HID-1 is required for homotypic fusion of immature secretory granules during maturation

[…] Overall, the experiments are well performed and of high quality, particularly the beautiful 3D-EM reconstructions. The conclusions are mostly supported by the data. The reviewers have specific suggestions to improve the manuscript before publications can be recommended.

Essential revisions:

1) While the STX6 data provides an interesting idea about a potential SNARE interaction, much work needs to be done before the HID1-STX6 link can be strengthened. For example, Further evidence for a role for STX6 in homotypic fusion of immature secretory vesicles would be interesting, and evidence that HID1 plays a role would be entirely novel. Similarly, the authors could strengthen their conjecture by demonstrating that an open form of STX6 bypasses a requirement for HID1, or that STX fails to be recruited to immature secretory vesicles in hid1 mutants or some other functional interaction. Considering the time it will take to do these experiments, the reviewers recommend the authors save the STX6 data for the next paper and eliminate them from the current manuscript.

We thank the reviewer for this constructive suggestion. In accordance with this suggestion we have eliminated the STX6 data from the current manuscript. We are currently testing the open form of STX6 and how STX6 is recruited to immature dense core vesicles.

2) ISGs in hid1 mutants. The authors argue that in hid1 KO cells, SGs persist in an immature form and mis-localize throughout the cell. However, the maturation status of the vesicles is monitored only by the presence of pro-insulin cargo and indirectly by size and electron-density in electron micrographs. It is possible that secretory granules in HID1 mutants partially mature. The authors should assess:

Do the secretory granules in hid1 KO cells retain known immature granule markers including synaptotagmin-IV and VAMP4 (Ahras et al., JCB 2006)?

Do the secretory granules in hid1 KO cells acquire known markers of maturity, such as synaptobrevin-2 (Walter et al., EMBO 2014)?

As suggested by the reviewer, we investigated whether VAMP4, synaptotagmin IV and VAMP2 were present on proinsulin-positive ISGs employing DeltaVision OMX V3 imaging system followed by deconvolution. As shown in Author response image 1A and C, VAMP4 and synaptotagmin IV both localized to the peri-nuclear compartment, similar as the distribution of proinsulin in WT β cells. VAMP4 and synaptotagmin IV partially co-localized with proinsulin puncta to a similar extent (similar Pearson coefficient). In HID-1 KO β cells, proinsulin and synaptotagmin IV staining dispersed throughout the cytosol and co-localized to a similar extent as in WT cells (Author response image 1A and C, bottom, B and D). In contrast, the localization of VAMP4 was unchanged and remained in the peri-nuclear region. As a consequence, the co-localization of proinsulin and VAMP4 was significantly reduced in KO cells (Author response image 1A and B). Thus, proinsulin-positive secretory granules in HID-1 KO cells seem to retain synaptotagmin IV but not VAMP4, suggesting VAMP4 recruitment or retaining to ISG depends on HID-1. Besides, previous study has shown that VAMP4 antibody did not inhibit homotypic fusion of PC12-derived ISGs, which indicates that VAMP4 is not an essential component for ISG-ISG fusion process (Wendler et al., 2001). As for MSG marker VAMP2, its dispersed localization was not altered by HID-1 KO and there was few co-localization between proinsulin and VAMP2 in WT and KO cells (Author response image 1E and F), suggesting VAMP2 is not recruited to proinsulin granules in either WT or KO cells. We are not sure whether to include these data or not and would like to hear the advice from the reviewers.

Author response image 1

Download asset Open asset

3) The Introduction needs to be rewritten with the following points in mind. The specific emphasis on SNAREs in the current Introduction, without the discussion of accessory proteins, is unintentionally misleading. Clarity about HID-1 could be provided by briefly discussing its domain structure and what is known about its ability to associate with membranes. A brief discussion of the granule proteases, ISG acidification, and/or other maturation factors would also be more relevant to the authors findings. Furthermore, HID-1 was also identified in a screen mislocalizing RBF-1 rabphilin, a RAB-27 effector and both of these proteins have important roles in DCV maturation and exocytosis. I think this literature might serve as better introduction to HID-1, and other maturation factors, as opposed to its significance to high-temperature-induced dauer formation and the suppression of HID-1 by the daf-16 transcription factor.

We rewrote the Introduction part as suggested.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

[…] Reviewer #1:

The authors have made a number of positive revisions and most of our major concerns were addressed, so we feel the study is ready for publication. Nice paper on a difficult subject – we applaud the authors.

However, there is one small concern that must be addressed: the revision of Figure 5D is confusing because the 3 micron scale bar did not increase in size with the higher magnification. Clearly, there is a minor error here that can be readily corrected.

We are sorry about the mistake. We have increased the length of the scale bar in Figure 5D accordingly.

Reviewer #3:

The revised manuscript addresses my concerns from the first review. There are three remaining concerns that warrant textual changes.

1) Homotypic fusion in endocrine cell. The authors argue that HID1 promotes homotypic fusion in β islet endocrine cells. However, the dogma is that homotypic fusion does not occur in endocrine cells, but rather only in neuroendocrine cells. On the other hand, the data against homotypic fusion in endocrine cells are not particularly strong, and that is in part why this manuscript is interesting. The potential conflict with the dogma should be mentioned in the Introduction and expanded upon in the Discussion. Evidence for homotypic fusion is observed in cells that form large dense-core granules such as neuroendocrine cells, including PC12 cells (1,2,3), mammotroph pituitary cells (4), mast cells (5), and in the formation of LDCVs in yeast (6). Evidence for an absence of homotypic fusion in endocrine cells rests on two observations: First, disrupting syntaxin-6 using a dominant negative construct did not exhibit a defect in insulin granule formation in INS-1 β-cells, but rather a delay in pro-insulin processing (7). Second, electron microscopy data in β cells (8, 9) does not reveal an increase in diameter in mature secretory granules. The strengths and weaknesses of these previous data should be mentioned in the Discussion.

We thank the reviewer for raising this important dogma and for providing detailed review of the references. As suggested, we have briefly introduced the dogma in the Introduction (first paragraph) and expanded in the Discussion section (subsection “Homotypic fusion in the SG maturation process”, first paragraph).

2) Granule diameter. The authors claim that previous researchers have noted that mature granules are larger in diameter (and are light density) and cite Furuta 1998, Noske 2008, and Wijesekara 2010 (subsection “HID-1 deficiency increases the number of ISGs”, second paragraph and subsection “HID-1 deficiency blocks homotypic fusion of ISGs”, first paragraph). Furuta 1998 and Wijesekara 2010 both note that immature granules are lighter, but did not report diameters. Noske et al. suggest that old granules are smaller, not larger. In fact, Noske et al. were not studying immature granules, they report a size change in older cells. Orci 1986 also states that maturation involves condensation of the matrix, and reduction in granule diameter. The authors observed the opposite – they see larger mature vesicles suggesting homotypic fusion. The authors may be the first to study granule maturation in any comprehensive way; block in the prohormone convertase or the zinc transporter are unlikely to affect membrane trafficking events.

We thank the reviewer for bringing up this important point. The reviewer is correct that there is controversy whether mature granules are larger in size or not. The papers we cited (Furuta et al., 1998; Wijesekara et al., 2010) noted the lighter ISG but did not report diameter. So we have changed the sentence to “It has been suggested that ISGs exhibit lighter dense cores than MSGs.” (subsection “HID-1 deficiency increases the number of ISGs”). Tooze et al. (Tooze et al., 1991) have reported the smaller size of ISG in PC12 cells as copied below, so we added this reference in the first paragraph of the subsection “Homotypic fusion in the SG maturation process”.

As the reviewer pointed out, the diameter of granules after maturation may not be always larger. As wrote by Tooze et al. (Tooze et al., 2001), “Maturation of secretory granules also involves a change in size, and MSGs can either become smaller or larger than ISGs. A decrease in size should be proportional to the number of vesicles budded from each ISG, although this has not been confirmed experimentally. MSGs become larger than ISGs through homotypic fusion of ISGs that probably precedes CCV formation”. We have added this possibility in the Discussion section (subsection “Homotypic fusion in the SG maturation process”, first paragraph).

We thank the reviewer for stating that we studied granule size in a comprehensive way. Previous analysis of granule size largely relies on EM of thin sections, which is problematic because planar sectioning at random depths through spherical granules tends to underestimate the diameter of granules and largely increase the variation. As we explained in the text:

“In this study we employed automatic FIB-SEM to visualize large volume (13.7x11.8x3~8 μm³) at nanometer resolution and analyze thousands of intact granules from 15 cells. We analyzed not only the diameter, but also the grey level of the cores and the subcellular localization of the granules. This comprehensive analysis enables us to clearly separate ISGs from MSGs and conclude that MSGs have larger size in β cells (Figure 4).”

3) Molecular nature of immature granules. The argument that granules are immature in the hid1 KO is solid. First, the matrix is lighter. Second, proinsulin is still present in the granules. Third, the presence of immature granule membrane proteins reinforces this conclusion, in particular synaptotagmin 4. Thus, the synaptotagmin 4 data should be included. It is unfortunate that the authors did not include syntaxin 6 staining, to demonstrate that proteins required for homotypic fusion were both still on the pro-insulin granules, which would provide evidence that granule maturation is blocked at a step before homotypic fusion. But that is a quibble; the manuscript should be published. VAMP4 was not mislocalized in the hid1 KO. This result does not challenge the author's conclusion about a role for HID1 in homotypic fusion. VAMP4 has not been shown to be required for fusion of immature secretory granules, whereas syntaxin-6 is required. The authors conclude that HID1 is required for VAMP4 inclusion in immature secretory granules. Alternatively, VAMP4 may be removed by an intact 'sorting-by-exit' mechanism in the hid1 KO. These data are an interesting facet of the KO phenotype and should be included. The mature granule marker VAMP2 is not observed on immature granules in hid1 KO cells. These data are consistent with their model that granules are not mature in hid1 KO cells. However, the authors did not validate that the VAMP2 antibody was associated with mature insulin granules in the wild type. Thus the negative result is not clearly interpretable and these data unfortunately should not be included.

Reviewer 1 didn’t think the new data should be included, whereas Reviewer 2 suggested including them as supplementary material. Reviewer 3 suggested including Syt-IV and VAMP4, but not VAMP2. We are not sure how we should respond and would like to hear the decision of the editor. For the time being, we included the data of Syt-IV and VAMP4 in the new Figure 4—figure supplement 1, described in the first paragraph of the subsection “HID-1 deficiency increases the number of ISGs”.