Reagents and antibodies.

Primary antibodies used for immunocytochemistry included the following: rabbit anti-E-cadherin (24E10; #3195; Cell Signaling Technology, Danvers, MA), rabbit anti-occludin (#31721; Abcam, Cambridge, UK), mouse anti-vinculin (#11194; Abcam), and rabbit anti-zonula occludens protein 1 (ZO-1; N-term; #40-2300; Invitrogen, Basel, Switzerland). Secondary antibodies used for immunocytochemistry included the following: Alexa goat anti-mouse 488 (#A11029; Invitrogen), Alexa goat anti-rabbit 633 (#A21070; Invitrogen), donkey anti-rabbit 488 (#703-505-155; Jackson ImmunoResearch, West Grove, PA), and goat anti-mouse 549 (#115-506-068; Jackson ImmunoResearch). Dyes and chemicals used for cellular staining included the following: Alexa Fluor 488 Phalloidin (#A12379; Invitrogen) and nuclear stain Hoechst 34580 (#H21486; Invitrogen).

Primary antibodies used for Western blotting included the following: mouse anti-claudin-2 (CLDN2) antibody (#32-5600; Invitrogen), mouse anti-lamin A/C (LMNA) antibody (#612162; BD Biosciences, Allschwil, Switzerland), rabbit anti-NF-κB p65 (Ser276) antibody (#3034; Cell Signaling Technology), rabbit anti-IκB-α antibody (#9242; Cell Signaling Technology), and mouse anti-human GAPDH (monoclonal; #MCA4740; AbD Serotec, Oxford, UK). Secondary antibodies used for Western blotting and conjugated to horseradish peroxidase included the following: donkey anti-rabbit IgG (#NA934; GE Healthcare, Glattbrugg, Switzerland) and goat anti-mouse (#A8924; Sigma-Aldrich, Buchs, Switzerland).

Cell culture.

Caco-2 cells (LGC Promochem, Molsheim, Switzerland) and derived clones were cultured in 5% CO₂, 37°C, in DMEM (GlutaMAX; Gibco, Invitrogen), with an additional 100 U/ml penicillin and 100 μg/ml streptomycin (Sigma-Aldrich), geneticin-selective antibiotic, 400 μg/ml (G418 Sulfate; Gibco, Invitrogen), and 10% FCS (2-01F120-I; Amimed; BioConcept, Allschwil, Switzerland).

pH shift experiments were carried out in serum-free RPMI medium (1-41F24-I; Amimed; BioConcept), supplemented with 2 mM GlutaMAX (35050-038; Gibco, Invitrogen) and 20 mM HEPES. For pH adjustment of the RPMI medium, the appropriate quantities of NaOH or HCl were added, and the medium was left to equilibrate in the CO₂ incubator for at least 36 h before using it for culture.

OGR1 overexpression.

The OGR1 construct (National Center for Biotechnology Information reference sequence NM_003485) was obtained through PCR using the following oligonucleotide primers: human (h)OGR1 forward 1, 5′ GGA TCC ATG AGG AGT GTG GCC CCT TCA GGC C; hOGR1 reverse 1, 5′ GAA TTC TAG GCC AAC CTG CCC GTG GGG AAC C (initiation and stop codons are underlined).

The OGR1 cDNA was obtained from human genomic DNA (Clontech, Mountain View, CA) using PCR and cloned into the mammalian expression plasmid pcDNA3.1(+) using the BamHI and EcoRI sites. All constructs were verified by DNA sequencing (Microsynth, Balgach, Switzerland). Cells were transfected with the OGR1 expression construct in serum-free DMEM by using the FuGENE HD Transfection Reagent (Roche Diagnostics, Basel, Switzerland). Stable transfectants were selected by treatment with 400 μg/ml G-418 for 2–3 wk. Cells exhibiting intracellular calcium signaling upon acidification were selected for further expansion. Further selection was based on significant inositol phosphate (IP) formation at acidic pH and confirmation of robust OGR1 mRNA expression compared with Caco-2 cells harboring the empty vector (Caco-2 vector control).

Calcium flux.

Calcium flux and IP formation experiments were carried out in HBSS containing 130 mM NaCl, 0.9 mM NaH₂PO₄, 5.4 mM KCl, 0.8 mM MgSO₄, 1.0 mM CaCl₂, and 25 mM glucose, buffered with 20 mM HEPES (15630-056; Invitrogen). The pH of all solutions was adjusted using a calibrated pH meter (Metrohm, Herisau, Switzerland). All data presented in this paper are referenced to pH measured at room temperature. To obtain pH at 37°C, 0.15 pH unit should be subtracted for HEPES buffers in the range of pH 6.8–7.8, according to our calibration experiments.

Cells were seeded in uncoated, 384-well, clear-bottom, black-walled imaging plates (#3712; Corning, Lowell, MA). Seeding density was 30,000 cells/well, resulting in ∼90% confluent monolayers on the day of Ca²⁺ imaging. After 24–36 h, cell culture medium was removed, and cells were loaded for 1 h at 37°C in a non-CO₂ incubator, with 20 μl Calcium 4 dye (prepared according to the manufacturer's instructions; Molecular Devices, Sunnyvale, CA) in pH 7.8 HBSS buffer, containing 2.5 mM Probenecid (#P-8761; Sigma-Aldrich), to reduce the baseline signal. Plates were then transferred to the FLIPR (Molecular Devices) or FDSS7000 (Hamamatsu, Shizuoka Prefecture, Japan) fluorescence plate reader and Ca²⁺ responses measured using a charge-coupled device camera with excitation at 480 nM and emission at 540 nM. Before any stimulation, a baseline was recorded over 5 s (F_b). For pH shift, another 10 μl pH 7.8 HBSS buffer was added as control, followed by 20 μl appropriate pH agonist buffers, resulting in the desired final pH. Fluorescence was read over 1.5 min. In routine experiments, peak fluorescence (F_peak) was determined and normalized to baseline. Data are reported as ΔF/F = (F_peak − F_b)/F_b. In specific cases, a calibration step was added to obtain maximum and minimum fluorescence (F_max and F_min, respectively) readings for calculation of free intracellular Ca²⁺ concentration ([Ca²⁺]_i; see Fig. 1B). To this end, ionomycin and EDTA were added as indicated. [Ca²⁺]_i was calculated using the following equation, with K_d = 345 nM for the Calcium 4 dye: [Ca²⁺]_i = K_d (F − F_min)/(F_max − F) (48).

IP formation assay.

Cells were seeded in 48-well plates (#3548; Corning) at an approximate density of 100,000 cells/well and grown to near confluence. IP formation was measured as described previously (21), using a 20-min incubation period and 20 mM LiCl to block inositol monophosphatase.

RNA extraction and quantitative real-time PCR.

Total RNA was isolated using the RNeasy Mini Kit in the automated QIAcube following the manufacturer's recommendations (Qiagen, Hombrechtikon, Switzerland). For removal of residual DNA, a DNase treatment was performed, according to the manufacturer's instructions, for 15 min at room temperature as part of the QIAcube program. Determination of mRNA expression was performed by quantitative real-time PCR on a 7900HT real-time PCR system (Applied Biosystems, Foster City, CA), under the following cycling conditions: 20 s at 95°C, then 45 cycles of 95°C for 1 s, and 60°C for 20 s with the TaqMan Fast Universal Master Mix. Samples were analyzed as triplicates. Relative mRNA expression was determined the by the ΔΔCt Method (20), which calculates the quantity of the target sequences relative to the endogenous control β-actin or GAPDH and a reference sample.

TaqMan Gene Expression Assays (all from Applied Biosystems), used in this study, were Hs 00268858-S1 GPR68 (OGR1), Hs 00270999-S1 GPR4, Hs 00269247-S1 GPR65 T cell death-associated gene 8 (TDAG8), human GAPDH Vic minor groove binder and nonfluor (4326317E), and human β-actin Vic TAMRA (4310881E).

Quantification of phosphorylated ERK1/2.

ERK1/2 phosphorylation levels were determined by cell-based sandwich immunoassay and quantified by homogenous time-resolved fluorescence (Cellul'erk phospho-Erk1/2; Cisbio Bioassays, Codolet, France). Caco-2 OGR1 clones U1 and U18 and vector control cells were seeded in 96-well plates at a density of 25,000 cells/well and cultured overnight at 37°C/5% CO₂. Cells were starved for 5 h in pH 7.6 serum-free medium, followed by treatment with pH 6.8 pre-equilibrated-buffered medium for 10 or 30 min at 37°C. Levels of phosphorylated ERK1/2 were determined following the protocol provided by Cisbio Bioassays. Data were normalized to the signals obtained at nonstimulating pH (pH 7.8 = 100%).

Luciferase reporter assays.

We used a serum response element (SRE)-luciferase reporter construct (Clontech) to study the effect of proton activation in OGR1-overexpressing Caco-2 cells on the SRF activation pathway. The SRE reporter construct, used here, encodes a firefly luciferase gene under the control of a minimal cytomegalovirus (CMV) promoter and tandem repeats of the SRE.

Cells were seeded in 48-well plates (#3548; Corning) at a density of 120,000 cells/well. The next day, cells were cotransfected with 500 ng each of the SRE-luciferase reporter construct or the control reporter plasmid pTAL-luc (Clontech) at a ratio of 3 μl FuGENE HD (Roche Diagnostics)/1 μg DNA. To control for variations in transfection efficiency, 100 ng of the pRL Renilla luciferase reporter plasmid (pRL-CMV; Promega, Madison, WI) was cotransfected in each well. Eighteen hours after transfection, the cells were starved in serum-free pH 7.8 RPMI medium for at least 5 h and then treated for 3 h in pH 6.8 or 7.8 pre-equilibrated serum-free RPMI medium. Cells were harvested in 1× passive lysis buffer (Promega), and luciferase activities were measured by using a GloMax-Multi Detection System (Promega). The pH 7.8 control condition values were set to 100%. The pCMV empty vector and the pTAL luciferase reporter vector were used as negative controls. Triplicate wells were measured for each transfection, and experiments were performed at least three times.

Transepithelial electrical resistance and epithelial cell permeability (paracellular flux).

Cells were seeded on noncoated Transwell Corning clear polyester membrane filters in 12-well format, 0.4 μm pore size (#3460; Sigma-Aldrich) at 125,000 cells/well. Cells were allowed to grow and differentiate for 3 wk before assays were performed. Culture medium was replaced 3×/wk and daily after 14 days.

Transepithelial electrical resistance (TEER) was measured using an equilibrated Epithelial Voltohmmeter (World Precision Instruments, Berlin, Germany) with “chopstick” electrodes.

Epithelial permeability was quantified by measuring the transepithelial flux of FITC-labeled dextran (40 kDa; Sigma-Aldrich), following confirmation of monolayer integrity by TEER. Starvation in serum-free RPMI medium for 24 h at pH 7.6 preceded simulation by acidic pH. For acidification, the apical medium was replaced with serum-free, pre-equilibrated RPMI at pH 6.7, containing 0.2 mg/ml FITC-labeled dextran, and incubated at 37°C, 5% CO₂. After 5 h incubation, 90 μl medium samples from the basal chambers were collected and analyzed for FITC-dextran by fluorometry (EnVision Multilabel Reader; 2104-0010; PerkinElmer, Waltham, MA), excitation 485 nM, and emission 540 nM. Data were normalized by defining flux at pH 7.8 as 100%.

Cell impedance measurements by electric cell-substrate impedance sensing.

Impedance of Caco-2 monolayers was monitored using electric cell-substrate impedance-sensing (ECIS) instruments (Applied BioPhysics, Troy, NY). Experiments were conducted at 37°C and 5% CO₂ in a humidified incubator. Cells were seeded in 8- or 96-well plates with electrodes embedded at the bottom of each well (ECIS 8W, 10E or 10E+; ECIS 96W, 10E+). Wells were pretreated with 10 mM L-cysteine. Seeding density was 150,000–250,000 cells/350 μl for eight-well arrays and 150,000–170,000 cells/200 μl for 96-well arrays.

Monolayers were subjected to a pH shift, at the earliest, 3 days after confluence was reached; thus total time after plating was 6–12 days. Growth medium was replaced with starvation medium (pH 7.8 serum-free RPMI medium plus glutamate, 20 mM HEPES) for 4–12 h to keep OGR1 in an inactive state. The baseline resistance was measured for 1 h before treatment. Starvation medium was removed and replaced with pre-equilibrated serum-free RPMI at different pH values in the absence or presence of PMA, as indicated, and impedance was measured in real time at multiple frequencies at a constant current of 1 μA at 20-s intervals.

The ECIS Z Theta instrument (Applied BioPhysics) allows impedance to be split into two components: resistance and capacitance. To determine the most suitable frequency for our experimental setup, multiple frequency scans were performed. As observed for other epithelial cell cultures, maximal effects were observed between 2,000 and 8,000 Hz, and we chose a frequency of 4,000 Hz for our standard experiments.

Automated wound-healing/cell-migration assay with ECIS.

ECIS allows the induction of a defined lesion into a confluent cell monolayer using an electric-field pulse; the repopulation and functional reorganization of the wounded area can be monitored by noninvasive impedance measurements using parameters, as described above. Lesions were set by applying three alternating current voltage pulses of 5.0 V amplitude, 64,000 Hz frequency, 20-s duration, at 2-min intervals. Impedance, resistance, and capacitance were measured at multiple frequencies every 20 s for 24 h.

Immunocytochemistry and confocal laser-scanning microscopy.

Cells were seeded at a density of 150,000–250,000 cells/350 μl per well in eight-well plates (#161107; ibidi, Martinsried, Germany). Plating was done in parallel when seeding the ECIS impedance arrays. Medium and culture conditions, including starvation before the pH change, were as described for the ECIS arrays, unless otherwise stated. Treatment of the epithelial cell monolayers was typically at 7–10 days and at least 3 days after confluency was reached. After low pH treatment, monolayers were rinsed three times with PBS and fixed in 4% paraformaldehyde, permeabilized with 0.1% Triton X-100 (Mallinckrodt, Paris, KY), and blocked for 30 min in blocking solution (5% BSA and 5% goat serum in 0.1% Triton X-100). Cell monolayers were blocked overnight in blocking solution (5% BSA and 5% goat serum) in 0.1% Triton X-100 in PBS. Primary antibodies were diluted (2 mg/ml) in blocking solution and incubated with monolayers for 1 h at room temperature. After washing three times, cell monolayers were incubated with secondary antibodies for 1 h, followed by incubation for 5 min with phalloidin conjugated to Alexa Fluor 488, diluted 1:100 (Alexa Fluor 488 Phalloidin; A12379; Invitrogen), which targets polymerized actin. Nuclei were counterstained with Hoechst 34580. Slides were stored at 4°C in the dark until imaged. Confocal data sets were sampled on a Zeiss LSM 700 inverted microscope; image acquisition and postprocessing were performed with proprietary Zeiss software, Zen 2012. Three random z-stacks/well were acquired and analyzed for each experimental condition. Each experiment contained duplicate wells and was carried out at least three times independently. F-Actin was quantified by measuring fluorescence intensity in ImageJ (42).

Protein extraction and Western blot analysis.

Caco-2 cells were seeded at 250,000 cells/well in six-well Transwell plates (#3450; Sigma-Aldrich) and cultured as described for the paracellular flux assays. Following stimulation as indicated, monolayers were washed twice with ice-cold PBS and scraped directly into ice-cold radioimmunoprecipitation buffer (Thermo Fisher Scientific, Waltham, MA), supplemented with complete protease inhibitor cocktail (Roche Diagnostics). Samples were stored at −20°C until further processing. After thawing, material was centrifuged for 10 min at 13,800 g, and the protein content of the cell lysate supernatants was determined by the bicinchoninic acid protein assay (Thermo Fisher Scientific). Western blotting was performed according to standard procedures, as published (47). Protein bands were visualized using the ECL Plus detection kit from Amersham Biosciences (Arlington Heights, IL). Equal loading of the samples was demonstrated by reprobing membranes with a GAPDH antibody (AbD Serotec). The level of protein expression was quantified by OptiQuant Analysis Software (Microsoft XP). Details on the antibodies used for Western blotting are provided above.

Expression profiling by microarrays.

Caco-2 cells were grown on Transwell plates, as described above for protein extraction, and monolayer integrity was determined by TEER. For pH shift, medium was replaced in the apical compartment with pre-equilibrated serum-free RPMI medium at indicated pH. The control condition was pH 7.8. Medium in the basal compartment was also replaced and maintained at nonstimulating pH 7.8. Treatment time for microarray samples was 24 h. Cells were collected and RNA samples prepared with Ovation PicoSL WTA System V2 (#3312-12; NuGEN, San Carlos, CA). Reverse transcription was carried out from 100 ng total RNA using single primer isothermal amplification, and a linear isothermal strand-displacement DNA amplification process was used to prepare micrograms of amplified cDNA (Ovation PicoSL WTA System V2; #3312-12; NuGEN). cDNA quality and quantity were determined using NanoDrop ND 1000 and Bioanalyzer 2100. Fragmented and biotin-labeled cDNA targets were generated starting with 3.5 μg cDNA with the Encore Biotin Module (#4200-12; NuGEN).

Biotin-labeled cDNA samples were hybridized to the GeneChip Human Gene 1.1 ST Array Strip (P/N 901626; Affymetrix, Santa Clara, CA), following protocols provided by Affymetrix. Data were summarized on gene level using Robust Multi-array Average. For each experimental group (vector control and OGR1-overexpressing clone U1), data quality was assessed using the bioconductor/R package “arrayQualityMetrics” (17), and reproducibility between replicates was assessed using Pearson's correlation for all of the filtered expression values and hierarchical clustering. For all pairwise comparisons, differentially expressed genes were selected using fold change and P value, as determined using ANOVA (as implemented by the R package, Linear Models for Microarray Data) and F-test for the complete experimental design. The results were compared on Venn diagrams and analyzed in the context of existing biological knowledge using MetaCore software for pathway and Gene Ontology Enrichment.

Data were deposited at the Gene Expression Omnibus under GSE60294 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?token=mzqjsmgmjdmvvef&acc=GSE60294).

Statistical analysis.

Statistical analyses were performed using GraphPad Prism 5 (GraphPad Software, San Diego, CA). Data are presented as means ± SE or SD (ECIS), and statistical significance was determined using unpaired two-tailed Student's t-test. P < 0.05 was considered significant. Where indicated, one-way ANOVA was performed, followed by the Tukey post hoc test.

Caco-2 cells overexpressing OGR1.

Overexpression of the pH-sensing receptor OGR1 in human Caco-2 epithelial cells and isolation of cell clones were performed, as described in materials and methods. OGR1 is a G_q-coupled receptor and is known to stimulate IP formation and intracellular calcium signaling upon exposure to neutral or slightly acidic pH. The cell clones U1 and U18 were selected based on their robust IP formation at pH 6.8 (Fig. 1A). As a consequence of stimulated phosphoinositide turnover, calcium is released from intracellular stores, and this response can be measured using calcium-sensitive fluorescent dyes. In cells expressing OGR1, a shift in extracellular pH to slightly acidic values induced robust increases of intracellular-free calcium concentrations that peaked at >100 nM within 2 min (Fig. 1B). The EC₅₀ for calcium release observed in our experiments was at pH 7.2 (Fig. 1C), well in line with earlier results for OGR1 (21). A significantly weaker response was observed in the vector control cells, which may be attributable to the endogenous expression of low levels of OGR1 in Caco-2 cells. Kinetics and amplitudes of intracellular pH (pHi) changes in response to pH shift were similar in cells expressing OGR compared with vector control, indicating that OGR1 does not interfere significantly with regulation of pHi (data not shown).

To assess the level of OGR1 mRNA expression in cells, we performed RT-PCR. Clones U1 and U18 showed a several hundred-fold increased mRNA expression relative to vector control (Fig. 1D). We also analyzed expression of the related pH sensors GPR4 and TDAG8 and found unchanged levels when comparing control cells with OGR1 transfectants (data not shown). Transfection of Caco-2 cells with an OGR1 receptor construct bearing a C-terminal myc tag showed expression of the protein predominantly at the plasma membrane (data not shown), as observed before for human embryonic kidney cells and fibroblasts (21).

We next proceeded to test for the level of ERK phosphorylation and SRF-dependent transcription, as OGR1-dependent stimulation of these signaling events was reported in other cell types (50).

OGR1 overexpression leads to increased phosphorylation of ERK1/2 and increased SRF-dependent gene transcription upon acidification.

Exposure of Caco-2 cells overexpressing OGR1 to acidic pH led to a significant phosphorylation of ERK1 and -2 isoforms within minutes (Fig. 1E). This was followed by increased expression of the SRF-dependent luciferase reporter (Fig. 1F). In clones U1 and U18, the signal increased 4.9- and 7.4-fold (P < 0.05 and P < 0.01), respectively, compared with the vector control.

Decreased paracellular permeability at acidic pH in cells overexpressing OGR1.

To investigate the influence of acidic pH on epithelial permeability, the diffusion of fluorescently labeled dextran was measured across the Caco-2 monolayers. We observed a significant decrease in the permeability of the OGR1-overexpressing clone U1 compared with the vector control monolayer, 6 h following the shift of extracellular pH from pH 7.6 to 6.7 (Fig. 2A).

We carried out further experiments to monitor epithelial function using measurements of cellular impedance. This technology allows for noninvasive monitoring of epithelial function over time under several experimental conditions in parallel. Typical time traces of cellular impedance for clones Caco-2 U1 and U18 compared with vector control are shown in Fig. 2, B and C, respectively. For both clones, exposure to acidic pH led to increased impedance that developed slowly over time, peaked ∼3–6 h, and thereafter declined. To exclude possible artifacts due to cell cloning, we confirmed an impedance increase in response to acidification in nonclonal cell populations isolated directly following transfection. As expected, impedance changes measured at 6 h after pH shift were smaller but significant (OGR1 population pH 7.7: 44.8 ± 20.5; OGR1 population pH 6.8: 197.8 ± 36.3; means + SE; n = 4; P < 0.05). For vector control, there was no significant change.

PMA, a potent and reversible PKC activator, was used as a positive control. PMA was shown to activate PKC and increase TEER in Caco-2 cells (52). As expected, PMA, applied at pH 7.6, induced an increase in impedance in all cells, including the vector control, with similar kinetics and amplitude, as observed for the response to low pH in the OGR1-transfected clones.

Figure 2D shows normalized impedance responses at 5 h, compiled from several experiments carried out independently. The response of OGR1-transfected Caco-2 cells to acidic pH was highly significant. The small response to pH shift in vector control cells may be due to the presence of low levels of endogenous OGR1.

Inhibition of cell migration at acidic pH in cells overexpressing OGR1.

The effect of stably expressed OGR1 on cell migration was assessed by an automated wound-healing/cell-migration assay with ECIS. Wounds were introduced as described in materials and methods, following cell culture at acidic pH for 6 h. Impedance, resistance, and capacitance readings were recorded at multiple frequencies to monitor how cells reoccupy the electrode area in culture wells. Figure 3A shows the vector control, where no significant change in impedance was observed among the experimental conditions tested (pH 7.6, 7.0, and 6.8), and no difference in the migration rates after wounding could be measured. On the other hand, cells expressing OGR1 exhibited the pH-dependent impedance increase, as described before, and following wounding, the recovery of impedance was inhibited at pH 7.0 and 6.8 compared with pH 7.6, indicating reduced cell migration (Fig. 3B).

Increased F-actin stress fibers and focal adhesions but no effect on tight junctions in OGR1-transfected Caco-2 cells exposed to acidic pH.

To gain a better understanding of the role of OGR1 in regulating cell–cell and cell–matrix communication in Caco-2 monolayers, we examined the status of actin and several cell junction proteins by confocal microscopy. Exposure of Caco-2 cells expressing OGR1 to slightly acidic pH led to a striking increase of F-actin stress fiber formation, which was nearly absent in vector control cells (Fig. 4). These fibers are basolateral, as shown in the orthogonal view, constructed from a series of Z-stack images of monolayers (Fig. 4B).

In contrast to the findings for polymerized actin, no significant increase or redistribution of occludin or ZO-1 was obvious by confocal microscopy analysis after low pH treatment (Figs. 4 and 5, and data not shown).

We also analyzed the status of cell junction proteins using Western blotting, following our standard experimental conditions for TEER measurements. In contrast to the observations by confocal microscopy, OGR1-overexpressing cells showed a significant increase in both occludin and ZO-1 overall protein content vs. control cells, reflecting the long-term growth of cells under standard cell culture conditions (pH 7.4 and lower), where OGR1 is partially activated. The reasons for this discrepancy are not clear; it may, in part, be due to the fact that Western blotting measures the overall protein pool, whereas immunocytochemistry, as applied here, only detects proteins expressed at the plasma membrane. No difference in expression of either of the tight junction proteins at nonstimulating vs. acidic pH was observed following the relatively short culture time under differentiating pH conditions (Fig. 6, A and B).

Interestingly, there was a significant reduction in expression of the pore-forming protein CLDN2 in cells overexpressing OGR1, which appears nicely complementary to the increase in tight junction proteins described above. Again, no difference in expression levels was detected following 5–6 h after pH shift (Fig. 6E).

Vinculin has been reported to couple integrins to the actomyosin cytoskeleton of focal adhesion proteins to the ECM (26, 32). In agreement with the striking upregulation of F-actin fibers, Western blotting for vinculin revealed a slight increase in protein expression at pH 6.8 vs. 7.7 (Fig. 6C). This result is also in agreement with the microarray gene expression analysis reported below, where an expression fold change of 2.29 was shown for vinculin mRNA upon OGR1 activation by acidic pH (Table 1).

Because lamin A/C was reported to regulate SRF activity via actin polymerization (13), we were interested in evaluating the status of lamin A/C protein levels under our experimental conditions. As observed for ZO-1 and occludin, lamin A/C was increased significantly in cells overexpressing OGR1; however, no influence of the short-term acidic pH shift was observed (Fig. 6D).

Microarray analysis.

To obtain a more comprehensive overview of gene and pathway regulation by acidic pH in Caco-2 cells expressing OGR1, we performed a DNA microarray analysis.

We compared gene expression patterns of vector control cells with those of OGR1-overexpressing cells (clone U1), following 24 h treatment under differentiating pH conditions. Figure 7A shows a heatmap of expression patterns, demonstrating a significant response to acidosis. A volcano plot, indicating the most significant expression changes, is presented in Fig. 7B. Highly regulated genes include genes involved in cell adhesion (carcinoembryonic antigen-related cell adhesion molecule 1 and 5); integrin alpha 2 (ITGA2); laminin gamma 2; vasoconstriction (endothelin-1); inflammatory processes (CXCL2); prostaglandin-endoperoxide synthase 2 (PTGS2), also known as cyclooxygenase-2 (COX-2); nuclear receptor subfamily 4, group A, member 1 (NR4A1); and the WNT signaling pathway inhibitor Dickkopf-related protein 1. Many differentially regulated genes are SRF target genes involved in regulation of the cytoskeleton [Miano et al. (25) and personal communication with J. M. Miano, 2013] (Tables 1 and 2).

Of particular interest is the highly significant regulation of the urokinase-type plasminogen activator receptor (PLAUR or uPAR); this receptor binds the serine protease uPA and regulates localized ECM remodeling. PLAUR is a SRF target gene also involved in regulation of the cytoskeleton (25). We identified antibodies useful to detect the protein in Western blot, and we could indeed demonstrate strong upregulation in cells overexpressing OGR1 compared with vector-transfected controls (Fig. 6F).