Protein arginine methylation is catalyzed by a family of enzymes called protein arginine methyltransferases (PRMTs). Three forms of methylarginine have been identified in eukaryotes: monomethylarginine (l-NMMA), asymmetric dimethylarginine (ADMA), and symmetric dimethylarginine (SDMA), all characterized by methylation of one or both guanidine nitrogen atoms of arginine. l-NMMA and ADMA, but not SDMA, are competitive inhibitors of all nitric oxide synthase isoforms. SDMA is eliminated almost entirely by renal excretion, whereas l-NMMA and ADMA are further metabolized by dimethylarginine dimethylaminohydrolase (DDAH). To explore the interplay between methylarginine synthesis and degradation in vivo, we determined PRMT expression and DDAH activity in mouse lung, heart, liver, and kidney homogenates. In addition, we employed HPLC-based quantification of protein-incorporated and free methylarginine, combined with immunoblotting for the assessment of tissue-specific patterns of arginine methylation. The salient findings of the present investigation can be summarized as follows: 1) pulmonary expression of type I PRMTs was correlated with enhanced protein arginine methylation; 2) pulmonary ADMA degradation was undertaken by DDAH1; 3) bronchoalveolar lavage fluid and serum exhibited almost identical ADMA/SDMA ratios, and 4) kidney and liver provide complementary routes for clearance and metabolic conversion of circulating ADMA. Together, these observations suggest that methylarginine metabolism by the pulmonary system significantly contributes to circulating ADMA and SDMA levels.
- protein arginine methyltransferases
- asymmetric dimethylarginine
- dimethylarginine dimethylaminohydrolase
three forms of methylarginine have been identified in eukaryotes: NG-monomethylarginine (l-NMMA), NGNG- (asymmetric) dimethylarginine (ADMA), and NGN′G- (symmetric) dimethylarginine (SDMA), all characterized by methylation of one or both guanidine nitrogen atoms of arginine. Free methylarginine is generated by proteolysis of posttranslationally methylated tissue proteins. l-NMMA and ADMA, but not SDMA, are competitive inhibitors of all nitric oxide synthase (NOS) isoforms (29, 36) and thus modulates the biological effects of nitric oxide (NO), particularly in the cardiovascular system (19, 20, 36). Several studies have suggested that ADMA plasma levels are a diagnostic marker of endothelial dysfunction in cardiovascular diseases (4, 24, 38).
Methylation of the terminal nitrogen atom(s) of arginine is catalyzed by a family of enzymes called protein arginine methyltransferases (PRMT). PRMT activity was initially identified by Paik and Kim (28), and PRMTs are now divided into two groups, type I or type II, according to substrate and product specificity. Type I and II PRMTs both form l-NMMA, possibly as an intermediary species before the formation of the dimethylated species, but differ in that type I enzymes produce ADMA, whereas type II enzymes produce SDMA (8, 21). Nine PRMT genes have been cloned to date and are termed PRMT1 to PRMT9. In vitro methyltransferase activity has been demonstrated for all PRMT gene products except PRMT2 (9, 15, 17).
Up to now, the biological impact of protein arginine methylation remains to be determined, but this process is thought to be involved in the regulation of RNA export, control of transcription, DNA repair, protein localization, protein-protein interaction, signal transduction, and recycling of receptors (17, 21, 35). In vivo substrates for type I PRMTs include histones and RNA-binding proteins including heterogeneous nuclear ribonuclear protein (hnRNP) A1, fibrillarin, and nucleolin. Methylation of RNA-binding proteins occurs in all cells. The hnRNPs are a group of proteins that regulate mRNA maturation, stability, and export to the cytoplasm, and their activity is regulated by methylation. Approximately 65% of total cellular ADMA occurs in hnRNPs (3). Large-scale proteomic approaches have revealed a potentially broad range of substrate proteins for PRMT methylation (5, 27), suggesting a significant role for arginine methylation in cellular processes. Synthesis and degradation of methylated proteins is closely coupled to protein synthesis and degradation (23). Thus it is assumed that upon proteolysis of these PRMT substrates, significant amounts of free methylarginine are released in the cytoplasm. Intracellular ADMA and SDMA levels may therefore be determined by PRMT activity and protein turnover as direct methylation of free arginines has thus far not be demonstrated.
To date, three pathways for the in vivo elimination of methylarginine have been identified: 1) renal excretion (11), 2) degradation of dimethylarginine by pyruvate aminotransferase (25), and 3) specific metabolic conversion of ADMA and l-NMMA to citrulline and dimethylamine or monomethylamine. The enzyme responsible for this specific pathway was purified from rat and human tissues and is called NG,NG-dimethylarginine dimethylaminohydrolase (DDAH) (13, 16). Subsequent studies have shown that >90% of ADMA is metabolized by DDAH. DDAH activity was found in kidney, pancreas, liver, and brain, and this correlates well with the high-protein expression levels as revealed by immunoblotting analyses in rat tissues (14). A study of DDAH expression in the rat kidney also revealed colocalization of DDAH with NOS at several anatomical sites (34). Thus the substrate specificity of DDAH and its distribution in NO-generating systems supported the idea that regulation of intracellular ADMA levels by DDAH might in turn regulate NOS activity.
To date, the interplay between methylarginine synthesis and degradation in vivo has not been described. Thus we set out to determine PRMT and DDAH activity in lung, heart, liver, and kidney homogenates. To this end, we employed HPLC-based quantification of protein-incorporated and free methylarginine, combined with immunoblotting and ADMA degradation assays, for the assessment of tissue-specific patterns of arginine methylation.
MATERIALS AND METHODS
Sample preparation and protein hydrolysis.
Tissues from C57/bl6 mice were surgically excised after thoracotomy and immediately homogenized in liquid nitrogen followed by addition of ice-cold cell lysis buffer [20 mM Tris·Cl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% (vol/vol) Triton X-100, 2 mM Na3VO4]. Homogenized tissue was incubated for 1 h on ice and centrifuged for 15 min at 16,000 g. The resulting supernatant was stored at −20°C. Proteins were then precipitated by mixing 100 μl of tissue extract with an equal volume of 20% (vol/vol) trichloroacetic acid for 20 min. After centrifugation at 16,000 g for 12 min, the supernatants were aspirated, and the protein pellets were washed with 100 μl of ice-cold acetone for 60 min at −20°C. The suspension was centrifuged (16,000 g for 12 min), and the resulting protein pellet was dissolved in 100 μl of distilled water. Protein concentrations of crude extracts and precipitated proteins were determined by Quick Start Bradford Dye Reagent using a SmartSpecTM 3000 spectrophotometer (both Bio-Rad Laboratories, Munich, Germany). Before protein hydrolysis, 20 μl of each sample was combined with 10 μl of l-homoarginine (4 pmol/μl) as an internal standard. Total hydrolysis of precipitated protein fractions was achieved by gas-phase hydrolysis with 6 M HCl (constant boiling, sequencing grade; Pierce, Bonn, Germany) at 110°C for 16 h. Samples were dried by use of a vacuum centrifuge and stored at −20°C until further analyzed.
Human serum and bronchoalveolar lavage.
This study was approved by the local ethics committee, and informed consent was obtained from each individual before entering into the study and subsequent bronchoscopy. Flexible fiberoptic bronchoscopy was performed in a standardized manner as previously described (10). The examined group consisted of seven healthy volunteers (age: 37 ± 2 yr; sex, M/F: 4/3) who had never smoked and had no history of cardiac or pulmonary disease. They also showed normal chest X-rays and pulmonary function tests.
Isolation of basic amino acids and derivatization.
Tissue extracts, serum, bronchoalveolar lavage (BAL) fluid, and amino acid hydrolysates were subjected to crude fractionation on Oasis MCX solid-phase extraction (SPE) cartridges (30 mg, 1 ml; Waters, Eschborn, Germany). For tissue extracts, 30 μl of each sample was combined with 10 μl of l-homoarginine (4 pmol/μl) as an internal standard and adjusted to a final volume of 1 ml with PBS (PAA Laboratories, Pasching, Austria). Amino acid hydrolysates were directly dissolved in 1 ml of PBS. All conditioning, washing, and elution steps were performed on a vacuum manifold with a capacity for 24 columns (Waters) at a flow rate of ∼0.5 ml/min. The SPE cartridges were conditioned with 2 ml of methanol/water/ammonia (50:45:5, vol/vol/vol) followed by 2 ml of PBS before sample application. Samples were passed through SPE cartridges, and contaminating components were removed with 2 ml of 0.1 M HCl followed by 2 ml of methanol. Basic compounds were eluted with 1 ml of methanol/water/ammonia (50:45:5, vol/vol/vol). Samples were dried in a vacuum centrifuge and stored at −20°C until further analyzed. Eluates were redissolved in 230 μl of distilled water and centrifuged at 14,000 g for 2 min to remove particulates before derivatization for HPLC.
Derivatization and chromatographic separation.
Ortho-phthaldialdehyde (OPA) reagent was freshly prepared in potassium borate buffer (both Grom, Rottenburg-Hailfingen, Germany) according to the manufacturer's instructions. Samples (115 μl) were combined with 50 μl of OPA reagent, immediately transferred to the auto sampler, and injected after exactly 2 min.
Quantification of amino acids was performed on an HPLC system consisting of an ASI-100 auto sampler, a model P680 gradient pump, a model RF-2000 fluorescence detector, and a data acquisition system (Chromeleon, version 6.60; all Dionex, Idstein, Germany). Separation was carried out according to the method described by Teerlink et al. (33) with slight modifications. Fluorescent amino acid derivatives were separated on a SunFire C18 column (4.6 × 150 mm; 3.5-μm particle size; 100 Å pore size) with a μBondapak C18 guard column at 30°C and a flow rate of 1.1 ml/min (all columns were from Waters). After sample injection (125 μl), separation was performed under isocratic conditions with 8.8% (vol/vol) acetonitrile in 25 mM potassium phosphate buffer (pH 6.8) as solvent. The isocratic conditions were maintained for 24 min. To elute strongly bound compounds, the column was flushed with acetonitrile/water (50:50, vol/vol) for 5 min and reequilibrated under isocratic conditions for 15 min before the next injection. Fluorescent derivatives were detected at excitation and emission wavelengths of 330 and 450 nm, respectively. l-Arginine, ADMA, and SDMA were quantified by two separation steps. For the detection of ADMA and SDMA, the gain of the detector was switched to a hundred-fold higher sensitivity.
Calibration was performed as described previously (6). Briefly, six combined standards spanning the range 1.5 to 450 pmol (60 nM to 18 μM) for l-Arg, 0.15 to 45 pmol (6 nM to 1.8 μM) for l-NMMA and ADMA, and 0.09 pmol to 9 pmol (3.6 nM to 0.36 μM) for SDMA (all Sigma, St. Louis, MO) were used for calibration. Standard solutions were combined with 15 pmol of the internal standard l-homoarginine and subjected to SPE, derivatization, and chromatography as described above. Calibration curves were calculated by plotting the ratio of the peak area of analyte (arginine, l-NMMA, ADMA or SDMA) to the area of the internal standard (l-homoarginine) vs. analyte quantity. Since l-homoarginine is highly abundant in crude liver extract, quantification of methylarginine was performed by external calibration. Statistical analysis was performed using Student's t-test.
Measurement of ADMA degradation.
For tissue extracts and serum, 25 μl of each sample was combined with 6 μl of combined ADMA/SDMA solution (500 pmol/μl each) and adjusted to a final volume of 0.5 ml with 0.1 M sodium phosphate buffer (pH 6.5) (26). After incubation for 2 h at 37°C, samples were directly subjected to crude fractionation on Oasis MCX cartridges, HPLC separation, and fluorescence detection as described above. As assay blank, ADMA and SDMA from the crude tissue extracts were directly quantified.
Determination of NO concentration.
The NO concentrations of tissue extracts were determined using QuantiChrom Nitric Oxide Assay Kit (BioAssay Systems, Hayward, CA) according to the manufacturer's instructions.
Western blot analysis.
Equal amounts of protein extracts (30 μg) were separated on 10% SDS-PAGE gels and transferred to PVDF-PLUS membranes (Osmonics, Moers, Germany). Western blots were performed with antibodies against PRMT1 (at a dilution of 1:2,000; Upstate, Dundee, UK), PRMT2 (1:1,000; Abcam, Cambridge, UK), PRMT3 (1:2,000; Upstate), PRMT4 (1:1,000; Upstate), PRMT5 (1:2,000; Upstate), PRMT6 (1:500; Imgenex, San Diego, CA), PRMT7 (1:1,000; Upstate), and DDAH1/2 [1:1,000 (22)]. After incubation with the respective secondary antibodies, specific bands were visualized by autoradiography using enhanced chemiluminescence according to the manufacturer's instructions (LUMIGEN; Amersham, Buckinghamshire, UK). The specific antibodies did not exhibit any cross-reactivity with other PRMT isoforms.
Densitometric analysis of autoradiographies was performed using a GS-800 Calibrated Densitometer and the 1-D analysis software Quantity One (both Bio-Rad Laboratories).
Methylarginine and NO content of mouse lung, heart, liver, and kidney.
We first sought to establish whether different organs exhibited distinct methylation characteristics. As such, we measured the concentration of free cellular and protein-incorporated methylarginine in crude tissue extracts and protein hydrolysates, respectively, derived from mouse lung, heart, liver, and kidney. As depicted in Fig. 1A, we detected similar free ADMA and SDMA levels in lung (ADMA: 0.010 ± 0.004 nmol/mg protein; SDMA: 0.0017 ± 0.0005 nmol/mg protein) and heart (ADMA: 0.014 ± 0.004 nmol/mg protein; SDMA: 0.0026 ± 0.0005 nmol/mg protein) crude tissue extracts, whereas kidney (ADMA: 0.181 ± 0.04 nmol/mg protein; SDMA: 0.033 ± 0.012 nmol/mg protein) and liver (ADMA: 0.088 ± 0.02 nmol/mg protein; SDMA: 0.011 ± 0.003 nmol/mg protein) exhibited significantly higher concentrations for both dimethylarginines. As expected, crude kidney extracts exhibited the highest levels of free ADMA and SDMA (n = 6, P < 0.001), supporting the idea that the kidneys provide the main route for clearance of both methylarginines. Surprisingly, liver extracts displayed four- to eightfold higher levels of free cellular ADMA and SDMA compared with lung and heart (n = 6, P < 0.001). Because of the low cellular arginine concentration, liver homogenates exhibited a dramatically increased arginine/ADMA ratio of ∼2, which is significantly higher than the ratios observed in lung, heart, and kidney.
In contrast, arginine residues in protein hydrolysates of the lung exhibited a fourfold higher degree of asymmetrical and a twofold higher degree of symmetrical dimethylation (ADMA: 4.23 ± 2.19 nmol/mg protein; SDMA: 0.37 ± 0.17 nmol/mg protein) compared with heart (ADMA: 1.11 ± 0.53 nmol/mg protein; SDMA: 0.22 ± 0.07 nmol/mg protein), kidney (ADMA: 1.41 ± 0.26 nmol/mg protein; SDMA: 0.27 ± 0.03 nmol/mg protein), or liver (ADMA: 1.05 ± 0.42 nmol/mg protein; SDMA: 0.21 ± 0.07 nmol/mg protein) hydrolysates (n = 6, P < 0.05), which exhibited almost identical levels of protein-incorporated ADMA and SDMA (Fig. 2). In conclusion, we found that, in the lung, 1.4% of all protein-incorporated arginine residues were asymmetrically dimethylated, whereas all other tissues represented levels of 0.3–0.6%. We were unable to detect any monomethylation of arginine residues in either hydrolysates or crude tissue lysates.
In addition, we measured the concentration of NO in crude tissue extracts. As depicted in Fig. 1B, we detected similar NO levels in lung (0.66 ± 0.23 nnmol/mg protein) and heart (0.81 ± 0.13 nnmol/mg protein) crude tissue extracts, whereas kidney extracts (0.32 ± 0.20 nnmol/mg protein) exhibited lower concentrations and liver extracts (1.10 ± 0.86 nnmol/mg) higher concentrations for NO.
Protein expression of PRMT1–7.
To determine whether the increased concentration of protein-incorporated ADMA observed in the mouse lung correlated with increased expression of type I PRMTs, immunoblot analysis was performed on crude tissue homogenates. The expression of PRMT isoforms in selected tissues is illustrated in Fig. 3. Densitometric analysis revealed that mouse lungs expressed significantly higher levels of PRMT1, 2, and 6 compared with heart, kidney, and liver (n = 3, P < 0.05).
Methylarginine content of serum and BAL fluid.
The increased type I PRMT expression and activity observed in the mouse lung may result in significant levels of free cellular ADMA after protein breakdown, leading to a release of ADMA in the BAL fluid. To determine whether this was the case, we quantified ADMA in mouse and human BAL fluid by HPLC analysis (Table 1). Both ADMA and SDMA were detected in mouse and human BAL fluid. Interestingly, the BAL fluid of mouse and human exhibited an ADMA/SDMA ratio similar to the ratio observed in the respective sera (see Table 1).
DDAH protein expression and ADMA degradation.
Western blot analysis of tissue homogenates indicated a tissue-specific expression of DDAH isoforms (Fig. 4A). Densitometry revealed that DDAH1 expression in kidney and liver was significantly higher compared with lung. DDAH1 was not detected in heart lysates. In contrast, DDAH2 was equally expressed in liver, lung, and heart but significantly less in kidney lysates (Fig. 4B). To determine whether the different patterns in DDAH protein expression were correlated with enzymatic activity, in vitro ADMA degradation was measured in crude tissue lysates (Fig. 5). Kidney lysates exhibited the highest activity compared with liver and lung. Moreover, degradation activity was found to be significantly higher in liver vs. lung. ADMA degradation was not detected in heart lysates and serum. Significant degradation of SDMA was not detected in any of the tissues investigated.
Arginine methylation of proteins is catalyzed by the action of PRMTs. There is potentially a broad range of target proteins for both types of PRMTs, and the enzymes and their substrates are widely distributed throughout the body (5, 37). In the cardiovascular system, the expression of PRMT has only been reported for type I PRMTs in the heart, smooth muscle cells, and endothelial cells (35). Initially, we characterized the expression of PRMTs in lung, heart, liver, and kidney. We found that mouse lungs expressed significantly higher levels of PRMT1, 2, and 6 compared with the heart, kidney, or liver. To elucidate whether the increased pulmonary expression of type I PRMTs correlated with increased asymmetrical dimethylation of lung proteins, protein hydrolysis of tissue proteins and HPLC-analysis were performed. In lung protein hydrolysates, we found that arginine residues exhibited an almost fourfold higher degree of asymmetrical and twofold higher degree of symmetrical dimethylation compared with arginine residues from heart, kidney, and liver tissue.
Free cellular methylarginine levels are dependent on PRMT activity, the rate of protein degradation, the rates of ADMA metabolism by DDAHs, degradation of ADMA and SDMA by pyruvate aminotransferase, and the rates of active methylarginine uptake and release. We found similar free cellular ADMA and SDMA levels in the lung and heart, whereas kidney and liver exhibited significantly higher concentrations for both dimethylarginines. Kidney lysates exhibited the highest levels of both dimethylarginines. Because kidney proteins did not display a higher degree of protein-incorporated methylarginine, the increased ADMA and SDMA levels may be a result of active renal methylarginine uptake, as suggested in previously published studies (1, 18, 36). Kidney homogenates exhibited significantly higher DDAH1 expression and ADMA degradation activity compared with liver, lung, or heart homogenate, suggesting that the kidney provides the main route for clearance and metabolic conversion of circulating methylarginines. Moreover, kidney lysates displayed the highest capacity for ADMA degradation, albeit DDAH2 expression was significantly lower than that observed in the liver, lung, or heart. Furthermore, no significant degradation of SDMA (as a consequence of pyruvate aminotransferase activity) was observed. These findings suggest that renal metabolism of ADMA is due to the metabolic action of DDAH1 and not DDAH2 and pyruvate aminotransferase.
This is the first report on the direct characterization of methylarginine metabolism in the liver. Liver lysates displayed four- to eightfold higher levels of free cellular ADMA and SDMA compared with the lung and heart. Because liver proteins did not exhibit a higher degree of protein-incorporated methylarginine, the increased ADMA and SDMA levels may be a result of active hepatic methylarginine uptake by the y+ transporter, as suggested in previously published studies (31, 32). Liver lysates also exhibited significantly higher DDAH1 expression and ADMA degradation activity than did lysates from lung and heart tissue. A previously published study on the characterization of ADMA clearance in rat plasma after nephrectomy has suggested that ADMA does not require the kidney for its elimination from the plasma (7). Moreover, transplanted liver graft is capable of clearing circulating ADMA in human patients (32). Thus our data suggest that the liver provides an alternative route for the clearance and metabolic conversion of circulating ADMA.
Together, kidney and liver tissues exhibited a high capacity for ADMA degradation, supporting the idea that both organs provide complementary routes for clearance and metabolic conversion of circulating ADMA.
ADMA is a competitive inhibitor of all NOS isoforms and thus modulates the formation of NO, particularly in the cardiovascular system. We found similar NO levels in lung and heart tissue extracts, whereas kidney extracts exhibited lower, and liver extracts exhibited higher, concentrations of NO. Taking the tissue-specific ADMA levels into account, a connection between NO formation and free cellular ADMA levels was not evident. Thus the increased methylarginine levels in the kidney and liver do not contribute to the regulation of NOS activity but rather reflect their function of metabolic conversion of circulating methylarginine.
Compared with the heart, pulmonary expression of PRMTs was significantly increased and correlated with enhanced asymmetrical and symmetrical dimethylation of proteins in the lung. In contrast, we found similar levels of free cellular ADMA and SDMA in lung and heart. Lung homogenates exhibited DDAH1 expression and ADMA degradation activity, which was not detected at all in heart lysates, suggesting that the lung is capable of metabolic conversion of free cellular methylarginine. Heart lysates did not display a capacity for ADMA degradation, although their DDAH2 expression was slightly higher than those of lung lysates. Therefore, relevant pulmonary degradation of ADMA is a result of DDAH1 activity, whereas no contribution of DDAH2 is evident. Furthermore, ADMA and also SDMA were detected in mouse and human BAL fluid. The BAL fluid of mouse and human exhibited an ADMA/SDMA ratio is similar to the ratio observed in the respective sera, suggesting that the similar levels are a result of diffusion through paracellular spaces. Circulating ADMA levels are raised in patients with pulmonary hypertension and in experimental models of pulmonary hypertension (2, 12, 22, 30). In the rat model of chronic hypoxia-induced pulmonary hypertension, the effect was caused by a decreased expression and activity of DDAH1 (22). Thus the overall pulmonary ADMA output reflects a balance of PRMT activity, rates of protein turnover, intracellular DDAH1 activity, and active extrusion from the cell.
The relative contribution of each component remains to be determined, but the salient findings of the present investigation can be summarized as: 1) pulmonary expression of type I PRMTs was correlated with enhanced protein arginine methylation of the lung proteome; 2) pulmonary ADMA degradation was undertaken by DDAH1; and 3) BAL fluid and serum exhibited almost identical ADMA/SDMA ratios. Together, these observations suggest that methylarginine metabolism by the lung significantly contributes to circulating ADMA and SDMA levels.
This study was supported by Deutsche Forschungsgemeinschaft Grant DFG-SFB547 (to F. Grimminger and O. Eickelberg) and the Alexander von Humboldt Foundation (Sofja Kovalevskaja Award) (to O. Eickelberg). D. Zakrzewicz and K. Kitowska were supported by predoctoral fellowships of the International Graduate Program Molecular Biology and Medicine of the Lung.
We are indebted to Drs. Werner Seeger and Rory E. Morty for critical reading of the manuscript; Günter Lochnit for performing protein hydrolysis; and all members of the Eickelberg Laboratory for valuable discussions.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society