these elements bind to a factor present in fibro- 
blasts as well as HeLa cell nuclear extracts and that 
this binding is competed for by the/o5 SRE. Thus it 
is likely that the serum response factor (SRF) that 
binds to the/o5 SRE mediates both serum and TPA 
inducibility of Egr-1. Evidence has emerged "from 
studies using a cDNA clone for SRF that show that 
the in t'/Yro-generated SRF protein binds to the 
Egr-1 upstream CArG boxes. However, this analysis 
does not exclude the possibility that other SRF-like 
proteins exist. Experiments intended to explore 
this hypothesis are in progress. 
With the Egr-1 5' region deletion mutants, Dr. 
Sukhatme's laboratory is also defining how Egr-1 
induction is mediated in the many non-growth- 
related situations mentioned above. Attempts to de- 
fine the mechanisms by which Egr-1 transcription is 
negatively regulated are also under way. 
C. Properties of the Egr-1 protein. During the last 
year, antisera against an Egr-l-specific peptide and 
a fusion protein have been generated. Both anti- 
sera give intense nuclear staining of serum-stimu- 
lated fibroblasts. Peak signal intensity is noted be- 
tween 2 and 3 hours, but the signal persists for at 
least 5 hours. Western analysis and immunopre- 
cipitation show an 80 kDa band in the serum-in- 
duced fibroblasts, a size in agreement with that 
seen for the in fzYro-translated product. The Egr-1 
protein is also rapidly phosphorylated, suggesting 
that this mechanism may be an additional means of 
modulating its activity. 
D. Developmental expression of Egr-1 shows co- 
regulation with c-fos. [These studies were con- 
ducted in collaboration with A. McMahon (Roche 
Institute of Molecular Biology).] Egr-1 expression is 
easily seen in total embryonic RNA by day 11. In 
situ hybridization shows striking localization to the 
perichondral regions of long bones and digits, to 
mesenchymal odontoblasts, and to roots of hair 
whiskers, and intense but patchy signal in various 
muscle groups. Several inferences may be derived 
from these developmental expression patterns: 1) 
The bone and tooth data are similar to those ob- 
tained for c-fos and point to the rather special rela- 
tionship between these two genes. 2) Possible tar- 
gets for Egr-1 action might include collagen type I, 
since a major function of odontoblasts and osteo- 
blasts is to secrete collagen. 3) Nerve growth factor 
(NGF) receptors have recently been localized to 
several nonneuronal tissues, including the dental 
ridge, muscle, and limb buds. Thus Egr-1 mRNA in- 
duction may be mediated by NGF in these tissues. 
4) Osteoblasts or cell lines derived from them 
might be a particularly interesting system in which 
to study Egr-1 (and c-fos) expression with an eye to 
asking what keeps Egr-1 expression at constitutively 
high levels in these cells. Findings in these cells 
could be contrasted to those derived from serum- 
stimulated fibroblasts, where expression of these 
genes is transient. 
E. EGRl expression in human tumors. To analyze 
whether EGRl is deregulated in abnormal cell 
growth. Northern blots of human tumors were pro- 
vided to us by Dr. D. Slamon (University of Califor- 
nia at Los Angeles). In over 50% of the cases, EGRl 
expression is reduced 3- to 10-fold in tumor tissue 
versus surrounding normal tissue from the resected 
human specimens. The tumors were derived from 
lung, colon, bladder, liver, breast, bone, and mus- 
cle. In the near future. Dr. Sukhatme and his col- 
leagues will look at whether these data can be sub- 
stantiated at the protein level. What might 
decreased tumor expression imply? One possibility 
is that EGRl functions as part of a negative growth 
regulatory pathway in a manner similar to the reti- 
noblastoma gene product. Earlier data from Dr. 
Sukhatme's laboratory on the 5q" deletions in ther- 
apy-related acute nonlymphocytic leukemia (ANLL) 
that map to the same regions as the chromosomal 
localization of EGRl would tend to support this 
view. Clearly, attempts at overexpressing and/or m- 
hibixXn^EGRl expression in both normal and trans- 
formed cells will be needed to answer these ques- 
tions. 
II. Genes Related to Egr-1. 
During the past year. Dr. Sukhatme's laboratory 
has used the finger domain probe of Egr-1 to re- 
screen the original serum plus cycloheximide fibro- 
blast cDNA library Three novel genes (Egr-2, -3, 
and -4) have been partially characterized. All share 
nearly identical DNA-binding domains but differ 
elsewhere in their sequence. Egr-2 [also known as 
Krox20; Drs. R. Bravo and R Charnay (European 
Molecular Biology Laboratory, Heidelberg)] and 
Egr-3 are both novel immediate-early response 
genes. Their transcripts [~3.2 kb (Egr-2) and ~5 kb 
(Egr-3)] show kinetics similar to those of Egr-1. The 
nonfinger regions of Egr-2 and Egr-3 appear to dif- 
fer substantially from those of Egr-1. Studies along 
the lines described above for Egr-1 are under way 
for these two genes as well. 
Continued 
310 
