The activity-dependent histone variant H2BE modulates the life span of olfactory neurons

  1. Stephen W Santoro
  2. Catherine Dulac  Is a corresponding author
  1. Harvard University, United States
14 figures and 1 additional file

Figures

Mouse H2BE is detected exclusively in chemosensory neurons. (A) Analysis of H2be mRNA in the MOE of an 8-week old mouse, showing expression limited to sensory neurons. Boxed region is magnified (right). (B) Analysis of H2be mRNA in the VNO of an 8-week old mouse, showing expression limited to sensory neurons, especially in the apical zone. (C) Analysis of H2be mRNA in a sagittal section of an E14.5 mouse embryo (image from genepaint.org; ID: ES2590). (D) Profile of H2be mRNA levels in 61 mouse tissues (listed, right), showing exclusive expression in MOE and VNO. Data from GNF, now maintained by BioGPS (http://biogps.org/). (E) Alignment of H2BE and canonical H2B sequences. H2BE variant positions are highlighted in red; H2B PTM sites supported by >5 or ≤5 reports are highlighted in dark and light gray, respectively (http://www.phosphosite.org). Scale bars for (A), 500 µm; (B), 100 µm; (C), 1000 µm.

https://doi.org/10.7554/eLife.00070.003

Generation of an H2be:Flag-H2be transgenic mouse (referred to as Flag-H2be) reveals that H2BE levels are stereotyped according to OR identity. (A) Flag-H2be transgenic construct, generated through modification of a BAC containing the H2be genomic region by insertion of a FLAG-encoding sequence immediately upstream of the H2be CDS. (B,C) Representative images of FLAG-H2BE in MOE (B) and VNO (C) from 10-week old Flag-H2be transgenic mice. (D) Quantitative PCR (qPCR) analysis of Flag-H2be transgene mRNA levels in whole MOE tissue from three-week old Flag-H2be(+/−) transgenic mice. Signals were normalized to a value of three, corresponding to a primer pair recognizing all three H2be mRNAs (two endogenous and one transgenic). A primer pair specific for the Flag-tagged transgenic allele produces a normalized signal of approximately 1, indicating similar per-allele expression levels for the transgenic and endogenous H2be alleles. Negative control samples (−RT) were prepared by omitting reverse transcriptase during cDNA synthesis. (E) Colocalization analysis of FLAG-H2BE protein and H2be mRNA in the MOE of a 3-week old Flag-H2be(+/−) transgenic mouse. Observation of occasional basally-located neurons that are H2be-mRNA-positive and FLAG-H2BE-negative is likely due to an expected lag in protein production and accumulation following H2be transcription onset during neuronal development. (F) Colocalization analysis of Olfr867 or Olfr1463 (arrowheads) and FLAG-H2BE, showing representative ORs associated with high or low levels of H2BE, respectively. Mouse age: 10 weeks. (G) Quantification of average H2BE levels in neurons expressing specific ORs (n = from 4 to 36 neurons per OR examined; mean, 18). Gene symbols are from the Mouse Genome Informatics database (MGI; http://www.informatics.jax.org/). Scale bars for (B, left), 500 µm; (B, right) and (F), 20 µm; (C) and (E), 100 µm.

https://doi.org/10.7554/eLife.00070.004

Generation of an H2be-KO/GAP43-mCherry-KI mouse line (referred to as H2be-KO) reveals that loss of H2be causes defects in olfaction. (A) H2be-KO allele, constructed through replacement of the endogenous H2be CDS with a membrane-targeted mCherry-encoding sequence (Gap43-mCherry). (B,C) Intrinsic GAP43-mCherry fluorescence in the MOE (B) and OB (C) of H2be-KO mice, showing GAP43-mCherry localization to the cell membranes and processes of olfactory neurons. Mouse ages: (B), 6 months; (C), 2 months. Scale bar for (B), 20 µm; (C), 200 µm. (D) Performance of approximately 3-month old water-restricted H2be-KO and control littermates in discriminating between hexanol/hexanoic acid (left) or (+)/(−)-carvone (right) odor pairs to obtain water (n = 5 per genotype). *p<0.05. (E) Effects of H2be loss-of-function on odor-evoked electrical responses in the MOE. Electro-olfactogram traces (black and red) represent average responses to a 0.5-s stream of air from the head space of a 1% solution of isoamyl-acetate in mineral oil. Gray traces show timing of switching between the delivery of clean, de-odorized air (low), and odor-containing air (high). Results shown are representative of multiple trials, odorants, and concentrations; experimental procedures were adapted from those described previously (Waggener and Coppola, 2007).

https://doi.org/10.7554/eLife.00070.005

Loss of H2be causes defects in gene expression and OR expression frequencies. (A) Enriched gene ontology categories among genes up- (top) or down-regulated (bottom) in 6-month old H2be-KO vs WT MOEs, based on microarray analyses of whole MOE tissue (n = 6 samples/genotype, 2 animals/sample). (B) Multiplex qPCR analysis of OR mRNA levels in MOE tissue of six-month old H2be-KO and WT mice. Signals represent normalized ratios of specific OR mRNAs to Cnga2, which is unaltered in H2be-KO mice and used as an internal control (n = 6). (C) Expression differences in H2be-KO and WT MOEs (based on microarray analysis), plotted as a function of age. For simplicity, only ORs with differences >20% at age 6 months are shown, with up- and down-regulated ORs shown in red and black, respectively; statistical analysis included all ORs interrogated. (D,E) Representative images (left) and quantification (right) of Olfr867 (D) and Olfr1463 (E) expression frequencies in 6-month old H2be-KO and WT littermates (n = 3 mice, 10 sections per mouse). Scale bars, 200 µm. (F) Relationship between OR gene expression defects in 6-month old male H2be-KO mice (based on microarray analysis) and stereotypical H2BE levels as measured in male Flag-H2be mice. Red line, best fit. *p<0.05, **p<0.01, ***p<0.001, ****p<10−30.

https://doi.org/10.7554/eLife.00070.006
Figure 4—source data 1

Effects of H2be loss of function on gene expression in the main olfactory epithelium of 6-month old mice.

https://doi.org/10.7554/eLife.00070.007

Generation of an Omp:Flag-H2be transgenic mouse (referred to as H2be-GF) reveals that ectopic overexpression of H2be in olfactory neurons alters gene expression and OR expression frequencies. (A) H2be-GF transgenic construct, generated through replacement of the Omp CDS with a FLAG-H2BE-encoding sequence in a vector containing the Omp genomic region. (B) Analysis of FLAG-H2BE in the MOE of a 5-week old H2be-GF mouse, showing high transgene expression in all mature neurons, except for a band near zone 2 (Sullivan et al., 1996). Boxed region is magnified (right). (C) Gene ontology categories enriched among genes up- (top) or down-regulated (bottom) in 5-week old H2be-GF vs WT MOEs, based on microarray analyses of whole MOE tissue (n = 4 samples, 3 animals per sample) and WT (n = 6 samples, 2 animals per sample) mice. (D) Percentage of OR genes with significantly differential expression (FDR-adjusted p<0.05) in 5-week old H2be-GF and WT mice. (E) Relationship between OR gene expression defects in H2be-GF mice and transgene co-expression penetrance. Red line, best fit. (F,H) Colocalization of Olfr1277 or Olfr293 (arrowheads) and FLAG-H2BE in 5-week old H2be-GF mice, showing representative ORs associated with high- or low transgene penetrance, respectively. (G,I) Representative images (left) and quantification (right) of Olfr1277 (G) and Olfr293 (I) expression frequencies in 5-week old H2be-GF and WT littermates (n = 3 mice; 10 sections per mouse). *p<0.05, **p<0.01. Scale bars for (B), 500 µm; (F) and (H), 20 µm; (G) and (I), 200 µm.

https://doi.org/10.7554/eLife.00070.008
Figure 5—source data 1

Effects of the ectopic over-expression of H2be (expressed from an Omp-promoter-driven transgene and tagged with a FLAG epitope) on gene expression in the main olfactory epithelium of 5-week old mice.

https://doi.org/10.7554/eLife.00070.009

The expression onset of H2be follows OR choice during neuronal development. (A) H2BE is undetectable in an Olfr827+ neuron that is newly-differentiated (note its basal position in the epithelium; boxed and magnified, right), but expressed at uniformly high levels in mature Olfr827+ neurons. (B,D,E) MOE expression patterns in adult (left) or embryo (right) of H2be relative to Neurod1, Gap43, and Omp, which have expression onsets prior to, concurrently with, and following OR choice, respectively. All NEUROD1+ cells (B) and a fraction of basal GAP43+ (newly-differentiated) neurons (C) are H2BE (arrowheads), whereas a fraction of H2BE+ neurons are OMP (E; arrowheads). (C) Colocalization of endogenous H2be and Neurod1 mRNAs in the MOE of a 3-week old WT mouse. Mouse ages: (A), 3 months; (B, left), 4 months; (C), 3 weeks; (D, left) and (E, left), 10 months. Scale bars for (A) to (E), 20 µm.

https://doi.org/10.7554/eLife.00070.010

H2BE affects olfactory neuronal longevity, not OR choice. (A) Representative images (left) and quantification (right) of Olfr1507 and Olfr727 co-expression frequencies in H2be-KO and control littermates (n = 10 sections). (B) Labeled axons of Olfr17-expressing neurons form indistinguishable glomeruli in the OBs of H2be-KO and control littermates. (C,E) Representative images (C, left; arrowheads) and quantification (C, right; E) of apoptosis in mature neurons of H2be-KO (C) or H2be-GF (E) mice compared to controls (C: n = 3 mice, 12 sections per mouse; E: n = 10 sections). (D,F) Schematic of experimental analysis timeline (left) and quantification (right) of relative BrdU+ neuron frequencies in H2be-KO (D) and H2be-GF (F) mice compared to controls, respectively (D: n = 3 mice per timepoint, 12 sections per mouse; F: n = 10 sections per timepoint). F-H: Flag-H2be. (G) Representative images (left) and quantification (right) of neurogenesis in H2be-GF (GF) mice and Flag-H2be (F-H) controls by analysis of frequencies of BrdU+ neurons (arrowheads) 15 days post-injection of BrdU (T = 0 timepoint; n = 10 sections). *p<0.05, ***p<0.001. Mouse ages: (A) and (E), 4 months; (B) and (G), 2 months; (C), 15 months. Scale bars for (A) and (C), 20 µm; (B), 200 µm; (G), 100 µm.

https://doi.org/10.7554/eLife.00070.011
Figure 8 with 1 supplement

H2be is regulated by activity. (A,B) Effects of unilateral naris occlusion (UNO) on H2BE level in the MOE. (A) Representative images of FLAG-H2BE and Olfr733 colocalization (boxed regions magnified, right; Olfr733+ neurons, arrowheads). (B) Distributions (main) and averages (inset) of H2BE level within nuclei of randomly-sampled (main; inset, left; n = 200) or Olfr733+ (inset, right; n = 10) neurons on the two sides of the MOE. (C) Effects of UNO on intrinsic GAP43-mCherry fluorescence in glomeruli (arrowheads) within the OB of an H2be-KO heterozygous mouse. Reduced tyrosine hydroxylase (TH; a marker of olfactory activity) staining on the closed side indicates completeness of naris closure. (DF) Effects of odor exposure on H2BE level in olfactory neurons. Representative images (D) and quantification (E, left) of FLAG-H2BE in Olfr2+ neurons (D, arrowheads) exposed to odors or mineral oil (no odor). (E, right) Quantification of FLAG-H2BE in Olfr653+ neurons (chosen randomly as a negative control) exposed to odors or mineral oil (no odor). (F) Average H2BE level within neurons expressing odor-stimulated or control ORs (n = 20–60). (G,H) Representative image (G; boxed region magnified, right) and quantification (H) of the relationship between GAP43-mCherry and tyrosine hydroxylase intensities within glomeruli of an H2be-KO heterozygous mouse. Red line, best fit. *p<0.05; **p<0.01; ****p<0.0001. Mouse ages: (A) and (C), 4 weeks; (D), 8 weeks; (G), 10 weeks. Scale bars for (A) and (D), 20 µm; (C) and (G), 100 µm.

https://doi.org/10.7554/eLife.00070.012
Figure 8—figure supplement 1

Distribution of relative H2BE levels within neurons expressing Olfr73, Olfr958, Olfr16, and Olfr167 for odor-exposed and control littermates. Stimulating odors are indicated, except for Olfr167, which serves as a negative control.

https://doi.org/10.7554/eLife.00070.013

H2BE levels are cAMP-, but not Ca2+-dependent. (A) Representative example of the effects of Adcy3 loss-of-function on H2BE levels in newborn (P0) mice. In Adcy3(+/+) mice, Olfr1508+ neurons contain extremely low H2BE levels (left, arrowheads), while in Adcy3(−/−) mice, Olfr1508+ neurons frequently contain extremely high levels (right, arrowhead), indicating that cAMP participates in the negative regulation of H2BE. Note: age P0 was chosen due to the low postnatal survival rate of Adcy3(−/−) mice. (B) Effects of Cnga2 loss-of-function on H2be expression. Mice in which Cnga2 (an X-chromosomal gene necessary for odor-evoked Ca2+ signaling) was replaced with the Tau-LacZ gene (Zhao and Reed, 2001) were crossed with H2be-KO mice to generate three-week old Cnga2-KO-Tau-LacZ(+/−)/H2be-KO-Gap43-mCherry(+/−) compound heterozygous females. In these mice, one half of new olfactory neurons express TAU-LACZ instead of CNGA2 and project to glomeruli distinct from neurons expressing Cnga2 (Zheng et al., 2000). Analysis of β-GAL+ (TAU-LACZ) and GAP43-mCherry intensities within glomeruli revealed that Cnga2 neurons (β-GAL+; arrowheads) do not have higher levels of GAP43-mCherry than Cnga2+ neurons (β-GAL), indicating that H2be expression is not negatively regulated by Ca2+ signaling. Scale bars for (A), 40 µm; (B), 100 µm.

https://doi.org/10.7554/eLife.00070.014

Unilateral naris occlusion (UNO) alters gene expression and OR expression frequencies. (A) Gene ontology (biological process) terms enriched at the top of a gene list ranked descendingly according to differential expression on the two MOE halves from 5-week old WT mice subjected to UNO (21 days), based on microarray analysis (n = 3 samples per MOE side, four animals per sample). (B) Percentage of OR genes with significantly differential (FDR-adjusted p<0.05) expression on the two sides of the MOE of WT mice after UNO (21 days; values from microarray data). (C,D) Representative images of Olfr1325 (C) and Olfr1336 (D) expression in the MOE after UNO. (E) Representative images of normal FLAG-H2BE levels in neurons associated with ORs that are down- (left) or up-regulated (right) in frequency after olfactory deprivation. (F) Relationship between UNO-mediated OR gene expression differences on the two sides of the MOE (values from microarray data for WT mice subjected to UNO for 21 days) and associated FLAG-H2BE levels measured in intact mice. Red line, best fit; ****p<0.0001. Mouse ages: (C) and (D), 5 weeks; (E), 12 weeks. Scale bars for (C) and (D), 200 µm; (E), 20 µm.

https://doi.org/10.7554/eLife.00070.015
Figure 10—source data 1

Effects of H2be loss of function on gene expression changes in the main olfactory epithelium (MOE) as a result of activity deprivation through unilateral naris occlusion (UNO) in 5-week old mice.

https://doi.org/10.7554/eLife.00070.016

H2be affects activity-dependent gene expression. (A) Gene ontology (biological process) terms enriched among genes with UNO-mediated expression differences in WT mice (log2 fold-change > 0.3; unadjusted p<0.02), but at least 20% less altered expression in H2be-KO compared to WT mice after UNO, based on microarray analysis of MOE halves from 5-week old WT and H2be-KO mice subjected to UNO (21 days; n = 3 samples per MOE side, four animals per sample). (B) Histograms of UNO-mediated OR expression differences on the closed and open sides of the MOE of H2be-KO mice as a percentage of the corresponding WT differences (normalized to 100% or −100%; red lines) for ORs significantly up- (left) or down-regulated (right) in WT mice after olfactory deprivation (FDR-adjusted p<0.05; values from microarray data). (C) Comparison of UNO-altered Olfr1336 and Olfr1313 frequencies in 5-week old WT and H2be-KO mice subjected to UNO (21 days). Values correspond to relative OR expression frequency differences on the two sides of the MOE according to the anterior (Ant)/ posterior (Post) position (n = 3 mice, five sections per region per mouse). Note: UNO appears to affect OR frequencies differently in the anterior and posterior regions of the MOE. *p<0.05; **p<0.01***p<0.001; ****p<0.0001.

https://doi.org/10.7554/eLife.00070.017

Model for the effects of neuronal activity on H2BE expression level, life span and resulting neuronal representation.

https://doi.org/10.7554/eLife.00070.018

Chromatin incorporation and localization of FLAG-H2BE. (A) Two-color western analysis of FLAG-H2BE and H3 in soluble nucleoplasm (sol) and chromatin (chr) fractions of unfixed MOE cell nuclei from 16-week old Flag-H2be, H2be-GF, and WT mice. (B) SDS-PAGE analysis (left), mass spectrometric identification (listed, middle), and western analysis (right) of proteins associated with immunoprecipitated FLAG-H2BE-containing native mononucleosomes from MOE tissue of Flag-H2be transgenic mice. Proteins identified by mass spectrometry are listed according to their approximate electrophoretic mobility. (C) Quantification of DNA immunoprecipitated from crosslinked and fragmented chromatin derived from MOE tissue of Flag-H2be transgenic mice. (D) Representative image of FLAG-H2BE localization within olfactory neurons of a 10-week old Flag-H2be mouse. Scale bar, 5 µm. (E) Genome-wide ChIP analysis of relative FLAG-H2BE levels with respect to distance from the transcript start sites (TSS) for all mouse genes (grey), and genes expressed at high (red) and low (black) levels in olfactory neurons. (F) Genome-wide ChIP analysis of relative FLAG-H2BE levels with respect to distance from the CDS start sites for mouse histone (left), OR (middle), and vomeronasal type 1 receptor (V1R; right) genes in comparison to all genes. (G) Quantitative PCR analysis of relative FLAG-H2BE levels in the protein-coding regions of representative histone and OR genes. Analyses were performed on ligation-mediated-PCR-amplified Flag-H2BE and H3 ChIP DNA samples. (H) Genome-wide ChIP analysis shows depleted FLAG-H2BE levels within representative OR CDS regions.

https://doi.org/10.7554/eLife.00070.019

H2BE's post-translational modifications (PTMs) differ from those of canonical H2B. (A,B,D,E) Representative images of H2B-Lys5-Me (A, B, D, and E) and FLAG-H2BE (A, B, and E) staining in the MOE of Flag-H2be (A and B), H2be-KO (D) or H2be-GF (E) mice. (B) High-magnification image of FLAG-H2BE and H2B-Lys5-Me colocalization shows that H2BE is depleted in nuclear regions enriched for H2B-Lys5-Me (arrowheads). Mouse ages: (A), (B), and (E), 10 weeks; (D), 34 weeks. (C) Confirmation of reactivity of the anti-H2B-Lys5-Me1 polyclonal antibody with the Lys5-Me PTM in the context of the H2BE protein sequence. Image (top) and quantification (bottom) of an ELISA assay for peptides corresponding to canonical H2B or H2BE and containing the Lys5-Me PTM. (F) Two-color fluorescent western analysis of Lys5-Me modification of FLAG-H2BE in MOE lysates from WT and Flag-H2be (F-H) mice. No detectable H2B-Lys5-Me staining of the FLAG-H2BE bands (red; arrowheads) is observed. Approximate molecular weights (kDa) are indicated (left). The bands observed at approximately 30–35 kDa likely correspond to histone dimers. (GI) Age-dependence of H2BE accumulation. (G and H) Images (left) and quantification (right) of FLAG-H2BE and H2B-Lys5-Me co-localization in 3- (G) and 34-week old (H) mice. Red lines, best fits. (I) Quantification of H2B-Lys5-Me levels in high-H2BE neurons (relative to apical sustentacular cells; n = 20 nuclei from two images per timepoint). (J) Two-color fluorescent western analysis (left) and quantification (right) of FLAG-H2BE relative to total H2B as a function of age in MOE lysates from WT and Flag-H2be (F-H) mice. Approximate molecular weights (kDa) are indicated (left). The bands observed at approximately 30–35 kDa likely correspond to histone dimers. **p<0.01; ****p<0.0001; n.s., not significant. Scale bar for (A), (D), (E), (G), and (H), 20 µm; (B), 5 µm.

https://doi.org/10.7554/eLife.00070.020

Additional files

Supplementary file 1

(A) Effects of H2be loss-of-function on gene expression, based on microarray analyses of whole MOE tissue from 6-month old H2be-KO and WT mice (n = 6 samples per genotype, two animals per sample). Top 100 genes that are up- (left) or down-regulated (right) in H2be-KO mice compared to WT controls are listed according to their unadjusted p-value rank. OR genes are highlighted. (B) Effects of ectopic over-expression of H2be on gene expression, based on microarray analyses of whole MOE tissue from 5-week old H2be-GF (n = 4 samples, three animals per sample) and WT (n = 6 samples, two animals per sample) mice. Top 100 genes that are up- (left) or down- (right) regulated in H2be-GF mice compared to WT controls are listed according to their unadjusted p-value rank. OR genes are highlighted.

https://doi.org/10.7554/eLife.00070.021

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Stephen W Santoro
  2. Catherine Dulac
(2012)
The activity-dependent histone variant H2BE modulates the life span of olfactory neurons
eLife 1:e00070.
https://doi.org/10.7554/eLife.00070