Clonal Stabilization Reveals a DAO/3MST‐Expressing MDCK Subpopulation With Robust d‐Cysteine‐Mediated H₂S Production

ABSTRACT

Hydrogen sulfide (H₂S) functions as a signaling molecule and cytoprotectant. In many tissues, H₂S is produced from l‐cysteine, whereas the kidney also produces H₂S from d‐cysteine by the sequential activities of d‐amino acid oxidase (DAO) and 3‐mercaptopyruvate sulfurtransferase (3MST). The d‐cysteine pathway produces H₂S more efficiently than the l‐cysteine pathway, and d‐cysteine significantly reduces ischemia–reperfusion injury in the renal cortex compared with l‐cysteine. However, the specific renal cell types responsible for the d‐cysteine pathway have not been clearly identified. Here, we show that a subpopulation of MDCK (NBL‐2) cells, a renal epithelial cell line, expresses both DAO and 3MST. MDCK cells initially produced H₂S from d‐cysteine, but this activity progressively declined with passage. In contrast, clonal cells expressing DAO and 3MST, established from this subpopulation, maintained stable H₂S production. The clonal cells exhibited two distinct patterns of DAO expression: one showed increased DAO levels after confluence, whereas the other displayed high DAO expression even before reaching confluence. These findings suggest that the d‐cysteine pathway operates in epithelial cells and plays distinct roles depending on cellular state and dynamics. Moreover, the established clonal cells may provide insight into the role of d‐cysteine‐mediated H₂S production in the kidney.

A subpopulation of MDCK (NBL‐2) cells, an epithelial cell line derived from the normal kidney, expresses DAO and 3MST. The DAO/3MST‐expressing cells are progressively eliminated during passaging; however, clonal isolation enables the establishment of sublines that maintain stable H₂S production. Clonal cells co‐expressing DAO and 3MST therefore provide a useful experimental system for investigating the regulation and cellular significance of d‐cysteine‐mediated H₂S production.

1. Introduction

Hydrogen sulfide (H₂S) is an important physiological mediator with diverse biological functions. It facilitates the induction of hippocampal long‐term potentiation by enhancing n ‐methyl‐d‐aspartate (NMDA) receptor activity in neurons [1] and induces Ca²⁺ influx in astrocytes, a type of glial cell that surrounds synapses [2]. H₂S also relaxes vascular smooth muscle by activating K⁺ channels and regulates insulin secretion from pancreatic β‐cells [3, 4, 5]. In addition to its signaling functions, H₂S protects neurons from oxidative stress by enhancing glutathione synthesis, scavenging reactive oxygen species, and suppressing excessive increases in intracellular Ca²⁺ [6, 7, 8]. H₂S also protects cardiac muscle from ischemia–reperfusion injury by preserving mitochondrial function and reducing apoptosis [9].

In mammals, H₂S is enzymatically produced by cystathionine β‐synthase (CBS), cystathionine γ‐lyase (CSE), and 3‐mercaptopyruvate sulfurtransferase (3MST). CBS and CSE produce H₂S from l‐cysteine [10, 11, 12], whereas 3MST produces H₂S from 3‐mercaptopyruvate (3MP), which is generated by cysteine aminotransferase (CAT) from l‐cysteine and α‐ketoglutarate (α‐KG) [13, 14]. 3MST reacts with 3MP to form a persulfide intermediate, from which dithiols such as thioredoxin and dihydrolipoic acid accept a sulfur atom that is subsequently attacked by another thiol to release H₂S [15].

In addition to the l‐cysteine pathway, an alternative pathway produces H₂S from d‐cysteine [16]. In this pathway, d‐amino acid oxidase (DAO) catalyzes the conversion of d‐cysteine to 3MP, which then serves as a substrate for 3MST. DAO has particularly high catalytic activity toward d‐cysteine compared with other neutral and lipophilic d‐amino acids [17]. It has been reported that d‐cysteine is generated endogenously from l‐cysteine by serine racemase [18]. DAO and 3MST are localized in different intracellular compartments: DAO is localized in peroxisomes, whereas 3MST is mainly found in mitochondria [13] [19]. Crosstalk between these two organelles enables production of H₂S from d‐cysteine in vivo.

Although the l‐cysteine pathway is found in many tissues, the d‐cysteine pathway is more restricted and primarily localized in the cerebellum and kidney. The kidney produces more than ten times the amount of H₂S from d‐cysteine than the cerebellum, and d‐cysteine provides greater protection against ischemia–reperfusion injury in the renal cortex than l‐cysteine [16]. These findings implicate a critical role of the d‐cysteine pathway in the kidney. However, the renal cell types responsible for H₂S production from d‐cysteine remain unclear.

In this study, we address the cellular basis of the renal d‐cysteine pathway by investigating MDCK (NBL‐2) cells, an epithelial cell line derived from the normal kidney. We show that a subpopulation of MDCK cells expresses both DAO and 3MST. We further demonstrate that the DAO/3MST‐expressing cells are progressively eliminated during passaging, but clonal isolation enables the establishment of sublines that maintain stable H₂S production.

2. Materials and Methods

2.1. Materials

The 3MST inhibitor I3MT‐3 was obtained from Selleck (Houston, TX); the DAO inhibitor indole‐2‐carboxylate (I2CA), Triton X‐100, sodium deoxycholate, and l‐homocysteine were obtained from Tokyo Chemical Industry (Tokyo, Japan). Other reagents were purchased from Fujifilm Wako (Osaka, Japan) or Sigma‐Aldrich (St. Louis, MO).

2.2. Cell Culture

MDCK (NBL‐2) cells were obtained from the Japanese Collection of Research Bioresources Cell Bank (JCRB, catalog no. JCRB9029) at passage 57. These cells were grown in AccuDia Eagle's MEM ① (Shimadzu Diagnostics, Tokyo, Japan) supplemented with 10% fetal bovine serum, 2 mM l‐glutamine, and 0.18% sodium bicarbonate and cultured in a humidified atmosphere at 37°C under 5% CO₂. The medium was typically changed every 3 days until the cells reached confluence, after which it was changed daily. For single‐cell cloning, MDCK cells were detached using 0.25% trypsin, and the cell suspension was centrifuged at 200 × g for 3 min. The cell pellet was resuspended in medium and passed through a 40 μm nylon cell strainer (pluriSelect Life Science, Leipzig, Germany). A volume of 100 μL of the cell suspension (at a concentration of one cell per 200 μL) was then seeded into each well of a 96‐well plate. Colonies developed in 21 wells. Six were discarded due to multiple colonies forming in the same well, leaving 15 clones for subsequent experiments.

LLC‐PK1 cells were obtained from JCRB (catalog no. JCRB0060). The cells were grown in medium 199 (Merck, Rahway, NJ) supplemented with 3% fetal bovine serum, 100 U/mL penicillin, and 100 μg/mL streptomycin and cultured in a humidified atmosphere at 37°C under 5% CO₂.

2.3. Western Blot Analysis

Cultured cells were washed with phosphate‐buffered saline (PBS, Shimadzu Diagnostics) and lysed in ice‐cold RIPA buffer [50 mM Tris/HCl (pH 7.4), 150 mM NaCl, 1% Triton X‐100, 0.5% sodium deoxycholate, 0.1% SDS, 1 mM EDTA, and protease inhibitor cocktail (Complete, EDTA‐free; Roche, Basel, Switzerland)]. The lysate was centrifuged at 1000 × g for 10 min, and the supernatant was recovered. Protein concentrations were quantified using the BCA protein assay kit (Takara Bio, Shiga, Japan). Five μg of protein per sample were separated by 10% or 15% SDS‐PAGE (DRC, Tokyo, Japan) and transferred to a polyvinylidene difluoride membrane with a pore size of 0.45 μm (Merck). The membrane was blocked with 2% albumin (Fujifilm Wako) or 2% skim milk (Becton Dickinson, Franklin Lakes, NJ) in PBS/0.1% Tween‐20 for 2 h at RT and incubated with primary antibodies: anti‐DAO rabbit polyclonal antibody (1:1500; 13 273–1‐AP, Proteintech, Rosemont, IL), anti‐MPST antibody (1:2500; HPA001240, Sigma, St Louis, MO), anti‐Gamma cystathionase polyclonal antibody (1:4000; 12 217–1‐AP, Proteintech), anti‐CBS polyclonal antibody (1:8000; 14 787–1‐AP, Proteintech), anti‐E‐cadherin polyclonal antibody (1:20000; 20 874–1‐AP, Proteintech), anti‐AQP1 Polyclonal antibody (1:20000; 20 333–1‐AP, Proteintech), or anti‐β‐actin rabbit recombinant antibody (1:20000; 81 115–1‐RR, Proteintech) overnight at 4°C. After incubation with secondary antibodies conjugated to horseradish peroxidase (Anti‐Rabbit IgG, HRP‐Linked F (ab’) 2 Fragment, 1:5000; Cytiva, Tokyo, Japan), the binding of antibodies was detected by chemiluminescence using ImmunoStar Zeta (Fujifilm Wako). Blot images were acquired using ChemiDoc Imaging System (Bio‐Rad, Hercules, CA). The intensity of each band was quantified using ImageJ software. The molecular sizes of chemiluminescent bands were determined in relation to BlueEasy Prestained Protein Marker (Nippon genetics Europe, Düren, Germany) or Precision Plus Protein Dual Color Standards (Bio‐Rad).

2.4. Determination of H₂S‐Producing Activity

Sub‐confluent or post‐confluent cell monolayers cultured in 12‐well plates or 6 cm dishes were washed, scraped into PBS, and precipitated by centrifugation at 1000 × g for 3 min. The cell precipitates were resuspended in lysis buffer consisting of 100 mM potassium phosphate (pH 7.4), 1% Triton X‐100, 1 mM dithiothreitol (DTT), and protease inhibitor cocktail (Complete, EDTA‐free). Cell suspensions were disrupted by sonication for 5 s using a sonifier (Branson Model 450, 10% output control; Branson Ultrasonics, Danbury, CT) or by Potter‐type glass homogenization with a Teflon pestle (700 rpm, 20 strokes). Following cell disruption, samples were centrifuged at 1000 × g for 10 min, and the resulting supernatants were collected as cell lysates. Protein concentrations were quantified using the DC protein assay kit (Bio‐Rad). For enzyme reactions, 11 μL of substrate solution was added to 100 μL of cell lysate in a 15 mL centrifuge tube. Final substrate concentration was 10 mM d‐cysteine, 10 mM l‐cysteine, 10 mM l‐cysteine combined with 2.5 mM α‐KG, 50 μM 3MP, or 10 mM l‐homocysteine, unless otherwise specified in the corresponding figure. Tubes were sealed with Parafilm and incubated at 37°C. Reactions containing d‐cysteine or 3MP were incubated for 30 min, whereas reactions containing l‐cysteine or l‐cysteine combined with α‐KG were incubated for 90 min unless otherwise indicated. To examine the inhibitory effects on DAO‐ and 3MST‐dependent H₂S production, the DAO inhibitor (I2CA) or the 3MST inhibitor I3MT‐3 was added at the final concentrations indicated in the corresponding figures. Following incubation, two volumes of 1 M sodium citrate buffer (pH 6.0) were added to each reaction mixture. Samples were further incubated at 37°C for 10 min with shaking at 150 rpm on a rotary shaker NR‐3 (TAITEC, Saitama, Japan) to facilitate the release of H₂S from the aqueous phase. Two mL of headspace gas was analyzed for H₂S using a gas chromatograph GC‐2014 or GC‐14B (Shimadzu, Kyoto, Japan) equipped with a flame photometric detector.

2.5. Reverse Transcription‐Quantitative PCR (RT‐qPCR)

Total RNA was isolated from cells using the NucleoSpin RNA kit (Takara Bio), and RNA concentration was determined using a DU730 UV/Vis spectrophotometer (Beckman Coulter, Brea, CA). Reverse transcription was performed with the PrimeScript FAST RT reagent Kit with gDNA Eraser (Takara Bio), according to the manufacturer's instructions. Briefly, genomic DNA was removed by incubating 1 μg of total RNA in a 16 μL reaction at room temperature for 5 min. Reverse transcription was then carried out in a final volume of 20 μL at 37°C for 10 min, followed by enzyme inactivation at 85°C for 5 s. For negative controls, reverse transcriptase was omitted to confirm that PCR amplification originated from cDNA rather than contaminating genomic DNA. Quantitative PCR was performed using TB Green Premix Ex Taq II (Fast qPCR; Takara Bio) in a 10 μL reaction containing 0.33 μL of the reverse transcription reaction mixture. Amplification was conducted under the following conditions: initial denaturation at 95°C for 30 s, followed by 40 cycles of denaturation at 95°C for 5 s and annealing/extension at 60°C for 10 s. A melting curve analysis was performed at the end of the amplification process to verify amplification specificity. All reactions were performed in duplicate. Primer sequences for RT‐qPCR are listed in Table S1. To select reference genes, the expression stability of the candidate housekeeping genes hypoxanthine‐guanine phosphoribosyltransferase (HPRT1), TATA‐binding protein (TBP), glyceraldehyde‐3‐phosphate dehydrogenase (GAPDH), and β‐actin (ACTB) was evaluated using the geNorm analysis. HPRT1 and TBP were identified as the most stably expressed genes across experimental conditions, including MDCK cells and clones C19 and C20 under both sub‐confluent and post‐confluent culture conditions. Accordingly, the geometric mean of HPRT1 and TBP was used for normalization. Relative gene expression levels were calculated using the 2^‐ΔΔCt method. For TBP measurements in post‐confluent C19 and C20, the difference in threshold cycle (Ct) values between duplicate reactions was 2.2 and 3.3, respectively. Because one replicate showed a markedly deviating Ct value relative to the other, it was excluded from subsequent analysis. Statistical analyses were performed on ΔCt values.

2.6. Fluorescence Imaging of H₂S

To detect intracellular H₂S production in live cells, we used the cell membrane‐permeable H₂S‐sensitive fluorescent probe HSip‐1 DA (Dojindo Laboratories, Kumamoto, Japan). Cells were seeded at a density of 2 × 10⁴ cells per well in a 96‐well optical‐bottom plate (Thermo Scientific, Waltham, MA) pre‐coated with 0.01% collagen (Atelocell IAC‐30 collagen, Koken, Tokyo, Japan). The cells were cultured overnight to obtain sub‐confluent conditions or maintained for 7 days after reaching confluence to obtain post‐confluent cultures. For fluorescence imaging, the culture medium was removed and the cells were washed twice with serum‐free MEM. Serum‐free MEM containing 5 μM HSip‐1 DA (200 μL per well) was then added, and the cells were incubated at 37°C for 30 min in a 5% CO₂ incubator. After probe loading, the medium was removed and the cells were washed twice with Hank's balanced salt solution (HBSS). Cells were subsequently incubated with MEM containing d‐cysteine or l‐cysteine (0, 1, or 10 mM) or sodium sulfide (Na₂S; 0.2 mM) as an H₂S donor at 37°C for 20 min. Following incubation, the cells were washed twice with HBSS, and 200 μL of FluoroBright DMEM (Thermo Scientific) was added to each well for imaging. Fluorescence images were acquired using a BIOREVO BZ‐X710 fluorescence microscope (KEYENCE, Osaka, Japan) with excitation and emission wavelengths of 470 and 525 nm, respectively.

2.7. Statistical Analysis

Statistical significance was determined using the unpaired two‐tailed Student's t‐test or one‐way ANOVA followed by Tukey's test. Statistical analysis was performed using Statcel add‐in software for Excel (third edition, OMS, Tokyo). p < 0.05 was considered statistically significant.

3. Results

3.1. MDCK (NBL‐2) Cells Produce H₂S From d‐Cysteine by DAO and 3MST

We previously found that 3MST produces H₂S from 3MP, which is generated from d‐cysteine by DAO, and that d‐cysteine attenuates ischemia–reperfusion injury in the renal cortex more than l‐cysteine [16]. The renal cortex is largely composed of epithelial cells capable of metabolizing d‐amino acid [20]. Thus, the d‐cysteine pathway may operate in renal epithelial cells. To investigate this possibility, we examined DAO and 3MST expression in MDCK (NBL‐2) cells, an epithelial cell line derived from the normal kidney [21]. Western blot analysis revealed that DAO expression increased after the cells reached confluence, whereas 3MST was already present under sub‐confluent conditions and continued to increase until day 16 post‐confluence (Figure 1B,C).

FIGURE 1.

DAO and 3MST mediate H₂S production from d‐cysteine in MDCK (NBL‐2) cells. (A) Schematic pathway of H₂S production from d‐cysteine. I2CA and I3MT‐3 inhibit DAO and 3MST, respectively. (B and C) MDCK cells were harvested at six different time points from sub‐confluence to post‐confluence. The cells were lysed, and proteins were separated by SDS‐PAGE. DAO (B) and 3MST (C) expression levels were analyzed by Western blotting. β‐Actin was used as a loading control. The relative expression levels of DAO and 3MST were normalized to β‐Actin. Data are means ± s.e.m. (n = 5). *p < 0.05, **p < 0.01 (one‐way ANOVA with Tukey's post hoc test). (D) Cell lysates prepared from 12‐well plates were incubated with 10 mM d‐cysteine, 10 mM l‐cysteine, 10 mM l‐cysteine combined with 2.5 mM α‐KG, or 50 μM 3MP. The released H₂S was measured by gas chromatography. Data are means ± s.e.m. (n = 5). For d‐cysteine, three out of five experiments were below the detection limit (0.2 μM) at sub‐confluence; the mean was calculated from the remaining two experiments. For l‐cysteine, three out of five experiments were below the detection limit at sub‐confluence and on day 0 post‐confluence, and two out of five at days 4, 8, 12, and 16 post‐confluence; means were calculated using only detectable values. **p < 0.01 vs. day 0 post‐confluence, #p < 0.05 vs. sub‐confluence, ##p < 0.01 vs. days 4 and 8 post‐confluence (one‐way ANOVA with Tukey's post hoc test). (E and F) On day 7 post‐confluence, cell lysates were incubated with 10 mM d‐cysteine or 50 μM 3mp in the presence or absence of the DAO inhibitor I2CA (E) or the 3MST inhibitor I3MT‐3 (F). H₂S production was measured by gas chromatography. Data are means ± s.e.m. (n = 5). **p < 0.01, ##p < 0.01 vs. control (one‐way ANOVA with Tukey's post hoc test).

We next examined the H₂S‐producing activity of MDCK cell lysates. H₂S production from d‐cysteine increased after confluence and reached a maximum at day 16 post‐confluence, whereas production from l‐cysteine or from l‐cysteine combined with α‐KG remained much lower and relatively constant throughout the culture period (Figure 1D). To determine whether H₂S production from d‐cysteine was mediated by DAO and 3MST, we tested the effects of enzyme inhibitors. The DAO inhibitor I2CA suppressed H₂S production from d‐cysteine but not from 3MP (Figure 1E). The 3MST inhibitor I3MT‐3 reduced H₂S production from both d‐cysteine and 3MP in a dose‐dependent manner (Figure 1F). These results indicate that MDCK cells produce H₂S from d‐cysteine by the sequential activities of DAO and 3MST.

3.2. Clonal Cells Expressing DAO and 3MST Maintain Stable H₂S Production From d‐Cysteine

When we examined H₂S‐producing activities of MDCK cells, we found that H₂S production from d‐cysteine progressively declined with passage, whereas production from 3MP remained relatively constant (Figure 2A). Accordingly, DAO expression decreased in a passage‐dependent manner, showing only a faint signal at passage 32, whereas 3MST expression remained stable (Figure 2B).

FIGURE 2.

Clonal cells expressing DAO and 3MST established from MDCK (NBL‐2) cells maintain H₂S production from d‐cysteine. (A) MDCK cells from different passages were harvested at day 7 post‐confluence in 12‐well plates. Lysates were incubated with 10 mM d‐cysteine, 10 mM l‐cysteine, 10 mM l‐cysteine combined with 2.5 mM α‐KG, or 50 μM 3MP, and released H₂S was measured by gas chromatography. Data are means ± s.e.m. (n = 5). **p < 0.01 vs. passage 8; n.s., not significant (one‐way ANOVA with Tukey's post hoc test). For d‐cysteine at passage 32 and l‐cysteine at 8, 24, and 32 passages, three out of five experiments were below the detection limit (0.2 μM) at sub‐confluence; means were calculated using only detectable values. (B) Western blotting for DAO and 3MST in MDCK cells at day 7 post‐confluence across different passages. β‐Actin was used as a loading control. (C) Western blotting for DAO and 3MST in 15 clonal cells established from MDCK cells. The cells were harvested at day 7 post‐confluence. β‐Actin was used as a loading control. (D and F) Lysates of clones C19 (D) and C20 (F) from different passages were prepared at day 7 post‐confluence in 12‐well plates and incubated with 10 mM d‐cysteine, 10 mM l‐cysteine, 10 mM l‐cysteine combined with 2.5 mM α‐KG, or 50 μM 3MP. The released H₂S was measured by gas chromatography. Data are means ± s.e.m. (n = 5). For l‐cysteine, all five experiments were below the detection limit. (E and G) Western blotting for DAO and 3MST in clone C19 (E) and C20 (G) at day 7 post‐confluence across passages. β‐Actin was used as a loading control.

MDCK (NBL‐2) cells are known to consist of subpopulations with distinct phenotypes [22, 23]. To determine whether the decline in H₂S production from d‐cysteine was due to cellular heterogeneity, we analyzed DAO and 3MST expression in clonal cells. Fifteen clones were established according to the order in which single‐cell‐derived colonies reached confluence. DAO and 3MST were detected only in two clones, C19 and C20, which reached confluence most slowly (Figure 2C). These two clones were then monitored over multiple passages, during which they maintained stable H₂S production from d‐cysteine through passage 32 (Figure 2D,F), with both DAO and 3MST expression remaining detectable (Figure 2E,G). Taken together, these findings suggest that the DAO/3MST‐expressing cells represent a subpopulation within MDCK cells, and that this subpopulation is eliminated during passaging.

3.3. Expression of d‐Cysteine‐Dependent H₂S‐Producing Enzymes in MDCK Cells and Derived Clonal Cell Lines

To assess clone‐dependent differences in enzymes involved in d‐cysteine‐dependent H₂S production, the expression of DAO and 3MST was examined in parental MDCK cells and clones C19 and C20 at the protein and mRNA levels under sub‐confluent and post‐confluent conditions. At the protein level, DAO expression under sub‐confluent conditions was significantly higher in C19 than in MDCK cells and C20 (Figure 3A). Under post‐confluent conditions, DAO protein levels were increased in C19 and C20 relative to MDCK cells. Consistent with these findings, DAO mRNA levels were higher in C19 than in MDCK cells and C20 under sub‐confluent conditions and were significantly increased in both clones compared with MDCK cells under post‐confluent conditions (Figure 3C). In contrast to DAO, 3MST protein levels under sub‐confluent conditions were comparable among MDCK cells and clones C19 and C20 (Figure 3B). Under post‐confluent conditions, however, MDCK cells exhibited significantly higher 3MST protein levels than either clonal cell line (Figure 3B). A similar pattern was observed at the mRNA level, with higher 3MST expression in MDCK cells than in C19 and C20 under post‐confluent conditions (Figure 3D).

FIGURE 3.

Expression of d‐cysteine‐dependent H₂S‐producing enzymes in MDCK (NBL‐2) cells and derived clones. (A and B) Protein levels of DAO (A) and 3MST (B) in MDCK cells and clones C19 and C20 at sub‐confluence and at day 7 post‐confluence were analyzed by Western blotting. β‐Actin was used as a loading control, and protein levels were normalized to β‐Actin. The values on the vertical axis indicate the relative expression levels compared to MDCK cells under sub‐confluent conditions. Data are means ± s.e.m. (n = 5). **p < 0.01, n.s., not significant (one‐way ANOVA with Tukey's post hoc test). (C and D) mRNA levels of DAO (C) and 3MST (D) in MDCK cells, C19, and C20 at sub‐confluence and at day 7 post‐confluence were analyzed by RT‐qPCR. Expression levels were normalized to the mean of HPRT1 and TBP as reference genes. Data are means ± s.e.m. (n = 3). *p < 0.05, **p < 0.01 (one‐way ANOVA with Tukey's post hoc test).

3.4. Expression of l‐Cysteine‐Dependent H₂S‐Producing Enzymes in MDCK Cells and Derived Clonal Cell Lines

MDCK cells are known to express enzymes involved in H₂S production from l‐cysteine [24], thus the expression of CSE and CBS was examined at the protein and mRNA levels in parental MDCK cells and the two clonal cell lines. At the protein level, CSE expression showed no significant differences among MDCK cells and clones C19 and C20 under either sub‐confluent or post‐confluent conditions (Figure 4A). CBS expression was comparable among the three cell types under sub‐confluent conditions; however, under post‐confluent conditions, it was significantly higher in MDCK cells than in either of the two clonal cell lines (Figure 4B). Based on the observation that CSE protein levels showed a decreasing pattern whereas CBS protein levels showed an increasing pattern under post‐confluent conditions across all cell types, mRNA expression was analyzed by comparing sub‐confluent and post‐confluent states within each cell type. CSE mRNA levels were significantly decreased under post‐confluent conditions compared with sub‐confluent conditions in all cell types examined (Figure 4C). CBS mRNA levels were increased under post‐confluent conditions in MDCK cells and C20, whereas no significant difference was observed between sub‐confluent and post‐confluent conditions in C19 (Figure 4D).

FIGURE 4.

Expression of l‐cysteine‐dependent H₂S‐producing enzymes in MDCK (NBL‐2) cells and derived clones. (A and B) Protein levels of CSE (A) and CBS (B) in MDCK cells and clones C19 and C20 at sub‐confluence and at day 7 post‐confluence were analyzed by Western blotting. β‐Actin was used as a loading control, and protein levels were normalized to β‐Actin. The values on the vertical axis indicate the relative expression levels compared to MDCK cells under sub‐confluent conditions. Data are means ± s.e.m. (n = 5). *p < 0.05, **p < 0.01, n.s., not significant (one‐way ANOVA with Tukey's post hoc test). (C and D) mRNA levels of CSE (C) and CBS (D) in MDCK cells, C19, and C20 at sub‐confluence and at day 7 post‐confluence were analyzed by RT‐qPCR. Expression levels were normalized to the mean of HPRT1 and TBP as reference genes. Data are means ± s.e.m. (n = 3). *p < 0.05, **p < 0.01.

3.5. l‐Cysteine‐Dependent H₂S‐Producing Activities in MDCK Cells

Although MDCK cells expressed CSE and CBS (Figure 4), little to no H₂S production was detected in cell lysates when l‐cysteine was supplied alone, even after incubation of 90 min (Figure 1D). The H₂S production assays were conducted using lysates with relatively low total protein concentrations (~0.5 mg protein/mL), conditions optimized for measuring the highly active d‐cysteine pathway. To determine whether the lack of detectable H₂S production resulted from dilution effects rather than an absence of enzymatic activity, experiments were performed using lysates with increased protein concentrations. Under conditions in which the protein concentration was increased fivefold relative to the initial assays, H₂S production from l‐cysteine alone became detectable at 15–30 min of incubation (Figure 5C). We also examined whether methodological factors influenced enzyme activity. Triton X‐100 has been reported to inhibit activities of certain enzymes, including cytochrome c oxidase [25], and ultrasonic disruption may also affect enzyme activity during cell extraction. When lysates were prepared in buffer containing Triton X‐100, sonication and Potter homogenization yielded comparable H₂S production from l‐cysteine (Figure 5A). Triton X‐100 did not inhibit H₂S production; rather, lysates with detergent produced higher H₂S levels from l‐cysteine than those prepared without detergent. These results indicate that the lack of H₂S production observed under the initial conditions was unlikely to result from the extraction procedure. Because CSE and CBS can catalyze reactions involving l‐cysteine and l‐homocysteine to produce H₂S [11, 12], we further examined the effect of l‐homocysteine. In its presence, l‐cysteine induced H₂S production in a dose‐dependent manner (Figure 5B). When both substrates were present, higher levels of H₂S than with l‐cysteine alone were detectable as early as 15 min after initiation of the reaction (Figure 5C).

FIGURE 5.

MDCK (NBL‐2) cell lysates produce H₂S from l‐cysteine. (A) Lysates of MDCK cells at day 7 post‐confluence, cultured in 6 cm dishes, were prepared using lysis buffer with or without Triton X‐100 and disrupted by either sonication or Potter‐type glass homogenization. Lysates were incubated with 10 mM l‐cysteine or 10 mM l‐cysteine combined with 2.5 mM α‐KG for 90 min, or with 10 mM d‐cysteine or 50 μM 3MP for 30 min. The released H₂S was measured by gas chromatography. Data are means ± s.e.m. (n = 3). *p < 0.05; n.s., not significant (one‐way ANOVA with Tukey's post hoc test). (B) Lysates of MDCK cells at day 7 post‐confluence were prepared using lysis buffer containing TritonX‐100 by sonication and were incubated with 10 mM l‐homocysteine combined with l‐cysteine for 90 min. The released H₂S was measured by gas chromatography. Data are means ± s.e.m. (n = 4). *p < 0.05, **p < 0.01 (one‐way ANOVA with Tukey's post hoc test). (C) Lysates prepared using lysis buffer containing TritonX‐100 by sonication were incubated with 10 mM l‐cysteine, 10 mM l‐homocysteine, or their combination. H₂S was measured by gas chromatography. Data are means ± s.e.m. (n = 3). TrX, Triton X‐100; l‐Hcys, l‐homocysteine.

3.6. Clonal Cells Expressing DAO and 3MST Retain Characteristics of Epithelial Cells

MDCK cells displayed heterogeneous morphologies, including cells with clearly visible nuclei, cells with small upper surface areas, polymorphic cells, and cells with poorly defined borders, which appeared to form clusters within the monolayer (Figure 6A). In contrast, clone C19 formed a uniform monolayer characterized by well‐defined intercellular spaces, polygonal cells with relatively small surface areas, and blister‐like structures. Clone C20 also formed monolayers, consisting of cells with larger surface areas and identifiable borders, although blister‐like structures were observed less frequently than in clone C19.

FIGURE 6.

Morphological and molecular characteristics of clonal cells expressing DAO and 3MST established from MDCK (NBL‐2) cells. (A) Phase‐contrast images of MDCK cells and clonal cells. The cells were analyzed at passage 10, whereas clones C19 and C20 were analyzed at passage 15 after isolation. (B and C) E‐cadherin (B) and AQP1 (C) expression levels in MDCK cells and clones C19 and C20 at sub‐confluence and at day 7 post‐confluence were analyzed by Western blotting. LLC‐PK1 cells were included as a control for proximal tubule cells. β‐Actin was used as a loading control. The relative expression levels of E‐cadherin and AQP1 were normalized to β‐Actin. Data are means ± s.e.m. (n = 5). *p < 0.05, **p < 0.01 (one‐way ANOVA with Tukey's post hoc test).

To further characterize the phenotype of C19 and C20, we analyzed the expression of epithelial marker proteins by Western blotting. Both clones, as well as parental MDCK cells, expressed E‐cadherin, a marker of epithelial cell–cell adhesion that is commonly detected in MDCK cells (Figure 6B) [26, 27]. Conversely, the proximal tubule marker aquaporin‐1 (AQP1) was absent in both clones and MDCK cells but was present in LLC‐PK1 cells, a porcine proximal tubule cell line (Figure 6C). These results suggest that C19 and C20 retain epithelial characteristics similar to those of MDCK cells rather than exhibiting features of proximal tubule cells.

3.7. Live‐Cell Imaging of H₂S Production in Clonal Cells

To evaluate H₂S‐producing activity via the d‐cysteine pathway in clones C19 and C20, we performed live‐cell imaging using the intracellular H₂S probe HSip‐1 DA. Under sub‐confluent conditions, treatment with d‐cysteine did not induce obvious fluorescence in C20 compared with the 0 mM control, whereas C19 showed detectable fluorescence following treatment with 10 mM d‐cysteine (Figure 7A). Under post‐confluent conditions, C19 predominantly displayed rounded fluorescence in the presence of d‐cysteine (Figure 7B). Similar rounded fluorescence patterns were also observed after treatment with sodium sulfide (Na₂S), an H₂S donor, as well as with l‐cysteine. In C20, treatment with 10 mM d‐cysteine induced only weak fluorescence, which was limited to a subset of cells attached to the culture plate.

FIGURE 7.

Live‐cell imaging of H₂S‐associated fluorescence in clonal cells derived from MDCK (NBL‐2) cells. (A and B) Fluorescence and phase‐contrast images of MDCK cells and clones C19 and C20 at sub‐confluence (A) and at day 7 post‐confluence (B). MDCK cells were analyzed at passage 8, and C19 and C20 were analyzed at passage 9–16 after isolation. Cells were loaded with the membrane‐permeable intracellular H₂S probe HSip‐1 DA (5 μM) for 30 min and subsequently incubated with d‐cysteine or l‐cysteine (0, 1, or 10 mM) or with sodium sulfide (Na₂S: 0.2 mM) as an H₂S donor. After 20 min of incubation, fluorescence images were acquired using 470/525 nm excitation/emission settings together with corresponding phase‐contrast images. Images were obtained at 20 × objective magnification. Arrowheads indicate rounded fluorescence in C19 under post‐confluent conditions.

4. Discussion

Many different strains have been derived from MDCK (NBL‐2) cells, which are not a single‐cell population but consist of a mixture of functionally distinct subpopulations [23]. In this study, we focused on this parental cell line and identified clonal cells expressing DAO and 3MST as one such subpopulation, distinguished by a slower proliferation to confluence compared with other clonal cells (Figure 2C). H₂S production from d‐cysteine in MDCK cells declined with passage (Figure 2A), but clones C19 and C20 maintained stable production (Figure 2D,F). These observations, together with previous findings that phenotypic traits of MDCK cells containing heterogeneous subpopulations can shift upon repeated culture [22] [28, 29], indicate that passaging may preferentially select for DAO‐negative cells, leading to the loss of the ability to produce H₂S from d‐cysteine. Cultured cells are generally maintained under atmospheric oxygen levels (~21%), which markedly exceed the physiological oxygen tensions in the kidney [30]. Although it was previously assumed that mammalian cells contained only l‐amino acids, recent studies have demonstrated the presence of endogenous d‐amino acids, including d‐serine and d‐alanine [31, 32, 33]. DAO catalyzes the oxidative deamination of neutral and basic d‐amino acids, generating the corresponding α‐keto acids, ammonia, and hydrogen peroxide [34]. Thus, the presence of endogenous d‐amino acids together with DAO activity under high oxygen tension may enhance oxidative stress. This enhanced oxidative burden could, in turn, inhibit cell proliferation and contribute to the selective loss of DAO‐positive cells during culture.

We confirmed that MDCK cells express CSE and CBS (Figure 4), enzymes known to produce H₂S from l‐cysteine. This observation is consistent with previous studies demonstrating that MDCK cells express CSE, CBS, and 3MST [24]. Under the initial conditions to assess H₂S‐producing activity, l‐cysteine alone generated little to no detectable H₂S (Figure 1D). However, these assays were performed using lysates with low protein concentrations because of the high activity of the d‐cysteine pathway. Increasing the protein concentration revealed measurable H₂S production from l‐cysteine (Figure 5A,C), indicating that the initial observation was primarily attributable to limited detectability under the diluted conditions rather than an absence of enzymatic activity of the l‐cysteine pathway. CSE and CBS catalyze multiple reactions capable of generating H₂S from l‐cysteine, l‐homocysteine, or combinations of these substrates [35]. While H₂S production observed with l‐cysteine and l‐homocysteine remained lower than that observed with d‐cysteine (Figure 5A,C), these results demonstrate that the l‐cysteine pathway is functionally active in MDCK cells. CBS can catalyze β‐replacement of l‐cysteine by water; however, kinetic analyses suggest that H₂S production by CBS predominantly proceeds via a β‐replacement reaction involving l‐cysteine and l‐homocysteine [11]. Although the relative contributions of CSE and CBS in MDCK cells were not resolved, the data support the catalytic activity of both enzymes.

Compared with parental MDCK cells, the protein levels of the H₂S‐producing enzymes 3MST, CSE, and CBS were comparable between clones C19 and C20 under both sub‐confluent and post‐confluent conditions (Figure 3B; Figure 4A,B). DAO protein levels were also similar between the two clones under post‐confluent conditions, whereas under sub‐confluent conditions DAO levels were significantly higher in C19 than in C20 (Figure 3A). This difference was consistent with the corresponding mRNA expression profiles (Figure 3C), suggesting that the observed clone‐specific difference in DAO expression under sub‐confluent conditions is likely regulated at the transcriptional level. Further studies will be required to elucidate the molecular mechanisms underlying this selective regulation of DAO expression during proliferative states and to clarify how this regulation contributes to cellular function. In addition, more comprehensive genetic or transcriptomic characterization should be pursued, particularly if these clones are to be developed into research resources.

DAO expression has been reported to increase after confluence in LLC‐PK1 cells, a porcine proximal tubule cell line in which cells are tightly packed and exhibit strong cell–cell adhesion [36, 37]. A similar confluence‐dependent increase in DAO expression was observed in MDCK cells and in clone C20; however, clone C19 expressed high levels of DAO even before reaching confluence (Figure 3A). Live‐cell imaging using the intracellular H₂S probe HSip‐1 DA revealed cell density‐dependent differences between C19 and C20. Under sub‐confluent conditions, treatment with 10 mM d‐cysteine induced detectable fluorescence in C19, with signals primarily localized to cells attached to the culture plate (Figure 7A). In contrast, C20 showed little or no detectable fluorescence under the same conditions. These observations suggest that the d‐cysteine‐dependent H₂S‐producing pathway is functionally active in C19 during the proliferative state, whereas this pathway appears to be minimally active or absent in C20 under comparable conditions.

Under post‐confluent conditions, the fluorescence pattern in C19 changed markedly. Following treatment with d‐cysteine, l‐cysteine, or the H₂S donor Na₂S, rounded fluorescence was observed predominantly, whereas cells attached to the culture plate showed little or no detectable signal (Figure 7B). The rounded fluorescence remained stationary upon gentle agitation, indicating that they were not free‐floating debris. Under confluent conditions, spherical cells are commonly observed in culture and may represent cells undergoing mitotic rounding or cells that have lost adhesion due to apoptosis or anoikis [38]. In such cellular states, alterations in morphology and membrane properties may influence the probe uptake, retention, or intracellular distribution, potentially resulting in enhanced fluorescence that does not directly reflect intracellular H₂S levels. In contrast to C19, C20 under post‐confluent conditions showed only weak fluorescence following d‐cysteine treatment, with intense fluorescence observed in only a small subset of rounded cells and minimal overall signal. In both clones, fluorescence associated with cells attached to the culture plate was not clearly detectable. Thus, under post‐confluent conditions, the imaging data do not support the conclusion that the d‐cysteine pathway operates in the same manner as in C19 under sub‐confluent conditions. The weak or absent fluorescence observed in adherent cells might reflect alterations in intracellular H₂S metabolism. Because H₂S can be stored intracellularly as bound sulfane sulfur, it is plausible that, in post‐confluent cells, bound sulfur species predominate over free H₂S, thereby limiting probe‐detectable H₂S. These observations highlight the importance of evaluating H₂S production in conjunction with its intracellular storage and metabolic fate.

H₂S is an endogenous signaling molecule with cytoprotective roles in the kidney, including attenuation of oxidative stress, suppression of inflammation and fibrosis, and preservation of mitochondrial function in experimental models of chronic kidney disease (CKD) [39, 40, 41]. In particular, studies of proteinuric and ischemic nephropathies have demonstrated that H₂S mitigates tubular injury and promotes recovery of renal function [39] [41]. In the present study, we observed that DAO‐positive cells were readily lost during repeated passaging in vitro (Figure 2A,B). This observation raises the possibility that, during chronic renal injury in vivo, tubular cells exposed to sustained oxidative stress, metabolic stress, and mitochondrial dysfunction may include DAO‐expressing, H₂S‐producing cells that are susceptible to damage or depletion. Such selective vulnerability could result in reduced local H₂S availability, thereby weakening endogenous protective mechanisms and potentially contributing to disease progression. In vivo, renal tubular epithelial cells reside within a highly regulated microenvironment characterized by controlled oxygen gradients, extracellular matrix interactions, and mechanical cues that help maintain nephron segment‐specific phenotypes [42]. These physiological conditions may contribute to stabilization of the d‐cysteine pathway under normal conditions, consistent with previous observations that renal DAO expression increases from birth to adulthood [16]. Moreover, d‐cysteine administration attenuates renal ischemia–reperfusion injury [16], supporting a protective role for this pathway in experimental models of kidney injury. In this context, the expression of DAO and 3MST in the kidney, including in humans [43, 44, 45], suggests potential translational relevance. The extent to which the d‐cysteine‐dependent H₂S‐producing pathway contributes to human kidney disease, particularly in chronic and proteinuric states, remains to be determined. Further studies will be required to clarify the regulation of this pathway during chronic renal injury and to evaluate its therapeutic potential in CKD.

In conclusion, our findings demonstrate that DAO and 3MST mediate H₂S production from d‐cysteine in MDCK cells. Although MDCK cells progressively lost this activity during passaging, clonal isolation enabled the establishment of sublines that stably maintain the d‐cysteine pathway. A limitation of this model is the canine origin of MDCK cells. Because the canine genome is less well annotated than the human or mouse genomes, transcriptomic and proteomic analyses can be more challenging. Given that MDCK cells are widely used as a model of renal epithelial cells, the clonal cells co‐expressing DAO and 3MST may provide an experimental system for investigating the regulation and cellular significance of d‐cysteine‐mediated H₂S production.

Author Contributions

Norihiro Shibuya conceived and supervised the project; Akari Miyamoto, Haruna Ueno, and Norihiro Shibuya designed and performed experiments and analyzed data; Tadashi Nakagawa assisted with experimental design; Norihiro Shibuya wrote the manuscript; and Norihiro Shibuya and Hideo Kimura made manuscript revisions.

Funding

This work was supported by Takeda Science Foundation (TSF) and Japan Society for the Promotion of Science (JSPS) KAKENHI Grant Number JP23K06155.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1: Primer sequences used for RT‐qPCR. Target genes: DAO, d‐amino acid oxidase; MPST, 3‐mercaptopyruvate sulfurtransferase (3MST); CTH, cystathionine γ‐lyase (CSE): CBS, cystathionine β‐synthase. Reference genes: HPRT1, hypoxanthine‐guanine phosphoribosyltransferase; TBP, TATA‐binding protein; GAPDH, glyceraldehyde‐3‐phosphate dehydrogenase; ACTB, β‐actin. Primer sequences for ACTB were obtained from a previous study (see Reference 1 in the Supporting Information) and were originally designed based on the canine ACTB sequence.

Data Availability Statement

The data supporting the findings of this study are available from the corresponding author upon reasonable request.