Proc. Nat!. Acad. Sci. USA
Vol. 77, No. 7, pp. 4287-4291, July 1980
Processing ofhuman f3-globin mRNA precursor to mRNA is
defective in three patients with#+-thalassemia
(nucleated erythroid cells/globin RNA pulse-chase kinetics/RNA splicing)
LYNNE E. MAQUAT*, ALAN J. KINNIBURGH*, LARRY R. BEACH*t, GEORGE R. HONIGt,
JACK LAZERSONO, WILLIAM B. ERSHLER11, AND JEFFREY Ross*
*McArdle Laboratory for Cancer Research, University of Wisconsin, 450 North Randall Avenue, Madison, Wisconsin 53706; *Children's Memorial Hospital, 2300
Children's Plaza, Chicago, Illinois 60614; §Milwaukee Children's Hospital, 1700 West Wisconsin Avenue, Milwaukee, Wisconsin 53233; and; HDepartment of
Medicine,Center for the HealthSciences, Madison, Wisconsin53706
Communicated by Howard M. Temin, March 10, 1980
individuals and from three patients with homozygous
lassemia were pulse-labeled with tritiated nucleosides. The
processing of the newly synthesized globin mRNA precursors
was monitored by inhibiting additional transcription with ac-
tinomycin D for 30 min. Human"globinmRNA is derived from
itsprecursor via a series of reactions that generate processing
intermediates. In nonthalassemic cells the precursor is prssed
efficiently to mature mRNA during the chase. In contrast, in
#+-thalassemiccells the processing of#-globinRNA is defective.
In one patient theAglobinmRNA precursor turns over during
the chase, but some of the interm
are not processed to mRNA. In two other patients a large frac-
tion of the precursor and intermediate RNAs is not processed
to mRNA. The a-globinmRNA precursor and intermediates are
processed efficiently to mRNA-sized molecules in thalassemic
and normal cells. The reduction in the rate of -globin but not
a-globin RNA processing accounts for the a/,B globin mRNA
imbalance in thalassemic erythroid cells. We discuss the pos-
sibility that the genetic lesions in +-thalassemia are at splicing
signal sites within intervening sequences of the ,-globin
Nucleated bone marrow cells from normal
iate RNAs accumulate and
Thef3-thalassemiasare a heterogeneous group of hereditary
anemias in man in which /3-globin protein synthesis is decreased
or absent and a-globin protein synthesis occurs at normal rates
(1-6). Erythrocytes from patients with homozygousf30-tha-
lassemia contain no detectablef3-globinprotein, and their re-
ticulocytes either lack f3-globin mRNA or containf3-globin
mRNA that is not translated into functional protein (7-9). The
f3-globin genes are intact in most, but not all, #°-thalassemic
patients (4, 5, 10). Erythroid cells from patients with f3+-tha-
lassemia contain structurally normalf3-globinprotein, but there
is significantly lessf3-globinthan a-globin protein and the re-
ticulocyte a/f3-globin mRNA ratio is increased (11-14).
f3-GlobinmRNAs in /3+-thalassemic and normal reticulocytes
are translated with equal efficiencies (15-17), and thef3-globin
genes in /3+-thalassemia do not contain detectable deletions (4,
Based on these data, the fl+-thalassemia phenotype must
result from mutations that specifically reduce the expression
of thef3-globinstructural genes. Such mutations might affect
gene transcription, processing of themRNA precursor, transport
of mRNA from the nucleus to the cytoplasm, or stability of
mature mRNA in the cytoplasm. The ratios of a-globin to
f3-globinmRNA sequences in nuclear RNA from normal indi-
viduals and from two patients with 3+-thalassemia were ap-
proximately equal (13). This result suggests that the tran-
scription rates of the two genes are similar in some patients.
We have investigated the possibility that the 3-globin mRNA
precursor is processed to mRNA inefficiently in the f+-tha-
lassemias. Nucleated bone marrow cells were pulse-labeled with
3H-labeled nucleosides and chased with actinomycin D, and
globin-specific RNA was analyzed. The results indicate that the
processing of f3-globin RNA, but not a-globin RNA, occurs
inefficiently in three unrelated patients with#+-thalassemia.
Globin Chain Synthesis, Cell Culture, and RNA Purifica-
tion. Peripheral blood cells were incubated with [14C]leucine.
Globin proteins were fractionated by CM-cellulose chroma-
Nucleated cells were obtained from bone marrow aspirates,
and all manipulations prior to cell culturing were performed
at 240C. Each aspirate was passed through a series of progres-
sively larger gauge needles and centrifuged for 10 min at 300
X g. The buffy coat was removed and brought to 5 ml with
growth medium (RPMI-1640 containing 20% fetal calf serum
and 50 units of penicillin and 50,gof streptomycin per ml).
Cells were homogenized by gentle pipetting, transferred to five
1-ml Wintrobe tubes, and centrifuged for 10 min at 300 X g.
The nucleated cells were pooled, brought to 10 ml with growth
medium, and centrifuged for 12 min at 200 X g. The cells were
resuspended in medium to 2-10 X 107 nucleated cells per ml
and incubated in a humidified atmosphere of 5% C02/95% air
at 370C for 1 hr, at which time [3H]uridine, [3H]cytidine, and
[3H]guanosine (ICN) were added to a final concentration of 500
.uCi/mleach (1 Ci = 3.7 X 1010 becquerels) (19). After 12 min,
actinomycin D (Calbiochem) was added to 10 jg/ml, and a
portion of the culture (the "pulse" fraction) was harvested. After
further incubation (10 or 30 min), an equal volume of culture
fluid was harvested (the "chase" fraction). Immediately after
the harvest, the cells were chilled, pelleted, and lysed, and total
cell RNA was extracted (19).
RNA Fractionation and Hybridization to Filter-Immo-
bilized DNA. Prior to electrophoresis, a portion of each RNA
sample was hybridized to [3H] human globin cDNA, and the
quantity of steady-state globin mRNA was calculated by Cot
analysis (Cot is concentration of nucleotide in mol/liter times
time in sec). RNA electrophoresis in 4.5% polyacrylamide/98%
The publication costs of this article were defrayed in part by page
charge payment. This article must therefore be hereby marked "ad-
vertisement" in accordance with 18 U. S. C. §1734 solely to indicate
Abbreviation: nt, nucleotides.
tPresent address: Department of Biochemistry, University of Wash-
ington, Seattle, WA 98195.
¶Present address: Department of Pediatrics, University of California
Davis Medical Center, Sacramento, CA 95817.
Medical Sciences: Maquat et al.
synthesis in PB
D. S. and D. K. receive transfusions monthly and were analyzed just prior to receiving blood. T. T. does not receive transfusions regularly
and received no blood during the period ofstudy. Hb, hemoglobin; MCV, mean cell volume; HbF, fetal human hemoglobin (a2y2);HbA, major
adult human hemoglobin(a2A2);HbA2, minor adult human hemoglobin (a262); E:M, erythroid cell-to-myeloid cell ratio; BM, bone marrow;
PB, peripheral blood.
formamide tube gels, gel fractionation, and RNA elution were
performed as described (20). The recovery of RNA from gel
slices did not vary with molecular weight (unpublished obser-
vations; see also ref. 20).
pMB9 DNA or recombinant pMB9 DNAs containing human
a- (JW101), A- (JW102), or -y-globin cDNA (JW151) (21) were
amplified and purified (22) under P2/EK1 containment. Pu-
rified plasmid DNA was immobilized on nitrocellulose filters
(12.5jigof DNA per 5-mm-diameter filter) (23). Each filter
contained at least 5-fold excess of cDNA sequences relative to
the quantity of globin mRNA used in each hybridization re-
action. Analytical RNA hybridization to filter-bound DNA was
performed as described (24), except that the hybridizing re-
action was performed at 65-690C, the RNase step was omitted,
and the hybridized RNA was eluted at 90-950C. Each reaction
tube contained three filters (5 mm diameter, circles): one with
pMB9 DNA to serve as a background for nonspecific binding
of RNA, one withJW101 (a) DNA, and one with eitherJW102
(j3) or JW151 ('y) DNA. The data in Figs. 1-4 represent the
radioactivity (dpm) bound to the cDNA filter minus that bound
to the pMB9 filter. For all hybridization reactions the pMB9
background was 0-20 dpm. The dpm were determined by the
external standard ratio method, and the counting efficiency was
45-50% for all samples. The extent of hybridization of human
RNA to filters was not determined, but approximately 80% of
mouse globin mRNAs formed hybrids with filter-bound mouse
globin cDNA under similar conditions (table 3 in ref. 24).
Hematological Characteristics of Patients. Hematologic
studies on three homozygous patients with f3+-thalassemia are
summarized in Table 1. D. S., a man of Greek origin, and D.
K., a girl of Italian-American descent, require monthly trans-
fusions to maintain a hemoglobin level of 9 g/dl (normal, 12-14
g/dl). T. T., a woman of Italian descent who was studied pre-
viously (25), has mild homozygous (3+-thalassemia and does not
receive transfusion regularly. Thea-globin/f3-globinprotein
synthesis imbalance is a characteristic feature of
Processing of Globin mRNA Precursors in Cells of Non-
thalassemic Individuals. In murine erythroid cells, the a- and
j3-globin genes are transcribed into large precursor RNAs that
are processed stoichiometrically to globin mRNA within 30 min
(19, 20, 26-28). Initial experiments with human cells were
performed to determine if human globin mRNA precursors
exist and, if so, to measure the efficiency with which they are
processed. Nucleated bone marrow cells from a patient with
drug-induced hemolytic anemia and from normal individuals
were pulse labeled for 12 min and chased for 30 min. RNA was
electrophoresed in denaturing gels, and the RNA eluted from
gel slices was hybridized to a- and 3-globincDNA filters. Peaks
of newly synthesized fl-globin RNA-approximately 1900-
2000 nucleotides (nt), 1550 nt, 1350 nt, 1150 nt, 950 nt, 900 nt,
and mRNA-sized molecules (680-780 nt)-were observed (Fig.
1C). Similar RNA profiles were observed in two normal indi-
Summary of pulse-chase experiments
Radioactivity in a-globin mRNA sequences:
Radioactivity infl-globinmRNA sequences:
Total radioactivity ratio:
Cells were pulse-labeled for 12 min and chased withactinomycinD for 30 min. Thepercentage (tothe nearest5%)ofa- andf3-globinmRNA
sequences in each RNA class was calculated fromFigs. 1-4,as follows: (dpmin RNA size class XMrofglobin mRNA/MrofRNA sizeclass).
*The total radioactivity in mRNA sequences is the sum of the mRNA-specific radioactivity in each RNA size class. The values reflect the
erythroid-to-myeloid cell ratio, the labeling efficiency of the various celltypesin themarrow,and the amount ofRNAelectrophoresed (see
legends to Figs. 1-4).
Proc. Natl. Acad. Sci. USA 77(1980)
Proc. Natl. Acad. Sc. USA 77 (1980)
Pulse-chase analysis of a- andfl-globinRNAs from bone
marrow cells of a nonthalassemic individual with drug-induced he-
molytic anemia. Approximately 6 X 108 nucleated bone marrow cells
were cultured at 2 X 107 cells per ml. The cells were pulse labeled and
chased, and the RNA was extracted and purified. RNA (175 Ag/gel)
was coelectrophoresed with 18S mouse rRNA (2000 nt), 16S Esche-
richia coli rRNA (1550 nt), and 12S brome mosaic virus component
4 RNA (800 nt). RNA eluted from each gel slice was hybridized to
DNAs (pMB9 and globin cDNA cloned into pMB9) immobilized on
filters. The radioactivity bound to the pMB9 filter (0-20 dpm) was
subtracted from that hybridized to the cloned cDNA filters. The
cDNA-specific dpm are shown. RNA larger than 3000 nt and smaller
than 450 nt failed to hybridize to the a or ,B filters (data not shown).
(Left) Hybridization of a-globin cDNA to RNA from cells pulse-
labeled for 12 min (A) and cells chased for 30 min (B). (Right) Hy-
bridization offl-globincDNA to RNA from cells pulse-labeled for 12
min (C) and cells chased for 30 min (D).
viduals who did not have hemolytic anemia (data not shown).
In some gels, one or more of the intermediate-sized RNAs mi-
grated as shoulders on adjacent RNA peaks. ,B-Globin RNAs
larger than 1900-2000 nt were not detected. Therefore, this
RNA is designated as the human ,B-globin mRNA precursor.
Approximately 60% of the pulse-labeled #-globin mRNA
sequences was detected in molecules that were larger than
mRNA (Fig. 1; Table 2). After the chase 85% of the f3-globin
radioactivity was present as mature mRNA (Fig. 1D). Little,
if any, precursor or intermediate-sized RNAs remained after
the chase. We conclude that the precursor and intermediate-
sized molecules in these nonthalassemic individuals are sub-
strates in stepwise processing reactions that generate ,B-globin
mRNA with high efficiency. The experiments do not allow us
to define precisely the processing pathway or its mechanism.
Peaks of a-globin-specific RNA from pulse-labeled cells were
also observed (Fig. 1A; similar data from two other normal
individuals are not shown). The largest RNA (approximately
1100 nt) is designated the precursor of a-globin mRNA. In-
termediate-sized molecules, approximately 900-1000 nt, and
mature a-globin mRNA (630-780 nt) were alsb detected (Fig.
1A). The precursor and intermediate-sized RNAs accounted
n elo 0mn() Rgt
Tale2. hs eul ndcte ta teRNsof9()100
Pulse-chase analysis of a- and
si-globin RNAs from bone
marrow cells of D. S. Nucleated bone marrow cells (5.3 X 108) were
cultured at 1.8 X iO7cells per ml, pulse-labeled, and chased, and the
RNA (12e noggel) was electrophoresed, fractionated, and hybridized
to a- and
pc-globin-specific cDNA probes. The cDNA-specific dpm
shown represent the radioactivity bound to the cDNA filter minus
that bound to the pMB9 filter (0-20 dpm). (Left) Hybridization of
a-globin cDNA toRNA from cells pulse-labeled for 12 mn (A), cells
chased for 10 mn (B), and cells chased for 30 min (C). (Right) Hy-
bridization of n)-globin cDNA to RNA from cells pulse-labeled for 12
(D), cells chased for 10 min (E), and cells chased for 30 mi
for 20% of the radioactive a-globin RNA, most of which was
processed to mRNA-sized moleculesduringthe chase(Fig. 1B;
Table 2). This result indicates that the RNAs of 900-1100 nt are
substrates in processing reactions that generate a-globin
Processingof (3.Globin xBRNA Precursor to mRNA Occurs
Inefficiently in Cells from Patients with#+-Thalassemia.
The profile ofnewly synthesized f3-globinRNA from D. S., a
Greek patient with severef3+-thalassemia,was similar to that
of the nonthalassemic individuals (compare Figs. IC and 2D).
Theprecursorand intermediate-sized RNAs accounted for 50%
of the radioactivej3-globinRNA (Table 2). Theprecursor was
not observed as a prominent peak after a 10-mmn chase, indi-
catingthat most of theprecursormolecules had beenprocessed
(Fig. 2E). However, alargefraction of the intermediate-sized
(900-1200 nt) RNAs was notprocessed to mRNA after chases
of 10 min (Fig. 2E) or 30 min (Fig. 2F), at which timethey
accounted for 55% of the radioactive#-globinRNA (Table 2).
These results indicate thatf3-globinmRNA processing inter-
mediates in D. S. are not convertedefficiently to mRNA.
Medical Sciences: Maquatet al.
Medical Sciences: Maquat et al.
The molecular weights of pulse-labeled a-globin RNAs in
cells of D. S. were similar to those in normal individuals (com-
pare Figs. 1A and 2A). The precursor and intermediate-sized
RNAsaccounted for 45% of the pulse-labeled a-globin mRNA.
After 10-min and 30-min chases, 70% and 95%, respectively,
of the newly synthesized a-globin RNA was processed to
mRNA (Fig. 2 B and C; Table 2). These results indicate that the
a-globin mRNA precursor and intermediates in D. S. are pro-
cessed to mRNA efficiently and without significant wastage.
In D. K., a patient with severe#+-thalassemia,pulse-labeled
fl-globinRNAs were comparable in size to those in nonthalas-
semic cells (Fig. 3C). The precursor and intermediate-sized
RNAs accounted for 65% of the newly synthesizedf3-globin
RNA (Table 2). A portion, but not all of the precursor RNA was
processed during the chase, in contrast to D. S. (Figs. 2F and
3D). The larger of the intermediate-sized RNAs (1300-1600
nt) also turned over, but the smaller intermediates (900-1200
nt) were not processed and accounted for 50% of the total ra-
dioactivef3-globinRNA. The quantity of newly synthesized
fl-globin mRNA-sized molecules increased very little (10%)
during the chase.
The sizes of pulse-labeled a-globin RNA molecules in D. K.
were similar to those of normal individuals (Fig. 3 A and B).
They did not turn over as rapidly in this patient as in the normal
control or in D. S. but they were processed more efficiently than
the#l-globinRNAs (Table 2). Pulse-labeled y-globin RNAs
were aso assayed, to investigate y-globin RNA processing and
to monitor the extent of cross-hybridization of y-specificRNA
to (3-globin cDNA. Because the quantity of y-globinRNA (Fig.
3E) was 5- to 10-fold lower than that offl-globinRNA (Fig. 3C),
cross-hybridization of y-globinRNA tof3-globincDNA could
not account for thef3-globinRNA peaks.
In the pulse-labeled cells of T. T., 45% of the radioactive
f3-globinRNA sequences was in precursor and intermediate-
sized RNAs (Fig. 4C; Table 2), most of which were not pro-
cessed and accounted for 80% of thef3-globinRNA after the
chase (Fig. 4D). Because the percentage of (3-globin mRNA-
sized molecules actually decreased to almost one-half during
the chase (Table 2), it will be of interest to determine if this
mRNA-sized material includes cleavage products generated
from precursor or intermediate molecules during the pulse; if
so, these products might be ligated during the chase to re-form
the larger RNAs (in progress). The ca-globin mRNA precursor
and intermediate-sized RNAs were processed as efficiently in
T. T. as in nonthalassemic individuals (Fig. 4 A and B). Pulse-
labeled y-globin RNAs were also processed efficiently (Fig. 4
E and F).
bone marrow cells ofD. K. Nucleated bone marrow cells (1.1 X 109)
were cultured (3.6 X 107 cells per ml) and further processed as de-
scribed in the legend to Fig. 1. The cDNA-specific dpm are shown and
represent the radioactivity bound to the cDNA filter minus that
bound to the pMB9 filter (0-20 dpm). RNA larger than 3000 nt or
smaller than 450 nt did not hybridize (data not shown). (Left) Hy-
bridization of a-globin cDNA to RNA from cells pulse-labeled for 12
min (A) and cells chased for 30 min (B). (Right) Hybridization of
f3-globincDNA to RNA from cells pulse-labeled for 12 min (C) and
cells chased for 30 min (D). Hybridization of y-globin cDNA toRNA
from cells pulse-labeled for 12 min (E).
Pulse-chase analysis of a-,
(3-, and y-globin RNAs from
bone marrow cells ofT. T. Nucleated cells (2.7 x 109) were cultured
(9 X 107 cells per ml) and analyzed as described in the legend to Fig.
1. The cDNA-specific dpm are shown and represent the radioactivity
bound to the cDNA filter minus that bound to the pMB9 filter (0-20
dpm). (Left) Hybridization of a-globin cDNA to RNA from cells
pulse-labeled for 12 min (A) and cells chased for 30 min (B). (Right)
Hybridization off3-globincDNA to RNA from cells pulse-labeled for
12 min (C) and cells chased for 30min (D). Hybridization of y-globin
cDNA toRNA from cells pulse-labeled for 12 min (E) and cells chased
for 30 min (F).
Pulse-chase analysis of a-, (3- and y-globin RNAs from
a- GLOBIN SPECIFIC RNA
200Ont I5OOnt SOOnt
- GLOBIN SPECIFIC RNA
a Iobin mRNA
a - globin mRNA
-Y-GLOBIN SPECIFIC RNA
Proc. Nati. Acad. Sci. USA 77(1980)
Proc. Natl. Acad. Sci. USA 77 (1980)
These data indicate that imbalanced globin chain synthesis in
these three#+thalassemic individuals can be accounted for by
defective processing of the (3-globin mRNA precursor to
mRNA. In three nonthalassemic individuals, 75-95% of the
radioactive a- andf3-globinmRNA sequences was detected in
mRNA-sized molecules after the 30-min chase (Fig. 1 and data
not shown). In contrast, 55% (at most) of the (3-globin RNA was
processed to mRNA-sized molecules in thalassemic cells (Figs.
2-4). The processing defect is restricted tofl-globin RNAs,
because a-globin (and, in T. T., y-globin) mRNA precursors
were processed efficiently to mRNA.
The quantities of newly synthesized fl-globinRNA in the
pulse and chase fractions were similar in most experiments
(Table 2). It is unclear why in some experiments there was al-
most 2-fold more globin mRNA-specific radioactivity in the
chase than in the pulse (Table 2; Figs. 2 D-F and 4A and B),
but similar observations have been noted in experiments with
mouse erythroid cells (19). It is also unclear why somefl-globin
precursor and intermediate molecules turn over during the
12-min pulse but others are not processed during-the subsequent
30 min. Perhaps the mutations in the paternal and maternal
fl-globinalleles of these patients are different, so that the pre-
cursors transcribed from each allele are processed differently.
It will be of interest to determine the ultimate fate of these
It is important to note that the precursor and the largest in-
termediate purified from cells of patient D. S. hybridize at least
10-fold more efficiently to
(unpublished observations). Therefore, it is unlikely that we are
analyzing 3-globin RNA processing. We propose that the ge-
netic defects in at least some
3+-thalassemias are mutations in
one or both of the intervening sequences (introns) of the
fl-globin gene (29-31). Assuming that the humanf3-globin
mRNA precursor undergoes multiple splicing reactions, as in
the mouse (24), the processing defect might vary from patient
to patient, accounting for the heterogeneity of the I3+-thalas-
semia syndromes and for the variety of processing patterns
observed in these experiments (Figs. 2-4). In other words, a
mutation at one splicing signal site may affect processing to a
greater extent than a mutation at another site. Processing defects
in some I3+-thalassemias could result also from mutations in
enzymes that processfl-globinRNA specifically (32,33) or in
structural proteins that bind exclusively to thefl-globinmRNA
precursor or cleavage intermediates (34). Moreover, we have
not excluded the possibility of mutations that reduce both the
rates of processing and of transcription.
We thank Milton Johnson for his contributions to the preliminary
aspects of this work and Lorri Steffen and Deborah Hass for excellent
technical assistance. We are grateful to Drs. Edward Prendergast and
Michael Borzy for performing bone marrow aspirations, Dr. John
Pellett for providing bone chips, Drs. Bernard Forget and Sherman
Weissman for human globin cDNA clones, Dr. Tom Maniatis for
human globin genomic clones, and Drs. Gary Altman and Paul
Kaesberg for brome mosaic virus. This investigation was supported by
Grants CA-07175 and CA-23076 from the National Institutes of Health
L.E.M., A.J.K., L.R.B., and G.R.H. were supported by Grants CA-
09230, CA-23076, CA-91935, and AM-19046, respectively, from the
National Institutes of Health.
Weatherall, D. J. & Clegg, J. B. (1979) Cell 16,467-479.
Bank, A. & Ramirez, F. (1978) CRC Crit. Rev. Biochem. 5,
/3- than to 6-globin genomic DNA
Forget, B. G. (1978) Trend Biochem. Sci. 3,86-89.
Flavell, R. A., Bernards, R., Kooter, J. M., de Boer, E., Little, P.
F. R., Annison, G. & Williamson, R. (1979) Nucleic Acids Res.
Orkin, S., Old, J. M., Weatherall, D. J. & Nathan, D. G. (1979)
Proc. Natl. Acad. Sci. USA 76,2400-2404.
Chang, J. C. & Kan, Y.-W. (1979) Proc. Natl. Acad. Sci. USA 76,
Ramirez, F., O'Donnell, J. V., Marks, P. A., Bank, A., Musumeci,
S., Schiliro, G., Pizzarelli, G., Russo, G., Luppis, B. &,Gambino,
R. (1976) Nature (London) 263,471-475.
Old, J. M., Proudfoot, N. J., Wood, W. G., Longley, J. I., Clegg,
J. B. & Weatherall, D. J. (1978) Cell 14,289-298.
Benz, E. J., Jr., Forget, B. G., Hillman, D. G., Cohen-Solal, M.,
Pritchard, J., Cavallesco, C., Prensky, W. & Housman, D. (1978)
Tolstoshev, P., Mitchell, J., Lanyon, G., Williamson, R., Otto-
lenghi, S., Comi, P., Giglioni, B., Masera, G., Modell, B.,
Weatherall, D. J. & Clegg, J. B. (1976) Nature (London) 259,
Housman, D., Forget, B. G., Skoultchi, A. & Benz, E. J., Jr. (1973)
Proc. Natl. Acad. Scd. USA 70,1809-1813.
Kacian, D. L., Gambino, R., Dow, L. W., Grossbard, E., Natta,
C., Ramirez, F., Spiegelman, S., Marks, P. A. & Bank, A. (1973)
Proc. Natl. Acad. Sci. USA 70,1886-1890.
Nienhuis, A. W., Turner, P. & Benz, E. J., Jr. (1977) Proc. Natl.
Acad. Scd. USA 74,3460-3464.
Kazian, H. H., Ginder, G. D., Snyder, P. G., van Beneden, R.
J. & Woodhead, A. P. (1975) Proc. Natl. Acad. Sci. USA 72,
Nienhuis, A. W. & Anderson, W. F. (1971) J. Clin. Invest. 50,
Benz, E. J. & Forget, B. G. (1971) J. Clin. Invest. 50, 2755-
Rieder, R. F. (1972) J. Clin. Invest. 51,364-372.
Honig, G. R., Mason, G. R., Tremaine, L. M. & Vida, L. N. (1978)
Am. J. Hematol. 5,335-340.
Ross, J. (1976)J.Mol. Biol. 106,403-420.
Ross, J. & Knecht, D. A. (1978)1. Mol. Biol. 119, 1-20.
Wilson, J. T., Wilson, L. B., deRiel, J. K., Villa-Komaroff, L.,
Forget, B. G. & Weissman, S. M. (1978) Nucleic Acids Res. 5,
Humphreys, G. O., Willshaw, G. A. & Anderson, E. S. (1975)
Biochem. Biophys. Acta 383,457-463.
Melli, M., Ginelli, E., Corneo, G. & Lernia, R. D. (1975)J.Mol.
Kinniburgh, A. J. & Ross, J. (1979) Cell 17,915-921.
Heller, P., Yakulis, V. J., Rosenzweig, A. I., Abildgaard, C. F. &
Rucknagel, D. L. (1966) Ann. Intern. Med. 64,52-61.
Curtis, P. J. & Weissmann, C. (1976) J. Mol. Biol. 106, 1061-
Kwan, S.-P., Wood, T. C. & Lingel, J. B. (1977) Proc. NdtI. Aced.
Sci. USA 74,178-182.
Bastos, R. N. & Aviv, H. (1977) Cell 11,641-650.
Lawn, R. M., Fritsch, E. F., Parker, R. C., Blake, G. & Maniatis,
T. (1978) Cell 15, 1157-1174.
Mears, J. G., Ramirez, F., Leibowitz, D. & Bank, A. (1978) Cell
Flavell, R. A., Kooter, J. M., deBoer, E., Little, P. F. R. & Wil-
liamson, R. (1978) Cell 15,25-41.
Hutchison, H. T., Hartwell, L. H. & McLaughlin, C. S. (1969)
J. Bacteriol. 99,807-814.
Hopper, A. K., Banks, F. & Evangelidis, V. (1978) Cell 14,
Toniolo, D. & Basilico, C. (1974) Nature (London) 248, 411-
Medical Sciences: Maquatet al.