J. biol. chem. 1986-fushitani-8414-23

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1986 by The American Society of Biological Chemists, Inc.

Vol. 261, No. 18, Issue of June 25, pp. 8414-8423 1986 Printed in L k A .

Oxygenation Properties of Hemoglobin from the Earthworm, Lumbricus terrestris EFFECTS OF pH, SALTS, AND TEMPERATURE*

(Received for publication, December 16,1985)

Kenzo FushitaniSBT, Kiyohiro ImaiB, and Austen F. RiggslI)) From the *Department of Biophysical Engineering, Faculty of Engineering Science, Osaka University, Toyonaka, Osaka 560, Japan, the $Department of Physicochemical Physwlogy, Medical School, Osaka University, Nakanoshima, Osaka 530, Japan, and the TDepartment of Zoology, University of Texas, Austin, Texas 78712

Oxygen equilibrium curves of the extracellular hemoglobin from Lumbricus terrestris were determined under a variety of conditions. These data were characterized by (i) a rather small free energy of cooperativity (1.6-2.8 kcal/mol), (ii) a large and strongly pH-dependent Hill coefficient with a maximum value of 7.9, (iii) a high sensitivity of the upper asymptote of the Hill plot to pH, and (iv) a maximum association constant as large as that of the free #I subunit of human hemoglobin A.

The effects of LiCl, KCl, NaCl, BaCl,, CaC12, SrC12, and MgClz on the oxygen equilibrium were measured. Cations, not C1-, were found to control oxygen binding. Divalent cations have a larger effect on oxygen affinity than monovalent cations, and their effectiveness decreased in the order listed above within each valence class. These specific effects depend in part on ionic radius and cannot be explained in terms of ionic strength. The data indicate that the oxygenation- linked binding of a Ca2+ ion is accompanied by the release of two protons; the binding of a Na+ ion is associated with the release of one proton. These findings indicate that the oxygenation-linked cation-binding site contains two acid groups that do not readily dissociate their protons except when replaced by cations.

Incubation at either pH 6.2 or 8.9 had no effect on subsequent measurements of oxygen equilibria at pH 7.8. The apparent heat of oxygenation was found to be -1 1.8, -7.3, and -9.3 kcal/mol at pH 9.0,7.4, and 6.6, respectively. These differences indicate that proton- binding processes contribute to the heat of oxygenation.

The extracellular hemoglobin of the earthworm, Lumbricus terrestris, consists of 12 subunits, arranged as two superimposed hexagonal disks (3), about 30 nm in diameter and 20 nm high (4) with a molecular weight of 3-4 X lo6. Although its chain composition and the spatial arrangement of the chains have not yet been determined, one of the polypeptide chains has been sequenced and shown to be homologous to

PCM 8202760 and DMB-8502857, Welch Foundation Grant F-213, * This work was supported by National Science Foundation Grants

and National Institutes of Health Grant GM28410. A preliminary account of some of this work has been presented (1,2). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduer- tisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

11 To whom reprint requests should be addressed.

chains of vertebrate hemoglobins (5). The whole molecule has been reconstituted from dissociation products and the shape has been compared with that of the native molecule by scan- ning transmission electron microscopy (6). Oxygen-binding properties of L. terrestris hemoglobin have been studied by several investigators (7-9). Giardina et al. (10) measured oxygen binding by earthworm’ hemoglobin under a variety of conditions and showed that the Hill coefficient, n, depends strongly on pH and has a maximum value of 4 at pH 7.8; the oxygen affinity varies between 0.4-7 mm Hg between pH 5.5 and 10. Vinogradov et al. (11) have reported similar values for hemoglobin from L. terrestris.

Recently, Weber (12) showed that cations control the oxygen affinity of the extracellular hemoglobin of Arenicolu marina. Addition of cations such as Na+ or Mg2+ enhanced cooperativity and raised the oxygen affinity by binding to hemoglobin at high levels of oxygenation. He also showed that protons lower the oxygen affinity by preferential binding to hemoglobin at high saturation levels. The mechanisms of cationic and protonic interaction in these extracellular hemoglobins must be quite different from those in human hemoglobin A, where anions and protons decrease the oxygen affinity by preferential binding mainly to molecules in the low affinity state (13, 14). The important study by Santucci et al. (15) has shown for hemoglobin of Octolasium complun- atum that oxygen binding becomes independent of pH between pH 7 and 8.5 at sufficiently low cation concentrations. The pH dependence of cooperativity and cationic control of oxygen affinity has also been found in several other extracellular hemoglobins of annelids (16-23).

In the present study we have measured oxygen binding between 1 and 99% saturation with high precision under a wide variety of conditions which includes different kinds and concentrations of salts, changes in pH, temperature, and protein concentration.

EXPERIMENTAL PROCEDURES

Preparation of Hemoglobin-Earthworms, originally obtained in Ontario, were purchased from the Wholesale Bait Co., Hamilton, OH 45015. They were cut with scissors at a position just anterior to their hearts and bled into CO-saturated 0.1 M sodium phosphate buffer (pH 7.0) containing 3 mM phenylmethylsulfonyl fluoride as a protease inhibitor. The crude solution was centrifuged to remove cellular matter. The hemoglobin was precipitated by adding polyethylene glycol (8,000 average molecular weight, Sigma) to a concentration of

Originally identified in 1975 as Lumbricus sp. (10) and recently reidentified (1983) as Octohium complnmtum (6) although the worms used by Giardina et al. (10) were referred to as L. terrestris in 1984 (15).

8414

Oxygenation Properties of Lumbricus terrestris Hemoglobin 8415

10%. The precipitate was redissolved in 0.05 M Tris-HC1, pH 8.0, and twice pelleted at 40,000 rpm for 2 h. Pellets were resuspended in this CO-saturated buffer and stored in liquid nitrogen. For oxygenation experiments, an aliquot of the frozen hemoglobin sample was thawed, centrifuged at low speed, and subjected to gel chromatography on a Sepharose CL-GB column equilibrated with 0.1 M sodium phosphate buffer (pH 7.7) containing 0.1 M NaCl saturated with CO. Less than 5% of the total hemoglobin was found to be dissociated into %z subunits during the freeze-thaw treatment on the basis of gel chromatography. The eluted fractions were concentrated to about 3.2% w/v by ultrafiltration (Toyo Roshi apparatus, UM-10 membrane). The concentrated hemoglobin solutions were dialyzed against 0.05 M Tris-HC1 (pH 7.0 at 25 "C) containing 0.1 M NaCl and stored on ice as the CO form.

Oxygenation Measurements-Oxygen equilibria were measured with an improved version of an automatic oxygenation apparatus (24) without the use of an enzymatic reducing system. Deoxygenation and/or reoxygenation data were acquired in real time by a model PDP-ll/vOB microcomputer (Digital Equipment Corp., Maynard, MA) and were stored on floppy disks. The absorbance value corresponding to 100% saturation with oxygen was obtained by extrapolating a &4 versus 1/P plot to 1/P = 0 where A.4 is the absorbance change upon oxygenation and P is the partial pressure of oxygen (13). The buffer was 0.05 M BisTris/propane' over the range pH 6.2-9.0. The pH was adjusted with concentrated HCl at the same temperatures as those used in the oxygenation experiments. The hemoglobin concentration was 60 FM on a heme basis, unless otherwise stated. The bound CO was removed from the hemoglobin by light while the sample was flushed with pure 0' in a rotating flask immersed in ice water just prior to the measurements.

Reproducibility of the Oxygenation Curve-Reproducibility of the oxygen equilibria is affected by two factors: the stability of the automatic oxygenation apparatus and that of the hemoglobin sample. Careful inspection of the data showed that oxygenation curves measured under the same conditions within a few days of one another could be superimposed over the whole saturation range between 1 and 99% with a variance of 1-2% of P W and n,- values where Pw is the oxygen pressure at half-saturation and n,- is the maximum value of the Hill coefficient. Therefore, the reproducibility of the oxygenation curve was found to depend mainly on irreversible alteration of the hemoglobin sample during storage such as autoxidation to a hemichrome. The PW value dropped by 3%, and n- decreased from 5.0 to 4.8 at pH 8.9 after 3 weeks. Purified CO-hemoglobin samples gave a variance of about 6% for PW values and 7% for n.-= values after storage on ice for 3 months.

Estimation of Methemoglobin Content-No complete set of absorption coefficients for estimating methemoglobin content over the pH range of 6.2-8.9 is available for this hemoglobin. We failed to determine the coefficients for methemoglobin because, as Ascoli et al. (25) reported, the spectrum of the L. terrestris hemoglobin changes upon oxidation from that of typical aquomethemoglobin to that of hemichrome. The millimolar extinction coefficient for oxyhemoglobin at 576 nm, obtained by the pyridine hemochromogen method, was 16.4, which is close to the value of 14.6 at 577 nm for the oxy form of human HbA (26). We used the value of 34.4 for the millimolar extinction coefficient of the pyridine hemochromogen at 557 nm. Methemoglobin content was estimated by using spectra taken before and after oxygenation measurements. The ratio of absorbance at a trough near 505 nm to that at a peak at 576 nm provides an index of methemoglobin. The value of this index for hemoglobin solutions between pH 7.0-7.8 before measurements was 2.99 0.03 S.D. (26 measurements). The index for human HbA obtained under similar conditions was 2.99 f 0.06 S.D. (10 measurements). The value of2.99 for human HbA corresponded to 2.8% f 1.1 S.D. methemoglobin (27). We estimated that 2-4% methemoglobin was present in the purified stock solutions of L. terrestris hemoglobin by these approxi- mate criteria. Methemoglobin content after oxygen equilibrium measurements was usually estimated to be 6-8% and never more than 10% except for three experiments carried out at heme concentrations under 20 pM (see Fig. 7). Estimation of MetHb content at other pH values was similar. The index decreased to 2.90 f 0.03 S.D. before measurement 3 months after purification.

'The abbreviations used are: BisTris, 2-[bis(2-hydroxyethyl)- amino]-2-(hydroxymethyl)-propane-1,3-diol; MetHb, methemoglobin; MWC, Monod-Wyman-Changeux.

Analysis-Oxygenation data were expressed in terms of the Hill plot (log(Y/(l- Y ) ) uersus log P ) where Y is fractional saturation of the hemoglobin with oxygen and P is the oxygen pressure in mm Hg. Overall oxygen affinity and cooperativity were characterized in terms of oxygen pressure at half-saturation (PW) and maximum slope of the Hill plot (hex) or slope of the Hill plot at half-saturation (nw). Cooperativity was also expressed in terms of a free energy change defined as AG, = RT ln(K,,,/KJ where K,,, and Kl are estimated association constants for the last and first oxygens bound to hemoglobin. Values of n,, were determined by plotting values of n against logp. Here, n is the slope of the line connecting two adjacent points on the Hill plot. The apparent association constants for the binding of the first and last oxygens to hemoglobin were estimated approximately with an m-step Adair's oxygenation scheme (28),

log(Y/(l - Y ) ) = log P + log Kl at P -+ o ( 1 )

log(Y/(l - Y ) ) = log P + log K,,, at P -+ m (2)

where Kl and K,,, are apparent association equilibrium constants for the first and last oxygens.

The magnitude of the Bohr effect was estimated by the following expression (29).

h = -Alog Pw/ApH (3)

Here, h is the number of protons released by Hb/oxygen bound. Similarly the magnitude of the effect of a given ion on oxygenation is expressed by

I = +Alog PW/Alog[ion] (4)

where I is the number of ions released by Hb/oxygen bound. The apparent enthalpy of oxygenation, A H , was calculated from

the slope of the plot of log PW versus 1/T between temperatures of 283 and 303 K.

The Monod-Wyman-Changeux (MWC) allosteric model (30) de- scribes the oxygenation of hemoglobin as,

Y = L K T P ( ~ + KTP)""' + K ~ p ( 1 + KRP)""' L(1 + KTP)" + (1 + KRP)'" (5)

where rn is the number of interacting binding sites for oxygen, KT and KR are the intrinsic association equilibrium constants for oxygen binding to the T state and the R state, respectively, and L is the allosteric constant. Recently, Decker et al. (31) introduced a conven- ient graphic method for analyzing oxygenation data on the basis of the MWC model. Imai and Yoshikawa (28) modified their formula slightly, as follows.

log Z = (m - 1) log X + log L (6)

Here, 2 = (KR - Q ) / ( Q - KT), Q = Y/((l - Y)P), and X = (1 + K T P ) / ( ~ + KRP). Z is the ratio of the difference in affinity between hemoglobin in the R state and hemoglobin at a given value of P to that between hemoglobin at the given value of P and the T state. X is the ratio of the binding polynomial for oxygenation of the T state to that for oxygenation of the R state. When KT and KR values are already known, log Z can be plotted against log X, yielding a straight line. The values of m and log L can then be determined from the slope and the intercept on the ordinate at log X = 0. In the present study, we used Kl and K,,, for KT and KR, respectively.

We also attempted to estimate the number of interacting binding sites, m, by using Kegeles' expression (321, m = n,-/y-, derived from the MWC model. However, the slope is too steep to permit accurate values to be obtained.

RESULTS

Oxygen-binding Properties: Effect of pH and Calcium-Hill plots of oxygen equilibria of L. terrestris hemoglobin are given in Fig. 1. Measurements were made between pH 6.2 and 8.9 using 0.05 M BisTris/propane with and without different additional salts. Values of Pw and n- obtained from these curves are listed in Table I together with other parameters and are plotted against pH in Fig. 2. Fig. 2 also includes data from human hemoglobin A (34) for comparison. The data for L. terrestris hemoglobin in Figs. 1 and 2 have four major features as described in the following paragraphs.

8416 Oxygenation Properties of Lumbricus terrestris Hemoglobin

log P FIG. 1. Hill plots of oxygen binding by hemoglobin of L. terrestris at different pH values and salt

concentrations. Symbols: Y, fractional saturation of hemoglobin with oxygen; P, partial pressure of oxygen in mm Hg. Conditions: 50 mM BisTris/propane, 3 mM NaCl, 25 "C, with various amounts of additional salts; hemoglobin concentration, 60 pM heme. A, no additional salt. pH from the left to right: 8.56,8.13, 7.75, 7.36, 7.00, 6.58, 6.23. B, additional salt, 0.1 M NaCl. pH from the left to right: 8.86, 8.52, 8.10, 7.75, 7.35, 6.95, 6.55, 6.16. C, additional salt, 0.1 M NaC1,25 mM CaC12. pH from the left to right: 8.88,8.48,8.10, 7.73,7.36,6.98,6.58, 6.20. The solid straight lines with a slope of unity indicate the lowest lower and highest upper asymptote of the Hill plots. The lowest lower asymptote was obtained from cunres in A, the highest upper asymptote from C. Their intercepts on the ordinate at log P = 0 give log K L =i -1.6 (KL = 0.024 mm Hg") and log KH = 0.58 (KH = 3.8 mm Hg-'), where K L and KH are the lowest and highest values, covered by the present experiments, corresponding to the association equilibrium constant for the low oxygen affinity state (the lowest first Adair constant, KJ and that for the high oxygen affinity (the highest last Adair constant, Km), respectively. The asymptotes for free B subunit and hemoglobin M Milwaukee corresponding to 4 and 0.004 mm Hg", respectively, are also shown for comparison (33).

1) The shape of the oxygen equilibrium curve varies greatly with pH. In the absence of calcium, the value of kx rises from 2.5-3.0 at pH 6.2-6.6 to a maximum of 6.5-7.2 near pH 8.1 and decreases to 5 at pH 8.9. The value of nmlu rises to a maximum value of 7.9 at pH 7.5 in the presence of calcium; values of n- at both extremes of pH remain unchanged.

2) Salts increase overall oxygen affinity. This effect increases with pH and is more pronounced with CaClz than NaC1. The pH of the maximum Bohr effect is decreased by salts. The maximum values of the Bohr coefficient (Alog Ps0/ ApH) are -0.35 (pH 8.3-8.5), -0.53 (pH 8.6-8.9), and -0.77 (pH 7.4) in the presence of no additional salt, 0.1 M NaCl, and 0.1 M NaC1, 25 mM CaC12, respectively.

3) The two asymptotes of the Hill plot depend on pH differently. The values of Kl and K, calculated from the extrapolated asymptotes are listed in Table I. Their pH dependence is given in Fig. 3. The value of Kl is about 0.024 mm Hg" between pH 6.2 and 8.2 in the absence of added salt. However, the value of K, increases from 0.43 mm Hg-l at pH 6.2 to 1.9 mm Hg-I at pH 8.6. NaCl and CaClz both enhance the pH dependence of K,. In the presence of CaCh (Table I, C), the value of K, reaches a plateau near pH 8 corresponding to a value of 3.8 mm Hg". The value of Kl increases only slightly with pH in the presence of Ca2+.

4) The overall free energy changes associated with cooper-

ative binding of oxygen are smaller than those of human HbA (33), although the values of the Hill coefficient are much larger (Fig. 2 and Table I). The lowest value for Kl, 0.024 mm Hg-l, was obtained in the absence of added salt (Fig. lA). The highest value for K,, 3.8 mm Hg-l, was obtained in the presence of 0.1 M NaC1, 25 mM CaClz (Fig. lc, Table I). The asymptotes corresponding to these values are shown in Fig. 1. Values of free energy of cooperativity (AG,) are listed in Table I. The pH dependence of AG, is similar to that of hax (Fig. 2). The maximum AG, value, 2900 cal/mol, was obtained in the presence of 0.1 M NaCl, pH 8.52. This value is about 80% of the maximum value, 3640 cal/mol, for human HbA in 0.1 M Cl-, 2 mM 2,3-diphosphoglycerate, pH 7.4 (33).

Effect of Different Salts on Oxygenation-Fig. 4 and Table I1 show the effects of chloride salts of Li+, Na+, K+, Mf, Ca2+, and of mixtures at pH 7.36, 25 "C. Monovalent salts have almost no effect on nmar or log Ps0 up to 0.1 M, and changes in the Hill coefficient are negligible up to 1 M. However, the oxygen affinity doubles between salt concentrations of 0.1 and 1 M. The effect of divalent salts is pronounced at much lower concentrations; the effect of 0.1 M CaC1, is similar to that of 1 M NaC1. The combination of Ca2+ and 0.1 M NaCl causes a concentration dependence of log Pm similar to that caused by Ca" alone, whereas ha= shows a quite different dependence on CaClz compared with NaC1.


TABLE I Values of oxygenation parameters for hemoglobin of L. terrestris

obtained under different conditions of salt and pH In 0.05 M BisTris/propane/HCl buffer, at 25 "C; hemoglobin con-

centration 60 PM on a heme basis. A, no additional salt; B, in the presence of added 0.1 M NaCI; C, in the presence of added 25 mM CaCL 0.1 M NaCl.

A 8.56 8.13 7.75 7.36 7.00 6.58 6.23

B 8.86 8.52 8.10 7.75 7.35 6.95 6.55 6.16

C 8.88 8.48 8.10 7.73 7.36 6.98 6.58 6.20

mm Hg

5.9 8.2

10.3 12.7 15.2 16.4 17.1

2.7 4.1 6.3 9.0

12.1 15.4 17.0 17.3

1.6 1.9 2.6 3.9 6.9

11.6 15.5 17.2

eallmol

6.3 5.4 2400 6.5 5.6 2400 6.0 5.2 2300 6.6 5.1 2300 4.1 3.6 2200 3.1 3.0 1700 2.5 2.3 1700

5.0 4.9 2300 6.9 6.6 2900 7.5 6.5 2800 7.2 5.6 2600 6.2 5.3 2400 4.5 3.5 2100 3.0 2.5 1700 2.7 2.2 1700

4.9 4.9 2500 5.8 5.5 2600 6.8 6.4 2800 7.9 6.9 2800 7.7 7.1 2600 5.1 4.2 2100 3.1 2.7 1900 2.6 2.3 1700

11.2 71 8.9 12.4 72 9.0 11.6 72 8.3 11.1 70 9.4 8.5 67 6.1 6.5 75 4.1 5.4 63 4.0

9.9 46 10.9 10.8 61 11.3 10.8 69 10.9 10.8 74 9.7 11.1 73 8.5 8.4 66 6.8 6.5 67 4.5 5.1 72 3.8

8.0 57 8.6 9.4 56 10.4

10.5 61 11.1 12.4 67 11.8 11.9 67 11.5 9.7 70 7.3 6.2 71 4.4 5.7 70 3.7

a Oxygen pressure at half-saturation. * Maximum slope in Hill plot.

Slope at half-saturation in Hill plot. Free energy of interaction; AG, = RTln(Km/Kl).

e Estimated number of interacting sites based on the linearized plot

f Oxygen saturation giving a n,-. (Equation 6).

Estimated interacting sites based on Kegeles' analysis.

Oxygenation in the presence of 3 mM NaC1, 25 mM CaCli (Fig. 5) shows that CaClz increases the oxygen affinity and cooperativity by shifting the upper asymptote to the left. Although 25 mM CaClz has no effect on the lower asymptote, 125 mM CaClz causes a significant shift to the left which is associated with a decrease in ha. Similar effects were observed with MgClz.

The effect of 25 mM S?+ and Ba" was also examined at pH 7.37. Positions of the lower asymptote of the oxygenation curves were similar to those for Ca" and M$+ whereas the position of the upper asymptote differed. The log Pm values were: 0.98,0.86, 0.80, and 0.74 for M$+, Sr2+, Ca", and Ba2+, respectively. Thus, Ba2+ with the largest ionic radius was the most effective and M$+, the smallest ion, had the least effect among the divalent cations so far examined, and Sr2+ and ea2+ with intermediate ionic radii had effects between those of Ba" and M$+. Although the difference between the log P m values for Sr2+ and Ca" or Ca" and Ba2+ are only 0.06 this is 15% in the P value and appears to be well beyond the estimated experimental error. Inositol hexaphosphate had no effect on oxygenation; the data obtained with 2 mM inositol hexaphosphate at pH 7.41 could be superimposed on those obtained in its absence.

Effect of Partial Oxidation-We examined the effect of partial formation of methemoglobin on oxygen affinity and

I I I

9

E c

----"""" i 'M T T

FIG. 2. The effect of pH on overall oxygen affinity (log P,) and cooperativity (%=) at 25 OC. Symbols: Pw, partial oxygen pressure at half-saturation; hx, maximal slope of the Hill plot. Values of these parameters were obtained from the Hill plots in Fig. 1. A, no salt added; 0 , O . l M NaCI; 0 , 2 5 mM CaC& with 0.1 M NaC1. Broken lines show a plot of log P, and for human HbA in 0.1 M NaCl(34). P,,, is the median oxygen pressure.

I ' 1 I (

-2.oL " . 8 I

6.0 7 .O 8.0 9.0 PH

FIG. 3. The effect of pH on K1 and K,. Kl and K,,, are the association equilibrium constants for the first and last oxygens to be bound. Values of these parameters were obtained from the Hill plots in Fig. 1. A, no salt added, 0,O.l M NaCI; 0 , 2 5 mM CaClz with 0.1 M NaCI; 0, estimated values of Kl and K, obtained by extrapolation to zero MetHb of the data obtained with par t idy oxidized hemoglobin (see text).

the Hill coefficient by measuring oxygen binding by hemoglobin which had been partially oxidized with potassium ferricyanide at pH 7.3 (Fig. 6). The ferricyanide was assumed to react completely with heme. Methemoglobin content up to 30% had no signifkant effect on the P m value which was 11.2-11.8 mm Hg (average, 11.5 mm Hg k 0.2 S.D.). The value increased slightly at 50% methemoglobin. However the

value decreased with an increase in methemoglobin (6% decrease with 10% methemoglobin). We estimated values of both Kl and K , at 0% methemoglobin by extrapolating plots of log Kl or log K , against MetHb content. The extrapolated values are -1.7 and 0.13 for log K1 and log K,, respectively (Fig. 3). These results indicate that Kl and K,,, were not affected significantly by the methemoglobin present under


c :m I

T T

Q! 0

- 0 0

Salt Concentration

FIG. 4. Effect of salts on overall oxygen affinity (log Pso) and cooperativity (n-). Symbols: PSO, partial pressure of oxygen at half-saturation; n,-, maximum slope of the Hill plot. Hemoglobin concentration, 60 p~ on a heme basis, 25 "C. A, effect of monovalent cations: 0, LiCl (pH 7.33-7.36); 0, NaCl (pH 7.38-7.39); A, KC1 (pH 7.36-7.40). B, effect of divalent cations: 0, MgClz in the presence of 0.1 M NaCl (pH 7.34-7.38); 0, CaClz in the presence of 0.1 M NaCl (pH 7.36-7.38); A, CaC12 only (pH 7.36-7.40). Broken lines show effect of monovalent salts; data from A included for comparison.

3 t

0

1

-3- - 1 0 1 2 3

log P

FIG. 5. Effect of calcium chloride on oxygen binding by hemoglobin of L. terrestris in terms of the Hill plot. Buffer: 0.05 M BisTris/propane/HCl and 3 mM NaCl (pH 7.36-7.40) at 25 "C. Hemoglobin concentration, 60 WM (heme basis). P, oxygen pressure, mm Hg; Y, fractional oxygen saturation. Ca2+ concentration: from the left to right, 125, 25, 0 mM.

the conditions used. Effect of Hemoglobin Concentration-Fig. 7 shows the effect

of increasing the hemoglobin concentration from 0.6 to 600 pM at pH 7.57 in the presence of 0.1 M NaC1. Values of log P50 and nmax are not of high accuracy at concentrations of 2 p~ or lower because of significant autoxidation. The 100-fold decrease in concentration from 600 to 6 PM is associated with an increase in Pw of about 7% and a decrease in nmar from 6 to 5. The upper and lower asymptotes of the Hill plots shift only slightly upon the dilution of hemoglobin sample. The small changes in log PSo and at or below 20 PM heme are very closely correlated with methemoglobin formation; the P50 'value is linearly related to the MetHb index with. a correlation coefficient of 0.97. The MetHb index is invariant between 60 and 600 pM heme. Although a small degree of subunit dissociation may accompany dilution, dissociation is

cg ~~ I 25°C

0'50 10 30 50 % Meihernoglobin

H b Concentration (pM heme)

FIG. 6 (left). Effect of partial oxidation on overall oxygen affinity (log Ps0) and cooperativity (ram-). Oxygenation curves were measured in 0.05 M BisTris/propane/HCl, 0.1 M NaCl, pH 7.30 at 25 "C. Hemoglobin concentration, 60 ptM on a heme basis. The abscissa shows the percentage of methemoglobin.

FIG. 7 (right). Effect of hemoglobin concentration on overall oxygen affinity (log P60) and cooperativity (k-). Oxygenation curves were measured in 0.05 M BisTris/propane/HCl buffer, pH 7.38-7.39, at a concentration of hemoglobin between 0.6 and 600 p~ on a heme basis.

TABLE I1 Effect of salts on overall oxygen affinity and number of cations taken

up upon oxygen binding In 0.05 M BisTris/propane/HCl buffer at 25 "C; pH 7.34-7.40;

hemoglobin concentration, 60 NM heme.

Salt" Cation bound/O.

mM Li+ 580 0.75 Na' 940 K+

0.52 820 0.45

M$+ + 0.1 M Na+ 110 0.33 Ca2+ 27 0.37 Ca2+ + 0.1 M Na+ 41 0.37

* Salt concentration needed t o double overall oxygen affinity.

not required to explain the very small shifts. Reversibility of Oxygen Equilibria with Changes of pH-

Giardina et al. (10) reported that the shape of the oxygen equilibrium curve of hemoglobin incubated at either pH 6.0 or 10 for 1 h is modified irreversibly. Several workers have reported similar phenomena for other extracellular annelid hemoglobins (35-37). We have re-examined this property with L. terrestris hemoglobin. Oxyhemoglobin solutions at pH 6.2, 7.8, and 8.9 were prepared as described under "Experimental Procedures." Each solution was incubated for 1 h at 25 "C and passed through a Sephadex G-25 column (pH 7.8) so that the three samples had the same fiial pH (Fig. 8, Table 111). The plots at the left in Fig. 8 show the Hill plots for a control experiment, where the curves for the hemoglobin samples untreated and treated (i.e. incubated and passed through the column) at pH 7.8 are superimposed. Likewise, the plots at the right in Fig. 8 show five curves. Two of these show data obtained at pH 6,2 or 8.9 and not brought back to pH 7.8. Three curves show data on samples treated at pH 6.2, 7.8, or 8.9 and brought back to pH 7.8. Over a range of 1-99% saturation, their agreement is excellent. It was found that addition of 25 mM CaClz to a solution incubated at pH 8.9 improved the reversibility even further for pH changes from


FIG. 8. Test of the reproducibility of oxygenation after changing the buffer. Hemoglobin solutions incubated for 1 h at each pH (untreated; shown with open symbols) were brought back to the same pH 7.8 with Sephadex G-25 (treated; shown with solid symbols). Left two curves are shown for control purposes.

-2 - I 1 I t I I I I

I 2 - I 0

TABLE I11 Reversibility of oxygenation parameters upon changing pH of

hemoglobin solution In 0.05 M BisTris/propane/HCl buffer with 0.1 M NaCl at 25 "C.

Hemoglobin concentration, 60 pM heme. Buffer was changed with Sephadex G-25. P50, oxygen pressure at half-saturation; n-, maximum slope in Hill plot.

Condition" PSO n,- MetHb indexb

7.8 (untreated)" 7.64 6.7 2.99 2.85 7 . h 7 . 8 7.59 6.0 2.81 2.67 6.2-7.8 7.83 5.9 2.81 2.62 8.9+7.gd 7.87 5.9 2.80 2.64 8.9-7.8' 7.83 5.1 2.95 2.42

Left column shows starting conditions. Right column shows conditions after change (see legend for Fig. 8).

Ratio of absorption at minimum value near 505 nm to that at the a peak in oxy form. Left, before measurement; right, after measurement.

No buffer change with Sephadex G-25. Incubated at pH 8.9 with 25 mM CaC12.

e Incubated at pH 8.9 without 25 mM CaC12.

8.9 to 7.8. The numerical data for these experiments (Table 111) show excellent reversibility with respect to pH change.

Effect of Temperature on Oxygenation-Oxygen equilibrium curves were determined at 10,15, 20,25, and 30 "C at pH 6.6, 7.4, and 9.0 in the presence of 0.1 M NaC1, but satisfactory data at pH 9.0 and 30 "C could not be obtained because of significant autoxidation. The dependence of oxygenation of L. terrestris hemoglobin on temperature is similar in extent to that for human HbA (38). Overall heat of oxygenation (AI&) was obtained by plotting log Pm against 1/T. The values are listed in Table IV, together with data for human HbA (27).

DISCUSSION

Oxygenation Characterization of L. terrestris Hemoglobin- Oxygenation parameters for L. terrestris hemoglobin in the present work are compared in Table V with those for the same hemoglobin obtained by other workers, for hemoglobin from two related earthworms, for chlorocruorin from Potam- ilk leptochaeta, and for human HbA. Our data on the strong

I I I -

n

pH 8.9 -7.8 0 8

0 .E *A

0 0

-I 0 I 2 log P

TABLE IV Apparent heat of oxygenation (kcal/mol)

In 50 mM BisTris/propane/HCl buffer with 0.1 M NaC1. Hemoglo- bin concentration, 60 pM heme. Heats of oxgen binding at oxygen saturations of 50%. Values were corrected for heat of solution of oxygen (3 kcal/mol).

pH L. terrestris Human Hb 9.0-9.1 -11.8 -15.3" 7.4 -7.5 -9.3 6.5-6.6 -9.3 -11.2

In 50 mM BisTris or Tris/HCl buffer with 0.1 M NaC1, at pH 6.5, 7.4, 9.1 (27).

pH dependence of n,,, are qualitatively in agreement with results obtained by Giardina et al. (lo), Santucci et al. ( E ) , and Vinogradov et al. (ll), although our n values are consist- ently higher even when expressed as nsO rather than n- (Table I). Qualitatively similar properties have been reported for extracellular hemoglobins from other annelids, both po- lychaetes and oligochaetes: Amphitrite ornata (20), Lumbri- nereis tertraura (19), Pheretima hilgedorfi (22), and Eiseniu foetida (23).

This pH dependence has been suggested to result from a greater sensitivity of the upper asymptote to pH than the lower one (20). Recently, Weber (12) clearly showed for A. marina hemoglobin that the upper asymptote in the Hill plot shifted to the left with an increase in pH but that the lower asymptote hardly moved. He suggested that the oxygen affinity of the high affinity states depends on cation and proton binding. Our data (see Fig. 3) for L. terrestris hemoglobin are completely consistent with this picture. The upper asymptote is more sensitive to pH than the lower one although the lower asymptote does shift slightly above pH 8.0 in the absence of salts and gradually shifts to the right in the presence of 0.1 M NaCl and 25 mM CaClz above pH 7.0.

The data in Fig. 3 may be interpreted in terms of at least two high affinity states in L. terrestris hemoglobin. One appears at low pH independent of the presence of salts (KmL = 0.43 mm Hg-') and the other at high pH and/or high salt concentrations (K," = 3.8 mm Hg-l, consequently equal to KH). Free energy differences between KL and KmL, and K,"


TABLE V Oxygenation parameters for extracellular hemglobin and chloroeruorin

Source" Temperature pH P, n Bohr effect AHb Reference oc

L. terrestris 25

7 20

10 10

15 15 15 25 25 25

22

0. complnnatom 20

E. foetida 20

P. leptochaeta 25

Human HbA 25

6.2-9.0

7.3 7.3

7.12 7.21

7.10 7.44 7.10 7.70 7.44 7.10

6-9

5.4-10.0

5.1-9.4

6.2-9.2

6.0-9.0

1.63-17.3

2 a

4.8

3.89 4.98 5.28 6.78

(2.7-15.8)

3.5

2.88

9.20

(2.2-6.9)

(0.59-4.4)

11-420

1.6-16

2.5-7.9

(3.4)'

(3.0) (2.3)

5.21 5.20 5.13 5.41 5.30 5.11

(1.8-5.0)

(1.6-4.2)

(1.8-3.5)

1.14-5.82

2.53-2.98

-0.35-0.77

-0.25

-0.4

-0.4

(-0.54)

(-0.64)

-0.44

-0.98

-0.53

-7.5 to -11.8

-8.0

-10.2 -9.1

-10.6

-6 to -13.7

-3.9

-11.2 to -15.3 ' P. leptochaeta is the source of chlorocruorin; the others are those of hemoglobin. Values have been corrected for the heat of solution of oxygen, 3.0 kcal/mol, except for those from Ref. 9. Values in parentheses are obtained bv reading or recalculating values from published data. See Footn>te 1.

-

and KmH are 1700 and 1300 cal/mol, respectively. The change of free energy of cooperativity with pH (Table I) is essentially parallel to that of n- under three sets of conditions. Thus, the pH dependence of the shape of the oxygen-binding curve between pH 6.2-9.0 results from the relative movements of the lower and upper asymptotes. The increase in at low pH appears to be due exclusively to an increase in the K,,, value with constant Kl) and the decrease in n,, at alkaline pH is due to a slight increase in K1 with the upper asymptote remaining unchanged. On the same basis, the shift of pH giving the maximum value in the presence of 0.1 M NaC1,25 mM CaClz can be explained by a sharp increase of the K, value to its maximum combined with a more gradual increase of Kl (Fig. 3). A similar effect has been observed in chlorocruorin (28).

Similar shapes of the pH dependence of n- have been reported for several extracellular hemoglobins and chlorocruorins from different annelid species (8-12, 16-23, 37, 39- 45) where the pH giving a maximum n varies from pH 7.5 to 9.0. Chiancone et al. (35) reported an exceptional case where the shape of pH dependence of the Hill coefficient is concave upward as observed for %z subunits from Affinis affiinis (44).

Effect of Salts-The present experiments show that the effect of different salts on oxygen binding are specific and depend exclusively on the cations. Comparison of data at the same c1- concentrations (Fig. 4B), one set at 50 mM CaClz and the other at 100 mM NaCl or at the corresponding ionic strength of NaCl (150 mM), shows clearly that the increase in oxygen affinity is caused by Ca2+, not by C1-. The data also show that the effects cannot be explained merely in terms of changes in ionic strength. The absence of an effect of C1- on oxygenation is supported by nuclear magnetic quadrupole relaxation experiments, where C1- binds to both liganded and unliganded forms of the hemoglobin with the same affinity (45). The stronger effect of divalent than monovalent cations is probably also true for other extracellular hemoglobins and chlorocruorin (12, 20, 28). Among divalent cations, Ca2+ has

a stronger effect than M$+ in E. foetida (23) and L. terrestris (present data), but in A. ornata (20) the opposite is true.

Similar effects of salt concentration on log P50 and (Fig. 4B) have been found in several other extracellular hemoglobins (11, 15, 18, 23). One interpretation is as follows: (i) cations bind to the liganded form of hemoglobin more than to the unliganded form at relatively low salt concentrations, resulting in a shift of upper asymptote leftward (Fig. 5); (ii) changes of n- value with cation concentration are due to relative movements of both the upper and lower asymptotes. The fact that the upper asymptote shifts toward the left with increased concentration of cations while the lower one re- mains almost unchanged (except at high cation concentration) indicates that cations bind to hemoglobin at late stages of oxygenation. These features are consistent with the results obtained for A. marina hemoglobin (12) and P. leptochaeta chlorocruorin (28) and seem to be a general characteristic of extracellular hemoglobins and chlorocruorins of annelids.

Magnetic quadrupole relaxation experiments suggest that Na+ may compete with Ca" for the same site in Lumbricus sp.' hemoglobin (45). The stoichiometry of Ca2+ binding was estimated to be 0.26-0.31 Caz+/heme. These values are very close to our values: 0.33 M$+ and 0.37 Ca2+ ions taken up per oxygen bound (Table 11). Chiancone et al. (20) reported that 1.6 oxygenation-linked Ca2+ ions/heme occur in the pH range 7.7-8.5 and suggested that carboxyl groups with abnormal pK values may be responsible for the Ca2+-binding site for A. ornuta hemoglobin. However, it seems rather unlikely that carboxyl groups themselves would have pK values this high. Makino (46) showed for hemocyanin from Dolabella auricu- laria that H+ and Ca2+ may compete for the same binding site on the basis of measurements of equilibrium dialysis and H+ titration. He also suggested that the Caz+-binding site(s) may include a histidine residue on the basis of the calculated pK value of the protonated site.

Our data (lower panel in Fig. 2) suggest that the maximum number of oxygenation-linked protons apparently depends


on ions such as Na+ or Ca2+ and reciprocally, the protons which are released in the presence of Na+ or Ca" depend on pH. Our observations are consistent with those of Santucci et al. (15) on the hemoglobin of Octolasium cornplanaturn. We suggest that both sets of data may be explained in the following way. We suggest that certain oxygenation-linked acid groups are present which have high pK values (at least 9.5 or higher ). If cations could bind only to the deprotonated group with even a low affinity the fraction of molecules with the acid groups dissociated would be greatly increased at a much lower pH than would otherwise be possible. Assume the following simple sequence for illustrative purposes,

where K, is the acid dissociation constant and K z is the binding constant for the cation. If the apparent dissociation constant in the presence of cations is given by Ki = ([Hb] + [HbNa]) [H+]/ [HbH+] then the apparent pK; will be given by

pK; = pKl - log(1 + Kz[Na+])

If Na+ = 0.3 M and we assume only a very modest binding constant for Na+, say 100, this would be sufficient to lower the pKl value by 1.5 units. The greater effects of Ca" could be explained by a much larger value of K2. Since the log P ~ o values at high and low pH in the presence of sufficient salt differ by about 1.1, the value of the product KlK2 would shift by the same amount because of the linked-function relationship. One cannot tell on the basis of the present data whether values of Kl and K, are both oxygenation dependent. This simple model is probably a considerable oversimplification but suffices to show that cations can lower the pH at which the Bohr effect could be observed. An alternative would be to assume that the cation binds first and that this results in a lowered pK for the acid groups. The net result would be the same.

The data indicate that approximately 0.77 protons are released and 0.37 Ca2+ are taken up per oxygen bound at pH 7.4 almost exactly 2 protons/calcium ion. Similarly, for monovalent ions, 0.52 Na+ ions are taken up and 0.54 protons released per 0 2 bound giving a ratio of 1:l. Although the Bohr effect in the presence of Li+ has not been measured, it is striking that the number of Li+ ions taken up per 0, bound at pH 7.4 is 0.75, close to the value for the Bohr effect in the presence of calcium. These results indicate the presence of two acid groups. Since a Ca2+ ion may be coordinated with up to 8 ligands (47) an attractive possibility is that two salt links of the form, -NH,+ -0OC-, are present. The pK value of the assumed -NH,+ would be raised to over 10 because of the influence of the -COO- group. Ionization of two such groups might lead to a binding site for Ca2+ with at least four groups.

-NH; i "OOC

The Ca2+ could also coordinate with other protein groups such as the carbonyl group of the peptide bond or other amino acid side chains and/or with water molecules (47). This model would explain nicely the observed relationship between the Bohr effect and the binding of both monovalent and divalent cations.

The assembly of extracellular hemoglobins from their constituent subunits is well known to depend on Ca2+. Rokosz and Viogradov (48) reported on the basis of x-ray fluores- cence that the whole molecule of Tubifex tubifex hemoglobin

contains 70 Ca2+ relative to 160 iron atoms. The sample was dialyzed exhaustively against distilled deionized water or against 10 mM EDTA, pH 7, followed by dialysis against water. Chiancone et al. (45) suggested that Ca2+ acts as a cross-linking agent between two carboxyl groups anchored at the interface between two "one-twelfth" subunits. The whole molecule dissociates to %z subunits at alkaline pH, but this dissociation is prevented by Ca2+ ions (49). The whole molecules of some hemoglobins dissociate even at neutral pH upon removing Ca" ions (20, 50). One can, therefore, ask whether "structural" Ca2+ and "functional" Ca" have the same or different binding sites. Chiancone et al. (20) suggested that structural binding sites differ from functional ones on the basis that a concentration of Ca2+ ions only slightly above 1 mM Ca2+ is enough to stabilize the whole structure whereas more than 10 mM Ca" is necessary to effect a change in the functional properties. It would, therefore, seem to be important to determine the level of subunit dissociation at which functional alteration by Ca2' can be observed. If L. terrestris hemoglobin were to have the same number of bound Ca" ions as T. tubifex hemoglobin, 0.44/heme (48), it would mean that deoxygenation would be associated with the dissociation of most of the Ca" since 0.37 Ca2+ ions become bound per oxygen bound or one Ca+ per 3 0,. If this were true then at least 84% of the Ca2+ would be functional and the possibility is raised that uniquely structural nonfunctional Ca2+ may not exist. The lowest concentration of added Ca" which we have used is 0.2 mM. The estimated amount of endogenous bound ca2' is no more than 26 pM (0.44 X 60 pM) or 13% of that added, an amount that appears to be too small to affect our Bohr proton calculations.

Weber and Olsen (51) have recently sought to explain the dependence of oxygen binding on cations by invoking a concept of "surface pH." They argue that the cations do not exert their effect by binding to specific sites but by altering the surface pH. The basis of this conclusion rests on a Gouy- Chapman planar model with a fixed uniform charge density on the surface. However, amino acid sequence and composition data indicate that the net negative charge results from only a small difference between large numbers of both posi- tively and negatively charged groups which have a very non- uniform distribution. This means that any electrostatic effects on ion distribution must be extremely local. Although such effects would give rise to local Debye-Hiickel distributions of ions, the concept of surface pH does not appear physically meaningful. Furthermore, the very high mobility of protons would tend to homogenize this effect. Weber and Olsen (51) excluded binding at specific sites largely on the basis that choline and Na+ appeared to have the same effect even though the two cations differ greatly in size. This could be fortuitous because our data on four different divalent cations (Ca2+, Sr2+, Ba", MP) show that each has a unique effect which cannot be explained on the basis of ionic strength. Rather, the effects appear to depend at least in part on ionic radius. Monovalent ions (Li+, Na+, K+) also have specific effects; the number bound per O2 bound is closely related to ionic radius: the smallest Li+ is bound to the greatest extent, Na+ is next, followed by the largest ion, K'. Although the effects of cations on annelid hemoglobins must involve electrostatic binding, this does not exclude specificity. Detailed studies of pH- dependent processes in many proteins (52) indicate that a primary role is played by electrostatic modifications of pK values and by conformation-dependent changes in hydrogen bonding.

Reversibility of Oxygenation Properties upon pH Change-Giardina et al. (10) concluded that "native" hemoglobin of the earthworm' was in a metastable conformation


which was converted irreversibly to a more stable state when the Hb solution was brought to neutral pH after incubation at either pH 6.0 or 10 for 1 h. The basis of this conclusion was the finding that this treatment appeared to decrease the value of the Hill coefficient irreversibly. Several workers (35- 37) have supported the idea of “a metastable state” by observ- ing such an irreversible decrease in n. This phenomenon was also believed to be supported by circular dichroism (53) and small-angle neutron-scattering experiments (36). However, our results (Fig. 8 and Table 111) show that the reduction in cooperativity upon exposure to acid or alkaline pH is completely reversible. The MetHb index shows that the manipu- lation involved in this experiment such as equilibration with Sephadex G-25 or incubation in dilute solution of HbOz at 25 “C increased the MetHb content compared to that of untreated Hb. In the case of incubation at pH 9:0 without Ca2+, the n value decreased significantly, apparently irreversibly. Such a drop in the MetHb index is serious. Fig. 6 shows that the presence of MetHb reduces the value but has little effect on the P, value. We conclude, therefore, that ”irreversible” decrease in n- at pH 9.0 without Ca2+, as observed in this experiment, was caused by the formation of MetHb. Presumably the MetHb formation would have bees even higher at pH 10. Our results are consistent with the report that Ca2+ prevents dissociation of the whole molecule and also protects against autoxidation (48,53-55).

Martel et al. (36) reported an irreversible decrease in the Hill coefficient for L. terrestris hemoglobin that they explained as resulting from mismatched reassembly upon re- turning the solution to neutral pH. Carbon monoxide was apparently not used as a protective agent during hernoglobin preparation and storage (56), and Ca2* appears not to have been used when hemoglobin was dialyzed against pH 9.0 buffer (36). A significant amount of methemoglobin can form under these conditions, as described above. We believe, therefore, that the apparently irreversible decrease in the value of n may well result from the formation of methemoglobin.

Circular dichroism experiments (53) showed spectral changes in the Soret region that were attributed to a local alteration of the heme environment. This local change, however, does not appear to exclude the possibility of MetHb formation for the following reasons. Harrington et al. (57) reported for L. terrestris hemoglobin that the CD spectrum has maxima at 412 and 422 nm in the oxy and cyanmet forms, respectively. Ascoli et al. (25) reported that the aquomethemoglobin form at pH 7.0 irreversibly changed to a hemichrome after bringing the pH to 6.0 or 7.6. This hemichrome may cause a change in the heme environment which would result in a different CD spectrum. Although small-angle neutron- scattering data (36) show a scattering curve different from that of the parent molecule, the difference is difficult to interpret unambiguously.

In contrast to the irreversible changes reported for “Lum- bricus SP.’’~ Hb (IO), Santucci et al. (15) and Chiancone et al. (20) have reported reversible changes in n with pH for the hemoglobins of both 0. complanatum and A. ornata. These findings together with our data suggest that the existence of a metastable state of ferrous hemoglobin is unlikely.

Temperature Effect-The variation of A H S O with pH is very similar to that of human HbA (Table IV). An enthalpy difference of 3.5 kcal/mol observed at pH 9.0 between L. terrestris hemoglobin and human HbA is larger than those at pH 6.6 (1.9 kcal/mol) and pH 7.4 (1.8 kcal/mol). Studies have shown (27, 58) that the intrinsic heat of heme oxygenation for human hemoglobin is -14.4 to -15.6 kcal/mol of oxygen bound. Variation in the apparent enthalpy appears to result entirely from proton and other nonheme ligands. The appar-

ent enthalpy of oxygenation in Lumbricus hemoglobin will also include contributions from proton reactions. Since Bohr proton release in Lumbricus hemoglobin is coupled to cation binding, the enthalpy of cation binding will also be involved.

Functional Unit-Since the hemoglobin of L. terrestris has about 200 02-binding subunits the task of describing the oxygen equilibrium directly in terms of a 200-step Adair model is clearly hopeless. However, if it could be demonstrated that cooperative oxygen binding involved only a small set of interacting groups, the problem would be easier to handle. Several attempts have been made to estimate the size of such a functional unit in annelid hemoglobins and chlorocruorins. Weber (16) reported that at least five 02-binding sites in A. marina hemoglobin constitute an interacting group. Wyman (29) reported at least 5 interacting sites for Spirographis spallanzanii chlorocruorin and Colosimo et al. (59) reported 10 sites for the same chlorocruorin. Imai and Yoshikawa (28) reported 6 interacting sites for P. leptochaeta chlorocruorin. All of these estimates have been based on the MWC model and/or deductions based on the Hill coefficient. We have also attempted to estimate the size of the functional unit on these bases. A plot of oxygen equilibrium data in terms of a linearized MWC equation (Equation 6) is shown in Fig. 9. The value of X in the figure is calculated from the relation, X = (1 + KTP)/(l + KRP). Although the data are nonlinear, the maximum slope in the central part was used to provide the values shown in Table I. These data show that the apparent number of interacting sites appears to vary between 5 and 12, depending on pH. Maximum values were obtained between pH 7.8 and 8.2. This corresponds to the pH range which gave the maximum values of n. Similar estimates of the number of interacting sites were obtained with Kegeles’ method (32) and are listed in Table I together with y,,,. The data suggest that the size of the functional unit depends not only on the pH but also on the kind and concentration of salt. Although the calculations give plausible numbers, all of which are very small compared to the total number of hemes/molecule, they are based on a two-state MWC model. We have shown, however, that a two-state model is not sufficient for the description of the data. The nonlinearity of the curves in Fig. 9 emphasizes this point. An alternative procedure would be to adopt a completely statistical approach to binding. More generally, without recourse to any model of hemoglobin be- havior, nmax can be shown to be proportional to the statistical

-2t i - -2 - I 0

log x FIG. 9. Plots of oxygen binding by L. terrestris hemoglobin

according to the linearized MWC equation (Equation 6 in text). Kl and K,,, (Table I) were used for KT and KB, respectively. Buffer: 0.05 M BisTris/propane/HCl, 0.1 M NaC1. pH from left to right: 8.10, 7.35,8.86, 6.55.


variance of the distribution of bound ligand among sites (60, 61), and so a high Hill coefficient reflects a highly nonrandom distribution of bound ligands among possible sites. At least four intrinsic 02-binding equilibria must be involved because 4 major kinds of subunits with unique 02-binding character- istics have been isolated (1). One can hope to establish a better basis upon which to describe the oxygen equilibria by obtaining detailed data on the subunits'in various states of assembly and so to obtain a description of the whole molecule in terms of its 4 major constituent subunits and their meas- urable interactions. This procedure will also help ascertain whether the minor chains V and VI (Ref. 6) play any important functional or structural role.

Acknowledgments-We wish to express our thanks to Dr. I. Tyuma of the Medical School at Osaka University and Dr. H. Morimoto, Faculty of Engineering Science, Osaka University for their interest and encouragement. We thank Dr. Walter J. Harman (Dept. of Zoology, Louisiana State University) for the identification of the worms.

REFERENCES 1. Fushitani, K., Imai, K., and Riggs, A. F. (1985) in Symposium on

Invertebrate Oxygen Carriers (Linzen, B., ed) Springer Verlag, New York, in press

2. Imai, K., Yoshikawa, S., Fushitani, K., Takizawa, H., Handa, T., and Kihara, H. (1985) in Symposium on Invertebrate Oxygen Carriers (Linzen, B., ed) Springer Verlag, New York, in press

3. Chung, M. C. M., and Ellerton, H. D. (1979) Prog. Biophys. Mol. Biol. 35,53-102

4. Kapp, 0. H., Vinogradov, S. N., Ohtsuki, M., and Crewe, A. V. (1982) Biochim. Biophys. Acta 704,546-548

5. Garlick, R. L., and Riggs, A. F. (1982) J. Biol. Chem. 257,9005- 9015

6. Kapp, 0. H., Polidor, G., Mainwaring, M. G., Crewe, A. V., and Vinogradov, S. N. (1984) J. Biol. Chem. 259,628-639

7. Haughton, T. M., Kerkut, G. A., and Munday, K. A. (1958) J.

8. Manwell, C. (1959) J. Cell. Comp. Physiol. 53, 61-74 9. Cosgrove, W. B., and Schwartz, J. B. (1965) Physiol. 2001. 38,

10. Giardina, B., Chiancone, E., and Antonini, E. (1975) J. Mol. Biol.

11. Vinogradov, S. N., Shlom, J. M., Hall, B. C., Kapp, 0. H., and

12. Weber, R. E. (1981) Nature 292,386-387 13. Imai, K. and Yonetani, T. (1975) J. Biol. Chem. 250,2227-2231 14. Imaizumi, K., Imai, K., and Tyuma, I. (1979) J. Biochem. (Tokyo)

15. Santucci, R., Chiancone, E., and Giardina, B. (1984) J. Mol. Biol.

16. Weber, R. E. (1970) Comp. Biochem. Physiol. 35,179-189 17. Terwilliger, R. C. (1974) C o w . Bwchem. Physiol. 48A, 745-755 18. Garlick, R. L., and Terwilliger, R. C. (1975) Comp. Biochem.

19. Weber, R. E. (1975) J. Comp. Physwl. 9 9 , 297-307 20. Chiancone, E., Femzzi, G., Bonaventura, C., and Bonaventura,

J. (1981) Biochim. Biophys. Acta 670,84-92 21. Chung, M. C. M., and Ellerton, H. D. (1982) Biochim. Biophys.

Acta 702,6-16 22. Ochiai, T. (1983) Arch. Biochem. Biophys. 226, 111-117 23. Ochiai, T. (1984) Arch. Biochem. Biophys. 231 , 136-143 24. Imai. K. (1981) Methods Enzvmol. 76.438-449

EXP. BWl, 35,360-368

206-212

93 , l -10

Mizukami, H. (1977) Biochim. Biophys. Acta 492 ,136155

86,1829-1840

179,713-727

Physiol. 5 1 A, 849-857

26. Antonini, E., and Brunori, M. (1971) in Hemoglobin and Myoglo- bin in the Reactions with Ligands, North-Holland Publishing Co., Amsterdam

27. Imai, K. (1979) J. Mol. Biol. 133,233-247 28. Imai, K., and Yoshikawa, S. (1985) Eur. J. Biochem. 147, 453-

29. Wyman, J. (1964) Ado. Protein Chem. 19, 223-286 30. Monod, J., Wyman, J., and Changuex, J. P. (1965) J. Mol. Biol.

31. Decker, I., Sabel, A., Linzen, B., and Van Holde, K. E. (1983) in Life Chemistry Reports (Wood, E. J., ed) Suppl. 1, pp. 251-256, Harwood Academic Publishers, London

463

12,88-118

32. Kegeles, G. (1979) FEBS Lett. 103,5-6 33. Imai, K. (1982) Allosteric Effects in Haemoglobin, pp. 113-114,

34. Imai, K., Yonetani, T., and Ikeda-Saito, M. (1977) J'? Mol. Biol.

35. Chiancone, E., Ascoli, F., Giardina, B., Vecchini, P., Antonini, E., Musmeci, M. T., Cinb, R., Zagra, M., D'Amelio, V., and De Leo, G. (1977) Biochim. Biophys. Acta 494 , 1-8

36. Martel, P., Powell, B. M., Kapp, 0. H., and Vinogradov, S. N. (1982) Biochim. Biophys. Acta 709,134-141

37. Frossard, P. (1982) Biochim. Biophys. Acta 704,524-534 38. Imai, K., and Yonetani, T. (1975) J. Biol. Chem. 250,7093-7098 39. Wood, E. J., Mosby, L. J., and Robinson, M. S. (1976) Biochem.

40. Bannister, J. V., Bannister, W. H., Anastasi, A., and Wood, E. J.

41. Garlick, R. C., and Terwilliger, R. C. (1977) Comp. Biochem.

42. Weber, R. E., Bonaventura, J., Sullivan, B., and Bonaventura, C.

43. Warren, L. M., Wells, R. M. G., and Weber, R. E. (1981) J. Exp.

44. Chung, M. C. M., and Ellerton, H. D. (1982) Biochim. Bwphys.

45. Chiancone, E., Bull, T. E., Norne, J. E., Forsen, S., and Antonini,

46. Makino, N. (1972) J. Biochem. (Tokyo) 71,987-991 47. Williams, R. J. P. (1976) Symp. Soc. Exp. Bwl. 30 , 1-17 48. Rokosz, M. J., and Vinogradov, S. N. (1982) Bwchim. Biophys.

49. David, M. M., and Daniel, E. (1974) J. Mol. Biol. 8 7 , 89-101 50. Terwilliger, R. C., Terwilliger, N. B., and Roxby, R. (1975) Comp.

51. Weber, R. E., and Olsen, L. F. (1984) Mol. Physiol. 6 , 1-8 52. Matthew, J. B., Gurd, F. R. N., Garcia-Moreno, E., Flanagan, M.

A., March, K. L., and Shire, S. J. (1985) CRC Crit. Rev. Bwchem.

53. Ascoli, F., Chiancone, E., and Antonini, E. (1976) J. Mol. Biol.

54. Swaney, J. B., and Klotz, I. M. (1971) Arch. Bwchem. Biophys.

55. Chiancone, E., Brenowitz, M., Ascoli, F., Bonaventura, C., and

56. Shlom, J. M., and Vinogradov, S. N. (1973) J. Biol. Chem. 248 ,

57. Harrington, J. P., Pandolfelli, E. R., and Herskovits, T. T. (1973)

58. Mills, F. C., Ackers, G. K., Gaud, H. T., and Gill, S. J. (1979) J.

59. Colosimo, A., Brunori, M., and Wyman, J. (1974) Biophys. Chem.

60. Cohn. E. J.. and Edsall. J. T. (1943) Proteins. Amino Acids and

Cambridge University Press, London

109,83-97

J. 153,589-596

(1976) Biochem. J. 159,35-42

Physiol. 57B, 177-184

(1978) J. Comp. Physiol. 123,177-184

Mar. Bwl. Ecol. 5 5 , 11-24

Acta 702,17-22

E. (1976) J. Mol. Biol. 107,25-34

Acta 707,291-293

Bwchem. Physiol. 50B, 225-232

18391-197

105,343-351

147,475-486

Bonaventura, J. (1980) Biochim. Biophys. Acta 623,146-162

7904-7912

Biochim. Bwphys. Acta 328 , 61-73

Biol. Chem 254,2875-2880

2,338-344

Peptides & Ions and Dipolarons,'pp. 462-463, Reinhold Pub-

25. Ascoli, F., Rossi-Fanelli, M. R., Chiancone, E., Vecchini, P., and 61. Edsall, J. T., and Gutfreund, H. (1983) Biothermodynumics, p. . . lishing Corp., New York

Antonini, E. (1978) J. Mol. Biol. 119, 191-202 201, John Wiley and Sons, New York

J. biol. chem. 1986-fushitani-8414-23

Documents

Transcript of J. biol. chem. 1986-fushitani-8414-23