(C) PLOS One
This story was originally published by PLOS One and is unaltered.
 . . . . . . . . . .



Organ transformation by environmental disruption of protein integrity and epigenetic memory in Drosophila [1]

['Orli Snir', 'Department Of Biomolecular Sciences', 'Weizmann Institute Of Science', 'Rehovot', 'Michael Elgart', 'Yulia Gnainsky', 'Moshe Goldsmith', 'Filippo Ciabrelli', 'Institute Of Human Genetics', 'Cnrs']

Date: 2024-06 

Despite significant progress in understanding epigenetic reprogramming of cells, the mechanistic basis of “organ reprogramming” by (epi-)gene–environment interactions remained largely obscure. Here, we use the ether-induced haltere-to-wing transformations in Drosophila as a model for epigenetic “reprogramming” at the whole organism level. Our findings support a mechanistic chain of events explaining why and how brief embryonic exposure to ether leads to haltere-to-wing transformations manifested at the larval stage and on. We show that ether interferes with protein integrity in the egg, leading to altered deployment of Hsp90 and widespread repression of Trithorax-mediated establishment of active H3K4me3 chromatin marks throughout the genome. Despite this global reduction, Ubx targets and wing development genes preferentially retain higher levels of H3K4me3 that predisposes these genes for later up-regulation in the larval haltere disc, hence the wing-like outcome. Consistent with compromised protein integrity during the exposure, the penetrance of bithorax transformations increases by genetic or chemical reduction of Hsp90 function. Moreover, joint reduction in Hsp90 and trx gene dosage can cause bithorax transformations even without exposure to ether, supporting an underlying epistasis between Hsp90 and trx loss-of-functions. These findings implicate environmental disruption of protein integrity at the onset of histone methylation with altered epigenetic regulation of developmental patterning genes. The emerging picture provides a unique example wherein the alleviation of the Hsp90 “capacitor function” by the environment drives a morphogenetic shift towards an ancestral-like body plan. The morphogenetic impact of chaperone response during a major setup of epigenetic patterns may be a general scheme for organ reprogramming by environmental cues.

By analyzing stage-dependent effects of exposure to ether, we provide evidence for a mechanistic chain of events connecting brief embryonic exposure to ether with organ transformation manifested at the larval stage and on. We show that ether disrupts the eggshell and interferes with native protein folding in the embryo. The induction of proteotoxic stress alters the profile of deployments of Hsp90 toward its clients, including Trx. This, in turn, leads to a decrease in Trx function and repression of H3K4 tri-methylation at a critical stage of development. We further found that the repression of active chromatin marks (H3K4 tri-methylation) is less pronounced in actively transcribed genes, including Ubx targets and wing development genes. The differentially higher postexposure levels of active chromatin marks in wing genes predisposes these genes for later up-regulation in the larval haltere disc. The joint contribution of interaction between proteotoxic stress and epigenetic patterning to bithorax induction was supported by impacts of reduced function of Hsp90. We found that the induction of bithorax phenocopies is enhanced by chemical or genetic reduction of Hsp90 and that joint reduction of Hsp90 and trx gene dosage causes spontaneous bithorax transformations, implicating the interaction between Hsp90 and Trx loss-of-functions in causing bithorax transformations. This is consistent with a reported dependence of Trx on Hsp90 function [ 33 ]. Altogether, these findings link the proteotoxic stress of ether to excess demand for Hsp90 function that interferes with initial epigenetic patterning and predisposes wing genes for later up-regulation in the haltere disc. Notably, this chain of events can also account for the reported induction of bithorax phenocopies in response to brief exposure to heat at this critical stage in development [ 32 , 34 ].

Combining genetic (and/or epigenetic) assimilation of environmental induction with resemblance to morphogenetic phenotypes of mutations in patterning genes gave rise to a nontraditional view of the environment as a driver of rapid morphological diversification (in addition to its contribution to selection) [ 3 , 23 – 30 ]. The potentially profound impact on evolution, however, often comes at the expense of generating deformed individuals (“hopeful monsters”). Suppressing these harmful changes and promoting reversal toward normal development requires detailed understanding of the teratologic process [ 3 , 31 ]. While effector genes have been identified in various cases, mechanistic understanding of the chain of events connecting the respective environmental disruptor with stage-specific changes that lead to disfigurement is generally lacking. Here, we use the classic example of bithorax induction as a model for investigating how brief exposure to ether during early embryogenesis modifies the adult body plan. The resemblance to phenotypes of Ubx and trx mutations [ 16 , 18 ] suggests that environmental induction of bithorax phenocopies may be mediated by reduced function of Ubx and/or trx. However, it is not clear how the function of these (and/or other) genes mediates the stage-dependent response to ether, why this induction requires exposure in a specific time window, and why a similar transformation is observed in response to heat exposure at the same time window [ 32 ].

In a landmark experiment, Waddington showed that environmental induction of gross morphological changes can be rapidly enhanced and “stabilized” (genetically assimilated) by successive exposures and selections of phenotypic individuals over a few generations [ 10 ]. Stabilization of altered phenotypes was indicated by spontaneous transformations (generation of transformed individuals without exposure to the environmental inducer). The validity and generality of this genetic assimilation were supported by: (i) identified accumulation of alleles contributing to stronger induction and generation of spontaneous transformations; (ii) demonstration of similar effects in other scenarios of environmental-induced transformations; and (iii) evidence for potential stabilization by epigenetic means [ 19 – 22 ].

Determination of cell and tissue identities in flies is established during embryonic development and maintained by epigenetic means, particularly by the Polycomb and Trithorax systems [ 1 , 2 ]. Early embryonic exposure to biotic and abiotic environmental stimuli (e.g., chemicals, heat, unusual diets, crowding, predation, electromagnetic fields, and many more) can interfere with tissue identities and induce homeotic transformations in a wide range of species [ 3 – 7 ]. Hallmark examples include environmental induction of haltere-to-wing phenocopies [ 8 – 10 ], antenna-to-leg, leg-to-wing, and eye-to-wing transformations by disruptions of primordial organs and boundaries [ 11 , 12 ], as well as limb malformations and/or organ failures by embryonic exposure to various substances (e.g., thalidomide) at a sensitive time window of development in human and other mammals [ 13 – 15 ].

(A) Transcriptional levels in the haltere (third instar larvae) for genes with the highest (top 10%, purple) and lowest (bottom 10%, blue) levels of H3K4me3 shortly after exposure to ether (4.5 h AED). Shown for yw (with and without dechorionation), trx 1-/+ , and trx 1-/- . Mean FPKM ± SE, n = 3. **p < 0.01, ***p < 0.001, two-way ANOVA followed by Tukey HSD test. (B) Cumulative distribution function (CDF) of embryonic H3K4me3 levels shown for up-regulated genes (purple), up-regulated Ubx targets (blue), up-regulated wing development genes (red), and all genes (gray). (C) Same as (B) for joint targets of Trx and Ubx (blue), targets of Trx that are not shared with Ubx (black), and all genes with detectable methylation (gray). Inset: Same for the indicated subsets of targets that are up- and down-regulated in the haltere disc (left and right, respectively). The data underlying this figure can be found in S4 Data .

While up-regulation of joint targets of Ubx and Trx in the haltere disc may account for the ectopic wing phenotypes, it is not clear how this up-regulation is promoted by brief exposure to ether during early embryogenesis. To investigate if the induction of wing fates could be specified by the impact of ether on H3K4 trimethylation in the early embryo, we examined correlations between H3K4me3 levels shortly after exposure and mRNA levels in the haltere discs of third instar larvae. We found that genes with high and low H3K4me3 levels in the embryo (top and bottom 10%) are expressed, respectively, at high and low levels in the haltere ( Fig 5A ). Reciprocal analysis of genes that are up-regulated in the haltere showed that the embryonic H3K4me3 levels of these genes are higher compared to bulk (Figs 5B and S8A and S5 Table ). The preferentially higher levels of H3K4me3 were even more pronounced in the subset of up-regulated wing genes versus the entire set of up-regulated genes ( Fig 5B ). The linkage between early H3K4me3 and later expression of wing genes in the haltere was further supported by differential methylation of distinct subsets of Trx targets ( Fig 5C ). In particular, the H3K4me3 of joint targets of Trx and Ubx was significantly shifted towards higher levels ( Fig 5C , left inset; p < 1E-5), while H3K4me3 of Trx targets that are not shared with Ubx was shifted toward lower levels ( Fig 5C , right inset; p < 1E-5). These differences in H3K4me3 were consistent with matching (positive and negative) transcriptional fold changes in the haltere disc (insets to Figs 5C , S8B , and S8C ). Taken together, these findings implicate wing genes-related differences in H3K4me3 shortly after exposure with later-stage induction of haltere-to-wing transformations.

(A) Functional enrichments of GO terms in genes that were significantly up-regulated by ether in haltere discs of third instar larvae from trx 1-/+ and trx 1-/- stocks, and a wild-type (yw) stock, with and without egg dechorionation prior to exposure (“Decho”). Based on David online tool. (B) Ether-induced mRNA fold-change of wingless in the haltere disc. Mean fold-change (ether vs. control) for the cases in (A). (C) Median levels of mRNA ± SE for the indicated subsets of genes. n = 3. Ether effect: p < 0.001, Genotype effect: p < 0.001, Ether-Genotype interaction: p < 0.001, two-way ANOVA (full set of ANOVA p-values provided in S3 Table ). (D) Enrichment of Trx and Ubx targets (inset) within up- and down-regulated genes in haltere disc (fold-change >1.5 and p < 0.05), Fisher exact test. (E, F) Same as (C, D) for wing development genes and their intersections with Ubx targets or Trx targets that are not shared with Ubx. Ether effect: p < 0.001, Genotype effect: p < 0.001, Ether-Genotype interaction: p < 0.001, two-way ANOVA (full set of ANOVA p-values provided in S4 Table ). mRNA analyses are based on 3 biological replicates for each genotype. The data underlying this figure can be found in S3 Data .

The higher penetrance of bithorax phenocopies in trx mutants versus wild type ( Fig 2A ) suggested a causal link between the repressive effect of ether on the (trx-mediated) H3K4me3 and the subsequent induction of haltere-to-wing transformation. To investigate this possibility, we analyzed mRNA changes in the haltere discs of third instar larvae from ether-exposed versus non-exposed wild-type and trx mutant stocks ( S6A Fig and S2 Spreadsheet). RNA-seq analysis of dissected discs revealed dosage-dependent induction of wing genes ( Fig 4A ), including wingless (wg), the master regulator of wing development ( Fig 4B ). Ubx targets and joint targets of Ubx and Trx were also up-regulated, whereas targets of Trx alone were down-regulated ( Fig 4C and 4D ). The relevance of these changes to the bithorax induction was further highlighted by preferential up-regulation of wing genes that are jointly targeted by Ubx, but not wing genes that are targeted by Trx alone ( Fig 4E and 4F ). Since Ubx is itself a target of Trx and a transcriptional repressor of wing genes in the haltere [ 44 , 45 ], the up-regulation of joint targets of Ubx and Trx is consistent with the induction of wing phenotypes. Notably, despite the significant changes in the levels of their targets, the mRNA levels of trx and Ubx, as well as the protein levels of Ubx, were not affected by exposure to ether ( S6B , S6C , and S7 Figs and S2 Spreadsheet ).

(A) Percentile-normalized numbers of H3K4me3 reads per 100 bp in ether-exposed (“Ether”) vs. non-exposed yw embryos (“Ctrl”). Red and purple dots correspond, respectively, to changes above and below 2 standard deviations from the mean. Inset: no. of regions with methylation levels above and below 2 standard deviations from the mean. (B) Same as (A) for normalized numbers of H3K27me3 reads per 1,000 bp. (C) Gene-specific H3K4me3/H3K27me3 ratios in ether-exposed vs. non-exposed embryos. *** p < 1E-19, Wilcoxon signed rank test. (D) mRNA fold-change (Ether/Ctrl) vs. H3K4me3 fold-change (Ether/Ctrl). Inset: numbers of differentially expressed genes (fold-change >1.5 and p < 0.05), based on 3 biological replicates. (E) Enrichments of GO terms in subsets of genes that are up- (brown) and down-regulated (blue) 1 h after exposure vs. no exposure. Based on the “DAVID” online tool. (F) Probability density functions (PDF) of H3K4me3 levels, demonstrating a shift in the distribution of “wing development” genes (red) toward higher levels compared to bulk (gray) in both ether-exposed and control embryos (solid and dotted lines, respectively). *** p < 1E-6. (G) Left: Enrichments of GO terms in genomic loci exhibiting the highest retention of H3K4me3 (top 10%) following exposure to ether. Based on the “DAVID” online tool. Gene-specific retention is defined by the ratio between H3K4me3 levels in ether-exposed and non-exposed embryos. Right: STRING Network analysis of genes with the highest retention of H3K4me3. Significance levels are indicated in S2 Table . (H) Cumulative distribution function (CDF) of mRNA levels, shown for all genes with a detectable level of H3K4me3 (gray) as well as for genes with 10% highest and lowest retention of H3K4me3 (blue and purple, respectively). *** p < 1E-6 corresponds to the difference between the bulk distribution and the distributions for genes with either lowest or highest retention. ChIP and mRNA analyses are based on 2 and 3 biological replicates, respectively.

To seek additional evidence for joint involvement of Hsp90 and trx loss-of-function in the induction of bithorax phenocopies, we tested if the exposure to ether compromises the activity of the Trx protein. Since Trx is responsible for the tri-methylation of histone H3K4 (H3K4me3) in the early embryo [ 47 – 49 ], we examined changes in histone methylation and gene expression shortly after exposure ( Fig 1A ). Active and repressive chromatin marks were analyzed by ChIP-seq using antibodies specific for H3K4me3 and tri-methylation of histone H3K27 (H3K27me3), respectively [ 50 ] ( Fig 3A and 3B ). Comparison between methylation states in different samples was preceded by global percentile normalization of read counts applied to 100 bp genomic segments of each sample [ 51 ]. Differences between normalized counts of exposed (“Ether”) versus non-exposed embryos (“Ctrl”) revealed a genome-wide (albeit not uniform) decrease of H3K4me3 levels, without a noticeable change in H3K27me3 ( Fig 3A versus Fig 3B ; S3 and S4A Figs). Predominant suppression of H3K4me3 was also reflected by site-specific ratio of H3K4me3 to H3K27me3, as determined by integrating normalized counts over each gene region and dividing the region-specific H3K4me3 level by the respective level of H3K27me3. The H3K4me3/H3K27me3 ratio in ether-exposed versus control embryos was lower in 88% of the genes and higher in only 5% of the genes ( Fig 3C ). Unlike the extensive suppression of H3K4me3, the immediate transcriptional response was very mild (Figs 3D and S4B–S4E and S1 Spreadsheet ). Moreover, the up-regulated genes were not enriched with wing genes ( Fig 3E ), suggesting that the induced predisposition towards wing might be primarily specified by site-specific changes in histone methylation. This was supported by significantly higher abundance of H3K4me3 at loci of wing genes (Figs 3F , S4F , and S4G ), as well as by enrichment of wing genes in genomic regions exhibiting the highest retention of H3K4me3 following exposure to ether ( Fig 3G and S2 Table ). High retention of H3K4me3 was significantly correlated with higher levels of mRNA (Figs 3H and S5A–S5D ), suggesting that active expression of wing genes during the exposure contributes to the higher levels of H3K4me3 in wing genes versus bulk. Given the reported contribution of H3K4me3 to long-term stability of active state of transcription [ 52 , 53 ], higher levels of H3K4me3 in wing genes at the end of exposure could predispose these genes for differentially higher expression at later stages.

(A) Penetrance of bithorax phenocopies in ether-exposed and non-exposed in trx 1-/+ , trx 1-/- and a wild-type (yw) stock, with and without egg dechorionation prior to exposure (“Decho”). Two-way ANOVA followed by Tukey HSD test. Groups denoted by different letters (a, b, c, d) are significantly different from each other (p-values provided in S1 Table ). Based on 3 biological replicates pooled. Inset: Exemplary image of ether-induced phenocopies in the trx 1 stock. (B) Penetrance of bithorax phenocopies in ether-exposed and non-exposed Hsp83 e6A+/- and wild-type (yw) stocks. ***p = 8.4e-11, Student’s t test. (C) Same as (B) on yw background, with and without treatment of dechorionated eggs with the Hsp90 inhibitor, Geldanamycin (yw (Decho) + GdA), prior to ether exposure, based on 3 biological replicates pooled. ***p = 3.7e-07, Student’s t test. (D) Percentages of Hsp83 e6A-/+ /trx 1-/+ , trx 1-/- , and Hsp83 e6A+/- flies exhibiting spontaneous bithorax phenocopies. Based on 4 biological replicates with approximately 100 flies each. ***p < 0.001. The data underlying this figure can be found in S2 Data .

To investigate the involvements of Trx and Hsp90 in the induction of bithorax phenocopies, we analyzed impacts of the ether on trx and Hsp90 mutant stocks versus wild type. The penetrance of bithorax phenocopies in ether-exposed, temperature-sensitive trx 1-/+ and trx 1-/- stocks increased with the mutation dosage. Notably, over 20% of trx 1-/- individuals exhibited spontaneous transformation without exposure to ether ( Fig 2A ). A significant increase in penetrance was also noted in Hsp90 heterozygotes (Hsp83 e6A -/+ ) versus wild-type control ( Fig 2B ). Similar aggravation was observed in ether-exposed wild-type embryos that were pretreated immediately after egg deposition with the Hsp90 inhibitor, Geldanamycin (GdA) ( Fig 2C ). Analysis of Hsp83 e6A-/+ /trx 1-/+ double heterozygotes showed that a combined reduction of trx and Hsp90 gene dosage can cause spontaneous transformations even in non-exposed embryos ( Fig 2D ). This was in notable contrast to lack of spontaneous transformations in trx and Hsp90 single mutants. Taken together with the dependence of trx on Hsp90 function [ 33 ], these findings suggest that loss-of-function of Hsp90 aggravates the bithorax induction by contributing to the reduction of trx function.

(A) Flowchart of experimental procedures and measurements. (B) Representative image of a severe transformation in an adult fly (yw line) that was exposed to ether during early embryogenesis. (C) Percentage of individuals exhibiting bithorax phenocopies (including abnormal pupae that failed to eclose). p < 0.001, Mann–Whitney. (D) Numbers of pupae formed with and without exposure to ether. p > 0.05, Student’s t test. (E) Effect of ether vapor on egg perimeter and transparency. Representative images (10×) of untreated eggs (top) and eggs that were exposed to ether vapor for 2 h (bottom). (F) Representative images of Acridine Orange-stained yw embryos for the following cases: no treatment (up left), 5-min immersion in Citrasolv solution [ 46 ] (up right), 5 min immersion in ether liquid (bottom left), and 1.5-h exposure to ether vapor (bottom right). (G) Circular dichroism (Far-UV) spectra of proteins extracted from yw embryos that were exposed or not exposed to ether vapor (solid and dotted lines, respectively). Displayed spectra (in [millidegrees*milliliter/milligram]) represent the average of 3 independent measurements. p < 0.05, Student’s t test. (H) Fluorescence intensity of lysates of embryos (yw line) stained with 8-anilino-1-naphthalene sulfonate (ANS), a fluorescent probe for protein conformational changes. Shown for untreated embryos (Ctrl), ether exposed embryos (Ether), and embryos that were denatured at 80°C. Mean intensity ± SE, based on 3 biological replicates. **p < 0.01, ***p < 0.001, one-way ANOVA following Dunnett’s test. (I) Representative images (left) and average fluorescence intensity in His2Av-mRFP1-tagged embryos (right), with and without exposure to ether vapor. Mean intensity per embryo ± SE, n = 32 (Ctrl), n = 29 (Ether). ***p < 1E- 4, Mann–Whitney test. (J) Trx protein levels following exposure vs. no exposure control, as determined by mass spectrometry analysis. Mean levels (relative to the average in untreated control) ± SE, based on 3 replicates of ether-exposed and control embryos. p > 0.05, Mann–Whitney. (K) Same as (J) for the Drosophila Hsp90 (Hsp83) protein level. Mean ± SE, n = 3. p > 0.05, Student’s t test. (L) Western blot analysis of Ubx protein levels in exposed embryos and non-exposed control. Mean ± SE, n = 3. p > 0.05, Student’s t test. (M) Median fold-change (ether/ctrl) of protein levels in the following subsets of Hsp90 clients (left to right): All clients (n = 280), Hsp90 w/o Trx or Ubx targets (n = 197), Hsp90 and Trx but not Ubx targets (n = 52), Hsp90 and Trx targets (n = 69), Hsp90 and Ubx targets (n = 31), Hsp90 and Trx and Ubx targets (n = 16). Based on mass spectrometry analysis with 3 replicates of ether vs. control. Subset-specific median fold-change (±SE) averaged over the proteins in the subset. Note the preferential increase of Hsp90 clients that are targeted jointly by Ubx and Trx. a, b, c, d: p-values based on Student’s t test. (N) GO enrichment analysis of proteins that increased (red) and decreased by ether, as determined by proteomics profiles of exposed and non-exposed embryos. Based on 3 biological replicates. ** p < 0.01, *** p < 0.001. The data underlying this figure can be found in S1 Data .

The proteotoxic stress caused by exposure to ether is expected to increase the workload on protein chaperones such as Hsp90, potentially altering their target deployment [ 42 , 43 ]. Moreover, mutations in Trx (a reported client of Hsp90 [ 33 ]) generate bithorax transformations [ 16 ], suggesting that the ether-induced phenocopies are mediated by impact of the proteotoxic stress on Hsp90 and Trx. To investigate the potential involvement of these genes in the response to ether, we analyzed the effect of ether on the protein levels of Hsp90 and its targets, including Trx. Proteomics analysis of embryos shortly after exposure revealed reduced levels of Trx ( Fig 1J ) without a change in the protein levels of Hsp90 (Figs 1K and S1 ). Western blot analysis of Ubx (a downstream target of Trx that specifies haltere fate by repressing the transcription of wing-related genes [ 18 , 44 , 45 ]) showed that Ubx levels are also unaffected by ether (Figs 1L and S2 ). In contrast to the lack of change in the levels of Hsp90 and Ubx, the protein levels of Hsp90 clients that are joint targets of Ubx and Trx were significantly elevated compared to other subsets of Hsp90 clients ( Fig 1M ). The response to ether was therefore accompanied by a differential increase of Hsp90 clients that are likely to favor bithorax transformation. Gene ontology analysis of ether-induced changes in the proteome profile further showed that elevated proteins are enriched with chaperone genes and the decreased proteins are enriched with proteasome genes ( Fig 1N ).

To investigate how embryonic exposure leads to homeotic transformations that are manifested at a later stage, we established effective conditions for bithorax induction by 30-min exposure of early-stage embryos to ether vapor ( Fig 1A ). To quantify the reaction, we scored the number and fraction of adult flies (and failed-to-eclose pupae) exhibiting bithorax phenocopies, defined as any morphological abnormality in the third thoracic segment that increases resemblance to the second thoracic segment [ 10 , 35 – 37 ] ( Fig 1B ). Bithorax phenocopies were induced in approximately 9% of the individuals without significant reduction in the number of pupae ( Fig 1C and 1D ). We then sought to investigate how exposure to ether causes these transformations. Since ether is an organic solvent, we suspected that its vapor may have a denaturing impact which could disrupt protein function and induce proteotoxic stress. To investigate this possibility, we first analyzed the effect of ether on the integrity of the eggshell. Inspection of embryos that were exposed for 2 h to the vapor revealed a gross increase in egg clarity ( Fig 1E ), suggesting elevated egg permeability and lipid precipitation. This was confirmed by increased penetration of the nucleic-acid stain, Acridine Orange [ 38 ], into eggs that were exposed to ether vapor ( Fig 1F ). To determine if the exposure to ether vapor affects the integrity of proteins in the embryo, we analyzed lysates of embryos by (Far-UV) circular dichroism (CD) [ 39 , 40 ]. Comparison of extracts from (in vivo) exposed versus non-exposed embryos revealed a significant decrease in the degree of light polarization produced by proteins in the sample ( Fig 1G ), suggesting a widespread impact on protein secondary structures, specifically α-helices. Staining with the protein conformation-sensitive probe, 8-anilino-1 naphthalene sulfonate (ANS) [ 41 ], revealed concordant differences between extracts from exposed versus non-exposed embryos ( Fig 1H ). Notably, the direction of change in response to ether was the same as in heat denaturation. Independent analysis of the effect of ether on fluorescence intensity in live, Histone-RFP tagged (His2Av-mRFP1) embryos revealed a significant reduction of RFP intensity in ether-exposed versus control embryos ( Fig 1I ). This reduction occurred without a change in His2Av mRNA (S1 Spreadsheet), indicating posttranscriptional impact of ether on the His2Av-RFP protein. These findings show that the exposure to ether vapor compromises the integrity of the eggshell and causes disruptions in protein structure and function.

Discussion

Fossil record evidence suggests that the common ancestor of all winged insects had 2 pairs of large membranous flight wings, located on the second and third thoracic segments [54]. In flies and various orders of insects, the hindwings evolved into organs with altered functions, such as the gyroscopic haltere in Drosophila. These alterations appear to have emerged as co-options of a wing program, as evidenced by reversals to a double pair of wings on the background of mutations in genes of the bithorax gene complex, such as Ubx and trx [16,17,19,55]. Trx is the methylase responsible for deposition of the histone H3K4me3 mark that contributes to maintenance of an active state of expression of many genes, including the homeotic genes that specify segment identities in the Drosophila embryo (e.g., the Antennapedia complex and Bithorax complex genes) [47–49]. Ubx, in turn, is a Trx target and a master regulator of haltere development that specifies haltere fates in Drosophila by repressing the transcription of multiple wing genes in the third thoracic segment. Loss-of-function mutations in these genes are therefore consistent with spontaneous haltere-to-wing transformations. In a seminal work on gene–environment interactions, Waddington demonstrated that haltere-to-wing transformations can also be induced by embryonic exposure of wild-type embryos to ether vapor [9,10]. Similar induction was demonstrated by exposure to heat at the same sensitivity time window [32,34,56], but the mechanistic basis of induction remained unknown for over 60 years. Our results show that brief exposure to ether at the time of cellularization (Fig 6) alleviates the suppression of Ubx-dependent wing regulators in the larval haltere disc. This “reprogramming” of the larval disc is mirrored in ether-exposed embryos by higher levels of H3K4 trimethylation of wing genes versus bulk. This is, in turn, expected to assist in preferential maintenance of their active state of transcription over time [52], which is consistent with the haltere-to-wing predisposition. While investigating how early exposure to ether creates this predisposition, we discovered that ether vapor dissolves the eggshell and compromises protein integrity in the embryo. This increases the demand for Hsp90, a central chaperone that assists the folding of a wide range of proteins (“Hsp90 clients”) during normal development, especially under stress. The chaperone activity of Hsp90 is required to support diverse functions, including transcriptional regulation, chromatin remodeling, and phenotypic buffering of genetic and epigenetic variations [21,22,57–65]. Since Hsp90 has also been implicated in the function of Drosophila Trx, as well as in the methyltransferase function of SMYD3 in mammals [33,66], we suspected that the proteotoxic stress in ether-exposed embryos compromises Trx function by altering the deployment of Hsp90. This was supported by several lines of evidence, including enhanced induction of bithorax phenocopies on the background of genetic or chemical reduction in Hsp90 and the epistasis between reduced functions of Hsp90 and trx. These findings implicate the environmental impact on chaperone function with an altered pattern of epigenetic memory that predisposes the haltere segment for a reversal toward wing.

By integrating our findings with evidence from previous studies, we propose a model that accounts for bithorax-like transformations in response to early embryonic exposure to both ether and heat (Fig 6). Exposure at around the time of cellularization creates proteomic stress and subsequent redeployment of the Drosophila Hsp90 (Hsp83) towards misfolded proteins. The reduced availability of Hsp90 for Trx at a stage in which Trx function is particularly required (Fig 6A and 6B [67,68]) interferes with the setup of H3K4me3. Genes with relatively high levels of H3K4me3 following the exposure (e.g., wing development genes that are targeted by Ubx) are expected to be predisposed for sustained expression at a later stage of development. This predisposition is most clearly manifested by alleviated suppression of Ubx targets in the third thoracic segment (whose identity is specified by Ubx), which is in turn, consistent with induction of wing phenotypes in the haltere.

In addition to providing a plausible explanation for the induction of bithorax phenocopies by ether (or heat), the chain of events identified in this work can also account for the lethality of exposure at an earlier stage as well as the lack of responsiveness to exposure a few hours later (Fig 6B). Previous work showed that the induction of bithorax phenocopies by either ether or heat is largely confined to a time window that starts shortly before cellularization and ends at around the stage of partial invagination of the anterior and posterior midgut. Exposure during syncytial stages (<2 h AED) leads to complete embryonic lethality, while exposure after furrow formation (>4 h AED) is no longer capable of inducing a phenocopy. In between, the survival increases as a function of the onset of exposure, while the penetrance increases to a peak at around the end of cellularization and gradually decreases at later onsets of exposure (Fig 6B). This phenomenology is fully consistent with the hypothesized involvement of Hsp90 and Trx functions; early embryonic stages are characterized by rapid divisions of nuclei within a large cytoplasmic compartment. This cytoplasm is initially loaded with very high levels of maternal transcripts of Hsp90 (modENCODE data [69]), which contribute to protein folding and functional integrity in this large compartment. The activity of Hsp90 at this stage may be particularly critical for maintaining the cytoplasmic protein gradients that specify the anterior-posterior and lateral-ventral axes [70]. Since the interruption of these gradients is lethal [71], a sufficient disruption of protein integrity can account for the lethality of exposure at that stage (Fig 6B). This was indeed supported by the 2 categories of defects that have been observed in the case of early exposure [35], namely: (i) failure to form a blastoderm, resulting in an undifferentiated-like mass with no recognizable structures; and (ii) emergence of anterior, posterior, or segmentation defects that are eventually followed by failure to hatch. As expected, these abnormalities are also more pronounced in embryos that were exposed at progressively earlier stages [35].

Altogether, these findings portray a causal chain of events, connecting environmental disruption of protein integrity at the onset of histone methylations with modified epigenetic patterning that supports a morphogenetic shift towards an ancestral-like body plan. The increased homeotic sensitivity to the reduction of Hsp90 and Trx functions is likely to be involved in other contexts of induced homeotic transformations [16,21] and can be utilized for fate manipulation and/or regeneration at the levels of tissues and organs.

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3002629

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/