Functional Overexpression of Vomeronasal Receptors Using a Herpes Simplex Virus Type 1 (HSV-1)-Derived Amplicon

In mice, social behaviors such as mating and aggression are mediated by pheromones and related chemosignals. The vomeronasal organ (VNO) detects olfactory information from other individuals by sensory neurons tuned to respond to specific chemical cues. Receptors expressed by vomeronasal neurons are implicated in selective detection of these cues. Nearly 400 receptor genes have been identified in the mouse VNO, but the tuning properties of individual receptors remain poorly understood, in part due to the lack of a robust heterologous expression system. Here we develop a herpes virus-based amplicon delivery system to overexpress three types of vomeronasal receptor genes and to characterize cell responses to their proposed ligands. Through Ca2+ imaging in native VNO cells we show that virus-induced overexpression of V1rj2, V2r1b or Fpr3 caused a pronounced increase of responsivity to sulfated steroids, MHC-binding peptide or the synthetic hexapeptide W-peptide, respectively. Other related ligands were not recognized by infected individual neurons, indicating a high degree of selectivity by the overexpressed receptor. Removal of G-protein signaling eliminates Ca2+ responses, indicating that the endogenous second messenger system is essential for observing receptor activation. Our results provide a novel expression system for vomeronasal receptors that should be useful for understanding the molecular logic of VNO ligand detection. Functional expression of vomeronasal receptors and their deorphanization provides an essential requirement for deciphering the neural mechanisms controlling behavior.


Introduction
In mice and other terrestrial vertebrates, social behavior and reproductive physiology are mediated by chemosignals that are recognized by the olfactory systems [1,2]. The accessory (or vomeronasal) olfactory system seems to specialize in social interactions such as those conveyed by pheromones, kairomones, and other socially relevant chemosignals. These molecular cues are detected in the vomeronasal organ (VNO) by sensory neurons that express three major families of G protein-coupled receptors (GPCRs), vomeronasal receptors (VRs) type 1 and 2 (V1Rs and V2Rs), and formyl peptide receptors (Fprs) [3][4][5][6]. These receptors have been implicated in recognizing a highly diverse collection of ligands and initiating transduction in vomeronasal sensory neurons (VSNs). However, only few receptors from a repertoire of nearly 400 functional genes have been paired with molecularly defined ligands. In part, this is a direct consequence of the lack of a suitable expression system to enable in vitro receptor analysis of ligand specificity. Gene transfer and functional expression of cloned vomeronasal receptors is a fundamental requirement for providing a causal link between receptor activation and ligand recognition and understanding the chemical receptive properties that, in turn, provide the foundation for neural coding of social odorants.
Identification of a few vomeronasal receptor-ligand pairings has been possible by using gene-targeted mice that express fluorescent markers in specific sensory neurons and, in some cases, through genetic deletion of receptors. Known receptor-ligand pairs include V1rj2 and V1rj3 paired with sulfated steroids [7], V1rb2 paired with 2-heptanone [8], V2rp5 paired with ESP1 peptide [9], and V2r1b and V2rf2 paired with 9-mer peptides [10,11]. Despite recent advances in genome editing, generation of knockin and knockout mice is still costly and time consuming, and in consequence impractical to cover the enormous vomeronasal receptor repertoire. As a rare exception, two V2Rs (V2rp1 and V2rp2, which also detect ESP peptides) and one Fpr (Fpr3, previously known as Fpr-rs1, detecting bacterial signal peptides) have been functionally expressed in heterologous cells [12][13][14][15]. To our knowledge, however, this approach was not effectively used with other VRs and Fprs, thus far. A rapid and uncomplicated expression system that covers any vomeronasal receptor type would provide an invaluable tool for understanding the logic of information processing in the VNO. A functional expression system for vomeronasal receptors requires first that the receptors are properly localized to the surface of the cells, and second that they couple efficiently with a second messenger system that generate a measurable response to ligand stimulation. We reasoned that native VSNs themselves would be the most competent cells for expressing, localizing, and coupling vomeronasal receptors. Further support for this assumptions comes from the existence of only two major signaling cascades in the VNO: one that involves the G-protein Gαi2 used by V1R-expressing VSNs that are localized in the apical VNO layer, and a second one that involves Gαo in basal, V2R-positive VSNs [16][17][18][19]. Fpr signaling also seems to follow similar segregation rules as determined by G-protein expression: Fpr-rs3, 4, 6 and 7 co-localize with Gαi2 whereas Fpr3 co-localizes with Gαo [20]. Here, we rely on the large number of native VSNs expressing either Gαi2 or Gαo to develop an expression system, driving the expression of selected members from each of the V1R, V2R or Fpr families by delivering them through a herpes virus amplicon vector and infecting VSNs in vitro. Infected individual cells are identified by bicistronic expression of marker green fluorescent protein (GFP), allowing responses to specific ligands to be tested in single cells. Ca 2+ imaging recordings show that ectopic expression of a single V1R, V2R or Fpr gene causes a significant increase in the number of responses to cognate ligands in VSNs overexpressing one of these receptors. Our results provide a simple and effective system for functional expression of three types of vomeronasal receptors that should be useful in the deorphanization effort of these important receptor families.  [21] and Gnao1/OMP-Cre conditional Gαo knockout [22][23][24] mice were used. Mice were 6-15 weeks old and were kept under standard light/dark cycle (12:12) with food and water ad libitum. Mice were euthanized by CO 2 asphyxiation followed by cervical dislocation as directed by the guidelines of the University of Saarland.

VSN primary culture
To obtain freshly dissociated VSNs, the mouse VNO epithelium was detached from the cartilage and minced in PBS at 4°C. The tissue was incubated (20 min at 37°C) in PBS supplemented with papain (0.22 U/ml) and DNase I (10 U/ml; Fermentas), gently washed and centrifuged (100 × g, 5 min) in Dulbecco's modified Eagle's medium (DMEM, Invitrogen) supplemented with 10% fetal bovine serum (FBS; Gibco). Dissociated cells were plated for 1 h on coverslips previously coated with concanavalin-A type V (0.5 mg/ml, overnight at 4°C; Sigma). Cells were either used immediately for imaging (Fig 1), or infected with virus and incubated at 37°C in FBS-supplemented DMEM medium for 20-24h before imaging.

Calcium imaging
Imaging was performed as described [22,24,25]. Briefly, cells were loaded with the Ca 2+ indicator fura-2/AM (5 μM; Invitrogen). Coverslips containing fura-2-loaded VSNs were placed in a laminar-flow chamber (Warner Instruments) and constantly perfused with extracellular Hank's balanced salt solution (HBSS) supplemented with 10mM Hepes (2-[4-(2-hydroxyethyl) piperazin-1-yl]ethanesulfonic acid). Calcium imaging on HSV-infected cells was performed 20-24 h post infection. Ratiometric (F340/F380) fura-2 Ca 2+ imaging was performed using an Olympus IX71 microscope equipped with a Hamamatsu camera. Image pairs were acquired at 0.25 Hz and analyzed using ImageJ (NIH). Chemostimuli were bath applied for 30 s. Image processing and data analysis, including region of interest (ROI) detection and signal analyses, were performed using ImageJ (NIH) and Origin8.6 (OriginLab). ROIs were selected manually and peak signals calculated from the temporal profiles of image ratio values. A response was defined as a stimulus-dependent deviation in fluorescence ratio that exceeds 1.5 the standard deviation of the mean of the baseline.

Single-cell RT-PCR
Single-cell cDNAs were reverse transcribed and amplified by a previously described procedure [26]. VSNs exhibiting a specific response profile identified in Ca 2+ imaging were picked using a glass capillary (~10 μm tip size) in 1 μl of Hepes-HBSS. Single cells were then transferred to a PCR tube containing 1μl of diethylpyrocarbonate (DEPC) treated water. Samples were immediately frozen on dry ice and kept at −80°C until use. Cells were lysed in 5μl of a NP-40 based lysis buffer by incubation at 65°C for 1min. Reverse transcription (RT) was performed with anchor-T primer (TATAGAATTCGCGGCCGCTCGCGA[ T]24) and Superscript III (Invitrogen) at 37°C for 90 min and at 50°C for 15 min. Reactions were stopped by heating at 70°C for 10 min. Next, a tailing reaction was performed using TdT, dATP and RNAseH (Roche) at 37°C for 20 min, followed by incubation at 65°C for 10 min. For analysis of gene expression, RTtailed products were first amplified by PCR reaction with EX-Taq HS DNA polymerase (Takara) and anchor-T primer by incubating at 95°C for 2 min, 37°C for 5 min, 72°C for 20 min, then 30 cycles of 95°C for 30 s, 67°C for 1 min, 72°C for 6 min plus 6 s extension for each cycle, then 72°C for 10 min. The second round of PCR was with the following primers: V1rj: CATCACTCCCAGTAAYTCTAAGTGGGC and GCTGGGMTCTCTTGTRRTGYCTGTAGAG; Gnai2: GAGCATGAAGCTGTTTGACAGC and CTCCTTGGTGTCTTTGCGC; Omp: GCACAGT TAGCAGGTTCAGCT and GGTTTGCAGTCCTGGCAGC; Gnao: CTCCACGAGGACGAAACCAC and GCCCCGGAGATTGTTGGCA. V1rj primers are designed to capture V1rj2 (Vmn1r89), V1rj3 (Vmn1r85), and Vmn1r86 genes. Total VNO mRNA was obtained from pooled C57Bl/6 adult mice (both genders) VNO tissue using PureLink RNA Mini Kit (Ambion) according to the manufacturer's instructions. Traces of genomic DNA were digested by incubation with 30U of DNase I (Fermentas).

Statistics
Independent Student's t-test was used for measuring the significance of difference between two independent distributions. Mann-Whitney test was used for comparative analysis of cell-to- cell response variability. Multiple groups were compared using either one-way or two-way analysis of variance (ANOVA) with Tukey's test as a post hoc comparison. Analysis was performed using Origin8.6 (OriginLab) software. Unless otherwise stated, results are presented as means ± SEM.

Receptor identification by Ca 2+ imaging and single-cell RT-PCR
Chemosignals present in urine and other secretions are detected by VSNs using a signal cascade that ultimately activates the ion channel Trpc2, causing robust intracellular Ca 2+ transients [3]. We used Ca 2+ imaging to identify the VRs responsible for detecting specific chemosignals. We first targeted VSNs that showed robust Ca 2+ responses to 1, 3, 5(10)-estratrien-3, 17β-diol disulfate (E1050) (Fig 1A and 1B), a urinary sulfated steroid previously shown to activate VSNs [7,31]. We used a dissociated VSN preparation [22,24,25] to have access to activated single neurons and capture the whole cell using a micro-capillary pipette (Fig 1B). Single cells were then lysed and used as substrate for reverse transcription (RT) and cDNA amplification, followed by PCR using primers for specific genes. Pooled VNO mRNA as well as randomly picked single cells were used as controls. We isolated five cells responding to E1050 (cells A to E), four of which yielded a PCR product that amplified Omp (Fig 1C), encoding an abundant cytosolic protein expressed in all mature VSNs [32]. In all cells, we also identified expression of Gnai2, which codes for G-protein subunit Gαi2 (Fig 1C), typically expressed together with V1R receptors in the apical layer. VSNs responding to E1050 have been previously shown to express V1Rs of the V1rj subfamily [7,31]. Thus, we used a subset of degenerate primers designed to capture genes of this family: V1rj2 (Vmn1r89), V1rj3 (Vmn1r85), and Vmn1r86. We identified V1rj transcripts (either V1rj2 or V1rj3) in four of the E1050-activated cells, but not in randomly picked cells (n/r) which were not activated by E1050 (Fig 1C). Cells were further screened for Gnao1 encoding Gαo-which is not present in V1R-expressing VSNs -and found no detectable expression. Thus, using a combination of Ca 2+ imaging and single cell RT-PCR we were able to identify specific vomeronasal receptors expressed in a given cell.

HSV-1 expression system in intact VSNs
Identifying cognate ligands for olfactory or vomeronasal receptors in vitro has been a major challenge [33], mostly because of the lack of a robust heterologous expression system. Among the different vomeronasal receptors, only two V2Rs (V2rp1 and V2rp2) and one Fpr (Fpr3, also known as Fpr-rs1) have been functionally expressed in HEK cells [12][13][14]. To determine whether heterologous expression of V1Rs is feasible, we cloned V1rj2 into a mammalian expression vector and transfected HEK cells. Further Ca 2+ imaging analysis revealed that the transfected cells-labeled with a GFP reporter-failed to show any response to a mix of sulfated steroids (E1050, E1100, E0893, E0588, and E1050, each at 100 μM) (Fig 2A), indicating that functional V1R expression was not effective in these cells. To circumvent this problem, we aimed at performing ectopic expression of vomeronasal receptors in their native cellular environment, i.e. in intact VSNs. This method has proven effective in the past by using transgenic mice [7]. However, production of genetically altered mice is laborious and time consuming. We reasoned that the use of a viral vector system in vitro could significantly simplify VR overexpression in native VSNs.
To test this possibility, we attempted to maintain acutely dissociated VSNs in vitro for a period of time sufficient to enable receptor expression, transport and localization to the plasma membrane. We were able to keep primary VSNs in culture for 24-48 h with almost intact cell viability using standard cell culture conditions (see Methods). Longer incubation periods resulted in increased cell death. Therefore, an efficient viral vector should enable fast and robust expression under these conditions. We then examined expression at different time points in HEK cells infected with a herpes simplex virus type 1 (HSV-1) amplicon vector. Amplicons are defective, helper-dependent, highly versatile gene transfer vectors in which, in contrast to other widely used viral vectors, the amplicon genome does not carry protein-encoding viral sequences. Consequently, they are safe for the host and non-toxic for the infected cells, yet with the ability to transduce cells at exceedingly high efficiencies. Furthermore, the complete absence of virus genes provides space to accommodate very large foreign DNA sequences, including complete genes [34]. HSV-1 is also specifically effective for infection of VSNs because, under natural conditions, it targets the mouse VNO as a route for neuroinvasion of the rodent central nervous system [35]. Thus, HSV-1 derived amplicons are adequate candidate vectors for fast and robust ectopic receptor expression in VSNs.
In HEK cells infected with a HSV-1 carrying GFP, we observed GFP fluorescence after 6 h (Fig 2B), with a peak in fluorescence intensity after 48 h (Fig 2C). Next, we infected acutely dissociated VSNs with the HSV-GFP virus and analyzed GFP fluorescence after 6, 24 and 48 h. In VSNs, we also observed GFP expression after 6 h following infection with little or no cytotoxic effects (Fig 2B). The relative number of cells expressing GFP increased sharply at 24 h (from 19% to 72%) and then stabilized between 24 and 48 h (Fig 2C). Maximum GFP fluorescence intensity values were observed after 24 h and decreased at 48 h (Fig 2C), perhaps as a result of reduced cell fitness. Therefore, in subsequent VSN experiments we used a 20-24 h interval for optimal infection efficiency.
We evaluated the integrity of the plasma membrane by loading the cells with the Ca 2+ indicator fura-2 and observed robust dye uptake after 30 min of incubation (Fig 2D). On average, we observed an infection rate of 23% of all dissociated VNO cells in culture, measured as the number of cells positive for GFP (Fig 2D). VNO dissociation yields a mixed culture that includes VSNs but also glial supporting cells and immature VSNs. To determine whether VSNs were indeed infected, we used dissociated VNOs taken from OMP-GFP mice [21], which express GFP in all mature VSNs. GFP was detectable in~80% of the cells in culture, thus unambiguously categorizing them as VSNs. Cells were then infected by HSV-1 carrying an mCherry expression cassette (Fig 2E). Among mCherry-positive cells, 76% were also positive for GFP ( Fig 2E). Hence, HSV-1 was capable of infecting both VSNs and non-VSN cells with no evident bias for either cell type.

Functional validation of V1R viral overexpression
Next, we asked whether V1Rs overexpressed in virus-infected VSNs are functionally responsive. To address this question, we cloned V1rj2 into an expression cassette of the HSV-1 amplicon (Fig 3A). To enable identification of infected cells, we included GFP in the expression cassette, inserting an internal ribosomal entry site (IRES) to produce a bicistronic transcript that results in the expression of V1R receptor and GFP as separate proteins in the same cells (Fig 3A). Ca 2+ imaging assays on cells infected with the V1rj2-IRES-GFP virus revealed a 7-fold increase in the number of responsive cells (p < 0.005) to a mix of sulfated steroids (E1100, E0893, E0588, and E1050, each at 100 μM) as compared to responsive cells infected with the vector expressing GFP but no V1rj2 (Fig 3B and 3E). Importantly, cells expressing V1rj2 did not show significant Ca 2+ responses to the related sulfated steroid E2734, major urinary proteins contained in the high molecular weight fraction of urine (HMW), or control buffer containing DMSO (p = 0.1-0.4) (Fig 3B, 3C and 3E). E2734 has been reported to show no activation of V1rj2-positive cells [31], and HMW was shown to activate V2R-positive VSNs [22]. Thus, the response profile of VSNs infected with V1rj2-encoding virus, at least within the scope of the ligands screened here, was relatively specific. We used a high (100 mM) K + solution as positive control to estimate the number of infected VSNs and their viability. Receptor-GFP and GFP-only infected cells showed similar responses to high K + solution (p = 0.12) indicating that ectopic expression of these receptors did not affect cell excitability and survival ( Fig  3E). Both high K + and E mix evoked similar calcium transient amplitudes in individual V1rj2virus infected VSNs (Fig 3C and 3D), indicating that cell-to-cell response variability to E mix was comparable to that observed for high K + .

V2Rs and Fprs functionally expressed in infected VSNs
Our results described thus far indicate that ectopic overexpression of a given V1R in VSNs via HSV-1 mediated gene transfer results in a greatly increased number of VSNs that respond to specific ligands for this receptor (Fig 3). On the basis of these results, we reasoned that overexpression of other types of vomeronasal receptors including V2Rs and Fprs should also lead to increased responsiveness to specific ligands recognized by these receptors. To determine whether this is the case, we cloned the receptor V2r1b (Vmn2r24) into the HSV-1 amplicon and monitored Ca 2+ responses to the class I MHC binding peptide SYFPEITHI. Previous studies in VNO tissue slices from V2r1b-IRES-tauGFP gene-targeted mice revealed that more than half of VSNs expressing this receptor are normally activated by SYFPEITHI [10,11]. Remarkably, in our V2r1b-GFP virus expression system, we observed a clear 6-fold increase in the number of cells responding to SYFPEITHI (10 −11 M) stimulation versus receptor-free GFP controls (p < 0.005) (Fig 4A and 4B). By contrast, other related ligands such as the mitochondria-derived formylated peptide f-MFFINTLTL (ND1) and ligands present in the HMW fraction of urine did not produce enhanced responsiveness (p = 0.37-0.65) (Fig 4B). These results strongly argue that our virus-based expression approach can not only be used for V1Rs but also for V2Rs.
Both ND1 and W-peptide are known to function as Fpr agonists [13]. Fprs are expressed in the VNO by a small subset of VSNs which do not seem to overlap with V1Ror V2R-expressing VSNs [20]. To assess whether virus-mediated overexpression could serve to functionally express Fprs, we infected VSNs with a HSV-1 amplicon encoding Fpr3. This receptor is specifically activated by W-peptide but not by ND1 [13]. Using Ca 2+ imaging we found that W-peptide activated a greatly increased number of Fpr3-GFP infected cells, by~20 fold (p < 0.001) compared to GFP controls (Fig 4D and 4E), whereas ND1 and HMW fraction failed to induce higher levels of activity (p = 0.83-0.63) (Fig 4E). Hence, HSV-1-mediated gene expression is also applicable to Fpr3.
Previous studies have shown that electrical and Ca 2+ responses in the VNO evoked by SYF-PEITHI and some formylated peptides require the participation of the G-protein subunit Gαo [22]. To gain insight into the signaling properties of the overexpressed V2r1b receptor and to obtain additional controls about the specificity of our expression system, we used VSNs obtained from the Gnao1/OMP-Cre conditional Gαo knockout mouse line [22][23][24]. When such VSNs were infected with V2r1b-GFP virus, Ca 2+ responses to SYFPEITHI were drastically reduced (2 out of 193 cells;~1%) ( Fig 4C), now reaching similar values (p = 0.96) as those evoked by HMW and ND1 (2 and 3 cells). This result indicates that the presence of Gαo is required for the generation of enhanced responsiveness to SYFPEITHI in VSNs after virusmediated V2r1b expression. Importantly, Gαo-deficient VSNs showed normal Ca 2+ transients to high K + solution, indicating intact viability and excitability (Fig 4C).

Discussion
We have developed a virus-based gene delivery procedure to construct an expression system for the investigation of receptor-ligand interactions in the VNO. By choosing a selected member from each of the three major vomeronasal receptor families-V1rj2, V2r1b and Fpr3 -we demonstrate here that virus-induced overexpression in native VNO cells is a useful approach to characterize and identify receptor-ligand pairs underlying molecular sensing in the VNO. In each case, receptor overexpression resulted in a markedly increased number of infected cells responding to specific ligands. These ligands have been linked to these receptors previously, either through heterologous expression [13,14] or through the use of genetically modified mouse lines [7,10,11]. Together with the direct expression system developed here, these results provide a firm foundation for understanding the detection capabilities of specific vomeronasal receptors.
Neither viral infection nor GFP alone were sufficient to elicit this gain in responsivity and even closely related ligands were not recognized by the overexpressed receptor, indicating a high specificity of our system. Because the infected cells expressed GFP, it was possible to identify individual positive neurons, allowing Ca 2+ responses to be tested in single cells (Figs 3 and  4). The ligand response in virally-transduced cells followed the pattern of transient intracellular Ca 2+ increases shown in untreated cells (Fig 1A and [7,10,22,25,36,37]), indicating that the endogenous second messenger system expressed by these neurons was most likely responsible for generating the response in the infected cells. This view is further strengthened by the observation that infection of VSNs lacking Gαo-a key transduction molecule in basal VSNs [22]with viruses encoding the V2r1b receptor failed to show an increase in the responses to SYF-PEITHI peptide (Fig 4C).
We observed a 6-to 20-fold gain in responsivity to specific ligands after infection, but the overall number of activated cells was relatively low (about 3-5%), even though as many as 76% of infected cells were neurons (Fig 2E). The reasons for this are not known, but the most parsimonious explanation would be limited access to the correct endogenous second messenger and trafficking systems for membrane expression. Only half of the neurons express a given G-protein, either Gαi2 or Gαo, and correct localization to the plasma membrane may require extra components specific for particular receptors or receptor families. This could be the case for at least certain V2Rs that seem to require calreticulin4 and/or other factors to display fully functional responsive features [11,12]. It is also conceivable that some vomeronasal receptors may form heterodimers or heteromultimers [10,38]. Thus, the possible interaction of the overexpressed receptor with an endogenous receptor may have an impact on the activation profile of infected cells. Further experiments overexpressing different receptor types simultaneously in the same cells will be needed in the future to address these questions.
Matching the immense repertoire of vomeronasal receptors to the even greater collection of putative stimuli (the vomeronasal ligandome) and systematic analysis of ligand-receptor interactions are fundamental steps towards understanding how chemical information is encoded by the olfactory systems. Future identification of ligand-receptor pairs will define chemical coding mechanisms during social communication, the neural circuits that are activated by specific receptors and their ligands, and their precise roles in the control of social behaviors.