CLIN.CHEM.22/12, 1986-1990 (1976)

1986 CLINICAL CHEMISTRY, Vol. 22, No. 12, 1976

A Mechanism for the Conversion of Oxyhemoglobin

to Methemoglobin by Nitrite

F. Lee Rodkey

Each mole of oxyhemoglobin iron converted to methe-moglobin causes the oxidation of 1.5 mol of nitrite to nitrateand consumes 1 mol of protons. No oxygen is liberated.The overall reaction has two simultaneously occurringparts. In the beginning the rate-limiting reaction convertingO2Hb to metHb is directly proportional to H and N02concentrations and is independent of metHb. The secondportion accounts in major part for the stoichiometry andrate of the overall reaction. In this portion O2Hb tetramersand metHbNO2 are the reactants. Essentially no reactiontakes place in the presence of CN, which displaces nitritefrom the metHbNO2, nor in the presence of 0.5 mol/literNat, which converts the O2Hb to af3-dimers. The auto-catalytic nature of the overall reaction in the presence ofexcess nitrite is the result of metHb, which is formed inboth parts of the reaction, associating with nitrite to in-crease the concentration of one reactant of the cyanide-sensitive part. The reaction rates at constant pH in excessnitrite are proportional to the product of the O2Hb con-centration and the square of the metHb concentration. Therate increases up to about 66% conversion of O2Hb fol-lowed by a decrease as the O2Hb becomes limiting. Thedissociation constant of metHbNO2 at 25 #{176}Cand pH =

6.4 was found to be 1.11 ± 0.11 mmol/liter.

The reactions of nitrite ion in the formation ofmethemoglobin from oxyhemoglobin are not well un-derstood. At low concentrations of nitrite the reactionis very slow, but after a period of time the rate of reac-tion increases in an “autocatalytic” fashion (1-3). Incontrast to oxidation with ferricyanide, essentially nooxygen gas is released from oxyhemoglobin by nitriteeven when conversion to methemoglobin is complete(3). Oxygen is essential for the formation of methemo-globin by nitrite as carboxyhemoglobin is unchangedand the reaction proceeds very slowly, if at all, in ni-trogen (2)’. It was observed by chance that oxyhemo-globin preparations treated with sodium nitrite showedan “induction period” that was inversely proportionalto the amount of methemoglobin in the original prep-aration. This report shows that the methemoglobinnitrite complex is required for the rapid conversion ofoxyhemoglobin to methemoglobin by nitrite.

The Laboratory of Analytical Biochemistry, Naval Medical Re-search Institute, National Naval Medical Center, Bethesda, Md.20014.

The opinions or assertions contained herein are those of the authorand are not to be construed as official or as reflecting the views of theNavy Department or the Naval Service at large.

Received Aug. 18, 1976; accepted Sept. 1, 1976.

Materials and MethodsHuman hemoglobin was prepared from the freshly

drawn heparinized or disodium EDTA-treated blood.Erythrocytes were washed three times in 20 volumes ofNaC1 solution (9 g/liter). The washed cells were hemo-lyzed with 10 volumes of de-ionized water at 0 #{176}C.Celldebris was removed by centrifugation (Model PR-2,Internatioal Equipment Co., Needham Heights, Mass.02194). The clear hemoglobin solution was kept on iceand diluted with phosphate or tris(hydroxymethyl)-aminomethane buffers as desired. Methemoglobin wasdetermined by the method of Rodkey and O’Neal (4).Total hemoglobin, as cyanmethemoglobin, was mea-sured by the method of Van Kampen and Zijlstra (5).Absorption measurements were obtained with a Beck-man DB-GT spectrophotometer (Beckman Instru-ments, Inc., Fullerton, Calif. 92634) with the cuvettecompartment maintained at 28 #{176}C.Nitrite concentra-tion was measured by the method of Schneider andYeary (6) by use of the neutral ZnSO4-Ba(OH)2 pro-tein-free filtrates as suggested by Wegner (7).

I estimated the conversion of O2Hb to metHb bycontinuous spectrophotometric measurement at 540nm, recording with a strip-chart recorder. The bufferedO2Hb solution, 3.0 ml, was placed in the cuvette and anoriginal absorbance obtained. The reaction was startedby adding 0.05 or 0.10 ml of freshly prepared aqueousNaNO2. The contents of the cuvette were mixed witha Teflon paddle and spectrophotometric recording wasstarted. When the reaction appeared complete, 2-3 mgof K3Fe(CN)6 was added to ensure that no O2Hb re-mained and the absorbance, A , was recorded. Finally,all metHb was converted to the cyanide derivative byadding 2-5 mg of KCN. The absorbance of the cyan-methemoglobin so obtained, ACN, was used to estimatethe total heme concentration and to calculate the the-oretical absorbance of the solution when only O2Hb waspresent, i.e., 1.3 X ACN. The percentage of the totalhemoglobin present as metHb was then calculated fromthe equation:

%metHb = l.3ACN - A 1001.3ACN - A

(1)

where A is the measured absorbance, 1.3 ACN is theabsorbance expected when all the heme is present asO2Hb, and A,. is the absorbance observed when all theheme is present as the metHb or metHbNO2 mix-ture.

I 40

8 0 2 14 14 8Tim - mm

Fig. 1. Effect of pH on the conversion of oxyhemoglobin tomethemoglobin by nitriteEach solution contained the same amount of hemoglobin, 50 tmol of heme perliter, in 90 mmol/liter phosphate buffer at the pH indicated. The reaction wasstarted by addingNaNO2to a concentration of 935 imol/liter

Results and Discussion

CLINICALCHEMISTRY,Vol. 22, No. 12, 1976 1987

Effect of pH: Conversion of oxyhemoglobin tomethemoglobin was tested as a function of pH in thepresence of ninefold excess of sodium nitrite over hemeiron. The oxyhemoglobin solution was diluted to ap-proximately 50 mol of heme per liter of 0.2 mol/literphosphate or tris(hydroxymethyl)aminomethanebuffers. Potassium was the counter ion for all phosphatebuffers and chloride was the counter ion in all tris(hy-droxymethyl)aminomethane buffers used. The for-mation of methemoglobin was complete in less than 1mm at pH 6.4 and about 13 mm at pH 7.4. At higher pHthe reaction was extremely slow. Only 3% of the O2Hbreacted in 15 mm at pH 8J while less than 2% of theO2Hb was converted to metHb at pH 9.0 in 30 mm. Itwas clear that the induction period for a given oxy-hemoglobin solution was considerably prolonged athigher pH. The freshly prepared hemoglobin solutionused in the above experiment contained less than 1% ofthe pigment as metHb. Air oxidation of the hemolysateat room temperature was permitted until the methe-moglobin increased to about 7% of the total pigment.Addition of the same ninefold excess of nitrite to the93% OHb/7% metHb mixture at pH 8.1 virtuallyeliminated the induction period and all O2Hb wasconverted to metHb in less than 4 mm.

The 93% O2Hb/7% metHb mixture was saturatedwith 100% CO to remove all O2Hb. Addition of nitriteto the COHb/metHb mixture at pH 8.1 did not causeany detectable formation of metHb in 30 mm, demon-strating that O2Hb is required for the reaction. A fewcrystals of KCN were added to the O2Hb/metHb mix-ture at pH 8.1, to convert the methemoglobin to cyan-methemoglobin. Nitrite was then added to this O2Hb/cyanmetHb mixture. No measurable conversion ofO2Hb to methemoglobin occurred in 2 h. On standingovernight at room temperature less than 5% of the O2Hbwas converted to metHb. Cyanide was effective inslowing methemoglobin formation in this mixture evenat pH 6.4, where the conversion was complete in less

than 30 s in the absence of cyanide.These experiments suggested that free metHb or a

metHb-nitrite complex was required as one reactant inthe conversion of O2Hb to metHb by nitrite. Both CNand high pH inhibited the oxidation of oxyhemoglo-bin.

I measured the effect of pH on the conversion ofO2Hb to metHb by nitrite in a more quantitative fash-ion in phosphate buffers containing 0.1 mol of phos-

phate per liter, with use of a nitrite/heme ratio of 17/1.The results in Figure 1 show that the rate of the reactionis slower at higher pH and the “induction period” isprolonged. Curves similar to those shown in Figure 1may be obtained at lower pH with less NO2- and athigher pH with greater NO2- concentration. In all cases

there is a continually increasing rate up to about 66%conversion and then a decreasing rate as the reactionapproached completion. It will be demonstrated laterthat the overall reaction as shown in Figure 1 is a result

of two reactions occurring together, one of which is in-sensitive to the presence of CN- and the other so sen-sitive as to be virtually stopped by cyanide.

Absorption curve of methemoglobin derivatives.Methemoglobin was prepared by adding sodium nitriteto oxyhemoglobin in 50-fold excess over heme iron. Themixture was allowed to stand at 25 #{176}Cfor 2 h, then di-alyzed exhaustively with a Bio-Rad hollow fiber (Bio-Rad Laboratories, Richmond, Calif. 94804) againstde-ionized water until no nitrite remained. The dialyzedmethemoglobin solution was centrifuged and the su-

pernate stored at 4 #{176}Cuntil used. The absorption curveof methemoglobin was measured in 20 mmol/liter po-tassium phosphate and tris(hydroxymethyl)amino-methane chloride buffers. Curves of identically pre-pared methemoglobin solutions containing sodium ni-trite at 85-fold the molar heme iron concentration werealso obtained.

Results shown in Figure 2 indicate that aquamethe-moglobin (the form present at low pH) has absorptionmaxima near 630 and 500 nm, with shoulders at 575 and540 nm. At alkaline pH, the methemoglobin hydroxidehas maxima only at 575 and 540 nm. The two isosbesticpoints of aquamethoglobin and methemoglobin hy-droxide at 615 and 522 nm observed by Austin andDrabkin (1) are apparent, with a third isosbestic pointnear 483 nm. Methemoglobin nitrite, the major formpresent at pH 6 in the presence of nitrite, has an ab-sorption peak near 625 nm and one nearly in commonwith one for methemoglobin hydroxide at about 538 nm.Methemoglobin nitrite absorbs considerably morestrongly at 575 nm than does aquamethemoglobin, but,unlike methemoglobin hydroxide, has no distinct ab-sorption maximum at this wavelength. The similarityof absorption spectra of methemoglobin nitrite andmethemoglobin hydroxide, together with the competi-tion of nitrite and hydroxide for the methemoglobin, ledVan Assendelft and Zijlstra (8) to conclude that theabsorption spectra of methemoglobin nitrite was in-dependent of pH.

Dissociation constant of methemoglobin nitrite. Thedissociation constant of methemoglobin hydroxide at20 #{176}Cis 1.13 X 10 mmol/liter (1, 9). Smith (10) esti-

‘I

I4)

4)

0

V

I

-I

0

0

S

2..

0- NoN02

-- K3Fe(CN)6

- Lmm4) -

,

W4) L094) -

6.0 70 7.5

1988 CLINICAL CHEMISTRY, Vol. 22, No. 12, 1976

Fig.2. Absorption curves for methemoglobin in the absence (left)and in the presence (right) of NaNO2Each solution contained heme, 73 zmol/llter as metHb, in 0.1 mel/liter phos-phate, pH 7.1, or tris(hyoxymethyI)aminomethane, pH 8.1 and 9.0. SodIumnitrite, where used, was dissolved in the buffers to 6.21 mmol/Iiter, 85-fold theheme concentration

mated the methemoglobin nitrite dissociation constantspectrophotometrically at 25 #{176}Cand pH 7.4. A disso-ciation constant of 3 mmol/liter was reported, althoughother data caused him to believe the value to be toolarge. A number of estimates of the methemoglobinnitrite dissociation constant were made in dilute po-tassium phosphate buffers (20 mmol/liter) where the

methemoglobin is expected to be predominately in thetetrameric state. Guidotti (12) has estimated the tet-ramer-dimer dissociation constants for cyanmethem-oglobin to be similar to those for O2Hb. The presentmeasurements were based on the decrease of absorb-ance at 630 nm when nitrite was added to aquamethe-moglobmn in known amounts. A nitrite concentration

800-fold that of the heme iron was used to establish the

absorbance of methemoglobin nitrite. Essentially nomethemoglobin hydroxide is present at this pH. Sevenmeasurements were made in which the methemoglobinnitrite varied from 38 to 84% of the total methemoglo-bin. The mean value observed was K = 1.11 ± 0.10 (std.error of the mean) mmol/liter. The reason for the dif-ference between this value and the value reported bySmith (10) is not apparent unless it is related to thewavelength and pH at which the measurements weremade. At 560 nm as used by Smith (10), the absorbancy

of methemoglobin nitrite is nearly identical to that ofmethemoglobin hydroxide, as shown in Figure 2, andboth complexes will be present in significant amounts

at pH 7.4.Stoichiometry of the reaction. A number of attempts

were made to determine the minimum amount of nitriterequired to convert O2Hb to metHb by limiting theamount of nitrite. The reaction was done in 20 mmol/liter phosphate buffer at pH 6.4. Although O2Hb wascompletely converted to metHb when the molar ratioof N02 to O2Hb (heme iron) was 1.5 or greater, thereaction was so slow that control air oxidation of thehemoglobin compromised the results. Solutions of O2Hb

pH

Fig. 3. Proton change when oxyhemoglobin is treated with NaNO2

or K3Fe(CN)6 at fourfold the molar heme concentration

in 20 mmol/liter phosphate buffer at pH 6.4 were thenused in which the initial ratio of NO2- to O2Hb variedfrom 2 to 5. When the reaction was complete, as judgedby the decrease of absorbance at 540 nm, the remainingnitrite was determined. Eight such experiments weredone, which showed that the moles of nitrite disap-

pearing correspond to 1.68 ± 0.15 (std. error of themean) times the moles of heme iron oxidized. Clearly,about three nitrite molecules disappear for each twomolecules of O2Hb iron converted to metHb.

Qualitative data had shown that the conversion ofO2Hb to metHb by nitrite in unbuffered solution re-sulted in a more alkaline pH. The extent of proton ab-sorption in this reaction was measured with a Radi-ometer pH Stat (The London Co., Westlake, Ohio44145) when O2Hb at various pH values was treatedwith nitrite. Similar measurements were made whenK;3Fe(CN)6 was used as oxidant. The results in Figure3 show that, within experimental error, the reaction withnitrite consumes one proton for each iron atom oxidized

at all pH values tested. In contrast, the oxidation withK3Fe(CN)6 actually produces protons if the pH exceeds7.4 but is quite similar to the nitrite reaction at pH 6.3.It is probable that K3Fe(CN)6 may produce protons byoxidation of thiol groups in addition to iron atoms (10,ii), a reaction that does not occur with nitrite.

All data available are consistent with the stoichiom-etry of the overall reaction according to the followingequation:

2O9Fe2 + 3N02 + 2H = 2Fe + 3NO + H2O

(2)

where O2Fe2+ and Fe represent the moles of hemeiron present as O2Hb and metHb, respectively. Equa-tion 2 also accounts for the failure of nitrite to release

molecular oxygen from O2Hb, because all oxygen isconsumed in the reaction.

Kinetics of the reaction. During these studies I ob-served that O2Hb is converted to metHb by nitrite muchmore slowly in the presence of cyanide. Furthermore,

O2Hb oxidized/mm, 11,2,mm

o 80I

0

0

60

404)

5 6

Time - Solid curve mm Dashed curves h

N14NO2 KCNmmol/lherpH

6.036.046.11

6.126.05

0.470.941.81

1.812.64

0.470.941.81

1.812.64

O2Hboxidized/mm,

8.225.327.821.330.8

mm

8.452.742.493.25

2.25

Fig. 4. Conversion of oxyhemoglobin to methemoglobin by nitriteat ph = 7.25Each solution contained phosphate buffer (18 mmol/Ilter) and hemoglobin (50zmol of home per liter). Solutions corresponding to the dashed curves alsocontained 0.5 mel of Nal per liter. The reaction was started by adding NaNO2,

up to 935 umol/Iiter, or a mixture of NaNO2 + KCN, up to 935 Mmol of each perliter. Note that on the abscissa, units of minutes are used for the solutions withphosphate alone (solid ctrves), but hours for the Nal-containing solutions (brokencurves)

00

80

0

0

‘ 60

40

20

4 6 8 0 12 4

CLINICAL CHEMISTRY. Vol. 22, No. 12, 1976 1989

Table 1. First-Order Velocity Constants for theCyanide-insensitive Conversion of O2Hb to metHb

by Nitrite a

pH

6.12 21.3 3.256.11 27.8 2.496.30 27.8 2.496.47 10.3 6.76.68 6.0 11.67.06 1.9 36.9

t = 28 #{176}C;IKCN] = 1.82 mrnol/liter, (NaNO2I 1.82 mmol/liter, and [021-lb+ metHb] 50 Mmol/liter (heme)

Table 2. First-Order Velocity Constants for theCyanide-Insensitive Conversion of O2Hb to metHb

by Nitrite a

a = 28 #{176}C,[O2Hb + metHb] = 50 tmol/liter (home)

the kinetics of the reaction were strictly first order inO2Hb concentration in the presence of cyanide andfollowed closely the “induction period” of the usualreaction represented in Figure 1. I studied this cya-nide-insensitive portion of the reaction by evaluatingthe first-order reaction rate constants as a function ofpH at constant nitrite concentration and of nitrite

concentration at constant pH. It was found that thereaction rate constant for the cyanide-insensitive re-

action was directly proportional to the proton concen-tration (Table 1) with constant nitrite. When pH wasmaintained constant, the reaction rate increased withincreasing nitrite concentration (Table 2). Variationsin the reaction velocity from direct proportionality wereobserved but are probably explained by the small dif-ferences in pH and in the mixing and starting times forthe rapid rates used. It was clear in all these studies that

no autocatalytic effect occurred in the presence of cy-anide at any pH or nitrite concentration used. The ex-treme effect of CN- on the overall reaction is illustratedin the solid curves of Figure 4, which shows a typicalautocatalytic reaction at pH 7.25 that was complete in

less than 10 mm. In the presence of CN- the autocata-lytic effect was completely eliminated and the cya-nide-insensitive reaction, t112 = 204 mm, accounted forless than 4% conversion in the same time.

The rates of the overall reaction, cyanide-insensitive

plus cyanide-sensitive, through the major portion of thereaction (20 to 90% metHb), were very closely repre-sented by the product of O2Hb concentration times thesquare of the metHb concentration. This suggested that

Time - m,n

Fig. 5. Effect of neutral salts on the conversion of oxyhemoglobinto methemoglobin by nitrite at pH = 7.25

Each solution contained phosphate buffer (18 mmol/liter) and the indicatedconcentration of NaCI or Nal. All solutions were 50 zmoI/liter in home iron ashemoglobin, and NaNO2 was added to a concentration of 935 zmol/liter at zerotime

the cyanide-sensitive portion of the reaction might re-sult from the interaction of one unit of O2Hb with twounits of methemoglobin nitrite. The previously deter-mined stoichiometry suggested that the O2Hb, at least,must be present as the tetrameric unit. I tested thishypothesis by measuring the rate of metHb formationin solutions of neutral salts known to dissociate O2Hbtetramers into the afl-dimers (12-14). In the presenceof 0.5 mol/liter Nal, oxyhemoglobin is about 96% dis-sociated into dimers (13). The conversion of such O2Hbdimers to metHb by nitrite is shown in Figure 4 andcompared with the same reaction in the absence of NaIwhere about 74% of the O2Hb is present as tetramers.The reaction in the presence of Nal was first order, witht 1,’2 = 77 mm, and showed no autocatalytic effect. Ad-dition of cyanide caused the reaction to start somewhat

1990 CLINICAL CHEMISTRY, Vol. 22, No. 12, 1976

more slowly, giving t112 = 86 mm for the cyanide-in-sensitive reaction.

Effects of increasing NaC1 concentration are shownin Figure 5. The inverse of the time required to go from20 to 90% metHb, 1/t20...0, was used to estimate the re-action rate. Tetramer-to-dimer dissociation constantsin 0.1, 1.0, and 2.0 mol/liter NaCl reported by Kellett(13) were used to calculate the initial concentrations ofO2Hb tetrainer in the solutions. In the absence of addedNaC1, with 74% of the total O2Hb present as tetramers,a value of t2oo = 1.4 mm was observed. Increasing theNaCI to 0.97 mol/liter, a concentration at which 52% ofO2Hb is present as tetramers, slowed the reaction tot20 = 1.95 mm. Further increase of NaCl to 1.97mol/liter, a concentration at which 39% of the O2Hb ispresent as tetramers, caused an additional decrease inrate to t2g0 = 2.9 mm. Thus with NaCl = 0.97 mol/literthe O2Hb tetramer concentration is 70% of that of thecontrol in the absence of NaC1 and the reaction rate is72% of the control. In 1.97 mol/liter NaCl the O2Hbtetramer is 52% of the control, while the reaction ratewas 48% of the control. These data suggest that theautocatalytic cyanide-sensitive portion of the reactionhas an obligatory requirement for O2Hb tetramers. Thecyanide-insensitive portion occurs with either the di-meric or tetrameric form, although the rates may differ.These suggestions are supported by the data in Figure4, which show that the reaction in the presence of 0.5mol/liter Nal (96% a-diniers) proceeds nearly the samein the presence and absence of cyanide.

Proposed reaction mechanism. Data presented aboveshow that the overall conversion of O2Hb to metHb inthe presence of excess nitrite may be explained as thesum of two reactions. One of these, the rate-limiting stepat the start of the reaction, involves the oxidation ofO2Hb to metllb and is unaffected by the presence of

CN. The rate of this reaction is directly proportionalto both W and NO2- concentration but is independentof metHb. The stoichiometry of the cyanide-insensitivereaction has not been determined. The reaction occurswhen the O2Hb is present as dimers or as tetramers,although there is some evidence that dimers react morerapidly than tetramers under otherwise comparableconditions.

The second reaction dominates both the stoichiom-etry and rate of the overall reaction. Each mole of O2Hbheme that is converted to metHb heme oxidizes 1.5 molof nitrite to nitrate and consumes 1 mol of H. The rateat constant pH in the presence of excess nitrite is pro-portional to the product (O2Hb)(metHb)2. The O2Hbmust be in the tetrameric form and metHb must bepresent as the nitrite derivative, because CN- com-pletely stops the reaction. The preferred tetramer-dimer state of the metHbNO2 is not known, as thedissociation constants have not been evaluated for this

compound. Guidotti (12) has shown that the tetra-mer-dimer dissociation constants for cyanmetHb andO2Hb are similar. This suggests that both O2Hb andmetHbNO2 must both be present as tetramers. Theautocatalytic nature of this reaction is due to the factthat all metHb present (either endogenous, that formedin the cyanide-insensitive reaction of O2Hb with NO2-,or that formed from the reaction of O2Hb withmetHbNO2j will associate with remaining excess NO2-and increase the concentration of this reactant. Thereaction rate thus continues to increase until the O2Hbbecomes limiting at about 66% conversion.

Supported by the Naval Medical Research and DevelopmentCommand, National Naval Medical Center, Department of the Navy,Research Task No. MRO41.01.01.01039. I wish to acknowledge theassistance of Dennis P. Nelson, LT, MSC, USN, ExperimentalMedicine Department, Naval Medical Research Institute, in themeasurement of the proton utilization during formation of methe-moglobin. I also wish to express my deep appreciation to Dr. LouisD. Homer, Clinical and Experimental Immunology Department,Naval Medical Research Institute, for valuable discussions on theinterpretationof the kineticdata.

References1. Austin, J. H., and Drabkin, D. L., Spectrophotometric studies III.Methemoglobin. J. Biol. Chern. 112,67 (1935).

2. Kiese, M., The biochemical production of ferrihemoglobin-formingderivatives from aromatic amines, and mechanisms of ferrihemoglobinformation. Pharmacol. Rev. 18, 1091 (1966).3. Smith, R. P., Some features of the reaction between cobaltinitriteand hemoglobin. Toxicol. App!. Pharmaco!. 17, 634 (1970).

4. Rodkey, F. L., and O’Neal, J. D., Effects of carboxyhemoglobin onthe determination of methemoglobin in blood. Biochern. Med. 9,261(1974).

5. Van Kampen, E. J., and Zijlstra, W. G., Standardization of hem-oglobinometry. II. The hemiglobincyanide method. Clin. Chim. Acta6,538 (1961).

6. Schneider, N. R., and Yeary, R. A., Measurement of nitrite andnitrate in blood. Am. J. Vet. Res. 34, 133 (1973).7. Wegner, T. N., Simple and sensitive procedure for determiningnitrate and nitrite in mixtures in biological fluids. J. Dairy Sci. 55,642 (1972).

8. Van Assendelft, 0. W., and Zijistra, W. G., The formation ofhemiglobin using nitrites. Clin. Chim. Acta 11,571(1965).

9 Brunori, M., Amiconi, G., Antonini, E., et al., The transition be-tween “acid” and “alkaline” heme proteins. Biochim. Biophys. Acta154, 315 (1968).10. Smith, R. P., The nitrite methemoglobin complex- its significancein methemoglobin analyses and its possible role in methemoglobi-nemia. Biochem. Pharmacol. 16, 1655 (1967).11. Smith, R. P., and Gosselin, R. E., On the mechanism of sulfideinactivation by methemoglobin. Toxicol. APP!. Pharmaco!. 8, 159(1966).

12. Guidotti, 0., Studies on the chemistry of hemoglobin II. The effectof salts on the dissociation of hemoglobin into subunits. J. Biol. Chem.242, 3685 (1967).

13. Kellett, G. L., Dissociation of hemoglobin into subunits. Li-gand-linked dissociation at neutral H. J. Mo!. Biol. 59, 401(1971).

14. Tomita, S., Maeda, N., Enoki, Y., and Okuda, T., An evidence fordissociation of hemoglobin into monomer in concentrated potassiumiodide. J. Mo!. Biol. 69, 303 (1972).