Elevated Cytosolic Cl⁻ Concentrations in Dendritic Knobs of Mouse Vomeronasal Sensory Neurons

Abstract

In rodents, the vomeronasal system controls social and sexual behavior. However, several mechanistic aspects of sensory signaling in the vomeronasal organ remain unclear. Here, we investigate the biophysical basis of a recently proposed vomeronasal signal transduction component—a Ca²⁺-activated Cl⁻ current. As the physiological role of such a current is a direct function of the Cl⁻ equilibrium potential, we determined the intracellular Cl⁻ concentration in dendritic knobs of vomeronasal neurons. Quantitative fluorescence lifetime imaging of a Cl⁻-sensitive dye at the apical surface of the intact vomeronasal neuroepithelium revealed increased cytosolic Cl⁻ levels in dendritic knobs, a substantially lower Cl⁻ concentration in vomeronasal sustentacular cells, and an apparent Cl⁻ gradient in vomeronasal neurons along their dendritic apicobasal axis. Together, our data provide a biophysical basis for sensory signal amplification in vomeronasal neuron microvilli by opening Ca²⁺-activated Cl⁻ channels.

Introduction

In most mammals, the vomeronasal organ (VNO) plays a critical role in semiochemical detection and social communication. As a paired tubular structure at the base of the anterior nasal septum (Meredith 1991; Halpern and Martínez-Marcos 2003) the rodent VNO harbors a few hundred thousand vomeronasal sensory neurons (VSNs). Each of these small bipolar neurons extends a single unbranched apical dendrite that terminates in a microvillous knob at the interface to the organ’s mucus-filled lumen. G protein-coupled vomeronasal chemoreceptor proteins are encoded by 3 unrelated multigene families—the V1r (Dulac and Axel 1995), V2r (Herrada and Dulac 1997; Matsunami and Buck 1997; Ryba and Tirindelli 1997), and Fpr-rs (Liberles et al. 2009; Rivière et al. 2009) genes. Monogenic receptor expression and type-specific coexpression of G protein α-subunits (G_αi2/G_αo) confer a distinct receptor “identity” to each VSN (Roppolo et al. 2007) and result in topographical segregation into at least 2 VSN subpopulations.

Vomeronasal receptor–ligand interaction triggers a metabotropic signaling cascade that, via activation of phospholipase C (PLC), elevates cellular concentrations of both diacylglycerol (DAG) and inositol-1,4,5-trisphosphate (IP₃) (Krieger et al. 1999; Holy et al. 2000). Various models for the actions of either or all products of PLC-dependent lipid turnover have been proposed (Spehr et al. 2002; Lucas et al. 2003; Zhang et al. 2008; Yang and Delay 2010; Kim et al. 2011, 2012; Dibattista et al. 2012). These concepts share a common denominator: a key role of cytosolic Ca²⁺ elevations and an important, though not indispensable (Kelliher et al. 2006; Yu 2015) function of the transient receptor potential (TRP) channel isoform TRPC2 (Liman et al. 1999). Cytosolic Ca²⁺ transients can affect various transduction cascade proteins, exerting both positive and negative feedback regulation (Chamero et al. 2012). Both mouse (Spehr et al. 2009) and hamster (Liman 2003) VSNs express a Ca²⁺-activated nonselective cation current that might boost sensory signals. By contrast, Ca²⁺/calmodulin mediates VSN adaptation and gain control by inhibition of TRPC2 (Spehr et al. 2009), analogous to effective inhibition of cyclic nucleotide-gated channels in sensory neurons of the main olfactory epithelium (Bradley et al. 2001, 2004; Munger et al. 2001; Song et al. 2008).

Similar to the Ca²⁺-dependent Cl⁻ conductance known to amplify transduction currents in canonical olfactory neurons (Pifferi et al. 2009; Stephan et al. 2009; Sagheddu et al. 2010; Billig et al. 2011; Dauner et al. 2012; Ponissery-Saidu et al. 2013; Henkel et al. 2014), a considerable amount of stimulus-evoked VSN activity seems to be carried by a Ca²⁺-activated Cl⁻ current (Yang and Delay 2010; Kim et al. 2011; Dibattista et al. 2012). Its molecular correlate has recently been identified as TMEM16A/anoctamin1 (Amjad et al. 2015), the founding member of the anoctamin family of Ca²⁺-activated Cl⁻ channels (Caputo et al. 2008; Schroeder et al. 2008; Yang et al. 2008). While different groups have demonstrated expression of both anoctamin1 and anoctamin2 in VSNs (Rasche et al. 2010; Billig et al. 2011; Dauner et al. 2012; Dibattista et al. 2012), vomeronasal Ca²⁺-activated Cl⁻ currents were abolished in conditional anoctamin1 null mice (Amjad et al. 2015). For interpretation of these results, however, it is essential to determine whether vomeronasal Cl⁻ channels contribute a strongly depolarizing current or, by contrast, if anoctamin1 activation mediates membrane hyperpolarization or shunting inhibition. Here, the Cl⁻ equilibrium potential (E _Cl) established at the microvillar VSN membrane in vivo is the key physiological determinant. Thus, the driving force for Cl⁻ at the apical VSN membrane depends on each neuron’s cytosolic Cl⁻ concentration ([Cl⁻]_c), a physiological parameter presently unknown.

Here, we used 2-photon fluorescence lifetime imaging microscopy (FLIM) of a Cl⁻-sensitive dye to measure the intracellular Cl⁻ levels in VSN dendritic knobs. Our data indicate a relatively homogeneous [Cl⁻]_c at rest that ranges from 31 to 63mM (25 and 75 percentile) and appears significantly increased compared to values measured in the apical compartment of VNO sustentacular cells. Moreover, our findings suggest an apparent Cl⁻ gradient along the VSN dendritic apicobasal axis, indicative of active Cl⁻ accumulation in the knob layer. Thus, at resting membrane potential, elevated [Cl⁻]_c in mouse VSN knobs likely results in an outwardly directed driving force for Cl⁻ and, consequently, response amplification upon opening of Ca²⁺-activated Cl⁻ channels.

Materials and methods

Animals

All animal procedures were approved by local authorities and in compliance with European Union legislation (Directive 86/609/EEC) and recommendations by the Federation of European Laboratory Animal Science Associations (FELASA). OMP-GFP mice (Potter et al. 2001) were housed in groups of both sexes (12:12h light-dark cycle; food and water available ad libitum). Experiments used young adults of either sex. We did not observe obvious gender-dependent differences.

Chemicals and solutions

The following solutions were used: (S₁) 4-(2-Hydroxyethyl)piperazine-1-ethanesulfonic acid (HEPES) buffered extracellular solution containing (in mM) 145 NaCl, 5 KCl, 1 CaCl₂, 1 MgCl₂, 10 HEPES; pH = 7.3 (adjusted with NaOH); osmolarity = 300 mOsm (adjusted with glucose). (S₂) HEPES-buffered calibration solution of different Cl⁻ concentrations containing (in mM) 140K⁺, 10 Na⁺, 10 HEPES, 10–80 Cl⁻, 70–140 gluconate; pH = 7.4 (adjusted with KOH); 310 mOsm (adjusted with K-gluconate). If not stated otherwise, chemicals were purchased from Sigma. Final solvent concentrations were ≤0.1%.

En face FLIM at the luminal surface of the VNO sensory epithelium

Preparation of intact vomeronasal sensory epithelia for en face imaging followed previously published protocols (Rivière et al. 2009; Cichy et al. 2015) with minor modifications. Briefly, the fur, lower jaw, and palate were removed and a sagittal hemisection was performed ~1mm lateral to midline. The hemisected head was transferred to a 35mm culture dish, embedded in agar (4%) and superfused continuously at room temperature. The cartilaginous capsule surrounding the VNO was opened laterally and the cavernous tissue and lateral blood vessel were gently removed to gain access to the luminal surface of the sensory epithelium. Next, the whole-mount preparation was incubated (30min; 37 °C) in physiological saline (S₁) containing 7mM 1-(ethoxycarbonylmethyl)-6-methoxyquinolinium bromide (MQAE; Sigma; Verkman 1990). After dye loading, the hemisected head—including the laterally opened VNO—was transferred into MQAE-free S₁ solution. The sample was then mounted on the stage of an upright scanning fluorescence microscope (A1 MP; Nikon Instruments Europe) equipped with a 25× objective (water immersion, numerical aperture 1.1; working distance 2mm; N25X-APO-MP; Nikon). Fluorescence was excited by 2-photon excitation (<100 fs light pulses; λ_exc = 750nm). Laser pulses were generated using a mode-locked Titan-Sapphire laser (MaiTai DeepSee; output power 2.3W at 750nm; Newport Spectra Physics) at a frequency of 80 MHz. Mean fluorescence lifetimes were measured using time-correlated single photon counting (TCSPC) in a volume of 0.08 µm³ per individual pixel. We obtained a 3-dimensional resolution of ~1 μm in the z axis and 0.3 µm in the x- and y-axes. The laser light was directed through the lens with reduced power of ~5 mW at the sample. MQAE fluorescence (λ_em,max: 460nm) was separated from tissue autofluorescence by a short pass filter (SP500, λ_obs < 510nm; Omega Optical) and recorded by a GaAsP hybrid photodetector (HPM-100–40; Becker & Hickl). TCSPC electronics and acquisition software (SPC-152; Becker & Hickl) were used for FLIM as previously described (Kaneko et al. 2004). SPCImage 4.8 software (Becker & Hickel) was used to fit bi-exponential functions to individual fluorescence decays (bin of 5). The average fluorescence lifetime was determined for each pixel from amplitudes and lifetimes of the 2 exponential functions as (a ₁ × τ₁ + a₂ × τ₂)/(a₁ + a₂). Experiments comparing [Cl⁻]_c in VSN knobs versus deeper dendritic compartments (Figure 2) were performed using older detection hardware (cooled photomultiplier PMC-100 and TCSPC-board SPC-730 [Becker & Hickl; details in Gilbert et al. 2007]). These data were analyzed by fitting bi-exponential functions to pixel fluorescence decays (bin of 0) using SPCImage 5.3 software (Becker & Hickel). Again, the average fluorescence lifetime was calculated for each pixel. Differences in both acquisition hardware and analysis software render the resulting fluorescence lifetimes quantitatively incomparable. Accordingly, results shown in Figure 2 cannot be translated into absolute [Cl⁻]_c values based on calibrations shown in Figure 1. Relative changes, however, provide reliable qualitative information on [Cl⁻]_c distributions along VSN apical dendrites.

Figure 2.

An apparent [Cl⁻]_c gradient along the apicobasal axes of VSN dendrites suggests active Cl⁻ accumulation in the knob layer. (A) Merged fluorescence intensity images (MQAE = red, EGFP = green) of the VSN knob layer (i) and—5 µm below—the dendritic / sustentacular cell layer (ii). VSN knobs (i) and dendrites (ii) become apparent as green/yellow dots, whereas sustentacular cells occupy relatively small (i) or large (ii) red areas. (B) MQAE fluorescence lifetime images corresponding to the VNO epithelial regions shown in A (lifetimes are depicted in pseudocolor; rainbow color map; 2.0–3.5 ns). (C) Histogram depicting the average fluorescence lifetime distribution of all VSN knobs shown in Ai (n = 140; green) and all VSN dendrites shown in Aii (n = 82; blue). Clustered values clearly identify 2 separate groups that are both well fit by Gaussian distributions (black curves). Inset shows the average fluorescence lifetime of all knobs (2.4±0.03 ns) and corresponding dendrites (2.9±0.03 ns). Data points are means ± SEM. Asterisk denotes statistical significance (Mann–Whitney rank sum test; P < 0.001). (D) Heat map of gene expression estimates from 30 single VSNs. Normalized RNAseq counts are represented in log₁₀ scale for canonical VSN markers (top) and 16 selected genes that encode Cl⁻ transporters (bottom). The VSNs are ordered by hierarchical clustering according to all data presented.

Figure 1.

Measuring [Cl⁻]_c within the apical layer of the mouse VNO sensory epithelium. (A and B) Image and schematic illustrating the en face multi-photon MQAE fluorescence lifetime imaging approach used for quantitative [Cl⁻]_c analysis in this study. (A) Merged macroscopic bright field and fluorescence images of the hemisected rostral head of an OMP-GFP mouse. The VNO capsule was opened laterally and the cavernous nonsensory tissue was removed to gain access to the surface of the sensory neuroepithelium. Note the bright green fluorescence of OMP-positive neurons in both the VNO and main olfactory epithelium. (B) Schematic drawing of the experimental setup used for en face FLIM. Advantages of this experimental approach include preservation of the vomeronasal mucus, structural integrity of the epithelium including its apical tight junction barrier, and intact axonal projections to the accessory olfactory bulb. Images in both (A) and (B) have been adapted and modified from Rivière et al. (2009). (C) Fluorescence intensity grayscale images of MQAE ([i] λ_exc = 750nm, λ_obs < 510nm) and EGFP ([ii]; λ_exc = 910nm, λ_obs > 500nm). The merged pseudocolor image ([iii]; MQAE = red, EGFP = green) identifies VSN dendritic knobs as round green/yellow structures, whereas sustentacular cells become apparent as smaller red spots. (D) En face MQAE fluorescence lifetime image of the VNO sensory surface (lifetimes are depicted in pseudocolor; rainbow color map; 1.2–3.1 ns). (E) Calibration curves showing the linear dependence of MQAE fluorescence lifetime and [Cl⁻]_c in both VSN knobs (K _SV = 7.1M⁻¹; R ² = 0.83; mean ± relative error) and sustentacular cells (K _SV = 7.5M⁻¹; R ² = 0.77; mean ± SD). (F) Histogram depicting the distribution of [Cl⁻]_c in individual VSN knobs shown in D. (G) Average [Cl⁻]_c of VNS knobs (42.1±1.1mM; 1563 neurons; 4 animals) and sustentacular cells (24.7±13.6mM; >240 cells; 6 experiments; 3 animals). Data points are means ± 95% confidence intervals. Apparent discrepancies in confidence interval magnitude result from different sample types (i.e., cells (VSN knobs) versus experiments (sustentacular cells); see Materials and methods). Asterisk denotes statistical significance (Mann–Whitney rank sum test; P = 0.045).

Cell identification

Individual VSN dendritic knobs were identified based on expression of enhanced green fluorescent protein (EGFP) under control of the olfactory marker protein (Omp) gene promoter (OMP-GFP; Potter et al. 2001). In these mice, all mature VSNs are labeled by intense EGFP fluorescence and, thus, individual VSN knobs are readily discernable from EGFP fluorescence intensity images (λ_exc = 910nm). Such images were routinely recorded before FLIM measurements were performed in identical fields of view. Each knob was defined as a region of interest (ROI) and the knob fluorescence lifetime was determined as the mean value of the average fluorescence lifetimes of all pixels in a given ROI. The lack of EGFP fluorescence in sustentacular cells renders them clearly distinguishable from VSNs, but—for the same reason—pinpointing individual sustentacular cells becomes challenging. Compared to VSNs, however, MQAE fluorescence intensities were uniformly increased in sustentacular cells. Thresholding of intensity images thus allowed a global analysis of most, if not all, sustentacular cells in a given FLIM image. Consequently, we obtained average [Cl⁻]_c values from 6 individual FLIM experiments, each analyzing between approximately 40 and 150 sustentacular cells. Image regions occupied by those few VSN knobs that showed unusually high fluorescence intensities were manually excluded.

Calibration of MQAE fluorescence lifetime and [Cl⁻]_c calculation

The fluorescence lifetime of MQAE in VSN dendritic knobs and sustentacular cells at different Cl⁻ concentrations was measured according to an established protocol (2-ionophore method; Krapf et al. 1988; Marandi et al. 2002; Kaneko et al. 2004). After FLIM recordings under control conditions (n = 2–3), samples were washed (4×) with S₂ containing defined chloride concentrations (10, 20, 60, and 80mM). Next, samples were washed (2×) with S₂ containing nigericin and tributyltin (10 µM each) to dissipate Cl⁻ gradients across the apical plasma membrane (Kovalchuk and Garaschuk 2012). MQAE is collisionally quenched by Cl⁻ resulting in a linear relationship between the inverse fluorescence lifetime and the Cl⁻ concentration:

$$\frac{{\tau }_0}{\tau }=1+K_{\text{SV}}{[{\text{Cl}}^{-}]}_{\text{int}}$$

with τ₀ and τ being the MQAE fluorescence lifetime in absence of Cl⁻ (i.e., [Cl⁻]_c = 0) and at a defined concentration (i.e., [Cl⁻]_c = 10, 20, 60, or 80mM), respectively. For each [Cl⁻]_c, inverse τ-values were multiplied with τ₁₀ and plotted. Next, τ₀ was determined from the intercept of a linear regression fit. When plotted in a Stern–Volmer representation (τ₀/τ as a function of [Cl⁻]_c), the calibrated fluorescence lifetimes result in a specific Stern–Volmer constant (K _SV) that is characteristic for the calibrated cell type. Accordingly, K _SV was used to calculate [Cl⁻]_c for both individual dendritic knobs and collective sustentacular cells. [Cl⁻]_c mean values were determined ±95% confidence intervals, which denote the range in which [Cl⁻]_c would be determined in repeated measurements performed under identical experimental conditions.

Single VSN RNA sequencing

VNOs were dissected and cells dissociated from 3 adult male heterozygous OMP-GFP mice (Potter et al. 2001). GFP-positive cells were enriched by fluorescence-activated cell sorting and collected on a C1 Single-Cell Auto Prep 96 well chip (Fluidigm). Next, cDNA libraries were prepared and sequenced as previously described (Saraiva et al. 2015). Across the 96 wells, 38 contained no cells, 2 contained cellular debris, and 56 contained single cells. The sequence fragments were aligned with STAR 2.3 (Dobin et al. 2013) to the GRCm38 mouse reference genome with some modifications (Saraiva et al. 2015). Notably, the vomeronasal receptor gene set was replaced with extended models (Ibarra-Soria et al. 2014). The fragments that uniquely aligned to each gene were counted using HTSeq (Anders et al. 2015). The DESeq2 package (Love et al. 2014) was used to normalize the data to account for variance in sequencing depth between cells. The 56 single cells were filtered through a conservative quality control process to remove those that did not express OMP-GFP and/or had a low proportion of uniquely mapped sequence data. This resulted in a final dataset of 30 single VSNs that had, on average, 3.6 million sequence reads of which 89.7% were uniquely mapped. The full sequence data are available from the European Nucleotide Archive (ENA) under study accession ERP004364.

Results

To determine [Cl⁻]_c in VSN dendritic knobs we performed en face 2-photon FLIM measurements from optical sections through the most apical layer of the VNO sensory neuroepithelium (Figure 1). The structurally intact epithelium was loaded with the Cl⁻-sensitive fluorescent reporter MQAE, a quinoline-based dye that interacts with Cl⁻ ions in its excited state and reduces both fluorescence intensity and fluorescence lifetime by collisional quenching (Verkman 1990; Kaneko et al. 2002; Kovalchuk and Garaschuk 2012). MQAE imaging has been used for [Cl⁻]_c analysis in acute brain slices (Marandi et al. 2002), frog taste disks (Li and Lindemann 2003), the main olfactory epithelium (Kaneko et al. 2004), cochlear outer hair cells (Song et al. 2005), neurons of dorsal root ganglia (Gilbert et al. 2007; Funk et al. 2008), and cockroach salivary glands (Hille et al. 2009). Using adult heterozygous OMP-GFP mice (Potter et al. 2001), we recorded both EGFP and MQAE fluorescence intensity (Figure 1C) as well as MQAE fluorescence lifetime (Figure 1D). As shown previously (Rivière et al. 2009), the neuron-specific EGFP signal illustrates a “mesh” of VSN dendritic knobs that appear as densely packed round spheres along the VNO sensory surface (Figure 1C-ii). The corresponding MQAE images show the same globose structures of varying intensity, clearly demarcated by narrow brighter regions that are devoid of EGFP fluorescence and thus correspond to VNO sustentacular cells (Figure 1C-i,C-iii).

For each experiment (n = 4), we determined every knob’s MQAE fluorescence lifetime (n = 1563) by defining individual ROIs based on the EGFP fluorescence intensity images (Figure 1C-ii) and by calculating the average lifetime of all pixels in a given ROI. To translate our findings into absolute steady-state [Cl⁻]_c values we next performed intracellular calibration of MQAE and calculated cell type-specific quenching constants (Figure 1E). Known quenching efficiency allows direct quantification of FLIM data irrespective of nonuniform dye loading or bleaching (Kovalchuk and Garaschuk 2012). Fluorescence lifetimes were recorded at defined [Cl⁻]_c. Incubation of corresponding calibration solutions in presence of the Cl⁻/OH antiporter tributyltin chloride (10 µM) and the K⁺/H⁺ ionophore nigericin (10 µM) dissipates Cl⁻ gradients across the plasma membrane and, thus, mediates Cl⁻ equilibration. When plotting the fluorescence lifetime ratio τ₀/τ as a function of [Cl⁻]_c (Stern–Volmer representation; Figure 1E), the slope of the resulting linear regression describes the cell type-specific Stern–Volmer constant (K _SV), which was 7.1M⁻¹ (R ² = 0.83) in VSN knobs and 7.5M⁻¹ (R ² = 0.77) in sustentacular cells, respectively. When plotted as a [Cl⁻]_c distribution histogram, individual knobs did not segregate into distinct groups. Rather, Cl⁻ levels were essentially normally distributed over a relatively wide range (Figure 1F). Together, calibration revealed average [Cl⁻]_c values of 42.1±1.1mM in VSN knobs and 24.7±13.6mM in sustentacular cells, respectively (±95% confidence intervals; Figure 1G). These results demonstrate a significant difference in [Cl⁻]_c between both VNO cell types and suggest distinct cytosolic Cl⁻ accumulation within the apical dendritic endings of vomeronasal neurons.

Prompted by both the previous report of a [Cl⁻]_c gradient in sensory neurons of the main olfactory epithelium (Kaneko et al. 2004) and the recent demonstration of increased resting [Cl⁻]_c in VSN cell bodies (Kim et al. 2015), we next determined dendritic [Cl⁻]_c values as a function of epithelial depth (Figure 2). Thus, we recorded MQAE fluorescence lifetime in the superficial epithelial layer (i.e., the knob layer; Figure 2A-i,B-i) versus a 4–7 µm deeper focal plane dominated by sustentacular cell bodies (Figure 2A-ii,B-ii). Both VSN knobs and apical dendrites are readily identified in both layers according to their EGFP fluorescence (Figure 2A). Qualitatively, we consistently observed an increase in neuronal MQAE fluorescence lifetime in deeper epithelial layers (n = 10 experiments; Figure 2B). When data from a representative experiment are plotted as an average fluorescence lifetime distribution histogram (Figure 2C), data points cluster as 2 distinct groups, depending on epithelial depth. While direct translation of these data into absolute [Cl⁻]_c values was not possible (see Materials and methods), average fluorescence lifetime significantly differed between VSN knobs and corresponding dendrites (Figure 2C, inset). The substantially shorter lifetime in the knob layer strongly suggests that VSNs establish a [Cl⁻]_c gradient along their apicobasal dendritic axes, indicative of active Cl⁻ accumulation in VSN knobs.

To determine the molecular identity of vomeronasal Cl⁻ transporters, we carried out single cell RNA sequencing (RNAseq) of fluorescent VSNs from adult heterozygous OMP-GFP mice. Cells were subjected to a capture, preparation, sequencing, and quality control process previously applied to olfactory neurons (Saraiva et al. 2015), resulting in transcriptome-wide expression estimates for 30 putative VSNs (as confirmed by canonical marker expression, i.e., Trpc2 and Omp; Figure 2D). While only 2 VSNs abundantly expressed a V2r gene in concert with Gnao1, the remaining 28 neurons co-expressed a V1r gene with Gnai2, indicating a bias toward “apical” VSNs during dissociation (Rivière et al. 2009). We next assessed the relative expression levels of those solute carrier (SLC) subfamily members that encode Cl⁻ transporters. Only Slc12a2 (encoding the Na–K–Cl cotransporter, Nkcc1) is present in all 30 VSNs (Figure 2D). Two others, Slc12a6 and Slc12a7 (encoding the K–Cl cotransporters, Kcc3 and Kcc4) are expressed in 57% and 83% of the VSNs, respectively. Nkcc1, Kcc3, and Kcc4 are thus prime candidates for generating [Cl⁻]_c gradients along VSN dendrites.

Discussion

Chemosensory transduction in vomeronasal neurons depends on phosphoinositide turnover, cytosolic Ca²⁺ elevations—via Ca²⁺-permeable plasma membrane channels and/or release from intracellular storage organelles—and, apparently, gating of a Ca²⁺-activated Cl⁻ channels. The physiological role of the latter is directly determined by the driving force for Cl⁻ at the apical VSN membrane and is, thus, a function of [Cl⁻]_c. By quantitative en face FLIM measurements of intracellular Cl⁻ levels in intact VSN dendritic knobs, we here report average [Cl⁻]_c values of 42.1±1.1mM at rest. If Cl⁻ concentrations in the vomeronasal mucus resemble typical interstitial levels (≤150mM) the observed Cl⁻ accumulation in VSN knobs will translate into E _Cl ≥ −34 mV, rendering vomeronasal anoctamins potent response amplifiers.

Our results complement and extent a recent report of increased resting [Cl⁻]_c values in VSN cell bodies recorded in acute VNO slices (Kim et al. 2015). In this study, technical challenges prevented the authors from tracking individual dendritic structures. In VSN somata, Kim et al. measure resting [Cl⁻]_c of ~85mM. Our own results indicate a lower, though compared to “regular” neuronal concentrations still markedly increased, average value. Several factors likely account for this discrepancy: 1) different recording sites (soma vs. knob); 2) different preparations (slices vs. intact whole-mount preparations); and 3) different methods of recording MQAE collisional quenching (fluorescence intensity vs. fluorescence lifetime). This last methodological difference might prove particularly relevant since FLIM measurements allow absolute quantification of [Cl⁻]_c independent of local cytosolic dye concentrations (Kaneko et al. 2004). Irrespective of the exact [Cl⁻]_c value, both this study and the work of Kim et al. (2015) consistently suggest a substantially elevated Cl⁻ level in VSNs that provides the electrochemical driving force necessary to “boost” sensory responses by a depolarizing Cl⁻ efflux, similar to the mechanisms implemented in canonical olfactory sensory neurons (Kleene 1993; Lowe and Gold 1993; Reuter et al. 1998; Reisert et al. 2003; Stephan et al. 2009).

Intriguingly, [Cl⁻]_c in the apical endings of VNO sustentacular cells is significantly lower than in VSN dendritic knobs (theoretical E _Cl in sustentacular cells: −48 mV), suggesting that cytosolic Cl⁻ accumulation is a specific property of neurons within the vomeronasal sensory epithelium. While the absence of anoctamin1 expression in the VNO microvillar layer after conditional knockout of Ano1 in OMP-positive mature sensory neurons indicates VSN-specific expression of this isoform (Amjad et al. 2015), other family members—for example, anoctamin2 (Rasche et al. 2010; Billig et al. 2011; Dauner et al. 2012)—could carry Ca²⁺-activated Cl⁻ currents in vomeronasal sustentacular cells. Therefore, members of the same ion channel family could exert opposite functions in the 2 main cell types of the VNO sensory epithelium.

The apparent [Cl⁻]_c gradient in VSN dendrites is reminiscent of active Cl⁻ accumulation in the dendritic endings of olfactory sensory neurons (Kaneko et al. 2004). Our transcriptome-wide single neuron analysis identified Slc12a2 (Nkcc1) as the Cl⁻ transporter gene, that is, 1) most prevalently expressed across VSNs and 2) most abundantly expressed within VSNs. This cation-coupled Cl⁻ cotransporter (Russell 2000) was previously detected in isolated VSNs by immunocytochemistry (Yang and Delay 2010). Nkcc1 is also abundantly expressed in the main olfactory epithelium and involved in olfactory neuron Cl⁻ accumulation (Kaneko et al. 2004; Reisert et al. 2005). In addition, Kcc3 and Kcc4 are also expressed in the majority of VSNs, indicating that either or both Cl⁻ cotransporter(s) could add to Cl⁻ accumulation in VSN knobs. In any case, the phenotypical similarity of Cl⁻-dependent response amplification provides another mechanistic signaling analogy between sensory neurons of the main and accessory olfactory system, including monoallelic receptor gene expression (Chess et al. 1994; Belluscio et al. 1999; Rodriguez et al. 1999), G protein-dependent transduction (Jones and Reed 1989; Chamero et al. 2011), or Ca²⁺-mediated response adaptation (Chen and Yau 1994; Spehr et al. 2009).

Together, using quantitative FLIM measurements at the apical surface of the intact vomeronasal neuroepithelium, we demonstrate differently regulated intracellular Cl⁻ concentrations in VSNs and sustentacular cells. VSNs appear to accumulate cytosolic Cl⁻ within the apical dendritic compartments, establishing a driving force for Cl⁻ ions that allows sensory response amplification by gating of Ca²⁺-activated Cl⁻ channels.

Funding

This work was funded by grants of the Volkswagen Foundation (I/83533), the Deutsche Forschungsgemeinschaft (SP724/9-1), and the Wellcome Trust (098051). M.S. is a Lichtenberg Professor of the Volkswagen Foundation.