Estrogens regulate early embryonic development of the olfactory sensory system via estrogen-responsive glia

ABSTRACT Estrogens are well-known to regulate development of sexual dimorphism of the brain; however, their role in embryonic brain development prior to sex-differentiation is unclear. Using estrogen biosensor zebrafish models, we found that estrogen activity in the embryonic brain occurs from early neurogenesis specifically in a type of glia in the olfactory bulb (OB), which we name estrogen-responsive olfactory bulb (EROB) cells. In response to estrogen, EROB cells overlay the outermost layer of the OB and interact tightly with olfactory sensory neurons at the olfactory glomeruli. Inhibiting estrogen activity using an estrogen receptor antagonist, ICI182,780 (ICI), and/or EROB cell ablation impedes olfactory glomerular development, including the topological organisation of olfactory glomeruli and inhibitory synaptogenesis in the OB. Furthermore, activation of estrogen signalling inhibits both intrinsic and olfaction-dependent neuronal activity in the OB, whereas ICI or EROB cell ablation results in the opposite effect on neuronal excitability. Altering the estrogen signalling disrupts olfaction-mediated behaviour in later larval stage. We propose that estrogens act on glia to regulate development of OB circuits, thereby modulating the local excitability in the OB and olfaction-mediated behaviour.


Experimental zebrafish lines
The elavl3(huC):GCaMP6s transgenic zebrafish line, Tg(elavl3:GCaMP6s), used in this study has been described previously in (Winter et al., 2017), and was originally supplied by Misha B. Ahrens (Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, Virginia, USA).The ERE:GFP line, Tg(ERE:Gal4ff; UAS:GFP), carrying two transgenes, ERE:Gal4ff and UAS:GFP, in a skin pigment free Casper background was developed at the University of Exeter (Green et al., 2016).ERE:mCherry line, Tg(ERE:Gal4ff; UAS:nfsBmCherry), was created for this study encoding an EREderived Gal4ff (ERE:Gal4ff) and a UAS-derived mCherry-nitroreductase fusion protein (UAS:nfsBmCherry).This was developed from our original ERE:GFP line by replacing UAS:GFP with UAS:nfsBmCherry.Briefly, homozygous Tg(ERE:Gal4ff; UAS:GFP) zebrafish and a TG zebrafish carrying fms:Gal4 and UAS:nfsBmCherry (originally supplied by Stephen A. Renshaw) (Gray et al., 2011) were pair-crossed to obtain offspring heterozygous for these four transgenes.The heterozygous offspring were raised to sexual maturity and in-crossed.From this in-crossed generation, embryos were collected and exposed to 50 ng/L 17α-ethinyloestradiol (EE2)(Sigma-Aldrich) from 0 to 3 days post fertilization (dpf) and screened for estrogen-dependent mCherry expression in estrogen responding tissues (Green et al., 2016;Lee et al., 2012) with an absence of either GFP expression in these tissues (UAS:GFP null) or mCherry expression in macrophages (fms:Gal4 null).These screened larvae were raised to sexual maturity and further in-crossed and following this F2 embryos were again screened for estrogen dependent mCherry expression and absence of other TG expression, as described above.Finally, homozygous progeny of Tg(ERE:Gal4ff; UAS:nfsBmCherry) was confirmed by pair-crossing with wild-type (WIK strain) adults and assessing the ratio of fluorescent/non-fluorescent offspring, subsequently establishing Tg(ERE:Gal4ff; UAS:nfsBmCherry), the ERE:mCherry line.Development: doi:10.1242/dev.199860: Supplementary information

Research Campus, Howard Hugh
s Medical Institute, Ashburn, Virginia, USA).The ERE:GFP line, Tg(ERE:Gal4ff; UAS:GFP), carrying two transgenes, ERE:Gal4ff and UAS:GFP, in a skin pigment free Casper background was developed at the University of Exeter (Green et al., 2016).ERE:mCherry line, Tg(ERE:Gal4ff; UAS:nfsBmCherry), was created for this study encoding an EREderived Gal4ff (ERE:Gal4ff) and a UAS-derived mCherry-nitroreductase fusion protein (UAS:nfsBmCherry).This was developed from our original ERE:GFP line by replacing UAS:GFP with UAS:nfsBmCherry.Briefly, homozygous Tg(ERE:Gal4ff; UAS:GFP) zebrafish and a TG zebrafish carrying fms:Gal4 and UAS:nfsBmCherry (originally supplied by Stephen A. Renshaw) (Gray et al., 2011) were pair-crossed to obtain offspring heterozygous for these four transgenes.The heterozygous offspring were raised to sexual maturity and in-crossed.From this in-crossed generation, embryos were collected and exposed to 50 ng/L 17α-ethinyloestradiol (EE2)(Sigma-Aldrich) from 0 to 3 days post fertilization (dpf) and screened for estrogen-dependent mCherry expression in estrogen responding tissues (Green et al., 2016;Lee et al., 2012) with an absence of either GFP expression in these tissues (UAS:GFP null) or mCherry expression in macrophages (fms:Gal4 null).These screened larvae were raised to sexual maturity and further in-crossed an

Supplementary Materials and Methods
s and Methods

Tg(ERE:Gal4ff;UAS:nfsBmCherry;cyp19a1b:GFP), the ERE:mCherry x cyp19:GFP line, was generated by natural-crossing of ERE:mCherry and cyp19a1b:GFP zebrafish (Brion et al., 2012)(the latter generously provided by Olivier Kah).The homozygous for three transgenes (ERE:Gal4ff, UAS:nfsBmCherry and cyp19a1b:GFP) were screened using a similar procedure to that described for the homozygous screening for ERE:mCherry.


Mosaic expression of DsRed in a subset of EROB cells.

To map the morphology of an individual EROB cell, pT2A UAS:D Tg(ERE:Gal4ff;UAS:nfsBmCherry;cyp19a1b:GFP), the ERE:mCherry x cyp19:GFP line, was generated by natural-crossing of ERE:mCherry and cyp19a1b:GFP zebrafish (Brion et al., 2012)(the latter generously provided by Olivier Kah).The homozygous for three transgenes (ERE:Gal4ff, UAS:nfsBmCherry and cyp19a1b:GFP) were screened using a similar procedure to that described for the homozygous screening for ERE:mCherry.

Mosaic expression of DsRed in a subset of EROB cells.
To map the morphology of an individual EROB cell, pT2A UAS:DsRed-Ex plasmid DNA (Miyasaka et al., 2014) (25 ng/l) and Tol2 transposase mRNA (25 ng/l) were co-injected into one cell stage ERE:GFP embryos and subsequently exposed to EE2 (100 ng/L) from 1 to 96 hpf.This rendered a mosaic expression of DsRed in a subset of oestrogen responding cells which was derived from oestrogen-dependent ERE:Gal4ff activation in these cells.The morphology of DsRed expressing EROB cells at 4 dpf was examined using a Zeiss LSM510 confocal microscope.

ed-Ex plasmid DNA (Mi
asaka et al., 2014) (25 ng/l) and Tol2 transposase mRNA (25 ng/l) were co- njected into one cell stage ERE:GFP embryos and subsequently exposed to EE2 (100 ng/L) from 1 to 96 hpf.This rendered a mosaic expression of DsRed in a subset of oestrogen responding cells which was derived from oestrogen-dependent ERE:Gal4ff activation in these cells.The morphology of DsRed expressing EROB cells at 4 dpf was examined using a Zeiss LSM510 confocal microscope.


Immunohistochemistry

Whole mount GFAP, Acetylated tubulin, SV2, Hemocyanin and mCherry staining:

The zebrafish embryo-larvae chemical exposure conditions are described in the "Methods" of the main manuscript.At 96 hpf, exposed embryo-larvae were washed, anaesthetised with 0.03% NM222 and fixed in 4% paraformaldehyde (PFA) in 1 x PBS at 4°C overnight.Fixed samples were then washed once with 0.5 % Triton in 1 x PBS (PBTr) and transferred to 50% methanol / 50% PBTr for 5 minutes, then to 100% methanol for 5 minutes, followed by placing them into fresh 100% methanol at -20°C (for between overnight -2 weeks).The samples were then rehydrated with 50% methanol / 50% PBTr for 5 minutes and subsequently with PBTr for 5 minutes.Embryos were then treated with 3 x proteinase K solution (Sigma-Aldrich) (final 30 µg/ml) for 30 minutes and then was

sher Scientific, A11001) and Alexa 594 conjugated ant
-rabbit IgG antibody (1:300 in blocking buffer, ThermoFisher Scientific, R37121) at 4°C overnight.The embryos were washed three times for 30 minutes with PBTr, and once with 1 x PBS for 10 minutes and then incubated with Hochest 33342 (1:3000 in 1 x PBS, ThermoFisher Scientific) for 30 minutes at room temperature to stain the nuclei.The stained embryos were washed twice for 10 minutes with 1 x PBS at room temperature and kept in 50% Glycerol in 1 x PBS at 4°C until imaging.


Whole mount HuC (Elavl3), SOX2 and GFP staining:

Chemical exposure conditions for ER

Whole mount HuC (Elavl3), SOX2 and GFP staining:
Chemical exposure conditions for ERE:GFP embryos were performed as described in "Methods" of the main manuscript.Fixation, permeabilisation and blocking of the embryos were conducted as described above.The blocked embryos were incubated with mouse anti-HuC antibody (the antibody against human ELAVL3 protein) (1:200 in blocking solution, ThermoFisher Scientific, A21271, 16A11) and rabbit anti-GFP (1:500 in blocking, AMS Biotechonology, TP401), or rabbit anti-SOX2 antibody (1 : 100 in blocking solution, GeneTex, GTX124477) and mouse anti-GFP (1 : 500 in blocking solution, GeneTex, GTX82564) at 4°C overnight.The embryos were then washed and incubated with secondary antibodies and Hoechst, as described above.

GFP embryos were p
rformed as described in "Methods" of the main manuscript.Fixation, permeabilisation and blocking of the embryos were conducted as described above.The blocked embryos were incubated with mouse anti-HuC antibody (the antibody against human ELAVL3 protein) (1:200 in blocking solution, ThermoFisher Scientific, A21271, 16A11) and rabbit anti-GFP (1:500 in blocking, AMS Biotechonology, TP401), or rabbit anti-SOX2 antibody (1 : 100 in blocking solution, GeneTex, GT

24477) and mouse anti-GFP (1 : 500 in blocking solution, GeneTex,
GTX82564) at 4°C overnight.The embryos were then washed and incubated with secondary antibodies and Hoechst, as described above.


Cryosection sample staining for OSNs and EROB cells:

The conditions for chemical exposure and for sample fixation were performed as described in the main manuscript Methods.Fixed embryos were cryo-protected in 30% sucrose in 1 x PBS and embedded in NEG-50 solution (ThermoFisher Scientific).Frontal sections of the forebrain of 4 dpf ERE:mCherry embr

Cryosection sample staining for OSNs and EROB cells:
The conditions for chemical exposure and for sample fixation were performed as described in the main manuscript Methods.Fixed embryos were cryo-protected in 30% sucrose in 1 x PBS and embedded in NEG-50 solution (ThermoFisher Scientific).Frontal sections of the forebrain of 4 dpf ERE:mCherry embryo (7 m thickness) were obtained using CM1950 cryostat (Leica).Sectioned samples were rehydrated with 0.2 % Triton in 1 x PBS (PBTr0.2) and blocked with 0.25 % milk in (1:25,000 in 1 x PBS, ThermoFisher Scientific) for 30 minutes.Finally, the samples were washed twice for 10 minutes with 1 x PBS and mounted with ProLong Gold antifade reagent (ThermoFisher scientific).

(7 m thickness) were obtained using CM1950 cryostat (Leica).Sectioned
samples were rehydrated with 0.2 % Triton in 1 x PBS (PBTr0.2) and blocked with 0.25 % milk in (1:25,000 in 1 x PBS, ThermoFisher Scientific) for 30 minutes.Finally, the samples were washed twice for 10 minutes with 1 x PBS and mounted with ProLong Gold antifade reagent (ThermoFisher scientific).


Confocal imaging:

Whole mount stained zebrafish embryos and immune-stained cyrosection samples of the zebrafish forebrain were analysed using laser scanning confocal microscopy (Zeiss LSM 510 and Zeiss LSM 810, Zeiss, Germany).For whole mount stained embryos, the samples were mounted in low melting agarose as described above.Confocal z stacks across the entire forebrain or the OBs were obtained using Zeiss LSM510 or Zeiss LSM810 keeping the optimal or at a set z step size.


Co-localisation analysis for EROB cells and GFAP or aromatase B:

Co-localisation of EROB cells and GFAP, or aromatase, were measured using a method described in (Steinfeld et al.

Confocal imaging:
Whole mount stained zebrafish embryos and immune-stained cyrosection samples of the zebrafish forebrain were analysed using laser scanning confocal microscopy (Zeiss LSM 510 and Zeiss LSM 810, Zeiss, Germany).For whole mount stained embryos, the samples were mounted in low melting agarose as described above.Confocal z stacks across the entire forebrain or the OBs were obtained using Zeiss LSM510 or Zeiss LSM810 keeping the optimal or at a set z step size.

Co-localisation analysis for EROB cells and GFAP or aromatase B:
Co-localisation of EROB cells and GFAP, or aromatase, were measured using a method described in (Steinfeld et al., 2015).Briefly, Pearson's correlation coefficients in two dimension were calculated between pixel intensities of two distinct channels over a range of image shift (dx) for each volume.Co-localisation of two signals results in a distinct peak at dx = 0.This was compared against control images where the image of one channel was flipped 90° (control flipped image).
2015).Briefly, Pearson's correlation coefficients in two dimension were calculated between pixel intensities of two distinct channels over a range of image shift (dx) for each volume.Co-localisation of two signals results in a distinct peak at dx = 0.This was compared against control images where the image of o e channel was flipped 90° (control flipped image).


3D reconstruction of confocal images and OB glomeruli volume analysis:

The confocal optical z section images were processed to obtain 3D images of the OB glomeruli by 3D Viewer in Fiji.The same data were also processed to quantify the volumes of glomeruli by measuring the areas of ROIs throughout the optical z section images

3D reconstruction of confocal images and OB glomeruli volume analysis:
The confocal optical z section images were processed to obtain 3D images of the OB glomeruli by 3D Viewer in Fiji.The same data were also processed to quantify the volumes of glomeruli by measuring the areas of ROIs throughout the optical z section images.Briefly, a set threshold (IJ_IsoData method) was applied to raw confocal optical z section images of SV2 stained forebrains, generating binary images of SV2+ glomeruli throughout the z slices.The outlines of five glomeruli (maG, dG, dlG, mdG3 and mdG1-6), which are located in dorsal areas of the OB and interact with EROB cells, were drawn manually using the freehand selection tool.The areas of those outlined glomeruli were measured individually from each z slice and the relative volume of a glomerulus was represented by the sum of the measured areas from the collective z slices.Statistics were performed with R (version 3.2.3-4).Normality of the data was confirmed with the Shapiro test and Levene's test showed homogeneity of variances and a linear model was built.The one-way ANOVA in conjunction with Tukey's post-hoc test was performed for pair-wise comparisons of the treatments, using the 'multcomp' package in R. Development: doi:10.1242/dev.199860: Supplementary information Briefly, a set threshold (IJ_IsoData method) was applied to raw confocal optical z section images of SV2 stained forebrains, generating binary images of SV2+ g omeruli throughout the z slices.The outlines of five glomeruli (maG, dG, dlG, mdG3 and mdG1-6), which are located in dorsal areas of the OB and interact with EROB cells, were drawn manually using the freehand selection tool.The areas of those outlined glomeruli were measured individually from each z slice and the relative volume of a glomerulus was represented by the sum of the measured areas from the collective z slices.Statistics were performed with R (version 3.2.3-4).Normality of the data was confirmed with the Shapiro test and Levene's test showed homogeneity of variances and a linear model was built.The one-way ANOVA in conjunction with Tukey's post-hoc test was performed for pair-wise comparisons of the treatments, using the 'multcomp' package in R. Development: doi:10.1242/dev.199860: Supplementary information


Development • Supplementary information pERK immunostaining

The conditions for chemical exposure and for sample fixation were performed as described in the main manuscript Methods.At 5 dpf, ERE:mCherry larvae were exposed with or without 100M cadavarine (D22606, Sigma-Aldrich) for 5 minutes at 28 °C (5 larvae/1ml of zebrafish egg water/ well of 12 well plate), and then immediately fixed by adding 1ml of 8% PFA in PBS to give final 4% PFA.The larvae in PFA were incubated for 2 hours at 28 °C, and subsequently overnight at 4°C.Fixed embryos were washed three times with tris-buffered saline (TBS) and treated with a sucrose gradient (10 %, 20 %, 30 %) and then stored in 30% sucrose at 4°C for cryoprotection.The samples were embedded in NEG-50 solution (ThermoFisher Scientific) and frozen minimally overnight at -80°C.The forebrain of the mounted larvae were then transversely sectioned at 10 m t

Development • Supplementary information pERK immunostaining
The conditions for chemical exposure and for sample fixation were performed as described in the main manuscript Methods.At 5 dpf, ERE:mCherry larvae were exposed with or without 100M cadavarine (D22606, Sigma-Aldrich) for 5 minutes at 28 °C (5 larvae/1ml of zebrafish egg water/ well of 12 well plate), and then immediately fixed by adding 1ml of 8% PFA in PBS to give final 4% PFA.The larvae in PFA were incubated for 2 hours at 28 °C, and subsequently overnight at 4°C.Fixed embryos were washed three times with tris-buffered saline (TBS) and treated with a sucrose gradient (10 %, 20 %, 30 %) and then stored in 30% sucrose at 4°C for cryoprotection.The samples were embedded in NEG-50 solution (ThermoFisher Scientific) and frozen minimally overnight at -80°C.The forebrain of the mounted larvae were then transversely sectioned at 10 m thickness consecutively from anterior to posterior using a CM1950 cryostat (Leica).Brain sections were transferred to SuperFrost Plus glass slides, (VWR), air-dried for 1 h and stored at -20°C with silicagel to remove air moisture.

ckness consecutively from anter
or to posterior using a CM1950 cryostat (Leica).Brain sections were transferred to SuperFrost Plus glass slides, (VWR), air-dried for 1 h and stored at -20°C with silicagel to remove air moisture.

Prior to staining, the slides were air dried for 1 h and washed with 0.1% Tween 20 in 1x TBS (TBST) for 10 minutes for rehydration.The samples were treated with a gradient of methanol (25, 50, 75 and 100% methanol) in TBST (for 5 min each) and then with 0.3% hydrogen peroxide in 100% methanol for 15 min to block endogenous peroxidase activity and reduce background signals in the Tyramide Signal Amplification (TSA) step (described below).The methanol exposure also removed all detectable ERE-mCherry signal.After washing with TBST, the slides were then washed with heat retrieval solution (Tris-EDTA buffer with 10 mM tris base, 1 mM EDTA, 0.05% Tween 20, pH 9.0) and immersed into a preheated retrieval solution for 10 minutes at room temperature.For antigen retrieval, the slides were then heated to approximately 95°C in heat retrieval solution for 10 minutes.The slides were washed 2 times for 15 minutes in TBS/0.1% Tween 20/0.1% Triton X-100 (TBSTwT) to ensure permeabilisation of the tissues, and then further washed with maleic buffer (150 mM maleic acid/100 mM NaCl/0.05% Tween 20, pH 7.4).Samples were then blocked with maleic blocking buffer (2% blocking reagent (Roche) in maleic buffer) for 1 h at room temperature.The Prior to staining, the slides were air dried for 1 h and washed with 0.1% Tween 20 in 1x TBS (TBST) for 10 minutes for rehydration.The samples were treated with a gradient of methanol (25, 50, 75 and 100% methanol) in TBST (for 5 min each) and then with 0.3% hydrogen peroxide in 100% methanol for 15 min to block endogenous peroxidase activity and reduce background signals in the Tyramide Signal Amplification (TSA) step (described below).The methanol exposure also removed all detectable ERE-mCherry signal.After washing with TBST, the slides were then washed with heat retrieval solution (Tris-EDTA buffer with 10 mM tris base, 1 mM EDTA, 0.05% Tween 20, pH 9.0) and immersed into a preheated retrieval solution for 10 minutes at room temperature.For antigen retrieval, the slides were then heated to approximately 95°C in heat retrieval solution for 10 minutes.The slides were washed 2 times for 15 minutes in TBS/0.1% Tween 20/0.1% Triton X-100 (TBSTwT) to ensure permeabilisation of the tissues, and then further washed with maleic buffer (150 mM maleic acid/100 mM NaCl/0.05% Tween 20, pH 7.4).Samples were then blocked with maleic blocking buffer (2% blocking reagent (Roche) in maleic buffer) for 1 h at room temperature.The primary antibody, anti-phospho p44/42 MAP Kinase (ERK1/2) rabbit polyclonal antibody (#4370, cell signaling), was then applied at 1:250 in the blocking buffer and the slides incubated overnight at 4°C.The slides were subsequently washed 2 times with TBSTwT, 1 time with maleic buffer and incubated with the maleic blocking buffer for 1h at room temperature for re-blocking.Slides were then incubated with 1:3000 goat anti-rabbit IgG-HRP (Dako) overnight at 4°C.The samples were then washed 3 times with maleic buffer for 30 min and once with TBS for     Signal/Noise Ratio (SNR) was calculated as "the mean of the odour-evoked pERK levels / the mean of the basal pERK levels" (shown below the x-axis).

rimary an
ibody, anti-phospho p44/42 MAP Kinase (ERK1/2) rabbit polyclonal antibody (#4370, cell signaling), was then applied at 1:250 in the blocking buffer and the slides incubated overnight at 4°C.The slides were subsequently washed 2 times with TBSTwT, 1 time with maleic buffer and incubated with the maleic blocking buffer for 1h at room temperature for re-blocking.Slides were then incubated with 1:3000 goat anti-rabbit IgG-HRP (Dako) overnight at 4°C.The samples were then washed 3 times with maleic buffer for 30 min and once with TBS for     Signal/Noise Ratio (SNR) was calculated as "the mean of the odour-evoked pERK levels / the mean of the basal pERK levels" (shown below the x-axis).



PBTr0.2 at room temperature for one hour.Blocked samples were then stained with rabbit anti-KLH (1:500 in the same blocking solution, Sigma-Aldrich, H0892) and mouse anti-mCherry antibody (1:500 in blocking, St. John's Laboratory, STJ34373) at 4°C overnight.The samples were washed three times for 15 minutes with PBTr0.2 and incubated with Alexa 594 conjugated antimouse IgG antibody (1:500 in blocking, ThermoFisher Scientific, A11012) and Alexa 488 conjugated anti-rabbit IgG antibody (1:500 in blocking, ThermoFisher Scientific, A11034) at room temperature for one hour.The stained samples were then washed three times for 15 minutesDevelopment: doi:10.1242/dev.199860:Supplementary information Development • Supplementary information with PBTr0.2 and once with 1 x PBS for 10 minutes and then incubated with Hochest 33342




Fig. S2.The projections of EROB cells interact with olfactory sensory neurons at olfactory glomeruli.(A) A 3D image of EROB cells (red) and acetylated tubulin positive axonal projections (green) in the OB (dorsal view).d, dorsal; v, ventral; l, lateral; a, anterior; p, posterior.(Bi-iv) Sequential confocal optical section images (3 µm steps) of EROB cells (red) and acetylated tubulin + axonal projections (green) in an mediodorsal area of the OB (indicated with white dotted rectangle in A) presenting along PBTr0.2 at room temperature for one hour.Blocked samples were then stained with rabbit anti-KLH (1:500 in the same blocking solution, Sigma-Aldrich, H0892) and mouse anti-mCherry antibody (1:500 in blocking, St. John's Laboratory, STJ34373) at 4°C overnight.The samples were washed three times for 15 minutes with PBTr0.2 and incubated with Alexa 594 conjugated antimouse IgG antibody (1:500 in blocking, ThermoFisher Scientific, A11012) and Alexa 488 conjugated anti-rabbit IgG antibody (1:500 in blocking, ThermoFisher Scientific, A11034) at room temperature for one hour.The stained samples were then washed three times for 15 minutesDevelopment: doi:10.1242/dev.199860:Supplementary information Development • Supplementary information with PBTr0.2 and once with 1 x PBS for 10 minutes and then incubated with Hochest 33342 Fig. S2.The projections of EROB cells interact with olfactory sensory neurons at olfactory glomeruli.(A) A 3D image of EROB cells (red) and acetylated tubulin positive axonal projections (green) in the OB (dorsal view).d, dorsal; v, ventral; l, lateral; a, anterior; p, posterior.(Bi-iv) Sequential confocal optical section images (3 µm steps) of EROB cells (red) and acetylated tubulin + axonal projections (green) in an mediodorsal area of the OB (indicated with white dotted rectangle in A) presenting along the dorso-ventral axis (top-bottom).White arrows, EROB cell-neuron contact sites in the OB glomeruli; white dotted lines, the outlines of glomeruli.(C) A 3D image of EROB cells (red) and olfactory glomeruli stained with SV2 antibody (cyan) (dorsal view).(Di-iv) Sequential confocal optical section images (3 µm steps) of EROB cells (red) and SV2+ olfactory glomeruli (cyan) in an mediodorsal area of the OB (indicated with white dotted rectangle in C) presenting along dorso-ventral axis (top-bottom).Arrows and lines indicate as above.

Fig. S5 .
Fig. S5.EE2 or ICI exposure do not affect the global development of the body or the brain size in 4 dpf zebrafish embryos.(A) Whole body confocal z projection images of control (top), EE2-(middle) and ICI-exposed (bottom) 4 dpf ERE:GFP embryos.Note that ERE:GFP embryos show a consistent basal GFP expression in the ganglions (marked as "G").EE2 exposure (from 1 to 96 hpf) induces GFP in the liver, heart and muscles, as previously reported.GFP expression in the brain is not easily visible in lateral orientation at this magnification (x10).ICI-exposure does not affect either the global development of the fish or the basal GFP expression.(B) Whole brain confocal z projection images of control, EE2-and ICI-exposed 4 dpf ERE:GFP embryos.Left end, a representative image showing the position of brain regions: olfactory bulb (OB); telencephalon, Tel; habenula, Ha; pineal, Pi; Optic tectum, Tec; cerebellum, Ce; medulla oblongata, MO.The outline of the brain is shown with a cyan line.EE2 exposure induces GFP expression predominantly in the OB (EROB cells)(the second image from right).(C-E) EE2 or ICI-exposure do not change the brain size.The width (D) and the A-P length (E) of brain regions were measured based on z projection images of Hoechst stained brain as indicated in C. Data were collected from control (white circles), EE2-(blue squares) and ICI-exposed (red triangles) 4 dpf ERE:GFP embryos.N=6-7.ANOVA with Tukey's test.

Fig. S6 .
Fig. S6.EE2 or ICI exposure do not affect the expression domains of Sox2 or HuC in 4 dpf zebrafish embryos.(A) Whole mount confocal z projection images of control (top), EE2-(middle) and ICI-exposed (bottom) 4 dpf ERE:GFP embryos stained with neural stem cell marker, Sox2 (red), pan-neuronal marker, HuC (green) and Hoechst for nuclei (blue).Expression domains of these markers as well as the morphology of the whole brain are similar among the experimental groups (N=3).