1997 NSF Participants-Department of Biological Sciences - Carnegie Mellon University

1997 National Science Foundation (NSF)-supported Participants

Eric Balicky, Allegheny College, Meadville, PA
(Advisor: Dr. Jonathan Minden)

RNA DIGE: A Novel Approach to the Differential Analysis of mRNA Expression
The use of difference gel electrophoresis (DIGE) to detect differences in the expression of biological macromolecules from any two specimens is a technique which is actively being explored in Dr. Minden's lab. The purpose of my project has been to develop a novel fluorescence-based gel technique to identify differences in mRNA expression from two differing cell or tissue types. To perform this technique, mRNA isolated from each sample is used as a template in a reverse transriptase reaction. Each reaction is primed with the use of an oligonucleotide primer that is labeled with one of two fluorescent dyes. The fluoresecently labeled products of the reverse transcription reaction are then utilized in a second reaction in which cDNA is synthesized. Differences in mRNA expression may then be analyzed by combining the two differentially labeled cDNA populations, one from each mRNA sample, and separating the products by gel electrophoresis. cDNA fragments of interest may then be isolated and sequenced. The focus of my research to date has been on the successful labeling and purification of fluorescently-tagged oligos and their use in the creation of a series of control cDNA libraries.


William Dirks, Western Washington Univ., Bellingham, WA
(Advisor: Dr. Robert F. Murphy)

A Wavelet Approach to Feature Extraction in Fluorescently Labeled Cells
An automated system able to classify cellular images based on protein localization patterns obtained via fluorescence microscopy could have important implications for biological research. Examples include the creation of a systematic protein localization database which could be used to study location/function relationships, and the ability to quantify affects of agents causing protein mislocalization. To develop such a system a mechanism to extract features which could group and differentiate images based on protein localization patterns is needed. Wavelet-based feature extraction was studied for this purpose. A light microscope was used to capture images of protein patterns in immunofluorescently labeled cells. These images were thresholded, normalized for intensity, rotation, and translation, and transformed using the Daubechies wavelet. A neural network classified images based on the calculated wavelet features with an 80% success rate. The results indicate that a small number of numerical features derived from a wavelet transform contain sufficient information to allow computer recognition of patterns typical of major cellular organelles.


Kate A. Dreher, Williams College, Williamstown, MA
Advisor: Dr. Elizabeth W. Jones)

Investigation of the Role of PEP3 in Vacuolar Biogenesis and Protein Targeting in Saccharomyces cerevisiae
The PEP3 gene in Saccharomyces cerevisiae encodes a 918 amino acid protein that is peripherally associated with the vacuolar membrane. pep3 mutants exhibit a vestigial vacuole morphology and missort inactive precursor forms of the soluble vacuolar hydrolases protease A, protease B, and carboxypeptidase Y (CpY). Many other proteins involved in the vacuolar branch of the yeast secretory pathway have been well characterized through extensive experimentation. Therefore, in order to clarify the role played by Pep3p in vacuolar biogenesis and protein targeting, a screen for high copy suppressors of the pep3ts phenotype was initiated in order to identify other proteins that interact with Pep3p. Eleven pep3ts alleles had previously been generated through in vitro PCR mutagenesis. Although these alleles resulted in wild type levels of CpY activity and near wild type growth on plates containing 5mM ZnCl2 or 300mM SrCl2 at the permissive temperature of 23C, these alleles caused deficiencies in CpY activity and/or growth sensitivity on plates containing 5mM ZnCl2 or 300mM SrCl2 at the restrictive temperature of 37C. Eight alleles that still caused the pep3ts phenotype following shuttling were prepared for integration into the yeast strain BJ8918 (PEP3 ). Two of these pep3ts alleles have been successfully integrated into the genome using the URA3 pop-in/pop-out method of integration and the integrants have acquired the temperature sensitive phenotypes. The six remaining alleles have been integrated into the genome and pop-out candidates are currently being screened for pep3ts phenotypes. A high copy yeast genomic DNA library has also been prepared for transformation into the pep3ts mutant strains. Future investigators will be able to transform the DNA library into the two pep3ts strains of yeast in order to perform a high copy suppression screen to identify proteins that interact with Pep3p.


Esteban Fernandez, Harvey Mudd College, Claremont, CA
(Advisor: Dr. Susan Henry)

Two-hybrid system confirms interaction between Ino2p and Ino4p in Saccharomyces cerevisiae
Inositol is a crucial precursor and signal molecule in phospholipid synthesis in the yeast Saccharomyces cerevisiae. Inositol can be produced through the conversion of glucose-6-phosphate into inositol-1-phosphate, which is subsequently dephosphorylated. The first step takes place through the action of the product of the INO1 gene, inositol 1-phosphate synthase. The regulation of INO1 is mediated by the products of the INO2, INO4, and OPI1 genes (Ino2p, Ino4p, and Opi1p). INO2 and INO4 code for basic helix-loop-helix proteins that interact to form a heterodimer which activates transcription of INO1. It has been proposed that Opi1p negatively regulates transcription of INO1 by binding to Ino4p and preventing the dimer from forming. Previous in vitro assays have shown that Ino2p and Ino4p interact and, together, can bind DNA. Opi1p and Ino4p may also interact in vitro but cannot bind DNA. This data, collected using in vitro translated Ino2p and Ino4p, supports the INO1 regulatory mechanism involving Ino2p and Ino4p interaction. A two-hybrid system using lacZ as the reporter gene was employed to study the interactions between Ino2p, Ino4p, and Opi1p in vivo. Qualitative and quantitative beta-galactosidase assays show that Ino2p and Ino4p do interact in vivo and that Ino2p can activate transcription. Further experiments using the two-hybrid system are now underway to study the interactions of Opi1p with Ino2p and Ino4p.


Lisa Glass, Trinity University, San Antonio, TX,
(Advisor: Dr. William E. Brown)

Biochemical Analysis of Lavage Fluid from Mice Exposed to 3-Methylindole
The overall aim of the project is to develop a less invasive method of assessing clinical damage than performing histology on the affected tissues. The compound 3-methylindole (3-MI), whose pathology in the respiratory system is well known, was selected for use. Several mice were injected with 3-MI interperitoneally (250 mg/kg and 300 mg/kg doses), and controls were given injections of plain corn oil, the solvent in which the 3-MI was dissolved. MRI was used to examine the nasal passages of the mice, and several biochemical assays were done to correlate and confirm the pathology that could be seen in the images. Biochemical analysis of bronchoalveolar lavage fluid (BAL) and nasal lavage fluid (NAL) were done at three time points after the injections (1 day, 3 days, 1 week). Control mice were used during each imaging procedure and in every biochemical assay. A Micro BCA protein assay was done to determine total protein content of the lavage samples, an ELISA assay was performed to determine the albumin content of the lavage fluids, and a lactate dehydrogenase (LDH) assay was performed to determine the activity of this enzyme in the samples. In the lung lavage samples at both concentrations of 3-MI, the level of LDH activity peaks at the one day time point and slowly decreases while the protein content peaks at 3 days and continues to remain high. This indicates that there is cell loss and damage occurring before the formation of lesions that allow circulatory proteins to cross the epithelial layer into the respiratory system. The nasal lavage samples at both concentrations of 3-MI fail to exhibit a detectable level of LDH activity. Howeverm the protein levels peak at one day and continue to remain high. The level of total protein is greater in the nasal lavages as compared to the lung lavages, probably due to the large amounts of mucosal secretions in the nasal passages that are seen in the MRI. In general, the biochemical data correlated well with the observations from the magnetic resonance images obtained of the mouse nasal passages.


Jane P. Kenney, Alma College, Alma, MI
(Advisor: Dr. Javier Lopez)

su(s) and Rbp-1: Two Potential Regulators of Ultrabithorax Alternative Splicing
Transcripts from the homeotic gene Ultrabithorax (Ubx) of Drosophila melanogaster are alternatively spliced into 6 different mRNA isoforms, each with a subtly different function in development. The isoforms are highly regulated for temporal and spatial specificity. This lab identified possible alternative splicing regulators by screening for chromosomal deficiencies that modified Ubx phenotypes. I studied two different regions identified in this screen. The first region contains suppressor of sable (su(s) ), which is known to suppress mutations caused by the transposable element 412 but has an unclear role in RNA metabolism. I crossed different alleles of su(s) to two alleles of Ubx, to test whether su(s) modified Ubx function and alternative splicing. Every mutant allele of su(s) tested enhanced the phenotype of Ubx9.22 (a deletion allele); most also enhanced the phenotype of Ubx195 (a stop codon in an alternative exon), but the weaker alleles suppressed this phenotype while enhancing that of Ubx9.22. The data, including preliminary RT-PCR analyses, suggest that su(s) is required for inclusion of the alternative exons in Ubx mRNAs. The second region I studied contains Rbp-1, which encodes a SR protein that is the homolog of mammalian SRp20. SR proteins facilitate the formation of splicing complexes and may regulate alternative splicing. There are no known point mutations in Rbp-1, so I began a local transposon hop of a nearby P-element into the Rbp-1 region using a screen designed to test genetically for Rbp-1 mutants. Thus far, I have identified 6 possible Rbp-1 mutants.


Adrienne McFadden, Univ. of Maryland, Baltimore County
(Advisor: Dr. John Pollock)

Three Dimensional Time Lapse Study of the Developing Fly Eye, with Pulse Labeling of the Cell Cycle
Eye development in Drosophilia has been extensively studied in the past two decades, but the precision and depth to which it has been studied has been limited to the "cut and paste" techniques that use fixed tissue. In order to acquire a more precise, detailed "story" behind the fly's eye development, we used live tissue kept vital in an organ culture chamber. Pilot experiments were carried out in the chamber to show that the tissue could be maintained for at least eighteen hours. Then, the experiments were repeated with the chamber on the Multimode microscope, where three dimensional time lapse movies were captured. These movies allowed us to visualize the typical patterns of movement and growth throughout an eighteen hour time period in the eye's development. As additional confirmation of robust growth, we incorporated pulse labeling of the eye disc cells with bromodeoxyuridine (BrdU) for two hours in the beginning and two hours at the end of the eighteen hour time period. BrdU only labels cells that are actively in the synthesis phase of the cell cycle at the time of pulse. If the tissue was in fact alive it would be indicated by the visualization of two populations of cells stained with BrdU. We have also assessed the vitality of the tissue by looking at green fluorescent protein (GFP) expression in a strain of lozenge flies homozygous for GFP expression in the eye and brain. If expression of GFP increased over time that would indicate growth and development. In the three movies we completed, we found that we were able to successfully keep the tissue vital and that it did in fact grow and develop. Currently, we are investigating these movies for additional information about the fly's eye development.


Joshua McFarlane, Rhodes College, Memphis, TN
(Advisor: Dr. Adam Linstedt)

Towards the Understanding of the Function of Giantin, a Golgi protein
Giantin is a C-terminally anchored Golgi protein possessing a large rod-like cytoplasmic domain suggesting a role in maintenance of Golgi structure. In order to investigate this possibility, we attempted to reduce giantin function in vivo by introducing anti-sense RNA and by introducing pieces of giantin's cytoplasmic domain (intended to function as competitive inhibitors by binding to substrates normally bound by giantin). Golgi structure was subsequently evaluated by immunoflourescence using both a polyclonal antibody for giantin which bound to the protein pieces introduced, and an antibody against GPP130, another Golgi protein, which allowed observation of Golgi structure independent of the protein under investigation. Technical difficulties with the anti-sense approach have so far prevented results from being obtained, but introduction of the cytoplasmic pieces was accomplished through transfection of portions of giantin cDNA. No changes in Golgi structure were associated with introduction of these portions, but reassembly of the Golgi following treatment with Brefeldin A (BFA) seems to have been impaired in transfected cells. BFA causes a reversible emptying of the Golgi into the endoplasmic reticulum (ER). As reversal depends on transport from the ER to the Golgi, this impairment suggests that giantin may have some involvement in vesicle mediated transport.


Jennifer E Pogoriler, Williams College, Williamstown, MA
(Advisor: Dr. John L. Woolford)

Identification of Regions in the Yeast Ribosomal Protein rp59 Involved in its Interaction with CRY2 pre-mRNA
The yeast ribosomal protein rp59, an essential component of the 40s subunit, is encoded by duplicate genes CRY1 and CRY2. It provides a model to study the balanced expression of duplicated and single copy ribosomal genes which results in equimolar accumulation of ribosomal proteins. In wild type cells, CRY1 is expressed at ten fold higher levels than CRY2; however, deletion of CRY1 leads to a five to ten fold increase in CRY2 expression. This and other experiments in which rp59 expressed from either gene was shown to repress expression of CRY2 suggest a feedback mechanism of regulation. The regulatory region of the CRY2 gene has been mapped to the 5' exon and first 62 nucleotides of the intron. Both the sequence and structure of the RNA encoded by this region are important for regulation, suggesting that rp59 interacts with the CRY2 pre-mRNA to post- transcriptionally control expression. This provides a mechanism for coupling expression of rp59 with assembly of ribosomal subunits, as excess unassembled rp59 could repress expression of CRY2 by binding to its pre-mRNA.

An in vivo interaction between the regulatory region of CRY2 pre-mRNA and rp59 has been demonstrated using the yeast three hybrid system. However, few RNA-protein interactions are well understood, and rp59 shows little homology to known RNA-binding proteins. To better understand this interaction we wished to determine which regions of the protein are involved in RNA binding. We looked for mutations in rp59 which allowed it to bind the CRY2 pre-mRNA more strongly by selecting mutants which showed increased reporter gene expression in the three hybrid system. Three mutations, D52G, E55G, and E55K were found which showed higher expression of both reporter genes, suggesting that this region may be involved in binding CRY2 pre-mRNA.


Matthew D. Zimmerman, Juniata College
(Advisor: Dr. Gordon Rule)

Cocrystallization Trials of the Binding Domain of Transcription Termination Factor Rho Complexed with Oligonucleotide
Rho is a hexameric protein involved in the termination of transcription of many genes in E. coli. Although no definitive atomic structure of rho has yet been produced, the RNA binding domain is similar to the family of cold-shock-domain DNA/RNA binding proteins (Briercheck, D.M., et al. J. Bio NMR. 8:429-444, 1996). We attempted to cocrystallize Rho130 (the N-terminal fragment containing only the RNA binding domain) with DNA oligonucleotide substrates (since Rho130 has equal affinity for RNA and DNA), to study the protein:nucleic acid interaction by X-ray crystallography. Oligonucleotides d(CCCC) and d(CCCCC) were each mixed with purified overexpressed Rho130 protein, with DNA in slight molar excess. The Rho130:DNA mixture was then sparse matrix screened (Jancarik, J. and Kim, S.H., J. Appl.Cryst. 24:409-411, 1991) by hanging drop vapor diffusion. Crystals of multiple morphologies were obtained, some of suitable size for X-ray diffraction. The diffraction data is pending.