1. Epigenetically Heritable Alteration of Fly Development in Response to Toxic Challenge 
Shay Stern, Yael Fridmann-Sirkis, Erez Braun, Yoav Soen 
Cell Reports 
Volume 1, Issue 5, Pages (May 2012)
DOI: /j.celrep
Copyright © 2012 The Authors Terms and Conditions

2. Cell Reports 2012 1, 528-542DOI: (10.1016/j.celrep.2012.03.012)
Copyright © 2012 The Authors Terms and Conditions

3. Figure 1 Presentation of a Toxic Challenge to Developing Fly Larvae
(A) Left view shows a scheme for creating artificial distribution of toxic stress across the larvae. Right view illustrates potential scenarios that include extension of the endogenous domain of activation of the developmental promoter (top), and no change (bottom).
(B) neoGFP is expressed using the GAL4-UAS system.
(C) Different distributions of toxic stress are generated by applying G418 to a library of fly lines, each expressing the resistance gene (neoGFP) under the regulation of a different promoter.
(D) A scheme of the basic experiment.
Cell Reports 2012 1, DOI: ( /j.celrep )
Copyright © 2012 The Authors Terms and Conditions

4. Figure 2 Challenge-Induced Modifications in Development
(A) Promoter-specific survival of fly lines. Survival ratio represents the number of adults developed with versus without G418. Data are represented as mean survival ratio ±SE measured in Nv vials pooled from Ne replicated experiments. Nv and Ne for each line are as follows: hairy (Nv = 33, Ne = 9); drm (24, 7); Hsp70 (9, 4); wg (9, 5); ptc (4, 2); elav (5, 3); dpp (7, 3); ftz (7, 3); WT (yw, 15, 4). Inset shows larvae carrying only the GAL4 driver (pool of six GAL4 lines, Nv = 22, Ne = 7), only the UAS-neoGFP transgene (Nv = 15, Ne = 4), and both (case of Act5C::neoGFP; Nv = 11, Ne = 4).
(B) Representative delay of 2–3 days in the development of challenged compared to unchallenged hairy::neoGFP larvae. Data are represented as mean fraction of pupae ±SE in six vials. Inset shows quantification of the mean delay ±SE in six replicated experiments.
(C) Representative images of induced expression of neoGFP (green) and the endogenous hairy protein (red) in the proventriculus of challenged third-instar hairy::neoGFP larva.
(D) Quantification of neoGFP expression (mean GFP intensity ±SE) in the proventriculi of challenged (n = 23) versus unchallenged (n = 31) hairy::neoGFP larvae pooled from six replicated experiments.
(E) Representative images (left) and quantification of neoGFP expression (right) in the midgut of challenged (n = 15) versus unchallenged (n = 23) drm::neoGFP larvae pooled from three replicated experiments. Inset shows quantification of developmental delay of challenged versus unchallenged larvae in seven replicated experiments.
(F) Same as (E) for challenged (n = 13) versus unchallenged (n = 18) hsp70::neoGFP larvae. Data are represented as mean area of expression ±SE. Inset shows quantification of developmental delay in five replicated experiments.
(G) Example of a dwarf drm::neoGFP adult fly that was exposed during development to 400 μg/ml of G418 (left) compared to a nonexposed fly (right).
(H) Histogram of adult lengths for challenged versus unchallenged drm::neoGFP flies.
(I) Representative image of challenged Hsp70::neoGFP fly with one abnormal wing and a corresponding induction of neoGFP expression at the base of the wing (inset).
(J) Incidence of wing abnormalities in challenged (200–400 μg/ml of G418) and unchallenged Hsp70::neoGFP adult flies pooled from seven experiments.
∗p < 0.05, ∗∗p < 0.001, ∗∗∗p < 10−5 (Student's t test).
See Figures S1 and S2 as well.
Cell Reports 2012 1, DOI: ( /j.celrep )
Copyright © 2012 The Authors Terms and Conditions

5. Figure 3
Modifications in the Regulation of hairy Are Mediated by Suppression of PcG Genes
(A) mRNA levels (mean ± SE) of PcG genes in the proventriculus of challenged versus unchallenged hairy::neoGFP third-instar larvae. mRNA was measured by Affymetrix microarrays (left, n = 2) and by qPCR (right, n = 3).
(B) Representative patterns of neoGFP expression in the proventriculus of challenged (top) and unchallenged (bottom) hairy::neoGFP larvae heterozygous for a mutation in Pc (Pc1, Pc3) or, alternatively, carrying a UAS-RNAi against E(z) (E(z)i/+).
(C) Statistics of the patterns of neoGFP expression (mean expression per proventriculus ±SE) based on the following numbers of proventriculi (n) pooled from at least three replicated experiments in each case (Unchallenged: +/+ n = 32, Pc1/+ n = 32, Pc3/+ n = 15, E(z)i/+ n = 10; Challenged: +/+ n = 23, Pc1/+ n = 17, Pc3/+ n = 25, E(z)i/+ n = 19).
(D) Left panel is normalized histograms of neoGFP expression in proventriculi of unchallenged and challenged nonmutant hairy::neoGFP larvae (blue and red, respectively), Pc1 larvae (black), and larvae carrying E(z) RNAi (gray). Right panel is the same as the left panel with challenged Pc1 (black), E(z) RNAi larvae (gray).
(E) Expression of neoGFP in proventriculi of unchallenged hairy::neoGFP larvae exposed to heat shock with or without hs-Pc (hs-Pc, n = 27; +/+, n = 10), on the background of basket mutation without heat shock (bsk1, n = 17), and in unchallenged nonmutant larvae (n = 23). Data are represented as mean expression ±SE based on the number of proventriculi (n) pooled from at least three replicated experiments in each case.
(F) Survival ratios for the cases in (E) under exposure to 400 μg/ml of G418 (hs-Pc, n = 12 vials; +/+ with heat shock, n = 10; bsk1/+, n = 9; +/+ without heat shock, n = 33). Data are represented as mean survival ratio ±SE in fly vials pooled from four to nine replicated experiments.
∗p < 0.05, ∗∗p < 10−3, ∗∗∗p < 10−5 (Student's t test). NS, not significant.
See Figure S3 as well.
Cell Reports 2012 1, DOI: ( /j.celrep )
Copyright © 2012 The Authors Terms and Conditions

6. Figure 4
Downregulation of PcG Genes Can Mimic the Challenge Phenotypes in drm::neoGFP Flies and Increase the Survival under Challenge
(A) Representative images of neoGFP expression in the midgut of unchallenged drm::neoGFP larvae (left), challenged larvae (center, 400 μg/ml of G418), and drm::neoGFP larvae expressing RNAi against ph-p under the control of the drm promoter (right).
(B) Statistics of neoGFP expression for the cases described in (A). Data are represented as mean expression ±SE measured in midgut tissues pooled from three replicated experiments.
(C) Left view shows representative delay in development of unchallenged drm::neoGFP larvae with ph-p RNAi (compared with +/+). Right view illustrates mean delay ±SE in four replicated experiments.
(D) A significant fraction of unchallenged larvae carrying RNAi against ph-p form very small pupae. Left view is a representative image of a small pupa (white arrow). Right view shows normalized length histograms for drm::neoGFP pupae with and without ph-p RNAi pooled from three replicated experiments.
(E) Same as (A) for the proventriculus of larvae carrying RNAi against pho (right).
(F) Statistics of neoGFP expression for the cases described in (E). Data are represented as mean expression ±SE measured in proventriculi pooled from three replicated experiments.
(G) Survival of drm::neoGFP larvae with or without pho RNAi (pho RNAi, n = 10 vials; +/+, n = 24) Data are represented as mean survival ratio (versus unchallenged) ±SE measured in vials (n) pooled from three to six replicated experiments.
(H) Hypothesized model.
∗p < 0.05, ∗∗p < 0.001, ∗∗∗p < 10−5 (Student's t test).
See Figure S4 as well.
Cell Reports 2012 1, DOI: ( /j.celrep )
Copyright © 2012 The Authors Terms and Conditions

7. Figure 5
Induced Developmental Patterns Are Epigenetically Inherited across Multiple Generations
(A) A scheme of the basic inheritance experiment.
(B) Representative images demonstrating elevated expression of neoGFP (green) and hairy (red) in the proventriculus of challenged F1 hairy::neoGFP larva (center versus left), and in unchallenged F9 offspring following eight generations without G418 (right versus left).
(C) Statistics of neoGFP expression in the proventriculi of unchallenged (n = 31) and challenged (n = 23) F1 hairy::neoGFP larvae, and in unchallenged F2, F3, and F6 generations with (n = 61, 20, 22, respectively) or without history of exposure on the F1 generation (n = 50, 18, 20). Data are represented as mean expression per proventriculus ±SE based on the noted numbers of proventriculi (n) pooled from three to five replicated experiments. The dashed lines were drawn to separate results obtained for the F1 generation from those of later generations.
(D) Representative delay in development of unchallenged F10 hairy::neoGFP larvae with past exposure on F1 compared to larvae without past exposure.
(E) Statistics of the delay (mean ± SE in three to four replicated experiments). The dashed lines were drawn to separate results obtained for the F1 generation from those of later generations.
(F) Example of reversal of inheritance in F6 indicated by mean neoGFP expression in reversed (n = 22) and nonreversed (n = 22) F6 hairy::neoGFP larvae with past exposure on F1 (left). Reversal of the inheritance of the delay in development for the F6 hairy::neoGFP larvae (right).
∗p < 0.05, ∗∗p < 10−3, ∗∗∗p < 10−5 (Student's t test).
See Figure S5 as well.
Cell Reports 2012 1, DOI: ( /j.celrep )
Copyright © 2012 The Authors Terms and Conditions

8. Figure 6
Inheritance of Additional Phenotypes in Other Scenarios of Challenge
(A) Representative images demonstrating elevated expression of neoGFP in the midgut of challenged F1 drm::neoGFP larvae (center versus left), and in unchallenged F3 offspring following two generations without G418 (right versus left).
(B) Mean neoGFP expression (±SE) in the midgut of unchallenged (n = 23) and challenged (n = 15) F1 drm::neoGFP larvae, and in unchallenged F2, F3 generations with (n = 12, 19, respectively) or without (n = 16, 20) past exposure on the F1 generation. The dashed lines were drawn to separate results obtained for the F1 generation from those of later generations.
(C) Representative delay in development of unchallenged F2 drm::neoGFP larvae with a history of exposure on F1 compared to larvae without past exposure.
(D) Mean delay (±SE) in unchallenged drm::neoGFP F2 and F3 generations with past exposure on F1 versus no past exposure. Based on three to four replicated experiments. The dashed lines were drawn to separate results obtained for the F1 generation from those of later generations.
(E) Representative images of abnormal wings in Hsp70::neoGFP adult fly that was challenged during its development (center) and in an unchallenged offspring following three generations without G418 (right). Insets display expression of neoGFP at the base of the abnormal wings.
(F) Incidence of wing abnormalities in unchallenged F1 Hsp70::neoGFP flies (n = 7 experiments, 0 of 615 flies), challenged F1 (200–400 μg/ml G418; n = 7, 31 of 240 flies, 13%), unchallenged offspring within a range of 1–6 generations after last exposure to G418 (F2–F7, n = 6, 11 of 192 flies, 6%), and flies from the same range of generations without past exposure in their ancestors (n = 4, 0 of 708 flies). The dashed lines were drawn to separate results obtained for the F1 generation from those of later generations.
(G) Representative delay in development of unchallenged F3 Hsp70::neoGFP larvae with history of exposure on the F1 ancestors compared to larvae without exposure on F1.
(H) Mean delay (±SE) in unchallenged F2 and F3 Hsp70::neoGFP larvae with history of exposure on the F1 ancestors. Based on two to four replicated experiments. The dashed lines were drawn to separate results obtained for the F1 generation from those of later generations.
∗p < 0.05, ∗∗p < 10−3, ∗∗∗p < 10−5 (Student's t test).
Cell Reports 2012 1, DOI: ( /j.celrep )
Copyright © 2012 The Authors Terms and Conditions

9. Figure 7
Heritable and Nonheritable Expression Changes in Response to the Challenge
(A) Nonheritable induction of expression of detoxification, GstD genes in proventriculi of challenged F1 hairy::neoGFP larvae compared to unchallenged larvae (based on microarray data). Data are represented as mRNA levels relative to unchallenged F1 ±SE.
(B) A list of general stress response genes compiled from the literature exhibited a small but significant overlap (12%, p < 10−11, hypergeometric test) with the response to the challenge in the proventriculi of hairy::neoGFP larvae.
(C) Lack of inheritance of expression changes of the 33 stress genes indicated in (B) (p < 10−22, Student's t test).
(D) mRNA expression fold change in the proventriculi of challenged F1 (compared to unchallenged hairy::neoGFP larvae; x axis) versus fold changes in unchallenged F3 with a history of exposure in F1 (compared to F3 without a history of exposure; y axis). Shown are genes with fold change above 1.6-fold in both F1 and F3 (189 genes). Red and blue data correspond to positive and negative correlation, respectively.
(E) Example of genes with heritably modified expression.
See Figure S6 and Table S1 as well.
Cell Reports 2012 1, DOI: ( /j.celrep )
Copyright © 2012 The Authors Terms and Conditions

10. Figure S1
Additional Activations of Developmental Promoters in Response to the Challenge, Related to Figure 2
(A) Expression of neoGFP in part of the midgut of challenged (400ug/ml of G418, bottom) and unchallenged hairy::neoGFP larvae (top). Shown are representative images.
(B) Histograms of neoGFP expression in proventriculi of unchallenged (blue, n = 32) and challenged hairy::neoGFP larvae (red, n = 23). Counts are normalized to the number of proventriculi dissected in each case.
(C) High resolution mapping of the enhancer trap hairy-GAL4 transgene (Brand and Perrimon, 1993) revealed that the GAL4 is located between the hairy promoter and hairy coding sequence.
(D) Representative images demonstrating changes in activation of elav (top) and ftz.ng promoters (bottom) in challenged (right) versus unchallenged larvae (left). Arrows indicate regions of neoGFP expression in the larval midgut. Left drawings designate the respective regions of expression in the midgut.
(E) Quantification of developmental delay in development of challenged versus unchallenged elav::neoGFP (n = 3), wg::neoGFP (n = 3), and ftz.ng::neoGFP (n = 2) larvae. Data represented as mean delay ± SE based on the number of replicated experiments (n). Significance analysis corresponds to the difference compared with unchallenged larvae from the same line. (∗∗) p-value < 0.001, and (∗) p-value < 0.05 (student's t test).
(F) Exposure of Hsp70::neoGFP larvae to 400ug/ml of G418 led to induction of neoGFP (green) in the midgut (left) but not in the proventriculus (center). In this case, the neoGFP was not under the control of the hairy promoter, but the hairy gene was nonetheless induced in the proventriculus (right), suggesting that its induction is independent of the state of expression of neoGFP.
(G) Representative images of Act5C::neoGFP adult fly expressing neoGFP ectopically in all tissues. None of the Act5C::neoGFP flies showed morphological abnormalities (0%, 0/459).
(H) Same as (G) for dissected larva.
Cell Reports 2012 1, DOI: ( /j.celrep )
Copyright © 2012 The Authors Terms and Conditions

11. Figure S2
Survival Is Correlated with the Ability to Express the Resistance Gene in Gut Regions, Related to Figure 2
Collection of fly lines, each expressing neoGFP in various tissues. Shown are (from left to right) GAL4 driver lines, list of tissues in which neoGFP is expressed, representative images of neoGFP expression, and survival ratios under exposure to 400ug/ml of G418.
(A–E) Lines with low or no expression of neoGFP in the gut and substantial expression of neoGFP in other tissues including: salivary glands (B–E), fat body (A and B), brain (A and E), and restricted expression in discs (C). Survival in all these lines does not exceed 12%.
(F–J) Lines expressing substantial levels of neoGFP in various regions within the midgut or foregut (with or without expression in other tissues listed in A–E). Survival of all these lines is above 30%.
Survival data represented as mean survival ratio ± SE.
Cell Reports 2012 1, DOI: ( /j.celrep )
Copyright © 2012 The Authors Terms and Conditions

12. Figure S3
HP1 and Su(var)3-9 Are Not Involved in the Response to the Challenge, Related to Figure 3
(A) Survival analysis of fly larvae overexpressing Su(var)3-9 and Pc under the control of a heat-shock inducible promoter (hs-Su(var)3-9-EGFP, n = 7 vials, hs-Pc data taken from Figure 3F). Unlike hs-Pc larvae, survival of hs-Su(var)3-9-EGFP hairy::neoGFP larvae was not compromised compared to challenged larvae lacking the hs-Su(var)3-9 transgene (+/+, n = 10 and n = 33 with and without heat-shock treatment, respectively). Data represented as mean survival ratio ± SE in fly vials.
(B) Quantification of neoGFP levels in the proventriculi of challenged and unchallenged hairy::neoGFP larvae heterozygous mutant for Su(var)3-9 (Su(var)3-92/+, n = 20 and n = 23 for challenged and unchallenged larvae, respectively). Unlike the Pc1 mutation (data adopted from Figure 3C), mutation in Su(var)3-9 do not cause elevated expression of neoGFP in challenged or unchallenged larvae. Data represented as mean expression per proventriculus ± SE.
(C) Quantification of neoGFP levels in the proventriculi of larvae overexpressing HP1 under the control of a heat-shock inducible promoter (hs-HP1, n = 20). In contrast to hs-Pc larvae (adopted from Figure 3E), the levels of neoGFP in hs-HP1 larvae were not reduced compared to unchallenged larvae lacking the hs-HP1 transgene (+/+, n = 10 and n = 23 with and without heat-shock treatment, respectively). Data represented as mean expression per proventriculus ± SE.
(D) Quantification of neoGFP levels in the proventriculi of challenged and unchallenged hairy::neoGFP larvae heterozygous mutant for HP1 (HP15/+, n = 18 and n = 19 for challenged and unchallenged larvae, respectively) or that carry HP1 RNAi (n = 18 and n = 19 for challenged and unchallenged larvae, respectively). Unlike the heterozygous Pc1 mutant, RNAi and mutation in HP1 do not show elevated expression of neoGFP. Data represented as mean expression per proventriculus ± SE (∗) p-value < 0.05, (∗∗∗) p-value < 10−5 (student's t test).
Cell Reports 2012 1, DOI: ( /j.celrep )
Copyright © 2012 The Authors Terms and Conditions

13. Figure S4
Developmental Promoters that Were Activated by the Challenge Are Located within Polycomb Regions, Related to Figure 4
Binding profiles of Ph and Pc, and profiles of H3 tri-methylation of lysine 27 (k27me3) at the genomic loci of drm, ftz, hairy and elav genes (Schuettengruber et al., 2009). Colored bars represent published classifications of epigenetic domains (Filion et al., 2010; Sexton et al., 2012), indicating that the promoters of the above genes are localized to Polycomb domains (Blue regions). Vertical dashed blue lines denote the putative borders of the consensus Polycomb domain. Black regions correspond to non-Polycomb domains (neutral –‘null’– and heterochromatic regions in Sexton et al. or Filion et al., respectively). Red and yellow regions correspond to euchromatic domains (in the Filion classification). Red line and dashed black line (top of hairy and elav genes) mark the gene and its 15kb upstream region, respectively. The black solid lines (top of drm and ftz genes) represent the regulatory regions of GAL4 in the respective non enhancer-trap lines.
Cell Reports 2012 1, DOI: ( /j.celrep )
Copyright © 2012 The Authors Terms and Conditions

14. Figure S5
Induction of Heritable Modifications in Homozygous hairy::neoGFP Flies and in Other Transgenic Lines, Related to Figure 5
(A) Representative images demonstrating elevated neoGFP expression in the proventriculus of unchallenged homozygous F3 hairy::neoGFP larva with past exposure of ancestors in F1 (bottom right) compared to larva without a history of exposure in the F1 ancestors (Top right). Left: Induction of neoGFP in the proventriculus of challenged versus unchallenged F1 larvae (bottom versus top).
(B) Quantification of neoGFP expression in the proventriculi of unchallenged homozygous F2 and F3 hairy::neoGFP larvae (n = 17, 4, respectively) with past exposure on the F1 generation. Expression of neoGFP in the proventriculi of unchallenged (n = 16) and challenged (n = 12) F1 homozygous are shown to the left. Expression is normalized to unchallenged F1 larvae. Data represented as mean expression ± SE based on the noted numbers of proventriculi (n) pooled from 2 experiments for each case. Significance analysis corresponds to the difference with respect to unchallenged F1 larvae (student's t test).
(C) Same as (A) for heterozygous, hairy::neoGFP larvae in which the UAS-neoGFP was inserted in a genomic location (line 7) that is different than the line used throughout this study (line 9).
(D) Right: Inheritance of the delay in development in unchallenged F6 offspring that no longer carry the hairy-GAL4 and UAS-neoGFP transgenes but with a history of ancestor exposure to G418 in F1. Data represented as mean fraction of pupae ± SE in 2-4 vials. Left: Schematic representation of the experiment.
(E) Survival of F3 larvae which do not carry the trans-genes (but nonetheless show the inheritance of developmental delay). These flies were subjected to the toxic challenge after 2 generations of inheritance (F3). Data was collected from 3 replicated experiments. None of these larvae survived the challenge without the hairy-GAL4 and UAS-neoGFP transgenes.
(F) The repression of PcG genes in the proventriculi of challenged F1 hairy::neoGFP larvae (blue) was not observed in proventriculi of unchallenged F3 larval offspring (black) which, nonetheless, inherited the delay in development and induction of neoGFP (data not shown). Shown are microarray (left) and real-time qPCR measurements (right). Data represented as mean expression ± SE (∗∗∗) p-value < 10−5, (∗∗) p-value < 10−3, (∗) p-value < (student's t test). NS = non-significant change (p-value > 0.05).
Cell Reports 2012 1, DOI: ( /j.celrep )
Copyright © 2012 The Authors Terms and Conditions

15. Figure S6
Delayed Development in Response to Reduced Temperatures Is Not Associated with Induction of hairy in the Proventriculus, Related to Figure 7
(A) Real-time qPCR-based verification of the non-heritable induction of GstD2, GstD4, and GstD5. Data represented as mean expression relative to unchallenged F1 hairy::neoGFP larvae ± SE.
(B) Left: ∼3 day delay in development of hairy::neoGFP larvae that were developed at 20°C compared to larvae that were kept at 25°C. Data represented as mean fraction of pupae ± SE in 2 vials. Right: The delay in response to reduced temperatures was not associated with induction of hairy in the proventriculus (representative image).
Cell Reports 2012 1, DOI: ( /j.celrep )
Copyright © 2012 The Authors Terms and Conditions

16. Figure S7
neoGFP Is Localized to Expressing Cells, Related to Experimental Procedures
High-resolution confocal microscopy image, demonstrating localization of the neoGFP to expressing cells (red), without leakage to the extracellular space or to neighboring cells (white).
Cell Reports 2012 1, DOI: ( /j.celrep )
Copyright © 2012 The Authors Terms and Conditions