key: cord-103396-jiuqk6kg
authors: Adler, Paul N.
title: Short distance non-autonomy and intercellular transfer of chitin synthase in Drosophila
date: 2020-05-26
journal: bioRxiv
DOI: 10.1101/2020.05.24.113803
sha: 
doc_id: 103396
cord_uid: jiuqk6kg

The complex structure of insect exoskeleton has inspired material scientists and engineers. Chitin is a major component of the cuticle and it is synthesized by the enzyme chitin synthase. There is a single chitin synthase gene (kkv) in Drosophila facilitating research on the function of chitin. Previous editing of kkv lead to the recovery of a viable hypomorphic allele. Experiments described in this paper argue that a reduction in chitin synthase leads to the shafts of sensory bristles becoming fragile and frequently breaking off as the animals age. This is likely due to reduced chitin levels and further suggests that chitin plays a role in resilience of insect cuticle. The different layers in cuticle are continuous across the many epidermal cells that secrete the cuticle that covers the body. Little is known about the mechanisms responsible for this continuity. Using genetic mosaics and scanning electron microscopy this paper establishes that kkv shows short range cell non-autonomy. It also provides evidence for 2 possible mechanisms. One is the intercellular transfer of Kkv protein from one cell to its neighbors and the second is the deposition of cuticular material across the boundaries of neighboring cells.

The complex structure of insect exoskeleton has inspired material scientists and engineers. Chitin is a 27 major component of the cuticle and it is synthesized by the enzyme chitin synthase. There is a single 28 chitin synthase gene (kkv) in Drosophila facilitating research on the function of chitin. Previous editing 29 of kkv lead to the recovery of a viable hypomorphic allele. Experiments described in this paper argue 30 that a reduction in chitin synthase leads to the shafts of sensory bristles becoming fragile and frequently 31 breaking off as the animals age. This is likely due to reduced chitin levels and further suggests that chitin 32 plays a role in resilience of insect cuticle. The different layers in cuticle are continuous across the many 33 epidermal cells that secrete the cuticle that covers the body. Little is known about the mechanisms 34 responsible for this continuity. Using genetic mosaics and scanning electron microscopy this paper 35 establishes that kkv shows short range cell non-autonomy. It also provides evidence for 2 possible 36 mechanisms. One is the intercellular transfer of Kkv protein from one cell to its neighbors and the 37 second is the deposition of cuticular material across the boundaries of neighboring cells. 38 39 insect cuticles. I will use the term envelope to describe the outermost layer that is lipid rich and serves 48 as a barrier to water loss and aids in hydrophobicity. I use the term epicuticle for the next layer. Both of 49 these are rather thin in the Drosophila cuticles studied in this paper. The innermost layer is the 50 procuticle, which is composed of multiple sub-layers of chitin and protein. The chitin fibrils are arranged 51 as parallel arrays and each layer is rotated with respect to its neighboring layers (Bouligand, 1972; 52 Moussian, 2013; Moussian et al., 2006) . The array rotation gives rise to the layered appearance of the 53 procuticle. The procuticle is typically the thickest layer and the arrays of chitin are thought to provide 54 strength and perhaps resilience to failure. In addition to these 3 layers various authors have proposed 55 an additional layer that is juxtaposed between the apical surface of the epidermal cells and the basal 56 region of the procuticle. This fourth layer was given the names such as the assembly zone or adhesion 57 layer (Fristrom et al., 1993; Locke, 1961 ; Moussian et al., 2005; Schmidt, 1956; Sobala and Adler, 2016) . 58

The suggested function can be gleamed from their names. My observations suggest that both of these 59 two types of layers can be found in at least some developing cuticles and that they are different from 60 one another (Sobala and Adler, 2016) . The insect cuticle is extremely variable in terms of its physical 61

properties. For example, the Young's modulus of insect cuticle varies over more than 6 orders of 62 magnitude (Vincent and Wegst, 2004) . Qualitative and quantitative differences in the molecular 63 components, their degree of cross linking and the degree of hydration are likely responsible for this 64 variation. To a first approximation, it is generally thought that the layers are synthesized in the order 65 envelope, exocuticle and procuticle but there is likely at least some "maturation" that takes place out of 66 order (Moussian et al., 2006; Sobala and Adler, 2016) . 67

Insect cuticle is a complicated material containing a large number of different components. For 68 example, more than 100 cuticle protein genes are found in the Drosophila genome (Ioannidou et al., 69 2014; Willis, 2010) and it has been suggested that this may be an underestimate (Sobala and Adler, 70 et al., 2015; Wong and Adler, 1993) . As morphogenesis continues the delayed hairs catch up to the 129 earlier forming ones. To determine if the deposition of cuticle followed a similar time course I used a 130 chitin reporter and examined wing cuticle deposition by in vivo imaging (Sobala et al., 2015) . At all times 131 examined (42, 44 and 46 hr awp (after white prepupae)) neighboring cells showed similar levels of the 132 reporter ( Fig 2C) . I carried out similar experiments where we localized Kkv::NG and similarly in all cases 133 neighboring cells always showed similar levels and localization of Kkv (Fig 2B) . We also examined 134 pigmentation in developing pupae and once again, neighboring cells showed a similar level of 135 pigmentation ( Fig S1) . Hence the mechanism used to initiate hair outgrowth appears to be different 136 from the mechanism used to initiate chitin deposition and Kkv synthesis. 137

The cuticle of a hypomorphic allele of kkv is less robust than wild type cuticle. 139

As described previously I used Crispr/Cas9 to tag the endogenous kkv gene with either Neon Green or 140 smFP (Adler, 2020) . Flies homozygous for each of the edits were viable and showed no dramatic 141 phenotype when examined in a stereomicroscope. However, when we mounted wings for examination 142 in the compound microscope we observed thin, bent and generally deformed wing hairs (trichomes) in 143 the smFP homozygotes but not the NG homozygotes (Adler, 2020) . We further established that the 144 Kkv::smFP protein accumulated to a much lower level than Kkv::NG. When we examined kkv::smFP/Df 145 flies the hair phenotype appeared slightly stronger. These two observations let me to conclude 146 kkv::smFP was a hypomorph due to the protein not accumulating efficiently (Adler, 2020) . I 147 subsequently noticed that a minority of the kkv::smFP/Df flies had deformed wings soon after or at 148 eclosion (Fig 3) . This suggested a problem in expansion of the wing or eclosion from the pupal case. We 149 found the frequency of flies showing the phenotype was significantly higher for kkv::smFP/Df , 150 kkv::smFP/kkv 1 (kkv 1 is listed as an amorphic allele on FyBase (Dos Santos et al., 2015)) and kkv::smFP 151 flies compared to Ore-R or kkv::NG flies (Fig 3) . We also noticed that some of the kkv::smFP containing 152 flies showed a slight downward curve to the wing. While the kkv hypomorphic individuals are able to 153 fly, by casual observation, they appeared less active than normal flies. 154

To test the hypothesis that the cuticle synthesized in a hypomorphic mutant was less robust to the wear 155 and tear of life we followed adult flies that were either wt or kkv hypomorphs over time to see if defects 156 arose more rapidly in the hypomorphs. We scored 3 phenotypes: life span, the loss of thoracic 157 macrocheatae or wing defects. All of the various mutants were compared to wild type Ore-R flies. To 158 reduce the length of time that these experiments took we kept the adult flies at 29.5 o C or 29.3 o C, which 159 shortens lifespan. We carried out 2 separate experiments using slightly different conditions (see 160 methods) and in both cases the Ore-R flies lived longer than the kkv edited flies (Fig S2) . The data 161 suggested that this was not simply due to decreased Kkv levels as kkv::smFP/Df flies lived slightly longer 162 than the kkv::smFP flies. The basis for the shorter life span is unclear. 163 A significant difference was seen for the loss of thoracic macrocheatae with aging that did appear to be 164 due to kkv levels as the strongest phenotypes were seen for kkv::smFP/Df and kkv::smFP/kkv 1 flies (Fig 4  165 H-K). Most of those flies that lived for 20 days or longer showed the loss of at least one of those bristles 166 and in most cases multiple bristle were lost. I also observed the loss of microcheatae (Fig 4) but did not 167 quantify this. In all, or at least most cases, the bristle shaft was lost but the socket cell remained (Fig  168   4G ). In addition to the two experiments where 10 females and 5 males were cultured together, we also 169 carried out a small scale experiment where we followed individual female flies for 20 days. This 170 experiment established that the loss was progressive. That is, we usually observed the loss of one or 171 two bristles followed by the subsequent loss of additional bristles (Fig 4DEF) . Bristle loss was less 172 frequent but still common in kkv::smFP homozygoes. In contrast, the loss of bristles associated with 173 aging was rarely seen for Ore-r (Fig 4 ABC, H-K) . 174

We also scored the flies in these experiments for loss of sections of the wing. The losses routinely began 175 at the wing margin. Surprisingly, this turned out to be substantially more common in Ore-R flies than in 176 the flies with edited kkv genes ( Fig S3) . Possible reasons for this are explored in the discussion. In an attempt to assay more directly for non-autonomy of kkv mutant clones I needed an assay that 200 allowed us to identify kkv clones and examine them at a higher resolution than is possible in the light 201 microscope. I found that I could fracture adult cuticle and by scanning electron microscopy detect a 202 layered structure. The layers were very distinctive in abdominal cuticle ( Fig S5) consistent with the 203 robust layering in abdominal procuticle seen by TEM (Fig 1) . I also attached wings to studs in a vertical 204 position, fractured them and then imaged the wings by scanning electron microscopy. In wing cuticle I 205 also detected a layered structure that presumably reflects the banding of chitin in the procuticle (Fig  206   5BC ). The layering was less distinctive (and sometimes hard to detect) than in the abdominal cuticle 207 consistent with TEM observations. We next fractured and observed wings carrying kkv loss of function 208 clones. We were able to identify mutant clones by the presence of the kkv flaccid hair phenotype ( wing cuticle thickness at wt-mutant clone boundaries ( Fig 5A) . In contrast, if there was a small degree of 211 cell non-autonomy we predicted that we would see a smooth change in cuticle thickness near the edge 212 of clones ( Fig 5A) . In all of the clones (n=9) we examined there was a smooth transition in cuticle 213 thickness ( Fig 5DE) . This transition zone appeared to be restricted to a mutant cell and its direct 214 neighbor. 215

Possible mechanism for the short-range non-cell autonomy of kkv. 217

In experiments where we imaged Kkv::NG in living pupae we noticed fluorescent puncta in the 219 extracellular space between the pupal cuticle and the epidermal cells that were in the process of 220 synthesizing the adult cuticle (Fig. 6BC ). To investigate this in more detail we obtained large Z stacks 221 that extended from the pupal cuticle to below the apical surface of the epidermal cells. Pupal cuticle in 222 50 hr pupae shows substantial autoflourescence so we examined and compared Ore-R and kkv::NG 223

pupae to determine what if any fluorescence was due to the presence of the Kkv-NG protein. The 224 autofluorescence of the thoracic pupal cuticle of Ore-R was spatially relatively even ( Fig 6D) . In contrast, 225 the fluorescence of pupal cuticle of kkv-NG flies was much more uneven with both puncta (arrow) and 226 lines (arrowhead) of bright fluorescence (Fig. 6A) . No fluorescence was observed in the region 227 between the pupal cuticle and the apical surface of the epithelial cells in Ore-R pupae (Fig. 6EF) . In 228 contrast, in this region many fluorescent puncta were observed in kkv::NG pupae (Fig. 6BC, arrows) . A 229 majority of these were located close to the pupal cuticle but some were observed throughout the 230 region. Most of the puncta located close to the pupal cuticle were stable but many of those located 231 lower were mobile (movie S1). Since the fluorescent puncta were only seen when the kkv::NG gene 232 was present we interpret the puncta as evidence of shed Kkv::NG. Since the puncta were located above 233 the impermeable adult cuticle (which is in the process of being synthesized), it seems likely that the 234 Kkv::NG was shed during or after the synthesis of the pupal cuticle and before the synthesis of the adult 235 cuticle began. Consistent with this hypothesis we observed puncta in 28hr pupae, well before the start 236 of adult cuticle deposition (Sobala and Adler, 2016) . We also observed puncta in very young pupae (20 237 hr awp) prior to the detachment of the epithelial cells from the pupal cuticle (Fig. S6B) . We observed 238 similar puncta when we examined ap>kkv::NG pupae ( Fig S6C) but not from ap>kkv-R896K::NG pupae 239 (this mutant protein does not localize correctly to the apical surface (Adler, 2020)) ( Fig S6D) , suggesting 240 that to be shed Kkv needs to be localized apically. 241

The highest concentration of puncta were over the dorsal thoracic midline (Fig S6ABC) . There were also 242 a large number of puncta over the dorsal abdomen and they were seen at a lower frequency in the 243 wing, legs and head. In the abdomen the puncta tended to align parallel to the segment boundary. 244

All of the experiments where we detected puncta in living pupae required imaging of the Kkv::NG fusion 245 protein. Experiments described earlier established that NG was a valid reporter for Kkv in bristles so it 246 seemed likely that it was also a valid reporter for Kkv in puncta (Adler, 2020) . To test this hypothesis I 247 immunostained pupal cuticle using both anti-NG and anti-Kkv-M antibodies. Among the puncta 248 detected 69.7% stained with both antibodies indicating most puncta contained both NG and Kkv (Fig S7,  249 arrows) supporting the idea that NG is an valid reporter for shed Kkv::NG. 250

The shedding of Kkv::NG could be specific for Kkv or it could reflect a process that leads to the shedding 251 of many if not all of the proteins located in the apical plasma membrane. To distinguish between these 252 two possibilities we examined live ap-Gal4/+; UAS-mCD8-GFP/+ pupae. These animals showed a large 253 number of fluorescent puncta present in the space between the pupal cuticle and the apical surface of 254 the epithelial cells ( Fig 6GHI) . As was the case for the puncta in kkv::NG pupae many of the puncta were 255 mobile. I concluded that the shedding of membrane proteins is not specific for Kkv or proteins involved 256 in cuticle deposition. 257

The pupal cuticle of Drosophila appears relatively transparent and uniform in bright field optics. 258

However, in carrying out these in vivo imaging experiments we observed that the autofluorescence of 259 the thoracic and abdominal pupal cuticles was quite distinct (Fig S8) . The autofluorescence of the 260 thoracic pupal cuticle was splotchy but without distinctive morphology. In contrast, the 261 autofluorescence of the abdominal pupal cuticle showed a pattern of bright elongated lines. Our first 262 thought when observing this was that the bright lines represented cell boundaries however attempts to 263 establish this were unsuccessful. 264 265 Flip out clones -Transfer of Kkv::NG puncta. 266

The observation that Kkv::NG was shed raised the possibility that this could be a mechanism to provide 267 for cell non-autonomy of kkv. One possible mechanism is the apical secretion of Kkv and its subsequent 268 lateral movement prior to it synthesizing chitin. An alternative possibility is the intercellular movement 269 of Kkv to neighboring cells followed by secretion or apical localization and chitin synthesis. In an 270 attempt to get evidence for either of these mechanisms we generated flip out clones comprised largely 271 of single cells that expressed Kkv::NG and then looked by in vivo confocal microscopy for lateral 272 movement of Kkv::NG beyond the clone cell. We observed 53 such clones during the deposition of the 273 procuticle and for 41 of these we observed Kkv::NG puncta beyond the lateral edge of the clone cell (Fig  274   7AB ). These puncta could be localized apically over the neighboring cells or in the neighboring cell. Another possible mechanism to explain the short distance cell non-autonomy is for a spread of apically 291 secreted chitin. With this in mind we examined thoracic epidermal cells during the synthesis of the 292 pupal cuticle by transmission electron microscopy (TEM). We could identify cell boundaries by the 293 presence of junctional complexes (Figure 8, asterisks) . We observed that the undulae found on most 294 epidermal cells during cuticle secretion were often bent over the position of the junctional complex (Fig.  295 8BEFI). We also often observed what appeared to be "trains" of secreted material between the 296 undualae and the cuticle. These "trains" were often curved and extended over the cell boundary to 297 above the neighboring cell ( Fig 8BEHFI) . These observations suggest that cuticle material secreted from 298 one cell can end up covering part of a neighbor. 299 

The importance of cuticle both as a barrier to the outside and as a support for movement makes its 336 integrity of paramount importance to the success of insects. Flight is a standard feature of insects and it 337 is extremely demanding in terms of energy and structural wear on the cuticle. I previously isolated a 338 hypomorphic allele of kkv (kkv::smFP) and here I described several additional phenotypes. One of these 339 was the loss of thoracic macrocheatae. We established that the loss of the bristle shaft increased over 340

time and it appears to be due to the shaft breaking off as the surrounding socket cell remained and 341 were made by the author in his lab and are described in (Adler, 2020) . 388

Immunostaining of fixed pupal epidermal cells during the deposition of cuticle is complicated by the 390 inability of the antibodies to penetrate cuticle after the early stages of its development. Thus, most of 391 the imaging experiments we carried out on Kkv in pupae were done by in vivo imaging of Kkv::NG. In a 392 small number of experiments we examined Kkv-NG in fixed tissue. Inthese experiments we carried out 393 anti-NG immunostaining. Such tissue was only weakly fixed and we did not use animals that were older 394 than around 48 hr after white prepupae (awp). Otherwise, immunostaining of pupal and larval tissues 395 were done as described previously (Nagaraj and Adler, 2012) . Imaging of live Kkv::NG containing pupae 396 was done on a Zeiss 780 confocal microscope in the Keck Center for Cellular Imaging. Stained dissected 397 samples were examined on the same microscope. 398

Wings were removed from two day old adult flies. In the experiments described in the paper the wings 400 contained flip out clones (AyGal4) that expressed an RNAi for kkv (Trip line -HMC.05880). The hair 401 phenotype overlapped with that seen previously in clones homozygous for kkv 1 with but with a smaller 402 fraction with the strongest phenotype (Ren et al., 2005) . The wings were attached to studs with a 403 vertical surface with conducting paint and they were then fractured with a tungsten needle. awp (after white prepupae) animals. Panels BEH are from animals 8 hr awp and panels CFI are from 483 animals 13 hr awp. Arrowheads point to undulate that extend over the junctional complex to be above 484 the neighboring cell. Arrows point to secreted material in "trains" that curve and join the forming 485 cuticle displaced laterally from the undulae it appears to derive from. This is most dramatic in the 13 hr 486 awp animals but is also clear in the 8 hr animals and there are hints in the 6 hr samples. The size 487 markers are 5 um. Panels A and D are shown at a higher magnification than the other images. Panel D 488 is shown to illustrate the distinctive dark/light/dark envelope. Not all micrographs show this as well. 489 490

The localization of chitin synthase mediates the patterned deposition of chitin in 500 developing Drosophila bristles

Ecdysone-responsive transcriptional regulation 502 determines the temporal expression of cuticular protein genes in wing discs of Bombyx mori

Flexibility and control of thorax deformation during hawkmoth flight

Twisted fibrous arrangements in biological materials and cholesteric mesophases

Elucidation of the 509 regulation of an adult cuticle gene Acp65A by the transcription factor Broad

A genome-wide transgenic RNAi library for conditional gene 513 inactivation in Drosophila

FlyBase: introduction of the Drosophila 516 melanogaster Release 6 reference genome assembly and large-scale migration of genome 517 annotations

Unexpected strength and toughness in chitosan-fibroin 519 laminates inspired by insect cuticle

The distribution of PS integrins, laminin A and F-actin 521 during key stages in Drosophila wing development

The 523 transcription factor Grainy head and the steroid hormone ecdysone cooperate during 524 differentiation of the skin of Drosophila melanogaster

Crustacean-derived biomimetic components and 526 nanostructured composites

CutProtFam-Pred: detection and classification of putative structural cuticular proteins from 529 sequence alone, based on profile hidden Markov models. Insect biochemistry and molecular 530 biology

Effects of geometry on stresses in discontinuous composite 532 materials

Drosophila DHR38 nuclear receptor is required for adult 534 cuticle integrity at eclosion

Novel transport function of adherens junction revealed by live imaging in Drosophila

Pore canals and realted structures in insect cuticle

The Drosophila planar polarity gene multiple wing hairs 541 directly regulates the actin cytoskeleton

Protein equilibration through somatic ring canals in Drosophila

Stages of cell hair construction in Drosophila

The apical plasma membrane of chitin-synthesizing epithelia

Involvement of chitin in 549 exoskeleton morphogenesis in Drosophila melanogaster

Cuticle differentiation during 551 Drosophila embryogenesis

Assembly of the Drosophila larval 553 exoskeleton requires controlled secretion and shaping of the apical plasma membrane

Dusky-like functions as a Rab11 effector for the deposition of cuticle 556 during Drosophila bristle development

Genetic control of epidermis differentiation in Drosophila. The International journal of 558 developmental biology

The Transgenic RNAi Project at Harvard Medical School: Resources 561 and Validation

THE EXCITATION AND CONTRACTION OF THE FLIGHT MUSCLES OF INSECTS

Numerical investigation of insect 565 wing fracture behaviour

Gene expression during Drosophila wing morphogenesis 567 and differentiation

Megalin-dependent yellow endocytosis restricts melanization 569 in the Drosophila cuticle

Influenza A virus uses intercellular connections 571 to spread to neighboring cells

Plasmodesmata at a glance

Observations on the subcuticular layer in the insect integument

The Gene Expression Program for the Formation of Wing Cuticle in 575 Drosophila

ChtVis-Tomato, a genetic reporter for in vivo visualization 577 of chitin deposition in Drosophila

Design and mechanical properties of insect cuticle. Arthropod 579 structure & development

In vivo time-resolved microtomography reveals the mechanics of 582 the blowfly flight motor

Ecdysone directly and indirectly 584 regulates a cuticle protein gene, BMWCP10, in the wing disc of Bombyx mori. Insect 585 biochemistry and molecular biology

Structural cuticular proteins from arthropods: annotation, nomenclature, and 587 sequence characteristics in the genomics era

Tissue polarity genes of Drosophila regulate the subcellular location 590 for prehair initiation in pupal wing cells

Adherens 592 junction-associated pores mediate the intercellular transport of endosomes and cytoplasmic 593 proteins