PHEb'r: II ”'1?ng l ,r 1c I an tat 20b Unigversity c This is to certify that the dissertation entitled ANALYSIS OF TRANSPORT AND SUB-CELLULAR LOCALIZATION OF AFLATOXIN BIOSYNTHETIC ENZYMES. VER-1 AND NOR-1, USING EGFP FUSIONS IN ASPERGILLUS PARASITICUS presented by SUNG-YONG HONG has been accepted towards fulfillment of the requirements for the Ph.D. degree in DEPARTMENT OF FOOD SCIENCE & HUMAN NUTRITION \7m 5 agm Major Professor’s Sflre H - l‘l —- Z 007 Date MSU is an affirmative-action, equal-opportunity employer 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 5/08 K‘IProi/AccaPres/CIRC/Daleoue indd ANALYSIS OF TRANSPORT AND SUB-CELLULAR LOCALIZATION OF AFLATOXIN BIOSYNTHETIC ENZYMES, VER-l AND NOR-l, USING EGFP FUSIONS IN ASPERGILL US PARASI TIC US By Sung-Yong Hong A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Food Science and Human Nutrition 2007 ABSTRACT ANALYSIS OF TRANSPORT AND SUB-CELLULAR LOCALIZATION OF AF LATOXIN BIOSYNTHETIC ENZYMES, VER—l AND NOR-1, USING EGFP FUSIONS IN ASPERGILL US PARASITIC US By Sung-Yong Hong Aflatoxins (AF) are toxic and carcinogenic secondary metabolites synthesized primarily by the filamentous fungi Aspergillus parasiticus and A. flavus when they grow on economically important food and feed crops including peanuts, treenuts, corn, and cottonseed. Aflatoxin contamination of food and feed results in huge economic losses and significant human and animal health risks throughout the world. Aflatoxin synthesis requires at least 24 enzyme activities encoded by up to 26 individual genes which are clustered within a 70 kb region on one chromosome. My research objective is the real- time sub-cellular localization of the enzymes encoded by the ver-I and nor-I genes involved in the aflatoxin B] (AF 8]) biosynthetic pathway in A. parasiticus. To monitor sub-cellular localization of aflatoxin enzymes in real time, we first developed an EGFP reporter system (pAPGFPVNB), which contains the ver-I promoter fiised to the egfp gene and a selectable marker (niaD, encodes nitrate reductase). EGFP was expressed in A. parasiticus, and the expression pattern of the ver-I promoter-driven EGFP was similar to the wild-type ver-I promoter and paralleled AFB] production in aflatoxin-inducing media. This reporter system was then modified to analyze the sub- cellular location of Ver-l and Nor-1. EGFP was fused to the N or C terminus of the aflatoxin enzymes with either a short or long hinge between the two proteins to ensure correct protein folding. These plasmids were transformed into A. parasiticus strains with one mutation in niaD and another in ver-I or nor-1. Transformants were screened for aflatoxin production using aflatoxin detection media CAM (coconut agar media) and for EGFP expression using fluorescence microscopy. Aflatoxin production was confirmed by TLC and ELISA. AF (+) and EGFP (+) transformants were then analyzed by Western blot, Southern hybridization, PCR, and nucleotide sequencing to demonstrate correct plasmid integration sites and correct expression of EGFP-tagged fusion proteins. Confocal laser scanning microscopy (CLSM) was used to track transport and location of these proteins in living cells. CLSM data indicated that N-terminally or C-terminally EGFP-tagged Ver-l and Nor-1 were localized in the cytoplasm and vacuoles (especially the vacuolar lumen) of fungal hyphae on aflatoxin-inducing solid media at maximum rates (48 h) of aflatoxin synthesis. Western blot analyses demonstrated that there was no evidence of turnover of the fusion proteins in the vacuoles. The data strongly suggest that the early and middle aflatoxin pathway enzymes, Nor-1 and Ver-l, were synthesized in the cytoplasm and transported to vacuoles of fungal hyphae for aflatoxin synthesis. ACKNOWLEDGMENTS First of all, I would like to thank my major adviser, Dr. John E. Linz, for his support, encouragement, and academic guidance throughout my study. He is a knowledgeable and great scientist. I have been very lucky because he has been my adviser. I am also grateful to Dr. Gale M. Strasburg, Dr. James J. Pestka, and Dr. Frances Trail for serving on my guidance committee. Special appreciation is also expressed to Dr. Melinda K. Frame and Dr. Shirley A. Owens for their assistance in confocal laser scanning microscopy. I would like to thank Dr. Ludmila V. Roze, Dr. Michael J. Miller, Dr. Ching-Hsun Chiou, Dr. Li-Wei Lee, Matthew Rarick, and other members in the Linz laboratory, and members of the Pestka laboratory for their help. Finally, I greatly appreciate my parents, family, and. friends for their emotional support and encouragement. iv TABLE OF CONTENTS LIST OF TABLES ......................................................................................................... vii LIST OF FIGURES ...................................................................................................... viii LIST OF ABBREVIATIONS ...................................................................................... xiv CHAPTER 1 LITERATURE REVIEW ................................................................................................ 1 AF LATOXINS Structure, toxicity, and possible roles of aflatoxins ........................................ 1 Elimination of aflatoxins from food and feed ................................................. 7 Aflatoxin biosynthetic pathway and genes ..................................................... 9 Regulation of aflatoxin biosynthesis ............................................................. 16 Sugar utilization gene cluster ........................................................................ 19 Gene cluster effects (positional effects) ........................................................ 19 Sub-cellular localization of enzymes involved in aflatoxin biosynthesis ..... 21 GREEN FLUORESCENT PROTEIN (GFP) Chracteristics of green fluorescent protein (GFP) ........................................ 22 Utilization of GFP in Aspergilli .................................................................... 26 VACUOLES The fungal vacuole ........................................................................................ 27 The plant vacuole .......................................................................................... 32 CHAPTER 2 UTILIZATION OF EGFP AS A REPORTER SYSTEM IN ASPERGILL US PARASI T ICUS .................................................................................................................. 37 ABSTRACT ........................................................................................................... 37 INTRODUCTION ................................................................................................. 38 MATERIALS AND METHODS ........................................................................... 40 RESULTS .............................................................................................................. 61 DISCUSSION ........................................................................................................ 85 ACKNOWLEDGMENTS ..................................................................................... 93 CHAPTER 3 SUB-CELLULAR LOCALIZATION OF THE VER-l PROTEIN IN ASPERGILLUS PARASI TIC US USING AN EGFP FUSION ..................................... 94 ABSTRACT ........................................................................................................... 94 INTRODUCTION ................................................................................................. 95 MATERIALS AND METHODS ........................................................................... 97 RESULTS ............................................................................................................ 113 DISCUSSION ...................................................................................................... 173 ACKNOWLEDGMENTS ................................................................................... 1 80 CHAPTER 4 SUB-CELLULAR LOCALIZATION OF THE NOR-l PROTEIN IN ASPERGILLUS PARASITICUS USING AN EGFP FUSION ................................... 181 ABSTRACT ......................................................................................................... 181 INTRODUCTION ............................................................................................... l 82 MATERIALS AND METHODS ......................................................................... 184 RESULTS ............................................................................................................ 199 DISCUSSION ...................................................................................................... 255 ACKNOWLEDGMENTS ................................................................................... 261 CONCLUSIONS ............................................................................................................ 262 CHAPTER 5 FUTURE STUDIES ....................................................................................................... 266 LIST OF REFERENCES .............................................................................................. 273 vi Table 2.1. Table 2.2. Table 3.1. Table 3.2. Table 3.3. Table 4.1. Table 4.2. Table 4.3. LIST OF TABLES Aspergillus parasiticus strains used in this study ...................................... 42 Primer sequences used in this study ........................................................... 50 Aspergillus parasiticus strains used in this study ...................................... 98 Primer sequences used in this study ......................................................... 104 Comparison of vacuole localization of EGFP in EGF P (+) transformant 33-15 with that of EGFP-tagged Ver-l in AF (+) and EGFP (+) transformants V86 and NV27 .................................................................. 170 Aspergillus parasiticus strains used in this study .................................... 185 Primer sequences used in this study ......................................................... 189 Comparison of vacuole localization of EGFP in EGF P (+) transforrnant B3-15 with that of EGFP-tagged Nor-1 in EGFP (+) transfonnant NN6 with non-detectable orange-pigment ........................................................ 252 vii Figure 1.1. Figure 1.2. Figure 1.3. Figure 1.4. Figure 1.5. Figure 1.6. Figure 1.7. Figure 2.1. Figure 2.2. Figure 2.3. Figure 2.4. Figure 2.5. LIST OF FIGURES Conidiophore characteristics of Aspergillus species .................................... 2 Chemical structures of the primary aflatoxins (Bl, G1, 82, and G2) .......... 4 Metabolic biotransforrnation pathways and harmful effects for aflatoxin BI .. .. 6 Cluster of aflatoxin biosynthetic pathway genes, corresponding enzymes, and precursors involved in the synthesis of aflatoxin B], G], B2, and 02 .. ....... 10 Schematic of the mechanism of green light emission from GFP .............. 23 Vacuolar transport pathways in yeast ........................................................ 29 Vacuolar transport pathways in plant cells ................................................ 34 Restriction endonuclease map of the A. parasiticus 8.2 kb Sall fragment that contains complete niaD and partial niiA genes ................................... 39 Restriction endonuclease map of plasmid, pAPGFPVNB ......................... 41 Comparison of selected sequences in pAPGFPVNB], 2, and 3 ............... 45 Restriction endonuclease maps of plasmids, pNEB-Nl and pNANG-3....46 Restriction endonuclease map of cosmid NorA ........................................ 48 viii Ill! 4 Figure 2.6. Figure 2.7. Figure 2.8. Figure 2.9. Figure 2.10. Figure 2.11. Figure 2.12. Figure 2.13. Figure 2.14. Figure 2.15. Figure 3.1. Figure 3.2. Restriction endonuclease map of plasmid, pEGFP-Nl (Clonetech Laboratories, Palo Alto, CA) ..................................................................... 54 Slide culture diagram ................................................................................. 56 Fluorescence microscopy of A. parasiticus expressing egfi) ..................... 62 Schematic for Southern hybridization and PCR analyses of integration sites of pAPGFPVNB into the chromosome ............................................. 64 Southern hybridization and PCR analyses of integration sites in transformants carrying pAPGFPVNB], 2, or 3 ......................................... 69 Southern hybridization and PCR analyses of integration sites in transformants carrying pAPGFPV1.1NBl ................................................ 74 EGFP expression and AFB1 production in EGFP (+) transformants containing different plasmids and the recipient strain NR-l ..................... 77 Comparison of EGFP fluorescence activity in EGF P (+) transformants Bl-86, B3-15,orLBl-101 ......................................................................... 82 Confocal laser scanning microscopy (CLSM) of EGF P (+) transforrnant 83-15 .......................................................................................................... 83 Sub-cellular localization of EGF P in EGF P (+) transforrnant B3-15 ........ 86 Restriction endonuclease map of plasmid, pAPCGFPVFNB .................. lOl Comparison of selected sequences in 3 plasmids carrying an egfp—tagged ver-I ........................................................................................................ l 02 ix Figure 3.3. Figure 3.4. Figure 3.5. Figure 3.6. Figure 3.7. Figure 3.8. Figure 3.9. Figure 3.10. Figure 3.11. Figure 3.12. Figure 3.13. Fluorescence microscopy of A. parasiticus CSlO-N2 expressing egfi) ...1 15 TLC analysis of extracts from transformants carrying pAPCGFPVFNB and the recipient strain CSlO-N2 ............................................................. 118 AFB] analysis of extracts from transformants carrying pAPCGFPVFNB, the recipient strain CSlO-N2, and the wild-type strain SU-l .................. 120 TLC analysis of extracts from transformants carrying pAPCGFPLVFNB and the recipient strain CS 1 0-N2 ............................................................. 121 TLC analysis of extracts from transformants carrying pAPNGFPVFNB and the recipient strain CSlO-N2 ............................................................. 124 AF B1 analysis of extracts from transformants carrying pAPNGFPVFNB, the recipient strain CSlO-N2, and the wild-type strain SU-l .................. 127 Western blot analysis of EGFP-tagged Ver-l from transformants carrying pAPCGFPVFNB, the recipient strain CSlO-N2, and the wild-type strain SU-l ......................................................................................................... 128 Western blot analysis of EGFP-tagged Ver-l from transformants carrying pAPCGFPLVFNB, the recipient strain CSlO-N2, and the wild-type strain SU-l ......................................................................................................... 132 Western blot analysis of EGFP-tagged Ver-l from transformants carrying pAPNGFPVFNB, the recipient strain CS10-N2, and the wild-type strain SU-l ......................................................................................................... 135 Comparison of selected nucleotide and amino acid sequences of ver-I A from SU-l and CSlO—N2 ......................................................................... 140 Schematic for production of EGFP-tagged functional Ver-l fusion protein depending on the integration site of pAPCGFPVFNB into the ver-I A locus ......................................................................................................... 142 Figure 3.14. Figure 3.15. Figure 3.16. Figure 3.17. Figure 3.18. Figure 3.19. Figure 3.20. Figure 3.21. Figure 3.22. Figure 4.1. Figure 4.2. Figure 4.3. Schematic for Southern hybridization and PCR analyses of integration sites of pAPCGFPVFNB or pAPCGFPLVFNB into the chromosome ...145 Southern hybridization and PCR analyses of integration sites in transformants carrying pAPCGFPVFNB ................................................ 150 Southern hybridization and PCR analyses of integration sites in transformants carrying pAPCGFPLVFNB .............................................. 152 Southern hybridization and PCR analyses of integration sites in transformants carrying pAPNGFPVFNB ................................................ 155 Sub-cellular localization of C-terminally EGFP-tagged Ver-l in AF (+) and EGF P (+) transfonnant V86 .............................................................. 159 Sub-cellular localization of N-terminally EGFP-tagged Ver-l in AF (+) and EGFP (+) transfonnant NV27 ........................................................... 164 AF B1 production in transfonnant V86 (EGFP-tagged Ver—l) and transfonnant 83-15 (EGF P) .................................................................... 171 AF31 production in transformant V86 (EGFP-tagged Ver-l) and transfonnant B3-15 (EGFP) in slide culture ............................................ 172 Schematic of Cvt and autophagy pathways for transport to the vacuoles .................................................................................................................. 178 Restriction endonuclease map of plasmid, pAPCGFPNFNB .................. 187 Comparison of selected sequences in 3 plasmids carrying an egfp-tagged nor-1 ........................................................................................................ 188 Fluorescence microscopy of A. parasiticus B62 expressing egfp ............ 200 xi Figure 4.4. Figure 4.5. Figure 4.6. Figure 4.7. Figure 4.8. Figure 4.9. Figure 4.10. Figure 4.1 1. Figure 4.12. Figure 4.13. Figure 4.14. TLC analysis of extracts from transformants carrying pAPCGFPNFNB and the recipient strain B62 ..................................................................... 204 TLC analysis of extracts from transformants carrying pAPCGFPLNFNB and the recipient strain B62 ..................................................................... 206 AFB] analysis of extracts from transformants carrying pAPCGFPLNFNB, the recipient strain B62, and the wild-type strain SU-l ........................... 207 TLC analysis of extracts from transformants carrying pAPNGFPNFNB and the recipient strain B62 ..................................................................... 208 AFB] analysis of extracts from transformants carrying pAPNGFPNFNB, the recipient strain B62, and the wild-type strain SU-l ........................... 211 Western blot analysis of EGFP-tagged Nor-1 from transformants carrying pAPCGFPNFNB, the recipient strain B62, and the wild-type strain SU-l .................................................................................................................. 212 Western blot analysis of EGFP-tagged Nor-l from transformants carrying pAPCGFPLNFNB, the recipient strain B62, and the wild-type strain SU-l .................................................................................................................. 216 Western blot analysis of EGFP-tagged Nor-l from transformants carrying pAPNGFPNFNB, the recipient strain B62, and the wild-type strain SU-l .................................................................................................................. 218 Comparison of selected nucleotide and amino acid sequences of nor-1 from SU-l and B62 .................................................................................. 222 Schematic for production of EGFP-tagged functional Nor-1 fusion protein depending on the integration site of pAPCGFPNFNB into the nor-1 locus ......................................................................................................... 223 Schematic for Southern hybridization analysis of integration sites of pAPCGFPNFNB into the chromosome ................................................... 227 xii Figure 4.15. Figure 4.16. Figure 4.17. Figure 4.18. Figure 4.19. Figure 4.20. Figure 4.21. Figure 4.22. Figure 4.23. Figure 5.1. Figure 5.2. Figure 5.3. Southern hybridization analysis of integration sites in transformants carrying pAPCGF PNFN B ....................................................................... 230 Schematic for Southern hybridization and PCR analyses of integration sites of pAPCGFPLNFN B into the chromosome .................................... 231 Southern hybridization analysis of integration sites in transformants carrying pAPCGFPLNFNB ..................................................................... 235 Southern hybridization analysis of integration sites in transformants carrying pAPNGF PNFN B ....................................................................... 237 Sub-cellular localization of C-terminally EGFP-tagged Nor-1 in light orange-pigmented EGFP (+) transfonnant LN196 .................................. 240 Sub-cellular localization of N-terminally EGFP-tagged Nor-1 in EGF P (+) transfonnant NN6 with non- detectable orange-pigment ........................ 246 AFB] production in transfonnant NN6 (EGFP-tagged Nor-1) and transfonnant B3-15 (EGF P) ..................................................................... 253 AFB] production in transformant NN6 (EGFP-tagged Nor-1) and transfonnant B3-15 (EGFP) in slide culture ............................................ 254 Proposed model for aflatoxin production in fungal cells ......................... 264 Restriction endonuclease map of plasmid, pAPCGFPOFNB .................. 267 Restriction endonuclease map of plasmid, pAPCGFPBFNB .................. 269 Western blot analysis of Ver-l from NV27 transfonnant, the recipient strain CSlO-N2, and the wild-type strain SU-l ....................................... 270 xiii EGFP GUS CAM LB YES YEP CZ PDA NOR, NA AVN VER A, VA DMST ST OMST OAVN ER CPY ALP API LV PSV PVC CCV DV PAC TLC ELISA NMR CLSM TEM CMAC DIC ORF LIST OF ABBREVIATIONS aflatoxin enhanced green fluorescent protein B-glucuronidase coconut agar medium Luria-Bertani medium yeast extract and sucrose medium yeast extract and peptone medium Czapek-Dox medium potato dextrose agar norsolorinic acid averantin versicolorin A demethylsterigmatocystin sterigmatocystin O-methylsterigmatocystin 5’-oxoaverantin endoplasmic reticulum carboxypeptidase Y alkaline phosphatase cytoplasm-to-vacuole targeting multivesicular body aminopeptidase I lytic vacuole protein-storage vacuole prevacuolar compartment clathrin-coated vesicle dense vesicle precursor-accumulating vesicle thin layer chromatography enzyme-linked immunosorbent assay nuclear magnetic resonance confocal laser scanning microscopy transmission electron microscopy 7-amino-4-chloromethylcoumarin differential interference contrast open reading frame xiv CHAPTER 1 LITERATURE REVIEW AFLATOXIN S Structure, toxicig, and possible roles of aflatoxins Aflatoxins are biologically active secondary metabolites produced by certain strains of Aspergillus parasiticus, A. flavus, A. nomius, A. pseudotamarii, A. bombycis, and A. ochraceoroseus (Cotty et al., 1994; Bhatnagar et a1. , 2003). These members of the Deuteromycotina (Fungi Imperfecti) lack a sexual stage and reproduce by conidia (Figure 1.1) (Moore-Landecker, 1990). These fungi are capable of infecting a wide variety of crops, including peanuts, cottonseed, corn, and tree nuts (Ellis et al. , 1991). Aflatoxin contamination in food and feed leads to both economic losses and human health risks throughout the world. Human exposure to aflatoxins occurs by ingestion of contaminated crops or products derived from animals given contaminated feed. It can lead to acute toxicity including hepatotoxicity, teratogenicity, immunotoxicity, and even death (Dvorackova, 1990). The potential toxicity of aflatoxins to human health was initially realized following an outbreak of acute hepatotoxic disease in poultry known as Turkey X disease in the United Kingdom, which resulted in the death of over 100,000 turkeys (Blount, 1961), and the etiologic agents were identified as A. flavus metabolites which were subsequently characterized and named aflatoxins (Asao et al. , 1963). Later human exposure to aflatoxins resulted in fatal aflatoxicosis in India (Krishnamachari er al., 1975) and West Africa (N gindu et al. , 1982). Since their discovery in the early \\\\\\‘§']‘"I$”//éé -: . . /.‘-\ . .. 3,, Conrdla EQ ‘\ [I] J/ . I. .90, . a. . : N I -'s:~-"(tu'lfy//... Stengmata 35$ ’4”? . "*‘S‘ 0,; . (Phialides) 2.2.-.3: ...::-:-:-2 o ‘3‘: £7: :g-E 5:. 3.: 3*. Metulae Z", «1* 9’, e‘?’ ”(fl/’11:- “6S5 3;”: L_ 9?- Vesicle 23" "v’l "‘ ‘ !- Stipe Foot cell Figure 1.1. Conidiophore characteristics of Aspergillus species. This schematic shows a single layer of phialides or sterigmata (uniseriate) and double layer of cells, phialides and metulae (biseriate). (Adapted from Klich and Pitt, 1988) 1960’s, the toxicity of aflatoxins has been studied extensively. In addition to hepatotoxicity, aflatoxins have been established as potent mutagens, carcinogens, and teratogens (McLean and Dutton, 1995). Aflatoxin B] (AFB]) has been proven to be a potent carcinogen in laboratory animals such as ducks (Canaghan, 1965), rainbow trout (Sinnhuber et al. , 1968), rats (Epstein et al., 1969), rhesus monkeys (Tilak, 1975), and tree shrews (Reddy et al. , 1976). Also, AFB] has been shown to be a potent teratogen in rats (Butler et al., 1966), hamsters (Elis and DiPaolo, 1967), and chick embryos (Verrett et al., 1973), and to produce immunotoxic effects resulting in increased susceptibility to diseases caused by bacteria, fungi, and viruses (Edds et al., 1973; Richard et al., 1973; Giambrone et al., 1985; Pestka and Bondy, 1990). Epidemiological data have linked AFB] with human liver cancer. AFB] has been suggested to have a synergistic effect with hepatitis B virus (HVB) in production of human primary liver cancer (PLC), and human hepatocellular carcinoma (HCC) (Goopman et al. , 1988; Hall and Wild, 1994). Seventeen different compounds were isolated and named as aflatoxins, but the most common aflatoxins are classified B], B2, G], G2 based on their fluorescent colors under long-wave UV. light (B for blue, G for green) (Figure 1.2). Toxigenic A. parasiticus, A. nomius, A. pseudotamarii, and A. bombycis produce both aflatoxin B and G groups while toxigenic A. flavus and A. ochraceoroseus only produce aflatoxin G group (Dvorackova, 1990; Bhatnagar et al. , 2003). The order of acute and chronic toxicity of the aflatoxins are AF B]>AFG]> AFB2>AFG2, suggesting that epoxidation of the 8, 9-double bond and also the presence of the cyclopentenone ring of the B group may 00 on O O OCH-,- Aflatoxin B2 0 O O O O O I O O O OCH3 O O OCH3 Aflatoxin Gr Aflatoxin G2 Figure 1.2. Chemical structures of the primary aflatoxins (B], G], B2, and Oz). play a major role in the harmful effects when compared with the six-membered lactone ring of the G group (McLean and Dutton, 1995). Toxicity of AFB] requires its bioactivation by microsomal cytochrome P-450, which converts AFB] to a variety of metabolites including AF B]-8, 9-epoxide (Figure 1.3) (Stark, 1980; Essigmann et al., 1982; Eaton et al. , 1994). The AFB]-8, 9-epoxide can form adducts with DNA and protein due to its electrophilicity, resulting in defective DNA repair, mutations, cancers, and enzyme malfunctions. It is known that the AF B ] -8, 9-epoxide binds covalently to the N7 position of guanine in DNA to form DNA adducts (Essigmann et al., 1982) and produces G-T transversions at the third base of codon 249 of the p53 tumor suppressor gene (Puisieux et al., 1991; Eaton and Gallagher, 1994). Because of the high level of concern about aflatoxin, the Inemational Agency for Research on Cancer (IARC) designated AFB] a probable htunan carcinogen and the US FDA set action levels of 20 p.p.b. for human food (except for milk, where the level is 0.5 p.p.b.) and 20-300 p.p.b. for most animal feeds (Labuza, 1983; Wogan, 1992; CAST, 2003). Some proposals for the biological function of aflatoxins include the removal of excess acetate when fungi grow on carbon-rich sources (Bu’Lock, 1965), chemical signals between species (Lillehoj, 1991), promotion of fungal development such as conidia or sclerotia (Cotty, 1988; Chang et al. , 2001), protection of fungi against soil microbial competitors, insect predators, or environmental stresses like UV. radiation (Drummond and Pinnock, 1990; Matsumura and Knight, 1967; Perez-Rodriguez et al. , 2003). Glucuronide Aflatoxin M 1'P1 conjugate Aflatoxicol H] x / Glucuronide Aflatoxin P] conjugates Aflatoxin Q] \ . \ Aflatoxin M] Aflatoxln B] DNA adducts Aflatoxicol M] / Aflatoxln-8,9-epoxide Aflatoxicol GSH conjugate I \ $ Protein adducts Glucuronide / \ conjugates Aflatoxin B]-8,9-dihydrodiol Aflatoxin 82a Figure 1.3. Metabolic biotransforrnation pathways and harmful effects for aflatoxin B]. (Adapted from Eaton et al., 1994) Elimination of aflatoxins from food and feed Approximately one-quarter of the food crops in the world are affected by mycotoxins annually (CAST, 2003). Therefore there is a huge worldwide economic cost that arises directly from crop and livestock losses and from regulatory programs designed to reduce health risks to animals and humans (Shane, 1994). Consequently, elimination of aflatoxins from the food and feed would reduce this economic cost. Two strategies have been used or proposed for reducing or eliminating aflatoxins from food and feed (Trail et al., 1995a). Preharvest strategies are designed to block fungal infection of the crops or to block the ability of the fungi to grow or synthesize aflatoxins on the crops (Darrah and Barry, 1991; Scott and Zummo, 1988). Postharvest strategies include aflatoxin screening/detection, removal/decontamination, or altered aflatoxin metabolism/DNA adduct formation (Chu, 1991; Ellis et al., 1991; Park and Liang, 1993), and these strategies can provide a safety net to remove low levels of aflatoxins that may escape preharvest control. Since aflatoxins are resistant to normal food processing such as milling and cooking, and postharvest controls are expensive and ineffective, preharvest strategies may present better approaches, and could obviate the need to detoxify large quantities of aflatoxin-contaminated seed material and avoid the uncertainties of gaining approval from regulatory agencies for the use of detoxified seed for animal or human food. These strategies include improved agronomic practices such as irrigation and application of fungicides and use of aflatoxin-resistant crop varieties by conventional breeding methods. Use of pesticides is environmentally unacceptable or too expensive and use of irrigation has demonstrated only limited potential for reducing aflatoxin levels in the field, especially in years of drought when environmental conditions are optimum for contamination. Also, genetically stable aflatoxin-resistant crops have not been successfully developed using conventional breeding methods. However, the application of molecular biology for preharvest strategies has been proposed (Bhatnagar er al., 1995). These approaches include development of genetically engineered crops to reduce fungal growth or inhibit aflatoxin synthesis, and utilization of genetically stable nontoxigenic strains of Aspergillus to competitively exclude toxigenic strains from infecting the crops. The following review will be limited to development of biocontrol strains. Use of nontoxigenic biocompetitive and native strains of A. flavus has demonstrated significant efficacy in reduction of aflatoxin contamination in cottonseed (Ehrlich, 1987; Cotty, 1990; Cotty, 1994; Antilla and Cotty, 2002), corn (Brown et al., 1991), and peanuts (Domer et al. , 1992; Cotty and Bhatnagar, 1994; Domer et al. , 1998; Doner, 2004) in green house and field studies. Two A. flavus isolates, AF 36 and NRRL 21882, were recently registered as biopesticides with the United States Environmental Protection Agency (US EPA) for the management of aflatoxin-producing fungi (Antilla and Cotty, 2002; Doner, 2004). However, one study suggested that naturally occurring nontoxigenic isolates of A. flavus may have the genetic capability to synthesize AFB] under as yet undetermined environmental conditions (Rarick et al., 1994). Therefore it is important to generate genetically stable nontoxigenic strains of Aspergillus by inhibition of aflatoxin biosynthetic or secretory processes through a molecular genetic approach (Bhatnagar et al., 1995; Cleveland et al. , 1997). This can be achieved from knowledge about the fundamental molecular and biological mechanisms that regulate the biosynthesis of aflatoxin and the processes of localization of the aflatoxin biosynthetic proteins by the fungi. My research focus is to identify the real-time sub-cellular location of key enzymes, Nor-1 and Ver-l, involved in AFB] synthesis. Aflatoxin biosynthetic pathway and genes Bioconversion experiments using aflatoxin-blocked mutants, metabolic inhibitors, and radiolabeled pathway intermediates revealed the generally accepted aflatoxin biosynthetic pathway (Figure 1.4) (Bennett et al., 1980; Bhatnagar et al. , 1987; Hsieh et al., 1973, Hsieh et al., 1976; McCormick et al., 1987; Shroeder et al. , 1974). The generally accepted AFB] biosynthetic pathway is as follows (Dutton, 1988; Bhatnagar et al., 1992; Minto and Townsend, 1997; Brown et al., 1999); acetyl CoA + 2 malonyl CoA —-> hexanoate + 7 malonyl CoA —) polyketide noranthrone precursor —> norsolorinic acid (NOR) —) averantin (AVN) —> 5’-hydroxyaverantin (HAVN) —> averufanin (AVNN) —-> averufm (AVF) —> versiconal hemiacetal acetate (VHA) —> versiconal (VAL) —> versicolorin B (V ER B) —+ versicolorin A (V ER A) —) demethylsterigmatocystin (DMST) —-) sterigmatocystin (ST) —+ O-methylsterigmatocystin (OMST) -—> aflatoxin B] (AF B]). It has been established that VER B is a branching point in the pathway leading to both AFB] or G] and AFB2 or G2 (Bhatnagar et al., 1991; Cleveland et al., 1987a; Yabe et al. , 1988) and that the B group and the G group are synthesized independently (Floyd et al., 1987; Henderberg et al. , 1988). The pathway leading to AF82 is as follows; VER B —+ dihydro-DMST (DHDMST) —> dihydro-ST (DHST) —+ dihydro-OMST Figure 1.4. Cluster of aflatoxin biosynthetic pathway genes, corresponding enzymes, and precursors involved in the synthesis of aflatoxin B], G], B2, and G2. (A) The horizontal line represents a 70 kb aflatoxin biosynthetic pathway gene cluster and a 12 kb sugar utilization gene cluster. Arrows along the horizontal line indicate the direction of transcription of the genes. The new names of the genes are shown below and the old names are shown above the horizontal line. (B) Horizontal arrows indicate the relationships between the genes and the enzymes they encode, from the enzymes to the conversion steps they are involved in, and from the intermediates to the products in the aflatoxin biosynthetic pathway. Abbreviations for the intermediates and the products are as follows; NOR, norsolorinic acid; AVN, averantin; HAVN, 5’-hydroxyaverantin; OAVN, oxoaverantin; AVNN, averufanin; AVF, averufin; VHA, versiconal hemiacetal acetate; VAL, versiconal; VER B, versicolorin B; VER A, versicolorin A; DMST, demethylsterigmatocystin; DHDMST, dihydrodemethylsterigmatocystin; ST, sterigmatocystin; DHST, dihydrosterigmatocystin; OMST, O-methylsterigmatocystin; DHOMST, dihydro-O-methylsterigmatocystin; AFB], aflatoxin B]; AFBZ, aflatoxin Bz; AFG], aflatoxin G]; AFGZ, aflatoxin Gz. (Adapted from Yu et al., 2004) 10 is 5% 9? Se as is Be Se 5? Se Se .89 Es Es E? as 2.9 Es he as 3% S? Be as Be .3839 sawsm ItFItflU-UHXUAUIHYHYHVUHYUUUDUUHYAUHVUHYUUHT .1... \Le V63 V3»: 595 Rake $98 $83 New: v29. v.3: ~33» 5? 7.3: V30 Mai. “.33 Vase has»: a? 3:8 EMS SE: 7.8: See 3? Task wreak. V53 5%» MS: Figure 1.4. (cont’d) aflA (firs-2) "F aflB (fas-I) —> aflC (pksA) -> aflD (nor-I) —> aflE (norA) —> aflF (norB) —> aflG (avnA) —> aflH (adhA) —> (1111 (ava) —-> aflJ (estA) —-> aflK (vbs) —> aflL (verB) —> aflM (ver-I) —> aflN (verA) —> aflO (omtB) —> aflP (omtA) —> aflQ (ordA) —> Fatty acid synthase (1 Fatty acid synthase [1 Polyketide synthase Reductase NOR-reductase Dehydrogenase P450 monooxygenase Alcohol dehydrogenase Oxidase Esterase VERB synthase Desaturase Dehydrogenase Monooxygenase O-methyltransferase B O-methyltransferase A Oxidoreductase 12 ACETATE I POLYKETIDE NOR AVN l. l: \H XH ¥¥¥¢ K \‘VER A 4 V 3‘ DMS?‘ DHDMST \ 8% \ DXST a? air“ AFB] AFG1AFB2 AFGz (DHOMST) —-> aflatoxin 3; (AF B2). Aflatoxin synthesis requires at least 24 enzyme activities encoded by up to 26 individual genes which are clustered within a 70 kb region on one chromosome (Figure 1.4) (Dutton, 1988; Bhatnagar et al., 1992; Trail et al. 1995b; Yu et al., 1995 ; Yu et al. , 2004). Several different strategies have been used to clone the biosynthetic genes including genetic complementation (Chang et al. , 1992, Chang et al. , 1993; Skory et al. , 1992; Payne et al., 1993), reverse genetics (Yu et al., 1993; Cary et al., 1996), and subtractive hybridization (Feng et al. , 1992). Since genetic transformation systems have been developed for A. flavus (Wolosuk et al., 1989) and A. parasiticus (Homg et al., 1990; Skory et al. , 1990), genetic complementation was used to clone nor-1 (Chang et al., 1992), which is a gene involved in the conversion of norsolorinic acid (NA) to averantin (AVN), ver-I (Skory et al. , 1992), which is a gene involved in the conversion of versicolorin A (VER A) to demethylsterigmatocystin (DMST), and aflR (Payne et al. , 1993; Chang et al., 1993), which is a pathway regulatory gene. These studies utilized aflatoxin-blocked mutants such as A. parasiticus B62 (nor-1, niaD, br-I), CSIO (ver-I , pyrG, wh-I), and RHNI (ordA, niaD) as recipient strains for complementation. A reverse genetics approach relying on purified pathway enzymes was used to clone omtA (Yu et al., 1993), involved in the conversion of stergrnatocystin (ST) to O-methylsterigmato- cystin (OMST), and NorA (Cary et a1. , 1996), involved in the conversion of norsolorinic acid (NA) to averantin (AVN). These researchers used antibodies raised against the purified enzymes to screen a cDNA expression library in E. coli. Also, vbs, a gene involved in the conversion of versiconal (VAL) to versicolorin B (VER B), was cloned using reverse transcriptase-mediated PCR (RT-PCR) and degenerate oligonucleotide 13 primers designed from the amino acid sequences of peptide fragments from purified VBS (Silva et al. , 1996; Silva and Townsend, 1996). Subtractive hybridization was used to clone several genes which are transcribed coordinately with aflatoxin production in A. parasiticus (Feng et al. , 1992). The goal of my research is to identify the real-time sub-cellular location of the enzymes encoded by ver- 1 and nor-1 genes involved in the AFB] biosynthetic pathway. The following review will be limited to ver- 1 and nor—I genes. 1. ver-I The ver-I gene was cloned by complementation of A. parasiticus CSlO (ver-I, pyrG, wh-I) that accumulated the yellow pigment versicolorin A (V ER A or VA) (Skory et al., 1992). Southern hybridization analysis showed that there were two copies of the ver-I gene (ver-I A and ver-I B) in A. parasiticus SU-l (Liang et al., 1996). Further investigation identified a 12 kb duplication of the aflatoxin gene cluster that spanned from aflR to ver—I plus omtB (Liang er al., 1996; Chang and Yu, 2002). The genes in the truncated aflatoxin gene cluster contained several mutations and seemed to be non- functional. This partial duplication of aflatoxin pathway genes in A. parasiticus was proposed as one possible explanation for the higher stability of the aflatoxin production in A. parasiticus as compared to A. flavus, in which no duplication of the aflatoxin gene cluster has been reported (Trail et al., 1995b; Cary et al. , 2000). More than 90 % of A. parasiticus isolates produce aflatoxin while 50 % of A. flavus isolates are toxigenic (Bennett and Papa, 1988). The predicted amino acid sequence of the 28 kDa Ver-l showed that ver-I B had 95 % identity with ver-I A but it contained a translational stop codon resulting in a truncated and nonfuntional polypeptide (Liang et al. , 1996). Inactivation of ver-I A in A. parasiticus NR-l resulted in accumulation of VER A. The deduced amino acid sequence of the ver-I gene revealed a striking similarity with ketoreductases involved in polyketide biosynthesis in Streptomyces (Skory et al. , 1992) and polyhydroxynaphthalene reductases involved in melanin biosynthesis in Magnaportha grisea (Liang et al. , 1996). These results were supported by the proposal that the ver-I gene product is responsible for the reduction of a dienone intermediate formed by epoxidation of the anthraquinone ring of VER A to a precursor in a Baeyer- Villiger oxidation reaction catalyzed by hypA (Ehrlich et al., 2005; Henry and Townsend, 2005). Also, the ver-I gene product is responsible for the deoxygenation of VER A to 6- deoxy VER A (Liang et al. , 1996). It is currently known that ver-I encodes an NADPH- dependent oxidoreductase which is involved in conversion of versicolorin A (VER A) to demethylsterigmatocystin (DMST) (Ehrlich et al., 2005 ; Henry and Townsend, 2005). 2. nor—I The nor-1 gene was cloned by complementation of A. parasiticus B62 (nor-1, niaD, br-I) that accumulated the brick-red pigment norsolorinic acid (NOR or NA) (Chang et al., 1992). The predicted amino acid sequence of the 31 kDa Nor-l showed 23% identity with several NAD(P)-binding dehydrogenase/reductases (Trail at al. , 1994). Inactivation of nor-1 in A. parasiticus NR-3 resulted in accumulation of NOR, but small quantities of aflatoxin were still produced (Trail 21 al. , 1994). This observation can be explained by an alternate route(s) or enzyme(s) that can synthesize AVF (averufin) from NOR (Bhatnagar et al. , 1992; Yabe et al., 1993; Cary et al. , 1996). It is currently known that nor-1 encodes an NADPH-dependent ketoreductase which is involved in conversion of norsolorinic acid (NOR) to averantin (AVN) (Zhou and Linz, 1999). Regglation of aflatoxin biosynthesis Aflatoxins are produced at maximum rates by A. parasiticus and A. flavus in aflatoxin—inducing liquid media during a transition (idiophase) from exponential grth to stationary phase when growth has slowed or ceased (Trail et al. , 1995a). Because the precursor (i.e. acetate) of aflatoxin biosynthesis is one product of primary metabolism, aflatoxin biosynthesis could be affected by factors which regulate primary metabolism such as enzyme cofactors (NADPH) and building blocks (acetyl CoA, malonyl CoA, S- adenosylrnethionine) (Luchese and Harrigan, 1993). It was reported that glucose induced certain enzymes in the glycolytic pathway and repressed enzymes in the hexose monophosphate shunt (HMS) and the tricarboxylic acid cycle (TCA) resulting in increased aflatoxin production (Buchanan and Lewis, 1984; Buchanan etal., 1985). It was proposed that the glucose effect was due to the shunt of acetate toward synthesis of polyketides and away from oxidation via the TCA cycle or use in fatty acid synthesis (Buchanan et al., 1987). Northern hybridization analysis and a B-glucuronidase (GUS) activity assay demonstrated that glucose stimulated the expression of aflatoxin genes such as nor-I and ver-I at the transcription level in A. parasiticus (Skory et al. , 1993; Liang et al., 1997). Also, sodium nitrate added as the sole nitrogen source completely inhibited aflatoxin biosynthesis, but ammonium nitrate induced the synthesis (Kacholz and Demain, 1983). It was proposed that nitrate stimulated glucose-6-phosphate dehydrogenase and mannitol dehydrogenase activities resulting in a high NADPH/NADP 16 ratio and increased fatty acid synthesis and decreased polyketide synthesis (N iehaus and Jiang, 1989). Northern hybridization analysis demonstrated that nitrate inhibited the accumulation of transcripts of aflatoxin genes such as nor-1, ver-I , and omtA in A. parasiticus (Chang et al., 1995). These results suggest that aflatoxin genes might be regulated by common regulatory factors that affect secondary metabolism as well as primary metabolism. The aflR gene has been demonstrated to encode a key protein in regulation of aflatoxin gene expression. Mutations in aflR inhibited aflatoxin synthesis and blocked the accumulation of transcripts and enzyme activities (Payne et al., 1993; Woloshuk et al., 1994). Complementation of an aflR mutation restored function of the pathway (Payne et al., 1993). Also, an aflR knockout mutant no longer produced aflatoxins nor did it produce any aflatoxin precursors (Cary et al. , 2002). Transformation of A. parasiticus with an additional copy of aflR led to increased aflatoxin production and elevated accumulation of nor-1, ver-I, and pksA transcripts (Chang et al. , 1995). Analysis of the predicted amino acid sequence of aflR showed a cysteine-rich binuclear zinc cluster DNA binding motif (Cys-Xaa2-Cys-Xaa6-Cys-Xaa6-Cys-Xaa2-Cys-Xaa6-Cys) (Chang et a1. , 1993; Woloshuk et al., 1994), which is characteristic of a group of fungal and yeast transcriptional activators including GAL 4 of Saccharomyces cerevisiae (Giniger, 1985; Johnston, 1987). Electrophoretic mobility shift assays (EMSA) demonstrated binding of AflR to a palindromic sequence (TTAGGCCTAA) in its own promoter in A. parasiticus (Chang et al., 1995) and to another conserved palindromic sequence (TCGN5CGA) in the promoters of 11 aflatoxin biosynthetic genes (Ehrlich et al. , 1999). I7 The effect of metal ions like zinc on aflatoxin production was reported (Tiwari et al. , 1986). The absence of zinc in a grth medium containing sucrose completely eliminated AFB] synthesis and severely reduced aflatoxin gene expression even though growth decreased by only 2 fold when compared to media with zinc (Bennett et al. , 1979; Miller, 2003). This observation can be explained by two possible mechanisms: 1) alteration of AflR activity because AflR is a cysteine-rich binuclear zinc cluster transcription factor; and 2) decreased aflatoxin production through increased reducing power (NADPH/NADP ratio) in the cell; because zinc inhibited mannitol dehydrogenase and glucose-6-phosphate dehydrogenase activity in A. parasiticus (N iehaus and Dilts, 1982; Niehaus and Dilts, 1984). Also, it was reported that AflR induced gene transcription in the aflatoxin biosynthetic cluster as well as for genes outside of the cluster like nadA in the sugar utilization gene cluster (Price et al., 2006). It is known that there is an involvement of an additional transcription factor, AflJ, in regulation of aflatoxin biosynthesis (Meyers et al., 1998; Chang, 2003). AflJ was known to be necessary for accumulation of early precursor metabolites involved in the aflatoxin biosynthetic pathway (Meyers et al. , 1998) and to modulate AflR activity by the interaction with AflR as a coactivator (Chang, 2003). The effects of other factors on aflatoxin synthesis have also been reported. Aflatoxins were synthesized at maximum levels in a temperature range from 25 to 30 0C (Northolt et al., 1977; Davis and Diener, 1986), in a pH range of 4.0-6.0 (Keller et al. , 1997), and in the dark (Bennett et al. , 1981) in A. parasiticus. Sugar utilization gene cluster It was reported that simple sugars such as glucose, sucrose, maltose, and galactose support aflatoxin production in A. parasiticus (Bhatnagar et al., 1992). Recently ordB and hypA were identified at one end of aflatoxin pathway gene cluster and their function was verified in aflatoxin biosynthesis (Cary et al. , 2006; Ehrlich et al. , 2005). Following hypA, four genes (nadA, hxtA, gch, and sugR) in another cluster related to sugar utilization were cloned from A. parasiticus (Figure 1.4) (Yu et al., 2000). Analysis of the deduced amino acid sequences of the genes involved in sugar utilization showed that nadA has similarities to genes encoding a NADH oxidase, that thA has similarities to genes encoding a hexose transporter protein, that gch has similarities to genes encoding a-1,4- or a-1,6-glucosidases, and that sug R has similarities to Cys-6-type transcription factor genes encoding sugar utilization regulatory proteins. An RT-PCR experiment showed that the hxtA of these genes was expressed concurrently with aflatoxin pathway genes in aflatoxin-inducing media. Also, a cDNA microarray experiment showed that the nadA was upregulated in aflatoxin-conducive conditions (Price et al., 2006). These results suggest that there is a linkage between the two gene clusters which supports the induction of aflatoxin biosynthesis by simple sugars such as glucose or sucrose although no AflR binding motif was identified in the promoters of the four sugar utilization genes (Yu et al. , 2000; Yu et a1. , 2004). Gene cluster effects (positional effects) The clustering of genes involved in the same secondary metabolic pathway is common in the filamentous fungi (Keller and Hohn, 1997). Early studies showed that 19 clustered genes could be coordinately expressed during fungal development and that placement of the clustered genes at ectopic chromosomal locations resulted in the loss of the coordination (Miller et al., 1987; Timberlake and Barnard, 1981). Also, it was reported that the chromosomal location of genes could play a role in regulation of trichothecene biosynthesis gene expression (Hohn et al. , 1999; Chen et al. , 2000). Other studies showed that integration of aflatoxin gene promoters into their homologous sites within the aflatoxin gene cluster functioned similarly to the native promoters, but not at sites outside of the cluster (Liang er al., 1997; Chiou et al. , 2002). For example, integration of the GUS reporter plasmid, in which ver-I promoter was fused to uidA (encodes B-glucuronidase) and niaD (encodes nitrate reductase) as a selectable marker, into the niaD locus resulted in a SOD-fold reduction in ver-I promoter activity when compared with integration into the ver-I locus within aflatoxin gene cluster (Liang et al. , 1997). Similarly, integration of the GUS reporter plasmid carrying a nor-1 promoter into the niaD or pyrG (encodes orotidine monophosphate decarboxylase) locus resulted in no detectable nor-1 promoter activity (Chiou et al., 2002). Also, Northern and RT-PCR analyses indicated that aflR 2 in the partially duplicated aflatoxin gene cluster was transcribed at extremely low levels compared to aflR I (Cary et al. , 2002). Another RT- PCR experiment demonstrated that estA 2 in the partially duplicated aflatoxin gene cluster did not have defects in its promoter region and coding region but it was not expressed under either aflatoxin-conducive or non-conducive conditions (Yu et al. , 2003). Therefore, aflatoxin genes may be subject to position-dependent regulation within the aflatoxin gene cluster in A. parasiticus. 20 Sub-cellular localization of eggmes involved in aflatoxin biosynthesis Previous studies were performed to identify the sub-cellular location of Ver-l , Nor-l, OmtA, and VBS proteins (Liang, 1996; Lee et al., 2004; Zhou, 1997; Lax et al., 1986; Cleveland et al. , 1987b; Chiou et al., 2004). Ver-l was observed to be associated with structures similar in size to lysosomes as well as in the cytoplasm using cell fractionation (Liang, 1996). Also, Ver-l was distributed evenly throughout substrate- level hyphae as well as aerial hyphae, vesicles, and conidiophores using immuno- fluorescence microscopy after paraffin embedment and was primarily detected in the cytoplasm using transmission electron microscopy (TEM) after immunogold labeling in 24-48 h fungal colonies (Lee eta1., 2004). Nor-l was observed to be associated with structures similar in size to ribosomes as well as in the cytoplasm using cell fractionation (Zhou, 1997). Also, similar to Ver-l , Nor-1 was distributed evenly throughout substrate- level hyphae as well as aerial hyphae, vesicles, and conidiophores using immuno- fluorescence microscopy after paraffin embedment and was primarily detected in the cytoplasm using TEM after immunogold labeling in 24-48 h fungal colonies (Lee et al. , 2004). OmtA was observed in the postrnicrosomal cytoplasmic fraction using cell fractionation (Lax et al., 1986; Cleveland et al. , 1987b). OmtA was clustered in patches and was distributed evenly throughout substrate-level hyphae as well as aerial hyphae, vesicles, and conidiophores using immunofluorescence microscopy in 24-48 h fungal colonies (Lee et al., 2004). Also, OmtA was found inside vacuole-like structures as well as in the cytoplasm using TEM in 24—48 h fungal cells located near the basal (substrate) surface of the colony (Lee et al. , 2004). VBS was clustered in patches and was distributed evenly throughout substrate-level hyphae as well as aerial hyphae, vesicles, 21 and conidiophores using immunofluorescence microscopy in 24-48 h (Chiou et al. , 2004). Also, VBS was observed in ring-like structures around nuclei as well as in the cytoplasm using TEM and immunofluorescence microscopy and was consistent with passage of this protein through the endoplasmic reticulum (ER) (Chiou et al. , 2004). GREEN FLUORESCENT PROTEIN (GFP) Characteristics of green fluorescent protein (GFP) Many cnidarians and ctenophores emit light when mechanically disturbed. Light from cnidarians is primarily green whereas light emitted from ctenophores is blue. In the cnidarian Aequorea, the GFP of the jellyfish, a protein of 238 amino acids, absorbs blue light with an excitation maximum of 395 nm and a smaller peak of 475 nm, and emits green light with an emission maximum of 509 nm (Morise et al. , 1974; Prasher et al. , 1992). The energy needed for the emission of light is produced when calcium binds to the photoprotern aequorrn; the major players in light emissron consrst of a Ca2 -b1nding apoaequorin, coelenterazine, and molecular oxygen. In this reaction coelenterazine is oxidized to coelenterarnide yielding a blue fluorescent protein, C02, and light as products. The energy is then transferred to the GFP molecule and the excited state of GFP emits green light (Figure 1.5). On the other hand, the GF P from Renilla, another cnidarian, receives energy from a luciferase-oxyluciferin excited state complex by a radiationless energy transfer mechanism (Ward and Cormier, 1976). The Aequorea GFP has been reported to be a 27 kDa monomer although it is 22 Aequorin Jr Ca2+ Blue fluorescent protein + hvxmaxmnm + CO; I + GFP hv3.ma11509nm Figure 1.5. Schematic of the mechanism of green light emission from GFP. The . . . 2+ . . . . photoprotein aequorin consrsts of a Ca -b1nd1ng apoaequorin, coelenterazrne, and molecular oxygen. When calcium binds to the photoprotein aequorin, coelenterazine is oxidized to coelenterarnide yielding a blue fluorescent protein, C02, and light as products. The energy is then transferred to the GFP and the excited state of GFP emits green light. (Adapted from Inouye and Tsuji, 1994) 23 known to self-associate whereas the Renilla GF P is a 54 kDa homodimer (Morise et al. , 1974; Ward and Bokman, 1982; Prendergast and Mann, 1978; Prasher et al., 1992; Ward and Cormier, 1979). These GFPs have different absorption spectra in that Renilla GFP has an excitation maximum of 498 nm but an identical emission spectrum (kmax=509 nm) to Aequorea GF P. The two proteins contain chromophores having the same structure. The GFP chromophore consists of an imidazolone ring formed by post- translational cyclization and oxidation of the amino acids Ser—65, Tyr-66, and Gly-67 (Shimomura, 1979; Ward et al., 1989; Cody et al., 1993; Cubbitt et al., 1995). It was reported that the cDNA for Aequorea Victoria GFP resulted in GFP expression and produced a strong green fluorescence in prokayotic and eukaryotic cells when excited by blue light (Prasher et al. , 1992; Chalfie et al. , 1994). The GFP expressed in both E. coli and C. elegans was quite stable (like the native protein) when illuminated with light at 450 to 490 nm, and the GF P expressed in E. coli showed fluorescence excitation and emission spectra identical to those of the native protein (Chalfie et al. , 1994; Inouye and Tsuji, 1994). Also, the fluorescence did not require any additional gene products from A. Victoria. Several reporter systems are available to monitor gene activity and protein localization in living organisms. These include fusion of promoters or native proteins with coding sequences for B-galactosidase (Drahos et al. , 1986), B-glucuronidase (GUS) (Roberts et al. , 1989), chloramphenicol acetyltransferase (CAT) (Hagedom, 1994), and luciferase (LUC) (Prosser, 1994). Because such methods require exogenous substrates or cofactors, they have limited use with living cells. However, because the detection of intracellular GFP requires only irradiation by near UV. or blue light, it is not limited by 24 substrate availability and is useful as a marker for gene expression and as a tag in studying protein localization in a variety of organisms (Errampalli et al. , 1999). Also, GFP can be detected without cell disruption and monitored in situ and in real time. GFP is heat (65 °C) and pH (6-12) stable, and resistant to denaturants such as 1 % sodium dodecyl sufate (SDS) or 6 M guanidium chloride and proteases, and does not require fixation or staining of cells and tissues (Cubbitt et al. , 1995; Ward and Bokman, 1982; Tsien, 1998). Modified forms of GFP were developed, that are more efficient in folding, generate increased fluorescence, and are subject to decreased photobleaching (Cubbitt et al., 1995; Crameri et al., 1996; Cormack et al., 1996). Most of these GFP variants (SGFP, yEGFP, and EGFP) contain a Ser-65 to Thr amino acid substitution (S65T) that causes a red shift from an excitation maximum of 395 nm to a maximum of 488 nm (Lorang et al. , 2001). Synthetic GF P (SGFP) contains the S65T mutation as well as plant-optimized codon usage that deletes a cryptic intron splice site reported to reduce GFP expression in Arabidopsis (Haseloff et al. , 1997). The yeast enhanced GF P (yEGF P) contains the S65T mutation and codon usage optimized specifically for Candida albicans (Cormack etal., 1997). The enhanced GFP (EGFP) contains the double amino acid substitutions, Phe-64 to Leu, and Ser—65 to Thr, and 190 silent base mutations corresponding to human codon usage preferences (Heim et al. , 1995; Cormack et al., 1996; Yang et al. , 1996; Stauber et al., 1998; Tsien, 1998). Consequently, EGFP fluoresces 35-fold more brightly than wt GFP. 25 Utilization of GFP in Aspeggilli GFP expression has been reported in A. nidulans (Suelmann et al. , 1997; Femandez-Abalos et al. , 1998; Suelmann and Fischer, 2000; Tavoularis et al. , 2001; Bok and Keller, 2004; Maggio-Hall and Keller, 2004), A. niger (Santerre Henriksen et al. , 1999; Gordon et al., 2000), A. oryzae (Ohneda et al. , 2002; Shoji et al. , 2006), and A. flavus (Du et al. , 1999). The 3gp variants driven by common promoters such as ach and gpd were expressed and targeted to the endoplasmic reticulum (ER) and mitochondria in A. nidulans (F emandez-Abalos et al., 1998; Suelmann and Fischer, 2000). Also, nuclear- targeted SGFP allowed real-time visualization of nuclear migration and mitosis in A. nidulans (Suelmann et al. , 1997; Femandez-Abalos et al., 1998; Bok and Keller, 2004). The sflps fused to a proline transporter were used to monitor plasma membrane localization of the fusion protein in A. nidulans (Tavoularis et al. , 2001). Also, GFP was used to monitor localization of mitochondria and peroxisomal proteins (Maggio-Hall and Keller, 2004). The sgfia under the control of the glucoamylase promoter was expressed in A. niger (Santerre Henriksen et al., 1999). The sgfps fused to glucoarnylases were used to monitor protein secretion in A. niger (Gordon et al. , 2000). The egfp fused to vacuolar carboxypeptidase Y (CPY) or vacuolar syntaxin was expressed in A. oryzae to study vacuolar morphology and functions, and vacuolar membrane dynamics (Ohneda et al. , 2002; Shoji er al., 2006). The egjp fused to a nuclear localization signal and driven by the human cytomegalovirus (cmv) promoter produced fluorescence sufficient for detection of A. flavus on a UV. light box without additional filters (Du et a1. , 1999). Therefore, we chose EGFP for our studies because it shows the highest fluorescence among the various modified GFPs and was expressed in A. flavus. 26 VACUOLES The fungal vacuole l. The role of the fungal vacuole The fungal vacuole is often described as an organelle that is functionally analogous to the mammalian lysosome and the vacuole of plant cells (Klionsky er al. , 1990; Kucharczyk and Rytka, 2001). As the main degradative site in the cell, the vacuole contains a variety of hydrolytic enzymes required for macromolecular degradation, including endo- and exoproteases, ribonucleases, polyphosphatases, or-mannosidase, trehalase, and alkaline phosphatase. The vacuolar proton pumping ATPase (vacuolar H+- ATPase) generates and maintains the acidic pH (pH 5-6) of this compartment. The vacuole also serves as a storage site for certain cellular nutrients or metabolites such as amino acids (primarily basic amino acids), purines, polyamines, S-adenosyl-L- methionine, and polyphosphates (Kucharczyk and Rytka, 2001). It also functions as a . . . + 2+ 2+ + 2+ reservorr for mono- and divalent cations such as K , Ca , Mg , an , and Co for pH and osmoregulation, and detoxification of the cytosol due to vacuolar sequestration of the toxic ions (Klionsky et al. , 1990). The electrochemical potential generated by the H+- ATPase on the vacuolar membrane drives the transport of amino acids and ions to the vacuole. 2. Vacuolar protein transport pathways in yeast In yeast cells it has been reported that there are several different vacuolar transport 27 pathways such as the CPY (carboxypeptidase Y) pathway, the ALP (alkaline phosphatase) pathway, the Cvt (cytoplasm-to-vacuole targeting) pathway as well as the autophagy and endocytosis pathways (Figure 1.6) (Kucharczyk and Rytka, 2001). In the CPY pathway newly synthesized vacuole resident proteases such as carboxypeptidase Y and proteinase A (PrA) are transported to the endoplasmic reticulum (ER) lumen or membrane and pass through the Golgi apparatus from the ER along the early stage of the secretory pathway. These enzymes are then sorted from the secretory proteins in the late Golgi and delivered to the multivesicular body (MVB), also known as the prevacuolar compartment (PVC) or late endosome. Plasma membrane proteins from the cell surface are also targeted to the vacuole by endocytosis. In the ALP pathway vacuolar proteins like alkaline phosphatase bypass the MVB to be transported directly to the vacuole. In the Cvt pathway vacuolar hydrolases such as arninopeptidase I (API) and a-mannosidase are directly delivered to the vacuole from the cytoplasm constitutively under nutrient—rich and -depleted conditions. Under nitrogen or carbon starvation conditions cytoplasmic proteins and organelles are nonselectively targeted to the vacuole by autophagy for degradation and turnover of their constituent components by vacuole resident hydrolases. The yeast vacuolar transport pathways are vesicle-mediated (Kucharczyk and Rytka, 2001). Small vesicle buds carrying cargo proteins are pinched off from the donor organelle membrane and then the transport vesicles deliver the cargo proteins to the target organelles. The protein coat on vesicles, which contains the coatomer protein (COP) (COP I or COP H) and the clathrin coat, together with adapter proteins provides selective sorting of cargo proteins. The delivery of cargo proteins by the vesicles to the target organelle is accomplished by a membrane recognition mechanism that includes 28 Figure 1.6. Vacuolar transport pathways in yeast. Vacuolar proteins transported by the carboxypeptidase Y (CPY) and alkaline phosphatase (ALP) pathways are sorted into vesicles at the late Golgi. CPY is transported to the vacuole via the multivesicular body (MVB), but ALP bypasses this structure and travels directly to the vacuole. Proteins endocytosed from the cell surface are delivered to the vacuole via the MVB. Vacuolar proteins transported by the cytoplasm-to-vacuole targeting (Cvt) pathway and autophagy are directly transported to the vacuole from the cytoplasm. (Adapted from Kucharczyk and Rytka, 2001) 29 tethering, followed by docking and fusion (Gupta and Heath, 2002). Vesicles are targeted to fusion sites via the action of the associated GTP-bound Rab, which recruits tethering factors and docking factors; these factors recognize and bind to a target specifier on the target membrane, thereby bringing the vesicle closer to the membrane. The v-SNARE (vesicle-synaptobrevin-related receptors) protein anchored on the vesicle membrane is paired with its cognate t-SNARE (target-syntaxin-related receptors) protein anchored on the target membrane and the formation of a stable four-helix bundle promotes mixing of the lipid bilayers at the membrane fusion stage. 3. Vacuolar sorting sequences in vacuolar proteins It was reported that the vacuolar targeting signal in fungal vacuolar proteins is contained within the polypeptide chain while in some mammalian lysosomal proteins targeting is mediated by N-linked glycosylation by mannose 6-phosphate (Klionsky et al. , 1990). In the CPY pathway the soluble carboxypeptidase (CPY) is encoded by the PRC 1 gene (Klionsky et al., 1990). CPY is synthesized as an inactive precursor, preproCPY containing an N-terminal signal sequence that directs it into the lumen of ER. In the ER the signal sequence (N -terrninal 20 amino acids) is cleaved and proCPY undergoes N- linked glycosylation resulting in the intermediate ER form, 67 kDa p1CPY. The p1CPY is then transported to the Golgi complex by vesicles. In the Golgi, p1CPY undergoes sequential oligosaccharide modifications resulting in the Golgi modified form, 69 kDa p2CPY. In the late Golgi, the vacuolar sorting sequence (QRPL, amino acid residues 24 to 27) in the propeptide region (amino acid residues 21 to 111) of p2CPY directs the protein to vacuoles via the MVB (Rothman et al., 1989; Valls et al. , 1990). Just before or 30 upon arrival in the vacuole, the propeptide is cleaved from p2CPY to produce the active mature form, 61 kDa mCPY, by vacuolar proteases (proteinase A and B). Cleavage occurs after dissociation of the protein from a sorting receptor in response to the acidic pH-induced affinity change between them. Another vacuolar sorting signal (NPFXD) in plasma membrane proteins of yeast directs the protein to vacuoles by endocytosis (Tan et al. , 1996). Also, monoubiquiti- nation of plasma membrane proteins targets the proteins to vacuoles via the MVB by endocytosis for degradation while polyubiquitination of cytosolic, ER, and nuclear proteins directs the proteins to proteasomes for degradation (Hershko et al. , 2000; Hicke, 2001). Ubiquitin is removed from cargo proteins prior to their entry into the MVB. Alkaline phosphatase (ALP), an integral vacuolar membrane protein, is encoded by the PH08 gene. ALP is synthesized as an inactive zymogen and undergoes proteolytic cleavage of the propeptide similar to CPY immediately prior to or upon arrival in the vacuole. ALP also initiates vacuolar trafficking at the ER and is then transported to the Golgi. It bypasses the MVB and is delivered directly to the vacuole. Unlike CPY, however, the transmembrane domain of ALP functions as an internal uncleaved signal sequence for translocation into the ER. ALP also differs from CPY in that its vacuolar sorting signal (N -terrninal 52 amino acids) is not contained in the C-terminal propeptide segment that is removed from the protein just before or upon delivery to the vacuole (Klionsky and Emr, 1990). Soluble aminopeptidase 1 (API) is encoded by the LAP4 (APEI) gene. API is synthesized on free ribosomes in the cytoplasm as an inactive zymogen, pro-API containing the vacuolar sorting signal in the N-terminal region (amino acid residues 1 to 31 45). This region is predicted to form two a-helices separated by a B-turn and the first amphipathic helix (N -terminal 20 amino acids) contains vacuolar sorting information for the enzyme (Oda et al., 1996). The newly synthesized pro-API (61 kDa) in the cytoplasm forms a homododecarner. This Cvt complex is packaged into double membrane-bound Cvt vesicles under conditions that promote active growth although the complex is enclosed by autophagosomes under starvation conditions (Baba et al. , 1997). The Cvt vesicles fuse with the vacuole to release single membrane-bound Cvt bodies into the vacuolar lumen. The pro-API is released into the vacuolar lumen by the lysis of the Cvt body and cleaved into the mature enzyme (50 kDa) by vacuolar proteases like proteinase B. Autophage non-specifically sequesters cytoplasmic components and organelles into large double membrane-bound vesicles known as autophagosomes under starvation conditions; these deliver the cargo to the vacuoles for degradation and recycling. Like Cvt vesicles, the outer autophagosomal membrane firses with the vacuole to release single membrane-bound autophagic bodies into the vacuolar lumen, resulting in the breakdown of the autophagic bodies by vacuolar hydrolases like proteinase B. The plant vacuole 1. The role of plant vacuoles Plant vacuoles contain ions, sugars, amino acids, organic acids, and diverse proteins such as hydrolases, lectins, enzyme inhibitors, and storage proteins (Wink 1993). Vacuoles also contain secondary metabolic products such as alkaloids, glycosides and glutathione conjugates and some of these are considered to be chemical defense compounds against wounding or infection by herbivores or microorganisms (N euhaus and 32 Rogers, 1998; Wink 1993). Because many secondary metabolites are also toxic to the plant cells themselves, they must be stored in separate compartments like vacuoles. The vacuole functions as a turgo pressure regulator. However, the function of the plant vacuole varies greatly depending on cell type and developmental stage. In plant cells the vacuolar ATPase and pyrophosphatase generate and maintain the acidic pH (pH 5-6.5) of this compartment. Small lipophilic compounds cross the tonoplast by simple diffusion whereas hydrophilic compounds such as amino acids, organic acids, ions, and polar secondary metabolites are taken up by transproters or perrneases through utilization of the proton motive force generated by the ATPase and pyrophosphatase (Wink 1993). The highly concentrated compounds can be maintained in the vacuole by several trapping mechanisms such as protonation, crystallization, conjugation, conformational changes, and complex formation (Wink 1993). 2. Vacuolar protein transport pathways in plant cells In contrast to yeast or mammals, plants carry at least two types of vacuoles; lytic vacuoles (LV), equivalent to the yeast vacuole or the mammalian lysosome, and protein- storage vacuoles (PSV) to prevent exposure of storage proteins to acidified vacuoles carrying active hydrolases (Vitale and Raikhel, 1999) (Figure 1.7). These two types of vacuoles fuse into a single large central vacuole when degradation of the storage proteins is required. The degraded amino acids serve as a nitrogen source during seedling growth. Vacuolar protein transport from the Golgi complex to the LV is mediated by clathrin- coated vesicles while transport from the Golgi to PSV is mediated by dense vesicles or precursor-accumulating vesicles. Also, unlike yeast, the prevacuolar compartment (PVC) 33 Protein-storage L © fl vacuole © \ Fused central vacuole / Lyfic vacuole CCV Prevacuolar compartment Figure 1.7. Vacuolar transport pathways in plant cells. In some plant cell types protein- storage and lytic vacuoles co-exist. Vacuolar proteins are transported to the vacuole by one of three routes. (a) Via precursor-accumulating vesicles (PAC) to the protein-storage vacuole (PSV). (b) Via dense vesicles (DV) and multivesicular bodies (MV B) to the protein-storage vacuole (PSV). (c) Via clathrin-coated vesicles (CCV) to the lytic vacuoles (LV). In mature cells the PSV and LV fuse into a single large central vacuole when conditions favor this event. (Adapted from Bassham and Raikhel, 2000) 34 in plants is involved with the LV whereas the mutivesicular body (MVB) is involved with the PSV. In another type of vacuolar protein transport pathway, protein aggregates in the ER of pumpkin cells bypass the Golgi to be transported to the PSV by precursor- accumulating vesicles (Hara—Nishimura et’ al., 1998). Like the Cvt pathway in yeast, [3- amylase, soybean lipoxygenase, and late embryogenesis—abundant proteins such as barley LEA3 protein are directly transported to vacuoles from the cytoplasm (N akamura and Matsuoka, 1993; Neuhaus and Rogers, 1998). 3. Vacuolar sorting sequences in vacuolar proteins It was reported that the vacuolar sorting signal in plant vacuolar proteins is contained within the polypeptide chain similar to the fungal vacuolar proteins discussed above (Chrispeels and Raikhel, 1992; Vitale and Raikhel, 1999). All of the identified N- terminal or C-terminal vacuolar sorting sequences are contained in the propeptide region, which are cleaved in the vacuole or the prevacuolar compartment (PVC) (Vitale and Raikhel, 1999). Uncleaved internal sorting sequences are contained within mature proteins. Vacuolar sorting signals in the N-terminal propeptides contain a NPIR or lex (x can be any amino acids) conserved sequence as in sweet potato sporanin and barley aleurain (Nakamura and Matsuoka, 1993; Neuhaus and Rogers, 1998). Sorting sequences in C-terminal propeptides are enriched in hydrophobic amino acids; examples include barley lectin, common bean phaseolin, and Brazil nut 2S albumin (Vitale and Raikhel, 1999). It was proposed that the NPIR-like conserved sequence in the N-terminal sorting signal directs the protein to the LV while less defined sequences in the C-terrninal or internal sorting signal direct the protein to the PSV (V itale and Raikhel, 1999; Neuhaus 35 and Rogers, 1998). Vacuolar sorting sequences of phytohemagglu-tinin, a common bean lectin, are contained in uncleaved internal sequences (LQRD, amino acid residues 38 to 41, and another region, amino acid residues 84 to 113) within the mature protein (Tague er al., 1990; Schaewen and Chrispeels, 1993). Therefore, multiple vacuolar targeting determinants may exist in some plant vacuolar proteins. Interestingly, the identified vacuolar sorting sequence (LORD) of phytohemagglutinin (PHA) is similar to the sequence ([L]QRPL) of carboxypeptidase Y (CPY) in yeast although the PHA sequence is located within the mature protein while the CPY sequence is located in the propeptide (Chrispeels and Raikhel, 1992). 36 CHAPTER 2 UTILIZATION OF EGF P AS A REPORTER SYSTEM IN ASPER GILL US PARASI T I C US ABSTRACT To monitor sub-cellular localization of aflatoxin enzymes in real time, we developed an EGFP reporter system (pAPGFPVNB), which contains the ver-I promoter fused to the egip gene and a selectable marker (niaD, encodes nitrate reductase). Transformation of pAPGFPVNB into A. parasiticus NR-l (niaD) resulted in 14 EGFP (+) transformants out of 3 12 transformants. The plasmid integration sites were analyzed by Southern hybridization and PCR. The plasmid was integrated into the ver-I A locus in all EGFP (+) transformants while it was integrated into the niaD locus in EGF P (-) transformants. A time-course experiment conducted on transformants carrying pAPGFPVNB showed that the expression pattern of the ver-I promoter-driven EGF P was similar to the wild-type ver-I promoter and paralleled AFB] production in aflatoxin- inducing media. Therefore, we concluded that EGFP was expressed by the ver-I promoter and can be used as a reporter gene for sub-cellular localization of aflatoxin enzymes in A. parasiticus. 37 INTRODUCTION Several reporters are available to monitor gene activity and protein localization in living organisms. These include B-galactosidase (Drahos et al. 1986), B-glucuronidase (GUS) (Roberts et al., 1989), chloramphenicol acetyltransferase (CAT) (Hagedom, 1994) and luciferase (LUC) (Prosser, 1994). However, these reporter systems have limited use with living cells because they require exogenous substrates or cofactors. On the contrary, green fluorescence protein (GF P) does not require substrates or cofactors because the detection of intracellular GFP requires only irradiation by near UV. or blue light (Errampalli et al. , 1999). Also, GFP can be detected without cell disruption and monitored in situ and in real time. Therefore, GFP is useful as a marker for gene expression and as a tag in studying protein localization in a variety of organisms. However, it was reported that wild-type gr}; is poorly expressed in many fungi (F emandez-Abalos et al., 1998). Therefore, in this study we determined if EGFP could be used as a reporter gene for sub-cellular localization under the control of the ver-I promoter in the filamentous fungus A. parasiticus; this protein showed the highest fluorescence among the modified GFPs. Choosing the optimtun size of the promoter is an issue. In previous studies using a GUS reporter, nor-1 promoter activity was decreased by 3 fold in a 332 bp promoter compared to activity in a 1.2 kb promoter. However, the activity was similar between 664 bp and 1.2 kb promoters (Miller, 2003). These observations suggested that expression of niiA carried on a 7.4 kb XhoI/Sall niaD fragment fused to the promoter in the opposite direction might negatively affect nor-1 promoter activity (Figure 2.1). 38 lkb S KXXh X HK BE B PK EEEB K B HS W <— > nI'IA niaD Figure 2.1. Restriction endonuclease map of the A. parasiticus 8.2 kb SalI fragment that contains complete niaD and partial niiA genes. Open boxes represent introns. Arrows indicate the location and direction of transcription of the nitrate reductase gene (niaD) and the nitrite reductase gene (niiA), respectively. Abbreviations for the restriction enzymes are as follows; B, BamHI; E, EcoRI; H, HindIII; K, Kpnl; P, Pstl; S, SalI; X, Xbal; Xh, Xhol. (Adapted from Chang et al. , 1996) 39 Therefore, in this study the activities of 0.6 kb and 1.1 kb ver-I promoter fragments were compared using an EGFP reporter gene (Figure 2.2). MATERIALS AND METHODS Strains and culture conditions Escherichia coli DHSor F’e [F’/endA1 hstI 70]; mg“) supE44 (hi-I recAI gyrA (Nalr)relA1 A(lacZYA-argF)u]69: (m80AlacZM15)] (Gibco BRL, Life Technologies Inc, Rockville, MD) was used to amplify plasmid DNA using standard procedures (Ausubel et al., 2003). A. parasiticus NR-l (niaD, encodes nitrate reductase), derived from A. parasiticus NRRL 5862 (SU-l; ATCC 56775) by spontaneous mutation using potassium chlorate selection (Homg et al., 1990), was used as the recipient strain for reporter plasmids containing the niaD selectable marker. A. parasiticus strains used in this study are listed in Table 2.1. E. coli was cultured in LB medium (Luria-Bertani; 1 % tryptone, 0.5 % yeast extract, 1 % sodimn chloride) for competent cell preparation and plasmid or cosmid isolation (Maniatis et a1. , 1989). A. parasiticus was cultured in yeast extract-sucrose liquid medium (YES; 2 % yeast extract, 6 % sucrose, pH 5.8; batch fermentation) at 30 0C in the dark with shaking at 150 rpm for genomic DNA isolation, total protein extraction, measurement of mycelial dry weight, analyses of EGFP fluorescence and aflatoxin concentration, and confocal laser scanning microscopy (CLSM) as described 40 Asc I 418 Pst I 1010 S011 1908 Sal I 2188 Fse I 2518 Pst I 11239 Sal I 11227 O. ‘ a e.“ r'.’ . I I' "-ii '34-. Amp ver-I In... fig... te rmina to r r/ Not I 3244 pAPGFPVNBegfi" (13533 bp) ,4. ver-I promoter Pac I 3859 Sal I/Xho I 3875 Pst I 7027 Figure 2.2. Restriction endonuclease map of plasmid, pAPGFPVNB. The 0.6 kb ver-I promoter was fused in frame to the 0.7 kb egfi? coding region, followed by the 2.1 kb ver-I terminator. The 7.4 kb niaD fragment was inserted as a selectable marker for transformation of the recipient strain NR-l (niaD). 41 Table 2.1. Aspergillus parasiticus strains used in this study. Strain Genotypea Phenotype Source NR-l niaD AF (+) J. E. Linz NR-l (pNiaD-Al) - AF (+) This study NR-l (pNANG3) uidAzmor—I t AF (+) This study B1-2 ver-I p::egfp AF (+), EGF P (-) This study Bl-49 ver-I p::egfp AF (+), EGF P (+) This study B1-78 vet-1 pzzegfp AF (+), EGFP (+) This study B1-86 ver-I p::egfp AF (+), EGFP (+) This study B2-1 ver-I p::egfp AF (+), EGFP (-) This study B2-9 ver-I p::egfp AF (+), EGFP (+) This study B2-20 ver-I p::egfp AF (+), EGFP (+) This study B3-l ver-I p::egfp AF (+), EGFP (-) This study B3-15 ver-I pzzegfp AF (+), EGFP (+) This study B3-46 ver-I p::egfp AF (+), EGFP (+) This study B3-l01 ver-I p::egfp AF (+), EGF P (+) This study B3-105 ver-I p::egfp AF (+), EGF P (+) This study B3-120 ver-I p::egfp AF (+), EGFP (+) This study B3-146 ver-I pizeg/p AF (+), EGFP (+) This study B3-160 ver-I p::egfp AF (+), EGFP (+) This study B3-186 ver-I p::egfp AF (+), EGFP (+) This study B3-194 ver-I p::egfp AF (+), EGFP (+) This study LBl-l ver-I pzzegfp AF (+), EGFP (-) This study LBl-ll ver-I pziegfp AF (+), EGFP (+) This study LB1-15 ver-I p::egfp AF (+), EGFP (+) This study LB1-46 ver-I p::egfp AF (+), EGFP (+) This study LB1-94 ver-I p::egfp AF (+), EGFP (+) This study 42 Table 2.1. (cont’d). Strain Genotypea Phenotype Source LB1-101 ver-I pzzegfp AF (+), EGFP (+) This study LBl-112 ver-I p::egfp AF (+), EGFP (+) This study LBl-119 ver-I p::egfp AF (+), EGFP (+) This study LB1-148 ver-I p::egfp AF (+), EGFP (+) This study LB1-170 ver-I p::egfp AF (+), EGFP (+) This study LB1-211 ver-I p::egfp AF (+), EGFP (+) This study a ver-I p and nor-I t represent ver-I promoter and nor-I terminater, respectively. :: represents gene fusions. 43 previously (Liang et al. , 1997). YES agar (1.5 % agar) was used for screening of EGF P producing fungal transformants and slide culture with a Nikon Eclipse E600 or Labophot fluorescence microscope (Nikon Inc., Melville, NY). Either YES or YEP (2 % yeast extract, 6 % peptone, pH 5.8) agar was used for slide culture which was observed using a Zeiss LSM 5 Pascal or Zeiss 510 Meta confocal laser scanning microscope (CLSM) (Carl Zeiss Inc., Germany). Either Czapek-Dox medium (CZ; Difco Laboratories, Detroit, MI) supplemented with 1 % peptone or YES medium was used for growth of the recipient strain NR-l for transformation. CZ agar supplemented with 20 % sucrose as an osmotic stabilizer was used as a selective medium for fungal transformation. Either potato dextrose agar (PDA; Difco Laboratories, Detroit, MI) or CZ agar was used for spore preparation (Skory et al. , 1992). Construction of pAPGFPVNB and pAPGFPV1.1NB Three closely related plasmid vectors, called pAPGFPVNBl, 2, and 3 (Figure 2.2 and 2.3), differing at the junction site (Natl) between the ver-I promoter and egfio coding region, were constructed using ver-I promoter and terminator fragments, an egjp gene fragment, and pNANG-3 (Figure 2.4) (Miller, 2003) as a plasmid backbone. Unique 8 base-cutting restriction enzyme sites allowed independent replacement of the promoter and terminator fragments. The 0.6 kb ver-I promoter and 2.1 kb ver-I terminator fragments were generated by PCR with Pfu DNA polymerase (Stratagene, La J olla, CA), appropriate primers, and cosmid NorA (Figure 2.5) (Liang et al., 1996) as a template using standard procedures (Maniatis et al., 1989). The primers contained restriction enzyme sites to facilitate cloning (Table 2.2). The 0.7 kb egfj) gene was generated by 44 A. pAPGFPVNBl 1 2 DNA ATG-Tgc-ggc-cgc-GTG ver-I promoter Natl egfp Amino acid Met-Cys- Gly-Arg-Val B. pAPGFPVNB; 1 2 3 4 5 6 DNA ATG-GTG-Agc-ggc-cgc-GAG Natl Amino acid Met- Val- Ser- Gly-Arg- Glu Original DNA AGC-AAG-GGC-GAG Original Amino acid Met- Val- Ser- Lys-Gly- Glu C. pAPGFPVNB3 1 2 DNA gc-ggc-cgc-ATG-GTG Natl Amino acid Met- Val Original DNA CC-GTC-AGC-ATG-GTG Figure 2.3. Comparison of selected sequences in pAPGFPVNBl, 2, and 3. (A) pAPGFPVNB]. (B) pAPGFPVNBZ. (C) pAPGFPVNB3. These plasmids differ at the junction site (Natl) between the ver-I promoter and egfp coding region. The start codon was designated amino acid residue 1 followed by the subsequent residues. Abbreviations for amino acids are as follows; Arg, arginine; Cys, cystein; Glu, glutamic acid; Gly, glycine; Lys, lysine; Met, methionine; Ser, serine; Val, valine. 45 Figure 2.4. Restriction endonuclease maps of plasmids, pNEB-Nl and pNANG-3. (A) We inserted a Natl linker into pNEB-Nl that replaces the BamHI site in pNEBl93 (New England Biolabs, Beverly, MA). (B) pNANG-3 was derived from pNEB-Nl and carries the niaD selectable marker and a small part of the nor-1 coding region (10 amino acids) fused to the B-glucuronidase (GUS) coding region (uidA); this is in turn fused to the 2 kb nor-1 terminator. 46 Asc I 418 Not I 426 Far: I 432 Sal I 447 Amp. pNEB-Nl (2700 bp) Asc I 418 Amp' . uidA-nor-I . ~ terminator ' pNANG-3 (13900 bp) [’31 I 11612 Sal I 11600 Not I 4226 Pac I 4232 Sal I/Xho I 4248 Pst I 7400 47 Figure 2.5. Restriction endonuclease map of cosmid NorA. Cosmid NorA was cloned from genomic DNA of the wild-type Aspergillus parasiticus SU-l. In addition to the pyrG selectable marker (encodes orotidine monophosphate decarboxylase), this cosmid carries kaA, nar—I,fas-2,fas-1 , aflR, aflJ, adhA, estA, narA, ver-I A, and verA coding regions fused to the cos site which is used to clone large DNA fragments. Arrows indicate the direction of transcription of the genes. Abbreviations for the restriction enzymes are as follows; B, BamHI; E, EcaRI; H, HindIII; Sac, Sacl; Sal, SalI; X, Xbal. 48 €5.N H .95... may V5. V TS: VS: Sue V53 as S? 73‘ N33 7.3: V33 DEE wmmmmmwm. I § w N m cum I: m Em m m mx w m m m m H m x m 8.0m SEN 6:.N Agni. axmé savages.— .Emfi 5:; 93.9 scam €56 <82 2880 Table 2.2. Primer sequences used in this study. Primer‘ Sequenceb Restriction enzyme site, annealing temp (°C) ver-I A 5’ CTCTTAATTAACAAATACACCTACTACACGAC 3’ PacI, promoter (B1)- F 55 ver-I A 5’ CTCGCGGCCGCACATGCTGACGGGATCGTG 3’ Natl, promoter (B1)- R 55 ver-I A 5’ CTCTTAATTAACAAATACACCTACTACACGAC 3’ PacI, promoter (B2)- F 55 ver-I A 5’ ACGGCGGCCGCTCACCATGCTGACGGGATCGTG 3’ Natl, promoter (B2)- R 55 ver-I A 5’ CTCTTAATTAACAAATACACCTACTACACGAC 3’ Fuel, promoter (B3)- F 50 ver-I A 5’ ATGCGGCCGCGATCGTGTATGGTAGAGATTT 3’ Not], promoter (B3)- R 50 ver-I A 5’ GTCGGCCGGCQTAAACCTTCACAGCTATATACTCG 3’ FseI, termmator- F ver-I A terminator- R egfp gene (Bl)- F egfp gene (Bl)- R efi'p gene (B2)- F 55 5’ GCCGGCGCGCCTGCTGATGGTGGGAAGAG 3’ Ascl, 55 5’ ATAGCGGCCGCGTGAGCAAGGGCGAGGAG 3’ Natl, 55 5’ GTCGGCCGGCCTTTACTTGTACAGCTCGTCCAT 3’ FseI, 55 5’ ATGGTGAGCGGCCGCGAGGAGCTG 3’ Natl, 55 50 Table 2.2. (cont’d). Primera Sequenceb Restriction enzyme site, annealing temp (0C) egfp gene 5’ GTCGGCCGGCCTTTACTTGTACAGCTCGTCCAT 3’ Fsel, (B2)- R 55 egfp gene 5’ CCGGGCGGCCGCATGGTGAGCAAGGGCGAG 3’ Natl, (B3)- F 55 egfp gene 5’ GTCGGCCGGCCTTTACTTGTACAGCTCGTCCAT 3’ Kid, (B3)- R 55 ver-I A 5’ CGGGTTAATTAAGATGCCGAACCATTTGAC 3’ Pacl, promoter (1.1)- F 55 ver-I A 5’ CTCGCGGCCGCACATGCTGACGGGATCGTG 3’ Natl, promoter (1.1)- R 55 egfp gene 5’ GGC-AAC-TAC-AAG-ACC-CGC-G 3’ None, (3’ integrant)- F 68 Downstream of 5’ AGC-CAC-CGT-GAG-CGT-CC 3’ None, ver-I A terminator 68 (3’ integrant)— R ver-I A gene 5’ ATC-CTG-ACC-AGC-TCT-AAC-ACC-G 3’ None, (3’ control)- F 68 Upstream of 5’ CAG-AGG-CTC-AGT-CAC-TTG-TTC 3’ None, ver-I A promoter (B3) 63 (5’ integrant)- F eflp gene 5’ TGC-GCT-CCT-GGA-CGT-AG 3’ None, (5’ integrant)- R 63 51 Table 2.2. (cont’d). Primera Sequenceb Restriction enzyme site, annealing temp (0C) ver-I A gene 5’ CAA-TTC-CAG-CGT-TCG-ATG 3’ None, (5’ control)- R 63 Upstream of 5’ GGT-ACT-GAG-CGA-GGA-GGA-A 3’ None, ver-I A promoter (1.1) 63 (5’ integrant)- F a F represents forward primers and R represents reverse primers. B1, B2, B3, and 1.1 represent pAPGFPVNB], pAPGFPVNBZ, pAPGFPVNB3, and pAPGFPVl.lNBl, respectively. b Underlined sequences show the position of the restriction enzyme sites. 52 PCR with Pfu DNA polymerase (Stratagene, LaJolla, CA) and appropriate primers using pEGFP-Nl (Clonetech Laboratories, Palo Alto, CA) as a template (Figure 2.6). The primers contained restriction enzyme sites to facilitate cloning (Table 2.2). PCR was performed in a Gene Amp PCR System 2400 thermal cycler (Perkin—Elmer life sciences Inc., Boston, MA). The reaction conditions for thermal cycling depended on the designed primers and the target size as follows: 94 0C for 5 min followed by 25 cycles of denaturation at 94 0C for 1 min, annealing for 1 min (Table 2.2), and extension at 72 0C for time in min that depended on PCR fragment sizes (2 min/1 kb). The reactions were completed with an extension at 72 0C for 10 min. The PCR fragments were digested with appropriate restriction enzymes after purification from an agarose gel using a Qiaex 11 gel extraction kit (Qiagen Inc.,Valencia, CA) and cloned into pNEB-Nl (Figure 2.4) (Miller, 2003) cutwith the same enzymes, resulting in pGFPV. DNA fragments containing the ver-I promoter, the egfia gene, and the ver-I terminator were subcloned from pGFPV into pNANG3.0 (Figure 2.4) cut with Pacl and Ascl, resulting in pAPGFPVNB]. The plasmids pAPGFPVNB2 and 3 were constructed by replacement of the ver-I promoter and the egfi) gene fragments in pAPGFPVNB] with modified ver-I promoter and the egfp gene fragments, which generated unique Natl junction sites as shown in Figure 2.3. Also, pAPGFPV1.1NB1 was constructed using pAPGFPVNB] and the 1.1 kb ver-I promoter PCR fragment in the same way. 53 Ase I 8 ApaL I 4362 SnaB I 341 P(:Mv IE pUC ori Ec00109I 3856 MCS .. HSV TK WA pEGFP-Nl (4700 bp) eat?) SV40 ‘\/Ber I 1389 Kan'l Neor Po'y A Not I 1402 Xba I 1412 SV40 M n m A II 1640 PSV40e p fl - Dra 111 1874 Stu l 2579 Figure 2.6. Restriction endonuclease map of plasmid, pEGFP-Nl (Clonetech Laboratories, Palo Alto, CA). 54 Transformation of E. coli and isolation of plasmids and cosmids The preparation and transformation of E. coli competent cells were conducted by a calcium chloride method (Ausubel et al. , 2003). The isolation of plasmids and cosmids was performed by an alkaline lysis method using a Wizard DNA purification kit (Promega Corporation, Madison, WI) or CsCl/ethidium bromide equilibrium centrifugation (Maniatis et al. , 1989). Transformation of A. parasiticus and screening for EGFP positive transformants Transformation of A. parasiticus NR-l with pAPGFPVNB or pAPGFPV1.1 NBl was performed by a polyethylene glycol method (Oakley et al. , 1987) with minor modifications as described previously (Skory et al. , 1990). Fifteen h afier inoculation of 108 conidia, protoplasts were generated by digestion of mycelia with lysing enzyme (25 mg/ml; Sigma Chemical Co., St. Louis, MO) and driselase (50 mg/ml; Sigma Chemical Co., St. Louis, MO). Transformants were selected on CZ agar supplemented with 20 % sucrose as an osmotic stabilizer. Transformants were transferred onto YES agar and then screened for EGFP (+) clones under a Nikon Eclipse E600 fluorescence microscope (Nikon Inc., Melville, NY) using a 450-490 nm excitation/ 515 nm emission filter after 2 day incubation at 30 0C. Conventional fluorescence microscopy Slide culture was performed by the method according to Harris (1986) with minor modifications as described previously (Liang, 1996) (Figure 2.7). Approximately 5x105 55 Figure 2.7. Slide culture diagram. Approximately 5x105 conidia were inoculated around the edge of the top surface (because they need oxygen for growth) of YES agar blocks (1 cm2) on sterile coverslips which were placed onto water agar plates. This was followed by placement of a second coverslip on top of the YES agar blocks. The curved arrow indicates the location of fungal spore inoculation. The cube and rhombuses indicate an agar block and coverslips, respectively. 56 conidia were inoculated around the edge of the top surface of YES agar blocks (1 cm2) on sterile coverslips which were placed onto water agar plates. This was followed by placement of a second coverslip on top of the YES agar blocks. The plates were incubated at 30 0C in the dark for 2 days. Coverslips were removed and washed three times with phosphate—buffered saline (PBS) and observed using a Nikon Labophot fluorescence microscope (Nikon Inc., Melville, NY) with a 450-490 nm excitation/ 520 nm emission filter. Genomic DNA isolation from A. parasiticus Genomic DNAs were isolated by a phenol-chloroform method (Ausubel et al. , 2003) with minor modifications as described previously (Skory et a1. , 1990). Approximately 2x106 conidia were cultured in 100 ml of YES media at 30 0C in the dark with shaking at 150 rpm for 48 h and then a phenol-chloroform protocol was used to isolate genomic DNA from mycelia after filtration through Miracloth (Calbiochem, La Jolla, CA). Southern hybridization and PCR analyses Southern hybridization analyses were conducted using standard procedures (Maniatis et al., 1989). Approximately 10 ug of genomic DNAs cut with HincII were separated by agarose gel electrophoresis and then transferred onto a nylon transfer membrane (Nytran supercharge membrane, Schleicher and Schell Inc., Keene, NH) by capillary action. Radiolabeled DNA probes were generated with [or-32P]dCTP (Perkin- 57 Elmer life sciences Inc., Boston, MA), the Random Primers DNA Labeling System (Invitrogen, Carlsbad, CA), and the 0.6 kb ver—I A promoter fragment using a procedure provided by the manufacturer. After the final wash, the membranes were exposed to X- ray film (Eastman Kodak company, Rochester, NY) at —80 0C. PCR analyses were performed with genomic DNA and specific primers to 3’ or 5’ ver-I to confirm integration sites of the plasmids. lt/HindIII (lnvitrogen, Carlsbad, CA) was used as a molecular size marker. Time-course of EGF P expression Approximately 2x106 conidia were cultured in 100 ml of YES media at 30 0C in the dark with shaking at 150 rpm as described previously (Liang et al. , 1997). Flasks were removed at different time points after inoculation for total protein extraction and analyses of mycelial dry weight and aflatoxin concentration. Mycelia were harvested by filtration through Miracloth (Calbiochem, La Jolla, CA), frozen in liquid nitrogen, and stored at —80 0C. Total protein extraction Mycelia were ground in liquid nitrogen with a mortar and pestle, suspended in TET buffer (100 mM Tris-HCl [pH7.5], 2.5 mM EDTA, 5 mM DTT, 1 % Triton X-100) ' containing 1 mM phenylmethylsunfonyl fluoride (PMSF) (Sigma Chemical Co., St. Louis, MO) and 4 % proteinase cocktail (Sigma Chemical Co., St. Louis, MO), and centrifuged at 10,000x g for 15 min at 4 0C (V aldez-Taubas et al. , 2000, Tavoularis et al. , 58 2001). The supematants were used for fluorometric measurement. Protein concentration in the supernatant was determined by a modified Bradford assay using a commercial Protein Assay reagent (Bio-Rad Laboratories, Hercules, CA) (Bradford, 1976). Measurement of EGFP fluorescence Supernatants containing fungal proteins were used for EGFP fluorescence measurement. Samples were dispensed into FluoroNunc TM Maxisorp 96-microwell plates (Nunc, Roskilde, Denmark) and analyzed with a cytofluor II (Biosearch Co., Bedford, MA) using 470 nm excitation/ 510 nm emission filters. The fluorescence values were normalized against the total protein concentration in the samples and expressed as relative EGFP fluorescence units per ug of protein. Measurement of mycelial dry weight and aflatoxin concentration Dry weight was determined after complete drying of the harvested mycelia at 100 oC. The AFB] concentration in the filtrate was determined by direct competitive enzyme-linked immunosorbent assay (ELISA) with AFB] monoclonal antibodies (kindly provided by J. Pestka, Michigan State University) as described previously (Pestka, 1988). Confocal Laser Scanning Microscopy (CLSM) Slide culture was performed as described above (Liang, 1996). For non-aflatoxin- inducing media, YEP was used instead of YES. The plates were incubated at 30 0C in the dark. The coverslips were removed at different time points after inoculation. Fungal 59 vacuoles were then stained with FM 4-64 or 7-amino-4-chloromethylcoumarin (CMAC) (Ohneda et al., 2002; Shoji et al. , 2006). FM 4-64 and CMAC were used for staining vacuolar membranes and lumens, respectively. The coverslips were placed in YES media containing 8 uM FM 4-64 or 10 uM CMAC. For FM 4-64, coverslips were incubated at 30 °c for 10 min and washed with fresh media without the dye for 30 min. For CMAC, coverslips were incubated at 30 0C for 30 min and washed with fresh media without the dye at 37 0C for 30 min. Coverslips were observed using a Zeiss LSM 5 Pascal or Zeiss LSM 510 Meta confocal laser scanning microscope (CLSM) (Carl Zeiss Inc., Germany). All images were captured using a Zeiss Plan-APOCHROMAT (63x/1.40 Oil) objective. EGFP fluorescence (488 nm excitation/ 509 nm emission) was detected using a BP 505- 530 emission filter set under excitation with the 488 nm argon-ion laser line. FM 4-64 fluorescence (558 nm excitation/ 734 nm emission) was detected using a LP 650 emission filter set under excitation with the 633 nm helium-neon laser line. CMAC fluorescence (353 nm excitation/ 466 nm emission) was detected using a BP 420-480 emission filter set under excitation with the 405 nm diode laser line. For liquid culture, approximately 2x106 conidia were cultured in 100 ml of YES media at 30 0C in the dark with shaking at 150 rpm as described previously (Liang et al., 1997). Flasks were removed at 24 h after inoculation. Fungal vacuoles were then stained in eppendorf tubes with FM 4-64 and fungal mycelia were observed using the Zeiss LSM 5 Pascal confocal laser scanning microscope (CLSM) (Carl Zeiss Inc., Germany). 60 RESULTS Transformation of A. parasiticus NR-l with pAPGFPVNB and pAPGFPV1.1NB1, and screening for EGFP positive transformants Five to ten ug of pAPGFPVNB was transformed into 108 A. parasiticus NR-l protoplasts, generating 312 transformants. All transformants selected on CZ agar were screened for EGFP expression on YES agar under the fluorescence microscope (Figure 2.8). Fourteen EGFP (+) transformants were identified (4.5%). Three EGF P fluorescent transformants carrying pAPGFPVNB] were identified out of 100 transformants screened. Two transformants carrying pAPGFPVNB2 were identified out of 14 transformants. Nine transformants carrying pAPGFPVNB3 were identified out of 198 transformants. Transformation of pAPGF PV1.1NB1 into A. parasiticus NR-l resulted in 10 EGFP (+) transformants out of 21 5 transformants (4.7 %). Determination of integration sites of pAPGFPVNB and pAPGFPV1.1NBl within the chromosome In order to determine the integration sites of pAPGFPVNB and pAPGFPV1.1NB1, Southern hybridization and PCR analyses were performed. Each EGF P reporter plasmid could theoretically integrate into the chromosome by homologous recombination at five sites: niaD, 3’ ver-I terminator within the ver-I A or ver-I B locus, or 5’ ver-I promoter within the ver-I A or ver-I B locus (Figure 2.9). Southern hybridization analysis showed that all 5 EGF P (+) transformants with pAPGFPVNBl or 2 had the plasmids integrated at the 3’ ver-I terminator within the ver-I A or 61 Figure 2.8. Fluorescence microscopy of A. parasiticus expressing EGFP. Transformants were inoculated on YES agar plates or agar blocks on coverslips and incubated at 30 0C for 2 days and observed using the Nikon Eclipse E600 or Labophot fluorescence microscope. (A) Transforrnant carrying pNiaD-Al (negative control). (100 x) (B) Transforrnant carrying pAPGFPVNB2. (100 x) (C) Transforrnant carrying pAPGFPVNB3. EGF P fluorescence was observed in conidia, vesicles, and conidiophores. However, EGF P fluorescence was not detected in some conidia but only in vesicles and conidiophores. (100 x) (D) Transforrnant carrying pAPGFPV1.1NB. EGFP fluorescence was observed in vesicles and conidiophores. (100 x) (E) Transforrnant carrying pAPGFPVNB]. EGFP fluorescence was observed in conidia, vesicles, conidiophores, and substrate-level mycelia. (400 x) (F) Conidia of transfonnant carrying pAPGFPVNB]. (1000 x) Images in this dissertation are presented in color. 62 63 Figure 2.9. Schematic for Southern hybridization and PCR analyses of integration sites of pAPGFPVNB into the chromosome. (A) ver-I A locus. (B) 3’ ver-I A integration. (C) 5’ ver-I A integration. (D) ver-I B locus. (E) 3’ ver-l B integration. (F) 5’ ver-I B integration. (G) niaD locus. (H) niaD integration. Genomic DNA was digested with HincII and probed with the ver-I A promoter. The restriction enzyme sites, probes, and expected band sizes in Southern hybridization analyses are shown. The primer positions in PCR analyses are also shown. In case of Southern hybridization analysis, integration of the plasmid pAPGFPVl.1NBl into 3’ ver-I A results in 2.8 kb and 8.7 kb bands, its integration into 5’ ver-I A results in 3.7 kb and 7.7 kb bands, and its integration into niaD results in an 8.7 kb band. Abbreviations for the restriction enzyme sites and DNA fragments are as follows; H, Hincll; S, SalI; Xh, X7101; ver-I p, ver-I promoter; ver-I t, ver-I terminator. 64 A. ver-I A locus ver-I p ver-I A ver-I t | (probe) 1 I 2.8 kb F B. 3’ ver-I A integration PCR H 4-1 Il-I HHll-I H H [H H vet-1 p ver-I A ver-I tamp niaD ver-I p egfp ver-I t l___l | l 2.8 kb I 8.2 kb I C. 5’ ver-I A integration PCR Ib‘l Ill HE HIl-I HHHIH H ver-I p egfp ver-I t amp niaD ver-I p ver-I A ver-I t I | I 3.7 kb I I 7.2 kb I 65 Figure 2.9. (cont’d) D. ver-I B locus 7’ m ver-I p ver-I B ver-I t (probe) 1.2 kb E. 3’ ver-I B integration III HHHH HIII H1] Il-I ver-I p ver-I B ver-I tamp niaD ver-I p egfii ver-I t l l 1.2 kb 8.2 kb F. 5’ ver-I B integration H H H HH HHH H I 13.11331 31.1.1 ver-I p egfp ver-I t amp niaD ver-I p ver-I B ver-l t l l l U 1.8 kb 7.2 kb 66 Figure 2.9. (cont’d) G. niaD locus niaD H. niaD integration Xh HS, H HH HS. H -I——( ‘ ' ‘ “Ii—Ml ’ ' ' I niaD amp ver-I t egfp ver-I p niaD I I ' 8.2 kb j 67 ver-I B locus, or niaD locus (isolates 81-49, 31-78, B1-86, 82-29, and B2-20) (Figure 2.10 A). PCR analysis showed that the 5 EGF P (+) transformants produced a 2.6 kb band with the eg/p primer and the 3’ ver-I A primer (Figure 2.9 and Table 2.2), indicating integration of the plasmids into 3’ ver-I terminator at the ver-I A locus (Figure 2.10 B). When taken together, Southern hybridization and PCR analyses confirmed that in all 5 EGFP (+) transformants carrying pAPGFPVNB] or 2, the plasmid was integrated into the ver-I terminator at the ver-I A locus in the chromosome (Figure 2.9 and 2.10 A and B). The data also suggested that the negative control pNiaD-Al was integrated into the niaD locus by double-crossover (gene replacement of niaD) and that the other negative control pNANG-3 was integrated into either the niaD locus by double-crossover or the nor-I terminator by single-crossover (Figure 2.9 and 2.10 A and B). Southern hybridization and PCR analyses also showed that in the 6 EGFP (+) transformants carrying pAPGFPVNB3, the plasmid was integrated into the ver-I terminator at the ver-I A locus (isolates B3-15, B3-101, B3-105, B3-146, B3-160, and 83-194). In the 3 EGFP (+) transformants carrying pAPGFPVNB3, the plasmid was integrated into the ver-I promoter at the ver-I A locus (isolates B3-46, B3-120, and B3-186) in the chromosome (Figure 2.9 and 2.10 C to E). EGFP (+) transfonnant B3-146 showed another band in addition to the expected bands (2.8 kb and 8.2 kb) based on 3’ ver-I A integration in Southern hybridization analysis (Figure 2.10 C). These data together with PCR analysis suggest multiple integration of the plasmid. Southern hybridization and PCR analyses showed that in 4 EGF P (+) transformants carrying pAPGFPV1.1NBl, the plasmid was integrated into the ver-I terminator at the ver-I A locus (isolates LB1-94, LBl-101, the plasmid was integrated LB1-119, and LB1-170) while in 6 EGF P (+) transformants 68 Figure 2.10. Southern hybridization and PCR analyses of integration sites in transformants carrying pAPGFPVNBl, 2, or 3. For Southern hybridization analyses, genomic DNA was isolated from A. parasitisus, digested with Hincll, and hybridized with the ver-I A promoter probe. For PCR analyses, genomic DNA was amplified with an egfi) primer and a second primer specific to a 3’ or 5’ ver-I A sequence as shown in Figure 2.9 and Table 2.2. (A) Southern hybridization analysis of transformants carrying pAPGFPVNB] and 2. 3’ ver-I A integrants resulted in 2.8 kb and 8.2 kb bands. (B) PCR analysis of 3’ ver-I A or niaD integrants of pAPGFPVNB] and 2. 3’ ver-I A integrants resulted in a 2.6 kb band. (C) Southern hybridization analysis of transformants carrying pAPGFPVNB3. 3’ ver-I A integrants resulted in 2.8 kb and 8.2 kb bands while 5’ ver-I A integrants resulted in 3.7 kb and 7.2 kb bands. (D) PCR analysis of 3’ ver-I A or niaD integrants of pAPGFPVNB3. 3’ ver-I A integrants resulted in a 2.6 kb band. (B) PCR analysis of 5’ ver-I A or niaD integrants of pAPGFPVNB3. 5’ ver-I A integrants resulted in a 0.95 kb band. The recipient strain NR-l and transformants carrying pNiaD— Al or pNANG-3 were used as negative controls. NR-l (control) was used as a positive control. For the positive control, the egffp primer was replaced with a ver-I A primer to generate the same fragment sizes as those in 3’ or 5’ ver-I A integrants (Figure 2.9). k/Hindlll was used as a molecular size marker. 69 68“ SO 0° ‘0 ‘3‘" @Gr‘x‘s’gfi q .8 eat-diet ex $49353” "2° '° e 3&09’49’9‘“ 70 Figure 2.10. (cont’d) 71 Figure 2.10. (cont’d) 72 carrying pAPGFPV1.lNB, into the ver-I promoter at the ver-I A locus (isolates LBl-l l, LBl-15, LB 1-46, LB 1-112, LB 1-148, and LB 1-211) in the chromosome (Figure 2.9 and 2.11). EGF P (+) transformant LBl-148 also showed another band in addition to the expected bands (3.7 kb and 7.7 kb) based on 5’ ver-I A integration in Southern hybridization analysis (Figure 2.9 and 2.11). Again, these data together with PCR analysis suggest multiple integration of the plasmid. The data indicated that pAPGFPVNB] in EGF P (-) transfonnant B1-2 was integrated into the niaD locus by double-crossover (gene replacement) and that pAPGFPVNB2, 3, and pAPGFPVl.1NBl in EGFP (-) transformants B2-l, 3-1, and LBl-l were integrated into the niaD locus by single-crossover (Figure 2.9, 2.10, and 2.11). Time-course of EGFP expression and AF B] production EGFP (+) transformants carrying plasmids integrated into 3’ ver—I A were cultured in aflatoxin-inducing media and EGFP fluorescence activity and AFB] production were analyzed after 24, 48, 72, and 96 h incubation (B3-15 and LB1-101 were not analyzed at 96 h incubation). Dry weights of the transformants and control strain NR-l were similar at each time point (Figure 2.12). A transition from active growth to stationary phase was observed between 48 and 72 h as described previously (Liang et al. , 1997; Chiou et al., 2002). AFB] was not detected in any transfonnant or the recipient strain NR-l at 24 h, but high levels of AFB] were detected at 48, 72, and 96 h as described previously (Figure 2.12) (Liang et al., 1997; Chiou et al., 2002). EGF P fluorescence activity was nearly absent in NR-l. EGFP fluorescence was relatively high in all transformants, but it was 73 Figure 2.11. Southern hybridization and PCR analyses of integration sites in transformants carrying pAPGFPV1.1NBl. For Southern hybridization analysis, genomic DNA was isolated from A. parasitisus, digested with HincII, and hybridized with the ver-I A promoter probe. For PCR analyses, genomic DNA was amplified by PCR with an egfp primer and a second primer specific to 3’ or 5’ ver-I A sequence as shown in Figure 2.9 and Table 2.2. (A) Southern hybridization analysis of transformants carrying pAPGFPV1.1NB1. 3’ ver-I A integrants resulted in 2.8 kb and 8.7 kb bands while 5’ ver-I A integrants resulted in 3.7 kb and 7.7 kb bands. (B) PCR analysis of 3’ ver-I A integrants oprPGFPV1.lNB1. 3’ ver-I A integrants resulted in a 2.6 kb band. (C) PCR analysis of 5’ ver-I A integrants of pAPGFPVl.1NB1. 5’ ver-I A integrants resulted in a 1.5 kb band. The recipient strain NR-l was used as a negative control. NR-l (control) was used as a positive control in PCR analyses. For the positive control, the egfi) primer was replaced with a ver-I A primer to generate the same fragment sizes as those in 3’ or 5’ ver-I A integrants (Figure 2.9). k/HindIII was used as a molecular size marker. 74 3.7 kb , 2.8 kb .. 1.2 kb " b R J 3 2 4.4 kh bb kk 22. 2.6 kb-° 75 Figure 2.11. (cont’d) + b k 5 l 76 Figure 2.12. EGF P expression and AFB] production in EGFP (+) transformants containing different plasmids and the recipient strain NR-l. EGFP fluorescence activity and AFB] from the EGF P (+) transfonnant (A) B1-78, (B) Bl-86, (C) B2-9, (D) 82-20, (B) 83-15, or (F) LBl-101, and NR-l were measured after 24, 48, 72, and 96 h incubation (B3-15 and LBl-lOI were not analyzed at 96 h incubation) at 30 0C with shaking at 150 rpm in YES medium. Dry Weight (g) Dry Weight (g) Time-course of the transformant 81-78 5 —e— NR—l (D.W) l -~o 131-78 (D.W) + NR-1(EGFP) .. -— % _ 60 ] .---0 31-78016") ..... " l " + NR-I (A8139 . ' ..... ' - 40 0.1 ~ I _. 0.05 § I II" 0.01 - ~ 20 0.005 E". 0i ' r i r r 0 0 20 40 60 80 100 Time (hr) Time-course of the transformant B1-86 5 + NR-1(D.W) . - o Bi-86(D.W) l .. + NR-1(EGFP) ' .. I - 60 ] -- D 31-86(EGFP) " E: + NR-1(AFBI) 0-5 - A Bl-86(A — 40 0.1 .. 0.05 0.01 - ....... - 20 0.005 0i I 1 I I I 0 0 20 40 60 so 100 Time (hr) 78 Fluorescence (U/ug) Fluorescence (U/ug) ~30 ~20 *10 [30 ~20 ~10 AFB] (ug/ml) AFBl (ug/ml) Figure 2.12. (cont’d) C. Dry Weight (3) Dry Weight (g) Time-course of the transformant 82-9 5 + NR-1(D.W) --0 82-9 (D.W) + NR-1(EGFP) _ 60 1 -4 "D 32.9 (EGFP) . + NR-l (AFBFF' 0.5 ”A m_9 (AFBI ~ 40 0.1 - 0.05 0.01 - “ 10 0.005 0 r.” . 4 . 0 0 20 80 100 Time (hr) Time-course of the transformant 82-20 5 —e— NR-1(D.W) -0 82-20 (D.W) - + NR-1(EGFP) o t 60 ] - to 82-20 (EGFP) ” + NR-I (A881? 05 -A 82-20(Ar8 L 40 0.1 - 0.05 0.01 - ' 20 0.005 0 SI. 1 i 0 0 20 80 100 Time (hr) 79 Fluorescence (U/ug) Fluorescence (U/ug) ~30 ~20 ~10 [30 ~20 ~10 AFBI (ug/ml) AFBl (ug/ml) Figure 2.12. (cont’d) E. Time-course of the transformant 83-15 5 —o— NR-1(D.W) -o 83-15(8.W) + NR-1(EGFP) - *0 _ 60 ] - -r:i B3-15(EGFP) + NR-1(AFBI) o--‘ 0.5 -A B3-15(AFB]) 3'9 '5. 0.1 - I 40 '3 B 0.05 2‘ D 0.01 — ” 20 0.005 _ -' . " 0 i . . t 0 0 20 40 60 80 Time (hr) F. Time-course of the transformant LB1-101 5 + NR-1(D.W) - o LBI-101(D.W) + NR—1(EGFP) _. o ................ W _ 60 ].- u L81-101(1-:GFP) + NR-1(AFBI) g ------ 0-5 - A L81-101(Ai~‘81)-’ 3 . '5'. 0.1 - _ 40 '8 3 0.05 b a 0.01 - ~ 20 0.005 0 H. I U“ l 2. l r 0 0 20 40 60 80 Time (hr) 80 Fluorescence (U/ug) Fluorescence (ll/ug) r30 ~20 ~10 ~20 ~10 AFB] (ug/ml) AF Bl (ug/ml) 2~3 fold lower in transformants containing pAPGFPVNB2 than in transformants carrying pAPGFPVNBl or 3, indicating that a new positively charged arginine in EGF P produced by the EGFP (+) transformants carrying pAPGFPVNB2 might negatively affect fluorescence (Figue 2.12 and 2.4). Also, we observed detectable EGFP fluorescence in all EGFP (+) transformants at 24 h and fluorescence increased to at least 72 h (with exception of B2-20) (Figure 2.12). Overall, the expression pattern of EGFP by the ver-I promoter in the EGFP reporter system (Figure 2.12) was similar to the wild-type ver-I promoter (Liang et al., 1997) and paralleled AFB] production (Figure 2.12). Also, EGF P fluorescence by the 1.1 kb and 0.6 kb ver-I promoters was similar at each time point (Figure 2.13) and was consistent with nor-I promoter results; in this analysis 1.2 kb and 0.66 kb nor-1 promoter activities were similar when measured by nor-1 promoter::GUS reporter gene (Miller, 2003). These observations suggest that the 0.6 kb promoter is large enough that its transcription is not inhibited by transcription of m'iA. In conclusion, EGF P was expressed by the ver-I A promoter in A. parasiticus when the expression plasmid was integrated into either the 5’ promoter or the 3’ terminator of the ver-I A locus. EGF P was not expressed at the niaD locus in EGF P (-) transformants. The expression pattern of EGFP by the ver-I A promoter was similar to the wild-type ver-I promoter under aflatoxin-inducing condition. Confocal Laser Scanning Microscopy (CLSM) We detected EGFP fluorescence in all EGF P (+) transformants at 24 h (Figure 2.12) and this observation was consistent with data generated by confocal laser scanning microscopy (CLSM) (Figure 2.14 B). At 24 h, most fungal hyphae did not produce EGFP 81 Time-course of the transformants 81-86, 83-15, and LBl-101 + 81-86(EGFP) -l:l- B3-15(EGFP) —a— 1.81-101 (EGFP) 60“ Fluorescence (U/ug) Time (hr) Figure 2.13. Comparison of EGF P fluorescence activity in EGF P (+) transformants B1- 86, B3-15, or LB1-101. EGFP fluorescence activity from the EGFP (+) transformants was measured after 24, 48, and 72 h incubation at 30 0C with shaking at 150 rpm in YES medium. 82 Figure 2.14. Confocal laser scanning microscopy (CLSM) of EGFP (+) transfonnant B3- 15. Fungal vacuoles were stained with 8 11M F M 4-64 and observed using a Zeiss LSM 5 Pascal confocal laser scanning microscope (CLSM) after 24 h incubation at 30 0C with shaking at 150 rpm in YES medium. (A) The recipient strain NR-l (negative control). Red fluorescent vacuolar membranes were observed in hyphae but green fluorescence was rarely detected. (B) 83-15. EGFP fluorescence was not detected in most fungal hyphae and very little fluorescence was detected. Each panel shows a red fluorescence image (FM 4-64) (top left), a green fluorescence image (EGFP) (top right), a bright field image (bottom left), and a merged image (bottom right). Scale bars, 10 um. Images in this dissertation are presented in color. 83 64 A. The recipient strain NR-l, YES, 24 h, FM 4 B3-15, YES, 24 h, FM 4-64 B. 84 fluorescence in EGFP (+) transfonnant 83-15 and little fluorescence was detected using CLSM (Figure 2.14 B). CLSM was also performed to determine the sub-cellular localization of EGFP in hyhae of EGFP (+) transfonnant B3-15 after 24, 48, and 72 h culture on aflatoxin—inducing media, YES. EGFP fluorescence was not detected at 24 h but was detected throughout the cytoplasm of B3-15 at 48 h (Figure 2.15 B, C, and D). However, EGFP was localized primarily in vacuoles of B3-l 5 at 72 h (Figure 2.15 E and F). Also, the EGFP positive transfonnant B3-15 did not produce any EGFP fluorescence when it was cultured on non-aflatoxin-inducing media, YEP (Figure 2.15 G). DISCUSSION Our experiments demanstrated that the ver-I promoter::EGFP reporter and wild- type ver-I genes are regulated in a similar manner and that the reporter gene could be used for sub-cellular localization of aflatoxin biosynthetic enzymes in A. parasiticus. Initially we used pAPGFPVNB] as an EGFP reporter gene for fungal transformation; in this plasmid the Natl restriction enzyme site was placed between amino acid codon 1 and 2 in the egp coding region. However, in preliminary experiments, we could not identify any EGF P (+) transformants. We speculated that the Natl site might have negatively affected correct EGFP folding which is likely required for EGFP fluorescence. Therefore, we constructed 2 additional plasmids. In pAPGFPVNB2, the Natl site replaced amino acid codon 3, 4, and 5 of the egfp coding region, resulting in a change of 2 amino acids. In pAPGFPVNB3, the Natl site was placed just upstream of the egfla 85 Figure 2.15. Sub-cellular localization of EGFP in EGFP (+) transformant B3-15. Fungal vacuoles were stained with 8 uM FM 4-64 or 10 11M CMAC and observed using a Zeiss LSM 5 Pascal or Zeiss 510 Meta confocal laser scanning microscope (CLSM) after 24, 48, and 72 h slide culture incubation at 30 0C on aflatoxin-inducing media, YES or non- aflatoxin-inducing media, YEP. (A) The recipient strain NR-l (negative control) stained with FM 4-64 at 48 h on YES. Red fluorescent vacuolar membranes were observed in hyphae but green fluorescence was not detected. (B) B3-15 stained with FM 4-64 at 24 h on YES. Red fluorescent vacuolar membranes were observed in hyphae but green fluorescence was not detected. (C) and (D) B3-15 stained with F M 4-64 at 48 h on YES. EGF P was localized in the cytoplasm. Green fluorescence was excluded from vacuoles. (E) B3-15 stained with FM 4-64 at 72 h on YES. EGFP was localized in vacuoles of hyphae and conidiophores, and in phialides. (F) B3-15 stained with CMAC at 72 h on YES. EGFP was localized in vacuoles. (G) B3-15 stained with FM 4-64 at 48 h on YEP. Red fluorescent vacuolar membranes were observed in hyphae but green fluorescence was not detected. Each panel shows a red fluorescence image (FM 4-64) or a blue fluorescence image (CMAC) (top left), a green fluorescence image (EGFP) (top right), a transmitted image (bright field or differential interference contrast [DIC]) (bottom left), and a merged image (bottom right) except for panel (D) in which only red and green fluorescence images are shown. Scale bars, 10 um. Images in this dissertation are presented in color. 86 A. The reci ient strain NR—l, YES, 48 h, FM 4—64 . .~ a .. ,2 S '23" w ’3. c. H lUum :i' i it, 1 Figure 2.15. (cont’d) C. B3-15 YES 48h FM 4-64 H 11] um D. B3—15, YES, 48 h, FM 4—64 88 Figure 2.15. (cont’d) E. B3-15, YES, 72 h, FM 4-64 F. B3-15, YES, 72 h, CMAC 89 Figure 2.15. (cont’d) G. 33-15, YEP, 48 h, FM 4-64 _ H . Ilium ' lflprrt 90 coding region. We then identified EGF P (+) transformants with pAPGFPVNBl as well as pAPGFPVNB2 and 3 using the fluorescence microscope although we could not identify any EGFP (+) transformants with pAPGFPVNBl using U.V. at 365 nm. Southern hybridization and PCR analyses showed that pAPGFPVNB] in EGF P (-) transfonnant Bl—2 was integrated into the niaD locus by double-crossover and that pAPGFPVNBZ, 3, and pAPGFPV1.1NB1 in EGFP (-) transformants B2-1, 3-1, and LB1- l was integrated into the niaD locus by single-crossover (Figure 2.9, 2.10, and 2.11). These data were consistent with previous observations in which chromosomal location plays a role in regulation of aflatoxin gene expression (Liang et al., 1997; Chiou et al. , 2002). Previously, Liang et al. (1997) reported a position-dependent aflatoxin gene expression using ver-I promoter::GUS reporter gene in which 500- to 1000-fold lower levels of GUS activity were observed when the plasmid was integrated at the niaD locus. Also, Chiou et al. (2002) reported similar observations using a nor-1 promoter::GUS reporter gene in which no detectable GUS activity was observed when the plasmid was integrated at the niaD or pyrG locus outside of the aflatoxin gene cluster. In time-course experiments, the pattern of AFB] production was consistent with the pattern of EGFP fluorescence in most EGF P (+) transformants at 48 and 72 h. Also, a time-course experiment and CLSM showed detectable EGFP fluorescence in all EGFP (+) transformants at 24 h (Figure 2.12 and 2.14 B) in contrast to wild-type Ver-l protein in Western blot analysis (Liang et al., 1997). These data might suggest that EGFP is more resistant to proteinases than wild-type Ver-l, resulting in the ability to detect low levels at 24 h. Alternatively, fluorescence may be more sensitive than Western blot analysis in detection of signals. The low level of the fluorescence in NR-l might be a 91 result of autofluorescence from fungal residual cell wall debris; this could also have contributed to the detectable EGF P fluorescence in all EGF P (+) transformants at 24 h. Also, time-course experiments showed that EGF P fluorescence was 2~3 fold lower in transformants carrying pAPGFPVNB2 than in transformants carrying pAPGFPVNB] or 3. According to Li et al. (1997) glutamic acid residues 6 and 7 in GFP are critical for GFP fluorescence. We speculate that a mutation from glycine to the positively charged arginine at amino acid residue 5 in EGFP produced by the EGFP (+) transformants carrying pAPGFPVNBZ might have negatively affected fluorescence through glutamic acid residues 6 or 7 (Figure 2.12 and 2.4). CLSM showed cytoplasm localization of EGF P in hyphae of EGF P (+) transformant B3-15 at 48 h (Figure 2.15 C and D). This was consistent with observations of other researchers in which they detected sGFP fluorescence in the cytoplasm (V aldez- Taubas et al., 2000; Fernandez-Abalos et al. , 1998). However, EGFP was localized in vacuoles of B3-15 at 72 h (Figure 2.15 E and F). One possible explanation for this result is degradation and recycling of EGFP under starvation conditions. According to Baba et al. (1997) cytoplasmic proteins and organelles are nonselectively targeted to the vacuole by the autophagy for degradation and turnover of their constituent components by vacuole resident hydrolases under nitrogen or carbon starvation conditions. Therefore, EGF P in 83-15 accumulates in the cytoplasm at 48 h and then is localized to vacuoles at 72 h. This suggests that B3-15 was exposed to starvation conditions at 72 h. Also, the EGFP (+) transfonnant B3-15 did not produce any EGFP fluorescence when it was cultured on non-aflatoxin-inducing media, YEP (Figure 2.15 G) and this was consistent with previous observations in which it was reported that peptone as a sole carbon source does not 92 support aflatoxin gene expression and aflatoxin production (Skory et al., 1993; Buchanan and Lewis, 1984). ACKNOWLEDGMENTS We thank Dr. Melinda K. Frame and Dr. Shirley A. Owens (Center for Advanced Microscopy at Michigan State University) for help in confocal laser scanning microscopy. All experiments described in Chapter 2 were performed by Sung-Yong Hang. Chapter 2 will be submitted by combining with Chapter 3 for publication in the near future. 93 CHAPTER 3 SUB-CELLULAR LOCALIZATION OF THE VER-1 PROTEIN IN ASPERGILL US PARASI TIC US USING AN EGF P FUSION ABSTRACT To identify the sub-cellular location of the middle aflatoxin pathway enzyme Ver-l in real time and to confirm previous observations using transmission electron microscopy (TEM) and cell fractionation, we developed plasmid constructs expressing N- terrninally or C-terminally EGFP-tagged Ver-l fusion protein. The egfi-tagged ver-I fusion plasmid carrying a niaD (encodes nitrate reductase) selectable marker was transformed into A. parasiticus CS10-N2 (ver-I , niaD, pyrG, wh-l). Transformants were screened for aflatoxin production using aflatoxin detection medium, CAM (coconut agar medium) and for EGFP expression using fluorescence microscopy. Aflatoxin production was confirmed by TLC and ELISA. The data indicated that N-terrninally or C-terminally EGFP-tagged Ver-l protein functionally complemented the non-firnctional Ver-l protein in all AF (+) and EGF P (+) transformants. The transformants expressing N-terrninally EGFP-tagged Ver-l protein showed lower levels of Ver-l activity compared to the wild- type strain SU-l while C-terminally EGFP-tagged Ver-l protein showed similar Ver-l activity as SU-l. Production of full-length EGFP-tagged Ver-l fusion protein analyzed by Western blot analysis using anti-Ver-l antibody or anti-EGFP antibody. The data 94 indicated that the EGFP-tagged Ver-l fusion protein was produced intact and full-length from all AF (+) and EGFP (+) transformants. The plasmid integration sites were analyzed by Southern hybridization and PCR. The plasmid was integrated into the ver-I A locus in all AF (+) and EGFP (+) transformants. Confocal laser scanning microscopy (CLSM) data indicated that N-terminally or C-terminally EGFP-tagged Ver-l was localized in the cytoplasm and vacuoles of fungal hyphae on aflatoxin-inducing solid media at 48 and 72 h and especially it was detected inside of the vacuoles. Time-course experiment data strongly suggested that the AF (+) and EGFP (+) transformants were exposed to starvation conditions in 72 h slide culture. Therefore, we concluded that Ver-l was synthesized in the cytoplasm and transported to vacuoles of fungal hyphae by specific targeting for aflatoxin synthesis. INTRODUCTION Aflatoxin biosynthesis is a complex process that requires at least 24 enzyme activities (Dutton, 1988; Bhatnagar et al., 1992; Trail at al. 1995b; Yu et al., 1995; Yu et al. , 2004). The location of aflatoxin synthesis within a fungal cell is not clear. It was previously reported that several enzyme activities involved in aflatoxin B] biosynthesis were detected in a microsomal fraction (Lax et al., 1986; Bhatnagar et al., 1989; Yabe et al. , 1989; Yabe et al., 1993; Yabe et al., 1999) while other activities were found in the cytoplasm or loosely bound to membranes (Lax et al. , 1986; Bhatnagar et al., 1989; Yabe and Harnasaki, 1993; Matsushima et al., 1994; Yabe et al., 1999; Sakuno et al., 2003). In 95 previous sub-cellular localization studies, Ver-l was observed to associate with structures similar in size to lysosomes as well as in the cytoplasm using cell fractionation (Liang, 1996). Also, our lab observed Ver-l distributed evenly throughout substrate-level hyphae as well as aerial hyphae, vesicles, and conidiophores using immunofluorescence microscopy after paraffin embedment; the protein was primarily detected in the cytoplasm using transmission electron microscopy (TEM) after immunogold labeling in 24-48 h fungal colonies (Lee et al., 2004). It is currently known that ver-I encodes an NADPH-dependent oxidoreductase which is involved in conversion of versicolorin A (V ER A or VA) to demethyl- sterigmatocystin (DMST) (Ehrlich et al. , 2005; Henry and Townsend, 2005). Versicolorin A (V ER A) and other intermediates after VER A in the AFB] biosynthetic pathway possess a bisfuran ring containing a double bond and are mutagenic and genotoxic (Mori et a1. , 1985). Therefore, it is reasonable to believe that toxic intermediates in the AFB] biosynthetic pathway like VER A may be compartmentalized in specific organelles to protect cells from their toxicity during aflatoxin biosynthesis. Based on these data, we hypothesized that Ver-l is localized within specific organelles. In this study we conducted confocal laser scanning microscopy (CLSM) using EGFP- tagged Ver-l to identify the sub-cellular location of Ver-l in real time and to confirm previous observations using TEM afier immunogold labeling and cell fractionation. The resulting information about the sub-cellular localization of Ver-l will provide efficient strategies to block aflatoxin production in fungi. 96 MATERIALS AND METHODS Strains and culture conditions Escherichia coli DHSor F,e [F’/endAI hstI 7(rk' mk+) supE44 thi-I recA] gyrA (Nalr)relA1 A(lacZYA-argF)u]69: (m80AlacZM15)] (Gibco BRL, Life Technologies Inc, Rockville, MD) was used to amplify plasmid DNA using standard procedures (Ausubel et al., 2003). A. parasiticus CSlO-N2 (ver-I, niaD, pyrG, wh-I) (kindly provided by P.-K. Chang, USDA/ARS Southern Regional Research Center) was used as the recipient strain for ver-I negfp fusion plasmids containing the niaD selectable marker. CS10-N2 was derived from A. parasiticus CS10 (ver-I , pyrG, wh-I) by spontaneous mutation using potassium chlorate selection (Takahashi et al. , 2002). CS10 was in turn derived from A. parasiticus ATCC3 6537 (ver-I, wh-I) by spontaneous mutation using N.T.G. (N -methyl-N’-nitro-N-nitrosoguanidine) (Skory et al. 1990). ATCC36537 was generated from A. parasiticus SU-l by UV irradiation (Bennett and Goldblatt, 1973; Lee eta1., 1975). A. parasiticus B3-15 generated in Chapter 2 was used as a control for time- course experiments. A. parasiticus strains used in this study are listed in Table 3.1. E. coli was cultured in LB medium (Luria-Bertani; 1 % tryptone, 0.5 % yeast extract, 1 % sodium chloride) for competent cell preparation and plasmid isolation (Maniatis et al. , 1989). A. parasiticus was cultured in yeast extract-sucrose liquid medium (YES; 2 % yeast extract, 6 % sucrose, pH 5.8; batch fermentation) (plus 20 mM uracil) at 30 °C in the dark with shaking at 150 rpm for genomic DNA isolation, total protein extraction, measurement of mycelial dry weight, and analysis of aflatoxin 97 Table 3.1. Aspergillus parasiticus strains used in this study. Strain Genotypea Phenotype Source CS10-N2 var-1, niaD, pyrG, Wh-I AF (-) P.-K. Chang V2 pyrG, wh-l; vet-Izzegfp AF (-), EGFP (-) This study V1 pyrG, wh-I; var-1::egfp AF (-), EGFP (+) This study V32 pyrG, wh-I; var-1::egfp AF (-), EGFP (+) This study V107 pyrG, wit-1 ; var-1::egfp AF (+), EGFP (-) This study V188 pyrG, wh-I; vet-1::egfp AF (+), EGFP (-) This study V86 pyrG, wh-I; var-1::egfp AF (+), EGFP (+) This study V152 pyrG, wh-I; ver-Izzegfp AF (+), EGFP (+) This study LV6 pyrG, wh-I; ver-I::egfp AF (-), EGFP (-) This study LV8 pyrG, wit-1; ver-I::egfp AF (-), EGF P (+) This study LV29 pyrG, wit-1; ver-1::egfp AF (-), EGFP (+) This study LV13 pyrG, wh-I; var-1::egfp AF (+), EGFP (-) This study LV44 pyrG, wh-I ; ver-1::egfp AF (+), EGFP (-) This study LV140 pyrG, wh-I; ver-I::egfp AF (+), EGFP (-) This study LV155 pyrG, wh-I ; ver-I negfp AF (+), EGFP (-) This study LV169 pyrG, wh-l; ver-I neg/p AF (+), EGFP (-) This study LV70 pyrG, wh-I; ver-I::egfp AF (+), EGF P (+) This study NVl pyrG, wh-I; egfp::ver-1 AF (-), EGFP (-) This study NV63 pyrG, wh-I; egfp::ver-l AF (-), EGFP (-) This study NV109 pyrG, wh-I; egfp::ver-I AF (-), EGF P (-) This study NV170 pyrG, wit-I; egfp::ver-1 AF (-), EGFP (+) This study NV196 pyrG, wit-1; egfp::ver-1 AF (-), EGFP (+) This study NV27 pyrG, wh-I ; egfp::ver-1 AF (+), EGFP (+) This study NV60 pyrG, wh-I; egfp::ver-1 AF (+), EGF P (+) This study NV67 pyrG, wh-I; egfp::ver-1 AF (+), EGFP (+) This study 98 Table 3.1. (cont’d). Strain Genotypea Phenotype Source NV79 pyrG, wh-I ; egfp::ver-1 AF (+), EGF P (+) This study NV165 pyrG, wh-I; egfp::ver-I AF (+), EGFP (+) This study NV195 pyrG, wh-I; egfpizver-I AF (+), EGFP (+) This study NV218 pyrG, wh-I; egfp::ver-1 AF (+), EGFP (+) This study B3-15 var-1 p::egfp AF (+), EGFP (+) Chapter 2 a var-I p represents ver-I promoter. :: represents gene fusions. 99 concentration as described previously (Liang et al. , 1997; Takahashi et al., 2002). Potato dextrose broth (PDB; Difco Laboratories, Detroit, MI) supplemented with 20 mM uracil was used for growth of the recipient strain CSlO—N2 for transformation (Takahashi et al. , 2002). CZ agar supplemented with 20 % sucrose as an osmotic stabilizer, 20 mM uracil, and Cove’s trace element solution (Cove, 1966) was used as a selective medium for fungal transformation. Either potato dextrose agar (PDA; Difco Laboratories, Detroit, MI) plus 20 mM uracil or CZ agar plus 20 mM uracil was used for spore preparation. Coconut agar medium (CAM) plus 20 mM uracil was used for screening of aflatoxin producing fungal transformants under UV. at 365 nm (Chang et al., 1992) and for EGFP producing transformants using a Nikon Eclipse E600 fluorescence microscope (Nikon Inc., Melville, NY). YES agar (1.5 % agar) (plus 20 mM uracil) was used for extraction of aflatoxins and aflatoxin intermediates for TLC and ELISA analyses. Either YES or YEP (2 % yeast extract, 6 % peptone, pH 5.8) agar (plus 20 m M uracil) was used for slide culture which was observed using a Zeiss LSM 5 Pascal or Zeiss LSM 510 Meta confocal laser scanning microscope (CLSM) (Carl Zeiss Inc., Germany). Construction of egfp-tagged ver-I plasmids The N-terminally or C-terrninally egfp-tagged ver-I plasmids (Figure 3.1 and 3.2), were constructed using a fused ver—I promoter/gene fragment, or separate ver-I promoter and gene fragments, an egfp gene fragment, and pAPGFPVNB3 as a plasmid backbone. These plasmids differed in the length of the hinge region between ver-I and em?) coding regions. For the C-terminally amp-tagged ver-I plasmid, the 2.0 kb ver-I promoter/gene fragment was generated by PCR with Pfu DNA polymerase (Stratagene, La Jolla, CA), 100 Asc I 418 / Pst I 1010 / Sal I 1908 Pst I 12639 S (I 12627 Am r ‘ , 54112188 a P vef-I / Fsel2518 terminator * " egfp Not I 3244 pAPCGFPV FNB L _ (14933 hp) ~ var-1 promoter, , lgene * Pac I 5259 Sal I/Xho I 5275 Pst I 8427 Figure 3.1. Restriction endonuclease map of plasmid, pAPCGFPVFNB. The 2.0 kb ver-I promoter/ gene was fused in frame to the 0.7 kb egfi) coding region, followed by the 2.1 kb ver-I terminator. The 7.4 kb niaD fragment was inserted as a selectable marker for transformation of the recipient strain CS10-N2 (ver-I , niaD, pyrG, wh-I). 101 A. pAPCGFPVFNB Ver-l Hinge region EGF P DNA CGA-gtg-gct-ggc-ggc—cgc-ATG Natl Amino acid Arg-Val-Ala-Gly-Gly-Arg-Met B. pAPCGFPLVFNB Ver—l Hinge region EGFP DNA CGA-gtg-gct-ggc-ggc-cgc-gga-gct—ggt-gca-ggc-gct-gga-gcc-ATG Natl Amino acid Arg—Val-Ala-Gly-Gly-Arg-GIy-Ala-GIy-Ala-Gly-Ala-Gly-Ala-Met C. pAPNGFPVFNB EGFP Hinge region Ver-l DNA AAG-ggc-gat—cgc-gga-gct—ggt-gca-ATG Sgfl Amino acid Lys-Gly-Asp-Arg-Gly-Ala-Gly-Ala-Met Figure 3.2. Comparison of selected sequences in 3 plasmids carrying an amp-tagged ver-I. (A) pAPCGFPVFNB. (B) pAPCGFPLVFNB. (C) pAPNGFPVFNB. These plasmids differ in the length of the hinge region between the ver-I and egfi) coding regions, and the relative location of these two regions. Natl and ngl sequences are shown in italic. Abbreviations for amino acids are as follows; Arg, arginine; Val, valine; Ala, alanine; Gly, glycine; Met, methionine; Lys, lysine; Asp, aspartic acid. The repeated Gly-Ala sequences are modified from a sequence kindly provided by S. Osmani, Ohio State University. 102 appropriate primers, and cosmid NorA (Figure 2.5) (Liang et al. , 1996) as a template using standard procedures (Maniatis et al., 1989). The primers contained restriction enzyme sites to facilitate cloning (Table 3.2). PCR was performed in a Gene Amp PCR System 2400 thermal cycler (Perkin-Elmer life sciences Inc., Boston, MA). The reaction conditions for thermal cycling depended on the designed primers and the target size as follows: 94 0C for 5 min followed by 25 cycles of denaturation at 94 0C for 1 min, annealing for 1 min (Table 3.2), and extension at 72 0C for time in min that depended on PCR fragment sizes (2 min/1 kb). The reactions were completed with an extension at 72 0C for 10 min. The PCR fragments were purified from an agarose gel using a Qiaex 11 gel extraction kit (Qiagen Inc., Valencia, CA), digested with PacI and Natl, and cloned into pAPGFPVNB3 (Figure 2.2) out with the same enzymes, resulting in pAPCGFPVFNB (Figure 3.1). To generate a C-terminally eflp-tagged ver-I plasmid containing a longer hinge region between ver-I gene and egfp gene, the 0.7 kb egfp gene was amplified by PCR with Pfu DNA polymerase and appropriate primers using pEGFP- Nl (Clonetech Laboratories, Palo Alto, CA) as a template (Figure 2.6). The primers contained restriction enzyme sites to facilitate cloning (Table 3.2). PCR was performed in a Gene Amp PCR System 2400 thermal cycler as described above. Plasmid pAPCGFPLVFNB containing a longer hinge region between ver-I gene and egfp gene was constructed by replacement of the egfp gene fragment in pAPCGFPVFNB with another eg'p gene fragment cut with Natl and Fsel to provide more space between Ver-l and EGF P for correct fusion protein folding. For the N-terrninally eg/p-tagged ver-I plasmid, the 0.9 kb ver-I gene fragment was generated by PCR with Pfix DNA 103 Table 3.2. Primer sequences used in this study. . a Primer b . . Sequence Restriction enzyme site, annealing temp (°C) ver-IA promoter 5’ TCCGGGTTAATTAAGATGCCGAACCATTTGAC 3’ Bad, lgene (C)- F 55 ver-IA promoter 5’ ACTATAGCGGCCGCCAGCCACTCGAAAAGCGCC- Natl, lgene (C)- R -ACC 3’ 55 egfp gene 5’ AGTGCGGCCGCGGAGCTGGTGCAGGCGCTGGA- Natl, (CL)- F -GCCATGGTGAGCAAGGGCGAGGAGCTGTTC 3’ 65 egfp gene 5’ GTCGGCCGGCCTTTACTTGTACAGCTCGTCCAT 3’ Fsel, (CL)- R 65 egfp gene 5’ CCGGGCGGCCGCATGGTGAGCAAGGGCGAG 3’ Natl, (N)- F 60 egfp gene 5’ GCCGCGATCGCCCTTGTACAGCTCGTCCATGCC 3’ Sg/I, (N)- R 60 ver-I A gene 5’ CTGGCGATCGCGGAGCTGGTGCAATGTCGGATA- Sgt], (N)- F -ATCACCGTTTAGAT 3’ 60 ver-I A gene 5’ AGCGGCCGGCCATTATCGAAAAGCGCCACC 3’ FseI, (N)- R 60 egfp gene 5’ GGC-AAC-TAC-AAG-ACC-CGC-G 3’ None, (3’ integrant)— F 68 Downstream of 5’ AGC-CAC-CGT-GAG-CGT-CC 3’ None, ver-I A terminator 68 (3’ integrant)- R 104 Table 3 .2. (cont’d). Primera Sequenceb Restriction enzyme site, annealing temp (”c1 Downstream of 5’ ACA-CAT-GAG-AGC-CAG-CAA-GAT-AA 3’ None, ver-I B terminator 68 (3’ integrant)- R ver-I A gene 5’ ATC-CTG-ACC-AGC-TCT-AAC-ACC-G 3’ None, (3’ control)- F 68 a C represents C-terminal egfp fusion, CL represents C-terminal egfp fusion containing a long hinge region, and N represents N-terminal egfp fusion. Also, F represents forward primers and R represents reverse primers. b Underlined sequences show the position of the restriction enzyme sites. 105 polymerase, appropriate primers, and cosmid NorA (Figure 2.5) (Liang et al. , 1996) as a template using standard procedures (Maniatis et al. , 1989). The primers contained restriction enzyme sites to facilitate cloning (Table 3.2). Also, the 0.7 kb egfp gene was generated by PCR with Pfu DNA polymerase and appropriate primers using pEGFP-Nl as a template (Figure 2.6). The primers contained restriction enzyme sites to facilitate cloning (Table 3.2). PCR was performed in a Gene Amp PCR System 2400 thermal cycler as described above. The PCR fragments were cloned into the SmaI site of pUC l 9, resulting in pUCGF P and pUCVER, after purification from an agarose gel using a Qiaex 11 gel extraction kit. DNA fragments containing the ver-I gene were subcloned from pUCVER into pUCGFP cut with Sgl and Sail, resulting in pUCGFPVER. DNA fragments containing the egp gene and the ver-I gene were then subcloned from pUCGFPVER into pAPGFPVNB3 (Figure 2.2) cut with Natl and Fsel, resulting in pAPNGFPVFNB. Transformation of E. coli and isolation of plasmids The preparation and transformation of E. coli competent cells were conducted by a calcium chloride method as described previously (Ausubel et al. , 2003). The isolation of plasmids was performed by an alkaline lysis method using a Wizard DNA purification kit (Promega Corporation, Madison, WI) or CsCl/ethidium bromide equilibrium centrifugation as described previously (Maniatis et a1. , 1989). 106 Transformation of A. parasiticus and screening for aflatoxin and EGFP in transformants Transformation of A. parasiticus CS10-N2 with pAPCGFPVFNB, pAPCGFPLVFNB, or pAPNGFPVFNB was performed by a polyethylene glycol method (Oakley et al. , 1987) with minor modifications as described previously (Skory et al. , 1990). Fifteen h after inoculation of 108 conidia, protoplasts were generated by digestion of mycelia with lysing enzyme (25 mg/ml; Sigma Chemical Co., St. Louis, MO) and driselase (50 mg/ml; Sigma Chemical Co., St. Louis, MO). Transformants were selected on CZ agar supplemented with 20 % sucrose as an osmotic stabilizer, 20 mM uracil, and Cove’s trace element solution. Transformants were transferred and cultured onto CAM plus 20 mM uracil at 30 0C for 4 days and then screened for a blue fluorescent halo (aflatoxin) around their colonies under UV. at 365 nm and for EGFP under a Nikon Eclipse E600 fluorescence microscope (Nikon Inc., Melville, NY) using a 450-490 nm excitation/ 515 nm emission filter. Conventional fluorescence microscopy Slide culture was performed by the method according to Harris (1986) with minor modifications as described previously (Liang, 1996) (Figure 2.7). Approximately 5x105 conidia were inoculated around the edge of the top surface of YES (plus 20 mM uracil) agar blocks (1 cm2) on sterile coverslips which were placed onto water agar plates. This was followed by placement of a second coverslip on top of the YES (plus 20 mM uracil) 107 agar blocks. The plates were incubated at 30 0C in the dark for 2 days. Coverslips were removed and washed three times with phosphate-buffered saline (PBS) and observed using a Nikon Labophot fluorescence microscope (Nikon Inc., Melville, NY) with a 450- 490 nm excitation/ 520 nm emission filter. Aflatoxin and aflatoxin intermediate analyses by TLC and ELISA Aflatoxin and aflatoxin intermediates were extracted by the method of Roze et al. (2004). Approximately 1x104 conidia were inoculated on YES (plus 20 mM uracil) agar plates and incubated at 30 0C in the dark for 4 days. Aflatoxin and aflatoxin intermediates were extracted from the colonies with 10 ml of chloroform and then 10 ml of acetone, dried by evaporation, and dissolved in 3 ml of 70 % methanol. TLC was conducted on aflatoxin and aflatoxin intermediate extracts using a TEA solvent system (toluene-ethyl acetate-acetic acid; 50:30:4 [v/v/v]) (Chang et al. , 2004). The AFB] concentration in the cell extracts was determined by direct competitive enzyme-linked immunosorbent assay (ELISA) with AFB] monoclonal antibodies (kindly provided by J. Pestka, Michigan State University) as described previously (Pestka, 1988). Total protein extraction Approximately 2x106 conidia were cultured in 100 ml of YES (plus 20 mM uracil) media at 30 0C in the dark with shaking at 150 rpm for 48 h. After filtration through Miracloth (Calbiochem, La Jolla, CA), mycelia were ground in liquid nitrogen 108 with a mortar and pestle, suspended in TET buffer (100 mM Tris-HCI [pH7.5], 2.5 mM EDTA, 5 mM DTT, 1 % Triton X-100) containing 1 mM phenylmethylsulfonyl fluoride (PMSF) (Sigma Chemical Co., St. Louis, MO) and 4 % proteinase cocktail (Sigma Chemical Co., St. Louis, MO), and centrifuged at 10,000x g for 15 min at 4 0C (V aldez- Taubas eta1., 2000, Tavoularis et al., 2001). The supematants were used for Western blot analysis. Protein concentration in the supernatant was determined by a modified Bradford assay using a commercial Protein Assay reagent (Bio-Rad Laboratories, Hercules, CA) (Bradford, 1976). Western blot analysis of EGFP-tagged Ver—l Approximately 30-50 ug of total proteins were separated by 12 % sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Gels were transferred onto Poly Screen polyvinylidene difluoride (PVDF) membranes (Perkin-Elmer life sciences, Boston, MA) using standard procedures (Ausubel et al. , 2003). After transfer, the gels were stained with Coomassie Brilliant Blue R-250 (Sigma Chemical Co., St. Louis, MO). Immunodetection was carried out with IgG antibody against Ver-l protein or EGFP (Clonetech Laboratories, Palo Alto, CA) as a primary antibody, goat anti-rabbit IgG alkaline phosphatase conjugate (Sigma Chemical Co., St. Louis, MO) as a secondary antibody, and a BCIP/NBT (5’-bromo-4-chloro-3-indolyl phosphate/nitroblue tetrazolium) colorimetric detection system (Roche Molecular Biochemicals, Indianapolis, IND). A Benchmark prestained protein ladder (Invitrogen, Carlsbad, CA) was used as a molecular mass marker. 109 Genomic DNA isolation from A. parasiticus Genomic DNAs were isolated by a phenol-chloroform method (Ausubel et al. , 2003) with minor modifications as described previously (Skory et al. , 1990). Approximately 2x106 conidia were cultured in 100 ml of YES plus 20 mM uracil media at 30 0C in the dark with shaking at 150 rpm for 48 h and then a phenol-chloroform protocol was used to isolate genomic DNA from mycelia after filtration through Miracloth (Calbiochem, La J olla, CA). Southern hybridization and PCR analyses Southern hybridization analyses were conducted using standard procedures as described previously (Maniatis et al., 1989). Approximately 10 ug of genomic DNAs cut with PstI were separated by agarose gel electrophoresis and then transferred onto a nylon transfer membrane (Nytran supercharge membrane, Schleicher and Schell Inc., Keene, NH) by capillary action. Radiolabeled DNA probes were generated with [or-32P]dCTP (Perkin-Elmer life sciences Inc., Boston, MA), the Random Primers DNA Labeling System (Invitrogen, Carlsbad, CA), and the 0.7 kb gfp gene fragment using a procedure provided by the manufacturer. After the final wash, the membranes were exposed to X- ray film (Eastman Kodak company, Rochester, NY) at —80 0C. PCR analyses were performed with genomic DNA to confirm integration sites of the plasmids and to amplify the ver-I A gene region fused to eg‘p gene in order to determine if the fusion protein carried functional wild-type Ver-l protein. DNA sequencing of the amplified ver-I A gene region fused to the egfi) gene was conducted at the Research Technology Support 110 Facility (RTSF; Macromolecular Structure, Sequencing and Synthesis facility) at Michigan State University. A/HindIII (Invitrogen, Carlsbad, CA) was used as a molecular size marker. Confocal Laser Scanning Microscopy (CLSM) Slide culture was performed as described above (Liang, 1996). For aflatoxin non- inducing media, YEP was used instead of YES. The plates were incubated at 30 0C in the dark. Coverslips were removed at different time points after inoculation. Fungal vacuoles were then stained with F M 4-64 or 7-amino-4-chloromethylcoumarin (CMAC) (Ohneda et al., 2002; Shoji et al. , 2006). FM 4-64 and CMAC were used for staining vacuolar membranes and lumens, respectively. The coverslips were put in YES (plus 20 mM uracil) media containing 8 11M FM 4-64 or 10 11M CMAC. For FM 4-64, coverslips were incubated at 30 0C for 10 min and washed with fresh media without the dye for 30 min. For CMAC, coverslips were incubated at 30 0C for 30 min and washed with fresh media without the dye at 37 0C for 30 min. Coverslips were observed using a Zeiss LSM 5 Pascal or Zeiss LSM 510 Meta confocal laser scanning microscope (CLSM) (Carl Zeiss Inc., Germany). All single optical sections and extended focus images from Z-stacks (Z- section interval: 0.46 um) were captured using a Zeiss Plan-APOCHROMAT (63x/1.40 Oil) objective. EGFP fluorescence (488 nm excitation/ 509 nm emission) was detected using a BP 505-530 emission filter set under excitation with the 488 nm argon-ion laser line. FM 4-64 fluorescence (558 nm excitation/ 734 nm emission) was detected using a 111 LP 650 emission filter set under excitation with the 633 nm helium-neon laser line. CMAC fluorescence (353 nm excitation/ 466 nm emission) was detected using a BP 420- 480 emission filter set under excitation with the 405 nm diode laser line. For counting of vacuole localization of green fluorescence, large- and mid-size vacuoles (>5 um) were counted in 2 or 3 hyphae from 1 microscopic field and this was repeated in a total 30 fields. The counting of green fluorescent vacuoles was statistically analyzed by two-way ANOVA followed by Tukey’s test for multiple comparisons using SigmaStat (SPSS Inc., Chicago, IL). Statistical significance among samples was defined by a P value not greater than 0.05 (P5005). Time-course of aflatoxin production Approximately 2x106 conidia were cultured in 100 ml of YES plus (20 mM uracil) media at 30 0C in the dark with shaking at 150 rpm as described previously (Liang et al., 1997). Flasks were removed at different time points after inoculation for analyses of mycelial dry weight and aflatoxin concentration. Mycelia were harvested by filtration through Miracloth (Calbiochem, La Jolla, CA), frozen in liquid nitrogen, and stored at — 80 0C. Slide culture was performed as described above. The coverslips were removed at different time points after inoculation for analysis of aflatoxin concentration. Measurement of mycelial dry weight and aflatoxin concentration Dry weight was determined after complete drying of the harvested mycelia at 100 0C. The AFB] concentration in the filtrate was determined by direct 112 competitive enzyme-linked immunosorbent assay (ELISA) with AFB] monoclonal antibodies (kindly provided by J. Pestka, Michigan State University) as described previously (Pestka, 1988). For slide culture, aflatoxins were extracted from the agar blocks with 5 ml of chloroform and then 5 ml of acetone, dried by evaporation, and dissolved in 1 ml of 70 % methanol. The AFB] concentration was determined by ELISA with AFB] monoclonal antibodies as described previously (Pestka, 1988). RESULTS Transformation of A. parasiticus CSlO—NZ with pAPCGFPVFNB, pAPCGFPLVFNB, and pAPNGFPVFNB, and screening for aflatoxin and EGFP in transformants Five to ten ug of pAPCGFPVFNB, pAPCGFPLVFNB, or pAPNGFPVFNB was transformed into 108 of A. parasiticus CS10-N2 protoplasts, generating 673 transformants. All transformants selected on CZ agar plus 20 mM uracil were screened for aflatoxin production on CAM plus 20 mM uracil under UV. at 365 nm. Four transformants carrying pAPCGFPVFNB produced blue fluorescent haloes around their colonies out of 210 transformants. All 4 aflatoxin producing transformants accumulated no detectable yellow pigment in the mycelia. This yellow pigment has been confirmed by NMR, TLC, and ELISA to contain predominantly versicolorin A (Lee et al. , 1975; Reynolds and Pestka, 1991). All 210 transformants were then screened for 113 EGFP expression under the fluorescence microscope (Figure 3.3). Two (isolates V86 and V152) of the 4 aflatoxin producing transformants produced EGF P fluorescence (EGFP [+] ). The other two (isolates V107 and V188) aflatoxin producing transformants did not produce EGFP fluorescence. Twenty-four EGFP (+) transformants were identified out of the non-aflatoxin producing transformants. Six transformants carrying pAPCGFPLVFN B produced blue fluorescent haloes around their colonies out of 233 transformants. All 6 aflatoxin producing transformants also accumultated no detectable yellow pigment in the mycelia. All 233 transformants were then screened for EGFP expression under the fluorescence microscope (Figure 3.3). One (isolate LV70) of 6 aflatoxin producing transformants produced EGF P fluorescence. The other five aflatoxin producing transformants (isolates LV13, LV44, LV140, LV155, and LV169) did not produce EGFP fluorescence. Twenty-four EGFP (+) transformants were identified out of the non-aflatoxin producing transformants. Seven transformants carrying pAPNGFPVFNB produced blue fluorescent haloes around their colonies out of 230 transformants. All 7 aflatoxin producing transformants accumultated a small quantity of detectable yellow pigment in the mycelia. All 230 transformants were then screened for EGF P expression under the fluorescence microscope (Figure 3.3). All 7 aflatoxin producing transformants (isolates NV27, NV60, NV67, NV79, NV165, NV195, and NV218) produced EGF P fluorescence. Twelve EGFP (+) transformants were identified out of the non-aflatoxin producing transformants. TLC and ELISA analyes of transformants To confirm complementation of non-functional Ver-l in A. parasiticus CS 1 0-N2 114 Figure 3.3. Fluorescence microscopy of A. parasiticus CSlO-N2 expressing EGF P. Transformants were inoculated on CAM plates or YES agar blocks on coverslips and incubated at 30 0C for 2 or 4 days and observed using a Nikon Labophot or Eclipse E600 fluorescence microscope. (A) Transformant carrying pAPCGFPVFNB. EGF P fluorescence was observed in vesicles and conidiophores. (400 x) (B) Transformant carrying pAPCGFPLVFNB. EGFP fluorescence was observed in conidia, vesicles and conidiophores. However, EGFP fluorescence was not detected in some conidia but only in vesicles and conidiophores. (100 x) (C) Transformant carrying pAPNGFPVFN B. EGF P fluorescence was observed in vesicles and conidiophores. (100 x) Images in this dissertation are presented in color. 115 116 by EGFP-tagged Ver-l , TLC analysis was performed using chloroform/acetone extracts from transformants carrying pAPCGFPVFNB. One AF (-) and EGFP (-) transfonnant (isolate V2) did not produce aflatoxins but produced versicolorin A (VA) and 2 AF (-) and EGFP (+) transformants (isolates V1 and V32) also did not produce aflatoxins but produced VA (Figure 3.4 A). However, 2 AF (+) and EGF P (-) transformants (isolates V107 and V188) produced aflatoxins but almost no VA (Figure 3.4 A). Two AF (+) and EGFP (+) transformants (isolates V86 and V152) produced aflatoxins but did not produce VA (Figure 3.4 B). To compare AFB] production from transformants with that from the wild-type stain SU-l , AFB] concentration was measured by ELISA. V107 and V86 produced similar amounts of AFB] to that from SU-l (Figure 3.5). TLC analysis was performed using chloroform/acetone extracts from the transformants carrying pAPCGFPLVFNB. One AF (-) and EGFP (-) transfonnant (isolate LV6) did not produce aflatoxins but produced versicolorin A (VA) and 2 AF (-) and EGF P (+) transformants (isolates LV8 and LV29) produced almost no aflatoxins but produced large amounts of VA (Figure 3.6 A). However, 5 AF (+) and EGF P (-) transformants (isolates LV13, LV44, LV140, LV155, and LV169) produced aflatoxins but almost no VA (Figure 3.6 A and B). One AF (+) and EGFP (+) transformants (isolate LV70) produced aflatoxins but did not produce VA (Figure 3.6 B). When taken together, TLC and ELISA analyses indicated that in all 3 AF (+) and EGFP (+) transformants carrying pAPCGFPVFNB or pAPCGFPLVFNB (V 86, V152, and LV70), the C- terminally EGFP-tagged Ver-l protein functionally complemented non-functional Ver-l in A. parasiticus CS10-N2 (Figure 3.4, 3.5, and 3.6). 117 Figure 3.4. TLC analysis of extracts from transformants carrying pAPCGFPVFNB and the recipient strain CSlO-N2. (A) TLC analysis of extracts from AF (-) and EGFP (-) (V2), AF (-) and EGFP (+) (V1 and V32), and AF (+) and EGFP (-) (V107 and V188) transformants. (B) TLC analysis of extracts from AF (+) and EGF P (+) transformants (V86 and V152). Aflatoxin B], B2, G], and G2 standard mixture and versicolorin A (VA) were used as standards. TEA (toluene-ethyl acetate-acetic acid; 50:30:4 [v/v/v]) was used as a solvent system. Fluorescence was detected under UV. at 365 nm. 118 AFB,» AFC," AF VA C810 V2 V1 V32 V107V188 AF Std Std -N2 Std VA-t AFB,» AFG," AF VA C510 V86 V152 AF Std Std -N2 St 119 AFB] analysis by ELISA 2500 AFB] (ug/ml) 1000 ‘ SU-l CSl0-N2 V2 V1 V107 V86 Transformants carrying pAPCGFPVFNB Figure 3.5. AFB] analysis of extracts from transformants carrying pAPCGFPVFNB, the recipient strain CS10-N2, and the wild-type strain SU-l. Extracts from AF (-) and EGFP (-) (V 2), AF (-) and EGF P (+) (V 1), AF (+) and EGFP (-) (V107), and AF (+) and EGFP (+) (V 86) transformants, CS10-N2, and SU-l were analyzed for AFB] production by ELISA. 120 Figure 3.6. TLC analysis of extracts from transformants carrying pAPCGFPLVFNB and the recipient strain CSlO-N2. (A) TLC analysis of extracts from AF (-) and EGFP (-) (LV6), AF (-) and EGF P (+) (LV8 and LV29), and AF (+) and EGFP (-) (LV13 and LV44) transformants. (B) TLC analysis of extracts from AF (+) and EGF P (-) (LV140, LV155, and LV169) and AF (+) and EGFP (+) transformants (LV70). Aflatoxin B], B2, G], and G2 standard mixture and versicolorin A (VA) were used as standards. TEA (toluene-ethyl acetate-acetic acid; 50:30:4 [v/v/v]) was used as a solvent system. Fluorescence was detected under UV. at 365 nm. 121 Ar8,—- AFG,-* AF VA C810 LV6 LV8 LV29 LV13 LV44 AF Std Std -N2 Std AF VA C810 LV140LV155LV169 LV70AF Std Std ~N2 Std 122 TLC analysis was performed using chloroform/acetone extracts from the transformants carrying pAPNGFPVFNB. Three AF (-) and EGFP (-) transformants (isolates NVl , NV63, and NV109) and 2 AF (-) and EGFP (+) transformants (isolates NV170 and NV196) did not produce aflatoxins nor did they produce versicolorin A (VA) (Figure 3.7 A and C). However, 7 AF (+) and EGF P (+) transformants (isolates NV27, NV60, NV67, NV79, NV165, NV195, and NV218) produced aflatoxins but still produced detectable amounts of VA (Figure 3.7 A and B). AFB] concentration was measured by ELISA. NV27, NV60, and NV67 produced 33% of AFB] produced by SU-l (Figure 3.8). When taken together, TLC and ELISA analyses indicated that in 3 AF (+) and EGF P (+) transformants carrying pAPNGFPVFNB, NV27, NV60, and NV67, N-terminally EGFP-tagged Ver-l protein functionally complemented the non-functional Ver-l in A. parasiticus CSlO-N2 although the transformants showed decreased amounts of AFB] production compared to SU-l (Figure 3.7 and 3.8) Western blot analysis of EGFP—tagged Ver—l To assure production of full-length EGFP-tagged Ver-l fusion protein in transformants, Western blot analysis was performed using anti-Ver-l antibody or anti- EGF P antibody. In transformants carrying pAPCGFPVFNB, 2 AF (+) and EGFP (+) transformants (isolates V86 and V152) produced a full length 55 kDa fusion protein (Figure 3.9 B and D). F iftyfive kDa is the expected molecular mass of fiill length fusion protein (28 kDa Ver-l and 27 kDa EGFP). Also, 2 AF (-) and EGFP (+) transformants (isolates V1 and V32) produced a 55 kDa fusion protein (Figure 3.9 A and C). However, 123 Figure 3.7. TLC analysis of extracts from transformants carrying pAPNGFPVFNB and the recipient strain CSlO-N2. (A) TLC analysis of extracts from AF (-) and EGFP (-) (NVl, NV63, and NV109), and AF (+) and EGFP (+) (NV 27 and NV60) transformants. (B) TLC analysis of extracts from AF (+) and EGFP (+) (NV67, NV79, NV165, NV195, and NV218) transformants. (C) TLC analysis of extracts from AF (-) and EGFP (+) (NV170 and NV196) transformants. Aflatoxin B], B2, G], and G2 standard mixture and versicolorin A (VA) were used as standards. TEA (toluene-ethyl acetate-acetic acid; 50:30:4 [v/v/v]) was used as a solvent system. Fluorescence was detected under UV. at 365 nm. 124 VA“ AFB,» AFGr‘ AF VA C810 NV1NV63NV109 NV27NV60 AF Std Std -N2 Std VA-‘ AFBf‘ AFC,» AF VA C810 NV67NV79NV165NV195NV218AF Std Std -N2 Std 125 Figure 3.7. (cont’d) AFB,» AFGf‘ AF VA C810 NV170NV196 AF Std Std -N2 Std 126 AFB] analysis by ELISA 2500 AFB, (ug/ml) I I I SU-l C810-N2 NVI NV170 NV27 NV60 NV67 Transformants carrying pAPCGFPVFNB Figure 3.8. AFB] analysis of extracts from transformants carrying pAPNGFPVFNB, the recipient strain CSlO-N2, and the wild-type strain SU-l. Extracts from AF (-) and EGF P (-) (NV 1), AF (-) and EGF P (+) (NV 170), and AF (+) and EGFP (+) (NV27, NV60, and NV67) transformants, CS10-N2, and SU-l were analyzed for AFB] production by ELISA. 127 Figure 3.9. Western blot analysis of EGFP-tagged Ver-l from transformants carrying pAPCGFPVFNB, the recipient strain CSlO-N2, and the wild-type strain SU-l. Fungal proteins were extracted from transformants, CSlO-N2, and SU-l grown in 100 ml of YES for 48 h at 30 °C with shaking at 150 rpm. Approximately 30-50 11g of proteins were separated by 12% SDS-PAGE, transferred onto PVDF membranes, and probed with (A) and (B) Ver-l polyclonal antibody, or (C) and (D) EGF P polyclonal antibody. EGFP- tagged Ver-l has a molecular mass of 55 kDa, and the 28 kDa protein represents Ver-l (with the exception of a functional Ver-l in SU-l). A Benchmark prestained protein ladder was used as a molecular mass marker (M). Recombinant EGFP (rEGF P) was used as a positive control and, the 30 kDa rEGFP contains a 27 kDa EGFP fused to a 3 kDa protein for affinity chromatography purification. 128 kDa 170.8 109.5 78.9 60.4 47.2 35.1 24.9 18.3 13.7 kDa 1 70.8 1 09.5 78.9 60.4 47.2 35.1 24.9 18.3 13.7 41» .5 é" tx % 859° 094'” 4‘494‘348’8 wit-om“... 129 "55 kDa «281011 "55 kDa *28kDa Figure 3.9. (cont’d) C. 4" as > at ’\ t], c, x 63° (5° 4" 4‘ 494‘“ 483’ 8“ kDa _ 170.8 .1 109.5 H 78.9 55 kDa» ~—- g3 60.4 it; 472 it! 35.1 30 kDa» .~ “i " a! 24.9 it". 18.3 ,1 13.7 D. e” as .5 s’ ‘3 C: 4 a? a 9° 8" 8 4 kDa 170.8 109.5 78.9 22'; *- .7 +55 kDa 35.1 i ° 0‘, i ‘30 kDa 24.9 , _ 18.3 13.7 130 1 AF (-) and EGF P (-) transfonnant (isolate V2), and 2 AF (+) and EGFP (-) transformants (isolates V107 and V188) did not produce a 55 kDa fusion protein (Figure 3.9 A and C). Western blot analysis also showed that all transformants and the recipient strain CSlO-N2 produced a 28 kDa Ver-l and that the wild-type strain SU-l produced a 28 kDa functional Ver-l (Figure 3.9 A and B). The analysis did not show any degradation products of EGFP-tagged Ver-l fusion protein using anti-EGFP antibody (Figure 3.9 C and D). These data indicated that EGFP-tagged Ver-l fusion protein is produced intact and full-length from the transformants analyzed. Western blot analysis was conducted on transformants carrying pAPCGFPLVFNB. One AF (+) and EGF P (+) transfonnant (isolate LV70) produced a 55 kDa fusion protein (Figure 3.10 B and D). Also, 2 AF (-) and EGFP (+) transformants (isolates LV8 and LV29) produced a 55 kDa fusion protein (Figure 3.10 A and C). However, 1 AF (-) and EGF P (-) transfonnant (isolate LV6) and 5 AF (+) and EGF P (-) transformants (isolates LV13, LV44, LV140, LV155, and LV169) did not produce a 55 kDa fusion protein (Figure 3.10 A to D). All transformants and the recipient strain C 810- N2 produced a 28 kDa Ver-l and the wild-type strain SU-l produced a 28 kDa functional Ver-l as previously shown (Figure 3.10 A and B). Again, the analysis did not show any degradation products of EGFP-tagged Ver—l fusion protein using anti-EGFP antibody (Figure 3.10 C and D). These data indicated that EGFP—tagged Ver-l fusion protein is produced intact and full-length in transformants. Western blot analysis also showed that in transformants carrying pAPNGFPVFNB, 7 AF (+) and EGFP (+) transformants (isolates NV27, NV60, NV67, NV79, NV165, NV195, and NV218) produced a 55 kDa fusion protein (Figure 3.11 A, B, 131 Figure 3.10. Western blot analysis of EGFP-tagged Ver-l from transformants carrying pAPCGFPLVFNB, the recipient strain CSlO-N2, and the wild-type strain SU-l. Fungal proteins were extracted from transformants, CSlO-N2, and SU-l grown in 100 ml of YES for 48 h at 30 °C with shaking at 150 rpm. Approximately 30-50 ug of proteins were separated by 12% SDS—PAGE, transferred onto PVDF membranes, and probed with (A) and (B) Ver-l polyclonal antibody, or (C) and (D) EGFP polyclonal antibody. EGFP— tagged Ver-l has a molecular mass of 55 kDa, and the 28 kDa protein represents Ver-l (with the exception of a functional Ver-l in SU-l). A Benchmark prestained protein ladder was used as a molecular mass marker (M). Recombinant EGFP (rEGF P) was used as a positive control and, the 30 kDa rEGFP contains a 27 kDa EGF P fused to a 3 kDa protein for affinity chromatography purification. 132 170.8 109.5 78.9 60.4 47.2 35.1 «55 kDa 24.9 " 28 kDa 18.3 13.7 . ‘ 170.8 109.5 78.9 60.4 47.2 35.1 in" '31 +55kDa 24.9 t;~“1'-"""~'-—i-_ . «28km 18.3 13.7 133 Figure 3.10. (cont’d) 170.8 109.5 78.9 60.4 47.2 35.1 24.9 18.3 13.7 170.8 109.5 78.9 60.4 47.2 35.1 24.9 18.3 13.7 >wa 480982 $49 090‘“; $366.4 "" " “*‘55kDa ”M*'30kDa 84 g 9 9: °l 3 \- e s ’s Q} a?“ A~ AN” A‘ 29 Q? 9 0 v v v V s 4‘ w *55kDa "" *30kDa 134 Figure 3.11. Western blot analysis of EGFP-tagged Ver-l from transformants carrying pAPNGFPVFN B, the recipient strain CS10-N2, and the wild-type strain SU-l. Fungal proteins were extracted from transformants, CSlO-N2, and SU-l grown in 100 ml of YES for 48 h at 30 0C with shaking at 150 rpm. Approximately 30-50 ug of proteins were separated by 12% SDS-PAGE, transferred onto PVDF membranes, and probed with (A), (B), and (C) Ver—l polyclonal antibody, or (D), (E), and (F) EGF P polyclonal antibody. EGFP-tagged Ver-l has a molecular mass of 55 kDa, and the 28 kDa protein represents Ver-l (with the exception of a functional Ver-l in SU-l). A Benchmark prestained protein ladder was used as a molecular mass marker (M). Recombinant EGF P (rEGFP) was used as a positive control and, the 30 kDa rEGFP contains 27 kDa EGFP fused to a 3 kDa protein for affinity chromatography purification. 135 1 70.8 109.5 78.9 60.4 f I I I 7 I I W gun-I <- 55 11011 47.2 35.1 24.9 i 18.3 13.7 170.8 109.5 78.9 60.4 4...... W“ 65—- 4..." W ---" ' w ' ‘- 55 kDa 47.2 35.1 W" “ +281t8a 24.9 18.3 13.7 136 Figure 3.1 1. (cont’d) 170.8 109.5 78.9 60.4 “ 55 kDa 47.2 35.1 24.9 can. “a” n G— 28 kDa 18.3 13.7 kDa 170.8 ~ ' 109.5 .. ~ - . a 78.9 = 60.4 “-4 3...... ‘7 55 kDa 47.2 - 35.1 24.9 . — 18.3 13.7 137 Figure 3.11. (cont’d) 170.8 . ’ 1" 109.5 - , 78.9 60.4 . ,2. : , p 55 km 47.2 , .. 35.1 - ‘~ 30 kDa 24.9 . 18.3 13.7 24.9 18.3 13.7 138 D, and E). Also, 2 AF (-) and EGF P (+) transformants (isolates NV170 and NV196) produced a 55 kDa fusion protein (Figure 3.11 C and F). However, 3 AF (-) and EGFP (-) transfonnant (isolates NVl, NV63, and NV109) did not produce a 55 kDa fusion protein (Figure 3.11 A and D). Western blot analysis also showed that all 7 AF (+) and EGFP (+) transformants and the recipient strain CS 1 0-N2 produced a 28 kDa Ver-l but all 3 AF (-) and EGFP (-) transformants produced very small amounts of a 28 kDa Ver—l, and the wild-type strain SU-l produced a 28 kDa functional Ver-l (Figure 3.11 A to C). Identification of a point mutation in ver-I A in C810-N2 The recipient strain CSlO-N2 (ver-I , niaD, pyrG, wh-I) was originally derived from ATCC3 6537 (ver-I , wh-I), which accumulates the aflatoxin pathway intermediate versicolorin A (VER A or VA). ATCC36537 was in turn derived from the wild-type strain SU-l by UV. irradiation (Bennett and Goldblatt, 1973; Lee et al. , 1975). To determine the mutation site in ver-I A in CS10-N2, PCR was performed and DNA sequencing of the PCR products was conducted. DNA sequencing showed a point mutation site (G —> A) at nucleotide residue 287 in ver-I A in CS10-N2 (Figure 3.12). These data indicate that ver-I A in CSlO-N2 contains a point mutation and encodes a non-functional Ver-l due to a change to the negatively charged amino acid glutamic acid from the uncharged amino acid glycine. Analysis of expression of EGFP-tagged Ver—l fusion protein carrying a functional wild-type Ver-l protein The recipient strain CSlO-N2 produces a non-functional Ver-l due to a point 139 Nucleotide sequence SU-l 277 tcgaacgctggaattgtatcg 297 CSlO-N2 277 tcgaacgctgaaattgtatcg 297 Amino acid sequence SU-l 93 Q 99 tn—tn 2-—-Z ?-—-F p]. H—H <:—< m—tn CSlO-NZ 93 99 Figure 3.12. Comparison of selected nucleotide and amino acid sequences of ver-I A from SU-l and CSlO—N2. Identical nucleotides or amino acids are indicated by lines and a point mutation is indicated by a dot. Nucleotide and amino acid sequences were aligned with the EMBOSS alignment tool (hgp://www.ebi.ac.uk/emboss/aligrl/index. html). Abbreviations for amino acids are as follows; 8, serine; N, asparagine; A, alanine; G, glycine; I, isoleucine; V, valine; E, glutamic acid. 140 mutation in ver-I A as described above (Figure 3.12). Therefore, transformants could theoretically produce either EGFP-tagged functional wild-type Ver-l fusion protein or EGFP-tagged non-functional Ver-l fusion depending on the plasmid integration site into the chromosome (Figure 3.13). To determine if the EGFP-tagged Ver-l fusion protein carried a fuctional wild-type Ver-l protein, PCR was performed with the egp primer and the 5’ or 3’ ver-I A primer (Figure 2.9, Table 2.2, and Table 3.2). DNA sequencing of the amplified ver-I A gene region fused to the egfp gene was performed. DNA sequence analysis showed that isolate V86 (AF [+] and EGF P [+]) carried wild-type ver-I A while isolate V152 (AF [+] and EGF P [+]) carried non-functional ver-I A (data not shown). Also, DNA sequence analysis showed that isolates NV27, NV60, NV67, and NV79 (AF [+] and EGFP [+]) transformed with pAPNGFPVFNB carried wild-type ver-I A (data not shown). PCR products for DNA sequence analysis could not be obtained from isolate LV70 (AF [+] and EGF P [+]) transformed with pAPCGFPLVFNB. Determination of integration sites of pAPCGFPVFNB, pAPCGFPLVFNB, and pAPNGFPVFNB within the chromosome In order to determine the integration sites of pAPCGFPVFNB, pAPCGFPLVFNB, and pAPNGFPVFN B, Southern hybridization and PCR analyses were performed. Each plasmid could theoretically be integrated into the chromosome by homologous recombination at five sites: niaD, 3’ ver-I terminator within the ver-I A or ver-I B locus, or the 5’ ver-I promoter/ ORF (open reading frame) region within the ver-I A or ver-I B locus (Figure 3.14). Southern hybridization analysis showed that isolates V86 and V152 (AF [+] and EGF P [+]) carrying pAPCGFPVFNB had the plasmids 141 Figure 3.13. Schematic for production of EGFP-tagged functional Ver-l fusion protein depending on the integration site of pAPCGFPVFNB into the ver-I A locus. (A) ver-I A locus. (B) Integration upstream of the point mutation in the ver-I A gene including 5’ ver-I A. (C) Integration downstream of the point mutation in the ver-I A gene. (D) 3’ ver-I A integration. Integration of pAPCGFPVFNB upstream of the point mutation in ver-I A gene including 5’ ver-I A or into 3’ ver-I A results in production of EGFP-tagged functional Ver-l fusion protein. Integration of pAPCGF PLVFN B into the ver-I A locus also results in the same pattern of fusion protein production as that of pAPCGFPVFNB. Integration of the plasmid pAPNGFPVFNB downstream of the point mutation in ver-I A gene including 3’ ver-I A or into 5’ ver-I A results in production of EGFP-tagged functional Ver-l fusion protein. M represents a point mutation. Abbreviations for the DNA fragments are as follows; ver-I p, ver-I promoter; ver-I t, ver-I terminator. 142 Amp” ver-I terminator ‘ ‘ 88mg pAPCGFPVFNB (14933 hp) f ' ver-I promoters. lgene niaD A. ver-I A locus M ver-I p ver-I A ver-I t 143 Figure 3.13. (cont’d) B. Integration upstream of the point mutation in the ver-I A gene including 5’ ver-I A M W; g _. n ver-I p ver-I A egfp ver-I t amp niaD ver-l p ver-I A ver-I t Ver—1+"' EGFP + Ver-l ' C. Integration downstream of the point mutation in the ver-I A gene M ver-I p ver-I A egfp ver-I t amp niaD ver-I p ver-l A ver-I t Ver-l '— EGFP + Ver-l + D. 3’ ver-I A integration M . ver-I p ver-I A ver-I t amp niaD ver-I p ver-I A egfp ver-I t Ver-l ' Ver-l +— EGFP + 144 Figure 3.14. Schematic for Southern hybridization and PCR analyses of integration sites of pAPCGFPVFNB or pAPCGFPLVFNB into the chromosome. (A) ver-I A locus. (B) 3’ ver-I A integration. (C) 5’ ver-I A/ ORF region integration. (D) ver-I B locus. (E) 3’ ver-I B integration. (F) 5’ ver-I B/ ORF region integration. (G) niaD locus. (H) niaD integration. Genomic DNA was digested with Pstl and probed with an egfp gene. The restriction enzyme sites, probes, and expected band sizes in Southern hybridization analysis are shown. The primer positions in PCR analysis are also shown. In Southern hybridization analysis, integration of the plasmid pAPNGFPVFNB at 3’ ver-I A results in the same pattern as its integration at the ver-I A ORF region. The plasmid contains a 0.6 kb ver-I promoter instead of the 1.1 kb ver-I promoter, resulting in a 7.0 kb band for its integration at 3’ ver-I A or B, or niaD. In PCR analysis of integration sites of pAPNGFPVFNB, 3’ ver-I A integrants produce a larger PCR product for the 0.9 kb ver-I A gene fragment than that from pAPCGFPVFNB or pAPCGFPLVFNB. Abbreviations for the restriction enzyme sites and DNA fragments are as follows; P, PstI; Xh, Xhal; ver-I p, ver-I promoter; ver-I t, ver-I terminator. 145 A. ver-l A locus ver-I p ver-l A ver-I t B. 3’ ver-I A integration PCR I-> 4—1 P P P P P " I [EA ver-I p ver-I A ver-I t amp niaD ver-I p ver-I A egfp ver-I t (probe) I J I 7.4 kb I C. 5’ ver-I A/ ORF region integration I" P P II’ P ver-l p ver-I A egfp ver-I t amp niaD ver-I p ver-I A ver-I t (probe) 1 ' 4.8 kb I 146 Figure 3.14. (cont’d) D. ver-I B locus P ._—__L vet-1 p ver-I B ver-I t E. 3’ ver-I B integration PCR H H I" P P II) P I M‘ '7 .' ' . I - i ver-I p ver-I B ver-I tamp niaD ver-I p ver-I B egfp ver-I t (probe) 1 l I 7.4 kb I F. 5’ ver-I Bl ORF region integration 1" P P 1" P ver-I p ver-I B egfp ver-I t amp niaD ver-I p ver-I B ver-I t (probe) 1 l I 8.9 kb I 147 Figure 3.14. (cont’d) G. niaD locus H. niaD integration P P P P . r. . _ . , .‘x t 7* .. . .' ~ >.... .. ~\A -.7‘ 4- g , -- niaD ‘ . g , I. A il m. I“; , )— T amp ver-I t egfp ver-I B ver-I p niaD (probe) 1 J I 7.4 kb I 148 integrated at 3’ ver-I terminator within the ver-I A or ver-I B locus, or niaD locus and that isolate V152 had another plasmid integrated at 5’ ver-I A/ ORF region (Figure 3.15 A). PCR analysis of isolate V86 produced a 2.6 kb band with the egp primer and the 3’ ver-I A primer (Figure 3.14 and Table 3.2), indicating that the plasmid was integrated into the 3’ ver-I terminator at the ver-I A locus (Figure 3.15 B). Isolate V152 did not produce any bands in the same assay with the eg/[D primer and the 3’ ver-I A or B primer (Figure 3.14 and Table 3.2), indicating that the plasmid was integrated into the niaD locus (Figure 3.15 B and data not shown). Southern hybridization and PCR analysis data suggest multiple integration of the plasmid in isolate V152. Southern hybridization and PCR analyses showed that isolates V107 and V188 (AF [+] and EGF P H) did not produce any bands, indicating a gene replacement event at the non-functional ver-I A or B genes (Figure 3.15 A and B). The analyses also showed that in isolate V1 (AF H and EGFP [+]) the plasmid was integrated at the 5’ ver-I A/ ORF region and that in isolate V32 (AF H and EGF P [+]) it was integrated at the 3’ ver-I A (Figure 3.15 A and B). Isolate V2 (AF H and EGFP [-]) produced a 7.4 kb band in Southern hybridization analysis but did not produce any bands with the egfp primer and the 3’ ver-I A or B primer (Figure 3.14 and Table 3.2), indicating that the plasmid was integrated into the niaD locus by single-crossover (Figure 3.15 A and B, and data not shown). Southern hybridization and PCR analyses showed that in isolate LV70 (AF [+] and EGF P [+]) transformed with pAPCGFPLVFNB, one plasmid was integrated into 3’ ver-I A terminator and another plasmid was integrated into 5’ ver-I A promoter (Figure 3.16 A and B). Southern hybridization and PCR analyses showed that isolates LV13, 149 Figure 3.15. Southern hybridization and PCR analyses of integration sites in transformants carrying pAPCGFPVFNB. For Southern hybridization analysis, genomic DNA was isolated from A. parasitisus, digested with PstI, and hybridized with an egfp gene probe. For PCR analysis, genomic DNA was amplified with the egfp primer and the 3’ ver-I A primers (Figure 3.14 and Table 3.2). (A) Southern hybridization analysis of pAPCGFPVFNB. 3’ ver-I A, B, or niaD integrants resulted in a 7.4 kb band and 5’ ver—I A/ ORF region integrants resulted in a 4.8 kb band. (B) PCR analysis of 3’ ver-I A, 5’ ver-I A/ ORF region, or niaD integrants of pAPCGFPVFNB. 3’ ver-I A integrants resulted in a 2.6 kb band. The recipient strain CSlO-NZ was used as a negative control. CSlO-N2 (control) was used as a positive control. For the positive control, the egfp primer was replaced with the ver-I A primer to generate the same fragment size as those in 3’ ver-I A integrants (Figure 3.14). k/HindIII was used as a molecular size marker. 150 N V ’\ % W 09‘ 4"! 4“ 4'5” 4N§ 4‘5 4%6 44, 7.4 kb " 4.8 kb "’ B. o“ 9 6° Wywéwg’é» (\ ¢§ » w ‘50 0°" (55‘ A” 4‘ 4’3" 4‘“ 46‘4“? 4‘6 ~39 « 23.1 kb .— 9.4 kb ‘- 6.6 kb « 4.4 kb 2.6 kb-’ U'G' 151 Figure 3.16. Southern hybridization and PCR analyses of integration sites in transformants carrying pAPCGFPLVFNB. For Southern hybridization analysis, genomic DNA was isolated from A. parasitisus, digested with PstI, and hybridized with an egfp gene probe. For PCR analysis, genomic DNA was amplified with the egfp primer and the 3’ ver-I A primer (Figure 3.14 and Table 3.2). (A) Southern hybridization analysis of pAPCGFPLVFNB. 3’ ver-I A, B, or niaD integrants resulted in a 7.4 kb band and 5’ ver-I A/ ORF region integrants resulted in a 4.8 kb band. (B) PCR analysis of 3’ ver-I A, 5’ ver-I A/ ORF region, or niaD integrants of pAPCGFPLVFNB. 3’ ver-I A integrants resulted in a 2.6 kb band. The recipient strain CSlO-N2 was used as a negative control. CSlO-N2 (control) was used as a positive control. For the positive controls the em primer was replaced with the ver-I A primer to generate the same fragment sizes as those in 3’ ver-I A integrants (Figure 3.14). k/HindIII was used as a molecular size marker. 152 é» nag Veefi">9‘")‘°«“ 69444444444 7.4 kb" ‘1" fl 4. + = » » 8kb 1” * B. \\ c°° ‘ <04” see 9‘ \°$’@’4§4% '9 ¢§~h~$~b A“ w etoscmvéc‘dddéde" 2.6 kb” 153 LV44, LV140, LV155, and LV169 (AF [+] and EGFP [-]) did not produce any bands, indicating a gene replacement at the non-functional ver-I A or B gene (Figure 3.16 A and B). The analyses also showed that in isolate LV8 (AF H and EGFP [+]) the plasmid was integrated at 5’ ver-I A/ ORF region and that in isolate LV29 (AF H and EGFP [+]) it was integrated at 3’ ver-I A (Figure 3.16 A and B). Isolate LV6 (AF H and EGFP [-]) did not produce any bands in Southern hybridization and PCR analysis, indicating a gene replacement event at the niaD locus (Figure 3.16 A and B). Southern hybridization and PCR analyses showed that in isolates NV27, NV60, NV67, NV79, NV165, NV195, and NV218 (AF [+] and EGFP [+]), pAPNGFPVFNB was inegrated into 3’ ver-I A terminator and that in isolate NV218 a second plasmid was integrated into 5’ ver-I B promoter/ ORF region (Figure 3.17 A and C). The analyses also showed that in isolates NV170 and NV196 (AF H and EGFP [+]) the plasmid was integrated at 3’ ver-I A/ ORF region and that in isolate NV170 a second plasmid was integrated at 5’ ver-I A (Figure 3.17 B and D). These data suggest multiple integration of the plasmid in isolate NV170. Isolates NVl, NV63, and NV109 (AF H and EGFP [-]) produced a 7.0 kb band in Southern hybridization analysis but did not produce any bands with the eg/p primer and the 3’ ver-I A or B primer (Figure 3.15 and Table 3.2), indicating that the plasmid was integrated into the niaD locus by single-crossover (Figure 3.17 A and C, and data not shown). Southern hybridization analysis also showed that in isolate NV109 (AF H and EGF P H) a second plasmid was integrated at 5’ ver-I A (Figure 3.17 B). These data suggest multiple plasmid integrations in NV109. l54 Figure 3.17. Southern hybridization and PCR analyses of integration sites in transformants carrying pAPNGFPVFNB. For Southern hybridization analysis, genomic DNA was isolated from A. parasitisus, digested with PstI, and hybridized with an egfp gene probe. For PCR analysis, genomic DNA was amplified with the egfp primer and the 3’ ver-I A primer (Figure 3.14 and Table 3.2). (A) and (B) Southern hybridization analysis of pAPNGFPVFNB. 3’ ver-I A or B/ ORF region, or niaD integrants resulted in a 7.0 kb band, 5’ ver-I A integrants resulted in a 4.8 kb band, and 5’ ver-I B integrants resulted in a 8.9 kb band. (C) and (D) PCR analysis of 3’ ver-I A/ ORF region, 5’ ver-I A, or niaD integrants of pAPNGFPVFNB. 3’ ver-I A/ ORF integrants resulted in a 3.5 kb band. The recipient strain CSlO-N2 was used as a negative control. A/HindIII was used as a molecular size marker. 155 156 Figure 3.17. (cont’d) “23.1 kb ‘— 9.4 kb *- 6.6 kb 35kb " 4.4kb " 2.3 kb “ 2.1kb & w c‘ ‘95 «6 «D 95 $5 9“ os‘é’s“ e“ 3.5 kb" 157 Confocal Laser Scanning Microscopy (CLSM) To identify the sub-cellular location of C-terminally or N-terminally EGFP-tagged Ver-l in hyhae of AF (+) and EGFP (+) transformants V86 or NV27, confocal laser scanning microscopy (CLSM) was performed after 24, 48, and 72 h culture on aflatoxin- inducing media, YES. EGF P fluorescence was not detected in V86 at 24 h (Figure 3.18 C). However, C-terminally EGFP-tagged Ver-l was localized in vacuoles and the cytoplasm of V86 at 48 and 72 h when either the vacuolar membrane staining dye FM 4- 64 or the vacuolar lumen staining dye CMAC was used (Figure 3.18 D to G). This vacuole localization of EGFP-tagged Ver-l was observed using the vacuolar membrane dye FM 4-64 and it was confirmed using the vacuolar lurnen-specific dye CMAC; it is known that FM 4-64 stains endosomal compartments like endosomes as well as vacuoles in Saccharomyces cerevisiae (Vida and Emr, 1995). Also, V86 did not produce any EGF P fluorescence when it was cultured on non-aflatoxin-inducing media, YEP (Figure 3.18 H). The pattern of localization of N-terminally EGFP-tagged Ver-l in NV27 was similar to that in V86. EGFP fluorescence was not detected in NV27 at 24 h (Figure 3.19 A). However, N—terminally EGFP-tagged Ver-l was localized in vacuoles and the cytoplasm of NV27 at 48 and 72 h when either the vacuolar membrane staining dye FM 4-64 or the vacuolar lumen staining dye CMAC was used (Figure 3.19 B to G). Also, NV27 did not produce any EGFP fluorescence when it was cultured on non-aflatoxin- inducing media, YEP (Figure 3.19 H). In Chapter 2, EGFP was localized in the cytoplasm of EGFP (+) transfonnant B3-15 at 48 h but localized in vacuoles at 72 h. Therefore, to compare vacuole localization of EGFP in B3-15 with that of EGFP-tagged Ver-l in V86 and NV27, green fluorescent vacuoles were counted at 48 and 72 h. 158 Figure 3.18. Sub-cellular localization of C-terminally EGFP-tagged Ver-l in AF (+) and EGF P (+) transfonnant V86. Fungal vacuoles were stained with 8 pM FM 4-64 or 10 uM CMAC and observed using a Zeiss LSM 5 Pascal or Zeiss LSM 510 Meta confocal laser scanning microscope (CLSM) after 24, 48, and 72 h incubation at 30 0C on aflatoxin-inducing media, YES or non-aflatoxin-inducing media, YEP, plus 20 mM uracil agar blocks. (A) and (B) The recipient strain CSlO-N2 (negative control) stained with CMAC at 48 h on YES plus 20 mM uracil. Blue fluorescent vacuoles were observed in hyphae and conidiophores but green fluorescence was not detected. (C) V86 stained with FM 4-64 at 24 h on YES plus 20 mM uracil. Red fluorescent vacuolar membranes were observed in hyphae but green fluorescence was not detected. (D) and (E) V86 stained with F M 4-64 at 48 h on YES plus 20 mM uracil. EGFP-tagged Ver-l was localized in vacuoles in panel (D) and in the cytoplasm in panel (B). (F) V86 stained with CMAC at 48 h on YES plus 20 mM uracil. EGFP-tagged Ver-l was localized in vacuoles in hyphae and conidiophores. (G) V86 stained with FM 4-64 at 72 h on YES plus 20 mM uracil. EGFP-tagged Ver—l was localized in vacuoles and the cytoplasm. (H) V86 stained with FM 4-64 at 72 h on YEP plus 20 mM uracil. Red fluorescent vacuolar membranes were observed in hyphae but green fluorescence was not detected. Each panel shows a red fluorescence image (FM 4-64) or a blue fluorescence image (CMAC) (top left), a green fluorescence image (EGFP) (top right), a transmitted image (bright field or differential interference contrast [DIC]) (bottom left), and a merged image (bottom right). Scale bars, 10 um. Images in this dissertation are presented in color. 159 A. The recipient strain CSlO—N2, YES plus 20 mM uracil, 48 h, CMAC 10 pm ' . ‘ x 11] um B. The reci sient strain CSlO-N2 ‘ YES lus 20 mM uracil 48 h, CMAC 1—1 1 [I urn . 10 pm 160 Figure 3.18. (cont’d) 24 h, FM 4-64 C. V86, YES plus 20 mM uracil, Figure 3.18. (cont’d) E. V86, YES plus 20 mM uracil, 48 h, FM 4-64 F. nlus 20 mM uracil 48 h CMAC Figure 3.18. (cont’d) G. V86, YES plus 20 mM uracil, 72 h, FM 4—64 Figure 3.19. Sub-cellular localization of N—terminally EGFP-tagged Ver-l in AF (+) and EGFP (+) transformant NV27. Fungal vacuoles were stained with 8 uM FM 4-64 or 10 uM CMAC and observed using a Zeiss LSM 5 Pascal or Zeiss LSM 510 Meta confocal laser scanning microscope (CLSM) after 24, 48, and 72 h incubation at 30 0C on aflatoxin-inducing media, YES or non-aflatoxin-inducing media, YEP, plus 20 mM uracil agar blocks. (A) NV27 stained with F M4-64 at 24 h on YES plus 20 mM uracil. Red fluorescent vacuolar membranes were observed in hyphae but very little green fluorescence was detected. (B), (C), and (D) NV27 stained with FM 4-64 at 48 h on YES plus 20 mM uracil. EGFP-tagged Ver-l was localized in vacuoles in panel (B) and in the cytoplasm in panel (C). Green fluorescence was observed in the vacuolar ltunen in panel (D) as shown by an extended focus image of Z-stacks (Z-section interval: 0.46 pm). (E) NV27 stained with CMAC at 48 h on YES plus 20 mM uracil. EGFP-tagged Ver-l was localized in vacuoles and the cytoplasm. (F) and (G) NV27 stained with FM 4-64 at 72 h on YES plus 20 mM uracil. EGFP-tagged Ver-l was localized in vacuoles. EGFP- tagged Ver-l was associated with the vacuolar membrane in panel (G). (H) NV27 stained with FM 4-64 at 72 h on YEP plus 20 mM uracil. Red fluorescent vacuolar membranes were observed in hyphae and conidiophores but green fluorescence was not detected. Each panel shows a red fluorescence image (FM 4-64) or a blue fluorescence image (CMAC) (top left), a green fluorescence image (EGFP) (top right), a transmitted image (bright field or differential interference contrast [DIC]) (bottom left), and a merged image (bottom right) except for panel (D) in which only a merged image is shown. Scale bars, 10 um. Images in this dissertation are presented in color. 164 A. NV27, YES plus 20 mM uracil, 24 h, FM 4-64 165 Figure 3.19. (cont’d) C. NV27, YES plus 20 mM uracil, 48 h, FM 4-64 166 Figure 3.19. (cont’d) E. NV27, YES plus 20 mM uracil, 48 h, CMAC F. NV27, YES plus 20 mM uracil, 72 h, FM 4-64 1 1 1 Figure 3.19. (cont’d) G. NV27, YES plus 20 mM uracil, 72 h, FM 4-64 168 In V86 and NV27, 72 % and 80 % of vacuoles showed localization of green fluorescence at 48 h, respectively while in B3-15 only 13 % of the vacuoles showed localization of green fluorescence at 48 h (Table 3.3). Pairwise multiple comparison confirmed that the percentage of vacuole localization of green fluorescence in B3-15 at 48 h is significantly less than in V86 and NV27 at 48 h (P<0.05). However, V86, NV27 and B3-15 showed 78.5 %, 86.5 %, and 62.5 % of vacuole localizations of green fluorescence at 72 h, respectively (Table 3.3). Time-course of AFB] production The above data (Table 3.3) of vacuole localization of green fluorescence suggest that V86 was under starvation conditions at 72 h. To confirm starvation conditions in slide culture at 72 h incubation, EGFP-tagged Ver-l producing transfonnant V86 and EGFP producing transfonnant B3-15 were cultured in YES (plus 20 mM uracil) media and on YES (plus 20 mM uracil) agar blocks. Dry weight and AF B1 production of the transformants were then analyzed after 24, 48, and 72 h incubation. Dry weights of these transformants were similar at each time point in liquid culture (Figure 3.20). A transition from active growth to stationary phase was observed between 48 and 72 h in liquid culture as described previously (Liang et al. , 1997; Chiou et al., 2002). AFB] was not detected in the transformants at 24 h, but high levels of AFB1 were detected at 48, and 72 h in both liquid and slide cultures as described previously (Figure 3.20 and 3.21) (Liang et al., 1997; Chiou et al., 2002). Therefore, these data strongly suggest that starvation conditions occur at 72 h incubation in both liquid and slide cultures. 169 Table 3.3. Comparison of vacuole localization of EGFP in EGFP (+) transfonnant 83-15 with that of EGFP-tagged Ver-l in AF (+) and EGFP (+) transformants V86 and NV27. * * * Time (h) 133-15 (% ) V86 (% ) NV27 (% ) 48 13.0%» 71013.0 80.0:I:l.0 72 62.5:I:3.5 7s.5¢1.5 86.5:t4.5 * % = # of green fluorescent vacuoles/ # of large- and mid-size of vacuoles (> 5 pm) in 30 microscopic fields (2 - 3 hyphae per field) 170 Time-course of transformants 83-15 and V86 5 30 —o— 3345 (D.W) . o V86(D.W) + 113-15mm!) 1 - . A V86(AFBI) 0.5 A ~ 20 3 i a ‘50 0.1 - 3 .3 V 3 0.05 E t“ u. a < » 10 0.01 « 0.005 0 b I I T 0 o 20 4o 60 so Time (hr) Figure 3.20. AF81 production in transfonnant V86 (EGFP-tagged Ver-l) and transfonnant B3-15 (EGFP). AF B1 was measured after 24, 48, and 72 h incubation at 30 0C with shaking at 150 rpm in YES medium. 171 Time-course of the transformants 83-15 and V86 in slide culture 100 + B3-15(AFBI) so _ ---n V86(AFBI) E 60 ‘ E = V E < 4° ’ 20 « 0 G x — I r 0 20 40 60 80 Time (br) Figure 3.21. AF B] production in transfonnant V86 (EGFP-tagged Ver-l) and transfonnant B3-15 (EGFP) in slide culture. AF Bl was measured after 24, 48, and 72 h incubation at 30 0C on YES agar. 172 DISCUSSION Our experiments demonstrated that N-terminally or C-terminally EGFP-tagged Ver-l fusion protein was enzymatically functional in A. parasiticus and that the Ver—l fusion protein was localized in the cytoplasm and vacuoles of hyphae at 48 and 72 h. Initially we used a C-terminally egfp-tagged ver-I plasmid (pAPCGFPVFNB) for fungal transformation. The hinge region between the ver-I and eg/p coding regions contained 5 amino acids and was designed to eliminate steric hindrance between the two proteins in the fusion and to promote efficient folding of the fusion protein. We decided to use 5 amino acids in the hinge region because other researchers have used 2-10 amino acids for this purpose (Valdez-Taubas et al., 2000; Bafiuelos et al., 2002). We used a C-terminally egfirtagged ver-I plasmid instead of an N- terminally egfiJ-tagged ver-I plasmid because we predicted that C-terminal EGF P would be more efficient at correct fusion protein folding. However, we did not identify any AF (+) and EGFP (+) transformants in this initial experiment. We speculated either the hinge region might not provide sufficient space between two proteins or that the C—terminus of Ver—l is critical to its activity or transport. Tavoularis et al. (2001) reported that the number of amino acids in the hinge region is critical for correct expression and translocation of the C-terminally SGFP-tagged fusion protein; in this study the fusion protein contained 4 amino acids in the hinge region and was functional. Therefore, we constructed 2 additional plasmids, one (pAPCGFPLVFNB) in which the hinge region contained 13 amino acids and the other (pAPNGFPVFNB) in which egfp was fused to the N-terminus of ver-I and the hinge region contained 7 amino acids. A subsequent fungal transformation with 173 pAPCGFPVFNB produced AF (+) and EGF P (+) transformants, indicating that 5 amino acids provide enough space between the two proteins in C-terminally EGFP-tagged Ver-l. TLC and ELISA analyses showed that in all AF (+) and EGF P (+) transformants, C-terminally and N-terminally EGFP-tagged Ver-l protein functionally complemented the non-functional Ver-l protein in A. parasiticus CSlO-N2 although the level of AF B1 production was lower in N-terminally EGFP-tagged Ver-l protein (Figure 3.4 to 3.8). ver-I encodes an NADPH-dependent oxidoreductase (Henry and Townsend, 2005) and we observed a conserved NADP binding motif (Gly-X-Gly-X-X-Ala-l 2X-Lys) starting at Gly, amino acid residue 18 in Ver-l (Skory et al., 1992). We speculate that this N- terminal functional domain may have been partially blocked by EGF P in N-terminally EGFP-tagged Ver-l, resulting in decreased Ver-l activity. TLC analysis was conducted on AF (-) and EGFP (+) transformants expressing C-terminally EGFP-tagged Ver—l which contains either a short or long hinge region. These transformants accumulated versicolorin A (VA) while AF (-) and EGFP (+) transformants expressing N-terminally EGFP-tagged Ver-l did not accumulate VA (Figure 3.4 A, 3.6 A, and 3.7 C). Although VA was converted to demethylsterigmatocystin (DMST) by N-terminally EGFP-tagged Ver-l in the AF (-) and EGFP (+) transformants, there might have been a blockage at any step downstream of the step catalyzed by Ver-l in aflatoxin pathway. Alternatively, there might have been a blockage at any step upstream of the step catalyzed by Ver-l in aflatoxin pathway. Western blot, Southern hybridization, and PCR analyses showed that all AF (-) and EGFP (+) transformants produced a 55 kDa fusion protein (Figure 3.9 A and C, 3.10 174 A and C, and 3.11 C and F) and that the plasmid was integrated into 3’ ver-I A or 5’ ver-I A/ ORF region in the transformants carrying pAPCGFPVFNB or pAPCGFPLVFNB and was integrated into 5’ ver-I A or 3’ ver-I A/ ORF region in the transformants carrying pAPNGFPVFNB (Figure 3.13, 3.14, and 3.15). The AF (-) and EGFP (+) phenotype might result from mislocalization of the fusion protein so that it is not transported to vacuoles for aflatoxin synthesis. Also, in Western blot analysis, all AF (~) and EGFP (-) transformants from N-terminally egfiv-tagged ver-I plasmid produced only a very faint 28 kDa non-functional Ver-l. This low level expression of the native Ver-l might result from decreased fimction of regulatory proteins like AflR. This possible explanation is expected from TLC results in which all AF (-) and EGFP (-) transformants did not accumulate VA nor did they produce aflatoxin (Figure 3.7 A). To determine if the EGFP-tagged Ver—l fusion protein carried fuctional wild-type Ver—l protein, PCR was conducted. However, no PCR products for DNA sequencing could be obtained from the AF (+) and EGF P (+) transfonnant LV70 carrying pAPCGF PLVFN B although several different primers or genomic DNAs were used. The reason remains unclear. DNA sequencing showed that glycine was mutated to glutamic acid at the deduced amino acid residue 96 of non-functional Ver-l in CSlO-N2 (Figure 3.12). It suggests that the amino acid change to the negatively charged amino acid from the uncharged amino acid may have generated conformational change to 3 dimentional structure of native Ver-l. Southern hybridization and PCR analyses showed that pAPCGFPVFNB in AF (-) and EGFP (-) transformant V2 was integrated into the niaD locus by single-crossover, 175 that pAPCGFPLVFNB in AF (—) and EGF P (-) transformant LV6 was integrated into the niaD locus by double-crossover (gene replacement), and that pAPNGFPVFNB in NV 1 , NV63, and NV109 was integrated into the niaD locus by single-crossover (Figure 3.15, 3.16, and 3.17). These data were consistent with previous observations in which no detectable aflatoxin gene activity was observed when the plasmid was integrated into either the niaD or pyrG locus outside of the aflatoxin gene cluster (Liang et al. , 1997; Chiou et al. , 2002). Also, Southern hybridization and PCR analyses showed that no plasmids were integrated into 3’ ver-I B locus. This result was expected because the ver—I B terminator shows the same nucleotide sequence as the ver-I A terminator only in a 0.56 kb region after the stop codon; this results in rare homologous recombination. Confocal laser scanning microscopy (CLSM) revealed cytoplasm and vacuole localization of N-terminally or C-terrninally EGFP-tagged Ver-l in hyphae of AF (+) and EGFP (+) transformants, V86 and NV27 at 48 h (Figure 3.18 D to F and 3.19 B to E). These data suggest that the Ver-l protein is synthesized in the cytoplasm and transported to vacuoles where aflatoxin is actively produced. In contrast, Lee et al. (2004) observed Ver—l primarily in the cytoplasm of 24-48 h old cells using transmission electron microscopy (TEM) after immunogold labeling. This observation might be explained because we used live samples in real time in this study instead of fixed samples in the previous study. Alternatively, the acidic pH (pH 5-6) of vacuoles may have negatively affected binding of Ver-l antibody to Ver-l localized in vacuoles in the previous study. However, vacuole localization of Ver-l was consistent with the previous observation that Ver-l was associated with structures similar in size to lysosomes and could be found in the cytoplasm fraction using cell fractionation (Liang, 1996). It was also consistent with 176 a recent observation using co-immunoprecipitation, which suggested that Ver-l, VBS, and OmtA may form a complex for aflatoxin synthesis (Chanda, unpublished data). Also, EGFP-tagged Ver-l was associated with vacuolar membranes as shown in figure 3.19 G, indicating fusion of the protein with the vacuolar membranes prior to transport into vacuolar lumen. Time-course experiments suggested that V86 was under starvation conditions in 72 h liquid and slide cultures (Figure 3.20 and 3.21). Therefore, these data suggest that EGFP-tagged Ver-l is transported to vacuoles by the Cvt pathway under conditions that promote active growth and later it is transported to vacuoles by autophagy under starvation conditions (Figure 3.22). In support of this idea we used prediction programs (WoLF PSORT, PSORT II, and iPSORT; http://www.psort.org) based on sequence analysis databases for fungi, yeast, animal, and plant to identify vacuole localization or vacuolar targeting signals in Ver-l. However, a possible vacuolar targeting motif was not found and possible sub-cellular localization sites were predicted as follows: cytoplasm, 60.9 %; mitochodria, 13.0 %; nucleus, 8.7 0/o; vacuole, 4.3 %; vesicles of secretory system, 4.3 %; extracellular including cell wall, 4.3 %; golgi, 4.3 %. Vacuolar sorting sequences from yeast or plant described in Chapter 1 were also not found in Ver-l. We speculate that the sequence analysis database for filamentous fungi is incomplete because a vacuolar targeting motif for OmtA was not found even though this protein is known to be localized in vacuole-like structures (Lee et al., 2004). Alternatively, aflatoxin proteins may use unique vacuolar sorting sequences. Also, CLSM data showed that AF (+) and EGFP (+) transformants, V86 and NV27 did not produce any EGFP fluorescence when they were cultured on non-aflatoxin-inducing media, YEP (Figure 3.18 H and 3.19 177 Figure 3.22. Schematic of Cvt and autophagy pathways for transport to the vacuoles. In the Cvt pathway, vacuolar resident enzymes like API form a Cvt complex (zymogen) in the cytoplasm. The Cvt complex is packaged into double membrane-bound Cvt vesicles under conditions that promote active growth. In contrast, under starvation conditions, the Cvt complex is taken up into larger double membrane-bound autophagosomes by the autophagy pathway. The Cvt vesicles or autophagosomes then fuse with the vacuole and the resulting Cvt bodies or autophagic bodies are broken down by vacuolar hydrolases to release zymogens into the vacuolar lumen. The zymogen is then processed into the mature enzyme. (Adapted from Baba et al., 1997) 178 H) and it was consistent with previous observations in which it was reported that peptone as a sole carbon source does not support aflatoxin gene expression and aflatoxin production (Skory et al., 1993; Buchanan and Lewis, 1984). In summary, N-terminally or C-terminally EGFP-tagged Ver-l protein fimctionally complemented the non-functional Ver—l in all AF (+) and EGFP (+) transformants. In transformants N-terminally EGFP-tagged Ver-l protein showed decreased Ver-l activity compared to wild-type strain SU-l while C-terminally EGFP- tagged Ver-l protein showed similar Ver-l activity as SU-l. Western blot analysis data indicated that the EGFP-tagged Ver-l fusion protein was produced intact and full-length form in all AF (+) and EGF P (+) transformants. Southern hybridization and PCR analyses showed that the plasmid was integrated into the ver-I A locus in all AF (+) and EGFP (+) transformants. Confocal laser scanning microscopy (CLSM) data indicated that N-terminally or C-terminally EGFP-tagged Ver-l was localized in the cytoplasm and vacuoles of fungal hyphae on aflatoxin-inducing solid media at 48 and 72 h and especially it was detected inside of the vacuoles. Time-course experiments strongly suggest that the AF (+) and EGFP (+) transformants were under starvation conditions in 72 h slide culture. Therefore, we concluded that Ver-l was synthesized in the cytoplasm and transported to vacuoles of fimgal hyphae by specific targeting for aflatoxin synthesis. 179 ACKNOWLEDGMENT We thank Dr. Melinda K. Frame and Dr. Shirley A. Owens (Center for Advanced Microscopy at Michigan State University) for help in confocal laser scanning microscopy, Dr. Pemg-Kuang Chang (U SDA/ARS Southern Regional Research Center) for providing the CSlO-N2 strain, and Dr. Stephen A. Osmani (Ohio State University) for providing a repeated Gly-Ala sequence for making egfia-tagged ver-I fusion plasmids. All experiments described in Chapter 3 were performed by Sung-Yong Hong. Chapter 3 will be submitted by combining with Chapter 2 for publication in the near future. 180 CHAPTER 4 SUB-CELLULAR LOCALIZATION OF THE NOR-l PROTEIN IN ASPERGILLUS PARASI TIC US USING AN EGFP FUSION ABSTRACT To identify the sub-cellular location of the early aflatoxin pathway enzyme Nor-1 in real time and to confirm previous observations by transmission electron microscopy (TEM) and cell fractionation, we developed plasmid constructs expressing N-terminally or C-terminally EGFP—tagged Nor-l fusion protein. The eg/p-tagged nor-I fusion plasmid carrying a niaD (encodes nitrate reductase) selectable marker was transformed into A. parasiticus B62 (nor-1, niaD, br-I). Transformants were screened for aflatoxin production using aflatoxin detection medium CAM (coconut agar medium) and for EGFP expression using fluorescence microscopy. Aflatoxin production was confirmed by TLC and ELISA. The data indicated that N-terminally or C-tenninally EGFP-tagged Nor-1 protein functionally complemented the non-functional Nor-l in all light or non-detectable orange-pigmented EGFP (+) transformants. In transformants N-terminally EGFP-tagged Nor-1 protein showed relatively similar levels of Nor-1 activity as the wild-type strain SU-l while C-terminally EGFP-tagged Nor-1 protein showed decreased levels of Nor-1 activity compared to SU-l. Production of full-length EGFP-tagged Nor-l fusion protein was analyzed by Western blot analysis using anti-Nor-l antibody or anti-EGFP antibody. The data indicated that the EGFP-tagged Nor-1 fusion protein was produced intact and full-length from all light or non-detectable orange-pigmented EGFP (+) transformants. 181 The plasmid integration sites were analyzed by Southern hybridization and PCR. The plasmid was integrated into the nor-I locus in all light or non-detectable orange- pigmented EGF P (+) transformants. Confocal laser scanning microscopy (CLSM) data indicated that N-terminally or C-terrninally EGFP-tagged Nor-l was localized in the cytoplasm and vacuoles of fungal hyphae on aflatoxin-inducing solid media at 48 and 72 h and especially it was detected inside of the vacuole. Time—course experiments strongly suggested that the light or non-detectable orange-pigmented EGFP (+) transformants were exposed to starvation conditions in 72 h slide culture. Therefore, we concluded that Nor-1 was synthesized in the cytoplasm and transported to vacuoles of fungal hyphae by specific targeting for aflatoxin synthesis. INTRODUCTION Aflatoxins (AF) are toxic and carcinogenic secondary metabolites synthesized primarily by the filamentous fungi Aspergillus parasiticus and A. flavus (Cotty et al., 1994; Bhatnagar et al. , 2003). Aflatoxin biosynthesis is a complex process that requires at least 24 enzyme activities (Dutton, 1988; Bhatnagar et al., 1992; Trail et al. 1995b; Yu et al., 1995; Yu et al. , 2004). The transport mechanism of aflatoxin biosynthetic enzymes to synthesize aflatoxin within a fungal cell is not completely understood. It was previously reported that several enzyme activities involved in aflatoxin B1 biosynthesis were detected in a microsomal fraction (Lax et al. , 1986; Bhatnagar et al. , 1989; Yabe et al., 1989, Yabe et al. , 1993; Yabe et al., 1999) while other enzymes were found in the cytoplasm or loosely bound to membranes (Lax et al. , 1986; Bhatnagar et al. , 1989; Yabe 182 and Hamasaki, 1993; Matsushima et al., 1994; Yabe et al., 1999; Sakuno et al., 2003). In previous sub-cellular localization studies, Nor-l was mainly detected in the cytoplasm and was associated with structures similar in size to ribosomes using cell fractionation (Zhou, 1997). This result was supported using TEM (Lee et al., 2004). It is currently known that nor-1 encodes an NADPH-dependent ketoreductase which is involved in conversion of norsolorinic acid (NOR or NA) to averantin (AVN) (Zhou and Linz, 1999). Norsolorinic acid (NOR) in the AF B1 biosynthetic pathway does not possess a bisfuran ring containing a double bond which generates mutation and cancer (Mori et al., 1985). Therefore, we speculate that it is not necessary that NOR is compartmentalized in specific organelles to protect cells from their toxicity during aflatoxin biosynthesis. Based on these data, we hypothesized that Nor-1 is localized in the cytoplasm. Therefore, we conducted confocal laser scatming microscopy (CLSM) using EGFP-tagged Nor-l in this study to identify the sub-cellular location of Nor-l in real time and to confirm previous observations using TEM after immunogold labeling and cell fractionation. The resulting information about the sub—cellular localization of Nor-l will provide important targets to block aflatoxin production in fungi. 183 MATERIALS AND METHODS Strains and culture conditions Escherichia coli DHSa F,e [F’/endAI hst17(rk- mk+) supE44 thi—I recAI gyrA (Nalr)relA1 A(IacZYA-argF)u169: (m80AlacZM15)] (Gibco BRL, Life Technologies Inc, Rockville, MD) was used to amplify plasmid DNA using standard procedures (Ausubel et al. , 2003). A. parasiticus B62 (nor-I , niaD, br-I) was used as the recipient strain for nor-1 neg/ti) fusion plasmids containing the niaD selectable marker. B62 was derived from A. parasiticus ATCC24690 (nor-I, br-I) by spontaneous mutation using potassium chlorate selection (Homg et al., 1990). ATCC24690 was in turn generated from A. parasiticus SU-l by UV. irradiation (Bennett and Goldblatt, 1973; Lee et al., 1971). A. parasiticus B3-15 generated in Chapter 2 was used as a control for time-course experiments. A. parasiticus strains used in this study are listed in Table 4.1. E. coli was cultured in LB medium (Luria-Bertani; 1 % tryptone, 0.5 % yeast extract, 1 % sodium chloride) for competent cell preparation and plasmid isolation (Maniatis etal., 1989). A. parasiticus was cultured in yeast extract—sucrose liquid medium (YES; 2 % yeast extract, 6 % sucrose, pH 5.8; batch fermentation) at 30 0C in the dark with shaking at 150 rpm for genomic DNA isolation, total protein extraction, measurement of mycelial dry weight, and analysis of aflatoxin concentration as described previously (Liang et al., 1997). Czapek-Dox medium (CZ; Difco Laboratories, Detroit, MI) supplemented with l % peptone was used for growth of the recipient strain B62 for transformation. CZ agar supplemented with 20 % sucrose as an osmotic stabilizer was 184 Table 4.1. Aspergillus parasiticus strains used in this study. Strain Genotypea Phenotype Source B62 nor-I, niaD, br-I AF (:1:) J. E. Linz N5 br-I; nor-1::egfp AF (i), EGFP (-) This study N7 br-I; nor-1::egfp AF (i), EGFP (+) This study N31 br-I; nor-1::egfp AF (:1:), EGFP (+) This study N43 br-I; nor-1::egfp AF (1:), EGFP (+) This study N51 br-I; nor-I neg/p AF (:t), EGFP (+) This study N60 br-I; nor-1::egfp AF (1:), EGFP (+) This study N70 br-I; nor-1::egfp AF (:t), EGFP (+) This study N128 br-I; nor-1::egfp AF (i), EGFP (+) This study LN2 br-I; nor-1::egfp AF (:lz), EGFP (-) This study LNl br-I; nor-1::egfp AF (21:), EGFP (+) This study LN6 br-I; nor-1::egfp AF (21:), EGFP (+) This study LN196 br-I; nor-1::egfp AF (+), EGF P (+) This study NNl br-I; egfp::nor—1 AF (:1:), EGFP (-) This study NN244 br-I; egfp::nor-1 AF (i), EGFP (-) This study NN145 br-I; egfpzmor-I AF (i), EGFP (+) This study NN183 br-I; egfp::nor-I AF (1:), EGFP (+) This study NN182 br-I; egfp::nor-I AF (+), EGFP (+) This study NN227 br-I; egfp::nor-I AF (+), EGFP (+) This study NN4 br-I; egfp::nor-I AF (+), EGFP (+) This study NN6 br-I; egfp::nor-I AF (+), EGF P (+) This study NN24 br-I; egfp::n0r-1 AF (+), EGFP (+) This study B3-15 ver-I p::egfp AF (+), EGF P (+) Chapter 2 a ver-I p represents ver-I promoter. :: represents gene fusions. 185 used as a selective medium for fungal transformation. Either potato dextrose agar (PDA; Difco Laboratories, Detroit, M1) or CZ agar was used for spore preparation. Coconut agar medium (CAM) was used for screening of aflatoxin producing fungal transformants under U.V at 365 nm (Chang et al., 1992) and for EGF P producing transformants using a Nikon Eclipse E600 fluorescence microscope (Nikon Inc., Melville, NY). YES agar (1.5 % agar) was used for extraction of aflatoxin and aflatoxin intermediates for TLC and ELISA analyses. Either YES or YEP (2 % yeast extract, 6 % peptone, pH 5.8) agar was used for slide culture which was observed using a Zeiss LSM 5 Pascal or Zeiss LSM 510 Meta confocal laser scanning microscope (CLSM) (Carl Zeiss Inc., Germany). Construction of egfp-tagged nor-1 plasmids The N-terminally or C-terminally egfp-tagged nor-1 plasmids (Figure 4.1 and 4.2) were constructed using a fused nor-1 promoter/gene fragment, or separate nor-1 promoter and gene fragments, an egfi) gene fragment, a nor-1 terminator fragment, and pAPCGFPVFNB as a plasmid backbone. These plasmids differed in the length of the hinge region between nor-1 and egfl coding regions. For the C-tenninally egfi-tagged nor-1 plasmid, the 2.4 kb nor-1 promoter/gene fragment and the 1.8 kb nor-1 terminator fragment were generated by PCR with Pfu DNA polymerase (Stratagene, La Jolla, CA), appropriate primers, and cosmid NorA (Figure 2.5) (Liang et al., 1996) as a template using standard procedures (Maniatis et al., 1989). The primers contained restriction enzyme sites to facilitate cloning (Table 4.2). PCR was performed in a Gene Amp PCR System 2400 thermal cycler (Perkin-Elmer life sciences Inc., Boston, MA). The reaction conditions for thermal cycling depended on the designed primers and the target size as 186 Asc I 418 /. Psi I 1528 Amprnor-I ” terminator egfp Pst I 12724 Sal I 12712 F se I 2248 / Not I 2974 Xho I 3534 pAPCGFPNFNB 7’ (15018 bp) nor-1 promoter . /gene ‘ Pac I 5344 Sal I/Xho I 5360 Pst I 8512 Figure 4.1. Restriction endonuclease map of plasmid, pAPCGFPNFNB. The 2.4 kb nor-I promoter/ gene was fused in flame to the 0.7 kb egfi) coding region, followed by the 1.8 kb nor-1 terminator. The 7.4 kb niaD fragment was inserted as a selectable marker for transformation of the recipient strain B62 (niaD). 187 A. pAPCGFPNFNB Nor-l Hinge region EGF P DNA TGG-gtg-gct-ggc-ggc-cgc—ATG Natl Amino acid Trp-Val-Ala-Gly-Gly-Arg—Met B. pAPCGFPLNFNB Nor-l Hinge region EGFP DNA TGG-gtg-gct-ggc-ggc-cgc-gga-gct-ggt-gca-ggc-gct-gga-gcc-ATG Natl Amino acid Trp-Val-Ala-Gly-Gly-Arg-Gly-Ala-Gly-Ala-Gly-Ala-GIy-Ala-Met C. LAPNGFPNFNB EGFP Hinge region Nor-l DNA AAG-ggc-gat-cgc-ggt-gca-ggc-gct-ATG Amino acid Lys-Gly-AiglArg-Gly-Ala-Gly-Ala-Met Figure 4.2. Comparison of selected sequences in 3 plasmids carrying an egfp-tagged nor-1. (A) pAPCGFPNFNB. (B) pAPCGFPLNFNB. (C) pAPNGFPNFNB. These plasmids differ in the length of the hinge region between the nor-1 and egfp coding regions, and the location of the two regions. Natl and ngl sequences are shown in italic. Abbreviations for amino acids are as follows; Trp, tryptophan; Arg, arginine; Val, valine; Ala, alanine; Gly, glycine; Met, methionine; Lys, lysine; Asp, aspartic acid. The repeated Gly-Ala sequences are modified from a sequence kindly provided by S. Osmani, Ohio State University. 188 Table 4.2. Primer sequences used in this study. . b . . Primera Sequence Restriction enzyme site, annealing temp (0C) nor-I promoter/ 5’ TCAGTCTTAATTAATTGTACAAAGGCCTCCTA 3’ Bad, gene (C)- F 50 nor-1 promoter/ 5’ TTCGTCGCGGCCGCCAGCCACCCAGGGGAGTT- Natl, gene (C)- R -GAGA 3’ 50 nor-1 5’ CTCAACGGCCGGCCTAGGACTCCAGTGACGAC- Fsel, terminator- F -GAA 3’ 55 nor-1 5’ ACGGACGGCGCGCCCCTCGATGATGATGCTCT 3’ Ascl, terminator- R 55 egfp gene 5’ AGTGCGGCCGCGGAGCTGGTGCAGGCGCTGGA- Natl, (CL)- F -GCCATGGTGAGCAAGGGCGAGGAGCTGTTC 3’ 65 egfp gene 5’ GTCGGCCGGCCTTTACTTGTACAGCTCGTCCAT 3’ Fsel, (CL)- R 65 nor-1 promoter 5’ TCAGTCTTAATTAATTGTACAAAGGCCTCCTA 3’ Fuel, (N)- F 50 nor-I promoter 5’ CTAAGGCGGCCGCTCATTATGTCACGG 3’ Natl, (N)- R 50 egfp gene 5’ CCGGGCGGCCGCATGGTGAGCAAGGGCGAG 3’ Natl, (N)- F 60 egfp gene 5’ GCCGCGATCGCCCTTGTACAGCTCGTCCATGCC 3’ ngl, (N)- R 60 nor-I gene 5’ GTGGCGATCGCGGTGCAGGCGCTATGAACGGA- ngl, (N)- F -TCACTTAGCCAGCAC 3’ 60 189 Table 4.2. (cont’d). b . . Primera Sequence Restriction enzyme site, annealing temp (°C) nor-1 gene 5’ GTCGGCCGGCCGCTACCAGGGGAGTTGAGATCC 3’Fsel, (N)- R 60 nor-1 promoter 5’ CAAAATTGACATGAGCAGATCT 3’ None, (PCR assay)- F 60 nor-1 terminator 5’ ACCGCGCCTTTCTGGACCA 3’ None, (PCR assay)- R 60 a C represents C-terminal egfp fusion, CL represents C-terminal egfp fusion containing a long hinge region, and N represents N-terminal egfp fusion. Also, F represents forward primers and R represents reverse primers. b Underlined sequences show the position of the restriction enzyme sites. 190 follows: 94 0C for 5 min followed by 25 cycles of denaturation at 94 0C for 1 min, annealing for 1 min (Table 4.2), and extension at 72 0C for time in min that depended on PCR fragment sizes (2 min/1 kb). The reactions were completed with an extension at 72 0C for 10 min. The PCR fragments for the nor-1 promoter were purified from an agarose gel using a Qiaex 11 gel extraction kit (Qiagen Inc., Valencia, CA), digested with PacI and Natl, and cloned into pAPCGFPVFNB (Figure 3.1) cut with the same enzymes, resulting in pAPCGFPNVNB. The PCR fragments for the nor-1 terminator were then purified from an agarose gel using a Qiaex H gel extraction kit, digested with FseI and AscI, and cloned into pAPCGFPNVNB cut with the same enzymes, resulting in pAPCGFPNFN B (Figure 4.1). To generate a C-terminally egfp-tagged nor-1 plasmid containing a longer hinge region between nor-1 gene and egfp gene, the 0.7 kb egp gene was amplified by PCR with Pfu DNA polymerase and appropriate primers using pEGFP-Nl (Clonetech Laboratories, Palo Alto, CA) as a template (Figure 2.6). The primers contained restriction enzyme sites to facilitate cloning (Table 4.2). PCR was performed in a Gene Amp PCR System 2400 thermal cycler as described above. Plasmid pAPCGFPLNFNB containing a longer hinge region between nor-1 gene and egfp gene was constructed by replacement of the egfp gene fragment in pAPCGFPNFNB with another egfp gene fragment cut with Natl and FseI to provide more space between Nor-1 and EGFP for correct fusion protein folding. For the N-terminally egfp-tagged nor-1 plasmid, the 1.3 kb nor-I promoter and the 1.0 kb nor-I gene fragments were generated by PCR with Pfu DNA polymerase, appropriate primers, and cosmid NorA (Figure 2.5) (Liang et al., 1996) as a template using standard procedures (Maniatis et al. , 1989). The primers contained 191 restriction enzyme sites to facilitate cloning (Table 4.2). Also, the 0.7 kb egp gene was generated by PCR with Pfu DNA polymerase and appropriate primers using pEGFP-Nl as a template (Figure 2.6). The primers contained restriction enzyme sites to facilitate cloning (Table 4.2). PCR was performed in a Gene Amp PCR System 2400 thermal cycler as described above. The PCR fragments were cloned into the SmaI site of pUC 1 9, resulting in pUCGFP, pUCNOR, and pUCNORP, after purification from an agarose gel using a Qiaex 11 gel extraction kit. DNA fragments containing the nor-1 gene were subcloned from pUCNOR into pUCGF P cut with sgl and S011, resulting in pUCGFPNOR. DNA fragments containing the nor-I promoter were then subcloned from pUCNORP into pUCGFPNOR cut with EcoRI and Natl, resulting in pUCNORPGFPNOR. Finally, DNA fragments containing the nor-I promoter, the egfp gene, and the nor-1 gene were subcloned from pUCNORPGFPNOR into pAPCGFPNFNB cut with Fuel and Fsel, resulting in pAPNGFPNFNB. Transformation of E. coli and isolation of plasmids The preparation and transformation of E. coli competent cells were conducted by a calcium chloride method as described previously (Ausubel et al. , 2003). The isolation of plasmids was performed by an alkaline lysis method using a Wizard DNA purification kit (Promega Corporation, Madison, WI) or CsCl/ethidium bromide equilibrium centrifugation as described previously (Maniatis et al. , 1989). 192 Transformation of A. parasiticus and screening for aflatoxin and EGFP in transformants Transformation of A. parasiticus B62 with pAPCGFPNFNB, pAPCGFPLNFNB, or pAPNGFPNFNB was performed by a polyethylene glycol method (Oakley et al. , 1987) with minor modifications as described previously (Skory et al. , 1990). Fifteen h after inoculation of 108 conidia, protoplasts were generated by digestion of mycelia with lysing enzyme (25 mg/ml; Sigma Chemical Co., St. Louis, MO) and driselase (50 mg/ml; Sigma Chemical Co., St. Louis, MO). Transformants were selected on CZ agar supplemented with 20 % sucrose as an osmotic stabilizer. Transformants were transferred and cultured onto CAM at 30 0C for 4 days and then screened for a blue fluorescent halo (aflatoxin) around their colonies under U.V. at 365 nm and for EGFP under a Nikon Eclipse E600 fluorescence microscope (Nikon Inc., Melville, NY) using a 450-490 nm excitation/ 515 nm emission filter. Conventional fluorescence microscopy Slide culture was performed by the method according to Harris (1986) with minor modifications as described previously (Liang, 1996) (Figure 2.7). Approximately 5x105 conidia were inoculated around the edge of the top surface of YES agar blocks (1 cm2) on sterile coverslips which were placed onto water agar plates. This was followed by placement of a second coverslip on top of the YES agar blocks. The plates were incubated at 30 0C in the dark for 2 days. Coverslips were removed and washed three 193 times with phosphate-buffered saline (PBS) and observed using a Nikon Labophot fluorescence microscope (Nikon Inc., Melville, NY) with a 450-490 nm excitation/ 520 nm emission filter. Aflatoxin and aflatoxin intermediate analyses by TLC and ELISA Aflatoxin and aflatoxin intermediates were extracted by the method of Roze et al. (2004). Approximately 1x104 conidia were inoculated on YES agar plates and incubated at 30 0C in the dark for 4 days. Aflatoxins and aflatoxin intermediates were extracted the colonies with 10 ml of chloroform and then 10 m1 of acetone from, dried by evaporation, and dissolved in 3 ml of 70 % methanol. TLC was conducted on cell extracts using a TEA solvent system (toluene-ethyl acetate-acetic acid; 50:30:4 [v/v/v]) (Chang et al., 2004). The AF B1 concentration in the aflatoxin and aflatoxin intermediate extracts was determined by direct competitive enzyme-linked immunosorbent assay (ELISA) with AF B1 monoclonal antibodies (kindly provided by J. Pestka, Michigan State University) as described previously (Pestka, 1988). Total protein extraction Approximately 2x106 conidia were cultured in 100ml of YES media at 30 0C in the dark with shaking at 150 rpm for 48 h. After filtration through Miracloth (Calbiochem, La J olla, CA), mycelia were ground in liquid nitrogen with a mortar and pestle, suspended in TET buffer (100 mM Tris-HCl [pH7.5], 2.5 mM EDTA, 5 mM DTT, 194 1 % Triton X-100) containing 1 mM phenylmethylsulfonyl fluoride (PMSF) (Sigma Chemical Co., St. Louis, MO) and 4 % proteinase cocktail (Sigma Chemical Co., St. Louis, MO), and centrifuged at 10,000x g for 15 min at 4 °C (Valdez-Taubas et al., 2000, Tavoularis et al. , 2001). The supematants were used for Western blot analysis. Protein concentration in the supernatant was determined by a modified Bradford assay using a commercial Protein Assay reagent (Bio-Rad Laboratories, Hercules, CA) (Bradford, 1976). Western blot analysis of EGFP-tagged Nor-l Approximately 30-50 pg of total proteins were separated by 12 % sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Gels were transferred onto Poly Screen polyvinylidene difluoride (PVDF) membranes (Perkin-Elmer life sciences, Boston, MA) using standard procedures (Ausubel et al., 2003). After transfer, the gels were stained with Coomassie Brilliant Blue R-250 (Sigma Chemical Co., St. Louis, MO). Immunodetection was carried out with IgG antibody against Nor-1 protein or EGFP (Clonetech Laboratories, Palo Alto, CA) as a primary antibody, goat anti-rabbit IgG alkaline phosphatase conjugate (Sigma Chemical Co., St. Louis, MO) as a secondary antibody, and a BCIP/NBT (5’-bromo-4-chloro-3-indolyl phosphate/nitroblue tetrazolium) colorimetric detection system (Roche Molecular Biochemicals, Indianapolis, IND). A Benchmark prestained protein ladder (Invitrogen, Carlsbad, CA) was used as a molecular mass marker. 195 Genomic DNA isolation from A. parasiticus Genomic DNAs were isolated by a phenol-chloroform method (Ausubel et al. , 2003) with minor modifications as described previously (Skory et al. , 1990). Approximately 2x106 conidia were cultured in 100 ml of YES media at 30 0C in the dark with shaking at 150 rpm for 48 h and then a phenol-chloroform protocol was used to isolate genomic DNA from mycelia after filtration through Miracloth (Calbiochem, La Jolla, CA). Southern hybridization and PCR analyses Southern hybridization analyses were conducted using standard procedures as described previously (Maniatis et al., 1989). Approximately 10 pg of genomic DNAs cut with Sall were separated by agarose gel electrophoresis and then transferred onto a nylon transfer membrane (Nytran supercharge membrane, Schleicher and Schell Inc., Keene, NH) by capillary action. Radiolabeled DNA probes were generated with [a-32P]dCTP (Perkin-Elmer life sciences Inc., Boston, MA), the Random Primers DNA Labeling System (Invitrogen, Carlsbad, CA), and the 0.7 kb egfi) gene fragment using a procedure provided by the manufacturer. Afier the final wash, the membranes were exposed to X- ray film (Eastman Kodak company, Rochester, NY) at —80 0C. PCR analyses were performed with genomic DNA to confirm integration sites of the plasmids and to amplify the nor-1 gene region fused to egfi) gene in order to determine if the fusion protein carried wild-type Nor-l protein. DNA sequencing of the amplified nor-1 gene region fused to the egfp gene was conducted at the Research Technology Support Facility (RTSF; 196 Macromolecular Structure, Sequencing and Synthesis facility) at Michigan State University. lt/Hindlll (lnvitrogen, Carlsbad, CA) was used as a molecular size marker. Confocal Laser Scanning Microscopy (CLSM) Slide culture was performed by the method according to Harris (1986) with minor modifications as described above (Liang, 1996). For aflatoxin non-inducing media, YEP was used instead of YES. The plates were incubated at 30 0C in the dark. The coverslips were removed at different time points after inoculation. Fungal vacuoles were then stained with F M 4-64 or 7-amino-4-chloromethylcoumarin (CMAC) (Ohneda et al. , 2002; Shoji et al., 2006). FM 4-64 and CMAC were used for staining vacuolar membranes and lumens, respectively. The coverslips were put in YES media containing 8 11M FM 4-64 or 10 11M CMAC. For FM 4-64, coverslips were incubated at 30 0C for 10 min and washed with fresh media without the dye for 30 min. For CMAC, coverslips were inCubated at 30 0C for 30 min and washed with fresh media without the dye at 37 0C for 30 min. The coverslips were observed using a Zeiss LSM 5 Pascal or Zeiss LSM 510 Meta confocal laser scanning microscope (CLSM) (Carl Zeiss Inc., Germany). All single optical sections and extended focus images from Z-stacks (Z-section interval: 0.46 mm) were captured using a Zeiss Plan-APOCHROMAT (63x/1.40 Oil) objective. EGF P fluorescence (488 nm excitation/ 509 nm emission) was detected using a BP 505-530 emission filter set under excitation with the 488 nmargon-ion laser line. FM 4-64 fluorescence (558 nm excitation/ 734 nm emission) was detected using a LP 650 emission filter set under excitation with the 633 nm helium-neon laser line. CMAC fluorescence 197 (353 nm excitation] 466 nm emission) was detected using a BP 420-480 emission filter set under excitation with the 405 nm diode laser line. For counting of vacuole localization of green fluorescence, large- and mid-size vacuoles (>5 um) were counted in 2 or 3 hyphae from 1 microscopic field and this was repeated in a total 40 fields. The counting of green fluorescent vacuoles was statistically analyzed by two-way ANOVA followed by Tukey’s test for multiple comparisons using SigmaStat (SPSS Inc., Chicago, IL). Statistical significance among samples was defined by a P value not greater than 0.05 (P5005). Time-course of aflatoxin production Approximately 2x106 conidia were cultured in YES media at 30 0C in the dark with shaking at 150 rpm as described previously (Liang et al. , 1997). Flasks were removed at different time points after inoculation for analyses of mycelial dry weight and aflatoxin concentration. Mycelia were harvested by filtration through Miracloth (Calbiochem, La Jolla, CA), frozen in liquid nitrogen, and stored at —80 0C. Slide culture was performed as described above. The coverslips were removed at different time points after inoculation for analysis of aflatoxin concentration. Measurement of mycelial dry weight and aflatoxin concentration Dry weight was determined after complete drying of the harvested mycelia at 100 0C. The AFB] concentration in the filtrate was determined by direct competitive enzyme-linked immunosorbent assay (ELISA) with AFB] monoclonal 198 antibodies (kindly provided by J. Pestka, Michigan State University) as described previously (Pestka, 1988). For slide culture, aflatoxins were extracted from the agar blocks with 5 m1 of chloroform and then 5 ml of acetone, dried by evaporation, and dissolved in 1 ml of 70 % methanol. The AFB] concentration was determined by ELISA with AFB] monoclonal antibodies as described previously (Pestka, 1988). RESULTS Transformation of A. parasiticus B62 with pAPCGFPNFNB, pAPCGFPLNFNB, and pAPNGFPNFNB, and screening for aflatoxin and EGFP in transformants Five to ten pg of pAPCGFPNFNB, pAPCGFPLNFNB, or pAPNGFPNFNB was transformed into 108 of A. parasiticus B62 protoplasts, generating 877 transformants. All transformants selected on CZ agar were screened for aflatoxin production on CAM under U.V. at 365 nm. None of 197 transformants carrying pAPCGFPNFNB produced bigger blue fluorescent haloes around their colonies than that from the recipient strain B62, in which small quantities of aflatoxin were still produced. All 197 transformants were then screened for EGF P expression under the fluorescence microscope (Figure 4.3). Seven transformants (isolates N7, N31, N43, N51, N60, N70, and N128) produced EGFP fluorescence (EGFP [+]). One transfonnant carrying pAPCGFPLNFNB out of 300 transformants produced 199 Figure 4.3. Fluorescence microscopy of A. parasiticus B62 expressing EGFP. Transformants were inoculated on CAM plates or YES agar blocks on coverslips and incubated at 30 0C for 3 or 4 days and observed using a Nikon Labophot or Eclipse E600 fluorescence microscope. (A) Transformant carrying pAPCGFPNFNB. EGF P fluorescence was observed in vesicles and conidiophores. (100 x) (B) Transformant carrying pAPCGFPLNFNB. EGFP fluorescence was observed in vesicles and conidiophores. Accumulation of orange-pigmented norsolorinic acid (NA) was observed in conidia. (400 x) (C) Transformant carrying pAPNGFPNFN B. EGFP fluorescence was observed in vesicles and conidiophores. (100 x) Images in this dissertation are presented in color. 200 a bigger blue fluorescent halo around the colony than that from the recipient strain B62. This transfonnant also accumultated less orange pigment in the mycelia than that in the recipient strain B62. This orange pigment has been confirmed by NMR, TLC, and ELISA to contain predominantly norsolorinic acid (Lee et al. , 1971; Reynolds and Pestka, 1991; Trail et a1. , 1994). All 300 transformants were then screened for EGF P expression under the fluorescence microscope (Figure 4.3). The transfonnant that produced increased levels of aflatoxin (isolate LN196) also produced EGF P fluorescence. F ourty- seven EGF P (+) transformants were identified out of the transformants that did not produce increased levels of aflatoxin. Fourty-four transformants carrying pAPNGFPNFNB out of 380 transformants produced bigger blue fluorescent haloes around their colonies than the halo from the recipient strain B62. F ourty-two out of the 44 transformants that produced higher levels of aflatoxin also accumultated no detectable orange pigment in the mycelia. However, 2 (isolates NN182 and NN227) of these transformants did accumulate a little orange pigment in the mycelia. All 380 transformants were then screened for EGFP expression under the fluorescence microscope (Figure 4.3). All 44 transformants that produced increased levels of aflatoxin also produced EGF P fluorescence. Eight EGFP (+) transformants were identified out of the transformants that did not produce increased levels of aflatoxin. TLC and ELISA analyes of transformants To check for complementation of non-functional Nor-1 in A. parasiticus B62 by EGFP-tagged Nor-1, TLC analysis was performed using chloroform/acetone extracts 202 from transformants carrying pAPCGFPNFNB. Seven EGF P (+) transformants (isolates N7, N31, N43, N51, N60, N70, and N128) and l EGFP (-) transfonnant (isolate N5) did not produce increased amounts of aflatoxins nor did they produce decreased amounts of norsolorinic acid (NA) than the recipient strain B62 (Figure 4.4). TLC analysis was performed using chloroform/acetone extracts from the transformants carrying pAPCGFPLNFNB. One EGFP (-) transformant (isolate LN2) and 2 EGFP (+) transformants (isolates LNl and LN6) did not produce increased amounts of aflatoxins nor did they produce decreased amounts of norsolorinic acid (NA) than the recipient strain B62 (Figure 4.5). However, 1 light orange-pigmented EGFP (+) transfonnant (isolate LN196) produced a little more aflatoxin than the recipient strain B62 but still produced NA (Figure 4.5). To compare AFB] production from transformants with those from the recipient strain B62 and the wild-type stain SU-l , AFB] concentration was measured by ELISA. LN196 produced 20 fold more AFB] than the recipient strain B62 (Figure 4.6). TLC analysis was performed using chloroform/ acetone extracts from transformants carrying pAPNGFPVFNB to confirm complementation of non-functional Nor-1 in A. parasiticus B62 by EGFP-tagged Nor-1. Two EGF P (-) transformants (isolates NNl and NN244) and 2 orange-pigmented EGF P (+) transformants (isolates NN145 and NN183) did not produce increased amounts of aflatoxin nor did they produce decreased amounts of norsolorinic acid (NA) than the recipient strain B62 (Figure 4.7 A). However, 2 light orange-pigmented EGFP (+) transformants (isolates NN182 and NN227) and 3 non-detectable orange-pigmented EGFP (+) transformants (isolates NN4, 203 Figure 4.4. TLC analysis of extracts from transformants carrying pAPCGFPNFNB and the recipient strain B62. (A) TLC analysis of extracts from EGF P (-) (N5), and EGFP (+) (N 7, N31, N43, and N51) transformants. (B) TLC analysis of extracts from EGFP (+) transformants (N 60, N70, and N128). Aflatoxin B], Bz, G], and G2 standard mixture and norsolorinic acid (NA) were used as standards. TEA (toluene-ethyl acetate-acetic acid; 50:30:4 [v/v/v]) was used as a solvent system. Fluorescence was detected under U.V. at 365 nm. 204 NA” Am," Arc,» AF NA B62 N5 N7 N31 N43 N51 AF Std Std Std AFBI" AFGf‘ AF NA B62 N60 N70 N128 AF Std Std Std 205 NA-‘ AFB] - AFGI" AF NA B62 LN2 LNl LN6 LN196 AF Std Std Std Figure 4.5. TLC analysis of extracts from transformants carrying pAPCGFPLNFNB and the recipient strain B62. TLC analysis of extracts from EGF P (-) (LN2), orange- pigmented EGFP (+) (LNl and LN6), and light orange-pigmented AF (+) and EGFP (+) (LN196) transformants. Aflatoxin B], B2, G], and G2 standard mixture and norsolorinic acid (NA) were used as standards. TEA (toluene-ethyl acetate-acetic acid; 50:30:4 [v/v/v]) was used as a solvent system. Fluorescence was detected under U.V. at 365 nm. 206 AFB] analysis by ELISA 2500 2000‘ 'l' 1500 ‘ AFB] (ug/ml) 1000‘ 500‘ 0 I l I T I j SU-l 862 LN2 LN 1 LN6 LN196 Transformants carrying pAPCGFPLNFNB Figure 4.6. AFB] analysis of extracts from transformants carrying pAPCGFPLNFNB, the recipient strain B62, and the wild-type strain SU-l. Extracts from EGFP (-) (LN2), orange-pigmented EGF P (+) (LNl and LN6), and light orange-pigmented (LN196) transformants, B62, and SU-l were analyzed for AFB] production by ELISA. 207 Figure 4.7. TLC analysis of extracts from transformants carrying pAPNGFPNFNB and the recipient strain B62. (A) TLC analysis of extracts from EGFP (-) (NN l and NN244), and orange-pigmented EGFP (+) (NN 145 and NN183) transformants. (B) TLC analysis of extracts from light or non-detectable orange-pigmented EGFP (+) transformants (NN 182, NN227, NN4, NN6, and NN24). Aflatoxin B], B2, G], and G2 standard mixture and norsolorinic acid (NA) were used as standards. TEA (toluene-ethyl acetate- acetic acid; 50:30:4 [v/v/v]) was used as a solvent system. Fluorescence was detected under U.V. at 365 nm. 208 NA" AFBl-o Arc: AF NA B62 NN1NN145NN183 NN244 AF Std Std Std NA‘ AFB] ~ A170," AF NA B62NN182 NN227 NN4 NN6 NN24 AF Std Std Std 209 NN6, and NN24) produced increased amounts of aflatoxin and decreased but still detectable amounts of NA than the recipient strain B62 (Figure 4.7 B). Three (isolates NN4, NN6, and NN24) non-detectable orange-pigmented EGFP (+) transformants produced less NA than the two light orange-pigmented EGFP (+) transformants (isolates NN182 and NN227) (Figure 4.7 B). To compare AFB] production from the transformants with that from the recipient strain B62 and the wild-type stain SU- 1 , AFB] concentration was measured by ELISA. NN4, NN6, and NN24 produced similar amounts of AFB] to SU-l, but NN182 and NN227 produced 50 % of AFB] produced by SU-l (Figure 4.8). When taken together, TLC and ELISA analyses indicated that in all 5 light or non-detectable orange-pigmented EGFP (+) transformants carrying pAPNGFPNFN B, the N-terminally EGFP-tagged Nor-1 protein functionally complemented non-functional Nor-l in A. parasiticus B62 although the 2 light orange-pigmented transformants showed decreased amounts of AFB] production compared to SU-l (Figure 4.7 and 4.8). Western blot analysis of EGFP-tagged Nor-l To check for production of full-length EGFP-tagged Nor-1 fusion protein in transformants, Western blot analysis was performed using anti-Nor-l antibody or anti- EGFP antibody. In transformants carrying pAPCGFPNFNB, 7 EGF P (+) transformants (isolates N7, N31, N43, N51, N60, N70, and N128) produced a full length 58 kDa fusion protein (Figure 4.9). F iftyeight kDa is the expected molecular mass of full length fusion protein (31 kDa Nor-l and 27 kDa EGFP). However, 1 EGFP (-) transformant (isolate N5) did not produce a 58 kDa fusion protein (Figure 4.9 A and C). Western blot analysis 210 AFB] analysis by ELISA 2500 2000 - 1500 - AFB] (ug/ml) 500‘ SU-l B62 NNI NN145 NN182 NN227 NN4 NN6 NN24 Transformants carrying pAPNGFPNFNB Figure 4.8. AFB] analysis of extracts from transformants carrying pAPNGFPNFNB, the recipient strain B62, and the wild-type strain SU-l. Extracts from EGFP (—) (NN 1), orange-pigmented EGFP (+) (NN145), and light or non-detectable orange-pigmented EGFP (+) (NN182, NN227, NN4, NN6, and NN24) transformants, B62, and SU-l were analyzed for AFB] production by ELISA. 211 Figure 4.9. Western blot analysis of EGFP-tagged Nor-1 from transformants carrying pAPCGFPNFNB, the recipient strain B62, and the wild-type strain SU—l. Fungal proteins were extracted from transformants, B62, and SU-l grown in 100 ml of YES for 48 h at 30 c’C with shaking at 150 rpm. Approximately 30—50 pg of proteins were separated by 12 % SDS-PAGE, transferred onto PVDF membranes, and probed with (A) and (B) Nor-l polyclonal antibody, or (C) and (D) EGFP polyclonal antibody. EGFP- tagged Nor-1 has a molecular mass of 58 kDa, and the 31 kDa protein represents Nor-1 (with the exception of functional Nor-1 from SU-l). A Benchmark prestained protein ladder was used as a molecular mass marker (M). Recombinant EGF P (rEGFP) was used as a positive control, and the 30 kDa rEGFP contains a 27 kDa EGFP fused to a 3 kDa protein for affinity chromatography purification. 212 at" e" e" 4'9 4" kDa 103 77 50 34 ”new SAW “N" “"w 29 21 58 kDa—v 3lkDa—'“~-““h 213 ,s “58 kDa “31 kDa 103 77 50 34 29 21 Figure 4.9. (cont’d) RD- 103 34 29 21 Q8 46° 4" 4"? 3” ._.-.__..~w +58kDa . *30kD: 214 also showed that all transformants and the recipient strain B62 produced 31 kDa Nor-1 and that the wild-type strain SU-l produced 31 kDa functional Nor-1 (Figure 4.9). These data indicated that the EGFP-tagged Nor-l fusion protein is produced intact and fill]- length in transformants. Western blot analysis was conducted on transformants carrying pAPCGFPLNFNB. One light orange- pigmented EGFP (+) transformant (isolate LN196) produced a 58 kDa fusion protein (Figure 4.10 A and B). Also, 2 orange- pigmented EGF P (+) transformants (isolates LNl and LN6) produced a 58 kDa fusion protein (Figure 4.10 A and B). However, 1 EGFP (-) transfonnant (isolate LN2) did not produce a 58 kDa fusion protein (Figure 4.10 A and B). All transformants except for LN196 and the recipient strain B62 produced 31 kDa Nor-1 and the wild-type strain SU-l produced 31 kDa functional Nor-1 as previously shown (Figure 4.10). These data indicated that EGFP-tagged Nor-l fusion protein is produced intact and full-length in transformants. Western blot analysis also showed that in transformants carrying pAPNGFPNFNB, 5 light or non-detectable orange-pigmented EGFP (+) transformants (isolates NN182, NN227, NN4, NN6, and NN24) produced a 58 kDa fusion protein (Figure 4.11 B and D). Also, 2 orange-pigmented EGF P (+) transformants (isolates NN145 and NN183) produced a 58 kDa fusion protein (Figure 4.11 A and C). However, 2 EGFP (-) transfonnant (isolates NNl and NN244) did not produce a 58 kDa fusion protein (Figure 4.11 A and C). Western blot analysis also showed that all transformants and the recipient strain B62 produced 31 kDa Nor-1 and that the wild-type strain SU-l produced 31 kDa functional Nor-1 protein (Figure 4.11 A and B). These data indicated 215 Figure 4.10. Western blot analysis of EGFP-tagged Nor-l from transformants carrying pAPCGFPLNFNB, the recipient strain B62, and the wild-type strain SU-l. Fungal proteins were extracted from transformants, B62, and SU-l grown in 100 m1 of YES for 48 h at 30 °C with shaking at 150 rpm. Approximately 30-50 pg of proteins were separated by 12 % SDS-PAGE, transferred onto PVDF membranes, and probed with (A) and (B) Nor-1 polyclonal antibody, or (C) and (D) EGF P polyclonal antibody. EGFP- tagged Nor-l has a molecular mass of 58 kDa, and the 31 kDa protein represents Nor-1 (with the exception of functional Nor-1 from SU-l). A Benchmark prestained protein ladder was used as a molecular mass marker (M). Recombinant EGF P (rEGFP) was used as a positive control, and the 30 kDa rEGFP contains a 27 kDa EGFP fused to a 3 kDa protein for affinity chromatography purification. 216 kDa 170.8 109.5 78.9 60.4 47.2 35.1 24.9 18.3 13.7 > 9“ 4‘ 4» as at" 0" 6" a" a" $0 é su- Am i .. .. A. ~ .. 1AA" «50km . ,. , t At r- 30 kDa arm-1i til, 217 Figure 4.11. Western blot analysis of EGFP-tagged Nor-1 from transformants carrying pAPNGFPNFNB, the recipient strain B62, and the wild-type strain SU-l. Fungal proteins were extracted from the transformants, B62, and SU-l grown in 100 ml of YES for 48 h at 30 °C with shaking at 150 rpm. Approximately 30-50 pg of proteins were separated by 12 % SDS-PAGE, transferred onto PVDF membranes, and probed with (A) and (B) Nor-l polyclonal antibody, or (C) and (D) EGFP polyclonal antibody. EGFP- tagged Nor-l has a molecular mass of 58 kDa, and the 31 kDa protein represents Nor-1 (with the exception of functional Nor-1 from SU-l). A Benchmark prestained protein ladder was used as a molecular mass marker (M). Recombinant EGFP (rEGFP) was used as a positive control, and the 30 kDa rEGFP contains a 27 kDa EGFP fused to a 3 kDa protein for affinity chromatography purification. 218 > wxx’f’ébww stafebxféééééfé 170.8 109.5 78.9 604 - '— 58 kDa 47.2 35.1 mmmmmmxm .1 «31 kDa 24.9 18.3 13.7 ‘9' q’,‘ t. > s q. a. b w 4* 6.9 <15" $484“ $484 kDa 170.3 109.5 78.9 20].: - , «53 kDa 35.1 ””Mrsw Mum-0“," “ 31 kDa 24.9 18.3 13.7 219 Figure 4.11. (cont’d) tb¢¢<$ $§§§§§A «58 kDa «30 kDa .8 @quchbw‘éq' ¢§$$$$$$3¢ kDa 170.8 ’ 109.5 78.9 ‘ 60.4 4* . . , ‘_ 58 kDa 47.2 ‘1‘ ' :E' 35.1 5" ‘3: ‘_ 30 kDa 24.9 A 18.3 A 13.7 - , 220 that EGFP-tagged Nor-1 fusion protein is produced intact and full-length in transformants. Identification of a point mutation in nor-I in B62 The recipient strain B62 (nor-1, niaD, br-I) was originally derived from A. parasiticus ATCC24690 (nor-1, br-I), which accumulates the aflatoxin pathway intermediate norsolorinic acid (NOR or NA); ATCC24690 was in turn derived from the wild-type strain SU-l by U.V. irradiation (Bennett and Goldblatt, 1973; Lee et al., 1971). To determine the mutation site in nor-I in B62, PCR was performed and DNA sequencing of the PCR products was conducted. DNA sequencing showed a point mutation site (T —> C) at nucleotide residue 790 in nor-I in B62 (Figure 4.12). These data indicate that nor-1 contains a point mutation and encodes a non-functional Nor-1 in B62 due to an amino acid change from leucine to proline. Analysis of expression of EGFP-tagged Nor-l fusion protein carrying functional wild-type Nor-1 protein The recipient strain B62 produces a non-functional Nor-l due to the point mutation in nor-1 as described above (Figure 4.12). Therefore, transformants could theoretically produce either EGFP-tagged fusion protein with functional Nor-lor EGFP- tagged fusion with non-functional Nor-1 depending on the plasmid integration site into the chromosome (Figure 4.13). To determine if the EGFP-tagged Nor-1 fusion protein carried a fuctional wild-type Nor-1 protein, PCR was performed with the egffp primer and the 5’ or 3’ nor-1 primer (Table 2.2 and Table 4.2). DNA sequencing of the amplified 221 Nucleotide sequence SU-l 783 gcgcggctgatgggccgtccg 803 362 783 gcgcggccgatgggccgtccg 803 Amino acid sequence SU-l 225 231 I." 3-—-3 61—6) ”—71 'U—‘U A R | 1 A R 231 I'd o 362 225 Figure 4.12. Comparison of selected nucleotide and amino acid sequences of nor—1 from SU-l and B62. Identical nucleotides or amino acids are indicated by lines and a point mutation is indicated by a dot. Nucleotide and amino acid sequences were aligned with the EMBOSS alignment tool (http://www.ebi.ac.uk/emboss/align/index.html). Abbreviations for amino acids are as follows; A, alanine; R, arginine; L, leucine; M, methionine; G, glycine; P, proline. 222 Figure 4.13. Schematic for production of an EGFP-tagged functional Nor-l fusion protein depending on the integration site of pAPCGFPNFNB into the nor-1 locus. (A) nor-1 locus. (B) Integration upstream of the point mutation in nor-1 gene including 5’ nor-1. (C) Integration downstream of the point mutation in nor-1 gene. (D) 3’ nor-1 integration. Integration of pAPCGFPNFNB upstream of the point mutation in nor-1 gene including 5’ nor-1 or into 3’ nor-1 results in production of an EGFP-tagged functional Nor-1 fusion protein. Integration of pAPCGFPLNFNB into the nor-1 locus also results the same pattern of fusion protein production as that of pAPCGFPNFNB. Integration of the plasmid pAPNGFPNFNB downstream of the point mutation in nor-1 gene including 3’ nor-1 or into 5’ nor-1 results in production of an EGFP-tagged functional Nor-1 fusion protein. M represents a point mutation. Abbreviations for the DNA fragments are as follows; nor-1 p, nor-1 promoter; nor-1 t, nor-1 terminator. 223 A. nar-I locus I' p nor-1 terminatezzlix pAPCGFPNFNB : . (15018 bp) nor-I promoter 1 /gene ' ‘ M nor-I p nor-1 nor-1 t 224 Figure 4.13. (cont’d) B. Integration upstream of the point mutation in nor-1 gene including 5’ nor-I M nor-1 pnar-I egfp nor-I t amp niaD nor-1 p nor-1 nor-1 t Nor-l +— EGFP + Nor-1 ' C. Integration downstream of the point mutation in nor-I gene M —-:rH-—@—<. A. . :.::;.:~:.. A: 2‘:- nar-I p nor-1 egfp nor-I t amp niaD nor-1 p nor-I nor-1 t Nor-1 '_ EGFP + Nor-l + D. 3’ nor-1 integration M ———-:-—-@——(‘x ' i J [E- nar-I p nor-1 nor-1 t amp niaD nor-1 p nor-I egfp nor-I t - + + Nor-1 Nor-l — EGF P 225 nor-1 gene region fused to the eg'p gene was performed. DNA sequence analysis showed that isolate LN196 (light orange-pigmented EGFP [+]) carried wild-type nor-1 and that isolates NN182, NN227, NN4, NN6, and NN24 (light or non-detectable orange pigmented EGFP [+]) carried wild-type nor-I (data not shown). Determination of integration sites of pAPCGFPNFNB, pAPCGFPLNFNB, and pAPNGFPNFNB within the chromosome In order to determine the integration sites of pAPCGFPNFNB, pAPCGFPLNFNB, and pAPNGFPNFNB, Southern hybridization and PCR analyses were performed. Each plasmid could theoretically be integrated into the chromosome by homologous recombination at three sites: niaD, 3’ nor-1 terminator, or 5’ nor-1 promoter/ ORF (open reding frame) region (Figure 4.14 and 4.16). Southern hybridization analysis showed that isolates N7 and N51 (EGF P [+] transformants) carrying pAPCGFPNFNB had the plasmids integrated at the 3’ nor-1 terminator while isolates N31, N43, N60, N70, and N128 (EGF P [+] transformants) had the plasmids integrated at the 5’ nor-1 promoter/ ORF region (Figure 4.15). The analysis also showed that in isolate N128, a second plasmid was integrated at the niaD locus by single- crossover and that in isolate N5 (EGF P H) the plasmid was integrated at the niaD locus by single-crossover (Figure 4.15). These data suggest multiple integration of the plasmid in N128. Southern hybridization and PCR analyses showed that in isolate LN196 (light orange-pigmented EGFP [+]) carrying pAPCGFPLNFN B, the plasmid was integrated at the 3’, 5’ nor-1, and niaD locus, indicating multiple integration of the plasmid (Figure 226 Figure 4.14. Schematic for Southern hybridization analysis of integration sites of pAPCGFPNFNB into the chromosome. (A) nor-1 locus. (B) 3’ nor-1 integration. (C) 5’ nor-I/ ORF region integration. (D) niaD locus. (H) niaD integration. Genomic DNA was digested with Sall and probed with an egfi) gene. The restriction enzyme sites, probes, and expected band sizes in Southern hybridization analyses are shown. Abbreviations for the restriction enzyme sites and DNA fragments are as follows; S, Sall; Xh, Xhol; nor-1 p, nor-1 promoter; nor-1 t, nor-I terminator. 227 A. nor-I locus Sui nor-I p nar-I nor-I t B. 3’ nor-I integration S S S ——-l-:-—@—|(‘17"T‘T"7°7' . ‘ .:E_ nor-1 p nor-I nor-1 t amp niaD nor-I p nor-I egfp nor-1 t (probe)l 13.3 kb 1 C. 5’ nar-I/ ORF region integration S S S nor-I p nor-1 egfp nor. 1 t amp niaD nor-I p nor-I nor-I t (probe) .L A' 9.1 kb 228 Figure 4.14. (cont’d) D. niaD locus niaD E. niaD integration Xh S S niaD amp nor-1 t nor-1 egfp nor-1 p niaD 1 (probe) 1 1 1 15.9 kb 229 $¢A§£§$§§ Figure 4.15. Southern hybridization analysis of integration sites in transformants carrying pAPCGFPNFNB. For Southern hybridization analysis, genomic DNA was isolated from A. parasitisus, digested with Sail, and hybridized with an ea?) gene probe. 3’ nor-1 integrants resulted in a 13.3 kb band, 5’ nor-I/ ORF region integrants resulted in a 9.1 kb band, and niaD integrants resulted in a 15.9 kb band . 230 Figure 4.16. Schematic for Southern hybridization and PCR analyses of integration sites of pAPCGFPLNFNB into the chromosome. (A) nor-1 locus. (B) 3’ nor-1 integration. (C) 5’ nor-1 / ORF region integration. (D) niaD locus. (H) niaD integration. Genomic DNA was digested with Sall and probed with an gfp gene. The restriction enzyme sites, probes, and expected band sizes in Southern hybridization analysis are shown. The primer positions in PCR analysis are also shown. In Southern hybridization analysis, integration of the plasmid pAPNGFPNFNB at 3’ nor-1 results in the same pattern as integration at the nor-1 ORF region. Abbreviations for the restriction enzyme sites and DNA fragments are as follows; S, SalI; Xh, Xhal; nor-I p, nor-1 promoter; nor-1 t, nor-1 terminator. 231 A. nor-1 locus nor-1 p nor-I nor-I t B. 3’ nor-I integration PCR PCR l-Dfl-l 1"} 4—1 S S S V nor-I p nor-I nor-1 t amp niaD nor-1 p nor-1 egfp nor-I t l (probe)l I 13.3 kb I C. 5’ nar-I/ ORF region integration PCR PCR 1'? {-1 l-V4-l nor-I p nor-I egfp nor-1 t amp niaD nor-I p nor-1 nor-I t (probe) 1 ' 9.1 kb ' 232 Figure 4.16. (cont’d) D. niaD locus E. niaD integration \Lm niaD PCR S niaD .. . .- ‘ I. - AT‘rf"'.'_‘I"HY mr.-- Ah: amp nor-1 t nor-1 egfp nor-1 p niaD L (probe) ' 15.9 kb 233 4.17 A and B). The analyses also showed that in isolate LNl (orange-pigmented EGFP [+]), the plasmid was integrated at the 5’ nor-I promoter/ ORF region while in isolate LN6 (orange-pigmented EGFP [+]), the plasmid was integrated at the 3’ nor-I terminator (Figure 4.17 A and B). Isolate LN2 (EGFP [-]) did not produce any bands in Southern hybridization and the same size band was observed as in B62 by PCR analysis, indicating integration of the plasmid into the niaD locus by double-crossover (gene replacement) (Figure 4.17 A and B). Southern hybridization and PCR analyses showed that in isolates NN182, NN227, NN4, NN6, and NN24 (light or non-detectable orange-pigmented EGFP [+]) carrying pAPNGFPNFNB, the plasmid was integrated at the 3’ nor-1 terminator/ ORF region and that in isolates NN182 and NN4 (light or non-detectable orange-pigmented EGF P [+]), a second plasmid was integrated at the niaD locus by single-crossover (Figure 4.18 A and B). The analyses also showed that in isolates NN145 and NN183 (orange-pigmented EGFP [+]), the plasmid was integrated at the 3’ nor-1 terminator/ ORF region and that in isolate NN244 (EGFP H), the plasmid was integrated at the niaD locus by single- crossover (Figure 4.18 A and B). Isolate NNl (EGF P H) did not produce any bands in Southern hybridization and the same size band was observed in B62 by PCR analysis, indicating integration of the plasmid into the niaD locus by double-crossover (gene replacement) (Figure 4.18 A and B). Confocal Laser Scanning Microscopy (CLSM) To identify the sub-cellular location of C-terminally or N-terminally EGFP-tagged Nor-1 in hyhae of EGFP (+) transformants producing increased amounts of aflatoxin 234 Figure 4.17. Southern hybridization analysis of integration sites in transformants carrying pAPCGFPLNFN B. For Southern hybridization analysis, genomic DNA was isolated from A. parasitisus, digested with Sall, and hybridized with an egfp gene probe. For PCR analysis, genomic DNA was amplified with the nor-1 promoter primer and the nor-l terminator primer (Figure 4.16 and Table 4.2). (A) Southern hybridization analysis of pAPCGFPLNFNB. 3’ nor-1 integrants resulted in a 13.3 kb band, 5’ nor-I/ ORF region integrants resulted in a 9.1 kb band, and niaD integrants resulted in a 15.9 kb band. (B) PCR analysis of 3’, 5’ nor-1 / ORF region, or niaD integrants of pAPCGFPLNFNB. 3’ or 5’ nor-I/ ORF region integrants resulted in 1.35 and 2.05 kb bands. The recipient strain B62 was used as a negative control. A/HindIII and ¢X174 RF/HaeIII were used as molecular size markers. 235 15.9 kb " 13.3 kb " 9.1 kb —‘ B. x «a?» «90° 98 ,6 9? § ’\ °l°'\" ,8 WAAAAAAAA I 23:11:: + 6.6 kb * 4.4 kb 2.3 kb 2.1 kb 1.35 kb 1.08 kb 0.87 kb “ 0.60 kb 11 2.05 kb -> 1.35 kb —. f f f 236 Figure 4.18. Southern hybridization analysis of integration sites in transformants carrying pAPNGFPNFNB. For Southern hybridization analysis, genomic DNA was isolated from A. parasitisus, digested with $011, and hybridized with an egfp gene probe. For PCR analysis, genomic DNA was amplified with the nor-1 promoter primer and the nor-1 terminator primer (Figure 4.16 and Table 4.2). (A) Southern hybridization analysis of pAPNGFPNFNB. 3’ nor-1 / ORF region integrants resulted in a 13.3 kb band, 5’ nor-1 integrants resulted in a 9.1 kb band, and niaD integrants resulted in a 15.9 kb band. (B) PCR analysis of 5’, 3’ nor-I/ ORF region, or niaD integrants of pAPNGFPNFNB. 5’ or 3’ nor-1 / ORF region integrants resulted in 1.35 and 2.05 kb bands. The recipient strain B62 was used as a negative control. A/HindIII and ¢X174 RF/HaeIII were used as molecular size markers. 237 ‘96 p ma ‘ N“\% W @w’l’u b '1’ fiaée$$éé§ééeééééé 15.9 kb —> , 13.3 kb * , B. Q 9 88° 39’ 6 h 9 ,p m ’\ k g ’\ s" 4"» {b '0' A. b w 4\ (A 9&8‘13‘9593 eéx§ $§$ $$$$ kb kb kb kb 2.05 kb 121: 1.35 kb 238 (isolate LN 1 96 or NN6) compared to the recipient strain B62, confocal laser scanning microscopy (CLSM) was performed after 24, 48, and 72 h culture on aflatoxin-inducing media, YES. EGFP fluorescence was not detected in LN196 at 24 h (Figure 4.19 C). However, C-terminally EGFP-tagged Nor-1 was localized in vacuoles and the cytoplasm of hyphae and the cytoplasm of conidiophores of LN196 at 48 and 72 h when either the vacuolar membrane staining dye FM 4-64 or the vacuolar lumen staining dye CMAC was used (Figure 4.19 D to I). This vacuole localization of EGFP-tagged Nor-1 was observed using the vacuolar membrane dye FM 4-64 and it was confirmed using the vacuolar lumen—specific dye CMAC; it is known that FM 4-64 stains endosomal compartments like endosomes as well as vacuoles in Saccharomyces cerevisiae (Vida and Emr, 1995). Also, LN196 did not produce any EGFP fluorescence when it was cultured on non- aflatoxin-inducing media, YEP (Figure 4.19 J). The pattern of localization of N- terrninally EGFP-tagged Nor-1 in NN6 was similar to that in LN196. Very little EGFP fluorescence was detected in NN6 at 24 b (Figure 4.20 A). However, N-terminally EGFP-tagged Nor-1 was localized in vacuoles and the cytoplasm of hypae and conidia of NN6 at 48 and 72 h when either the vacuolar membrane staining dye FM 4-64 or the vacuolar lumen staining dye CMAC was used (Figure 4.20 B to F). Also, NN6 did not produce any EGFP fluorescence when it was cultured on non-aflatoxin-inducing media, YEP (Figure 4.20 G). In Chapter 2, EGFP itself was localized in the cytoplasm of EGF P (+) transfonnant B3-15 at 48 h but localized in vacuoles at 72 h. Therefore, to compare vacuole localization of EGFP in B3-15 with that of EGFP-tagged Nor-1 in NN6, green fluorescent vacuoles were counted at 48 and 72 h. In NN6 75.5 ”/0 of vacuoles showed localization of green fluorescence at 48 h while in B3-15 only 7.0 % of vacuoles showed 239 Figure 4.19. Sub-cellular localization of C-terminally EGFP-tagged Nor-1 in light orange-pigmented EGFP (+) transformant LN196. Fungal vacuoles were stained with 8 pM FM 4-64 or 10 pM CMAC and observed using a Zeiss LSM 5 Pascal or Zeiss LSM 510 Meta confocal laser scanning microscope (CLSM) after 24, 48, and 72 h incubation at 30 0C on aflatoxin-inducing media, YES or non-aflatoxin-inducing media, YEP, agar blocks. (A) and (B) The recipient strain B62 (negative control) stained with CMAC at 48 h on YES. Blue fluorescent vacuoles were observed in hyphae but green fluorescence was not detected in panel (A). Small round structures were observed outside of the firngal cell wall and in the media near the fungal cell wall in panel (B). (C) LN196 stained with FM 4-64 at 24 h on YES. Red fluorescent vacuolar membranes were observed in hyphae but green fluorescence was not detected. (D), (E), and (F) LN196 stained with FM 4-64 at 48 h on YES. EGFP-tagged Nor-1 was localized in vacuoles in panel (D), in the cytoplasm of hyphae in panel (B), and in the cytoplasm of conidiophores and in phialides in panel (F). (G) and (H) LN196 stained with CMAC at 48 h on YES. EGFP-tagged Nor-1 was localized in large vacuoles in panel (G) and in small vacuoles in panel (H). The EGFP-tagged Nor-l localized in the small vacuoles was associated with the cytoplasmic membrane as seen in panel (H). (I) LN196 stained with FM 4-64 at 72 h on YES. EGFP-tagged Nor-l was localized in vacuoles. (J) LN196 stained with FM 4-64 at 72 h on YEP. Red fluorescent vacuolar membranes were observed in hyphae but very little green fluorescence was detected. Each panel shows a red fluorescence image (FM 4-64) or a blue fluorescence image (CMAC) (top left), a green fluorescence image (EGFP) (top right), a transmitted image (bright field or differential interference contrast [DIC]) (bottom left), and a merged image (bottom right) except for panel (B) in which only a merged image is shown. Scale bars, 10 pm. Images in this dissertation are presented in color. 240 The recipient strain B62, YES, 48 h, CMAC A. a ,l rt ,3 Y t :P 1‘111171 241 Figure 4.19. (cont’d) C. LN196, YES, 24 h, FM 4-64 1 I] 11 rn - ~ , - : 1 [1 1‘ m Figure 4.19. (cont’d) E. LN196, YES, 48 h, FM 4-64 H 1 [1 pm F. Figure 4.19. (cont’d) G. LN196, YES, 48 h, CMAC 1|] pm " '11] pm H. LN196, YES, 48 h, CMAC Figure 4.19. (cont’d) 72 h, FM 4—64 9 LN 196, YES I. YEP, 72 h, FM 4-64 9 LN196 l. 245 Figure 4.20. Sub-cellular localization of N-terminally EGFP-tagged Nor-1 in EGFP (+) transfonnant NN6 with non-detectable orange-pigment. Fungal vacuoles were stained with 8 pM F M 4-64 or 10 pM CMAC and observed using a Zeiss LSM 5 Pascal or Zeiss LSM 510 Meta confocal laser scanning microscope (CLSM) after 24, 48, and 72 h . . o . . . . . . . . 1ncubat10n at 30 C on aflatoxrn-rnducrng media, YES or non-aflatoxm-rnducrng media, YEP, agar blocks. (A) NN6 stained with FM 4-64 at 24 h on YES. Red fluorescent vacuolar membranes were observed in hyphae but very little green fluorescence was detected. (B), (C), and (D) NN6 stained with FM 4-64 at 48 h on YES. EGFP-tagged Nor-1 was localized in vacuoles and the cytoplasm. The EGFP-tagged Nor-1 localized in small vacuoles was associated with the cytoplasmic membrane in panel (C). Green fluorescence was observed in the vacuolar lumen in panel (D) as shown by an extended focus image of Z-stacks (Z—section interval: 0.46 pm). (E) NN6 stained with CMAC at 48 h on YES. EGFP-tagged Nor-1 was localized in vacuoles. The EGFP-tagged Nor-1 localized in small vacuoles was associated with the cytoplasmic membrane. (F) NN6 stained with FM 4-64 at 72 h on YES. EGFP-tagged Nor-1 was localized in vacuoles of hyphae and conidia. (G) NN6 stained with FM 4-64 at 72 h on YEP. Red fluorescent vacuolar membranes were observed in hyphae but green fluorescence was not detected. Each panel shows a red fluorescence image (FM 4-64) or a blue fluorescence image (CMAC) (top left), a green fluorescence image (EGF P) (top right), a transmitted image (bright field or differential interference contrast [DIC]) (bottom left), and a merged image (bottom right) except for panel (D) in which only a merged image is shown. Scale bars, 10 pm. Images in this dissertation are presented in color. 246 64 NN6, YES, 24 h, FM 4- A. FM 4-64 NN6, YES, 48 h, B. 247 Figure 4.20. (cont’d) FM 4—64 7 C. NN6, YES, 48 h .x. . Jul. . /...Pr.l..r.’..\u. m raw. . bfluw? l...... A. . .. 7;. A —64 (A... ......-..- 2.. . . ,48h,FM4 , YES D. NN6 248 Figure 4.20. (cont’d) E. NN6, YES, 48 h, CMAC F. Figure 4.20. (cont’d) G. NN6, YEP, 72 h, FM 4-64 250 localization of green fluorescence at 48 h (Table 4.3). Pairwise multiple comparison confirmed that the percentage of vacuole localization of green fluorescence in B3-15 at 48 h is significantly less than in NN6 at 48 h (P<0.05). However, NN6 and B3-15 showed 83.5 % and 83.0 % of vacuole localization of green fluorescence at 72 h, respectively (Table 4.3). Time-course of AFB] production The above data (Table 4.3) of vacuole localization of green fluorescence suggest that NN6 was under starvation conditions at 72 h. To confirm starvation conditions in slide culture at 72 h incubation, EGFP-tagged Nor-1 producing transformant NN6 and EGFP producing transfonnant B3-15 were cultured in YES media and on YES agar blocks. Dry weight and AFB] production of the transformants were then analyzed after 24, 48, and 72 h incubation. Dry weights of these transformants were similar at each time point in liquid culture (Figure 4.21). A transition from active growth to stationary phase was observed between 48 and 72 h in liquid culture as described previously (Liang et al. , 1997; Chiou et al., 2002). AFB] was not detected in the transformants at 24 h, but high levels of AFB] were detected at 48, and 72 h in both liquid and slide cultures as described previously (Figure 4.21 and 4.22) (Liang et al. , 1997; Chiou et al. , 2002). Therefore, the time-course experiments suggested that starvation conditions occurred at 72 h incubation in both liquid and slide cultures. 251 Table 4.3. Comparison of vacuole localization of EGFP in EGF P (+) transfonnant B3-15 with that of EGFP-tagged Nor-1 in EGFP (+) transfonnant NN6 with non-detectable orange-pigment. Time (h) 133-15 (% ) NN6 (% ) 48 7.0:I:6.0 75.5:135 72 83.0:1:5.0 83.5:1:2.5 * % = # of green fluorescent vacuoles/ # of large- and mid-size of vacuoles (> 5 pm) in 40 microscopic fields (2 - 3 hyphae per field) 252 Time-course of transformants 133-15 and NN6 —o— B3-IS(D.W) -o-- NN6(D.W) + B3-15(AFBI) -A-- NN6(AFBI) 0.5 ~ 20 0.1 4 0.05 Dry Weight (g) AFB] (ug/ml) - 10 0.01 - 0.005 Time (hr) Figure 4.21. AFB] production in transformant NN6 (EGFP-tagged Nor-1) and transformant B3-15 (EGFP). AFB] was measured after 24, 48, and 72 h incubation at 30 0C with shaking at 150 rpm in YES medium. 253 Time-course of transformants B3-15 and NN6 on slide culture 100 + B3-15(AFBI) 80 _ --1:1-- NN6(AFBI) E 60 ‘ a. 3 _ E < ‘° ’ 20 -‘ o C I I 0 20 40 60 80 Time (hr) Figure 4.22. AFB] production in transfonnant NN6 (EGFP—tagged Nor-1) and transfonnant B3-15 (EGF P) on slide culture. AFB] was measured after 24, 48, and 72 h incubation at 30 0C on YES agar. 254 DISCUSSION We demonstrated that N-terminally or C-terminally EGFP-tagged Nor-1 fusion protein was enzymatically functional in A. parasiticus and that the Nor-l fusion protein was localized in the cytoplasm and vacuoles of the hyphae at 48 and 72 h contrary to our hypothesis. Initially we used a C-terminally egfp-tagged nor-1 plasmid (pAPCGFPNFNB) for fungal transformation. The hinge region between the nor-1 and egfp coding regions contains 5 amino acids. We decided to use 5 amino acids in the hinge region because other researchers have used 2-10 amino acids in the region (V aldez- Taubas et al. , 2000; Bafiuelos et al., 2002). We used a C-terminally egyp-tagged nor-I plasmid instead of an N—terminally egp-tagged nor-1 plasmid because we predicted that C-terminal EGFP would be more efficient at correct fusion protein folding. However, we did not identify EGFP (+) transformants producing increased amounts of aflatoxin compared to the recipient strain B62 even though a 58 kDa fusion protein was produced and the plasmid was integrated into 3’ nor-1 or 5’ nor-1 / ORF region in EGFP (+) transformants (Figure 4.4, 4.9, and 4.13). We speculated that either the hinge region might not provide sufficient space between two proteins or that the C-terminus of Nor-1 is critical to its activity or transport. Tavoularis et al. (2001) reported that the number of amino acids in the hinge region is critical for correct expression and translocation of the C-terminally SGFP-tagged fusion protein; in this study the fusion protein contained 4 amino acids in the hinge region and was functional. Therefore, we constructed 2 additional plasmids, one (pAPCGFPLNFNB) in which the hinge region contained 13 amino acids and the other (pAPNGFPNFNB) in which egfi was fused to the N-terminus 255 of nor—1 and the hinge region contained 7 amino acids. Fungal transformation with pAPCGFPLNFNB or pAPNGFPNFNB generated EGFP (+) transformants producing increased amounts of aflatoxin compared to B62, indicating that either 5 amino acids did not provide an enough space between the two proteins in C-terminally EGFP-tagged Nor-1 or C-terrninal EGFP negatively affected Nor-1 activity in C-terrninally EGFP- tagged Nor-l. TLC and ELISA analyses showed that 1 light orange-pigmented EGF P (+) transformant carrying pAPCGFPLNFNB produced 20 fold more AFB] compared to the recipient strain B62 but produced only 10 % of AFB] produced by the wild-type strain SU-l (Figure 4.5 and 4.6). The analyses also showed that N-terminally EGFP-tagged Nor-1 protein functionally complemented the non-functional Nor-1 protein in A. parasiticus B62 although light orange-pigmented EGFP (+) transformants produced less AFB] than SU-l (Figure 4.7 and 4.8). nor-1 encodes an NADPH-dependent ketoreductase (Zhou and Linz, 1999) and we observed an amino acid motif (Gly-X-Gly- X-X-Leu) similar to the conserved NADP binding motif (Gly-X-Gly-X-X-Ala) starting at Gly, (amino acid residue 35) and another amino acid motif (Tyr-Gly-Val-Ser—Lys-Leu— Ala-Ala-Asn-Tyr-Met) found in the family of short-chain alcohol dehydrogenases starting at Tyr, (amino acid residue 291) (Trail et al. , 1994). We speculate that the N-terminal domain may have been partially blocked by EGFP in N-terminally EGFP-tagged Nor-1, resulting in a slight decrease in Nor-l activity. Also, EGF P may interfere with the C- terrninal domain in C-terminally EGFP-tagged Nor-l , resulting in a much greater decrease in Nor-1 activity. Western blot, Southern hybridization, and PCR analyses showed that orange- 256 pigemented EGF P (+) transformants produced a 58 kDa firsion protein (Figure 4.9, 4.10, and 4.11 A and C) and that the plasmid was integrated into the 3’ or 5’ nor-1 / ORF region in transformants carryng pAPCGF PNFN B or pAPCGF PLNFN B and was integrated into the 3’ nor-1 / ORF region in transformants carryng pAPNGFPNFNB (Figure 4.15, 4.17, and 4.18). The orange-pigmented EGFP (+) phenotype might result from mislocalization of the fusion protein blocking transport to vacuoles for aflatoxin synthesis. Western blot, Southern hybridization, and PCR analyses showed that light orange- pigmented EGFP (+) transformant LN196 produced a 58 kDa fusion protein but did not produce a 31 kDa non-functional Nor-1 (Figure 4.10 A), that multiple plasmids were integrated into 3’ and 5’ nor-1 / ORF, and niaD regions in the transfonnant, and that a 1.35 kb band was not produced from the non-functional nor-1 gene region (Figure 4.17 B). These data suggest that the production of the 31 kDa protein was negatively affected by multiple integration of the plasmid. Also, Southern hybridization and PCR analyses showed that in N5 (EGFP [-]), pAPCGFPNFNB was integrated into the niaD locus by single-crossover, that in LN2 (EGF P H), pAPCGFPLNFNB was integrated into the niaD locus by double-crossover (gene replacement), that in NNl (EGF P H), pAPNGFPNFNB was integrated into the niaD locus by double-crossover (gene replacement), and that in NN244 (EGFP [-]), pAPNGFPNFNB was integrated into the niaD locus by single- crossover (Figure 4.15, 4.17, and 4.18). These data were consistent with previous observations in which no detectable aflatoxin gene activity was observed when the plasmid was integrated into either the niaD or pyrG locus outside of the aflatoxin gene cluster (Liang et al. , 1997; Chiou et al. , 2002). 257 In PCR analyses of the samples used in this study, an unexpected 1.7 kb band was produced (Figure 4.17 B and 4.18 B). This band might result from non—specific binding of primers to the newly introduced plasmid region and to the native DNA because the same size PCR product was not produced when plasmids alone were used as a template for PCR reaction (data not shown). To determine if the EGFP-tagged Nor-1 fusion protein carried fuctional wild-type Nor-l protein, PCR was conducted. DNA sequencing showed that leucine was mutated to proline at the deduced amino acid residue 227 of non-firnctional Nor-1 in B62 (Figure 4.12). It suggests that the amino acid change may have generated conformational change to secondary structure of native Nor-1. Confocal laser scanning microscopy (CLSM) revealed cytoplasm and vacuole localization of N-terminally or C-terrninally EGFP-tagged Nor-l in hyphae of light or non-detectable orange-pigmented EGF P (+) transformants, LN196 and NN6, at 48 h (Figure 4.19 D to H and 4.20 B to E). These data suggest that the Nor-1 protein is synthesized in the cytoplasm and transported to vacuoles where aflatoxin is actively produced. This result is different from previous observations, in which Nor-1 was primarily observed in the cytoplasm at 24- 48 h culture using transmission electron microscopy (TEM) after immunogold labeling (Lee et al., 2004) and cell fractionation (Zhou, 1997). This observation might be explained because we used live samples in real time in this study instead of fixed samples in the previous study. Alternatively, the acidic pH (pH 5—6) of vacuoles may have negatively affected binding of Nor-l antibody to Nor- ] localized in vacuoles in the previous study. Also, grinding of mycelia under liquid 258 nitrogen for cell fiactionation may have disrupted intact vacuoles in the cell fractionation study. Time-course experiments suggested that NN6 was under starvation conditions in 72 h liquid and slide cultures (Figure 4.21 and 4.22). Therefore, these data suggest that EGFP-tagged Nor-1 is transported to vacuoles by the Cvt pathway under conditions that promote active grth and later it is transported to vacuoles by autophagy under starvation conditions (Figure 3.22). In support of this idea, we used prediction programs (WoLF PSORT, PSORT II, and iPSORT; http://www.psort.org) based on sequence analysis databases for fungi, yeast, animal, and plant to identify vacuolar localization or targeting signals in Nor-1. However, a possible vacuolar targeting motif was not found and possible sub-cellular localization sites were predicted as follows: cytoplasm, 65.2 %; nucleus, 17.4 %; vacuole, 4.3 %; vesicles of secretory system, 4.3 %; cytoskeletal, 4.3%; mitochondria, 4.3%. Vacuolar sorting sequences from yeast or plant described in Chapter 1 were also not found in Nor-l although LQRP at amino acid residue 45 was similar to the LQRD motif found in phytohemagglutinin. We speculate that the sequence analysis database for filamentous fungi is incomplete because a vacuolar targeting motif for OmtA, which is known to be localized in vacuole-like structures (Lee et al., 2004), was not found and sub-cellular localization to vacuoles was not predicted using these programs. The recipient strain B62 produces a non-functional Nor-1, but small amounts of aflatoxin are still produced by this strain as shown in TLC analyses. We observed small round structures outside of the fungal cell wall and in the media near the fungal cell wall (Figure 4.19 B). The size of the small round structures was similar to that of the small 259 vacuoles or vesicles in the hyphae (Figure 4.19 B). We speculate that exocytosis generated those small vacuoles or vesicles as part of secretion of aflatoxin into the media. Also, we observed small vacuoles associated with the cytoplasmic membrane (Figure 4.19 H) and these appear to have budded from large vacuoles for exocytosis. LN196 stained with CMAC at 48 h showed black dots in some of the vacuoles (Figure 4.19 G). The black dots might be vesicles carrying vacuolar resident proteins or autophagosomes in which their membranes were fused with vacuolar membranes (Figure 1.6). Also, CLSM data showed that LN196 and NN6 (light or non-detectable orange- pigmented EGFP [+]) did not produce any EGF P fluorescence when they were cultured on non-aflatoxin-inducing media, YEP (Figure 4.19 J and 4.20 G). It was consistent with previous observations in which it was reported that peptone as a sole carbon source does not support aflatoxin gene expression and aflatoxin production (Skory et al. , 1993; Buchanan and Lewis, 1984). In summary, N-terminally or C-terminally EGFP-tagged Nor-1 protein functionally complemented the non-functional Nor-1 in all light or non-detectable orange- pigrnented EGFP (+) transformants. In transformants N-terminally EGFP-tagged Nor-l protein showed relatively similar levels of Nor-1 activity as the wild-type strain SU-l while C-terminally EGFP-tagged Nor-1 protein showed decreased levels of Nor-1 activity compared to SU- 1. Western blot analysis data indicated that the EGFP-tagged Nor-1 fusion protein was produced intact and full-length form in all light or non-detectable orange-pigmented EGF P (+) transformants. Southern hybridization and PCR analyses showed that the plasmid was integrated into the nor-1 locus in all light or non-detectable orange-pigmented EGFP (+) transformants. Confocal laser scanning microscopy (CLSM) 260 data indicated that N-terminally or C-terminally EGFP-tagged Nor-1 was localized in the cytoplasm and vacuoles of fungal hyphae on aflatoxin-inducing solid media at 48 and 72 h and especially it was detected inside of the vacuole. Time-course experiments strongly suggested that the light or non-detectable orange-pigmented EGF P (+) transformants were exposed to starvation conditions in 72 h slide culture. Therefore, we concluded that Nor-1 was synthesized in the cytoplasm and transported to vacuoles of fungal hyphae by specific targeting for aflatoxin synthesis. ACKNOWLEDGMENT We thank Dr. Melinda K. Frame and Dr. Shirley A. Owens (Center for Advanced Microscopy at Michigan State University) for help in confocal laser scanning microscopy and Dr. Stephen A. Osmani (Ohio State University) for providing a repeated Gly-Ala sequence for making eg/p-tagged nor-I fusion plasmids. All experiments described in Chapter 4 were performed by Sung-Yong Hong. Chapter 4 will be submitted for publication in the near future. 261 CONCLUSIONS The long-term goal of our laboratory is to eliminate aflatoxin B] from the food chain by understanding the molecular mechanisms that regulate the expression of key genes involved in AFB] biosynthesis. My research is focused on the sub-cellular localization of aflatoxin biosynthetic enzymes in A. parasiticus. Information about the sub-cellular localization of enzymes involved in aflatoxin biosynthesis will allow us to design safe, practical, and inexpensive strategies to block aflatoxin synthesis in the field and during storage of nuts, seeds, and grains. In this dissertation we demonstrated that the EGFP reporter system could be used to monitor sub-cellular localization of aflatoxin biosynthetic enzymes in A. parasiticus and that EGFP-tagged Ver-l or Nor-1 was localized in the cytoplasm and vacuoles of the fungal hyphae at 48 and 72 h on aflatoxin-inducing solid media. The data proved that EGFP was expressed by the ver-I promoter in A. parasiticus and that its expression pattern by the ver-I promoter was similar to the wild-type ver-I promoter. Also, the data confirmed that EGFP-tagged Ver-l or Nor-1 was transported from cytoplasm to vacuoles, especially the vacuolar lumen, of the fungal hyphae during a time at which maximum rates of aflatoxin synthesis are observed. There was no evidence of turnover of the fusion proteins in vacuoles at this time as proved by Western blot analysis. Therefore, the data indicate that the early aflatoxin pathway enzyme Nor-1 and the middle aflatoxin pathway enzyme Ver-l are synthesized in the cytoplasm and then transported to vacuoles for aflatoxin synthesis. 262 The data also suggest that norsolorinic acid (NA) is transported to vacuoles by small vesicles and then converted to averantin (AVN) by Nor-1 during aflatoxin biosynthesis. We predict that Ver-l and OmtA are also transported to vacuoles by the Cvt pathway and involved in aflatoxin synthesis. Finally, our data suggest that aflatoxin is released to the media through vesicles or small vacuoles by exocytosis (Figure 4.23). 263 Figure 4.23. Proposed model for aflatoxin production in fungal cells. N, R, Va, and Ve represent nucleus, ribosomes, vacuoles, and vesicles, respectively. AF represents aflatoxins. Black arrows indicate transport of the aflatoxin enzymes and white arrows indicate transport of substrates and end products. 264 CHAPTER 5 FUTURE STUDIES We demonstrated that Ver-l and Nor-1 were synthesized in the cytoplasm and transported to vacuoles of fungal hyphae using EGFP fusions. However, the enzymatic function of Ver-l and Nor-1 in vacuoles for aflatoxin synthesis must be confirmed. Therefore, substrate-feeding experiments using purified vacuoles would provide supporting evidence of their function for aflatoxin synthesis in vacuoles. In addition to egfi9-tagged ver-I and nor-I fusion constructs, we also constructed 3 plasmids designed to express N-terminally or C-terminally eg/p-tagged omtA (Figure 5.1). We have not completed this study to date because of problems encountered in construction of the recipient strain. To complement non-functional Ver-l in A. parasiticus LWl432-N33 (ver-I, omtA, niaD, wh-I), we first transformed A. parasiticus LWl432-N33 (ver-I, omtA, niaD, wh-1)with pVer-Ben (Liang et al., 1996). LWl432- N33 (ver-I , omtA, niaD, wh-I) was derived from LWl432 (ver-I, omtA, wh-I) (Lee et al. , 2002) by spontaneous mutation using potassium chlorate selection. TLC analysis showed that isolate Ben 4 produced sterigmatocystin as expected. Also, Southern hybridization showed that the plasmid pVer-Ben was integrated into the ver-I A locus in transformant Ben4. The N-terminally or C-terminally egfiJ-tagged omtA plasmids were then transformed into Ben4 to complement non-functional OmtA. However, out of 400 transformants, none produced blue fluorescent haloes around their colonies on CAM (coconut agar medium) nor did they produce visible aflatoxin bands in TLC. The 266 Asc I418 Fse I 1238 [’31 I 11974 Nat 11964 Sat] 11962 Amp, omtA " terminator ! _ egfp pAPC GFPOFNB 1 (14268 bp) ' omtA promoter lgene Pst r 4424 Pac I 4594 Sal I/Xha I 4610 Pst I 7762 Figure 5.1. Restriction endonuclease map of plasmid, pAPCGFPOFNB. The 2.6 kb omtA promoter/gene was fused in frame to the 0.7 kb egfp coding region, followed by the 0.8 kb omtA terminator. The 7.4 kb niaD fragment was inserted as a selectable marker for transformation of the recipient strain Ben4 (omtA, niaD, wh-I). 267 difficulty in this experiment lies in use of the double mutant A. parasiticus LWl432 (ver-I , omtA, wh-I). Failure of the complementation to produce aflatoxin could be a problem in complementation by either Ver-l or EGFP-tagged OmtA. Therefore, generation of an omtA knockout mutant of A. parasiticus NR-l and transformation of the mutant with the egfp-tagged omtA plasmids would provide a better strategy for sub- cellular localization of EGFP-tagged OmtA. We also constructed one C-terrninally egfp-tagged vbs plasmid (Figure 5.2). This could be used for sub-cellular localization of VBS if one utilizes vbs knockout mutant VBS-DDl-16, VBS-DDl-23, or VBS-DDl-27 (Sakuno et al., 2005). However, one must generate a mutation at the niaD locus to use one of these strains as a recipient strain for transformation. We demonstrated that Ver-l and Nor-1 were targeted to vacuoles in this study. OmtA was previously shown to be targeted to vacuoles (Lee et al., 2004). Based on these data, one would predict that each of enzymes would contain vacuolar sorting sequences for targeting to vacuoles. Western blot analyses of Ver-l, Nor-1, OmtA, or VBS from wild-type strain SU-l after 48 h culture showed that each aflatoxin protein appears as a doublet band (V er—l, 28 kDa and 25 kDa; Nor-l, 31 kDa and 28 kDa; OmtA, 45 kDa and 42 kDa; VBS, 78—79 kDa and 72 kDa) (Liang et al., 1997; Zhou 1997; Yu et al., 1993; Lee et al., 2002; Chiou et al. , 2004). For Ver-l, a 28 kDa and 25 kDa doublet was detected by anti-Ver-l antibody in SU-l, CSlO-N2, and NV27 at 72 h (Figure 5.3). Others reported that a 45 kDa OmtA undergoes N-terminal cleavage between amino acid residues 41 and 42 to produce a 42 kDa OmtA (Keller et al. , 1993; Yu et al. , 1993) and that a 78-79 kDa VBS (versiconal cyclase or VER B synthase) undergoes N-terminal 268 Asc I 418 Pst I 468 Pst I 12784 Sal I 12772 Fse I 1928 Amp'vbs h‘ J ‘ Not 1 2654 terminator egfir ' pAPCGFPBFNB ' (15078 bp) vbs promoter lgene “ ' niaD Pac I 5404 Sal I/Xha I 5420 Pst I 8572 Figure 5.2. Restriction endonuclease map of plasmid, pAPCGFPBFNB. The 2.8 kb vbs promoter/gene was fused in frame to the 0.7 kb egp coding region, followed by the 1.5 kb vbs terminator. The 7.4 kb niaD fragment was inserted as a selectable marker for fungal transformation. 269 60-4 i ' “55m:- I'“'" m“ #28 kDa 24.9 “25 kDa Figure 5.3. Western blot analysis of Ver-l from transfonnant NV27, the recipient strain CS10-N2, and the wild-type strain SU-l. Fungal proteins were extracted from NV27, CSlO-N2, and SU-l grown in 100 ml of YES for 72 h at 30 °C with shaking at 150 rpm. Approximately 30-50 pg of proteins were separated by 12% SDS-PAGE, transferred onto PVDF membranes, and probed with Ver—l polyclonal antibody. EGFP-tagged Ver-l has a molecular mass of 55 kDa. The 28 kDa protein represents a non-functional Ver-l (with the exception of a functional Ver-l from SU-l) and the 25 kDa protein represents a cleavage product of the 28 kDa protein. A Benchmark prestained protein ladder was used as a molecular mass marker (M). 270 cleavage between amino acid residues 41 and 42 to produce OAVN (5’-oxoaverantin) cyclase. VBS and OAVN cyclase catalyze different biochemical steps in the aflatoxin biosynthetic pathway (Sakuno et al., 2005). Like VBS, both 45 kDa and 42 kDa OmtA proteins produce functional proteins, but unlike VBS both OmtA proteins catalyze the same step in the aflatoxin pathway (Keller et al. , 1993; Yu et al. , 1993). Also, both 31 kDa and 28 kDa Nor-1 proteins were not produced from nor-1 knockout mutant (Zhou, 1997). Therefore, we speculate that the 28 kDa Ver-l may undergo N-terminal cleavage to produce the 25 kDa Ver-l. Similarly we speculate that N-terminally EGFP-tagged Ver—l fusion protein may be cleaved to release N-terrninal EGF P (in this case we predict both cleaved N-terminal EGF P and Ver-l will be localized in vacuoles because the cleavage occurs just before or in vacuoles). We attempted to sequence the N-terminus of the 25 kDa Ver-l to address this speculation. However, sequencing of 25 kDa Ver-l was not successful. We speculate that the 28 kDa Ver-l might undergo N-terminal cleavage between amino acid residues 41 and 42 because both OmtA and VBS undergo N-terminal cleavage at this location. Also, we must analyze the characteristics of amino acids in N- terrninal sequences of the enzymes to identify potential cleavage sites. Based on our data, we predict that Ver-l , Nor-1, and OmtA contain vacuolar sorting sequences for targeting to vacuoles in N-terminal sequences. We are also aware of the possibility that these enzymes catalyzed reactions at different steps in addition to known steps in the aflatoxin biosynthetic pathway like VBS. To prove that the aflatoxin enzymes are targeted to vacuoles by N-terminal sequences, these sequences could be deleted or site-directed mutations could be introduced into the cleavage recognition site to block localization of the enzymes to vacuole. We predict that aflatoxins will not be produced when 271 localization of the enzyme to vacuoles is blocked. Analysis of complex formation between aflatoxin biosynthetic enzymes is also another area for firture experiment. To accomplish this we could use co-localization of GFP-tagged enzyme (OmtA for example) and either RF P (red fluorescent protein)- or BFP (blue fluorescent protein)-tagged enzyme (Nor-l for example) by FRET (fluorescence resonance energy transfer) which is used for monitoring protein-protein interactions. Determination of EGFP expression at the ver-I B locus and the sugar utilization gene cluster might also be an attractive experiment to analyze the molecular mechanisms that mediate position-dependent regulation of aflatoxin genes associated with the aflatoxin gene cluster in the chromosome. 272 LIST OF REFERENCES Antilla, L., and P. J. Cotty. 2002. The ARS-ACRPC partnership to control aflatoxin in Arizona cotton: current status. Mycopathologia 155:64. Asao, T., G. Buchi, M. M. Abdel-Kader, S. B. Chang, E. L. Wick, and G. N. Wogan. 1963. Aflatoxins B and G. J. Am. Chem. Soc. 85:1706-1707. Ausubel, F. M., R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, J. A. Smith, and K. Strubl (Eds). 2003. Current protocols in molecular biology. John Wiley & Sons, New York. Baba, M., M. Osumi, S. V. Scott, D. J. Klionsky, and Y. Ohsumi. 1997. Two distinct pathways for targeting proteins from the cytoplasm to the vacuole/lysosome. J. Cell Biol. 139: 1687-1695. Bafiuelos, O., J. Casqueiro, S. Steidl, S. Gutierrez, A. Brakhage, J. F. Martin. 2002. Subcellular localization of the homocitrate synthase in Penicillium chrysagenum. Mol. Genet. Genomics 266:7] 1-719. Bassham D. C. and N. V. Raikhel. 2000. Unique features of the plant vacuolar sorting machinery. Curr. Opin. Cell Biol. 12:491-495. Bennett, J. W., J. J. Dunn, and C. I. Goldsman. 1981. Influence of white light on production of aflatoxins and anthraquinones in Aspergillus parasiticus. Appl. Environ. Microbiol. 41:488-491. Bennett, J. W., and L. A. Goldblatt. 1973. The isolation of mutants of Aspergillus flavus and A. parasiticus with altered aflatoxin producing ability. Sabouraudia 11:235-241. Bennett, J. W., L. S. Lee, S. M. Shoss, and G. H. Boudreaux. 1980. Identification of averantin as an aflatoxin B] precursor: placement in the biosynthetic pathway. Appl. Environ. Microbiol. 39:835-839. 273 Bennett, J. W., and K. E. Papa. 1988. The aflatoxigenic Aspergillus spp. Adv. Plant Pathol. 6:263-280. Bennett, J. W., P. L. Rubin, L. S. Lee, and P. N. Chen. 1979. Influence of trace elements and nitrogen sources on versicolorin production by a mutant strain of Aspergillus parasiticus. Mycopathologia 69: 161-166. Bhatnagar, D., T. E. Cleveland, and D. G. I. Kingston. 1991. Enzymological evidence for separate pathway for aflatoxin B] and B; biosynthesis. Biochem. 30:4343-4350. Bhatnagar, D., T. E. Cleveland, and E. B. Lillehoj. 1989. Enzymes in aflatoxin B1 biosynthesis: strategies for identifying pertinent genes. Mycopathologia 107:75- 83. Bhatnagar, D., K. C. Ehrlich, and T. E. Cleveland. 1992. Oxidation-reduction reaction in biosynthesis of secondary metabolites, p. 255-286. In Bhatnagar, D., E. B. Lillehoj, and D. K. Aroa. (ed.), Handbook of applied mycology. Marcel Dekker Inc., New York. Bhatnagar, D., K. C. Ehrlich, and T. E. Cleveland. 2003. Molecular genetic analysis and regulation of aflatoxin biosynthesis. Appl. Microbiol. Biotechnol. 61:83-93. Bhatnagar, D., S. P. McCormick, L. S. Lee, and R. A. Hill. 1987. Identification of 0- methylsterigmatocystin as an aflatoxin B] and G] precursor in Aspergillus parasiticus. Appl. Environ. Microbiol. 53:1028-1033. Bhatnagar, D., G. A. Payne, J. E. Linz, and T. E. Cleveland. 1995. Molecular biology to eliminate aflatoxin. Inform 6:262-271. Blount, W. P. 1961. Turkey “X” disease. J. Brit. Turkey Fed. 9:55-58. Bok, J. W., and N. P. Keller. 2004. LaeA, a regulator of secondary metabolism in Aspergillus spp. Eukaryot. Cell 3:527-535. 274 Bradford, M. M. 1976. A rapid and sensitive method for the quantification of rnicrogram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72:248-254. Brown, M. P., C. S. Brown-Jenco, and G. A. Pane. 1999. Genetic and molecular analysis of aflatoxin biosynthesis. Fungal Genet. Biol. 26:81-98. Brown, R. L., P. J. Cotty, and T. E. Cleveland. 1991. Reduction in aflatoxin content of maize by atoxigenic strains of Aspergillus flavus. J. Food Prot. 54:623-626. Buchanan, R. L., D. Federowicz, and H. G. Stahl. 1985. Activities of tricarboxylic acid cycle enzymes in an aflatoxicgenic strain of Aspergillus parasiticus after a pepton to glucose carbon source shift. Trans. Brit. Mycol. Soc. 84:267-275. Buchanan, R. L., S. B. Jones, W. V. Gerasimowicz, L. L. Zaika, H. G. Stahl, and L. A. Ocker. 1987. Regulation of aflatoxin biosynthesis: assessment of the role of cellular energy status as a regulator of the induction of aflatoxin production. Appl. Environ. Microbiol. 53: 1224-123 1 . Buchanan, R. L., and D. F. Lewis. 1984. Regulation of aflatoxin biosynthesis: effect of glucose on activities of various glycolytic enzymes. Appl. Environ. Microbiol. 48:306-310. Bu’Lock, J. D. 1965. The biosynthesis of natural products. McGraw-Hill, New York. Butler, W. H., and J. S. Wigglesworth. 1966. Effects of aflatoxin B1 on the pregnant rat. Br. J. Exp. Pathol. 47:242-247. Carnaghan, R. B. A. 1965. Hepatic tumors in ducks fed a low level of toxin groundnut meal. Nature 208:308. Cary, J. W., J. M. Dyer, K. C. Ehrlich, M. S. Wright, S. H. Liang, and J. E. Linz. 2002. Molecular and functional characterization of a second copy of the aflatoxin regulatory gene, aflR-2, from Aspergillus parasiticus. Biochim. Biophys. Acta 1576:316-323. 275 Cary, J. W., K. C. Ehrlich, J. M. Bland, and B. G. Montalbano. 2006. The aflatoxin biosynthesis cluster gene, ale, encodes an oxidoreductase involved in conversion of versicolorin A to demethylsterigmatocystin. Appl. Environ. Microbiol. 72:1096-1101. Cary, J. W., J. E. Linz, and D. Bhatnagar. 2000. Aflatoxins: biological significance and regulation of synthesis, p. 317-361. In Cary, J. W., D. Bhatnagar, and J. Linz. (ed.), Microbial foodbom disease: mechanisms of pathogenesis and toxin synthesis. Technomic Publishing, Lancaster, PA. Cary, J. W., M. Wright, D. Bhatnagar, R. Lee, and F. S. Chu. 1996. Molecular characterization of an Aspergillus parasiticus dehydrogenase gene, norA, located on the aflatoxin biosynthesis gene cluster. Appl. Environ. Microbiol. 62:360- 366. Chalfie, M., Y. Tu, G. Euskirchen, W.W. Ward, and D. C. Prasher. 1994. Green fluorescent protein as a marker for gene expression. Science 263:802-805. Chang, P.-K. 2003. The Aspergillus parasiticus protein AflJ interacts with the aflatoxin pathway-specific regulator AflR. Mol. Gen. Genomics 268:711-719. Chang, P.-K., J. W. Bennett, and P. J. Cotty. 2001. Association of aflatoxin biosynthesis and sclerotial development in Aspergillus parasiticus. Mycopathologia 153:41-48. Chang, P.-K., J. W. Cary, D. Bhatnagar, T. E. Cleveland, J. W. Bennett, J. E. Linz, C. P. Woloshuk, and G. A. Paynne. 1993. Cloning of the Aspergillus parasiticus apa-2 gene associated with the regulation of aflatoxin biosynthesis. Appl. Environ. Microbiol. 59:3273-3279. Chang, P.-K., K. C. Ehrlich, J.-J. Yu, D. Bhatnagar, and T. E. Cleveland. 1995. Increased expression of Aspergillus parasiticus aflR, encoding a sequence- specific DNA-binding protein, relieves nitrate inhibition of aflatoxin biosynthesis. Appl. Environ. Microbiol. 61:2372-2377. Chang, P.-K., C. D. Skory, and J. E. Linz. 1992. Cloning of a gene associated with aflatoxin B] biosynthesis in Aspergillus parasiticus. Curr. Genet. 21:231-233. 276 Chang, P.-K., K. Yabe, and J. Yu. 2004. The Aspergillus parasiticus estA-encoded esterase converts versiconal hemiacetal acetate to versiconal and versiconol acetate to versiconol in aflatoxin biosynthesis. Appl. Environ. Microbiol. 70:3593-3599. Chang, P.-K., and J.-J. Yu. 2002. Characterization of a partial duplication of the aflatoxin gene cluster in Aspergillus parasiticus ATCC 56775. Appl. Microbiol. Biotechnol. 58:632-636. Chen, L., S. P. McCormick, and T. M. Hohn. 2000. Altered regulation of 15- acetyldeoxynivalenol production in F usarium graminearum. Appl. Environ. Microbiol. 66:2062-2065. Chiou, C.-H., L.-W. Lee, S. A. Owens, J. H. Whallon, K. L. Klomparens, C. A. Townsend, and J. E. Linz. 2004. Distribution and sub-cellular localization of The afiatoxin enzyme versicolorin B synthase in time-fractionated colonies of Aspergillus parasiticus. Arch. Microbiol. 182:67-79. Chiou, C.-H., M. Miller, D. L. Willson, F. Trail, and J. E. Linz. 2002. Chromosomal location plays a role in regulation of aflatoxin gene expression in Aspergillus parasiticus. Appl. Environ. Microbiol. 68:306-315. Chrispeels, M. J., and N. V. Raikhel. 1992. Short peptide domains target proteins to plant vacuoles. Cell 68:613-616. Cleveland, T. E., D. Bhatnagar, C. J. Foell, and S. P. McCormick. 1987a. Conversion of a new metabolite to aflatoxin B; by Aspergillus parasiticus. Appl. Environ. Microbiol. 53:2804-2807. Cleveland, T. E., J. W. Cary, D. Bhatnagar, J.-J. Yu, P.-K. Chang, C. A. Chlan, and K. Rajasekaran. 1997. Use of biotechnology to eliminate aflatoxin in preharvest crops. Bull. Inst. Compr. Agr. Sci. 5:75-90. Cleveland, T. E., A. R. Lax, L. S. Lee, and D. Bhatnagar. 1987b. Appearance of enzyme activities catalyzing conversion of sterigmatocystin to aflatoxin B1 in late-growth-phase Aspergillus parasiticus cultures. Appl. Environ. Microbiol. 53:1711-1713. 277 Chu, F. S. 1991. Mycotoxins : food contamination, mechanism, carcinogenic potential and preventive measures. Mutat. Res. 259:291-306. Cody, C. W., D. C. Prasher, W. M. Westler, F. G. Prendergast, and W. W. Ward. 1993. Chemical structure of the hexapeptide chromophore of the Aequorea green- fluorescent protein. Biochemistry 32: 1212-1218. Cohen, G., D. Dembiec, and J. Marcus. 1970. Measurement of catalase activity in tissue extracts. Anal. Biochem. 34:30-38. Cormack, B. P., G. Bertram, M. Egerton, N. A. R. Gow, S. Falkow, and A. J. P. Brown. 1997. Yeast-enhanced green fluorescent protein (yEGF P): a reporter of gene expression in Candida albicans. Microbiology 143:303-311. Cormack, B. P., R. H. Valdivia, and S. Falkow. 1996. F ACS-optimized mutants of the green fluorescent protein (GF P). Gene 173:33-38. Cotty, P. J. 1988. Aflatoxin and sclerotial production by Aspergillusflavus: influence of pH. Phytopathology 78:1250-1253. Cotty, P. J. 1990. Effect of atoxigenic strains of Aspergillusflavus on aflatoxin contamination of developing cottonseed. Plant Dis. 64:233-235. Cotty, P. J. 1994. Influence of field application of an atoxigenic strain Aspergillus flavus on the population of A. flavus infecting cotton bolls and on the aflatoxin content of cottonseed. Phytopathology 84:1270-1277. Cotty, P. J., P. Bayman, D. S. Egel and D. S. Elias. 1994. Agriculture, aflatoxins and Aspergillus, p.1-27. In K. A. Powell, A. F enwick, and J. F. Peberdy (ed.), The genus Aspergillus. Plenum Press, New York, NY. Cotty, P. J., and D. Bhatnagar. 1994. Variability among toxigenic Aspergillusflavus strains in ability to prevent aflatoxin contamination and production of aflatoxin biosynthetic pathway enzymes. Appl. Environ. Microbiol. 60:2248-2251. Council for Agricultural Science and Technology (CAST). 2003. Mycotoxins: 278 risks in plant, animal, and human systems. Report no. 1 39. Ames, Iowa. Cove, D. J. 1966. The induction and repression of nitrate reductase in the firngus Aspergillus nidulans. Biochim. Biophys. Acta 113:51-56. Crameri, A., E. A. Whitehorn, E. Tate, and W. P. Stemmer. 1996. Improved green fluorescent protein by molecular evolution using DNA shuffling. Nat. Biotechnol. 14:315-319. Cubitt, A. B., R. Heim, S. R. Adams, A. E. Boyd, L. A. Gross, and R. Y. Tsien. 1995. Understanding, improving and using green fluorescent proteins. Trends Biochem. Sci. 20:448-455. Darrah, L., and D. Barry. 1991. Reduction of preharvest aflatoxin in corn, p. 288-310. In G. A. Bray and D. H. Ryan (ed.), Mycotoxins, cancer, and health. Louisiana State University Press, Baton Rouge. Davis, N. D., and U. L. Diener. 1986. Biology of A. flavus and A. parasiticus : some characteristics of toxigenic and nontoxingenic isolates of Aspergillus flavus and Aspergillus parasiticus, p. 33-40. Proceedings of the workshops sponsored by CIMMYT, UNDP, and USAIS. Dorner, J. W. 2004. Combined effects of biological control formulations, cultivars, and fungicides on preharvest colonization and aflatoxin contamination of peanuts by Aspergillus species. Peanut Sci. 31:79-86. Dorner, J. W., R. J. Cole, and P. D. Blankenship. 1992. Use of a biocompetitive agent to control preharvest aflatoxin in drought stressed peanuts. J. Food Prot. 55:888-892. Dorner, J. W., R. J. Cole, and P. D. Blankenship. 1998. Effect of inoculum rate of biological control agents on preharvest aflatoxin contamination of peanuts. Biol. Control 12:171-176. Drahos, D., B. C. Hemming, and S. McPherson. 1986. Tracking recombinant organisms in the environment: B-galactosidase as a selectable non-antibiotic marker for fluorescent pseudomonads. Bio. Technology 4:439-444. 279 Drummond, J., and D. E. Pinnock. 1990. Aflatoxin production by entomopathogenic isolates of Aspergillus parasiticus and Aspergillusflavus. J. Invertebr. Pathol. 55:332-336. Du, W., Z. Huang, J. E. Flaherty, K. Wells, and G. A. Payne. 1999. Green fluorescent protein as a reporter to monitor gene expression and food colonization by Aspergillusflavus. Appl. Environ. Microbiol. 65:834-836. Dutton, M. F. 1988. Enzymes and aflatoxin biosynthesis. Microbiol. Rev. 52:274-295. Dvorackova, I. 1990. Aflatoxins and human health. CRC Press, Boca Raton, FL. Eaton, D. L., and E. P. Gallagher. 1994. Mechanisms of aflatoxin carcinogenesis. Annu. Rev. Pharmacol. Toxicol. 34:135-172. Eaton, D. L., H. S. Ramsdell, and G. E. Neal. 1994. Biotransformation of aflatoxins, p.45-72. In Eaton, D. L., and J. D. Groopman (ed.), The toxicology of aflatoxins: human health, veterinary and agricultural significance. Academic Press, San Diego. Edds, G. T., K. P. C. Nair, and C. F. Simpson. 1973. Effect of aflatoxin B1 on resistance in poultry against cecal coccidiosis and Marek's disease. Am. J. Vet. Res. 34:819-826. Ehrlich, K. 1987. Effect on aflatoxin production of competition between wild-type and mutant strains of Aspergillus parasiticus. Mycopathologia 97 :93-96. Ehrlich, K. C., B. G. Montalbano, and J. W. Cary. 1999. Binding of the C6-zinc cluster protein, AF LR, to the promoters of aflatoxin pathway biosynthesis genes in Aspergillus parasiticus. Gene 230:249-257. Ehrlich, K. C., B. G. Montalbano, S. M. Boué, and D. Bhatnagar. 2005. An aflatoxin biosynthesis cluster gene encodes a novel oxidase required for conversion of versicolorin A to sterigmatocystin. Appl. Environ. Microbiol. 71:8963-8965. 280 Elis, J., and J. A. DiPaolo. 1967. Aflatoxin B]. Induction of malformations. Arch. Pathol. 83:53-57. Ellis, W. O., J. P. Smith, B. K. Simpson, and J. H. Oldham. 1991. Aflatoxin in food: occurrence, biosynthesis, effects on organisms, detection, and methods of control. Crit. Rev. Food Sci. Nutr. 30:403-439. Epstein, S. M., B. Bartus, and E. Farber. 1969. Renal epithelial neoplasms induced in male Wistar rats by oral aflatoxin B]. Cancer Res. 29: 1045-1050. Errampalli, D., K. Leung, M. B. Cassidy, M. Kostrzynska, M. Blears, H. Lee, and J. T. Trevors. 1999. Applications of the green fluorescent protein as a molecular marker in environmental microorganisms. J. Microbiol. Methods 35:187-199. Essigmann, J. M., R. G. Croy, R. A. Bennett, and G. N. Wogan. 1982. Metabolic activation of aflatoxin B] : patterns of DNA adduct formation, removal, and excretion on relation to carcinogenesis. Drug. Metab. Rev. 13:581-602. F eng, G. H., F. S. Chu, and T. J. Leonard. 1992. Molecular cloning of genes related to aflatoxin biosynthesis by differential screening. Appl. Environ. Microbiol. 58:455-460. Fernandez-Abalos, J. M., H. Fox, C. Pitt, B. Wells, and J. H. Doonan. 1998. Plant- adapted green fluorescent protein is a versatile vital reporter for gene expression, protein localization and mitosis in the filamentous fungus, Aspergillus nidulans. Mol. Microbiol. 27:121-130. Floyd, J. C., J. Mills, and J. W. Bennett. 1987 . Biotransformation of sterigmatocystin and absence of aflatoxin biotransformation by blocked mutants of Aspergillus parasiticus. Exp. Mycol. 11:109-114. Giambrone, J. J., U. L. Diener, N. D. Davis, V. S. Panagala, and F. J. Hoerr. 1985. Effects of purified aflatoxin on broiler chickens. Poult. Sci. 64:852-858. Giniger, E., S. M. Varnum, and M. Ptashne. 1985. Specific DNA-binding of GAL4, a positive regulatory protein of yeast. Cell 40:767-774. 281 Gordon, C. L., V. Khalaj, A. F. J. Ram, D. B. Archer, J. L. Brookman, A. P. J. Trinci, D. J. Jeenes, J. H. Doonan, B. Wells, P. J. Punt, C. A. M. J. J. van den Hondel, and G. D. Robson. 2000. Glucoamylasez: green fluorescent protein fusions to monitor protein secretion in Aspergillus nigar. Microbiology 146:415- 426. Groopman, J. D., L. G. Cain, and T. W. Kensler. 1988. Aflatoxin exposure in human populations: measurements and relationship to cancer. Crit. Rev. Toxicol. 19:1 13-145. Gupta, G. D., and I. B. Heath. 2002. Predicting the distribution, conservation, and functions of SNAREs and related proteins in fungi. Fungal Genet. Biol. 36:1-21 Hagedorn, C. 1994. Spontaneous and intrinsic antibiotic resistance markers, p. 575- 591. In Weaver, R. S., S. Angle, and P. Bottomley. (ed.), Methods of soil analysis, part 2, microbiological and chemical properties. Soil Science of America Inc. Madison, WI. Hall, A. J, and C. P. Wild. 1994. Epidemiolgy of aflatoxin-related diseases, p.233-258. In Eaton, D. L., and J. D. Groopman (ed.), The toxicology of aflatoxins: human health, veterinary and agricultural significance. Academic Press, San Diego. Hara-Nishimura, I., T. Shimada, K. Hatano, Y. Takeuchi, and M. Nishimura. 1998. Transport of storage proteins to protein storage vacuoles is mediated by large precursor-accumulating vesicles. Plant Cell 10:825-836. Harris, J. L. 1986. Modified method for firngal slide culture. J. Clin. Microbiol. 24:460-461. Haseloff, J., K. R. Siemering, D. C. Prasher, and S. Hodge. 1997. Removal of a cryptic intron and subcellular localization of green fluorescent protein are required to mark transgenic Arabidopsis plants brightly. Proc. Natl. Acad. Sci. USA 94:2122-2127. Heim, R., A. B. Cubitt, and R. Y. Tsien. 1995. Improved green fluorescence. Nature 282 373:663-664. Henderberg, A., J. W. Bennett, and L. S. Lee. 1988. Biosynthetic origin of aflatoxin G]: confirmation of sterigmatocystin and lack of confirmation of aflatoxin B] as precursors. J. Gen. Microbiol. 134:661-667. Henry, K. M., and C. A. Townsend. 2005. Ordering the reductive and cytochrome P450 oxidative steps in demethylsterigmatocystin formation yields general insights into the biosynthesis of aflatoxin and related fungal metabolites. J. Am. Chem. Soc. 127:3724-3733. Hershko, A., A. Ciechanover, and A. Varshavsky. 2000. The ubiquitin system. Nat. Med. 6:1073-1081. Hicke, L. 2001. A new ticket for entry into budding vesicles- ubiquitin. Cell 106:527- 530. Hohn, T. M., R. Krishna, and R. H. Proctor. 1999. Characterization of a transcriptional activator controlling trichothecene toxin biosynthesis. Fungal Genet. Biol. 26:224-235. Horng, J. S., P.-K. Chang, J. J. Pestka, and J. E. Linz. 1990. Development of a homologous transformation system for Aspergillus parasiticus with the gene encoding nitrate reductase. Mol. Gene Genet. 224:294-296. Hsieh, D. P. H., M. T. Lin, and R. C. Yao. 1973. Conversion of sterigmatocystin to aflatoxin B] by Aspergillus parasiticus. Biochem. Biophys. Res. Commun. 52:992-997. Hsieh, D. P. H., M. T. Lin, R. C. Yao, and R. Singh. 1976. Biosynthesis of aflatoxins: conversion of norsolonic acid and other hypothetical intermediates into aflatoxin B] by Aspergillus parasiticus. J. Agric. Food Chem. 24:1167-1170. Inouye, S., and F. I., Tsuji. 1994. Aequorea green fluorescent protein. FEBS Lett. 341:277-280. 283 Johnston, M. 1987. A model fungal gene regulatory mechanism: the GAL genes of Saccharomyces cerevisiae. Microbiol. Rev. 51:458-476. Kachholz, T., and A. L. Demain. 1983. Nitrate repression of averufin and aflatoxin biosynthesis. J. Natural Products 46:499-506. Keller, N. P., H. C. Dischinger Jr., D. Bhatnagar, T. E. Cleveland, and A. H. J. Ullah. 1993. Purification of 40-kilodalton methyltransferase active in the aflatoxin biosynthetic patheway. Appl. Environ. Mocrobiol. 59:479-484. Keller, N. P., and T. M. Hohn. 1997. Metabolic pathway gene clusters in filamentous fungi. Fungal Genet. Biol. 21:17-29. Keller, N. P., C. Nesbitt, B. Sarr, T. D. Phillips, and G. D. Burrow. 1997. pH regulation of sterigmatocystin and aflatoxin biosynthesis in Aspergillus spp. Phytopathology 87:643-648. Klich, M. A., and J. I. Pitt. 1988. A laboratory guide to common Aspergillus species and their telemorphs. Commonwealth Scientific and Industrial Research Org., Div. Of Food Proc., North Ryde, New South Wales, Australia. Klionsky, D. J., and S. D. Emr. 1990. A new class of lysosomal/vacuolar protein sorting signals. J. Biol. Chem. 265:5349-5352. Klionsky, D. J., P. K. Herman, and S. D. EMR. 1990. The fungal vacuole: composition, function, and biogenesis. Microbiol. Rev 54:266-292. Krishnamachari, K. A. V. R., R. V. Bhat, V. Nagarajan, and T. B. G. Telak. 1975. Hepatitis due to aflatoxicosis. An outbreak in Western India. Lancet 1:106]- 1063. Kucharczyk, R., and J. Rytka. 2001. Saccharamyces cerevisiae- a model organism for the studies on vacuolar transport. Acta Biochim. Pol. 48:1025-1042. Labuza, T. P. 1983. Regulation of mycotoxins in food. Ames: international association of milk, food, and environmental Sanitarians. J. Food Prot. 46:260-265. 284 Lax, A. R., T. E. Cleveland, D. Bhatnagar, L. S. Lee. 1986. Enzymatic conversion of sterigmatocystin to aflatoxin B] through O-methylsterigmatocystin. Plant Physiol. 80(suppl.): 19. Lee, L. S., J. W. Bennett, A. F. Cucullu, and J. B. Stanley. 1975. Synthesis of versicolorin A by a mutant strain of Aspergillus parasiticus deficient in Aflatoxin production. J. Agric. Food Chem. 23:1132-1134. Lee, L. S., J. W. Bennett, L. A. Goldblatt, and R. E. Lundin. 1971. Norsolorinic acid from a mutant strain of Aspergillus parasiticus. J. Am. Oil Chem Soc. 48:93-94. Lee, L.-W., C.-H. Chiou, K. L. Klomparens, J. W. Cary, and J. E. Linz. 2004. Sub- cellular localization of aflatoxin biosynthetic enzymes Nor-l , Ver-l , and OmtA in time-dependent fractionated colonies of Aspergillus parasiticus. Arch. Microbiol. 181:204-214. Lee, L.-W., C.-H. Chiou, and J. E. Linz. 2002. Function of native OmtA in vivo and expression and distribution of this protein in colonies of Aspergillus parasiticus. Appl. Environ. Microbiol. 68:5718-5727. Li, X., G. Zhang, N. Ngo, X. Zhao, S. R. Kain, C.-C. Huang. 1997. Deletions of the Aequorea Victoria green fluorescent protein define the minimal domain required for fluorescence. J. Biol. Chem. 272:28545-28549. Liang, S.-H. 1996. The function and expression of the ver-I gene and localization of the Ver-l protein involved in aflatoxin biosynthesis in Aspergillus parasiticus. PhD Dissertation, Michgan State Unviersity. Liang, S.-H., C. D. Skory, and J. E. Linz. 1996. Characterization of the function of the ver-I A and ver-I B genes involved in aflatoxin biosynthesis in Aspergillus parasiticus. Appl. Environ. Microbiol. 62:4568-4575. Liang, S.-H., T.-S. Wu, R. Lee, F. S. Chu, and J. E. Linz. 1997. Analysis of mechanisms regulating the expression of the ver-I gene, involved in aflatoxin biosynthesis. Appl. Environ. Microbiol. 63: 1058-1065. 285 Lillehoj, E. B. 1991. Aflatoxin: an ecologically elicited genetic/activation signal, p.2-30 In Smith, J. E., and R. S. Henderson (ed.), Mycotoxins and Animal foods. CRC Press, Boca Raton, FL. Lorang, J. M., R. P. Tuori, J. P. Martinez, T. L. Sawyer, R. S. Redman, J. A. Rollins, T. J. Wolpert, K. B. Johnson, R. J. Rodriguez, M. B. Dickman, and L. M. Ciuffetti. 2001. Green fluorescent protein is lighting up fungal biology. Appl. Environ. Microbiol. 67: 1987-1994. Luchese, R. H., and W. F. Harrigan. 1993. Biosynthesis of aflatoxin: the role of nutritional factors. J. Appl. Bacteriol. 74:5-14. Maggio-Hall, L. A., and N. P. Keller. 2004. Mitochondrial B-oxidation in Aspergillus nidulans. Mol. Microbiol. 54: 1 173-1 185. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1989. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Matsushima, K., Y. Ando, T. Hamasaki, and K. Yabe. 1994. Purification and characterization of two versiconal hemiacetal acetate reductases involved in aflatoxin biosynthesis. Appl. Environ. Microbiol. 60:2561-2567. Matsumura, F., and S. G. Knight. 1967. Toxicity and chemosterilizing activity of aflatoxin against insects. J. Econ. Entomol. 60:871-872. McCormick, S. P., D. Bhatnagar, and L. S. Lee. 1987. Averufanin is an aflatoxin B] precursor between averantin and averufin in the biosynthetic pathway. Appl. Eviron. Microbiol. 53:14-16. McLean, M., and M. F. Dutton. 1995. Cellular interactions and metabolism of aflatoxin: an update. Pharmac. Ther. 65:163-192. Meyers, D. M., G. Obrian, W. L. Du, D. Bhatnagar, and G. A. Payne. 1998. Characterization of aflJ, a gene required for conversion of pathway intermediates to aflatoxin. Appl. Environ. Microbiol. 64:3713-3717. 286 Miller, B. L., K. Y. Miller, K. A. Roberti, and W. E. Timberlake. 1987. Position- dependent and -independent mechanisms regulate cell-specific expression of the SpaCI gene cluster of Aspergillus nidulans. Mol. Cell. Biol. 7 :427-434. Miller, M. J. 2003. Transcriptional regulation of the Aspergillus parasiticus aflatoxin biosynthetic pathway gene nor-1. Ph.D. Dissertation, Michgan State Unviersity. Minto, R. E., and C. A. Townsend. 1997. Enzymology and molecular biology of aflatoxin biosynthesis. Chem. Rev. 97:2537-2555. Moore-Landecker, E. 1990. F undarnentals of the firngi. Prentice Hall, Englewood Cliffs, New Jersey. Mori, H., J. Kitamura, S. Sugie, K. Kawai, and T. Hamasaki. 1985. Genotoxicity of fungal metabolites related to aflatoxin B] biosynthesis. Mutat. Res. 143:121- 125. Morise, J. G., O. Shimomura, F. H. Johnson, and J. Winant. 1974. Intermolecular energy transfer in the bioluminescent system of Aequorea. Biochemistry 13:2656-2662. Nakamura K., and K. Matsuoka. 1993. Protein targeting to the vacuole in plant cells. Plant Physiol. 101:1-5. Neuhaus, J.-M., and J. C. Rogers. 1998. Sorting of proteins to vacuoles in plant cells. Plant Mol. Biol. 38:127-144. Ngindu, A., P. R. Kenya, D. M. Ocheng, T. N. Omondi, W. Ngure, D. Gatei, B. K. Johnson, J. A. Ngira, H. N andwa, A. J. Jansen, J. N. Kaviti, and T. A. Siongok. 1982. Outbreak of acute hepatitis caused by aflatoxin poisoning in Kenya. Lancet 1: 1 346-1348. Niehaus, W. G., and R. P. Dilts. 1982. Biosynthesis of the aflatoxin precursor sterigmatocystin by Aspergillus parasiticus. J. Bacteriol. 151:243-250. 287 Niehaus, W. G., and R. P. Dilts. 1984. Purification and characterization of glucose-6- phosphate dehydrogenase from Aspergillus parasiticus. Arch. Biochem. Biophys. 228:113-119. Niehaus, W. G. Jr., and W. Jiang. 1989. Nitrate induces enzymes of the mannitol cycle and suppresses versicolorin synthesis in Aspergillus parasiticus. Mycopathologia 107: 131-137. Northolt, M. D., H. P. Van Egmond, and W. E. Paulsch. 1977. Differences between Aspergillus/laws strains in growth and aflatoxin B] production in relation to water activity and temperature. J. Food Prot. 40:778. Oakley, B. R., J. E. Rinehart, B. L. Mitchell, C. E. Oakley, C. Carmona, G. L. Gray, and G. S. May. 1987. Cloning, mapping and molecular analysis of the pyrG (orotidine-S’phosphate decarboxylase) gene of Aspergillus nidulans. Gene 61:385-3 99. Oda, M. N., S. V. Scott, A. Hefner-Gravink, A. D. Caffarelli, and D. J. Klionsky. 1996. Identification of a cytoplasm to vacuole targeting determinant in aminopeptidase I. J. Cell Biol. 132:999-1010. Ohneda, M., M. Arioka, H. Nakajima, and K. Kitamoto. 2002. Visualization of vacuoles in Aspergillus oryzae by expression of CPY-EGF P. Fungal Genet. Biol. 37 :29-38. Park, D. L., and B. Liang. 1993. Perspectives on aflatoxin control for human food and animal feed. Trends Food Sci. Technol. 4:334-342. Payne, G. A., G. J. Nystrom, D. Bhatnagar, T. E. Cleveland, and C. P. Woloshuk. 1993. Cloning of the afl-2 gene involved in aflatoxin biosynthesis from Aspergillusflavus. Appl. Environ. Microbiol. 59:156-162. Perez-Rodriguez, E., J. Aguilera, and F. L. Figueroa. 2003. Tissular localization of coumarins in the green alga Dasycladus vermicularis (Scopoli) Krasser: a photoprotective role. J. Exp. Bot. 54:1093-1100. Pestka, J. J. 1988. Enhanced surveillance of foodbome mycotoxins by 288 immunochemical assay. J. Assoc. Off. Anal. Chem. 71:1075-1081. Pestka, J. J., and G. S. Bondy. 1990. Alteration of immune function following dietary mycotoxin exposure. Can. J. Physiol. Pharmac. 68: 1009-1016. Prasher, D. C., V. K. Eckenrode, W. W. Ward, F. G. Prendergast, and M. J. Cormier. 1992. Primary structure of the Aequorea victoria green-fluorescent protein. Gene 111:229-233. Prendergast, F. G., and K. G. Mann. 1978. Chemical and physical properties of aequorin and the green-fluorescent protein isolated from Aequoreaforskalea. Biochemistry. 17:3448-3453. Price, M. S., J. Yu, W. C. Nierman, H. S. Kim, B. Pritchard, C. A. Jacobus, D. Bhatnagar, T. E. Cleveland, and G. A. Payne. 2006. The aflatoxin pathway regulator AflR induces gene transcription inside and outside of the aflatoxin biosynthetic cluster. FEMS Microbiol. Lett. 255:275-279. Prosser, J. I. 1994. Molecular marker systems for detection of genetically engineered micro-organisms in the environment. Microbiology 140:5-7. Puisieux, A., S. Lim, J. Groopman, and M. Ozturk. 1991. Selective targeting of p53 gene mutational hotspots in human cancers by etiologically defined carcinogenes. Cancer Res. 51:6185-6189. Rarick, M., R. ijmas, C. D. Skory, and J. E. Linz. 1994. Identification and analysis of the expression of the aflatoxin biosynthesis genes nor-I and ver-I in the commercial species Aspergillus sojae and A. aryzae as well as toxigenic and nontoxigenic strains of A. flavus. In Proceedings, Current issues in food safety. Toxicology Center, Michigan State University, MI. Reddy, J. K., D. J. Svoboda, and M. S. Rao. 1976. Induction of liver tumors by aflatoxin B] in the tree shrew (T upaia glis), a nonhuman primate. Cancer Res. 36:151-160. Reynolds, G. N., and J. J. Pestka. 1991. Enzyme-linked immunosorbent assay of versicolorin A and related aflatoxin biosynthetic precursors. J. Food Prot. 289 54:105-108. Richard, J. L., A. C. Pier, S. J. Cysewski, and C. K. Graham. 1973. Effect of aflatoxin and aspergillosis on turkey poults. Avian Dis. 17:111-121. Roberts, 1. N., R. P. Oliver, P. J. Punt, and CA. M. J. J. van den Hondel. 1989. Expression of Escherichia coli B-glucuronidase gene in industrial and phytopathogenic filamentous fungi. Curr. Genet. 15:177-180. Rothman, J. H., C. T. Yamashiro, P. M. Kane, and T. H. Stevens. 1989. Protein targeting to the yeast vacuole. Trends Biochem. Sci. 14:347-3 50. Roze, L. V., R. M. Beaudry, N. P. Keller, and J. E. Linz. 2004. Regulation of aflatoxin synthesis by F adA/cAMP/protein kinase A signaling in Aspergillus parasiticus. Mycopathologia 158:219-232. Sakuno, E., Y. Wen, H. Hatabayashi, H. Arai, C. Aoki, K. Yabe, and H. Nakajima. 2005. Aspergillus parasiticus cyclase catalyzes two dehydration steps in aflatoxin biosynthesis. Appl. Environ. Microbiol. 71:2999-3006. Sakuno, E., K. Yabe, and H. Nakajima. 2003. Involvement of two cytosolic enzymes and a novel intermediate, 5’-oxoaverantin, in the pathway from 5’- hydroxy- averantin to averufin in aflatoxin biosynthesis. Appl. Environ. Microbiol. 69:6418-6426. Santerre Henriksen, A. L., S. Even, C. Miiller, P. J. Punt, C. A. M. J. J. van den Hondel, and J. Nielsen. 1999. Study of the glucoamylase promoter in Aspergillus niger using green fluorescent protein. Microbiology. 145:729-734. Schaewen, A. V., and M. J. Chrispeels. 1993. Identification of vacuolar sorting information in phytohemagglutinin, an unprocessed vacuolar protein. J. Exp. Bot. 44(suppl.):339-342. Scott, G. E., and N. Zummo. 1988. Sources of resistance in maize to kernel infection by Aspergillus/lavas in the field. Crop Sci. 28:504-507. 290 Shane, S. M. 1994. Economic issues associated with aflatoxins, p. 513-527. In D. L. Eaton and J. D. Groopman (ed.), The toxicology of aflatoxins: human health, veterinary and agricultural significance. Academic Press, New York. Shimomura, O. 1979. Structure of the chromophore of Aequorea green fluorescent protein. FEBS Lett. 104:220-222. Shoji, J.-Y., M. Arioka, and K. Kitamoto. 2006. Vacuolar membrane dynamics in the filamentous fungus Aspergillus aryzae. Eukaryot. Cell 5:411-421. Shroeder, H. W., R. J. Cole, R. D. Grigsby, and H. Hein, Jr. 1974. Inhibition of aflatoxin production and tentative identification of an aflatoxin intermediate 'versiconal acetate' from treatment with dichlorvus. Appl. Microbiol. 27:394- 399. Silva, J. C., R. E. Minto, C. E. Barry, K. A. Holland, and C. A. Townsend. 1996. Isolation and characterization of the versicolorin B synthase gene from Aspergillus parasiticus: expansion of the aflatoxin B1 biosynthetic gene cluster. J. Biol. Chem. 271:13600-13608. Silva, J. C., and C. A. Townsend. 1996. Heterologous espression, isolation, and characterization of versicolorin B synthase from Aspergillus parasiticus. J. Biol. Chem. 272:804-813. Sinnhuber, R. O., J. H. Wales, J. L. Ayres, R. H. Engebrecht, and D. L. Amend. 1968. Dietary factors and hepatoma in rainbow trout (Salmo gairdner). Aflatoxins in vegetable protein feedstuffs. J. Natl. Cancer Inst. 41:711-718. Skory, C. D., P.-K. Chang, J. W. Cary, and J. E. Linz. 1992. Isolation and characterization of a gene from Aspergillus parasiticus associated with conversion of versicolorin A to sterigmatocystin in aflatoxin biosynthesis. Appl. Environ. Microbiol. 58:3527-3537. Skory, C. D., P.-K. Chang, and J. E. Linz. 1993. Regulated expression of the nor-1 and ver-I genes associated with aflatoxin biosynthesis. Appl. Environ. Microbiol. 59: 1642-1646. 291 Skory, C. D., J. S. Horng, J. J. Pestka, and J. E. Linz. 1990. Transformation of Aspergillus parasiticus with a homologous gene (pyrG) involved in pyrimidine biosynthesis. Appl. Environ. Microbiol. 56:3315-3320. Stark, A. 1980. Mutagenicity and carcinogenicity of mycotoxins: DNA binding as a possible mode of action. Annu. Rev. Microbiol. 34:235-262. Stauber, R. H., K. Hone, P. Carner, E. A. Hudson, N. I. Tarasova, G. A. Gaitanaris, and G. N. Pavlakis. 1998. Development and applications of enhanced green fluorescent protein mutants. Biotechniques 24:462-466. Suelmann, R., and R. Fischer. 2000. Mitochondrial movement and morphology depend on an intact actin cytoskeleton in Aspergillus nidulans. Cell Motil. Cytoskelet. 45:42-50. Suelmann, R., N. Sievers, and R. Fischer. 1997. Nuclear traffic in fungal hyphae: in vivo study of nuclear migration and positioning in Aspergillus nidulans. Mol. Microbiol. 25:757-769. Tague, B. W., C. D. Dickinson, and M. J. Chrispeels. 1990. A short domain of the plant vacuolar protein phytohemagglutinin targtets invertase to the yeast vacuole. Plant Cell 2:533-546. Takahashi, T., P.-K. Chang, K. Matsushima, J. Yu, K. Abe, D. Bhatnagar, T. E. Cleveland, and Y. Koyama. 2002. Nonfunctionality of Aspergillus sojae aflR in a strain of Aspergillus parasiticus with a disrupted aflR gene. Appl. Environ. | Microbiol. 68:3737-3 743. 1 Tan, P. K., J. P. Howard, and G. S. Payne. 1996. The sequence NPFXD defines a new class of endocytosis signal in Saccharomyces cerevisiae. J. Cell Biol. 135: 1789-1800. Tavoularis, S., C. Scazzocchio, and V. Sophianopoulou. 2001. Functional expression and cellular localization of a green fluorescent protein-tagged proline transporter in Aspergillus nidulans. Fungal Genet. Biol. 33:115-125. 292 Tilak, T. B. 1975. Induction of cholangiocarcinoma following treatment of Rhesus monkey with aflatoxin. Food Cosmet Toxicol. 13:247-249. Timberlake, W. E., and E. C. Barnard. 1981. Organization of a gene cluster expressed specifically in the asexual spores of Aspergillus. nidulans. Cell 26:29- 37. Tiwari, R. P., V. Mittal, T. C. Bhalla, S. S. Saini, G. Singh, and D. V. Vadehra. 1 1986. Effect of metal ions on aflatoxin production by Aspergillus parasiticus. 1 F olia Microbiol. 31: 124-128. Trail, F., P.-K. Chang, J. Cary, and J. E. Linz. 1994. Structural and functional analysis of the nor-1 gene involved in the biosynthesis of aflatoxins by Aspergillus parasiticus. Appl. Environ. Microbiol. 60:4078-4085. Trail, F., N. Mahanti, and J. E. Linz. 1995a. Molecular biology of aflatoxin biosynthesis. Microbiology 141:755-765. Trail, F., N. Mahanti, M. Rarick, R. Mehigh, S. H. Liang, R. Zhou, and J. E. Linz. 1995b. Physical and transcriptional map of an aflatoxin gene cluster in Aspergillus parasiticus and functional disruption of a gene involved early in the aflatoxin pathway. Appl. Environ. Microbiol. 61:2665-2673. Tsien, R. Y. 1998. The green fluorescent protein. Annu. Rev. Biochem. 67 :509-544. Valdez-Taubas, J., G. Diallias, C. Scazzcchio, and A. L. Rosa. 2000. Protein expression and subcellular localization of the general purine transporter UapC from Aspergillus nidulans. Fungal Genet. Biol. 30:105-113. Valls, L. A., J. R. Winther, and T. H. Stevens. 1990. Yeast carboxypeptidase Y vacuolar targeting signal is defined by four propeptide amino acids. J. Cell Biol. 111:361-368. Verrett, M. J., J. Winbush, E. F. Reynaldo, and W. F. Scott. 1973. Collaborative study of the chicken embryo bioassay for aflatoxin B], Ass. Offic. Anal. Chem. J. 56:901-904. 293 Vida, T. A., and S. D. Emr. 1995. A new vital stain for visualizing vacuolar membrane dynamics and endocytosis in yeast. J. Cell Biol. 128:779-792. Vitale, A., and N. V. Raikhel. 1999. What do proteins need to reach different vacuoles. Trends Plant Sci. 4:149-155. Ward, W. W., and S. H. Bokman. 1982. Reversible denaturation of Aequorea green- fluorescent protein: physical separation and characterization of the renatured protein. Biochemistry 21:4535-4550. Ward, W. W., C. W. Cody, D. C. Prasher, and F. G. Prendergast. 1989. Sequence of the chemical structure of the hexapeptide chromophore of Aequorea green- fluorescent protein. Photochem. Photobiol. 49:62S Ward, W. W., and M. J. Cormier. 1976. In vitro energy transfer protein in Renilla bioluminescence. J. Phys. Chem. 80:2289-2291. Ward, W. W., and M. J. Cormier. 1979. An energy transfer protein in coelenterate bioluminescence. J. Biol. Chem. 254:781-788. Ward, W. W., H. J. Prentice, A. F. Roth, C. W. Cody, and S. C. Reeves. 1982. Spectral perturbations of the Aequorea green-fluorescenct protein. Photochem. Photobiol. 35:803-808. Wink, M. 1993. The plant vacuole: a multifunctional compartment. J. Exp. Bot. 44(suppl.):23 1-246. Wogan, G. N. 1992. Aflatoxins as risk factors for hepatocellular carcinoma in humans. Cancer Res. 52:2114S-2118S. - Woloshuk, C. P., K. R. Foutz, J. F. Brewer, D. Bhatnagar, T. E. Cleveland, and G. A. Payne. 1994. Molecular characterization of aflR, a regulatory locus for aflatoxin biosynthesis. Appl. Environ. Microbiol. 60:2408-2414. Woloshuk, C., E. Seip, G. Payne, and C. Adkins. 1989. Gene transforamation system for the aflatoxin-producing fungus Aspergillus/lavas. Appl. Environ. 294 F: I_, _ " .'___.r ”.4"; Microbiol. 55:86-90. Yabe, K., Y. Ando, and T. Hamasaki. 1988. Biosynthesis relationship among aflatoxin B], Bz, G], G2. Appl. Environ. Microbiol. 54:2101-2106. Yabe, K., Y. Ando, J. Hashimoto, and T. Hamasaki. 1989. Two distinct 0- methyltransferases in aflatoxin biosynthesis. Appl. Environ. Microbiol. 55:2172- 2177. Yabe, K., and T. Hamasaki. 1993. Stereochemistry during aflatoxin biosynthesis: cyclase reaction in the conversion of versiconal to versicolorin B and racemization of versiconal hemiacetal acetate. Appl. Environ. Microbiol. 59:2486-2492. Yabe, K., Y. Matsuyama, Y. Ando, H. Nakajima, and T. Hamasaki. 1993. Stereochemistry during aflatoxin biosynthesis: conversion of norsolorinic acid to averufin. Appl. Environ. Microbiol. 59:2486-2492. Yabe, K., M. Nakamura, and T. Hamasaki. 1999. Enzymatic formation of G-group aflatoxins and biosynthetic relationship between G- and B-group aflatoxins. Appl. Environ. Microbiol. 65:3867-3872. Yang, T.-T., L. Cheng, and S. R. Kain. 1996. Optimized codon usage and chromophore mutations provide enhanced sensitivity with the green fluorescent protein. Nucleic Acids Res. 24:4592-4593. Yu, J., D. Bhatnagar, and T.E. Cleveland. 2004. Completed sequence of aflatoxin pathway gene cluster in Aspergillus parasiticus. FEBS Lett. 564:126-130. Yu, J., J. W. Cary, D. Bhatnagar, T. E. Cleveland, N. P. Keller, and F. S. Chu. 1993. Cloning and characterization of a cDNA from Aspergillus parasiticus encoding an O-methyltransferase involved in aflatoxin biosynthesis. Appl. Environ. Microbiol. 59:3564-3571. Yu, J., P.-K. Chang, D. Bhatnagar, and T. E. Cleveland. 2000. Cloning of a sugar utilization gene cluster in Aspergillus parasiticus. Biochim. Biophys. Acta 295 1493:211-214. Yu, J., P.-K. Chang, D. Bhatnagar, and T. E. Cleveland. 2003. Cloning and functional expression of an esterase gene in Aspergillus parasiticus. Mycopathologia 156:227-234. Yu, J., P.-K. Chang, K. C. Ehrlich, J. W. Cary, D. Bhatnagar, T. E. Cleveland, G. A. Payne, J. E. Linz, C. P. Woloshuk, and J. W. Bennett. 2004. Clustered pathway genes in aflatoxin biosynthesis. Appl. Environ. Microbiol. 70: 1253- 1262. Yu, J., P.-K. Chang, M. Wright, D. Bhatnagar, T. E. Cleveland, G. A. Payne, and J. ’ E. Linz. 1995. Comparative mapping of aflatoxin pathway gene clusters in i Aspergillus parasiticus and Aspergillusflavus. Appl. Environ. Microbiol. 61:2365-2371. Zhou, R. 1997. The function, expression, and localization of the Nor-1 protein involved in aflatoxin B] biosynthesis: the function of fluP, a gene associated with sporulation in Aspergillus parasiticus. Ph.D. Dissertation, Michgan State Unviersity. Zhou, R., and J. E. Linz. 1999. Enzymatic function of the Nor-l protein in aflatoxin biosynthesis in Aspergillus parasiticus. Appl. Environ. Microbiol. 65:5639-5641. 296 1|111111111111111111111111111111111111 3 1293 02956 6506