Splice-dependent trans-synaptic PTPδ-IL1RAPL1 interaction regulates synapse formation and non-REM sleep

Abstract

Alternative splicing regulates trans-synaptic adhesions and synapse development, but supporting in vivo evidence is limited. PTPδ, a receptor tyrosine phosphatase adhering to multiple synaptic adhesion molecules, is associated with various neuropsychiatric disorders; however, its in vivo functions remain unclear. Here, we show that PTPδ is mainly present at excitatory presynaptic sites by endogenous PTPδ tagging. Global PTPδ deletion in mice leads to input-specific decreases in excitatory synapse development and strength. This involves tyrosine dephosphorylation and synaptic loss of IL1RAPL1, a postsynaptic partner of PTPδ requiring the PTPδ-meA splice insert for binding. Importantly, PTPδ-mutant mice lacking the PTPδ-meA insert, and thus lacking the PTPδ interaction with IL1RAPL1 but not other postsynaptic partners, recapitulate biochemical and synaptic phenotypes of global PTPδ-mutant mice. Behaviorally, both global and meA-specific PTPδ-mutant mice display abnormal sleep behavior and non-REM rhythms. Therefore, alternative splicing in PTPδ regulates excitatory synapse development and sleep by modulating a specific trans-synaptic adhesion.

Keywords: alternative splicing; receptor tyrosine phosphatase; sleep behavior and rhythm; synapse development; synaptic adhesion.

Figure 1. PTPδ‐tdTomato protein mainly localizes to presynaptic sites at excitatory, but not inhibitory, synapses in the mouse hippocampus

1. A

   Transgenic strategy for generating PTPδ‐tdTomato reporter mice used to visualize the distribution patterns of endogenous PTPδ protein. CRISPR/Cas9 was used to fuse tdTomato in frame to the C‐terminus of the PTPδ protein, encoded by the last coding exon (exon 21) of the Ptprd gene. HA, homology arm; CDS, coding domain sequence; UTR, untranslated region; Ig, Immunoglobulin domain; FnIII, fibronectin 3‐like domain; D, phosphatase domain; td, tdTomato.

2. B

   Verification of PTPδ‐tdTomato reporter mice by Western blot analysis of whole‐brain lysates from heterozygote (HT) and homozygous (Ho) mice (P21) and PTPδ using mCherry antibodies. Note that levels of endogenous PTPδ protein (˜85 kDa) are decreased, and that PTPδ‐tdTomato protein (˜140 kDa) is detectable only in reporter mice (HT and Ho). Note also that the 85‐kDa band represents the transmembrane domain + cytoplasmic fragment of the full‐length PTPδ protein that undergoes proteolytic cleavage at extracellular membrane‐proximal sites, which can only be recognized by the PTPδ C‐terminal antibody that targets the cytoplasmic region. The PTPδ‐tdTomato protein (˜140 kDa) contains the membrane domain + cytoplasmic fragment of PTPδ (˜85 kDa) fused to tdTomato (55 kDa). The N‐terminal antibody recognizes equal amounts of the C‐terminal cleavage product derived from PTPδ and PTPδ‐tdTomato proteins.

3. C

   Representative coronal, horizontal, and sagittal sections from PTPδ‐tdTomato mice. Enlarged windows are demarcated by black dotted lines (Fig EV1A–C). Red tdTomato images were converted to grayscale images for clarity. ACA, anterior cingulate area; ACB, nucleus accumbens; BLA, basolateral amygdala; BMA, basomedial amygdala; CA1, cornu ammonis area 1; CA2, cornu ammonis area 2; CA3, cornu ammonis area 3; CB, cerebellum; CC, corpus callosum; CEA, central amygdala; CP, caudate putamen; CST, corticospinal tract; DG, dentate gyrus; EP, endopiriform nucleus, HY, hypothalamus; L1, cortical layer 1; L2/3, cortical layer 2/3; LA, lateral amygdala; lENT, lateral entorhinal area; LH, lateral habenula; MB, midbrain; MEA, medial amygdala; mENT, medial entorhinal area; MH, medial habenula; PIR1, piriform molecular layer; PIR2, piriform pyramidal layer; PIR3, piriform polymorph layer; PL, prelimbic; PONS, pons; PRE, presubiculum; RSP, retrosplenial area; SS, somatosensory area; SUB, subiculum; TEa, temporal association area; TH, thalamus; VIS, visual area; Scale bar, 500 □m.

4. D

   PTPδ‐tdTomato protein is detected in tau‐positive axonal compartments, but not in MAP2‐positive dendritic compartments in cultured neurons (entorhinal cortex + hippocampus; 1:2 mixture; days in vitro or DIV 24) derived from Ptprd ^−/− mouse embryos. Note that the tdTomato signals envelop, but not overlap with, MAP2‐delineated dendrites, although this does not necessarily suggest that tdTomato signals not in dendrites. Scale bars, 10 μm in main images and 2 μm in enlarged images.

5. E

   Presynaptic PTPδ‐tdTomato clusters colocalize more strongly with PSD‐95 clusters (excitatory postsynaptic marker) relative to gephyrin (inhibitory postsynaptic marker) clusters in cultured neurons (entorhinal cortex + hippocampus; DIV24) derived from Ptprd ^−/− mouse embryos. See Fig EV1F for the quantification of the results. Scale bars, 10 μm in main images and 2 μm in enlarged images.

6. F–I

   Ultrastructural localization of PTPδ‐tdTomato protein (red arrows; DAB staining) at axon terminals juxtaposed to electron‐dense postsynaptic densities (purple arrows; PSDs) and also at vGlut2‐positive excitatory synaptic axon terminals (green arrows; immunogold staining) but not at GAD67/65‐positive inhibitory synaptic axon terminals (blue arrows; immunogold staining) in the stratum lacunosum‐moleculare (SLM) region of the CA1 hippocampal region (P21). Note that ˜85% of PTPδ‐tdTomato signals are present in excitatory presynaptic terminals apposed to PSDs, ˜10% at presynaptic terminals not apposed to PSDs, and ˜5% in neuropils other than presynaptic terminals. Axon terminals are indicated by pink shades. Scale bar, 500 nm (n = 4 areas from 2 mice [WT and KO]).

Source data are available online for this figure.

Figure EV1. Distribution patterns of PTPδ‐tdTomato protein in brain regions and at subcellular and ultrastructural sites in PTPδ‐tdTomato reporter mice

1. A–C

   Distribution patterns of PTPδ‐tdTomato protein revealed by confocal microscopic imaging of coronal, horizontal, and sagittal sections of brains from PTPδ‐tdTomato reporter mice (P21). The imaging used the PTPδ‐tdTomato signals from unstained brain slices. Red tdTomato signals were converted to grayscale for enhanced visibility. Dotted boxes indicate location of corresponding panels of Fig 1C. Scale bar, 2 mm.

2. D

   PTPδ‐tdTomato signals are mainly detected in axonal compartments of mature (DIV 24) cultured hippocampal and cortical (entorhinal cortex; 2:1 mixture) glutamatergic and GABAergic neurons, marked by vGluT1 and GAD67, respectively. Scale bar, 10 μm in main images and 2 μm in enlarged images.

3. E

   PTPδ‐tdTomato signals are detected in both tau‐positive axons and MAP2‐positive dendrites in immature (DIV 6) cultured hippocampal and cortical (entorhinal cortex; 2:1 mixture) neurons. Scale bar, 50 μm in main images and 10 μm in enlarged images.

4. F

   Quantification of the results in Fig 1E for colocalization of PTPδ‐tdTomato with PSD‐95 and gephyrin clusters. For quantification, the area, average intensity, or total intensity of tdTomato signals in areas of tdTomato clusters overlapping with the areas of PSD‐95/gephyrin clusters was normalized to those (area/average intensity/total intensity) of total tdTomato clusters to obtain the PTPδ area/intensity/total ratio (y‐axis) of WT and KI at excitatory/inhibitory synapses (n = 15 images [WT and KO], mean ± SEM, ***P < 0.001, ns, not significant, Mann‐Whitney U test).

5. G–I

   Electron micrographs showing the ultrastructural distribution patterns of PTPδ‐tdTomato DAB signals in the SLM region of the hippocampus (P21) juxtaposed to PSD structures and colocalized with vGlut2 immunogold particles (excitatory axon terminals) but not with GAD67/65 immunogold particles (inhibitory axon terminals). Inset regions (red squares) are enlarged in main Fig 1F–H. Scale bar, 500 nm.

6. J

   PTPδ and PTPδ‐tdTomato proteins similarly distribute to subcellular brain fractions, as shown by immunoblot analysis of subcellular fractions from the whole brain of PTPδ‐tdTomato reporter mice (P21). Analysis of immunoblots reveals no difference in the subcellular localization of PTPδ‐tdTomato fusion proteins as compared to wild‐type PTPδ. H, Total homogenate; P1, nuclei and large debris; S2, supernatant after P2 precipitation; P2, crude synaptosomes; S3, cytosol; P3, small membrane; LP1, synaptosomal membrane; LS2, synaptic cytosol; LP2, synaptic vesicles (n = 3 mice for each fraction, mean ± SEM, ns, not significant, two‐way RM ANOVA with Holm‐Sidak test).

Source data are available online for this figure.

Figure 2. Ptprd deletion leads to input‐specific suppression of excitatory synapse density and strength in the hippocampal SLM layer

1. A

   Strong PTPδ‐tdTomato (red) signals in the SLM layer of the CA1 region and the MO layer of the DG in the hippocampus (2 months), revealed by quantitative analysis of PTPδ‐tdTomato fluorescence across hippocampal lamina. DAPI was used for nuclear staining. CC, corpus callosum; SO, stratum oriens; SP, stratum pyramidale; SR, stratum radiatum; SLM, stratum lacunosum moleculare; MO, molecular layer; SG, granule cell layer; PO, polymorph layer. Scale bar 500 μm.

2. B

   Biochemical enrichment of PTPδ protein in the hippocampal SLM layer relative to the SR layer. Micro‐dissected SLM and SR lysates (2 months) were immunoblotted for PTPδ (C‐terminal antibodies) and α‐tubulin (control). The asterisk indicates a non‐specific band. Analysis of immunoblots reveals significant enrichment of PTPδ in the SLM compared to the SR (n = 4 mice for SR and SLM, mean ± SEM, ***P < 0.001, Student's t‐test; see Dataset EV2 for statistical details for these and all following results).

3. C

   Schematic diagram for global Ptprd KO in mice, targeting exon 3. HTNC (His‐tagged, TAT‐fusion Cre with nuclear localization signal), carrying a membrane‐permeable Cre recombinase, was used to remove exon 13 in fertilized eggs (see Methods for details). Sites on the PTPδ protein corresponding to epitopes recognized by anti‐PTPδ antibodies (Ig1 domain [exon 13]) and N‐ and C‐termini) are indicated.

4. D

   Gene dose‐dependent reductions in PTPδ protein levels in whole brains of Ptprd ^+/− (HT) and Ptprd ^−/− (KO) mice relative to Ptprd ^+/+ (WT) mice (P21), shown by immunoblot analyses. Note that the C‐terminal antibody also detects a non‐specific band of similar size indicated by an asterisk, possibly either PTP( or LAR.

5. E–J

   (E, G, and I) Suppressed basal excitatory synaptic transmission in the Ptprd ^−/− hippocampal SLM layer, but not SO or SR layer (2 months; male), as shown by the I/O ratio (n = 14 slices from 5 mice [WT‐SO], 14, 5 [KO‐SO], 11, 4 [WT‐SR], 12, 5 [KO‐SR], 14, 6 [WT‐SLM], 12, 6 [KO‐SLM], mean ± SEM, *P < 0.05; **P < 0.01, two‐way repeated‐measures/RM ANOVA, Holm‐Sidak post‐hoc multiple comparisons test). (F, H, and J) Normal presynaptic release in the Ptprd ^−/− hippocampal SLM, SO, and SR layers (P22–27; male), as shown by the PPF ratio (n = 13 slices from 5 mice [WT‐SO], 13, 5 [KO‐SO], 10,4 [WT‐SR], 12, 5 [KO‐SR], 15, 6 [WT‐SLM], 14, 6 [KO‐SLM], mean ± SEM, ns, not significant, two‐way RM ANOVA).

6. K–O

   Decreased PSD density, but not PSD length, thickness, or perforation (%), in the Ptprd ^−/− SLM (P21; male), revealed by EM. Arrows in panel K indicate PSDs (n = 18 images from 3 mice [WT and KO], ***P < 0.001, ns, not significant, Student's t‐test for L, O‐R, Mann‐Whitney U test for M, N). Scale bar, 500 nm.

7. P–R

   Decreased density of presynaptic nerve terminals but normal area of nerve terminals and normal density of presynaptic vesicles in the Ptprd ^−/− SLM (P21; male), revealed by EM (n = 3 mice [WT and KO], mean ± SEM, ***P < 0.001, ns, not significant, Student's t‐test).

Source data are available online for this figure.

Figure EV2. Normal levels of spontaneous synaptic transmission in Ptprd ^−/− CA1 neurons and GABAergic synapse density and structure in the Ptprd ^−/− SLM

1. A

   Schematic diagrams for the recording of spontaneous synaptic transmission in CA1 pyramidal neurons.

2. B

   Normal mEPSC frequency and amplitude in Ptprd ^−/− CA1 pyramidal neurons (P21–24). Analysis of frequency and amplitude of mEPSC peaks reveal no statistical difference between WT and Ptprd ^−/− mice (n = 17 neurons from 3 mice [WT], 16, 3 [KO], mean ± SEM, ns, not significant, Student's t‐test).

3. C

   Normal mIPSCs in CA1 pyramidal neurons of Ptprd ^−/− mice (P22–24). Analysis of frequency and amplitude of mIPSC peaks reveal no statistical difference between WT and Ptprd ^−/− mice (n = 16 neurons from 3 mice [WT], 13, 3 [KO], mean ± SEM, ns, not significant, Student's t‐test).

4. D

   Normal sEPSCs in CA1 pyramidal neurons of Ptprd ^−/− mice (P21–25). Analysis of frequency and amplitude of sEPSC peaks reveal no statistical difference between WT and Ptprd ^−/− mice (n = 16 neurons from 4 mice [WT], 17, 4 [KO], mean ± SEM, ns, not significant, Mann‐Whitney U test).

5. E–H

   Normal density, length, and thickness of GABAergic synapses in the Ptprd ^−/− SLM (P21), as revealed by EM measurement of PSDs (pink arrowheads) apposed to GABA‐positive axon terminals, marked by immunogold staining for GABA (red arrows) (n = 36 images from 3 mice [WT], 36, 3 [KO], mean ± SEM, ns, not significant, Student's t‐test, Mann‐Whitney U test). Scale bar, 500 nm.

Figure EV3. Largely normal total levels of synaptic proteins in hippocampal Ptprd ^−/− SLM and SR regions and unaltered synaptic enrichment of Slitrk3, SALM3, and NGL‐3 in the Ptprd ^−/− brain

1. A

   Schematic diagram for the dissection of SR and SLM layers in the hippocampus (P21–27).

2. B–F

   Representative immunoblots of total lysates from the SR and SLM layers for the tested synaptic proteins, including proteins known to be enriched in the SLM (NGL‐1 and HCN1), the PTPδ relative PTPσ, postsynaptic partners of PTPδ (IL1RAPL1, Slitrk2/3, and NGL‐3), presynaptic scaffolds/adaptors (Bassoon and liprin‐α), postsynaptic scaffolds/adaptors (CaMKIIα/β, PSD‐95, SynGAP1), postsynaptic receptors (GluA1/2, GluN1, GluN2A/B), and signaling molecules (phospho‐Src). α‐Tubulin was used as a control (n = 4 mice for WT and KO, mean ± SEM, ***P < 0.001, ns, not significant, two‐way ANOVA, Tukey's HSD post‐hoc test).

3. G

   Normal synaptic levels of other PTPδ‐binding partners (Slitrk3, SALM3, and NGL‐3) in the Ptprd ^−/− brain (P21–27), as revealed by immunoblotting of crude synaptosomal (P2), synaptic plasma membrane (SPM), and PSD (PSD II) fractions (n = 3 mice [WT and KO] for each fraction [P2, SPM, and PSD], mean ± SEM, ns, not significant, Student's t‐test).

4. H

   Normal synaptic levels of other PTPδ‐binding partners (Slitrk3, SALM3, and NGL‐3) in the Ptprd‐meA ^−/− brain (P21–27), as revealed by immunoblotting of P2, SPM, and PSD (PSD II) fractions (n = 3 mice [Ptprd‐meA ^+/+ and Ptprd‐meA ^−/−] for each fraction [P2, SPM, and PSD], mean ± SEM, ns, not significant, Student's t‐test).

Source data are available online for this figure.

Figure 3. Altered pTyr levels in multiple pre‐ and postsynaptic proteins in the Ptprd ^−/− brain

1. A

   A volcano plot of proteins that showed significant up‐ or downregulation of pTyr levels (absolute fold change > 1.5, P < 0.05) in the whole brain of Ptprd ^−/− mice (P21). Note that IL1RAPL1 is the protein with the strongest decrease in pTyr levels (n = 3 for WT and 3 for KO). See Dataset EV1, “PhosphoScan Results” tab for a full list of significantly changed proteins.

2. B

   SynGO analysis (https://www.syngoportal.org/) of the 167 proteins that show significant (P < 0.05; fold changes were not considered for maximal inclusion of the proteins) up‐ or downregulation of pTyr levels in Ptprd ^−/− mice (P21) for their synaptic functions. See Dataset EV1, “SynGO Ontologies” and “SynGO Annotation” tabs for a full list of SynGO annotations and genes.

3. C

   SynGO analysis of the proteins with significant (P < 0.05) up‐ or downregulation of pTyr levels at their pre‐ and postsynaptic localizations.

4. D

   A volcano plot of presynaptic proteins included in the SynGO list to show that they are significantly changed in pTyr levels (P < 0.05).

5. E

   A volcano plot of postsynaptic proteins included in the SynGO list to show that they are significantly changed in pTyr levels (P < 0.05).

Source data are available online for this figure.

Figure 4. PTPδ trans‐synaptically regulates tyrosine phosphorylation and synaptic localization of IL1RAPL1 and excitatory synaptic strength

1. A

   Decreased pTyr levels, but normal total levels of IL1RAPL1 protein, in the Ptprd ^−/− brain (P21–27), as revealed by immunoprecipitation of IL1RAPL1 from whole‐brain lysates followed by immunoblotting for total IL1RAPL1 protein and pTyr levels (4G10 antibody).

2. B

   Decreased levels of synaptic IL1RAPL1 protein in the Ptprd ^−/− brain (P21–27), as revealed by immunoblotting of crude synaptosomal (P2), synaptic plasma membrane (SPM), and PSD (PSD II) fractions. Analysis of immunoblots reveals significant decrease of IL1RAPL1 in the SPM and PSD fractions of Ptprd ^−/− brain samples (n = 3 mice [WT and KO] for each fraction [P2, SPM, and PSD], mean ± SEM, **P < 0.01, Student's t‐test).

3. C

   Schematic diagram showing that the PTPδ‐meA splice insert is important for the interaction between PTPδ and IL1RAPL1. Note that the meA splice insert in the Ig2 domain of PTPδ interacts with the Ig1 domain of IL1RAPL1.

4. D

   Schematic diagram showing the strategy for generating mice carrying a deletion of the PTPδ‐meA splice insert encoded by exons 15 and 16 (Ptprd‐meA ^−/−).

5. E

   Normal levels of the PTPδ protein in whole‐brain lysates of Ptprd‐meA ^−/− mice (P21–27), which contrasts with the Ptprd ^−/− brain, where PTPδ protein is undetectable.

6. F

   Strongly decreased levels of tyrosine‐phosphorylated PTPδ protein in whole‐brain lysates of Ptprd‐meA ^−/− mice (P21–27), as shown by the immunoprecipitation of IL1RAPL1 protein from whole‐brain lysates followed by immunoblotting for total and tyrosine‐phosphorylated PTPδ protein.

7. G

   Decreased levels of synaptic IL1RAPL1 protein in the Ptprd‐meA ^−/− brain (P21–27), as revealed by immunoblotting P2, SPM, and PSD (PSD II) fractions. Analysis of immunoblots reveals significant decrease of IL1RAPL1 in the SPM and PSD fractions of Ptprd‐meA ^−/− brain samples (n = 3 mice [Ptprd‐meA ^+/+ and Ptprd‐meA ^−/−] and [P2, SPM, and PSD], mean ± SEM, *P < 0.05, ***P < 0.001, Student's t‐test). Note that, while unchanged in the P2 and SPM fractions, PTPδ is significantly decreased in the PSD of Ptprd‐meA ^−/− samples.

8. H

   Decreased basal excitatory synaptic transmission in the hippocampal SLM region of Ptprd‐meA ^−/− mice, as shown by the I/O ratio at the TA‐SLM pathway (P20–24) (n = 15 slice from 5 mice [Ptprd‐meA ^+/+], 12, 4 [Ptprd‐meA ^−/−], mean ± SEM, *P < 0.05; ***P < 0.001, two‐way RM ANOVA with Holm‐Sidak test).

9. I

   Normal presynaptic release in the hippocampal SLM region of Ptprd‐meA ^−/− mice, as shown by the PPF ratio at the TA‐SLM pathway (P20–24) (n = 15 slice from 5 mice [Ptprd‐meA ^+/+] and 13, 4 [Ptprd‐meA ^−/−], mean ± SEM, ns, not significant, two‐way RM ANOVA).

Source data are available online for this figure.

Figure 5. Shared hyperactivity and sleep disturbance phenotypes in Ptprd ^−/−, Emx1‐Cre;Ptprd ^fl/fl and Ptprd‐meA ^−/− mice

1. A

   Schematic diagram showing how mouse movements are measured in Laboras cages for four consecutive days and analyzed. Data from the first 24 h, when the mouse is not fully familiarized with the environment, were not included in the analysis. Each 24 h was sub‐divided into 3‐h time bins and variables averaged across matching time bins into a representative 24‐h period.

2. B, C

   Hyperactivity of Ptprd ^−/− mice (2.5–4 months) in Laboras cages observed during light‐off and light‐on periods was strongly mimicked by Emx1‐cKO mice, but more weakly by Viaat‐cKO mice and not at all by Camk2a‐cKO mice. Note also that immobility, a measure that better reflects sleep behavior, showed similar changes among different mouse lines (n = 14 mice [WT], 14 [global KO], 16 [Emx1‐cWT], 16 [Emx1‐cKO], 11 [Viaat‐cWT], 13 [Viaat‐cKO], 18 [Camk2a‐cWT] and 13 [Camk2a‐cKO], mean ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001; ns, not significant, two‐way RM ANOVA with Holm‐Sidak test).

3. D, E

   Hyperactive behavior and decreased immobility of Ptprd‐meA ^−/− mice (2.5–4 months) in Laboras cages observed during light‐off periods and light‐on periods, similar to Ptprd ^−/− mice (n = 7 mice for Ptprd‐meA ^+/+ and 7 mice for Ptprd‐meA ^−/−, mean ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001, ns, not significant, two‐way RM ANOVA with Holm‐Sidak test).

Figure EV4. Hyperactivity and anxiety‐like behaviors of Ptprd ^−/−, Emx1‐cKO, Viaat‐cKO, CaMKII‐cKO, and Ptprd‐meA ^−/− mice in open‐field, elevated plus‐maze, and light–dark tests

1. A–D

   Hyperactivity of global Ptprd ^−/−, Emx1‐cKO, Viaat‐cKO, Camk2a‐cKO, and Ptprd‐meA ^−/− mice (2.5–4 months) in the open‐field test, indicated by distance moved and total distance moved. Note that the hyperactivity of global Ptprd ^−/− mice is mimicked by Emx1‐cKO mice, but not by Viaat‐cKO, Camk2a‐cKO, or Ptprd‐meA ^−/− mice. Note also that Emx1‐cKO, but not other mice, showed enhanced anxiety‐like behaviors, as indicated by the time spent in the center region of the open‐field arena (n = 13 mice [WT], 12 [global KO], 24 [Emx1‐cWT], 19 [Emx1‐cKO], 18 [Viaat‐cWT], 20 [Viaat‐cKO], 7 [Camk2a‐cWT] and 9 [Camk2a‐cKO], 8 [Ptprd‐meA ^+/+] and 8 [Ptprd‐meA ^−/−], mean ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001, ns, not significant, two‐way RM ANOVA with Holm‐Sidak test for A and B, Student's t‐test for all pairs in B and D expect for Mann‐Whiney U test for the Emx1 pair).

2. E, F

   Anxiety‐like behaviors in Ptprd ^−/−, Emx1‐cKO, Viaat‐cKO, Camk2a‐cKO, and Ptprd‐meA ^−/− mice (2.5–4 months) in the elevated plus‐maze test, as indicated by time spent in closed/open arms and the center. Note that Ptprd ^−/−, Emx1‐cKO, and Viaat‐cKO mice, but not Camk2a‐cKO or Ptprd‐meA ^−/−, mice show slight anxiolytic‐like behavior. Note also that Emx1‐cKO and Ptprd‐meA ^−/− mice show hyperactivity in the EPM (F) (n = 13 mice [WT], 12 [global KO], 24 [Emx1‐cWT], 19 [Emx1‐cKO], 18 [Viaat‐cWT], 20 [Viaat‐cKO], 7 [Camk2a‐cWT] and 9 [Camk2a‐cKO], 20 [Ptprd‐meA ^+/+], and 19 [Ptprd‐meA ^−/−], mean ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001, ns, not significant, two‐way RM ANOVA with Holm‐Sidak test for E, Student's t‐test for all pairs in F except for Mann‐Whitney U test for the Emx1 pair).

3. G

   Anxiety‐like behaviors in Ptprd ^−/−, Emx1‐cKO, Viaat‐cKO, Camk2a‐cKO, and Ptprd‐meA ^−/− mice (2.5–4 months) in the light–dark test, as indicated by time spent in light chamber. Note that Ptprd ^−/− and Emx1‐cKO mice, but not Viaat‐cKO, Camk2a‐cKO, or Ptprd‐meA ^−/− mice, show enhanced anxiety‐like behavior (n = 16 mice [WT], 20 [global KO], 28 [Emx1‐cWT], 31 [Emx1‐cKO], 18 [Viaat‐cWT], 23 [Viaat‐cKO], 16 [Camk2a‐cWT] and 16 [Camk2a‐cKO], 20 [Ptprd‐meA ^+/+], and 20 [Ptprd‐meA ^−/−], mean ± SEM, ***P < 0.001, ns, not significant, Student's t‐test for Viaat, Camk2a, and Ptprd‐meA pairs, Mann‐Whitney U test for Ptprd ^−/− and Emx1 pairs).

Figure 6. Shared decreased NREM sleep and delta power phenotypes in Emx1‐Cre;Ptprd ^fl/fl and Ptprd‐meA ^−/− mice

1. A

   Schematic diagram showing the sites of electroencephalography (EEG) and electromyography (EMG) surgery (frontal [Ch1], parietal [Ch2], EEG reference [Ref_E], trapezius muscle [EMG], and EMG reference [Ref_M] probes), and examples of EEG and EMG traces during WAKE, NREM, and REM states. Gr, animal ground.

2. B

   Increased WAKE duration and decreased NREM duration during the first 3 h of the light‐on period in Emx1‐Cre;Ptprd ^fl/fl mice (2.5–4 months) (n = 9 mice [Emx1‐cWT] and 9 mice [Emx1‐cKO], mean ± SEM, *P < 0.05, **P < 0.01, ns, not significant, two‐way RM ANOVA with Holm‐Sidak test).

3. C

   Decreased delta power (normalized to total power) in NREM, but not WAKE, oscillations in Emx1‐Cre;Ptprd ^fl/fl mice (2.5–4 months). Note also the normal theta power (normalized to total power) in REM oscillations in Emx1‐Cre;Ptprd ^fl/fl mice (n = 9 [Emx1‐cWT] and 9 [Emx1‐cKO], mean ± SEM, *P < 0.05, **P < 0.01, ns, not significant, two‐way RM ANOVA with Holm‐Sidak test).

4. D

   Increased WAKE duration and decreased NREM duration during the first 3 h of the light‐on period and decreased REM duration in the first half of the light‐on period in Ptprd‐meA ^−/− mice (2.5–4 months). Note also that Ptprd‐meA ^−/− mice also display increased WAKE duration and decreased NREM duration during the light‐off period (n = 8 [Ptprd‐meA ^+/+] and 8 [Ptprd‐meA ^−/−], mean ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001, ns, not significant, two‐way RM ANOVA with Holm‐Sidak test).

5. E

   Decreased delta power (normalized to total power) in NREM (both light‐off and light‐on periods) and WAKE (light‐on period) oscillations, and decreased theta power in REM (mainly light‐on period) oscillations in Ptprd‐meA ^−/− mice (2.5–4 months) (n = 8 [Ptprd‐meA ^+/+] and 8 [Ptprd‐meA ^−/−], mean ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001, ns, not significant, two‐way RM ANOVA with Holm‐Sidak test).

Figure EV5. Raw EEG data from Emx1‐Cre;Ptprd ^fl/fl and Ptprd‐meA ^−/− mice

1. A, B

   Normalized power spectrogram (0–20 Hz) of each time‐bin (rows) and brain states in Emx1‐cKO mice (A) and Ptprd‐meA ^−/− mice (B) (2.5–4 months). Gray backgrounds in A and B panels denote light‐off hours (n = 9 mice [Emx1‐cWT], 9 [Emx1‐cKO], 8 [Ptprd‐meA ^+/+], and 8 [Ptprd‐meA ^−/−]).