APPLIED AND ENVIRONMENTAL MICROBIOLOGY, May 1993, p.

1507-1514

Vol. 59, No. 5

0099-2240/93/051507-08$02.00/0

Effects of Chemical Speciation in Growth Media on the Toxicity of Mercury(II)t

RICHARD E. FARRELL,* JAMES J. GERMIDA, AND P. MING HUANG Department of Soil Science, University of Saskatchewan, Saskatoon, Saskatchewan 57N OWO, Canada
Received 12 October 1992/Accepted 9 March 1993

The toxicity of metals, including mercury, is expressed differently in different media, and the addition of soluble organics to the growth medium can have a significant impact on bioassay results. Although the effect of medium composition on metal toxicity is generally attributed to its effect on metal speciation (i.e., the chemical form in which the metal occurs), the importance of individual metal-ligand species remains largely unclear. Here, we report the results of a study that investigated, both experimentally and from a modeling perspective, the effects of complex soluble organic supplements on the acute toxicity (i.e., 50%o inhibitory concentration [ICsaJ) of mercury to a Pseudomonasfluorescens isolate in chemically well-defined synthetic growth media (M-IIX). The media consisted of a basal inorganic salts medium supplemented with glycerol (0.1%, vol/vol) and a variety of common protein hydrolysates (0.1%, vol/vol), i.e., Difco beef extract (X = B), Casamino Acids (X = C), peptone (X = P), soytone (X = S), tryptone (X = T), and yeast extract (X = Y). These were analyzed to obtain cation, anion, and amino acid profiles and the results were used to compute the aqueous speciation of Hg(II) in the media. Respirometric bioassays were performed and IC,Os were calculated. Medium components varied significantly in their effects on the acute toxicity of Hg(II) to the P. fluorescens isolate. IC50s ranged from 1.48 to 14.54 ,ig of Hg ml-,, and the acute toxicity of Hg(H) in the different media decreased in the order M-IIC > M-IIP > M-IIB > M-IIT > M-IIS >>> M-IIY. The calculated ICsas were related to the aqueous speciation of Hg(II), and a significant negative correlation (r = -0.904) (P = 0.05) between the ICso and the mole fraction of Hg(II) bound in Hg-chloride complexes was observed. This was particularly noteworthy because the only source of chloride in the M-IIX media was the organic supplements. There also was a significant positive correlation (r = 0.831) (P = 0.05) between the IC50 and the mole fraction of free Ca2+ plus Mg2' in the media, suggesting that these ions have a moderating effect on the toxicity of Hg(II). Overall, differences in the mole fractions of Hg-chloride complexes (primarily HgCl+, HgCl20, and HgCIOH) and free Ca21 plus Mg2e accounted for about 96% of the variation in the acute toxicity of mercury (i.e., R2 = 0.955) (P = 0.01).

It is well documented that the bioavailability and toxicity of trace metals to aquatic microbiota are influenced by the chemical form (i.e., aqueous species) of the metal (4, 9, 15). Indeed, the response of an organism to a specific form of a metal, rather than its total concentration, is an important consideration in the development of water quality standards (24). In the case of mercury, bioavailability and toxicity are dependent on the speciation of the mercuric ion (Hg2+) which, in turn, is a function of the concentrations of all potential coordinating ligands in solution. Thus, given the complexity of the growth media used in most toxicity testing schemes, it is not surprising that bioassays for mercury toxicity rarely take Hg(II) speciation effects into consideration. The interpretation of bioassay results, and the extrapolation of these results to in situ conditions, is often complicated by the inclusion of complex soluble organics in the bioassay media. Complex ingredients, such as yeast extract and proteose peptone, have different affinities for Hg(II) (20) and can have a significant effect on the toxicity of the metal (5). The effects of medium composition on Hg(II) toxicity are usually attributed to the impact of the medium on the availability (activity) of the free aquo ion (Hg2+). It has been demonstrated, however, that some complexes of Hg(II) with

* Corresponding author. t Contribution R707 of Saskatchewan Institute of Pedology.

1507

inorganic and low-molecular-weight organic ligands can contribute significantly to the overall toxicity of Hg(II) (6, 10, 12, 17, 19). Consequently, any attempt to elucidate the biological effects of Hg(II) should include a careful consideration of the composition of the growth medium and its impact on the chemical speciation of the Hg(II) ion. It is much more difficult to quantitatively determine the individual metal-ligand species in a sample than to determine the total concentrations of metals and ligands. Moreover, it is unrealistic to think that the quantitative determination of all metal-ligand species in a sample can be incorporated into toxicity testing schemes on a routine basis. Computer modeling of metal speciation in natural and synthetic media, however, provides a relatively rapid and inexpensive alternative to the quantitative determination of individual metalligand species. Computer programs such as MINTEQA2 (1) and GEOCHEM-PC (18) have been developed for calculating the equilibrium composition of dilute aqueous solutions and can be used to calculate the distribution of chemical species in natural and synthetic systems. Models such as these provide us with the ability to predict the chemical form of trace metals in complex media, provided, of course, that the chemical composition of the medium is well-defined and that valid stability constants for the relevant metal-ligand species are available. Computer modeling of metal speciation in bioassay media, conducted in parallel with toxicity tests, may be helpful in interpreting the biological effects of heavy metals.

1508

FARRELL ET AL.

APPL. ENvIRON. MICROBIOL.

Previously (12), we used the GEOCHEM-PC program to calculate the effects of pH and added ligands on the aqueous speciation of Hg(II) salts in a synthetic medium containing yeast extract and used the results to evaluate the influence of Hg(II) speciation on the toxicity of mercury. Complexes of Hg(II) with chloride and various amino acids were identified as being positively correlated with the acute toxicity of Hg(II). Prompted by these results and the affinity studies of Ramamoorthy and Kushner (20), we speculated that the toxicity of Hg(II) in a basal salts medium enriched with different soluble organic supplements would vary and that differences in toxicity would reflect variations in the aqueous speciation of the Hg(II). That is, both the affinity of the medium supplements for Hg(II) and the acute toxicity of Hg(II) should vary as functions of the aqueous speciation of the metal. Thus, the present investigation was initiated (i) to evaluate the effects of complex soluble organics on the acute toxicity of Hg(II) to a Pseudomonas fluorescens isolate in chemically well-defined synthetic growth media, (ii) to compute the effects of these organics on the aqueous speciation of Hg(II) in the media, and (iii) to ascertain the dependence of toxicity on the chemical speciation of Hg(II).
MATERIALS AND METHODS Bacterial culture. The test organism was a P. fluorescens isolate obtained from a composite sample of surface sediments collected at Buffalo Pound Lake in the upper Qu'Appelle River basin, Saskatchewan, Canada. A pure culture of the isolate (BPL85-48) was obtained as described by Farrell et al. (12) and was tentatively identified by using the API 20E system (3). Stock cultures were prepared by streaking the isolate onto tryptic soy agar slants, which were incubated for 48 h at 25C and then stored at 4C. Working cultures were prepared by streaking the stock culture onto fresh tryptic soy agar plates, which were incubated for 48 h at 25C and then stored at 4C. Fresh working cultures were prepared every 6 to 8 weeks. Toxicity assessment. Bioassays based on the measurement of respiration (i.e., CO2 production) were used to evaluate the acute toxicity of mercuric nitrate in a variety of synthetic growth media. Media used in this study included a basal inorganic salts medium (M-I), the M-I medium supplemented to contain 0.1% (vol/vol) glycerol (M-II), and the M-II medium supplemented to contain 0.1% (wt/vol) of the various soluble organic components (M-IIX, where X is the code designation for the organic component). The organic components (Difco Laboratories, Detroit, Mich.) evaluated were beef extract (X = B), Casamino Acids (X = C), peptone (X = P), soytone (X = S), tryptone (X = T), and yeast extract (X = Y). The M-I medium was based on the minimal salts medium of Chan (11) but was prepared Cl free by substituting appropriate nitrate salts. The M-I medium consisted of the following (in milligrams per liter of deionized water): KN03 (1,000), KH2P04 (500), K2HPO4 (500), (NH4)2SO4 (1,000), CaSO4 2H20 (50), MgSO4 7H20 (50), MnSO4
H20 (10), and FeSO4 7H20 (10).

The bioassay procedure was as follows: (i) inoculum (sediment isolate BPL85-48) was grown in 50 ml of the appropriate bioassay medium (e.g., M-IIY) at 25C for 16 h; (ii) cells (in early to mid-log phase) were harvested by centrifugation (1,000 x g for 10 min), washed, and resuspended in 50 ml of M-I medium; (iii) the optical density of the resuspended cells was adjusted to a value of 105 + 5 Klett units with sterile M-I, yielding a cell density of approximately 109 cells ml-1; (iv) 1.0-ml aliquots of the cells were

added to 15-cm3 serum bottles containing 1.0 ml of M-II medium supplemented with the appropriate organic compound (0.2%, wt/vol) and mercuric nitrate (0 to 30 ,ug of Hg ml-1); and (v) the bottles were sealed with rubber septa and triplicate samples were incubated at 25C for 6 h, at which time the percent CO2 was determined by injecting 1.0 cm3 of the serum bottle atmosphere into a Fisher-Hamilton gas partitioner equipped with a thermal conductivity detector and connected to a Hewlett-Packard 3390A integrator. Controls were prepared by inoculating Hg-free media and were included in each trial. Blanks consisting of uninoculated media were used to measure the amount of ambient CO2 in the headspace of the serum bottles under no-growth conditions. Respiration was calculated as a fraction of the control and plotted against the total Hg concentration. The concentrations of total Hg that inhibit respiration by 50% (IC50s) were calculated from equations obtained by regression analysis of the CO2 versus Hg concentration data. The pH of the bioassay media was adjusted to 8.05 + 0.05 prior to the start of the toxicity tests and was measured again immediately after the serum bottle atmosphere was sampled. Tukey's honest significant difference was used to analyze the data from the toxicity assessment study. The study consisted of two trials with three replicates for each of the six treatments. Correlations between the IC50 and the activity or mole fraction of various Hg(II) species were assessed by using the Pearson correlation coefficient. Regression analysis was used to determine the nature of the relationship for comparisons yielding a significant correlation coefficient. Analyses of the organic medium components. The organic medium supplements were analyzed for (i) amino acid content (free and hydrolyzable) and ammonia, (ii) major and trace cations, and (iii) major and trace anions. Amino acid analyses were performed by using the Pico tag amino acid analysis system (25) and were carried out by the Industrial Microbiology Research Laboratory of the Department of Applied Microbiology and Food Science, University of Saskatchewan. Cation and anion profiles were determined in accordance with the protocols set forth by the American Public Health Association (2). Major and trace cations were determined by inductively coupled plasma-atomic emission spectroscopy at the Analytical Services Division of the Saskatchewan Research Council. Major and trace anions were determined at the Saskatchewan Soil Testing Laboratory by ion chromatography (Cl- and S042-) and titration (HCO3-) and by using a Technicon autoanalyzer (NO3- and PC43-). The cation, anion, and dissolved free amino acid profiles of the various organic medium components are presented in Tables 1 and 2. Mercury speciation. The aqueous speciation of mercury [i.e., the distribution of aqueous Hg(II)-ligand species] in the bioassay media was calculated by using the computer program GEOCHEM-PC (18). The variables used to compute Hg(II) speciation were pH (both initial and final), total metal concentration, total ligand concentration, and ionic strength. The input species were identified as the free metal cations, H+, and the free ligands. The data base (GEODATA) had been modified previously by Farrell et al. (12) to include stability constants for the various combinations of metals and ligands listed in Tables 1 and 2. The output data were expressed in terms of activity and mole fraction. Activity refers to the effective concentration of the various complexes in solution and is defined mathematically as follows: ai = fci, where ai is the activity of the aqueous species i, ci is the actual concentration of that species, and f is the activity coefficient. In general, the mole fraction (X) of a

VOL. 59, 1993

GROWTH MEDIUM EFFECTS ON Hg(II) BIOASSAYS

1509

TABLE 1. Elemental composition of the organic growth medium components

Analatea

Concn in medium component (,ug g-l)b

Beef extract

Casamino Acids

Peptone

Soytone

Tryptone

Yeast extract

Metals Calcium (Ca2") Magnesium (Mg2l) Potassium (K+) Sodium (Na+) Iron (Fe2+) Manganese (Mn2+) Copper (Cu2+) Barium (Ba2+)

56.01 157.1 73,118

Cadmium (Cd2+) Zinc (Zn2 2) Nickel (Ni +) Lead (Pb2+) Cobalt (Co2+) Chromium (Cr3+) Aluminum (A13+) Titanium (TiO2+)
Ligandsc
Carbonate (CO32-) Sulfate (SO42-) Chloride (Cl-) Ammonia (NH30) Phosphate (PO43-) Silica [SiO2(OH)22-] Boron [B(OH)4-j Molybdate (MoO42-)

46,671 40.45 1.87 1.24 0.62 2.18 2.18

1.24

140.5 113.7 6,355 96,990 14.72

6.69 0.67 0.33 18.73 4.35

40.16

2,075 16,734 11.38

1.34 1.34

188.7 1,799 26,555 45,423 62.89 1.75 3.84 0.70 11.88 16.95 8.74 8.56 28.13 1.40

20.08 1.34

679.4 475.5 2,174 40,761 37.36 1.70 2.72 7.81 0.34 47.55 0.68
2.04 37.36 1.36

940.5 957.3 60,464 3,829 50.39 1.01 3.53 2.35 62.14 3.02 2.69 0.67 9.91 0.34
48,035 5,374 3,359

12.45 4.98
1.87 17.11 0.62

3.01 7.36 0.33

1.34 10.38 0.34 32,129 2,677

69,384

1,867 37,337

21,780
124.5 2.49

37,458 1,672 107,692 7,510 10,368

1.67

12,383 1,473
100.4

54,158 51,013 3,145

67,255 1,359 4,416

2,568 314.5 69.88 5.24

10,870 203.8

12,261 177.6

a Symbols in parentheses represent the aqueous chemical species entered into the GEOCHEM-PC program. b Values are not reported for metals and ligands present at concentrations less than the limits of detection. c Nitrate (NO3-) was not detected in any of the medium components.

substance is defined as the ratio of moles of one substance to the total moles of all substances in a mixture; here we define the mole fraction of a metal-ligand species, i, as the ratio of the molar concentration of that species (ci) to the total concentration of the metal (CT); i.e., Xi =c'CFC.

RESULTS ducted to determine the effects of soluble organic growth medium supplements on the acute toxicity of mercury(II) to P. fluorescens isolate BPL85-48. Fresh, actively growing cells are generally more susceptible to toxicants; therefore, inocula were prepared from cultures in early to mid-log phase of growth. A 6-h incubation was chosen to encompass the periods of active cell growth (3 to 4 h) and maximum respiratory activity (data not shown). Typical Hg(II) response curves for the isolate are presented in Fig. 1, and the results of the bioassays, reported in terms of the IC50s, are summarized in Table 3. The organic supplements varied significantly in their effect on both microbial activity (as indicated by the maximum amounts of CO2 produced) and the acute toxicity of Hg(II). In the absence of added Hg(II), microbial activity increased in the order M-IIP = M-IIC < M-IIB < M-IIY = M-IIS < M-IIT. On the other hand, IC50s increased in the order M-IIC < M-IIP < M-IIB < M-IIT < M-IIS <<< M-IIY. It was interesting that the lowest IC50s and control (maximum) respiration occurred in the media supplemented with Casamino Acids (M-IIC) and peptone (M-IIP) but that the highest IC50 (M-IIY) did not correspond to the medium with the greatest CO2 maximum, i.e., the M-IIT medium. Indeed, although the IC50 for the M-IIY

Toxicity assessment. Respirometric bioassays were con-

medium was more than double that obtained for the M-IIT medium (Table 3), the maximum amount of CO2 produced by the P. fluorescens isolate in the M-IIY medium was about 30% less than that produced in the M-IIT medium. Likewise, although the maximum amounts of CO2 produced in the M-IIS and M-IIY media were essentially the same, the IC50 for the M-IIS medium was about 45% less than that obtained for the M-IIY medium (Table 3). Similar results were observed for the M-IIC and M-IIP media (Table 3). Mercury(H) speciation. The computer program GEOCHEM-PC (18) was used to calculate Hg(II) speciation in the different growth media. The free aquo ion (Hg2+) and 52 Hg(II)-ligand species were considered in the calculations. The primary Hg(II)-ligand species in each of the growth media (those present at concentrations of .0.01 mol% of the total mercury) are listed in Table 4 (calculated mole fractions based on the final pH of the media). Whereas the distribution of Hg(II) among the primary species varied between growth media, the 17 species listed in Table 4 accounted for more than 99.7 mol% of the total Hg(II). In all cases, the pH of the M-IIX media decreased during the 6-h incubation (Table 4), with the ApH ranging from -0.34 to -0.55 pH unit. Based on pH effects, the primary Hg(II)-ligand species were classified as being either group I species (activity increases as pH increases) or group II species (activity increases as pH decreases). At the lower pH values, small increases in the amounts of Hg(II) bound in group II complexes [i.e., HgCl20, HgClOH, HgH(HIS)2', and HgH2(HIS)22+] were offset by equally small decreases in the amounts of Hg(II) bound in complexes with the remaining (group I) species. Repartitioning of the Hg(II) had no

1510

FARRELL ET AL.

APPL. ENVIRON. MICROBIOL.

TABLE 2. Dissolved free amino acid profiles of the organic growth medium components
Amino acida
Beef extract

Concn in medium component (mg of amino acid g-l)b

Casamino Acids Peptone

Soytone

Tryptone

Yeast extract

Aspartate (ASP-) Asparagine (ASN-) Glutamate (GLU2-) Glutamine (GLN-) Serine (SER-) Glycine (GLY-) Histidine (HIS-) Threonine (THR-) Alanine (ALA-) Arginine (ARG-) Proline (PRO-) Tyrosine (TYR2-) Valine (VAL-) Methionine (MET-) Cysteine (CYS2-) Isoleucine (ILE-) Leucine (LEU-) Phenylalanine (PHE-) Tryptophan (TRP-) Ornithine (ORN-) Lysine (LYS-)

1.63 0.90 17.20 9.32 3.49 1.46 13.54 2.34 25.12 2.79 1.84 1.72 2.80 0.80 3.15 1.91 3.59 1.51 1.38 1.23 3.04 128.1 280.2

91.56
179.31 38.56 50.17 12.84 11.65 40.46 38.80 32.15 122.25 4.81 55.84 24.26 0.68 41.13 84.03 31.10

0.19 81.01
940.8 869.8

4.94 2.97 11.80 16.11 5.95 6.72 1.79 1.62 10.19 31.62 2.77 6.07 8.09 2.92 1.59 6.66 19.07 13.24 3.13 0.66 12.61 170.5 557.8

2.19 3.79 4.85 4.63 4.20 1.17 2.69

2.57 25.53 0.45 11.64 4.01 2.54 0.56 5.11 20.98 14.70 3.15 0.27 16.34 131.4 426.3

5.42 6.32 13.95 5.43 10.29 0.12 7.49 9.28 8.97 35.04 1.77 7.22 21.96 16.51 2.94 18.39 79.85 33.08 11.21 0.69 63.08
362.0
737.6

39.41 21.66

90.91
36.04 29.23 7.04 8.85 5.01 58.66 13.74 15.78 10.71 38.37 11.03 1.24 34.78 58.90 30.07 7.81 10.98 33.74

Dissolved free amino acids Total hydrolyzable amino acids

558.6
492.6

a Symbols in parentheses represent the aqueous chemical species entered into the GEOCHEM-PC program. Amino acids listed in italics are those for which stability constant data for the Hg(II)-amino acid species were not available. They were, however, included in the GEOCHEM-PC computations. b Values are not reported for amino acids present at concentrations less than the limit of detection.

significant effect on the total amount of Hg(II) bound by the primary species. The total amount of Hg(II) bound in complexes with the dissolved free amino acids was also relatively unaffected by the change in pH (Fig. 2) and could account for about 18 to 80 mol% of the total Hg(II). Correlations between the IC50 and the total concentrations of all metals and ligands and between the IC50 and the activities or mole fractions of selected Hg(II) species were evaluated by using the Pearson correlation (21). All data were log transformed before the correlation procedure was carried out, and only those Hg(II)-ligand species present at activities greater than that of the free Hg2+ were included in the correlation matrix. Of the nearly 100 correlations examined, only 9 were significant (P c 0.10). The correlation
1.8
I

coefficient data (calculated activities and mole fractions based on the final pHs of the media) are summarized in Table 5. Similar results were obtained when the comparisons were made by using calculated activities (mole fractions) based on the initial pHs of the media, except that the comparisons between IC50s and activities and mole fractions of the free Ca2' and Mg2+ were not significant. As we noted in previous studies (12, 13), the IC50 was more highly correlated to the mole fraction of a Hg(II)-ligand complex than to the activity (or concentration) of the complex. This reflects the fact that the toxicity of a metal is a
TABLE 3. Effect of bioassay medium on microbial respiration and the acute toxicity of mercuric nitrate to

P. fluorescens isolate BPL85-48

Maximum respirationb

A
1.5
1.2 1.2

Bioassay mediuma

(% CO2)

IC50 (xg of
Hg ml-')

0
1

04

++A
-V+

-e

*U 0.6 1.2~~~~~~~~2 0
00* o 0
5
0.3
0

0.9

U
0.9
8

++
0.6
0.3 0

,,o 0.,+ .. ,,......, A

0

VA

10

15

20

:V*+. 0 10

t
20

+ A

+-

M-IIS M-IIP M-IIC M-IIT M-IIY M-IIB

1.189 0.802 0.815 1.671 1.178 1.012 0.166

8.03 4.15 1.48 6.98 14.54 5.00

30

HSD (0.01)c

0.60

9g Hg(Il) ml' FIG. 1. Effects of mercury(II) concentration on the respiratory activity of P. fluorescens isolate BPL85-48 during a 6-h incubation in M-II medium supplemented (0.1%, wt/vol) with soytone (-), peptone (0), or Casamino Acids (*) (A) and with tryptone (A), yeast extract (+), or beef extract (V) (B). Each point is the average of six
replicate
determinations.

gg Hg(ll)

ml"'

a Organic supplements added (0.1%, wt/vol) to the M-II medium were soytone (S), peptone (P), Casamino Acids (C), tryptone (T), yeast extract (Y), and beef extract (B). b Percent CO2 measured in the control (0 ,ug of Hg ml-') vial at the completion of the 6-h incubation period. The reported values were corrected for the amounts of CO2 measured in the control blanks (i.e., medium only). c HSD, Tukey's honest significant difference, at the 1% level of probability, between mean values within columns.

VOL. 59, 1993

GROWTH MEDIUM EFFECTS ON Hg(II) BIOASSAYS

TABLE 4. Computed primary Hg(II)-ligand species present in the bioassay media
Mol% of species in the indicated bioassay mediuma

1511

Hg(II)-ligand species

M-IIC (1.48) [8.11; 7.77]

M-IIP (4.15) [8.03; 7.661

M-IIB (5.00) [8.08; 7.57]

M-IIT (6.98) [8.07; 7.65]

M-IIS (8.03) [8.08; 7.601

M-IIY (14.54)

[8.05; 7.50]

HgCl20
HgClOH

Hg(NH3)2+

Hg(GLY)20
Hg(GLY)OH Hg(CYS)22Hg(H-HIS)HIS+

Hg(H-HIS)22+ Hg(SER)20 Hg(ALA)20

Hg(ALA)OH

Hg(VAL)20
Hg(VAL)OH

Hg(THR)20 Hg(PHE)20

Hg(PRO)20
Hg(OH)20

1.94 0.75 9.62 0.52 1.95 38.08 4.74 0.04 1.21 13.44 7.87 4.02 6.83 0.58 0.76 0.33 7.14

0.17 0.44 17.87 0.57 4.01 31.79 0.57 0.01 0.07 3.58 7.96 0.34 3.86 0.56 0.05 28.15

0.84 0.60 6.74 0.01 0.31 52.15 13.04

0.02 0.13 14.31

0.16 0.01 7.98 7.30 0.02 0.50

0.06 34.77 6.75 0.07 0.16 2.17 5.63 1.90 8.36 0.10 2.72
22.64

0.02 0.23 31.22 0.03 1.23

5.74

2.07 0.02
0.06 0.38

3.50 0.14 3.35

1.21 50.70

0.01 0.07 7.21 0.01 0.40 7.06 5.81 0.09 0.66 39.32 16.83 2.86 7.22 0.01 1.13 11.18

10.34

a Speciation calculations based on the final pHs of the bioassay media. Organic supplements added (0.1%, wt/vol) to the M-II medium were Casamino Acids (C), peptone (P), beef extract (B), tryptone (T), soytone (S), and yeast extract (Y). Numbers in parentheses are the IC50s (in micrograms of mercury per milliliter); numbers in brackets are the initial and final pHs of the medium.

function of both the relative toxicity of individual metalligand species and the relative amounts of each species present in solution. Thus, mole fraction can be considered to be a relative intensity factor which describes the effects of species interactions and, as such, only mole fraction data were considered in the final statistical analyses. Regression analysis was used to further investigate the relationships between the IC50 and the variables that yielded significant Pearson correlation coefficients. Because the computed activities and mole fractions of the various Hg(II)chloride species were very collinear (i.e., highly correlated to each other), they were considered as a group. Likewise, because of collinearity between the activities of Hg(II)chloride species and the total concentration of chloride, only

1.0
0

the combined Hg(II)-chloride species were considered in the regression analysis, the results of which are presented in Fig. 3A. As expected, there was a strong negative linear correlation between the IC50 and the total mole fraction of Hg(II) bound in complexes with chloride (r = -0.904*). The significant correlations between the IC50 and the mole fractions of free Ca2" and Mg2" (Table 5) led us to investigate the relationship between the mole fractions of Hg(II)chloro complexes and free Ca2" and Mg2". It was observed, however, that the Ca2" and Mg2" were also collinear and, hence, they too were combined into a single variable (Ca2" plus Mg2+) for the regression analysis. Not unexpectedly, there was a very strong negative linear correlation between the IC50 and the mole fraction ratio of Hg(II)-chloro complexes to free Ca2+ plus Mg2+ (r = -0.917**; Fig. 3C). Multiple linear regression was also used to evaluate the dependence of the IC50 on the mole fractions of Hg(II)chloro complexes and free Ca2+ plus Mg2+. The two variables were not significantly correlated, and the results of the
TABLE 5. Pearson coefficients (r) computed for correlations between the IC50 and selected metal, ligand, and Hg(II)-ligand species

.E
0

0.8

0.6
.-

0.4
I 0

._C

m: co

cm

0.2
0.0

Speciesa Hg2+ HgCl+

Activity r,bc 0.925**

Mole fraction

rc

0.673NS

HgC120 HgClOH
M-IIC M-IIP M-IIB M-IIT M-IIS M-IIY

0.267NS

-0.751t
-0.097NS -0.893* 0.602NS 0.609NS 0.906*

-0.764t -0.884* -0.870*

Growth Medium FIG. 2. Mole fraction of the total Hg(II) bound in complexes with dissolved free amino acids. The M-II medium was supplemented with Casamino Acids (C), peptone (P), beef extract (B), tryptone (T), soytone (S), or yeast extract (Y). Computed mole fractions are based on the initial and final pHs of the media. Values in parentheses are the IC50s calculated for P. fluorescens isolate BPL85-48 in the different media.

Ca +
a

[Cl]T [Ca]T [Mg]T

Mg2+

0.673NS

0.838*

0.875*

Concentrations, activities, and mole fractions of species were calculated

on the basis of the final pHs of the media.

b Correlations between the IC50 and species listed in brackets involve the total concentration of the metal (Ca and Mg) or ligand (Cl). c Significance at the P = 0.10 (t), P = 0.05 (*), and P = 0.01 (**) levels of probability is indicated. NS, not significant.

1512

FARRELL ET AL.
1.2
1.0

APPL. ENVIRON. MICROBIOL.

1.2
1.0

B a

/
a *

0.8
0

g8 0.6
0.4 0.2 0.0

0.8 0.6
0.4

1.2 1.0 0.8 0.6

0.4
0.2

y--0.472 -0.510x R, R-0.817'

0.2 0.0

y . 5.642 + 9.749x
*

R2

0.691
0.0

y - 0.202 .

R2

0.504x 0.841 *
-2.0
-1.5

Mg2') log Mole Fraction [HgClx (Ca2* + Mgt')1 FIG. 3. Relationshi? between the toxicity index, log IC50, and the logarithm of the mole fraction of Hg(II) bound in complexes with chloride (A) or free Ca + plus Mg2" (B) and the ratio of Hg(II)-chloride to free Ca2" plus Mg2" (C). Each point is the average of six replicate
+

-3.5 -3.0 -2.5 -2.0 -1.5 log Mole Fraction HgClx

-0.54 -0.51 -0.48 log Mole Fraction (Ca2

-0.45

-3.0

-2.5

-1.0

determinations.

analysis are summarized in Table 6. Clearly, the dominant factor affecting the IC50 was the mole fraction of Hg-chloride complexes (F = 54.99**). The inclusion of additional variables in the regression analysis (using a stepwise procedure) failed to produce any significant improvement in the coefficient of multiple determination (R2); thus, we believe that differences in the toxicity of mercuric nitrate in the synthetic M-IIX media are controlled primarily by differences in the amounts of Hg(II)-chloro complexes and free Ca2" and Mg2+ in the media.

DISCUSSION Bioassay media should not be considered chemically inert systems whose sole function is to support microbial growth (7). Indeed, the composition of the medium can have a significant effect on the reactivity and bioavailability of a metal, and a given metal may exhibit vastly different toxicities when cells are grown in different media (4, 7, 16). Consequently, the interpretation of metal toxicity data is often clouded by the effects of medium composition on bioavailability. This is especially true for bioassay media that contain complex soluble organics. Ramamoorthy and Kushner (20) reported that the affinity for Hg(II) of several complex soluble organics increased in the order Bacto Peptone < Bacto Tryptone < yeast extract < proteose peptone << Casamino Acids. Thus, one could reasonably expect to find that the toxicity of Hg(II) in a simple growth medium (e.g., the M-II medium) supplemented with equal amounts of these organics would decrease in a similar order. Thus, the present investigation was conducted to determine the effects of several common growth medium supplements on the acute toxicity of Hg(II) and to ascertain the dependence of toxicity on Hg(II) speciation. Our initial hypothesis was that the acute toxicity of Hg(II) would mirror the affinity
sequence.
a

We found that the composition of the growth medium had significant effect on the acute toxicity of mercuric nitrate to

P. fluorescens isolate BPL85-48. Although there was an apparent tendency for the IC50 to increase as the ability of the medium to support growth increased, there was no significant correlation between the IC50 and the amounts of CO2 produced by the bacterial cultures. Likewise, although the organic supplements contained various amounts of amino acids (Table 2), there was no significant correlation between the IC50 and either the total or dissolved free amino acid content of the media. These results suggest that even though the media varied in their nutrient content and availability, nutritional status was not the major factor influencing Hg(II) toxicity. Because the various soluble organics have different affinities for Hg(II) (16, 20), it is often assumed that complexation of Hg(II) by the organics is the mechanism responsible for the differences in toxicity expressed in the different media. However, our data demonstrated that there was no significant relationship between the amounts of Hg(II) bound by the amino acids and the acute toxicity of mercury. For example, although the amount of Hg(II) bound in complexes with amino acids was essentially the same for the M-IIC, M-IIB, and M-IIY media (Fig. 2), mercuric nitrate exhibited vastly different toxicities in these media: acute toxicity was greatest in the M-IIC medium (IC50 = 1.48 ,ug of Hg ml-'), intermediate in the M-IIB medium (IC50 = 5.00 ,ug of Hg ml-l), and least in the M-IIY medium (IC50 = 14.54 jig of Hg ml-'). Likewise, although the distribution of Hg(II)-amino acid species varied between growth media (Table 4), there were no significant correlations between the IC50 and the mole fractions of any of the individual Hg(II)-amino acid species. Similar results (data not shown) were obtained when the total hydrolyzable amino acid profiles were used in the speciation model. These results indicate that something other than the amino acids was controlling the toxicity of Hg(II) in the M-IIX media. The acute toxicity of mercury increased as the total concentration of chloride in the media increased. Total

TABLE 6. Multiple regression analysisa of log IC50, log (mole fraction of Hg-chloride), and log (mole fraction of Ca2" plus Mg2")
Source

Total

SS 0.557347
0.532518 0.455127 0.077391 0.024829

df

MS
0.266259 0.455127 0.077391 0.008276

5 2 1 1 3 32.17 54.99 9.35 0.0094 0.0051 0.0551

Regressionb log (mole fraction of Hg-chloride) log (mole fraction of Ca2+ + Mg +) Error

a Terms of the multiple regression analysis: SS, sum of squares; df, degrees of freedom; MS, mean square; F, F statistic; P, significance probability associated with the F statistic. b log IC50 = 2.602 - 0.358 log (mole fraction of Hg-chloride) + 5.392 log (mole fraction of Ca2+ + Mg2+); R2 = 0.955**.

VOL. 59, 1993

GROWTH MEDIUM EFFECTS ON Hg(II) BIOASSAYS

1513

chloride concentrations (the only source of which was the organic supplements) ranged from 0.089 mM in the M-IIS medium to 3.03 mM in the M-IIC medium and were well within tolerable limits for the P. fluorescens isolate. That is, experiments with a set of chloride controls (M-IIX plus KCI [1.0 to 10 mM] but without Hg; data not presented) demonstrated that chloride alone had no significant effect on the growth or activity of the bacterial culture. These results suggest that the decrease in IC50 associated with increased chloride levels was not due to the chloride itself but to the interaction of Hg(II) with the chloride, i.e., the formation of Hg(II)-chloro complexes. Chloride-induced toxicity enhancement was the factor most closely linked to the differences in acute Hg(II) toxicity expressed in the M-IIX media. Statistical analysis of the toxicity and speciation data demonstrated that the acute toxicity of Hg(II) was highly correlated to the mole fraction of the Hg(II)-chloro complexes, particularly HgCl+, HgCl2 , and HgCIOH (Table 5). These results are similar to those reported earlier (12, 13). Because the activities and mole fractions of the various Hg(II)-chloro species were highly collinear, it was not possible to quantitatively determine the relative toxicity of individual species. Nevertheless, it was postulated that HgCl+ (because of its cationic nature and because the Hg2' ion is still free to interact directly with other ligands) and HgCl20 (because of its high degree of permeability through lipid bilayer membranes [14]) were the most toxic of the Hg(II)-chloro species. Not only is the toxicity of metals influenced by the organic and inorganic ligands in the bioassay medium, but it is also influenced by the type and concentration of inorganic cations in the medium (5). These secondary, or background, cations can affect Hg(II) speciation by competing for the organic and inorganic ligands in the bioassay medium and may affect the toxicity of Hg(II) directly by competing for sites on cell surfaces. Although there was no significant correlation between the IC50 and the total concentration of any metal in the media, our data indicated that the IC50 was significantly correlated to the activities of free Ca2' and Mg2+ (Table 5). Soluble complexes of calcium and magnesium with the dissolved free amino acids or chloride were generally negligible, accounting for no more than 0.13 mol% of the metals, 0.06 mol% of the amino acids, or 0.02 mol% of the chloride. Consequently, the protective effect of calcium and magnesium was most likely due to the interaction of free Ca2+ and Mg2+ with the bacteria themselves. Rai et al. (19) reported that Ca2+ and Mg2+ reduced the toxicity of Hg(II) to Chlorella vulgaris; however, no mechanism for the protective effect of these metals was suggested. Likewise, Shuttleworth and Unz (23) reported that free Ca2+ and Mg2+ reduced the toxicities of copper, nickel, and zinc to filamentous bacteria (Thiothrzx spp.). In general, inorganic cations are considered to protect bacteria and algae against metal toxicity by outcompeting the free metal ions for sites on the cell surfaces. In the case of mercury, however, this would assume that the toxic effects of the metal are due solely to free Hg2+. Our data do not justify this assumption, and we believe that the ameliorating effects of Ca2+ and Mg2+ were due to the capacity of the metals to (i) outcompete the cationic HgCl+ complex for sites on the cell surfaces and (ii) alter the cell membrane so that it was less permeable to HgCl20. In the latter instance, decreased permeability could reflect either the accumulation of positive charges within the cell membrane, resulting in decreased movement of HgCl20 across the membrane (8), or a change in the molecular architecture of the membrane, which is responsible for the

permeability characteristics of the membrane and is directly influenced by the concentration of free Ca2+ in the surrounding medium (22). Although the activity and mole fraction of free Hg2+ in the M-IIX media increased as the IC50 increased, this merely reflects the dependence of the free Hg2+ concentration on the total concentration of mercury in the media. However, when the total concentration of mercury in the M-IIX media was held constant (5 ,ug ml-'; data not shown), we found that there was no significant correlation between respiratory inhibition and the activity or mole fraction of free Hg2+ in the media. This suggests that the free Hg2+ contributes little to the overall (acute) toxicity of the metal and is in disagreement with the generally accepted view that the free aquo ion is the most toxic form of a metal. This lack of effect of free Hg2+ is presumably related to the extremely low activities (ranging from 5.7 x 10-16 to 4.8 x 10-14 M) of free aquo ion in the media. Furthermore, the low Hg2+/(Ca2+ + Mg2+) ratios (ca. 10-1") suggest that any toxic effects of free Hg2+ may have been mediated by the protective effects of free Ca + and Mg2+. Given that environmental conditions rarely favor the formation of free Hg2+ (9) and that Ca2+ and Mg2+ are the predominant cations in freshwater systems, it appears unlikely that the free Hg(II) ion will play a significant role in regulating the acute toxicity of mercury in these systems. In summary, this investigation was conducted by using experimental and modeling techniques to elucidate the critical parameters that affect the chemical form (species) and toxicity of mercury in a synthetic growth medium (M-II) supplemented with various complex soluble organics. We had previously demonstrated (12, 13) that chloride additions enhanced the toxicity of Hg(II) in M-IIY medium; consequently, chloride salts were not included in the basal salts medium. Although the acute toxicity of Hg(II) to P. fluorescens isolate BPL85-48 was clearly dependent upon the chemical composition of the growth medium, we found that there was virtually no relationship between the affinity of the organic supplements for Hg(II) and the acute toxicity of Hg(II) in the M-IIX media. Instead, we found that the medium supplements contained various amounts of chloride which greatly affected the acute toxicity of the Hg(II). Statistical analysis of the Hg(II) speciation and bioassay data indicated that complexes of mercury with chloride (principally HgC1+, HgCl20, and HgClOH) were the Hg(II) species primarily responsible for the acute toxicity of mercuric nitrate. Our data also indicated that the toxicity of Hg(II) was inversely related to the amounts of free Ca2+ and Mg2+ in the media. These results corroborate our earlier findings (12, 13) that 1:1 and 1:2 Hg-chloro complexes play a significant role in determining the relative toxicity of Hg(II) and demonstrate that these observations hold true for many types of media. Moreover, our results support the hypothesis that the biological effects of mercury are more closely related to the chemical form of the metal than to its total

concentration.

Ultimately, bioassays for evaluating the toxicity of metals must be related to the physicochemical speciation of the metal. However, until analytical techniques for quantifying metal-ligand species become less expensive and more refined, they are unlikely to be included in toxicity testing schemes on a routine basis. In the meantime, computer modeling of metal speciation represents an important tool that can help researchers make better judgments regarding the bioavailability and toxicity of metals in natural and synthetic systems. Although it is difficult to extrapolate the

1514

FARRELL ET AL.

APPL. ENvIRON. MICROBIOL.

results of in vitro studies with a single test organism and synthetic growth media to natural systems, it is not unreasonable to assume that metal-ligand species which are highly correlated to toxicity in artificial systems will also be correlated to toxicity in the natural environment. However, additional research is needed to determine the role of various Hg(II)-ligand species in the ecotoxicology and biogeochemistry of mercury.
ACKNOWLEDGMENT This study was supported by funds from the Natural Sciences and Engineering Research Council of Canada. REFERENCES 1. Allison, J. D., D. S. Brown, and K. J. Novo-Gradac. 1990. MINTEQA2/PRODEFA2, a geochemical assessment model for environmental systems: version 3.0. Environmental Research Laboratory, Office of Research and Development, U.S. Environmental Protection Agency, Athens, Ga. 2. American Public Health Association. 1989. Standard methods for the examination of water and wastewater, 17th ed. American Public Health Association, Washington, D.C. 3. Analytab Products. 1979. Analytical profile index: Enterobacteriaceae and other gram-negative bacteria. API Division of Ayerst Laboratories, Plainview, N.Y. 4. Babich, M., and G. Stotzky. 1983. Influence of chemical speciation on the toxicity of heavy metals to the microbiota, p. 1-46. In J. 0. Nriagu (ed.), Advances in environmental science and technology, vol. 13. Aquatic toxicology. John Wiley & Sons, Inc., New York. 5. Babich, M., and G. Stotzky. 1986. Environmental factors affecting the utility of microbial assays for the toxicity and mutagenicity of chemical pollutants, p. 9-42. In B. J. Dutka and G. Bitton (ed.), Toxicity testing using microorganisms, vol. 2. CRC Press, Inc., Boca Raton, Fla. 6. Baker, M. D., W. E. Innis, C. I. Mayfield, and P. T. S. Wong. 1982. Effects of pH on the growth and activity of heterotrophic sediment microorganisms. Chemosphere 11:973-983. 7. Bird, N. P., J. G. Chambers, R. W. Leech, and D. Cummins. 1985. A note on the use of metal species in microbiological tests involving growth media. J. Appl. Bacteriol. 59:353-355. 8. Boudou, A., D. Georgescauld, and J. P. Desmozes. 1983. Ecotoxicological role of the membrane barriers in transport and bioaccumulation of mercury compounds, p. 117-136. In J. 0. Nriagu (ed.), Advances in environmental science and technology, vol. 13. Aquatic toxicology. John Wiley & Sons, Inc., New York. 9. Campbell, P. G. C., A. G. Lewis, P. M. Chapman, A. A. Crowder, W. K. Fletcher, B. Imber, S. N. Luoma, P. M. Stokes,

10. 11.

12.

13.

14.

15. 16. 17. 18.

19.

20.
21. 22.

23.
24.

25.

and M. Winfrey. 1988. Biologically available metals in sediments. NRCC publication no. 27694. Publications, National Research Council of Canada, Ottawa, Ontario. Carter, J. W., and I. L. Cameron. 1973. Toxicity bioassay of heavy metals in water using Tetrahymena pynformis. Water Res. 7:951-961. Chan, E. C. S. 1964. Morphological aberration of Arthrobacter globiformis cells due to biotin deficiency. J. Bacteriol. 87:641651. Farrell, R. E., J. J. Germida, and P. M. Huang. 1990. Biotoxicity of mercury as influenced by mercury(II) speciation. Appl. Environ. Microbiol. 56:3006-3016. Farrell, R. E., P. M. Huang, J. J. Germida, U. T. Hammer, and W. K. Liaw. 1991. Influence of mercury(II) speciation on the methylation of mercury by Pseudomonas fluorescens. Verh. Int. Verein. Limnol. 24:2190-2198. Gutknecht, J. 1981. Inorganic mercury (Hg2+) transport through lipid bilayer membranes. J. Membr. Biol. 61:61-66. Hughes, M. N., and R. K. Poole. 1989. Metals and microorganisms. Chapman and Hall, London. Hughes, M. N., and R. K. Poole. 1991. Metal speciation and microbial growth-the hard (and soft) facts. J. Gen. Microbiol. 137:725-734. Nuzzi, R. 1972. Toxicity of mercury to phytoplankton. Nature (London) 237:38-40. Parker, D. R., W. A. Norvell, and R. L. Chaney. GEOCHEMPC: a chemical speciation program for IBM and compatible personal computers. In R. H. Loeppert et al. (ed.), Soil chemical equilibrium and reaction models, in press. American Society of Agronomy, Soil Science Society of America, Madison, Wis. Rai, L. C., J. P. Gaur, and H. D. Kumar. 1981. Protective effects of certain environmental factors on the toxicity of zinc, mercury, and methylmercury to Chlorella vulgaris. Environ. Res. 25:250-259. Ramamoorthy, S., and D. J. Kushner. 1975. Binding of mercuric and other heavy metal ions by microbial growth media. Microb. Ecol. 2:162-176. SAS Institute Inc. 1990. SAS procedures guide, version 6, 3rd ed. SAS Institute Inc., Cary, N.C. Schonffeniels, E. 1967. Cellular aspects of membrane permeability. Pergamon Press, New York. Shuttleworth, K. L., and R. F. Unz. 1991. Influence of metals and metal speciation on the growth of filamentous bacteria. Water Res. 25:1177-1186. U.S. Environmental Protection Agency. 1976. Quality criteria for water. U.S. Government Printing Office, Washington, D.C. Waters Chromatography Division. 1986. Pico tag amino acid analysis system. Operator's manual. Manual no. 88140, 3rd revision. Waters Chromatography Division, Milford, Mass.