THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1984 by The American Soeiety of Biological Chemists, Inc.

Vol. 259, No. 11, Issue of June 10, pp. 7325-7330,1984 Printed in U. S. A .

Human Hemoglobin Portland I1 (&&) ISOLATION AND CHARACTERIZATION OF PORTLAND HEMOGLOBIN COMPONENTS AND THEIR CONSTITUENT GLOBIN CHAINS*

(Received for publication, September 6, 1983)

Zafar I. RandhawaS, Richard T. Jones, and Luan E. Lie-Injos From the Department of Biochemistry, School of Medicine, Oregon Health Sciences University, Portland, Oregon 97201 and the §Department of Epidemiology and International Health, University of California Medical School, San Francisco, California 94143

Two types of embryonic hemoglobins (Hb) containing { chains have been identified in the blood of several neonates of Chinese origin with homozygous a-thalas- semia. In addition to Hb Portland I ( { z ~ z ) which was previously reported, another embryonic hemoglobin has been detected and found to contain { chains and j3 chains. It is being designated Hb Portland I1 and has the formula ({zj3z). It has a mobility slightly slower than that of Hb A on starch gel electrophoresis at pH 8.6 and has been found in the hemolysates of blood of some but not all hydropic infants. Another component with a mobility faster than that of Hb Az on starch gel has been isolated from the blood of some hydropic neonates. This latter component is postulated to be {z&. The occurrence of Hb Portland I and Hb Portland I1 in these hydropic neonates is consistent with the hypoth- esis that, in the absence of normal a chain production, { chains are continued to be produced at later states of development than normal and form tetramers with each of the &like globin chains. Because Hb Portland I1 has not been found in blood from all hydropic neo- nates, we postulate that the presence of this hemoglo- bin in these fetuses may be correlated with the gesta- tional age of the fetus at the time of birth.

The existence of embryonic hemoglobins in human embryos were first described by Huehns et al. (1). Two embryonic hemoglobins, Gower I (first thought to be c4 but later identi- fied as {2t2) and Gower I1 (a2t2) predominate only during the first 10 to 12 weeks of fetal development (1-4). It is believed that these hemoglobins are synthesized by erythroid cells derived from the yolk sac (5). They are also found in trace amounts in normal cord blood and in some K562 cell line cultures which have been induced with hemin (6). Hb’ Port- land I, another embryonic hemoglobin, is composed of two a- like chains (embryonic { chains) and two y chains. Hb Port- land I ( S;y2) was discovered in a female Chinese infant having multiple congenital anomalies and complex autosomal mosai-

* This work was supported in part by National Institutes of Health Grants AM 17850 and AM 21571. Part of these results were presented as an abstract for the Twenty-fourth Annual Meeting of the American Society of Hematology, Washington, D.C., December 4-7, 1982. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$Present address, Bruce Lyon Memorial Research Laboratory, Children’s Hospital of Northern California, 51st and Grove Streets, Oakland, CA 94609. To whom reprint requests should be sent.

The abbreviations used are: Hb, hemoglobin; HPLC, high pres- sure liquid chromatography; NP40, Nonidet P-40; RT, retention time.

cism (7, 8) and later reported in infants with hydrops fetalis due to homozygous a-thalassemia (9, lo), a disease condition first described by Lie-Injo and Jo (11). Capp et al. (8) provided the structural evidence that demonstrated a new hemoglobin chain which they designated the { chain. Their findings suggested not only another hemoglobin chain but at least one more globin gene locus. The similarity in the structure of human a and { globin chains led to the conclusion that this { chain was probably an a-like embryonic chain (12). Based on the findings of Kamuzora and Lehmann (13) and Huehns and Farooqui (14), it can be postulated that { chain should com- bine with all P-like chains (i.e. t, y, @, and 6). We have looked into this possibility by attempting to study all the hemoglobin components in the hemolysates of blood from several neonates with homozygous a-thalassemia.

MATERIALS AND METHODS

Blood Samples-Blood from stillborn neonates with hydrops fetalis due to homozygous a-thalassemia was obtained from the delivery room of the General Hospital, Kuala Lumpur, Malaysia. The washed red blood cells were packed in Dry Ice and shipped by air to the United States. Hemolysates were prepared from these samples by a standard method using deionized distilled water and toluene (15).

Starch Get Electrophoresis-Hemoglobin Portland I, Hb Bart’s, and other hemoglobin components were separated by starch gel electrophoresis (16). Starch gel electrophoresis was carried out in Tris-EDTA-boric acid buffer, pH 8.8. The hemoglobin components were eluted from the starch according to the method described by Lie-Injo (17).

Polyacrylamide Gel Electrofocusing-Hemoglobin Portland, Hb Bart’s, and other components were separated by isoelectric focusing using both analytical and preparative techniques. The analytical isoelectric focusing was performed in a 2-mm thick slab of polyacryl- amide gel containing 2.4% Ampholine carrier ampholyte in the pH range 6-8 without glycerol. After a 30-min prerun, 10-20 pl of 1 mg/ ml of Hb sample predialyzed against 1% Ampholine solution were applied to the gel surface by means of a paper wick (14 X 8 mm) and the hemoglobin separation was completed after 12 h at 4 “C with an LKB 2117 Multiphor Cell, using 10 watts from a constant wattage power supply. Measurement of the pH of the gel was performed using an isoelectric focusing combination surface electrode (Ingold) before staining the polyacrylamide gel. The hemoglobin staining was per- formed according to LKB Notes 138 and 250 using bromphenol blue.

Separation of Globin Chains by Reverse Phose HPLC-The HPLC system was assembled from the modular components: Altex-Beckman model 332 gradient liquid chromotograph, an Altex-Hitachi model 100-30 UV-VIS variable wavelength detector with 8-p1 analytical flow cell, and a Hewlett-Packard recorder of a HP5830A gas chromoto- graph system which was programmed to compute the per cent of area of the peaks and their retention times. HPLC grade acetonitrile was obtained from Burdick and Jackson and methanol from J. T. Baker Chemical Co. All other chemicals were reagent grade.

Globin was prepared by acid acetone precipitation (18) from the whole hemolysate or from the purified hemoglobin components. For globin separation both analytical (3.9 mm X 30 cm) and preparative (7.8 mm X 30 cm) C-18 Bondapak columns (Waters) were used. The

7325

7326 Hemoglobins Containing Chains mobile phase solvents "A" and "B" were those described by Shelton and co-workers (18). The gradient programs for separating globin chains by the short duration (50 rnin) and long duration (180 min) procedures using analytical and preparative columns are described in detail elsewhere (19) but are presented in summary with the results.

Polyacrylamide Gel Electrophoresis-Globin chains were separated by the method of Alter et 01. (20) using electrophoresis on 12% polyacrylamide gels containing 6 M urea and 2% Triton X-100 in 5% acetic acid.

Slab gels (12 X 16 cm X 1.5 mm) with 6 or 12 spaces were employed for analytical electrophoresis using 5% acetic acid as electrophoresis buffer. The sample buffer consisted of 5 ml of deionized 8 M urea, 0.5 ml of glacial acetic acid, 0.5 ml of 2-mercaptoethanol, and 2 mg of pyronin y. Five to ten pg of globin were mixed with 20 pl of sample buffer. Electrophoresis was performed for 17 h at a constant current of 8.5 mA at room temperature.

The gels were stained for 30 min with 0.5% Coomassie brilliant blue in 7% acetic acid, 30% methanol and destained with the acetic acid-methanol by diffusion. The gels were dried by placing in a Bio- Rad slab dryer (model 224) for 1 h at 45 "C.

Polyacrylamide Isoelectric Focusing in NP40, 8 M Urea-Lyophi- lized globin samples of 50-200 pg were dissolved in 8 M urea, 3% NP40, 10% 2-mercaptoethanol, at a concentration of 5 mg/ml, fo- cused in 6% acrylamide slab gels (12.5 x 2.6 cm x 1.5 mm) containing 8 M deionized urea, 2% pH 6-8 and 0.2% pH 3.5-10 Ampholine (LKB), 3% NP40 and stained with Coomassie brilliant blue G-250. The gels were prefocused for 2 h a t 200 V and finally focused for 2 h, 30 min after the application of globin samples. Although these con- ditions are very similar to those recently described by Gianni et al. (21), they were developed independently.

RESULTS

The results described here were obtained on samples from several neonates with homozygous a-thalassemia hydrops fetalis syndrome. All these infants were of Malaysian-Chinese origin. The results obtained from the separation of various hemoglobin components and their constituent globin chains (Figs. 1-6 and Tables I and 11) indicate that there was no detectable a chain in blood from any of these infants.

Starch gel electrophoresis of the whole hemolysate of a neonate with homozygous a-thalassemia at pH 8.8 produced four bands. These bands are designated components I (major band), 11,111, and IV (a faint slow moving band) as shown in Fig. 1.

Component I had a mobility identical to Hb Bart's on starch gel electrophoresis. Its identity as Hb Bart's was con- firmed (19, 22) by (a) separating the G, and A, chains by HPLC procedures, (b) isoelectric focusing of G, and A, globin chains, ( c ) globin chain electrophoresis on polyacrylamide gel of globin samples from Hb Bart's, ( d ) examinations of tryptic peptides of globin samples using HPLC procedures and Ami- nex A5 cation exchange chromatography,2 and (e) amino acid composition of the G, zone separated by HPLC, which agreed well with published data.

Component I1 had a mobility on starch gel electrophoresis at pH 8.6 faster (more anodic) than Hb A while component 111 had a mobility slightly slower than Hb A. The hemolysate of blood from one hydropic stillborn produced only two visible bands corresponding to components I and I11 whereas other stillborns produced all of the four bands. That the starch gel component I1 is identical to Hb Portland I ({*y2) was con- firmed by (a) separating the G,, A,, and { chains by HPLC procedures, (b) separating the y tryptic peptides by HPLC, and ( c ) amino acid composition of the y and {chains (19,22).

The tryptic digests of the globin from components I1 and 111 when separated by A-5 chromatography were found to contain a tryptic peptide unique to the { chain as identified by Capp et al. (zone XIX of Fig. IC in Ref. 8). The amino acid

* Z. I. Randhawa and R. T. Jones, unpublished data.

-(COMP I) Hb Bart's ('4)

-(COMI?II) Hb Portland I('& Y2)

Hb F- -(COMf?III) Hb Portland II (52p2)

-(COME' E ) H b PortlandllI (52%)

Hb A2-

@ A B I C I

6 f $

FIG. 1. Separation of various hemoglobin components by starch gel electrophresis using Tris-EDTA-boric acid buffer, pH 8.6. After separation the gel was stained with benzidine and o- dianisidine. The hemolysates were obtained from ( A ) normal adult; ( B ) stillborn infants with homozygous a-thalassemia (hydrops fe- talis); and (C) newborn infant with Hb Bart's.

composition of the {-like chain was that of {from Hb Portland I (22). In addition, several other differences were found for the tryptic digests of components I1 and I11 compared to the y chain tryptic digest. Hence it was concluded that these components were hemoglobins which contained { chains. A hemoglobin with the composition of Hb Portland I was also purified from two other hydropic neonates as component I1 by starch gel electrophoresis.

Examination of component I11 from two hydropic stillborn fetuses by using a short duration HPLC method (Fig. 3) also confirmed the presence of { chains by the characteristic retention time of 37 min. The evidence that the HPLC zone with a retention time of 37 min is { chain has been given in a previous report (19,22). In addition to very small amounts of G, and A, (which are considered to be contaminants) what appeared to be a unique globin chain with a retention time of about 15 min was also present in component 111. The elution position of the chain with retention time of about 15 min is identical to that of the /3 globin chain of Hb A. That this chain is the same as /3 of normal Hb A is based upon the following results: (a) separation of /3 and { chains from the purified hemoglobin component by polyacrylamide isoelectric focusing (Fig. 4B); (b) globin chain electrophoresis on poly- acrylamide gel of material present in the zone with retention time of 15 min; it had a mobility identical to that of /3 A chain (Fig. 5B); (c) examination of its tryptic peptides which had a pattern identical to those of the /3 chain (Fig. 6).

The presence of /3 chain was further confirmed from the amino acid compositions of two typtic peptides which corre-

Hemoglobins Containing { Chains

6.37 -

6.50 -

6.63 -

6.85 -

6.98 - 7.13 -

Band I (Hb Portland I )

Band II

Band III (Hb Bart's)

Band IZL

Hb A Hb F

I I A b c

FIG. 2. Analytical polyacrylamide gel electrofocusing (pH range 6-8) of hemoglobins from (A) Hb A; (B) cord blood; (C) neonate with hydrops fetalis due to homozygous a-thalasse- mia. Isoelectric points were determined prior to hemoglobin staining which was performed by using bromphenol blue. Bands I and 111 were identified as Hb Portland I and Hb Bart's. Band 11 is postulated as Hb Portland I1 whereas Band IV is a y chain-containing hemoglobin component (see text for details).

spond to /3 T-3 and /3 T-5 (Table I) both of which are unique to the /3 chain. The /3 and {chains appeared to be present in approximately equal molar amounts in component 111. Com- ponent I11 from starch gel was identified as Hb Portland I1 with the composition of c2/3* in three different hydropic fe- tuses.

A zone from the HPLC separation of chains of Hb Portland I1 (Fig. 3) with a retention time of 33 min was observed ahead of the main { chain zone (RT of 37 min). This was also observed in component I11 from two other hydropic neonates. After aminoethylation and tryptic hydrolysis, both of these zones gave identical peptide patterns by HPLC. No differ- ences in amino acid composition could be detected between RT 33 and RT 37 (22). Whether or not these two zones from the { chain are due to chemical modifications or sequence heterogeneity remains to be shown.

The analytical isoelectric focusing pattern (Fig. 2) of the hemolysate from the suspected hydropic neonates confirms the absence of Hb A, Hb F, Hb A2, or any other a chain containing hemoglobin. Lane C of Fig. 2 also shows separation of four bands (Bands I, 11,111, and I V ) in the whole hemolysate from a hydropic infant. Bands I and I11 were identified as Hb Portland I and Hb Bart's. This was established (a) by exam- ining by isoelectric focusing the behavior of components I and I1 separated and purified by starch gel electrophoresis and (b ) by separating tryptic peptides using A-5 chromatography. Band I and I1 were found to contain the unique peptide { TI by ion exchange chromatography and hence are considered 3

I

I I

I I

I I

I I

I I

I I

I I

I I I

I"

" I + I O

TIME IN MINUTES FIG. 3. Preparative separation of Hb Portland I1 (f&*) glo-

bin on Waters pBondapak C-18 preparative column (7.8 mm X 30 cm). The hemoglobin was obtained from the CKL hydrops as component I11 on starch gel electrophoresis. The reverse phase col- umn was attached with a guard column (containing C-18 Porasil packing), and the separation was made at a flow rate of 3.5 ml/min using a 55-min program. The gradient program and the column cleaning procedure are given in Ref. 19. About 6 mg of the globin was dissolved in 12 ml of phosphate solution and stirred for 1% h at room temperature prior to loading sample after centrifugation. The loading procedure is also given in Ref. 19. Similar results were obtained from component I11 of the CTY hydrops (22).

0 - , - - --Origin I

A B 6 FIG. 4. Analytical isoelectric focusing of hydrops globin

chains. The globin was made by acid-acetone precipitation method of the following CKL hydrops samples: A, whole hemolysate; B, Hb Portland I1 separated by starch gel electrophoresis; and C, Hb Bart's separated by starch gel electrophoresis. About 50-100 pg of the globin were dissolved in sample solution and focused in pH 6-8 gel contain- ing 8 M urea and NP40. The gel was stained with Coomassie brilliant blue, destained, and dried.

chain-containing hemoglobins. Since Band I has been shown to be Hb Portland I with an isoelectric point of 6.37, it is postulated that Band I1 is Hb Portland 11. Band I V which was not observed in all hydropic neonates contained only y chains and is postulated to be some modified form of Hb Bart's.

We found that the blood hemolysates from hydropic neo- nates which were sent from Malaysia frozen in Dry Ice did not give satisfactory separations by starch gel electrophoresis

Hemoglobins Containing {Chains 7328

0

5 - AY - GY

P -

-

P-

o

FIG. 5. Electrophoresis of globin chains on Triton polyac- rylamide gel. About 5-20 pg of globin were applied to each well. The gel was stained with Coomassie brilliant blue R. A, Hb A; B, RT 15 (CKL) zone purified by HPLC; C, Hb A; D, RT 37 (CKL) zone purified by HPLC; E, purified Hb Bart’s (component I); F, Hb F; and C, RT 37 (GTT) zone purified by HPLC short duration program. For details see Fig. 3 of Ref. 19.

FIG. 6. Separation of the tryptic peptides of about 0.5 mg of aminaethylated X chain obtained from Hb Portland I1 globin of CKL hydrops by HPLC. The separation was made on an Altex Lichrosorb C18 RP column (4.6 mm X 25 cm) using an ammonium acetate-acetonitrile gradient with a flow rate of 1 ml/mm. Both @T3 and @T5 tryptic peptides are unique to the B chains.

or isoelectric focusing on polyacrylamide gels if stored for several weeks. For this reason we could not purify various hemoglobin components from all of the blood samples. How- ever, the whole globin from this stored material could be used to separate the chains by HPLC procedures as shown in Table 11. The tryptic peptides of the globin chains of Hb Bart’s (component I), Hb Portland I (component 11), and Hb Port- land I1 (component 111) were investigated by converting pure hemoglobin components to globin.

DISCUSSION

The results obtained by electrophoresis and analysis of the constituent globin chains of the hemoglobin components from hemolysates of blood from infants with hydrops fetalis indi- cate a total absence of a-chain production. The results pre- sented here also provide clear evidence that in addition to Hb Portland I ( c2y2) identified earlier by Capp et al. (7) most of these neonates with homozygous a-thalassemia produce an- other hemoglobin-containing { chain with the composition of {&. Preliminary studies on the globin chain which elutes at a retention time of 15 min in an HPLC chain separation procedure have identified it as /3 chain by the presence of tryptic peptides f l -3 and f l -5 and the similarity of its tryptic peptide pattern to that of the B chain of Hb A. These findings rule out the possibility that this chain is an c chain as found in Hb Gower I, t2c2.

All 10 infants with homozygous a-thalassemia that were studied had detectable to significant amounts of {chains and most also had &globin chains in their cord blood. The con- clusion that all these neonates were producing Hb Portland I (l2y2), Hb Portland I1 ({2j32), or both of these hemoglobins in detectable amounts in addition to predominant amounts of Hb Bart’s is based upon starch gel electrophoresis and HPLC separations. The presence of Hb Portland I1 in some hydropic neonates has also been confirmed by separating this hemoglo- bin as component I11 on starch gel followed by subsequent analysis of the globin chains from three different hydropic fetuses. One hydropic fetus (CKL) which did not have any significant amount of Hb Portland I contained about 15% of Hb Portland 11. The cord blood of another hydropic neonate (GTT) did not have any detectable amounts of B chain in the whole globin preparation or in any of its hemoglobin compo- nents but rather contained only Hb Portland I in addition to the predominant amount of Hb Bart’s. Similar results were also obtained from hydrops CFM.

Hemoglobin Portland I ( {2yz) was separated as component I1 by starch gel electrophoresis. It moves faster than Hb A but slower than Hb Bart’s (Fig. 1). A number of authors have detected Hb Portland I in samples from early human embryos, hydropic neonates, and human leukemic cell lines K562 (6,7,

Hemoglobin X separated by Todd et al. (10) by starch gel electrophoresis using Tris-EDTA-borate buffer at pH 8.6 may have been a mixture of Hb Portland I and Hb Portland 11. The electrophoretic mobility of their Hb X appears to be identical to Hb Portland I1 (i.e. component I11 as shown in Fig. 1). This conclusion was strengthened further by the compositional analysis of their Hb X which Todd et al. found consisted of y chains, { chains, and B chains. In view of the data presented here, it seems very likely that the hemoglobin designated Hb X by Todd et al. (10) contained Hb Portland 11.

Hemoglobin components with the mobility of Hb Portland I (starch gel component 11) and Hb Portland I1 (starch gel component 111) were also separated by Pootrakul et al. in a survey of hemoglobins (26) from a neonate with Hb Bart’s hydrops in Thailand. In that survey a hemoglobin component separated by starch gel electrophoresis with mobility identical to that of Hb Portland I1 was assumed to be Hb F (a2y2). The presence of Hb F as any fraction in hydropic blood must be incorrect because hydrops fetalis results from complete dele- tion of a genes (27, 28) and, therefore, a complete lack of production of a chains. Thus we propose that the hemoglobin component identified by Pootrakul et al. (26) as Hb F was in fact Hb Portland 11.

Hemoglobin Portland I is now recognized as a normal

23-25).

Hemoglobins Containing { Chains 7329

TABLE I Amino acid composition of two tryptic peptides of 8-globin chain of Hb Portland II fS;BJ from CTY and CKL

hydropic infant (A) fl-3-like peptide (B) m-5-like peptide

Amino Acid R T R T R T R T R T R T R T

CKL 39.7

CTY 39.9 39.7 33.2 ET-3 dT-3 66.8

f l - 3 y T - 3 CTY" m - 5 y T - 5 67.4 77.0 rT-5 dT-5

Lysine 4.5 (1) 1 1 1 1 Arginine 6.3 (1) 3.8 (1) 1 1 1 1 Aspartic acid 13.3 (2) 7.1 (2) 2 2 1 2 11.7 (3) 3 3 3 3 Threonine 0 0 0 1 0 0 4.0 (1) 1 0 0 0 Serine 8.3 (2) 2 4 3 3 Glutamic acid 12.9 (2) 8.5 (2) 2 2 3 1 4.1 (1) 1 0 1 1 Proline 7.8 (2) 2 1 2 2 Glycine 19.5 (3) 11.1 (3) 3 3 3 3 10.4 (2) 2 2 2 2 Alanine 6.6 (1) 4.7 (1) 1 1 2 2 5.0 (1) 1 2 1 1 Valineb 16.5 (3) 9.8 (3) 3 2 1 3 5.2 (1) 1 0 1 1

0 Methionine 0 0 0 1 0 3.1 (1) 1 1 1 1 Isoleucine 0 0 1 0 0 Leucine 6.5 (1) 4.0 (1) 1 1 1 1 5.1 (1) 1 1 1 1 Phenylalanine 11.5 (3) 3 3 3 3 Average residue 6.3 (13) 3.8 (13) 13 13 13 13 4.1 (19) 19 19 19 19

(nm)/total Peptide zone (RT 67.1) from CKL hydrops @ chain had identical amino acid composition. 22-h acid hydrolysate.

TABLE I1 Per cent area of the globin chains from the HPLC separation of 1 mg

of whole globin from hydropic neonates R?" R?" R'l" R?" RT"

4 G, A, ? < Hydropic 15 min 19 min 21 min 23 min 37 min no. neonate

1 CKL 6.5 50 37 2 GTT 1 48 25 17 9

6.5

3 WM 6.6 49 37 4 YLL 10

7.4 52 34

5 3.0

TAM 9.7 62 18 10.3 6 CH 7.4 67 18 7.6

"These values represent percentage of the total area due to the absorption at 220 nm and not the actual amount of the globin chain (HPLC).

embryonic human hemoglobin. Its presence in trace amounts in normal newborn babies and significant proportions during early embryonic life has been reported by several investigators (5, 29). We postulate that Hb Portland I1 ( { 2 p 2 ) may not be detectable in normal newborn neonates because the normal infant produces significant amounts of a-chain and the affin- ity of a-chain for adult &chain may be greater than {-chain for p. In addition, because the {-chain in normal newborn neonates is produced in very small amounts, it would be difficult to detect in trace amounts even if Hb Portland I1 is produced. Therefore, it can be argued that Hb Portland I1 can only be detected in special conditions where the {and p globin chain synthesis is active, and synthesis of the LY chain is either absent or markedly reduced.

Because a-thalassemic hydropic neonates cannot make a- chains but continue to synthesize {-chain, we postulate that during early embryonic life these fetuses may produce Hb Gower I ( c2tz) followed by Hb Portland I ( c2yz). From the fact that some hydropic infants are found to have significant amounts of Hb Portland I1 ( {2;P2) in addition to large amounts of Hb Bart's (y4), it is concluded that these fetuses were producing p chain before birth. The fact that there was no Hb Portland I found in these hydropic fetuses in detectable amounts leads to the suggestion that p chains may have greater affinity for {chain than y chains have for {. Based on this assumption it is also speculated that those hydropic

neonates which are born closer to full term have more active synthesis of adult p chains and thus produce Hb Portland 11. Furthermore, we propose that hydropic infants which are born prematurely produce mainly Hb Portland I because little or no p chain is synthesized prior to 30 weeks of intrauterine life. In short, when no a chains are available to form tetramers with p, y, and 6 chains then { chains may form these tetra- mers. The proportion of the various tetramers depends upon the relative abundance of p, y, and 6 chains and their relative affinity for the { chains.

Starch gel component IV has not been investigated further as yet. It is unlikely to be Hb Gower I ( ( 2 4 because t-globin synthesis is active only during early embryonic life and most of the hydropic neonates included in this study were born close to full term. Moreover, based on the electrophoretic mobility of component IV on starch gel, it is unlikely to be Gower I because Gower I moves slower than Hb Az, whereas component IV moves faster than Hb A2. It is postulated that component IV may have the composition {2&. The existence of component IV as a {- and &containing hemoglobin com- ponent would be more consistent with adult 6 chain produc- tion closer to full term after adult p chain synthesis is fully activated. A hemoglobin designated X with a composition y2t2 has recently been reported in some embryonic neonates and in human K562 cell lines when induced by hemin (6). Com- ponents I11 or IV are unlikely to be Hb X (y2t2) because no t chain could be detected in samples from these hydropic neo- nates.

The presence of Hb Portland I and Hb Portland I1 suggests that { is a-like in its association with non-a chains. The chain forms similar products with y and p chains during fetal development as fetal type Hb F (azyz) and adult Hb A (a&).

REFERENCES 1. Huehns, E. R., Flynn, F. V., Butler, E. A., and Beaven, G. H.

(1961) Nature (Lord.) 189,496-497 2. Huehns, E., Dance, N., Beaven, G. H., Hecht, F., and Motulsky,

A. G. (1964) Cold Spring Harbor Symp. Qwnt. Biol. 29, 327-331

3. Hecht, F., Motulsky, A. G., Lemire, R. J., and Shepard, T. E.

4. Lorkin, P. A. (1973) J. Med. Genet. 10, 50-64 (1966) Science 152, 91-92

7330 Hemoglobins Containing ( Chains 5.

6.

7.

8.

9.

10.

11. 12.

13.

14.

15.

16. 17. 18.

Weatherall, D. J., Pembrey, M. E., and Pritchard, J. (1974) Clin. Hemoglobin 3, 353-358 Haematol. 3. 467-470 19. Randhawa, Z. I., Jones, R. T., and Lie-Injo, L. E. (1983) Anal.

Rutherford, T: R., Clegg, J. B., Higgs, D. R., James, R. W., Thompson, J., and Weatherall, D. J. (1981) Proc. Natl. Acad. Sci. U. S. A. 78, 348-352

Capp, G. L., Rigas, D. A., and Jones, R. T . (1967) Science 157, 65-66

Cam. G. L.. Rieas. D. A., and Jones, R. T. (1970) Nature (Lord.) 228,2781280

Weatherall. D. J.. Cleee. J. B.. and Boon. W. H. (1970) Br. J. . , Haematol. 18, 357-367

Br. J . Haematol. 19, 27-31 Todd, D., Lai, M. C. S., Beaven, G. H., and Huehns, E. R. (1970)

Lie-Injo, L. E., and Jo, B. H. (1960) Nature (Lond.) 185, 698 Kamuzora, H., Jones, R. T., and Lehmann, H. (1974) FEES Lett.

Kamuzora, H., and Lehmann, H. (1975) Nature (Lond.) 256, 511-513

Huehns, E. R., and Farooqui, A. M. (1975) Nature (Lond.) 254, 335-337

Schroeder, W. A,, Shelton, J. R., Shelton, J. B., Cormick, J., and Jones, R. T. (1963) Biochemistry 2, 992-1008

Smithies, 0. (1959) Biochem. J . 71,585-587 Lie-Injo, L. E. (1976) J. Chromatogr. 117, 53-58 Shelton, J. B., Shelton, J. R., and Schroeder, W. A. (1979)

46, 195-199

20.

21.

22.

23.

24. 25.

26.

27.

28.

29.

Biochem. 129, 184-191 Alter, B. P., Goff, S. C., Efremov, G. D., Gravely, M. E., and

Huisman, T. H. J. (1980) Br. J. Haematol. 44, 527-534 Gianni, A. M., Polli, E., Giglioni, B., Comi, P., Ottolenghi, S.,

Ferrari, M., Furbetta, M., Angius, A,, and Cao, A. (1981) Hemoglobin 5(4), 349-356

Randhawa, Z. I. (1981) Ph.D. thesis, Oregon Health Sciences University

Kaltsoya, A., Fessas, P., and StavroDoulous, A. (1966) Science 153; 1417-1418

Lie-Inio. L. E. (1972) Med. J. Mahvs. 27. 120-124 Presscy; L., Higgs, 'D. R., Clegg, 2. B., 'and Weatherall, D. J .

(1980) Proc. Natl. Acad. Sci. U. S. A. 77, 3586-3589 Pootrakul, S., Wasi, P., Pornpatkul, M., and Na-Nakorn, S.

(1970) J. Med. Assoc. Thail. 53, 250-262 Taylor, J. M., Dozy, A., Kan, Y. W., Varmus, H. E., Lie-Injo, L.

E., Ganesan, J., and Todd, D. (1974) Nature (Lond.) 251,392- 393

Ottolenghi, S., Laynon, W. G., Paul, J., Williamson, R., Weath- erall, D. J., Clegg, J. B., Pritchard, J., Pootrakul, S., and Boon, W. H. (1974) Nature (Lond.) 251,389-391

Hecht, F., Jones, R. T., and Koler, R. D. (1967) Ann. Hum. Genet. 31,215-218