Endocrinology and Metabolism

Knockout of the murine cysteine dioxygenase gene results in severe impairment in ability to synthesize taurine and an increased catabolism of cysteine to hydrogen sulfide

Iori Ueki, Heather B. Roman, Alessandro Valli, Krista Fieselmann, Jimmy Lam, Rachel Peters, Lawrence L. Hirschberger, Martha H. Stipanuk


Cysteine homeostasis is dependent on the regulation of cysteine dioxygenase (CDO) in response to changes in sulfur amino acid intake. CDO oxidizes cysteine to cysteinesulfinate, which is further metabolized to either taurine or to pyruvate plus sulfate. To gain insight into the physiological function of CDO and the consequence of a loss of CDO activity, mice carrying a null CDO allele (CDO+/− mice) were crossed to generate CDO−/−, CDO+/−, and CDO+/+ mice. CDO−/− mice exhibited postnatal mortality, growth deficit, and connective tissue pathology. CDO−/− mice had extremely low taurine levels and somewhat elevated cysteine levels, consistent with the lack of flux through CDO-dependent catabolic pathways. However, plasma sulfate levels were slightly higher in CDO−/− mice than in CDO+/− or CDO+/+ mice, and tissue levels of acid-labile sulfide were elevated, indicating an increase in cysteine catabolism by cysteine desulfhydration pathways. Null mice had lower hepatic cytochrome c oxidase levels, suggesting impaired electron transport capacity. Supplementation of mice with taurine improved survival of male pups but otherwise had little effect on the phenotype of the CDO−/− mice. H2S has been identified as an important gaseous signaling molecule as well as a toxicant, and pathology may be due to dysregulation of H2S production. Control of cysteine levels by regulation of CDO may be necessary to maintain low H2S/sulfane sulfur levels and facilitate the use of H2S as a signaling molecule.

  • cystathionine γ-lyase
  • cystathionine β-synthase
  • sulfane
  • sulfate

in mammals, the major route of cysteine catabolism is via an oxidative cysteinesulfinate-dependent pathway. As shown in Fig. 1, dioxygen is added to cysteine in a reaction catalyzed by cysteine dioxygenase (CDO), forming cysteinesulfinate (cysteinesulfinic acid). Cysteinesulfinate undergoes further metabolism by two competing pathways. First, cysteinesulfinate undergoes decarboxylation to form hypotaurine, which is then oxidized to taurine. Taurine biosynthesis from cysteine is thought to occur predominantly by this pathway. Second, cysteinesulfinate is metabolized via transamination to form 3-sulfinylpyruvate, which is unstable and gives rise to pyruvate and sulfite, which exists predominantly as the HSO3 and SO32− ions at physiological pH. Sulfite is further oxidized to sulfate (SO42−) by sulfite oxidase (60, 62). Dominy et al. (19) demonstrated that changes in CDO levels contribute to the control of intracellular cysteine concentrations. CDO concentration and activity undergo large increases in response to increased intake of protein or sulfur-containing amino acids and consequently accounts for a large fraction of cysteine catabolic flux when cysteine availability is near or above the animal's requirement level (1719, 64).

Fig. 1.

Pathways of cysteine catabolism. Cysteine is metabolized by cysteine dioxygenase (CDO)-dependent pathways to taurine (via cysteinesulfinate decarboxylation) and sulfate (via cysteinesulfinate transamination). Cysteine also undergoes desulfhydration that is catalyzed by the methionine transsulfuration pathway enzymes. The reduced sulfur released in these pathways can be further oxidized to sulfate.

Cysteine may also undergo catabolism by several enzymes, especially cystathionine γ-lyase and cystathionine β-synthase, that cleave the sulfur from cysteine prior to cysteine oxidation. These two enzymes catalyze the two steps of methionine transsulfuration (i.e., the synthesis of cystathionine from homocysteine and serine and the cleavage of cystathionine to yield cysteine, α-ketobutyrate, and ammonia). In addition, both enzymes have been shown to be responsible for sulfide (H2S, or HS) production from cysteine, with H2S believed to be an important signaling molecule (9, 10, 58, 61). Sulfide is further oxidized to sulfate by a complex series of reactions. First, two molecules of sulfide are converted to thiosulfate by the mitochondrial sulfide:quinine oxidoreductase/sulfur dioxygenase/sulfur transferase system. This is followed by the glutathione-dependent cleavage of thiosulfate to yield sulfite plus H2S, which is catalyzed by a thiosulfate reductase and then the oxidation of sulfite to sulfate by mitochondrial sulfite oxidase (26, 36, 38). The flux of cysteine through these desulfuration pathways is normally low and insensitive to changes in sulfur amino acid availability (2). Thus, it has been assumed that cysteine levels in the body are controlled predominantly via regulation of CDO concentration and activity (62, 64).

The possible importance of CDO in human health has been suggested by indirect evidence for abnormal or deficient CDO activity in individuals with several autoimmune and neurodegenerative diseases (6, 11, 23, 35, 79). Investigators conducting clinical studies on these patient populations have looked primarily at impairments in the conversion of cysteine to sulfate. For example, individuals with rheumatoid arthritis exhibited depressed levels of sulfate in plasma, elevated fasting plasma cysteine concentrations, elevated plasma cysteine-to-sulfate ratios, and lower sulfate concentrations in synovial fluid, all of which are consistent with impaired cysteine oxidation (6, 11). The hypothesis of a link between CDO loss-of-function mutations and the incidence or severity of human disease has been strengthened by the report of a much higher frequency of polymorphisms in the CDO1 gene in patients with rheumatoid arthritis and in people with a family history of rheumatoid arthritis compared with the control population (77). Those CDO polymorphisms that were associated with rheumatoid arthritis included frame-shift insertions, mutations involving proline residues that would be predicted to cause overall abnormalities in protein folding, and single nucleotide substitutions that resulted in mutation of residues known to be critical for CDO activity. The etiology of these diseases could be linked to functional impairment of cysteine dioxygenase, which might lead to elevated levels of cysteine and H2S as well as lowered concentrations of hypotaurine, taurine, and sulfate.

To gain more insight into the physiological function of CDO and the consequences of a loss of CDO activity, we generated mice carrying a null CDO allele (CDO+/− mice) and crossed these mice to generate CDO−/−, CDO+/−, and CDO+/+ mice for this study. Growth, fertility, clinical phenotype, and metabolic phenotype were assessed. Because CDO−/− mice had extremely low plasma and tissue taurine levels, the effect of supplemental taurine was assessed. The possibility that mice lacking CDO catabolized more cysteine by the desulfhydration pathways was also investigated.


Generation of CDO targeting construct and CDO+/Flox mice.

The mouse chromosome 18 sequence from the Ensembl database was used to design a targeting vector with loxP sites flanking exon 3 of the murine CDO1 gene (i.e., the floxed CDO1 gene). Deletion of exon 3 leads to loss of active site resides and also shifts the reading frame for exons 4 and 5. Briefly, the BAC clone RP23-394I6 was used to generate homologous arms and the conditional knockout region of the vector. As shown in Fig. 2A, the final targeting vector contained loxP sequences flanking the conditional knockout region, Frt sequences flanking the Neo expression cassette, and a DTA expression cassette. The sequence of the final vector was confirmed both by restriction digestion and end sequencing analysis. NotI was used to linearize the vector, and vector DNA was electroporated into C57BL/6 embryonic stem (ES) cells. Positive clones containing the targeting construct were selected using G418 (genticin sulfate) followed by Southern analysis (Fig. 2B). These clones were injected into blastocysts, and the blastocysts were transferred into pseudopregnant females to generate chimeras. Chimeras were bred with C57BL/6N Tac wild-type mice to generate C57BL/6 mice that were heterozygous for the CDO floxed gene (CDO+/Flox). This work was performed by Caliper Life Sciences (formerly Xenogen Biosciences) under contract by Cornell University.

Fig. 2.

Generation and analysis of the CDO null mutation. A: schematic representation of the wild-type (WT) CDO allele (+), the targeted CDO allele (Floxed) with exon 3, and the neo cassette flanked by loxP sites, and the null CDO allele (−) obtained by Cre recombinase-mediated deletion of the targeted region of the floxed allele. B: Southern blots of tail DNA from CDO+/+ and CDO+Flox mice. Restriction sites used for screening (Bg, BglII; B, BamHI; E, EcoRI) and probes are shown in A. C: genotyping of tail DNA from mice of various genotypes using primers indicated at the left (see Table 1). Sizes of expected PCR products are shown on the right.

Generation of CDO−/− mice.

Heterozygous C57BL/6 mice carrying one functional CDO allele and one floxed CDO allele (CDO+/Flox) were bred to generate mice homozygous for the floxed CDO allele (CDOFlox/Flox). Then, CDOFlox/Flox mice were crossed with CMV-cre (+) transgenic mice [B6.C-Tg6(CMV-cre)1 Cgn/J; Jackson Laboratory] to generate mice that were heterozygous for the conditional CDO allele and positive for the Cre recombinase allele (CDO+/FloxCre+). CDO+/FloxCre+ mice were bred to wild-type mice to generate mice that carried the null mutation in their germ-line and that did not carry the Cre allele (CDO+/−). The colony was expanded by breeding these CDO+/− mice.

Determination of genotype and sex of mice.

The presence of the wild-type, targeted/floxed, and disrupted/null CDO alleles and of the Cre allele was determined by PCR analysis of genomic DNA samples obtained from tail snips (Fig. 2C), using primers that yielded specific products for each allele (Table 1). For mice that died before sex could be visually distinguished, sex determination was performed via PCR with SRY primers (Table 1).

View this table:
Table 1.

Primers used for genotyping

Animal procedures.

Experimental procedures involving live animals were conducted with the approval of the Cornell University Institutional Animal Care and Use Committee. Mice were housed in a pathogen-free barrier facility. All animals had access to an irradiated standard nonpurified rodent diet (7012 Teklad LM-485; Harlan Laboratories) and water ad libitum. CDO+/− mice were crossed to generate mice that were homozygous for the null CDO gene (CDO−/−), as well as wild-type (CDO+/+) and heterozygous (CDO+/−) mice. Animals were genotyped at postnatal day 14 (P14) and weaned at day P28. Mice were weighed daily from P1 to P35 and every 2–4 days thereafter.

In the initial characterization studies, dams and pups were fed the standard nonpurified rodent diet throughout the entire experimental period. At 8 wk of age, animals used for initial characterization of growth and survival were evaluated by a basic SHIRPA primary screening (52; http://empress.har.mre.ac.uk). For analysis of metabolites and histopathology, mice were killed at 8 wk of age. Feed intake was measured for a separate group of mice that were caged separately by sex and genotype following weaning. In a follow-up study to assess the effect of taurine supplementation, taurine was added to the drinking water (2 g/100 ml) of the taurine-supplemented dams from the time of parturition until termination of the experiment at P21.

In another series of follow-up studies to further investigate the catabolism of cysteine to reduced sulfur compounds, mice were fed a semipurified diet based on the AIN93G formulation with either an adequate diet that contained 200 g of isolated soy protein plus 3.1 g of l-methionine per kilogram (8.0 g methionine equivalents/kg) or a sulfur amino acid-enriched diet that contained 200 g vitamin-free casein + 4.2 g l-cystine + 1.5 g l-methionine (12.3 g methionine equivalents/kg) for 2 wk prior to sample collection. (For reference, the standard AIN93G diet formulation for rodents provides 9.3 g methionine equivalents/kg, with 1 g cystine = 1.24 g methionine equivalents.)

The Glucometer Elite (Bayer, Elkhard, IN) was used to measure blood glucose in tail blood of unanesthetized mice that were in the postabsorptive state. Immediately following euthanization of mice with CO2, nose-to-anus length was measured and blood and tissues were collected. Tissues were either frozen immediately in liquid nitrogen or fixed for histopathological analysis. Frozen tissues and plasma were stored at −80°C until analysis. For a subset of mice (8 of each genotype), bone mineral density was measured with the Lunar PIXImus Densitometer (GE Medical Systems).

Histological analysis and skeletal preparation.

Tissues isolated from euthanized CDO+/+ and CDO−/− mice were fixed in 4% (wt/vol) phosphate-buffered paraformaldehyde for 48 h. Fixed tissues were embedded, sectioned, and differentially stained with hematoxylin and eosin (H&E) for histopathological analysis by the histology core facility in the College of Veterinary Medicine at Cornell University. For measurements on the skeleton, whole skeletons were prepared and stained with alcian blue and alizarin red for cartilage and bone by the method of Depew (12).

Additional mice were used to obtain lung tissue for histological analysis. Mice were anesthetized with an intraperitoneal injection of Avertin (tribromoethanol) and then euthanized by opening the abdominal cavity and cutting the diaphragm. Lungs were then perfused intratracheally with a blunt butterfly catheter connected to a column of 4% (vol/vol) paraformaldehyde suspended 25 cm above the dissecting surface. When the lungs were fully inflated, the trachea was clamped, and the entire mass of thoracic viscera was removed intact and placed in 4% paraformaldehyde for 24 h. Samples from two or more lobes from each mouse were submitted for histological processing. Semiserial 5-μm sections from paraffin-embedded tissues were processed with Verhoeff VanGeison (VVG) and Masson Trichrome (MTC) stains for connective tissue components. Analysis of differentially stained sections was performed using an Olympus BX50 microscope with a Moticam 2300 cMOS camera (Motic North America, Richmond, BC, Canada). Lung sections were evaluated for mean linear intercept (MLI) by the method of Dunnill (22) with modifications by Hamelet et al. (34). Image analysis was performed at ×200 magnification using JMicroVision 1.2.7 (http://www.jmicrovision.com). Images were overlayed with a scaled 50-μm grid (∼10 vertical lines × 10 horizontal lines), and the sum of the number of alveolar walls intersected by lines in both the horizontal and vertical direction was divided by the total length of the lines on each image, yielding the MLI value. The MLI value for each mouse was the average of 5–9 images.

Measurement of taurine, hypotaurine, cysteine, glutathione, and sulfate in plasma and liver.

For the analysis of taurine and hypotaurine, total glutathione (GSH) and total cysteine, plasma, and tissue homogenates were prepared as described previously (72). The protein concentration of tissue homogenates was measured using the BCA protein assay kit (Pierce). Homogenates and plasma were acidified with sulfosalicylic acid (SSA) to a final concentration of 2% (wt/vol). Precipitated proteins were removed by centrifugation at 16,000 g to obtain acid supernatants.

For the measurement of taurine and hypotaurine, acid supernatants were diluted in 200 mM borate buffer, pH 10.4, and then analyzed using pre-column derivatization with o-phthaldialdehyde (OPA) and HPLC as described previously (17).

To measure total GSH, 5 μl of the acid supernatant was reduced by addition of 2 μl of 200 mM tris(2-carboxyethyl)phosphine and 113 μl of 0.2 M borate buffer (pH 10.4), followed by incubation for 10 min at room temperature. The sample was then derivatized by addition of 120 μl of 37.3 mM OPA solution (50 mg of OPA dissolved in 0.5 ml of methanol and brought up to a final volume of 10 ml by addition of 0.2 M borate buffer, pH 10.4) and incubation for 5 min. Next, 960 μl of 0.5 mM phosphate buffer, pH 7, was added to each derivatized sample, and 30 μl of the final mixture was separated by reversed-phase HPLC using a Nova-Pak C18 4-μm 3.9 × 150 mm column (Waters). The separation was carried out by a gradient elution at a flow rate of 1 ml/min. The mobile phase was started at 95% eluent A (25 mM phosphate buffer, pH 7) and 5% eluent B (100% methanol), and the proportion of eluent B was increased linearly to 25% eluent B over 5 min and then to 90% eluent B over an additional 3 min and maintained at 90% eluent B for 4 min. The gradient was then switched back to 95% eluent A/5% eluent B over 4 min, and the column was allowed to equilibrate for 8 min prior to the next injection. Derivatized products were detected by a multiple wavelength fluorescence detector (model 2475, Waters, Milford, MA) at an excitation wavelength of 340 nm and an emission wavelength of 420 nm. Integration of the chromatographic peaks and interpolation of GSH concentrations were determined using Empower 2 software (Waters).

For measurement of total cysteine, the pH of the acid supernatant was adjusted to ∼8.3 with sodium hydroxide. Dithiothreitol was then added to a final concentration of 12.5 mM with thorough mixing followed by incubation for 10 min at room temperature to reduce disulfide bonds prior to proceeding with the assay. Cysteine was assayed by the acid ninhydrin assay of Gaitonde (28) as modified by Dominy et al. (19).

Plasma sulfate levels were measured by HPLC separation and conductivity detection as described by Hoffman et al. (37). Plasma was centrifuged using a 3,000-MW centrifugal filter (Amicon Ultracel, Millipore) to remove proteins prior to application to the column.


Frozen tissue was homogenized in lysis buffer [50 mM tris buffer, pH 7.5, 1% (vol/vol) Nonidet P-40, 2 mM EDTA, 150 mM NaCl supplemented with 1× mammalian protease inhibitor cocktail (Sigma-Aldrich)]. Homogenates were centrifuged at ∼14,000 g at 4°C for 20 min, and the total protein content of the supernatant fraction was measured using the BCA protein assay kit (Pierce). Soluble proteins were separated by SDS-PAGE and transferred to a 0.45-μm Immobilon-P polyvinylidene difluoride membrane (Millipore). Membranes were blocked using blocking buffer for near-infrared fluorescent Westerns (LI-COR Biosciences) and blotted for immunoreactive proteins. Antibodies against CDO and cysteamine dioxygenase (2-aminoethanethiol dioxygenase, ADO) were previously generated in our laboratory (26, 27). Antibody against cysteinesulfinate decarboxylase (CSAD) was the generous gift of Dr. Marcel Tappaz (Institut National de la Santé et de la Recherche Médicale, France) (54). Anti-actin was obtained from Cell Signaling Technology (Danvers, MA), anti-cystathionine γ-lyase (CSE) from Abnova (Taiwan), and anti-cystathionine β-synthase (CBS) from Santa Cruz Biotechnology (Santa Cruz, CA). Immunoreactive proteins were detected using an infrared fluorescent dye-labeled secondary antibody (IRDye, LI-COR Biosciences) as directed by the manufacturer and the Odyssey direct infrared imaging system (LI-COR Biosciences). The intensity of the signal was quantified using Odyssey software (LI-COR Biosciences). Actin was used as a loading control to normalize values for proteins of interest.

Measurement of acid-labile H2S and mitochondrial cytochrome c oxidase activity.

Acid-labile sulfur was measured in homogenates of pancreas, liver, lung, and kidney, following treatment with dithiothreitol to reduce disulfide bonds, by the method of Siegel (57), as previously described (10). Mitochondria were purified from mouse liver using a commercial isolation kit (no. 89801; Thermo Scientific, Rockford, IL). Cytochrome c oxidase activity in the isolated mitochondria was assessed using a commercial activity assay kit (no. KC310100; BioChain Institute, Hayward, CA).

Plasma IGF-I and blood acylcarnitine measurements.

Plasma IGF-I was measured using the Luminex system and Millipore Milliplex MAP Rat/Mouse IGF-I Assay (Vanderbilt Diabetes Center Core). Blood acylcarnitines were quantified by tandem mass spectrometry (Institute of Metabolic Disease, Baylor Research Institute) using dried blood spots.

Statistical analyses. Data are expressed as means ± SE unless indicated otherwise. Data were analyzed by SAS (SAS Institute, Cary, NC) using a general linear model accounting for genotype (G), sex (S), and their interaction (G × S). For the taurine study, data were analyzed by sex using a model accounting for genotype (G), dietary treatment with taurine (T), and their interaction (G × T). The differential effects of the slopes within each growth interval were analyzed using a mixed model in SAS that included the fixed effects of interval, day nested within interval, genotype, and all their interactions. Survival data were plotted by the Kaplan-Meier method and analyzed by the log-rank test using GraphPad Prism 5 Software. Results were accepted as significantly different at P ≤ 0.05 for main effects and P ≤ 0.10 for interactions.


CDO−/− mice exhibit high postnatal mortality and growth delay.

Male and female CDO+/− mice grew normally and were fertile. Pups born to CDO+/− females mated with CDO+/− males were distributed according to expected Mendelian ratios (observed, CDO+/+ 49, CDO+/− 135, CDO−/− 55; χ2 test, P > 0.1), indicating that CDO gene disruption did not affect prenatal survival in the CDO+/− dams. Early postnatal mortality (prior to weaning at P21) was high for CDO−/− pups, with only 71% of females and 83% of males surviving to P21 (Fig. 3A) opposed to 100% of CDO+/− pups and >99% (one death) of CDO+/+ offspring. No deaths occurred in any genotype from weaning to the end of the experiment at P56. CDO−/− mice weighed 90.5% as much as CDO+/+ mice (P < 0.001) at P1 (birth = P0) and continued to exhibit a growth deficit through P56 (Fig. 3B). By P21, differences in body weight had increased to a maximum difference: on average, at P21, CDO−/− mice weighed only 58% as much as CDO+/+ mice (5.6 vs. 9.7 g, respectively, P < 0.001). Throughout the postweaning period, body weight of CDO−/− mice remained low, being 66% of CDO+/+ at P35 (P < 0.001) and 77% of CDO+/+ at P56 (P < 0.001). The weights of CDO+/− mice were similar to those of CDO+/+ mice at all time points.

Fig. 3.

CDO-null mice have reduced mortality and grow at a reduced rate after birth. A: effect of genotype on survival of mice between postnatal day (P)1 and P21. Survival of Null (CDO−/−) mice was significantly less than that of WT (CDO+/+) and Het (CDO+/−) mice [log rank test: males, P < 0.05 (*) and females, P < 0.001 (**)]. Number of mice in each group at day 1 is shown in parentheses. B: effect of genotype on body weight of mice between P1 and P56. Data for male and female mice are plotted separately (each data point represents 4–11 mice for WT, 24–29 mice for Het, and 9–13 mice for Null). Null CDO mice were smaller than WT and Het mice at P1 (P < 0.001) and remained significantly smaller for the duration of the study. C: effect of genotype on weight gain per day. Daily weight gain from P2 to P35 was divided into 5 intervals, as shown by vertical dotted lines, and slopes (weight gain per day/postnatal day) for each interval were compared for each genotype × sex. Intervals at which slopes for Null mice differed from those of WT and Het mice are shown by * (P < 0.01) or ** (P < 0.001). D: effect of genotype on body length. Body weight, body length, and body weight/body length on P56 are shown. Significant effects of genotype (G), sex (S), and their interaction (G × S) are reported. Each bar represents the mean ± SE of 6–8 mice.

Daily weight gain, calculated as the change in body weight in a 24-h period, was consistently lower for both male and female CDO−/− mice than for CDO+/+ or CDO+/− mice. As shown in Fig. 3C, CDO+/+ and CDO+/− mice underwent their exponential phase of growth during the P18–P20 interval, whereas CDO−/− mice gained no weight during this same time period. Following the exponential growth phase, daily weight gain in CDO+/+ and CDO+/− mice leveled off at 0.8 g per day from P20 to P30 and then gradually decreased to ∼0.3 g per day. CDO−/− mice underwent a delayed exponential growth phase from P23 to P28, but, in contrast to the CDO+/+ and CDO+/− mice, CDO−/− mice did not have an extended period of weight gain at the higher rate but immediately began the gradual decrease in rate of weight gain along with the CDO+/+ and CDO+/− mice.

Average feed intake was similar for CDO−/− and CDO+/+ males but 15% higher for CDO−/− females than for CDO+/+ females, ruling out differences in feed intake as the cause of reduced growth in CDO−/− mice. When expressed as average intake per gram body weight, CDO−/− females consumed an average of 31% more feed per gram body weight than did CDO+/+ females, and CDO−/− males consumed 8% more feed per gram body weight than did CDO+/+ males.

As shown in Fig. 3D, the length of CDO−/− mice was less than that of CDO+/+ and CDO+/− mice. At P56, the length of null mice was 93% that of CDO+/+ mice for both males and females (P < 0.001), whereas weight-for-length of null mice was 80 and 87% that of CDO+/+ mice for males and females, respectively (P < 0.001). Organ weights are shown in Table 2. The percentage of total body weight accounted for by kidney and heart was greater, whereas that accounted for by gastrocnemius muscle was less, for CDO−/− mice than for CDO+/+ and CDO+/− mice.

View this table:
Table 2.

Effect of CDO genotype on body length and organ weights in CDO+/+ (WT), CDO+/− (Het), and CDO−/− (Null) mice

CDO−/− mice display a characteristic phenotype at 8 wk of age.

SHIRPA primary screening of mice at 8 wk of age revealed several characteristics observed exclusively in CDO−/− animals. CDO−/− mice demonstrated reduced spontaneous activity and reduced irritability. Eyes of CDO−/− mice exhibited excess lacrimation and partial palpebral closure in 72% of null females and 40% of null males. CDO−/− mice consistently demonstrated kyphosis accompanied by an irregularly shaped cranium with a foreshortened nose, sometimes accompanied by a transverse crossbite (malocclusion) of varying severity (Fig. 4A). Further evaluation of whole mount 8 wk old skulls revealed obvious lateral displacement of the nasal bone (Fig. 4B) in 85% of CDO−/− mice. Paw abnormalities included a plantigrade stance with digits having bulbous tips and extreme hyperlaxity (Fig. 4D). Surprisingly, postmortem gross and histopathological evaluation of 8-wk-old animals detected no overt pathologies in organs of CDO−/− mice compared with CDO+/− or CDO+/+ littermates. However, bone mineral density was decreased in CDO−/− mice compared with CDO+/+ mice (P < 0.001; CDO+/+, 0.045 ± 0.001 g/cm2 vs. CDO−/−, 0.038 ± 0.001 g/cm2; n = 8 per group). Blood glucose was 23% lower in CDO−/− mice (P < 0.01) than in CDO+/+ mice, whereas blood glucose in CDO+/− mice was not different from that in CDO+/+mice (Fig. 5). Plasma IGF-I levels were 70% lower in CDO−/− mice than in CDO+/+ mice (Fig. 5).

Fig. 4.

Eight-week-old CDO-null mice have a distinctive phenotype. A: CDO-null mice show excess lacrimation and partial palpebral closure, plantigrade stance, hyperextensible toes, and less-erect ears. Kyphosis is apparent in vivo and was confirmed in whole mount specimens (not shown). B: cleared whole mount crania from 8-wk-old WT and Null mice. Alcian blue and alizarin red staining did not reveal any significant differences in ossification patterns, but lateral displacement of the nasal bone was frequently apparent, even without malocclusion in vivo (arrow). C: transient joint contracture was observed in forelimbs of Null pups during early postnatal period. D: joint contracture was fully resolved in adult animals but toe abnormalities remained.

Fig. 5.

Blood glucose and Plasma IGF levels are lower in CDO-null mice. Blood glucose level was measured by glucose meter, and plasma IGF-I concentration was measured using a luminex assay. Significant effects of G (WT, Het, Null) and S (male, female) are reported. Each bar represents the mean ± SE of 6–8 mice.

CDO−/− mice display an early postnatal phenotype.

In addition to their small size, 62% of CDO−/− pups presented with a contracture of one or both distal forelimbs beginning P3 and resolving no later than P11 (Fig. 4, C and D). During this early postnatal period, CDO−/− animals could also be visually distinguished from their CDO+/− and CDO+/+ littermates by their lack of erect pinnae and the presence of bulbous toes and extreme joint hyperlaxity. Eye opening was consistently delayed by 2–3 days in all CDO−/− animals. None of these conditions was observed in any CDO+/− or CDO+/+ littermates.

CDO−/− mice exhibit impaired reproductive capacity.

Although CDO+/− mice reproduced normally, giving rise to offspring with a Mendelian distribution of genotypes, CDO−/− mice had reproductive impairments. A limited assessment of fertility of CDO−/− mice was conducted by mating several CDO−/− males and females with CDO+/+ mice. CDO−/− female mice were fertile and carried their pregnancies to term, although their offspring did not survive. On the other hand, no pregnancies resulted from repeated mating of CDO−/− male mice with CDO+/+ females. Histological examination of the testes and epididymides of 8-wk-old null males revealed the presence of mature sperm.

CDO+/− mice maintain relatively normal levels of hepatic CDO.

As expected, CDO was undetectable in the liver of CDO−/− mice (Fig. 6). The CDO abundance in livers of CDO+/− mice was not significantly different from that of CDO+/+ mice, although there was a trend toward lower levels in the CDO+/− mice. There was a significant sex difference in hepatic CDO concentration in both the CDO+/− and CDO+/+ mice, with hepatic CDO abundance in females being 1.9 times that in males.

Fig. 6.

Hepatic abundance of enzymes in the taurine synthetic pathway. Livers were collected from mice at P56 and were analyzed for CDO, (CSAD), and (ADO) abundance by Western blotting. Western blots of pooled samples are shown (top) and quantified (bar graphs). Actin was used as a loading control. Significant effects of G (WT, Het, Null) and S (male, female) and their interaction (G × S) are reported. Each bar represents the mean ± SE of 6–8 mice.

Also shown in Fig. 6 are the results of immunoblotting for amounts of CSAD, the second enzyme in the pathway of taurine synthesis from cysteine, and ADO, the enzyme that oxidizes cysteamine to hypotaurine in the alternative pathway for taurine synthesis. In CDO−/− mice, the abundance of CSAD was 2.3 times that present in CDO+/+ mice. CSAD levels in livers of CDO+/− mice were similar to those in CDO+/+ mice. In addition, CSAD was more abundant in liver of male than in liver of female CDO+/− or CDO+/+ mice but was similarly high in liver of CDO−/− mice. There were no effects of sex or genotype on the level of ADO.

CDO−/− mice have elevated levels of cysteine and sulfate and low levels of taurine.

In CDO−/− mice, the hepatic cysteine levels were 49% higher (P < 0.001) and the plasma cysteine levels were double (P < 0.001) those in CDO+/+ mice (Fig. 7A). Only a trend toward elevation in hepatic cysteine level (+14%, P < 0.10) was observed in CDO+/− mice compared with CDO+/+ mice, and plasma cysteine levels were not significantly different between CDO+/+ and CDO+/− mice.

Fig. 7.

Effect of null CDO mutation on plasma and liver metabolites involved in cysteine metabolism. A: plasma and liver cysteine. B: plasma and liver GSH. C: plasma and liver taurine. D: plasma hypotaurine. E: plasma sulfate. Significant effects of G (WT, Het, Null), S (M, F), and their interaction (G × S) are reported. Each bar represents the mean ± SE of 6–8 mice.

GSH levels were similarly elevated in liver and plasma of both CDO−/− and CDO+/− mice compared with CDO+/+ mice (Fig. 7B). Compared with the CDO+/+ mice, GSH levels in CDO−/− and CDO+/− mice were 41 and 60% greater, respectively, in the liver (G, P < 0.001) and 52 and 47% greater, respectively, in the plasma (G, P < 0.001).

CDO−/− mice exhibited a drastic decrease in hepatic taurine level to only 2.2% of that in CDO+/+ mice (Fig. 7C). CDO+/− mice had intermediate levels of hepatic taurine. Female mice had significantly higher levels of hepatic taurine than did male mice of the same genotype. Plasma taurine levels in CDO−/− mice were only 7.3% of those in CDO+/+ mice. Hypotaurine levels were less than 2 nmol/mg protein in liver of mice of all genotypes and both sexes. Plasma hypotaurine levels were elevated in CDO−/− mice, with levels being 4.3 times (P < 0.001) and 3.5 times (P = 0.08) those of CDO+/+ mice for males and females, respectively (Fig. 7D).

Plasma sulfate concentration was 33% higher in CDO−/− mice than in CDO+/+ mice (P < 0.001) (Fig. 7E). Plasma sulfate levels in CDO+/− mice tended to be greater (12%) than those in CDO+/+ mice (P < 0.10).

Taurine supplementation restored taurine levels and improved preweaning survival for CDO−/− mice but did not reverse the phenotype.

Supplementation of lactating dams with taurine restored liver and plasma taurine levels in CDO−/− pups to levels similar to or greater than those of CDO+/+ pups of unsupplemented dams (Fig. 8, A and B). Taurine supplementation had no effect on hepatic CDO abundance but significantly decreased hepatic CSAD abundance in both male and female CDO−/− pups and in female CDO+/+ pups (Fig. 8, C and D). The appearance/clinical characteristics of CDO−/− mice did not appear to be affected by taurine supplementation of the dams, being similar to the early postnatal phenotype described above for mice born to unsupplemented dams.

Fig. 8.

Effect of taurine supplementation of dams during lactation on plasma and liver taurine levels and abundance of taurine synthetic enzymes in offspring at P21. Dams and their litters had access to unsupplemented water or water supplemented with 2% (wt/vol) taurine from the day pups were born until pups were euthanized at P21. A: hepatic taurine levels in male and female Null and WT pups. B: plasma taurine levels in male and female Null and WT pups. C: CDO abundance in male and female WT pups. D: CSAD abundance in male and female Null and WT pups, measured by Western blotting. C and D: each value in bar graphs is the mean ± SE for measurements on samples from individual mice. Western blots on the right are for pooled samples run on a single gel; white lines indicate where lanes for CDO+/− mice were removed from the image. A–D: mice were separated by sex for statistical analysis. Significant effects of G (WT, Null), taurine (Control unsupplemented, Taurine supplemented), and their interaction (G × T) are reported. Each bar represents the mean ± SE of 4–9 mice.

Supplementation of dams with taurine completely reversed the mortality of male CDO−/− pups but did not significantly improve survival of female CDO−/− pups. As shown in Fig. 9, taurine supplementation improved survival of male CDO−/− pups to P21 (100% for pups of taurine-supplemented dams vs. 79% for pups of unsupplemented dams), whereas taurine supplementation did not significantly improve survival of female CDO−/− pups (73% for pups of taurine-supplemented dams vs. 62% for pups of unsupplemented dams; Fig. 9B). Supplementation of dams with taurine did not improve growth (either weight or length) of CDO−/− offspring (Fig. 9, C and D). In fact, taurine supplementation of dams actually had a negative effect on weight (P < 0.01) and length (P < 0.05) of male offspring regardless of genotype.

Fig. 9.

Effect of taurine supplementation of dams during lactation on survival and growth of offspring. Dams and their litters had access to unsupplemented water or water supplemented with 2% (wt/vol) taurine from the day pups were born to P21. A: comparison of survival of male offspring of unsupplemented and taurine-supplemented dams. Survival of taurine-supplemented Null offspring was not different from that of WT and Het pups regardless of whether drinking water of dams and pups was supplemented with taurine. Unsupplemented male offspring exhibited significantly higher mortality [log-rank test, P < 0.05 (*)]. B: comparison of survival of female offspring of unsupplemented and taurine-supplemented dams. Survival of female offspring was less than that of WT and Het offspring [log-rank test, P < 0.05 (*) and P < 0.001(**)] and was not significantly improved by taurine supplementation. A and B: number of mice in each group at P0 is shown in parentheses. C: growth curves, body weight at P21, and body length at P21 for male Null and WT offspring with and without taurine supplementation. D: growth curves, body weight at P21, and body length at P21 for female Null and WT offspring with and without taurine supplementation. C and D: significant effects of G (WT vs. Null), T (control or unsupplemented vs. taurine supplemented), and their interaction (G × T) are reported. Each bar represents the mean ± SE of 4-t9 mice.

Maternal taurine supplementation significantly (P < 0.05) decreased cysteine levels of female CDO−/− pups by 37.2% in the liver and 40.3% in plasma but not those of female CDO+/+ pups or those of male pups of either genotype (Fig. 10, A and B). Taurine supplementation did not affect plasma sulfate levels in either CDO−/− or CDO+/+ pups (Fig. 10, A and B).

Fig. 10.

Effect of taurine supplementation of dams during lactation on plasma and hepatic cysteine and plasma sulfate levels of offspring at P21. Dams and their litters had access to unsupplemented water or water supplemented with 2% (w/vol) taurine from the day pups were born until P21 when pups were euthanized. A and B: plasma cysteine, hepatic cysteine, and plasma sulfate levels in control unsupplemented and taurine-supplemented male (A) and female (B) pups. Mice were separated by sex for statistical analysis. Taurine had effects on plasma and hepatic cysteine levels in female, but not male, mice. Taurine had no effect on plasma sulfate levels. Significant effects of G (WT vs. Null), T (control or unsupplemented vs. taurine-supplemented), and their interaction (G × T) are reported. Each bar represents the mean ± SE of 4–9 mice.

CDO−/− mice exhibited elevated levels of acid-labile sulfide in tissues, loss of mitochondrial cytochrome c oxidase activity, and elevated levels of blood acylcarnitines.

Tissues of 8-wk-old mice fed the sulfur amino acid-supplemented diet had elevated levels of acid-labile sulfur, which includes sulfide itself but also sulfur in persulfide or polysulfide linkages. As shown in Fig. 11, acid-labile sulfide in the pancreas of CDO−/− mice was greater than that of CDO+/+ mice (P < 0.001), with male CDO−/− mice having 2.5 times and female CDO−/− mice having 1.9 times as much acid-labile sulfur as CDO+/+ mice of the same sex. Acid-labile sulfide in lungs of male and female CDO−/− mice was 1.6 and 2.6 times that of CDO+/+ mice, respectively (P < 0.08). There was a tendency for CDO−/− mice to have a higher level of hepatic acid-labile sulfur than CDO+/+ mice (P = 0.06).

Fig. 11.

Acid-labile sulfide and hepatic mitochondrial cytochrome c oxidase activity in CDO+/+ (WT) and CDO−/− (Null) mice. WT and Null male and female mice were fed a sulfur amino acid enriched diet (20% casein + 1.5% l-methionine and 4.2% l-cystine) for 2 wk (from P49 to P63) prior to tissue collection. A: acid-labile sulfur (H2S) was measured in homogenates of liver, lung, and pancreas. B: cytochrome c oxidase activity was measured in mitochondria isolated from fresh liver samples. A and B: significant effects of G (WT, Null) and S (male, female) are reported. Each bar represents the mean ± SE of 5–9 mice.

In mammals, sulfide production occurs when CBS uses cysteine plus homocysteine and when CSE uses cyst(e)ine as substrate. Assessment of abundance of these two cysteine desulfhydrases in liver, kidney, and lung indicated that the amounts of CBS and CSE proteins were not affected by genotype.

Additional evidence in support of excess hydrogen sulfide production includes the inhibition of mitochondrial cytochrome c oxidase and short-chain acyl-CoA dehydrogenase activities, both of which are strongly inhibited by hydrogen sulfide. As shown in Fig. 10, cytochrome c oxidase activity in liver mitochondria was low in CDO−/− mice, being only 54% of that in CDO+/+ mice (P < 0.05). CDO−/− mice had significantly higher levels of C4-hydroxy, C5, C5-hydroxy, and C3-dicarboxylic acylcarnitines than did CDO+/+ mice, and this effect was not dependent on the sulfur amino acid content of the diet (Table 3). Long-chain acylcarnitines were also higher in CDO−/− mice than in CDO+/+ mice, regardless of the diet fed, and this was particularly the case for C18, C18 hydroxy, C18:1, C18:1 hydroxy, and C18:2, suggesting that CDO−/− mice may have an impaired capacity for long-chain fatty acid oxidation, not just the oxidation of short-chain fatty acids.

View this table:
Table 3.

Effect of dietary sulfur amino acid level on blood acylcarnitine profiles in WT and CDO Null mice

Histology of lungs of mice fed diet enriched in sulfur-containing amino acids.

Analysis of differentially stained (H&E) lung sections of 8-wk-old CDO−/− mice showed enlarged alveolar airspaces. The MLI for CDO−/− animals was significantly smaller than for CDO+/+mice, indicating larger alveolar airspaces in CDO−/− mice than in CDO+/+ mice (P < 0.03, n = 9 per genotype). Representative images are shown in Fig. 12A. Comparison of MTC-stained slides revealed no striking differences in collagen deposition (not shown), but organization of elastic fibers, as observed with VVG staining, was clearly abnormal in CDO−/− mice. Intima of both intra-acinar pulmonary vessels (Fig. 12B, top) and larger preacinar pulmonary vessels (Fig. 12B, bottom) exhibited thickened, fragmented, and disorganized elastic fibers in CDO−/− mice. Fragmentation and tangled disarray of elastic fibers also was observed throughout the lung parenchyma in CDO−/− animals (Fig. 12C).

Fig. 12.

A: photomicrographs of lung parenchyma demonstrating increased alveolar airspaces in lung of CDO−/− mice compared with lung of CDO+/+ mice. B: small (top) and large (bottom) pleural blood vessels, demonstrating fragmentation and disarray of elastic fibers in null mice compared with wild-type mice. C: Null lung parenchyma with blood vessel showing fragmented elastic fibers and disorganized architecture of lung parenchyma. A–C: Verhoeff Van Geison stain; bar, 50 μm for all micrographs.


Taurine levels in CDO−/− mice.

In our initial hypothesis, we predicted that the absence of the CDO pathway, which normally leads to the production of taurine, would decrease plasma and tissue levels of taurine and hypotaurine. Although the ADO-catalyzed dioxygenation of cysteamine can also produce hypotaurine, the CDO/CSAD pathway is believed to be the major pathway involved in the synthesis of hypotaurine in the body (60, 59). Hypotaurine is readily converted to taurine such that hypotaurine concentrations are maintained very low relative to taurine levels. Indeed, our CDO−/− mice exhibited dramatically reduced hepatic and plasma taurine levels compared with CDO+/+ or CDO+/− mice. The plasma taurine level in 8-wk-old CDO−/− mice was 7% of the wild-type level, and the hepatic taurine level was only 2% of that for wild-type mice. These observations confirm the dominant role of the CDO/CSAD pathway in de novo taurine biosynthesis from cysteine. Taurine production from cysteamine, which is a product of coenzyme A turnover, via the ADO pathway is clearly insufficient to maintain tissue taurine levels in the absence of dietary taurine. The small amount of taurine present in CDO−/− mice may result from efficient retention of taurine that was transferred from the dam to the pup during pregnancy or lactation, or it may represent taurine synthesized by the ADO pathway.

Taurine serves several roles in the body. It is known to play roles in conjugation of bile acids, regulation of cell volume, modulation of neurotransmitter function, antioxidation, and removal of hypochlorous acid. Reproductive impairments have been reported for taurine-deficient mice and cats (15, 65, 78). Relatively high levels of taurine and hypotaurine are found in spermatozoa, seminal plasma, and the epididymis (7). Lack of taurine could be related to our findings of male reproductive impairments in CDO−/− mice. Similarly, taurine has been shown to be needed for normal growth in some animal models. Adult TauT−/− mice lacking the taurine transporter TauT exhibited a 25% lower body mass (78), and reduced activity of placental TauT was associated with intrauterine growth restriction in mice (49). In addition, taurine levels in placentas of low-birth-weight human infants were shown to be significantly lower than that those in placentas of normal-birth-weight infants (32).

Overall, our results indicate that the lack of taurine alone does not explain the phenotype of CDO−/− mice. Supplementation of dams with taurine during lactation resulted in tissue and plasma taurine levels in CDO−/− pups that were equivalent to or greater than those of CDO+/+ pups of unsupplemented dams. However, despite restoration of normal taurine levels in the pups, taurine supplementation had little effect on mortality and no effect on growth (length or body weight) of female CDO−/− pups. Furthermore, although supplementation of dams with taurine improved the survival of male CDO−/− pups, it significantly reduced the growth of these pups. In addition, the characteristic appearance of CDO−/− offspring, other than the size of male pups, was unaffected by taurine supplementation of the dams. Reproductive performance of taurine-supplemented CDO−/− mice has not yet been tested.

Reciprocal relationship of CSAD abundance and taurine concentration in murine liver.

Our laboratory has previously reported a reciprocal relationship between changes in CDO and CSAD protein abundance in liver of male rats in response to changes in dietary protein level (1, 35). Data from this study with CDO−/− mice provide further support for this relationship and show that the reciprocal relationship between hepatic CSAD and CDO not only occurs in response to diet but is affected by sex and CDO genotype. In all cases, CSAD abundance was high when CDO abundance was low. For the effect of sex, female mice had significantly lower hepatic CSAD levels than did male mice (0.43 and 0.64 times those of males for the CDO+/+ and CDO+/− mice, respectively), whereas they had higher hepatic CDO levels than did male mice (1.8 and 2.1 times those of males for the CDO+/+ and CDO+/− mice, respectively). CDO−/− mice had no CDO, but had the highest levels of hepatic CSAD. Interestingly, no sex-dependent differences in CSAD were observed in CDO−/− mice, suggesting that the sex difference in hepatic CSAD expression may be secondary to the sex differences in hepatic CDO abundance.

In this study, we further show a direct association between CDO levels and hepatic taurine levels. Female CDO+/+ and CDO+/− mice had significantly higher levels of hepatic taurine than did male mice of the same genotype, consistent with higher CDO levels in female mice. Likewise, CDO genotypes that expressed more CDO had higher taurine levels; taurine levels were highest in CDO+/+ mice, intermediate in CDO+/− mice, and very low in CDO−/− mice. The association between CDO levels and taurine levels is consistent with our previous observations showing that an increase in CDO concentration or activity, even if CSAD activity decreases, will lead to a parallel increase in taurine synthesis by hepatocytes of intact rats (2, 62, 63). This is consistent with a major role of CDO activity, or the rate of cysteinesulfinate formation, in determination of the rate of taurine synthesis in the animal.

CSAD levels may be regulated in response to changes in taurine levels rather than to changes in taurine synthetic capacity (i.e., CDO levels) per se. The lack of CDO in the CDO−/− mice, which was associated with extremely low hepatic taurine levels, was associated with the highest observed levels of hepatic CSAD expression. Both CDO and taurine levels were higher in liver of female mice, whereas CSAD levels were lower in liver of female mice than in male mice. The idea that CSAD activity is regulated in response to changes in tissue taurine level is further supported by our finding that taurine supplementation of both CDO+/+ and CDO−/− mice resulted in lower CSAD abundance. Thus, the reciprocal relationship between CDO and CSAD levels appears to be due to the regulatory role of CDO on overall taurine biosynthesis coupled with the downregulation of CSAD when taurine levels rise. Thus, when the tissue taurine level is low, CSAD is upregulated in an effort to increase the taurine synthetic capacity. This conclusion is also supported by the previous observation that hepatic CSAD activity was increased in kittens fed a taurine-free diet for 6 wk (53).

Cysteine and sulfate levels in CDO−/− mice.

As anticipated, due to a block in cysteine catabolism by the CDO-dependent pathway, hepatic and plasma cysteine levels were higher in CDO−/− mice than in CDO+/+ mice. The plasma and hepatic cysteine levels of CDO−/− mice were 2 and 1.5 times, respectively, those of CDO+/+ mice. Similarly, hepatic and plasma GSH levels were higher in CDO−/− mice than in CDO+/+ mice, reflecting the greater availability of cysteine as substrate for GSH synthesis (17, 19, 63). Although the higher cysteine and GSH levels are consistent with the block in CDO-dependent catabolism of cysteine in CDO−/− mice, cysteine and GSH levels were only modestly elevated. Additionally, despite the block in CDO-dependent catabolism of cysteine, sulfate levels were not reduced, being somewhat higher in CDO−/− mice than in CDO+/+ mice. These observations together suggest that there was increased catabolism of cysteine by the CDO-independent desulfuration pathways in CDO−/− mice, with further oxidation of the sulfide to sulfate (62, 70).

Cysteine catabolism by desulfhydration in CDO−/− mice.

Further support for an increase in cysteine metabolism by cysteine desulfuration pathways was obtained from assessment of sulfide production by CDO−/− mice fed a diet enriched in sulfur amino acids. Lower levels of cytochrome c oxidase activity in liver mitochondria, higher levels of acid-labile sulfide in tissues (liver, pancreas, lung), and higher levels of C4, C4-hydroxy, C5, C5-hydroxy, and C3-dicarboxylic acylcarnitines in blood were observed in CDO−/− mice compared with CDO+/+ mice. The higher levels of acid-labile sulfide indicate that more cysteine was metabolized by desulfhydration pathways in CDO−/− mice, leading to sulfur accumulation at the S−2 or S0 oxidation states in compounds such as persulfides and polysulfides as well as HS. Similarly, low levels of cytochrome c oxidase in CDO−/− mice are consistent with the presence of higher levels of H2S and its consequent inhibition of cytochrome c oxidase (13). The presence of short-chain fatty acids with hydroxyacyl groups is consistent with the well-established rate-limiting nature of the hydroxyacyl-CoA dehydrogenase step of β-oxidation. The ω-oxidation of fatty acids to dicarboxylic acids and the transesterification of acyl groups with carnitine are likely due to the contributions of alternative pathways of metabolism for fatty acids when β-oxidation is blocked (66, 70). Other blood acylcarnitine concentrations were higher in CDO−/− mice than in CDO+/+ mice, particularly C18, C18 hydroxy, C18:1,C18:1 hydroxy, and C18:2 acylcarnitines. This unanticipated observation suggests that CDO−/− mice might have an impaired capacity for long-chain fatty acid oxidation, not just the oxidation of short-chain fatty acids.

It seems probable that the increases in intracellular cysteine concentration alone were sufficient to increase flux through the desulfhydration pathways in CDO−/− mice. We saw no change in abundance of either CBS or CSE, the two main enzymes responsible for cysteine desulfhydration in mammals. However, it is possible that cysteine desulfhydration capacity was increased due to changes in activity state of either enzyme, although no evidence for such changes has been reported.

Similarities between lack of CDO and lack of Ethe1 expression or activity.

The abnormalities exhibited by CDO−/− mice have some similarities to those reported for the rare human disease ethylmalonic encephalopathy caused by mutations of the Ethe1 gene as well as for the Ethe1−/− mouse model. Ethe1 was recently shown to encode the mitochondrial sulfur dioxygenase (44), which catalyzes a step in the mitochondrial pathway for oxidation of sulfide to thiosulfate (12). Both human and murine Ethe1 mutations present with similar biochemical and clinical features, which include elevated tissue sulfide levels, elevated levels of thiosulfate and ethylmalonic acid in the urine, developmental delays, growth retardation, articular hyperlaxity, and elevated blood C4 and C5 acylcarnitine levels (31, 47, 50, 69, 70). Petechiae, ecchymoses, joint hyperlaxity, and MRI abnormalities consistent with previous brain vascular insults are observed in patients, suggesting a link between sulfide production and connective tissue abnormalities (14).

Although CDO−/− mice have no genetic impairment in the removal or oxidation of sulfide, their phenotypic similarity to Ethe1−/− animals may be due to an increase in sulfide/sulfane sulfur production via increased flux of cysteine through desulfhydration pathways. H2S or another reduced sulfur compound could be a toxic mediator of the phenotypic changes common to both CDO−/− and Ethe1−/− gene disruptions. H2S is known to react with the heme of cytochrome c oxidase to inhibit mitochondrial respiration, with inhibition of cytochrome c oxidase leading to its accelerated degradation (13). Cytochrome c oxidase activity was markedly reduced in muscle, brain, and colon of individuals with ethylmalonic encephalopathy and in muscle, brain, and liver of Ethe1−/− mice (45, 70). Inhibition of short-chain acyl-CoA dehydrogenase activity by H2S is thought to be the basis of elevated levels of short-chain acylcarnitines in blood of patients or animals that lack Ethe1/sulfur dioxygenase function (66, 70).

Contribution of H2S production to tissue pathology in CDO−/− mice.

Because CDO−/− mice metabolize much more cysteine by desulfhydration pathways than do wild-type mice, excess production of H2S could be responsible for the clinical and histopathological abnormalities observed in CDO−/− mice. In mammals, H2S is produced by catabolism of cysteine by cystathionine β-synthase, cystathionine γ-lyase, or cysteine transamination plus β-mercaptopyruvate sulfurtransferase (56, 58, 61). Over the past two decades, a substantial body of work has indicated that H2S is a physiologically important signaling molecule in eukaryotes (43, 44, 75). H2S relaxes smooth muscle, at least partially due to opening of KATP channels, thereby regulating vascular tone, intestinal contractility, and myocardial contractility (27, 29, 73, 75, 76). In addition, H2S has been shown to inhibit insulin secretion by pancreatic β-cells associated with increased KATP channel activity (39, 83). In the central nervous system, H2S increases the sensitivity of N-methyl-d-aspartate receptors to glutamate in hippocampal neurons to enhance synaptic transmission (16, 40, 51). In addition, H2S may activate anti-inflammatory and antioxidant pathways and promote vascular smooth muscle cell apoptosis (41, 74, 80). Although the signaling mechanism of H2S is not fully understood, H2S is known to react with targeted cysteine residues of a variety of proteins to form cysteine persulfide residues (−CyS-SH) capable of altering the protein's biological activity (48). H2S has been shown to have protective effects against oxidative stress (55, 67), ischemia-reperfusion injury (8, 30, 84), and certain toxins (25, 68) in mammals and to increase thermotolerance and lifespan in Caenorhabditis elegans (46).

At the same time, H2S has long been known to be an extremely toxic substance for aerobic organisms, at least partially due to its ability to react with cytochrome c oxidase and other heme- or disulfide-containing proteins (21, 33, 42). H2S poisoning leads to inhibition of the mitochondrial electron transport chain and can lead to death through respiratory paralysis and pulmonary edema. Pathogenic roles of moderately elevated levels of metabolically produced H2S in disease seem likely, given the recent identification of the Ethe1 gene, which is defective in individuals with the invariably fatal disease ethylmalonic encephalopathy, as encoding the mitochondrial sulfur dioxygenase. A connection between H2S and connective tissue pathology is also suggested by similar clinical observations in patients with disorders of collagen and elastin synthesis (e.g., Ehlers-Danlos syndrome) and those with a deficiency of the mitochondrial sulfur dioxygenase (14). The possibility that the enlarged alveolar air spaces and abnormal organization of elastic fibers in the vasculature and parenchyma of lungs from CDO−/− mice is due to the endogenously produced H2S should be further investigated.

Relevance of CDO knockout model and future questions.

Studies of H2S physiology in mammals have so far focused on use of the cystathionine β-synthase and cystathionine γ-lyase knockout mouse models with underproduction of H2S. Mutant mice lacking cystathionine γ-lyase display pronounced hypertension, diminished endothelium-dependent vasorelaxation, and increased smooth muscle proliferation rates (81, 82). Mice lacking cystathionine β-synthase have altered hippocampal long-term potentiation (24). The recently reported Ethe1 knockout mouse provides a model for sulfide accumulation under conditions in which sulfide oxidation or removal is impaired (70). Whereas the knockout of cystathionine β-synthase and cystathionine γ-lyase result in the accumulation of transsulfuration pathway metabolites (i.e., homocysteine and cystathionine) along with decreased synthesis of H2S, knockout of CDO results in decreased synthesis of taurine along with increased synthesis of H2S and mild accumulation of cysteine. Additional studies will be required to sort out the relative contributions of deficient taurine production vs. excess H2S production or elevated cysteine levels on the observed outcomes for the CDO−/− mouse. Dietary manipulations to reduce the sulfur amino acid load or to add supplemental taurine should facilitate these studies.

The hypothesized link between loss-of-function mutations in CDO and the incidence of rheumatoid arthritis was largely based on observed increases in the cysteine-to-sulfate ratio in the plasma and low levels of sulfate in synovial fluid of patients with this disease (6, 11, 23, 35). The observations on the CDO−/− mouse, however, suggest that loss-of-function mutations of CDO do not cause a depletion of sulfate, at least in mice. Cysteine-to-sulfate ratios in CDO−/− mice did increase slightly, due to the greater accumulation of cysteine than of sulfate, whereas the cysteine-to-taurine ratios were dramatically higher in CDO−/− mice and would seem to be a better indicator of lack of CDO activity.

Also of potential interest for future studies is the strong sexual dimorphism of cysteine metabolism in male and female mice that was observed in this study. In wild-type mice, CDO and taurine levels were higher in female mice, suggesting that more cysteine is catabolized by this pathway in female mice. When the CDO gene was disrupted, female mice displayed a notably higher incidence of postnatal mortality than did male mice. Supplementation of dams with taurine improved weight gain and lowered cysteine levels in female CDO−/− offspring but not in male CDO−/− offspring. Whether sex differences in cysteine metabolism could play any role in the higher incidence of rheumatoid arthritis and certain other autoimmune diseases in women than in men would be interesting to consider.


This project was supported by Grant DK-056649 from the National Institute of Diabetes and Digestive and Kidney Diseases. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.


No conflicts of onterest are reported by the author(s).


Dr. Rachel Peters is currently at Tufts Cummings School of Veterinary Medicine, North Grafton, MA 01536.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
  70. 70.
  71. 71.
  72. 72.
  73. 73.
  74. 74.
  75. 75.
  76. 76.
  77. 77.
  78. 78.
  79. 79.
  80. 80.
  81. 81.
  82. 82.
  83. 83.
  84. 84.
View Abstract