Receptor protein tyrosine phosphatase delta is not essential for synapse maintenance or transmission at hippocampal synapses

Members of the leukocyte common antigen-related receptor protein tyrosine phosphatase (LAR-RPTP) family, comprising PTPσ, PTPδ and LAR, are key hubs for presynaptic assembly and differentiation in vertebrate neurons. However, roles of individual LAR-RPTP members have not been investigated using member-specific conditional knockout mice. Here, we show that loss of PTPδ had no overt effect on synapse development in mouse cultured hippocampal neurons. Moreover, loss of PTPδ in presynaptic CA1 hippocampal neurons did not influence neurotransmitter release in subicular pyramidal neurons, suggesting that PTPδ is not critical for presynaptic function in vivo. Our results demonstrate that PTPδ is not essential for synapse maintenance or transmission, at least in the mouse hippocampus, and underscore the importance of using sophisticated genetic approaches to confirm the roles of synaptic proteins.


Introduction
Leukocyte common antigen-related receptor tyrosine phosphatases (LAR-RPTPs) are evolutionarily conserved synaptic organizers in vertebrate neurons [1][2][3]. LAR-RPTPs bind various postsynaptic ligands to organize multifarious synaptic adhesion pathways that are required for presynaptic assembly and a subset of postsynaptic functions [4][5][6][7][8][9]. Accumulating evidence from loss-of-function approaches, including constitutive knockout (KO) mice and shorthairpin RNA (shRNA)-mediated knockdown (KD), has lent support to this idea [2,3,10,11]. PTPδ is strongly expressed in brains and constitutive PTPδ KO has been reported to impair learning, alter long-term synaptic plasticity, and cause abnormal dendritic arborization [12,13]. Moreover, previous studies have reported that PTPδ KD produces pleiotropic effects on both presynaptic and postsynaptic processes [4,6,8,14]. Although all three members of the vertebrate LAR-RPTP family (PTPσ, PTPδ and LAR) share similar structural, biochemical and ligand-binding profiles [3], individual LAR-RPTP members exhibit unique features. For example, PTPσ, but not PTPδ or LAR, binds to TrkC receptor tyrosine kinase, whereas PTPδ specifically binds to IL1RAPL1 (interleukin 1 receptor accessory protein like 1). Moreover, KD analyses in cultured hippocampal neurons have revealed dichotomous roles of PTPσ and PTPδ in regulating maintenance of excitatory and inhibitory synapse development, respectively [2,5,15,16]. Although these findings clearly indicate the significance of LAR-RPTPs as important synaptic organizers, a more sophisticated approach using conditional KO (cKO) mice deficient in individual LAR-RPTP members is required to precisely evaluate the synaptic role of vertebrate LAR-RPTPs in vivo. A recent paper showed that all three LAR-RPTPs regulate postsynaptic responses mediated by N-methyl-D-aspartate receptors (NMDA-type glutamate receptors) through trans-synaptic mechanism(s), but this study did not examine the presynaptic role of individual LAR-RPTPs [17].
In the present study, we generated mutant mice carrying cKO alleles of PTPδ, and used a variety of experimental approaches employing both cultured hippocampal neurons and acute hippocampal slice preparations. We found that, in contrast to KD effects [4], conditional genetic deletion of PTPδ did not affect synapse numbers, synaptic transmission, or vesicular organization in presynaptic boutons. Moreover, loss of PTPδ in hippocampal CA1 pyramidal neurons did not affect neurotransmitter release at either excitatory or inhibitory synapses formed on subicular pyramidal neurons. Viewed together, our results suggest that PTPδ is not essential for synapse maintenance, synaptic transmission, or neurotransmitter release at hippocampal synapses.

Animals
The use and care of animals complied with the guidelines and protocols (DGIST-IACUC-17122104-01) for rodent experimentation approved by the Institutional Animal Care and Use Committee of DGIST under standard, temperature-controlled laboratory conditions. PTPδ conditional knockout mice were generated at Biocytogen Co., Ltd. (Beijing, China). Nestin-Cre (003771, Jackson Research Laboratories) mice were the gift of Dr. Albert Chen (DUKE-NUS, Singapore). Mice were kept on a 12:12 light/dark cycle (lights on at 9:00 am), and received water and food ad libitum. Floxed PTPδ (PTPδ f/f ) was generated by flanking exon 12 with loxP sites (See Fig. 1a for details). Nestin-Cre driver line was crossed with PTPδ f/f mice to generate pan-neuronal PTPδ knockout. Mice were maintained in the C57BL/6 N background. All experimental procedures were performed on male mice, using littermate control without Cre expression.
Cultured neurons were infected with lentiviruses at DIV3-4. For immunocytochemistry, cultured neurons were fixed with 4% paraformaldehyde/4% sucrose in PBS for 10 or 30 min at 4°C, and permeabilized with 0.2% Triton X-100 in PBS for 10 or 30 min at 4°C. Neurons were blocked with 3% horse serum/0.1% BSA in PBS for 15 min at room temperature and incubated with primary and secondary antibodies in blocking solution for 70 min at room temperature. The primary antibodies were used in these experiments included anti-VGLUT1 (Synaptic Systems; 1:700), anti-GABA A Rγ2 (Synaptic Systems; 1: 500), anti-GluA1 (1193; 1:200), anti-Gephyrin (Synaptic Systems; 1:100), and anti-pan-Shank (1172; 1:200). Images of randomly selected neurons were acquired using a confocal microscope (LSM800, Carl Zeiss) with a 63 × objective lens; all image settings were kept constant during image acquisition. Z-stack images obtained by confocal microscopy were converted to maximal projections, and puncta size and the density of the indicated presynaptic marker proteins were analyzed in a blinded manner using MetaMorph software (Molecular Devices Corp.). Relative levels of PTPσ, PTPδ, and LAR mRNAs were measured in cultured cortical neurons infected with lentiviruses expressing Cre-recombinase. Data are means ± SEMs (n = 3-4 independent experiments; *p < 0.05; Mann Whitney U-test). c, d, Representative immunoblot image (c) and quantitative analyses (d) of the level of PTPδ protein in cultured cortical neurons infected with lentiviruses expressing Cre-recombinase. Data are means ± SEMs (n = 6 independent experiments; **p < 0.01; Mann Whitney U-test). e, Representative immunoblot analysis of the level of PTPδ protein in brain homogenates from 8-week old control and PTPδ-cKO mice. Levels of PTPδ protein were measured in PTPδ fl/fl mice (Ctrl) or PTPδ fl/fl mice crossed with Nestin-Cre mice (Nestin-PTPδ). Arrows indicate band(s) that show immunoreactivity to PTPδ-specific (JK179) antibodies. f, An image illustrating the body size of littermate control (Ctrl) and Nestin-PTPδ mice at 2 months of age. Nestin-PTPδ mice were significantly smaller than age-and sex-matched Ctrl mice

Production of adeno-associated viruses
HEK293T cells were co-transfected with the indicated AAV vectors, pHelper and AAV1.0 (serotype 2/9) capsids vectors. After 72 h, the transfected HEK293T cells were collected, and resuspended in PBS, and lysed by subjecting them to four freeze-thaw cycles in an ethanol/dry ice bath (7 min each) and a 37°C water bath (5 min). The lysates were centrifuged and the supernatants were mixed with 40% polyethylene glycol and 2.5 M NaCl and centrifuged at 2000×g for 30 min. The cell pellets were resuspended in HEPES buffer (20 mM HEPES, 115 mM NaCl, 1.2 mM CaCl 2 , 1.2 mM MgCl 2 , and 2.4 mM KH 2 PO 4 , pH 8.0) to which was added an equal volume of chloroform. The mixture was centrifuged at 400×g for 5 min and concentrated three times with a Centriprep centrifugal filter (Cat. 4310, Millipore) at 1220×g (20 min each) and an Amicon Ultra centrifugal filter (Cat. UFC500396, Millipore) at 16,000×g for 30 min. AAVs were titered by treating 1 μl of concentrated, filter-sterilized AAVs with 1 μl of DNase I (AMPD1; Sigma) for 30 min at 37°C to eliminate any contaminating plasmid DNA. After treatment with 1 μl of stop solution (50 mM ethylenediaminetetraacetic acid) for 10 min at 65°C, 10 μg of protease K (Cat. P2308; Sigma) was added and the sample was incubated for 1 h at 50°C. Reactions were stopped by heat inactivation at 95°C for 20 min. The final virus titer was quantified by qRT-PCR. Empty AAV vector was used to generate a standard curve for qRT-PCRs targeting GFP sequences.

Electron microscopy
E17 embryonic hippocampi of PTPδ floxed mice were seeded onto 18 mm coverslips at densities of 40,000 cells/well. The neurons were infected with lentiviral vectors expressing ΔCre or Cre at DIV4. At DIV14, cultured neurons were fixed in 2% glutaraldehyde, 0.1 M Na-cacodylate buffer, pH 7.4, for 1 h at room temperature and overnight at 4°C. The cells were postfixed in 0.5% OsO 4 (osmium tetroxide), 0.8% K ferricyanide at room temperature for 60 min. All specimens were stained en bloc with 2% aqueous uranyl acetate for 30 min, dehydrated in a graded ethanol series up to 100%, embedded in Embed 812 resin (Electron Microscopy Science, PA), and polymerized overnight in a 60°C oven. Thin sections (50-60 nm) were cut with a Leica ultramicrotome and post-stained with uranyl acetate and lead citrate. Sample grids were examined using a FEI Tecnai BioTWIN transmission electron microscope running at accelerating voltage of 80 kV. Images were recorded with a Morada CCD camera and iTEM (Olympus) software. This protocol allowed the unambiguous staining of membranes of synaptic vesicles as well as of pre-and post-synaptic compartments, resulting in accurate measurements of the nanoscale organization of the synaptic vesicles within nerve endings. To analyze synapse ultrastructure, the lengths of active zone and PSD, tethered vesicles, the membrane proximal vesicles, and total vesicle numbers were quantified using MetaMorph software (Molecular Devices). The numbers of total vesicles and docked vesicles were counted manually, and the distances from the active zone and the PSD to the vesicle center were measured. Vesicles located below 200 nm were considered membrane-proximal vesicles.

Stereotaxic surgery and virus injections
4-5-week-old mice were anesthetized by intraperitoneal injection of 2% 2,2,2-tribromoethanol (Sigma), dissolved in saline, and secured in a stereotaxic apparatus. Viral solutions were injected using a Nanoliter 2010 Injector (World Precision Instruments), including a NanoFil syringe and 33 gauge needle, at a flow rate of 50 nl/min (injected volume, 500 nl). The coordinates used for stereotaxic injections targeting the ventral hippocampal CA1 were, relative to the bregma, anteroposterior (AP)

Data analysis and statistics
All data are expressed as means ± SEM. All experiments were repeated using at least three independent cultures, and data were statistically evaluated using a Mann-Whitney U test, analysis of variance (ANOVA) followed by Tukey's post hoc test, Kruskal-Wallis test (one-way ANOVA on ranks) followed by Dunn's pairwise post hoc test, or paired two-tailed t-test, as appropriate. Prism7.0 (GraphPad Software) was used for analysis of data and preparation of bar graphs. P-values < 0.05 were considered statistically significant (individual p-values are presented in figure legends).

Availability of data and materials
All data generated or analyzed during this study are included in this published article. Any additional information related to the current study is available from the corresponding author on reasonable request.

Generation of PTPδ-cKO mice
To determine the in vivo synaptic functions of PTPδ, we generated transgenic mice with a deletion of PTPδ by crossing PTPδ f/f mice (with exon 12 flanked by loxP sites) with a Cre recombinase driver line in which Cre is expressed under control of the Nestin promoter (Nestin-Cre) (Fig. 1a). To assess the effects of endogenous PTPδ deletion, we induced a global loss of PTPδ in all neurons by infecting cultured hippocampal neurons from PTPδ-cKO mice at 3-4 days in vitro (DIV) with lentiviruses expressing EGFP-fused nuclear Cre-recombinase, which exhibits high infection efficiency. Infection with a nonfunctional mutant version of Cre-recombinase (ΔCre) was used as a control. Global expression of Cre-recombinase caused a nearly complete and specific loss of PTPδ mRNA expression in PTPδ cKO neurons analyzed at DIV13-14, without affecting PTPσ or LAR mRNA (Fig. 1b). Immunoblotting also confirmed complete elimination of PTPδ protein expression in cultured cortical neurons (Fig. 1c, d) and brain lysates (Fig. 1e) of PTPδ cKO mice. PTPδ cKO mice, in which PTPδ was deleted from the entire brain, were viable and fertile, but showed a modest reduction in body size (Fig. 1f). There were no alterations in expression of various classes of synaptic proteins linked to presynaptic functions (Fig. S1).

Loss of PTPδ does not impair synaptic transmission
To assess whether PTPδ loss affects synaptic transmission, we performed an electrophysiological assessment of cultured hippocampal neurons (Fig. 2c-f). Cultured hippocampal PTPδ floxed neurons were infected with lentiviruses expressing either ΔCre (control) or Cre recombinase at DIV3, and miniature excitatory postsynaptic currents (mEPSCs) and miniature inhibitory postsynaptic currents (mIPSCs) were recorded at DIV14-16 (Fig. 2c-f). In line with results from imaging experiments (Fig. 2a, b), lentivirus-mediated global loss of PTPδ had no effect on baseline synaptic transmission at excitatory or inhibitory synapses (Fig. 2c-f). Again, these findings are different from previous reported results for PTPδ KD showing specific reductions in inhibitory synaptic transmission [4].

Loss of PTPδ does not alter synaptic vesicle tethering or active zone (AZ) architectures
To further probe whether PTPδ loss affects synaptic structures, we imaged presynaptic terminals and postsynaptic densities of chemically fixed, cultured hippocampal neurons from PTPδ floxed mice infected with lentiviruses expressing either ΔCre (control) or Cre recombinase using transmission electron microscopy (TEM), as previously described [21] (Fig. 3). TEM analyses of cultured neurons showed that length of active zone (AZ) and postsynaptic density (PSD) were similar between PTPδ-deficient and control presynaptic terminals (Fig. 3a-c). Moreover, PTPδ-deficient and control presynaptic terminals contained similar numbers of total vesicles, and that the distribution of vesicles in PTPδ-deficient terminals was similar that in control presynaptic terminals (Fig. 3d). In addition, the numbers of membrane-proximal vesicles (defined as vesicles < 100 nm from the presynaptic AZ membrane) in PTPσ-deficient presynaptic terminals were also comparable to those in control presynaptic terminals (Fig. 3e). Moreover, the number of tethered vesicles, defined as vesicular structures docked at presynaptic AZ membranes, was also similar (Fig. 3f). Overall, these results suggest that PTPδ is not crucial for maintaining the structural organization of either presynaptic AZs or PSDs in hippocampal neurons. Lastly, we sought to address the role of PTPδ in regulating presynaptic neurotransmitter release in acute hippocampal slice preparation. In these experiments, we targeted hippocampal CA1-to-subicular synapses for analyses because of its amenability for presynaptic manipulations by stereotactic injection of viruses into the hippocampal CA1 region [22,23]. We stereotactically injected adenoassociated viruses (AAVs) expressing either ΔCre (Control) or Cre recombinase into hippocampal CA1 region of PTPδ f/f mice. PTPδ KO in vivo was validated by quantitative immunoblot analyses using hippocampal CA1 tissue lysates of AAV-injected mice (Fig. 4a, b). We then recorded the paired-pulse ratios (PPRs) of excitatory (EPSC-PPRs) or inhibitory (IPSC-PPRs) postsynaptic currents in subicular pyramidal neurons. Again, neither EPSC-PPRs nor IPSC-PPRs were altered in subicular neurons from Cre-infected PTPδ f/f mice (Fig. 4c-f), suggesting that PTPδ is not essential for neurotransmitter release at excitatory or inhibitory synapses in the hippocampus in vivo.

Discussion
There is extensive literature support for the idea that PTPδ is important for synapse development [2,24]. The current study using PTPδ floxed mice infected with Cre recombinase, in conjunction with imaging, electrophysiology and electron microscopy, clearly demonstrated that conditional deletion of PTPδ does not significantly affect most synaptic parameters. These results were quite surprising and unexpected. Some discrepancies between the results of this study and previous studies-even including those by our laboratory-may be attributable to differences in developmental changes caused by constitutive KO and the well-documented off-target effects of RNA interference. Secondary effects triggered by KD may also have contributed to previously observed severe phenotypes; for example, PTPδ KD induced PTPσ upregulation in neurons, and vice versa, whereas PTPδ KO did not [4] (Fig. S1). It is also plausible that functional redundancies between LAR-RPTP proteins contribute to masking PTPδ loss-induced phenotypes in our experimental approaches. A recent study employing triple-cKO mice lacking all three LAR-RPTPs showed that, overall, the effects of LAR-RPTP triple cKO on cultured hippocampal neurons and hippocampal Schaffer-collateral synapses were marginal [17]. Remarkably, this latter study showed that triple LAR-RPTP cKO in presynaptic CA3 neurons strongly suppressed NMDA-type postsynaptic responses in postsynaptic CA1 neurons [17]. Our recent results using PTPσ-cKO mice did not reveal a similar trans-synaptic effect on postsynaptic responses [25]. Thus, testing whether presynaptic PTPδ is mainly involved in contributing to the reported trans-synaptic regulation of postsynaptic responses is warranted. However, it is equally likely that all three LAR-RPTPs are required for operation of this process in vivo. Regardless of precise answer(s), the current study indicates the importance of confirmatory analyses using a sophisticated system and approaches in elucidating functions of synaptic proteins in vivo.
As limitations, it could be argued that the approaches employed in the current study are not sufficiently sensitive to detect subtle changes in synapse properties. Moreover, the selection of experimental preparations may preclude detection of synaptic roles of PTPδ at in vivo synapses. Although PTPδ mRNA is widely distributed across rodent brain areas [26], it remains to be determined whether endogenous PTPδ protein is widely expressed or is restricted to a subset of certain neuron types or neural circuits. The lack of this information still hinders our ability to fully address the synaptic roles of PTPδ proteins. Development of effective PTPδ-specific antibodies suitable for highresolution in vivo imaging is also warranted, as has similarly been shown for other cell-adhesion proteins [27]. With such tools and information in hand, it should be possible to use the PTPδ-cKO mice developed in the current study to unravel the canonical and non-canonical roles of PTPδ at appropriate in vivo neural circuits. Doing so may provide a better understanding of how LAR-RPTPs act as multivalent signaling platforms in presynaptic neurons.