fax?! ' . . r... .. u 12. v :2! 1.»...«8 4.2.... E: ,- .E—Qu. v . 5.: { 5\.IV.T 7: .12 . >n. 6(3: ,.. a. v}. .. I l [(7)33 you 29 9 This is to certify that the thesis entitled CHARACTERIZATION OF HDC1 AND HDC2, TWO HISTONE DEACETYLASES FROM COCHLIOBOLUS CARBONUM, A FUNGAL PATHOGEN OF MAIZE LIBRARY Michigan State Universl presented by OSCAR CABALLERO has been accepted towards fulfillment of the requirements for the MASTER OF degree in CELL AND MOLECULAR SCIENCE BIOLOGY %"WAA W Major Professor’s Signature ll lo / 0 L Date MSU is an Afi‘irmative Action/Equal Opportunity Institution PLACE IN RETURN Box to remove this checkout from your record. To AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 6/01 cJCIRC/DateDue.p65-p.15 CHARACTERIZATION OF HDC1 AND HDCZ, TWO HISTONE DEACETYLASES FROM COCHLIOBOLUS CARBONUM, A FUNGAL PATHOGEN OF MAIZE By Oscar Caballero A THESIS Submitted to Michigan State University In partial fulfillment of the requirements For the degree of MASTER OF SCIENCE Cell and Molecular Biology Program 2002 ABSTRACT CHARACTERIZATION OF HDC1 AND HDC2, TWO HISTONE DEACETYLASES FROM COCHLIOBOLUS CARBONUM, A FUNGAL PATHOGEN OF MAIZE By Oscar Caballero Histone deacetylases (HDAC) are enzymes, predominantly nuclear, known to help regulate transcription in eukaryotes. In order to understand the function of HDACs in the filamentous fungus Cochliobolus carbonum, a genetic approach was undertaken. Targeted gene disruptions of HDC1 and HDC2, via homologous recombination, were attempted in C. carbonum. Whereas HDC1 was mutated successfiilly, HDCZ was not despite various attempts. These results suggest that HDCZ is either essential for survival in C. carbonum or the chromatin structure surrounding HDCZ is refractory to recombination. Disruption of HDC1 revealed a novel phenotype in C. carbonum. The hch mutant of C. carbonum displayed delayed growth on polysaccharides and complex carbohydrates. The 12ch conidia were smaller and less septate than the wild type. On maize plants, the hch mutant caused only small lesions and was unable to develop full disease symptoms as the wild type, which ultimately kills the plant. Overexpression of HDCZ in C. carbonum had no apparent phenotype. There was no difference in virulence levels between the HDCZ over-expressor and the wild type. Biochemical analysis of the HDCZ over-expressor provided evidence that the HDCZ product contains HDAC activity in C. carbonum. I would very much like to dedicate this work, however insignificant, to the native inhabitants of my beautiful homeland, Colombia. T o the Colombian indians, those forgotten by time and put into oblivion by an invading culture that never knew how to appreciate who they are and, because of ignorance, has tried at all costs to obliterate them. To the Pijaos, the Guambianos, the Chibchas, the Huitotos, and many, many more who have been there since the inception of time and cared for the land they inhabited with respect and love. This is nothing but a small token of appreciation, nothing but a small way to apologize for taking their rights away and exploiting them, as the white man did and still does. For trying to deprive their people from living ofl the land that once belonged only to them, although I believe there is no way we can be forgiven. Because you were the one and only rightful owners and landlords. For the profound admiration that I feel towards these natives of Colombia, and all the respect that you deserve, today and always. T 0 you and for you. iii ACKNOWLEDGEMENTS I am deeply indebted to my mentor, Dr. Jonathan D. Walton, for his guidance, help, and support throughout my time in his laboratory. My work could not have been possible without many “waltonians” along the way: Dipnath Baidyaroy, John Scott-Craig, and Kerry F. Pedley. I am also thanka to past and present members of the Plant Research Laboratory at large: Ahmed Faik, Rodrigo Sarria-Millan, Ivan Delgado, and the secretarial staff. I appreciate the help and guidance given by my thesis committee members, Dr. Triezenberg, Dr. He, Dr. Kuo, and Dr. Fullbright. I am also appreciative of the advice and help given by Dr. Conrad, director of the Cell and Molecular Biology program. Finally, I am forever grateful to my family; my father Angel Maria, my mother Maria del Pilar, my brothers Miguel Antonio and Angel Maria Jr., for always showing me that “the glass is not half-way empty, but half-way full.” iv TABLE OF CONTENTS LIST OF FIGURES ................................................................................ vi LITERATURE REVIEW .......................................................................... 1 CHAPTER ONE CHARACTERIZATION OF THE HISTONE DEACETYLASE HDC1 FROM COCHLIOBOL US CARBONUM ................................................................ 21 Introduction ................................................................................ 21 Results ...................................................................................... 22 Discussion .................................................................................. 33 Materials and Methods ................................................................... 36 CHAPTER TWO STUDIES ON THE HISTONE DEACETYLASE HDC2 FROM COCHLIOBOL US CARBON UM ........................................................................................ 39 Introduction ................................................................................ 39 Results ...................................................................................... 41 Discussion ................................................................................. 55 Materials and Methods ................................................................... 59 CONCLUSIONS AND PERSPECTIVES ...................................................... 62 LITERATURE CITED ............................................................................ 68 LIST OF FIGURES Figure 1. Unrooted cladogram showing relatedness among histone deacetylases from different fungi ............................................... 16 Figure 2. Amino acid sequence alignment of three HDAC proteins using CLUSTAL W ....................................................... 23 Figure 3. Construction and analysis of the hdcl disruption transformants .................. 25 Figure 4. Microscopical analysis of the hch conidia .......................................... 27 Figure 5. Analysis of HDAC activity in the hch mutant ...................................... 28 Figure 6. Growth comparisons between the hch mutant and wild type ..................... 30 Figure 7. RNA expression profiles of selected genes in the hch mutant versus wild type ....................................................... 31 Figure 8. Time course study of Hch synthesis in E. coli ..................................... 43 Figure 9. Integration of HDC2 at the genomic PGNI locus .......................... 45 and 46 Figure 10. RT-PCR expression analysis of HDC2 in the T709 and wild type strains ................................................................................ 47 Figure 11. RNA gel blot analysis of expression of HDC2 in T709 versus wild type strains ..................................................................... 49 Figure 12. The effect of HC-toxin on HDAC activity comparing T709 versus wild type strains as a function of grth ......................................... 51 Figure 13. HPLC separation of HDAC activity comparing T709 and wild type strains ......................................................................... 53 Figure 14. Sensitivity of the T709 and wild type major HDAC activity peaks to HC-toxin ......................................................................... 56 Figure 15. Virulence assays comparing wild type versus T709 strains .................................................................................................. 57 vi LITERATURE REVIEW Among all plant diseases, the ones caused by fungal pathogens are the most devastating type. Fungi can be defined as small, generally microscopic spore-bearing organisms that lack chlorophyll and have cell walls that contain chitin, cellulose, or both. Most of the 100,000 fungal species known are strictly saprophytic, living on dead organic matter, which they help decompose. About 50 species cause diseases in humans, and about as many cause diseases in animals, most of them superficial diseases of the skin or its appendages. More than 8,000 species of fungi, however, can cause diseases in plants. All of the species of flowering plants (over 300,000) are attacked by pathogenic fungi. However, a single plant species can be host to only a few fungal species, and similarly, most fungi usually have a limited host range. Throughout history, the vast majority of important crop diseases have been caused by phytopathogenic fungi. Significant annual crop yield losses have made fungal pathogens of plants a serious economic factor. For instance, 15% ($33 billion) of the total rice production between 1988 and 1990 were lost to fungal diseases, 12.4% ($14 billion) in the case of wheat, and 10.9% ($7.8 billion) in the case of maize (Oerke et al., 1994). When a fungal spore encounters a plant, it must be able to either penetrate the host tissue or tap an external source of the host’s nutrients if it is to survive. Penetration, the method employed by most phytopathogenic fungi, can occur either enzymatically or mechanically. Some fungi secrete a large variety of enzymes that can break down the plant cell wall, including cutinases, cellulases, pectinases, and proteases. Cell wall degrading enzymes (CWDE) probably did not evolve particularly as pathogenicity factors. All fungal species that live in a saprophytic fashion can also secrete enzymes necessary for the digestion of plant cell wall polymers even though they are not pathogenic. It is believed that CWDEs contribute to the virulence of the fungal pathogen but are not essential pathogenicity factors. Removal of a regulatory factor that controls CWDE expression in the fungus Cochliobolus carbonum shut off their transcription and the mutant displayed a significant reduction in virulence (Tonukari et al., 2000). Other fungi exert tremendous amounts of pressure and penetrate by sheer mechanical force. Magnaporthe grisea, the causal agent of rice blast disease, requires formation of an appressorium for plant infection. During penetration, M. grisea appressoria can generate as much as 8.0MPa of turgor pressure. This is the result of synthesis of large quantities of glycerol, a compatible solute, in the appressorium (deJong et al., 1997). Turgor is translated into mechanical force and this forces a thin penetration hypha through the plant cuticle. Other fungal species, including some rusts, have not evolved a direct penetration mechanism and instead bypass the plant cuticle and outer cell wall by entering through stomates. Stomates are the small openings on the leaf epidermis important for gas exchange in the plant. The bean rust Uromyces appendiculatus, for instance, uses a thigmo-responsive mechanism that allows it to find its way into the stomates (Correa and Hoch, 1995). After penetration, many fungi secrete toxins or plant hormone-like compounds that manipulate the plant’s physiology to the pathogen’s benefit. The end result may simply be host cell death for the purpose of nutrient uptake, or a more subtle redirecting of the cellular machinery via the production of phytotoxins with varying degrees of specificity toward different plants. Some toxins are host selective, whereas others are active in a wide range of plant species. Host-selective toxins (HST) can be generally described as low molecular weight compounds with diverse structures that act as positive agents of virulence or pathogenicity (Walton, 1996). HSTs can help determine host range or specificity in that plant species, such that genotypes sensitive to an HST are found to be susceptible to the producing pathogen. All HSTs known are produced by fungi and most of them can be classified as secondary metabolites. They are low molecular weight compounds of diverse structure that are restricted in their taxonomic distribution and are not necessary for normal survival and reproduction. HSTs are active at concentrations ranging from 10 pM to luM, and their degree of specificity (host selectivity) ranges from lOO-fold to >106 fold (Walton et al., 1985). Despite the diverse array of secondary metabolites produced by fungi, they are all thought to be synthesized from a limited number of primary metabolites modified in unique ways. The major biosynthetic pathways for secondary metabolites include the isoprenoid pathway, the polyketide pathway, the shikimate pathway, and the use of amino acids as precursors (Bentley, 1999). In addition, some compounds are derived from carbohydrates, intermediates of the tricarboxylic acid cycle, and combinations of multiple pathways. The Model System The Walton laboratory is studying the interaction between the filamentous fungus Cochliobolus carbonum and the maize plant (Zea mays L.), causing the disease commonly known as northern leaf spot and ear mold. On a sensitive corn variety (Pr), race 1 of C. carbonum will give rise to well-defined, zonate, rapidly spreading lesions on the foliage and pronounced black mycelium on the kernels of infected ears. Race 2, the other race of C. carbonum, is much less virulent when compared to race I, unable to colonize much beyond the site of penetration, and causing only mild chlorotic-necrotic flecks on the leaves. This fungus was originally called Helminthosporium carbonum until 1959, when its sexual stage was discovered. It was then determined that this pathogen was an ascomycete belonging to the genus Cochliobolus (Nelson, 1959). Genetic crosses between race 1 and race 2 isolates of C. carbonum revealed that virulence is determined by a single genetic locus, T 0X2, which also confers the ability to produce the secondary metabolite (HC)-toxin, where HC stands for flelminthosporium garbonum. HC-toxin was proven to be a bona-fide HST as it can inhibit root growth in susceptible maize but not in other related plant species (Scheffer and Ullstrup, 1965). HC-toxin is a cyclic tetrapeptide with the structure cyclo (D-Pro-L-Ala-D-Ala-L-Aeo), where Aeo stands for 2-amino- 9,10-epoxi-8-oxodecanoic acid (Walton et al., 1982). Molecular analysis of the T 0X2 locus led to the discovery of some of the genes required by race 1 isolates of C. carbonum to produce HC-toxin. The genes of the T 0X2 locus are loosely clustered within ~54O Kb of DNA and are unique to race 1 isolates (Ahn and Walton, 1996); (Ahn et al., 2002). This large and complex locus contains multiple copies of all the genes that appear to be needed for HC-toxin biosynthesis. The cyclic peptide synthetase, encoded by H T S] , is a non-ribosomal peptide synthetase required for the synthesis of HC-toxin. The open reading frame is 15.7 Kb in size, there are two copies in race 1 isolates of C. carbonum, and the enzyme encoded has a molecular weight of 570 KD (Scott-Craig et al., 1992). Located upstream of HT SI there is another gene, T OXA, which is also present only in HC-toxin producing isolates. The predicted product of TOXA exhibits a high degree of similarity to members of the major facilitator superfamily (MF S), which encode membrane-localized antibiotic efflux pumps (Pitkin et al., 1996). Three other genes, T OXC, T 0207 , and T OXG, are also unique to race 1 isolates and have been shown to be essential for HC-toxin biosynthesis. The predicted product of TOX C is highly similar to the B-subunit of fatty acid synthases from several lower eukaryotes, and contains domains predicted to encode acetyl transferase, enoyl reductase, dehydratase, and malonyl-palmityl transferase (Ahn and Walton, 1997). T OXF is predicted to encode a protein with moderate homology to many known or putative branched-chain-amino-acid transarninases (Cheng et al., 1999). Genetic analyses demonstrated that both T 0X C and T OXF are required for HC-toxin production and C. carbonum virulence. Another gene, T 0X G, was found to encode an alanine racemase, based on sequence comparisons and biochemical evidence (Cheng and Walton, 2000). Also within the T 0X2 locus are T 0M) and T OXE . Targeted disruption of T OXD did not unveil any change in either HC-toxin synthesis or pathogenicity (Y .Q. Cheng and JD. Walton, unpublished results). Interestingly, T OXE seems to be the regulatory factor for at least part of the T 0X2 locus. Deletion of T OXE resulted in loss of HC-toxin production and reduced virulence. In addition, transcripts of T OXA, T OXC, T 020), T0”, and T 0X G are down-regulated in the T OXE mutant (Ahn and Walton, 1998). T OXE has four ankyrin repeats and a basic region similar to those found in basic leucine zipper (bZIP) proteins, but lacks any apparent leucine zipper. It was demonstrated that TOXE is a DNA- binding protein that recognizes a ten-base motif (the “tox-box”) without dyad symmetry that is present in the promoters of all of the known genes present in the T 0X2 locus (Pedley and Walton, 2001 ). In maize, a single dominant gene, HM], governs resistance and confers complete protection at all stages of growth against C. carbonum (Nelson and Ullstrup, 1964). Using a cell-free extract from a resistant maize genotype (Hm I/hml), it was shown that HC-toxin could be inactivated via a reduction of the 8-carbonyl group at the Aeo group by means of an enzymatic activity that was NADPH dependent (Meeley and Walton, 1991). Subsequently, this HC-toxin reductase activity (HCTR) was shown to be present in all maize extracts from resistant genotypes (Hm I/hmI, HmI/HmI) tested, but absent in susceptible ones (hm 1/hm1) (Meeley et al., 1992). The HM] gene was cloned and proven to be similar to known NADPH-dependent reductases (Johal and Briggs, 1992). Furthermore, the cloning of HM 1 revealed that HCTR activity alone is sufficient to prevent severe infection by C. carbonum race 1 (Meeley et al., 1992). HCTR activity is detectable in extracts of several other grasses (e.g., barley, oats, and wheat) and hence may represent an ancient resistance strategy within the Poaceae against HC-toxin and similar compounds. The strong DNA sequence similarities between various HCTRs supports the idea that the function may be the same among different plant species (Han et al., 1997). The biological role of HC-toxin in C. carbonum that allows the fungus to colonize maize is not yet known. HC-toxin is unique because when compared with other phytotoxins, it appears to be cytostatic rather than cytotoxic (Wolf and Earle, 1991). Instead of killing the host cells ahead of the growing hyphae, HC-toxin is thought to suppress the active defense responses that will typically be mounted by the host against pathogen attack (Cantone and Dunkle, 1990). A bioassay for testing HC-toxin activity consists of root growth inhibition in germinating maize seedlings. However, unlike most other phytotoxins, HC-toxin only weakly promotes ion leakage (Y oder, 1980). HC-toxin is soluble in chloroform and water but not in solvents of intermediate polarity such as ether. The biological significance of this is that HC-toxin should be able to move readily through both the hydrophobic and hydrophilic domains of plant and fungal tissues. HC- toxin is also active against mammalian cells (Walton et al., 1997). The first major advance in understanding the mode of action of HC-toxin came from studies on the mode of action of trapoxin, an Aeo-containing fungal secondary metabolite (Itazaki et al., 1990). Trapoxin was shown to induce morphological reversion from transformed to normal in sis-transforrned NIH—3T3 fibroblasts (Kijima et al., 1993). In addition, trapoxin was found to cause accumulation of highly acetylated histones by binding irreversibly to histone deacetylases (HDAC), a family of enzymes primarily localized to the nucleus. Based on this knowledge, Schreiber and colleagues synthesized in vitro a trapoxin affinity matrix and used it to isolate and later clone the gene for the first mammalian HDAC, a human ortholog of the yeast transcriptional regulator RPD3. This gene, initially called HDI, is now called HDA C1, and its product has intrinsic HDAC activity (Taunton et al., 1996). Histone Deaceglases In all eukaryotes, the nuclear genetic material is arranged in a highly complex structure made up of histones and DNA called chromatin. The basic unit of structure in chromatin is called the nucleosome. Each nucleosome contains 147 base pairs of DNA and a histone octamer. The histone octamer is composed of a dimer of each of the core histone proteins: H3, H4, H2A, and H2B (Kornberg and Lorch, 1999). Histones are small basic proteins consisting of a globular domain and a more flexible and charged amino terminus (histone “tail”) that protrudes from the nucleosome. The function of HDACs is to remove the acetyl moieties from the a-amino group of specific lysine residues present at histone tails previously acetylated. The acetylation of histone tails is enzymatically mediated by a family of enzymes known as histone acetyl-transferases (HAT). HATS take Acetyl-CoA as substrate and attach the acetyl moiety to the e-amino group of lysine residues. This type of reaction has been estimated to have a large negative AG value (-7.5 kcal/mol), which makes it thermodynamically favorable and able to occur spontaneously (Stryer, 1995). Because HDACs are part of nuclear complexes known to repress gene transcription, they are called “co-repressors.” HATS, on the other hand, are called “co- activators” as they have been implicated in activation of gene expression. Because HC-toxin, like trapoxin, contains an Aeo tail, it was inferred that HC- toxin could be also an HDAC inhibitor. Walton and collaborators tested this hypothesis and found that HC-toxin can inhibit HDAC activity not only from maize but also chicken, the myxomycete Physarum polycephalum, and the yeast Saccharomyces cerevisiae (Brosch et al., 1995). In addition, they found that HC-toxin is an noncompetitive inhibitor and its binding is reversible. In a follow-up study, Ransom and Walton (1997) discovered that treatment in vivo of maize embryos and tissue cultures with HC-toxin leads to an accumulation of hyperacetylated forms of histones H3 and H4, but not H2A and HZB, in Pr, the sensitive genotype. Further, the fact that accumulation of hyperacetylated histones began 24 hours after inoculating maize leaves with C. carbonum, which is before development of visible disease symptoms, argues that inhibition of HDACs by HC-toxin is necessary for C. carbonum pathogenesis (Ransom and Walton, 1997). Histone Deacetylase Families HDACs have been organized into three distinct classes. Class I includes mammalian HDAC 1, 2, 3, and 8, which are related to RPD3 from S. cerevisiae. These HDACs have been found to be part of similar repressor complexes, have similar sensitivity levels to trichostatin A (TSA, an HDAC inhibitor), are similar in size, and share a well-conserved catalytic domain (Rundlett et al., 1996). Class 11 includes mammalian HDAC 4, 5, 6, and 7, which are related to HDAI, HOS], H052, and H083 from S. cerevisiae, respectively (Grozinger et al., 1999). Members of class II are much greater in size than members from class I and also exhibit much greater sensitivity to TSA (Khochbin et al., 2001). Class III, which is composed of HDACs related to SIR2 from S. cerevisiae, is involved in heterochromatin silencing at silent mating loci, telomeres, and ribosomal DNA (Moazed, 2001a). Unlike the other two classes, SIR2 has in vitro NAD+-dependent HDAC activity as well as ADP-ribosyltransferase activity (Imai et al., 2000). Heterochromatin can be defined as the densely staining regions of the nucleus that generally contain condensed, trancriptionally inactive regions of the genome. Euchromatin, on the other hand, contains the decondensed, transcriptionally active regions of the genome. A novel plant-specific class of HDACs has been discovered recently. HDZ, a member of this new class, was isolated from maize embryos and immunologically localized to the nucleolus (Lusser et al., 1997). HDZ is a nucleolar phosphoprotein that might regulate ribosomal chromatin structure and function. In Arabidopsis thaliana, two genes with high similarity to HD2 from maize were identified. It was shown that these HD2-like HDACs not only are expressed in various Arabidopsis organs but also appear to be important in its reproductive development (Wu et al., 2000). Histone Deaceglase Function The action of HDACs, in theory, causes a stronger attraction between the histone tails and DNA phosphate backbone via stronger ionic interactions. At promoter regions, this process could potentially tighten up chromatin structure so that genomic DNA becomes less accessible. Therefore, the function of HDACs could contribute to repression of gene expression (N g and Bird, 2000). The first association between histone deacetylation and transcriptional repression came from genetic studies on nuclear repressor complexes. The yeast SIN3/RPD3 complex, for instance, has been studied in detail. The complex is about 600 KD in size. SIN3 contains several paired amphipathic helix (PAH) domains postulated to be involved in protein-protein interactions (Halleck et al., 1995). SIN3 is thought to act as a molecular scaffold for assembly of the other proteins involved in the complex (Kasten et al., 1997). Struhl and colleagues showed that localized histone deacetylation on a repressed yeast promoter depends on the recruitment of the SIN3/RPD3 complex by UME6, a DNA-binding transcriptional repressor (Kadosh and Struhl, 1997, 1998). More recent work has now shed light on the mechanism by which histone deacetylation may lead to repression of transcription. The evidence presented suggests that de3 -dependent repression is associated with decreased occupancy by TATA binding protein (TBP), the Swi/Snf nucleosome-remodeling complex, and the SAGA histone acetylase complex (Deckert and Struhl, 2002). The 10 authors concluded that the domain of localized histone deacetylation generated by recruitment of de3 might mediate repression by inhibiting recruitment of chromatin- modifying activities and TBP. Another well-characterized complex is the Ssn6/Tupl repressor complex. TUPl interacts directly with the amino-terminal tail domains of histones H3 and H4 in vitro (Edmondson et al., 1996). These tail domains are both necessary and sufficient for Tupl binding. Moreover, the region of Tupl that interacts with the histones closely coincides with a domain that can confer repression independently when fused to LexA (Tzamarias and Struhl, 1994), indicating that the function of Tupl in vivo depends on its interactions with histones. Indeed, mutations in these histone domains synergistically reduce repression of multiple classes of Tupl-regulated genes in vivo (Huang et al., 1997). Tupl binds poorly to highly acetylated forms of H3 and H4 in vitro but interacts very well with unacetylated isoforms. In addition, genes repressed by Tupl in yeast are associated with unacetylated forms of histones H3 and H4 in vivo (Bone and Roth, 2001). Roth and coworkers demonstrated that histone hyperacetylation caused by combined mutations in the HDAC genes RPD3, HOS], and HOS2 abolishes Ssn6/Tupl repression in yeast (Watson et al., 2000). Further, they showed that the Ssn6/Tupl complex can interact with at least two different HDAC proteins, de3 and H032. Tupl was also found to interact with Hdal, which brings specific deacetylation to histones H3 and H2B in vivo (Wu et al., 2001), and this interaction is required for gene repression. Neither Ssn6 nor Tupl can bind DNA directly. They are recruited to individual promoters through interactions with DNA-bound repressors, such as a2/Mcml for repression of STEZ and ST E6, two mating- specific genes (Komachi et al., 1995), and Migl and Skol for repression of ENAI, which 11 encodes a membrane ATPase involved in sodium efflux from the cytoplasm (Proft and Serrano, 1999). Other Histone Modifications In addition to acetylation, histones undergo a variety of post-translational modifications. Histones may be phosphorylated, methylated, ubiquitinated, and nucleosomes may be remodeled in an ATP-dependent fashion (Berger, 2001). Phosphorylation of histones is critical for regulation of several genes involved in the cell cycle control. Rsk2, for instance, is an H3 kinase that when mutated is associated with Coffin-Lowry syndrome in humans (Sassone-Corsi et al., 1999). Transcriptional activation in response to mitogenic and other stimuli are altered in Coffin-Lowry cells, indicating an important role for H3 phosphorylation in regulating gene transcription (De Cesare et al., 1998). Histone methylation is regarded as a more long-term epigenetic mark than other histone modifications, which is consistent with the relatively low turnover of the methyl group (Jenuwein, 2001). This histone modification, which occurs on arginine and lysine residues, is very stable and can be maintained from one generation to the next. Bulk histone methylation steadily increases during the S and G2 phases of the cell cycle, consistent with a role in preparing chromatin for mitotic condensation (Byvoet et al., 1972). As a chemical modification, methylation is not significant in the sense that the overall charge of a lysine residue remains unchanged. However, methylation creates a binding site for heterochromatic proteins such as HPl that contain a chromodomain (Lachner et al., 2001). HPl is involved in propagation of heterochromatic subdomains 12 (Nakayama et al., 2001). All histone methyltransferases (HMT) contain the SET motif, a highly conserved domain 130 amino acids in length responsible for the catalysis of this reaction (Jenuwein, 2001). It has now become evident that specific methylation patterns are correlated with gene activity. H3-K9 methylation seems to be primarily associated with heterochromatin (Noma et al., 2001), whereas H3-K4 methylation (in higher eukaryotes) is observed in transcriptionally active regions (Strahl et al., 1999). DNA methylation has recently been shown to be dependent on histone methylation in some fungi. Methylation of DNA at CpG sites is also an epigenetic mark, which can be inherited through mitosis, and often through meiosis. In Neurospora crassa, it was shown that all DNA methylation is dependent on H3-K9 methylation, suggesting that methylation of H3-K9 occurs prior to DNA methylation (Tamaru and Selker, 2001). Nucleosome remodeling complexes use energy from ATP to modify chromatin structure in a noncovalent manner. Their function is to increase accessibility of nucleosomal DNA, a fundamental requirement for several steps in transcription. Each remodeling complex contains a central ATPase subunit which can alter chromatin structure in the absence of the remaining subunits (Kingston and Narlikar, 1999). The role of the remaining subunits is thought to be targeting and modulation of the activity of the ATPase subunit. There are two main families of nucleosome remodeling complexes: Swi/Snf and Iswi. Swi/Snf can remodel nucleosomes in the absence of histone tails whereas Iswi can not (Langst and Becker, 2001). Swi/Snf complexes can be stimulated similarly by nucleosomes and naked DNA, while Nurf (an Iswi complex) is stimulated significantly better by nucleosomes than by naked DNA (Tsukiyama and Wu, 1995). Two main mechanisms have been proposed as to how chromatin remodeling may occur. 13 Sliding is the translational repositioning of histone octamers, so that the DNA that was originally interacting with histones becomes non-nucleosomal. However, this mechanism can not explain all scenarios of nucleosome displacement, such as the way in which substantial tracts of DNA can be made accessible in regions of tightly spaced nucleosomes. Hence, a second mechanism must exist, whereby the conformation of the histone octamer is changed without displacement, facilitating DNA exposure in regions of closely packed nucleosomes by bringing DNA to the surface of the histone octamer. In addition, it has been shown that RSC, an abundant chromatin remodeling complex, can transfer a histone octamer from a nucleosome core particle to naked DNA (Lorch et al., 1999). Histone Code Hypothesis Histones are integral and dynamic components of the machinery responsible for regulating gene transcription. The histone tails, which protrude from the surface of the chromatin polymer, are subject to a diverse array of covalent modifications. The combinations of these modifications are specifically recognized by individual transcriptional regulatory protein modules. These combinations act as a code or “language” that somehow must be deciphered or “read” by the corresponding protein module or modules (Strahl and Allis, 2000). The histone code hypothesis fits well in instances where research results previously seemed conflicting. For instance, as mentioned before, H3-K9 HMTases such as the Su(var)3-9 methylate histones generating an affinity for HP], which leads to heterochromatin-induced gene silencing in Drosophila (Schotta et al., 2002). By immunological co-localization, it was shown that Su(var)3-9 l4 and HPl have the same site of action. Set-9/Set-7, a H3-K4 HMTase, on the other hand, stimulates transcriptional activation both by competing with HDACs as well as precluding H3-K9 methylation by Suv39h1 (N ishioka et al., 2002). The major point to be drawn from these and similar results is that the specificity of histone residues modified and their combinations serve as platforms for binding transcriptional regulators that will in turn determine whether genes will be silenced or highly expressed. The histone code hypothesis also predicts that histone modifications can influence one another in either a synergistic or antagonistic way, providing a mechanism to generate and stabilize specific imprints (Jenuwein and Allis, 2001). Further support for the hypothesis was provided by Struhl and colleagues. They showed that an increase in acetylation levels at various promoters is not necessarily associated with an increase in transcriptional activation (Deckert and Struhl, 2001). Me Deacetylases in C. carbonum The genome of C. carbonum contains a total of four HDACs, excluding the ones that are NAD-dependent. This information was inferred by examination of a very closely related species of Cochliobolus called C. heterostrophus. The genome of C. heterostrophus was fully sequenced by the company Syngenta Biotechnology and found to contain four HDACs. Each one of the HDACs has its counterpart in C. carbonum and no other HDACs were found in C. heterostrophus. Hence, it is assumed from these data that there are no extra HDACs in C. carbonum. The C. carbonum HDACs all have counterparts in the yeast S. cerevisiae (F igurel). HDAl is part of HDA, a 350KD complex that exhibits greater sensitivity to TSA than the HDB complex 15 No oortig 1.1059 MM CoHdo‘l Sol-00d ._ MdeA 6014602 , Nooorflgtfi‘l 303963 Sol-I081 ”0001119137 A .... MW 1 I 809631 001-1604 |__.[l MHOSB Mommas: '- Sam 242.0 , , r . . 200 150 100 50 0 Figure 1. Unrooted Cladogram showing relatedness among histone deacetylases from different fungi. An, Aspergillus nidulans; Sc, Saccharomyces cerevisiae; Cc, Cochliobolus carbonum; Nc, Neurospora crassa. All No Contigs represent predicted proteins from the eponymous genomic DNA sequences of N. crassa. The units on the scale at bottom are numbers 'of substitution events. Cladogram was created using MegAlign by the DNAStar software package (version 5.03). 16 (Rundlett et al., 1996). The HDA complex contains three HDAC proteins, Hdal , Hda2, and Hda3. When the HDA] gene is disrupted in yeast, the nucleosomes of Tupl- regulated genes are hyperacetylated specifically at histones H3 and H2B. It was also shown that Tupl interacts with Hdal in vitro, suggesting that Tupl recruits Hdal to promoters of Tupl-regulated genes (Wu et al., 2001). HDC3 is the C. carbonum ortholog of HDA] . HDC3 was disrupted in C. carbonum but there was no detectable phenotype (Baidyaroy and Walton, unpublished results). The yeast H053 gene has 38.9% similarity to RPD3 over 271 amino acids and its disruption was found to increase histone H4 acetylation in yeast cell extracts (Cannon et al., 1999). Rather than being part of a large complex, Hos3 was purified as a homo-dimer, and its activity is relatively insensitive to TSA, unlike Hdal and de3 (Carmen et al., 1999). HDC4 is the C. carbonum ortholog of H053. Disruption of HDC4 also results in no detectable phenotype (Baidyaroy and Walton, unpublished results). The other two HDAC genes from C. carbonum, HDC1 and HDC2, are the subjects of this dissertation work. HDC1 is the ortholog of the yeast H052 gene. H052 stands for HDA Qne Similar and was discovered when the histone deacetylase-A (HDA) complex was first purified (Rundlett et al., 1996). The Hdal protein sequence was compared with sequences from GenBank and the best scores obtained included three newly-sequenced open reading frames (ORF) that Grunstein and coworkers termed hosl, and hosZ, and hos3. Despite having no visible phenotype, the hos2 yeast strain was shown, by microarray analysis, to be hyperacetylated specifically at the ribosomal protein genes (Robyr et al., 2002). In addition, H032 was shown to be required for deacetylation of histone H4 K12. HDC2 is the C. carbonum ortholog of RPD3. RPD3 stands for Reduced Potassium Dependency and the rpd3 yeast strain was 17 found in a screen for mutants that were able to grow well under low potassium concentrations (Vidal et al., 1990; Vidal and Gaber, 1991). Microarray analyses have shown that genes up-regulated by RPD3 deletion correspond to cell cycle-regulated genes (Bernstein et al., 2000). However, deletion of RPD3 also down-regulated certain genes, indicating that de3 may also activate transcription. In fact, 40% of endogenous genes located within 20 Kb of telomeres are down-regulated by RPD3 deletion. Unlike H082, it was demonstrated that de3 affects the regulation of genes in virtually all cellular pathways, with. a modest over- representation of genes that take part in sporulation, germination, and meiosis (Robyr et aL,2002) C. carbonum, as any other eukaryote, contains HDACs to modify its own histones. This raises the question: why are the HDACs of C. carbonum not affected by its own toxin? How does C. carbonum protect itself against its own toxin? TOXA is one of the genes found at the T 0X2 locus in the genome of C. carbonum. The function of T OXA is unknown but based on its amino acid sequence it appears to encode a putative HC- toxin efflux pump. The inability to recover toxA knockouts of C. carbonum race 1 isolates supports the idea that ToxA may secrete HC-toxin out of the cytoplasm and is essential for the survival of C. carbonum (Pitkin et al., 1996). This kind of efflux carriers belongs to the Major Facilitator Superfamily (MF S) of transporters (Del Sorbo et al., 2000). MFS transporters do not hydrolyze ATP. Transport of compounds by MFS transporters through membranes is driven by an electrochemical proton gradient. It is now clear that MF S transporters are involved in secretion of HSTs and non-HSTs in several species of plant pathogens. For instance, the soybean pathogen Cercospora 18 kikuchii produces the toxin cercosporin. Callahan and colleagues identified and cloned CFP, a gene from C. kikuchii with similarity to several MFS transporters. Targeted disruption of CFP (Cercosporin Facilitator Protein) resulted in drastic reduction in cercosporin production, greatly reduced virulence, and increased sensitivity to exogenous cercosporin (Callahan et al., 1999). Hence, CFP most likely encodes a cercosporin transporter that contributes resistance to cercosporin by actively exporting the toxin and maintaining low cellular concentrations. This and other results—cg. (Alexander et al., 1999)—support the hypothesis that active secretion of toxins by MFS transporters in fungi is likely to be a common virulence factor. Other self-protection mechanisms seem plausible in addition to TOM. The C. carbonum HDACs may be intrinsically resistant to HC-toxin, having slight changes in amino acid composition that lead to changes in protein folding. Then, HC-toxin would be unable to cause inhibition because its binding site has been altered. In addition, it is conceivable that resistance may be extrinsic. A protein factor unique to C. carbonum could bind HC-toxin and thus prevent it from inhibiting C. carbonum HDACs. Alternatively, to abrogate HDAC inhibition, such protein factor could bind HDACs and cause a conformational change such that HC-toxin can no longer bind. Some evidence supporting the “intrinsic” hypothesis was presented previously (Brosch et al., 2001). Brosch and collaborators have partially purified and characterized two HDAC complexes from C. carbonum. One of them is 60 KD in size with HDAC activity resistant to high concentrations of HC-toxin and TSA. This HDAC activity unique to C. carbonum appears to be dependent on HC-toxin production. More recent work has shown evidence for the “extrinsic” hypothesis. Isolates of C. carbonum that do not produce HC-toxin are 19 sensitive to the toxin when applied exogenously, whereas all toxin-producing isolates are resistant. HDAC extracts from resistant strains were found to protect sensitive extracts when mixed together (Baidyaroy et al., 2002). This protection was specific to C. carbonum and could not be achieved with other Cochliobolus species or Neurospora crassa. In conclusion, it appears that C. carbonum has multiple mechanisms of self- protection against HC-toxin. The objective of this study is to better understand the function of HDACs in filamentous fungi in general and in C. carbonum in particular. It is also important to better understand the differences and similarities between the C. carbonum HDACs and their yeast counterparts. The work presented here undertook both a biochemical and genetic approach to studying specifically two HDACs from C. carbonum, HDCI and HDC2. These two genes are orthologs of the yeast HDACs H052 and RPD3, respectively. Targeted disruption of HDC1 was found to cause a significant reduction in C. carbonum virulence against maize, altered conidia morphology, and poor growth on complex polysaccharides when compared to the parental wild type. When placed on maize leaves, the hdc] strain was unable to penetrate the cell wall. Indeed, expression of various CWDE transcripts was highly reduced. HDC2, on the other hand, could not be disrupted despite many attempts, most likely because this gene is essential for the survival of C. carbonum. Overexpression of Hdc2, instead, proved that it has intrinsic HDAC activity, just as Hdcl. However, characterization of this strain showed no significant difference from its parental wild type. 20 CHAPTER ONE CHARACTERIZATION OF THE HISTONE DEACETYLASE HDC1 FROM COCHLIOBOLUS CARBONUM Introduction The gene HDC1 from C. carbonum most closely resembles the sequence of H052 from yeast. As it is the case with other histone deacetylases (HDAC), H052 plays a role in repression of transcription (Watson et al., 2000). H052 is part of large complexes that can remodel chromatin at promoter regions leading to repression of gene expression. For instance, H032 was recently found to be a member of the Set3 complex, named after Set3, a histone methyltransferase (Pijnappel et al., 2001). The Set3 complex also includes Hstl, a member of the Sir2 class of NAD-dependent deacetylases (Smith et al., 2000). When H032 was removed from the Set3 complex, repression of meiosis-dependent genes was abolished. Both the hos2 and set3 strains underwent normal premeiotic DNA synthesis but showed a faster progression through meiosis. Hence, H032 is essential for the normal function of the Set3 repressor complex. Genome-wide expression studies have proven important to unveil the molecular role of HDACs in yeast. Using acetylation microarrays, Grunstein and coworkers showed that HOS deacetylase genes (H051, H052, and H053) are required for the deacetylation of histone H4 K12 preferentially at a very limited number of intergenic sites mainly on chromosome XII-R (Robyr et al., 2002). Further, using the hosZ strain, they showed that H052 is required for the preferential deacetylation of ribosomal protein genes. 21 To understand the biological role of HDACs in filamentous fungi, the structure and function of HDAC genes in C. carbonum were investigated. Using C. carbonum as a model organism makes this a unique study since C. carbonum produces a potent HDAC inhibitor. The report presented here focuses on the mutational analysis of HDC1 , a gene from C. carbonum whose ortholog is the yeast H052. B£§u_|t§ HDC1 was the second gene encoding an HDAC to be identified in C. carbonum. HDC1 was originally isolated using polymerase chain reaction (PCR) primers based on amino acid sequences that are conserved in known HDAC genes from other organisms (Hassig et al., 1998). These PCR primers were used to amplify a putative HDC1 fragment using C. carbonum genomic DNA as template (Baidyaroy et al., 2001). The PCR product was radiolabeled and used as a probe to screen a genomic library of C. carbonum (Scott- Craig et al., 1990). The gene contains no introns based on comparison of its genomic and cDNA sequences. The closest matches of Hdcl to proteins in the public databases were the products of HOSA, a gene encoding for a HDAC in Aspergillus nidulans (Graessle et al., 2000), followed by H052 of yeast (Rundlett et al., 1996). Hdcl has an overall amino acid identity of 46% to HosA from A. nidulans, 44% to H032, and 38% to de3 from yeast (Figure 2). The predicted molecular masses and p1 values of Hdcl , HosA, and H032 are similar: 56.9 KD and 5.7, 53.4 KD and 5.9, and 51.4 KD and 5.1, respectively. As expected, Hdcl contains all seven of the motifs characteristic of class I HDACs, but lacks the characteristic features of class II HDACs. 22 H082 HOSA HDC1 H082 HOSA HDC1 HOSZ HOSA HDC1 H082 HOSA HDC1 HOSZ HOSA HDC1 HOS2 HOSA HDC1 H082 HOSA HDC1 H052 HOSA HDC1 H082 HOSA HDC1 Figure 2. Amino acid sequence alignment of three HDAC proteins using CLUSTAL W (Thompson et al., 1994). H032 from S. cerevisiae, HosA from A. nidulans, and Hdcl from C. carbonum. Amino acids that are identical in all three proteins are indicated by asterisks. Sixteen of the seventeen amino acids that are highly conserved in all de3-like ----------------------------------- MSGTFSYDVKTKENEPLFEFNSAYS ------------------------ MPRSAIVQEYNPTPTLSVDRKSAQHQHEGIIARPSG MRTPGYIPSNGPANNIVQEYLSPDDAYPASFQDFKAMTDAEKIEVIAREAEEHGLERPKG PRVSYHFNSKVSHYHYGVKHPMKPFRLMLTDHLVSSYGLHKIMDLYETRSATRDELLQFH YRVSWHANPAVELHHFGQSHPMKPWRLTLTKQLVLAYGMHHAMDLYHCRAATVEELSDFH WNVSFHYNPHVEYHHFGSSHPMKPWRLTLTKQLVLAYGLEYTMDLFEPRPANFNELAIFH *9 t O t i * *iiti it it fit it it. t i it it SEDYVNFLSKVSPENANKLPRGTLEN----FNIGDDCPIFQNLYDYTTLYTGASLDATRK TSDYLDFLQTVVPGDMNDAQSKDFSENIVRFNFGDDCPIFDGLFQYCSLYAGASLDAARK DREYLSYLSKITPQNAQPDDPQYITYG--FGGBSNDCPVFDGLWNYVSLYTGATCSATWN * i i tit i i 1' it it i LINNQSDIAINWSGGLHHAKKNSPSGFCYVNDIVLSILNLLRYHPRILYIDIDLHHGDGV LCNNOADIAN-WSGGLHHAKKAEASGFSYVNDIVLAILQLLRIHPRVMYIDIDVHHGDGV LLNSRSDIAINWSGGLHHAKKNLASQFQYVNDIVIAIQLLLSHHQRVLYIQIQVHHGDQV t ‘0 iii fittttiiiti tifiittitii * ti i O fitfiti itifiti oaAgvrrpayFrLgrngNc ----- BFFEGTGDLTEIG----CDKGKHFALNVPLEDGID EQAEWSTDRVLTVSFHKYDKE----NFFEGTGPLDSTGPTHPLNPGAHHAVNVPLHDGID EQAEESTDRVFTLSYHKYGIDKHGYPFFEGTGNIDETGPTDARNPGKAHSLNIPIDDGID 1* i*** t 9 iii *ifiii'. * it i 1' fit! DDSYINLFKSIVDPLIMTFKPTLIVQQCGADSLGHDRLGCFNLNIKAHGECVKFVKSFGL DESYVQLFKDVVGACVSKFRPAAIVLQCGADSLGCDRLGCFNLNVAAHGACVAYTKTEGL DEQYKWLFKTVTSAVIEKYNPTAIVLQSGADSLGGDRLGRFNLNIKAHGFCVETVKAYGR t i it. i ii * tittii flit. fit! iii if i * PMLVVGGGGYTPRNVSRLWTYETGILNDVLLPED----IPEDIPFRDSFGPDYSLYPMLD PMLVVGGGGYTPRNVSRAWAHETSILIDAQDKINP--VIPSNVAFRNHFGPDFSLFPPLS PLLIIGGGGYTPRNVARTWCHETSVCVGAQLHNELPAHVPYLQAFQGAENGDGVLYPDLH t t titifitifiii i t it i t i t t * DL--YENKNSKKLLEDIRIRCLENIR-YLQGAPSVRMDAECIPTQDISALTE----EEDK EMRKLENKNSRAYLATIVQTITEQLRRYLQAAPSVQMSVIPPDLLGLREETEKEIEEEIA NIKRHENLNSQAKLHKLIEQALENLR-YLEGAPSVTVDTRGISLEEIMKVREMID---QE it ti 0 i i ** **** t IIQEMNEETEADSSNRLEEMEKENSGLIAFS- 452 KLEEKREEAEGGKNSRRRDAEKGAGLRGELYS 481 LEDEAEDRARLTVENSRRKKERNVGGRNERR- 505 i * HDACs and related bacterial proteins are underlined. 23 25 36 60 85 96 120 141 156 178 201 215 238 252 271 298 312 331 358 368 389 418 421 449 474 Strains of C. carbonum that were specifically mutated in HDC1 were constructed by targeted gene replacement. Two fragments of HDC1 flanking the cassette encoding resistance to the drug hygromycin were cloned into the shuttle vector pSP72 and the resultant plasmid (see Materials and Methods) was used to transform C. carbonum wild type strains 367-2 and 164R1 (Figure 3A). Whereas 367-2 is a toxin-producing strain (Tox+), 164R1 is a naturally occurring toxin-non-producing strain (Tox-). 164R1 lacks the genes required for HC-toxin biosynthesis (Panaccione et al., 1992). The rationale behind the disruption of HDC1 in a Tox- background is that Hdcl may be the only toxin- resistant HDAC that binds most of the toxin, and its deletion may cause self-inhibition as the toxin no longer binds Hdcl but will inhibit the other C. carbonum HDACs. If this hypothesis is correct, deleting HDC1 in 164R] should not cause any phenotype, whereas there will be a phenotype in 376-2. Otherwise, the hypothesis is incorrect and any phenotype observed in 367-2 can be attributable specifically to the mutation of HDC1 and not to a secondary effect caused by self-inhibition of other HDACs by HC-toxin. The HDC1 mutational analysis on 367-2 has been described elsewhere (Baidyaroy et al., 2001). The results presented here focus primarily on 164R]. Putative mutants were screened for their ability to grow in the presence of the drug hygromycin. Five independent transformants were obtained and purified by two rounds of single-spore isolation to nuclear homogeneity. The aim of this process is to ensure that the genotype of every mutant spore is identical. These putative mutants were confirmed by Southern hybridization (Figure 3B and Figure 3C). All five isolates showed the expected pattern of DNA hybridization for single gene replacement. 24 A Wild Type 1kb l———l F 1.9 kb : E B H P C H x E Gene Replacement Mutant |————— 33 kb I E B H H HX E .l—J—i—| HPH {—l—I—L WT T1 T2 T3 T4 T5 Q B MFW ..L~"-- afir'W-{ia g;- '- .‘ ' 6.4 ‘5 -4.8 - 3.7 WT T1 T2 T3 T4 T5 Q C - i ~- ~ ' - 3.7 Figure 3. Construction and analysis of the hdc] disruption transformants. A. Restriction map of the genomic region of wild-type HDC1 and predicted map of the gene replacement mutant. The open box in the wild-type map indicates the fragment of HDC1 DNA replaced by the HPH gene encoding hygromycin phosphotransferase in the mutant. The arrow indicates the location of the HDC1 coding region. E, EcoRI; B, BamI-H; H, HindIII; P, PstI; C, Clal; X, Xhol. B. DNA gel blot probed with a fi'agment of HPH. WT, Wild Type; T1, T717-1; T2, T717-2; T3, T717-3; T4, T717-4; T5, T717-5. C. The same blot was stripped and then probed with the deleted segment of HDC1. Sizes of DNA markers in kilo-bases are shown at right. DNA was digested with EcoRI. 25 The phenotypes observed for the mutants in the Tox+ background (T702) were identical to those observed in the Tox- background (T717). Mutation of HDC1 resulted in a striking developmental defect. Filamentous ascomycetes such as C. carbonum multiply by dispersing non-motile asexual spores called conidia. The conidia in the hdc] mutant were reduced in size and septum number (Figure 4). However, the germination rates in vitro were the same as in the wild type. Furthermore, the degree of curvature present in the wild type conidia is reduced in the conidia of hdc] . HDAC activity in crude extracts of hdc] is approximately half of that in the wild type (Figure 5). This result remained consistent after the assays were repeated several times. This finding strongly suggests that HDC1 encodes a functional HDAC. However, because HDACs associate with each other in mammals and yeast, an indirect effect of disrupting HDC1 on other HDAC activities can not be excluded (Grozinger et al., 1999). As further assurance that the HDAC activity phenotype was attributable specifically to the mutation of HDC1, HDAC activity was measured in the ccsnfl mutant of C. carbonum. As it turns out, the ccsnfl strain shares many of the same phenotypes found in hdc] as described below (Tonukari et al., 2000). HDAC activity in ccsnf] (strainT688) was not significantly different from that in the wild type (Figure 5). This result indicates that the reduced HDAC activity in hdc] was not a side effect of any of the phenotypic abnormalities shared by the two mutants. To test the ability of hdc] to grow on complex carbohydrates, growth rates were measured against the wild type as a control. The C. carbonum strains were grown on agar plates containing one of the following carbon sources: sucrose, glucose, arabinose, xylan, 26 Figure 4. Microscopical analysis of the hdc] conidia. A. Ungemrinated wild type conidia. B. Ungerminated conidia of T7l7-1. C. Wild type conidia 6 hours afler germination on glass slides. D. Conidia of T717-l 6 hours after germination on glass slides. Bar in A. and B. is 12.5 pm; bar in C. and D. is 50 um. 27 6000 . H l--I 5000 i .5 O O O HDAC Activity (dpm) o: O O O 2000 a 1000 r 0 “‘— n r. r T l 1 WT T71 7-1 T71 7-2 T71 7-3 1717-4 T717-5 ccsnf1 Figure 5. Analysis of HDAC activity in the hdc] mutant. HDAC activity measured in crude extracts of 367-2 (WT), five hdc] mutants (T7l7-1, T717-2, T717-3, T717-4, T717-5), and the ccsnf1 mutant. HDAC activity is measured by the amount of 3H-labelled acetate released (in dpms) by the enzymatic activity of HDACs when incubated with 3H- labelled chicken histones. 28 pectin, or corn cell walls. Growth of the hdc] mutant on sucrose and glucose was very similar to that of the wild type. However, growth of the hdc] mutant on corn cell walls, pectin, and arabinose was reduced by at least 50% when compared with the growth of the wild type (Figure 6). Growth of the hdc] mutant on xylan was reduced by about 30% when compared to the wild type. The degree of growth reduction on the various substrates tested was similar for the ccsnf1 and hdc] mutants. For instance, the growth of both mutants was severely affected on arabinose, pectin, and maize cell walls (Figure 6) (Tonukari et al., 2000), while on glucose and sucrose there were no significant differences. One possible explanation for the reduced growth of hdc] on complex carbohydrates and not on glucose is that HDC1 is required for the expression of cell-wall- degrading genes, of which C. carbonum contains an abundant variety. To test this possibility, RNA was extracted from the hdc] mutant and wild type after growing on various liquid media for 7 days. The RNA samples were blotted and probed for several different transcripts. The steady-state levels of mRNA of EXG], encoding exo-B-1,3- glucanase (V anhoof et al., 1991), PGN], encoding endo-polygalacturonase (Scott-Craig et al., 1998), and XYLI, XYLZ, and XYL3, encoding endo-xylanases (Apel-Birkhold and Walton, 1996), were either decreased or entirely down-regulated in the hdc] mutant (Figure 7). Therefore, HDC1, like SNF] , is required for the expression of at least some glucose-repressed genes in C. carbonum. The virulence in T717 was not tested because the parental line does not produce HC-toxin. There is no significant virulence in natural isolates of C. carbonum that do not 29 i awnd Type so 1 Ith1 8 8 8 Growth (mm) ll|||||||||l|||||||||||||||||||l||||||lllllllllllllll|l||||||||||l|l||||||||||||||| ll|||||||||||||||||||||l||lllllllll||ll||||Ill||||l||||||||l|||||||||||||l|| W //////////////////////. W N Glucose 1. > “g, lll|||||||||||||||||||||||||||||||||||l|||||||||||||l|l|||||llllllllllll O s Pectin Walls Figure 6. Growth comparisons between the hdc] mutant and wild type. Various carbon sources were used: simple sugars such as glucose and sucrose, or complex carbohydrates such as arabinose, xylan, corn cell walls, and pectin. Growth was measured in millimeters as a radius, by day 7, from the spot of inoculation at one extreme of the petri plate. 30 WTltdcl Q PGN] "' XYL 1 X112 XYL3 ccSNFI GPD Figure 7. RNA expression profiles of selected genes in the hdc] mutant versus wild type. Total RNA was extracted from wild type and T717 strains. The ftmgus was grown on maize cell walls as the sole carbon source for EXG] expression, on pectin for PGNI, on xylan for XYL], XYL2, and XYL3, and on sucrose for ccSNF] and GPD. EXG] encodes exo-B-l,3-glucanase; PGN] encodes endo-or-l,4-polygalacturonase; XYL], XYLZ, and XYL3 encode endo-B-1,4-xylanases; ccSNF] encodes a protein kinase; GPD encodes glyceraldehyde 3-phosphate dehydrogenase and is used here as a loading control. 31 synthesize HC-toxin. Virulence of T702 was tested as its parental line makes HC-toxin and is fully virulent on maize. Virulence of T702 was greatly reduced as indicated by a reduction in the number of lesions formed on maize leaves (Baidyaroy etal., 2001). Lesions that developed had similar morphology and rates of expansion, as did those lesions caused by the wild type. Even at high inoculation densities (105 conidia/mL) and extended periods of disease development (more than 14 days), T702 never killed plants, unlike the wild type, which eventually colonized and killed seedlings (Baidyaroy et al., 2001). Conidia of T702 could be seen microscopically to adhere efficiently to maize leaves, indicating that there is no defect on appressorium formation. However, Baidyaroy and coworkers found that all lesions formed by T702 were associated with clumps of conidia. These results suggest that T702 is deficient in the production of a virulence factor or factors and only many T702 conidia in close proximity can compensate for such deficiency. Therefore, the product of HDC1 may be considered a virulence factor necessary for successful penetration of the maize epidermis. Is the reduced virulence of T702 caused by a reduced efficiency of germination or appressorium formation on leaves? Baidyaroy and coworkers answered this question by examining inoculated leaves using scanning electron microscopy. They found no differences between the germination rates of T702 and wild type conidia. The T702 hyphae were able to develop appressoria of wild type morphology preferentially at the junctions between leaf epidermal cells, as has been reported for wild-type C. carbonum and other species of Cochliobolus (Jennings and Ullstrup, 1957; Murray and Maxwell, 1975). This finding indicates that the defect in virulence of T702 is at a stage after germination and appressorium formation. Confocal microscopy with reconstructed cross- 32 sectional views also indicated that although hdc] conidia germinated and grew along the surface of the maize leaf, they did not penetrate efficiently (Baidyaroy et al., 2001). HC-toxin is an essential virulence determinant for C. carbonum (Walton, 1996). In vitro, the hdc] mutant T702 is able to synthesize HC-toxin at concentrations similar to those in the wild type (Baidyaroy et al., 2001). Therefore, the reduction of virulence by T702 can not be attributed to a decrease in HC-toxin production, and HDC1 does not regulate HC-toxin biosynthesis. Discussion The HDC1 gene product is related to many characterized HDAC proteins and is most closely related to the yeast H032. Reduction in HDAC activity by as much as 50% in total extracts of the hdc] strain of C. carbonum when compared to the wild type indicates that Hdcl is likely to be a histone deacetylase. Further biochemical studies are needed for full confirmation of this hypothesis. Viability of hdc] mutants is surprising considering that HDAC function is expected to have critical effects on global gene expression. However, Grunstein and colleagues have recently used chromatin immunoprecipitation and intergenic microarrays to demonstrate that H032 preferentially affects ribosomal protein genes (Robyr et al., 2002). Hence, it is possible that HDC1 is not essential for the survival of C. carbonum, unlike other HDACs (e.g. RPD3) that may affect a much wider variety of genes (Kurdistani et al., 2002). Further, at least in the case of yeast, it was found that resulting phenotypes due to H052 gene disruption are dependent on strain background (Bilsland et al., 1998). As in the case of C. carbonum, one yeast hos2 strain in particular displayed a slow-grth phenotype. 33 Strains of the plant pathogenic firngus C. carbonum mutated in the HDC] gene encoding a putative HDAC are viable but have several significant phenotypes. The hdc] mutants originated from a Tox+ background (T702) are severely reduced in virulence. Evidence suggests that the reduced virulence is attributable not to reduced conidia] germination or appressorium formation in vitro or in vivo, but to a decreased efficiency in penetration of the maize leaf epidermis (Baidyaroy et al., 2001). Since the lesions caused by T702 have normal morphology, HDC1 does not appear to be important for ramification within the maize leaf. C. carbonum probably breaches the maize epidermis by enzymatic action and not mechanical force (Horwitz et al., 1999) and, as shown here, HDC1 is required for expression of at least some of the genes encoding glucose-repressed extracellular enzymes that can break down the cell wall. Hence, the inability of T702 to penetrate leaves might be due to its lack of synthesis of some of these enzymes. The fact that clustering of T702 conidia causes small lesions may mean that extracellular depolymerases are virulence factors in C. carbonum. However, it is possible that HDC1 controls other kinds of virulence factors. The decreased growth of the hdc] mutant on complex polysaccharides can be accounted for by decreased production of the polysaccharide depolymerases and/or enzymes needed for uptake or metabolism of alternative sugars, which in turn can be attributed to decreased expression of the encoding genes. No strict correspondence was found between mRN A levels and grth on the corresponding substrates in the hdc] mutant (Figures 6 and 7). That is, no sound correlation was observed between the inability of the hdc] mutant to grow on a certain carbohydrate and expression in the hdc] 34 mutant of the gene presumed to be involved in degradation of such carbohydrate. It is possible that the utilization of a complex carbohydrate requires the uptake and metabolism of the released sugars. For instance, complete down-regulation of the major xylose uptake carrier would have a severe effect on the growth on xylan regardless of the expression levels of xylanase. All known depolymerases are redundant in C. carbonum. Hence, other xylanase genes in addition to XYL], XYL2, and XYL3 might permit some amount of growth on xylan despite down—regulation of XYL I , XYLZ, and XYL3 in the hdc] mutant (Figure 7). Interestingly, there is overlap between the phenotypes of the C. carbonum hdc] and snf] mutants. The similarities found suggest that the two gene products, Hdcl and Snfl , may be part of the same signaling cascade in C. carbonum. The yeast Tupl has been shown to recruit Hdal to promoters of Tupl-regulated genes (Wu et al., 2001). Tupl itself is recruited to promoters via Migl , a zinc-finger protein, in association with Ssn6 (Treitel and Carlson, 1995; Smith and Johnson, 2000). Filamentous fungi have orthologs of TUP] and M10], although MIG] goes by the name of CREA (Ebbole, 1998). Hence, when yeast is grown on glucose as carbon source, genes required for grth on alternative sugar sources are repressed by the Migl/Tupl-Ssn6/Hdal complex. Migl itself is regulated by Snfl, a protein kinase (Treitel et al., 1998). In the absence of glucose, Snfl phosphorylates Migl , thereby causing it to dissociate from the promoters of glucose-repressed genes (Carlson, 1999). Recently, it was shown that Snfl can also activate transcription, in response to glucose limitation, by directly interacting with the RNA polymerase 11 holoenzyme in yeast (Kuchin et al., 2000). 35 The yeast regulatory circuit logically predicts that the disruption of H052, an HDAC gene whose product interacts with Tupl , should cause de-repression of glucose- repressed genes. Indeed, it was observed that mutation of H052 (in an rpd3/hos1 background) causes de-repression of SUC2, which encodes a sucrose invertase, even under repressing (high-glucose) conditions (Watson et al., 2000). In contrast, the yeast logic is not consistent with the results obtained from mutational analysis of HDC1 . If HDC1 encodes a co-repressor, its disruption would be predicted to result in the de- repression of glucose-repressed genes. In fact, the exact opposite is observed. This finding suggests that the biological role of HDC1 is markedly different from that of H052. Materials and Methods Fungal Cultures, Media, and Growth Conditions: The wild type HC-toxin producing isolate of C. carbonum, 367-2A, was derived from isolate 8B1 11 (ATCC 90305) and maintained on V8 juice-agar plates. The wild-type Tox- isolate was 164R1, a progeny of SBl 11 (Walton, 1987). The fungus was grown in liquid media or agar plates containing mineral salts, 0.2% yeast extract, and trace elements (van Hoof et al., 1991). Carbon sources were 2% (w/v) glucose, sucrose, oat spelt xylan (fluka, Buchs, Switzerland), citrus pectin (catalog no. P-9l35; sigma), or maize cell walls (Sposato et al., 1995). For quantifying growth on agar plates, 5 1.1L of a conidial suspension (104 conidia per mL) in 0.1% Tween 20 was spotted on one extreme end of the plate. Plates were incubated under fluorescent lights at 21°C. Growth was measured in millimeters every day as a linear progression from the point that was initially spotted. 36 flmption of HDC1 and Nucleic Acid Manipulatiogsz T 0 construct the replacement vector pAJ63, pSP72 (Promega) was cut with SphI, blunted, and cut again with EcoRV to eliminate the multiple cloning sites between SphI and EcoRV. The resulting plasmid was cut with HindIII and ligated with a 1.3-kb HindIII fragment of HDC1.This plasmid was then cut with PstI and ClaI, and the deleted fragment was replaced with a PstI-Clal fragment from pHYG4, which contains the HPH gene for hygromycin resistance from pCB 1004 (Carroll et al., 1994) sub-cloned into the Smal site of pBluescript (KS) +. pAJ63 was then linearized at the unique PstI site prior to transformation of C. carbonum wild type strains 367-2A and 164R1. Transformation was performed exactly as described previously (Apel et al., 1993). Transformants were then purified to nuclear homogeneity by two rounds of single-spore isolation. The C. carbonum genomic and cDNA libraries have been described previously (Scott-Craig et al., 1990). DNA and total RNA were extracted from lyophilized mats after 7 days of growth in still liquid culture (Apel et al., 1993). The methods used for DNA and RNA electrophoresis, gel blotting, probe labeling, and hybridization have been described elsewhere (Apel-Birkhold and Walton, 1996). HDAC Assay: HDAC activity was measured using 3H-acetate-labeled chicken reticulocyte histones (Kolle et al., 1998). Freeze-dried tissue (0.5g) from mycelial mats grown in still culture for 7 days was ground in liquid nitrogen and re-suspended by vortexing in 4.0 mL of extraction buffer (15 mM Tris-HCl, pH 7.3, 10 mM NaCl, 0.25 mM EDTA, 10% [v/v] glycerol, and 1 mM B-mercaptoethanol) containing one protease inhibitor tablet (Roche, Mannheim, Germany) per 30 mL of buffer. Afier centrifugation 37 at 11,000g for 15 min, 3 mL of the supernatant was de—salted by gel filtration (Econo-Pak 10 DO; Bio-Rad, Richmond, CA). Fifty micro-liters of protein extract and 5 uL of tritiated histones (40,000 dpm) were incubated for 2 hr at 23°C, 35 uL of 1N HCl was added, and the released acetate was extracted twice with ethyl acetate, first with 0.8 mL (removing 0.6 mL) and then with 0.6 mL (removing 0.7 mL). The ethyl acetate fractions were combined and counted by scintillation spectroscopy. 38 CHAPTER 2 STUDIES ON THE HISTONE DEACETYLASE HDC2 FROM C 0CHLI OBOL US CARBON UM Introduction A fundamental aspect of eukaryotic gene regulation is the ability of DNA-binding activators and repressors to recruit chromatin-modifying activities to specific promoters. Once recruited, such modifying activities generate local domains of altered chromatin structure that influence the level of gene activity. The yeast repressor Ume6, for instance, specifically binds DNA sequences (URSl) in a variety of promoters and inhibits transcription by recruiting the de3 HDAC complex (Kadosh and Struhl, 1997). Recruitment occurs through an interaction between the Ume6 repression domain and Sin3, a component of the de3 complex (Washbum and Esposito, 2001). Targeted recruitment of de3 leads to localized deacetylation of the N-terminal tails of histones H3 and H4 over a range of one to two nucleosomes (Kadosh and Struhl, 1998; Rundlett et al., 1998). Genome-wide binding maps of de3 and its associated factor Umel have shown that this HDAC complex is common to a large and diverse set of promoters (Kurdistani et al., 2002). While de3 affects the acetylation of promoters of genes that are part of many cellular pathways, there is a modest over-representation of genes involved in sporulation, germination, and meiosis (Robyr et al., 2002). There is also a significant preference for genes throughout the genome that are involved in carbohydrate utilization. These include genes involved in carbohydrate transport and metabolism, as well as energy reserves. 39 Although many intergenic regions that are under the regulation of de3 contain URSl (the Ume6 binding site), just as many have been found not to contain the Ume6 recognition site (Robyr et al., 2002). Hence, many of the promoters affected by de3 must use other mechanisms to recruit the deacetylase. Microarray studies in yeast have shown a significant overlap in the genes that are regulated by de3 and Sin3. At least 107 genes in common are up-regulated, and 198 genes in common are down-regulated (at least two-fold) (Bernstein et al., 2000). Loss of RPD3 and SIN3 results in the two-fold down-regulation of 264 and 269 transcripts, respectively. Two conclusions can be drawn from these results. First, de3 and Sin3 functions are linked and loss of one protein results in complete loss of the linked function. Second, de3 may also activate transcription in addition to being a co- repressor. Treatment of yeast cells with the HDAC inhibitor TSA results in down- regulation of many of the same genes as those found in the rpd3 strain, thus indicating that HDACs may function as direct transcriptional activators (Bernstein et al., 2000). Transcription profiles have also demonstrated that 40% of endogenous genes located within 20 kb of yeast telomeres are down-regulated by RPD3 deletion. One possible model is that de3 might activate telomeric genes repressed by SIR proteins directly by deacetylating H4 K12 (Hecht et al., 1995; Strahl-Bolsinger et al., 1997; Moazed, 2001b). H4 K12 is acetylated at silenced loci in yeast, and this appears to facilitate interaction with Sir3, which leads to silencing (Braunstein et al., 1996). Since de3 is known to deacetylate at H4 K12, this action in certain cases may activate transcription simply by preventing binding of the repressive SIR complex. 40 Recently, it was shown that the domain of localized histone deacetylation generated by recruitment of de3 mediates repression by inhibiting recruitment of chromatin-modifying activities and the TATA binding protein (TBP) (Deckert and Struhl, 2002). Further, Struhl and colleagues showed that repression by de3 depends on the activator and the level of activation, not the extent of histone deacetylation. They were able to abolish repression by direct recruitment of TBP, but not Pol II, to the H153 promoter. In C. carbonum, the gene HDC2 is the ortholog of the yeast RPD3. The first gene encoding an HDAC found in C. carbonum, HDC2 has been studied using genetic as well as biochemical strategies. The results of this work are presented here. 32% The gene HDC2 from C. carbonum was isolated using degenerate PCR primers that were designed from the DNA sequence of the yeast RPD3 gene. The product generated was radioactively labelled and used to screen a genomic library of C. carbonum in a similar fashion as it was done for HDC1 described in chapter one (S. Wegener and J .D. Walton, unpublished results). Disruption of HDC2 was attempted several times by several scientists and all attempts were unsuccessful. These results indicate that HDC2 may be required for the survival of C. carbonum or, alternatively, the chromatin surrounding the HDC2 locus is refractory to integration of foreign DNA. In order to find out whether HC-toxin can inhibit purified Hdc2, in the absence of other C. carbonum HDACs, the gene HDC2 was cloned and expressed in Escherichia coli. The HDC2 open reading frame was cloned into plasmid pQE3O and the resulting 41 construct was transformed into E. coli M15 cells. High levels of stably expressed Hch protein were obtained by inducing actively growing cells with IPTG (see Materials and Methods). Expression of Hch was determined by denaturing polyacrylamide gel electrophoresis on a time course, using the pQE30 vector without insert as a control (Figure 8). High levels of accumulation of Hdc2, a protein of 75 kD, were found two hours after induction with IPTG. To maintain HDAC activity in recombinant Hdc2, zinc was added to the growth media. Zinc is known to be a co-factor required for optimal HDAC activity (Hassig et al., 1998). Most of Hch was found in the soluble fraction (data not shown). This fraction was assayed for HDAC activity but none could be found. It is possible that the recombinant Hch protein is quickly degraded since it has been found to be sensitive to proteases (Brosch et al., 2001). Alternatively, Hch may have specific pH, salt, or temperature requirements for optimal function that are not yet known. A different strategy to study the function of HDC2 is to over-express the gene in vivo. The objective was to insert a second copy of HDC2 and then remove the native HDC2 from this new C. carbonum strain. The second copy of HDC2 would be under the control of a promoter that can be easily induced and regulated. Since it is likely to be an essential gene, the only effective way to delete the native copy of HDC2 requires the ability to control efficiently the levels of expression of the chimeric copy of HDC2. The levels of expression of HDC2 may be modulated by the kind and concentration of carbon source added to the grth medium. Only by this strategy will it be feasible to unveil any phenotypes that may be a consequence of a deficiency in HDC2 expression, while maintaining C. carbonum viable. 42 0030005 Home E N 3 N00: .h ”00.50005 baa E N an 30:00 .0 ”003265 coax E _ 8 N00: .m 60:26.: “can E _ we 30000 .v ”00:03:: coca .3 m6 3 N00: .m ”H m6 3 9.8me 500:3 .880.» 023293 30:00 .N ”0000205 00 N00: — Eco—.82 .2 .m0» 88.508 5805 £00 .m 5 0.600090 N003 m0 0:5 08:00 08:. .w 9...»:— -93 m. n -98 y”..- ,, Quasi. m V: a... . a... 3.3!. 4 -033: b e m v m N. tau 2 43 For this purpose, the HDC2 open reading frame was firsed to the B-xylosidase (XYPI) promoter and the resulting chimera was cloned into the pKP5 vector. pKP5 is a vector for C. carbonum transformation, which contains the hygromycin resistance cassette and the PGNI target locus. This target locus allows homologous recombination to occur, by a single cross-over event, at the genomic PGN] site, thereby integrating the foreign DNA into the PGN] locus at high efficiency. Disruption of the PGN] locus does not cause any apparent phenotype in C. carbonum (Scott-Craig et al., 1990), so it is deemed safe to use this locus as an integration site. The promoter of XYP] was chosen because, as any other glucose-repressed gene, XYP] is easily induced by growing the fungus on either corn cell walls or pure xylose as a carbon source (Wegener et al., 1999). From this transformation event, four independent transformants were obtained and two of these four (T709-2 and T709-3) had the correct DNA pattern expected for true transformants (Figure 9). These C. carbonum mutants (strains T709-2 and T709-3) contain two copies of HDC2, one under the control of the native HDC2 promoter and the other (chimera) under the control of the XYPI promoter. Reverse transcriptase PCR (RT-PCR) reactions were performed to determine the expression pattern of HDC2 in T709 (strain T709-3), using the wild type parent as a control. Sets of primers were used to amplify specifically three different cDNA strands: the native HDC2, the native XYPI , or the XYPI-HDC2 chimera. The RT-PCR results demonstrate the non-quantitative nature of the RT-PCR technique (Figure 10). The control lane, with no reverse transcriptase added, shows that there was no DNA contamination. As expected, there was no amplification product present in the chimera lane of the wild type. A chimera product is present in T709, both when grown on sucrose 44 40000050 00w:0_ 0: 0:5 .0E000w .20: 0:0 00:00:00 06 800.0 .20: .500 0080000 93% 005 0002 0:02 9G... 05 00 0003005 00% 002 00—02 x 0003 0:0 00.50000 00:00:03 05 .00 002 .< .800 000.0000 29.0 a .3 .00005080000 Homo—080: 00> 000000 0000000 00:. .0002 NZOR omfioaow 05 00 NOQQ .00 00:30:: .a 0.5mm..— 5 a. a h .3 in M. E3.“ Ham N62 Ham Ham ESE Bum H8: 53m _ _ _$ Earn _- 890 TE 30 me. .200 ; F a. 3. « F 03.0 he. “ _ 88. n _ E 35300 X 88. 3.0. 30 me. _ =03 _ 3.523“. < 5 4 0.83. .3. ”08:. .2. N030. .2. 2-8:. 4 H .200m 505 0000090 003 80 0_ 008 NOQE 000850 08 0:0 NOQI 0300: 08 ”08009.0 000 0085 03h doem. 50$ 0800mm. 003 » 0:00 0: .025. 000550 NDQméoSN 0.0 00 000000 05 3:0 >595 0300 0000 058:0 030 05 5000003080 000820 How .00000 003 0&5 PM 0: £00008 30:00 00m .0505 093 0:3 0:0 002. 05 E NUQE .00 203000 02000098 MUAQM .3 PEER -32: 7 4 L508 3 3 33 30 mo Ho mm Hmm Hmm mm.“ xao [ONT umww. Jami { { { ~6ch Néchh MEWH. Ana—>9 Hmogx e\e~ HmOy—Uim gm HmOMUDm .xfi and xylose. This result was unexpected because XYPI is not supposed to be actively transcribed when C. carbonum is grown on sucrose. The drawback with RT-PCR is that even only one XYPI-HDCZ RNA molecule present can be reverse transcribed, and the product after the amplification steps will appear as a prominent band just as if the XYPI- HDCZ message were abundant. It seems that the native HDC2 is expressed under sucrose and xylose with no significant differences, both in T709 and in the wild type. A similar pattern of expression is observed for the native XYPI . A PCR product is present when T709 is grown on both sucrose and xylose. In conclusion, the XYPI-HDCZ chimera is expressed at normal levels in T709, but the XYPI promoter is not nearly as tightly regulated as previously thought. Northern blot analysis was performed also to examine the levels of expression of HDC2 in T709. Both T709 and wild type strains were grown for seven days on various carbon sources. Total RNA was extracted and blotted onto nitro-cellulose membranes. The carbon sources used include sucrose, xylose, glucose, and a combination of sucrose and xylose. Blots were probed with a piece of the HDC2 cDNA as well as the untranslated region (5’-UTR) of the XYPI promoter (Figure 11A). Because one of the project goals was to precisely modulate HDC2 expression, it was critical to determine how well regulated the XYPI promoter was under various carbon sources. The expression level of the XYPI-HDCZ chimera was highest when the fungus was grown in a mixture of sucrose and xylose (Figure 11B). However, the chimera was also expressed, at lower levels, when the fungus was grown in glucose and sucrose. Therefore, the XYPI promoter is not under tight regulation, confirming the RT-PCR results. Interestingly, the behavior of the XYPI promoter appeared to be different in the chimera gene from that of the native 48 l—p —_—l XYPlp I xvm }—___—— 1—9 2% Sucrose 2% Glucose 2% Xylose 2% Xylose 0.2% Sucrose I II II II L WTT2 T3 WT T2 T3WTT2 T3 WTT2 T3 4.4 kb - 2.37 kb - 1.35 kb - C 2% Sucrose 2% Glucose 2% Xylose L lWT T3 I I WT T3 I I WT T3 1.35 kb - Figure 11. RNA gel blot analysis of expression of HDC2 in T709 versus wild type strains. A. Map describing the sizes (in kb) of the expected transcripts. B. RNA blot probed with a segment of HDC2 (500 bp). As expected, the wild type has one transcript, the native HDC2, whereas the transformants have an additional, chimera transcript. C. RNA blot probed with the 5’-UTR of XYPI. Being a much smaller probe (70 bp), the signal was not as strong. As expected, the native XYPI was most highly expressed on xylose as the carbon source. L, ladder; WT, wild type; T2, T709-2; T3, T709-3. XYPI gene. The chimeric promoter could not be strongly induced under xylose, but the native one was (Figure 11C). There are two possible explanations for this phenomenon. First, the chimeric XYPI promoter may be missing some regulatory elements that are present in the native XYPI promoter. The XYPI promoter used in the chimera is 500 base pairs in length and there could be other regulatory elements further upstream in the native XYPI promoter. Second, the site of integration for the chimera might be different from that of the native XYPI locus, and thus the new chromatin environment may alter the pattern of expression of the XYPI promoter. Does the C. carbonum strain T709 have more Hdc2 enzyme activity? Total HDACs were extracted from T709 and the wild type (367-2), and their HDAC activity was assayed. As shown on Figure 12, afier 5 days of growth on xylose, there is no significant difference in total HDAC activity between T709 and the wild type. Afier 10 days of growth, total HDAC activity actually diminished by 50% in T709 when compared with the wild type (Figure 12). However, after only 3 days of growth, the total HDAC activity of T709 appears to be more resistant to HC-toxin than the wild type. The HDAC activity in T709 after 3 days is only inhibited by about 20% in the presence of HC-toxin, while the HDAC activity in the wild type is reduced by about 80% due to HC- toxin inhibition (Figure 12). Resistance to the toxin in the wild type develops at 5 days of growth and is maintained thereafter. In T709, a significant amount of resistance is already present at 3 days of grth and maintained thereafter as well. In conclusion, it seems that over-production of Hdc2 renders C. carbonum resistant to exogenous HC-toxin at an earlier growth stage. The results also indicate that the total levels HDAC activity may not be necessarily higher in T709, but indeed, the level of toxin-resistant HDAC activity is 50 [111130 y E5 Days 88 Days E10 Days \\\\‘\\\\\\\\\\\\ |IIIIII|||||||||lIIIIllI|||||||||||||||||||I|||I|IIIII|I|||Illlllllllm/IIIIII/ffiI/IIIIIIII/I/I/III IIIIImmmmnmmumunnnunumuIlllllllfifififififim nmnmlInmIIIIIIIIIIIIIIIIII/I/III/I/lmmnnlnlIIIIIIIIWIII/lll/II/ll/Ifi W //////// / /////§- I|||||||||||||II|II||||||||||||||||I|||I||II|lII|IIII||||||||I||lIIIII||||||||||||||||||lI|I|II|||||||||||I||I|||IIII|I||||||l|||IIg mmmmm MMMMMMMMM (wdp) Stunoo T709, Toxin T709, Water Wild Type, Toxin a function of growth. Fungi were ng xylose as the carbon source. Total HDAC activity was measured from crude HDAC activity comparing T709 versus wild type strains as HDAC extracts. Each assay was performed in replicates, two sets with water (control) and two with HC-toxin. of still culture usi Figure 12. The effect of HC-toxin on grown for 3, 5, 8, and 10 days higher. This phenomenon might be explained by a feedback regulatory mechanism that would hypothetically control the total amount of HDAC activity in C. carbonum. More production and activity of Hch in C. carbonum thus would lead to lower activity levels by other HDACs as a compensatory mechanism. However, it is not yet understood why total HDAC activity is diminished at 10 days of growth in T709. Possibly, as the fungus ages, it begins to degrade the HDACs that may no longer need to survive, targeting them for the proteasome via the ubiquitination pathway for protein degradation. For a more precise biochemical dissection of Hdc2, total HDAC extracts of the T709 and wild type (367-2) strains were fractionated by anion-exchange high- performance liquid chromatography (HPLC). Two major peaks of HDAC activity are obtained after HPLC fractionation. In the C. carbonum wild-type strain SB] 1 l, the first peak is almost completely resistant to HC-toxin, whereas the second, broader peak is almost completely sensitive (Figure 13A). This HPLC pattern is standard in fractionation of C. carbonum HDACs. The first peak is assayed under conditions of higher substrate and enzyme concentration and longer incubation time, thus its activity is much weaker than that of the second peak. The activity in the second peak is inhibited by salt, so the dilution increases its apparent activity. The first peak is comprised of HPLC fractions 9, 10, 11, and 12; the second peak is comprised of fractions 13, 14, 15, and 16 (Figure 13B). It is clear that the second peak, when T709 is grown on xylose, is almost twice the size of the second peak in the wild type (367-2A) when grown on xylose. The first peak in T709 is not significantly different from the first peak in the wild type. However, the wild type’s second peak, when grown on sucrose, is almost the same as that when grown on xylose. On the other hand, there is a very drastic difference in the size of T709’s second peak 52 +851“ peak 2 6°00 I:|50114 E g _ 9?. peak1 '5- 4000 - ‘ U ( c- O I 0 18 20 222420 2830 32 Retentionlmomiin HPLC Fractionation at HDAC Activity 8000 ~ , 7 -- - W - ,.,__ , — WW - ~WWW _, — W . W + Wild Type, 2% Sucrose I + Wild Type. 2% Xylose I 7000 ‘ + T709. 2% Sucrose 1 —u- T709. 2% Xylose I 6000 I 5000 7 I I I 8 4000 1 3000 < 2000 / - ‘5. ‘ , 1000 o I C Fraction 9 Fraction 10 Fraction 11 Fraction 12 Fraction 13 Fraction 14 Fraction 15 Fraction 16 iiii i 1 II _ 367-2A Figure 13. HPLC separation of HDAC activity comparing T709 and wild type strains. A. Standard pattern of HPLC separation of HDACs (from Baidyaroy et al., 2002). 367-2A, 881 1 l, and SB114 are wild-type isolates of C. carbonum. B. HPLC fractionation of HDACs followed by assays of each fraction, strains 367-2A and T709-3 grown in sucrose versus xylose. C. Western blot analysis of the HPLC fi'actions, hybridized with an antibody raised against the C-terminus of Hdc2. $81 1 l= control, total HDAC extracts. 367-2A and T709-3 were grown on xylose. 53 when grown on xylose, as compared to sucrose (Figure 138). Under xylose, T709’s second peak is about twice as large as the second peak when T709 is grown on sucrose. Western analysis indicates that the majority of the Hdc2 protein seems to be contained within the second HPLC peak in the T709 strain (Figure 13C). Polyclonal antibodies raised against the unique carboxyl terminus of Hdc2 (~200 amino acids) were obtained from Brosch and collaborators at the University of Innsbruck, Austria. HPLC fractions 9 through 16 were concentrated with TCA and loaded on gradient SDS-PAGE gels (Biorad). The western blot was hybridized with the a-Hdc2 antibodies (1:3,000) overnight. Hdc2 could be detected quite well in fractions 13 through 16 on T709, but it is absent in fractions 9 through 12. There is no strict correlation between the amount of protein in the fractions and the corresponding HDAC activity. The reason may be that Hdc2 in fraction 16, for instance, could be inactive, yet still recognized by the a-Hdc2 antibodies. Surprisingly, Hdc2 could not be detected in any fraction in the wild type strain 367—2A. Total HDAC extracts from the wild type strain SBl 11 were used as a control. As a way to ensure that the fraction samples contained a similar concentration of proteins, the HPLC traces were found to contain areas under the major peaks of similar size, meaning that most likely there was not over-loading of any one sample (data not shown). Interestingly, 831 11 shows a prominent band indicating that adequate amounts of protein were loaded. However, it is not yet understood why the a-Hch antibody recognizes Hdc2 in the wild type SBl 11 but not in the wild type 367-2A strains. In conclusion, the antibody used indicates that Hdc2 is present abundantly in the second peak of T709 but is unable to detect Hdc2 in the first peak. In the wild type, Hdc2 can not 54 be detected by this particular antibody in any of the two peaks. It appears that the antibody is only recognizing the recombinant Hdc2 while the native protein is not being recognized. It is possible that the native Hch is bound by other proteins such that the antibody can not bind to it. Alternatively, a post-translational modification (e. g. phosphorylation) occurs uniquely to the native Hch and not the recombinant one, so the antibody is unable to bind the native Hdc2. The two HDAC peaks were analyzed for sensitivity to exogenous HC-toxin. The first peak in both wild type and T709 was almost completely resistant to the toxin (Figure 14). The second peak, on the other hand, was inhibited by almost 80% in both the wild type and T709. These results indicate that Hdc2 is present mostly in the second peak and this HDAC is particularly sensitive to HC-toxin. Does over-expression of an HDAC lead to more virulence in C. carbonum? To answer this question, three-week-old maize seedlings were inoculated with spore suspensions (104/mL in O. 1% Tween 20) of T709-3 and wild type as a control. Maize leaves were analyzed 48 hours after inoculation. As seen on Figure 15, there is no significant distinction between the damage caused by wild type versus that caused by T709. Therefore, it appears that over-expression of an HDAC, HDC2, has no effect on virulence in C. carbonum. Discussion The filamentous ascomycete C. carbonum, a maize pathogen, probably has multiple mechanisms to protect itself against HC-toxin. One of these mechanisms of self- protection may involve its own HDAC activity being insensitive to HC-toxin when 55 Sensitivity of HDAC Peeks to Hc-Toxln ‘20 a Wlld Type ‘ [I] T 709 Percentage Resletance First Peak Second Peak Figure 14. Sensitivity of the T709 and wild type major HDAC activity peaks to HC- toxin. The two fractions that represent the two major peaks were tested for sensitivity to HC—toxin. Both fractions, in both wild type (367-2A) and T709 strains, seem to have similar levels of sensitivity to the toxin. The first peak is mostly resistant while the second peak is mostly sensitive. 56 douflsoofi Home Ego: we ecu—S 29$ 8555 do? .8 A