Cyanide degradation by an Escherichia coli strain

Cyanide degradation by an Escherichia coli strain

Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by 192.64.10.76 on 06/05/13 For personal use only.

Marianne M. Figueira, Virginia S.T. Ciminelli, Mabel C. de Andrade, and Valter R. Linardi

Abstract: Chemical formation of a glucose-cyanide complex was necessary for metabolic degradation of cyanide at concentrations up to 50.0 mg/L by a strain of Escherichia coli isolated from gold extraction circuit liquids. Ammonia accumulating during the culture log phase as the sole nitrogen by-product was further utilized for bacterial growth. Washed (intact) cells, harvested at different periods of bacterial growth on cyanide, consumed oxygen in presence of cyanide. These findings suggest that metabolism of cyanide involved a dioxygenase enzyme that converted cyanide directly to ammonia, without the formation of cyanate. Key words: cyanide, cyanide metabolism, cyanide oxygenase, Escherichia coli.

Resume : La formation clinique d'un complexe glucose-cyanure est un prCalable 2 la dkgradation mCtabolique du cyanure a des concentrations allant jusqu'i 50.0 mg/L par une souche d'Escherichia coli isolCe des liquides de procCdC d'un circuit d'extraction d'or. L'ammoniaque qui s'accumule durant la phase logarithmique de croissance comme source unique de produit dkrivC azotC sert par la suite 2 la croissance bacttrienne. Des cellules lavkes (intactes), recoltkes B diffkrentes Ctapes de croissance en prksence de cyanure, consomment de I'oxygkne en prksence de cyanure. Ces rksultats suggkrent que le mCtabolisme du cyanure implique une enzyme dioxygknase qui transforme le cyanure directement en ammoniaque sans formation de cyanate.

Mots elks : cyanure, mCtabolisme du cyanure, cyanure oxygCnase, Escherichia coli. [Traduit par la rkdaction]

Cyanide is commonly used by industries in the production of chemicals and synthesis of metacrylates, dyes, synthetic fibers, and agricultural products (Basheer et al. 1992; Raybuck 1992). Gold mining and electroplating industries are responsible for releasing large amounts of cyanide into the environment (Knowles and Bunch 1986). Owing to their high toxicity, cyanide-containing waste waters must be treated before release. Many chemical processes are proposed for the decomposition of cyanide, but some processes require special equipment and, in many cases, do not degrade all cyanide complexes. Consequently, biological treatment has been investigated as a way to achieve high degradation efficiency at low costs. A full-scale biological treatment facility is currently being used to treat cyanide waste waters at Homestake Mining Co. (Lead, S. Dak.): cyanide compounds and thiocyanate are oxidized to ammonia and carbonate by a microbial consortium containing Pseudomonas paucimobilis, specially acclimated to high cyanide concentrations, and immobilized in RBC (rotating biological contactor) reactors (Mudder and Whitlock 1984; Whitlock 1989; Smith and Mudder 1991). The process reduces 95-98% Received May 19, 1995. Revision received December 19, 1995. Accepted January 4, 1996.

M.M. Figueira and V.S.T. Ciminelli. Departamento de Engenharia Metaltirgica, Escola de Engenharia, Universidade Federal de Minas Gerais, 3016-030-Belo Horizonte, Brazil. M.C. Andrade and V.R. Linardi.l Departamento de Microbiologia, Universidade Federal de Minas Gerais, Cx.P. 486, 3 1270-901 -Belo Horizonte, Brazil. Author to whom all correspondence should be addressed. Can. J. Microbiol. 42: 519-523 (1996). Printed in Canada / ImprimC au Canada

of the initial cyanides and metals concentrations in a daily discharge of 4 million U.S. gal. of waste waters (1 U.S. gal. = 3.785 dm3) (Whitlock 1990). Metabolism of cyanides by strains of Pseudomonas, Acinetobacter, Bacillus, and Alcaligenes has been the subject of various studies during the last decade (Harris and Knowles 1 9 8 3 ~Finnegan ; et al. 1991; Meyers et al. 1991; Ingvorsen et al. 1991). The metabolism of cyanide by Pseudomonasfluorescens NCIMB 11764 was described by Harris and Knowles (1983a), who proposed that cyanide metabolism involved the activity of an oxygenase enzyme (Harris and Knowles 1983b). Kunz and Nagappan (1989) reported the presence of cyanase activity in P.fluorescens NCIMB 11764, but cyanase and cyanide activities are not coinduced when this bacterium is grown on KCN, suggesting the presence of a dioxygenase capable of transforming cyanide into C 0 2 and NH3 without formation of cyanate as an intermediate (Dorr and Knowles 1989). Studies by Kunz et al. (1994) found that growth of strain NCIMB 11764 on cyanide involved several cyanide-degrading components (CDAs) (i.e., cyanide oxygenase, cyanide nitrilase, and cyanide hydratase), with cyanide oxygenase being the most important in initiating substrate attack. Cyanide resistance and metabolism in Escherichia coli has not been completely defined (Pudek and Bragg 1974; Ashcroft and Haddock 1975; Haddock and Jones 1977). This study describes cyanide degradation by an E. coli strain isolated from gold extraction circuit liquids and suggests a potential metabolic pathway. Five samples taken from the gold extraction circuit of MineraqBo Morro Velho (Nova Lima, Brazil) were inoculated

Can. J. Microbiol. Vol. 42, 1996


Fig. 1. Cyanide-glucose complexation and ammonia formation in a chemically defined medium (pH 7.0) in cotton-plugged ( a ) and rubber-plugged (b) flasks. 0,total cyanide concentration; 0, free cyanide; A, ammonia. Initial [CN-] = 100 mg/L.

lncubation time (h)

0 0

20

40

60

80

100

120

140

lncubation time (h)

into sterile brain heart infusion (BHI) and thioglycolate broth and incubated for 48 h at 35°C. After growth, isolated bacterial colonies were obtained by spreading 0.2 mL broth onto bloodagar plates. Morphologically distinct colonies were isolated and identified by the Laboratbrio de Bacteriologia de Aerbbios, Federal University of Minas Gerais. All the isolates were assayed for cyanide degradation ability (Harris and Knowles 1983~).Each isolate was spread onto a buffered minimal medium consisting of 0.03 g MgS04.7H20, 0.3 g NaC1, 1.0 g glucose, and 0.15 g agar dissolved in 100.0 mL of a 67 mM KH2P04-K2HP04 buffer, (pH 7.0). Filter-sterilized KCN solution (2.504 g/L) was added (0.2 mL) daily to a filter paper placed on the top of each Petri dish. The dishes were incubated at 35°C for up to 15 days. Seventeen bacterial strains belonging the genera: Escherichia, Pseudomonas, Klebsiella, Staphylococcus,Alcaligenes, and Providencia were isolated from a gold extraction circuit. Only nine of these strains were able to grow with cyanide vapors as the sole source of nitrogen. Since growth could be due to the presence of some other nitrogen sources, carried over from the rich culture broth to the plates,

or by the accumulation of endogenous nitrogen compounds present in the cells, the procedure was repeated for all positive growth to eliminate this possibility. At this stage, only six strains retained the ability to grow in cyanide vapor after an incubation period of 7 days. One of the isolates capable of growing on cyanide vapor and identified as Escherichia coli (strain BCN6) was used for the following experiments. The pK, for KCN is 9.2, and thus, the concentration of HCN increases with decreasing pH. Since HCN is a volatile complex, it could be lost during incubation of the bacteria. The daily addition of fresh cyanide solutions to the medium was proposed by Harris and Knowles ( 1 9 8 3 ~and ) Kang and Kim (1993) to avoid this problem. However, we found that bacteria were unable to grow when cyanide solutions were added to the flasks, unless the cyanide addition was stopped for 48 h. This suggested that the E. coli strain selected for further studies tolerated cyanide (even after 72 h of daily addition of cyanide, the cells were viable), but it was not capable of degrading CN-. Moreover, cyanide ion could be complexed to another substance of the medium, thus forming a less toxic compound. Experiments using buffered solutions of cyanide and one of each medium component revealed that glucose was the sole component capable of reducing the cyanide concentration, determined as CN- and HCN. It was assumed that CN- could react with glucose, probably producing cyanohydrin. Furthermore, the cyanide might be hydrolysed to ammonia and Dglucoheptonic acid at high pH values (Hope and Knowles 1991). Because of these reasons, the next step was to verify the rate of loss of cyanide as HCN to the atmosphere and possible formation of a cyanide-glucose complex in the buffered minimal medium. Cyanide-glucose complex formation and cyanide stripping were observed in buffered minimal medium (described above) containing 1% glucose. An aliquot of 1.0 mL of a filter-sterile KCN solution (6.25 g/L) was added to sterilized medium (25 mL) in Erlemmeyer flasks to reach the desired CN- concentration of 100.0 mg/L. In one experiment the flasks were sealed with rubber plugs and in a second study they were sealed with cotton plugs. The flasks were incubated at 30°C on a reciprocal shaker (150 rpm) and concentrations of cyanide and ammonia were monitored during a 120-h period. The cyanide concentration was determined as free (CN- plus HCN) and total (CN-, HCN, and complexed forms) cyanide. Cyanide was determined according to the titration method described by APHA (1985), total cyanide being determined after the previous distillation of samples. Ammonia was assayed as described by Fawcett and Scott (1960). It was observed that in cotton-plugged flasks the loss of volatile forms of cyanide was almost complete after 48 h. Furthermore, only a small amount of ammonia (0.524 mg/L) was produced during 120 h of incubation (Fig. la). While in the rubber plug-sealed flasks, the total cyanide concentration remained the same and the free cyanide decreased with time. This result suggested the reaction of cyanide with glucose to form cyanohydrin as proposed by Hope and Knowles (1991). After 24 h of incubation, about 40% of the present cyanide was consumed by this reaction and only a small amount of ammonia was detected (0.937 mg/L after 120 h of incubation) (Fig. 1b). The rate of cyanide-glucose complex formation after 24 h of incubation sharply decreased. After 120 h of incubation, about 50% of cyanide was consumed to form that compound. The results obtained with the special

Notes


Fig. 2. Cyanide formation in a chemically defined medium free cyanide; A , ammonia. Initial (pH 7.0). o,total cyanide; 0, [CN-] = 100 mg/L.

Fig. 3. Growth of Escherichia coli BCN6 in different cyanide concentrations: preinoculum with ammonia (w) or cyanide (0) as nitrogen source.

[Cyanide] (mg/L) Incubation time (h)

procedure for cyanide-glucose complex formation revealed that the proportion of cyanide complexation was similar to the above: about 40% of the added CN- reacted with glucose (Fig. 2). The stripping of the remaining free cyanide was rapid, and at 24 h after the substitution of the rubber for a cotton plug, almost all of the free form of cyanide had disappeared. A small amount of ammonia was detected after 120 h of incubation (0.873 mg/L). The same proportion of complexed cyanide was found when mantaining the glucose concentration and modifing the cyanide concentration. For bacterial inoculated experiments, a defined volume KCN solution was added to a flask containing 25 mL of medium to obtain a 2.5-fold molar excess over the desired final CNconcentration (e.g., 2.5 mL for a desired final concentration of 100.0 mg CN-/L). The flasks were sealed with rubber plugs and incubated for 24 h on a reciprocal shaker. The plugs were then replaced by cotton plugs and the flasks were agitated for 2 h before inoculation with bacterial cells. A 0.5-mL volume of cell suspension (OD540nm = 1.O or 0.044 mg dry weightImL) of the E. coli BCN6 grown in buffered minimal medium containing NH4C1(5.0 g/L) was used as a preinoculum in the cyanidecontaining medium. In one experiment the bacteria were inoculated into medium containing different CN- concentrations, and in a second set of experiments the bacteria were transferred to medium with increasing concentrations of cyanide: 25.0 mg CN-/L for 3 days and then to 50.0 mg CN-/L for 3 days up to 200.0 mg CN-/L. For both studies the flasks were incubated at 30°C on a reciprocal shaker and bacterial growth was monitored by OD determinations at 540 nm. Growth of E. coli BCN6 using different concentrations of cyanide (Fig. 3) revealed that, using a preinoculum that had been grown on ammonia, the maximum biomass production was obtained at a cyanide concentration of 25.0 mg/L. The growth was progressively less at higher concentrations. A different effect was observed when the bacterium was preincubated on cyanide as the sole source of nitrogen: the best growth was at cyanide concentration of 50.0 mg/L.

The growth curve of E. coli strain was assessed in flasks containing buffered minimal medium supplemented with 50.0 mg cyanidell. The flask was inoculated with 0.5 mL of a 48-h bacteria culture grown in 50.0 mg CN-/L. The flasks were incubated at 30°C on a reciprocal shaker and growth was monitored by OD540nm. Cyanide, ammonia, cyanate, and formate concentrations in the supernatant were determined at defined intervals. Standardized procedures for the cyanase activity (Anderson 1980) and cyanide-oxygenase activity (Rollinson et al. 1987) were used for the determination of such enzymatic activities in culture samples. The cells were centrifuged at 2287 x g for 20 min at 5°C and washed twice in a phosphate buffer (67 mM, pH 7.0); the pellet was resuspended in a phosphate buffer to a final OD = 1.O. The cyanase activity was determined by adding 1.0 mL of cell suspension and 1.0 mL of a 6 mM NaHC03 solution to a centrifuge tube. A 0.1-mL aliquot of a NaOCN solution (1624.47 mg/L) was added to the tube, which was incubated in a water bath (37°C) for 10 min. The reaction was terminated by centrifuging the cells. A 1.O-mL aliquot of the supernatant was used for ammonia determination. One unit of activity was defined as the activity amount capable of catalyzing the hydrolysis of 1 mmol cyanate/min under the experimental conditions. Cyanideoxygenase activity was determinated with a 3.0-mL aliquot of washed cells, incubated in a polarograph cube (YNI model 53). A solution (1.0 mL) of KCN in phosphate buffer (CN- final concentration, 100.0 mg/L) was added to the suspension and oxygen uptake was monitored after 10 min of incubation (30°C under agitation) by a specific O2 electrode. Controls were carried out with cells in (i) buffered solution or (ii) buffered cyanide solution. The specific enzymatic activity was determined as the amount of O2 (mmol) consumed in 1 min of reaction by each milligram of dried cells under the experimental conditions. Cyanate was determined by the spectrophotometric method (Guilloton and Karst 1985). Escherichia coli BCN6 strain was incubated at the optimal cyanide concentration indicated by the previous experiments (50.0 mg/L). A long lag phase of about 30 h resulted during which more than 90% of total cyanide was consumed

Can. J. Microbial. Vol. 42, 1996


Fig. 4. Growth curve of Escherichia coli BCN6 in 50.0 mg/L of cyanide as nitrogen source. A, bacterial growth; m, total cyanide; 0, ammonia; and 0,specific enzymatic activity.

Incubation time (h)

or otherwise lost from the medium, although no microbial growth was detected (Fig. 4). Ammonia was the only nitrogen compound detected, which accumulated in the medium until the cells entered the log phase that lasted for about 41 h. During this period most of the ammonia was consumed. The stationary phase of bacterial growth was reached after 72 h of incubation. The biomass concentration reached approximately 0.15 mg/mL (dry weight). Cyanase activity and oxygen consumption activity were determined for cells collected at various intervals during growth. No cyanase activity (responsible for the transformation of cyanate into ammonia and bicarbonate) was detected. Another possible mechanism of cyanide degradation is the production of formamide, which would be subsequently converted to ammonia and formate (determined according to Wood and Gest 1957). However, no formate was detected in the medium, although formate could have been further metabolized to release carbon dioxide. Oxygen consumption activity was detected at different growth curve intervals (data not shown). The highest activity (34.124 mmol O2 consumed/(min.mg dried biomass)) was observed for lag phase cells, from the start of the experiment. The enzymatic activity decreased in the early log phase, almost disappearing by the end of the experiment. Escherichia coli BCN6 grew in a cyanide-containing medium only after almost complete transformation of cyanide into ammonia. This fact suggests that the bacterium first converted the compound into a less toxic form. Once in the cell, although presumably complexed to glucose, cyanide could be released into its free form, which could inhibit the metabolism of the cell. Its complete conversion into ammonia could function as a mechanism of detoxification of the medium. The culture grew only after the disappearance of the cyanide. Escherichia coli strains are well recognized as cyanase producers (Anderson 1980). It is reported that about 60% of the naturally occurring strains of E. coli are able to produce cyanase. However, no cyanase activity was detected in whole cells of the BCN6 strain. This finding led us to suggest a mechanism of cyanide conversion by which no cyanate is formed. Two possible mechanisms may be involved in cyanide degradation: (i) its conversion by a cyanide oxygenase to ammonia or (ii) the production of cyanide nitrilase, since the

complex formed by cyanide and glucose is a nitrile. This high initial activity determined as cyanide oxygenase can be justified by the previous growth of bacteria on cyanide, which induced the cell enzymatic system. The cyanide-oxygenase activity decreased in the early log phase, almost disappearing by the end of the experiment. Such behavior is expected for inducible enzymatic processes: in the absence of substrate, en~~meproductibn drops, leading to a fall in its specific activity. The cyanide-oxygenase activity of P. fluorescens was reported by Harris and Knowles (1983b) who observed that the strain degraded cyanide, producing and consuming stoichiometric amounts of ammonia and oxygen, respectively. The results obtained in the present work suggest that the enzyme involved in cyanide degradation by E. coli BCN6 could also be a dioxygenase, capable of oxidizing cyanide directly to ammonia and carbon dioxide, since oxygen consumption was observed, without previous formation of cyanate. Similar results have been obtained by Dorr and Knowles (1989) for l? fluorescens NCIMB 11674. These authors reported that this strain produced cyanase only when it grew in the presence of cyanate. Cyanide oxygenase was the sole enzyme detected when the microorganism grew on cyanide. These results led them to the conclusion that the enzymes are never coinduced. Several authors have discussed the cyanide metabolism by bacteria, but only two studies have pointed at a possible "complexation" between cyanide and glucose (Raef et al. 1977; Hope and Knowles 1991), the latter paper describing cyanide production under anaerobic conditions and its subsequent hansformation into ammonia, which was finally metabolized by the culture. The results concerning cyanide formation in a defined medium reported in the present work are not consistent with those reported in the literature, since the ammonia production is due tomicrobial activity. This may be due to differences in the chemical composition of the medium used, as well as to the distinct physical conditions. It may be necessary to determine a possible reaction between cyanide and carbohydrates present in the medium, in addition to its possible conversion to ammonia by an abiotic process, to guarantee that a microbial activity is taking place. Also, it is necessary to acertain that the method used for cyanide determination is satisfactory to confirm its presence, even when it is complexed to aldoses. This study illustrates the possible mechanism involved in cyanide degradation by E. coli. Although this bacterium is able to degrade cyanate (a hypothesis that has been proposed as a mechanism by which this microorganism could be capable of degrading cyanide), it was observed that the strain BCN6 does not convert cyanide into an intermediate compound to form ammonia. ~ikewise,it was noticed that this strain requires a complexed cyanide form to grow under such conditions. However, it does not require this complex to be recognized by its enzyme. This could be explained by the possibility of reducing the toxicity of free cyanide, instead of the requirement of a special configuration by which the enzyme distinguishes the substrate. The ability of E. coli BCN6 strain to degrade most of the cyanide in the medium before its growth commences suggests that it might be used in special bioreactors with immobilized cells where the biomass production would not be excessive. The requirement for glucose or other reducing sugar could be

Notes

satisfied by addition of molasses or different sugar containing wastes, such as those from breweries, canneries, cereal grain processing plants, pulp and paper mills, and others.

Acknowledgements


The authors gratefully acknowledge Dr. B. Volesky for kindly helping to review the manuscript, R. Liberato and W. Moura (Mineraqiio Morro Velho, Brazil) for providing liquid samples, and A. Villela for her extensive help with chemical analysis. We are also thankful for financial support from Conselho Nacional de Pesquisa, Funda~aode Amparo e Pesquisa do Estado de Minas Gerais, and Pr6-Reitoria de Pesquisa da Universidade Federal de Minas Gerais.

References

Anderson, P.M. 1980. Purification and properties of the inducible enzyme cyanase. Biochemistry, 19: 2882-2888. APHA (American Public Health Association). 1985. Standard methods for the examination of water and wastewater. APHA, Washington. pp. 327-352. Ashcroft, J.R., and Haddock, B.A. 1975. Synthesis of alternative membrane-bound redox carriers during aerobic growth of Escherichia coli in the presence of potassium cyanide. Biochem. J. 148: 349-352. Basheer, S., Kut, O.M., Prenosil, J.E., and Bourne, J.R. 1992. Kinetics of enzymatic degradation of cyanide. Biotechnol. Bioeng. 39: 629-634. Dorr, P.K., and Knowles, C.J. 1989. Cyanide oxygenase and cyanase activities of Pseudomonas fluorescens NCIMB 11764. FEMS Microbiol. Lett. 60: 289-294. Fawcett, J.K., and Scott, J.E. 1960. A rapid and precise method for the determination of urea. J. Clin. Pathol. 13: 156-159. Finnegan, I., Toerien, S., Abbot, L., Smit, F., and Raubenheimer, H.G. 1991. Identification and characterisation of a Acinetobacter sp. capable of assimilation of a range of cyano-metal complexes, free cyanide ions and simple organic nitriles. Appl. Microbiol. Biotechnol. 36: 142-144. Guilloton, M, and Karst, F. 1985. A spectrophotometric determination of cyanide using reaction with 2-aminobenzoic acid. Anal. Biochem. 149: 291-295. Haddock, B.A., and Jones, C.W. 1977. Bacterial respiration. Bacterial. Rev. 41: 47-99. Harris, R., and Knowles, C.J. 1983a. Isolation and growth of a Pseudomonas species that utilizes cyanide as a source of nitrogen. J. Gen. Microbiol. 129: 1005- 101 1.

Harris, R., and Knowles, C.J. 1983b. The conversion of cyanide to ammonia by extracts of a strain of Pseudomonas fluorescens that utilizes cyanide as a source of nitrogen for growth. FEMS Microbiol. Lett. 20: 337-341. Hope, K.M., and Knowles, C. J. 1991. The anaerobic utilisation of cyanides in the presence of sugars by microbial cultures can involve an abiotic process. FEMS Microbiol. Lett. 80: 217-220. Ingvorsen, K., Hojer-Pedersen, B., and Godtfredsen, S.E. 1991. Novel cyanide-hydrolyzing enzyme from Alcaligenes xylosoxidans subsp. denitrGcans. Appl. Environ. Microbiol. 57: 1783-1789. Kang, S.M., andKim, D. J. 1993. Degradation of cyanide by abacterial mixture composed of new types of cyanide degrading bacteria. Biotechnol. Lett. 15: 201-206. Knowles, C.J., and Bunch, A.W. 1986. Microbial cyanide metabolism. Adv. Microbiol. Physiol. 27: 73-1 11. Kunz, D.A., and Nagappan, 0. 1989. Cyanase-mediated utilization of cyanate in Pseudomonasjuorescens NCIMB 11764. Appl. Environ. Microbiol. 55: 256-258. Kunz, D.A., Wang, C., and Chen, J.L. 1994. Alternative routes of enzyme cyanide metabolism in Pseudomonas fluorescens NCIMB 11764. Microbiology, 140 : 1705-1712. Meyers, P.R., Gokool, P., Rawlings, D.E., and Woods, D.R. 1991. An efficient cyanide-degrading Bacillus pumilus strain. J. Gen. Microbiol. 137: 1397-1400. Mudder, T.I., and Whitlock, J.L. 1984. Method for the biological removal of free and complex cyanides and thiocyanates from water. U.S. Pat. 4 440 644. Pudek, M.R., and Bragg, P.D. 1974. Inhibition by cyanide of the respiratory chain oxydases of Escherichia coli. Arch. Biochem. Biophys. 164: 682-693. Raef, S.F., Characklis, W.G., Kessick, M.A., and Ward, C.H. 1977. Fate of cyanide reation and related compounds in aerobic microbial systems. 1. Chemical reaction with substrate and physical removal. Water Res. 11: 477-483. Raybuck, S.A. 1992. Microbes and microbial enzymes for cyanide degradation. Biodegradation, 3: 3-18. Rollinson, G., Jones, R., Meadows, M.P., Harris, R.E., and Knowles, C. J. 1987. The growth of a cyanide-utilising strain of Pseudomonasfluorescens in a liquid culture on nickel cyanide as a source of nitrogen. FEMS Microbiol. Lett. 40: 199-205. Smith, T., and Mudder, T. 1991. The chemistry and treatment of cyanidation wastes. Mining Journal Books, London. Whitlock, J. 1989. The advantages of biodegradation of cyanides. J. Metals, 41: 46-47. Whitlock, J. 1990. Biological detoxification of precious metals processing wastewaters. Geomicrobiol. J. 8: 241-249. Wood, H.G., and Gest, H. 1957. Determination of formate. Methods Enzymol. 3: 285-289.