Multiplex DNA fluorescence in situ hybridization to analyze maternal vs. paternal C. elegans chromosomes

  1. Silvia Gutnik
  2. Ahilya Sawh
  3. Susan E. Mango

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

Recent advances in high-throughput microscopy have paved the way to study chromosome organization at the single-molecule level and have led to a better understanding of genome organization in space and time. During development, distinct maternal and paternal contributions ensure the formation of an embryo proper, yet little is known about the organization of chromosomes inherited from mothers versus fathers. To tackle this question, we have modified single-molecule chromosome tracing to distinguish between the chromosomes of two well-studied strains of C. elegans called Bristol and Hawai’ian. We find that chromosomes from these two strains have similar folding patterns in homozygous hermaphrodites. However, crosses between Bristol and Hawai’ian animals reveal that the paternal chromosome adopts the folding parameters of the maternal chromosome in embryos. This is accomplished by an increase in the polymer step size and decompaction of the chromosome. The data indicate that factors from the mother impact chromosome folding in trans. We also characterize the degree of intermixing between homologues within the chromosome territories. Sister chromosomes overlap frequently in C. elegans embryos, but pairing between homologues is rare, suggesting that transvection is unlikely to occur. This method constitutes a powerful tool to investigate chromosome architecture from mothers and fathers.

Article activity feed

  1. Review Commons

    Note: This rebuttal was posted by the corresponding author to Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Reply to the reviewers

    'The authors do not wish to provide a response at this time.' The response has been included in a PDF

  2. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #3

    Evidence, reproducibility and clarity

    Summary:

    This manuscript attempts to answer how the maternal and paternal chromosomes are organized, and probe the determinants of this organization.

    The authors use two divergent strains of worms - Bristol(N2) and Hawaiian - and hybrids progeny to study this as there are large regions of sequence variants between chromosome V in the two strains, making it an ideal candidate to design specific FISH probes. The authors build on their previously published work and optimize a protocol to trace chromosomes by using a multiple-probe FISH approach to investigate chromosome architecture. This approach is well illustrated in Figure 1.

    Overall, this manuscript does a good job of describing a potentially useful technique with wide application. The claims about differences (and similarities) require statistical analysis to be appreciated, and much work is necessary to make the analysis approachable to readers outside the immediate field.

    Major comments:

    Nearly all claims regarding the organization, compactness and pair-wise distances of chromosomes lack any statistical measures of significance. This is particularly important for the clustering and scaling analysis. This makes interpretation of the claims made in the text impossible. For example, claims such as "the step size remained virtually unchanged" or "the paternal chromosomes adopt the maternal conformation in hybrids" cannot be currently analyzed.

    Throughout the manuscript (including in the abstract), the authors use the term "sister chromosomes" to (presumably) refer to the maternal and paternal chromosomes. This is a confusing term, since "sister chromatids" usually refers to the identical products of DNA replication, and "homologous chromosomes" is usually used to describe the parental chromosomes. The term would ideally be changed, and at the very least, it should be clearly defined.

    Presentation: The manuscript in its current state (excluding figure one) is essentially impossible to interpret by readers who are unfamiliar with this subfield. The authors could include a blurb on the methodology behind each data type to help the manuscript reach a larger audience. The pipeline, meaning, and potential caveats of the clustering analysis should also be explicated.

    Other suggestions:

    The use of Hi-C-like heatmaps is good, since they are commonly used and are clear and easy to understand. However, it would be best to explain how the FISH data were used to construct the maps. The implications of the map could also be better explained (e.g., that the red cluster means relatively looser regions, and the blue means more tightly compacted ones).

    The last sentence in the Introduction does not make clear sense. How does the similarity between N2 & HI open up the possibility of interrogating inheritance effects?

    Several of the additional analysis that could improve the paper are:

    1. Measure the nucleus diameter in N2/HI hybrid and HI/N2 hybrid.
    2. Normalize the spatial distance to the nucleus size, rather than directly using the distance.
    3. Explore some of the patterns in the spatial distance plots (e.g., red/blue lines and boxes). Are the sequences that are in them any different between N2 and HI in a way that might be able to account for these patterns?

    Significance

    The manuscript introduces a technical advance in the study of chromosomal territories - an important area of study that benefitted from recent advances in microscopy and in development of FISH approaches. However, it lacks mechanistic analysis and remains almost purely descriptive. It is also not clear what motivated the work beyond the technical feasibility. These issues make it impossible to assign biological significance to the seemingly minor differences that are documented.

    However, as a report of a technical advance, it could be useful to many chromosome biologists who might apply it to diverse organisms and biological questions.

    I am a chromosome biologist working on worms. However, my research does not directly deal with parental effects or with the development of novel FISH methodologies, so I did not examine claims regarding these specific points.

  3. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #2

    Evidence, reproducibility and clarity

    Building on their previous work using sequential fluorescent in situ hybridization (FISH) to follow the path of entire chromosomes in C. elegans (Sawh et al. 2020; Sawh and Mango 2020), the authors developed and then applied a method that distinguishes maternal and paternal homologs. In particular, they designed probes specific for ChrV in two strains, N2 and HI, and then demonstrated their effectiveness in homozygotes and hybrids. They found that HI ChrV is more compact in HI homozygotes as compared to N2 homozygotes, but decompacts in the F1 of N2 hermaphrodites x HI males. A different outcome was observed with respect to decompaction in the reverse cross, where both N2p and HIm ChrV chromosomes (p=paternal, m=maternal) exhibit decompaction. Through unsupervised clustering, the authors further found both dominant and minor patterns of chromosome-wide organization, with maternal chromosomes similar in terms of their dominant clusters regardless of the direction of the cross, but paternal chromosomes less so. Finally, the authors measured the degree to which homologous chromosomes interact, concluding that, although homologs overlap quite frequently, they rarely align (pair).

    In sum, the significance of the authors' work lies in its chromosome-level observations and the application of computationally designed probes that distinguish homologs by targeting strain-specific insertions. While homolog distinction by FISH has been previously demonstrated, this study is the first to demonstrate this approach in C. elegans as well as implement it via insertions in a chromosome-wide manner. As such, the manuscripts should be of interest to a broad range of researchers, especially those in the fields of genetics, genomics, and 3D genome organization. That said, the study falls short in several ways, and we recommend the authors i) present a more thorough summary of what is known about homolog positioning across species, ii) describe how homologs have been previously distinguished by FISH and, thus, more clearly elucidate the specific advances they have enabled, iii) ground their work in more rigorous quantitation, and iii) provide a better description of the technologies (strengths as well as limitations) carried over from their previous studies so that readers can better evaluate the current study. We detail our suggestions and questions, below:

    Major Comments:

    1. Page 1: We suggest the authors broaden the reach of their introduction to include well-known examples outside of C. elegans of the impact of parent-of-origin on 3D genome organization. Such examples would include X-inactivation, selective silencing of paternal genomes, the physical elimination of paternal chromosomes, and the like.
    2. Page 1: We also suggest the authors consider including mention of observations from the following research publications and review:

    Mayer W et al. Spatial separation of parental genomes in preimplantation mouse embryos. 2003 PMC2169371

    Reichmann J et al. Dual-spindle formation in zygotes keeps parental genomes apart in early mammalian embryos. 2018 PMID: 30002254

    Nagele R, Freeman T, McMorrow L, Lee HY. Precise spatial positioning of chromosomes during prometaphase: evidence for chromosomal order. 1995 PMID: 8525379

    Hua LL et al. Mitotic antipairing of homologous and sex chromosomes via spatial restriction of two haploid sets. 2018 PMID: 30530674

    Hua LL, Casas CJ, Mikawa T. Mitotic antipairing of homologous chromosomes. 2022 PMC9731508

    1. Page 5: The authors designed their strain-specific probes to target 172 insertions that are over 1 kb in size and distributed across ChrV. We ask the authors to describe these insertions in greater detail, especially as, later in the manuscript, the authors touch on the possibility that sequence differences between the two strains may account for differences in chromatin architecture. How many insertions were on the N2 and HI chromosomes, respectively? What is the range and distribution of insertion sizes for all insertions as well as specifically for the N2 and HI ChrV chromosome? Do the insertions contain repetitive sequences, or are they predominantly composed of unique sequences? What is their distribution with respect to genes, active regions, TADs? Is there an explanation for why some are clustered? If they contain genes, are the genes enriched in certain GO categories? Do the insertions differ in their characteristics across the different chromosomes? This information could be included as graphs and/or tables.
    2. Page 5: Given that N2 and HI ChrV chromosomes differ by the number of insertions and, ultimately, probes, how might these differences have skewed the authors' results, especially with respect to overlap between the homologs? Here, simulations of different ratios of insertions between N2 and HI could be clarifying.
    3. What difficulties were encountered when tracing the paths of overlapping homologs, and how were these difficulties accounted for and/or solved? Did the different numbers and distributions of insertions between N2 and HI exacerbate the challenge? What confidence levels accompanied their findings?
    4. Page 4: It would be helpful if the authors to put their insertion-based method into the context of other studies that have developed and used FISH to distinguish homologs.
    5. Pag 6: It would clarifying if the authors provided details about the mis-annotation of the Thompson genome and how it is pertains to probe design.
    6. Page 6: The authors state that "...N2 and HI...harbor sequence differences, some of which are predicted to affect chromatin architecture" and that HI lacks ppw-1. We ask the authors to provide a more thorough discussion. To what extent might such predictions rest on the insertional differences between the strains?
    7. Page 7: How much smaller is the HI genome and what percent of this difference is due to insertions (deletions)? Related to this, how much smaller is the HI ChrV chromosome as compared to N2?
    8. Page 7 and throughout: For those readers who have not read Sawh and Mango 2020 and Sawh et al. 2020, or who are unfamiliar with the broader category of imaging technologies that support the current study, we ask the authors to provide much more background and citations to key methods. Without this information, many readers will neither sufficiently understand the strategies used for imaging acquisition, processing, and analysis nor grasp the relevance of terms such as step size, polymer step size, power-law fitting, etc. and therefore be less able to assess the authors' data and conclusions.
    9. General statement: When the authors infer similarity or differences between power-law fittings, scaling exponents, step sizes, etc. what is the statistical significance of those comparisons?
    10. Experiments in general: We suggest the authors provide considerably more quantitation. For example, we urge them to provide the number of trials, sample sizes, numbers of embryos examined for all experiments. Equally important would be information regarding the stage of embryos examined and, where more than one embryonic stage was involved, the number of embryos for each stage. Are there stage-specific changes? We are also concerned about the impact of mixed populations of embryos on studies using unsupervised clustering. In other words, what was the contribution of developmental stage to the outcome of the clustering? Furthermore, if not all nuclei in an embryo were captured, we ask the authors to give the percent of nuclei captured from an embryo and reasons why only a subset of nuclei were included in the analysis.
    11. With regard to cluster analyses both here and elsewhere, will the authors please include statements of statistical significance whenever they note differences and/or similarities?
    12. Page 8: It would be helpful if the authors explained how they implemented the nearest-neighbor approach, including caveats and limitations (success rates, drop-out rates, etc.) and providing statistical assessment wherever possible.
    13. Page 8: "In some instances (6-8% of traces), traces were ambiguous and excluded from further analysis". We ask the authors to provide more detail. For example, what does ambiguous mean and, with respect to the 6-8% value, what was the total number of traces? What was the distribution of all traces (prior to filtering) with respect to percent of targets detected? Also, how coincident were the two homologs of a nucleus to each other in terms of capturing all the targets?
    14. Page 8: "...we classified traces into N2 or HI based on whether the trace was located closest to a strain marking volume for N2 or HI." Will the authors please quantify "closest" and explain what this means, whether there was a cut-off and, if there had been, how it was determined and implemented? What percent of cells were problematic, and were there traces that did not overlap the strain marking volumes at all? As stated in the Materials and Methods, only a subset of traced chromosomes were analyzed for overlap - why were only a subset analyzed for overlap, how was the subset selected, and how many/what percentage did these traces represent? It would also be helpful if the authors provided a quantitative summary of the traces.
    15. Page 8: Did the authors account for chromatic aberration and, if so, what protocol did they use?
    16. Page 8: "...counting how often more than two N2 or HI traces were detected in one nucleus." This is puzzling, and we suggest that the authors include explanations for how this might have happened. Did the nuclei not contain signal from the other strain marking probe at all? Was there a bias for this to happen with N2 or HI chromosomes? Could this have been a consequence of biology or of the algorithm for tracing? The authors' observations are reminiscent of the many implications raised by Jia et al. (2023; A spatial genome aligner for resolving chromatin architectures from multiplexed DNA FISH. PMID: 36593410), and we ask the authors to comment on the relevance of their observations to those in this recent publication.
    17. Page 8: "We found only a minority of 2% of HI traces and 7% of N2 traces were mis-assigned and excluded these from downstream analysis." 2% and 7% of what total?
    18. Page 8: How do pairwise distances remain almost identical between N2 and HI and yet generate different scaling exponents?
    19. Page 9: Figure 1F shows images of embryos derived from N2 hermaphrodites x HI males. It would be helpful if the authors added analogous images from the reciprocal cross as a supplementary figure.
    20. Page 2, 9, 9, and 12: The authors make several comments regarding the action of factors in trans: "... factors from the mother impact chromosome folding in trans (p. 2)"; "... the HIp decompacts when subjected to the N2m environment and implies that the paternal chromosome is influenced by the maternal environment in trans (p. 9)"; "...N2 chromosomes influence HI chromosomes in trans, while N2 chromosome structure seems to be resistant to influences by the HI chromosomes (p. 9)"; and "...implicating maternal factors that act in trans (p. 12). While provocative, these statements call for more concrete consideration. Are the authors using "in trans" in lieu of "indirectly", or are they alluding to factors, such as ppw-1, or direct physical contact? Without further substantiation or argument, mention of in trans activity might best be reserved for the Discussion.
    21. Page 10: "While HIp subpopulations were characterized by folding of one or the other chromosome arm, N2p clusters were more open and a subpopulation with a highly folded right arm was not present" (Figure 5CD). Was there a significant correlation between left vs. right arm folding and overall genome organization and function?
    22. Page 11: Will the authors please provide a more explicit definition of alignment as well as a more detailed description of how alignment is quantified in the main text?
    23. Page 11: With respect to direct physical contact, the authors mention transvection, which they conclude in the abstract is unlikely because pairing between homologs was observed to be rare. As transvection and pairing can both be short-lived, and the data are not compelling, the statement may need to be toned down considerably and/or be moved to the discussion.
    24. Page 11: The authors draw a distinction between % territory overlap and physical distances between homologs. In particular, the differences between % overlap across different stages is quite interesting and potentially suggests embryo stage-specific changes. Could the authors explore this further by breaking down mean pairwise distances into different embryo stages (Figure 6B and C)?
    25. Pages 2, 11-13: When the authors use "sisters", do they mean "homologs"? If the latter, we recommend they use "homologs", only, as "sisters" refers to the sister chromatids after replication. If, however, the authors are using "sisters" to mean sister chromatids, will they please explain how their data can distinguish sisters?
    26. Discussion: We encourage the authors to speculate further regarding the basis of decompaction. Is it a hybrid-specific phenomenon?

    Minor comments:

    1. Page 1: Although the introduction focuses on C. elegans, the genome length that is mentioned (2 meters) is more aligned with that of mammalian species. The authors could cite a range of lengths or make more clear which species is being discussed.
    2. Page 5: As the figures label the strains as Hawaiian and Bristol, the authors might wish to include this nomenclature in the main text. Curiously, the authors use several different spellings for the Hawaiian/Hawai'ian/Hawaiin/Hawaii strain.
    3. Page 5: Will the authors please explain why the common whole-chromosome tracing probes have only one tail, while the strain-specific probes have two tails, as shown in Figure 1D?
    4. Figures in general: The axes of a number graphs and heat maps need to be labeled.
    5. Materials and Methods: The section on "Cluster analysis" is missing units for the resolution.
    6. Page 8: Will the authors please give a reference for and explain watershed segmentation?

    Significance

    In sum, the significance of the authors' work lies in its chromosome-level observations and the application of computationally designed probes that distinguish homologs by targeting strain-specific insertions. While homolog distinction by FISH has been previously demonstrated, this study is the first to demonstrate this approach in C. elegans as well as implement it via insertions in a chromosome-wide manner. As such, the manuscript should be of interest to a broad range of researchers, especially those in the fields of genetics, genomics, and 3D genome organization. That said, the study falls short in several ways, and we recommend the authors i) present a more thorough summary of what is known about homolog positioning across species, ii) describe how homologs have been previously distinguished by FISH and, thus, more clearly elucidate the specific advances they have enabled, iii) ground their work in more rigorous quantitation, and iii) provide a better description of the technologies (strengths as well as limitations) carried over from their previous studies so that readers can better evaluate the current study.

  4. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #1

    Evidence, reproducibility and clarity

    The authors designed an elegant series of FISH probes taking advantage of insertions that are divergent between HI and N2 strains of C. elegans to identify the maternal versus paternal chromosomes in hybrid embryos. Overall, the conclusions are well supported by the data. I only have minor comments, which are numbered below in relation to each figure.

    In figure 1 they demonstrate that the probes can specifically recognize their corresponding chromosome in hybrid embryos

    In figure 2 they demonstrate that overall chromosome 5 adopts a similar shape from both strains.

    In figure 3, they demonstrate that in the hybrids, the chromosomes are generally the same as in the homozygous embryos. However, when the HI chromosome is brought in paternally in the hybrids, it is more decompacted. In this cross, the maternal N2 chromosome is normal. In figure 5, when the N2 chromosome is now brought in paternally, it is similarly decompacted. However, in this cross, the HI chromosome that is brought in maternally is also decompacted. This appears to be the biggest difference, but is somewhat mitigated by a decrease in the scaling component, which I believe means it takes a straighter path? This is in contrast to the reciprocal cross where the maternal N2 chromosome is normal.

    1. In figures 2-4, it is important to note what stages of embryos are being analyzed and whether any analysis was done to determine if the chromosomes varied with embryonic stage?
    2. The authors need to clearly define chromosomal step size, scaling coefficient and pair wise distance and then describe the difference between these measurements. For example, it would be nice in relation to figure 4F if it was described exactly what it means to have an increase in step size along with a decrease in scaling exponent. This will enable the reader to more easily interpret the results.
    3. Is there any way to determine if changes in the step size and scaling are significant? It would be good to know if the changes are actually significant.

    In figure 5, the authors examine sub-clusters of where individual chromosomes locate.

    1. In figure 5 (and maybe in earlier figures as well), it would also be helpful to mark the different clusters in the figures to show what is meant by a folder arm or a compacted central domain and refer to every panel in the text (only some descriptions in the text reference specific panels).

    In figure 6, the authors determine whether the two chromosomes 5's overlap in nuclear territory and whether they the align along the length of the chromosome. From this analysis, they conclude that the chromosomes overlap a fair amount of time, but do not align. This makes it unlikely that transvection might occur.

    1. In figure 6, the authors need to do a better job of describing how the data is being presented. The % overlap is being graphed by density, but density of what? Also, the text mentions the total number of nuclei that overlap, but how is that number derived from the presented data? Finally, the data are broken down by embryonic stage, but there is no mention of this in the text. It is not mentioned until the discussion. Overall, this makes it very difficult to determine what the data are showing.
    2. In the discussion, the authors perhaps should spend more time interpreting their results in light of others work on the maternal and paternal inheritance of chromatin in C. elegans. For example, Arico et al 2011 Plos Genetics from the Kelly Lab. In addition to examining chromatin in the early embryo, in this paper the authors examine translocations, which might be interesting to look at using the technique presented here. Also, a number of recent papers from the Strome lab have examined chromatin inheritance from sperm. It would be interesting to interpret the finding that paternal chromosomes are influenced by the maternal environment, in light of this work.

    Significance

    These studies are the important extension of the elegant chromosome tracing that the Mango lab has pioneered. The authors have clearly demonstrated that the technique works well to identify individual chromosomes in a hybrid background. This provides the opportunity for the system to be used in numerous different ways, not just in C. elegans. This is the most significant advance of this paper and should be of interest to a fairly broad audience. However, using this technique, the authors also provide the initial characterization of sister chromosomes in C. elegans embryos and draw important initial conclusions, such as finding that the chromosomes do not pair, as they do in Drosophila. This makes the paper of interest to the C. elegans audience, as well as the general field of chromosomal organization. My expertise is in C. elegans biology and chromatin biology in general. I also am familiar with the field of chromosome biology. As a result, I believe I am capable of judging the significance of this paper in these areas.

  5. Version published to 10.1101/2022.11.01.514763v1 on bioRxiv