A COMPLEX FAMILY OF IMMUNOGLOBULIN HEAVY-CHAIN ENHANCER-BINDING PROTEINS 
Thomas R. Kadesch, Ph.D., Assistant Investigator 
During the past year, Dr. Kadesch 's laboratory 
has used two genes to explore the nature of tissue- 
specific gene expression. The first gene, which com- 
mands the primary focus of the laboratory and is 
discussed in this report, is the immunoglobulin 
heavy-chain gene. It has been known for some time 
that B cell-specific expression of all immunoglobu- 
lin genes is controlled at the level of transcription. 
For the heavy-chain gene, much of this transcrip- 
tional control (activation) is thought to be due to a 
B cell-specific enhancer located in the intron be- 
tween the J and C segments of the gene. The labo- 
ratory is attempting to identify and characterize 
the proteins that bind to the enhancer and medi- 
ate its activity. The second gene being studied is 
the human liver/bone/kidney alkaline phosphatase 
(LBK AP) gene. The LBKAP gene is also expressed 
at different levels in different cell types. It is ex- 
pressed at very high levels in osteoblasts and at 
moderate levels in most other cell types. Hence the 
gene has two modes of expression: one (high) that 
reflects a specialized cell function (bone mineraliza- 
tion) and another (low) that possibly reflects a 
"housekeeping" function for the gene. The basic 
parameters that underlie this differential expres- 
sion are being defined. 
The efforts of many groups have led to the identi- 
fication of several major protein-binding sites 
within the IgH enhancer. Most of these sites are im- 
portant for enhancer activity. Hence the proteins 
that bind to them must be responsible for mediat- 
ing enhancer function. Although this conclusion 
may be straightforward, the implications for the 
enhancer's cell-type specificity are not. Only one of 
the enhancer-binding proteins (Oct-2) is clearly re- 
stricted to B cells; the others appear to be present 
in a wide variety of cell types. Furthermore, the site 
that binds Oct-2 plays only a minor role in overall 
enhancer activity, and its absence does not affect 
cell- type-specific expression. The question remains: 
How is B cell-specific enhancer activity mediated 
by ubiquitous DNA-binding proteins? A recent se- 
ries of experiments carried out in Dr. Kadesch's lab- 
oratory have begun to shed light on this issue. One 
approach has been to define systematically the pro- 
tein-binding sites with regard to their influence on 
B cell-specific expression. Another has involved the 
direct cloning of cDNAs that encode enhancer-bind- 
ing proteins. 
Several years ago experiments with deletion mu- 
tants of the enhancer hinted that negative regu- 
lation may play at least a minor role in dictating 
B cell-specific expression. This has now been con- 
firmed. Moreover, two protein-binding sites have 
been identified that recapitulate the initial results 
dramatically. One site, |jlE3, binds a ubiquitous 
DNA-binding protein. An oligonucleotide carrying 
only this site is capable of activating transcription 
from a linked promoter in both B cells and non-B 
cells. The other site, |jlE5, has no activity on its own. 
However, when the |jlE5 site is linked to the [xE5 
site (as it is normally found in the enhancer), activ- 
ity is increased in B cells and totally abolished in 
non-B cells. Mutations introduced into the |jlE5 site 
restore expression in non-B cells. Hence the |xE5 
site stimulates activity of the |jlE3 site in B cells and 
inhibits its activity in non-B cells. 
A cDNA clone, designated X-3, that encodes a 
|xE3-binding protein has been isolated from a 
B cell-derived cDNA library. The encoded protein 
binds to the |jlE3 site in a way that is indistinguish- 
able from the binding activity seen with crude nu- 
clear extracts. Furthermore, when this cDNA is 
overexpressed in mammalian cells, it specifically 
stimulates transcription of a reporter gene linked to 
a |jLE3-binding site. This confirms that the X-3 cDNA 
encodes a |jlE3 transcription factor (TFE3). The pre- 
dicted amino acid sequence of TFE3 identifies two 
regions that may specify protein oligomerization 
motifs. One is homologous to a region of c-myc and 
defines a putative helix-turn-helix motif. The other 
is a putative leucine zipper that lies adjacent to the 
helix-turn-helix region. Hence one question is 
whether TFE3 is able to form functional hetero- 
dimers with other related proteins. Heterodimer 
formation would not be required for activity of the 
TFE3 protein, however, because the transcription 
activation domain of the protein has been mapped 
to a region distinct from the presumed protein di- 
merization motif Perhaps heterodimer formation is 
required for inhibiting the activity of TFE3. 
Clones corresponding to |jLE5-binding proteins 
have also been isolated. The multiple cDNAs iso- 
lated thus far are encoded by at least three, and 
possibly four, distinct genes. Two of these have 
been isolated from a B cell-derived library and one 
(or two) from a non-B cell (HeLa) library. The two 
isolated from the B cell-derived library have been 
characterized in detail. Both are related at the level 
of amino acid sequence, suggesting that they are 
Continued 
225 
