Tetanus Toxin Is Transported in a Novel Neuronal Compartment Characterized by a Specialized pH Regulation*

Tetanus toxin binds specifically to motor neurons at the neuromuscular junction. There, it is internalized into vesicular carriers undergoing fast retrograde transport to the spinal cord. Despite the importance of this axonal transport pathway in health and disease, its molecular and biophysical characterization is presently lacking. We sought to fill this gap by determining the pH regulation of this compartment in living motor neurons using a chimera of the tetanus toxin binding fragment (TeNT HC) and a pH-sensitive variant of the green fluorescent protein (ratiometric pHluorin). We have demonstrated that moving retrograde carriers display a narrow range of neutral pH values, which is kept constant during transport. Stationary TeNT HC-positive organelles instead exhibit a wide spectrum of pH values, ranging from acidic to neutral. This distinct pH regulation is due to a differential targeting of the vacuolar (H+) ATPase, which is not present on moving TeNT HC compartments. Accordingly, inhibition of the vacuolar (H+) ATPase under conditions that completely abolish the intracellular accumulation of acidotrophic dyes does not affect axonal retrograde transport of TeNT HC.However, a functional vacuolar (H+) ATPase is required for early steps of TeNT HC trafficking following endocytosis, and it is localized to axonal vesicles containing TeNT HC. Altogether, these findings indicate that the vacuolar (H+) ATPase plays a specific role in early sorting events directing TeNT HC to axonal carriers but not in their subsequent progression along the retrograde transport route, which escapes acidification and targeting to degradative organelles.

Tetanus toxin binds specifically to motor neurons at the neuromuscular junction. There, it is internalized into vesicular carriers undergoing fast retrograde transport to the spinal cord. Despite the importance of this axonal transport pathway in health and disease, its molecular and biophysical characterization is presently lacking. We sought to fill this gap by determining the pH regulation of this compartment in living motor neurons using a chimera of the tetanus toxin binding fragment (TeNT H C ) and a pH-sensitive variant of the green fluorescent protein (ratiometric pHluorin). We have demonstrated that moving retrograde carriers display a narrow range of neutral pH values, which is kept constant during transport. Stationary TeNT H C -positive organelles instead exhibit a wide spectrum of pH values, ranging from acidic to neutral. This distinct pH regulation is due to a differential targeting of the vacuolar (H ؉ ) ATPase, which is not present on moving TeNT H C compartments. Accordingly, inhibition of the vacuolar (H ؉ ) ATPase under conditions that completely abolish the intracellular accumulation of acidotrophic dyes does not affect axonal retrograde transport of TeNT H C . However, a functional vacuolar (H ؉ ) ATPase is required for early steps of TeNT H C trafficking following endocytosis, and it is localized to axonal vesicles containing TeNT H C . Altogether, these findings indicate that the vacuolar (H ؉ ) ATPase plays a specific role in early sorting events directing TeNT H C to axonal carriers but not in their subsequent progression along the retrograde transport route, which escapes acidification and targeting to degradative organelles.
The clostridial neurotoxin family is formed by tetanus (TeNT) 2 and seven serotypes of botulinus neurotoxins (BoNTs, named A-G). They all share an identical structural organization, comprising a 100-kDa heavy (H) chain and a 50-kDa light (L) chain linked via a disulfide bond and other noncovalent interactions (1, 2) (Fig. 1a). The carboxyl-terminal part of the heavy chain (H C , 50 kDa) is responsible for membrane binding and internalization, whereas the amino-terminal domain (H N , 50 kDa) mediates membrane translocation of the L chain into the cytosol (1,3). The L chain is a highly specific zinc endoprotease, which is responsible for the cleavage of synaptic SNARE proteins necessary for neurotransmitter release (2). TeNT binds to motor neurons (MNs) at the neuromuscular junction. Upon internalization, TeNT is retrogradely transported toward the MN cell body, where it is released into the extracellular medium and enters adjacent inhibitory interneurons (4). In these cells, TeNT blocks the release of inhibitory neurotransmitters by cleaving VAMP/synaptobrevin, a member of the SNARE superfamily (1,3).
We have recently developed an assay to follow the retrograde transport of TeNT in MNs using fluorescently tagged versions of the nontoxic TeNT H C binding fragment (5). In these cells, TeNT and TeNT H C are internalized and transported in morphologically identical organelles with overlapping speed distributions (6). These carriers are powered for their retrograde movement by distinct molecular motors, including cytoplasmic dynein and myosin Va (6 -8). TeNT H C shares this compartment with the nerve growth factor and the low affinity neurotrophin receptor p75 NTR (5). These findings validate TeNT H C as an ideal tool for dissecting the molecular machinery controlling axonal retrograde transport and the trafficking of neurotrophin receptors and their ligands in living MNs.
To date, the precise characterization of the intraluminal pH of this axonal retrograde pathway and its relevance in the regulation of transport in MNs remain unclear. These features are particularly important, because acidic pH triggers a conformational change in TeNT and other CNTs, allowing the active subunit to translocate through the endosomal membrane into the cytosol (9, 10). Moreover, acidic pH is known to drive the dissociations of ligands, including growth factors, from their receptors and influence their signaling and intracellular targeting (11,12). We sought to address these important questions by preparing chimeras of TeNT H C and ratiometric pHluorin, a pH-sensitive green fluorescent protein mutant (13). Several pH-sensitive variants of green fluorescent protein have been developed by independent mutagenesis approaches and used as pH reporters to follow vesicular transport dynamics (14,15). Whereas the wild-type green fluorescent protein has a largely unaltered spectrum between pH 5.5 and 10, the excitation spectrum of the ratiometric pHluorin shows two main peaks at 395 and 475 nm, the intensity of which reciprocally varies as a function of pH because of the protonation of its chromophore (13).
In this study, we demonstrated that moving axonal retrograde TeNT H C carriers display a neutral pH, which is kept constant during transport. In contrast, stationary TeNT H C organelles located in axons and in somas exhibit a wide range of pH values. The distinct pH regulation of these compartments is due to a differential targeting of the vacuolar (H ϩ ) ATPase (vATPase), which activity is required for early events in the formation and/or sorting of axonal TeNT H C vesicles but not for their subsequent axonal transport. * This work was supported by Cancer Research UK. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. □ S The on-line version of this article (available at http://www.jbc.org) contains supplemental Figs. S1-S4 and Videos 1 (a-c) and 2 (a-e). 1

MATERIALS AND METHODS
Chemicals and Fluorescent Probes-Chemicals were purchased from Sigma and BDH unless stated otherwise. Restriction endonucleases and DNA-modifying enzymes were from New England Biolabs. pGEX-4T-VSV-G-Kin-TeNT H C encoding the residues 855-1,314 of TeNT fused at the amino terminus with a protein kinase A phosphorylation site, the vesicular stomatitis virus protein G (VSV-G) epitope, and glutathione S-transferase (GST) (3) were linearized with NdeI. This vector and the insert encoding ratiometric pHluorin, obtained by treating the pGEX-2T-AF058694 construct (13) with AgeI, were blunt-ended using the (3Ј35Ј-exo Ϫ ) Klenow fragment of DNA polymerase I, purified and then ligated. The resulting pGEX-4T-3-VSV-G-Kin-pHluorin-TeNT H C construct was shown to correspond to the published TeNT H C and pHluorin sequences, except for the absence of the mutation L220F in the latter (13). After transformation into FB810-competent bacteria, GST-VSV-G-Kin-pHluorin-TeNT H C and GST-pHluorin were expressed and purified as previously described (16). Cysteine-tagged TeNT H C was expressed using pGEX-4T3-Cys-TeNT H C and labeled with AlexaFluor-488-maleimide (Molecular Probes) as previously described (5). Fluorescent probes contained on average 1.8 mol of dye/ mol of TeNT H C .
Immunofluorescence and Western Blot Analysis-Primary MNs were isolated from Sprague-Dawley (embryonic day 14) rat embryos and cultured following previously described protocols (17). Embryo-tested bovine serum albumin was dialyzed at 4°C against phosphate-buffered saline (PBS) for 24 h and then for a further 48 h against Leibovitz L-15 medium (Invitrogen), pH 7.3, using Spectra/Por membranes with a 25-kDa cutoff (Pierce). All experiments were performed using MNs differentiated for 5-8 days in vitro.
For immunofluorescence (IF) experiments, MNs were incubated with either 15 M DAMP (Oxford Biomedical Research; stock solution 3 mM) and/or 15 nM TeNT H C -AlexaFluor-488 for 2 or 45 min at 37°C. The cells were then fixed in 4% paraformaldehyde, 20% sucrose in PBS, permeabilized with 0.1% Triton X-100, and blocked with 2% embryotested bovine serum albumin, 0.25% porcine skin gelatin, 0.2% glycine, and 15% goat serum in PBS. Primary antibodies were diluted in blocking solution and incubated with the fixed cells for 45 min at room temperature. DAMP was detected using monoclonal ␣-2,4-dinitrophenol or 2,4-dinitrophenyl for 45 min at room temperature (Oxford Biomedical Research). After rinsing, AlexaFluor goat ␣-mouse, ␣-rabbit, or ␣-chicken IgG (Molecular Probes) were diluted 1:300 in blocking solution and incubated for 30 min. After extensive washing, the coverslips were mounted with Mowiol 4 -88 (Harco). Fluorescent images were acquired using a Zeiss LSM 510 confocal microscope equipped with a Zeiss ϫ63, 1.40 numerical aperture, difference interference contrast Plan-Apochromat, oil immersion objective. A region of interest was chosen and simultaneously excited using the 488-and 543-nm lines of a krypton-argon and a helium-neon laser, respectively. Images were collected by averaging eight times at a single focal plane.
The polyclonal rabbit ␣ 31-kDa vATPase (a kind gift from Prof. S. Breton, Massachusetts General Hospital, Charlestown, MA) was used 1:300 for IF and 1:4,000 for Western blot (WB). The monoclonal ␣ 60-kDa vATPase antibody (2E7; a kind gift from Prof. H. Sze, University of Maryland, College Park, MD) (18) was diluted 1:10 for IF and 1:100 for WB. The chicken ␣ 75-kDa vATPase (ATP6E; GenWay, San Diego, CA) was used 1:100 for IF and 1:1000 for WB. Polyclonal rabbit ␣ 116-kDa vATPase (Synaptic Systems, Goettingen, Germany) was used 1:300 for IF and 1:3000 for WB. Rat brain extracts (100 g/lane) were incubated for 10 min at room temperature with loading buffer, separated on a 12.5% SDS gel, and transferred onto nitrocellulose (Schleicher & Schuell). After blocking for 1 h in PBS containing 5% skimmed milk, the membranes were incubated with primary antibody for 1 h in the same buffer, washed with PBS, and incubated with horseradish peroxidase-conjugated secondary antibody. Immunoreactivity was detected using enhanced chemiluminescence (ECL, Amersham Biosciences).
Endocytosis Assays-MNs were treated with 0.5 nM BafA1 for 15 min at 37°C and then incubated for 30 min with VSV-G-Kin-TeNT H C (20 nM) or TeNT H C labeled with disulfide-linked biotin 3 (30 nM). After VSV-G-Kin-TeNT H C incubation, the cells were fixed in 4% paraformaldehyde in PBS and blocked with 2% embryo-tested bovine serum albumin, 0.25% porcine skin gelatin, 0.2% glycine, and 15% goat serum in PBS. The rabbit ␣-VSV-G tag antibody (BioGenes, Berlin, Germany) diluted 1:80 in blocking solution was incubated with the fixed cells for 30 min at room temperature. The cells were washed twice with PBS and then permeabilized with 0.1% Triton X-100 in blocking solution and blocked for a further 30 min. MNs were incubated with a primary mouse ␣-VSV-G (Cancer Research UK, London, UK) for 30 min, washed twice in PBS, incubated for 30 min with AlexaFluor goat ␣-mouse and ␣-rabbit IgG (Molecular Probes), and diluted 1:300 in blocking solution. After extensive washing with PBS, coverslips were mounted with Mowiol 4 -88. Alternatively, MNs incubated with TeNT H C -biotin were cooled on ice and then treated three times with ice-cold 20 mM sodium 2-mercaptoethanesulfonate in neurobasal medium (Invitrogen), pH 8.3, for 15 min to cleave the biotin moiety from surface-bound TeNT H C . The cells were washed three times in neurobasal medium, pH 8.3, once in PBS fixed in 4% paraformaldehyde in PBS and permeabilized with 0.1% Triton X-100 and then blocked in blocking solution. TeNT H C -biotin was revealed using streptavidin-AlexaFluor-488 1:500 in blocking solution. The MNs were extensively washed in PBS and mounted in Mowiol 4 -88.
Fluorescence Excitation Spectra of pHluorin-TeNT H C in Vitro and in Vivo-Samples containing 0.5 M pHluorin-TeNT H C or GST-pHluorin in 50 mM sodium cacodylate, 100 mM NaCl, 1 mM CaCl 2 , and 1 mM MgCl 2 were adjusted to the indicated pH values with NaOH or HCl. Excitation spectrums (350 -500 nm) were acquired using a 710 Photomultiplier detection system (Photon Technology International) at a fixed emission of 508 nm. As expected, the fluorescence excitation spectrum of GST-pHluorin was identical to that of pHluorin-TeNT H C .
MNs were incubated with 40 nM pHluorin-TeNT H C in neurobasal medium for 30 min at 37°C and 7.5% CO 2 . The medium was then replaced with a buffer containing 120 mM KCl, 20 mM NaCl, 0.5 mM CaCl 2 , 0.5 mM MgSO 4 , 20 mM HEPES, adjusted to a pH between 5.0 and 8.0 with NaOH or HCl. 20 min before imaging, the ionophores monensin and nigericin (both 10 g/ml; Calbiochem), were added to equilibrate the intracellular pH to that of the external buffer (19). The emission intensities of pHluorin-TeNT H C compartments in axons and soma were measured for each pH at 508 nm upon excitation at 405 and 488 nm. The background was then subtracted, and the corrected emission values for the 405 nm excitation were divided by the corresponding emission intensities obtained by exciting at 488 nm (R405/488).
Microscopy and Data Quantification-To monitor fluid phase uptake, MNs were incubated with 6 M GST-pHluorin in neurobasal medium (Invitrogen) for 1 h at 37°C and 7.5% CO 2 , washed with nonfluorescent Dulbecco's minimum essential medium (without phenol red), riboflavin, folic acid, penicillin/streptomycin, and supplemented with 30 mM HEPES-NaOH, pH 7.3 (DMEM Ϫ ), and immediately imaged by low light time lapse microscopy. Alternatively, MNs were transferred to neurobasal medium, incubated for a further 30 min, washed, and then imaged.
MNs were incubated with 40 nM pHluorin-TeNT H C or 40 nM TeNT H C -AlexaFluor-488 in complete medium for 30 min at 37°C, followed by three washes with DMEM Ϫ . The cells were placed in a humidified chamber maintained at 37°C and after 20 min, imaged every 5 s. In selected samples, where axonal transport of TeNT H C -AlexaFluor-488 was observed, concanamycin A (ConA) (Alexis Biochemicals) or bafilomycin A1 (BafA1) (Alexis Biochemicals) were added at a final concentration of 10 and 0.5 nM, respectively, and the same cell was imaged after 5, 15, and 30 min of incubation. The ConA analysis shown in Fig. 4 is representative of five independent experiments using MNs from two different preparations, whereas the BafA1 effects have been evaluated in three independent experiments using MNs isolated from the same preparation. The functionality of the proton pump inhibitors was tested by applying 50 nM Lysotracker Red DND-99 (Molecular Probes) for 15 min on untreated MNs or MNs pre-incubated with either 10 nM ConA for 15 min or 0.5 nM BafA1 for 10 min and then imaged by confocal microscopy. For quantification purposes, the colocalization macro of the LSM510 software (Zeiss) has been used following the manufacturer's instructions.
For pHluorin-TeNT H C , images were taken every 5 s with a Nikon Diaphot 300 inverted microscope equipped with a Nikon ϫ100, 1.3 numerical aperture Plan Fluor oil immersion objective. Specimens were sequentially excited at 488 and then at 405 nm for 333 ms using the excitation filters 485DF15, 4000DF10, the dichroic filter 490 -575DBDR, and the emission filter 528 -633DBDR (Omega Optical). Images were acquired with a Hamamatsu C4742-95 Orcal cooled charge-coupled device camera (Hamamatsu Photonic Systems) controlled by the Kinetic Acquisition Manager 2000 software (Kinetic Imaging). Two channel series were processed with Lucida, version 4.0 (Kinetic Imaging). For time lapse low light microscopy of TeNT H C -AlexaFluor-488-treated cells, images were acquired every 5 s with a Nikon Diaphot 200 inverted microscope equipped with Nikon ϫ100, 1.25 numerical aperture, difference interference contrast, oil immersion objective, using a standard Nikon B-2A filter and exposure times between 111 and 222 ms.
The motion analysis software (Kinetic Imaging) was used for carrier tracking. Only moving carriers that could be tracked for at least four time points were analyzed. Single movements of TeNT H C carriers, which are described by their progress between two consecutive frames, were used to determine the speed of the carriers. Statistical analysis and curve fitting were performed using Microsoft Excel®. Single frame images were processed for presentation in Adobe® Photoshop®, version 5.5. Kymographs were generated using MetaMorph (Universal Imaging Corporation TM ) after rotation of the image stack to align the neuronal process vertically. Vertical single line scans through the center of each process were plotted sequentially for every frame in the time series.

RESULTS
Axonal TeNT H C carriers have recently been shown to avoid accumulating the acidotrophic probe Lysotracker Red DND-99 in their lumen, suggesting that these organelles escape acidification (5). However, no conclusions could be drawn from previous work about the pH dynamics of the retrograde transport compartment or its regulation in the axon and soma. To address this issue, we chose a strategy based on bacterially expressed ratiometric pHluorin fused with TeNT H C . First of all, we tested whether pHluorin still retains its described optical properties (13) once tagged to TeNT H C by diluting it with buffers ranging from pH 5.0 to 8.0 and monitoring its excitation spectrum at an emission wavelength of 508 nm (Fig. 1b). As expected, the fluorescence spectrum of pHluorin-TeNT H C shows two peaks at 395 and 475 nm. The 395-nm peak was more intense at a neutral pH and decreased at acidic pH values, whereas the 475-nm peak showed an opposite behavior. To establish a pH calibration curve in this system, MNs were pretreated with 40 nM pHluorin-TeNT H C at 37°C for 30 min, washed, and then incubated with buffers at different pH values containing monensin and nigericin. Following treatment with these ionophores, all cellular compartments have been observed to re-equilibrate to the pH of the external medium (19). Ratiometric images were taken, and the 405/488 ratios for each pH, averaged from randomly chosen pHluorin-TeNT H C -positive organelles in axons and somas (n ϭ 24 to 52), were used to build a cellular calibration curve for pHluorin-TeNT H C (Fig. 1c). To test the ability of pHluorin to monitor pH in living cells, we incubated MNs with an excess of GST-pHluorin (6 M) under conditions allowing the labeling of all endocytic compartments by fluid phase uptake. Following incubation for 1 h at 37°C, MNs were washed and imaged immediately or after a 30-min chase. In the absence of chase, the labeled organelles (n ϭ 119) showed a broad pH spectrum, ranging from acidic to neutral (Fig. 1d), which was in agreement with the pH heterogeneity observed experimentally in the endosomal pathway (20). This distribution drastically changed after a 30-min chase, when mainly acidic structures were observed (n ϭ 96). Approximately 70% of the organelles had a luminal pH below 5.5 (Fig. 1d), which reflects the accumulation of GST-pHluorin in late endosomes and lysosomes (20).
These combined results indicate that pHluorin can be used as a functional pH reporter to assess the pH of axonal retrograde carriers in our cellular system. To this end, we incubated rat MNs with 40 nM pHluorin-TeNT H C at 37°C for 30 min. Binding experiments confirmed that pHluorin-TeNT H C binds specifically to the MN surface (data not shown), as reported for untagged TeNT H C (21). The cells were then washed, and after 20 min, low light time lapse imaging started in the 488-nm channel, followed by the 405-nm channel. Ratiometric images (R405/488) were obtained from these two time series (Fig. 2a) (see also supplemental Video 1, a-c). The movement of a retrograde pHluorin-TeNT H C carrier is indicated by arrowheads, whereas a stationary compartment is labeled with asterisks (Fig. 2a). TeNT and TeNT H C have been shown to undergo axonal retrograde transport with a similar speed range both in vivo (0.8 -3.6 m/s) (4) and in vitro (0.2-3.6 m/s) (5, 6). To investigate whether pHluorin-TeNT H C carriers are transported at the same rate as the TeNT compartments, we determined their speed distribution profile. Both the average speed and the overall curve shape observed for pHluorin-TeNT H C carriers fit well with the previously reported data (5,6). In particular, the average speed observed for round vesicles (n ϭ 33, 402 movements) was 0.82 Ϯ 0.25 m/s, whereas tubules (n ϭ 34, 334 movements) had an average speed of 1.18 Ϯ 0.29 m/s (supplemental Fig. S1).
The pH dynamics of axonal retrograde carriers were then followed by monitoring the 405/488 emission ratio of TeNT H C compartments during transport. As shown in Fig. 2b, the pH of both types of carriers remained relatively constant during transport. Based on this finding, we then measured the pH of these organelles by ratiometric imaging only in the first  DECEMBER 23, 2005 • VOLUME 280 • NUMBER 51 frame of the videos, thus avoiding possible artifacts because of phototoxicity and photobleaching. To unravel a possible correlation between intraluminal pH and the average speed of axonal pHluorin-TeNT H C compartments, the velocity of the single movements (n ϭ 735; 67 carriers) was plotted against their corresponding pH (Fig. 2c). The large majority of moving organelles was neutral, their pH falling in the 7.0 -7.5 (42% of the total) or 7.5-8.0 range (26% of the total). This neutral population of axonal carriers included both round vesicles and tubules and presented a speed distribution profile overlapping with that observed previously (compare Fig. 2c with supplemental Fig. S1) (5,6). Only a few pHluorin-TeNT H C -positive compartments displayed a pH below 7.0 (14% of the total), whereas no moving carriers having an acidic pH (Յ6.0) were detected. In sharp contrast, stationary pHluorin-TeNT H C -positive organelles in somas and axons displayed a wide range of pH values, from acidic to neutral (Fig. 2d). This result was confirmed by the partial co-localization of these stationary compartments in the soma with Lysotracker Red DND-99 using time lapse confocal microscopy (supplemental Fig. S2). Organelle acidification in the cell body has also been observed in real time, although with extremely low frequency (data not shown).

pH Regulation of the Axonal Retrograde Pathway
What is the cause of the different pH distribution in moving and stationary pHluorin-TeNT H C -positive organelles? At least three formal explanations might account for this phenomenon: (i) the acidification machinery is excluded from or not acquired by transported organelles, (ii) its activity is specifically inhibited, or (iii) the axonal retrograde carriers containing TeNT H C rapidly dissipate the pH gradient due to a high permeability to protons. A major cellular complex involved in organelle acidification is the vATPase. vATPases are found on a variety of intracellular compartments, including clathrin-coated vesicles, synaptic vesicles, and secretory granules (22) and are responsible for the acidification of degradative organelles, such as lysosomes and phagosomes (22,23). We tested the presence of the vAT-Pase on axonal and somatic organelles containing TeNT H C using antibodies against subunits B and E in the V 1 domain of the vATPase and the a subunit in the V 0 domain (Fig. 3a). All antibodies specifically recognized their corresponding subunits in rat brain extracts (Fig. 3b). Immunofluorescence experiments performed by fixing MNs upon incubation for 45 min with TeNT H C -AlexaFluor-488 revealed that vATPase and TeNT H Cpositive transport compartments were morphologically very different both in axons (Fig. 3c) and soma (Fig. 3d). A very limited overlapping of the two staining patterns was seen (Յ6% at late internalization points; see Fig. 5e), suggesting that the vATPase is not present on TeNT H C carriers. The frequency of the structures in which vATPase colocalized with TeNT H C (Fig. 3) resembled those observed for the stationary TeNT H C organelles, which can be acidified (Fig. 2d). We were also able to partially localize the vATPase with DAMP, an acidic granule marker (supplemental Fig. S3, a  S3c), thus confirming that these antibodies are able to detect functional vATPase on acidic organelles in neurons (22).
A major caveat of an experimental approach based on immunolocalization is the limit imposed by the sensitivity of the staining antibodies. To exclude a role for the vATPase activity in axonal TeNT transport, we specifically blocked its catalytic function using ConA and BafA1. Both drugs bind with high affinity to the subunit c of the vATPase V 0 domain (IC 50 ϭ 10 nM for ConA and 0.5 nM for BafA1) (24). At these concentrations, ConA and BafA1 are effective in blocking Lysotracker Red DND-99 accumulation in acidic organelles in MNs (Fig. 4a), demonstrating that the activity of the vATPase was inhibited under these experimental conditions. However, direct comparison of the speed distribution profile of TeNT H C carriers in untreated MNs (Fig. 4, b and c, solid line) with those of treated MNs at different time points (Fig. 4, b  and c, dotted lines) indicates that neither of the drugs alter axonal retrograde transport (see also supplemental Video 2, a-e and Fig. S4a). TeNT H C transport was also unaltered by long term incubation with ConA (Ն90 min). In treated MNs, we observed a slight reduction in speed values and a decrease in carrier frequency at later time points (Fig.  4, b and c, empty circles). We attribute these minor variations to the repetitive, long term imaging of the same axon, because untreated MNs showed the same overall alterations of AlexaFluor488-TeNT H C transport (supplemental Fig. S4b).
To test whether the vATPase is involved in an early step of TeNT H C trafficking, we modified our experimental protocol by pretreating MNs for 15 min with BafA1 and ConA before adding AlexaFluor-488-TeNT H C (Fig. 5a). Imaging started 20 min after removal of the fluorescent TeNT H C (t ϭ 0 min). Only a faint, rather homogeneous membrane staining of MN axons was observed at t ϭ 30 min (Fig. 5b). As shown by the kymographs in Fig. 5b, no moving TeNT H C carriers or stationary TeNT H C -positive compartments were detected, strongly indicating that this probe was not sorted to the retrograde transport pathway. In contrast, kymographs of MNs treated with vATPase inhibitors after endocytosis and sorting of TeNT H C following the protocol described in Fig. 4b (t ϭ 30 min) displayed several progressing carriers (Fig. 5c, arrowheads) and stationary compartments (Fig. 5c, asterisks). Interestingly, we were able to observe a significant overlap between AlexaFluor-488-TeNT H C and vATPase at early time points of internalization (t ϭ 2 min) (Fig. 5, d and e), in sharp contrast to that observed at late time points (t ϭ 45 min) (Fig. 5e).
Is TeNT H C internalization impaired by treatment with vATPase inhibitors? Two independent lines of evidence shown in Fig. 6 argue pH Regulation of the Axonal Retrograde Pathway DECEMBER 23, 2005 • VOLUME 280 • NUMBER 51 against this possibility. Treatment with 0.5 nM BafA1 prior to incubation with VSV-G-tagged TeNT H C followed by sequential staining in the absence and the presence of permeabilization (Fig. 6a) revealed that a pool of TeNT H C is internalized under these conditions. Furthermore, BafA1 application did not impair the entry of biotinylated TeNT H C into MNs (Fig. 6b). These findings indicate that a functional vATPase is not required for TeNT H C endocytosis but plays a specific role in an early sorting step(s) targeting TeNT H C to axonal transport carriers.

DISCUSSION
In this study, we monitored the pH dynamics of TeNT H C -positive compartments in real time using a pHluorin-TeNT H C fusion protein.
Chimeras of pHluorin have been used as pH probes in a variety of biological systems. In neurons, pHluorin-based sensors have been exploited to follow the presynaptic activity within neuronal networks (14) to report synaptic vesicle fusion (25) and to assess the molecular identity of different synaptic compartments during synaptic vesicle recycling (26). On this basis, we chose a novel fusion protein between ratiometric pHluorin and TeNT H C as a noninvasive method to investigate the pH of the fast axonal retrograde compartment in MNs. Here, we showed that these endocytic carriers have a neutral pH, which is kept constant during axonal movement. This finding is in contrast with the pH dynamics described for the classical endosomal pathway, which undergoes a rapid acidification upon internalization (20). The entry of TeNT into a neutral endocytic compartment has important mechanistic effects on the pathogenesis of tetanus. In fact, acidic pH triggers a conformational change of TeNT, which allows its membrane insertion and the translocation of the L chain into the cytosol, where it cleaves VAMP (2). This phenomenon occurs in hippocampal neurons in vitro (27) and in spinal cord inhibitory interneurons in vivo (1,28). In artificial liposomes, the pH initiating the membrane insertion of TeNT is ϳ5.0 (29), which is within the pH range of the endosomal pathway. By dem- onstrating that the fast axonal retrograde compartment in MNs displays a neutral pH, we provide an explanation of the entrapment of TeNT in the lumen of the retrograde carrier during axonal transport to the soma, a journey shared with neurotrophins and their receptors (5). The lack of acidification of the fast retrograde transport carriers might contribute to the low degradative capability of this compartment, which ensures the integrity of its endogenous and exogenous cargoes. This is in agreement with the long half-life of internalized TeNT in vivo (Ն5 days in mouse sciatic nerve) (28).
Several pathogens and virulence factors have been shown to either transit or accumulate in nonacidic cellular compartments. This is the case for shiga-like (SLT-1B) and cholera toxin (CT-B) B subunits, which have been extensively used as probes to uncover Golgi-dependent and -independent pathways from the plasma membrane to the endoplasmic reticulum (30,31). Chlamydia pneumoniae, an intracellular parasite, inhabits a nonacidic vacuole, which is distinct from late endosomes and lysosomes (32). Furthermore, echovirus 1 and simian virus 40 (SV40) enter a nonacidic compartment after leaving the plasma membrane in caveolin-enriched structures termed "caveosomes," which lack markers for endosomes, lysosomes, endoplasmic reticulum, or Golgi apparatus (33)(34)(35). In contrast with the retrograde transport carriers, stationary TeNT H C -positive structures, which are distributed in soma and neurites, have a much broader pH range spanning from pH 5.0 -7.5. This observation is in agreement with the reported trimodal frequency distribution of the endocytic organelle pH in axon shafts of sympathetic neurons (36). Furthermore, it suggests that the machinery determining the intraluminal pH of these axonal compartments is strictly regulated and coordinated to the activity of the motor complexes responsible for their retrograde movement (37).
What is the mechanistic basis of this differential pH regulation? A likely possibility is that this is due to the sorting of the vATPase away from this compartment. Acidification is essential for diverse cellular processes, such as protein targeting along the secretory pathway, recycling of receptors to the plasma membrane, and protein degradation in the endomembrane lumen (22,38). Inhibition of the vATPase activity perturbs some, if not all of these functions (39,40). Consequently, we FIGURE 6. BafA1 does not inhibit TeNT H C internalization. a, MNs were treated with 0.5 nM BafA1 for 15 min at 37°C, followed by 30 min of incubation with VSV-G-Kin-TeNT H C. The cells were fixed and surface-bound TeNT H C was detected using an ␣-rabbit VSV-G antibody (green). MNs were then permeabilized and internalized TeNT H C was detected using an ␣-mouse VSV-G antibody (red). See "Materials and Methods." A scheme of the experiment is shown on top. Scale bar, 5 m. b, after BafA1 treatment (0.5 nM), cells were incubated with TeNT H C labeled with disulfide-linked biotin at 37°C for 30 min, washed, and incubated with a membrane-impermeable reducing agent (sodium 2-mercapto-ethanesulfonate) on ice.
Internalized TeNT H C was detected using streptavidin-AlexaFluor-488. For the control, MNs were incubated with TeNT H C -biotin on ice for 20 min prior to sodium 2-mercapto-ethanesulfonate treatment. Time scale of the experiment is on the top. Scale bars, 10 m. tested for the presence of the vATPase by using several antibodies against different subunits of its V 1 and V 0 domains. The experiments revealed very little colocalization between TeNT H C -and vATPasepositive organelles, suggesting that the lack of acidification is due to the absence of the vATPase from the retrograde carriers and not to an altered proton permeability of their membrane bilayer (41). In contrast, the vATPase is likely to be present on stationary organelles, which can be acidified. An inducible, organelle-specific vATPase sorting has been previously demonstrated for phagosomes containing Mycobacterium avium, which fail to acidify due to the exclusion of vATPase from their delimiting membranes (42). The notion that vATPase activity is dispensable for fast retrograde transport in MNs has been confirmed by the lack of effect on this process of vATPase inhibitors used at concentrations abolishing the accumulation of acidotrophic dyes. In contrast, pretreatment of MNs with vATPase inhibitors blocked TeNT accumulation in endosomal compartments and its sorting to the retrograde transport route. Accordingly, we found that TeNT organelles may contain vATPase at very early time points after their endocytosis. This is in agreement with previous findings where the vATPase has been detected in clathrin-coated vesicles (22). Interestingly, TeNT has been found in clathrin-coated vesicles in spinal cord neurons (43), and it is internalized via a clathrin-dependent pathway in isolated MNs. 4 Clathrin-coated vesicles and other early endocytic organelles, which may coincide with the stationary TeNT H C structures, might act as initial sorting stations where axonal carriers are formed and their content sorted for transport. Mild acidic pH values might play a role in the early steps of the endocytic pathway of TeNT by inducing a partial conformational change, which, although not sufficient to trigger the insertion of the molecule in the inner core of the lipid bilayer, might promote an enhanced interaction with the membrane surface and facilitate the initial sorting of TeNT H C .
In conclusion, we demonstrated that a functional vATPase is transiently required for the initial sorting of TeNT H C to its retrograde transport route, and it is then sorted away from the axonal carriers at later stages. As a consequence, these retrograde organelles display a neutral luminal pH, which might protect the native conformation of acid-labile cargoes and stabilize receptor-ligand interactions. In particular, these conditions might allow the sustained interaction of some neurotrophins (such as nerve growth factor, which has been found to enter TeNT H C carriers (5), and its receptors), providing additional basis for the long range axonal neurotrophin signaling (44). Future experiments will directly address the molecular composition of these retrograde carriers and the activation status of neurotrophin-receptor cargo complexes during transport.