Molecular analysis of the beta thalassemia phenotype associated with inheritance of hemoglobin E (alpha 2 beta2(26)Glu leads to Lys)

Molecular analysis of the beta-thalassemia

phenotype associated with inheritance of

hemoglobin E (alpha 2 beta2(26)Glu leads to

Lys).

E J Benz Jr, … , A Altman, J G Adams 3rd

J Clin Invest.

1981;68(1):118-126. https://doi.org/10.1172/JCI110226.

Inheritance of the gene for betaE-globin is associated with hypochromia and microcytosis,

reminiscent of typical heterozygous beta-thalassemia. Patients with hemoglobin

(Hb)E-beta-thalassemia exhibit clinical phenotypes of severe (Hb)E-beta-thalassemia, a circumstance

not encountered in other compound heterozygous states for structural beta-chain mutations

and beta-thalassemia. We have analyzed the kinetics of globin synthesis and the levels of

globin messenger (m) RNA accumulation in patients with Hb E-beta-thalassemia and Hb E

trait. The initial rate of beta-globin synthesis (betaE/alpha=0.20-0.34) was less than

expected on the basis of gene dosage, or comparable studies of other compound

heterozygous states for beta-thalassemia and structurally abnormal beta-chains.

betaE-globin synthesis was not only reduced during short-term incubations (1-5 min), but also

remained relatively unchanged during long-term pulse or chase incubations up to 5h.

Analysis of globin mRNA by cell-free translation and molecular hybridization confirmed that

the unexpectedly low levels of betaE-globin synthesis were associated with comparable

reduction in the levels of beta-globin mRNA. In Hb E-beta-thalassemia the betaA + betaE

(alpha globin nRNA ratio observed were substantially lower than those obtained from

reticulocytes of patients with heterozygous beta-thalassemia, or Hb S-betaO-thalassemia,

while in Hb E trait, the betaA + betaE/alpha mRNA ratio was in the ranged observed for

beta-thalassemia trait. The globin gene specifies reduced accumulation of

betaE-globin mRNA, a property characteristic of other forms of thalassemia. The

beta-thalassemia […]

Research Article

Find the latest version:

http://jci.me/110226/pdf

Molecular Analysis of the

83-Thalassemia

Phenotype Associated with Inheritance of

Hemoglobin

E

(a2f3226Glu

->

Lys)

EDWARDJ. BENZ,JR.,BRIAN W.BERMAN, BARRY L. TONKONOW, and ELAINE COUPAL, Hematology Section, Departments of Internal Medicine and Pediatrics, Yale UniversitySchool of Medicine, New Haven, Connecticut 06510

THOMAS COATES and LAURENCE A. BOXER, Division ofPediatric Hematology, Indiana University SchoolofMedicine andJamesWhitcomb Riley Hospital

for Children, Indianapolis, Indiana 46223

ARNOLD ALTMAN,Division of Pediatric Hematology, University ofConnecticut

School of Medicine, Farmington, Connecticut 06032

JUNIUS G. ADAMS III,Hemoglobin Research Laboratory, Veteran's Administration

Hospital, University of Mississippi Medical Center,Jackson, Mississippi39216

AB S T R A C T

Inheritance

of the

gene

for 8E_globin

is

associated with hypochromia and

microcytosis, rem-iniscent

of typical heterozygous

8-thalassemia.

Pa-tients

with

hemoglobin

(Hb)E-,B-thalassemia

exhibit

clinical

phenotypes of

severe

8-thalassemia,

a circum-stance not

encountered

in

other compound

heterozy-gous states

for structural

p-chain

mutations

and

8-thalassemia.

We

have analyzed

the kinetics of globin

synthesis and the levels

of globin

messenger(m) RNA

accumulation

inpatients

with Hb

E-p-thalassemia

and

Hb

E trait.

The initial

rate

of

8-globin

synthesis

(pE/a

=

0.20-0.34)

was

less than expected

on

the basis

of

gene

dosage,

or

comparable studies of other

com-pound heterozygous

states

_for,

-thalassemia and

struc-turally abnormal 8-chains.

pE_globin

synthesis

was

not

only reduced during short-term incubations (1-5

min),

but

also remained relatively

unchanged during

long-term

pulse or chase

incubations

up to 5 h.

Analysis

of

globin

mRNA

by cell-free translation and

molec-ular

hybridization

confirmed

that

the

unexpectedly

low levels of

pE_globin

synthesis were

associated

with

Preliminaryfindings fromthese studieswerepresentedat

the 37th Annual Meeting of the American Federation for Clinical Research, 10-12 May 1980, in Washington, D. C.

(Clin. Res. 28: 305A).

Address all correspondence to Dr. Benz, Hematology

Section, Department of Internal Medicine, Yale University

School of Medicine,New Haven, Conn. 06510.

Receivedfor publication 15August 1980 and in revised form20February1981.

118

comparable

reductions

in

the levels

of,8-globin

mRNA.

In

Hb

E-,p-thalassemia

the 8A

+

pE/a

globin

mRNA ratios

observed

were substantially lower than those

obtained

from reticulocytes of

patients

with

heterozy-gous

p-thalassemia,

or

Hb

S-lo-thalassemia, while

in

Hb

Etrait,

the 8A

+

pE/a

mRNA ratio was in

the

range

observed

for p-thalassemia

trait.

The

pE-globin

gene

specifies reduced accumulation of

,E_globin

mRNA, a property

characteristic

of other forms

of,p-thalassemia.

The

p-thalassemia

phenotype associated

with

inheri-tance

of Hb

E is

thus

determined

at

the level

of

8-globin

mRNA

metabolism.

INTRODUCTION

Hemoglobin (Hb)l

E

(f826

Glu-*LYS)

isthe

third

most

com-mon

hemoglobin

variant

worldwide;

it is

particularly

prevalent

among

Southeast

Asian

individuals

(1-3).

In-dividuals

heterozygous

(Hb

E

trait)

or

homozygous

(Hb

E

disease) for

Hb E

exhibit

microcytosis

and

erythrocyte morphology

resembling

p-thalassemia

trait

(1-3). Double heterozygosity for Hb

E

and

p-thalas-semia is

frequently associated with

a

moderately

severe

,p-thalassemia

syndrome, the molecular basis of which

has

not

been

well defined

(2-5).

Ithas

been

suggested

that the

thalassemia-like

phenotype of Hb

E

disorders

1Abbreviations usedinthispaper: c,complementary(e.g., cDNA); Hb, hemoglobin; m, messenger(e.g.,mRNA);MCV,

meancorpuscular volume.

results from impaired synthesis

of

iE_globin (6-10),

or,

perhaps from I3E_globin

chain

instability (11, 12).

Either

mechanism could

result in a

limited supply of the

ab-normal

f8-globin chain and lead

to

decreased net

ac-cumulation of Hb E. To define more

precisely the

pathogenesis of

Hb E

disorders,

we

have analyzed the

kinetics

of

_jIE-globin

synthesis

and the amounts

of

total

/8 messenger

(m)

RNA (/3A +1E

mRNA)

in two patients with Hb

_{E-f3-thalassemia,}

and a

patient

with

Hb

E trait. Our results support

the concept that

1E_

globin

is

synthesized initially

at a

moderately

de-creased rate.

Although 1A_

and 3E-mRNA could not be

individually distinguished and quantified, the total

,8-mRNA

(

,A

+ 3E

mRNA)

level was

markedly reduced,

suggesting that diminished

PE-globin synthesis

was

due

to a comparable

reduction

in

the

amount of

PE-globin

mRNA.

The

gene

coding for this

_,8-globin

struc-tural variant thus exhibits

features associated with

most other forms of /8-thalassemia genes.

METHODS

Patients. TwounrelatedSoutheastAsian patientswith Hb

E-,f-thalassemia

were studied. Hematologic parameters and

familystudiesaresummarizedinTableI.Patient1,a

Vietnam-ese, had severe, transfusion-dependent anemia

phenotypi-callysimulatingclassicalhomozygous f3-thalassemia.Patient

2,of Thai extraction, hada13-thalassemiaintermedia pheno-type, with marked hypochromic, microcytic anemia and hepatosplenomegaly, but no chronic transfusion require-ment. Neither patient had undergone splenectomy. Both exhibit typical thalassemic facies, growth retardation, and erythroid hyperplasia. Patientswith Hb

S-,8-thalassemia

and "classical"

/3-thalassemia

who were studied have been de-scribed previously (13-16).Inaddition, the father ofpatient 1,

aheterozygote for HbE, was studied (Table I). Other family members were unavailable for detailed study.

Preparation and radioactive labeling of reticulocytes. Erthrocytes from50mlofperipheralbloodwerecollected by

centrifugation at2,000 rpm for10 minat2-4°C inaSorvall

RC-3 centrifuge (DuPont Instruments, Newtown, Conn.);

the cellswere washed threetimes in sterile isotonicsaline.

About5 mlof cellswassetasidein an icebath for incubation witheither[uS]methionineor[3H]leucine.Theremainderof

the cell pellet was suspended in ice-cold5 mMmagnesium chloride, shaken vigorously, and incubatedfor 1 minonice.

The hemolysates were then extracted withphenol for prep-aration of reticulocyte RNA, using methods previously de-scribed(17).

The kineticsofglobin synthesis were measured by pulse-chase incubationswith[3H]leucineor[35S]methionine as

fol-lows: aliquots of cells were resuspended to hematocrits of 30-35 in an incubation mixture containing Krebs-Ringer phosphate buffer, glucose, iron, and AB negative dialyzed plasma, exactly asdescribed(18);the suspension was prein-cubated for 5 min at370C with gentle shaking prior tothe

addition of isotope. For short-term, pulse-chase labeling studies,200

_ACi

of[3H]leucine wasadded to 3cm3 of pre-warmed cell suspension. The incubation was allowed to

proceed for 1 or 5 min at 37°C in a shakingwaterbath, at

which time aliquots of200 ,ul were removed and quickly frozeninacetone-dryicefor later globin analysis. 10 mM

non-radioactive leucinewasaddedtothe remaining cell

suspen-sion,andthe samplerapidlychilled in anice-waterbath.The

cells were pelleted at 0°C for5 min at 2,000 rpm in aSorvall RC-3centrifuge, and the cell pellet was resuspendedin pre-warmed incubation mixture containing 10 mM nonradio-active leucine, at a hematocritof30-35%. "Chase" incuba-tions werethen conductedat37°Cinashaking waterbath for thetimes indicatedinthe text. Ateachtimepoint,a400-,ul aliquot was removed,quick-frozen, and storedat -800C for globin chain analysis. Long-term incubations of reticulocytes from patient1wereconducted similarly, except that the

radio-active precursor was

[u5S]methionine

(sp act 4 mmol, 200 ,uCi/mlcell suspension).

Globin chain analysis. Each cell pellet from the pulse-chase labelingexperiment wasthawedin anequal volumeof

ice-cold 5mMmagnesium chloride,and theresulting

hemoly-sate either used directly to prepare globin or clarified by centrifugation for30 s at12,000 rpmin anEppendorftabletop centrifuge (EppendorfGeratebau,Netherler,WestGermany). Globin was prepared by acid-acetone precipitation, and globin chains were separated and quantitated by carboxy-methylcellulose chromatography, using methods which

we have described in detail elsewhere (17, 19). In

pre-liminary experiments, weverified that the globin chain bio-syntheticratios ofclarified lysateswere essentiallyequalto

TABLE I

Hb electrophoresis Reticulocyte

Hematocrit count Erythrocytes MCV HbA HbF HbE +Hb_Al

% % per100 WBC fl % %*

Patient 1 21 6 6 75* 69* 16* 15*

Father 46 2.6 0 84 63 37

Mother 37 1.2 0 69 94 6

Brother 41 0.3 0 71 68 32

Brother 40 0.3 0 83 97 3

Patient 2 17.2 10.6 4 52 <5 20 75

Hematologic findings in patients with Hb

E-,f-thalassemia

and available family members. None of patient 2'sfamily members were availableforstudy. However, family historywasnegativefor

trans-missionof severeanemia, except for theproband. Thefatherofpatient 1 hadthalassemic erythrocyte morphology despite an MCVof 84.WBC, leukocytes.

*_{Reflects contribution of transfused blood.}

those of the total hemolysate and membrane fractions

regard-less of whether samples from the Hb E patients or from nonthalassemic controls were examined. Over 90% of the radioactive globin was in the clarified lysate. No preferential accumulation of globin chains in the insoluble membrane frac-tion was observed (see Results, Table III). Globin chain biosynthetic ratios were calculated by comparing the total radioactivity present within optical density peaks defined by authentic nonradioactive globin chain markers included in thechromatographic analysis. In the case of [35S]methionine studies, the measured non-a/a ratios were doubled, to account for the presence of two methionine ratios in a-globin, but only one in the relevant non-a globins (20).

The production of 3E_globin by both patients was also verified by "fingerprint" mapping of peptides resulting from tryptic digests of nonradioactive globin from each patient, using methods previously described in detail (21). Globin from both patients yielded the characteristic fingerprint of patientsproducing both pA- and1E_globins: loss of intensity from the (iT3 peptide spot,and generation oftwonewspots, oneof whichisdetectableby arginine staining.Inbothcases, the two new spots migratedinthe characteristic positions of

,3E_globin.

These results were entirely typical of patients heterozygous forpA-and1E_globin. In addition,aportion of

theradioactive globinfrom both patients comigratedwith

au-thentic

,3E_globin

markeroncarboxymethylcellulosecolumns. Extractionandanalysis ofreticulocyteRNA. Total

retic-ulocyte cellularRNA waspreparedby phenolextractionand alcohol precipitationasdescribed previously by this laboratory (17, 19). RNA was desalted by centrifugation through a

Sephadex G-25 (coarse) syringe column (22), and used directly for further analysis without purification of globin mRNA

fractions. Cell-free translations inwheat germ extracts were

performed utilizing methods describedindetailelsewhere,as

was molecular hybridization analysis, utilizing radioactive, purifieda-and

,8-globin

cDNAhybridizationprobes (13-17, 19). Globin chainssynthesizedincell-free translation systems primed with reticulocyte RNA were analyzed by carboxy-methylcellulose chromatography after addition of authentic

0.8

0.6

C,,

0

Nl1 0.4

0.2

0

globin chainmarkers,asdescribedinthe precedingsection.

,3/a mRNAratios assessedby molecular hybridization were

obtained bysaturationhybridization analysis,comparingthe slopesobtained fromtitrationofequal amountsofa-and /8-cDNA by increasing amounts of each RNA sample, as

described elsewhere (14-16). However, hybridization reac-tions were conducted under conditions which we have previously shown (14-16) to prohibit cross-hybridization

among /8-, 8-, e-, and -y-mRNA and cDNA sequences. The molecular hybridization analyses donotdistinguish between

,A_

and 3E_mRNA,sincethesemRNA mostlikely differ only

inone or a fewbase pairs. Traegerandco-workers (23) have recently observed thatf3E_mRNAhybridizes efficientlyto

/8-cDNA probes with properties identical to ours. The

hybridization assay thus measures the total content of/8-mRNA sequences relative to a-mRNA(i.e., /3A +/E/a mRNA) and

doesnotallow individual quantitation oftherelativeamounts

of/3A_ and /E-mRNA(i.e.,/E//3AmRNA). Therefore, deficiency

of,BE_mRNA

canbeinferredonly ifthe/3A+/3E/a mRNA ratio is too low to be due solely to the I3AmRNA produced by the normal (/3A) or 3thaL_gene present in trans to the 3E_gene.

RESULTS

Characteristics of patients and initial rates

of

globin biosynthesis.

Table I shows the

hematologic

characteristics of the probands and their available relatives. Thetworelatives ofpatient 1with HbE trait

have typical hemoglobin electrophoretic values for Hb E (37 and 32%), well within the ranges reported by others (1-10). The mother ofpatient 1 has

8-thal-assemia trait with anelevated HbA2. Patient 2hadno

relatives available for study but exhibits typical

fea-tures of Hb

E-,f-thalassemia

with a low output

/3-thalassemia allele inherited in trans to the I83E-globin

allele. Thehighpercentage of HbA,low percentage of

8

a

6 2

~~~~~~~I

Io

0

0

0.

10 20 30 40 50 60

FRACTION NUMBER

FIGURE 1 Globin synthesis by intact reticulocytes from patient 1 with Hb E-13-thalassemia. Reticulocytes wereincubatedwith[35S]methioninefor1 min and the radioactiveproteins

frac-tionatedbycarboxymethylcellulose

chromatography,

asdescribedinthetext.The closed circles

indicate the opticaldensity at280nmofauthenticy-,

3A_,

8E_,anda-globin mRNA.Theopen circles indicate 35S countsperminute ineachfraction.

a

r

kA

11

1~3

E

INC U BAT ION 10 min 120min

28.0 - 56.0

(o-o)

- _(0-.) 240 - 48.0

20.0 - 40.0

3Hcpm

x

₁₀₃

Fraction

12.0

-8.0 _

4.0

0

24.0

16.0

8.0

0

30 40

FRACTION NUMBER

FIGURE 2 Globin synthesis by intact reticulocytes from patient 2 with Hb E-pB-thalassemia. Experimental conditionswereessentiallyasdescribedinFig. 1,except that[3H]leucinewasused

asthe isotope, and the incubationswerefor 10 min (opencircles)and 120 min (closedcircles).

HbE,and relatively high MCVofpatient1are dueto

thepresenceof transfused normal cells. Patients1and

2, and the father ofpatient 1, with Hb E trait,

con-sentedto furtherstudy.

Results obtained by analysis of globin radioactivity

after short-term incubations ofintactreticulocytes are

shown in Fig. 1 for patient 1 and Fig. 2for patient 2.

Patient 1 appears to have inherited a f+-thalassemia

geneintrans tothe

,IE_globin

gene,asindicated bythe

consistentPA/a biosynthetic ratios of 0.1-0.2 (Figs. 1

and 3). The 8-thalassemia gene in patient 2 is a

,13-thalassemia gene associated with lower p-globin out-put(BA/a= 0.02-0.05), _as shown in Figs. 2 and 4. In

both cases, the initial levels of 3BE_globin synthesis

during 1-or5-min"pulse" incubations (I3E/a = 0.20in

patient 1, and 0.33 in patient 2) were lower than

ex-pected on the basis of inheritance of a normally

functioning,3 globingeneintrans tothe

,8-thalassemia

allele; the ratios were comparable to results reported byothers(1-10).AsindicatedinTableII,weobserved

,8/a

ratios of >0.4 attributable to the normally func-tioning A_ or

8s-globin

gene inherited in trans to

3-thalassemia genes inpatients with 83-thalassemiatrait,

or Hb

S-,8-thalassemia,

or the f8s- and 83c-genes

co-inheritedinHbsickle cell disease. These findingsare

alsoconsistentwiththose of otherlaboratories (2, 3, 7, 24-29). As shown in Table II, total /8-globin chain synthesis (/3A +pE/a) observed in reticulocytes of the

two patients with Hb

_{E-fl-thalassemia}

fell into the

range observed among patients with homozygous 38-thalassemia.

As shown in Table III, >90% of the radioactive

globin synthesized in our experiments could be

re-coveredin clarified supernatant fractions. The globin biosynthetic ratios were nearly identical in total

hemolysates, membrane pellets, andsupernatant

frac-0.6 r

0.4

0.2

0 0 Intactcells A A mRNAtranslation

0 30 60 90

INCUBATION TIME (min)

FIGURE3 Kinetics ofpA_ and 'E-globin synthesisinpatient 1 withHb E-13-thalassemia. Datawereobtainedasdescribed inthetextandareillustratedinFig. 1. Opensymbols indicate

'BE/aratios,andclosedsymbols, PBA/airatios. Results obtained

from intact reticulocyte incubations are shown by circles;

resultsobtained by analysis of [35S]methionine protein; syn-thesized in a wheat germ cell-free system primed with

reticulocyte mRNAfrompatient 1areshown by triangles.

3-Thalassemia Phenotype Associated with HemoglobinE

0.7

0.6

0.5

A280

04

Units 0

(-)

0.3

0.2

0.1

0 0

I

0.6r

0.5

0.4

0.2-0.1

p As E/

PulsePoints * 0

Chase Points * 0

D o o

0/a

0.3 -0

-°

-0~~~~~~~~~~~~~~~

0

30 0 9 30

30 60 90 120 i50

1A80

210

240

270 300

INCUBATION TIME (minutes)

FIGURE4 Kinetics of /A and fE_globin synthesis in intact

reticulocytes from patient 2 with Hb E-,8-thalassemia. Data

were obtainedasdescribed inthetextandareillustrated in Fig. 2. /3A/a ratios are indicated by closed symbols, and

'BE/a

ratios, by open symbols. Pulse incubations (circles) were conducted for the times indicated on the abscissa. Chase incubations (squares)wereconducted under conditions described in Methods after a 5-min pulse incubation; the chase time is plotted along the abscissa exclusive of the 5-minpulseincubation time.

TABLE III

Total /35/a 1,3/a pA/,3E [3H]globin

Totalhemolysate 0.42 0.26 0.61 100

Clarified supernate 0.41 0.27 0.64 93.4

Membranefraction 0.51 0.28 0.55 5.8

Globin synthesis in Hb E trait. Peripheral blood (2 ml) from thefather of patient 1 was incubated with [3H]leucine and the cells lysed as described in Methods. Globin was prepared from total hemolysate, the clarified lysate (supernatant frac-tion)after centrifugationat 12,000 gfor 10 min, and the pellet

(membrane

fraction). Individual [3H]globins were analyzed

from each fraction by carboxymethylcellulose chromatog-raphy as described in Methods. "Total globin" refers to the percentage of total radioactive [3H]globin countsper minute (cpm) (sum of net cpmin 8A, 3E, anda-globin peaks) ineach

fraction,using the totalhemolysatecpm as anarbitrary standard

of 100%. Similar studies of patients 1 and 2 showed no detectable 8A or fE_globin in the membrane fraction at 1 or60 min.

tions. No selective precipitation of 3E_globin or Hb E was apparent.

The father ofpatient 1, who has HbE trait, synthe-sized less 3E-globin (,BE/a = 0.26) than 83Aglobin (PA/a = 0.42),asshownin TableIII. Nosequestration

of8E_globin on membranes was observed. The 3E/pA

biosynthetic ratio after 30 min incubation was 0.61

(Table III), a value which correlated well with the

relative proportions of Hb E and HbA in circulating reticulocytes (Hb E/HbA = 0.59). Similarcomparisons

ofbiosyntheticratios tocirculating Hb A and Hb E in

thetwoHb

E-,8-thalassemia

patients couldnotbemade because of the presence of transfused Hb A

erythro-cytes inpatient 1 and theverysmall amountsof Hb A

and 8A_globin synthesisinpatient 2. The results argue

against a posttranslational mechanism in Hb E trait

whichwould yielda 8E/pA biosyntheticratioof1.0;the

lower HbE/HbA ratio inthe bloodwould result froma

selective decline ofBE_globin subsequent to biosyn-thesis, aphenomenon not supportedbyour data.

Kinetics of 83E_globin synthesis in intact

reticulo-cytes. The possibility ofselective post-translational

catabolism

of,/E_globin

wasexaminedbykinetic

analy-sisofglobin synthesis. Globinchainbiosyntheticratios

obtained from each patientatvaryingtimesofpulseand chase incubation are displayed in Figs. 3 and 4. The

relative ratios of a, -A_, and 1E_globin synthesis

remained essentially unchanged duringbothpulseand

TABLE II

,3/aratio Patients

Syndrome studied ,3B/a 1c/a 13Ea/e s/a

Normal Hb A 5 0.97-1.0

Sickle cell anemia 4 0.99- 1.1

Hb SCdisease 1 0.50 0.51

Hb_{S-,&-thalassemia} 3 0.00 0.43-0.49

,3-Thalassemia

trait 4 0.42-0.55

Hb E trait 1 0.42 0.26

Hb

E-/8-thalassemia,

Pt.#1 1 0.15 0.24

Hb

_{E-,3-thalassemia,}

Pt.#2 1 0.05 0.33

Homozygous/3-thalassemia 13 0-0.38

Synthesis

off3A_globin

and

,/-globin

variantsinvariousnonthalassemic and

,B-thalassemic

reticulocytes. Data were obtained as described in the textand illustrated in Figs. 1 and2.Some of these results have been presentedinearliercommunications (13-17).

Datashowninthetable represententirerange of values obtained from30,90,or120-min incubations with radioactive amino acid.

chase incubationsofupto5 h. There was noevidence

for atime-dependent selective decline of IE_globin, a

trend which should be apparent for a structurally

abnormal globin exhibiting post-translational

in-stability. Thus, these data donotsupportthehypothesis that defective IE_globin production is mediated

primarilyatthe level ofposttranslational instability. In

all of these experiments, the amounts of y-globin synthesis observed were low despite relatively high

levels of HbFintheperipheral blood oftheprobands (Table I). These findings are also routinely observed

duringourstudies ofmanypatients withahomozygous

f3-thalassemia (data not shown), and probably reflect

the selective survival of Hb F-bearing cells (whichare

thus older, nonglobin synthesizing erythrocytes) in

nonsplenectomized

,3-thalassemic

patients.

Analysis of reticulocyte globin mRNA. Total cellu-lar RNA prepared from reticulocytes of each patient

was translated into globin chains in a wheat germ

extract cell-free system, as noted in Methods. The

globin biosynthetic ratios observed bythis technique

were similar to those observed in intact cells for

pa-tient 1 (cf. Fig. 3). Precise quantitation of globin biosynthetic ratios could not be computed for patient 2 because of relatively high background nonglobin radioactivity comigrating in this region of

chromato-gram. However, the 'BE/a ratio, including the

non-globin peaks, remained clearly <0.30 with thismRNA preparation. These data are summarized in Table IV.

The level of 13E_globin synthesis in the cell-free translation system primed with mRNA from the

pro-band was substantially lower than the levels of fs3

globin synthesisobtained from studies oftwopatients

with Hb S-f3-thalassemia (see also reference 30, 31). Theseexperiments indicate thatadefectintheamount

and/or function of I3E_globin mRNA occurs in these

reticulocytes, and is sufficiently severeto account for reduced I3E-globin production.

Toobtainmorereliablequantitativedataconcerning

the actual chemicalamounts ofa-and

,B-globin

mRNA

in reticulocytes from the probands, saturation hy-bridization analysis was performed as described in

Methods and shown in Fig. 5. Under the conditions of hybridization employed, cross-hybridization

be-tween f3E-globin mRNA and IB3A-globin cDNA is

virtually 100%,since a single base mutation does not

appreciably alter the stability of RNA-cDNA hybrids

in this system (23). The assay does not permit

individual quantitation of I3A_mRNA relative to 8E_

mRNA. In contrast, we have repeatedly shown

(13-16) that the hybridization conditions employed

re-producibly detect 83-mRNA without

cross-hybridiza-tion to y-mRNA. RNA from both patients exhibited a

striking decrease in the total /3 globin mRNA content (i.e., /3A+ JJE/a ratio). The total /3a mRNA ratios

ob-served were 0.28forpatient 1, 0.20forpatient2, and

0.49 in the patient with Hb E trait (Fig. 6). These results are compared with those of patients with

typical heterozygous and homozygous thalassemic disorders, and withnonthalassemic reticulocyte RNA

preparations in Figs. 6 and 7. The values observed forpatients with Hb

E-,/-thalassemia

fall inthe range typically observed for patients with homozygous ,3-thalassemia. The ratioobtained in Hb Etraitmatches

mostclosely those obtained forpatients heterozygous

for

,-thalassemia.

One caninfer a deficitin pE-globin

mRNAin thesepatients despite the indirectnature of the measurements, since, in each case, the 8-globin

allele inheritedin trans cannotbe solely responsible

for the lowratios observed. DISCUSSION

Ourstudies demonstrate that the /3-thalassemia

pheno-typeassociated with Hb E is due to aprimary

reduc-tion of I3E_globin synthesis resulting from decreased

TABLE IV

,8/aratios Patients

Syndrome studied 8A/a f'/a 3aE/a

NormalHb A 5 1.1-1.8

Sickle cell anemia 4 1.09-1.6

HbSCdisease 1 0.59 0.48

Hb

_{S-,3-thalassemia}

3 0.00 0.43-0.60

/3-thalassemiatrait 1 0.46

Hb

E-f8-thalassemia

(Patient 1) 1 0.15 0.23 HbE-,8-thalassemia(Patient 2) 1 <0.05 <0.30*

Homozygous 8-thalassemia 14 0-0.47

Globin messengerRNAtranslation in various nonthalassemic and/3-thalassemicconditions. mRNAtranslation analysis was conducted asdescribed in Methods. Data shown represent the range of8/a globin ratios obtained aftercell-free translation from all patients of each category studied.

* IE_globin product peak partially obscured by nonglobin background contaminant. Total

radioactivity under

3E_globin

peak was -28% of a-radioactivity.

100

50

z

0

N 0

0

Patient *2

100 0

0~~~~~~~~~~~~~~

50

o V

30

RNA INPUT (ng)

60

FIGURE5 Molecular hybridization analysis of Hb

E-,/-thal-assemia reticulocyte RNA. Reticulocyte RNA from each

pa-tient was prepared and analyzed as described in Methods. Each reaction contained the indicated amount ofRNA, and 1,000cpm ofa-cDNA (closed circles) or

f3-cDNA

(open cir-cles).Thespecificactivityof each cDNAwas50.4 Ci/mmol. The a-cDNA contained 5% f3-cDNA sequences, and the f8-cDNA,18% a-cDNAsequences.The early slopetransitions in

each,8-cDNAcurverepresenthybridization ofa-mRNAinthe RNA samplestothe a-cDNAcontaminant of

_jl-cDNA.

RNA from patient 2 achieved 70-75% saturation of the f8-cDNA probeat120 and 240ngRNAinputs. 8A+ 1E/amRNAratios

werecalculated from theRNAinputsrequiredtoachieve half saturationofthea-and /-cDNA probes. The values ofpercent

hybridization achieved at saturation with these mRNA and cDNAareidenticaltothoseobservedinparallelexperiments using these cDNA and well-characterized a-, /8-, and nonthalassemic mRNA (cf. reference 13-16, Fig.6).

accumulation

of,3E_globin

mRNA. We have eliminated

posttranslational mechanisms as possible causes for

the reduced Hb E production. Since the original sub-mission ofthis manuscript, Traeger et al. (23) have

reported similar conclusions. Both their results and

ours suggest that the I3E_globin gene behaves like a

mild /3-thalassemia gene, since f3-mRNA levels range

nearthe upperlimits observed for patients with

com-parablenumbers

of,/-thalassemia

genes(Fig. 7,Table IV). These observations are consistent with the mild

hypochromic anemia seen in homozygotes for Hb E

(1-10). However, the /3E-gene is sufficiently

"thalas-semic" to interact with a true ,8-thalassemia allele to

produce severe clinical phenotypes seen in Hb

E-,8-thalassemia. The pE_globin gene thus resembles the mild

,3-thalassemia

allele possessed by"silent carriers" of/8-thalassemia (reference 2, 3).

124 Benzet al.

10

5

C

u

0 ~~~~~~~0

0

(,~~~8A

+,

_P)

/a mRNA=049

I,

0 5 1

0 50 100 150 200

RNA INPUT (nanograms)

FIGURE 6 Molecular hybridization analysis of Hb E trait

reticulocyte RNA. RNApreparation andhybridization anal-ysiswereperformedasdescribed inFig. 5. Nonthalassemic

reticulocytes(A)wereobtainedfromapatientwith paroxysmal nocturnalhemoglobinuria, HbE trait reticulocytes (C)from

the father ofpatient 1, and /3-thalassemiatrait reticulocytes

(B)from anadult with typicalheterozygous /B-thalassemia.

Because of the recent influx of Southeast Asian

immigrants, Hb E is being encountered with in-creasing frequencyinWesternnations. /8-Thalassemia

is also very common in these populations (2, 3). For

purposes of genetic counseling, therefore, it is

im-portant to recognize the IE_globin gene as a mild

,8-thalassemia allele. The potential genetic combination

of Hb E and /8-thalassemia is of far greater clinical significance than potential homozygosity for Hb E. Hb E itself is apparently compatible with normal

oxygen delivery and erythrocyte homeostasis (9, 10), despitemildinstabilitydemonstrablein vitro (32, 33).

Our studies offamily1(Figs. 1 and3,TableIII)

sug-gestlower levels

of,3E_globin

synthesisthan inpatient

2 (Figs. 2 and 4). Similar heterogeneity of fEE/a

bio-A

100

50

0

I0

*_ a CDNA

0-o X3 CDNA

/a mRNA=1.0

B

100

50

z

0

N 0

m I O-R

I.0

r

0

I-cr

4

z

E a

0 0 0 0 S 0

0.51-0

0

0@

0 S 0

0

0 0

0 0 0 0"

I I I II _

NORMAL /-THAL HbS HbE HbE /-THAL TRAIT 3-THAL TRAIT 8-THAL MAJOR

FIGURE 7 /3 mRNA content in various /8-thalassemia syn-dromes.Datawereobtained by molecular hybridization anal-ysis as illustrated in Figs. 5 and 6. Some of the data for

patientswith homozygous 3-thalassemiawere derived from

our previous studies (13-16). Nonthalassemic (3/a mRNA ratios range from 0.8 to 1.0, reflecting the normal 5-20% excessofa-mRNAinreticulocytes.

synthetic

ratios

(0.2-0.4) has been encountered

by

others (1-10, 23). However, Hb E invariably

con-stitutes

30-39% of total circulating hemoglobin

in

heterozygotes (1-10, 23, present study), suggesting

technical

variation

rather than

true genetic

hetero-geneity of

3E_globin

alleles in

different

families. It

is

likely that

the

variability

results

from

the

close

proximity

of

the

/3E_globin

peak

to a-

and

pre-a-(carbamylated a-chains) peaks, which renders

precise quantitation

difficult. The

presence

of a-thalassemia

among

these

patients

could

also contribute

to

hetero-geneity,

but

cannot

explain the

apparent,8-thalassemic

behavior of the

/BE-gene

reported

in this communica-tion.

Indeed, a-thalassemia

would tend to "mask" the lowered

/3E_globin

and

/3E-globin

mRNA levels

ob-served

by lowering a-globin

and a-globin mRNA. The

recently described

condition in which more than the

normal

number of

a-genes are inherited (34)

could

conceivably

account

for

our

data

if the extra a-genes

were

functional.

However, these genesdo not appear to altera- and

/8-globin

balance inpatients studied thus

far (35).

In any event, this hypothesis cannot explain

the lower

production of

/3E_globin

than

/3A_globin

in

heterozygotes.

Hb E is an example of an increasing number of

structurally

abnormal globins that are also synthesized

at

abnormally

low rates. The Hb Lepore and Hb

Con-stant Spring

hemoglobins (2,

3, 35) are

the best known

examples. More recently, Hb Indianapolis, a /-chain mutant, has

been shown

to

produce

/8-thalassemia by

profound posttranslational

instability

of the newly

assembled ,8-variant, preventing combination with a-globin to form hemoglobin tetramers (18).

Finally,

wehave described a profound reduction in the mRNA

coding

for the

,8-chain

of Hb

Vicksburg

(/375beu0

_),

_which

exhibits a

/8-thalassemia phenotype

(39).

Our

studies do

not

explain

the mechanism for the

thalassemia-like

behavior of

/3E_globin

genes. It is

dif-ficult

to envision a precise

mechanism whereby

the

presumed single nucleotide

change responsible

for

the

amino

acid substitution

atposition26

should

cause

lower globin

mRNA

production.

However, the location

of the

mutation in codon

26,

near the junction of mRNA

coding and

intervening sequencesat

codon 30,

could lead

to

disruption

of intranuclear mRNA proc-essing, a

mechanism

nowknowntobe

responsible

for some forms

of

/8-thalassemia (36-38). Alternatively,

the

mutation

could disrupt the

secondary

structureof

/3E_mRNA,

causing

instability

or decreased

template

capacity with consequent

instability, mechanisms

also

implicated

in

/-thalassemia

(35).

Finally,

Hb E could represent a

"double

mutation,"

i.e.,

base substitution

occurring on a pre-existing

/3-thalassemia allele,

a

/8-thalassemia

occurring on a

/3E_globin

allele,

or a

re-combination

event between two such alleles. These

occurrences

would

not

be

surprising within a

popula-tion

where both conditions

are so

frequent.

The

high

frequency

of Hb E in Southeast Asiacould

conceivably

result

from

selection

for

the

/3-thalassemia phenotype.

The

amino

acid

substitution in Hb E may

be

an

incidental feature "carried" through

evolution by the

same

mechanism that

preserves

/-thalassemia

inmany groups.

Further

exploration of these possibilities

can

be achieved by studies of

mRNA metabolism

and

restriction

endonuclease polymorphisms,

using gene

blotting techniques (35).

ACKNOWLE DGMENTS

We are indebtedto Dr. Bernard G.Forget for muchhelpful

advice,encouragement, and supportthroughoutthesestudies,

toAlphonse L.Scarpafor excellenttechnical assistance,and to Ms. NancyGrinnell for preparation of the manuscript.

This work was supported by grant HL24385 from the National InstitutesofHealth, andaBasilO'Connor Fellow-ship to EdwardJ. Benz, Jr. from the National Foundation-March ofDimes.

REFERENCES

1. Fairbanks, V. F., G. S. Gilchrist, B. Brimhall, J. A.

Joreb, and E. G. Goldstein. 1979. Hemoglobin E trait reexamined: a cause ofmicrocytosis and erythrocytosis.

Blood. 53: 109-115.

2. Bunn, H. F., B. G. Forget, and H. M. Ranney. 1972. Hu-man Hemoglobins. W. B. Saunders Co., Philadelphia,

Pa. 140-192.

3. Weatherall,D.J.,andJ. B.Clegg. 1972. TheThalassemia

Syndromes. Blackwell Scientific Publishers, Oxford,

England. 75-144, 163-167.

4. Sturgeon, P., H. A. Itano, and W. R. Bergen. 1955. Clinical

manifestations ofinherited abnormal hemoglobins. II. InteractionsofhemoglobinEandthalassemiatrait.Blood.

10:396-404.

5. Chemoff,A. I., V. Minnich, S.Na-Nakom, S. Tuchinda, C. Kashemsant, and R. P. Chemoff. 1956. Studies of hemoglobin E. I. The clinical, laboratory, and genetic

characteristics of hemoglobin E syndrome.J. Lab. Clin. Med.47: 455-498.

6. Pagnier, J., H. Wajeman,and A. Labie. 1974. Defect in

hemoglobin synthesis possibly due to a disturbed asso-ciation. FEBS (Fed. Eur. Biochem. Soc.) Lett. 35: 252-255.

7. Weatherall,D.J., J.B.Clegg, S.Na-Nakorn,and P.Wasi. 1969. The pattern of disordered haemoglobin synthesis in homozygous andheterozygous 3-thalassaemia. Br. J. Haematol. 16: 251-257.

8. Feldman, R., and R. Rieder. 1973. The interaction of

hemoglobinEwithf3-thalassemia:astudy ofhemoglobin

synthesis in a family of mixed Burmese and Iranian origin.Blood. 42: 783-791.

9. Testa, U., A. Dubart, N. Hinard, F. Galacteros, W.

Vainchencker, P. Rouyer-Fessard, Y. Beuzard, and J. Rosa. 1980. f30-Thalassemia/Hb E association. Acta

Haematol. 64:42-52.

10. Fairbanks, V. F., R. Oliveros, J. H. Brandabor, R. R.

Willis, andR. F. Fiester. 1980.HomozygousHbEmimics

,3-thalassemia

minorwithout hemolysis or anemia. Am. J.Hematol. 8: 109-121.

11. Finer, M., J. R. Humbert, and S. L. Friedman, 1979.

Globin synthesisinhemoglobin E-,3o-thalassemia. Blood.

54: 53A.

12. Ruymann, F. B., L. A. Popejoy, and R. B. Bronillard. 1978. Splenic sequestration and ineffective

erythro-poiesis inhemoglobin

E-,8-thalassemia

disease.Ped.Res. 12: 1020-1023.

13. Benz, E. J. Jr., P. S. Swerdlow, and B. G.Forget. 1975.

Absence of functional messenger RNA activity for ,B

globin chainsynthesisin18-thalassemia.Blood.45: 1- 10. 14. Housman, D., B. G. Forget, A.Skoultchi, andE.J.Benz,

Jr. 1973. Quantitativedeficiency of chain-specific globin

messengerribonucleic acidsinthe thalassemia syndrome.

Proc.Natl. Acad. Sci. U. S. A.70: 1809-1813.

15. Forget, B.G.,D.Housman,E.J.Benz,Jr.,andR.P.

Mc-Caifrey. 1975. Synthesis of DNA complementary to

separated humanaandf8globinmessenger RNAs. Proc.

Natl.Acad. Sci.U. S.A. 72:984-988.

16. Benz, E. J., Jr., B. G. Forget,D. G.Hillman,M. Cohen-Solal, J. Pritchard, C. Cavallesco, W. Prensky, and D. Housman. 1978. Variability in the amount off globin

messenger RNA in,o-thalassemia. Cell. 14: 299-312. 17. Benz, E. J., Jr., and B. G. Forget. 1971. Defect in messenger

RNA forhuman hemoglobin synthesisin

f3-thalassemia.

J. Clin. Invest. 50:2755-2760.

18. Adams,J.G., III,L. A.Boxer,R.L.Baehner,B.G.Forget,

G. A.Tsistrakis, andM.H.Steinberg.1979.Hemoglobin Indianapolis

(,3112[G14]arginine):

an unstable ,B chain

variantproducing thephenotypeofseveref3-thalassemia.

J.Clin. Invest. 63:931-938.

19. Benz, E. J., Jr.,P. S.Swerdlow, and B. G. Forget. 1973.

Globinmessenger RNAinhemoglobinHdisease.Blood.

42: 825-833.

20. Nathan,D.G., H. F.Lodish,Y. W. Kan, and D. Housman. 1971.

,B-Thalassemia

andtranslationofglobinmessenger RNA. Proc.Natl.Acad.Sci. U. S. A. 68: 2514-2518.

21. Adams,J.G.,W.P. Winter, K.Tausk,and P.Heller. 1974.

Hemoglobin Rush (/3101[G3]glutamine): a new unstable hemoglobin causing mild hemolytic anemia. Blood. 43: 261-269.

22. Benz, E. J., Jr., C. E.Geist, A. W. Steggles, J. E. Barker,

and A. W. Nienhuis. 1977. Hemoglobin switching in

sheep and goats. VII. Purification and characterization ofcDNAs complementary to sheep a,,f,and y mRNAs of

sheep.J. Biol. Chem. 252: 1908-1916.

23. Traeger, J., W. G. Wood,J. B. Glegg, D. J. Weatherall,

and P. Wasi. 1980. Defective synthesis of Hb E is dueto reduced levels of

'BE

mRNA. Nature (Lond.). 288:

497-499.

24. Rieder, R. F. 1975. Variations inthesynthesis andcellular content of structurally distinct human hemoglobins. In

Sickle Cell Anemia and Other Hemoglobinopathies. Academic Press,Inc., NewYork. 25-49.

25. Kim, H. C., R. G. Weierbach, S. Friedman, and E. Schwartz. 1977. Detection ofsickle aor,-thalassemia

bystudies of globin biosynthesis. Blood. 49: 785-792. 26. Kim, H. C., R. G. Weierbach, S. Friedman, and E.

Schwartz. 1977.Globin biosynthesisinsicklecell,HbSC,

and Hb C diseases. J.Pediatr. 91: 13-18.

27. Steinberg, M. H., and B. J. Dreiling. 1976. Clinical,

hematologic and biosynthetic studies in sickle

cell-,80-thalassemia: acomparison with sickle cellanemia. Am. J. Hematol. 1: 35-44.

28. White,J. M. 1971. The synthesis ofabnormal hemoglo-bins. Semin.Haematol. 4: 116-130.

29. Benz,E. J., Jr., and B. G. Forget. 1974.Thebiosynthesis ofhemoglobin.Semin.Hematol. 11:463-523.

30. Nienhuis, A. W., P. H.Canfield,and W. F. Anderson. 1973.

Hemoglobinmessenger RNAfrom human bonemarrow:

isolationandtranslationinhomozygous and heterozygous

,f-thalassemia.

J. Clin.Invest. 52: 1735-1745.

31. Natta, C.,J. Banks, G.Niazi, P. A. Marks,and A. Bank.

1973.Decreased8globin'MRNA activity inhomozygous and heterozygous 3-thalassemia. Nature (New Biol.).

244:280-281.

32. Frischer, H., and J. Bowman. 1975. Hemoglobin E, an

oxidatively unstable mutation. J. Lab. Clin. Med. 85; 531-539.

33. Ali,M. A.M., A.Quinlan,and S. C. Wong. 1980. Identifica-tionofhemoglobin Eby theisopropanolsolubilitytest. Clin. Biochem. 13: 146-148.

34. Higgs, D.R.,J. M.Old,L.Pressley,J. B.Clegg,and D. J.

Weatherall. 1980. Anovelaglobingene arrangement in man.Nature(Lond.).284: 632-635.

35. Benz, E. J., Jr., andB.G.Forget. 1980.Thepathogenesis of the thalassemia syndromes. Pathobiol. Rep. 10: 1-33. 36. Maquat,L.E., A. J.Kinniburgh,L.R.Beach,G. R.Honig,

J.Lazerson, W. B.Ershlev,and J. Ross. 1980.Processing of

_P3

globin mRNA precursor to mRNA is defective in

three patients with

f3+-thalassemia.

Proc. Natl. Acad. Sci. U. S. A. 77: 4287-4291.

37. Kantor, J. A., P. H. Turner, and A. W. Nienhuis 1980. ,8-Thalassemia: mutations which affectprocessingof the /3globinmRNA precursor.Cell.21: 149-157.

38. Benz,E.J., Jr.,A.L.Scarpa,J.K.deRiel,B. L.Tonkonow,

A.K.Ritchey, and B.G.Forget. 1981.Globinmessenger RNA metabolism in normal and thalassemic erythro-blasts.In Hemoglobinsin Differentiation and

Develop-ment. A. W. Nienhuis and G. Stamatoyannopoulos,

editors. AlanR. Liss,Inc.,NewYork. In press.

39. Adams, J. G. III, M. H. Steinberg, M. V. Newman,

T: W. Morrison, E. J. Benz, Jr., and R. Iyer. 1981. 13-Thalassemia present in cis to a new/8 chain structural variant: Hb Vicksburg (/375[E19]Leu-*0). Proc. Natl.

Acad. Sci. U.S. A.78: 469-473.