|
The Giovanni Armenise-Harvard Foundation Cancer Biology, Genomics, and Post-Genomics Presentations - Day Two
For a generation of biologists, sequencing the human genome was the most important challenge in their field. Since initial sequencing was completed in June 2000, the focus has shifted to another monumental task: identifying all the proteins encoded by genes and defining their function. As the first step toward this goal, the Institute of Proteomics at Harvard Medical School is building a large scale, state-of-the-art repository of all known genes, and predicted open reading frames (sequences that encode proteins), from humans and from other important organisms. Thousands of human and animal genes will be packaged in a uniform, easy-to-use format called FLEXGeneä (Full-Length Expression-Ready). Researchers can order hundreds or thousands of genes from the FLEXGeneä Repository and transfer them, in a single overnight step, into whatever expression vectors their experiments require. "What took weeks can now be done overnight," Dr. LaBaer said. A major roadblock in high-throughput protein analysis has been the labor-intensive task of expressing all the needed proteins in a way appropriate for the experiment. The old approach required having the desired cDNAs available, then sub-cloning them in-frame in appropriate expression vectors. Using traditional methods, this step is tedious for a single gene, and simply not feasible for most labs if hundreds of genes were needed. For genes in the FLEXGeneä Repository, however, large numbers could be done very quickly. Moreover, the cDNAs are configured to allow the preparation of proteins with or without N- and/or C-terminal fusion peptides as needed. Stock in the Repository is growing at the rate of about 500 clones per week, Dr. LaBaer said, drawn from the human, drosophila, malaria, and P. aeruginosa genomes. The rate-limiting step for preparing FLEXGeneä plasmids turns out to be sequence validation, which is essential for quality control. More than 40 institutions and companies have joined Harvard Medical School in backing this project.
There are competing theories about how the mammalian cerebral cortex takes shape during early embryonic development. Proponents of the "protomap" model hold that the primordial cortex holds all the information needed to make a cortex, while those in the "protocortex" camp see the primordium as raw material that is shaped into a cortex by external signals. Reality, Dr. Mallamaci said in prefacing his remarks, will probably turn out to be somewhere in the middle. His laboratory has analyzed how manipulating the expression of two homeogenes, Emx2 and Pax6, affects primary neurogenesis and shapes the profile of the cortical area. The researchers suspected that these genes might help organize the cerebral cortex because they are expressed in a graded fashion in neocortex. In a knockout mouse with no Emx2 expression, Dr. Mallamaci and his colleagues saw an obvious enlargement of frontal-motor areas as well as a dramatic reduction of occipital-visual areas at the end of cortical primary neurogenesis. Distortions continued later in development as well. When Pax6 was disabled, the researchers observed complementary abnormalities; for example, the occipital-visual region was abnormally large, not undersized as it was in mice without Emx2. A series of experiments with knockout mice lacking both Emx2 and Pax6 led the researchers to several conclusions about the role of their gene products in early neurogenesis. It appears that the two work together in a kind of "balancing act," Dr. Mallamaci said, each shaping part of the neocortex while at the same time down-regulating activity of the other, so that different regions develop in the correct proportions to one another. The absence of functional EMX2 or PAX6 proteins not only reduced size of the caudal-medial and rostral-lateral regions, as predicted, but also impaired the WNT signaling center at the medial-caudal edge of the cortical field. These results suggest that pre-neurogenic cortical regionalization may rely on interactions between these two transcription factors and that later abnormalities in the double knockout mouse model may arise from mistakes in the molecular protomap and distortion of the cortical growth profile.
Scientists interested in infectious disease have unprecedented tools available in the post-genome era, and these may enable them to determine exactly why a pathogen that causes mild symptoms in one person may cause life-threatening illness in another. Dr. Lory's group focuses on interactions between Pseudomonas aeruginosa and cultured cells from the human respiratory system, a clinically important interface because P. aeruginosa respiratory infections account for more morbidity and mortality among cystic fibrosis patients than any other cause. Two resources allow them to scrutinize both sides of the infection story. The genome of P. aeruginosa has been completely sequenced, and Dr. Lory's team used this database to create DNA microarrays for analyzing global patterns of gene expression during P. aeruginosa infection of susceptible individuals. The researchers exposed these microarrays to environmental conditions that mimic the respiratory mucosa, which influenced the expression of intracellular effector proteins secreted by P. aeruginosa. These experiments revealed a complex regulatory network , which activated or repressed several known genes and a large number of unknown ones. Using a series of P. aeruginosa mutants with defects in several virulence traits, comparisons of mRNA data indicated that various combinations altered the activity of key organelles, such as secretory machinery that manufacture enzymes for penetrating host cells. Libraries of expressed human genes enabled the researchers to see what infection looked like from the host's point of view. Dr. Lory and his colleagues used DNA arrays of human genes to determine what signaling mechanisms were mobilized by respiratory epithelial cells following P. aeruginosa adhesion. Engineered mutants of P. aeruginosa were used to assess the contribution of individual virulence factors to the overall host response. This approach will be also used to define the defects in host response in various disease states, such as chronic respiratory infection in individuals with cystic fibrosis. The long-term goals of this research are to identify innate defense mechanisms that might be boosted to help susceptible people fight off respiratory infections.
The use of viruses as delivery vehicles for gene therapy and immunization has become increasingly important in recent years. In Dr. De Palma's laboratory, a specially packaged form of HIV is being used to learn more about a specific type of cell that is a key player in cancer. Angiogenesis, the formation of new blood vessels, is now seen as the rate-limiting step in tumor growth. Unlike tumor cells, the cells in angiogenetic vessels appear to be the same no matter what type of tumor the vessels supply with blood. Because they are so distinctive, these vascular cells appear to be a likely target for gene therapy aimed at fighting cancer. In order to design effective therapy, however, scientists need to know where the cells lining tumor vessels originate and how they grow within the tumor. In order to explore these questions, Dr. De Palma and his colleagues engineered a vector that will selectively express the genes it carries in the endothelial cells of tumor vessels. The vector incorporates a marker gene as well as structural genes from HIV, which facilitate incorporation into the nucleus, all wrapped in the envelope of VSV (vesticular stomatitis virus). In mice with tumors, the researchers used this new vector to selectively mark ECs engaged in tumor angiogenesis and to trace their origin and growth pattern in vivo. As expected, IV infusion of this vector lit up the tumor vessels themselves; it also expressed in cells in the tumor periphery. More surprisingly, Dr. De Palma reported that it penetrated hematopoetic precursors of these endothelial cells. This is the first indication that bone marrow cells contribute to angiogenesis, a finding that will be pursued in future studies.
Epstein-Barr Virus Nuclear Protein 1 (EBNA1) enables the EBV genome to persist as a transcriptionally active episome, a packet of bacterial DNA can replicate outside the nucleus, in proliferating lymphoblastoid cell lines (LCLs). Dr. Kieff's laboratory has a long-standing interest in EBV nuclear proteins in a less benign role, as contributors to the transformation of normal lymphocytes into hyper-proliferative cells. When one of his colleagues raised the possibility of using EBNA1 and its cognate cis acting element, oriP, in human gene therapy, this raised an obvious safety issue. In an effort to determine whether EBNA1 and oriP are safe for therapeutic us of r. Kieff's group began by determining which components of EBNA1 mediate episome DNA replication and which bind the viral DNA to human chromosomes. The researchers found that EBNA1 C-terminus, aa 379-641 could not bind to chromosomes, but was needed Ü along with oriP Ü for episome copying. The EBNA1 N-terminus, aa 1-386, mediated chromosome association and episome persistence. At least two large domains, aa33-88 and 328-382, are both required for wild type chromosome binding. The EBNA1 distribution on chromosomes and salt elution from chromosomes were similar to HMG-I and Histone H1. Dr. Kieff's team engineered chimeric HMG (aa 1-90)-or Histone H1-EBNA1 C-terminus fusion proteins that could substitute for the EBNA1 N-terminus in diffuse, tight, association with mitotic chromosomes. In short- and long-term assays, these fusion proteins mediated tight association of oriP episomes with chromosomes and functioned much like EBNA1 in maintaining episome replication. These experiments open the possibility of creating a "humanized" episome maintenance system, which could be useful in gene therapy. They also evaluated EBNA1's oncogenic potential by transgenic expression of EBNA1 in T and B lymphocytes from three lineages of FVB mice. After 18 months or more, there was no increase in lymphomas or enlarged lymphoid organs compared with age-matched control animals. Finally, they tested the potential impact of EBNA1 on lymphoid cell line growth and survival by conditionally expressing dominant negative EBNA1 (DNE) in an cell line where the EBV genome was integrated into cellular DNA. DNE did not affect EBV gene expression, cell growth and survival, or cell gene expression. Neither DNE nor EBNA1 had any effect on oriP dependent luciferase expression from plasmids that had integrated into cell DNA, whereas Gal4-VP16 activated expression from episomes or from integrated DNA. This is important additional evidence that EBNA1 is unlikely to have any direct effect on cell gene transcription, and that therapeutic use is unlikely to cause malignancy.
Muscle wasting is a physiological response to fasting, and a characteristic, debilitating feature of cancer cachexia, diabetes mellitus, sepsis, chronic renal failure, spinal injury, and other physiologically stressful conditions. The common thread that links these disease states is that muscle proteins are rapidly lost due to accelerated degradation by the ubiquitin-proteasome pathway. Exactly how skeletal muscle proteins are marked for proteolysis, and what ubiquitinating enzymes are involved, is poorly understood. In order to gain a comprehensive understanding of transcriptional adaptations leading to enhanced proteolysis during various types of atrophy, Dr. Goldberg's team used cDNA microarrays to compare normal and atrophying muscles. Fasted mice were initially analyzed because the various changes in energy metabolism and protein breakdown in fasting have been well characterized. Surprisingly, food deprivation had no impact on 95% of gene transcripts in muscle. As expected, the researchers saw expected increases in mRNA levels for familiar genes involved in ubiquitination and proteolysis. The most exciting new finding from the microarrays was a group of novel genes they call atrogins (for atrophy-specific genes), including one that was expressed at high levels in skeletal muscle during fasting, but which fell within hours when feeding was resumed. Dr. Goldberg's group cloned this gene, which they designated atrogin-1. It encodes a protein that contains an F-box domain, which typically functions as a substrate-binding subunit of SCF-type ubiquitin protein ligases. Experiments confirmed that the F-box of atrogin-1 does bind a central component of SCF complexes. Although atrogin-1 is expressed primarily in skeletal muscle, it is also found in cardiac muscle. In addition to being triggered by food deprivation, its mRNA also increases 7-10 fold in muscles atrophying due to cancer cachexia, diabetes, or renal failure (despite normal food intake). Glucocorticoids, which are essential for the activation of proteolysis in these states, can also induce its expression. These findings indicate that atrogin-1 plays a central role in muscle wasting in several clinically important disease states, Dr. Goldberg concluded.
During normal brain development, cyclin-dependent kinase 5 (CDK5) teams up with p35, a regulatory protein, to guide newborn neurons into their proper place in the brain. But when CDK5 pairs up with a p25, a fragment of p35, Dr. Musacchio and collaborators at Harvard Medical School have found, the combination is hazardous to the brain. For starters, the two transform a normal brain protein, tau, into a potent toxin that is associated with neuron death in stroke and Alzheimer's. The CDK-p25 duo no doubt alters other proteins as well, and the researchers are working to discover them. As a structural biologist, Dr. Musacchio's focus is on the crystal structure of the CDK5-p25 complex. It contains a single copy of a helical assembly similar to the Cyclin-box fold, and although this bears a structural resemblance to other cyclins, it displays an unprecedented mechanism for the regulation of a cyclin-dependent kinase. Binding to p25 tethers the unphosphorylated T-loop of CDK5 in a conformation typical of active proline-directed kinases. In the normal pairing of CDK5 and p35 during brain development, residue Ser159 in this loop contributes to the specificity of the CDK5-p35 interaction. When this serine is replaced with threonine, p35 binding is prevented, while the presence of alanine neither affects binding nor kinase activity. His analysis also showed involvement of the activator subunit in substrate recognition, and provided evidence that the CDK5-p25 complex uses a novel mechanism to establish substrate specificity.
It is well known that steroids work by turning on receptors that are members of a large superfamily of ligand-activated transcription factors. Progesterone is known to cross the plasma membrane, bind and activate the nuclear progesterone receptor, and activate (or repress) the transcription of specific genes important for the growth and functioning of target cells and tissues. Dr. Ruderman's research focuses on a second, less appreciated scenario for progesterone activity. In addition to regulating gene transcription, steroids such as progesterone can be important players in cytoplasmic signal transduction pathways. The oocytes of the Xenopus frog sit quietly in G2 arrest for long periods, until they are activated by progesterone. The hormone breaks the G2-arrest of the meiotic oocyte and sets in motion a poorly understood chain of events that involves translational activation of most mRNA, MAP kinase activation, and cell cycle re-entry. These events prepare the egg for a meeting with the sperm, and all of this occurs independently of transcription. Other researchers have shown what when estrogen and progesterone activate signal transduction pathways in somatic cells, they do so via the conventional estrogen or progesterone receptors. But what happens in germ cells? For years, the identity of the progesterone receptor in Xenopus oocytes has been elusive. But the discovery that other steroid receptors might be dual-function structures led Dr. Ruderman to ask whether the conventional progesterone receptor might activate cytoplasmic signaling in frog oocytes. Her team cloned what appears to be a conventional Xenopus PR (XPR-1) that also has the ability to signal, and their experiments locate it in the cytoplasm, rather than on the surface as expected. Now the researchers are screening for proteins that may function in a signaling pathway downstream of XPR-1, as well as exploring possible cross-talk with membrane receptors.
Hepatocyte growth factor, more recently known as plasminogen-related growth factor (PRGF), is among the scatter factors that control events involved in normal growth and in malignant transformation and metastasis. In epithelial cells, PRGF activates a genetic program in which cells dissociate from one another ("scattering"), then grow and become invasive. Although other growth factors trigger scattering, Dr. Medico realized that PRGF specifically stimulates invasive growth. He set out to learn why. To identify genes involved in the onset of invasive growth, Dr. Medico's team compared the transcriptional response of mouse liver cells to PRGF and to epidermal growth factor (EGF). Like PRGF, EGF activates a tyrosine kinase receptor in liver cells; unlike PRGF, it does not trigger invasive growth. Dr. Medico's team used two different commercial microarray platforms to sort through some 20,000 genes that might be involved in PRGF-triggered invasive growth. One technology used high-density spotted cDNAs and the other relied on in-situ synthesized oligonucleotides. Comparing their results winnowed the 20,000 candidate genes down to about 1,000 transcriptionally regulated sequences, with a surprisingly high overlap between the PRGF and EGF responses. If PRGF and EGF act on so many of the same genes, why does one cause invasive growth and the other not? One clue came from Dr. Medico's functional studies of a major transcriptional target, the extracellular matrix protein osteopontin (OPN). While both PRGF and EGF induce OPN, experiments in cultured cells showed that only PRGF promoted adhesion of the cells to newly synthesized OPN through the CD44 receptor. In these experiments, the researchers could see PRGF stimulate the growth of tendrils that extended across the surface and scattered. EGF-stimulated cells put out small fingers, but their growth soon stopped. Dr. Medico concluded that PRGF's ability to stimulate invasive growth does not rely directly on transcriptional regulation, but rather on functional interplay among the products of the regulated genes and the activated PRGF receptor. Symposium Pages
|
Manipulating the Proteome
