The yeast Saccharomyces cerevisiae (baker's yeast)
reproduces by budding. The shape of the new bud and its
placement on the cell surface are governed by a complex
series of interactions among intracellular signaling
proteins responsible for regulating organization of the
cytoskeleton. Many of these yeast signaling proteins have been
identified and have been found to possess higher
eukaryotic relatives that are similarly responsible for controlling
cell shape and division properties in mammalian cells.
Significantly, in a number of cases, these mammalian genes
are oncogenes and antioncogenes, implying the essential
importance of regulation of cell shape control as a decision
point in neoplastic transformation.
Our laboratory has used several approaches to explore the connections between yeast budding control and regulation of human cell growth. We expressed a library of human proteins in yeast and identified human genes capable of converting yeast budding pattern from vegetative to "pseudohyphal," a process that is sensitive to nitrogen limitation (a nutritional stressor) and involves basic changes in both cell division polarity and cell elongation control. From this screen, we isolated one gene, HEF1, that appears to be a key intermediate in mammalian growth control signaling; and a second gene, hsRPB7, that may act specifically to protect the transcription apparatus during stress. In complementary work, we used a two-hybrid approach to identify novel interacting partners of Krev-1 (an antioncogene whose yeast homolog, BUD1, controls bud placement). From this screen we obtained a novel gene, Krit1, that may itself be a cell shape regulator. Characterization of HEF1, hsRPB7 and its partner protein hsRPB4, and Krit1 is the current focus of this laboratory.
Human enhancer of filamentation 1 (HEF1) was isolated in a genetic screen as a gene that constitutively induced pseudohyphal formation in diploid yeast. The original 900 bp partial cDNA isolate was used as a probe in a combination of screening approaches, and an ~3.7kb clone, which encodes a single continuous open reading frame of 835 amino acids, was subsequently isolated. The predicted HEF1 protein contains an amino-terminal SH3 domain, an adjacent domain containing multiple SH2 binding motifs, and the carboxy terminal domain originally isolated in the pseudohyphal screen. The amino acid sequence of HEF1 is 64% similar to the recently described rat protein p130Cas (Sakai et al., EMBO J. 13: 3748, 1994), with greatest similarity in the SH3 and C-terminal domains. p130Cas is highly phosphorylated in cells transformed by the oncoprotein Crk and is an in vitro substrate for Abl and Src kinases. The abundance of SH2 consensus recognition sites on p130Cas suggests that this protein functions as a general adaptor for input of growth regulatory signals. Such adaptors facilitate the physical association of signaling molecules such as kinases with effectors and cytoskeletal components.
Southern blot analysis indicates that HEF1 is highly conserved in mammals. We isolated a partial clone of a murine HEF1-related gene (mHEF1) that is almost identical with HEF1 in the carboxy-terminal region, confirming that HEF1 is not just a human homolog of p130Cas. Northern blot analysis of human tissues indicates that HEF1 is expressed as two predominant transcripts approximately 3.8 and 5.4 kb in length. Although present in all tissues examined, these transcripts are detected at significantly higher levels in kidney, lung, and placenta, in contrast to the more uniform distribution and single transcript size reported for p130Cas.
Structural predictions for the carboxy-terminal domain of HEF1 protein suggest a continuous helical region that might mediate oligomerization with other proteins, thereby regulating its ability to modulate cellular morphology. Interaction trap/two hybrid and immunoprecipitation approaches confirmed that the carboxy terminus of HEF1 protein homodimerizes with itself and heterodimerizes with the same domain of p130Cas, suggesting an interrelated function for the two proteins. The identification of HEF1 signaling partners was further investigated using HEF1-specific antiserum to immunoprecipitate HEF1 from normal and v-Abl-transformed NIH3T3 cells. These immunoprecipitates were probed with an antibody to phosphotyrosine. A species migrating at ~130-140 kD was observed only in v-Abl-transformed fibroblasts. This species may represent HEF1 phosphorylated by v-Abl, or p130Cas phosphorylated by v-Abl and immunoprecipitated because of association with HEF1. These two possibilities are currently under investigation. As a further investigation of HEF1 signaling partners, interaction trap/two-hybrid screens were conducted using the HEF1 SH3 domain as a bait. Of the isolated clones, 15/20 consisted of a small protein of uncharacterized function containing an SH3 binding site consensus The remaining 5/20 were partial cDNAs of human focal adhesion kinase (FAK). FAK is an important component of focal adhesions, which allow cells to organize their cytoskeleton in response to cell-substratum contact. The study of these and other potential HEF1 associated signaling partners remains a major future goal.
Our data support the interpretation that the HEF1 carboxy terminus region (and the corresponding region of p130Cas) acts as an effector domain to enhance actin cytoskeletal polarization, while the amino terminal SH3 domains anchor FAK to the complex. In this model, the extensive SH2-binding domain would negatively regulate the function of the HEF1 carboxy terminus when unphosphorylated, while phosphorylation on tyrosine as a result of growth signals from phosphotyrosine kinases such as Src and Abl would induce the HEF1 carboxy terminus to stimulate cell polarization. This process may be important in the acquisition of invasive characteristics by transformed malignant cells.
Using a screen for human cDNAs that enhance the formation of pseudohyphae in S.cerevisiae, we obtained a cDNA encoding hsRPB7, a novel human homolog of the seventh largest subunit of S.cerevisiae RNA polymerase II (pol II). In yeast, the RPB7 protein forms a stable complex with another pol II subunit, RPB4. This complex has been shown to play a key role in the process of stress response and increases the viability of yeast cells in stationary phase (e.g., Choder and Young, Mol. Cell. Biol. 13: 6984, 1993). The fact that hsRPB7 is able to substitute successfully for some key functions of RPB7 in yeast (Khazak et al., Mol.Biol.Cell 6: 7599, 1995) suggests that it may work similarly in protecting RNA pol II during stress conditions in human cells. If so, this might imply the existence of a human homolog of RPB4. Using a HeLa cDNA library and a hsRPB7 gene as a probe, we performed a screen for "hsRPB4" using the interaction trap approach. From this screen we obtained multiple isolates of two cDNAs, one of which appears to be a novel human homolog of the yeast RPB4 gene.
The hsRPB4 cDNA clone isolated in the two-hybrid screen contained 1855 bp and contained an open reading frame (ORF) of 381 bp (127 aa). The putative protein encoded by the ORF had strong homology with the C-terminal region of yeast RPB4 (40% identity) but did not contain an initiating ATG codon. Combining different approaches, we extended the cDNA to 1920 bp; the extended clone contains a candidate start methionine. The cDNA clone of hsRPB4 was used to probe a multiple tissue Northern blot. This analysis revealed two cross-hybridizing mRNA species of ~2 kb and 5.5 kb. The extended hsRPB4 cDNA may encode the smaller mRNA. We are using several approaches to determine if hsRPB4 sequences related to the N-terminal region of RPB4 protein exist; and if so, to clone them. Although these two RNA species are present in all tissues, they are most abundant in skeletal muscle and heart in a pattern similar to that of the hsRPB7 transcript. The hsRPB4 cDNA has been mapped to chromosome 2q21. Using a two-hybrid based approach, we determined that activation domain-tagged hsRPB4 interacts strongly and specifically with hsRPB7 but not with RPB7 or a series of negative controls. We also made LexA fusions to hsRPB4 and to the yeast RPB4 protein and determined that both proteins are strong activators of transcription in yeast.
The ultimate goal of this project is to establish whether these two subunits protect pol II during stress conditions in humans as RPB4/RPB7 do in yeast. To this end, we made epitope-tagged versions of both hsRPB4 and hsRPB7 in mammalian vectors. We are in the process of generating antibodies to an hsRPB4-derived peptide and hsRPB4-GST fusion protein. Future studies will address the role of hsRPB4 and hsRPB7 in mammalian cells undergoing various types of stresses, including elevated temperature, nutritional starvation, and acute or chronic exposure to toxic agents.
The human gene Krev-1/rap1a codes for a small GTP-binding protein related to the Ras oncogene. Because overexpression of Krev-1 can partially revert Ras transformed cells and because Krev-1 can bind and potentially sequester the Ras GTPase-activating protein (GAP) in vitro, Krev-1 was initially thought to work as an antioncogene. More recently, work by some groups has suggested that Krev-1 may also function as an oncogene in some cellular contexts. Significantly, when TSC2 was identified as the genetic lesion in tuberous sclerosis type 2, it was also shown to function as a GAP protein for Krev-1. Thus, mutation of a gene whose normal function is to regulate the catalytic activity of Krev-1 promotes abnormal unregulated cell growth, supporting the idea that Krev-1 may also act oncogenically in humans and that its function is likely to be complex. The yeast homolog of Krev-1, RSR1 or BUD1, possesses highly conserved sequence and partially conserved function with its human counterpart. Work by a number of groups suggested that BUD1 plays a key regulatory role both in GTPase cascades controlling cell polarity and morphology and in regulation of MAPK pathways required for mating and response to environmental signaling. The function of this yeast protein may provide a key paradigm for its mammalian homolog, Krev-1. We thus initiated a search for proteins that interacted with Krev-1 as an approach to elucidating both the basic regulation of cell growth controls and to understanding the means by which Krev-1 can apparently function both as an oncogene and an antioncogene.
Using a LexA-fused Krev-1 protein in a two-hybrid screen, we isolated two groups of proteins that specifically interacted with Krev-1 but not with 24 other LexA-fused proteins. One class corresponded to the oncogene Raf, which was already identified as a Krev-1 partner protein. The second class was represented by 7 independent partial clones of a novel gene, which we have designated Krit1 (for Krev-1 Interaction Trap 1). We currently possess a 2.0 kb Krit1 clone encoding a 529 amino acid open reading frame, which probably corresponds to most of the complete protein sequence of this gene. Probe of a multi-tissue Northern blot with Krit1-complementary sequences revealed a complex expression pattern with one strongly hybridizing transcript unique to brain, and a group of moderately hybridizing transcripts predominant in heart and muscle. The amino-terminal region of Krit1 encodes 4 ankyrin repeat (protein-interaction mediating) motifs. The carboxy-terminal region of Krit1, containing the Krev-1 interacting domain, possesses striking similarity to a family of cytoskeletal/membrane junction proteins comprising moesin, radixin, ezrin and merlin; merlin is also known as the antioncogene neurofibromatosis 2, or NF2. Investigators in the laboratory of Dr. Joseph Testa at FCCC used fluorescent in situ hybridization to map the Krit1 gene to chromosome 7q22, a common site of breaks and deletions in human leukemias and other cancers, and a region hypothesized to contain an antioncogene. In collaboration with several laboratories, we are investigating whether Krit1 maps within small defined mutations in the 7q22 region. To facilitate this effort, as well as to understand Krit1 structure better, we recently cloned a genomic Krit1 clone that spans the complete coding sequence of the gene. Finally, we continue to build reagents that will enable us to ask key questions about Krit1 function. At present, we believe Krit1 is a good candidate to regulate Krev-1 function directly and may itself be antioncogenic and function in tumors involving NF1 and Ras mutations.
PUBLICATIONSESTOJAK, J., BRENT, R., GOLEMIS, E.A. Correlation of two-hybrid affinity data with in vitro measurements. Mol. Cell. Biol. 15: 5820-5829, 1995.
GOLEMIS, E.A., KHAZAK, V. Alternative yeast two-hybrid systems: The interaction trap and interaction mating. In Methods in Molecular Biology, Chapter 60, edited by R. Tuan. Humana Press, Clifton (in press).
GOLEMIS, E.A., GYURIS, J., BRENT, R. Interaction trap/two-hybrid systems to identify interacting proteins. Unit 20.1.1-20.1.28. In Current Protocols in Molecular Biology, edited by F.M. Ausubel, et al. John Wiley & Sons, New York, 1996.
KHAZAK, V., SADHALE, P., WOYCHIK, N.A., BRENT, R., GOLEMIS, E.A. Human RNA polymerase II subunit hsRPB7 functions in yeast and influences yeast cell morphology and stress survival. Mol. Biol. Cell. 6: 759-775, 1995.
LAW, S.F., ESTOJAK, J., WANG, B., MYSLIWIEC, T., KRUH, G.D., GOLEMIS, E.A. Human Enhancer of Filamentation 1 (HEF1), a novel p130Cas-like docking protein, associates with FAK and induces pseudohyphal growth in yeast. Mol. Cell. Biol. (in press).
WANG, B., MYSLIWIEC, T., KRAINC, D., JENSEN, R.A., SONODA, G., TESTA, J.R., GOLEMIS, E.A., KRUH, G.D. Identification of ArgBP1, an Arg protein tyrosine kinase binding protein that is the human homologue of a CNS-specific Xenopus gene. Oncogene (in press).
Paper in press at time of previous report:
DATTA, K., FRANKE, T.F., CHAN, T.O., MAKRIS, A., YANG, S.-I., KAPLAN, D., MORRISON, D.K., GOLEMIS, E.A., TSICHLIS, P.N. AH/PH domain mediated dimerization of Akt and its potential role in Akt regulation. Mol. Cell. Biol. 15: 2304-2310, 1995.
KLICKSTEIN, L.B., NEVE, R.L., GOLEMIS, E.A., GYURIS, J. Conversion of mRNA into double-stranded cDNA. Unit 5.5.1-5.5.13. In Current Protocols in Molecular Biology, edited by F.M. Ausubel, et al. John Wiley & Sons, New York, 1995.
* Fox Chase researcher

Seated left to right: Susan Law and Erica Golemis
Standing left to right: Ilya Serebriiski, Ying-Tong Wang,
and Vladimir Khazak