APPLIED MICROBIOLOGY, Nov. 1974, p. 872-880Copyright 0 1974 American Society for Microbiology

Vol. 28, No.5Printed in U.S.A.

Direct Method for Continuous Determination of Iron Oxidationby Autotrophic Bacteria

MICHAEL STEINER AND NORMAN LAZAROFF

Department of Biological Sciences, State University of New York, Binghamton, New York 13901

Received for publication 24 July 1974

A method for direct, continuous determination of ferric ions produced inautotrophic iron oxidation, which depends upon the measurement of ferric ionabsorbance at 304 nm, is described. The use of initial rates is shown tocompensate for such changes in extinction during oxidation, which are due todependence of the extinction coefficient on the ratio of complexing anions toferric ions. A graphical method and a computer method are given for determina-tion of absolute ferric ion concentration, at any time interval, in reactionmixtures containing Thiobacillus ferrooxidans and ferrous ions at known levels ofSO,2+ and hydrogen ion concentrations. Some examples are discussed of theapplicability of these methods to study of the rates of ferrous ion oxidationrelated to sulfate concentration.

Thiobacillus ferrooxidans is an acidophilicbacterium capable of chemolithotrophic growthusing carbon dioxide as the sole source of carbonwhile obtaining its metabolic energy from theoxidation of ferrous iron or the sulfur atoms ofthiosulfate and elemental sulfur. The oxidationof ferrous iron by T. ferrooxidans is not only ofintrinsic interest due to its biological unique-ness, but is also of economic significance be-cause of the ecological impact of such processesin natural environments.

In nature, the oxidation of iron by T.ferrooxidans in mine drainage waters results inthe formation of sulfuric acid and ferric ironprecipitates. These products account for thesevere pollution problems that occur in surfacewaters which receive discharges from geologicalsubstrata containing reduced iron and sulfurminerals.Various methods -have been employed to

study the oxidation of iron by T. ferrooxidans.Manometric techniques were used by Templeand Colmer (15), Silverman and Lundgren (13),Beck (2), Landesman et al. (7), and others. Thedata of these workers expressed ferrous ironoxidation as a function of oxygen uptake undervarious experimental conditions.An additional approach has been to measure

ferric iron concentration with colorimetricmethods. For instance, Lazaroff (8), and Raz-zell and Trussell (10) calorimetrically measuredthe concentration of ferric-thiocyanate complex.Schnaitman et al. (11) reported a simplifiedcalorimetric assay in which ferric iron was

measured as the yellow chloride complex at 410nm.

All of the colorimetric assays mentioned havethe disadvantage of requiring the introductionof color reagents to the samples which areremoved from the reaction mixtures for analy-sis. This necessitates stopping the reaction priorto the determination of absorbance.

It is essential for physiological studies thatthe absorbance of the colored iron complexesnot be altered indeterminately by the experi-mental variables imposed on the bacteriologicalsystem under study. Such alterations which dooccur must be compensated for by appropriatecontrols to obtain an accurate determination ofoxidized iron. For example, if the effect ofsulfate ions on bacterial iron oxidation is thesubject of inquiry, it would be prerequisite todetermine the effects of sulfate concentration onthe molar extinction of the absorbing ferric ironspecies before the amount of oxidized iron couldbe accurately measured.Hayon and Weiss (5) have used the ultravio-

let absorbance of ferric ions as the basis for themeasurement of ferrous iron oxidation in irradi-ated nonbiological systems. Previously, Whi-teker and Davidson (17), Bastian et al. (1), andHeidt et al. (6) had described the ultravioletphotochemistry of ferrous and ferric ions ininorganic solution. Our investigations showedthat the methods employed by these workerscould be adapted to the study of the biologicaloxidation of ferrous iron by T. ferrooxidans. Incontrast to previously used calorimetric tech-

872

on July 28, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

IRON OXIDATION BY AUTOTROPHIC BACTERIA

niques, this adaptation has permitted the con-tinuous, kinetic determination of ferric ironformation in a reaction mixture without stop-ping the reaction or adding colorimetric rea-gents during the course of reactions.

MATERIALS AND METHODSStrain used. The strain of T. ferrooxidans used in

this study was obtained from D. G. Lundgren (Syra-cuse University, Syracuse, N.Y.). Stock cultures weremaintained aseptically in the 9K medium of Silver-man and Lundgren (12). Working stocks for inoculat-ing mass cultures were prepared from these andmaintained aseptically in 250-ml shaking flasks atroom temperature and were transferred weekly. Thebasal constituents of the 9K media were sterilized byautoclaving at 15 lb/in2 for 15 min. Sterilization of theferrous iron constituent was accomplished by forcingit through Metricel GA-8 filters (0.20-Mm pore size)with a syringe.

Culture and purification of cells. Mass culture ofthe bacteria to obtain large quantities of cells forexperimental purposes did not require aseptic condi-tions since the pH of 2.5 and the inorganic nature ofthe 9K medium inhibited growth of contaminatingorganisms. For mass cultures, 20-liter carboys eachcontaining 10 liters of 9K medium were inoculatedusing the contents of an actively growing 250-mlshaken flask. Mass cultures were grown at roomtemperature and rigorously aerated by bubbling withcompressed air through sintered-glass spargers.One week after inoculation, the contents of the

carboys were filtered through Whatman no. 1 filterpaper to remove precipitated iron compounds. Thecells were separated from the filtrate by continuous-flow centrifugation at 20,000 x g.The sedimented cells were washed by suspension

and sedimentation in dilute sulfuric acid (pH 2.5) at5 C. Repetition of this procedure produced a cellsuspension relatively free of solid contaminants. Thefinal, washed cell suspensions contained approxi-mately 3.0 x 1010 cells/ml, as determined microscopi-cally with a Petroff-Hausser counting chamber.

Determination of iron absorption. A Cary model15, double-beam, recording spectrophotometer wasused for the determination of ferrous and ferric ironabsorption spectra by scanning the ultraviolet-absorbing region of the spectrum between 340 and 220nm. This instrument was also used to record thecontinuous production of ferric ions by measuringchanges in absorbance at 304 nm. Using the samplecuvette as a reaction vessel containing the desiredamounts of bacterial cells and substrate, kinetic datawere obtained by comparing the sample at 304 nm toa reference cuvette which contained the appropriateconcentrations of bacterial cells suspended in wateracidified to pH 2.5 with H2SO4. Absorbance increasesdue to ferric iron production were plotted directly,versus elapsed time, on the strip-chart read-out of theinstrument. Reaction rates were obtained from theslopes of these responses and plotted against theparticular experimental variable employed.When manometric determinations were made, the

standard Warburg technique for oxygen uptake wasused as described by Umbreit et al. (16).

RESULTS

The absorption maxima of ferric and ferrousions are shown in Fig. 1. Since sulfate anionsaffect the rate of bacterial iron oxidation, it wasimportant to note the differences in absorbancedue to the replacement of sulfate by chlorideanions. Ferric ions absorb strongly in the 295- to304-nm region, in contrast to ferrous ions withmajor absorbance in the 220- to 250-nm region.In both cases, chloride anions cause shifts of theabsorption peaks to lower wavelengths andoverall increases in absorptivity. It is importantto note that the extinction due to ferric com-plexes at 304 nm are approximately 300 timesthat of the corresponding ferrous species at thesame wavelengths, a factor which permits thespectrophotometric determination of smallamounts of ferric ions in solutions containinghigh concentrations of ferrous ions.The small absorbance of ferrous iron solutions

initially observed in the 300-nm region wasvirtually eliminated by treatment with stan-nous chloride, indicating that the absorbancewas due to contaminating ferric ions (Fig. 2).

FIG. Ia

043

0.5[

0.3

8 0.1.0 Ih.0

4 0.7

0.5

0.3

0.1

/ WITH/ No Cl

/

297

- 304 285,.

FIG. lb WITH / 220

NaCI A/WT 224

/ 24342/4841243 224

/

_ //302

308 302

340 320 300 280 260 240 220

Wavelength (nm)

FIG. 1. Absorption spectra of Fe23,(SO4)3 andFe2+SO4. (a) 1.2 x 10-5M Fe23+(SO4)3 alone and withequimolar NaCI. (b) 3.23 x 10-3 M Fe2+SO4 aloneand with equimolar NaCI. All solutions adjusted topH2.5 with HS04.

VOL. 28, 1974 873

0°91 -

on July 28, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

STEINER AND LAZAROFF

0.6

OAL

"~02

310 m 33) 2) as

Wavelength( nm)FIG. 2. Absorbance spectra of 3.23 x 10-3 M

FeSO4 solutions treated as follows: (i) oxidized by a

cell suspension of T. ferrooxidans, removed by filtra-tion prior to determination; (ii) oxidized with 1 dropof 0.3% HaOa; (3) oxidized by irradiation with a 100-Wmercury discharge lamp; (4) reduced with 9.7 gmol ofSnCI per cuvette containing 3 ml of sample solution;and (5) untreated. All solutions at pH 2.5.

Conversely, the absorption of ferric solutions atwavelengths shorter than 250 nm was clearlynot due to ferrous iron contamination. This isindicated by the disparity of the ferrous andferric extinction coefficients, and by the stabil-ity of the absorption peaks in the 250- to 200-nmregion after ferric solutions had been treatedwith hydrogen peroxide or by ultraviolet irradi-ation.Because of the overlap of the ferrous and

ferric iron absorption peaks in the region below250 nm, as well as the large effects of complex-ing anions on short-wavelength absorbance, itwas preferable to measure the formation offerric ions in the 295- to 305-nm region. Anyabsorbance in this region is clearly attributableto ferric complexes with minimal interferencefrom other species present in the reaction mix-tures. This conclusion is in agreement with theobservations of Bastian et al. (1), Whiteker andDavidson (17), and Heidt et al. (6).A comparison was made of absorption spectra

when identical ferrous sulfate solutions wereoxidized biotically by a cell suspension of T.ferrooxidans and abiotically with hydrogen per-oxide or by ultraviolet irradiation. These spec-tra indicate that the end products of ferrous ionoxidation are spectrophotometrically similar inall cases (Fig. 2). Therefore, it was feasible tomeasure directly the bacterial oxidation of fer-rous ions by a spectrophotometric method de-

pendent on ultraviolet absorbance specific forferric ions. Any absorbance due to the cellsthemselves could be blanked out by using a cellsuspension of T. ferrooxidans without substratein the reference cuvette.To obtain ferric iron formation in absolute

units the molar extinction coefficients of ferricions at 304 nm were determined. For the pur-pose of our investigations of bacterial ironoxidation, the effect of sulfate anion concentra-tion on the molar extinction of ferric ions had tobe considered.

Ferric ion absorbance was measured at 304nm when ferrous solutions were oxidized withH202 in the presence of various concentrationsof sulfate (Table 1). The oxidized solutions wereallowed to stand overnight, and then werediluted 1:10 with H2S04 (pH 2.5) prior to thedetermination of absorbance and calculation ofthe molar extinction coefficients.The elevation of sulfate concentrations in-

creased the absorbance at 304 nm. It wasnecessary to distinguish between two possiblemeans by which sulfate anions could influenceabsorption: (i) higher sulfate concentrationscould affect the equilibrium of the oxidationreaction, resulting in an increased yield of ferricions; or (ii) the amounts of ferric ions formedwere the same in each case, but increasing levelsof sulfate resulted in sulfate-ferric ion com-plexes with greater specific extinctions.To determine which of the above effects were

operative, ferrous sulfate solutions were oxi-dized by the procedure described above, exceptthat the sulfate levels were adjusted 1 h afterthe solutions had been oxidized. A comparisonof these data shows that both procedures re-sulted in equivalent ferric ion absorbances at

TABLE 1. Molar extinction coefficients (at 304 nm) of3.23 x 10- sM FeSO4 solutions (pH 2.5) oxidized with

1 drop of H2,Oa

Total e= A/Mlsulfatefrom [SO2 -V Optical Na2SO, Na2SO f

NaaSO, [Fe +] density added addedNaSO, ~~~~before after(X 10-aI M) oxidation oxidation

3.23 1 6.10 1,888 1,873 +1516.56 5.1 6.40 1,981 1,935 +4656.56 17.5 6.65 2,059 2,043 +16216.56 67.0 7.60 2,353 2,353 0669.89 207.4 8.30 2,570 2,554 +16

a Various concentrations of sulfate were added as Na2SO4before and after iron oxidation. Total reaction mixturevolumes were 3.0 ml.

IA paired t test value of t = 2.64 indicates that thedifferences between the two procedures are not statisticallysignificant at the 5% level.

874 APPL. MICROBIOL.

on July 28, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

IRON OXIDATION BY AUTOTROPHIC BACTERIA

the same sulfate levels (Table 1). This indicatesthat increasing the sulfate concentration in-creases the molar extinction of the absorbingspecies, most probably by altering the nature ofthe sulfate-ferric ion complex in solution. Ourmolar extinction coefficients are in agreementwith those of Heidt et al. (6), who showed asimilar range of values for oxidized, acidic,ferrous sulfate solutions.

Figure 3 is a graphic presentation of therelationship between the molar extinction coef-ficients of ferric ions and the relative propor-tions of SO2- to Fe3+. This curve can be used tocalculate the absolute quantities of ferric ionsproduced at different sulfate concentrations.For example, a typical substrate for oxidation

consisted of 3.23 x 10-1 M (final concentration)FeSO, adjusted to pH 2.5 with 0.42 x 10-a MH2SO4, yielding a total sulfate concentration of3.65 x 10-3 M. For any concentration of ferricions up to the maximum of 3.23 x 10-3 M, theSO2- to Fe3+ ratio can be calculated and usedto obtain an appropriate molar extinction valuefrom the curve shown in Fig. 3. By plotting Fe3+concentration versus absorbance (where absorb-ance = A = eM, a curve is obtained which canbe used to convert absorbance values to Fe3+concentration. Such a curve is shown in Fig. 4.It may be noted that the relationship betweenthe absorbance and molarity of Fe3+ is nearlylinear at the lowest values. With increasingFe3+, the line curves slightly and then becomeslinear again. This reflects the comparativelysmall change in extinction at high [SO2- V[Fe3+ ] ratios and the diminishing change in the[SO2-1 [Fe3+] ratio as it approaches unity athigher Fe3+ molarity.

FIG. 3. Effect of sulfate concentration on molarextinction coefficient of ferric ions (3.23 x 10'- MFe2+ [as FeSO4] oxidized with I drop of30% H202 anddiluted 1 to 10 in H2S0, [pH 2.5] for absorbancedeterminations). Symbols: (@), sulfate added beforeoxidation; (x), sulfate added after oxidation. Allsolutions were at pH 2.5.

FIG. 4. The calculated relationship of absorbanceat 304 nm to increasing ferric ion concentration at fivelevels of sulfate ion concentration: (1) 1.0 x 10-1 MSO,2-; (2) 1.0 x 10-2 M S042-; (3) 3.65 x 10-3 MS042-; (4) 1.0 x 10-3 M S042-; (5) 1.0 x 10-4 MS042- (used experimentally).

Figure 5 shows this even more strikingly byrelating the calculated changes in the extinctioncoefficient, the [SO,2- ]/[Fe3+ ] ratio, absorb-ance, and micromoles of Fe3+ formed. Since theextinction coefficient is maximal at the begin-ning of the reaction (due to the high ratio ofsulfate to ferric ions), the use of initial rates forkinetics determinations appears amply justi-fied. In addition, actual observations of absorb-ance changes measured by continuous spectro-photometric analysis are nearly linear fromtime zero during the first few minutes of oxida-tion, depending on the oxidation rate.The practical limits for [SO2- V [Fe3+] ratios

which may be treated in this method of analysisare fixed at an upper value by the solubilityproduct for ferric sulfate in acid solution and at alower value by a ratio which is slightly greaterthan the equivalence of sulfate to ferric ions.The latter situation would be represented by aferrous sulfate solution containing a smallamount of additional sulfate originating frompH adjustment to 2.5 with H2SO4. Lower ratiosthan 1 cannot be treated with our extinctioncoefficients, since anions other than sulfatewould change the relationship shown in Fig. 3.(Although the extinction values in the presenceof other anions can be determined, it should berecognized that the extinction coefficient wouldchange very little unless competing anions werepresent in concentrations approaching the sul-fate ion.)

Figure 4 shows the relationship between ab-sorbance and Fe3+ concentrations in solutionscontaining four different levels of sulfate. Thevalue for [Fe3+] may be approximated for ab-sorbance measurements at any sulfate level by

VOL. 28, 1974 875

on July 28, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

STEINER AND LAZAROFF

minutes

FIG. 5. The calculated relationship of Fe3+ forma-tion, [SO42-1 to [Fe2+] ratio, e3O4, and absorbance totime, in a system containing 3.23 x 10-' MFeSO4 and0.42 x 10-3 M H2S04 (3.65 x 10- M [SO42- ]), with1.0 x 10-2 umol of Fe3+ formed/min. Key to lines:

), Amol x 10- 2 Fe3+ formed; (-----),absorbance at 304 nm; (- - -), molar extinction coeffi-cient (fS34); ( --), [SO42-Y[Fe2+] ratio.

interpolation between the curves. However, amore precise value can be obtained by thegraphical method described previously.

In addition, it is possible to obtain an equa-tion which approximately expresses the curve inFig. 3, relating the molar extinction coefficient(e) of ferric ions to the [SO,2- V [Fe'+ ] ratio. Thecorrespondence of the experimental relationshipto the least-squares curve (Fig. 6) is expressedby

Y = eCO + ClnX + C2 (lnX)2 + C (lnX)3

where Y = , the molar extinction coefficient; X= S/F, S = [S042- ], F = [Fe'+]; C0 = 7.797, C1= 0.243988, C2 = 0.08920, and C3 = 0.00782431.Since Y = A/F 1, where A = absorbance, F =molar concentration of ferric ions, and l = pathlength of cuvette in centimeters, then:A304= Fl x eCu + C, (In S/F) + C2 (in S/F)2 + C. (n S/F)3Thus, the absorbance can be calculated di-

rectly for any ferric ion concentration if thesulfate concentration is known. However, amore important application of this equation isto use a computer to evaluate F ( [Fe3+ ]) for anygiven absorbance and sulfate level. This elimi-nates the cumbersome operations involved inthe construction of curves and the application ofthe graphic method for determining [Fe3+ ].Under conditions employed for spectrophoto-

metric determination, the rates of bacterial ironoxidation were proportional to the concentra-tion of bacterial cells present in the reactioncuvette (Fig. 7). This indicates that neither theconcentration of ferrous iron substrate nor oxy-gen availability were rate-limiting.

Figure 8 shows a Lineweaver-Burk treatmentof data from an experiment in which the bacte-rial cell concentrations were constant but theferrous sulfate concentrations were varied. TheKm and maximal velocity were calculated to be5.68 x 10- I mol/liter and 2.08 x 10-I2 gtmol/min,respectively.The spectrophotometric method was applied

in experiments to determine the effects ofsulfate ion concentration on bacterial iron oxi-dation. By increasing the [SO42- [Fe2+] ratiofrom 1 to 14, stimulation of ferric iron formationoccurred (Fig. 9).A comparison was made of these data with

manometric determinations in which the con-centrations of bacterial cells and ferrous ironsubstrate were constant and the concentrationof sulfate was varied (Fig. 9). The manometricdata indicated that a [SO42- V[Fe2+ ] ratio of 1.5produced the greatest rate of oxygen uptake,but the further addition of sulfate sharplydecreased this activity.

This pattern of response differed from thespectrophotometric determination of Fe3+ for-mation, prompting additional determinationsof ferric iron formation at higher sulfate levels(Table 2). It can be seen that the rate of ferrous

280O F

24001

£ 23001.

200

+

1*

4-4-

0 100[

ISCQ1F-t

30D 40D

FIG. 6. A comparison of the relationship between[S042- V[Fe3+1] to molar extinction (ESO4) and theleast-squares function

y = e C0 + C, inX + C. (inX)2 + C, (InX)3

where: y = (304; X = [SO42-V[Fe3+]; Co = 7.797; C, =

0.243988; C2 = 0.08920; and C3 = 0.00782431. Sym-bols: (+) experimental data and (*) least-squaresfunction.

876 APPL. MICROBIOL.

on July 28, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

IRON OXIDATION BY AUTOTROPHIC BACTERIA

4.0

3.5

, 3.0 -

C:.° 2.5 -o/0E

1.5 -

E CD 1.0 -/

a) .54-,0

I:

1. 2.0 3D 4.0 5.0 DConcentration of bacterialcells (xlO 8 /ml )

FIG. 7. Dependence of rate of bacterial iron oxida-tion (measured as absorbance change at 304 nm) onconcentration of bacterial cells. Substrate solutionscontained 3.23 x 10s M FeSO4 adjusted to pH 2.5with HS04. The rate of ferric ion formation isexpressed as micromoles x 10- 'per 3 ml of reactionmixture (in cuvette) per min. Bacterial concentrationis given as cells x 108 per ml of reaction mixture.

iron oxidation does decrease at elevated sulfatelevels. However, this inhibition occurs at muchhigher sulfate concentrations than those whichinhibit oxygen uptake in manometric determi-nations.

It was important to compare the spectropho-tometric kinetics of bacterial iron oxidationwith similar data describing the nonbiologicaloxidation of ferrous iron. Solutions containingidentical concentrations of ferrous iron, butwith varying sulfate levels, were placed inquartz cuvettes and irradiated with the fullspectrum of a 100-W mercury discharge lamp.At timed intervals, the cuvettes were placed inthe spectrophotometer and differentiallyscanned against distilled water. It was apparentthat much higher [S0'2- [Fe2+ ] ratios arerequired for maximal rates of photochemicaloxidation than for maximal rates of oxidationby T. ferrooxidans (Fig. 10).

DISCUSSIONThe ultraviolet absorbance which results

when T. ferrooxidans oxidizes ferrous ions coin-cides with that obtained from the nonbiologicalformation of ferric ions under comparable con-ditions of pH and anionic environment (1, 6,17). Since our values obtained for the molarextinction coefficient of ferric ions at 304 nm are

in agreement with the generally accepted valuesof others (6), we believe that these values maybe used for converting spectrophotometric ab-sorbance at 304 nm to absolute units of ferricion formation during autotrophic iron oxida-tion.The derived data indicate that the system

exhibits consistent Michaelis-Menten kineticseven though the Km of 5.68 x 10-4 M differsfrom that of Schnaitman et al. (11), who re-ported a value of 5.4 x 10-' M. This 10-folddiscrepancy may reflect several differences inthe systems used for studying the rates ofautotrophic, ferric-ion formation.Probably most important is the temperature

at which rate determinations were made. Ourdeterminations were made at 22 C, in deferenceto the conditions under which T. ferrooxidansseems to grow best in nature and in the labora-tory, whereas those of Schnaitman et al. (11)were conducted at 35 C. This 130 temperaturedifference would be expected to greatly alter theoxidation rate and to account in large measurefor the observed differences in the maximalvelocity and Km.Din and Suzuki (4) working with cell-free

extracts, in which the reduction of cytochrome cby Fe'+ was measured spectrophotometrically,obtained a Km of 5.9 x 10-4 M for Fe'+, a valuewhich closely approximates our result. Theirdeterminations were carried out at approxi-mately 25 C, although at a pH of 7.0. If we notethat Schnaitman et al. found that the Km ofautotrophic oxidation was virtually independ-ent of pH, then it may be concluded that ourmeasurements agree precisely with the kinetics

1/[ FeSO4] x10-3M

FIG. 8. Lineweaver-Burk plot showing relationshipof rate of bacterial ferric iron formation at 304 nm toinitial ferrous ion concentration. Temperature ofdeterminations was 22 C. All cuveates contained 2.55x 108 bacterial cells; all substrate solutions wereadjusted to pH 2.5 with HS04.

VOL. 28, 1974 877

on July 28, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

STEINER AND LAZAROFF

total [SO;]/[Fer

FIG. 9. Effect of increasing sulfate concentrationon the rate of oxygen uptake and ferric ion formationby Thiobacillus ferrooxidans. Ferric ion formation(-v-) determined spectrophotometrically at 304nm. Initial concentration of Fe2+ (as FeSO4) was 3.23X 10'- M. 02 uptake (-x-) was determined mano-metrically; initial concentration of Fe'+ (as FeSOJwas 4.44 x 10-2 M. For both determinations, the pH-was adjusted to 2.5 with H2SO4, and additionalsulfate was added as Na2SO4.

of Fe-2+cytochrome c reduction of Din andSuzuki.

It is pertinent to compare the spectrophoto-metric methods used previously for studying theautotrophic oxidation of ferrous ions. The thi-ocyanate method, whereby the absorbance of a

ferric thiocyanate complex is measured, resultsin erroneous measurements due to instability ofthe complex in the presence of various anions.Stopping the reaction for sampling is a problemsince the addition of hydrochloric acid or mostother mineral reagents would decrease extinc-tion due to the thiocyanate complex. The needfor intermittent sampling or replication of reac-

tion mixtures is cumbersome and prone tosampling error. This also applies to the chlo-ride-complex method of Schnaitman et al.Neither of these approaches allows measure-

ment of initial rates. There is a large initialinflection of rate which could be caused byproblems of mixing (11) or of nonconformanceto the Beer-Lambert Law at low Fe3+ concen-tration. The latter possibility seems plausiblewhen the data of Schnaitman et al. are exam-ined, and it is noted that, in the course of 10min, the amount of ferric ions increases almost300-fold. Moreover, the ratio of complexingchloride ions to ferric ions or the presence ofother anions would affect the specific molarextinction at the wavelength employed (410

nm). This can be concluded from the determi-nations of Rabinowitch and Stockmayer (9) ofthe extinction values of ferric ions between 400and 460 nm at several HCl concentrations. The410-nm extinction value of ferric ions at the HClconcentrations used by Schnaitman et al. wouldshow considerable fluctuation with compara-tively small changes in acidity and anionicconcentration.Our data at 304 nm allow linear extrapolation

to zero time and can be reliably converted toabsolute amounts of ferric ions formed at anypoint during the time course. The effects ofvariations in the anionic environment on ferricabsorbance at 304 nm are comparatively small,but can be compensated for when absolutemeasurements are required by using an appro-priate calibration of extinction coefficient ver-sus anionic molarity.

It might seem that continuous measurementof ferric ion formation in a cuvette withoutprovision of aeration would become quicklylimited by the rate of oxygen diffusion. How-ever, our determinations are proportional to cellconcentration and are linear with time, showingthat the rates are independent of oxygen con-centration. This finding allowed us to dispensewith the aeration used by Schnaitman et al. andto measure the oxidation as it occurred.The significance of this lack of requirement

for aeration is not completely understood, al-though our data have shown that oxygen utili-zation and ferric ion formation in autotrophiciron oxidation are not tightly coupled (Fig. 9).The stimulatory effects of sulfate upon bacte-

rial iron oxidation are in agreement with thoseof Lazaroff (8), Silverman (14), Schnaitman etal. (11), and others. In this respect, our resultsclosely compare with those of Schnaitman et al.(11). These workers had shown that there is apoint beyond which the addition of sulfate failsto produce continued stimulation of iron oxida-tion.

It is further indicated that sulfate concentra-tions greater than those previously employed

TABLE 2. Relative effect of high sulfate concentrationon the oxidation of 3.23 x 10-'MFeSO4 solutions by

identical bacterial suspensionsa

Total sulfate fromFeSO4 and Na,S04 [80S [foredpeJi

(x 10' M) oFermm

56.56 17.5 0.267216.56 67.0 0.210669.89 207.4 0.088

aAll reaction mixtures were at pH 2.5.

878 APPL. MICROBIOL.

on July 28, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

IRON OXIDATION BY AUTOTROPHIC BACTERIA

FIG. 10. The relationship of LSO42-V[Fe2+] toamounts of ferric ions, formed from FeSO4 solutionafter 5 min (-x-) and 10 min of ultraviolet irradia-tion (-0-). The Fe2+ substrate was 3.23 x 10-3M(as FeSOJ. Reaction mixtures were adjusted to pH2.5.

begin to inhibit iron oxidation in a patternsimilar to that observed using manometric de-terminations of oxygen uptake. However, our

measurements and those of Schnaitman et al.(11) show that the levels of sulfate (relative toferrous iron substrate) required for maximalstimulation of bacterial iron oxidation are muchgreater when Fe'+ formation is determinedspectrophotometrically, than when oxygen con-

sumption is determined manometrically. Thisagain suggests that biological oxidation of fer-rous iron by the autotrophic system is notalways paralleled by oxygen reduction.The effects of sulfate on ferrous iron oxidation

in biological and nonbiological systems aresimilar but not identical. The stimulatory effectof sulfate occurs in both the biological andnonbiological systems. However, bacterial ironoxidation was inhibited at sulfate concentra-tions above 0.216 M. Similar solutions oxidizedby ultraviolet irradiation showed a decelerationof the rate of oxidation rather than a netinhibitory effect.The information obtained so far is consistent

with the interpretation that the important roleof sulfate in bacterial iron oxidation is a conse-quence of its inorganic complexing of iron.Increasing concentrations of sulfate result indifferent complexes with either the ferrous ironsubstrate or the ferric iron product. This couldfavor the oxidation reaction thermodynamicallyand therefore accelerate the observed rates inboth biological and nonbiological systems.However, this does not preclude the participa-tion of sulfate in some additional role in bacte-rial iron oxidation as suggested by Lazaroff (8)

or Dugan and Lundgren (5).In conclusion, we think that the method

developed for the direct spectrophotometricdetermination of autotrophic iron oxidationpossesses a number of advantages over previouscolorimetric techniques. (i) The analysis offerric ion production can be carried out contin-uously as a monitoring technique which doesnot require that the reaction be stopped foranalysis or that color reagents be added. Thisgreatly facilitates -laboratory studies, allowingrapid analysis of oxidation kinetics with max-imum precision due to elimination of manipula-tive and sampling error. (ii) The extinctioncoefficient for ferric ions is consistent for initialrate determinations in all experimental proto-cols employed. In such systems, therefore, ini-tial increases in absorbance will be strictly dueto ferric ion formation. On the other hand, ifcomparisons of ferric ion concentrations inabsolute terms are desired for systems whichdiffer greatly in anionic composition, tempera-ture, or pH, it is possible to calibrate theextinction coefficient accordingly. (iii) The highmolar extinction coefficient of ferric ions at 304nm makes this technique extremely sensitive,as compared with previous colorimetricmethods. (iv) A mathematical computer solu-tion has been developed that can be used toconvert absorbance values to absolute [Fe'+I insystems of known sulfate concentrations.

ACKNOWLEDGMENTSWe gratefully acknowledge the services of Anne Kellerman

(Computer Center, State University of New York, Bingham-ton, N.Y.), who determined the least-squares function fittingthe relationship of extinction coefficient to the ratio of[SO42-]to [Fe'+].

LITERATURE CITED

1. Bastian, R., R Weberling, and F. Palilla. 1953. Spectro-photometric determination of iron as ferric sulfatecomplex. Anal. Chem. 25:284-288.

2. Beck, J. V. 1960. A ferrous-ion-oxidizing bacterium. I.Isolation and some general physiological characteris-tics. J. Bacteriol. 79:502-509.

3. Beck, J. V., and F. M. Shafia. 1964. Effect of phosphateion and 2.4-dinitrophenol on the activity of intact cellsof Thiobacilkusferrooxidans. J. Bacteriol. 88:850-857.

4. Din, G. A., and I. Suzuki. 1967. Mechanism of Fe'+-cyto-chrome c reductase of Ferrobacillus ferrooxidans. Can.J. Biochem. 45:1542-1556.

5. Hayon, E., and J. Weiss. 1960. Photochemical decompo-sition of water by ferrous ions. J. Chem. Soc.1960(3):3866-3872.

6. Heidt, L. J., M. G. Mullin, W. B. Martin, Jr., and A. M.J. Beatty. 1962. Gross and net quantum yields at 2537 Afor ferrous to ferric in aqueous sulfuric acid and theaccompanying reduction of water to gaseous hydrogen.J. Phys. Chem. 66:336-341.

7. Landesman, J. D., D. W. Duncan, and C. C. Walden.

879VOL. 28, 1974

on July 28, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

880 STEINER AND LAZAROFF

1966. Iron oxidation by washed cell suspensions of thechemoautotroph Thiobacillus ferrooxidans. Can. J. Mi-crobiol. 12:25-33.

8. Lazaroff, N. 1963. Sulfate requirement for iron oxidationby Thiobacillus ferrooxidans. J. Bacteriol. 85:78-83.

9. Rabinowitch, E., and W. H. Stockmayer. 1942. Associa-tion of ferric ions with chloride, bromide and hydroxylions (a spectroscopic study). J. Amer. Chem. Soc.64:335-347.

10. Razzell, W. E., and P. C. Trussell. 1963. Isolation andproperties of an iron-oxidizing Thiobacillus. J. Bacte-riol. 85:595-603.

11. Schnaitman, C. A., M. S. Korczynski, and D. G. Lund-gren. 1969. Kinetic studies of iron oxidation by wholecells of Ferrobacillu ferrooxidans. J. Bacteriol. 99:552-557.

12. Silverman, M. P., and D. G. Lundgren. 1959a. Studies onthe chemoautotrophic iron bacterium Ferrobacillus

APPL. MICROBIOL.

ferrooxidans. I. An improved medium and harvestingprocedure for securing high cell yields. J. Bacteriol.77:642-647.

13. Silverman, M. P., and D. G. Lundgren. 1959b. Studies onthe chemoautotrophic iron bacterium Ferrobacillusferrooxidans. H. Manometric studies. J. Bacteriol.78:326-331.

14. Silverman, M. P. 1967. Mechanism of bacterial pyriteoxidation. J. Bacteriol. 94:1046-1051.

15. Temple, K. L., and A. R. Colmer. 1951. The autotrophicoxidation of iron by a new bacterium, Thiobacillusferrooxidans. J. Bacteriol. 62:605-611.

16. Umbreit, W. W., R. H. Burris, and J. F. Stauffer. 1957.Manometric techniques. Burgess Publishing Co., Min-neapolis.

17. Whiteker, R. A., and N. Davidson. 1953. Ion-exchangeand spectrophotometric investigation of iron (III) sul-fate complex ions. J. Amer. Chem. Soc. 75:3081-3085.

on July 28, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from