David Grant Medical Center, Travis Air Force Base, California. Eugene ... The developer was 0.05 M Tris buffer, ... by using the same developer to which 0.5 M NaCl ..... G. L. Haycraft, David Grant Medical Center (MAC), Travis Air Force Base.
Eur. J. Biochem. 77, 561 - 566 (1977)
Hemoglobin S Travis : a Sickling Hemoglobin with Two Amino Acid Substitutions [p6(A3)Glutamic Acid and p 142(H20) Alanine -+ Valine]
-+
Valine
Winston F. MOO-PENN, Robert M. SCHMIDT, Danny L. JUE, Katherine C. BECHTEL, and Jane M. WRIGHT Hematology Division, Center for Disease Control, Public Health Service, Department of Health, Education, and Welfare, Atlanta, Georgia McDonald K. HORNE, 111 and Gordon L. HAYCRAFT David Grant Medical Center, Travis Air Force Base, California Eugene F. ROTH, Jr and Ronald L. NAGEL Division of Hematology, Department of Medicine, Albert Einstein College of Medicine, Bronx, New York (Received March 3, 1977)
Hb S Travis is a previously undescribed sickling hemoglobin with two amino acid substitutions in the p chain: p6 Glu-Val and p142 Ala-tVal. The p6 Glu-Val mutation imparts to H b S Travis the characteristic properties of sickling hemoglobin, namely its association with erythrocyte sickling, the insolubility of the hemoglobin in the reduced form, and a minimum gelling concentration value identical to Hb S. Unlike Hb S, Hb S Travis exhibits an increased oxygen affinity and a decreased affinity for 2,3-bisphosphoglycerate and inositol hexakisphosphate. In addition, the variant hemoglobin’s tendency to autoxidize and its mechanical precipitability suggest that there are conformational differences between Hb S and Hb S Travis. Most common human hemoglobin variants arise from a single mutation with the substitution of one amino acid residue for another. Rarer mutational events resulting in deletion, insertion or addition of amino acids, or fusion of parts of different globin chains have also been documented [l]. In addition, two different substitutions in the chain have been described for Hb C Harlem [2], Hb C Ziguinchor [3], and Hb Arlington-Park [4]. These mutants could have been derived either from a second mutation or from homologous crossing over between genes, each carrying one of the mutations. The appropriate fl chain substitutions (S, Korle Bu, Dhofar, and N Baltimore [l]) occur at relatively high frequency in several black populations and crossing over to produce a fl chain with two substitutions could have occurred from random matings. In this report we describe the structural characterization and some of the functional properties of Hb S Travis, a variant with both the sickle mutation at p6 and a new substitution, p142(H20) Ala+Val. MATERIALS AND METHODS Electrophoresis and other methods were used for detecting abnormal hemoglobins [5]. Hemoglobins Abbreviation. P2-glycerate, 2,3-bisphosphoglycerate. Enzyme. Trypsin (EC 3.4.21.4).
(A, S, C Harlem and S Travis) in this study were purified by column chromatography on both DEAESephadex [6] and DEAE-cellulose (Whatman) ion exchangers. In the procedure with DEAE-cellulose exchanger a special Chromo-Flex column (Kontes, Vineland, N. J.), which has two separable sections, was employed. The developer was 0.05 M Tris buffer, pH 8.1. The rate of elution was 1 ml/min and the entire separation took about 3 h. When Hb S Travis had migrated to the lower section, the column was separated and the top (containing Hb A) and bottom (containing Hb S Travis) sections were each eluted by using the same developer to which 0.5 M NaCl had been added. This procedure allowed collection of a concentrated solution of hemoglobin (generally over 1 g/dl) and avoided possible denaturation due to dilution, which is generally encountered with more classical chromatographic methods. After heme was removed [7] the chains were separated by the method of Clegg et al. [8]. The abnormal chain was aminoethylated [9] and digested with trypsin [lo] treated with L-l-tosylamido-2-phenylethyl chloromethyl ketone for 4 h at 37°C. The peptides were resolved by ion-exchange chromatography on a column (0.9 x 23 cm) of Aminex A-5 resin (Bio-Rad Lab.) using pyridine acetate buffers [ll]. Amino acid compositions of the peptides were determined on a Beckman Model 121 amino acid
562
Hb S Travis: A Sickling Hemoglobin with Two Substitutions in the /j Chain
analyzer after hydrolysis for 24 h in 6 M HCl in YUCUO [12]. Sequence data were obtained with a Beckman model 890C sequencer by using the programs (no. 122947 and 102974) provided by the manufacturer [13]. Before sequencing, the peptide was reacted with 4-sulfophenylisothiocyanateto reduce the solubility of the peptide in organic solvents [14]. The phenythiohydantoin derivatives were identified by gas [15] and high-pressure liquid chromatography [16]. The latter procedure permitted accurate quantification of lysine, histidine, arginine, glutamine and asparagine. Minimum gelling concentrations were determined for purified Hb S Travis and for mixtures of Hb S Travis and Hb A by the method described elsewhere [17]. Autoxidation studies were carried out by incubating the purified hemoglobin solutions in a tonometer immersed in a 37 "C thermoregulated bath. No agitation of the tonometer was allowed (for fear of inducing mechanical denaturation of Hb S or Hb S Travis), but the volume used and the shape of the tonometer permitted rapid and constant temperature regulation. The percentage of methemoglobin formation was calculated by observing the changes at 576 nm. For purposes of calculation, 100 % methemoglobin formation was obtained experimentally by converting an aliquot of the sample to the ferric state with ferricyanide. All experiments were conducted in 0.04 M potassium phosphate buffer, pH 7.1. Mechanical instability of Hbs A, S, and S Travis, isolated by column chromatography, was studied by a method described elsewhere [18]. All supernatant concentrations were determined after converting the solutions to cyanmethemoglobin by using Drabkin's reagent. The method used for studying the mechanical instability of deoxygenated samples has also been previously described [19]. Heat-denaturation studies were done by incubating a 1 g/dl solution of hemoglobin in 0.1 M NaCl, pH 7.0, at 65 "C. Samples were removed at various time points, pipetted into ice-chilled centrifuge tubes and centrifuged at 31 000 x g for 25 min. The supernatant was removed and the absorbance measured at 577 nm to determine the percentage of undenatured hemoglobin. Oxygen equilibria studies were done by the spectrophotometric technique of Riggs and Wolbach [20]. Hemoglobins were separated at 4 "C and stripped of organic phosphates by passage through a column of mixed-bed resin (AG 501- x 8, Bio-Rad Lab.). Deoxygenation was carried out by alternately evacuating and flushing the tonometers with nitrogen. P2-Glycerate and inositol hexakisphosphate were prepared as previously reported [21]. Hb A purified from the same column as Hb S Travis was used for all control functional studies. Hematological data and clinical findings will be reported elsewhere.
Trypsin was purchased from Worthington Biochemical Corporation, and P2-glycerate and inositol hexakisphosphate were obtained from Sigma. All other chemicals were of analytical grade.
RESULTS Hb S Travis was detected by electrophoresis on cellulose acetate (pH 8.4) and citrate agar (pH 6.2). On cellulose acetate, bands corresponding to Hbs A and S were present, but the amount in the Hb S position appeared less than would be expected for sickle cell trait. Electrophoresis on citrate agar also revealed two bands, one in the position of Hb A and another minor band with a mobility between Hbs A and S. Visual examination of the abnormal hemoglobin bands on both support media indicated approximately equivalent percentages. The solubility test [22] for sickling hemoglobin was positive, and globin chain electrophoresis in acidic and alkaline buffers revealed an abnormal p chain in the PS position. The pedigree of this family is shown in Fig. 1. The family study indicates that the propositus' mother is the primary carrier of Hb S Travis, which is present in her progeny from two different matings. The five carriers are healthy and completely asymptomatic. Fig. 2 shows the chromatographic separation of Hb S Travis on a DEAE-Sephadex column. The percentages of hemoglobins from this separation were: Hb A2 3.8"/I, Hb S Travis 14.1 %, and Hb A 82.1 The peak trailing Hb A was not analyzed but probably represented Hb A1 and some degradation products [6]. The value of Hb F, as determined by the alkali denaturation test, was 0.5 %. The mobility of the abnormal P chain, isolated from purified Hb S Travis by the Clegg procedure, was identical to that of Hb S. Peptide fingerprinting produced a normal map, except that PTp I migrated in the position of the PS peptide (since the second substitution, P142 Ala+Val involved neutral amino acids this result would be expected). Fig. 3 shows the elution chromatogram of a tryptic digest of the abnormal fi chain. Tryptic peptide I eluted slower than the normal peptide but still chromatographed between /ITp I1 and PTp XIV-XV. The amino acid composition of abnormal BTp I isolated by column chromatography and by elution from the peptide map indicated a substitution of valine for glutamic acid (Table 1). Amino acid analyses of all other peptides were normal except for the loss of one residue of alanine with a corresponding gain of one residue of valine in PTp XIV-XV (Table 1). This substitution did not alter the rate of elution of the combined peptide from the ion exchanger. Results of sequence analyses of intact p chain and PTp XIV-XV from Hb S Travis are given in Table 2.
x.
563
W. F. Moo-Penn et al.
I
II
93h 5
I
0Male, not studied 0Female
0 Hb S Travis trait @HbA
\ Propositus Fig. 1. Pedigree offamily w5ith Hh S Travis
The data confirm the j 6 Glu+Val substitution and indicate that alanine at position 142 is replaced by valine. The minimum gelling concentration of purified Hb S Travis was found to be 25.1 g/dl, a value not significantly different from that of 24.3 g/dl for a Hb S sample examined simultaneously. A mixture of 40 % Hb S Travis and 60 % Hb A yielded a minimum gelling concentration of 32.5 g/dl, a value indistinguishable from the minimum gelling concentration of 32 g/dl obtained for the same proportion of Hb S and H b A. The study of the rate of autoxidation at 37 "C and pH 7.1 demonstrated clear differences between H b S Travis, on the one hand, and Hb S and H b C Harlem on the other. As seen in Fig.4, purified samples of Hbs S and C Harlem have an autoxidation rate about
Table 1. Amino acid composition of OTp I and OTp X I V - XV from Hb S Travis Peptides were isolated by chromatography on a column (0.9 x 23 cm) of Aminex A-5. The /ITp I and /ITp XIV-XV peptides were hydrolyzed in vacua at 110 "C in 6 M HC1 for 24 and 48 h respectively. The results are expressed as molar ratios Amino acid
Lysine Histidine Aspartic acid" Threonine Glutamic acid Proline G1ycin e Alanine Valine Leucine Tyrosine
Effluent volume (ml)
Fig.2. Chromatographic separation o j Hh S Travis on DEALSephadex. Approximately 3 g hemoglobin were applied to a column (2.5 x 60 cm) equilibrated with 0.05 M Tris buffer (pH 8.4). The various hemoglobins were eluted with a pH gradient of the same buffer (pH 8.4-6.4). Fractions were collected at a flow rate of 20 ml/h
1
0
'
8
20
,
,
40
.
,
60
.
I
80
.
I
100
.
PTP 1
/ITp XIV
~
XV
cxpected
found
expected
found
1 1
1 .0 0.8
1 2 1
1.3 1.7 1.0
1 2 1
1.0 1.0 0.8
1 4 3 1 1
1.2 3.1 3.7 1.2 0.7
1.7 1.4
1 1
Amides are converted to the acid during hydrolysis
,
.
,
120 140 Tube number
.
,
160
.
,
180
.
I
200
i
.
2;O
240
Fig. 3. ChromatoRruphic separation of an aminoethylated tryptic digest of'the/I chainfrom Hh S Travis. A linear gradient of pyridine acetate (0.2 M, pH 3.1 -2.0 M, pH 5.0) was used to elute the peptides from a column (0.9 x 23 cm) of Aminex A-5 resin (Bio-Rad Lab.) at 52 "C and at a flow rate of 30 ml/h. 10 % of the effluent was removed and reacted with ninhydrin to allow for automated monitoring at 570 nm
564
Hb S Travis: A Sickling Hemoglobin with Two Substitutions in the
fi Chain
Table 2. Sequence analysis of Hb S Travis Initial samples were 700 nmol chain and 600 nmol PTp XIV-XV. All phenylthiohydantoin derivatives were identified by gas chromatography unless otherwise stated. Repetitive yield between valine at positions 1 and 6 is 98.7 %. Repetitive yield between valine at positions 134 and 142 is 89.5 "/, 86 Glu-tVal
b142 Ala-Val
sequence no.
cycle no.
identification
1
1
2 3 4 5 6 7
2 3 4 5 6 7
Val His" Leu Thr Pro Val Glu
amount
sequence no.
cycle no.
identification
nmol
nmol
a
281.9 123.0 436.2 168.3 170.5 264.4 191.5
amount
Val' Val Ala GlY Val Ala Asn a Ala Leu Val His"
1 2 3 4 5 6 7 8 9 10 11
133 134 135 136 137 138 139 140 141 142 143
-
530.0 343.7 287.7 366.6 266.3 202.7 205.9 178.5 218.8 3 64.8
Quantified by high-pressure liquid chromatography. Quantified by on-column silylation. Modified with 4-sulfophenylisothiocyanateand not determined.
\ 0 Time (h)
2
4
6
a
10
Time (min)
Fig.4. Rates of autoxidation at 37 " Cfor Hb S Travis (@), Hb S (A) and Hb C Harlem (W). (0, A , 0 ) Correspond to values for Hb A isolated simultaneously from the hemolysate of the heterozygous carriers of these mutants. Phosphate buffer, 0.04 M, pH 7.1
Fig. 5. Mechanical precipitability of column-purified Hb A (O), S ( A ) and S Travis (@) in the oxyform. Aliquots of each hemoglobin in 0.15 M phosphate buffer, pH 7.35, were agitated at 24-26 "C for 2-min periods'and the amount of hemoglobin remaining in the supernatant was &ermined as cyanmethemoglobin. The initial concentration for all hemoglobins was 200 mgjdl
double that of Hb A. Hb S Travis has a rate approximately twice that of Hb S and Hb C Harlem. The mechanical instability of the oxyhemoglobins S, S Travis and A are shown in Fig. 5. All samples were purified by column chromatography before mechanical agitation. As previously established [18,23], oxyhemoglobin S is more mechanically unstable than oxyhemoglobin A. The present study shows that Hb S Travis is even more unstable than Hb S. In this respect
it resembles Hb C Harlem (fi6Va1, p73 Asn), another p chain mutant with two substitutions which is also more mechanically unstable than oxyhemoglobin S [23] (data not shown). It was also instructive to compare stabilities in the deoxy conformation in order to determine if ligandlinked conformational changes played a role in the mechanical stabilities of these mutants (Table 3). In this study constant volumes of hemoglobin solutions
565
W. F. Moo-Penn et a1
Table 3. Mechanical instability of oxy and deoxyhemoglobin mutants Values represent the percentage of hemoglobin remaining in the supernatant after 2 min of mechanical agitation in 0.15 M potassium phosphate buffer, pH 7.35. Three determinations were done for each hemoglobin (oxy and deoxy). The initial concentration of all hemoglobin samples was 200 mg/dl Hemoglobin
Hb A Hb S Hb S Travis
Oxy
Deoxy
x
range
x
range
96.2 83.3 59.9
(94.4-99.1) (80.4-84.9) (54.3-65.8)
98.3 88.6 79.2
(95.1 -99.8) (86.6-92.2) (75.6-83.8)
O!
in either air or 100% helium were shaken for 2 min. The hemoglobin remaining in the supernatant after shaking was expressed as a percentage of the initial concentration. Under these conditions approximately 4.4 times more oxyhemoglobin S was precipitated than oxyhemoglobin A, whereas approximately 10 times more oxyhemoglobin S Travis was precipitated. In the deoxy form all the hemoglobins are correspondingly more stable, with the biggest change seen in deoxyhemoglobin S Travis, which is nearly twice as stable as its oxy form. Nevertheless, deoxyhemoglobin S Travis remains more unstable than deoxyhemoglobin S by a factor of 1.8. Heat denaturation studies (Fig. 6) indicate that Hb S Travis is mildly unstable and the isopropanol test also indicated similar instability. The oxygen equilibria data (Fig. 7) indicate that 'stripped' Hb S Travis has a higher affinity for oxygen and a lower affinity for Pz-glycerate and inositol hexakisphosphate than Hb A. The high oxygen affinity of Hb S Travis is particularly marked at acid pH. The degree of cooperativity of the variant in the absence of allosteric effectors, as reflected in the values of Hill's constant n, is similar to that of Hb A. In the presence of these effectors cooperativity is reduced. The results can probably be interpreted in terms of the proximity of the mutation at P142 to both a heme contact and cofactor binding site. The Bohr effect is unchanged.
DISCUSSION Differences in the properties of Hb S and Hb S Travis can be attributed to the second substitution at position 142 in the /3 chain. The oxygen equilibria data show that Hb S Travis has a higher oxygen affinity than Hbs A and S . Both 'stripped' Hbs A and S have similar oxygen affinities. Substitution of the bulky amino acid valine for the smaller alanine could produce destabilization of the contact between neighboring leucine-141 (H19) and heme in the interior of the molecule, as well as lead to changes in the
0
I
5
10
20 Time (min)
30
40
Fig. 6. Kinetics of heat denaturation of Hb A ( 0 ) and Hb S Travis (A) at 6 5 ° C . Both hemoglobins were isolated from the same chromatographic column. A 1 g/dl solution of hemoglobin in 0.1 M NaC1, pH 7.0, was heated at 65 "C. Samples were removed at the indicated times to determine the amount of denatured hemoglobin. A hemolysate of H b A gave results identical to purified Hb A
Fig. 7. The p H dependence of the oxygen equilibria und Hill's constant, n, of Hb S Travis. The hemoglobin concentration was between 50- 60 pM in 0.05 M [bis(2-hydroxyethyl)amino]tris(hydroxymethy1)methane (Bistris) or Tris buffers. Incubation was at 20 "C. Pz-glycerate and inositol hexakisphosphate (inositol-Ps) were added in 100-fold molar excess over tetramer. (0)Stripped Hb S Travis; (A) Pz-glycerate; (H) inositol hexakisphosphate; (----) Hb A
+
+
orientation of the side-chains of tyrosine-145 (HC2) and histidine-146 (HC3), residues that are involved in the stabilization of the deoxy or T conformation [24, 251. Partial destabilization of the deoxy quaternary
566
W. F. Moo-Penn et al. : Hb S Travis A Sickling Hemoglobin with Two Substitutions in the 4 Chain
structure would thereby lead to increased oxygen affinity. An increased tendency for Hb S Travis to dissociate into subunits would also contribute to its increased oxygen affinity. Kinetic studies to determine the degree of subunit dissociation are in progress and will be reported elsewhere. It should also be noted that the other known mutants in this region of the H helix all show increased oxygen affinity [26 - 321. Of further interest is the lowered affinity of Pz-glycerate and inositol hexakisphosphate for Hb S Travis. The binding of P2-glycerate and inositol hexakisphosphate has been shown to involve eight cationic sites of the twoJ! / chains [33]. These include the two pl valine, 82 histidine and p143 histidine sites and one of the j8 2 lysine residues from either chain. It appears that the substitution of valine for alanine at position 142, adjacent to histidine-143, may also disrupt the P2-glycerate and inositol hexakisphosphate binding site in the central cavity, and this would be reflected in the lowered affinity of the allosteric effectors for Hb S Travis. The increase in the rate of autoxidation and mechanical precipitability of Hb S Travis (as compared to Hb A and S) suggest that conformational differences exist between H b S and Hb S Travis. These properties, in addition to the mild heat instability, could possibly be related to stereochemical interference (a result of the 142 Ala-Val substitution) of the adjacent heme contact at leucine-141. Only X-ray crystallographic studies (in progress) will provide definitive evidence for the conformational changes proposed here. We thank Drs J. Bonaventura, C. Bonaventura and Bolling Sullivan for permission to use their data on the effect of inositol hexakisphosphate on Hb S Travis. We also thank Dr Goldstein, Miami Beach, Florida for his help with the family study.
REFERENCES 1. Lehmann, H. (1975) in Abnormal Haemoglobins and Thalassaemia (Schmidt, R. M., ed.) pp. 323 - 366, Academic Press, New York. 2. Bookchin, R. M., Nagel, R. L. & Ranney, H. M. (1967) J . Biol. Chem. 242, 248-255. 3. Goossens, M., Garel, M. C., Auvinet, J., Basset, P., Ferreira Comes, P. & Rosa, J. (2975) FEES Lett. 58,149- 154. 4. Adams, J. G. & Heller, P. (1973) Blood, 42, 990 A40. 5. Schmidt, R. M. & Brosious, E. M. (1975) Basic Laborutory Methods ojHemoglobinopathy Detection, 7th edn, pp. 26 37, HEW Publication no. (CDC) 76-8266, Atlanta. -
6. Huisman, T. H. J. & Dozy, A. M. (1965) J . Chromatogr. 19, 160- 169. 7. Schmidt, R. M., Moo-Penn, W. F. & Brosious, E. M. (1975) Advanced Laboratory Methods of Hemoglobinopathy Detection, 2nd edn, pp. 200-201, HEW Publication no. (CDC) 75-8296, Atlanta. 8. Clegg, J. B., Naughton, M. A. & Weatherall, D. J. (1966) J . Mol. Biol. 19, 91-308. 9. Raftery, M. A. & Cole, R. D. (1963) Biochem. Biophys. Rcs. Commun. 10,461- 472. 10. Smyth, D. C. (1967) Methods Enzymol. I ! , 216-222. 11. Jones, R. T. (1964) Cold Spring Harbor Symp. Quant. Bid. 29, 297 - 308. 12. Spackman, D. H., Stein, W. H. & Moore, S. (1958) Anal. Chew. 30,1190-1206. 13. Beckman Instruments (1975) in Sequence, 7 , 19- 19b. 14. Inman, J. D., Hannon, J. E. & Apella, E. (1972) Biochem. Biophys. Res. Commun. 46,2075 - 2081. 15. Pisano, J. J. (1972) Methods Enzymol. 25, 27-43. 16. Moo-Penn, W. F., Jue, D. L., Bechtel, K. C., Johnson, M. H., Schmidt, R. M., McCurdy, P. R., Fox, J., Bonaventura, J., Sullivan, B. & Bonaventura, C. (1976) J. Biol. Chem. 251, 7557-7562. 17. Bookchin, R. M. & Nagel, R. L. (1971) J . Mol. B i d . 60, 263 - 210. 18. Roth, E. F., Jr, Elbaum, D. & Nagel, R. L. (1975) Blood, 45, 377- 386. 19. Roth, E. F., Jr, Elbaum, D., Bookchin, R. M. & Nagel, R. L. (1976) Blood, 48, 265-271. 20. Riggs, A. F. & Wolbach, R. A. (1956) J . Gen. Physiol. 39, 585-605. 21. Bonaventura, J., Bonaventura, C., Amiconi, G., Antonini, E. & Brunori, M. (1974) Arch. Biochem. Biophys. 161,328-332. 22. Itano, H. A. (1953) Arch. Biochem. Biophys. 47, 148- 156. 23. Asakura, T., Ohnishi, T., Friedman, S. & Schwartz, E. (1974) Proc. Natl Acud. Sci. U.S.A. 71, 1594- 1598. 24. Perutz, M. F., Muirhead, H., Cox. J. M. & Goaman, L. C. G. (1968) Nature (L0nd.J 219, 131-139. 25. Perutz, M. F. (1970) Nature (Land.) 228, 726-739. 26. Bonaventura, C., Bonaventura, J., Amiconi, G., Tentori, L., Brunori, M. & Antonini, E. (1975) J . Biol. Chem. 250, 6273 6277. 27. Bare, G. H., Alben, J. O., Bromberg, P. A,, Jones, R. T., Brimhall, B. & Padilla, F. (1974) J . Biol. Chem. 249, 773779. 28. Jensen, M., Oski, F. A,, Nathan, D. G. & Bunn, H. F. (1975) J . Clin. Invest. 55, 469-477. 29. Zak, S. J., Brimhall, B., Jones, R. T. & Kaplan, M. E. (1974) Blood, 44, 543 - 549. 30. Hayashi, A,, Stamatoyannopoulos, G., Yoshida, A. & Adamson, J. (1971) Nature (Lond.) 230, 264-267. 31. Charache, S., Brimhall, B. &Jones, R. T. (1975) Johns Hopkins Med. J . 136, 132-136. 32. Perutz, M. F., del Pulsinelli, P., Ten Eyck, L., Kilmartin, J. V., Shibata, S., Tuchi, I., Miydji, T. & Hamilton, H. B. (1971) Nature (L,ond.j 232, 147- 149. 33. Arnone, A. (1972) Nature (Lond.) 237, 146- 149. -
W. F. Moo-Penn, R. M. Schmidt, D. L. Jue, K. C. Bechtel, and J. M . Wright, Hematology Division, Center for Disease Control, Public Health Service, Department of Health, Education, and Welfare, Atlanta, Georgia, U.S.A. 30333 M. K . Horne, 111, College of Medicine, Upstate Medical Center at Syracuse, State University of New York, Syracuse, New York, U.S.A. 13210 G. L. Haycraft, David Grant Medical Center (MAC), Travis Air Force Base. California, U.S.A. 94535
E. F. Roth, Jr and R. I,. Nagel, Division of Hematology, Department of Medicine, Albert Einstein College of Medicine, Yeshiva University, 1300 Morris-Park Avenue, Bronx, New York, U.S.A. 10461
