only the fibril-forming proteins, while others are aggregated, for simplicity) in the main text and related to the related volume frac3ons occupied by the different ECM structures as observed by cryo-ET.
We thank the reviewer for this comment.The cita3ons had ini3ally been stated in Table S1 alone and have now been added also to the main text for clarifica3on, with a reference to Table S1 for more details: "Most strikingly, our data does not contain any other clearly discernible ECM fiber types, as one would expect based on our proteomics data or previous reports of the filamentous assemblies present within the ECM (Baldock et al., 2003;Lansky et al., 2019a;SherraY et al., 2001;Früh et al., 2015) (such as FN fibers or Col-VI, see also Table S1 for published reports on filament spacing and periodicity)." Structure predic3ons of fibronec3n, Col VI, and fibrillin are unfortunately not a viable op3on to define the exact structural conforma3on of these proteins.For example, the AlphaFold predic3on of Fibronec3n does not show a fibril arrangement, that could be easily compared to filaments observed in our tomograms.(hSps://alphafold.ebi.ac.uk/entry/P02751).Moreover, these fibers are supposed to be flexible, adding another layer of difficulty to the predic3ons, as one would not know which state of the protein will be predicted, and which one we might actually see in our samples.Given these reasons, we did not include any AlphaFold predic3ons in our manuscript.
For example, Table 2 "ECM proteins in TIFF CDMs iden3fied by MS" could perhaps be presented in a plot form that also reflects rela3ve abundance (possibly only the fibril-forming proteins, while others are aggregated, for simplicity) in the main text and related to the related volume frac3ons occupied by the different ECM structures as observed by cryo-ET.
The sugges3on to plot the MS data reflec3ng the rela3ve MS abundance and to relate this to the volume frac3ons of the different en33es in our tomograms is very interes3ng.While, this indeed could help to get a beSer understanding of how much the different components contribute to form structures observed in our data, there are several limita3ons associated with this approach, which we discuss below.
We have generated plots based on our proteomics data showing the rela3ve abundance of fibrilforming ECM proteins with respect to each other, and also with respect to all (fibril-forming and non-fibril forming) ECM proteins.We note, that our proteomics data allows to determine rela3ve abundance, and with this in mind we created a semi-quan3ta3ve comparison between the different CDM components.We have added this comparison as two panels in Figure 1: We now include these plots as new panels D and E in an updated version of Figure 1, as they indeed allow a straighNorward interpreta3on of ECM content.
However, concerning the quan3fica3on of volume frac3ons in our cryo-electron tomograms, we need to point out that the tomograms presented in this study were acquired at selected sites, based on the quality (such as thickness) of the lamella and what we defined as sites of interest when visualizing lamellae at lower magnifica3on.This means that our selec3on of acquisi3on sites was influenced by the visibility of features of interest, such as fibrillary ECM structures, or intriguing sites of cell-ECM interac3on.
Hence, the tomograms do not allow a generaliza3on of the distribu3on of the individual ECM components, but instead are aimed at describing their molecular architecture and interplay.Thus, a comparison between tomogram volume frac3ons and MS volume frac3ons could be poten3ally misleading.We have chosen to include a quan3fica3on comparing volume frac3ons in our tomograms for the reviewer (as described below), but would not include them in the manuscript: Importantly, with the workflow now being established, such a quan3ta3ve analysis as suggested by the reviewer is definitely something that could be pursued in future work, keeping the limita3ons of comparing volume data (of proteins with different density) to proteomics quan3fica3ons in mind.Indeed, in our new data acquisi3ons we already take this into considera3on (e.g. by performing unbiased montage cryo-ET to image en3re lamellae).
We now men3on in the Discussion how montage cryo-ET can increase throughput, together with the men3oning of the Waffle-method and Serial-liQout (please see answer below).
I recommend that the authors include in their discussion/conclusion some men3on on the poten3al of "waffle" or "Serial liQ-out" methodologies to gain higher throughput and more insights into similar model systems given the dimensions of their specimens.
We now discuss both methods in the revised manuscript.

"A combina8on of the workflow shown in this study and recently introduced methodologies such as the "Waffle method" (Kelley et al., 2022), the "Serial li6-out" technique ( Schiøtz et al., 2023) and montage cryo-electron tomography (Peck et al., 2022; Yang et al., 2023; Chua et al., 2024) could substan8ally improve throughput. In turn, with increased dataset sizes this could grant more detailed and poten8ally also quan8ta8ve insights into CDMs and their components."
Movies S4 and S4: some of the fuzzy amorphous density appears to be con3nuous with the Col-I fibrils.It might be useful for the authors to men3on this if this appears to be a common feature in other data as it might indicate that these structures are precursors of Col-I assembly.
The amorphous density that appears to be following/preceding the Col-I fibrils can be observed in several instances, oQen when the fibrils are running in transverse.However, this is not the case for all fibrils.This makes the amorphous density closely associated with some Col-I fibrils difficult to interpret.It could show poten3al precursor structures, but it could also show bound sugars and other proteins.Hence, we refrain from specula3ng about it.Corrected.
Methods, cryo-FIB: unclear at which point the half-moon grid are mounted into autogrids.Please add.
A sentence was added to clarify.

"Prior to sample loading, a half-moon grid with 4 finger-like extensions for li6-out aLachment (Ted Pella, #10GC04) was clipped into an AutoGrid. For clipping, all finger-like extensions of the half-moon grid were oriented in line with the milling window to allow for low angle sample thinning."
Data availability: it would be extremely valuable for the community if the authors deposit the tomograms and segmenta3ons associated with this work.I would even strongly recommend deposi3on of the full data, from raw movies to segmenta3ons in EMPIAR.
We have deposited the raw frames, 3lt series and isonet-corrected tomograms to EMPIAR.They will be available via accession code EMPIAR-11897 upon publica3on of this manuscript.We hope that this data will represent a valuable resource to the scien3fic community interested in ECM and cryo-electron tomography.
Reviewer #2 (Comments to the Authors (Required)): In this ar3cle en3tled "Unveiling the ultrastructural landscape of extracellular matrix via liQ-out cryo-FIBSEM and cryo-ET", Zens et al. repor3ng a technical development based on cryoEM and FIBSEM technologies.They adapted and combined methodologies (references 24, 28, 31 and 36) to enable visualiza3on of cell derived matrices in a hydrated context.This is of outmost relevance considering the deleterious effects of classic protocols including chemical fixa3on and dehydra3on steps.
This ar3cle is well wriSen and well-illustrated.It has an appreciable part of "trial and error" repor3ng of the scien3fic process, and results in establishment of a successful and promising protocol for the study and for future studies.This is the primary merit of the study and as such may become seminal in the field.
Limita3ons of the manuscript are related to the rela3vely low number of conclusive points that are made by the authors.Indeed, the use of a single biological condi3on (TIFF derived CDM) is somewhat limi3ng the general value of the study.Having said that, owing to the laborious nature of this approach, limi3ng efforts to a single sample 3me is understandable.
Major comments: The authors should consider toning down this statement: "Another aggrava3ng factor for the structural annota3on of ECMs is the heterogeneity of 3ssuederived ECM material.In contrast, cell derived matrices (CDMs) are a highly adaptable and versa3le tool that is increasingly used to recapitulate the complexity of na3ve 3ssue ECM (29-31)." I reality, CMD are not recapitula3ng the complexity of na3ve 3ssue, which is the results of many cell types interac3ng together and jointly contribu3ng to the matrix composi3ons and remodelling.Only in vivo observa3on can fully recapitulate in vivo complexity and CDM is rather a (valuable) reduc3onist in vitro model system to study matrix biology.
The statement was rephrased and toned down as follows: "Another aggrava6ng factor for the structural annota6on of ECMs is the heterogeneity of 6ssuederived ECM material.In contrast, cell derived matrices (CDMs) are a highly adaptable and versa6le tool that is increasingly used to mimic fundamental aspects of na6ve 6ssue ECM and study its structure and func8on (Hakkinen et al., 2011;Petrie and Yamada, 2016;Kaukonen et al., 2017;Cukierman et al., 2001)." The authors are taking care not to overinterpret their findings and this is commendable.However, it would be beneficial for the reader to if some more concrete conclusions could be drawn.some conclusions out of this work.
Related to this topic, the "beaded filaments" and the absence of fibronec3n visualiza3on are suprising and need more explana3on.
Can the authors rule out the possibility that "amorphous ECM" is in fact not well-preserved during sample prepara3on?Can the rare fibers that are observed in the amorphous ECM be a low percentage of preserved material?
To our knowledge, our work is the first study to employ cryo-electron tomography on na3vely preserved (i.e.fully vitrified) ECM at this resolu3on.There are no other studies we can directly compare our data to.Also, there are currently no high-resolu3on structures available for fibronec3n fibers, fibrillin fibers, or collagen VI fibers that we could use as gold-standard for comparison.As explained above, Alphafold predic3ons are also not providing unambiguous structures for comparison.
While we aimed to avoid any damage to our specimen and to keep it as na3vely preserved as possible, we naturally cannot exclude (as is the case for any experimental approach) that specimen prepara3on induces minor altera3ons.However, we are confident that the workflow we have established is as op3mal as currently possible.Specifically, we do not fix or otherwise chemically treat our specimen during prepara3on.Also, we have tested different buffer condi3ons for cryo-protec3on, including one that contained no addi3onal cryo-protectant but solely cell culture medium (which showed very good contrast, but as noted good but s3ll incomplete vitrifica3on).In all these condi3ons the amorphous ECM is present.
In summary, all of the above leads us to the belief that the amorphous density is not an artefact created by our sample prepara3on process.
Can this be fibronec3n filaments?Can the beaded filaments be fibronec3n filaments?
We would like to answer both these ques3ons together.It is possible that the rare filaments in the amorphous ECM are fibronec3n filaments and this has been a thought we have discussed mul3ple 3mes.However, due to the lack of structural knowledge of fibronec3n we cannot make any definite statements.Similarly, we also cannot exclude the possibility that the beaded filaments are indeed fibronec3n.However, as stated in the manuscript the periodicity we observed fits beSer to fibrillin fibers.
There are different experimental approaches that we could take in the future to further iden3fy these filaments.This could for example include immunogold labeling approaches.We did indeed try this with an an3-Fibronec3n an3body, but unfortunately without achieving sufficient labelling efficiency and specificity to make an unambiguous statement.Future work could invest into op3mizing this approach, requiring substan3al experimental efforts, which we believe is beyond the scope of this work.Specifically, based on our experience with this approach we are not convinced that immunogold labeling would be a straighNorward method for iden3fying single structures in highly complex, 3ghtly packed environments such as our CDM.
Another approach could be to employ more advanced super-resolu3on cryo-fluorescence microscopy methods once they become available (as we would require resolu3on in the sub-20 nanometer range in 3D to achieve unambiguous iden3fica3on).
Finally, as already pointed out by the reviewer in the next comment, gene3c engineering of CDMs via knockout of specific proteins could be used as an indirect experimental approach to iden3fy filaments (e.g. via the absence of specific features upon KO of Fibronec3n, Fibrillin etc..).As stated below, this is indeed an avenue we are following up, but is beyond the scope of this manuscript.
One major limita3on of the study is the lack of comparisons between for example condi3on or treatments that are expected to influence the CDM.For example, gene3c manipula3on of the TIFF or CDM made by another cell type.If these are beyond the scope of this study, the authors should discuss this point and men3on this limita3on.
We fully agree that the gene3c manipula3on of TIFF cells for the CDM produc3on or the use of another cell type would be an important future step for this project.The 3me-consuming and complex process for sample genera3on and data acquisi3on render these op3ons being out of scope for this manuscript (as also acknowledged by this reviewer).Indeed, we had already tried to address these future op3ons in the use of CDMs in the main text.
We have now added that these plans are beyond the scope of the current manuscript (changes in bold): "As also revealed by our data, ECM filaments represent a challenging target for structure determina6on approaches due to their apparent variability.Future experiments using lig-out cryo-FIBSEM and cryo-ET, combined with novel image processing tools based on neural networks (for example (Rice et al., 2023)) or func6onal studies using gene6c knockouts will be required to extend and annotate the gallery of ECM structures." "Specifically, based on the hypothesis that changes in matrisome composi6on of ECMs of different origin are reflected on a structural level, cryo-ET of CDMs now allows studying how ECM-specific topologies define 6ssue homeostasis and underlying cellular behavior.Hence, CDMs derived from cell types of different origin, e.g.skin, lung, mammary or cancer-associated fibroblasts (CAF), offer an appealing avenue for compara6ve analysis of ECMs to obtain a clearer depic6on of the role of individual ECM components in defining specialized 6ssue-specific matrices.Specifically, this could be achieved via an integra6ve approach using a combina6on of molecular imaging via cryo-ET, proteomics analysis via mass spectrometry (MS) followed by func6onally characterizing the role of specific matrix components using CRISPR-Cas9 knockout approaches.Together with pharmacological treatment to target specific ECM component this offers the possibility to manipulate ECM produc6on.While being beyond the scope of this work, these approaches offer exci9ng future avenues for a more detailed understanding of the ECM."

Minor comments:
The authors abstract first paragraph is somewhat misleading, especially when men3oning Col VI assembly as remaining "enigma3c", considering that there is no conclusive statement made in the manuscript regarding Col VI.
We have removed this statement.These expressions/sentences are unclear or inappropriate to my opinion."There is no unambiguous consensus..." "molecular sociology..." The sentences were rephrased to avoid the men3oned expressions: "current literature does not allow for an unambiguous consensus on the exact molecular assembly of many of the other ECM fibers" "This includes the structural and func3onal characteriza3on of single components and their interac9ons in na3vely preserved ECM." Reviewer #3 (Comments to the Authors (Required)): This study is an exquisite example of how new technology can facilitate novel informa3on.The only areas of poten3al improvement are some missing cita3ons to credit the original CDM work and methodology (and similar) as well as a sugges3on to combine the Cryo-generated data with SHG (second harmonic genera3on of polarized light) using the same 3D sample.The laSer is merely a sugges3on that could significantly improve the interpreta3on of the data (regarding collagen fibers).
We appreciate the reviewer's posi3ve comments.
Specific MINOR sugges3ons: 1.In the introduc3on, when CDMs are first described as a valid in vivo mimicry approach, the authors may want to cite the original paper and methods depic3ng these cell/ECM func3onal units: a We apologize for not having cited these papers already in the first version of our manuscript.We have now adapted the text to include these cita3ons and men3on cancer research in this context: "The high versa6lity and adaptability of CDMs to dis6nct research ques6ons has also made them a rou6nely used tool in cancer research (Franco-Barraza et al., 2017;Malik et al., 2019;Padhi et al., 2020;Francescone et al., 2021;Jones et al., 2022).The use of cancer-associated fibroblasts (CAFs) for CDM genera6on results in a close mimic of the tumor-associated stroma (Amatangelo et al., 2005) and has aided in the iden6fica6on of novel regulators of ECM alignment (Jones et al., 2022) as well as poten6al therapy targets (Francescone et al., 2021)." 3. Technically, in results shown in Figure 1C, the day 14 fibronec3n staining looks as if the immunofluorescence followed a process whereby fixing of the cells was done first and permeabiliza3on followed.The original methods have provided troubleshoo3ng for this process (to avoid the apparent void in fibers) by conduc3ng a simultaneous fix/permeabiliza3on step followed by added fixing (published in this version of CDM produc3on methods and quality control analyses: PMCID: PMC5058441).The authors may want to repeat this using this advice and poten3ally generate a higher quality image/figure.
We thank the reviewer for poin3ng us towards these papers and the protocol for a beSer way of fixa3on/permeabiliza3on.We have repeated the day 14 staining of Figure 1C accordingly.This indeed substan3ally improved the image quality.We have updated Figure 1, Panel C and describe this also in the methods.Figure 1 has been included in the answer to Reviewer #1, including Panel C that was updated (see above).
4. Regarding data in supplemental figure 2, CDMs have been shown to mature enough (e.g., incorporate enough fibrous collagen) in vitro to include collagen fibers that are detected using the second harmonic genera3on of polarized light (obtained with most mul3photon microscopes: PMCID: PMC7442735).Hence, the authors may want to include this orthogonal approach to support their interpreta3on of the data (to support that the 25-60 nm filaments are indeed collagen fibers as suggested) and demonstrate the high quality of their CDMs.
As suggested, we have performed second harmonic genera3on of our CDMs to visualize collagen fibers.Specifically, we have combined our CNA35-EGFP live staining of collagen with second harmonic genera3on of polarized light.
As expected, the SHG imaging confirms the results obtained by using the collagen live-stain.The collagen shown by SHG imaging overlaps with our fluorescent signal for CNA35-EGFP, highligh3ng that both imaging methods visualize the same fibers.
Considering the fluorescence microscopy data and SHG imaging data of our CDMs, combined with the clearly visible D-spacing in the collagen fibers in our cryo-electron tomograms, we are confident that the filaments we visualize are indeed collagen fibers.
The main advantage of the CNA35-EGFP imaging is that it is compa3ble with our cryofluorescence approach that we use prior to ion-beam milling and liQ-out experiments, and that it requires substan3ally less laser power which is beneficial for specimen preserva3on.Given that SHG and fluorescent imaging achieve approximately the same resolu3on and due to space constraints with showing supplementary data, we have refrained from adding the SHG imaging to the manuscript.5.The statement sugges3ng that the growth of new CDM (fibers) preferen3ally takes place on the top layers is not well jus3fied and should probably be omiSed, while the density of fibers statement could prevail.
As suggested by the reviewer, we have now rephrased this sec3on and removed the men3oning of CDM growth in the top layers.Instead, we just men3on an apparent decrease in fiber density towards the top.Please also see our response to reviewer #1.8) References: There is no limit to the number of references cited in a manuscript.References should be cited parenthetically in the text by author and year of publication.Abbreviate the names of journals according to PubMed.9) Supplemental materials: Reports generally have up to 3 supplemental figures and 10 videos.You currently exceed this limit but, in this case, we will be able to give you the extra space.Please also note that tables, like figures, should be provided as individual, editable files.A summary of all supplemental material should appear at the end of the Materials and methods section.
Please include one brief sentence per item.12) Conflict of interest statement: JCB requires inclusion of a statement in the acknowledgements regarding competing financial interests.If no competing financial interests exist, please include the following statement: "The authors declare no competing financial interests."If competing interests are declared, please follow your statement of these competing interests with the following statement: "The authors declare no further competing financial interests." 13) A separate author contribution section is required following the Acknowledgments in all research manuscripts.All authors should be mentioned and designated by their first and middle initials and full surnames.We encourage use of the CRediT nomenclature (https://casrai.org/credit/).
14) ORCID IDs: ORCID IDs are unique identifiers allowing researchers to create a record of their various scholarly contributions in a single place.Please note that ORCID IDs are required for all authors.At resubmission of your final files, please be sure to provide your ORCID ID and those of all co-authors.
15) Journal of Cell Biology now requires a data availability statement for all research article submissions.These statements will be published in the article directly above the Acknowledgments.The statement should address all data underlying the research presented in the manuscript.Please visit the JCB instructions for authors for guidelines and examples of statements at (https://rupress.org/jcb/pages/editorial-policies#data-availability-statement).

B. FINAL FILES:
Please upload the following materials to our online submission system.These items are required prior to acceptance.If you have any questions, contact JCB's Managing Editor, Lindsey Hollander (lhollander@rockefeller.edu).
--An editable version of the final text (.DOC or .DOCX) is needed for copyediting (no PDFs).
--Cover images: If you have any striking images related to this story, we would be happy to consider them for inclusion on the journal cover.Submitted images may also be chosen for highlighting on the journal table of contents or JCB homepage carousel.
Images should be uploaded as TIFF or EPS files and must be at least 300 dpi resolution.
**It is JCB policy that if requested, original data images must be made available to the editors.Failure to provide original images upon request will result in unavoidable delays in publication.Please ensure that you have access to all original data images prior to final submission.****The license to publish form must be signed before your manuscript can be sent to production.A link to the electronic license to publish form will be sent to the corresponding author only.Please take a moment to check your funder requirements before choosing the appropriate license.**Additionally, JCB encourages authors to submit a short video summary of their work.These videos are intended to convey the main messages of the study to a non-specialist, scientific audience.Think of them as an extended version of your abstract, or a short poster presentation.We encourage first authors to present the results to increase their visibility.The videos will be shared on social media to promote your work.For more detailed guidelines and tips on preparing your video, please visit https://rupress.org/jcb/pages/submission-guidelines#videoSummaries.
Thank you for your attention to these final processing requirements.Please revise and format the manuscript and upload materials within 7 days.If complications arising from measures taken to prevent the spread of COVID-19 will prevent you from meeting this deadline (e.g. if you cannot retrieve necessary files from your laboratory, etc.), please let us know and we can work with you to determine a suitable revision period.
Please contact the journal office with any questions at cellbio@rockefeller.edu.

Figure
Figure S1: scale bar size missing in b.

Fig S1 :
Fig S1 : Please include original images of panel A, before annotation.Fig S6 : "enodcytosis" should be corrected to "endocytosis".Fig S9 : If available, please show low magnification pictures for panel D. The same way it has been done in Fig S6 A, for instance.These expressions/sentences are unclear or inappropriate to my opinion."There is no unambiguous consensus..." "molecular sociology..."

Figure R1 :
Figure R1: 10 exemplary tomograms were chosen from the 42 available tomograms, 5 of which had mostly a crosssec(on view of ECM fibers and 5 a transverse view.For each tomogram, 5 central slices were manually measured for the area occupied by collagen, small ECM fibers, amorphous density, and seemingly empty area in the ECM.The calculated frac(ons per tomogram were averaged.Then, an average of all 10 tomograms was calculated and is displayed as a pie chart.

Figure S1 :
Figure S1: scale bar size missing in b.

Fig S1 :
Fig S1 : Please include original images of panel A, before annota3on.

Figure R2 :
Figure R2: On-grid imaging of collagen in CDMs.TIFF CDMs were grown for 14 days and stained with CNA35-EGFP as described in the manuscript.Specimens were then imaged on a Leica SP8 DIVE Mul(photon and confocal microscope, using a HC FUOTAR L 25x / 0.95 W objec(ve.Two exemplary sites of two different CDMs are shown here, imaged first by second-harmonic genera(on of polarized light microscopy (SHG) and then by fluorescence microscopy.Scale bar dimensions are shown in the figure.

10)
Video legends: Should describe what is being shown, the cell type or tissue being viewed (including relevant cell treatments, concentration and duration, or transfection), the imaging method (e.g., time-lapse epifluorescence microscopy), what each color represents, how often frames were collected, the frames/second display rate, and the number of any figure that has related video stills or images.11) eTOC summary: A ~40-50 word summary that describes the context and significance of the findings for a general readership should be included on the title page.The statement should be written in the present tense and refer to the work in the third person.It should begin with "First author name(s) et al..." to match our preferred style.

and a decrease in ECM fibre density from boLom towards top of the CDM could be observed in a majority of lamellae."
"Empty areas in extracellular space devoid of any structures were also observed, Thank you for submitting your revised manuscript entitled "Unveiling the ultrastructural landscape of extracellular matrix via liftout cryo-FIBSEM and cryo-ET".We would be happy to publish your paper in JCB pending final revisions necessary to meet our formatting guidelines (see details below).To avoid unnecessary delays in the acceptance and publication of your paper, please read the following information carefully.A. MANUSCRIPT ORGANIZATION AND FORMATTING:Full guidelines are available on our Instructions for Authors page, https://jcb.rupress.org/submission-guidelines#revised.**Submission of a paper that does not conform to JCB guidelines will delay the acceptance of your manuscript.** 1) Text limits: Character count for Reports is < 20,000, not including spaces.Count includes title page, abstract, introduction, combined results & discussion, and acknowledgments.Count does not include materials and methods, figure legends, references, tables, or supplemental legends.2) Figure formatting: Reports may have up to 5 main text figures.Scale bars must be present on all microscopy images, including inset magnifications.Please add a scale bar for the magnification image in FigureS1C.Also, please avoid pairing red and green for images and graphs to ensure legibility for color-blind readers.If red and green are paired for images, please ensure that the particular red and green hues used in micrographs are distinctive with any of the colorblind types.If not, please modify colors accordingly or provide separate images of the individual channels.3)Statisticalanalysis:Errorbars on graphic representations of numerical data must be clearly described in the figure legend.The number of independent data points (n) represented in a graph must be indicated in the legend.Please, indicate whether 'n' refers to technical or biological replicates (i.e.number of analyzed cells, samples or animals, number of independent experiments).If independent experiments with multiple biological replicates have been performed, we recommend using distribution-reproducibility SuperPlots (please seeLord et al., JCB 2020)to better display the distribution of the entire dataset, and report statistics (such as means, error bars, and P values) that address the reproducibility of the findings.Statistical methods should be explained in full in the materials and methods.For figures presenting pooled data the statistical measure should be defined in the figure legends.Please also be sure to indicate the statistical tests used in each of your experiments (both in the figure legend itself and in a separate methods section) as well as the parameters of the test (for example, if you ran a t-test, please indicate if it was one-or two-sided, etc.).Also, if you used parametric tests, please indicate if the data distribution was tested for normality (and if so, how).If not, you must state something to the effect that "Data distribution was assumed to be normal but this was not formally tested."4) Title: To make the title more concise and match our preferred style we suggest a slight change to the following: "Lift-out cryo-FIBSEM and cryo-ET reveal the ultrastructural landscape of extracellular matrix" 5) Materials and methods: Should be comprehensive and not simply reference a previous publication for details on how an experiment was performed.Please provide full descriptions (at least in brief) in the text for readers who may not have access to referenced manuscripts.The text should not refer to methods "...as previously described."Any software used for image processing subsequent to data acquisition.Please include details and types of operations involved (e.g., type of deconvolution, 3D reconstitutions, surface or volume rendering, gamma adjustments, etc.).
6) For all cell lines, vectors, constructs/cDNAs, etc. -all genetic material: please include database / vendor ID (e.g., Addgene, ATCC, etc.) or if unavailable, please briefly describe their basic genetic features, even if described in other published work or gifted to you by other investigators (and provide references where appropriate).Please be sure to provide the sequences for all of your oligos: primers, si/shRNA, RNAi, gRNAs, etc. in the materials and methods.You must also indicate in the methods the source, species, and catalog numbers/vendor identifiers (where appropriate) for all of your antibodies, including secondary.If antibodies are not commercial, please add a reference citation if possible.
Thank you for this interesting contribution, we look forward to publishing your paper in Journal of Cell Biology.