XENOBIOTICA,
1975, VOL. 5,
NO.
9, 563-571
Reduction of Azo Food Dyes in Cultures of
Proteus vulgaris P. D U B I N and K. L. W R I G H T
Xenobiotica Downloaded from informahealthcare.com by University of Queensland on 04/23/13 For personal use only.
Dynapol, Palo Alto, California, U S A .
(Received 2 March 1975)
1. Rates of reduction of a number of azo food dyes were measured in anaerobic cultures of Proteus vulgaris. The rates of colour loss were found to be zero order under conditions in which the concentration of viable cells remained constant. 2. A significant increase in the rate of reduction followed the onset of cell mortality. 3. The zero-order rates correlate with the redox potentials of the dyes. A mechanism consistent with these observations involves an extracellular nonenzymic reducing agent which acts as an electron shuttle between dye and cellular reducing enzymes.
Introduction Bacterial reduction of azo dyes takes place both in vitro (Roxon, Ryan & Wright, 1 9 6 7 a ; Scheline, Nygaard & Longberg, 1970) and in the gut of animals (Radomski & Mellinger, 1962 ; Schroder & Johansson, 1973). These studies have dealt primarily with identification of metabolic products and elucidation of the biochemical reduction mechanism (Fore, Walker & Golberg, 1967 ; Roxon, Ryan & Wright, 1967 b ; Walker, Gingell & Murrells, 1971 ; Gingell & Walker, 1971 ; Gingell & Bridges, 1973). While some attempts have been made to relate the ease of reduction of azo dyes to molecular parameters (Walker & Ryan, 1971), such structure-activity relationships have been difficult to establish, in part because reduction rates in bacterial cultures are not easy to measure in a reproducible manner. In contrast to the status of investigations on bacterial reduction, rates of reduction of azo dyes in irradiated solutions (Giles, Hojiwala & Shah, 1974 ; Bridgeman & Peters, 1974) and in the presence of various reducing agents (Conant & Pratt, 1926; Brownley & Lachman, 1963) have been measured extensively. T h e rates of photochemical reduction of these dyes in the presence and absence of reducing agents depend on the reduction potential of the dye (Kozeny & Velich, 1960 ; Rao & Hayon, 1973). I n the present study, reduction rates of a number of water-soluble dyes were measured in whole cell cultures of P. vulgaris. T h e rate constants of the various dyes correlate with their redox potentials. A reduction mechanism is suggested that explains both this correlation and the observed zero-order reduction kinetics.
P. Dubin and K . L. Wright
5 64
Experimental Materials Orange I1 was synthesized by diazotization of sulphanilic acid and coupling to P-naphthol. Ponceau 3R ( C I 16155) was obtained from Pfalz & Bauer, Inc. (Flushing, New York). All other dyes (see Fig. 1) were certified colour
Xenobiotica Downloaded from informahealthcare.com by University of Queensland on 04/23/13 For personal use only.
additives (Allied Chemical Co., Morristown, New Jersey). Dye samples were not further purified before use. T h e purity of the dyes was determined by azo reduction with excess TiCl, under N, followed by back-titration of untreated Ti3+with Fe3-'-(Siggia, 1963).
0
S03Na
S03Na
N \N
N,\N
Na03S Amaranth
Orange
II
Ponceau 3 R
0 Q S03Na
S03Na
kN
Ponceax SX
Sunset Yellow
Fig. 1.
Tortrazine
Structures of azo dyes used.
Cell culture Proteus vulgaris ATCC (8427) was maintained on a slope of Difco tryptocase soy agar. A 50 ml overnight culture in Difco tryptocase soy broth was grown from the slope. This standing culture (10 ml) was then used to inoculate 500 ml of tryptocase soy broth. T h e final culture was grown without agitation for 12 h at 37" to yield a viable cell density of 2-4 x 108 cells/ml. T h e concentration of viable cells in mixtures of such cultures with dye solutions remained constant under anaerobic conditions for 16 h or more. When cultures with appreciably higher cell densities were used, the viable cell count declined rapidly once anaerobic conditions were attained. Cell counts were determined by serial dilution and plating on tryptocase soy agar.
Xenobiotica Downloaded from informahealthcare.com by University of Queensland on 04/23/13 For personal use only.
Reduction of Axo Dyes by P. vulgaris
565
Dye incubation Dye solutions (ca. 1.5 x M ) were prepared in deionized water and sterilized by autoclave (121", 20 min). Aliquots (8 ml) of the P. vulgaris culture were dispensed into sterile screw-cap culture tubes. Usually 2-0 ml of sterile dye solution was then added to give an azo concentration ca. 3 x lo5 M. Tubes containing bacterial culture alone were retained for later use as spectrophotometric blanks. Samples and blanks were incubated in the dark at 37" in BBL Gas Pak@ anaerobic jars (BioQuest, Cockeysville, Maryland). Anaerobic conditions were attained in about two hours as determined by the colour change of the methylene blue indicator. Samples to be analysed at a given incubation time were contained in a single jar in order to maintain anaerobic conditions for the other samples. Measurement of azo reduction rates Cells were removed from the incubation mixtures by centrifugation at 12 000 g and 4" for 10 min. T h e visible absorption spectrum of the transparent supernatant was measured against the appropriate reference with a Varian Techtron 635 Spectrophotometer. T h e reference solution consisted of the supernatant from culture grown in the absence of dye, diluted with deionized water to correct for addition of the dye solution to the sample. This blank nullified any contribution of the spent medium to the visible spectrum of the supernatant. Chromophore concentrations were determined from the blankof the given dye, relative to the absorbance of corrected absorbance at the A,, an unincubated mixture of dye and medium. T h e use of this reference compensated for indirect effects of the medium on the dye spectrum. I n separate experiments, the removal of dye by binding to precipitated cells was found to be negligibly small. Rates of azo reduction were determined graphically from the dependence of chromophore concentration on incubation time. T h e absolute rate values were sensitive to assay conditions, e.g., the reduction rate for Sunset Yellow (FD&C Yellow No. 6) determined in nine separate assays varied from 5.0 to 7.5 x lo-' mol 1-1 h-l. T h e data from assays with individual dyes were standardized by measuring rates relative to Sunset Yellow, which was included in each assay.
Results and discussion Cell viability T h e viable cell densities of the incubation mixtures were measured as a function of incubation time. Under the conditions described, viable cell density was constant during the first 18-24 h. During the subsequent 12 h an order-ofmagnitude decrease in viable cell count was typically observed. An abrupt increase in dye reduction rate was also found to occur within the period of cell mortality, as shown in Fig. 2. Reduction rates reported in the following always refer to data obtained during the first 24 h of anaerobic incubation. Azo reduction rates Azo reduction rates were determined from plots of chromophore concentration against anaerobic incubation time. Representative data are shown in Fig. 3.
P. Dubin and K. L. Wright
Xenobiotica Downloaded from informahealthcare.com by University of Queensland on 04/23/13 For personal use only.
5 66
Anaerobic incubation time ( h )
Fig. 2.
Dependence on anaerobic incubation time of viable cell density ( 0) and of chromophore concentration for Sunset Yellow ( A) and tartrazine (0). Dyes were incubated with anaerobic cultures of P . vulgaris for the indicated times. The cells were removed by centrifugation and the dye concentration in the supernatant was determined from the visible spectrum measured against the appropriate blank (Sunset Yellow, ; tartrazine, 0).Viable cell densities ( 0 )were obtained by serial dilution and plating on tryptocase soy agar of a random sampling of dye incubation mixtures.
L
c c>
I
1
I
I
I
Anaerobic incubction time (h)
Fig. 3.
Representative plots of dye concentration against anaerobic incubation time used to determine rates of reduction. ( 0) Sunset Yellow. (0) Ponceau 3R. ( A )Amaranth.
Reduction of A x o Dyes by P. vulgaris
5 67
Table 1. Reduction of water-soluble azo dyes in P. vulgaris cultures, relative to Sunset Yellow
Xenobiotica Downloaded from informahealthcare.com by University of Queensland on 04/23/13 For personal use only.
Dye
Orange I1 Amaranth Ponceau SX Sunset Yellow Ponceau 3R Tartrazine
CI Number
Reduction rate relative to Sunset Yellow (R’)
15518 16185 14700 15985 16155 19140
6.3 2.0 1.5 1*o 0.55 0.41
T h e linear dependence of chromophore concentration on time indicates reduction rates zero order with respect to dye concentration. Furthermore, identical reduction rates were observed for Sunset Yellow at initial concentrations differing by a factor of two. T h e values of R‘, the zero order reduction rate relative to Sunset Yellow, are given in Table 1. T h e initial dye concentrations were within the concentration range bracketed by the two assays for Sunset Yellow shown in Fig. 3. Relationship of reduction rate ( R ‘ ) to dye reduction potential T h e rate of photochemical reduction of azo dyes depends on the dye redox potential (Rao & Hayon, 1973 ; Kozeny & Velich, 1960). A similar correlation might be expected for bacterial azo reduction if the rate-controlling step involved a redox equilibrium between dye and an extracellular reducing agent. It appears likely that soluble flavins can play the role of the latter (Roxon et al., 1967 b ; Gingell & Walker, 1971). Other observations in the literature are in accord with a mechanism involving redox equilibrium. Reduction rates correlate with the redox potential of the local environment, both in vivo (Schroder & Johansson, 1973) and in vitro (Gingell & Walker, 1971). T h e effects of substituents on the reduction rate have been rationalized on the basis of effects on the electron density a t the azo bond (Walker & Ryan, 1971). T h e redox potentials of azo dyes may be estimated from the polarographic half-wave potentials. T h e relationship between the redox potential, E,, and the half-wave potential, Eliz, is given by the following expression :
(see for example Willard, Merritt & Dean, 1968), where n is the number of electrons transferred in the redox process and D is the diffusion coefficient of either the oxidized or reduced member of the redox pair. It is reasonable to assume that the diffusion coefficients of the azo and hydrazo species are nearly equal and hence to neglect the second term in the above expression for Ellz. T h e dependence of the relative reduction rate in whole-cell P. vulgaris cultures ( R ’ )on the polarographic half-wave potential, El/,, is shown in Fig. 4 for the dyes of this study. (Redox potentials in biological systems are generally reported
P . Dubin and K . L. Wright
Xenobiotica Downloaded from informahealthcare.com by University of Queensland on 04/23/13 For personal use only.
568
relative to the standard hydrogen electrode (S.H.E.) ; the usual polarographic reference is the saturated calomel electrode (S.C.E.). We have chosen for clarity to adhere to the latter convention, noting that E(S.C.E.) =E(S.H.E.) 0.246 V. I t is also appropriate to note the marked p H dependence typically reported for Ell2. T h e values used here were either measured at or interpolated to pH 7.) T h e logarithm of the relative rate constant varies in a linear manner with the redox potential of the dye. T h is correlation between reduction rate and an electrochemical property of the dye suggests that the rate-determining step in the bacterial reduction of the chromophore does not involve a structurespecific phenomenon such as selective membrane permeation or enzyme binding.
t
80/
:,
60
L?
/
"/
I
- 700
'
I
I
I
- 600
- 500
- 400
EL (pH 7,vs S C.E 1 2
(mV)
Fig. 4. Dependence of relative veduction rate (R') on dye redox potential. Reduction rates relative to Sunset Yellow were determined graphically as shown in Fig. 3. Dye redox potentials were obtained from literature polarographic data (at pH 7, vs. S.C.E.) from the following references : ( 0 )Orange I 1 (Florence, 1964 a) ; (A) Amaranth (Iijima, 1967 ; Mizunoya & Kita, 1965 ; McKeown & Thomson, 1954) ; (0) Sunset Yellow (Nazario & Zenebon, 1972) ; ( 0 )Ponceau SX (Mizunoya & Kita, 1965) ; ( A ) Ponceau 3R (hlizunoya & Kita, 1965) ; (m) Tartrazine (Florence, 1964 a ; Mizunoya & Kita, 1965 ; Nazario & Zenebon, 1972).
Zero-order dependence of reduction rate on dye concentration T h e rate of colour loss was zero order with respect to dye, over a three- to four-fold range in azo dye concentration. T h e absence of any effect of dye concentration on reduction rate is incompatible with simple diffusion of dye through bacterial membrane as a rate controlling step. Two types of interactions between dye and bacterial culture consistent with the observed kinetics may be proposed : (1) saturation by the dye of either active membrane transport sites or the binding sites of a general azo reductase,
Reduction of Azo Dyes by P. vulgaris
5 69
Xenobiotica Downloaded from informahealthcare.com by University of Queensland on 04/23/13 For personal use only.
(2) extracellular (presumably nonenzymic) dye reduction, the rate of which is controlled by a process independent of dye concentration. The transport or binding processes involved in the first hypothesis imply a structural specificity not in accord with the simple dependence on redox potential shown above ; hence, mechanism (2) is to be preferred on the basis of the present results. Postulated mechanism Experimental evidence in support of mechanism (2) has also been observed by others. Roxon et al. (1967 b) reported that the presence of a flavin in the medium was essential for the reduction of tartrazine in whole-cell suspensions of P. vulgaris. Gingell & Walker (1971) also noted that the addition of soluble flavin accelerated the reduction of Red 2G in a cell-free preparation of Streptococcus faecalis. In addition, the rate of reduction was found to be zero order with respect to dye and to depend on the reduction potential of the system. On the basis of these results, Gingell & Walker proposed that the soluble flavin acts as an electron shuttle between the dye and a reducing enzyme, the rate being controlled by generation of reduced flavin. We will show that this type of mechanism is consistent with the kinetics and the dependence on dye redox potential observed here with whole cells of P . vulgaris. The enzymic generation and subsequent oxidation by dye of a low molecular weight reducing agent may be written as k,
B+EH, -+ BH,+E slow kz
BH,+ RN=NROH
k-a
B + R(NH),ROH
(2)
where BH,/B and EH,/E are the redox couples for the low molecular weight electron carrier and enzyme, respectively. The measured rate reflects only the loss in azo concentration and is insensitive to the subsequent fate of the hydrazo compound. It may be noted, however, that in both potentiometric (Conant & Pratt, 1926) and polarographic (Florence, 1964 b) reduction the phenolic hydrazo compound undergoes subsequent disproportionation and reduction : R(NH),R'OH
+RNH, ka
+ R' /O LNH
slow
BH, + R' No
NNH
e" R /OH '-4
+B
\NH,
Steps (1) and (2) may be written as a redox cycle :
X.B.
2P
570
P . Dubin and K. L. Wright
with reducing power provided by bacterial metabolism. No net reduction of dye occurs until the concentration of BH, is such that the reduction potential of the BH,/B couple approaches that of the dye, i.e.,
E = EO - 0.030 log
0= En, (dye).
Xenobiotica Downloaded from informahealthcare.com by University of Queensland on 04/23/13 For personal use only.
(B) Subsequently, the reduction rate, i.e., colour loss, will depend on step ( l ) , the rate of formation of BH,. T h e rate of BH, production may be seen from equation (1) to depend on the concentration of B. If the sum (BH,) +(B) is constant, the concentration of B is controlled by the ratio (BH,)/(B), which, according to equation (5) is determined by the dye redox potential. Thus, the electrochemical properties of the dye may indirectly control the reduction rate by determining the concentration of B in the system, although the dye molecule is not itself involved in the rate-determining step. T h e rate of colour loss would then be zero order with respect to dye (until extremely low dye concentrations are reached) and dependent on dye redox potential, in accord with the results cited above. We have considered only steps (1) and (2) so far. I t is generally thought that step ( 3 ) is slow and rate-controlling for the complete polarographic or chemical reduction of hydroxy azo compounds to primary amines (Conant & Pratt, 1926 ; Florence, 1964 b ; Florence, 1974). However, the rate of this unimolecular disproportionation would not affect the rate of colour loss if formation of BH, were the slowest process, as would be expected if step (1) involved diffusion through a cell membrane. I n the presence of a species capable of reducing the azo bond, the reduction of the imino-quinone (step (4)) is expected to be rapid and complete (Conant & Pratt, 1926). Step (4) can then only influence the rate of colour loss via consumption of BH,. Thus a decrease in the rate of colour loss might be expected when azo concentration is sufficiently low for the rate of BH, consumption in step (4) to effectively compete with that in step (2). Implicit in the foregoing analysis is the assumption of constant (BH,) + (B). Introduction of additional reducing agent would not alter the reaction order or the dependence of rate on dye redox potential, but would increase the rate of colour loss. While no such effect was observed under experimental conditions such that viable cell count was constant, acceleration of the rate was noted subsequent to cell death. Effect of cell mortality on dye reduction rate T h e increase in dye reduction rate observed upon cell death, as shown in Fig. 2, could arise from release into the medium by lysed cells of agents that provide an alternative reduction mechanism. However, the reduction process after cell death apparently remains zero order in dye ; this would be unexpected if the increase in reduction rate corresponded to a large increase in the concentration of metabolites capable of directly reducing the dye. On the other hand, the observed enhancement in reduction could correspond to an increase in the efficiency of the postulated mechanism arising from the disruption of membrane barriers to permeability of reducing agent. Such change in accessibility of cellular enzymes to extracellular reducing agents would effectively increase the
Reduction of A x o Dyes by P. vulgaris
571
Xenobiotica Downloaded from informahealthcare.com by University of Queensland on 04/23/13 For personal use only.
forward rate constant of step (1 ) and accelerate reduction without altering reaction order, as is observed. Roxon, Ryan & Wright (1966, 1967 b) noted that starved cultures of P. vulgaris are much more active than young cultures in reducing tartrazine. They ascribed this effect to an increase in cell permeability to tartrazine (Roxon et al., 1967 b). However, if transport of dye across the cell membrane were intrinsic to the reduction process, it would be expected that the rate of reduction would be influenced by both the concentration and the shape and ionic charge of the dye, in sharp contrast to the present results.
Acknowledgments T h e authors express their appreciation of helpful comments from Dr. J. P. Brown. T h e assistance of Dave Cartopassi is gratefully acknowledged.
References BRIDGEMAN, I. & PETERS,A. T. (1974). Text. Res., 44, 639. BROWNLEY,C. A., Jr. & LACHMAN, L. (1963). J . Pharm. Sci., 52, 86. CONANT, J. B. & PRATT,M. F. (1926). J. A m . chem. Soc., 48, 2468. FLORENCE, T. M. (1964 a). Aust. J. Chem., 18, 609. FLORENCE, T. M. (1964 b). Aust. J . Chem., 18, 619. FLORENCE, T. M. (1974). J. electroanal. Chem., 52, 115. FORE,H., WALKER, R. & GOLBERG, L. (1967). Fd. C o m e t . Toxic., 5, 459. GILES,C. H., HOJIWALA, B. J., & SHAH,C. D. (1974). J . SOC.Dyers Colour., 90,45. R. & BRIDGES, J. E. (1973). Xenobiotica, 3, 599. GINGELL, GINGELL, R. &WALKER, R. (1971). Xenobiotica, 1, 231. IIJIMA,T. (1967). Rev. Polarogr. (Japan), 14, 317. KOZENY, M. & VELICH,V. (1960). Coll. Czech. chern. Commun., 25, 1031. C. G. & THOMSON, J. L. (1954). Can.J . Chem., 32, 1025. MCKEOWN, Y. & KITA,T. (1965). Japan Analyst, 14, 437. MIZUNOYA, G. & ZENEBON, 0. (1972). Reota. Inst. Adolfo Lutz, 32, 101. NAZARIO, J. L. & MELLINGER, T. J. (1962). J . Pharmac. exp. They., 136, 259. RADOMSKI, RAo, P. S. & HAYON, E. (1973). J . phys. Chem., 77, 2753. ROXON, J. J., RYAN,A. J. & WRIGHT,S. E. (1966). Fd. Cosmet. Toxic., 4, 419. J. J., RYAN,A. J. & WRIGHT,S. E. (1967 a). Fd. Cosmet. Toxic., 5, 367. ROXON, J. J., RYAN,A. J. & WRGIHT,S. E. (1967 b). Fd. Cosmet. Toxic., 5, 645. ROXON, SCHELINE, R. R., NYGAARD, R. T. & LONGBERG, B. (1970). Fd. C o m e t . Toxic., 8, 55. SCHRODER, H. & JOHANSSON, A. K. (1973). Xenobiotica, 3, 233. S. (1973). Quantitative organic analysis, p. 526, New York : John Wiley and Sons. SIGGIA, WALKER, R., GINGELL, R. & MURRELLS, D. F. (1971). Xenobiotica, 3, 221. R. & RYAN,A. J. (1971). Xenobiotica, 1, 483. WALKER, H., MERRITT,L. L. & DEAN,J. A. (1968). Instrumental methods of analysis, WILLARD, p. 684, Princetown : Van Nostrand Co.
2P2
1975, VOL. 5,
NO.
9, 563-571
Reduction of Azo Food Dyes in Cultures of
Proteus vulgaris P. D U B I N and K. L. W R I G H T
Xenobiotica Downloaded from informahealthcare.com by University of Queensland on 04/23/13 For personal use only.
Dynapol, Palo Alto, California, U S A .
(Received 2 March 1975)
1. Rates of reduction of a number of azo food dyes were measured in anaerobic cultures of Proteus vulgaris. The rates of colour loss were found to be zero order under conditions in which the concentration of viable cells remained constant. 2. A significant increase in the rate of reduction followed the onset of cell mortality. 3. The zero-order rates correlate with the redox potentials of the dyes. A mechanism consistent with these observations involves an extracellular nonenzymic reducing agent which acts as an electron shuttle between dye and cellular reducing enzymes.
Introduction Bacterial reduction of azo dyes takes place both in vitro (Roxon, Ryan & Wright, 1 9 6 7 a ; Scheline, Nygaard & Longberg, 1970) and in the gut of animals (Radomski & Mellinger, 1962 ; Schroder & Johansson, 1973). These studies have dealt primarily with identification of metabolic products and elucidation of the biochemical reduction mechanism (Fore, Walker & Golberg, 1967 ; Roxon, Ryan & Wright, 1967 b ; Walker, Gingell & Murrells, 1971 ; Gingell & Walker, 1971 ; Gingell & Bridges, 1973). While some attempts have been made to relate the ease of reduction of azo dyes to molecular parameters (Walker & Ryan, 1971), such structure-activity relationships have been difficult to establish, in part because reduction rates in bacterial cultures are not easy to measure in a reproducible manner. In contrast to the status of investigations on bacterial reduction, rates of reduction of azo dyes in irradiated solutions (Giles, Hojiwala & Shah, 1974 ; Bridgeman & Peters, 1974) and in the presence of various reducing agents (Conant & Pratt, 1926; Brownley & Lachman, 1963) have been measured extensively. T h e rates of photochemical reduction of these dyes in the presence and absence of reducing agents depend on the reduction potential of the dye (Kozeny & Velich, 1960 ; Rao & Hayon, 1973). I n the present study, reduction rates of a number of water-soluble dyes were measured in whole cell cultures of P. vulgaris. T h e rate constants of the various dyes correlate with their redox potentials. A reduction mechanism is suggested that explains both this correlation and the observed zero-order reduction kinetics.
P. Dubin and K . L. Wright
5 64
Experimental Materials Orange I1 was synthesized by diazotization of sulphanilic acid and coupling to P-naphthol. Ponceau 3R ( C I 16155) was obtained from Pfalz & Bauer, Inc. (Flushing, New York). All other dyes (see Fig. 1) were certified colour
Xenobiotica Downloaded from informahealthcare.com by University of Queensland on 04/23/13 For personal use only.
additives (Allied Chemical Co., Morristown, New Jersey). Dye samples were not further purified before use. T h e purity of the dyes was determined by azo reduction with excess TiCl, under N, followed by back-titration of untreated Ti3+with Fe3-'-(Siggia, 1963).
0
S03Na
S03Na
N \N
N,\N
Na03S Amaranth
Orange
II
Ponceau 3 R
0 Q S03Na
S03Na
kN
Ponceax SX
Sunset Yellow
Fig. 1.
Tortrazine
Structures of azo dyes used.
Cell culture Proteus vulgaris ATCC (8427) was maintained on a slope of Difco tryptocase soy agar. A 50 ml overnight culture in Difco tryptocase soy broth was grown from the slope. This standing culture (10 ml) was then used to inoculate 500 ml of tryptocase soy broth. T h e final culture was grown without agitation for 12 h at 37" to yield a viable cell density of 2-4 x 108 cells/ml. T h e concentration of viable cells in mixtures of such cultures with dye solutions remained constant under anaerobic conditions for 16 h or more. When cultures with appreciably higher cell densities were used, the viable cell count declined rapidly once anaerobic conditions were attained. Cell counts were determined by serial dilution and plating on tryptocase soy agar.
Xenobiotica Downloaded from informahealthcare.com by University of Queensland on 04/23/13 For personal use only.
Reduction of Axo Dyes by P. vulgaris
565
Dye incubation Dye solutions (ca. 1.5 x M ) were prepared in deionized water and sterilized by autoclave (121", 20 min). Aliquots (8 ml) of the P. vulgaris culture were dispensed into sterile screw-cap culture tubes. Usually 2-0 ml of sterile dye solution was then added to give an azo concentration ca. 3 x lo5 M. Tubes containing bacterial culture alone were retained for later use as spectrophotometric blanks. Samples and blanks were incubated in the dark at 37" in BBL Gas Pak@ anaerobic jars (BioQuest, Cockeysville, Maryland). Anaerobic conditions were attained in about two hours as determined by the colour change of the methylene blue indicator. Samples to be analysed at a given incubation time were contained in a single jar in order to maintain anaerobic conditions for the other samples. Measurement of azo reduction rates Cells were removed from the incubation mixtures by centrifugation at 12 000 g and 4" for 10 min. T h e visible absorption spectrum of the transparent supernatant was measured against the appropriate reference with a Varian Techtron 635 Spectrophotometer. T h e reference solution consisted of the supernatant from culture grown in the absence of dye, diluted with deionized water to correct for addition of the dye solution to the sample. This blank nullified any contribution of the spent medium to the visible spectrum of the supernatant. Chromophore concentrations were determined from the blankof the given dye, relative to the absorbance of corrected absorbance at the A,, an unincubated mixture of dye and medium. T h e use of this reference compensated for indirect effects of the medium on the dye spectrum. I n separate experiments, the removal of dye by binding to precipitated cells was found to be negligibly small. Rates of azo reduction were determined graphically from the dependence of chromophore concentration on incubation time. T h e absolute rate values were sensitive to assay conditions, e.g., the reduction rate for Sunset Yellow (FD&C Yellow No. 6) determined in nine separate assays varied from 5.0 to 7.5 x lo-' mol 1-1 h-l. T h e data from assays with individual dyes were standardized by measuring rates relative to Sunset Yellow, which was included in each assay.
Results and discussion Cell viability T h e viable cell densities of the incubation mixtures were measured as a function of incubation time. Under the conditions described, viable cell density was constant during the first 18-24 h. During the subsequent 12 h an order-ofmagnitude decrease in viable cell count was typically observed. An abrupt increase in dye reduction rate was also found to occur within the period of cell mortality, as shown in Fig. 2. Reduction rates reported in the following always refer to data obtained during the first 24 h of anaerobic incubation. Azo reduction rates Azo reduction rates were determined from plots of chromophore concentration against anaerobic incubation time. Representative data are shown in Fig. 3.
P. Dubin and K. L. Wright
Xenobiotica Downloaded from informahealthcare.com by University of Queensland on 04/23/13 For personal use only.
5 66
Anaerobic incubation time ( h )
Fig. 2.
Dependence on anaerobic incubation time of viable cell density ( 0) and of chromophore concentration for Sunset Yellow ( A) and tartrazine (0). Dyes were incubated with anaerobic cultures of P . vulgaris for the indicated times. The cells were removed by centrifugation and the dye concentration in the supernatant was determined from the visible spectrum measured against the appropriate blank (Sunset Yellow, ; tartrazine, 0).Viable cell densities ( 0 )were obtained by serial dilution and plating on tryptocase soy agar of a random sampling of dye incubation mixtures.
L
c c>
I
1
I
I
I
Anaerobic incubction time (h)
Fig. 3.
Representative plots of dye concentration against anaerobic incubation time used to determine rates of reduction. ( 0) Sunset Yellow. (0) Ponceau 3R. ( A )Amaranth.
Reduction of A x o Dyes by P. vulgaris
5 67
Table 1. Reduction of water-soluble azo dyes in P. vulgaris cultures, relative to Sunset Yellow
Xenobiotica Downloaded from informahealthcare.com by University of Queensland on 04/23/13 For personal use only.
Dye
Orange I1 Amaranth Ponceau SX Sunset Yellow Ponceau 3R Tartrazine
CI Number
Reduction rate relative to Sunset Yellow (R’)
15518 16185 14700 15985 16155 19140
6.3 2.0 1.5 1*o 0.55 0.41
T h e linear dependence of chromophore concentration on time indicates reduction rates zero order with respect to dye concentration. Furthermore, identical reduction rates were observed for Sunset Yellow at initial concentrations differing by a factor of two. T h e values of R‘, the zero order reduction rate relative to Sunset Yellow, are given in Table 1. T h e initial dye concentrations were within the concentration range bracketed by the two assays for Sunset Yellow shown in Fig. 3. Relationship of reduction rate ( R ‘ ) to dye reduction potential T h e rate of photochemical reduction of azo dyes depends on the dye redox potential (Rao & Hayon, 1973 ; Kozeny & Velich, 1960). A similar correlation might be expected for bacterial azo reduction if the rate-controlling step involved a redox equilibrium between dye and an extracellular reducing agent. It appears likely that soluble flavins can play the role of the latter (Roxon et al., 1967 b ; Gingell & Walker, 1971). Other observations in the literature are in accord with a mechanism involving redox equilibrium. Reduction rates correlate with the redox potential of the local environment, both in vivo (Schroder & Johansson, 1973) and in vitro (Gingell & Walker, 1971). T h e effects of substituents on the reduction rate have been rationalized on the basis of effects on the electron density a t the azo bond (Walker & Ryan, 1971). T h e redox potentials of azo dyes may be estimated from the polarographic half-wave potentials. T h e relationship between the redox potential, E,, and the half-wave potential, Eliz, is given by the following expression :
(see for example Willard, Merritt & Dean, 1968), where n is the number of electrons transferred in the redox process and D is the diffusion coefficient of either the oxidized or reduced member of the redox pair. It is reasonable to assume that the diffusion coefficients of the azo and hydrazo species are nearly equal and hence to neglect the second term in the above expression for Ellz. T h e dependence of the relative reduction rate in whole-cell P. vulgaris cultures ( R ’ )on the polarographic half-wave potential, El/,, is shown in Fig. 4 for the dyes of this study. (Redox potentials in biological systems are generally reported
P . Dubin and K . L. Wright
Xenobiotica Downloaded from informahealthcare.com by University of Queensland on 04/23/13 For personal use only.
568
relative to the standard hydrogen electrode (S.H.E.) ; the usual polarographic reference is the saturated calomel electrode (S.C.E.). We have chosen for clarity to adhere to the latter convention, noting that E(S.C.E.) =E(S.H.E.) 0.246 V. I t is also appropriate to note the marked p H dependence typically reported for Ell2. T h e values used here were either measured at or interpolated to pH 7.) T h e logarithm of the relative rate constant varies in a linear manner with the redox potential of the dye. T h is correlation between reduction rate and an electrochemical property of the dye suggests that the rate-determining step in the bacterial reduction of the chromophore does not involve a structurespecific phenomenon such as selective membrane permeation or enzyme binding.
t
80/
:,
60
L?
/
"/
I
- 700
'
I
I
I
- 600
- 500
- 400
EL (pH 7,vs S C.E 1 2
(mV)
Fig. 4. Dependence of relative veduction rate (R') on dye redox potential. Reduction rates relative to Sunset Yellow were determined graphically as shown in Fig. 3. Dye redox potentials were obtained from literature polarographic data (at pH 7, vs. S.C.E.) from the following references : ( 0 )Orange I 1 (Florence, 1964 a) ; (A) Amaranth (Iijima, 1967 ; Mizunoya & Kita, 1965 ; McKeown & Thomson, 1954) ; (0) Sunset Yellow (Nazario & Zenebon, 1972) ; ( 0 )Ponceau SX (Mizunoya & Kita, 1965) ; ( A ) Ponceau 3R (hlizunoya & Kita, 1965) ; (m) Tartrazine (Florence, 1964 a ; Mizunoya & Kita, 1965 ; Nazario & Zenebon, 1972).
Zero-order dependence of reduction rate on dye concentration T h e rate of colour loss was zero order with respect to dye, over a three- to four-fold range in azo dye concentration. T h e absence of any effect of dye concentration on reduction rate is incompatible with simple diffusion of dye through bacterial membrane as a rate controlling step. Two types of interactions between dye and bacterial culture consistent with the observed kinetics may be proposed : (1) saturation by the dye of either active membrane transport sites or the binding sites of a general azo reductase,
Reduction of Azo Dyes by P. vulgaris
5 69
Xenobiotica Downloaded from informahealthcare.com by University of Queensland on 04/23/13 For personal use only.
(2) extracellular (presumably nonenzymic) dye reduction, the rate of which is controlled by a process independent of dye concentration. The transport or binding processes involved in the first hypothesis imply a structural specificity not in accord with the simple dependence on redox potential shown above ; hence, mechanism (2) is to be preferred on the basis of the present results. Postulated mechanism Experimental evidence in support of mechanism (2) has also been observed by others. Roxon et al. (1967 b) reported that the presence of a flavin in the medium was essential for the reduction of tartrazine in whole-cell suspensions of P. vulgaris. Gingell & Walker (1971) also noted that the addition of soluble flavin accelerated the reduction of Red 2G in a cell-free preparation of Streptococcus faecalis. In addition, the rate of reduction was found to be zero order with respect to dye and to depend on the reduction potential of the system. On the basis of these results, Gingell & Walker proposed that the soluble flavin acts as an electron shuttle between the dye and a reducing enzyme, the rate being controlled by generation of reduced flavin. We will show that this type of mechanism is consistent with the kinetics and the dependence on dye redox potential observed here with whole cells of P . vulgaris. The enzymic generation and subsequent oxidation by dye of a low molecular weight reducing agent may be written as k,
B+EH, -+ BH,+E slow kz
BH,+ RN=NROH
k-a
B + R(NH),ROH
(2)
where BH,/B and EH,/E are the redox couples for the low molecular weight electron carrier and enzyme, respectively. The measured rate reflects only the loss in azo concentration and is insensitive to the subsequent fate of the hydrazo compound. It may be noted, however, that in both potentiometric (Conant & Pratt, 1926) and polarographic (Florence, 1964 b) reduction the phenolic hydrazo compound undergoes subsequent disproportionation and reduction : R(NH),R'OH
+RNH, ka
+ R' /O LNH
slow
BH, + R' No
NNH
e" R /OH '-4
+B
\NH,
Steps (1) and (2) may be written as a redox cycle :
X.B.
2P
570
P . Dubin and K. L. Wright
with reducing power provided by bacterial metabolism. No net reduction of dye occurs until the concentration of BH, is such that the reduction potential of the BH,/B couple approaches that of the dye, i.e.,
E = EO - 0.030 log
0= En, (dye).
Xenobiotica Downloaded from informahealthcare.com by University of Queensland on 04/23/13 For personal use only.
(B) Subsequently, the reduction rate, i.e., colour loss, will depend on step ( l ) , the rate of formation of BH,. T h e rate of BH, production may be seen from equation (1) to depend on the concentration of B. If the sum (BH,) +(B) is constant, the concentration of B is controlled by the ratio (BH,)/(B), which, according to equation (5) is determined by the dye redox potential. Thus, the electrochemical properties of the dye may indirectly control the reduction rate by determining the concentration of B in the system, although the dye molecule is not itself involved in the rate-determining step. T h e rate of colour loss would then be zero order with respect to dye (until extremely low dye concentrations are reached) and dependent on dye redox potential, in accord with the results cited above. We have considered only steps (1) and (2) so far. I t is generally thought that step ( 3 ) is slow and rate-controlling for the complete polarographic or chemical reduction of hydroxy azo compounds to primary amines (Conant & Pratt, 1926 ; Florence, 1964 b ; Florence, 1974). However, the rate of this unimolecular disproportionation would not affect the rate of colour loss if formation of BH, were the slowest process, as would be expected if step (1) involved diffusion through a cell membrane. I n the presence of a species capable of reducing the azo bond, the reduction of the imino-quinone (step (4)) is expected to be rapid and complete (Conant & Pratt, 1926). Step (4) can then only influence the rate of colour loss via consumption of BH,. Thus a decrease in the rate of colour loss might be expected when azo concentration is sufficiently low for the rate of BH, consumption in step (4) to effectively compete with that in step (2). Implicit in the foregoing analysis is the assumption of constant (BH,) + (B). Introduction of additional reducing agent would not alter the reaction order or the dependence of rate on dye redox potential, but would increase the rate of colour loss. While no such effect was observed under experimental conditions such that viable cell count was constant, acceleration of the rate was noted subsequent to cell death. Effect of cell mortality on dye reduction rate T h e increase in dye reduction rate observed upon cell death, as shown in Fig. 2, could arise from release into the medium by lysed cells of agents that provide an alternative reduction mechanism. However, the reduction process after cell death apparently remains zero order in dye ; this would be unexpected if the increase in reduction rate corresponded to a large increase in the concentration of metabolites capable of directly reducing the dye. On the other hand, the observed enhancement in reduction could correspond to an increase in the efficiency of the postulated mechanism arising from the disruption of membrane barriers to permeability of reducing agent. Such change in accessibility of cellular enzymes to extracellular reducing agents would effectively increase the
Reduction of A x o Dyes by P. vulgaris
571
Xenobiotica Downloaded from informahealthcare.com by University of Queensland on 04/23/13 For personal use only.
forward rate constant of step (1 ) and accelerate reduction without altering reaction order, as is observed. Roxon, Ryan & Wright (1966, 1967 b) noted that starved cultures of P. vulgaris are much more active than young cultures in reducing tartrazine. They ascribed this effect to an increase in cell permeability to tartrazine (Roxon et al., 1967 b). However, if transport of dye across the cell membrane were intrinsic to the reduction process, it would be expected that the rate of reduction would be influenced by both the concentration and the shape and ionic charge of the dye, in sharp contrast to the present results.
Acknowledgments T h e authors express their appreciation of helpful comments from Dr. J. P. Brown. T h e assistance of Dave Cartopassi is gratefully acknowledged.
References BRIDGEMAN, I. & PETERS,A. T. (1974). Text. Res., 44, 639. BROWNLEY,C. A., Jr. & LACHMAN, L. (1963). J . Pharm. Sci., 52, 86. CONANT, J. B. & PRATT,M. F. (1926). J. A m . chem. Soc., 48, 2468. FLORENCE, T. M. (1964 a). Aust. J. Chem., 18, 609. FLORENCE, T. M. (1964 b). Aust. J . Chem., 18, 619. FLORENCE, T. M. (1974). J. electroanal. Chem., 52, 115. FORE,H., WALKER, R. & GOLBERG, L. (1967). Fd. C o m e t . Toxic., 5, 459. GILES,C. H., HOJIWALA, B. J., & SHAH,C. D. (1974). J . SOC.Dyers Colour., 90,45. R. & BRIDGES, J. E. (1973). Xenobiotica, 3, 599. GINGELL, GINGELL, R. &WALKER, R. (1971). Xenobiotica, 1, 231. IIJIMA,T. (1967). Rev. Polarogr. (Japan), 14, 317. KOZENY, M. & VELICH,V. (1960). Coll. Czech. chern. Commun., 25, 1031. C. G. & THOMSON, J. L. (1954). Can.J . Chem., 32, 1025. MCKEOWN, Y. & KITA,T. (1965). Japan Analyst, 14, 437. MIZUNOYA, G. & ZENEBON, 0. (1972). Reota. Inst. Adolfo Lutz, 32, 101. NAZARIO, J. L. & MELLINGER, T. J. (1962). J . Pharmac. exp. They., 136, 259. RADOMSKI, RAo, P. S. & HAYON, E. (1973). J . phys. Chem., 77, 2753. ROXON, J. J., RYAN,A. J. & WRIGHT,S. E. (1966). Fd. Cosmet. Toxic., 4, 419. J. J., RYAN,A. J. & WRIGHT,S. E. (1967 a). Fd. Cosmet. Toxic., 5, 367. ROXON, J. J., RYAN,A. J. & WRGIHT,S. E. (1967 b). Fd. Cosmet. Toxic., 5, 645. ROXON, SCHELINE, R. R., NYGAARD, R. T. & LONGBERG, B. (1970). Fd. C o m e t . Toxic., 8, 55. SCHRODER, H. & JOHANSSON, A. K. (1973). Xenobiotica, 3, 233. S. (1973). Quantitative organic analysis, p. 526, New York : John Wiley and Sons. SIGGIA, WALKER, R., GINGELL, R. & MURRELLS, D. F. (1971). Xenobiotica, 3, 221. R. & RYAN,A. J. (1971). Xenobiotica, 1, 483. WALKER, H., MERRITT,L. L. & DEAN,J. A. (1968). Instrumental methods of analysis, WILLARD, p. 684, Princetown : Van Nostrand Co.
2P2
