Biochimica et Biophysica Acta, 427 (1976) 57-69

© Elsevier Scientific Publishing Company, Amsterdam -- Printed in The Netherlands BBA 37276 S U R F A C E ACTIVITY OF H E M O G L O B I N S A N D O T H E R H U M A N HEMOG L O B I N VARIANTS

DANEK ELBAUM, JOHN HARRINGTON, EUGENE F. ROTH, Jr. and RONALD L. NAGEL Department of Medicine, Division of Hematology, Albert Einstein College of Medicine, Bronx, N.Y. 10461 (U.S.A.)

(Received July 10th, 1975)

SUMMARY The kinetics of surface pressure change (A~ vs. t isotherms) were determined for several single point mutations of the human hemoglobin system. It was observed that hemoglobin S and hemoglobin CHar~em (both containing r6 Glu -~- Val substitutions) have a specific behavior at the water-air interface: their extent of surface pressure change is larger than for hemoglobin A, hemoglobin C and hemoglobin Korle Bu @73 Asp ~ Asn). In addition, hemoglobin S seems to occupy a larger area per molecule than hemoglobin A. The conformational requirements for this property, in addition to the r 6 Val substitution, appear to be the liganded state of the fls chain in the tetramer. Electrostatic, hydrogen bonding and hydrophobic interactions are involved in determining the surface activity of a hemoglobin molecule. The differences between the surface activity o f oxyhemoglobin S and oxyhemoglobin A could be the basis for their differences in mechanical precipitability, although other factors may play a role.

INTRODUCTION Since the elucidation of the primary structure of human hemoglobins, attention has moved to aspects o f tertiary and quaternary structure in this tetrameric protein. In particular, the capacity of hemoglobin S to polymerize has become of special interest, and investigations into the properties of this protein have multiplied in recent years. The following studies on the surface activity of different human hemoglobin mutants contribute to the description of new properties added to the molecule by single point mutations. Interest in the surface activity of proteins began in the last century and a comprehensive review of the early work in this field has been published by Neurath and Bull [1]. More recent invescigations have been discussed by James and Augenstein [2]. Wu and Ling [3] were the first to examine hemoglobin in terms of its surface activity. They studied the coagulation of purified oxy- and methemoglobin solutions induced by shaking. Jonxis [4] studied the spreading of different animal hemoglobins at air-wa~er

58 interfaces. His main conclusion was that the time of unfolding for different animal hemoglobins is different and that the ligand state of the hemoglobin has a marked influence on the rate of spreading. More recently, the findings of Asakura et al. [5] and Roth et al. [6], have shown that oxyhemoglobin S has an increased mechanical precipitability as compared with oxyhemoglobin A. It seems, then, of interest to examine the surface activity of different human hemoglobin variants and to undertake an inquiry of the structural basis of this property. The possible relation of surface activity to the mechanical precipitability of hemoglobin S will also be discussed. A preliminary account of some of the work presented here has been recently published [7]. MATERIALS AND METHODS Hemoglobin preparations. All experiments were performed on hemoglobin A solutions prepared on the same day according to the method of Drabkin [8]. Hemoglobin S was prepared from the blood of homozygous donors and used without further purification (93 ~ pure) within 2 days, or stored in liquid nitrogen until used. Hemoglobin C was prepared from the blood of a homozygous donor by the same procedure and used immediately without further purification (95 ~ pure). Hemoglobin Cnar~em and hemoglobin Korle Bu were isolated chromatographically on DE-52 (Whatman) cellulose developed with a 0.05 M Tris-HC1 buffer, pH 8.0-8.2 at 4 °C. The hemoglobin Korle Bu heterozygous sample was shipped to us in an iced container by Dr. Graham Serjeant and used within 1 week of its being drawn. The preparation of isolated ~ and /3 chains were carried out as described by Geraci et al. [9] by the blocking o f - S H groups using p-hydroxymercuribenzoate (pOH-HgBzO-). Mixed liganded hybrids were prepared as previously described by Bookchin and Nagel [10]. All hemoglobin solutions were less than 24-h-old hemolysates, unless indicated in the text. The carbonyl derivative of hemoglobin S was prepared by passing CO gas through a solution of hemoglobin until the spectrophotometric properties were constant. The cyanmet derivative of hemoglobin S was prepared by addition of 5-fold excess of K3Fe(CN)6 and K C N followed by passage on a Sephadex G-25 column to remove the excess oxidizing agent. Az~ vs. t isotherms. Protein films at air-water interface were studied by following the rate of adsorption or film formation after injection of protein solution into the subphase. Surface tension measurements were made by means of the Wilhelmy plate technique as described by Colacicco and Rapport [11]. The change in surface pressure (A~ ~ ~-70) corresponds to the difference between the surface tension in the presence of the protein film (7) and in its absence (7o). This term was preferred over the conventional ~r (er = 70 -- 7) because it reflects better the actual observed phenomena: the decrease of surface tension by protein film formation. A platinum blade, 1 cm length, suspended from a tension balance (RG Automatic Electrobalance, Cahn Instrument Corp.) was employed. The platinum blade was flamed and the surface of the aqueous phase was cleaned with a capillary tube attached to a suction pump before every measurement. The alterations of surface tension values were monitored

59 on a chart recorder as a function of time. Measurements were made in a circular glass dish of a constant surface area of 22.42 cm 2, after injecting 10/zl of a 3.2 g/100 ml hemoglobin solution into a subphase of 30 ml of 0.15 M potassium phosphate buffer, pH 7.35, in the presence of constant and gentle mechanical stirring. All measurements were obtained at 22 4- 1 °C unless indicated in the text. The measurements were arbitrarily stopped at 60 min due to concern over the nativeness of hemoglobin solutions kept for longer times at room temperature. In most instances, An changes at 60 rain were less than 1 dyne/cm per 10 min. These apparent equilibria values were reproducible within 1.5 dynes/cm. Dynamic surface pressure (Az~ vs. A) curves. Droplets were delivered from the tip of a Hamilton microsyringe by touching the interface. After 30 min, needed for equilibration and formation of the film on the surface, compression of the protein films between 67.56 and 13.51 cm z was accomplished using a moveable barrier monitored with a two-channel recorder. All measurements were performed at room temperature, 22 ± 1 °C. Area per molecule determination with labelled hemoglobin. Hemoglobin A and S were labelled by carbamylation with [14C]cyanate as described by Jansen et al. [12]. Under the conditions used, the label corresponds almost exclusively to the carbamylation of a-amino group in both a and/3 chains. The amount of protein present in the air-water interface was determined by comparing the number of counts in the subphase at time 0 with the subphase concentration after the surface film had reached apparent equilibrium. Deuterium exchange studies. Hemoglobin S solutions were exchanged by 24-28 h dialysis against 2H20 in an Amicon Model 8MC microultrafiltration system equipped with a type UM-10 membrane (lot No. 397, Amicon Corp., Lexington, Mass.). This deuterated hemoglobin S was injected into a subphase of 2HzO (New England Nuclear Corp.). As controls, non-exchanged hemoglobin S was studied in a pure ZH20phase and in a pure water phase. RESULTS Surface activity at the air-water interface seems to be dependent upon the primary structure of hemoglobin and the single amino acid substitution of hemoglobin S (f16 Glu -+ Val) confers on the oxy form of this protein the ability to be absorbed into the interface faster than oxy Hb A, Hb C (f16 G l u - + Lys) and Hb Korle Bu (fl73 Asp -+ Asn) and to reach a lower Az~ at apparent equilibrium (Fig. 1). Similar behavior is found for Hb Cnarlem where the r6 (Glu --~ Val) substitution is accompanied by a second substitution at/373 (Asp --~ Asn) [13]. We have presented the results of two hemoglobin A preparations in Fig. 1. One, hemoglobin A corresponds to freshly prepared normal hemolyzate. Hemoglobin A (c) is the corresponding control to the chromatographed hemoglobin Korle Bu. While both exhibit slower initial kinetics than fresh hemoglobin A, the apparent equilibria values were indistinguishable for these three hemoglobins. To determine if this difference in surface activity between hemoglobin S and A is dependent on the tetramer or if the protein is present at lower structural levels, we studied the surface activity of isolated a A, /3A and/3s chains, treated with p-hydroxymercuribenzoate (Fig. 2). While all isolated chains exhibited faster kinetics than

60

-4 -6

-B

-10 -12 -14

-16 -18

-20 -22

BUk)

-24

)xy Hb C

-26 Oxy HbA(c] Oxy HbA

-28 -30

Oxy Hb C HARLEM(c)

-32 -34 0

10

20

30

40

50

60

TIME (rain)

Fig. I. Pressure vs. time isotherm of various h u m a n h e m o g l o b i n mutants. Proteins were injected into a subphase o f 0.15 M potassium phosphate buffer, p H 7.35, at 22 i 1 °C. Total hemoglobin concentration, 10#g/ml. (c), isolated by chromatography. -

0

~,-14 9-18 -22 -24 -28 -5C

x l x ~~ , _ _ .•

I

I I0

I

I 20

I

I

x'x=A-PUa

I

I

30 40 TIME (min)

I

I 50

I

I 60

Fig. 2. Effect of isolated, p O H - H g B z O - - t r e a t e d cd', ilk, flS, chains on the kinetics o f protein film formation (dzt vs. t isotherms). All measurements were done in 0.15 M potassium phosphate buffer, p H 7.35, at 22 :k 1 °C. Total hemoglobin concentration, 10 #g/ml.

61 hemoglobin tetramers, the oxy fls-S-HgBzO- chain is considerably faster and reaches equilibrium at a lower surface pressure than flA-S-HgBzO-. The next question is how the quaternary conformational states, R (oxy) and T (deoxy) influence film formation. In Fig. 3, it can be observed that the rate of deoxyhemoglobin S film formation is much slower and reaches apparent equilibrium at a lower surface pressure (--19 dynes/cm) than oxyhemoglobin S (--28 dynes/cm), Tiffs

-8

-10

"~u -12 ~ -14

-

2

8

~

-30 0

10

20

30

40

50

60

TIME (Minutes)

Fig. 3. Effect of the ligand state of the hemoglobin on the surface activity (Azt vs. t isotherm) of hemoglobin S and A. All subphase solutions contained 0.15 M potassium phosphate buffer, pH 7.35. Temperature, 22 + 1 °C. Total hemoglobin concentration, 10/zg/ml. difference is significantly greater than that found between deoxyhemoglobin A (--22 dynes/cm) and oxyhemoglobin A (--24 dynes/cm). The effect of ligands within the R conformational state of hemoglobin S on the kinetics of surface film formation showed no pronounced differences. Methemoglobin S and CO hemoglobin S are indistinguishable from oxyhemoglobin S (Fig. 4). The question of the conformation requirements for increased surface activity o f hemoglobin S was approached by the study of mixed liganded hybrids. The forms * 2S and azfl 2*S (in which the * chain is fixed in the cyanmet form) were:studied both a2/~ in deoxy and oxy state. As it can be seen in Fig. 5, both oxy hybrids have high surface activity. The deoxy a 2* f12S form, in contrast, has a low surface activity comparable to deoxyhemoglobin S. We turn now to the effect of changes in the solvent on the surface activity of

62

_i -10

-~ .u -12

~ -16 o

-18 -20 -22 -24 -26 10

20

30 TIME (rain)

40

50

60

Fig. 4. Pressure vs. time isotherm of hemoglobin S combined with different ligands. All subphase solutions contained 0.15 M potassium phosphate buffer, pH 7.35. Temperature, 22 + 1 °C. Total hemoglobin concentration, 10/~g/ml. Hemoglobin solution was kept for 48 h at 4 °C before injection into the subphase.

hemoglobin S as compared to hemoglobin A. In Fig. 6, the effect of temperature is depicted on the rate of film formation comparing oxyhemoglobin A and oxyhemoglobin S at 7.6 and 23 °C. Low temperature (7.6 °C) reduces the surface activity of oxyhemoglobin S from (--28 dynes/cm to almost --21 dynes/cm). Hemoglobin A is also reduced in its apparent equilibrium An but only from --23 to --21 dynes/cm. The electrostatic interactions involved in stabilizing the protein at the air-water interface are altered by the addition of salt. The surface activity of oxyhemoglobin S is enhanced in the presence of increasing concentrations of NaC1 (Fig. 7). The opposite phenomenon occurs when ZH20 is exchanged for water (Fig. 8). Dialysis of ~Hexchanged oxyhemoglobin S against 2H20 resulted in decreased surface activity (--20 dynes/cm at 50 min) compared to the activity of oxyhemoglobin S in water or in 2H20 (--25 dynes/cm at 50 rain). As the rate of change of surface pressure (Az0 is influenced by the extent of unfolding of the protein during the film formation, the area occupied per molecule on the surface for the oxy forms of hemoglobin A, and S, were determined. Surface pressure vs. area curves (Fig. 9) were obtained from compression of the protein monolayer. The area per molecule is obtained by extrapolating the approximately linear portion of the force vs. area curve to zero pressure [4]. Molecules of hemoglobin S were found to occupy a larger area (8000/X,Z/molecule) than hemoglobin A (5000 Az/ molecule) at the air-water interface. This implies a greater or different degree of unfolding of oxyhemoglobin S. Utilizing [14C]cyanate for the labeling of the NHzterminal amino groups of the a and fl chains of these hemoglobins similar results

63

-6

-8 -10 -12

-14 -16 ~. - 1 8 t::: - 2 0