Anal. Chem. 1999, 71, 4886-4891
A Technique for the Determination of Trimethylamine-N-oxide in Natural Waters and Biological Media Angela D. Hatton*,†,§ and Stuart W. Gibb‡
School of Environmental Sciences, University of East Anglia, Norwich, NR5 7TJ, U.K., Scottish Association for Marine Science, P.O. Box 3, Oban Argyll, PA 344AD Scotland, and Plymouth Marine Laboratory, Prospect Place, Plymouth, PL1 3DH, U.K.
Trimethylamine-N-oxide (TMAO) is a nitrogenous osmolyte widely distributed in marine organisms. The reduction of TMAO to TMA has long been implicated as characteristic reaction associated with fish and seafood spoilage. However, it is now apparent that, in the marine environment, TMAO can act as precursor to a range of reduced nitrogenous biogases that can play a significant role in the biogeochemical cycle of nitrogen and in the regulation of atmospheric pH. Although methods exist for the analysis of TMAO in some biological samples, they lack the sensitivity required for measurement of TMAO in natural waters. Here we present a new, safe and sensitive method for the determination of TMAO in aqueous and biological media, where TMAO is enzymatically reduced to TMA and subsequently quantified using Flow Injection Gas Diffusion-Ion Chromatography (Gibb et al. J. Autom. Chem. 1995, 17 (6), 205-212). The limit of detection was calculated to be 1.35 nmol dm-3 TMAO, and the response was linear for both fresh and seawater (R2 ) 0.996 and 0.993, respectively). Precision (RSD) for standards in the range 40-600 nmol dm-3 was within 3%. The specificity and competitive inhibition of the enzyme are addressed and the applicability of the technique demonstrated through analysis of a number of natural water and biological samples. Trimethylamine-N-oxide (TMAO) is widely distributed in marine organisms such as bacteria, algae, zooplankton and fish where together with quaternary amines (QAs) such as choline, glycine betaine (GBT), and methylamines (MAs) it is proposed to function in osmoregulation.2-4 Like several amino acids, TMAO is accumulated by cells in response to salinity or water stresses, * Corresponding author. Fax: 44 1631 565 518. † University of Easy Anglia. ‡ Plymouth Marine Laboratory. § Current address: Scottish Association for Marine Science (SAMS), Dunstaffnage Marine Laboratory, P.O. Box 3, Oban, Argyll, PA34 4AD, U.K. (1) Gibb, S. W.; Wood, J. W.; Mantoura, R. F. C. J. Autom. Chem., 1995, 17 (6), 205-212. (2) King, G. M. In Nitrogen Cycling in Coastal Marine Environments; Blackburn, T. H., Sorenson, J., Eds.; John Wiley & Sons: Chichester, 1988. (3) Yancey, P. H.; Clarke, M. E.; Hand, S. C.; Bowlus, R. B.; Somero, G. B. Science (Washington, D.C.) 1982, 217, 1214-1222. (4) Hachachka, P. W.; Somero, G. N. In Biochemical Adaptation; Princeton University Press: Princeton, 1984.
4886 Analytical Chemistry, Vol. 71, No. 21, November 1, 1999
and there it acts as a compatible solute that does not disrupt protein structure or inhibit enzyme activity at high concentration.2 TMAO is also thought to be involved in the regulation of cellular nitrogen toxicity,2 to contribute to the positive buoyancy of marine animals,5 and to act as a cryo-protectant in cold-acclimated fish.6 Over many decades, much interest has been focused on the occurrence and significance of TMAO in marine fish and shellfish. During storage, TMAO can be reduced to trimethylamine (TMA) through its role as an electron acceptor in the anaerobic metabolism of a number of marine bacteria, including Pseudomonas and Alteromonas sp.7,8 In addition, TMAO can be demethylated to dimethylamine (DMA) by enzymes endogenous in fish muscle.9,10 TMA and DMA are two of the volatile bases responsible for the characteristic smell of spoiled sea fish and, hence, often used as subjective organoleptic indicators of seafood quality. Thus the determination of TMAO may represent a useful quality control index of spoilage of commercially important fish and shellfish, especially as TMAO levels are not affected by washing the fish, unlike other commonly used spoilage indicators, such as total volatile bases (TVBs) which may be less reliable.11 More recently, in studies of the biogeochemical cycling of nitrogen in the marine environment, it has been suggested that the breakdown of TMAO and quaternary amines (e.g., choline and glycine betaine) may be a significant source of ammonia (NH3) and methylamines (MAs). NH3 and MAs are biogenic, reduced nitrogenous compounds widely distributed in the marine environment and intimately involved in oceanic nitrogen fertility.2,12-13 In addition, they are capable of evasion across the seaair interface and may be an important source of alkali to the troposphere, thus playing a significant role in the regulation of atmospheric pH, particularly in remote marine environments.14-16 (5) Withers, P. C.; Morrison, G.; Hefter, G. T.; Pang, T. J. Exp. Biol. 1994, 188, 175-189. (6) Raymond, J. A. Fish Physiol. Biochem.; 1994, 13, 13-22. (7) Beatty, S. A. J. Fish Res. Board Can. 1938, 4, 63-68. (8) Barrett, E. L.; Kwan H. S. Ann. Rev. Microbiol. 1985, 39, 131-149. (9) Banda, M. C. M.; Hultin, H. O. J. Food Process. Preserv. 1983, 7, 221-236. (10) Sotelo, C. G.; Pioneer, C.; Perezmartin, R. I. Z. Lebensm.-Unters.-Forsch. 1995, 1, 14-23. (11) Sadok, S.; Uglow, R. F.; Haswell, S. J. Anal. Chim. Acta 1996, 334, 279285. (12) Carpenter, E. J.; Capone, D. G. In Nitrogen in The Marine Environment; Academic Press: London, 1983. (13) Lee, C. In Nitrogen Cycling in Coastal Marine Environments; Blackburn, T. H., Sorenson, J., Eds.; John Wiley: Chichester, 1988. 10.1021/ac990366y CCC: $18.00
© 1999 American Chemical Society Published on Web 10/01/1999
Although TMAO would appear to be a significant component in the marine nitrogen cycle, its role has yet to be defined since few reliable measurements exist. Methods for the determination of TMAO in biological samples include the Conway microdiffusion assay,17 proton nuclear magnetic resonance spectroscopy (NMR),18 and gas chromatography.19,20 Alternatively, chemical reduction of TMAO to TMA using acidified titanium chloride (TiCl3)11 or ferrous sulfate (FeSO4),21 with subsequent analysis of the TMA by flow injection, may be used. However, all these techniques lack the sensitivity required for the measurement of TMAO in many environmental matrixes, including natural waters. Here we report on the development and application of a highly specific enzyme-linked method for the reliable quantification of TMAO at nanomolar concentrations in natural waters and in environmental and biological media. EXPERIMENTAL SECTION Standards and Reagents. All reagents and standards were prepared in water taken freshly from a Milli-Q Water Purification System (Millipore, U.K., specific resistivity of >18 MΩ). Primary and secondary standards of TMAO (0.1 mol dm-3 and 0.1 mmol dm-3, respectively, TMAO; Sigma, UK) were prepared and further diluted to give working standards of 20-1200 nmol dm-3 TMAO for use in calibration. Ammonia and methylamine stock solutions (0.10 mol dm-3) were prepared from their hydrochloride salts (Fluka). Single and mixed standards were prepared daily from these stocks. Sec-butylamine (s-BA), an internal standard, was prepared from serial dilution of laboratory-grade reagents (Sigma, U.K.). The alkaline-chelating reagent, 1.1 mol dm-3 ethylenediamine N,N,N′,N′, tetraacetic acid (EDTA)/0.11 mol dm-3 sodium hydroxide (NaOH), was prepared from the tetra-sodium salt of EDTA and ACS grade NaOH pellets (Sigma, U.K.) and filtered (GF/F; Whatman) before use. Eluent (40 mmol dm-3 methane sulphonic acid) was routinely prepared via a 1.0 mol dm-3 stock from “Aristar” grade (Fluka). All standard solutions were stored in volumetric flasks in the dark for a maximum of one week. A mixed reducing solution of flavin mononucleotide (FMN, 0.54 µmol dm-3) and EDTA (30 mmol dm-3) was prepared from the FMN sodium salt and EDTA disodium salt (99%). A number of compounds were tested for possible interference with the enzyme-linked reduction of TMAO to TMA. These were dimethyl sulfoxide (DMSO), GBT, and choline. FNM, EDTA, DMSO, GBT, and choline were obtained from Sigma, U.K. Purification of the Enzyme. Enzymes that catalyze the reduction of TMAO have previously been purified from the periplasmic space of both Escherichia coli22 and S. putrefaciens.23 Therefore, work initially focused on the partial purification of (14) Quinn, P. K.; Charlson, R. J.; Zoller, W. H. Tellus 1987, 39B, 413-425. (15) Quinn, G. K.; Charlson, R. J.; Bates, T. S. Nature (London) 1988, 335, 120121. (16) Van Neste, A.; Duce, R. A.; Lee, C. Geophys. Res. Lett. 1987, 14, 711-714. (17) Conway, E. J.; Byrne, A. Biochem. J. 1933, 27, 419-429. (18) Anthoni, U.; Chritophersen, C.; Nielsen, P. H. J. Agric. Food Chem. 1989, 37, 705-707. (19) Fiddler, W.; Doerr, R. C.; Gates, R. A. J. Assoc. Off. Anal. Chem. 1991, 74, 400-403. (20) Zhang, A. Q.; Mitchell, S. C.; Ayesh, R.; Smith, R. L. J. Chromatogr.-Biomed. Appl. 1992, 584, 141-145. (21) Wekell J. C.; Barnett, H. J. J. Food Sci. 1991, 56, 132-135. (22) Weiner, J. H.; MacIsaac, D. P.; Bishop, R. E.; Bilous, P. T. J. Bacteriol. 1988, 170, 1505-1510. (23) Clarke, G. J.; Ward, F. B. J. Gen. Microbiol. 1988, 133, 379-386.
TMAO reductase from these two bacteria. Cultures of E. coli and S. putrefaciens were grown anaerobically in the presence of 30 mmol dm-3 TMAO. E. coli HB101 was grown on LB medium (pH 7.0) containing bacto tryptone 1%, yeast extract 0.5%, and NaCl 1%, and S. putrefaciens strain NCIMB 12577 was grown in a medium23 (pH 7.2) containing (g dm-3) NaCl, 20; K2HPO4, 1.0; MgSO4, 1.0; peptone, 5.0; yeast extract, 2.0 (Sigma, U.K.). Cells were harvested by centrifugation (8000g, 20 min, 4 °C). After washing once with 50 mmol dm-3 Tris-HCl, pH 8.0, the cells were fractionated using lysozyme/EDTA buffer as described in McEwan.24 The periplasm of the cells was then slowly brought to 50% saturation with ammonium sulfate (grade III, Sigma, U.K.). Following centrifugation for 20 min at 8000g and 4 °C the supernatant was removed and brought to 75% saturation with ammonium sulfate (grade III) and recentrifuged. The resultant 50-75% pellet was suspended in a minimum volume of 50 mmol dm-3 Tris/HCl, pH 8.0, and dialyzed against 3 × 100 volumes of the same buffer. The enzymes contained within the sample were then concentrated by ultrafiltration through an Amicon PM 10 membrane. TMAO reductase activity was measured as described by McEwan,24 with dithionite-reduced methyl viologen as electron donor. Protein determination was performed by the method of Bradford.25 Results indicated that the activities of the TMAO reductase from both E. coli and S. putrefaciens were too low for our needs (0.24 and 0.86 µM methyl viologen oxidized min-1 mg-1 of enzyme respectively). Since it had been previously suggested that the dimethyl sulfoxide (DMSO) reductase could also reduce TMAO,8,24,26 this enzyme was tested using both DMSO and TMAO as substrates. The DMSO reductase enzyme was purified from Rhodobacter capsulatus strain H123 following the method outlined by McEwan,24 with the only exception being that any ammonium sulfate was removed from the enzyme by dialyzing against Tris buffer (50 mmol dm-3, pH 8). This was to reduce the levels of ammonium present, which would be problematic for the FIGD-IC analysis. The purified enzyme appeared as a single band on a silver stained gel following standard protein gel electrophoresis (data not shown). The purified enzyme had a protein content of 40 mg cm-3 and a specific activity of 41.5 µmol methyl viologen oxidized min-1 mg-1 for either DMSO or TMAO. Analytical Procedure. Reduction of TMAO to TMA was achieved through addition of the purified DMSO reductase enzyme and the EDTA/FMN reducing solution. Sample or standard (55 cm3) was sparged with nitrogen (OFN grade; 100 cm3 min-1) to remove any oxygen present and added to a 60-mL vial containing 2 cm3 of the reducing solution. The vial was crimpsealed and the headspace sparged with nitrogen (OFN grade) via a pair of hypodermic needles for 2-3 min, to maintain the semi-anaerobic conditions required. An aliquot of DMSO reductase (2 µl) was then added and the vial illuminated with three 60 W daylight bulbs for 20 min, while being continuously sparged (Figure 1). Under these conditions, EDTA acts as an effective photoreductant reducing FMN to FMNH2 (eq 1).27 FMNH2 then (24) McEwan, A. G.; Wetzstein, H. G.; Ferguson, S. J.; Jackson, J. B. Biochim. Biophy. Acta 1985, 806, 410-417. (25) Bradford, M. M. Anal. Biochem. 1976, 15, 248-254. (26) McEwan, A. G.; Ferguson, S. J.; Jackson, J. B. Biochem. J. 1991, 274, 305307. (27) Massey, V.; Stankovich, M.; Hemmerich, P. Biochemistry 1978, 17, 1-8.
Analytical Chemistry, Vol. 71, No. 21, November 1, 1999
4887
Figure 1. Schematic representation of the enzyme reaction apparatus and chromatography system.
acts as an electron donor to DMSO reductase, catalyzing the reduction of TMAO to TMA (eq 2). H2O
FMNox + EDTA 98 FMNH2 + ED-triacetate + CH2O + CO2 (1) (CH3)3N ) O + FMNH2 f (CH3)3N + FMN + H2O (2) After 20 min, the sample was analyzed for TMA using flow injection gas diffusion-ion chromatography (FIGD-IC) similar to that described in Gibb et al.1,28 In summary, the sample is pumped continuously and treated with the mixed EDTA (1.0 mol dm-3)/ NaOH (0.11 mol dm-3) reagent to chelate alkaline earth cations and concurrently raise the pH of the mixture to >12. Under these conditions, >95% of the total dissolved MA cations (e.g., TMA.H+(aq)) are deprotonated to their volatile gaseous forms (e.g., TMA(g)) and are capable of diffusion from the sample stream, across a Goretex membrane, into an acidic acceptor stream (40 mmol dm-3 MSA + sec-BA) in which they are reprotonated. Recycling the acceptor (20 min) promotes selective accumulation of the analytes in the acceptor stream, which is then transferred to a Dionex DX100 ion chromatograph (IC) and 200 µl is injected. TMA.H+ and other low molecular weight amine cations were resolved isocratically (40 mmol dm-3 MSA, 1 cm3 min-1) by three IonPac CG-10 cation-exchange columns (Dionex, U.K.) using a Cation SelfRegenerating Suppressor (CSRS) and current controller (Dionex, U.K.) to suppress the background conductivity. The analogue signal (which is proportional to the analyte concentration) was digitized by a data collection unit (DCU) and collected and processed by PC based ATI-Unicam 4880 software. The configuration of the equipment is shown schematically in Figure 1. The observed peak area of TMA was used to calculate the concentration of TMAO in the sample via response factors calculated from the analysis of standards. (28) Gibb, S. W.; Mantoura, R. F. C.; Liss, P. S. Anal. Chim. Acta 1995, 316, 291-304.
4888
Analytical Chemistry, Vol. 71, No. 21, November 1, 1999
Interferences. To examine the potential of competitive inhibition of the enzyme by DMSO, which is thought to be ubiquitous in seawater and a known substrate for DMSO reductase,29 an equimolar solution of DMSO and TMAO (300 nmol dm-3) was prepared. Results from analysis of this solution showed no decrease from the expected concentration, indicating complete conversion of TMAO to TMA. QAs, structural analogues of TMAO, and known precursors of TMA in the marine environment,2 were also tested for interference, i.e., whether they would generate TMA through reaction with the enzyme and reducing solution. Equimolar standards of TMAO and the QAs GBT and choline (300 mmol dm-3; Sigma) were prepared, enzymatically treated, and analyzed as above. No additional TMA response was recorded, indicating that the enzymatic reduction was specific for the conversion of TMAO to TMA. In addition, enzymatic treatment of TMA standards resulted in no change in response to TMA, indicating that TMA was not lost during treatment. Optimization and Performance. To evaluate the efficiency of the reduction step and, thus, the amount of enzyme required for optimum performance, a series of analyses were performed using varying amounts of the DMSO reductase enzyme and different incubation times. Figure 2 shows that the addition of
