In the Interaction Trap/two hybrid system, protein-protein interaction is
assayed using a genetic strategy in which a protein (P1), or the "bait," is
expressed as a fusion protein with a DNA-binding domain (LexA); and a second
protein (P2) is expressed as a fusion protein with a transcriptional activation
domain. Coexpression of the two protein fusions in yeast in which the binding
site for the DNA binding domain is located upstream of a reporter gene will
result in transcriptional activation of the reporter gene by the P2-fused
activation domain if P1 and P2 are able to associate. This two-hybrid approach
has gained considerable popularity because it provides a rapid and convenient
assay for a large number of interactions without the need to purify proteins.
In addition, we have shown that by substituting a cDNA library for a defined
P2, the two-hybrid approach can be readily converted into a screening system
for novel interacting proteins. We are currently conducting experiments to
characterize and improve the basic functioning of the Interaction Trap. As one
approach, we are attempting to introduce an additional "decoy" reporter system
to eliminate the tedious tasks of isolating and eliminating false positive cDNA
s obtained in library screens. In a continuing collaboration with Roger Brent,
we also have performed a series of calibration experiments to determine how
the affinity of two proteins as measured by two hybrid system correlates with
in vitro biochemical measures of affinity.
Diploid S. cerevisiae undergo a conversion to pseudohyphal
(filamentous) budding when grown on severely nitrogen-limited media. In
this conversion, the budding pattern changes from bipolar to unipolar with
a concomitant alteration in cell morphology from rounded to extremely
elongated. Colonies of pseudo-hyphal cells have a stellate appearance that is
easily distinguishable from vegetatively growing cells. The yeast budding
pathway contains functional homologs of a number of human oncogenes and
antioncogenes, suggesting that human genes that can specifically affect
budding control might function as regulators of cell transformation and
metastasis. In addition, because pseudohyphal conversion is induced
by nutritional stress, human genes that enhance stress response might also
enhance formation of pseudohyphae. We transformed yeast with a human cDNA
expression library, then screened the transformed yeast and isolated several
human genes that enhanced pseudohyphal formation.
One human gene identified in the pseudohyphal enhancement screen possesses
significant sequence homology (21% to 53% identity) to a family of RNA
polymerase subunits of which the most well characterized member is the yeast
RPB7 subunit (McKune et al., Yeast 9: 295, 1993), leading us to
designate it hsRPB7. Deletion of RPB7 in yeast is normally lethal. HsRPB7 can
functionally replace its yeast counterpart and allow growth at moderate
temperatures; we have used antibody directed against the largest subunit of
pol II in co-immunoprecipitation experiments to demonstrate that hsRPB7 is
capable of assembling into complete RNA pol II. HsRPB7-containing yeast are
strikingly deficient for survival relative to wild-type yeast during culture
at high and low temperatures, after heat shock, or when maintained at
stationary phase. We postulated that this reduced ability to resist stress
might be due to reduced ability of hsRPB7 to associate with another RNA pol II
subunit, RPB4. Work from other laboratories had previously shown that an
RPB4/RPB7 subcomplex was specifically required for effective stress response
(Choder and Young, Mol. Cell. Biol. 13: 6984, 1993). Using the
Interaction Trap, we have determined that hsRPB7 is able to interact with RPB4
in vivo, but does so with a lower affinity than the yeast RPB4-RPB7
subunit pair. This property provides a possible mechanism for the defective
stress response of hsRPB7-dependent yeast.
The above data and other experiments indicated that the RPB7 subunit in
yeast plays a role in both stress response and cell morphology control, and
that the human gene preserves at least some of these functions. This
observation suggests a parallel mechanism to some prokaryotic systems in
which specific sigma factors associate with RNA polymerase during stress
response and function, in part, by reprogramming cell shape to one better
suited for survival (reviewed in Hengge-Aronis, Cell 72: 165, 1993).
Our main interest is to determine whether hsRPB7 similarly functions to control
stress response and morphology in human cells. We have begun a series of
experiments to address these issues.
To date, we have found that the RNA encoding hsRPB7 is expressed at
dramatically different levels (80-fold variation) in different tissues;
the highest levels are in heart, muscle, kidney, and liver. We developed an
antibody to hsRPB7 and used it to assay for protein expression in mammalian
cells. In a Western blot of kidney cell proteins, the antibody detects one
band of ~18 kD, which corresponds to the size of hsRPB7 predicted from its
176 amino acid coding sequence, and a second band of ~24 kD, which may
correspond to a phosphorylated or otherwise modified form of the protein.
Using hsRPB7 in Interaction Trap library screening, we obtained multiple
isolates of two cDNAs encoding proteins that interacted strongly and
specifically with LexA-hsRPB7. The first of these is a human homolog of the
RPB4 subunit, which we have designated hsRPB4. Preliminary Northern analysis
indicates that expression of this gene is also tissue specific. The highest
levels of RNA were detected in the heart and muscle tissue, which parallels
the expression pattern of hsRPB7. The second cDNA encodes a novel member of
a family that includes the Arabidopsis (plant) EMB30 gene. The EMB30
gene was isolated because plants that have this gene deleted cannot form roots
as a result of deficiencies in cell division and cell expansion (Shevell et
al., Cell 77: 1051, 1994)morphological phenotypes
reminiscent of those involved in pseudohyphal transition. Over the next year,
our goal is to explore in mammalian cells the increasingly intriguing
relationship between hsRPB7 expression and modification, and the transcription
of genes required for stress response and cell growth control.
HEF1 (Human Enhancer of Filamentous growth 1) is the sole cDNA isolated
from the screen that induces constitutive pseudohyphal growth under all
nutrient conditions. The original HEF1 clone is 900 bp long and contains an
open reading frame (ORF) that encodes a putative protein of 186 amino acids.
We have recently isolated a murine genomic clone for HEF1 (MEF1). Analysis of
MEF1 sequence indicates it is likely that the HEF1 clone represents one exon
of a longer gene. The MEF1 and HEF1 amino acid sequences are 96 percent
identical. The high degree of conservation suggests that these sequences
encode an important function. Using a Northern blot of RNAs from several
tissues, a HEF1-complementary probe identified a number of cross-hybridizing
bands, including an ~3.4 kb and a 5 kb RNA species that were present in all
tissues, but most abundant in kidney, lung, and placenta; and two other weakly
cross-reacting species, one of ~8 kb that was detected only in kidney and
lung, and one of ~1.2 kb that was detected only in liver.
Using a HEF1-specific probe, we have isolated a 3.7 kb cDNA sequence with
an ORF coding for ~830 amino acids that contain an SH3 domain and multiple
SH2-binding motifs. We found very recently that HEF1 is closely related to the
gene for p130Cas, which was cloned based on its association with the oncogene
v-crk (Sakai et al, EMBO J. 13: 3748, 1994). P130Cas appears to associate with and act as a substrate
for multiple tyrosine-kinase containing oncogenes. The current model for
the function of this protein suggests that it may act as a central assembly
point for signals relevant to growth control and carcinogenesis. Preliminary
immunohistochemical experiments with an antibody generated using a HEF1-
specific peptide suggest that HEF1 associates with the actin cytoskeleton.
Cumulatively, these data suggest that HEF1 may directly modulate cell growth
function in response to oncogenic signals. The short ORF originally isolated
in the pseudohyphal screen may correspond to an unregulated effector domain.
Work on this gene remains a high priority.
DATTA, K., T.F. FRANKE, T.O. CHAN, A. MAKRIS, S.-L. YANG, D. KAPLAN,
D.K. MORRISON, E.A. GOLEMIS, and P.N TSICHLIS. AH/PH domain mediated
dimerization of Akt and its potential role in Akt regulation. Mol. Cell.
Biol. (in press).
GOLEMIS, E., J. GYURIS, and R. BRENT. Interaction
trap/Two-hybrid system to identify interacting proteins. Unit
13.14.113.14.17. In Current Protocols in Molecular Biology,
edited by F.M. Ausubel, et al. John Wiley & Sons, NY, 1994.
KLICKSTEIN, L.B., R.L. NEVE, E. GOLEMIS, and J. GYURIS. Conversion of mRNA
into double-stranded cDNA. Unit 5.5.15.5.13. In Current
Protocols in Molecular Biology, edited by F.M. Ausubel, et al. John Wiley
& Sons, NY, 1995.
SATO, T., M. HANADA, S. BODRUG, S. IRIE, N. IWAMA, L, BOISE, C. THOMPSON,
E. GOLEMIS, L. FONG, H.-G. WANG, and J.C REED. Interactions among members of
the BCL-2 protein family analyzed with a yeast two-hybrid system.
Proc. Natl. Acad. Sci. USA 91: 92389242, 1994.
HUMAN ENHANCERS OF FILAMENTOUS GROWTH IN YEAST
ERICA GOLEMIS, Ph.D., Associate Member
VLADIMIR KHAZAK, Ph.D., Postdoctoral Associate (from April 1994)
SUSAN LAW, Ph.D., Postdoctoral Associate (from July 1994)
JOANNE ESTOJAK, B.A., Scientific Technician
YING-TONG WANG, Technical Aide (from April 1994)
AILEEN BILYAK, Student Assistant, Haverford College, Haverford, PA
Over the last forty years, genetic and
biochemical studies of the simple eukaryote Saccharomyces cerevisiae
have established paradigms for cell growth control, gene expression, and
metabolism. These paradigms were found to apply to more complicated eukaryotes
up through and including humans, and have been used to identify and
characterize key regulators of cell division control, greatly advancing our
general understanding of the process of cancer. We have postulated that if
parts of a genetic pathway are conserved between humans and yeast, it might be
possible to identify novel important human genes by virtue of their ability to
alter the normal functioning of the pathway in yeast. The primary focus of my
laboratory is to characterize a number of human genes that, based on their
ability to enhance formation of pseudohyphae in yeast, are potential regulators
of cell growth control during carcinogenesis or during stress response. A
second focus in my laboratory is the continuation of work on a protein-protein
interaction system, called the Interaction Trap, to improve the utility of
this yeast-based technology for identification and characterization of novel
proteins.
THE INTERACTION TRAP. GOLEMIS, KHAZAK, ESTOJAK
RNA POLYMERASE SUBUNITS THAT MAY REGULATE CELL MORPHOLOGY AND STRESS
RESPONSE. KHAZAK, ESTOJAK, in collaboration with SADHALE, WOYCHIK
HEF1. LAW, ESTOJAK, BILYAK, WANG
PUBLICATIONS