Corneal lens curvature depends on localized chitin secretion

  1. Neha Ghosh
  2. Jessica E. Treisman

Reviewed by

Read the full article See related articles

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

Log in to save this article

Abstract

The Drosophila corneal lens is an apical extracellular matrix structure with a biconvex shape that enables it to focus light. Here we investigated how this shape is influenced by the source of one of its major components, the polysaccharide chitin. Knocking down the chitin synthase Krotzkopf verkehrt reduced corneal lens thickness and curvature. Conversely, enhancing chitin export by overexpressing Rebuf expanded the corneal lens. We found that the cone and primary pigment cells in the center of each ommatidium produce most of the chitin, and preventing chitin synthesis by these central cells reduced corneal lens curvature. Increasing chitin export from central cells increased the thickness of the central corneal lens, while increasing export from peripheral lattice cells expanded its edges. The biconvex shape thus results from high levels of chitin production by central cells relative to peripheral cells, indicating that localized chitin secretion is critical for normal corneal lens curvature.

Article activity feed

  1. Review Commons

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

    Learn more at Review Commons

    Reply to the reviewers

    Reviewer 1, point 1: In general, the statistical analysis is not transparent. The size of the sample, i.e. the number of observations or data points, is never specified. This information is essential for further evaluation of the statistical details.

    The size of each sample quantified, given as number of ommatidia/number of retinas, is indicated in the figure legends. This must have escaped the attention of reviewer 1, so we have added a sentence in the legend of Fig. 2 to state it more clearly. We think that the figure legends are the best place to put this information for ease of comparison to the figures.

    *Reviewer 1, point 2: To gain a better understanding of chitin deposition, it would be beneficial to have data on Kkv overexpression in cone cells versus outer pigment cells. Does it cause reb/exp-like effects on chitin deposition and corneal lens formation? Furthermore, can the authors rule out the involvement of chitin synthase 2 in chitin matrix formation and the retention of the matrix in kkv knockdowns? *

    We will generate clones of cells that over-express Kkv in either central cells (cone and primary pigment cells) or lattice cells (secondary and tertiary pigment cells), using the same drivers that we used to over-express Reb, and will examine chitin secretion at 54 h after puparium formation (APF) and in adults.

    As there are no available mutations in Chitin synthase 2 (Chs2), we will knock it down with RNAi in all retinal cells using lGMR-GAL4 and look for corneal lens defects. However, we think that Chs2 is unlikely to contribute chitin to the corneal lens, because its expression is restricted to the digestive system, and because *kkv *knockdown essentially eliminates chitin from the corneal lens.

    *Reviewer 1, point 3: Recent results published by the authors regarding ZP domain proteins, such as dusky-like (dyl), have not been adequately discussed in the context of chitin secretion and Kkv expression, a matter that must be addressed. It has been demonstrated that dyl mutants do not affect Kkv expression, but chitin levels are reduced. Does Dyl exhibit Kkv-like phenotypes? Furthermore, what is the expression of Dyl or Dmupy in Kkv knockdowns? Is there any interaction between the ZP domain protein matrix and the chitin matrix required for lens formation? *

    In *dyl *mutants, chitin deposition is delayed, but it does accumulate later in development, so the phenotype is different from *kkv *mutants. We have clarified this in the manuscript (p. 6). To address the other points, we will examine the expression of Dyl and of Dumpy-YFP in mid-pupal and late pupal retinas in which *kkv *is knocked down in all cells with lGMR-GAL4. The ZP protein matrix is originally deposited before chitin secretion begins, so we will examine whether loss of chitin affects its later maintenance.

    *Reviewer 1, point 4: What is retained in the chitin matrix if chitin is missing in kkv knockdown? Is it the ZP domain matrix (see the above question) or are the chitin matrix proteins also involved, such as Obst-A, Obst-C (Gasp), Knk and others? Obst proteins are particularly essential for the regular packaging of chitin and thus for the formation of the chitin layer, which is shown in Fig. 1. Beyond this story, it would also be interesting to see how the aforementioned chitin matrix proteins (Obst-A, Obst-C (Gasp), Knk and others) impact lens formation. *

    Adult corneal lenses derived from *kkv *knockdown retinas do not contain chitin, but there is remaining corneal lens material. We do not think that this is the ZP domain matrix, as this is normally lost in late pupal development, but we will check whether Dpy-YFP is retained in *kkv *knockdown adults. We will try to detect Obst-A and Gasp proteins using available antibodies. However, this may not be successful, as we have found that antibodies do not penetrate the corneal lens well. Our transcriptomic studies have identified numerous secreted proteins that are expressed at high levels in the mid-pupal retina and could be components of the corneal lens. We may be able to detect some of these using fluorescently tagged forms, but it is possible that the currently available tools will not be sufficient to answer this question.

    We have begun to work on how some of these proteins affect corneal lens structure, but this will take a significant amount of time and we think it would work better as a separate manuscript. We see our current manuscript as a short and focused story about the importance of the source of chitin in determining corneal lens shape.

    *Reviewer 1, minor comment 1: Figure 1 is not easily comprehensible for those who are not already familiar with the subject of eye development. Fig -1A' please label the cone cells and pigment cells. *

    We have labeled these cells in Fig. 1A’’.

    *Reviewer 1, minor comment 2: Fig. 1H - The meaning of the abbreviations and numbers is not given in the legend. It would also be beneficial to include a meaningful cartoon illustrating the corneal lens situation before and after chitin secretion, as shown in Figure 3.

    We have defined the abbreviations in the figure legend. Fig. 1H did show the corneal lens situation before, during and after chitin secretion, but we have added the cone and pigment cells to the 72 h APF and adult diagrams to make them more meaningful (now Fig. 1I).

    *Reviewer 1, minor comment 3: Fig.1 F when does the authors recognize a first chitin assembly as initial corneal lens at the eye and how does it look like? Chitin expression is high already at 54h APF, which means 20 hours earlier. *

    We think that the reviewer is asking when the chitin first starts to form a dome shape. We have added an orthogonal view of chitin in a 54 h APF retina viewed with LIGHTNING microscopy, showing that the external curvature is already present at this stage (new Fig. 1F).

    *Reviewer 1, minor comment 4: Page 6 / Fig 2E: cells autonomously synthesize chitin and no lateral diffusion. Please label which lens contains chitin and which not *

    Fig. 2E shows part of a retina in which kkv has been knocked down in all cells, so none of the corneal lenses contain chitin. We have clarified this in the legend to Fig. 2.

    *Reviewer 1, minor comment 5: Page 7: The authors state that reb/exp knockdown affects external and internal curvature. However, Fig. S1 statistics does not support this statement. *

    We were referring to the double knockdown, which Fig. 2L, M show is significant, and not to the single knockdowns quantified in Fig. S1. We have clarified this in the text.

    *Reviewer 1, minor comment 6: Fig.2 and Fig. S1: what is Chp (Chaoptin)? *

    We have stated in the legend to Fig. 2 that Chaoptin is a component of photoreceptor rhabdomeres.

    *Reviewer 1, minor comment 7: Fig. S1E,I: which part of the eye is marked by the chitin staining outside the cone and pigment cells?

    Chitin is still present in the mechanosensory bristles in Fig.S1I, as these do not express lGMR-GAL4. We have stated this in the figure legend.

    *Reviewer 1, minor comment 8: Fig. 2 L,M, Why do exp/reb show different statistical results at outer angle in exp and reb knockdown when compared with the IGMR driver line, although chitin reduction is eliminated in exp knockdown already from 54h APF onwards?

    The double knockdown of *exp *and *reb *has a more significant effect on the adult corneal lens outer angle than the single *exp *knockdown, even though the exp knockdown lacks chitin at 54 h APF. We believe that this is because Reb is sufficient for some chitin synthesis at later stages of development. This was mentioned in the text (p. 6) and we have added further clarification in the legend to Fig. S1.

    *Reviewer 1, minor comment 9: Fig 3 G-H: please clarify where the chitin reduction can be observed at the edge of adult corneal lens and provide comparable wt staining's. Fig. S2 D. What was the normalization and the sample number? *

    We have added a high magnification image of a mosaic ommatidium with one wild-type and one kkv knockdown edge, showing the region at the edge of the corneal lens in which chitin fluorescence was quantified and the central region used for the normalization (Fig. 3I). The sample numbers are given in the legend to Fig. S2D.

    *Reviewer 1, minor comment 10: Page 6, last paragraph: I fully agree that ZP domain proteins may retain other corneal lens components. But deeper discussion is missing. It should be noted that the authors hypothesis fits well to the proposed function of the ZP matrix in providing chitin matrix adhesion to the underlying cell surface. A loss of the ZP domain protein Piopio causes loss of the chitin matrix as show recently in trachea and at epidermal tendon cells (Göpfert et al., 2025; https://www.sciencedirect.com/science/article/pii/S1742706125003733). Furthermore, a recent publication identifies ZPD proteins as modular units that establish the mechanical environment essential for nanoscale morphogenesis (Itakura et al., https://www.biorxiv.org/content/10.1101/2024.08.20.608778v1.full.pdf). This should be cited and discussed accordingly.

    It could be that outer and inner part of the chitin is different in ultrastructure due to expression pattern. In dragonfly the surface morphology analysis by scanning electron microscopy revealed that the outer part of corneal lenses consisted of long chitin fibrils with regular arrays of papillary structures while the smoother inner part had concentric lamellated chitin formation with shorter chitin nanofibrils (Kaya et al., 2016; https://www.sciencedirect.com/science/article/pii/S0141813016303646?via%3Dihub#fig0020) . Thus, a ultrastructure analyses would be very beneficial, or at least a detailed discussion.

    We have added a discussion of these points and papers to the text (p. 6 and 9). Although we are not specifically addressing differences between the inner and outer parts of the corneal lens in this manuscript, we have now included a high-resolution LIGHTNING image showing how the layered structure of the corneal lens is affected when chitin production by central cells is increased (Fig. 4F).

    *Reviewer 2, point 1: Adult corneal lenses lacking chitin still form a thin structure in kkv RNAi. The authors suggest that this may be due to the presence of the ZP domain proteins Dyl, Dpy and Pio. Immunostaining for these ZP domain proteins could provide supporting evidence. *

    To clarify, we meant to say that the earlier presence of the ZP domain matrix could retain components other than chitin in the corneal lens. The ZP domain proteins are no longer present in the adult. We have made this clearer in the text. As described under reviewer 1, points 3 and 4, we will examine Dyl and Dpy-YFP expression in kkv knockdown retinas at mid-pupal and adult stages, and we will also look at the expression of another ZP domain protein, Piopio.

    *Reviewer 2, minor comment 1: At 50 h APF, Kkv (Fig. 2B, B') and Reb (Fig. S1A, A') appear to be expressed at higher levels in lattice cells than in central cells, even though chitin is mainly present in the central cells at this time (Fig. 1B-B'). Discuss possible explanation for their expression pattern and their roles at this stage. *

    We agree that this is a surprising result. We have added a discussion of possible explanations, such as the lack of another component necessary for chitin secretion in lattice cells at this stage, or the presence of high levels of chitinases (p. 7).

    *Reviewer 2, minor comment 2: Fig. 1F and G: Indicate that the cryosection images represent single ommatidia, and label "external" and "internal" to help orient readers. *

    We have made these changes to the figure panels (now G and H), and indicated in the legend that they are single ommatidia.

    *Reviewer 2, minor comment 3: Figure 2. The cartoon diagram showing the angle measurement (currently Fig S1K) should be moved to the main figure to help readers understand the quantifications. *

    We have moved this diagram to Figure 2L.

    *Reviewer 2, minor comment 4: Figure 3H. It would be helpful to clearly mark the edge of the corneal lens in the chitin intensity image. *

    As described under reviewer 1, minor comment 9, we have added a high magnification picture showing the edge region used for chitin quantification (Fig. 3I), which should also address reviewer 2’s concern.

  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 #2

    Evidence, reproducibility and clarity

    Summary:

    The manuscript by Ghosh and Treisman demonstrates that localized chitin secretion shapes the Drosophila corneal lens. Building on their previous work showing that zona pellucida domain proteins influence corneal lens architecture, which correlates with a delay in chitin deposition, the authors investigate how chitin and its secretion contribute to defining the lens curvature. Through cell-type-specific RNAi and overexpression experiments combined with beautiful imaging and quantifications, they provide convincing evidence that the central cone and primary pigment cells are the principal sources of chitin in the middle region of the corneal lens. Overall, this study offers strong evidence that localized chitin secretion and restricted diffusion underlie the precise shaping of the corneal lens. I have only one major comment and a few relatively minor suggestions to improve clarity.

    Major comments:

    • Adult corneal lenses lacking chitin still form a thin structure in kkv RNAi. The authors suggest that this may be due to the presence of the ZP domain proteins Dyl, Dpy and Pio. Immunostaining for these ZP domain proteins could provide supporting evidence.

    Minor comments:

    • At 50 h APF, Kkv (Fig. 2B, B') and Reb (Fig. S1A, A') appear to be expressed at higher levels in lattice cells than in central cells, even though chitin is mainly present in the central cells at this time (Fig. 1B-B'). Discuss possible explanation for their expression pattern and their roles at this stage.
    • Fig. 1F and G: Indicate that the cryosection images represent single ommatidia, and label "external" and "internal" to help orient readers.
    • Figure 2. The cartoon diagram showing the angle measurement (currently Fig S1K) should be moved to the main figure to help readers understand the quantifications.
    • Figure 3H. It would be helpful to clearly mark the edge of the corneal lens in the chitin intensity image.

    Significance

    This study provides novel insights into how the differential secretion of a polysaccharide determines the curvature of a complex optical structure. The elegant use of cell-type-specific genetic manipulations, together with high-quality imaging and rigorous quantification is the key strength.

    The study advances our understanding of how chitin secretion and limited diffusion shape apical ECM structures during tissue morphogenesis. It also extends findings from the tracheal and cuticular chitin systems into a new optical context.

    The manuscript will be of interest to developmental biologists, particularly those studying epithelial tissue morphogenesis and apical ECM organization.

    I have expertise in Drosophila epithelial morphogenesis.

  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 #1

    Evidence, reproducibility and clarity

    Summary:

    Chitin plays a crucial role in the morphogenesis of the Drosophila corneal lens by supporting the structural integrity and biconvex shape of the lens. The Drosophila corneal lens is a biconvex structure that focuses light. Chitin, a major component, is produced mainly by the central cone and primary pigment cells. The production and arrangement of chitin by central cells directly impacts the thickness and curvature of the lens. Adequate chitin secretion is necessary to ensure the correct shape and function of the corneal lens, while disturbances in chitin production can lead to deformed lenses. Blocking chitin synthesis leads to a significant reduction in chitin deposition in the corneal lens, resulting in a thinner and deformed lens. In particular, the corneal lens shows reduced outer and inner curvature, which compromises its biconvex shape. These changes in chitin production and arrangement result in abnormal morphology of the corneal lens in the adult stage. The key messages of the paper's results are: The Drosophila corneal lens is a biconvex structure that focuses light. 2.) chitin, a significant component, is produced mainly by central cells (cone and primary pigment cells). 3.) Downregulation of the chitin synthase gene Krotzkopf reduces lens thickness and curvature. 4.) Overexpression of Rebuf increases chitin secretion and lens thickness. 5.) Localized chitin secretion is crucial for the typical shape of the corneal lens.

    Comments

    Main comments

    The manuscript provides an exciting insight into how the formation of the lens is regulated by the secretion of chitin. However, the data set appears to have shortcomings that must be considered for the next steps. 1.) In general, the statistical analysis is not transparent. The size of the sample, i.e. the number of observations or data points, is never specified. This information is essential for further evaluation of the statistical details.

    2.) To gain a better understanding of chitin deposition, it would be beneficial to have data on Kkv overexpression in cone cells versus outer pigment cells. Does it cause reb/exp-like effects on chitin deposition and corneal lens formation? Furthermore, can the authors rule out the involvement of chitin synthase 2 in chitin matrix formation and the retention of the matrix in kkv knockdowns?

    3.) Recent results published by the authors regarding ZP domain proteins, such as dusky-like (dyl), have not been adequately discussed in the context of chitin secretion and Kkv expression, a matter that must be addressed. It has been demonstrated that dyl mutants do not affect Kkv expression, but chitin levels are reduced. Does Dyl exhibit Kkv-like phenotypes? Furthermore, what is the expression of Dyl or Dmupy in Kkv knockdowns? Is there any interaction between the ZP domain protein matrix and the chitin matrix required for lens formation?

    4.) What is retained in the chitin matrix if chitin is missing in kkv knockdown? Is it the ZP domain matrix (see the above question) or are the chitin matrix proteins also involved, such as Obst-A, Obst-C (Gasp), Knk and others? Obst proteins are particularly essential for the regular packaging of chitin and thus for the formation of the chitin layer, which is shown in Fig. 1. Beyond this story, it would also be interesting to see how the aforementioned chitin matrix proteins impact lens formation.

    Minor comments:

    Page 6: Figure 1 is not easily comprehensible for those who are not already familiar with the subject of eye development.

    Fig -1A' please label the cone cells and pigment cells.

    Fig. 1H - The meaning of the abbreviations and numbers is not given in the legend. It would also be beneficial to include a meaningful cartoon illustrating the corneal lens situation before and after chitin secretion, as shown in Figure 3.

    Fig.1 F when does the authors recognize a first chitin assembly as initial corneal lens at the eye and how does it look like? Chitin expression is high already at 54h APF, which means 20 hours earlier.

    Page 6 / Fig 2E: cells autonomously synthesize chitin and no lateral diffusion. Please label which lens contains chitin and which not

    Page 7: The authors state that reb/exp knockdown affects external and internal curvature. However, Fig. S1 statistics does not support this statement.

    Fig.2 and Fig. S1: what is Chp (Chaoptin)?

    Fig. S1E,I: which part of the eye is marked by the chitin staining outside the cone and pigment cells?

    Fig. 2 L,M, Why do exp/reb show different statistical results at outer angle in exp and reb knockdown when compared with the IGMR driver line, although chitin reduction is eliminated in exp knockdown already from 54h APF onwards?

    Fig 3 G-H: please clarify where the chitin reduction can be observed at the edge of adult corneal lens and provide comparable wt staining's. Fig. S2 D. What was the normalization and the sample number?

    Page 6, last paragraph: I fully agree that ZP domain proteins may retain other corneal lens components. But deeper discussion is missing. It should be noted that the authors hypothesis fits well to the proposed function of the ZP matrix in providing chitin matrix adhesion to the underlying cell surface. A loss of the ZP domain protein Piopio causes loss of the chitin matrix as show recently in trachea and at epidermal tendon cells (Göpfert et al., 2025; https://www.sciencedirect.com/science/article/pii/S1742706125003733). Furthermore, a recent publication identifies ZPD proteins as modular units that establish the mechanical environment essential for nanoscale morphogenesis (Itakura et al., https://www.biorxiv.org/content/10.1101/2024.08.20.608778v1.full.pdf). This should be cited and discussed accordingly.

    It could be that outer and inner part of the chitin is different in ultrastructure due to expression pattern. In dragonfly the surface morphology analysis by scanning electron microscopy revealed that the outer part of corneal lenses consisted of long chitin fibrils with regular arrays of papillary structures while the smoother inner part had concentric lamellated chitin formation with shorter chitin nanofibrils (Kaya et al., 2016; https://www.sciencedirect.com/science/article/pii/S0141813016303646?via%3Dihub#fig0020) . Thus, a ultrastructure analyses would be very beneficial, or at least a detailed discussion.

    Significance

    The manuscript's strength and most important aspects are the genetic expression, and localization studies of the chitin under control of the chitin synthase kkv, reb and exp in Drosophila pupal and adult eye . However, beyond this manuscript, the development of mechanistic details, such as interaction partners that trigger secretion and action at the ZP matrix and adjacent apical membranes will be interesting.

    The manuscript uses nice genetics tools to describe the Chitin secretion differences in Drosophila eye and their specific impact on corneal lens formation. Such a precise molecular analysis has not been investigated before in insects. Therefore, the study deeply extends knowledge about the role of Chitin synthases and chitin secretion in insect eye.

    The audience will not only rather specialized in basic research in zoology, developmental biology, and cell biology in terms of how the Chitin synthases produce chitin. Nevertheless, as chitin is relevant to material research and medical and immunological aspects, the manuscript will be interesting beyond the specific field and thus for a broader audience.

    I'm working on chitin in the tracheal system and epidermis in Drosophila.

  4. Version published to 10.1101/2025.08.22.671854 on bioRxiv