Nature Reference Number: 2002-04169B; File: MisslerNeurexinMsSupplMaterialsFinal; 3/15/03
Supplementary Materials
M. Missler et al., "a-Neurexins Couple Ca2+-Channels to Synaptic Vesicle Exocytosis"
Note: reference numbers refer to references in the main manuscript.
TEXT
Brain structure in a-neurexin KO mice. At birth, the body and brain weights of triple KO and control mice did not differ significantly, but a-neurexin mutant pups frequently showed a lack of milk in their stomachs (see Table S1). We observed no anatomical differences between triple KO mice lacking all a-neurexins and control mice in Nissl-stained brain sections (Fig. S3), and found no changes in the immunocytochemical distribution of synaptic proteins (data not shown). Major fiber tracts (e.g., the corpus callosum and anterior commissure) were present, without evidence for axonal pathfinding abnormalities. Furthermore, we detected no apparent changes in neurite outgrowth and axonal branching in cultured hippocampal neurons from a-neurexin deficient mice (Fig. S4), and found no increase in apoptotic cell death by light and electron microscopy or TUNEL analysis in triple KO mice (Fig. S3). Quantitations of the levels of 22 neuronal proteins in mice lacking all a-neurexins also detected no major changes (see Table S2).
Biophysical parameters of the electrophysiological recordings in brainstem slices. Capacitance values for neurons in the pre-Bötzinger complex in pF [number of cells/number of mice]: wild type, 27.3 ± 1.7 [41/35]; 2α KO, 26.3 ± 2.5 [14/14]; 2a/3a double KO, 27.6 ± 2.3 [23/23]; 1a/2a double KO, 28.3 ± 2.7 [21/21]; triple KO, 28.0 ± 2.1 [32/32]). In hypoglossal motor neurons: mean capacitance ± SEMs [number of mice analysed]: WT, 32.6 ± 3.0 pF [5]; 2α KO, 29.7 ± 4.7 pF [10]; 2a/3a double KO, 32.9 ± 7.0 pF [5]; 1a/2a double KO, 30.9 ± 4.3 pF [5]; and triple KO, 30.1 ± 7.10 pF [6]). No significant differences in serial resistance or input resistance were detected.
Ca2+-current recordings of neurons in slices are notoriously difficult because the extensive arborization of most neurons causes enormous voltage clamp problems. As a result, traditionally most Ca2+-current analyses are carried out on cultured immature neurons. However, such analyses could be potentially useless in a case such as that of the a-neurexin KO mice. In these mice, the phenotype is presumably coupled to synapse formation which is absent or abnormal in the immature cultured neurons that are used for biophysical analysis of Ca2+-channels. For our analyses we therefore carried out the Ca2+-current recordings in whole-cell mode in neurons of the brainstem that are relatively small and have less voltage clamp problems than most neurons in acute slice preparations. In our recordings, the serial resistance (8-20 MW) and the membrane resistance (usually between 0.8 - 1.2 GW) exhibited ratios of at least 1:40 (usually better than 1:50) which is reasonably close the ideal ratio of 1:100 and allowed conclusions based on the maximal amplitude of the Ca2+-currents, although the biophysical properties of the Ca2+-currents were difficult to assess in these measurements because of the space clamp problems.
EXPERIMENTAL PROCEDURES
Generation and breeding of a-neurexin KO and transgenic mice. a-Neurexin KO mice generated by deleting the large first exons of the murine a-neurexin genes were maintained on a mixed SV129/C57bl6 background, and genotyped by PCR essentially as described13,46 (see Fig. S1; Sequences of oligonucleotides used for genotyping: neurexin 1α WT reaction: oligo A (CGAGCCTCCCAACAGC GGTGGCGGCGGGA) vs. oligo B (CTGATGGTACAGGGCAGTAGA GGACCA) (440 bp product); neurexin 1α KO reaction: oligo C (GAGCGCGCGCGGCGGAGTTGTTGAC) vs. oligo B (390 bp product); neurexin 2α WT reaction: oligo D (AGATGCAGGTGGCCAGCGACTGTTC) vs. oligo E (CACTGCAGAAACTTGCCGCCAAATC) (305 bp product); neurexin 2α KO reaction: oligo F (CCTTCGATTAAACCTTTCCTC) vs. oligo G (GGATGCGGTGGGCTCT ATGGCTTCTGA) (220 bp product); neurexin 3α WT reaction, oligo H (CCTGAGTCCTCAGTTGTTTGC) vs. oligo I (CTGAGTGGAGGGTAAAGCTCA) (650 bp product); neurexin 3α KO reaction, oligo H vs. oligo J (CGCCGCTCCCGATTCGCAGCGCAT) (610 bp product)).
Analysis of respiration: Baseline respiratory patterns were measured by whole body plethysmography: Unanesthetized newborn pups (P1) were placed in a 15 ml closed chamber connected to a differential pressure transducer. The analog signal was amplified, digitized, and analysed off-line.
Transgenic mice production was achieved as described47 using a chimeric HRP-neurexin 1a vector encoding neurexin 1a with an inserted HRP-tag as an epitope (A.R., W.Z., R.E.H., T.C.S., and M.M., in preparation). All analyses were carried out with heterozygous transgenic mice on wild type or a-neurexin KO backgrounds. Transgenic neurexin-HRP fusion proteins were probed with affinity-purified HRP antibodies.
Breeding of KO mice: To obtain sufficient numbers of mice for experimental analysis in spite of the mortality of double and triple a-neurexin KO mice (see Fig. 1 and Table S1), we performed all analyses on littermate mice that were descendents of triple heterozygote matings. These matings led to separate sublines that were used as parental lines for experiments: lines of wild type controls with matching single a-neurexin KO mice, and a line that is heterozygous for the neurexin 1a and 3a KOs and homozygous for the neurexin 2a KO (the KO with the least phenotype). In addition, single neurexin 3α KOs and double neurexin 1α/3α KOs derived from the triple heterozygote matings were analysed. In all experiments, heterozygous mice were pooled with wild type mice as controls for a given neurexin. Since heterozygous KO mice probably already have a phenotype given the dose-dependence of the various phenotypes, this analysis likely underestimates the phenotype. In all experiments except where indicated, "n" refers to number of mice analysed, and in most experiments the experimenters were blinded with respect to the genotype of the mice. Survival curves were obtained by sacrificing complete litters at the time points indicated in Fig. 1b, and the ratios obtained at P30 were independently confirmed by analysis of the offspring of triple heterozygous matings (n = 1018). This cumbersome approach was necessary because continuous counting of the surviving pubs in a litter was not feasible since the handling of the mice during counting caused a severe bias due to the behavioral sensitivity of the a-neurexin KO mice.
Morphology and image analysis. Brains from newborn mice (P1) were immersion-fixed, and adult mice were perfusion-fixed, in 0.1 M phosphate buffer containing 2% paraformaldehyde for light microscopy, or 2.5% paraformaldehyde / 2.5% glutaraldehyde for electron microscopy. Serial coronal or horizontal cryostat sections mounted on glass slides were examined by histological stainings (cresylviolet, Nissl), TUNEL analysis (Roche), or immunohistochemistry using antibodies to synaptic, growth cone, dendritic, and axonal proteins. For quantitative analysis of immunofluorescence images, neighboring vibratome sections (~150 μm) were fixed briefly, cryoprotected, and sections parallel to the surface were cut on a cryostat. After double-immunolabeling with antibodies to the vesicular GABA (VGAT, 1:100; Synaptic Systems) or glutamate transporter (VGLU2, 1:200; Synaptic Systems) and to synaptophysin (Sy38, Dako), sections were probed with Alexa-coupled secondary antibodies (Molecular Probes), and viewed in a Zeiss Axioskop 2 with an AxioCam digital camera or a Zeiss LSM510 confocal microscope. Confocal laserscans were transferred into Adobe Photoshop to separate different color channels. Subsequently NIH Image 1.61 software was used to quantitate area densities of the stained terminal clusters.Ultrastructural analyses were performed on counter-stained ultramicrotome sections (60-80 nm) in a Joel 1200 electron microscope. Random anonymised electron micrographs from the nucleus tractus solitarius of the brainstem of wild type and triple KO mice or area Oc1 of the neocortex from wild type and neurexins 1a/2a double KO mice were examined. High magnification pictures were screened for changes in synaptic structure such as active zone length, number of vesicles, etc.
Neocortical slice culture and electrophysiology. Thick slices (0.5 mm) from P1 mice were cultured 25-40 days in vitro on teflone membranes (Millicell-CM, 0.4 µm, from Millipore Corp.). Incubation was done at 37°C and 5 % CO2 in medium consisting of: Minimum Essential Medium (50 %, 2x), Basal Medium Eagle (25 %, 1x), and horse serum (25 %, 1x) with addition of L-glutamine and glucose. To inhibit glia proliferation, ARA-C (10 µM) was added once after 1 day in vitro, and the medium was changed every 2-3 days. Whole-cell voltage clamp recordings were performed at a holding potential of -80 mV at 28-30 °C in a submerged chamber perfused with gassed artificial cerebrospinal fluid (ACSF) containing (in mM): 119 NaCl, 2.5 KCl, 1.5 MgCl2, 2.5 Ca2+, 1 Na2HPO4, 26 NaHCO3, 10 Glucose. Patch pipettes (7-10 MW) were pulled from borosilicate glass (GC150TF-10, Harvard Apparatus Ltd.) filled with a solution containing (in mM): 135 CsCl, 20 TEA, 2 MgCl2, 10 HEPES-KOH pH 7.3. Minis were recorded in 5 mM Ca2+ and 1 μM tetrodotoxin with 10 μM DNQX (GABAA-dependent minis) or 0.1 mM picrotoxin (AMPA-dependent minis). In paired recordings, presynaptic action potentials were elicited by 4 depolarizations to +10 mV (0.5-1 ms) at 20 Hz with 5 sec intervals (30 repetitions). Postsynaptic GABAergic responses were identified by their kinetic properties at membrane potentials of -80 mV and +40 mV and by reversible blocks with bicuculline (25 µM). Sucrose responses were recorded in ACSF containing 50 μM D-APV, 10 μM DNQX, and 1 µM tetrodotoxin. Pressure-application (~400 mm Hg) of hypertonic sucrose (0.5 M sucrose, 50 µM D-APV, 10 µM DNQX and 1 µM tetrodotoxin) to the cell soma was done with a patch pipette (2 µm diameter) for 4 sec with 45 sec intervals. All recordings were made using a HEKA EPC-7 patch-clamp amplifier, filtered at 3kHz, and sampled at 10kHz (2 kHz for sucrose application) with a TL-1 interface using pClamp 6.0.4Software, and analysed off-line with AUTESP and Clampfit Software (Axon Instruments, Union City, CA).
Detailed description of brainstem electrophysiology. General. All electrophysiological analysis of brainstem neurons were performed on brainstem slices of littermate mice whose genotype was unknown to the experimenter. Mice were analysed at the day of birth (P1). The dissection of brainstem slice were performed essential as described48. In detail, coronal sections of the brainstem (200 µm, P1) were cut in a rostral to caudal direction on a vibratome until the nucleus ambiguus and inferior olive were seen at the rostral boundary of the pre-Bötzinger-Complex. For all experiments, slices containing the pre-Bötzinger-Complex and hypoglossal motor nucleus (see Fig. S7 for a diagram of the recording area) were immediately transferred into a recording chamber and submerged and superfused (flow rate 10 ml/min) with recording solution (composition in mM: 118 NaCl, 3 KCl, 1.5 CaCl2, 1 MgCl2, 25 NaHCO3, 1 NaH2PO4, 5 glucose pH 7.4) aerated with 95% O2 and 5% CO2 an kept at 28-30 ºC. Slices were approximately 200 μm thick, except for recordings from hypoglossal nerve rootlets where slices were approximately 700 µm thick. For whole-cell recordings, all neurons were approached under visual control (Axioscop, Zeiss, Germany) with infrared optics (C2400, Hamamatsu Photonics, Enfield, UK). In all experiments, stock solutions of tetrodotoxin, nifedipine, w-conotoxin and w-agatoxin were dissolved in bath solution and superfused over the slice. Recordings were started 1.5 min after the drug application. After establishing the whole-cell configuration, membrane capacitance, serial and membrane resistances were estimated from the current transient induced by 20 mV hyperpolarization voltage commands from a holding potential of -70 mV (Gillis K.D. (1995) Techniques for membrane capacitance measurements. In: B. Sakmann & E. Neher, eds., Single-channel recordings, Plenum Press, New York, pp. 155-198). While the serial resistance was obtained from the voltage step (20 mV) and the “instantaneous current” just after the jump, the membrane resistance was estimated from the voltage step (20 mV) and the steady-state current (A. Marty and E. Neher (1995) Tight-seal whole-cell recording. In: B. Sakmann & E. Neher, eds., Single-channel recordings, Plenum Press, New York, pp. 31-52). Values of serial resistance were checked and corrected repeatedly during each experiments. At least 80% of the value were compensated. Generally, no significant differences in serial resistance (8 to 20 MW) or membrane resistance (0.8 to 1.2 GW) could be found between different genotypes. In experiments, cells were excluded from the analysis if the serial resistance exceed 20 MW, the membrane resistance was below 0.8 GW or the leak currents exceed 150 pA.
Recording of spontaneous miniature postsynaptic currents. For “mini”-experiments, patch electrodes were filled with a solution containing (mM): 140 KCl, 1 CaCl2, 10 EGTA, 2 MgCl2, 4 Na3ATP, 0.5 Na3GTP, 10 HEPES-KOH pH 7.3. As the concentration of chloride-ions was similar between intra- and extracellular solutions, the reversal potential of chlorid was close to 0 mV. In control and after drug applications, minis were recorded in presence of 10 µM CNQX for 3 minutes, during which up to 300 events were collected. Minis could be blocked by 2 μM bicuculline and 2 μM strychnine, demonstrating that they are GABAergic. The mini events were detected when the amplitudes were at least 2.5 times higher than the background noise. The statistical significance of minis were tested in each experiment with Kolmogorov-Smirnow test using commercial software (MiniAnalysis, Synaptosoft, USA).
Recording of evoked EPSCs. For experiments of evoked EPSCs, patch electrodes were filled with a solution containing (mM): 140 KGluconate, 1 CaCl2, 10 EGTA, 2 MgCl2, 4 Na3ATP, 0.5 Na3GTP, 10 HEPES-KOH pH 7.3. EPSCs were elicited by electrical stimulation of axons of the pre-Bötzinger complex, and recorded from hypoglossal motoneurons in the presence of 2 µM strychnine and 2 µM bicuculline. For stimulation, we used bipolar platinum wire electrodes (30 µm diameter, Degussa, Germany) positioned in a theta-style two channel borosilicate glass capillary with a 10 µm distance between wires (Clark Electromedical, UK). Current pulses (0.5-10 mA; 50 μsec duration, 0.1 Hz, 25 pulses total) were applied with a commercial isolation-unit IsoFlex (A.M.P.I., USA) and a custom-build power supply was used. The stimulation strengths given in the paper were read-outs from the potentio-meter on front of IsoFlex. Commercial software was used for data acquisition and analysis (pClamp 6.0 and AxoGraph 3.5, Axon Instrument Inc., USA, MiniAnalysis, SynaptoSoft, USA, and GraphPad Software, USA). The peak amplitude was derived by averaging 25 consecutive EPSCs in control and after application of drugs. To monitor changes in input resistance, current responses to a –10 mV voltage step (20 msec) from a holding potential of –70 mV were recorded before every fifth stimulus. In all experiments the distance between the stimulation and recording electrodes was similar between slices of different genotypes and we could not found any significant differences in latency of evoked responses between the genotypes. For the experiments described in Figs. 4c and 4d, different stimulation strengths were applied for each KO tested in order to compensate for the decreased KO response. Stimulation currents were increased until a 50% failure rate was achieved, and then additionally increased by 20% for the actual experiments. For some triple KO mice, the maximal stimulation with the available instrumentation was insufficient for reaching a 50% response, in which case the maximal possible current was used (10 mA).