main (locus control region) in transgenic mice. The 
human a-globin transgenes in these mice are appro- 
priately expressed, the ^globin gene during the em- 
bryonic period and the a-globin gene in the fetal/ 
adult stage. Analysis of mouse lines carrying 
more-limited segments of the cluster demonstrate 
that this developmental control is autonomous; i.e., 
the normal expression pattern of the a-globin gene 
is not dependent on the presence of the embryonic 
(r)-globin gene in cis. 
These studies have been extended by generating 
transgenic lines carrying more-limited segments of 
the a- and ^globin genes or chimeric genes in 
which the a- and f-globin gene promoters are either 
reciprocally exchanged or placed under the tran- 
scriptional control of a nonerythroid reporter gene. 
These studies are aimed at defining the limits of the 
cis-acting developmental control elements. 
Additional insight into the developmental switch- 
ing in the a-globin gene cluster has been derived 
from concurrent analysis of naturally occurring a- 
thalassemic mutations. In these studies it was shown 
that a subset of large deletions within the a-globin 
cluster can result in persistence of f-globin gene ex- 
pression. These data suggest that some of the deter- 
minants of ^globin gene silencing maybe located as 
much as 1 0 kb 3' of the ^globin gene itself. 
The selective enrichment of globin mRNAs in the 
developing erythroblast is dependent on the un- 
usual stability of globin mRNAs. Based on the analy- 
sis of informative a-thalassemic mutations that de- 
stabilize globin mRNA, it was hypothesized that 
critical determinants of a-globin mRNA stability are 
located in the 3'-nontranslated region. This hypoth- 
esis was confirmed by studies using a gene transfec- 
tion system in which normal or mutated human a- 
globin genes could be expressed and the stability of 
their encoded mRNAs determined. Stability determi- 
nants mapped by two approaches, a ribosome inter- 
ference assay and linker scanning mutagenesis, gave 
mutually confirming results that revealed the exis- 
tence of at least two closely spaced stability determi- 
nants. Their mode of action is now being studied. 
Structure and Function of the Human Growth 
Hormone Genes 
As with the globin gene system, the human growth 
hormone (hGH) gene cluster is highly regulated. 
The cluster contains the single gene hGH-N, ex- 
pressed in the pituitary, and four genes expressed in 
the placenta, hGH V, hCS-A, hCS-B, and hCS-L. Two 
of these five genes, hGH V and bCS-L, initially 
thought to be pseudogenes, have now been shown 
to be expressed placentally, based on work from this 
laboratory. By a series of receptor-binding and 
bioactivity assays, hGH V was demonstrated to 
represent the major somatogen hormone of preg- 
nancy. hCS-L was demonstrated to be expressed in 
the placenta in a number of alternatively spliced 
forms, at least one of which is likely to be expressed 
as a secreted hormone. 
Since the placenta expresses a spectrum of hGHs, 
it is possible that it also contains a corresponding set 
of hGH receptor isoforms. Consistent with this hy- 
pothesis, work from Dr. Liebhaber's laboratory has 
revealed the existence of an hGH receptor isoform 
generated by alternative splicing of exon 3 of the 
hGHR transcript. The deletion of exon 3 from the 
placentally expressed hGHR (hGHRdS) may be 
quite significant in that the deleted segment is lo- 
cated in the extracellular domain and may affect li- 
gand binding. This alternative splice is tissue spe- 
cific. Studies now in progress are aimed at defining 
whether hGHRdS demonstrates a binding prefer- 
ence for any of the hGH-related hormones and 
whether it might in this way serve as the efi'ector arm 
of a placental autocrine/paracrine loop. (This work 
was carried out in collaboration with Dr. Nancy 
Cooke, with funding from the National Institutes of 
Health.) 
RNA Structure and Function 
An underlying interest of this laboratory that over- 
laps with the studies of the globin and hGH gene 
systems is the study of RNA structure-function rela- 
tionships. A recently completed analysis of alterna- 
tive splicing of the hGH gene transcript demon- 
strated a direct controlling effect of transcript 
secondary structure. A specific stem that encom- 
passes the major splice acceptor of the hGH-N trun- 
script and its probable lariat branch point favors the 
utilization of a more distal acceptor. 
Secondary structures may also affect mRNA trans- 
lation. This has been documented by demonstrating 
that the effects of intramolecular mRNA duplexes on 
translation are position dependent. Intramolecular 
duplexes within the 5'-nontranslated region or in 
the 5' proximal coding region block assembly of the 
complete SOS ribosome, while duplexes located 
within the coding region well 3' to the AUG are effi- 
ciently melted out by the elongating ribosome. 
These data suggest that antisense strategies aimed at 
translational inhibition must be appropriately tar- 
geted. 
In addition to their effects on mRNA processing 
and function, secondary structures can result in the 
alteration of primary structure. It was demonstrated 
that intramolecular duplexes in an mRNA can target 
a cellular adenosine inosine converting activity. 
This editing activity can be targeted to a specific 
222 
