A new tool to sense pH changes at the neuromuscular junction synaptic cleft

Synaptic transmission triggers transient acidification of the synaptic cleft. Recently, it has been shown that pH affects the opening of postsynaptic channels and therefore the production of tools that allow to study these behaviors should result of paramount value. We fused α-bungarotoxin, a neurotoxin derived from the snake Bungarus multicinctus that binds irreversibly to the acetylcholine receptor extracellular domain, to the pH sensitive GFP Super Ecliptic pHluorin, and efficiently expressed it in Pichia pastoris. This sensor allows synaptic changes in pH to be measured without the need of incorporating transgenes into animal cells. Here, we show that incubation of the mouse levator auris muscle with a solution containing this recombinant protein is enough to fluorescently label the endplate post synaptic membrane. Furthermore, we could physiologically alter and measure post synaptic pH by evaluating changes in the fluorescent signal of pHluorin molecules bound to acetylcholine receptors. In fact, using this tool we were able to detect a drop in 0.01 to 0.05 pH units in the vicinity of the acetylcholine receptors following vesicle exocytosis triggered by nerve electrical stimulation. Further experiments will allow to learn the precise changes in pH during and after synaptic activation.

The content of synaptic vesicles has been shown to be acid 1 , depending on a proton ATPase that causes the electrochemical gradient necessary to transport neurotransmitters into the vesicle 2 . During synaptic transmission, the synaptic vesicle fuses to the presynaptic membrane and its content is released into the synaptic cleft. During this event, vesicular pH rises and falls again after endocytosis 1 , which is consistent with proton release into the synaptic cleft. Taking into account the very small volume of a synaptic vesicle (2 × 10 −20 L) and the estimated pH before exocytosis (around 5.5), it has been calculated that there is less than one free proton per vesicle 3 . However, upon vesicle opening a process which includes deprotonation of proteins and neurotransmitters takes place and over 50 protons are released per vesicle 3 . Therefore, the synaptic cleft, which is a space 50 nM wide, is exposed to changes in pH due to the release of vesicle content and also to transmembrane fluxes owing to Ca 2+ / H + exchange by the plasma membrane Ca 2+ ATPase 4 . A transient drop in extracellular pH has been reported during synaptic activity in hippocampal sections 5 . It has been proposed that in photoreceptor terminals the liberation of vesicular protons results in an acidification of the synaptic cleft from 7.5 to 6.9, depending on the feedback inhibition of Ca 2+ channels 6 . This study also showed that pH oscillations depend on the concentration of protons in the extracellular solution and that acidification is insignificant in acutely dissociated terminals, demonstrating that an intact synaptic cleft is essential.
Recently, Super Ecliptic pHluorin (SEP) 1 , a pH-sensitive green fluorescent protein, was fused to the extracellular domain of a postsynaptic membrane protein and used in lateral amygdala pyramidal neurons to report changes in extracellular pH at dendrite synapses 7 . After stimulation, the transfected pyramidal neurons transiently acidified neighboring dendrites and spines, a phenomenon which was followed by a slower increase  8 . Therefore, the exact changes in pH during synaptic transmission are not fully resolved. Changes in pH at the synaptic cleft at the mammalian NMJ have not been documented. With this aim, we developed a probe by fusing α bungarotoxin (BTX), a neurotoxin derived from the snake Bungarus multicinctus that binds irreversibly to the acetylcholine receptor extracellular domain, to SEP 1 and efficiently expressed this fusion protein in Pichia pastoris. Using this tool, we were able to detect a drop in pH in the vicinity of the acetylcholine receptors following vesicle exocytosis triggered by nerve electrical stimulation.

Results
SEP-BTX was cloned in pPic9, expressed in P. pastoris and purified, resulting in an active protein of the expected size for de fusion (37 kDa, Fig. 1A,B). Muscles incubated with SEP-BTX washed and mounted in the recording chamber were screened looking for NMJ´s following the thin nerve branches with Nomarsky interference optics. NMJ´s were illuminated with a 488 nM wavelength light source to excite the SEP-BTX fusion protein attached to the postsynaptic acetylcholine receptor. Typical mouse NMJ pretzel like images as those displayed in Figs. 1C and 2A were observed.
The sensitivity to pH of the SEP-BTX fusion protein probe was investigated by measuring the fluorescence intensity of the NMJ incubated at different pHs in a high capacity buffer (HEPES buffer 10 mM). A fluorescence vs pH curve was constructed taken the value at pH 9 as the maximal fluorescent output ( Fig. 2A,B) and the following equation Y = y 0 + y max /(1 + 10 (pK-pH) ) was fitted to the data, where y 0 represents the offset, y max the dynamic range, and pK the pH where fluorescence is half maximal, a value that corresponds to minus the logarithm of the equilibrium constant for protonation 9 . The pK from the fit was 7.36.
The effect of nerve stimulation on the NMJ fluorescence was studied measuring the fluorescence intensity before, during and after nerve stimulation at 50, 100 or 300 Hz. As shown in Fig. 3A, a drop in the fluorescence was recorded during stimulation at 100 and 300 Hz in a bicarbonate buffer solution. The maximum pH change was estimated based on the calibration curve to be of 0.024 + / − 0.006 (n:4) for the 300 Hz stimulation and less than 0.02 pH unit for the 100 Hz stimulation. In another series of experiments the synaptic output was increased by incubating the nerve muscle preparation with the K + channel blocker diaminopiridine (DAP, 100 µM, Sigma www.nature.com/scientificreports/ Aldrich, USA), known to increase the amplitude and prolong the time course of transmitter release 10 . As a result of DAP addition, fluorescence changes were significatively stronger (Fig. 3A).
To demonstrate that the observed changes in fluorescence were due to a change in pH in the synaptic cleft, the buffer capacity of the bathing solution was increased by replacing the bicarbonate buffer with a HEPES 20 mM pH 7.3 solution in the presence of DAP. In these experimental conditions, fluorescence changes due to nerve stimulation were not observed (Fig. 3B). In contrast, when the muscle was incubated with a lower HEPES buffer concentration (0.8 mM), the stimulation of the nerve induced a clear reduction in fluorescence, even at a 50 Hz stimulation, which corresponded to a drop in pH of 0.032 + / − 0.008 (Fig. 3B, n:3).

Discussion
The experiments reported here have demonstrated that it is feasible to measure pH changes at the synaptic cleft using a fusion between a compound which binds to the postsynaptic receptor with high affinity and a pH-sensitive probe ready to capture the protons released into the synaptic cleft. This sensor allows synaptic changes in pH to be measured obviating the need of incorporating transgenes into animal cells, a process that can potentially lead to perturbations associated with their exogenous expression.
Our experimental results using SEP-BTX have shown that acidification takes place at the NMJ synaptic cleft, confirming previous reports in other synapses 3,[11][12][13][14][15][16] . A drop in pH during synaptic activity at the NMJ was suggested to be associated to the modulatory effect of the Acid Sensing Ion Channels (ASIC) on neuromuscular transmission 17 . Activation of ASICs requires changes in pH of 0.5 units or more, a value not reached in our measurements. One possibility is that the sensor also binds to other acetylcholine receptors located outside the synaptic cleft. Since these regions would not be exposed to the release of synaptic vesicle content, this event could mask the changes in pH reported by our SEP-BTX sensor in the synaptic cleft. However, it has been established that extra synaptic acetylcholine receptors are not expressed in normal adult mouse muscle both, by iontophoretic application of acetylcholine 18,19 and by using labelled bungarotoxin 20 . Moreover, we focused the detection of fluorescent changes at the synaptic spot where only synaptic acetylcholine receptors are located. Secondly, it is worth mentioning that our measurements are highly dependent on the distance of the probe in relation with the source of the protons, which in our case is the synaptic cleft width. The high mobility of protons 21 and the presence of buffers in the media could result in a very fast transient non accumulative increase in proton concentration not resolved with our detection system. Furthermore, the rapid dissipating proton concentration during the intervals between evoked transmitter release (10 ms at 100 Hz and 3.3 ms at 300 Hz) will strongly down average the fluorescent signal.
In summary, using a novel approach, in contrast to other reports showing an alkalization at CNS synapses 8,22,23 , we were able to show for the first time an acidification process due to transmitter release at the mouse NMJ. www.nature.com/scientificreports/ Synaptic cleft acidification and alkalization are probably not mutually exclusive processes. While they could operate in different types of synapses, it is also feasible that they occur at the same synapsis but with different kinetics. For example, following a rapid drop in pH, an over compensating mechanism driven by the Ca 2+ /H + antiporter activity of plasma membrane Ca 2+ ATPase 4 and aimed at maintaining proton homeostasis could result in transient alkalinization 8 . Further experimental development of sensors located at the presynaptic membrane close to the vesicle release site and a fast image recording system are needed to obtain more realistic measurements of changes in the pH at the synaptic cleft.    AAC CTG TGC  TAC AGA AAA ATG TGG TGT GAT GCC TTC TGT TCT TCA CGT GGT AAA GTT GTC GAG CTG GGA TGC GCT  GCC ACC TGT CCT AGT AAG AAA CCT TAT GAA GAA GTA ACA TGT TGC TCT ACA GAT AAA TGT AAT CCA  CAT CCT AAA CAA CGT CCT GGT TCG CGA GAG AAC CTG TAC TTT CAA GGA CCC GGG CAT CAT CAT CAC  CAT CAT TAA ATT TAA ATG AATTC3′. This fragment was cloned into the EcoRI site of pPic9 (Invitrogen Life Technologies Inc.) in frame with Saccharomyces cerevisiae α-factor signal sequence.

Methods
Expression system. The protocol for protein expression in P. pastoris was previously described 24 . Basically, pPic9-SEP-BTX-TEV-6XHis vector was linearized with DraI and used for transformation of P. pastoris strain GS115 (Invitrogen Life Technologies) by electroporation, and recombinant clones reverting histidine auxotrophy were selected on minimal MD plates (0.34% yeast nitrogen base without amino acids, 10 g/L (NH 4 ) 2 SO 4 , 2% dextrose and 2% agar). Selection of clones expressing and secreting SEP-BTX was performed by transferring colonies to a nitrocellulose membrane placed on plates of MM minimal medium (0.34% yeast nitrogen base without amino acids, 10 g/L (NH 4 ) 2 SO 4 , and 2% agar) with methanol as sole carbon source, for induction of the AOX1 promoter (P. pastoris alcohol oxidase 1). After 4 days of growth at 30 °C, the nitrocellulose membranes were washed with distilled water and the secreted protein was revealed by western blot using a polyclonal rabbit anti-HIS antibody (Genescript, USA). For SEP-BTX production, recombinant clones were grown in 50 mL of BMGY medium (1% yeast extract, 2% peptone, 0.34% yeast nitrogen base without amino acids, 10 g/L (NH 4 ) 2 SO 4 , 400 mg/L biotin, 1% glycerol, 100 mM potassium phosphate buffer, pH 6.0) for 48 h at 30 °C and 220 rpm. Cells were harvested by centrifugation for 5 min at 1500 g and resuspended in BMMY medium (1% yeast extract, 2% peptone, 100 mM potassium phosphate buffer, pH 6.0, 0.34% yeast nitrogen base without amino acids, 10 g/L (NH 4 ) 2 SO 4 , 400 mg/L biotin, 200 µM CuSO 4 and 3% sorbitol) to a final OD 600 nm = 10 and cultivated in 1000 mL shake flasks at 28 °C and 220 rpm. Sterile methanol (0.5% final) was added every 24 h to maintain induction conditions.

Purification of recombinant SEP-BTX fusion protein and western blot.
The protocol for protein purification from P. pastoris was previously described 24 . Briefly, P. pastoris cultures were harvested after 4 days of induction in BMMY and centrifuged at 1500 g for 10 min. The extracellular supernatant was concentrated by ultrafiltration (30 kDa MWCO, Amicon Ultra, Merck Millipore) and buffer exchanged to 300 mM NaCl, 50 mM sodium phosphate buffer, pH 8, containing 1 mM PMSF. SEP-BTX was purified by gravity flow Ni-NTA affinity chromatography using a "His select" nickel affinity gel (Sigma Chemical Co., USA), eluted with 300 mM NaCl, 50 mM sodium phosphate buffer, pH 6.5, 250 mM imidazole, 1 mM PMSF. Fractions containing SEP-BTX were pooled and buffer exchanged to 1× PBS by ultrafiltration (10 kDa MWCO, Amicon Ultra, Merck Millipore). Recombinant SEP-BTX was revealed by western blot. Culture supernatants and purification fractions were separated by reducing 12% SDS-PAGE and transferred to 0.45 μm nitrocellulose membranes (Bio-Rad Laboratories Inc, USA). Western blot was performed by probing the membranes with 0.1 μg/mL of polyclonal rabbit anti-HIS antibody (Genescript, USA) or a 1:1000 dilution of mouse anti-GFP antibody (Abcam) followed by 1:15,000 dilution of alkaline phosphatase-linked goat anti-rabbit antibody or goat anti-mouse antibody (Sigma Chemical Co., USA). Phosphatase activity was revealed by a chromogenic reaction using 5 bromo-4 chloro-3 indolyl phosphate and nitroblue tetrazolium as substrates (Sigma Chemical Co., USA).
Animal protocols. Experiments were carried out on the left Levator auris longus (LAL) muscle of male C57BL/6 J mice (https:// www. jax. org/ strain/ 000664) as previously described 25 . Briefly, animals were supplied by the animal house of the FCEN-UBA. Animals were cared for in accordance with national guidelines for the human treatment of laboratory animals, similar to those of the US National Institutes of Health. All experimental protocols were approved by the Institutional Commission for the Care and Use of Laboratory Animals (CICUAL), FCEN-UBA. Animals were anesthetized with an overdose of 2% tribromoethanol (0.15 mL/10 g body weight) injected in the peritoneal cavity and exsanguinated immediately. The muscle with its nerve supply was excised and dissected on a Sylgard-coated Petri dish containing physiological saline solution of the following composition (in mM): 137 NaCl, 5 KCl, 2 CaCl 2 , 1 MgSO 4 , 12 NaHCO 3 , 1 Na 2 HPO 4 and 11 glucose; continuously bubbled with 95% O 2 /5% CO 2 ; pH 7.3 as previously described 25 .
LAL muscles were incubated in Ringer´s saline solution with the SEP-BTX sensor for 2 h and washed for 30 min. For each batch of SEP-BTX produced, containing 30-50 ng/µL of purified recombinant protein, preliminary experiments were done to identify the dilution (1/5 to 1/20) necessary to obtain a low background strong labelled NMJ. The preparation was then transferred to a 1.5 mL recording chamber. Experiments were performed at room temperature (20-23ºC). NMJs were visualized with Nomarsky interference optics and illuminated with a 488 nM light source (T.l.L.L Photonics, Germany).
The nerve was stimulated via two platinum electrodes isolated with vacuum grease (Down Corning, USA) coupled to a pulse generator (Grass S88, Grass Inst., USA). Supra threshold high frequency (50, 100 or 300 Hz) nerve stimulation was applied during 250 ms. To avoid using high concentrations of the reporter and to prevent any remaining muscle contraction during stimulation, curare up to 5 × 10 -6 g/mL (Sigma Aldrich, USA) was used during the experiment.
Live imaging acquisition and analysis. For fluorescent image acquisition, we used a BX51WI upright microscope with a 60X objective lens (Olympus) and an electron multiplying CCD camera (Andor iXon, Oxford Instruments), together with cell^M System Coordinator/cell^R real time controller software.