Citation
Molecular characterization of autophagy in methylotrophic yeast Pichia pastoris

Material Information

Title:
Molecular characterization of autophagy in methylotrophic yeast Pichia pastoris
Alternate title:
Molecular characterization of autophagy in methylotrophic yeasts Pichia pastoris
Creator:
Yuan, Weiping, 1964-
Publication Date:
Language:
English
Physical Description:
xii, 151 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Complementation ( jstor )
DNA ( jstor )
Enzymes ( jstor )
Ethanol ( jstor )
Glycolysis ( jstor )
Peroxisomes ( jstor )
Pichia ( jstor )
Proteins ( jstor )
Vacuoles ( jstor )
Yeasts ( jstor )
Autophagocytosis ( mesh )
Department of Anatomy and Cell Biology thesis Ph.D ( mesh )
Dissertations, Academic -- College of Medicine -- Department of Anatomy and Cell Biology -- UF ( mesh )
Microbodies ( mesh )
Phosphofructokinase-1 ( mesh )
Pichia ( mesh )
Research ( mesh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1998.
Bibliography:
Bibliography: leaves 141-150.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Weiping Yuan.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
029484030 ( ALEPH )
51173778 ( OCLC )

Downloads

This item has the following downloads:


Full Text











MOLECULAR CHARACTERIZATION OF AUTOPHAGY IN
METHYLOTROPHIC YEAST PICHIA PASTORIS












By


WEIPING YUAN


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE
UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY





UNIVERSITY OF FLORIDA


1998














ACKNOWLEDGMENTS


I would like to thank my mentor, Dr. William A. Dunn, Jr., for his

excellent guidance and training in my 5-year graduate study. I would also like to extend my gratitude to my committee members, Drs. John P. Aris, Gudrun S. Bennett and Alfred S. Lewin, for their support, help and encouragement of my research. I am in great debt to Dr. Dan Tuttle for his pioneer work in the lab and Dr. Jim Cregg for providing yeast strains, plasmids and the helpful conversations. I would also like to thank the faculty, staff and all other persons at the Department of Anatomy and Cell Biology for the help they rendered and for the stimulating intellectual environment they provided.

My special thanks are reserved for my wife Ying Shen for her wholehearted support in these five years.














TABLE OF CONTENTS


ACKNOWLEDGMENTS.................................. ................ ii

LIST OF TABLES................... . .... .. ..... ................... vi

LIST O F FIG URES .................................................................................. vii

KEY TO ABBREVIATIONS ...................................................................... x

A BST RA CT ............................................................................................ xi

CHAPTERS

1 INTRODUCTION AND REVIEW OF THE LITERATURE................ 1

The Peroxisome and its Characteristics...................................... 2
Peroxisome Biogenesis ...................................................... 2
Peroxisom e Functions ...................................................... 3
Methanol Metabolism in Pichia Pastoris............................ 5
Autophagy ................................................................................... 7
Non-selective Autophagy in Mammalian Cells and Yeast..... 8 Selective Autophagy in Mammalian Cells and Yeast ............ 11
Genetic Analysis of Pichia pastoris................................................ 15
Pichia pastoris as a Study Model for Autophagy.................. 15
Classical Genetics ............................................................ 16
M olecular Biology ............................................................ 16
Previous Research of Peroxisome Degradation in Our Lab ........... 19
Chapter Summary ........................................................................ 26

2 MATERIALS AND METHODS ..................................................... 27

Yeast, Bacterial Strains and Media ............................................. 27
Enzym e Assays ........................................................................... 29
Isolation of Glucose-Induced Selective Autophagy-Deficient
(gsa) M utants ......................................................... 31
Mutant Generation and Isolation....................................... 31









M utant Backcrossing............................................................ 32
Mutant Complementation Analysis..................................... 33
M olecular B iology......................................................................... 34
Yeast Transformation........................................................ 34
Plasmid Isolation and DNA Sequencing.............................. 35
Northern, Southern and Western Blot Analysis.................. 37
Isolation of PFK1 and GSA7 Knockouts.............................. 38
Site-directed Mutagenesis of PFK1 ...................................... 39
Generation of a HA Epitope Tag in GSA7 and YHR171w
ge ne ....................................................................... 4 0
Fluorescence Microscopy and Electron Microscopy.................... 41

3 CHARACTERIZATION OF PICHIA PASTORIS MUTANTS
DEFECTIVE IN GLUCOSE-INDUCED SELECTIVE
A UTO PHA G Y ................................................................... 43

Introd uctio n .................................................................................... 4 3
Complementation Group Identification.......................................... 45
Screening for gsa M utants............................................................. 48
Glucose Induced Microautophagy is Defective in gsa Mutants....... 49 Morphological Studies of gsa Mutants........................................... 57
C hapter Sum m ary ......................................................................... 64

4 GSA1 PROTEIN IS PPF1 PROTEIN........................................... 68

Introduction ................................................................................... 68
Gsal-1 Signaling an Upstream Event of Microautophagy.............. 68
Microautophagy of Peroxisomes is Defective in gsal-1....... 68
Morphological Study Revealed That a Step Before
Peroxisome Sequestration Is Blocked in gsal-1........ 70
pDLT1 Complemented gsal-1 Phenotype................................. 73
Rescue Study of Different Fragments in the Insert of pDLT1 ......... 74 Sequencing and Sequence Analysis of pDLT1 Insert.................... 75
Northern Blot Analysis of gsal-1 Mutant...................................... 79
Phosphofructokinase Activity Assay in gsa I ................................. 80
Glycolysis Pathway and gsa Mutants...................................... 87
Glycolysis and gsa Mutants .......................................................... 91
Chapter Summary ......................................................................... 93

5 PFK1 PROTEIN IS REQUIRED FOR THE INITIATION OF
PEROXISOME MICROAUTOPHAGY.................................. 96

Introduction ............................................................................... 96









Degradation of Peroxisomes by Microautophagy Requires
P F K 1 .......................................................................... 9 7
PFK1 Gene Disruption and the Verification......................... 97
The PFK1 Knockout Has the Same Phenotype as gsal -1.... 97
Verification by Site Directed Mutagenesis of P. pastoris PFK1
Gene of the Distinction between Rescue Function
and PFKp A ctivity............................................... 100
Normal Pfklp and Catalytically-inactive pfklp
Com plement Apfkl................................................... 100
Restoration of the Degradation Ability of Apfkl by PFK1 ......104
Morphological Characterization of Apfkl..................................... 106
C hapter Sum m ary........................................................................ 109

6 CHARACTERIZATION OF GSA7 MUTANTS........................... 111

Introd uction ............................... .............................................. 1 1 1
Morphological Studies of gsa7................................................... 111
Recovery of GSA7 Gene and its Verification............................. 112
pYWP7-4 Rescues gsa7 phenotype.............................. 112
Identification of G SA 7........................................................ 114
Sequence Analysis of GSA7............................................. 116
Agsa7 Generation and its Phenotype Studies.................... 119
A HA Tagged Gsa7p Rescues P. pastoris................................. 120
Chapter Summary .................................................................... 122

7 CONCLUSIONS AND PROSPECTS........................................ 127

Introduction ................................................................................ 127
Characterization of gsa Mutants............................................. 131
Regulation of the Signaling of Microautophagy by Gsal p......... 132
Regulation of a Homotypic Vacuolar Membrane Fusion
Event of Microautophagy by Gsa7p.................................. 135
Prospects and Conclusion........................................................ 137

R E FE R E N C ES ................................................ ............................ 14 1

BIOGRAPHICAL SKETCH .......................................................... 151














LIST OF TABLES


Tables Page

2-1 Parental and mutant strains of Pichia pastoris............................. 28

3-1 Biochemical profiles of gsa mutants............................................ 53














LIST OF FIGURES


Figure Pae

1-1 Compartmentalization of the pathways involved in methanol
metabolism in methylotrophic yeast ...................................... 6

1-2 Loss of peroxisomal and cytosolic enzymes during ethanol and
glucose adaptation .............................................................. 20

1-3 Morphology characterization of peroxisomes during glucose and
ethanol adaptation................................................................... 21

1-4 Model of glucose-induced microautophagy pathway............................ 24

2-1 Maps of plasmids pYM8 and pYM4..................................................... 36

3-1 A flow chart of mutants characterization .......................................... 44

3-2 Identification of gsa mutants complementation groups......................... 47

3-3 AOX and FDH activities of gsa mutants under glucose adaptation....... 51 3-4 AOX and FDH activities of gsa mutants under glucose adaptation....... 52 3-5 AOX degradation and histidine addition during glucose adaptation... 56 3-6 Ultrastructural studies of gsal-1, gsa2, gsa3, gsa6, gsa7 and gsa8
mutants under glucose adaptation.................................... 62

3-7 Ultrastructural studies of gsa4 and gsa5 mutants under glucose
adaptation.................. . ..................... 63

3-8 Microautophagy of peroxisomes and gsa mutants in P. pastoris........ 65 4-1 Glucose and ethanol adaptation in parental GS115 and
W DY2 (gsal-1)......... .... .... .......................................... 69









4-2 Morphology of gsal-1 and gsa2 under glucose adaptation at
Oh and 3h tim e points................................................................... 72

4-3 pDLT1 rescues gsal-1 phenotype........................... ................... 76

4-4 Identification of genomic DNA that rescues gsal-1 during glucose
adaptation ................................................................................ 77

4-5 Northern blot analyses of GS115, WDY1 (gsal 1) and WDY2 (gsal 1-1)..... 81 4-6 PFK activity is greatly reduced in WDY2 (gsal 1-1) cells.......................... 82

4-7 Nucleotide and predicted amino acid sequence of P. pastoris PFK1
g e n e ................................................ ....................................... 8 5

4-8 Comparison of PFK1 genes ................................................................... 86

4-9 Glycolysis pathway and the entry point of carbon metabolites................ 89

4-10 Metabolites in the glycolysis pathway and their induction of the
degradation of AOX .................................................................. 90

4-11 Glycolysis enzyme activities and gsa mutants.................................... 92

5-1 Generation of the PFK1 knockout ......................................................... 98

5-2 Glucose-induced degradation of AOX and FDH in PPF1 and WDKO1..99 5-3 Partial sequences comparison of Pfkl1p in P. pastoris (PP),
S. cerevisiae (SC) and Human sapiens (HU)............................... 100

5-4 Site-directed mutagenesis of P. pastoris PFK1 gene............................... 102

5-5 Morphology of GS115, gsal 1-1, and Apfkl during glucose adaptation ...... 107 5-6 Morphology of WDKO1 (Apfkl) and its transformants during
glucose adaptation .................................................................... 108

6-1 Morphology of gsa3 and gsa7 mutants during glucose adaptation.......... 113

6-2 Verification of GSA 7................... ................................................ 115

6-3 Complete nucleotide and amino acid sequence of GSA7...................... 117

VIII









6-4 A protein sequence comparison of GSA7 in P. pastoris, S. cerevisiae,
S. pom be and H. sapiens.......................................................... 118

6-5 Agsa7 generation and its verification.................................................... 121

6-6 Detection of a HA epitope tagged Gsa7p and Yhrl71wp in
transform ed gsa7 ..................................................................... 123

7-1 Pathways of peroxisome induction and degradation under glucose and
ethanol adaptation in methylotrophic yeast Pichia pastoris..... 129 7-2 Microautophagy and macroautophagy of peroxisomes
in P. pastoris ............................................................................. 130

7-3 Putative substrate binding and regulatory sites in Pfkl p...................... 134

7-4 Pathway of glucose induced peroxisome microautophagy and its
relationship with gsa mutants......................... ................... 139













KEY TO ABBREVIATIONS


AOX: alcohol oxidase BSA: bovine serum albumin CCO: cytochrome c oxidase DSM: diploid selection medium EAM: ethanol adaptation medium Fl3: mitochondrial F1 ATPase, P subunit FBP: fructose-1,6- bisphosphatase FDH: formate dehydrogenase GAM: glucose adaptation medium gsa: glucose-induced selective autophagy deficient mutant Gsap: Gsa protein GSA: GSA gene Agsal: gsal knockout strain Agsa7: gsa7 knockout strain mRNA: message RNA MIM: methanol induction medium ORF: open reading frame PFKI: phosphofructokinase 1 gene Pfklp: phosphofructokinase 1 protein pfkl: mutated Pfkl protein pfkl: mutated PFK1 gene SM: sporulation medium TM: transformation medium YND: yeast nitrogen base medium with glucose YNM: yeast nitrogen base medium with methanol YPD: glucose-containing complete medium













Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

MOLECULAR CHARACTERIZATION OF AUTOPHAGY IN
METHYLOTROPHIC YEASTS PICHIA PASTORIS

By

Weiping Yuan

December 1998

Chairman: William A. Dunn, Jr.
Major Department: Anatomy and Cell Biology

Eukaryotic cells adapt to environmental changes by synthesizing and degrading cellular proteins and organelles. Although much is known regarding the biogenesis of cell organelles, the mechanism of turnover of many organelles and proteins through autophagy remains unclear. Our lab has utilized the yeast P. pastoris as a study model for the autophagy of peroxisomes during metabolic adaptation. Peroxisomes are induced when P. pastoris is grown on methanol. During the adaptation to glucose, peroxisomes are degraded within the yeast vacuole by microautophagy. This process includes glucose signaling, vacuolar recognition and sequestration of peroxisomes, homotypic vacuolar membrane fusion and finally vacuolar degradation of peroxisomes.








In order to better understand the microautophagy process, glucoseinduced selective autophagy-deficient (gsa) mutants were generated and isolated by their inability to degrade peroxisomal alcohol oxidase (AOX) during glucose adaptation. Eight gsa complementation groups (gsal to gsa8) have been identified and they represent different gene products that control microautophagy processes. These gsa mutants fall into four major groups. Gsal belongs to microautophagy initiation mutants. Gsa4 and gsa8 are possible autophagy recognition mutants. Gsa2, gsa3, gsa6, and gsa7 are possible homotypic vacuolar membrane fusion mutants. Gsa5 is also a homotypic fusion mutant but its macroautophagy pathway has been turned on during the methanol induction.

Gsal-1 was transformed with a Pichia pastoris genomic DNA library. The gene that complemented the gsal-1 phenotype was identified as phosphofructokinase 1 (PFK1). Cellular levels of both PFK1 mRNA and PFKp activity were greatly reduced in gsal-1. The inability of Apfkl to degrade AOX could be rescued by either normal PFK1 or mutant pfkl whose catalytic site had been inactivated by a single amino acid mutation. It suggests that the degradation of peroxisomes does not require a catalytically active PFKp. Pfkl p might regulate microautophagy independent of its PFK activity.

WDY7 (gsa7) was the second mutant I chose for further genetic study. I recovered a plasmid pYWP7-4 that complemented gsa7. The gene was identified as GSA7. A knockout strain Agsa7 failed to degrade AOX confirmed

xii









GSA7 is essential for the microautophagy. Gsa7p might participate in a ubiquitin-like pathway to regulate the vacuole membrane fusion.

The study of microautophagy in Pichia pastoris provides us a unique

opportunity to look into the mechanisms of protein degradation via autophagy. The study I have done and the tools that have become available in recent years will be useful to further elucidate the molecular mechanism of microautophagy in eukaryotic cells.














CHAPTER 1
INTRODUCTION AND REVIEW OF THE LITERATURE



Some yeast species such as Candida boidinii, Hansenula polymorpha, and Pichia pastoris are able to utilize methanol as the sole carbon and energy source for their growth and thus are called methylotrophic yeasts. These methylotrophic yeasts metabolize the methanol by enzymes in the peroxisomes and cytoplasm whose synthesis is induced under methanol-growth conditions (Hill et al., 1985). These enzymes as well as the whole peroxisomes are rapidly degraded upon a carbon source switch to such as glucose or ethanol. The study of the biogenesis and degradation processes of peroxisomes under different growth conditions in Pichia pastoris and other yeast species has gained great interest in recent years because in humans, an inability to import proteins into the peroxisomes has serious health consequences (Moser and Moser, 1996). Genetic screens for peroxisomal assembly (pas) (Gould et al., 1992), deficiency (per) (van der Klei et al., 1991), and degradation (gsa, ppd) (Tuttle and Dunn, 1995, Titorenko et al., 1995) mutants in yeast have been successfully utilized to identify many genes required for these events. Moreover, several genes cloned in these mutants have been found to have human homologous counterparts (Dodt and Gould, 1996). These findings emphasize the usefulness of the yeast









2

mutant study model and the significance of these studies. In the following sections, I will discuss the existing knowledge relate to peroxisome biogenesis, degradation and autophagy. I will also discuss the methods available to study the mechanism of autophagy in yeast peroxisomes and the previous work that has been done in our lab.



The Peroxisome and its Characteristics



Peroxisome Bioqenesis

Peroxisomes are ubiquitous, single-membrane-bound organelles that are involved in many important cellular activities. They were first discovered thirty years ago in a centrifugation fraction of cell components (de Duve, 1965). The peroxisomes are thought to arise by budding or fission from preexisting peroxisomes although evidence for de novo synthesis has been presented. Peroxisomes contain neither DNA nor ribosomes. Their proteins are encoded by nuclear genes and synthesized in cytoplasm on free polyribosomes and then post-translationally imported into the preexisting organellar matrix or surrounding membrane (Liu et al., 1995). Almost all peroxisomal proteins are synthesized at their final size. Two peroxisomal targeting sequences (PTS) responsible for correct delivery of matrix proteins to the organelles have been identified. The majority of peroxisomal matrix proteins contain PTS1, a C-terminal tripeptide of the sequence SKL-COOH (Keller et al., 1991). PTS2, with the consensus










sequence RLX5H/QL is located at N-terminal of peroxisomal proteins (Swinkels et al., 1991). Many studies focused on the biogenesis of peroxisomes and the defects in peroxisome assembly and their protein import. In P. pastoris, 15 complementation groups of peroxisome assembly mutants (PAS mutant) have been identified (Subramani, 1993). Some of the PAS genes have been sequenced. PAS4 gene was identified as a ubiquitin-conjugating enzyme required for peroxisome assembly (Crane et al., 1994). It is an interesting finding since the ubiquitin mediated protein degradation pathway is responsible for protein turnover. We do not yet know if PAS4 has a role in the peroxisome protein degradation pathway through targeted degradation by the proteasome. McCollum et al. (1993) reported cloning the PAS8 gene and demonstrated that it is a PTS1 import factor. Subsequent studies showed that it is a PTS1 import receptor and the tetratricopeptide repeat domain of PAS8p is identified as the PTS1 binding region (Terlecky et al., 1995). This finding linked peroxisome protein import and targeting and the peroxisome assembly pathway in the peroxisome biogenesis process.

Peroxisome Functions

Two prominent functions of peroxisomes are the 13-oxidation of fatty acids and the oxidation of substrates by different H202-generating oxidases (Subramani, 1993). Other functions of peroxisomes include lipid biosynthesis, cholesterol biosynthesis and peroxisomal purine metabolism. Peroxisomes are unique in their functional diversity when compared with other cell organelles











such as mitochondria, the nucleus, and organelles of the secretory pathway. Their functions are dispensable in certain nutritional environments and the specific metabolic pathways found in the organelle vary depending upon the organism, the tissue and its environment. This was verified by observations that mutants defective in peroxisome assembly in yeast and human fibroblast cell lines were still viable (Subramani, 1993). Nevertheless, the peroxisome is essential for normal human development, since defects in peroxisome assembly have been identified as the cause of the peroxisome biogenesis diseases.

In methylotrophic yeast P. pastoris, when cells were grown on nutrition rich YPD media, yeast maintained a low number of peroxisomes per cell. The size of the peroxisomes is also small. However, peroxisomes can also be made to proliferate or degrade in response to nutritional cues such as methanol or oleic acid (Liu et al., 1992, Subramani, 1993). The process of peroxisome degradation is not as well understood as biogenesis. Cells respond to nutritional clues to reduce the number of peroxisomes via degradation. Such events are exemplified by supplying methanol as growth carbon source to induce peroxisome production and then abolishing the induction by a carbon source switch such as glucose or ethanol. The whole process usually takes less than eight hours to finish (Tuttle and Dunn, 1995). This degradation event provides us a unique opportunity to investigate the mechanisms of peroxisome degradation.










Methanol Metabolism in Pichia pastoris

When methanol is added to the P. pastoris growth media as the only

carbon source, methanol readily enters the cell and peroxisome for metabolism. The peroxisomal enzyme alcohol oxidase (AOX) is the first enzyme in the methanol assimilation pathway. AOX is a homooctameric protein of ~ 74 kD, each subunit containing a flavin adenine dinucleotide molecule (FAD) as a prosthetic group (van der Klei et al., 1991). It uses 02 as the electron acceptor, oxidizes methanol and produces hydrogen peroxide and formaldehyde (Gleeson and Sudbery, 1988). The hydrogen peroxide is reduced by peroxisomal catalase to H20 and 02. Formaldehyde exits the peroxisome and is catalyzed in cytosol by formaldehyde dehydrogenase and formate dehydrogenase (FDH) to CO2 providing an energy source in the form of NADH2 (Tolbert, 1981). Formaldehyde enters the xylulose 5-phosphate cycle to provide a carbon source for amino acid synthesis (Fig. 1-1).

One characteristic of AOX is that it is not detectable in glucose-grown

cells but inducible under methanol growth condition. Another is that it has a low affinity for its substrate and thus large quantities of the enzyme (up to 30% of the total cellular soluble proteins) are needed for yeast to grow on methanol as the sole carbon source. When the carbon source for growth is switched from methanol to glucose or ethanol (glucose or ethanol adaptation), yeast peroxisomes and thereby also AOX are degraded rapidly in its vacuole via autophagy (Tuttle et al., 1993). The inactivation of AOX activity is paralleled by


















CH3OH CH3OH GSH
, ,--------GSH 1 ~D
H202 :- D
I � HCHO - -HCHO GS-CH2OH NAD HCOOH C02
1 ND 1 NAD "1,
0+ H20 NADH NADH
Xu5P
5 rearrangement
reactions
GAP DHA
1 // 1 cell
; GAP - constituents
DHA DHAP \c
ADP F3P F6P
GAP I





Figure 1-1. Compartmentalization of the pathways involved in methanol
metabolism in methylotrophic yeast. Methanol is metabolized first in
peroxisome and then in the cytosol for further degradation. Its metabolism
pathway is shown in the figure above and the enzyme names are listed
below. The pathway is responsible for the generation of GAP for the
production of biomass and the NADH2 for energy generation. 1 = Alcohol
oxidase; 2 = catalase; 3 = formaldehyde dehydrogenase; 4 = formate
dehydrogenase; 5 = dihydroxyacetone synthase; 6 = dihydroxyacetone
kinase; 7 = fructose 1,6-bisphosphate aldolase; 8 = fructose 1,6bisphosphate phosphatase (Adapted from van der Klei et al., 1991)










its degradation of peroxisomes in this process (Veenhuis et al., 1983) and the disappearance of AOX activity correlates with the degradation of AOX protein and autophagic loss of peroxisomes. The degradation of peroxisome is proteinase A and B dependent (Tuttle and Dunn, 1995). These characteristics of AOX provide us an excellent biochemical index for studying peroxisome degradation under different nutritional adaptation conditions. It also provides us an easy enzyme assay to screen autophagy mutants.



AutoDhagv



Cells respond to changes in the environment by eliminating cell

components that are no longer needed. The amino acids then are released and used as an energy source or for the synthesis of new proteins. This provides the organism with added adaptability that outweighs the metabolic costs such as ATP used in protein degradation pathways (Olson et al., 1992). The continual turnover of cellular proteins also prevents the accumulation of a variety of deleterious nonenzymatic modifications such as oxidation, deamination, glycosylation of the abnormal proteins that are deleterious to the cells (Klionsky, 1997). The degradation of protein in the cell is mainly via lysosomal proteolytic pathways, proteolytic pathways of non-lysosomal organelles and cytosolic proteolytic pathways. In the following section, I will mainly discuss the existing knowledge of autophagy pathways in mammalian cells and yeast cells.










Non-Selective Autophaqy in Mammalian Cells and Yeast

Autophagy means self-eating. It is used to refer to the process whereby cytoplasmic constituents are degraded in a vacuole-dependent manner. It has been recognized as a highly regulated non-selective and sometimes selective process for the degradation of cellular proteins and organelles in eukaryotic cells. Mammalian cells respond to amino acid starvation and other environmental changes (e.g., heat shock) by activating the degradation of proteins via non-selective autophagy (Lardeux and Mortimore, 1987, Mortimore et al., 1989, Kopitz et al., 1990, Dunn et al., 1994). Both microautophagy and macroautophagy have been observed in mammalian cells. Microautophagy is the process whereby cellular components are sequestered directly into the lysosome. Regions of the cytoplasm are surrounded by invagination of the lysosomal membrane or by finger-like protrusions of the lysosome. Upon fusion of the lysosomal membrane, intralysosomal vesicles are formed containing the sequestered cellular components (Ahlberg and Glaumann, 1985, Mortimore et al., 1989). It is believed that multivesicular bodies arise from such autophagic events. Macroautophagy is the process whereby cellular components are sequestered first within an autophagosome (Dunn, 1990a). The autophagosome originates from the rough endoplasmic reticulum in mammalian cells and, then fuses with a lysosome. Once within the lysosome, the proteins are degraded to their monomeric subunits by the concerted action of exopeptidases and endopeptidases (Dunn, 1990a, 1990b).











Many strains of yeast are capable of activating autophagy in response to nutritional changes (Veenhuis et al., 1983, Tuttle et al., 1993, Baba et al., 1994, Tuttle and Dunn, 1995, Chiang et al., 1996). The main organelle for autophagy in yeast is its vacuole, the equivalent of lysosome in mammalian cells. In order to better define the molecular events of autophagy, mutants defective in autophagy have been used to study the mechanism of autophagy in yeast S. cerevisiae, P. pastoris and H. polymorpha. During nitrogen starvation, cells of a normal strain of S. cerevisiae respond to this stress by degrading cellular contents non-specifically via autophagy. Ohsumi's group found in a S. cerevisiae mutant lacking proteinase B, "autophagic bodies" containing cytosolic contents were accumulated in the vacuoles during nitrogen starvation. (Takeshige et al., 1992). They (Tsukada and Ohsumi, 1993) used this strain to generated apg (autophagy) mutants which failed to accumulate autophagic bodies within the yeast vacuole. Strains with apg mutation failed to deliver celluar contents to the yeast vacuole and are thus the non-selective autophagy mutants. Their ultrastructural data suggested that the autophagic process in yeast is essentially similar to that of the lysosomal system in mammalian cells (Baba et al., 1994). The apg mutants fell into 15 complementation groups. The characterization of these mutants is underway and several novel genes have been identified, sequenced and their functions in the process of autophagy are being studied (Shirahama et al., 1997; Matsuura et al., 1997; Noda and Ohsumi, 1998).










The aut (autophagy) mutants identified by a colony screening procedure by Thumm et al. (1994) are defective in the degradation of cytosolic fatty acid synthase during carbon starvation. This served as a selection marker for rapid isolation of autophagocytosis yeast mutants defective in vacuolar breakdown of fatty acid synthase during starvation. Further identification and characterization of the genes of apg and aut mutants will reveal the mechanisms and function of autophagocytosis. It will also reveal why all autophagy mutants have defects in their sporulation process. It appears that some of these autophagy mutations are allelic (Harding et al., 1996, Scott et al., 1996). However, it is not known whether or not these mutations affect macroautophagy induced by readministering glucose in S. cerevisiae. It will be interesting to see if there is an overlap in the degradation pathway in non-selective and selective autophagy pathways and the diverging point of these two pathways.

Several labs including ours have observed similar autophagy processes in methylotrophic yeast during the protein and cell organelle degradation process. Veenhuis et al (1978, 1981) observed in methanol-induced H. polymorpha cells, when the carbon source was switched to ethanol, the earliest change was the appearance of a variable number (2 to 12) layers of electrondense membranes surrounding peroxisomes. These membranes surrounded individual peroxisomes within a cluster and appeared to sequester a given peroxisome into an autophagosome and then deliver its contents to the yeast vacuole. The process is analogous to macroautophagy in mammalian cells.










Tuttle and Dunn (1995) found that when growth carbon source was switched from methanol to ethanol, the process of peroxisome autophagy in P. pastoris includes an intermediate stage in which individual peroxisomes were sequestered into autophagosomes by wrapping membranes and the autophagosomes then fused with the vacuole, a process similar to macroautophagy. In the glucose-induced peroxisome degradation pathway, he observed that autophagy began with the engulfment of clusters of peroxisomes by finger-like protrusions of the vacuole and the contents were degraded, a process analogous to microautophagy in mammalian cells. Selective Autophaqy in Mammalian Cells and Yeast

Although non-selective autophagy is considered to be responsible for the bulk turnover of proteins that occurs in response to nutritional starvation or developmental stress (Takeshige et al., 1992), several studies suggested that selective autophagy also plays an important role in regulation of protein and organelle levels in mammalian and yeast cells. Zellweger syndrome is a wellknown prototypic peroxisomal disorder in the newborn (Lazarow and Fujiki, 1985). Normal peroxisomes appeared to be absent in cells of these patients and peroxisomal enzymes remained in the cytosol due to impaired assembly of peroxisomes (Shimozawa et al., 1992). However, the fibroblasts from these patients were propagated readily in tissue culture. Meijer's group found that in Zellweger fibroblasts, most peroxisomal ghosts in fact contained lysosomal hydrolases (Heikoop et al., 1992). The treatment with the autophagy inhibitor 3-










methyladenine caused an increase in the number of peroxisome ghosts though peroxisomal functions were impaired. It suggested that peroxisomal ghosts of Zellweger cells were selectively degraded by autophagy. In another study (Luiken et al., 1992) in, they found that peroxisomal enzymes fatty acyl-CoA oxidase and catalase were preferentially degraded in the isolated hepatocytes of clofibrate-fed rats. This increased degradation of the peroxisomal enzymes was prevented by 3-methyladenine, an inhibitor of macroautophagic sequestration. Long-chain fatty acid could inhibit this degradation too. They concluded that preferential autophagy of peroxisomes exists.

In yeasts, independent studies showed evidence that selective autophagy plays an important role to adjust their cellular composition according to environment changes. Veenhuis and coworkers studied selective inactivation of AOX in two peroxisome-deficient (per) mutants in Hansenula (Veenhuis et al., 1983, van der Klei et al., 1991, Titorenko et al., 1995). They found that these two per mutants synthesized AOX but most of AOX remained in cytosol. Only a small percentage of AOX was degraded under the glucose adaptation condition. These data indicated that degradative inactivation of AOX in H. polymorpha is strictly dependent on the localization of the enzyme inside peroxisomes and the mechanism triggering this process is not directed against AOX protein, but instead, to the membrane surrounding the organelle. In another study, they also observed that not all peroxisomes were degraded in H. polymorpha during the carbon source switch (Veenhuis et al., 1983). One or few small peroxisomes









13

escaped degradation and subsequently served as the target organelle for newly synthesized matrix proteins in the new growth environment. In their recent study (Titorenko et al., 1995), they have generated mutants impaired in the selective degradation of peroxisomes in H. polymorpha (peroxisome degradationdeficient, pdd). Of the seven mutants they studied, two complementation groups have been identified with two possible responsible genes PDD1 and PDD2. The function of their products in selective autophagy of AOX-containing peroxisomes is under study. The PDD1 p is involved in the initial signaling events for sequestration of the organelle since no multilayer membrane sequestration of peroxisomes was observed in pddl when compared with wild type cells under ETOH adaptation. In pdd2 mutants, sequestration did occur since the multilayer membrane appeared and wrapped peroxisomes but subsequent fusion of vacuole membrane did not occur. Immunocytochemical detection of alcohol oxidase protein in the vacuole was not observed in any cases. They concluded that PDD2p is essential for mediating the second step in selective peroxisome degradation, namely, fusion and subsequent uptake of sequestered organelles into the vacuole. The model they established for the study of peroxisome maintenance and proliferation, sorting, folding, assembly of matrix proteins, synthesis and function of the peroxisome membrane, and peroxisome turnover in H. polymorpha (van der Klei and Veenhuis, 1996) is similar to what we have established in P. pastoris (Tuttle and Dunn, 1995).










In S. cerevisiae, carbon and nitrogen starvation stimulates non-selective macroautophagy (Takeshige et al., 1992, Egner et al., 1993, Baba et al., 1994). However, data for selective degradation by microautophagy have been reported when yeast cells was switched from oleic acid medium to the one containing glucose (Chiang and Schekman, 1991). When cells were replenished with glucose, a key gluconeogenic enzyme, fructose- 1,6-bisphosphatase (FBPase) was selectively targeted from the cytosol to the yeast lysosome (vacuole) for degradation (Chiang et al., 1996). Peroxisomes have been identified as the target organelles to be delivered to the vacuole for degradation when cells were replenished with glucose. They have generated vacuolar import and degradation-deficient (vid) mutants and these mutants were placed into 20 complementation groups. FBPase degradation was blocked in all these mutants (Hoffman and Chiang, 1996). They identified a novel type of vesicles in the cytosol specific to vacuolar protein degradation pathway. These vesicles are intermediate in the FBPase degradation pathway which is different to all established vesicles (Huang and Chiang, 1997). One novel responsible gene VID24 has been recovered (Chiang, 1997) Vip24p is a 41kD protein and is induced by glucose addition and localized to the intermediate vesicles as a peripheral protein. In the absence of Vid24p, FBPase accumulates in the intermediate vesicles. It seems that this protein plays a critical role to deliver FBPase from the intermediate vesicles for vacuolar degradation. The study clearly showed that the selective protein degradation is present in S. cerevisiae.










Our lab was the first to study selective autophagy in P. pastoris using microautophagy mutants. As stated in the beginning of this chapter, we have identified two distinct peroxisome degradation pathways namely, microautophagy and macroautophagy. We have shown that the autophagy of peroxisome needs synthesis of new molecules.



Genetic Analysis of Pichia pastoris



Pichia pastoris as a Study Model for Autophagqy

Yeast is a much simpler system to study autophagy processes than mammalian cells systems. Yeast cells are easy to grow and handle. Their biochemical and morphological characterization can be done quickly. They are one cell eukaryotic organisms and have a much smaller genome and thus fewer genes involved in the autophagy process. When comparing the cells of P. pastoris with S. cerevisiae in the peroxisome degradation process, under glucose induction, peroxisomes are larger, fewer in number, and much easier for biochemical and morphological studies in P. pastoris than S. cerevisiae. The peroxisome degradation process is also much faster in P. pastoris. Eight hour glucose adaptation results in complete peroxisome degradation while in S. cerevisiae, the degradation of peroxisomes under glucose induction needs at least 24 hrs. The measurement of degradation is difficult in S. cerevisiae because the growth of S. cerevisiae under glucose induction results in a dilution










of peroxisomes per cell in a 24 hour period (Chiang et al., 1996; Tuttle, unpublished data).

Classical Genetics

Pichia pastoris has many advantages in genetic studies. It is

ascosporogenic and exists in one of two mating types. The life cycle of Pichia pastoris is characterized by defined haploid and diploid stages (Gleeson and Sudbery, 1988, Cregg et al., 1990) and it can be maintained indefinitely as vegetative haploids. However, upon nutritional limitation, particularly for nitrogen, mating occurs and diploid cells are formed. The mating type can switch between two opposite strains under poor nutritional conditions at high frequency. These key features allow the isolation and phenotypic characterization of mutants in the chemically mutagenized P. pastoris. It also allows us to characterize the phenotypic identification and the complementation groups of mutants (Cregg et al., 1990). Dr. Cregg's group (Liu et al., 1992) developed an efficient screening method for peroxisome-deficient mutants in P. pastoris. The screen relies on the unusual ability of P. pastoris to grow on two carbon sources, methanol and oleic acid, both of which absolutely require peroxisomes to be degraded. A collection of 280 methanol utilization-defective (mut) mutants were isolated, and organized into 46 complementation groups. Molecular Bioloqy

P. pastoris has also been developed to be an efficient transformation system for the introduction of replicating plasmids by using E. coli-P. pastoris










shuttle vectors. The system is based on a histidinol dehydrogenase-defective mutant host of P. pastoris and a modified version of the spheroplast fusion-gene transfer procedure (Cregg et al., 1985). The His4 gene of P. pastoris and the autonomous replication sequences (ARS) were isolated and used to construct the plasmid that has about 10s5/pg of transformation frequencies and can be maintained as extrachromosomal elements. For example, the plasmid pYM8 is widely used as a shuttle vector in gene complementation studies. P. pastoris is amenable for transformation by both the electroporation and spheroplast generation method (Cregg et al., 1985, 1993). A number of genes have been cloned and sequenced in P. pastoris due to the availability of a set of auxotrophic strains with essentially wild type genetic background, the plasmids that act as shuttle vectors and the genomic DNA libraries for the isolation of P. pastoris genes by functional complementation of mutants. The genes cloned by this method encode a varied set of proteins such as putative ATPase, peroxisomal integral membrane proteins, proteins related to ubiquitinconjugating enzymes and proteins that might be constituents of the peroxisomal protein import machinery (Nuttley et al., 1995).

Peroxisome assembly factor-1 (PAF1) is a well known protein

responsible for the human Zellweger syndrome (Goldfischer, 1996). Dr. Cregg's group has cloned this gene's homolog in H. polymorpha (Waterham et al., 1996a, 1996b). Future studies of the function of this gene in yeast will be interesting. It will also be more interesting if we can test if human PAF-1 could










rescue H. polymorpha mutant or vice versa and thus to probe the functional conservation in eukayotes. Gould and his coworkers (Dodt et al., 1996) developed a method to isolate human peroxisome biogenesis disorder genes by computer-based homology probing of the dBEST database. They found PXR1 is a human orthologue of the P. pastoris PAS8 gene and PXAAA1 (Yahraus et al., 1996) as a homologue of P. pastoris PAS5 gene. The same group of Dr. Gould (Kalish et al., 1996) also developed a visual screening method using a peroxisomal form of the green fluorescent protein (GFP). Wild-type cells expressing PTS1-GFP were chemically mutagenized and the mutagenized strains unable to import PTS1-GFP into peroxisomes were identified by fluorescence microscopy. This technique provides an effective visual marker for peroxisomal protein import in living cells. The PTS-GFP containing wild-type cells can be used to regenerate peroxisome mutants.

The available molecular technologies also made P. pastoris cells as a host system for the large scale heterologous protein production for industrial usage (Sreekrishna et al., 1997) as well as to attain a certain scale amount of protein for functional study of such proteins. Most of the experiments utilized AOX1 gene promoter to induce gene expression in P. pastoris after AOX1 gene promoter containing plasmid was introduced into P. pastoris. The induction of desired protein synthesis is initiated by methanol addition. The same is true when gene repression is induced by addition of glucose. Dr. Cregg also developed a P. pastoris vector containing a promoter of the GAPDH gene










(Waterham et al., 1997). This promoter can induce expression of the gene constructed behind it. This adds convenience for over-expressing the desired protein for observation or other uses. Direct PCR screening of P. pastoris clones after yeast transformation to verify the right transformant was also developed recently so that less work is needed to verify the expression of the desired gene in P. pastoris.



Previous Research of Peroxisome Degqradation in Our Lab



In our lab, we used methanol induction and subsequent glucose or ethanol adaptation to characterize peroxisomal and cytosolic protein degradation pathways in P. pastoris (Tuttle et al., 1993, Tuttle and Dunn, 1995). Parental strain GS115 cells under glucose and ethanol adaptation showed different protein degradation approaches biochemically (Fig. 1-2). The activities of peroxisomal alcohol oxidase (AOX) and cytosolic formate dehydrogenase (FDH) were only about 20% of their zero time activity after a 6 hour glucose adaptation. A 6-hour time course of monitoring AOX and FDH activity and the protein level of AOX and FDH in glucose adaptation showed good correlation and verified that the loss of activity was mainly due to loss of AOX and FDH protein (Tuttle and Dunn, 1995). The activities of cytochrome c oxidase (COO), a mitochondria protein and activities fructose-1,6-bisphosphatase (FBP), a cytosolic protein remains high after 6 hour glucose adaptation (data not shown).


















Loss of AOX and FDH during Glucose
and Ethanol Adapations


120


0


o,- 0 I-


SAOX wlglucose --..
-A- FDH wlglucose .
-0 AOX wlethanol
--- FDH wlethanol

0 1 2 3 4 5 6 7

Time (hrs)


Figure 1-2. Loss of peroxisomal and cytosolic enzymes during ethanol and glucose adaptation. P. pastoris (GS1 15) were cultured in methanol induction medium until stationary, at which time ethanol or glucose were added to begin adaptation (time zero), cell-free extracts were prepared at 0- and 6-hour time point and alcohol oxidase (AOX) and formate dehydrogenase (FDH) activities were assayed. Values are represented as a percentage of the 0-hour activity. All assays were measured at least three times and the values shown on the graph represent the average � sd.


A
















P




MP
V


Nu
A


Nu







B V




V
V







C F


Figure 1-3. Morphological characterization of peroxisomes during glucose and ethanol adaptation. P. pastoris cells (GS115) were induced in methanol and then switch to either ethanol (A-C) or glucose (D-F) for one hour. Under glucose adaptation, peroxisomes were engulfed by the yeast vacuole and contents get degraded. However, under ethanol adaptation, peroxisomes were first wrapped by several layers of membrane to form autophagosome, then fused with the yeast vacuole. Bar in panel A inlet: 25nm, Bar in other panels: 0.5pM. (use of this figure is permitted by Dr. D. Tuttle)









22

Also in the glucose-induced pathway, the degradation process requires protein synthesis since protein synthesis inhibitor cycloheximide prevented peroxisome degradation in GS115 cells. Under ethanol adaptation, AOX activities reduced significantly as it did in glucose adaptation in a 6-hour time course while FDH activity remained unchanged. The degradation of peroxisomes in ethanolinduced pathway in P. pastoris is independent of protein synthesis since cycloheximide did not prevent ethanol-induced peroxisome degradation. In the same study (Tuttle and Dunn, 1995), they found proteinases A and B were required for the degradation of peroxisomes and FDH. Peroxisomes in PrA and PrB mutants were induced by methanol. However, during the subsequent adaptation to glucose or ethanol, the mutants showed no significant degradation of AOX and FDH proteins as well as peroxisomes. The combined results suggested that the microautophagy of peroxisomal AOX and cytosolic protein FDH is selective under glucose adaptation condition and regulatory molecules are needed to modulate this process.

The previous observation was also supported by the morphology studies in light microscopy and electron microscopy. Immunofluorescence microscopy studies were used to screen or verify that the degradation process was disrupted in the mutants (see below). Using quinacrine that labels specifically the yeast vacuole, we can follow the generation and degradation of peroxisomes in parental strains and mutants when we combine the observation results of fluorescence microscopy of the yeast vacuole and the phase-contrast








23

microscope image of yeast peroxisomes. Immunofluorescence microscopy can follow specifically the alcohol oxidase generation and degradation in peroxisomes as well as other proteins that could detected by the antibodies. However, to characterize the specific detailed structures and degradation process of peroxisomes, electron microscopy is needed. As shown in the Fig.13, at 1 hour of the glucose adaptation in parental GS115 cells, peroxisomes without additional membrane layers were seen in close association with the yeast vacuole (panel D). The extensions or arms of the vacuole were observed surrounding clusters of peroxisomes (panel E) and the vacuolar membrane protrusions fused to engulf the peroxisomes, a homotypic event of fusion. Peroxisomes were seen being degraded inside the vacuole (panel F). However, after 1 hour of ethanol adaptation, peroxisomes were seen first to be wrapped by several layers of membrane (panel A) to form autophagosomes. The autophagosomes then fused with the yeast vacuole (panel B) and the contents in peroxisome were degraded (panel C). We do not know the origin of these layers of membrane. It is possible that they are synthesized de novo. The autophagy pathway under ethanol adaptation is analogous to the macroautophagy that has been characterized in mammalian system (see previous discussion in autophagy section).

Through these observations, we proposed that the microautophagy

mechanism of peroxisomes proceeds via a sequence of events: environmental signaling, peroxisomal recognition, peroxisomal sequestration & homotypic












GLUCOSE SIGNALING .

FDH
Peroxlsome q M.



FDH



SEQUESTRATION AND FUSION






PEROXISOME
DEGRADATION


RECOGNITION




H~Vacuole


Figure 1-4. Model of glucose-induced microautophagy pathway. The yeast responds to glucose adaptation by degrading peroxisomal alcohol oxidase (AOX) and cytosolic formate dehydrogenase (FDH). Peroxisomes and FDH are recognized by and sequestered within the yeast vacuole. Once within the vacuole, these proteins are degraded by a mechanism dependent upon the actions of proteinases A (PrA) and B (PrB).










vacuole fusion, and vacuolar degradation (Fig. 1-4). We hypothesized that cytosolic, peroxisomal or vacuolar membrane regulatory proteins need to be synthesized during the peroxisome autophagy in yeast cells. The interaction and coordination of these regulatory proteins lead to the degradation of specific proteins or organelles of the cells (Fig. 1-4). To approach this question, Dr. Tuttle has generated glucose-induced selective autophagy (gsa) mutants deficient specifically in AOX degradation during the glucose adaptation. Mutants were screened by an colorimetric AOX direct colony assay for their inability to degrade AOX during glucose adaptation and verified by a liquid media AOX assay. We tried to identify molecules regulating the autophagy process as well as to elucidate the mechanisms of the autophagy process. Indeed, in Tuttle's study (Tuttle and Dunn, 1995), he has identified two complementation groups (gsa I and gsa2) that were defective in the peroxisome degradation pathway during glucose adaptation. The degradation of AOX and FDH in response to glucose was inhibited by 70-90% in both mutants while the degradation of AOX proceeded normally during ethanol adaptation. This suggested that the vacuolar proteinases activities were normal in these mutants and the mutation did not affect vacuolar function but inhibited an event upstream of vacuolar degradation. The ultrastructure studies confirmed that peroxisome entry to the vacuole was defective in these mutants. This also provides further evidence of divergence between these two degradative pathways.










Chapter Summary



The methylotrophic yeast P. pastoris readily grows in media containing

methanol as the sole carbon and energy source. During this growth stage, they synthesize the large amount of peroxisomal and cytosolic enzymes that are necessary for the utilization of methanol. Upon adaptation to an alternative carbon source such as glucose or ethanol, the peroxisomes are rapidly targeted to and degraded in the vacuole. The peroxisomes are large enough (0.5 to 1 um) to be identified under light and electron microscopy. The degradation process can also be easily detected by a simple colorimetric assay of AOX activities. Based on previous studies in our lab, we hypothesize that the degradation of peroxisomes during glucose adaptation requires regulatory molecules for the signaling, recognition, peroxisomal sequestration & homotypic vacuolar membrane fusion, and degradation steps in the autophagy of peroxisomes. To approach this problem, gsa mutants that are defective in AOX degradation during glucose adaptation were generated. I tried to define the molecular events of autophagy in P. pastoris by analyzing these mutants. The Pichia pastoris is a proven study model for peroxisome assembly, import and peroxisome degradation. The existing knowledge of peroxisome degradation and the available molecular tools combined with biochemical and ultrastructural methods enable us to examine the microautophagy of peroxisomes and to understand the mechanisms of this autophagy in yeast at the molecular level.













CHAPTER 2
MATERIALS AND METHODS



Yeast and Bacterial Strains and Media



All Pichia pastoris parental yeast strains were the generous gifts of Dr. J. M. Cregg (Oregon Graduate Institute, Beaverton, OR). The wild type strains, GS115 (his4), GS190 (arg4), WP1 (ade4), WP2 (met4) and PPF1 (his4, arg4), were routinely cultured at 300C in YPD (1% Bacto yeast extract, 2% Bacto peptone, and 2% dextrose). YNM and YND media refer to the growth media containing yeast nitrogen without amino acids plus methanol or glucose respectively. If histidine (40mg/L) is added, the media are called YNMH or YNDH. The sporulation and mating medium (SM media) is composed of 0.5% sodium acetate, 1.0% KCI, 1.0% glucose and 2.0% agar. The electroporation medium is composed of 1M Sorbitol, 2% glucose, 0.1% yeast nitrogen base (YNB), 0.4 mg/L biotin and 2% of agar. Bacto peptone, yeast extract and media agar were purchased from Fisher Co. LB broth media is composed of 0.5% yeast extract, 1% tryptone, 0.5% NaCI. Ampicillin is added to LB media at a final concentration of 100 pg/ml when selection is needed for E. coi growth. WDY (gsa) mutants were generated by Dr. Dan Tuttle by selecting














Table 2-1 P. pastoris Strains Available for Current Studies


Strain(Species) Complementation group Genotype

GS115 Parental strain his4 GS190 Parental strain arg4 WP1 Parental strain ade4 WP2 Parental strain met4 KB1 Parental strain ade4, his4 JC205 Parental strain met4, his4 PPF1 Parental strain his4, arg4 WDY1(12.1) gsa2-1 his4, gsa2-1 WDY2(4.3) gsal-1 his4, gsal-1 WDKO1 Apfkl pfkl::ARG4, his4 WDY3(13.1) gsal 1-2 his4, gsal-2 WDY4(18.1) gsa4 his4, gsa4 WDY5(6.3) gsa5 his4, gsa5 WDY6(4.1) gsa6 his4, gsa6 WDY7(13.2) gsa7 his4, gsa7 WDKO7 Agsa7 gsa7::ARG4, his4 WDY8(17.3) gsal 1-3 his4, gsal-3 WDY9(15.2) gsa? his4, gsa? WDY10(35.3) gsa8 his4, gsa8 WDY11(35.2) gsa? his4, gsa? WDY12(9.3) gsa? his4, gsa? WDY13(32.1) gsa3 his4, gsa3 WDY14(15.1) gsa? his4, gsa? WDY15(18.3) gsa? his4, gsa? WDY16(32.2) gsa? his4, gsa?








29

glucose-induced selective autophagy deficient strains in chemically mutagenized parental GS115. WDKO1 (Apfkl) is a PFK1 gene knockout strain derived from PPF1 by replacing its PFK1 gene with a S. cerevisiae ARG4 gene. WDKO7 (Agsa7) is a knockout strain of GSA7 from PPFI. Both Escherichia col DH5a and Epicurian coli� XL1-blue (Stratagene, Co.) were used to amplify plasmids. A list of yeast strains used in the study is included in the Table 2-1.



Enzyme Assays



Yeast cells were first cultured in the methanol induction medium. Glucose or ethanol was then added for adaptation. The yeast cells were harvested at stationary stage (zero hour time point of adaptation) and then six-hours after adaptation (six hour time point of adaptation). Two ml of yeast cells were pelleted at 40C, 2500 rpm. The cells were resuspended in 1 ml ice cold breaking buffer (20 mM Tris/CI, pH 7.5, 50 mM NaCI, 1mM EDTA). Cells were broken by vortexing 1 min for 3 times mixed with 500pm diameter glass beads (Tuttle et al., 1993). Phenylmethylsulfonyl fluoride (PMSF, 1 mM) was added to prevent proteolysis. After centrifugation at 4�C, 2500 rpm for 5 min, the supernatant was aspirated and stored at 4�C for subsequent enzyme assays.

Alcohol oxidase activity was measured in a reaction coupled with horseradish peroxidase and the oxidation of 2,2'-azino-bio(3-ethylbenzthiazoline-6-sulfonic acid), ABTS. The reaction mixture contained horseradish








30

peroxidase and ABTS in 33 mM potassium phosphate buffer (Tuttle et al., 1993). The samples were incubated at 370C until green color developed. The color was measured at 410 nm. The activity of formate dehydrogenase was determined by measuring the formation of NADH during the oxidation of formate by absorption at 340 nm (Kato, 1990). The reaction mix was made of 33 mM potassium phosphate buffer, pH 7.5, 2 mM NAD*, 167 mM sodium formate per reaction plus cell free sample. The rate of change of absorbance at 340nm was followed and used as a measure of enzyme activity.

Phosphofructokinase activity and other glycolysis enzyme activities were measured in a coupled enzyme assay resulting in the oxidation of NADH/ NADPH or reduction of NAD/NADP. The NADP and NADH cocktail solution for enzyme activity assays were composed of 98 mL 50mM triethanolamine, 10mM MgCI2 pH7.4, 2 mL NADP 14mg/mL or 0.75 mL NADH (30 mg/mL in 0.1 M Tris pH 7.6) respectively. Other substrates and enzymes were added according to the reaction requirement and finally extracts were added. The activities recorded were adjusted by the extract protein concentration. For example, hexokinase was assayed by a cocktail containing NADP, 0.1 M fructose, 1 mg/mL phososphoglucose-isomerase, 1 mg/mL of glucose-6-phosphate dehydrogenase, yeast cell extract, and the reaction was started with 0.1 M ATP. The production of NADPH was monitored by a kinetic program with an absorbance reading set at 340nm. Pyruvate kinase was assayed by a cocktail containing NADH, 0.1 M ADP, 0.1 M fructose-1,6-bisphosphate, lactate dehydrogenase, the cell free










extract and beginning the reaction by addition of 0.1 M Phosphoenolpyruvate. The production of NAD was measured by a kinetic program at OD 340nm (Dyson et al., 1975). The glycolysis enzymes tested were GK: glucokinase; HK: hexokinase; PGI: phosphoglucose isomerase; PFK: phosphofructokinase; ALD: aldolase; TIM: Triosphosphate isomerase; PGM: phosphoglycerate mutase; ENO: enolase; and PK: pyruvate kinase. Protein concentrations were measured using crystalline bovine serum albumin as a standard (Bradford, 1976).



Isolation of Glucose-Induced Selective Autophaqy-Deficient (qsa) Mutants



Mutant Generation and Isolation

WDY mutant strains were generated by mutagenizing parental GS115 cells with nitrosoguanidine at 100 pg/ml concentration (Tuttle and Dunn, 1995). The mutagenized cells were screened for the loss of the ability to degrade alcohol oxidase in response to a shift in a carbon source from methanol to glucose with a direct colony assay (Tomlison, 1992, Tuttle and Dunn, 1995). Briefly, the mutagenized cells grown on YPD plates were replica-plated to YNMH plates (6.7 g/L yeast nitrogen base without amino acids, 0.5% methanol,

0.4 mg/L biotin, and 40 pg/ml histidine) and the colonies were allowed to grow for 4-5 days. The colonies were replica-plated onto nitrocellulose and placed on YNDH plates (6.7 g/L yeast nitrogen base without amino acids, 2% glucose, 0.4 mg/L biotin, and 40 pg/ml histidine ) for 12-16 hours. Those putative mutant










colonies which retained AOX despite glucose adaptation were identified by the purple color reaction product of the direct colony assay (Tuttle and Dunn, 1995). The principle of reaction is the same as the alcohol oxidase assay. Those mutant strains appeared purple were isolated and their inability to degrade AOX during glucose adaptation verified in liquid cultures. This was done by first growing the cells for 24-36 hours in 20 mls of methanol induction medium that consisted of 6.7 g/L yeast nitrogen base without amino acids (Difco), 0.4 mg/L biotin, 40 pg/ml L-histidine, and 0.5% methanol. Glucose was then added to a final concentration of 2% at the beginning of glucose adaptation. Cells were harvested at Oh and 6h of glucose adaptation and lysed by vortexing with acidwashed glass beads. The resulting homogenates were then assayed for AOX activity according to the procedures of Tuttle and Dunn (1995). Mutant Backcrossing

To further characterize the mutant strains, gsa mutants that have been identified were backcrossed to GS190 (arg4) and GS115 (his4) for generating gsa specific mutants as well as for identifying complementation groups. The haploid mutants were sequentially mated to essentially wild type strains (GS190 and GS115) containing complementary auxotrophic markers and haploid progeny recovered. Several rounds of backcrossings resulted in a particular mutant strain with a mutated gene of our interest. One round of backcrossing was accomplished as follows: the mutant strains were streaked onto YPD plates in patches and the parental strains of the complementary auxotrophy was spread










on YPD plates to form a lawn of cells and both plates were incubated at 300C overnight. The lawn and the patch were both replica plated to a single sporulation media (SM) plate and incubated overnight at 300C to induce mating. These plates were then replicated to YND plates without amino acids and incubated for 2 to 3 days until diploid colonies appeared. Colonies from these plates were streaked onto a fresh YND plates and incubated for 1 to 2 days and then were streaked on YPD plate. Cells from this plate were then streaked to SM plates and incubated four days at 300C to induce meiosis and sporulation. Cell spores were harvested with an inoculation loop and the remaining diploid cells were killed by etherization. An aliquot from this mixture was diluted 1 to 100 and 100 pl plated on YPD plates for two days for the haploid cells to grow. The cells were then washed off and plated on YNMH or YNMA plate for a direct colony assay for mutants screen. The verified mutant via this assay and the subsequent AOX liquid media assay was selected and used for the next round of backcrossing with a parental strain that has the opposite auxotroph. Mutant Complementation Analysis

To determine the number of defective genes present in the gsa mutants in my possession, the complementation groups of gsa mutants had to be identified first. To achieve this, different auxotroph strains containing mutants (for example, a histidine auxotroph and an arginine auxotroph mutant) are mated and a prototroph (diploid) of them is generated. The ability of this diploid to degrade peroxisomes can be tested in a glucose adaptation assay. If the









34

peroxisome degradation can proceed in this diploid, these two strains are said to complement each other. Complementation indicates the defective genes in these two mutants are different and the mutants belong to different complementation groups. If the diploid still could not degrade peroxisomes, the two mutant strains presumably contain different mutant alleles of the same gene and they belong to the same complementation group.



Molecular Biology



Yeast Transformation

The genomic library was a gift of Dr. J.M. Cregg (Oregon Graduate

Institute). It was constructed by a partial digestion of genomic DNA with Sau 3A I first. The digestion products were then run on a DNA gel and the DNA of 5-10 kb size was collected, purified and ligated into BamH I site of plasmid pYM8. Pichia pastoris cells (his4) were grown in YPD to an OD600nm = 1.0 harvested, and transformed by electroporation at 1.5 kV, 25 pF, 400 ) (Gene Pulser, BioRad Corporation) with 5 to 10 pg of genomic DNA library. The transformed cells were grown for 4-5 days on electroporation plates, and the colonies were replica-plated to nitrocellulose and placed on YND-aa plates for 12-16 hours. Those colonies which appeared purple or white (according to experiment purposes) upon a direct colony assay were isolated and their abilities to degrade AOX during glucose adaptation verified in liquid media as described above.










Stable transformation with vectors pYM8 and pYM4 containing specific fragments of genomic DNA was also done by electroporation. pYM8 was derived from pBR322 and contains the 164 bp autonomous replicating sequence from P. pastoris and the HIS4 gene of S. cerevisiae. pYM4 was also derived from pBR322 and contains a HIS4 gene of P. pastoris, but lacking an autonomous replicating sequence. pYM25 was derived from pBR322 and contains an ARG4 gene of S. cerevisiae. Both pYM4 and pYM25 are nonepisomal vectors. Stable integration of the episomal pYM8 constructs was accomplished by two cycles of growth under nonselection and selection conditions. Stable integration of non-episomal pYM4 or pYM25 constructs was promoted by a single cut with restriction enzyme Stu I within the HIS4 locus. This cut should also direct the integration into His4 locus in the transformed cell genome. The plasmid maps of pYM8 and pYM4 is shown in Fig. 2-1. Plasmid Isolation and DNA Sequencing

The rescued colonies that had been complemented with the genomic DNA library were grown overnight in 2 ml of YPD medium. The cells were pelleted and resuspended into 0.2 ml of 2% Triton X-100, 100 mM NaCI, 1 mM EDTA, 1% SDS, and 10 mM Tris/HCI, pH 8.0 and 0.2 ml of phenol:chloroform: isoamyl alcohol (25:24:1). The cells were then disrupted by vortexing in the presence of 0.5 ml acid-washed glass beads. After centrifugation, 2 HI of the upper phase containing the plasmid was used to transform E. coli DH5a cells (Ausubel et al., 1988). The plasmid was then isolated by Wizard Plus minipreps








































NhcI (1178)


Sal (1615)

pYM4 PHIS4 NdeI (494 7013bp


Pvull (47
BamHI (2670)
Pvull (2850)
EagI (35 Sphl (3218)




Figure 2-1. Maps of plasmids pYM8 and pYM4. pYM8 is constructed based on pBR322 with a S. cerevisiae HIS4 gene and an autonomous replication sequence (ARS). pYM4 is also based on pBR322 but lacking an ARS. pYM4 contains a P. pastoris HIS4 gene. The main restriction enzyme sites are also indicated in the two plasmid maps.








37

of Promega and restriction analysis was done to identify the genomic insert size and sites. The genomic DNA fragment was excised from the pYM8 shuttle vector and engineered into pBluescript/KS for sequencing by nested deletion (Ausubel et al., 1988). Direct sequencing of pDLT1 was also done using self generated primers that flank the desired region. The primers for sequencing were synthesized by DNA synthesis lab, University of Florida. All sequencing was performed by University of Florida Sequencing Core via the dideoxy chain termination method. The insert of pDLT1 was sequenced from both plus/minus strands. The DNA sequences were assembled in DNAman� or Gene RunnerO programs and all six reading frames were compared to protein sequence databases using the Blastx program of the National Center for Biotechnology Information (NCBI).

Northern, Southern and Western Blot Analysis

Yeast cells in exponential growth were pelleted and resuspended in 1% SDS, 10 mM EDTA, and 50 mM sodium acetate, pH 5 solution. Total RNA was extracted with addition of hot phenol/chloroform/ isoamyl alcohol (25:24:1). About 10 pg of total RNA was loaded on glyoxal agarose gels. The RNA was then transferred to Maximum Strength Nytran" (Schlelcher & Schuell Inc.) by overnight blotting (Ausubel et al., 1988). The RNA in the blot was cross-linked by UV for 12 seconds and then subjected to methylene blue staining for 2 min. The staining pattern of ribosomal RNA was photographed or scanned as loading control. A 2.5 kb DNA fragment in PFK1 gene was used as template for 32p_










labeling by random priming reaction with NEB's NEBIOtTM Kit. The probe was cleaned by NucTrap� Push Columns (Stratagene Inc.). Northern blotting using labeled DNA probes were done as described (Ausubel et al., 1988). An equivalent 10 million counts of radiolabeled DNA probes were used for one typical labeling.

For Southern blot purpose, genomic DNA was isolated from stationary

growth cells. It was then digested with BamH I, separated on agarose gels, and then transferred to nylon (Kaiser et al., 1994). 32P-labeled DNA probe labeling was done as described in Northern blotting.

To prepare Western blot samples, P. pastoris were grown on YPD or on methanol and then adapted to glucose. The yeast cells were collected and prepared for SDS gel according to Kaiser (Kaiser et al., 1994). The samples were separated on a SDS minigel (Bio-Rad). The proteins were transferred to the nitrocellulose and incubated with anti-HA.11 mAb (against mouse) or polyclonal antibody (against rabbit) (Babco). The blot were then incubated with secondary anti-mouse or anti-rabbit antibodies conjugated with HRP. Protein bands was detected by ECL method (Amersham). Isolation of PFK1 and GSA7 Knockouts

The construction of WDKO1 (his4, pfkl::ARG4) cell was done by disrupting the PFK1 gene of PPF1 (his4, arg4). First, the 1.2 kb Hind III fragment within the open reading frame of P. pastoris PFK1 was replaced with the 3.0 kb S. cerevisiae ARG4 gene from pYM25. A linear 5.6 kb fragment with








39

the ARG4 gene flanked by PFK1 was excised from the shuttle vector by BamH I digestion and used directly to transform PPF1 cells. Cells were grown on transformation plates supplemented with histidine (40 pg/ml). Colonies were isolated and their ability to degrade AOX during glucose adaptation examined (see Mutants Generation and Isolation section). PFK1 gene disruption was verified by measuring phosphofructokinase activity assay (Blangy et al., 1968) and by Southern blotting (see above). The construction of WDKO7 was basically the same as for WDKO1. The S. cerevisiae ARG4 gene was put in between Hind I1l and Bgl II sites of the GSA7 gene and the knockout fragment was cut out by Apa I and Sca I restriction enzymes and used to transform PPF1 cells. The transformants were first screened with an AOX direct colony assay. The colonies that appeared purple in the assay should lose AOX degradation ability and are thus the putative "knockouts". Possible knockouts were tested for the correct gene targeting by PCR the genomic DNA extracted from these clones. The clone has the expected PCR fragment (4.9 kb instead of 2.5 kb in a normal strain) is the true knockout and was named as WDKO7 (Agsa7). Site-directed Mutagenesis of PFK1

The aspartic acid at position 362 of Pfkl protein was changed to a serine by site directed mutagenesis. This was done using the Stratagene@ QuikChange"M Site-Directed Mutagenesis Kit. Mutated pfkl was prepared by PCR utilizing two complementary oligonucleotide primers (synthesized at the University of Florida Oligonucleotide Core) that contained the desired two base








40

mutations, Pfu DNA polymerase, and pDLT1 containing the wild type PFK1 as a template. This resulted in "nicked" plasmid containing pfkl. The template DNA was digested by Dpn I endonuclease and the nicked plasmid was repaired and amplified in Epicurian Coli� XL1-Blue cells. Pfkl was then subcloned into pYM4 (pWP-pfk). Sequencing in both directions of the pWP-pfk was performed at the University of Florida Sequencing Core by the dideoxy chain termination method and desired bases change was verified (see Fig. 5-4). pWP-pfk was then introduced into gsal-1 and WDKO1 by electroporation. The transformants were selected based on the integration of HIS4 gene of plasmid into the genome. The linearization of the plasmid inside HIS4 gene promoted integration into HIS4 gene locus of the genome.

Generation of a HA Epitope Taq in GSA7 and YHR171w Gene

Two primers were designed to PCR the GSA7 gene plus a HA epitope tag at the C terminus of the protein. The 5' primer started 350 bases before the start codon of GSA7 in order to include its endogenous promoter, while the 3' primer was at the end of the coding sequence plus a HA epitope tag. The HA epitope tagged GSA7 gene was amplified by PCR with ID-PROOFTM polymerase (ID Labs Biotechnology). The successful construct was verified and subcloned into vector pYM4 Cla I and EcoR V sites (pYWP7-HA). The construct pYWP7-HA was linearized in HIS4 gene or GSA7 gene and used to transform gsa7 cells. In S. cerevisiae, the same strategy was used to generate a HA epitope tagged YHR171w and put into pYM4. The YHR171w gene was also put in the pHWO10









41

vector behind a GAP promoter (Waterham et al., 1997) and in the pPIC3 vector behind an AOX promoter (both vectors are a gift of Dr. JM Cregg). These constructs were used to transform S. cerevisiae as well as P. pastoris. The expression of HA tagged protein was confirmed by Western blot by HA antibodies.



Fluorescence Microscopy and Electron Microscopy



Yeast cells were grown on YPD, glucose, ethanol or methanol media for different experiment setting. The cells were then fixed in 3% (w/v) of paraformaldeyde (pFA) for 10 min at room temperature. The yeast cells were centrifuged and resuspended in 50mM potassium phosphate buffer (Kpi) I 3% pFA at room temperature for another 30 min with gentle rotation. The cells were washed with Kpi and resuspended in SPC buffer (50 mM K2HPO4, 16 mM citric acid, 1M sobitol, pH 5.8). The cells were spheroplasted with Zymolase 20T in SPC buffer for 15 min and the digestion was monitored every 5 min for cell wall digestion completion. After digestion, the cells were washed with SPC buffer and put on a 15-well slide precoated with polylycine (ICN Biomedicals). The fluorescence containing media such as phalloidin, quinacrine were applied to the slide for direct observation or the slides were placed in the blocking medium for immunofluorescence microscopy study (5% normal goat serum in PBS). For this purpose the cells on the wells were incubated with primary antibody for 1-2 hr or








42

overnight and then with secondary antibody for one hour. The observation was done with a Zeiss axiophot photomicroscope using a Zeiss 100OX oil emersion.

Ultrastructural analysis was performed using a potassium permanganate fixation protocol (Veenhuis et al., 1983). This procedure effectively delineates membrane structures in methylotrophic yeasts. Briefly, cells were harvested by centrifugation, washed in water, and fixed in 1.5% KMnO4 in veronal-acetate buffer (0.3 mM sodium acetate; 0.3 mM sodium barbital, pH 7.6) for 20 min at room temperature. The specimens were then washed three times and dehydrated in increasing concentrations of ethanol wash. This was followed by 100% propylene oxide twice. The cells were then infiltrated with a 50:50 mix of propylene oxide and the POLY/BED 812 (Polysciences, Inc., Warrington, PA) for two days. The preparations were dried by a vacuum overnight, infiltrated with 100% POLY/BED with accelerator 2,4,6-Tri(dimethylaminomethyl) phenol (DMP-30@, Polysciences, Inc.) for another two days, and then incubated in an oven overnight at 600C. The resulting samples were then mounted on the blocks and the blocks were then sectioned (by D. Player, Department of Anatomy and Cell Biology, University of Florida College of Medicine) and prepared for examination on a JEOL 100CX II transmission electron microscope.













CHAPTER 3
CHARACTERIZATION OF PICHIA PASTORIS MUTANTS DEFECTIVE
IN GLUCOSE-INDUCED MICROAUTOPHAGY



Introduction



Methylotrophic yeast P. pastoris can utilize methanol as a sole carbon source by synthesizing those peroxisomal and cytosolic enzymes necessary to assimilate methanol. When P. pastoris cells were grown on methanol, their peroxisomes were induced in large size and numbers. During the subsequent adaptation to glucose or ethanol, their peroxisomes were rapidly degraded (Tuttle et al., 1993). P. pastoris has been an ideal model to investigate the molecular events of peroxisome biogenesis (Subramani, 1996) as well as peroxisome degradation (Tuttle and Dunn, 1995).

Our lab has identified two distinct pathways for the degradation of

peroxisomes in P. pastoris (Tuttle and Dunn, 1995) namely, microautophagy and macroautophagy. These pathways are independently regulated, but share at least one common event, vacuolar degradation. In order to better define the molecular aspects of peroxisome turnover in P. pastoris, glucose-induced selective autophagy deficient (gsa) mutants have been generated from the











Steps for the Identification of those
Proteins Required for Microautophagy

Mutagenesis of Parental Strain


Isolation and Selection of Mutants


Backcross and Complementation Analysis



Biochemical and Morphologic
Characterization of Mutants


Rescue of Mutants with Genomic DNA


Determination of Genomic DNA Sequence Further Studies


Figure 3-1. A flow chart of mutants characterization. A mutant was selected for its inability to degrade AOX during glucose adaptation. Once its biochemical and morphological profile had been identified, it went through backcrossing and complementation analysis. The mutant was then transformed with a P. pastoris genomic DNA library to identify the responsible gene. Once the gene was recovered and verified, the gene was sequenced and, according the sequencing result, the subsequent studies were determined and carried out.










parental strain GS115 by chemical mutagenesis (Tuttle and Dunn, 1995) for their inability to degrade peroxisomal alcohol oxidase (AOX). The putative mutants that appeared purple in an AOX direct colony assay were selected. They were then subjected to liquid media AOX and FDH assays for further verification. The mutant strains were then backcrossed with parental GS190 (an arginine auxotroph) and then mated with another mutant in a histidine auxotrophic background (see below), to identify their complementation groups. The selected mutants were also backcrossed with parental GS190 and GS115 two to four times to eliminate background mutations that were generated in the chemical mutagenesis unrelated to gsa phenotype. A flow chart of mutant generation, identification and characterization steps has been included in Fig. 31. In this chapter, I will detail the genetic, biochemical and morphological characterization of gsa mutants.



Complementation Group Identification



The purpose of the complementation analysis of mutant strains is to

provide an indication of the genetic relationship of the mutations involved and thus to further characterize the gsa mutants. Complementation groups were identified by backcrossing different gsa mutants with an arginine auxotroph GS190. The resulting arginine mutants were then mated with other gsa mutants (histidine auxotrophs) as well as parental GS115. Their progeny were analyzed










for their peroxisomal AOX degradation ability. If the progeny (diploid cells) rescued the defect of both mutant strains (regaining AOX degradation ability), then these two strains represent two different complementation groups and two unique genes in each of the mutants is defective. If they complemented each other, they belong to the same complementation group with possible different mutated alleles. A sample complementation study graph of strain WDY4, WDY8 and WDY10 is shown in Fig. 3-2. AOX activities of six hour glucose adaptation were presented as a percentage of zero hour activities. As shown in the graph, the progeny of WDY8 and WDY4 could degrade AOX and little AOX activity remained after 6 hour glucose adaptation, and thus they complemented each other. WDY4 and WDY8 did not complemented WDY10. The same was true for WDY10 and GS115. The result of AOX activity assays of these diploids clearly showed that WDY4, WDY8 and WDY10 are three separate gsa groups. The fact that WDY10 was not complemented by GS115 made it clear that WDY10 has a dominant trait. Currently, I have identified eight complementation groups. All gsa mutants except gsa8 showed recessive traits while gsa8 showed a dominant trait. WDY2, WDY3 and WDY8 belong to the same gsal mutant group and are thus named as gsal-1, gsal-2 and gsal-3 respectively. A list of gsa mutants is shown in Table 3-1. Some of the mutants showed variable AOX degradation results, such as WDY9, 11, 12, 14 and WDY15. This variability may be due to several mutations in one single mutant strain and thus this strain might need to be backcrossed before complementation analysis. Since there











gsa Mutants Complementation Analyses


90 80 70 S60 o 50. 40 1. 30


0 -


WDY10 WDY4 WDY8 WDY4 WDY4 WDY8
X X X X X X
GS115 GS115 GS115 WDY8 WDY10 WDY10

Matings



Figure 3-2. Identification of gsa mutants complementation groups. The figure is used to illustrate how complementation groups were identified. WDY4, WDY8 and WDY10 were screened from frozen stocks and identified as gsa mutants. They were then subjected to backcrossing. The resultant WDY8al, WDY10al, WDY4al (Arg auxotroph, generation one) were then mated with GS115 cells and themselves. The progenies of mating were subjected to AOX assay. Values of AOX activities after 6 hour glucose adaptation were represented as a percentage of the 0-hour activity. They were measured at least three times in the separate experiments and were shown on the graph as average � sd. The result showed that WDY4, WDY8 and WDY10 belong to three new complementation groups. However, WDY10 showed a dominant trait since the prototroph of WDY10 and other mutants as well as GS115 showed no rescue effect. WDY4 and WDY10 were subsequently named as gsa4 and gsa8. WDY8 was subsequently found to be in the same gsa group as gsal-1 (data not shown) and was thus named as gsal 1-3 (WDY3 has been identified as gsal 1-2).


T








48

are still more potential mutants and their complementation groups have not been identified, more work needs to be done on the remaining mutants. Although the identified eight gsa complementation groups cover only part of the regulatory molecules that control the process of autophagy, further characterization of these mutants now is possible because of the establishment of these complementation groups. Future studies of the remaining mutants will yield information to better cover all the molecules participating in the peroxisome degradation process as well as to better understand the autophagy mechanism.



Screening for qsa Mutants



At the time of complementation analysis of gsa mutants, the gsa mutants were also going through backcrossing cycles (see Materials and Methods for backcrossing steps). The purpose of backcrossing was to prepare gsa mutants for further genetic study. Because the mutant strains were generated by random chemical mutation, the mutations occurred at various sites in the genome. By backcrossing two to four times back & forth from histidine auxotroph background to arginine auxotroph background, most background mutations that were not related to gsa phenotype were eliminated.

In eight gsa complementation groups (gsal to gsa8) (see table 2-1) that have been identified, gsal-1 and gsa2 have been successfully backcrossed four times. Gsa3, gsa5, gsa7 and gsa8 have been backcrossed two times and gsa4








49

has been backcrossed one time, but the phenotype of this strain is not stable in its arginine auxotroph background. Gsa6 has a stable and strong phenotype biochemically. However, we have failed in three separate attempts to backcross it. Once it was placed into the arginine auxotroph background, the phenotype was lost. It is possible that the phenotype effect of gsa6 is an accumulated one of several defective genes. Backcrossing that leads to the loss of any one of them shows no phenotype in gsa6. Gsal-1 and gsa7 were later chosen for further genetic analysis.



Glucose Induced Microautophagy is Defective in gsa Mutants



Previously, we have shown that parental strain GS115 undergoes microautophagy during glucose adaptation. The loss of AOX activity, the degradation of AOX protein and the degradation of peroxisomes were compatible (Tuttle and Dunn, 1995). We selected gsa mutants on the basis that the mutant could not degrade AOX and peroxisomes. Of the collection of possible mutants, I have identified eight gsa complementation groups. All mutants were unable to degrade peroxisomal AOX during glucose adaptation while their ethanol-induced macroautophagy proceeded normally (Fig. 3-3 and Fig. 3-4). The biogenesis of peroxisomes observed during methanol induction was intact in all these mutants as determined by light microscopy. The degradation of AOX and FDH during glucose adaptation was rapid in parental










GS115 cells (also see Fig. 1-2). Less than 20% of the activities of AOX and FDH was left in GS115 after six hour glucose adaptation. However, in mutants presented in these two graphs, they lacked the ability to degrade AOX that has been induced under methanol growth condition. Compare to parental GS1 15 cells, AOX degradation in response to glucose adaptation in all mutants was inhibited to various extents. The degradation of FDH, a cytosolic protein that required for methanol utilization, was also shut down in most gsa mutants. However, in gsa4 and gsa8, FDH degradation proceeds nearly normal when compared with parental GS1 15. Since FDH is a cytosolic protein, one possibility is that a mutation block an early event of peroxisome microautophagy pathway in this mutant but the early cytosolic protein degradation pathway may be controlled differently from that of peroxisome degradation and thus the FDH degradation process is intact. Interestingly, in the complementation analysis done by Mr. Kendal (Kendal, unpublished data, 1993), gsa4 strain did not complement either gsal-1 or gsa2 cells. This may be either due to more than one mutant gene in it since gsa4 was not backcrossed at the time of experiment or to some unknown reasons. As for gsa8 cells, I found it is a dominant mutant in our complementation analysis. We still need to define the relationship of dominant gsa8 and FDH degradation.

Ethanol induced peroxisome degradation in all mutants proceeded

normally (see Table 3-1). This confirmed that our gsa mutants only affected early events of microautophagy of peroxisomes since vacuolar degradation of

















W 100


S80

60


" 40 20


AOX and FDH Activities in gsa Mutants
during Glucose Adaptation


GS115 gsal-3 gsa4 gsa5 gsa6 gsa8


Figure 3-3. AOX and FDH activities of gsa mutants under glucose adaptation. P. pastoris (GS115) and gsa mutants were cultured in methanol induction medium until stationary, at which time glucose was added to begin adaptation (time zero). Cell-free extracts were prepared at 0- and 6-hour time point and alcohol oxidase (AOX) and formate dehydrogenase (FDH) activities were measured. GS1 15, gsal-3, gsa4, gsa5, gsa6 and gsa8 are included in this figure. Values are presented as a percentage of the 0-hour activity. All assays were measured at least three times and the values shown on the graph represent the average � sd.


















AOX and FDH Activities in gsa Mutants
during Glucose Adaptation


0
0



- 0



Z
0




Z
W�


GS115 gsal-1 gsa2 gsa3 gsa7


Figure 3-4. AOX and FDH activities of gsa mutants under glucose adaptation. P. pastoris (GS115) and gsal-1, 2, 3 and gsa7 were cultured in methanol induction medium until stationary, at which time glucose was added to begin adaptation (time zero). Cell-free extracts were prepared at 0- and 6-hour time point and alcohol oxidase (AOX) and formate dehydrogenase (FDH) activities were measured according to Materials and Methods. Values are presented as a percentage of the 0-hour activity. All assays were measured at least three times and the values shown on the graph represent the average + sd.



















Table 3-1 Characterization of gsa Mutants


Genotype


Phenotype
Glucose ETOH AOX FDH AOX FDH


GS115 Parental strain his4 + + + GS190 Parental strain ar94 + + + PPF1 Parental strain his4, arg4 + + + WDY1 gsa2 his4, gsa2 - - + WDY2 gsal-1 his4, gsal-1 - - + WDY3 gsal 1-2 his4, gsal 1-2 - - + WDY4 gsa4 his4, gsa4 - + + WDY5 gsa5 his4, gsa5 - - + WDY6 gsa6 his4, gsa6 - - + WDY7 gsa7 his4, gsa7 - - + WDY8 gsal 1-3 his4, gsal 1-3 - - + WDY10 gsa8 his4, gsa8 - + + WDY13 gsa3 his4, gsa3 - - + -


Strains


gsa group










peroxisomes was intact in ethanol induced macroautophagy which is shared with microautophagy. Although my work focused on glucose-induced microautophagy, the studies of ethanol-induced macroautophagy are interesting and important. This is because most of autophagy mutants in S. cerevisiae and H. polymorpha are macroautophagy mutants. Only P. pastoris can undergo both degradation pathways during glucose and ethanol adaptation. A list of complete biochemical profiles of gsa mutants and parental strains are compiled in Table 3-1. GS115 is the parental strain used for chemical mutagenesis. GS190 is used for backcrossing. PPF1 is a double auxotroph used later in the experiment. Their biochemical profiles are provided. All identified gsa mutants are included in the table. Some of them are not listed here due to the variable results of their biochemical studies such as WDY9, 11, 12, 14 and WDY15.

While I was studying gsal-1, I also backcrossed WDY1 (gsa2) four times to prepare for further genetic analysis. After I finished biochemical and morphological studies of gsa2, I transformed gsa2 cells with a genomic DNA library to identify GSA2 gene. The colonies appeared white in an AOX direct colony assay regained their AOX degradation ability and thus should bear rescue plasmids with GSA2 gene. Much to my surprise, about 10% of the colonies appeared white on the AOX direct colony assay and thus were putative rescue clones. The positive and negative control strains tested in the direct colony assay worked fine which meant the detection of white rescue colonies was not due to technical errors. I eliminated some of the false positive rescue










colonies by using liquid media AOX assay. However, I still have too many promising colonies. Twelve plasmids were recovered from 12 of these colonies. Each plasmid appeared different based on restriction enzyme analysis. Some of them had no genomic DNA insert but still rescued gsa2. Although I could not exclude the possibility that there are GSA2 or GSA2 suppressor genes in my collections of "rescue colonies", I could not identify them.

I tested different growth conditions in order to seek optimal conditions to grow and to characterize the rescue clones. One surprising phenomenon was that the gsa2 mutant was rescued by addition of 80 pl of 10mg/mL histidine at the beginning of glucose adaptation during liquid media AOX assay. Subsequent studies verified that this was a true phenomenon and the doubling of histidine amount at the beginning of methanol induction did not affect this result. This fact promoted me to re-consider the failed gsa2 rescue. In the rescue study, only mutants bearing pYM8 containing plasmids and expressing histidine could survive and grow. Since gsa2 can be complemented by histidine, the expression of histidine in the complemented strains partially or completely offset the phenotype of gsa2 and thus resulting in a lot of rescue colonies. The expression level of histidine might be quite different in each clone and thus not all strains bearing pYM8 could rescue gsa2. We do not know why histidine rescues gsa2 when it was added upon glucose adaptation. A review of published references did not help us either. One possibility is the addition of histidine at the time of glucose adaptation kicks in a nonspecific















1 00

llll glucose
-E glucose and histidine 80

bk

60
A

0 40






0
0
GS115 gsal-1 gsa2 gsa4 gsa5 gsa8








C
S 100
ll glucose
Glucose and histidine
I
! - 80



II o: TT





r, 20
40



20



GS115 gsal-2 gsal-3 gsa3 gsa6 gsa7







Figure 3-5. AOX degradation and histidine addition during glucose adaptation. The test was used to identify gsa mutants that are not affected by histidine addition The experiment was done as a liquid media AOX assay described in Materials and Methods except that at the zero time point of glucose addition, glucose or glucose/histidine were added to each gsa mutant to compare the effect of histidine addition. All assays were measured at least three times and the values shown on the graph represent the average � sd.










protein degradation pathway. It is also possible that histidine may directly activate macroautophagy. To further probe this problem, I first screened all the mutants that I have and test their ability to degrade AOX when histidine was added at the time of glucose adaptation. It was not surprising to find that gsal was not rescued by histidine. Using non-backcrossed gsa2 strain for the same experiment, I verified that the histidine rescue effect exists and concluded that the phenomenon was not due to improper selection during the backcrossing. All other gsa mutants were tested for the effect of histidine addition at the beginning of glucose adaptation and Fig. 3-5 showed the AOX activities remaining after a six-hour glucose adaptation in these gsa mutants. The effect of histidine addition was compared with control groups grown in the same condition except no histidine addition at the onset of glucose adaptation. As shown in the upper panel, histidine restored AOX degradation ability (or partially restored) in gsa2, 4, 5 and gsa8 while in gsal, gsa3, gsa6 and gsa7, histidine has no effect on their AOX degradation. However, the rescue effect varied in these three strains which may be a reflection of the expression level of histidine. Gsal, gsa3, gsa6 and gsa7 were chosen for further genetic studies.



Morpholoqical Studies of gsa Mutants



The ultrastructural analysis was done by the potassium permanganate fixation protocol (Veenhuis, et al., 1983). The basic phenomena in all these










mutants were their inability to degrade peroxisomes in a prompt manner albeit these eight mutants showed distinct presentations (Fig. 3-6 and Fig. 3-7). In gsal-1 mutant, the vacuole appeared round and small. It lacked the finger-like extensions or invaginations present in glucose-adapting GS115 cells (see Fig. 1-3). The morphological presentation of the vacuole did not change much during glucose adaptation. The vacuole was not active since there were few autophagosomes and little cell debris inside the vacuole when compared to GS115 cells. All this suggested that gsal-1 is required for the initiation of microautophagy and the molecule mutated in gsal-1 may control an early step of microautophagy, that is, the initiation of autophagy. The limited degradation of AOX and FDH probably is due to inefficient degradation of peroxisome via macroautophagy under glucose adaptation conditions.

In gsa4, the peroxisomes were identified outside of the yeast vacuole

after 3 hour's glucose adaptation while autophagosomes were seen associated with the yeast vacuole. Although AOX degradation was defective in gsa4 cells (see Fig. 3-7), the extent of the defect, when judged by AOX degradation, was much less when compared with other mutants (Fig. 3-3). This was in agreement with the morphological data. We knew in gsa4, cytosolic protein FDH's degradation proceeds normally. It is possible that the defect in gsa4 was mild and although degradation of AOX ability was inhibited, the cytosolic protein FDH degradation still could proceed via impaired microautophagy or inefficient macroautophagy. It is also possible that the mutant has a recognition defect








59

and this defect disrupted the microautophagy of peroxisomes while its cytosolic protein degradation in gsa4 cells was intact. It should be interesting to pursue gsa4 to identify the defective gene of this mutant.

The genetics data showed that gsa8 has a dominant trait. When

examined under electron microscopy, it bore similarity to gsa4 cells albeit to less extent than to the active level of the yeast vacuole. During glucose adaptation, autophagosomes with cell debris or cytosolic contents were seen around and inside the yeast vacuole (Fig. 3-6). However, large peroxisomes outside of the yeast vacuole were easily identified, and we did not observe vacuolar degradation of peroxisomes. Our biochemical data showed that the AOX degradation was defective. However, FDH degradation proceeded normally. I suspected that the slow progress of macroautophagy leads to a much greater loss of cytosolic protein FDH while the defect in microautophagy keeps AOX level relatively higher in gsa8. Gsa4 and gsa8 should be in the same group and their defective gene products control the recognition step in the process of peroxisome microautophagy downstream of Gsalp.

In gsa2 cells, finger-like extensions of the vacuole were found

surrounding the peroxisomes after 3 hour glucose adaptation and this was not seen at the time of methanol induction. However, it appeared that fusion of vacuolar membrane did not occur, and thus, sequestration of peroxisomes was not complete (Tuttle and Dunn, 1995). Although the morphology suggested that microautophagy was shut down in gsa2 cells, we did observe some small








60

autophagosomes were in contact with the yeast vacuole. The biochemical data showing that AOX degradation was defective corroborated this observation. However, slow degradation of AOX and FDH did exist and it appeared that the gsa2 cells might be able to adapt to their deficiency and the environmental situation by turning on the macroautophagy pathway albeit much less efficiently. I suspected a fusion step of vacuolar membrane during the engulfment of peroxisomes is defective in gsa2 mutants and Gsa2p should act downstream of Gsalp.

Morphologically, gsa3, gsa6 and gsa7 showed similar characteristics as gsa2. The initiation of autophagy was turned on during the glucose adaptation. The vacuole has formed protrusions that surround peroxisomes. However, we did not observe peroxisome degradation in the yeast vacuole. The biochemical data of AOX and FDH activities during glucose adaptation also supported that peroxisome degradation was impaired in all these three mutants. It appears these mutants are defective in an event at or before the fusion of protrusions of vacuolar membrane, a homotypic vacuolar membrane fusion event. I speculated the molecules that are defective in these mutants may either coordinate the fusion steps in a complex or interact in a step by step manner. The proteins may also reside at different compartments such as peroxisomal and vacuolar membrane or in cytosol. In conclusion, gsa2, gsa3, gsa6 and gsa7 are classified in a mutant group with defects in homotypic vacuolar membrane fusion.










In gsa5 cells, during methanol induction, we observed peroxisomes,

mitochondrial and cell debris were wrapped with several high density layers of membrane, a hallmark of macroautophagy. This means that macroautophagy has been turned on in this particular mutant during methanol induction, which is unusual. Some of these autophagosomes were in contact with the yeast vacuole. When the cells were examined at 3 hour of glucose adaptation, we observed the same phenomena as in methanol induction. The macroautophagy was apparently still functioning since multiple membrane layers of cell contents could be easily identified. The yeast vacuole has formed membranous protrusions and the sequestration of peroxisomes by vacuolar membrane protrusions was initiated. However, a subsequent step was defective since no peroxisome degradation was observed. The biochemical data showed that AOX and FDH degradation was defective in gsa5 and supported our conclusion that gsa5 was a very distinct mutant defective in microautophagy. I speculated that the defective gene product may inversely control a common event of microautophagy and macroautophagy. Defects in this gene product turned on macroautophagy and interfered with microautophagy process during glucose adaptation. However, this is only a simplified explanation. Although I am aware of this mutants is unique in its phenotypic presentation, I put gsa5 in the same group of vacuolar membrane fusion defect mutants. Further genetic studies of this mutant will reveal the function as well the role of the defective gene product that control the microautophagy and macroautophagy process.















































Figure 3-6. Ultrastructural studies of gsal-1, gsa2, gsa3, gsa6, gsa7 and gsa8 mutants under glucose adaptation. Gsa mutants were grown on methanol medium and then adapted to glucose for 3 hours. Cells were prepared by a potassium permanganate protocol for ultrastructural analysis (Tuttle and Dunn, 1995). The representative morphology of mutants at 3 hour adaptation is shown here. The detailed description is in the text. N, nucleus; P, peroxisome; V, vacuole and M, mitochondria. Bar: 0.9 pM.








































Figure 3-7. Ultrastructural studies of gsa4 and gsa5 mutants under glucose adaptation. Gsa4 and gsa5 cells were grown on methanol and then adapted to glucose for 3 hours. Cells of Oh and 3h adaptation were prepared by a potassium permanganate protocol for ultrastructural analysis (Tuttle and Dunn, 1995). The representative morphology of gsa4 and gsa5 is shown here. The detailed description is in the text. N, nucleus; P, peroxisome; V, vacuole and M, mitochondria. Bar: 0.7pM.










In conclusion, The morphological and biochemical characteristics in all eight gsa mutants clearly suggested that the gene products defective in these gsa mutants regulate different steps in peroxisome microautophagy. Further genetic studies will reveal their identities and help us to understand the mechanism of autophagy.



Chapter Summary




In this chapter, I have characterized most of the available mutants in our stocks and eight gsa complementation groups (gsal to gsa8) have been identified. Strain gsal-1, 2, 3, 5, 7, and gsa8 have been backcrossed two to four times and are ready for further genetic studies. Gsal-1 to gsa8 have also been characterized biochemically and morphologically. All gsa mutants lacked the ability to degrade peroxisomes during glucose adaptation. However, their abilities to degrade peroxisome during ethanol adaptation were not affected. I have divided these mutants into four groups. The first group includes gsal, which is defective in the initiation of microautophagy. The second group includes gsa4 and gsa8 that have selective degradation defect in AOX but not FDH in peroxisome autophagy. A recognition step might be defective in these two mutants. The third group includes gsa2, gsa3, gsa6 and gsa7. Our data suggested that they are homotypic vacuolar membrane fusion mutants











P. pastoris Methanol - Peroxisomes induction


gsal -



gsa4,gsa8 ) gsa2, gsa3, gsa5, .
gsa6, gsa7


Glucose Signaling


Peroxisome
Sequestration


Vacuolar Membrane Membrane Fusion


pra, prb 3) Vacuolar Degradation of Peroxisomes



Figure 3-8. Microautophagy of peroxisomes and gsa mutants in P. pastoris. The process of microautophagy of peroxisomes is depicted here based on the biochemical and morphological evidence we had (see text). Mutations in specific GSA genes acted at the specific stages of microautophagy are also indicated. All GSA genes seems act at a step before vacuolar degradation. The vacuolar degradation mutants pra and prb are also shown in the diagram to complete the microautophagy pathway.










downstream of Gsalp. The final group includes gsa5. In this group, macroautophagy is abnormally turned on under both methanol and glucose growth condition while a fusion step of vacuolar membrane protrusions of microautophagy is shut down as in group 3. As shown in the diagram, all mutants are defective at a step before vacuolar degradation. The representative vacuolar protein degradation mutant pra and prb are also included in the diagram. Although all defects in gsa mutants are specific for glucose-induced microautophagy, the other vacuolar degradation pathway(s) might share a common degradation step or, a common regulatory molecules with the microautophagy pathway. An example of this is microautophagy and macroautophagy at least share a common step of vacuolar degradation and in gsa5, the mutated gene product regulates a step in microautophagy buy might also control a step in macroautophagy. The distinct morphological and biochemical phenotypes described in this chapter hint that the defective genes in these gsa mutants may regulate different steps of microautophagy of peroxisomes (Fig. 3-8). The next step after mutant characterization is to identify the defective genes in the mutants and to study the functions of their gene products in the microautophagy process. Some of the mutants in our original possession such as WDY9, 11, 12, 14 and WDY15 showed variable biochemical results and thus we chose not to study these mutants further at this time. Gsa6 cannot be backcrossed to arginine auxotrophs. However, since it is a stable mutant,








67

we can rescue it without backcrossing. Gsa8 has a dominant trait and thus classic rescue study is not applicable for it. A clone bank could be made from this strain and the gene can be identified by screens for autophagy defects after transformation. In the following chapters, I will discuss the genetic studies conducted in gsal-1 and gsa7 mutants that we believed control the steps of initiation and vacuolar membrane fusion events of microautophagy of peroxisome.














CHAPTER 4
GSA1 PROTEIN IS PFK1 PROTEIN



Introduction



Our studies presented in chapter 3 showed WDY2 (gsal-1) is a

genetically stable and a morphologically distinct mutant. It probably signals the initiation of peroxisome microautophagy. We chose it as the first mutant for further genetic studies. As mentioned in the chapter 3, we followed the experimental steps (see Fig. 3-1) by first backcrossing WDY2 four times. This eliminated the multiple mutations generated during chemical mutagenesis that were not related to gsal-1 phenotype. We then complemented gsal-1 with a genomic DNA library to recover the gene that corrected the defect in gsal-1. In the following two chapters, I will detail my pursuit of Gsalp.



Gsal-1 Siqnaling an Upstream Event of Microautophaqy



Microautophaqy of Peroxisomes is Defective in qsal-1

As mentioned in chapter 2, I used peroxisome alcohol oxidase (AOX)

and cytosolic formate dehydrogenase (FDH) as a quantitative index to monitor



































GS115


WDY2 GS115 WDY2


GLUCOSE


ETHANOL


Figure 4-1 Glucose and ethanol adaptation in parental GS115 and WDY2 (gsal-1) mutant. The production of peroxisomes was induced first by culturing the cells in methanol induction medium. After 24-36 hours, glucose or ethanol was added to the medium at final concentrations of 2% and 0.5%, respectively. Cell free extracts were prepared at 0 and 6 hours of adaptation. Alcohol oxidase (AOX) and formate dehydrogenase (FDH) activities were assayed as described in Materials and Methods and presented as a percentage of the activity measured at 0 hour. The values represent the mean � s.d. of three or more determinations.


- AOX
FDH


120 100


0










peroxisome microautophagy of gsal-1 and GS115 during glucose adaptation. As shown in Fig. 4-1, the glucose-induced degradation of both AOX and FDH was dramatically reduced in gsal-1 when compared with that of parental GS115. Meanwhile, the loss of AOX in gsal-1 cells during six hours of ethanol adaptation was comparable to that observed in GS115 cells. FDH degradation was not observed under ethanol adaptation as in GS1 15. The data suggested that the mutation in gsal-1 affected only the glucose-mediated degradation pathway, while its ethanol-mediated degradation pathway proceeded normally. In addition, vacuolar function appeared normal in the gsal-1 cells since under the ethanol-mediated degradation pathway, the vacuolar degradation of peroxisome is intact. Our previous studies have shown that both degradation pathways shared a common vacuolar degradation step. In conclusion, our biochemical data supported that gsal-1 is a glucose-mediated peroxisome autophagy mutant with an early event defect. Morphological Study Revealed That a Step Before Peroxisome Sequestration Is Blocked in gsal-1

We next examined the ultrastructure of gsal-1 cells during glucose

adaptation. Large peroxisomes were induced and the yeast vacuole was small and round in parental GS115 cells at zero hour of adaptation. When GS115 cells were adapted to glucose, the peroxisomes were few in number when compared with 0 hour glucose adaptation. They were found to be associated with the vacuole. The vacuole contained cellular structures such as









71

peroxisomes and cell debris (Fig. 1-3). In addition, finger-like extensions of the vacuole surrounding peroxisomes were common. In gsal-1 cells at 0 hour of adaptation, gsal-1 cells had large peroxisomes that have been induced by methanol while the yeast vacuole was small and round. After three hour of glucose adaption, the peroxisomes were found outside the vacuole in gsal-1. The vacuole appeared "inactive", that is, it lacked cellular contents, membrane invaginations, and finger-like extensions. Also shown in Fig. 4-2 is the WDY1 (gsa2) mutant. When comparing the morphology of these two mutants, it supported the hypothesis that gsa2 is different from gsal-1 and Gsalp should act before Gsa2p.

The ultrastructural data and the biochemical data of gsal-1 and gsa2 suggested that both gene products GSA1 and GSA2 are required for the delivery of peroxisomes to the vacuole for degradation during glucose adaptation. I suggested that microautophagy was mainly shut down in both mutants in early steps of microautophagy and that the limited autophagy observed in our biochemical analysis occurred by inefficient macroautophagy under glucose adaptation condition. The different morphological presentations of gsal-1 and gsa2 showed that their gene products control different events of peroxisome microautophagy. My observations supported the conclusion that Gsalp controls the initiation of microautophagy while Gsa2p controls a downstream event of microautophagy after Gsalp such as the fusion step of vacuolar membrane during the engulfment of peroxisomes by yeast vacuole.









































Figure 4-2. Morphology of gsal-1 and gsa2 under glucose adaptation at Oh and 3h time points. The ultrastructural analysis of gsal-1 and gsa2 was done by a potassium permanganate fixation protocol of yeast cells at Oh and 3h of glucose adaptation. At 3h of glucose adaptation, gsal-1 showed no morphological change when compared to Oh. However, in gsa2, yeast vacuole formed protrusions to wrap the peroxisomes. This is an indication that the defective gene in gsa2 probably controls a step downstream of gsal-1 in the microautophagy process. M: mitochondrion; N: nucleus; P: peroxisome; V: vacuole. Bar: 0.9 pM.










pDLT1 Complemented qgsal-1 Phenotype



The biochemical and morphological studies have suggested that Gsalp control the initiation step in microautophagy. I chose gsal-1 for further genetic analysis. The main purpose of genetic study is to recover the gene that is defective in a specific gsa mutant. A P. pastoris genomic DNA library (a generous gift of Dr. J. M. Cregg, Oregon Graduate Institute) was made and put into an episomal shuttle vector pYM8. WDY2 (his4, gsal-1) was transformed with the genomic library and plated on transformation medium lacking histidine. The resulting transformants were screened by an AOX direct colony assay to test their ability to degrade AOX in gsal-1 cells (Tuttle and Dunn, 1995). If the cell acquired the GSA1 gene and the gene were transcribed, it should rescue the defect of gsal-1 and thus regain the ability to degrade AOX. When examined on nitrocellulose by direct colony assay, the rescued colonies should appear white while non-rescued should appear purple. Of the fourteen putative rescued colonies that were isolated by Dr. Dan Tuttle, only two maintained the wild type ability to degrade AOX when assayed in liquid media. The plasmid DNA isolated from these two clones was amplified in E. coi DH5a cells. Both vectors contained a 5.7 kb genomic DNA insert that displayed identical restriction maps. I next tried to verify the ability of this plasmid, pDLT1, to complement the mutant gsal-1 cells by reintroducing it back to gsal-1. A 3.8 kb BamH I fragment was also subcloned into pYM8 (pDLT-BB) and used to










transform gsal-1. Since the instability of the episomal vector, pYM8, contributed to inconsistent measurements of AOX degradation, we induced stable integration of these constructs prior to measuring their ability to complement the gsal-1 phenotype through YPD-YND-YPD non-selection and selection media growth procedure. Those gsal-1 cells stably integrated with pDLT1 and pDLT-BB efficiently degraded AOX during glucose adaptation (Fig. 4-3 and 4-4) in a manner comparable to GS115. This result showed that 5.7 kb insert containing pDLT1 could restore AOX degradation ability of gsal-1 cells.



Rescue Study of Different Fraqments in the Insert of pDLT1



We next tried to further define regions of the 5.7 kb DNA insert in pDLT-1 that was responsible for complementing the defect of gsal-1 (Fig. 4-3 and Fig. 4-4). Restriction fragments of the 5.7 kb genomic DNA were subcloned into pYM8 and pYM4 (Fig. 2-1). Using these constructs, stable transformants of gsal-1 were isolated and their ability to degrade AOX during glucose adaptation examined. Stable integration of the pDLT1 constructs would presumably occur at the locus homologous to the genomic DNA insert, since pYM8 contained only 164 nucleotides of DNA homologous to P. pastoris. A 3.8 kb BamH I fragment (pDLT-BB) rescued the gsal-1 phenotype, while the DNA insert without this 3.8 kb fragment (pDLT-AB) did not complement the mutation. I also examined whether smaller fragments from within this 3.8 kb region could








75

rescue gsal-1 or not. My results showed that all fragments tested such as the

2.3 kb Hind III / BamH I fragment (pDLT-HB) did not rescue gsal-1 and the minimum rescue fragment is 3.8 kb BamH I fragment. Since pDLT-HB and pDLT-XB did not rescue gsa 1-1, the critical rescue part probably lies between BamH 1(644) to Hind III (2260) sites. Once the smallest rescue DNA fragment was identified, the next step was to sequence this 3.8 kb fragment. The result of sequence is shown in next section.



Sequencingq and Sequence Analysis of pDLT1 Insert



The 3.8 kb BamH I fragment was subcloned into pBluescript and

sequenced in both directions by nested deletion methods. The sequence revealed one long open reading frame without a start codon and five short open reading frames encoding proteins of 70 - 115 amino acids (Fig. 4-4). The short open reading frames did not match any amino acid sequences in protein databases accessible to NCBI, but the large open reading frame showed close amino acid homology to the a-subunit of phosphofructokinase (Pfkl p) of S. cerevisiae and Kluyveromyces lactis. We next sequenced the initial 644 nucleotides of the genomic DNA insert of pDLT-1. The start codon of PFK1 was identified in addition to 161 base pairs of non-coding 5' upstream DNA. No additional open reading frames were revealed.



















CA
S100
o 90
80

.- 70
60 50
� 40
30 20 10 0
O GS115 WDY2


C6




Figure 4-3. pDLT1 rescues the gsal-1 phenotype. pDLT1 was recovered by Dr. Dan Tuttle by transforming a genomic DNA library of P. pastoris and then screened the rescue colonies that appeared white on the direct colony assay. The recovered plasmid was identified as pDLT1 and it rescued gsal-1 by reintroducing it into gsal-1 cells. A 3.8kb BamH I fragment cut from pDLTI's
5.7 kb insert also rescued the mutant phenotype in vector pYM8 (pDLT-BB) but not pYM4 (pWP-BB). The rescue result of a full length PFK1 gene (pWP-PFK) is also shown here. A Hind III - BamH I fragment (see next figure) did not rescue gsal-1. AOX activities were tested in a six-hour glucose adaptation. The values of AOX activities were represented as a percentage of 0 hour values. All assays were measured at least three times and the values shown on the graph represent the average + sd.












BamH I
(644)
1


Spe I (1598)
1


Hind III (2260)
1


Xho I (3191) Spe I (3206)
1


(162) PFK 1 (3134)


Vector pDLT-1 pDLT-AB

pDLT-BB pDLT-HB pDLT-AS pDLT-SS
pDLT-XB


pWP-PFK pWP-BB pWP-HB


BamH I (4512)


5.7 kb





Rescues WDY2 Yes


Figure 4-4. Identification of the genomic DNA that rescues gsal-1 during glucose adaptation. Genomic DNA fragments were subcloned into pYM8 (pDLT) or pYM4 (pWP) vectors and used to stably transform gsal-1. GS115, gsal-1 and transformed gsal-1 cells were grown on methanol for 24-36 hours then adapted to 2% glucose medium. Cell free extracts were prepared at 0 and
6 hours of adaptation and assayed for alcohol oxidase (AOX) activities. Glucose-induce degradation of AOX in those gsal-1 cells that had been rescued by complementing DNA was rapid and comparable to parental GS115 cells (e.g., pDLT-BB). However, when gsal-1 cells were transformed with noncomplementing DNA, AOX degradation was slow and comparable to nontransformed gsal-1 mutants (e.g., pDLT-HB).










The smallest complementing DNA had five short putative open reading frames with start codons. Three of these open reading frames resided within pDLT-HB while two resided in pDLT-SS. Neither construct was capable of rescuing gsal-1 (Fig. 4-3). Since we were stably integrating the genomic DNA prior to testing its ability to complement gsal-1, the smallest complementing DNA does not necessarily represent the entire open reading frame of the rescue gene in gsal-1. Indeed, we suggest that the 3.8 kb BamH I fragment complemented gsal-1 by correcting the mutation upon homologous recombination into the PFK1 locus.

In a parallel experiment, I also integrated the complementing DNA

fragments into the HIS4 locus of gsal-1 and tested their ability to rescue gsal1. This was done by constructing the 2.3 kb Hind Ill / BamH I fragment, the 3.8 kb BamH I fragment and the complete PFK1 gene including 161 nucleotides of 5' upstream DNA into pYM4, which contains a P. pastoris HIS4 gene. Prior to transformation these vectors were linearized by a single cut within the HIS4 gene to direct their integration to the HIS4 locus. Under these conditions, neither pWP-HB nor pWP-BB rescued gsal-1 (Fig. 4-4). We propose that the inability of pWP-BB to complement was due to the absence of the PFK1 start codon when the vector integrated into the HIS4 locus. Indeed, when the start codon was present, PFK1 (pWP-PFK) complemented the gsal-1 phenotype (Fig. 4-4). This experiment showed that PFK is responsible for the rescue effect of gsal-1. Our morphological studies of mutants and their transformants








79

showed that PFK restored the ability of gsal-1 cells to degrade peroxisome (Fig 5-5 and Fig 5-6). The results that Pfkl1p might be Gsalp was a surprise for me since PFKp has identified function in glycolysis. Before I could reach the conclusion that Pfklp is Gsalp, I needed more data about gsal-1.



Northern Blot Analysis of qsal-1 Mutants



At the time of identifying genomic DNA that rescues gsal-1, I also tried to use Northern analysis to 1) identify possible open reading frames that might transcribe in this 3.8 kb region and 2) observe the message RNA (mRNA) level of PFK1 in different yeast strains. A 2.5 kb BamH I / Xho I DNA fragment inside the pDLT insert was used as a probe to identify on Northern blots those mRNA's of GS115, WDY1(gsa2) and WDY2 (gsal-1) that were being transcribed in situ. One predominant mRNA at 3.2 kb and two minor mRNA's at

2.7 and 1.5 kb were identified in GS115 cells (Fig. 4-5). The 3.2 kb mRNA matched the predicted size for PFK1 mRNA. Neither 2.7 kb nor 1.5 kb mRNA corresponded with any predicted open reading frames within the PFK1 locus (Fig. 4-4). We believed they bound to the membrane nonspecifically. The 3.2 kb mRNA was also present in gsa2, but was greatly diminished in gsal-1 cells. These results could not be attributed to sample loading since the amounts of 18S and 28S rRNA loaded onto the lanes were comparable. It is possible that the 3.2 kb mRNA was neither transcribed nor stable in gsal-1 cells. A mutation









80

within the promoter region may suppress transcription, while a premature stop codon would terminate translation causing a destabilization of the mRNA. Our studies suggested that the mutation was not within the upstream 5' region but within the PFK1 encoding region, since pDLT-BB which lacked bases -161 through +483 of the PFK1 gene was able to rescue gsal-1 when stably integrated (see Fig. 4-4). This experiment showed that PFKI is the only transcribing gene in this region and PFK mRNA was greatly reduced in gsal-1 cells and the further evidence that PFK1 gene is related to gsal-1 mutation.



Phosphofructokinase Activity Assay in qsal



The absence of the 3.2 kb PFK1 mRNA in gsal-1 led me to measure the phosphofructokinase activities in these cells (see Materials and Methods). We found that WDY2 (gsa 1-1) cells had less than 15% of the phosphofructokinase activity present in parental GS115 cells whereas WDY1 (gsa2) cells had comparable PFK activity level to that of GS115 (Fig. 4-6). The data were consistent with GSA 1 encodes Pfklp and gsal-1 is a pfkl mutant.

Phosphofructokinase is a heterooctamer composed of four a-subunits encoded by PFK1 and four 3-subunits encoded by PFK2 in yeast cells (Berger and Evans, 1992, Arvanitidis and Heinisch, 1994). The P. pastoris PFK1 gene is 2.97 kb long encoding a protein of 989 amino acids with a predicted molecular mass of 108.7 kDa (Fig. 4-7). The PFK protein catalyzes the


















PFK mRNA











+- 28S rRNA

18S rRNA


Figure 4-5. Northern blot analyses of GS115, WDY1 (gsa2), and WDY2 (gsal-1) cells. Cells were grown in YPD medium, harvested and total RNA prepared. The RNA was separated on an agarose gel and transferred to a nylon membrane. After staining with methylene blue to identify rRNA (lower panel), the membrane was probed with a 32P-labeled 2.5 kb BamH I / Xho I fragment from the original 5.7 kb genomic DNA (see Fig. 4-4). This fragment spans 85% of the PFK1 open reading frame.















Phosphofructokinase Activity in gsa Mutants


500 400 300 < 200

N
W 100


0 +-


GS115 gsa2 gsal-1 gsal-2 gsal-3



Figure 4-6. PFK activity is greatly reduced in WDY2 (gsal-1). Since the sequence data and Northern blot showed that PFK1 might be GSA 1, we measured the activity of PFK protein. The parental GS115 and gsa2 showed normal PFK activity while the activity of PFK protein in gsal-1 reduced greatly (less than 10%). This clearly showed that we recover the right gene. The PFK activities of gsal-2 and gsal-3 are also included. Interestingly, the activity levels in these three gsal mutants are totally different. All assays are measured at least three times and the values shown in the graph represent the average � sd.








83

conversion of fructose 6-phosphate to fructose 1,6-bisphosphate. The enzyme controls one of the rate limiting steps in glycolysis and is highly regulated by many metabolites including fructose 2,6-bisphosphate, AMP and citrate. The catalytic and regulatory sites of the Pfkl protein are highly conserved between P. pastoris, S. cerevisiae and H. sapiens (Fig. 4-8). For example, there exists 60-85% amino acid identity within the fructose 6-phosphate binding site, the ATP binding site, the fructose 2,6-bisphosphate binding site, the AMP binding site, and the citrate binding site (Arvanitidis and Heinisch, 1994). Interestingly WDY3 (gsal-2) and WDY8 (gsal-3) previously identified as the same gsa group of WDY2 (gsal-1) also showed decrease of PFK activity. This also verified that they belong to the same gsal complementation group. However, gsal-2 showed only a two-thirds decrease of the activity when compared with GS115 cells and in gsal-3, about 20% activity remained. I speculate that different activity levels of PFK reflected different missense or nonsense mutation sites in these mutants. For example, in gsal-2 and gsal-3, the mutation may well be missense but at important positions such as a substrate binding site or an ATP binding sites. If PFK protein in P. pastoris is also composed of a- and -subunits as in S. cerevisiae, the mutations in gsal-2 and gsal-3 could also be in PFK2 gene. These strains are very useful for future studies. Based on the ability of genomic DNA fragments to rescue the gsal-1 cells and the absence of PFK1 mRNA, we speculate that the gsal-1 mutant has a premature stop codon near the N-terminus of the PFK1 gene. This would








84

result in the synthesis of a truncated mRNA which may lead to the instability of PFK1 mRNA as we observed on our PFK mRNA study. The translation could also be prematurely stopped.

Carbohydrate metabolism is complicated in yeast and some of the

pathways are still not fully understood because different yeast strains show different pathways and abilities to metabolize sugars (Wills, 1990, Wills, 1996). However, evidence has accumulated that S. cerevisiae can grow on glucose in pfk mutants (Lobo and Maitra, 1983, Schmitt, et al, 1984). Glucose is metabolized via the pentose phosphate pathway in pfk mutants (Schmitt, et al, 1984, Heinisch and Zimmermann, 1985, Jacoby et al, 1993, Boles, et al, 1993). This pathway then joins the glycolysis at glyceraldehyde-3-phosphate level (Wills, 1990). This may also be the case in P. pastoris. Sibirny et al. (1987) found that when Pichia pinus cells were grown on methanol, the enzymes for methanol utilization were greatly induced. The subsequent glucose induced catabolite repression of transcription of methanol metabolism enzymes is controlled by the glucose catabolite repression gene 1 (GCR1). By using glycolysis enzyme assay, they found that the gcrl mutant has low level of PFK activity. They have evidence that in gcrl mutant, the loss of repression was not due to the damage of the glycolysis pathway. However, no follow-up studies of this gcrl mutant can be found. This was an indirect evidence that Pfkl1p might participate in glucose repression of AOX protein via autophagy.















GGATCTTTCTTCTCTTGCTATAAATCAAACATTATTCATACAGAGTTTAATCGA TTCAACAAASCATAACATTGATTGCAATTGGTTCCGTCTTGAACTGCTAAGGAG AACAATCATAAATTAGTATTTTGTTTGCTTGTTAGACTCAAATCGAATTACAGA
M
TGCCAGAACCATCTATAAGTGCACTTTCCTTCACTTCGTTTGTCACTAATGATG
P E P S I S A L S F T S F V T N D D ACAAACTGTTTGAAGAGACTTTCAATTTTTACACGAAGTTGGGCTTCCACGCAA
K L F E E T F N F Y TK L G F H A T CACGCTCATATGTTAAAGACAACCGGTCAGACTTTGAATTGACGGGGATTTCCA
RSY V K DNRS D F E L T GIS T CGGATTCAATCAAGGAAATCTGGCTGGAAAGTTTCCCACTATCTGAAGTGGTCG
D SI K El N L ES F P L SE V E AAACGTCAGCTGGTAGAGAGTTGAGAAAACCACTGCAAGAATCTGTGGGCTACC
T S A G R E L RK PL 0 ES V G Y 0 AATCTGAASCTCTTCTGGGATATTCTCCCTACCAGAGTGACGGTGTTGTTATAA
S E A L L G Y SP Y Q S D G V VI K AATTAAGGTTATCAAATCATGACCTTCAGAAAAACAAAGACTTGCCCGGTGAAG
L R L S N H D L 0 K N K D L P G E V TTACGTTTTTCACCGCTAGTATCGACAAATTAAGASCTAAACTCATTGAAATTG
T F F T A S 0 K L R A K L I E I G GTGCTGAGATAATTCCCTCAGAAATAGACCTTGTTGAATTTTCAACCAAGGATC
AE I I P SE I D L VE F S T K D P CTATGGGCGACGTCATTAGCTTTTCTTCTTATCCCTCTTTGAGTTCCAAGAAGA
M G D V I S F S S Y P S L S S K K K AGATTACCTCTCCAGACTTTTTCCTCCACCCTAAGAAGGAAGTACGCTCCCAAG
I T S P D F F L H P K K E V R S Q E AATCAATAGTTGAGCAGGTTAAATCTGAAGAAGGTAAGAAGAAGATTGCCATCA
SI V E 0 V KS E E G K K K IA I I TAACTTCAGGTGGAGACGCACCGGGAATGAATGCTGCAGTAAGGGCTGTGACAA
T S G G D A P G M N A A V R A V T R GAGCCGGTATTTTCTATGGCTGTAAAGTTTACGCTTGTTATGAAGGTTACACTG
A G I FY G C K V Y A C Y E G Y T G GACTGGTTAAGGGTGGTGATATGTTAAAGGAACTGCAGTGGCAAGATGTCCGTG
L V K G G 0D M L K E L Q W Q D V R G GTTTACTTTCCATTGGTGGTACCATAATTGGTACTGCAAGAAGTAAGGAATTCA
L L S I G G T I I G T A R S K E F R GAGAACGATGGGGCCGTCTTCAAGCTTGCTACAATATGGTCAGCAATGGTATTG
E R W G R L A C Y N M V S N G ID ATGCGTTAGTTGTTTGTGGAGGTGACGGATCTCTTACAGGTGCCGATCTATTTC
A L V V C G G D G S L T G A 0 L F R GAAATGAATGGCCTGAACTGATAAAGGAACTTTTGGGTGAGGGCAAAATTACAA
N E W P E L I K E L L G E G K I T K AAGAACAATATGAAACACACAGAAACTTGACAATCGTAGGTCTCGTTGGTTCTA
E 0 Y E T H R N L T I V G L V G S I TCGATAACGATATGTGCGGAACTGATTCCACAATTGGTGCTTATTCCTCATTGG
D N D M C G T D S T I G A Y S S L E AGAGAATCATAGASCTGGTAGACTACATCGATGCTACTGCCGCCTCCCATTCAC
R I I E L V 0 Y I D A T A A S H S R GAGCCTTCGTGGTGGAAGTCATGGGTAGACATTGTGGATGGTTAGGTTTAATGT
A F V V E V M G R H C G W L G L M S CCGGAATTGCTACTGGAGCTGATTACATTTTCATCCCTGAAAGACCTCCAAGTG
G I A T G A D Y I F I P E R P P S E AAACAAACTGGAAGGACGACTTGAAGAAAGTCTGTTTGAGACATAGAGAGAAAG
T N N K D D L K K V C L R H R E K G GACGCAGGAAGACCACCGTTATTGTTGCTGAAGGTGCTATTGATGACCAACTGA
R R K T T V I V A E G A I 0 0 0 L N ACCCTATCACTTCTGAAGAGGTGAAAGATGTACTAGTGGAGATTGGTTTGGACA
P I T S E E V K D V L V E I G L 0 T


85

CTCGTATTACCCGTCTAGGACATGTCCAAAGAGGTGGAGCTCCGTGTGCTTTTG R I T R L G H V 0 R G G A P C A F D ATAGATTCTTGGCCACTGTTCAAGGTGTTGATGCTGTTAGGGCTGTTTTAGAAA
R F L A T V 0 G V D A V R A V L E S GCACCCCAGCAATTCCTTCTCCTGTCATCAGCATTTTGGAGAACAAAATTGTTC T PA I P S P V I S I L E N K I V R GCCAGCCGTTGGTGGAATCTGTTGCTCAAACAAAGACTGTCAGTGATGCTATCG
0 PL V E S VA 0 T K T V S D A I E AGGCCAAGGATTTCGATAAASCTTTGAAATTAAGAGACCAAGATTTGCCACAT
AK 0 F D K A L K L R 0 0 E F A T S CATATGAGAGCTTCCTGTCCGTTTCCAAGTATGACGATGGATCATATCTAGTAC
Y E S F L S V S K Y D D G S Y L V P CAGAGAGCTCAAGATTAAATATTGCCATCATCCATGTGGGAGCTCCAACATCTG
E S S R LN I A I I H V G A P T S A CGTTGAATCCTGCCACAAGAGTTGCTACTTTGAACTCGTTGGCAAAAGGACACA
LNP A T R V A T L N S L A K G HR GAGTTTTTGCTATTCGAAACGGATTGCAGGATTAATTCGCCACGGCGCTGTAC
V F A I RN G F A G LI RH GA V R GAGAGCTCAACTGGATTGATGTTGAGGACTGGCACAACACAGGTGGGTCGGAGA
E L N W D VE 0 W H NT G G S E I TTGGCACCAACAGAAGTCTTCCTAGTGATGATATGGGCACTGCGGCGTACTACT
G T N R S L P S 0 D M G T A A Y Y F TCCAGCAATACAAGTTTGATGGTCTTATTATTATCGGNGGATTTGAAGCTTTCA
0 0 Y K F D G L III I GG F E A F T CAGCTCTGTACCAGCTGGACGCAGCTCGCGCTCAGCATCCTATCTTCAATATTC
ALY 0 L D A A R A 0 H P I F N I P CAATGTGTTGNCTTCCAGCTACTGTTTCTAATAACGTTCCTGGTACCGAGTATT
MC X L P A T V S N N V P G T E Y S CCTTAGGGTCTGACACATGTCTAAACACCTTGTCTGGATACTGTGATGCTGTGA
L G S D TC L N T L S G Y C D A V K AACAATCTGCTTCTCTTAGTAGAAGAAGAACATTTGTTGTGGAAGTTCAAGGTG
a S A S A S R R R T F V VE V 0 G G GATACTCAGGATATCTTGCCAGCTACGCTGGTCTGATCACAGGAGCTTTGGCTG
Y SG Y L A S Y A G L I T G A L A V TTTATACTCCTGAAAACCCAATCAACCTTCAAACAGTGCAGGAAGACATTGAAT
Y T P E N P I N L 0 T V 0 ED I EL TGTTGACTCGAACATACGAGGAAGACGATGGTAAGAACAGATCGGGTAAAATCT
L T R T Y E E D D G K N R S G K I F TTATTCATAATGAAAAGGCTTCAAAGGTTTACACCACGGATCTGATTGCTGCTA
I H NE KASK V Y T T D L I A A I TCATAGGTGAAGCTGGAAAGGGTACGTTTGAGAGCCGTACTGCCGTGCCTGGTC
I G E A G K G T F E S R T A V P G H ATGTACAACAGGGTAAATCTCCCTCATCTATTGACCGGGTTAATGCCTGCAGAC
V 0 G K S P S S I D R V N A C R L TGGCTATCAAATGTTGTAACTTCATCGAGGACGCCAATTTCCAGGTGAAACACA
Al I K C C N F I E 0 AN F 0 V K H N ATGCCAATTTGAGCGCCSACGAACGTCATTTGAGATTCTTTTGCGATGACGGAG
ANL S ADE RH L R F F C D D G V TTAAGACATCTGCAGTGAGCGGCAATCTTCCGTGATAGATGATAACACGTCAG
K T S A V S G K S S V I D D N T S V TGGTCATTGGAATCCAAGGTTCCGAGGTTACATTCACTCCTGTAAAACAGCTAT
VI G IQ G S E V T F T P V K 0 L W GGGAGAAGGAAACTCATCATAAGTGGCGAAAGGGTAAGAACGTTCATTGGGAGC
E K E T H H K W R K G K NY V H W E 0 AGTTGAACATTGTCTCTGACCTCTTGAGTGGTCGTTTGTCTATTCGTACCACGT
LN IV S D L L S G R LS I R T T * AAAAGACGGATCAAATCGGTTGTTTGGGTACTAAAGACAATCCATTTTTTTTCT TNCTCTCGAGCTGGATGAAACTAGTGCATGTACGAATCCGCGTGTAATCTACTS


Figure 4-7. Nucleotide and predicted amino acid sequences of P. pastoris PFK1 gene. The 5' upstream region in the original 5.7 kb clone included 161 bases. The Pfkl protein is predicted to be 989 amino acids with a molecular mass of 108.7 kD. The sequence of PFK1 has a GenBank accession number of U73376.















PP MPEPSISALSFTSFV..TNDDKLFEETFNFYTKLGFHATRSYVKDNRSDFELTGIST..D 56 SC -QSQDSCYGVAFRSIITNDEALFKKTIHFYH-LGFATVKDFNKFKHGENSL-SSGTSQDS 60 HU -DADDSR-PKGSLRKFLEHLSGAGKAIGVL.............................. 30
PP SIKEIWLESFPLSEWETSAGRELRKPLQESVGYQSEALLGYSPYQSDGWIKLRLSNHD 116 SC LREVWLESFKLSEVDASGFRIPQQEATNKAQSQGALLKIRLVMSAPI-ETFDTNETATIT 120 HU ............................................................ 30

PP LQKNKDLPGEVTFFTASIDKLRAKLIEIGAEIIPSEIDLVEFSTKDPMGDVISFSSYPSL 176 SC YFSTDLNKIVEK-PKQAEKLSDTLVFL.................-.... NN-T--GLANA 163
HU ............................................................ 30

PP SSKKKITSPDFFLHPKKEVRSQESIVEQVKSEEG.......KKKIAIITSGGDAPGMNAA 229 SC TDSAPTSKDA-LEATSEDEIISRASSDASDLLRQTLGSSQK----- VM -----S------ 223
HU ................................................ .....Q.....----- 42
PP VRAVTRAGIFYGCKVYACYEGYTGLVKGGDMLKELQWQDVRGLLSIGGTIIGTARSKEFR 289 SC ----V-T--HF--D-F-V----E--LR--KY--KMA-E----W--E---L ------ M--- 283
HU ----V-M--YV-A---FI ----Q-M-D--SNIA-AD-ES-SSI-QV------S--CQA-- 102
PP ERWGRLQACYNMVSNGIDALWVVCGGDGSLTGADLFRNEWPELIKELLGEGKITKEQYETH 349 SC K-E--R--AG-LI-Q--------------------- H--S-VD--VA--RF---EVAPY 343
HU T-E---K-AC-LLQR--TN-C-I---------N---K--SG-LE--ARN-Q-D--AVQKY 162
ATY
PP RNLTIVGLVGSIDNDMCGTDSTIGAYSSLERIIELVDYIDATAASHSRAFWEVMGRHCG 409 SC K--S -----------S---------A--A --C-M -------- K ----------------- 403
HU AY-NV--M----- F ---- M---TD-A-H ----V--A-MT--Q--Q-T--L-------- 222
F-S-r
PP WLGLMSGIATGADYIFIPERPPSETNWKDDLKKVCLRHREKGRRKTTVIVAEGAIDDQLN 469 SC --A--A------------- AVPHGK-Q-E--E--Q---S ----NN-I ------L----- 463
HU y-A-V-AL-C---WV-L--S--E-GWEEQMCV-LSENRAR-K-LNIIIVAEGAIDTQNKP 282
PP PITSEEVKDVLVEIGLDTRITRLGHVQRGGAPCAFDRFLATVQGVDAVRAVLESTPAIPS 529 SC -V-AND---A-I-L----KV-I --------TAV-H--W---L---- K---- F--ET-- 523
HU ITSEKIKEL-VTQL-Y---V-I--------T-S----I --SRM--E--I-L--A--DT-A 342

PP PVISILENKIVRQPLVESVAQTKTVSDAIEAKDFDKALKLREQEFATSYESFLSVSKYDD 589 SC -L-G------I-M ------KL--S-AT---N ------IS--DT--IEL--N---TTVK-- 583
HU C-V-LNG-HA--L--M-C-QM-QD-QK-MDERR-QD-VR--GRS--GNLNTYKRLAIKLP 402

PP GSYLVPESSRLNIAIIHVGAPTSALNPATRVATLNSLAKGHRVFAIRNGFAGLIRHGAVR 649 SC --E-L-V-D----G-V-----SA ---A---A---YC-SH--KPY--M---S---QT-E-K 643
HU DDQIPKTNCNVAVIN..----AAGM-A-V-S-VRVGI-D---ML--YD--D-FAKGQIKE 460

PP ELNWIDVEDWHNTGGSEIGTNRSLPSDDMGTVAYYFQQYKFDGLIIIGGFEAFTALYQLD 709 SC --S-----N---L--------- VA-E-L--I-----KN-L--- L ---- G-RS-K--R 703
HU IGWT.--GG-TGQ---IL--K-V--GKYLEEI-TQMRTHSINA-L-------YLG-LE-S 519

PP AARAQYPIFNIPMCCLPATVSNNVPGTEYSLGSDTCLNTLSGYCDAVKQSASASRRRTFV 769 SC DG-T-H --------LI ---------------V -----A-VN-T-DI ------ T---V-- 763
HU ---EKHEE-CV--VNV--------- SDF-I-A--A---ITDT--RI -----GTK--V-I 579
r-2,O- L itrate PP VEVQGGYSGYLASYAGLITGALAVYTPENPINLQTVQEDIELLTRTYEEDDGKNRSGKIF 829 SC C----- H---I--FT------VS-----KK-D-ASIR---T--KENFRH-K-E--N--LL 823
HU I-TM---C ----NMG--AA--D-A-IF-E-FDIRDL-SNV-H--EKMKTTIQRGLVLR.. 637

PP IHNEKASKVYTTDLIAAIIGEAGKGRFESRTAVPGHVQQGKSPSSIDRVNACRLAIKCCN 889 SC VR--Q--S--S-Q-L-D--S--S--K-GV---I -------GV---K---T-S-F-V--IK 883
HU ..--SC-EN ---- F-YQLYS-E---V-DC-KN-L--M---GA--PF--NFGTKISARAME 695

PP FIEDANFQVKHNANLSADERHLRFFCDDGVKTSAVSGKSSVIDDNTSWIGIQGSEVTFT 949 SC ---QW-KKNEASP-TD-KVLRFK-DTHGEKVPTVEHEDD-AA...... --CVN--H-S-K 937
HU W-TAKLKEARGRGKKFTTDDSICVLGISKRNVIFQPVAELKKQTDFEHR-PKEQWWLKLR 755
PP PVKQLWEKETHHKWRKGKNVHWEQLNIVSDLLSGRLSIRTT 990 SC -IAN---N--NVEL---FE---AEY-KIG-I-----KL-AEVAALAAENK 987
HU -LMKILA-YKASYDVSDSGQLEHVQPWSV 784


Figure 4-8. Comparison of PFK1 genes. The sequence of Pfkl1p (PP)
was aligned to those PFK1 sequences of S. cerevisiae (SC) and H.
sapiens (HU). Gaps are represented by "dots" (....) and amino acid
identity represented by "dashed lines" (--). Conserved catalytic and
regulatory domains sites are also included (Arvanitidis and Heinisch,
1994).










Glycolysis Pathway and gsa Mutants



Glucose signals the onset of peroxisome degradation in the glucose

adaptation pathway. It may either directly send a signal to initiate the onset of microautophagy of peroxisomal and cytosolic proteins or exert that effect through a subsequent glycolysis enzyme or product. The discovery that GSA1 is PFK1 prompted me to probe this question in gsal-1 as well other gsa mutants. I want to know If glucose is the direct signaling molecule and if other carbon sources could signal the initiation of peroxisome degradation. We asked ourselves which entry point in the glycolysis pathway is related to the autophagy signaling process and if this is related to Pfkl function in the glycolysis pathway. By providing different carbon sources for yeast to grow, we found that P. pastoris could utilize fructose, mannose, glycerol and pyruvate as the sole carbon sources for growth. However, when these carbon sugars were used as adaptation carbon source, we observe pyruvate could not initiate peroxisomal AOX degradation in parental GS115 strain while fructose, mannose and glycerol could do the job as efficiently as glucose in GS115 (Fig 4-9 and Fig 4-10). Pyruvate, fructose and mannose could not initiate microautophagy in gsal-1 and gsa2 mutants while glycerol did initiate the degradation process in GS115 cells as well as in gsal and gsa2 mutants. In a PFK1 knockout, a null mutant of gsal-1 I isolated later in the experiment also could undergo glycerol induced autophagy (see chapter 5).




Full Text

PAGE 1

MOLECULAR CHARACTERIZATION OF AUTOPHAGY IN METHYLOTROPHIC YEAST PICHIA PASTORIS By WEIPING YUAN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1998

PAGE 2

ACKNOWLEDGMENTS I would like to thank my mentor, Dr. William A. Dunn, Jr., for his excellent guidance and training in my 5-year graduate study. I would also like to extend my gratitude to my committee members, Drs. John P. Aris, Gudrun S. Bennett and Alfred S. Lewin, for their support, help and encouragement of my research. I am in great debt to Dr. Dan Tuttle for his pioneer work in the lab and Dr. Jim Cregg for providing yeast strains, plasmids and the helpful conversations. I would also like to thank the faculty, staff and all other persons at the Department of Anatomy and Cell Biology for the help they rendered and for the stimulating intellectual environment they provided. My special thanks are reserved for my wife Ying Shen for her wholehearted support in these five years. 11

PAGE 3

TABLE OF CONTENTS ACKNOWLEDGMENTS ii LIST OF TABLES vi LIST OF FIGURES vil KEY TO ABBREVIATIONS x ABSTRACT xi CHAPTERS 1 INTRODUCTION AND REVIEW OF THE LITERATURE 1 The Peroxisome and its Characteristics 2 Peroxisome Biogenesis 2 Peroxisome Functions 3 Methanol Metabolism in Pichia Pastoris 5 Autophagy 7 Non-selective Autophagy in Mammalian Cells and Yeast 8 Selective Autophagy in Mammalian Cells and Yeast 1 1 Genetic Analysis of Pichia pastoris 15 Piciiia pastoris as a Study Model for Autophagy 15 Classical Genetics 16 Molecular Biology 16 Previous Research of Peroxisome Degradation in Our Lab 19 Chapter Summary 26 2 MATERIALS AND METHODS 27 Yeast, Bacterial Strains and Media 27 Enzyme Assays 29 Isolation of Glucose-Induced Selective Autophagy-Deficient (gsa) Mutants 31 Mutant Generation and Isolation 31 iii

PAGE 4

Mutant Backcrossing 32 Mutant Complementation Analysis 33 Molecular Biology 34 Yeast Transfomiation 34 Plasmid Isolation and DNA Sequencing 35 Northern, Southern and Western Blot Analysis 37 Isolation of PFK1 and GSA7 Knockouts 38 Site-directed Mutagenesis of PFK1 39 Generation of a HA Epitope Tag in GSA7 and YHR171w gene 40 Fluorescence Microscopy and Electron Microscopy 41 3 CHARACTERIZATION OF PICHIA PASTORIS MUTANTS DEFECTIVE IN GLUCOSE-INDUCED SELECTIVE AUTOPHAGY 43 Introduction 43 Complementation Group Identification 45 Screening for gsa Mutants 48 Glucose Induced Microautophagy is Defective in gsa Mutants 49 Morphological Studies of gsa Mutants 57 Chapter Summary 64 4 GSA1 PROTEIN IS PPF1 PROTEIN 68 Introduction 68 Gsa1-1 Signaling an Upstream Event of Microautophagy 68 Microautophagy of Peroxisomes is Defective in gsa1-1 68 Morphological Study Revealed That a Step Before Peroxisome Sequestration Is Blocked in gsa1-1 70 pDLTI Complemented gsa f--/ Phenotype 73 Rescue Study of Different Fragments in the Insert of pDLT1 74 Sequencing and Sequence Analysis of pDLT1 Insert 75 Northern Blot Analysis of gsa1-1 Mutant 79 Phosphofructokinase Activity Assay in gsa1 80 Glycolysis Pathway and gsa Mutants 87 Glycolysis and gsa Mutants 91 Chapter Summary 93 5 PFK1 PROTEIN IS REQUIRED FOR THE INITIATION OF PEROXISOME MICROAUTOPHAGY 96 Introduction 96 Iv

PAGE 5

Degradation of Peroxisomes by Microautophagy Requires PFK1 97 PFK1 Gene Disruption and the Verification 97 The PFK1 Knockout Has the Same Phenotype as gsa1-1.... 97 Verification by Site Directed Mutagenesis of P. pastoris PFK1 Gene of the Distinction between Rescue Function and PFKp Activity 100 Normal Pfk1p and Catalytically-inactive pfk1p Complement ^p/?cf 100 Restoration of the Degradation Ability of Apfk1 by PFK1 1 04 Morphological Characterization oi Apfk1 106 Chapter Summary 109 6 CHARACTERIZATION OF GSA7 MUTANTS Ill Introduction Ill Morphological Studies of gsaJ 1 1 1 Recovery of GSA 7 Gene and its Verification 112 pYWP7-4 Rescues gsa? phenotype 1 1 2 Identification of GSA7. 1 1 4 Sequence Analysis of GSA7 116 Agsa7 Generation and its Phenotype Studies 119 A HA Tagged Gsa7p Rescues P. pastoris 1 20 Chapter Summary 1 22 7 CONCLUSIONS AND PROSPECTS 127 Introduction 127 Characterization of gsa Mutants 131 Regulation of the Signaling of Microautophagy by Gsal p 1 32 Regulation of a Homotypic Vacuolar Membrane Fusion Event of Microautophagy by Gsa7p 1 35 Prospects and Conclusion 137 REFERENCES 141 BIOGRAPHICAL SKETCH 151 V

PAGE 6

LIST OF TABLES Tables Page 21 Parental and mutant strains of Pichia pastoris 28 31 Biochemical profiles of gsa mutants 53 vi

PAGE 7

LIST OF FIGURES Figure Page 11 Compartmentalization of the pathways involved in methanol metabolism in methylotrophic yeast 6 1 -2 Loss of peroxisomal and cytosolic enzymes during ethanol and glucose adaptation 20 1 -3 Morphology characterization of peroxisomes during glucose and ethanol adaptation 21 1 -4 Model of glucose-induced microautophagy pathway 24 21 Maps of plasmids pYM8 and pYM4 36 31 A flow chart of mutants characterization 44 3-2 Identification of gsa mutants complementation groups 47 3-3 AOX and FDH activities of gsa mutants under glucose adaptation 51 3-4 AOX and FDH activities of gsa mutants under glucose adaptation 52 3-5 AOX degradation and histidine addition during glucose adaptation... 56 3-6 Ultrastructural studies o1 gsa1-1, gsa2, gsa3, gsa6, gsaZand gsa8 mutants under glucose adaptation 62 3-7 Ultrastructural studies of gsa4 and gsa5 mutants under glucose adaptation 63 38 Microautophagy of peroxisomes and gsa mutants in P. pastoris 65 41 Glucose and ethanol adaptation in parental GS1 15 and WDY2 {gsa1-1) 69 vii

PAGE 8

4-2 Morphology of gsa1-1 and gsa2 under glucose adaptation at Oh and 3h time points 72 4-3 pDLT1 rescues gsa1-1 phenotype 76 4-4 Identification of genomic DNA that rescues gsa1-1 during glucose adaptation 77 4-5 Northern blot analyses of GS1 15, WDY1 (gsa1) and WDY2 (gsa1-1) 81 4-6 PFK activity is greatly reduced in WDY2 {gsa1-1) cells 82 4-7 Nucleotide and predicted amino acid sequence of P. pastoris PFK1 gene 85 4-8 Comparison of PFK1 genes 86 4-9 Glycolysis pathway and the entry point of carbon metabolites 89 4-1 0 Metabolites in the glycolysis pathway and their induction of the degradation of AOX 90 41 1 Glycolysis enzyme activities and gsa mutants 92 51 Generation of the PFKf knockout 98 5-2 Glucose-induced degradation of AOX and FDH in PPF1 and WDK01 ..99 5-3 Partial sequences comparison of Pfkl p in P. pastoris (PP), S. cerevisiae (SC) and Human sapiens (HU) 100 5-4 Site-directed mutagenesis of P. pastoris PFK1 gene 1 02 5-5 Morphology of GS1 1 5, gsa1-1, and Apfk1 during glucose adaptation 1 07 56 Morphology of WDK01 (Apfkl) and its transformants during glucose adaptation 108 61 Morphology of gsa3 and gsa7 mutants during glucose adaptation 113 6-2 Verification of GS/ii 7. 115 6-3 Complete nucleotide and amino acid sequence of GSA7. 1 1 7 viii

PAGE 9

6-4 A protein sequence comparison of GSA7 in P. pastoris, S. cerevisiae, S. pombe and H. sapiens 118 6-5 idgsaZ generation and its verification 121 66 Detection of a HA epitope tagged Gsa7p and Yhr171wp in transformed gsa7 1 23 71 Pathways of peroxisome induction and degradation under glucose and ethanoi adaptation in methylotrophic yeast Pichia pastoris 129 7-2 Microautophagy and macroautophagy of peroxisomes in P. pastoris 1 30 7-3 Putative substrate binding and regulatory sites in Pfkl p 1 34 7-4 Pathway of glucose induced peroxisome microautophagy and its relationship with gsa mutants 1 39 ix

PAGE 10

KEY TO ABBREVIATIONS AOX: alcohol oxidase BSA: bovine serum albumin CCO: cytochrome c oxidase DSM: diploid selection medium EAM: ethanol adaptation medium F1[3: mitochondrial F1 ATPase, p subunit FBP: fructose1 ,6bisphosphatase FDH: formate dehydrogenase GAM: glucose adaptation medium gsa: glucose-induced selective autophagy deficient mutant Gsap: Gsa protein GSA: GSA gene Agsa1: gsa1 knockout strain Agsa7: gsaJ knockout strain mRNA: message RNA MIM: methanol induction medium ORF: open reading frame PFK1: phosphofructokinase 1 gene Pfk1p: phosphofructokinase 1 protein pfk1: mutated Pfk1 protein pfk1\ mutated PFK1 gene SM: sporulation medium TM: transformation medium YND: yeast nitrogen base medium with glucose YNM: yeast nitrogen base medium with methanol YPD: glucose-containing complete medium X

PAGE 11

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MOLECULAR CHARACTERIZATION OF AUTOPHAGY IN METHYLOTROPHIC YEASTS PICHIA PASTORIS By Weiping Yuan December 1998 Chairman: William A. Dunn, Jr. Major Department: Anatomy and Cell Biology Eukaryotic cells adapt to environmental changes by synthesizing and degrading cellular proteins and organelles. Although much is known regarding the biogenesis of cell organelles, the mechanism of turnover of many organelles and proteins through autophagy remains unclear. Our lab has utilized the yeast P. pastoris as a study model for the autophagy of peroxisomes during metabolic adaptation. Peroxisomes are induced when P. pastoris is grown on methanol. During the adaptation to glucose, peroxisomes are degraded within the yeast vacuole by microautophagy. This process includes glucose signaling, vacuolar recognition and sequestration of peroxisomes, homotypic vacuolar membrane fusion and finally vacuolar degradation of peroxisomes. xi

PAGE 12

In order to better understand the microautophagy process, glucoseinduced selective autophagy-deficient (gsa) mutants were generated and isolated by their inability to degrade peroxisomal alcohol oxidase (AOX) during glucose adaptation. Eight gsa complementation groups {gsa1 to gsaS) have been identified and they represent different gene products that control microautophagy processes. These gsa mutants fall into four major groups. Gsa1 belongs to microautophagy initiation mutants. Gsa4 and gsaS are possible autophagy recognition mutants. Gsa2, gsa3, gsa6, and gsa? are possible homotypic vacuolar membrane fusion mutants. GsaS is also a homotypic fusion mutant but its macroautophagy pathway has been turned on during the methanol induction. Gsa1-1 was transformed with a Pichia pastoris genomic DNA library. The gene that complemented the gsa1-1 phenotype was identified as phosphofructokinase 1 (PFK1). Cellular levels of both PFK1 mRNA and PFKp activity were greatly reduced in gsa1-1. The inability o\ t^pfkl to degrade AOX could be rescued by either normal PFK1 or mutant pfk1 whose catalytic site had been inactivated by a single amino acid mutation. It suggests that the degradation of peroxisomes does not require a catalytically active PFKp. Pfk1p might regulate microautophagy independent of its PFK activity. WDY7 {gsa?) was the second mutant I chose for further genetic study. I recovered a plasmid pYWP7-4 that complemented gsaJ. The gene was identified as GSA7. A knockout strain AgsaZ failed to degrade AOX confirmed xii

PAGE 13

GSA7 is essential for the microautophagy. Gsa7p might participate in a ubiquitin-like pathway to regulate the vacuole membrane fusion. The study of microautophagy in Pichia pastoris provides us a unique opportunity to look into the mechanisms of protein degradation via autophagy. The study I have done and the tools that have become available in recent years will be useful to further elucidate the molecular mechanism of microautophagy in eukaryotic cells. xiii

PAGE 14

CHAPTER 1 INTRODUCTION AND REVIEW OF THE LITERATURE Some yeast species such as Candida boidinii, Hansenula polymorpha, and Pichia pastoris are able to utilize methanol as the sole carbon and energy source for their growth and thus are called methylotrophic yeasts. These methylotrophic yeasts metabolize the methanol by enzymes in the peroxisomes and cytoplasm whose synthesis is induced under methanol-growth conditions (Hill et al., 1985). These enzymes as well as the whole peroxisomes are rapidly degraded upon a carbon source switch to such as glucose or ethanol. The study of the biogenesis and degradation processes of peroxisomes under different growth conditions in Pichia pastoris and other yeast species has gained great interest in recent years because in humans, an inability to import proteins into the peroxisomes has serious health consequences (Moser and Moser, 1996). Genetic screens for peroxisomal assembly (pas) (Gould et al., 1992), deficiency {per) (van der Klei et al., 1991), and degradation {gsa, ppd) (Tuttle and Dunn, 1995, Titorenko et al., 1995) mutants in yeast have been successfully utilized to identify many genes required for these events. Moreover, several genes cloned in these mutants have been found to have human homologous counterparts (Dodt and Gould, 1996). These findings emphasize the usefulness of the yeast 1

PAGE 15

2 mutant study model and the significance of these studies. In the following sections, I will discuss the existing knowledge relate to peroxisome biogenesis, degradation and autophagy. I will also discuss the methods available to study the mechanism of autophagy in yeast peroxisomes and the previous work that has been done in our lab. The Peroxisome and its Characteristics Peroxisome Biogenesis Peroxisomes are ubiquitous, single-membrane-bound organelles that are involved in many important cellular activities. They were first discovered thirty years ago in a centrifugation fraction of cell components (de Duve, 1965). The peroxisomes are thought to arise by budding or fission from preexisting peroxisomes although evidence for de novo synthesis has been presented. Peroxisomes contain neither DNA nor ribosomes. Their proteins are encoded by nuclear genes and synthesized in cytoplasm on free polyribosomes and then post-translationally imported into the preexisting organellar matrix or surrounding membrane (Liu et al., 1995). Almost all peroxisomal proteins are synthesized at their final size. Two peroxisomal targeting sequences (PTS) responsible for correct delivery of matrix proteins to the organelles have been identified. The majority of peroxisomal matrix proteins contain PTS1, a C-terminal tripeptide of the sequence SKL-COOH (Keller et al., 1991). PTS2, with the consensus

PAGE 16

3 sequence RLX5H/QL is located at N-terminal of peroxisomal proteins (Swinkels et a!., 1 991 ). Many studies focused on the biogenesis of peroxisomes and the defects in peroxisome assembly and their protein import. In P. pastoris, 1 5 complementation groups of peroxisome assembly mutants {PAS mutant) have been identified (Subramani, 1993). Some of the PAS genes have been sequenced. PAS4 gene was identified as a ubiquitin-conjugating enzyme required for peroxisome assembly (Crane et a!., 1994). It is an interesting finding since the ubiquitin mediated protein degradation pathway is responsible for protein turnover. We do not yet know if PAS4 has a role in the peroxisome protein degradation pathway through targeted degradation by the proteasome. McCollum et al. (1993) reported cloning the PAS8 gene and demonstrated that it is a PTS1 import factor. Subsequent studies showed that it is a PTS1 import receptor and the tetratricopeptide repeat domain of PAS8p is identified as the PTS1 binding region (Terlecky et al., 1995). This finding linked peroxisome protein import and targeting and the peroxisome assembly pathway in the peroxisome biogenesis process. Peroxisome Functions Two prominent functions of peroxisomes are the P-oxidation of fatty acids and the oxidation of substrates by different HjOj-generating oxidases (Subramani, 1993). Other functions of peroxisomes include lipid biosynthesis, cholesterol biosynthesis and peroxisomal purine metabolism. Peroxisomes are unique in their functional diversity when compared with other cell organelles

PAGE 17

4 such as mitochondria, the nucleus, and organelles of the secretory pathway. Their functions are dispensable in certain nutritional environments and the specific metabolic pathways found in the organelle vary depending upon the organism, the tissue and its environment. This was verified by observations that mutants defective in peroxisome assembly in yeast and human fibroblast cell lines were still viable (Subramani, 1993). Nevertheless, the peroxisome is essential for normal human development, since defects in peroxisome assembly have been identified as the cause of the peroxisome biogenesis diseases. In methylotrophic yeast P. pastoris, when cells were grown on nutrition rich YPD media, yeast maintained a low number of peroxisomes per cell. The size of the peroxisomes is also small. However, peroxisomes can also be made to proliferate or degrade in response to nutritional cues such as methanol or oleic acid (Liu et al., 1992, Subramani, 1993). The process of peroxisome degradation is not as well understood as biogenesis. Cells respond to nutritional clues to reduce the number of peroxisomes via degradation. Such events are exemplified by supplying methanol as growth carbon source to induce peroxisome production and then abolishing the induction by a carbon source switch such as glucose or ethanol. The whole process usually takes less than eight hours to finish (Tuttle and Dunn, 1995). This degradation event provides us a unique opportunity to investigate the mechanisms of peroxisome degradation.

PAGE 18

5 Methanol Metabolism in Pichia pastoris When methanol is added to the P. pastoris growth media as the only carbon source, methanol readily enters the cell and peroxisome for metabolism. The peroxisomal enzyme alcohol oxidase (AOX) is the first enzyme in the methanol assimilation pathway. AOX is a homooctameric protein of ~ 74 kD, each subunit containing a flavin adenine dinucleotide molecule (FAD) as a prosthetic group (van der Klei et al., 1991). It uses O2 as the electron acceptor, oxidizes methanol and produces hydrogen peroxide and formaldehyde (Gleeson and Sudbery, 1988). The hydrogen peroxide is reduced by peroxisomal catalase to H2O and O2. Formaldehyde exits the peroxisome and is catalyzed in cytosol by formaldehyde dehydrogenase and formate dehydrogenase (FDH) to CO2 providing an energy source in the form of NADHj (Tolbert, 1981). Formaldehyde enters the xylulose 5-phosphate cycle to provide a carbon source for amino acid synthesis (Fig. 1 -1 ). One characteristic of AOX is that it is not detectable in glucose-grown cells but inducible under methanol growth condition. Another is that it has a low affinity for its substrate and thus large quantities of the enzyme (up to 30% of the total cellular soluble proteins) are needed for yeast to grow on methanol as the sole carbon source. When the carbon source for growth is switched from methanol to glucose or ethanol (glucose or ethanol adaptation), yeast peroxisomes and thereby also AOX are degraded rapidly in its vacuole via autophagy (Tuttle et al., 1993). The inactivation of AOX activity is paralleled by

PAGE 19

6 (r PEROXISOME -CH3OH H2O2., ^, © HCHO ^02 + H2O r® GAP DHA CYTOSOL GSH *® HCHO^^-^-GS-CHoOH — / < * HCOOH — / v. * COj NADH NADH Xu5P J DHA ATP ® rearrangement reactions ^3 GAP cell constituents ADP -»• GAP •F3P -F6P J) Figure 1-1 . Compartmentalization of the pathways involved in methanol metabolism in methylotrophic yeast. Methanol is metabolized first in peroxisome and then in the cytosol for further degradation. Its metabolism pathway is shown in the figure above and the enzyme names are listed below. The pathway is responsible for the generation of GAP for the production of biomass and the NADHj for energy generation. 1 = Alcohol oxidase; 2 = catalase; 3 = formaldehyde dehydrogenase; 4 = formate dehydrogenase; 5 = dihydroxyacetone synthase; 6 = dihydroxyacetone kinase; 7 = fructose 1 ,6-bisphosphate aldolase; 8 = fructose 1,6bisphosphate phosphatase (Adapted from van der Klei et al., 1991)

PAGE 20

7 its degradation of peroxisomes in this process (Veenhuis et al., 1983) and the disappearance of AOX activity correlates with the degradation of AOX protein and autophagic loss of peroxisomes. The degradation of peroxisome is proteinase A and B dependent (Tuttle and Dunn, 1995). These characteristics of AOX provide us an excellent biochemical index for studying peroxisome degradation under different nutritional adaptation conditions. It also provides us an easy enzyme assay to screen autophagy mutants. Autophaqy Cells respond to changes in the environment by eliminating cell components that are no longer needed. The amino acids then are released and used as an energy source or for the synthesis of new proteins. This provides the organism with added adaptability that outweighs the metabolic costs such as ATP used in protein degradation pathways (Olson et al., 1992). The continual turnover of cellular proteins also prevents the accumulation of a variety of deleterious nonenzymatic modifications such as oxidation, deamination, glycosylation of the abnormal proteins that are deleterious to the cells (Klionsky, 1997). The degradation of protein in the cell is mainly via lysosomal proteolytic pathways, proteolytic pathways of non-lysosomal organelles and cytosolic proteolytic pathways. In the following section, I will mainly discuss the existing knowledge of autophagy pathways in mammalian cells and yeast cells.

PAGE 21

8 Non-Selective Autophaqy in Mammalian Cells and Yeast Autophagy means self-eating. It is used to refer to the process whereby cytoplasmic constituents are degraded in a vacuoie-dependent manner. It has been recognized as a highly regulated non-selective and sometimes selective process for the degradation of cellular proteins and organelles in eukaryotic cells. Mammalian cells respond to amino acid starvation and other environmental changes (e.g., heat shock) by activating the degradation of proteins via non-selective autophagy (Lardeux and Mortimore, 1987, Mortimore et al., 1989, Kopitz et al., 1990, Dunn et al., 1994). Both microautophagy and macroautophagy have been observed in mammalian cells. Microautophagy is the process whereby cellular components are sequestered directly into the lysosome. Regions of the cytoplasm are surrounded by invagination of the lysosomal membrane or by finger-like protrusions of the lysosome. Upon fusion of the lysosomal membrane, intralysosomal vesicles are formed containing the sequestered cellular components (Ahlberg and Glaumann, 1985, Mortimore et al., 1989). It is believed that multivesicular bodies arise from such autophagic events. Macroautophagy is the process whereby cellular components are sequestered first within an autophagosome (Dunn, 1990a). The autophagosome originates from the rough endoplasmic reticulum in mammalian cells and, then fuses with a lysosome. Once within the lysosome, the proteins are degraded to their monomeric subunits by the concerted action of exopeptidases and endopeptidases (Dunn, 1990a, 1990b).

PAGE 22

9 Many strains of yeast are capable of activating autophagy in response to nutritional changes (Veenhuis et al., 1983, Tuttle et al., 1993, Baba et al., 1994, Tuttle and Dunn, 1995, Chiang et al., 1996). The main organelle for autophagy in yeast is its vacuole, the equivalent of lysosome in mammalian cells. In order to better define the molecular events of autophagy, mutants defective in autophagy have been used to study the mechanism of autophagy in yeast S. cerevisiae, P. pastoris and H. polymorpha. During nitrogen starvation, cells of a normal strain of S. cerevisiae respond to this stress by degrading cellular contents non-specifically via autophagy. Ohsumi's group found in a S. cerevisiae mutant lacking proteinase B, "autophagic bodies" containing cytosolic contents were accumulated in the vacuoles during nitrogen starvation. (Takeshige et al., 1992). They (Tsukada and Ohsumi, 1993) used this strain to generated apg (autophagy) mutants v\/hich failed to accumulate autophagic bodies within the yeast vacuole. Strains with apg mutation failed to deliver celluar contents to the yeast vacuole and are thus the non-selective autophagy mutants. Their ultrastructural data suggested that the autophagic process in yeast is essentially similar to that of the lysosomal system in mammalian cells (Baba et al., 1994). The apg mutants fell into 15 complementation groups. The characterization of these mutants is underway and several novel genes have been identified, sequenced and their functions in the process of autophagy are being studied (Shirahama et al., 1997; Matsuura et al., 1997; Noda and Ohsumi, 1998).

PAGE 23

10 The aut (autophagy) mutants identified by a colony screening procedure by Thumm et al. (1994) are defective in the degradation of cytosolic fatty acid synthase during carbon starvation. This served as a selection marker for rapid isolation of autophagocytosis yeast mutants defective in vacuolar breakdown of fatty acid synthase during starvation. Further identification and characterization of the genes of apg and aut mutants will reveal the mechanisms and function of autophagocytosis. It will also reveal why all autophagy mutants have defects in their sporulation process. It appears that some of these autophagy mutations are allelic (Harding et al., 1996, Scott et al., 1996). However, It is not known whether or not these mutations affect macroautophagy induced by readministering glucose in S. cerevisiae. It will be interesting to see if there is an overlap in the degradation pathway in non-selective and selective autophagy pathways and the diverging point of these two pathways. Several labs including ours have observed similar autophagy processes in methylotrophic yeast during the protein and cell organelle degradation process. Veenhuis et al (1978, 1981) observed in methanol-induced H. polymorpha cells, when the carbon source was switched to ethanol, the earliest change was the appearance of a variable number (2 to 12) layers of electrondense membranes surrounding peroxisomes. These membranes surrounded individual peroxisomes within a cluster and appeared to sequester a given peroxisome into an autophagosome and then deliver its contents to the yeast vacuole. The process is analogous to macroautophagy in mammalian cells.

PAGE 24

11 Tuttle and Dunn (1995) found that when growth carbon source was switched from methanol to ethanol, the process of peroxisome autophagy in P. pastoris includes an intermediate stage in which individual peroxisomes were sequestered into autophagosomes by wrapping membranes and the autophagosomes then fused with the vacuole, a process similar to macroautophagy. In the glucose-induced peroxisome degradation pathway, he observed that autophagy began with the engulfment of clusters of peroxisomes by finger-like protrusions of the vacuole and the contents were degraded, a process analogous to microautophagy in mammalian cells. Selective Autophagy in Mammalian Cells and Yeast Although non-selective autophagy is considered to be responsible for the bulk turnover of proteins that occurs in response to nutritional starvation or developmental stress (Takeshige et al., 1992), several studies suggested that selective autophagy also plays an important role in regulation of protein and organelle levels in mammalian and yeast cells. Zellweger syndrome is a wellknown prototypic peroxisomal disorder in the newborn (Lazarow and Fujiki, 1 985). Normal peroxisomes appeared to be absent in cells of these patients and peroxisomal enzymes remained in the cytosol due to impaired assembly of peroxisomes (Shimozawa et al., 1992). However, the fibroblasts from these patients were propagated readily in tissue culture. Meijer's group found that in Zellweger fibroblasts, most peroxisomal ghosts in fact contained lysosomal hydrolases (Heikoop et al., 1992). The treatment with the autophagy inhibitor 3-

PAGE 25

12 methyladenine caused an increase in tine number of peroxisome ghosts though peroxisomal functions were impaired. It suggested that peroxisomal ghosts of Zellweger cells were selectively degraded by autophagy. In another study (Luiken et a!., 1992) in, they found that peroxisomal enzymes fatty acyl-CoA oxidase and catalase were preferentially degraded in the isolated hepatocytes of clofibrate-fed rats. This increased degradation of the peroxisomal enzymes was prevented by 3-methyladenine, an inhibitor of macroautophagic sequestration. Long-chain fatty acid could inhibit this degradation too. They concluded that preferential autophagy of peroxisomes exists. In yeasts, independent studies showed evidence that selective autophagy plays an important role to adjust their cellular composition according to environment changes. Veenhuis and coworkers studied selective inactivation of AOX in two peroxisome-deficient {per) mutants in Hansenula (Veenhuis et al., 1983, van der Klei et a!., 1991, Titorenko et al., 1995). They found that these two per mutants synthesized AOX but most of AOX remained in cytosol. Only a small percentage of AOX was degraded under the glucose adaptation condition. These data indicated that degradative inactivation of AOX in H. polymorpha is strictly dependent on the localization of the enzyme inside peroxisomes and the mechanism triggering this process is not directed against AOX protein, but instead, to the membrane surrounding the organelle. In another study, they also observed that not all peroxisomes were degraded in H. polymorpha during the carbon source switch (Veenhuis et al., 1983). One or few small peroxisomes

PAGE 26

13 escaped degradation and subsequently served as the target organelle for newly synthesized matrix proteins in the new growth environment. In their recent study (Titorenko et al., 1995), they have generated mutants impaired in the selective degradation of peroxisomes in H. polymorpha (peroxisome degradationdeficient, pdd). Of the seven mutants they studied, two complementation groups have been identified with two possible responsible genes PDD1 and PDD2. The function of their products in selective autophagy of AOX-containing peroxisomes is under study. The PDD1p is involved in the initial signaling events for sequestration of the organelle since no multilayer membrane sequestration of peroxisomes was observed in pdd1 when compared with wild type cells under ETOH adaptation. In pdd2 mutants, sequestration did occur since the multilayer membrane appeared and wrapped peroxisomes but subsequent fusion of vacuole membrane did not occur. Immunocytochemical detection of alcohol oxidase protein in the vacuole was not observed in any cases. They concluded that PDD2p is essential for mediating the second step in selective peroxisome degradation, namely, fusion and subsequent uptake of sequestered organelles into the vacuole. The model they established for the study of peroxisome maintenance and proliferation, sorting, folding, assembly of matrix proteins, synthesis and function of the peroxisome membrane, and peroxisome turnover in H. polymorpha (van der Klei and Veenhuis, 1996) is similar to what we have established in P. pastoris (Tuttle and Dunn, 1995).

PAGE 27

14 In S. cerevisiae, carbon and nitrogen starvation stimulates non-selective macroautophagy (Takeshige et al., 1992, Egner et al., 1993, Baba at al., 1994). However, data for selective degradation by microautophagy have been reported when yeast cells was switched from oleic acid medium to the one containing glucose (Chiang and Schekman, 1991). When cells were replenished with glucose, a key gluconeogenic enzyme, fructose-1 ,6-bisphosphatase (FBPase) was selectively targeted from the cytosol to the yeast lysosome (vacuole) for degradation (Chiang et al., 1996). Peroxisomes have been identified as the target organelles to be delivered to the vacuole for degradation when cells were replenished with glucose. They have generated vacuolar import and degradation-deficient (vid) mutants and these mutants were placed into 20 complementation groups. FBPase degradation was blocked in all these mutants (Hoffman and Chiang, 1996). They identified a novel type of vesicles in the cytosol specific to vacuolar protein degradation pathway. These vesicles are intermediate in the FBPase degradation pathway which is different to all established vesicles (Huang and Chiang, 1997). One novel responsible gene VID24 has been recovered (Chiang, 1 997) Vip24p is a 41 kD protein and is induced by glucose addition and localized to the intermediate vesicles as a peripheral protein. In the absence of Vid24p, FBPase accumulates in the intermediate vesicles. It seems that this protein plays a critical role to deliver FBPase from the intermediate vesicles for vacuolar degradation. The study clearly showed that the selective protein degradation is present in S. cerevisiae.

PAGE 28

15 Our lab was the first to study selective autophagy in P. pastoris using microautophagy mutants. As stated in the beginning of this chapter, we have identified two distinct peroxisome degradation pathways namely, microautophagy and macroautophagy. We have shown that the autophagy of peroxisome needs synthesis of new molecules. Genetic Analysis of Pichia pastoris Pichia pastoris as a Study Model for Autophagy Yeast is a much simpler system to study autophagy processes than mammalian cells systems. Yeast cells are easy to grow and handle. Their biochemical and morphological characterization can be done quickly. They are one cell eukaryotic organisms and have a much smaller genome and thus fewer genes involved in the autophagy process. When comparing the cells of P. pastoris with S. cerevisiae in the peroxisome degradation process, under glucose induction, peroxisomes are larger, fewer in number, and much easier for biochemical and morphological studies in P. pastoris than S. cerevisiae. The peroxisome degradation process is also much faster in P. pastoris. Eight hour glucose adaptation results in complete peroxisome degradation while in S. cerevisiae, the degradation of peroxisomes under glucose induction needs at least 24 hrs. The measurement of degradation is difficult in S. cerevisiae because the growth of S. cerevisiae under glucose induction results in a dilution

PAGE 29

16 of peroxisomes per cell in a 24 hour period (Chiang et al., 1996; Tuttle, unpublished data). Classical Genetics Pichia pastoris has many advantages in genetic studies. It is ascosporogenic and exists in one of two mating types. The life cycle of Pichia pastoris is characterized by defined haploid and diploid stages (Gleeson and Sudbery, 1988, Cregg et al., 1990) and it can be maintained indefinitely as vegetative haploids. However, upon nutritional limitation, particularly for nitrogen, mating occurs and diploid cells are formed. The mating type can switch between two opposite strains under poor nutritional conditions at high frequency. These key features allow the isolation and phenotypic characterization of mutants in the chemically mutagenized P. pastoris. It also allows us to characterize the phenotypic identification and the complementation groups of mutants (Cregg et al., 1990). Dr. Cregg's group (Liu et al., 1992) developed an efficient screening method for peroxisome-deficient mutants in P. pastoris. The screen relies on the unusual ability of P. pastoris to grow on two carbon sources, methanol and oleic acid, both of which absolutely require peroxisomes to be degraded. A collection of 280 methanol utilization-defective {mut) mutants were isolated, and organized into 46 complementation groups. Molecular Biology P. pastoris has also been developed to be an efficient transformation system for the introduction of replicating plasmids by using E. coli-P. pastoris

PAGE 30

17 shuttle vectors. The system is based on a histidinol dehydrogenase-defective mutant host of P. pastoris and a modified version of the spheroplast fusion-gene transfer procedure (Cregg et al., 1985). The His4 gene of P. pastoris and the autonomous replication sequences (ARS) were isolated and used to construct the plasmid that has about 10^/|jg of transformation frequencies and can be maintained as extrachromosomal elements. For example, the plasmid pYM8 is widely used as a shuttle vector in gene complementation studies. P. pastoris is amenable for transformation by both the electroporation and spheroplast generation method (Cregg et al., 1985, 1993). A number of genes have been cloned and sequenced in P. pastoris due to the availability of a set of auxotrophic strains with essentially wild type genetic background, the plasmids that act as shuttle vectors and the genomic DNA libraries for the isolation of P. pastoris genes by functional complementation of mutants. The genes cloned by this method encode a varied set of proteins such as putative ATPase, peroxisomal integral membrane proteins, proteins related to ubiquitinconjugating enzymes and proteins that might be constituents of the peroxisomal protein import machinery (Nuttley et al., 1995). Peroxisome assembly factor-1 (PAF1 ) is a well known protein responsible for the human Zellweger syndrome (Goldfischer, 1996). Dr. Gregg's group has cloned this gene's homolog in H. poiymorptia (Waterham et al., 1996a, 1996b). Future studies of the function of this gene in yeast will be interesting. It will also be more interesting if we can test if human PAF-1 could

PAGE 31

18 rescue H. polymorpha mutant or vice versa and thus to probe the functional conservation in eukayotes. Gould and his coworkers (Dodt et a!., 1996) developed a method to isolate human peroxisome biogenesis disorder genes by computer-based homology probing of the dBEST database. They found PXR1 is a human orthologue of the P. pastoris PASS gene and PXAAA1 (Yahraus et al., 1996) as a homologue of P. pastoris PASS gene. The same group of Dr. Gould (Kalish et al., 1996) also developed a visual screening method using a peroxisomal form of the green fluorescent protein (GFP). Wild-type cells expressing PTS1-GFP were chemically mutagenized and the mutagenized strains unable to import PTS1-GFP into peroxisomes were identified by fluorescence microscopy. This technique provides an effective visual marker for peroxisomal protein import in living cells. The PTS-GFP containing wild-type cells can be used to regenerate peroxisome mutants. The available molecular technologies also made P. pastoris cells as a host system for the large scale heterologous protein production for industrial usage (Sreekhshna et al., 1997) as well as to attain a certain scale amount of protein for functional study of such proteins. Most of the experiments utilized A0X1 gene promoter to induce gene expression in P. pastoris after A0X1 gene promoter containing plasmid was introduced into P. pastoris. The induction of desired protein synthesis is initiated by methanol addition. The same is true when gene repression is induced by addition of glucose. Dr. Cregg also developed a P. pastoris vector containing a promoter of the GAPDH gene

PAGE 32

19 (Waterham et al., 1997). This promoter can induce expression of the gene constructed behind it. This adds convenience for over-expressing the desired protein for observation or other uses. Direct PGR screening of P. pastoris clones after yeast transformation to verify the right transformant was also developed recently so that less work is needed to verify the expression of the desired gene in P. pastoris. Previous Research of Peroxisome Degradation in Our Lab In our lab, we used methanol induction and subsequent glucose or ethanol adaptation to characterize peroxisomal and cytosolic protein degradation pathways in P. pastoris (Tuttle et al., 1993, Tuttle and Dunn, 1995). Parental strain GS1 15 cells under glucose and ethanol adaptation showed different protein degradation approaches biochemically (Fig. 1-2). The activities of peroxisomal alcohol oxidase (AOX) and cytosolic formate dehydrogenase (FDH) were only about 20% of their zero time activity after a 6 hour glucose adaptation. A 6-hour time course of monitoring AOX and FDH activity and the protein level of AOX and FDH in glucose adaptation showed good correlation and verified that the loss of activity was mainly due to loss of AOX and FDH protein (Tuttle and Dunn, 1995). The activities of cytochrome c oxidase (COO), a mitochondria protein and activities fructose-1 ,6-bisphosphatase (FBP), a cytosolic protein remains high after 6 hour glucose adaptation (data not shown).

PAGE 33

20 Loss of AOX and FDH during Glucose and Ethanol Adapations Figure 1-2. Loss of peroxisomal and cytosolic enzymes during ethanol and glucose adaptation. P. pastoris (GS115) were cultured in methanol induction medium until stationary, at which time ethanol or glucose were added to begin adaptation (time zero), cell-free extracts were prepared at 0and 6-hour time point and alcohol oxidase (AOX) and formate dehydrogenase (FDH) activities were assayed. Values are represented as a percentage of the 0-hour activity. All assays were measured at least three times and the values shown on the graph represent the average ± sd.

PAGE 34

Nu Figure 1-3. Morphological characterization of peroxisomes during glucose and ethanol adaptation. P. pastoris cells (GS115) were induced in methanol and then switch to either ethanol (A-C) or glucose (D-F) for one hour. Under glucose adaptation, peroxisomes were engulfed by the yeast vacuole and contents get degraded. However, under ethanol adaptation, peroxisomes were first wrapped by several layers of membrane to form autophagosome, then fused with the yeast vacuole. Bar in panel A inlet: 25nm, Bar in other panels: 0.5|jM. (use of this figure is permitted by Dr. D. Tuttle)

PAGE 35

22 Also in the glucose-induced pathway, the degradation process requires protein synthesis since protein synthesis inhibitor cycloheximide prevented peroxisome degradation in GS1 15 cells. Under ethanol adaptation, AOX activities reduced significantly as it did in glucose adaptation in a 6-hour time course while FDH activity remained unchanged. The degradation of peroxisomes in ethanolinduced pathway in P. pastoris is independent of protein synthesis since cycloheximide did not prevent ethanol-induced peroxisome degradation. In the same study (Tuttle and Dunn, 1995), they found proteinases A and B were required for the degradation of peroxisomes and FDH. Peroxisomes in PrA and PrB mutants were induced by methanol. However, during the subsequent adaptation to glucose or ethanol, the mutants showed no significant degradation of AOX and FDH proteins as well as peroxisomes. The combined results suggested that the microautophagy of peroxisomal AOX and cytosolic protein FDH is selective under glucose adaptation condition and regulatory molecules are needed to modulate this process. The previous observation was also supported by the morphology studies in light microscopy and electron microscopy. Immunofluorescence microscopy studies were used to screen or verify that the degradation process was disrupted in the mutants (see below). Using quinacrine that labels specifically the yeast vacuole, we can follow the generation and degradation of peroxisomes in parental strains and mutants when we combine the observation results of fluorescence microscopy of the yeast vacuole and the phase-contrast

PAGE 36

23 microscope image of yeast peroxisomes. Immunofluorescence microscopy can follow specifically the alcohol oxidase generation and degradation in peroxisomes as well as other proteins that could detected by the antibodies. However, to characterize the specific detailed structures and degradation process of peroxisomes, electron microscopy is needed. As shown in the Fig.13, at 1 hour of the glucose adaptation in parental GS115 cells, peroxisomes without additional membrane layers were seen in close association with the yeast vacuole (panel D). The extensions or arms of the vacuole were observed surrounding clusters of peroxisomes (panel E) and the vacuolar membrane protrusions fused to engulf the peroxisomes, a homotypic event effusion. Peroxisomes were seen being degraded inside the vacuole (panel F). However, after 1 hour of ethanol adaptation, peroxisomes were seen first to be wrapped by several layers of membrane (panel A) to form autophagosomes. The autophagosomes then fused with the yeast vacuole (panel B) and the contents in peroxisome were degraded (panel C). We do not know the origin of these layers of membrane. It is possible that they are synthesized de novo. The autophagy pathway under ethanol adaptation is analogous to the macroautophagy that has been characterized in mammalian system (see previous discussion in autophagy section). Through these observations, we proposed that the microautophagy mechanism of peroxisomes proceeds via a sequence of events: environmental signaling, peroxisomal recognition, peroxisomal sequestration & homotypic

PAGE 37

24 GLUCOSE SIGNALING Peroxisome . FDH ^ J SEQUESTRATION AND FUSION Vacuole Jflfi '"^ ISfdhS'1 PEROXISOME DEGRADATION Figure 1-4. Model of glucose-induced microautophagy pathway. The yeast responds to glucose adaptation by degrading peroxisomal alcohol oxidase (AOX) and cytosolic formate dehydrogenase (FDH). Peroxisomes and FDH are recognized by and sequestered within the yeast vacuole. Once within the vacuole, these proteins are degraded by a mechanism dependent upon the actions of proteinases A (PrA) and B (PrB).

PAGE 38

25 vacuole fusion, and vacuolar degradation (Fig. 1-4). We hypothesized that cytosolic, peroxisomal or vacuolar membrane regulatory proteins need to be synthesized during the peroxisome autophagy in yeast cells. The interaction and coordination of these regulatory proteins lead to the degradation of specific proteins or organelles of the cells (Fig. 1-4). To approach this question, Dr. Tuttle has generated glucose-induced selective autophagy (gsa) mutants deficient specifically in AOX degradation during the glucose adaptation. Mutants were screened by an colorimetric AOX direct colony assay for their inability to degrade AOX during glucose adaptation and verified by a liquid media AOX assay. We tried to identify molecules regulating the autophagy process as well as to elucidate the mechanisms of the autophagy process. Indeed, in Tuttle's study (Tuttle and Dunn, 1995), he has identified two complementation groups {gsa1 and gsa2) that were defective in the peroxisome degradation pathway during glucose adaptation. The degradation of AOX and FDH in response to glucose was inhibited by 70-90% in both mutants while the degradation of AOX proceeded normally during ethanol adaptation. This suggested that the vacuolar proteinases activities were normal in these mutants and the mutation did not affect vacuolar function but inhibited an event upstream of vacuolar degradation. The ultrastructure studies confirmed that peroxisome entry to the vacuole was defective in these mutants. This also provides further evidence of divergence between these two degradative pathways.

PAGE 39

Chapter Summary 26 The methylotrophic yeast P. pastoris readily grows in media containing methanol as the sole carbon and energy source. During this growth stage, they synthesize the large amount of peroxisomal and cytosolic enzymes that are necessary for the utilization of methanol. Upon adaptation to an alternative carbon source such as glucose or ethanol, the peroxisomes are rapidly targeted to and degraded in the vacuole. The peroxisomes are large enough (0.5 to 1 um) to be identified under light and electron microscopy. The degradation process can also be easily detected by a simple colorimetric assay of AOX activities. Based on previous studies in our lab, we hypothesize that the degradation of peroxisomes during glucose adaptation requires regulatory molecules for the signaling, recognition, peroxisomal sequestration & homotypic vacuolar membrane fusion, and degradation steps in the autophagy of peroxisomes. To approach this problem, gsa mutants that are defective in AOX degradation during glucose adaptation were generated. I tried to define the molecular events of autophagy in P. pastoris by analyzing these mutants. The Pichia pastoris is a proven study model for peroxisome assembly, import and peroxisome degradation. The existing knowledge of peroxisome degradation and the available molecular tools combined with biochemical and ultrastructural methods enable us to examine the microautophagy of peroxisomes and to understand the mechanisms of this autophagy in yeast at the molecular level.

PAGE 40

CHAPTER 2 MATERIALS AND METHODS Yeast and Bacterial Strains and Media All Pichia pastoris parental yeast strains were the generous gifts of Dr. J. M. Cregg (Oregon Graduate Institute, Beaverton, OR). The wild type strains, GS115 {his4), GS190 {arg4), WP1 {ade4), WP2 {met4) and PPF1 {his4, arg4), were routinely cultured at 30°C in YPD (1% Bacto yeast extract, 2% Bacto peptone, and 2% dextrose). YNM and YND media refer to the growth media containing yeast nitrogen without amino acids plus methanol or glucose respectively. If histidine (40mg/L) is added, the media are called YNMH or YNDH. The sporulation and mating medium (SM media) is composed of 0.5% sodium acetate, 1.0% KCI, 1.0% glucose and 2.0% agar. The electroporation medium is composed of 1M Sorbitol, 2% glucose, 0.1% yeast nitrogen base (YNB), 0.4 mg/L biotin and 2% of agar. Bacto peptone, yeast extract and media agar were purchased from Fisher Co. LB broth media is composed of 0.5% yeast extract, 1% tryptone, 0.5% NaCI. Ampicillin is added to LB media at a final concentration of 100 |jg/ml when selection is needed for E. coli growth. WDY (gsa) mutants were generated by Dr. Dan Tuttle by selecting 27

PAGE 41

Table 2-1 P. pastoris Strains Available for Current Studies Strain(Species) Complementation group Genotype GS115 Parental strain his4 GS190 Parental strain arg4 WP1 Parental strain ade4 WP2 Parental strain met4 KB1 Parental strain ade4, his4 JC205 Parental strain met4, his4 PPF1 Parental strain his4, arg4 WDY1(12.1) gsa2-1 his4, gsa2-1 WDY2(4.3) gsa 1-1 his4, gsa1-1 WDK01 Apfkl pfk1::ARG4, his4 WDY3(13.1) gsa1-2 his4, gsa1-2 WDY4(18.1) gsa4 his4, gsa4 WDY5(6.3) gsa5 his4, gsa5 WDY6(4.1) gsa6 his4, gsa6 WDY7(13.2) gsa? his4, gsa7 WDK07 AgsaY gsa7::ARG4, his4 WDY8(17.3) gsa 1-3 his4, gsa1-3 WDY9(15.2) gsa? his4, gsa? WDY1 0(35.3) gsa 8 his4, gsaS WDY11(35.2) gsa? his4, gsa? WDY1 2(9.3) gsa? his4, gsa? WDY1 3(32.1) gsa3 his4, gsa3 WDY14(15.1) gsa? his4, gsa? WDY15(18.3) gsa? his4, gsa? WDY1 6(32.2) gsa? his4, gsa?

PAGE 42

29 glucose-inducecl_selective autophagy deficient strains in chemically mutagenized parental GS1 1 5. WDK01 {Apfl<^ ) is a PFK1 gene knockout strain derived from PPF1 by replacing its PFK1 gene with a S. cerevisiae ARG4 gene. WDK07 (AgsaT) is a knockout strain of GS/A7 from PPF1 . Both Escherichia coli DH5a and Epicurian coli® XL1-blue (Stratagene, Co.) were used to amplify plasmids. A list of yeast strains used in the study is included in the Table 2-1 . Enzyme Assays Yeast cells were first cultured in the methanol induction medium. Glucose or ethanol was then added for adaptation. The yeast cells were harvested at stationary stage (zero hour time point of adaptation) and then six-hours after adaptation (six hour time point of adaptation). Two ml of yeast cells were pelleted at 4°C, 2500 rpm. The cells were resuspended in 1 ml ice cold breaking buffer (20 mM Tris/CI, pH 7.5, 50 mM NaCI, 1mM EDTA). Cells were broken by vortexing 1 min for 3 times mixed with 500pm diameter glass beads (Tuttle et al., 1993). Phenylmethylsulfonyl fluoride (PMSF, ImM) was added to prevent proteolysis. After centrifugation at 4°C, 2500 rpm for 5 min, the supernatant was aspirated and stored at 4°C for subsequent enzyme assays. Alcohol oxidase activity was measured in a reaction coupled with horseradish peroxidase and the oxidation of 2,2'-azino-bio(3-ethylbenzthiazoline-6-sulfonic acid), ABTS. The reaction mixture contained horseradish

PAGE 43

30 peroxidase and ABTS in 33 mM potassium phosphate buffer (Tuttle et al., 1993). The samples were incubated at 37°C until green color developed. The color was measured at 410 nm. The activity of formate dehydrogenase was determined by measuring the formation of NADH during the oxidation of formate by absorption at 340 nm (Kato, 1990). The reaction mix was made of 33 mM potassium phosphate buffer, pH 7.5, 2 mM NAD*, 167 mM sodium formate per reaction plus cell free sample. The rate of change of absorbance at 340nm was followed and used as a measure of enzyme activity. Phosphofructokinase activity and other glycolysis enzyme activities were measured in a coupled enzyme assay resulting in the oxidation of NADH/ NADPH or reduction of NAD/NADP. The NADP and NADH cocktail solution for enzyme activity assays were composed of 98 mL 50mM triethanolamine, lOmM MgCl2 pH7.4, 2 mL NADP 14mg/mL or 0.75 mL NADH (30 mg/mL in 0.1 M Tris pH 7.6) respectively. Other substrates and enzymes were added according to the reaction requirement and finally extracts were added. The activities recorded were adjusted by the extract protein concentration. For example, hexokinase was assayed by a cocktail containing NADP, 0.1 M fructose, 1 mg/mL phososphoglucose-isomerase, 1 mg/mL of glucose-6-phosphate dehydrogenase, yeast cell extract, and the reaction was started with 0.1 M ATP. The production of NADPH was monitored by a kinetic program with an absorbance reading set at 340nm. Pyruvate kinase was assayed by a cocktail containing NADH, 0.1 M ADP, 0.1 M fructose-1,6-bisphosphate, lactate dehydrogenase, the cell free

PAGE 44

31 extract and beginning the reaction by addition of 0.1 M Phosphoenolpyruvate. The production of NAD was measured by a kinetic program at OD 340nm (Dyson et a!., 1975). The glycolysis enzymes tested were GK: glucokinase; HK: hexokinase; PGI: phosphoglucose isomerase; PFK: phosphofructokinase; ALD: aldolase; TIM: Triosphosphate isomerase; PGM: phosphoglycerate mutase; ENO: enolase; and PK: pyruvate kinase. Protein concentrations were measured using crystalline bovine serum albumin as a standard (Bradford, 1976). Isolation of Glucose-Induced Selective Autophaav-Deficient (asa) Mutants Mutant Generation and Isolation WDY mutant strains were generated by mutagenizing parental GS1 15 cells with nitrosoguanidine at 100 [jg/ml concentration (Tuttle and Dunn, 1995). The mutagenized cells were screened for the loss of the ability to degrade alcohol oxidase in response to a shift in a carbon source from methanol to glucose with a direct colony assay (Tomlison, 1992, Tuttle and Dunn, 1995). Briefly, the mutagenized cells grown on YPD plates were replica-plated to YNMH plates (6.7 g/L yeast nitrogen base without amino acids, 0.5% methanol, 0.4 mg/L biotin, and 40 pg/ml histidine) and the colonies were allowed to grow for 4-5 days. The colonies were replica-plated onto nitrocellulose and placed on YNDH plates (6.7 g/L yeast nitrogen base without amino acids, 2% glucose, 0.4 mg/L biotin, and 40 |jg/ml histidine ) for 12-16 hours. Those putative mutant

PAGE 45

32 colonies which retained AOX despite glucose adaptation were identified by the purple color reaction product of the direct colony assay (Tuttle and Dunn, 1995). The principle of reaction is the same as the alcohol oxidase assay. Those mutant strains appeared purple were isolated and their inability to degrade AOX during glucose adaptation verified in liquid cultures. This was done by first growing the cells for 24-36 hours in 20 mis of methanol induction medium that consisted of 6.7 g/L yeast nitrogen base without amino acids (Difco), 0.4 mg/L biotin, 40 pg/ml L-histidine, and 0.5% methanol. Glucose was then added to a final concentration of 2% at the beginning of glucose adaptation. Cells were harvested at Oh and 6h of glucose adaptation and lysed by vortexing with acidwashed glass beads. The resulting homogenates were then assayed for AOX activity according to the procedures of Tuttle and Dunn (1995). Mutant Backcrossing To further characterize the mutant strains, gsa mutants that have been identified were backcrossed to GS190 {arg4) and GS1 15 {his4) for generating gsa specific mutants as well as for identifying complementation groups. The haploid mutants were sequentially mated to essentially wild type strains (GS190 and GS1 15) containing complementary auxotrophic markers and haploid progeny recovered. Several rounds of backcrossings resulted in a particular mutant strain with a mutated gene of our interest. One round of backcrossing was accomplished as follows: the mutant strains were streaked onto YPD plates in patches and the parental strains of the complementary auxotrophy was spread

PAGE 46

33 on YPD plates to form a lawn of cells and both plates were incubated at 30°C overnight. The lawn and the patch were both replica plated to a single sporuiation media (SM) plate and incubated overnight at SCC to induce mating. These plates were then replicated to YND plates without amino acids and incubated for 2 to 3 days until diploid colonies appeared. Colonies from these plates were streaked onto a fresh YND plates and incubated for 1 to 2 days and then were streaked on YPD plate. Cells from this plate were then streaked to SM plates and incubated four days at 30°C to induce meiosis and sporuiation. Cell spores were harvested with an inoculation loop and the remaining diploid cells were killed by etherization. An aliquot from this mixture was diluted 1 to 100 and 100 pi plated on YPD plates for two days for the haploid cells to grow. The cells were then washed off and plated on YNMH or YNMA plate for a direct colony assay for mutants screen. The verified mutant via this assay and the subsequent AOX liquid media assay was selected and used for the next round of backcrossing with a parental strain that has the opposite auxotroph. Mutant Complementation Analysis To determine the number of defective genes present in the gsa mutants in my possession, the complementation groups of gsa mutants had to be identified first. To achieve this, different auxotroph strains containing mutants (for example, a histidine auxotroph and an arginine auxotroph mutant) are mated and a prototroph (diploid) of them is generated. The ability of this diploid to degrade peroxisomes can be tested in a glucose adaptation assay. If the

PAGE 47

34 peroxisome degradation can proceed in this diploid, these two strains are said to complement each other. Complementation indicates the defective genes in these two mutants are different and the mutants belong to different complementation groups. If the diploid still could not degrade peroxisomes, the two mutant strains presumably contain different mutant alleles of the same gene and they belong to the same complementation group. Molecular Biology Yeast Transformation The genomic library was a gift of Dr. J.M. Cregg (Oregon Graduate Institute). It was constructed by a partial digestion of genomic DNA with Sau 3A I first. The digestion products were then run on a DNA gel and the DNA of 5-10 kb size was collected, purified and ligated into BamH I site of plasmid pYM8. Pichia pastoris cells {his4) were grown in YPD to an ODSOOnm = 1 .0 harvested, and transformed by electroporation at 1.5 kV, 25 mF, 400 Q (Gene Pulser, BioRad Corporation) with 5 to 10 pg of genomic DNA library. The transformed cells were grown for 4-5 days on electroporation plates, and the colonies were replica-plated to nitrocellulose and placed on YND-aa plates for 12-16 hours. Those colonies which appeared purple or white (according to experiment purposes) upon a direct colony assay were isolated and their abilities to degrade AOX during glucose adaptation verified in liquid media as described above.

PAGE 48

35 Stable transformation with vectors pYM8 and pYM4 containing specific fragments of genomic DNA was also done by electroporation. pYM8 was derived from pBR322 and contains the 164 bp autonomous replicating sequence from P. pastoris and the HIS4 gene of S. cerevisiae. pYM4 was also derived from pBR322 and contains a HIS4 gene of P. pastoris, but lacking an autonomous replicating sequence. pYM25 was derived from pBR322 and contains an ARG4 gene of S. cerevisiae. Both pYM4 and pYM25 are nonepisomal vectors. Stable integration of the episomal pYM8 constructs was accomplished by two cycles of growth under nonselection and selection conditions. Stable integration of non-episomal pYM4 or pYM25 constructs was promoted by a single cut with restriction enzyme Stu I within the HIS4 locus. This cut should also direct the integration into His4 locus in the transformed cell genome. The plasmid maps of pYMS and pYM4 is shown in Fig. 2-1 . Plasmid Isolation and DNA Sequencino The rescued colonies that had been complemented with the genomic DNA library were grown overnight in 2 ml of YPD medium. The cells were pelleted and resuspended into 0.2 ml of 2% Triton X-100, 100 mM NaCI, 1 mM EDTA, 1% SDS, and 10 mM Tris/HCI, pH 8.0 and 0.2 ml of phenolxhloroform: isoamyl alcohol (25:24:1). The cells were then disrupted by vortexing in the presence of 0.5 ml acid-washed glass beads. After centrifugation, 2 fx\ of the upper phase containing the plasmid was used to transform E. coll DH5a cells (Ausubel et al., 1988). The plasmid was then isolated by Wizard Plus minipreps

PAGE 49

36 Figure 2-1 . Maps of plasmids pYM8 and pYM4. pYM8 is constructed based on pBR322 with a S. cerevisiae HIS4 gene and an autonomous replication sequence (ARS). pYM4 is also based on pBR322 but lacking an ARS. pYM4 contains a P. pastoris HIS4 gene. The main restriction enzyme sites are also indicated in the two plasmid maps.

PAGE 50

37 of Promega and restriction analysis was done to identify the genomic insert size and sites. The genomic DNA fragment was excised from the pYM8 shuttle vector and engineered into pBluescript/KS for sequencing by nested deletion (Ausubel et al., 1988). Direct sequencing of pDLT1 was also done using self generated primers that flank the desired region. The primers for sequencing were synthesized by DNA synthesis lab, University of Florida. All sequencing was performed by University of Florida Sequencing Core via the dideoxy chain termination method. The insert of pDLT1 was sequenced from both plus/minus strands. The DNA sequences were assembled in DNAman® or Gene Runner® programs and all six reading frames were compared to protein sequence databases using the Blastx program of the National Center for Biotechnology Information (NCBI). Northern. Southern and Western Blot Analysis Yeast cells in exponential growth were pelleted and resuspended in 1% SDS, 10 mM EDTA, and 50 mM sodium acetate, pH 5 solution. Total RNA was extracted with addition of hot phenol/chloroform/ isoamyl alcohol (25:24:1). About 10 |jg of total RNA was loaded on glyoxal agarose gels. The RNA was then transferred to Maximum Strength Nytran® (Schleicher & Schuell Inc.) by overnight blotting (Ausubel et al., 1988). The RNA in the blot was cross-linked by UV for 12 seconds and then subjected to methylene blue staining for 2 min. The staining pattern of ribosomal RNA was photographed or scanned as loading control. A 2.5 kb DNA fragment in PFK1 gene was used as template for ^^P-

PAGE 51

38 labeling by random priming reaction with NEB's NEBIot™ Kit. The probe was cleaned by NucTrap® Push Columns (Stratagene Inc.). Northern blotting using labeled DNA probes were done as described (Ausubel et al., 1988). An equivalent 10 million counts of radiolabeled DNA probes were used for one typical labeling. For Southern blot purpose, genomic DNA was isolated from stationary growth cells. It was then digested with BamH I, separated on agarose gels, and then transferred to nylon (Kaiser et al., 1994). ^^P-labeled DNA probe labeling was done as described in Northern blotting. To prepare Western blot samples, P. pastoris were grown on YPD or on methanol and then adapted to glucose. The yeast cells were collected and prepared for SDS gel according to Kaiser (Kaiser et al., 1994). The samples were separated on a SDS minigel (Bio-Rad). The proteins were transferred to the nitrocellulose and incubated with anti-HA.1 1 mAb (against mouse) or polyclonal antibody (against rabbit) (Babco). The blot were then incubated with secondary anti-mouse or anti-rabbit antibodies conjugated with HRP. Protein bands was detected by ECL method (Amersham). Isolation of PFK1 and GSA7 Knockouts The construction of WDK01 {his4, pfk1::ARG4) cell was done by disrupting the PFK1 gene of PPF1 {his4, arg4). First, the 1.2 kb Hind III fragment within the open reading frame of P. pastoris PFK1 was replaced with the 3.0 kb S. cerevisiae ARG4 gene from pYM25. A linear 5.6 kb fragment with

PAGE 52

39 the ARG4 gene flanked by PFK1 was excised from the shuttle vector by BamH I digestion and used directly to transform PPF1 cells. Cells were grown on transformation plates supplemented with histidine (40 |jg/ml). Colonies were isolated and their ability to degrade AOX during glucose adaptation examined (see Mutants Generation and Isolation section). PFK1 gene disruption was verified by measuring phosphofructokinase activity assay (Blangy et al., 1968) and by Southern blotting (see above). The construction of WDK07 was basically the same as for WDK01 . The S. cerevisiae ARG4 gene was put in between Hind III and Bgl II sites of the GSA7 gene and the knockout fragment was cut out by Apa I and Sea I restriction enzymes and used to transform PPF1 cells. The transformants were first screened with an AOX direct colony assay. The colonies that appeared purple in the assay should lose AOX degradation ability and are thus the putative "knockouts". Possible knockouts were tested for the correct gene targeting by PGR the genomic DNA extracted from these clones. The clone has the expected PCR fragment (4.9 kb instead of 2.5 kb in a normal strain) is the true knockout and was named as WDK07 (Agsa7). Site-directed Mutagenesis of PFK1 The aspartic acid at position 362 of Pfk1 protein was changed to a serine by site directed mutagenesis. This was done using the Stratagene® QuikChange™ Site-Directed Mutagenesis Kit. Mutated pfk1 was prepared by PCR utilizing two complementary oligonucleotide primers (synthesized at the University of Florida Oligonucleotide Core) that contained the desired two base

PAGE 53

40 mutations, Pfu DNA polymerase, and pDLT1 containing the wild type PFK1 as a template. This resulted in "nicked" plasmid containing pfkt The template DNA was digested by Dpn I endonuclease and the nicked plasmid was repaired and amplified in Epicurian Coli® XL1 -Blue cells. Pfk1 was then subcloned into pYM4 (pWP-pfk). Sequencing in both directions of the pWP-pfk was performed at the University of Florida Sequencing Core by the dideoxy chain termination method and desired bases change was verified (see Fig. 5-4). pWP-pfk was then introduced into gsa1-1 and WDK01 by electroporation. The transformants were selected based on the integration of HIS4 gene of plasmid into the genome. The linearization of the plasmid inside HIS4 gene promoted integration into HIS4 gene locus of the genome. Generation of a HA Epitope Tag in GSA7 and YHR171w Gene Two primers were designed to PGR the GSA7 gene plus a HA epitope tag at the C terminus of the protein. The 5' primer started 350 bases before the start codon of GSA7 in order to include its endogenous promoter, while the 3' primer was at the end of the coding sequence plus a HA epitope tag. The HA epitope tagged GSA7 gene was amplified by PGR with ID-PROOF™ polymerase (ID Labs Biotechnology). The successful construct was verified and subcloned into vector pYM4 Cla I and EcoR V sites (pYWP7-HA). The construct pYWP7-HA was linearized in HIS4 gene or GSA7 gene and used to transform gsa7 cells. In S. cerevisiae, the same strategy was used to generate a HA epitope tagged YHR171wan6 put into pYM4. The YH R1 71 w gene v^as also put in the pHWOlO

PAGE 54

41 vector behind a GAP promoter (Waterham et al., 1997) and in the pPIC3 vector behind an AOX promoter (both vectors are a gift of Dr. JM Cregg). These constructs were used to transform S. cerevisiae as well as P. pastoris. The expression of HA tagged protein was confirmed by Western blot by HA antibodies. Fluorescence Microscopy and Electron Microscopy Yeast cells were grown on YPD, glucose, ethanol or methanol media for different experiment setting. The cells were then fixed in 3% (w/v) of paraformaldeyde (pFA) for 10 min at room temperature. The yeast cells were centrifuged and resuspended in 50mM potassium phosphate buffer (Kpi) / 3% pFA at room temperature for another 30 min with gentle rotation. The cells were washed with Kpi and resuspended in SPC buffer (50 mM K2HPO4, 16 mM citric acid, 1M sobitol, pH 5.8). The cells were spheroplasted with Zymolase 20T in SPC buffer for 15 min and the digestion was monitored every 5 min for cell wall digestion completion. After digestion, the cells were washed with SPC buffer and put on a 15-well slide precoated with polylycine (ICN Biomedicals). The fluorescence containing media such as phalloidin, quinacrine were applied to the slide for direct observation or the slides were placed in the blocking medium for immunofluorescence microscopy study (5% normal goat serum in PBS). For this purpose the cells on the wells were incubated with primary antibody for 1-2 hr or

PAGE 55

42 overnight and then with secondary antibody for one hour. The observation was done with a Zeiss axiophot photomicroscope using a Zeiss 100X oil emersion. Ultrastructural analysis was performed using a potassium permanganate fixation protocol (Veenhuis et al., 1 983). This procedure effectively delineates membrane structures in methylotrophic yeasts. Briefly, cells were harvested by centrifugation, washed in water, and fixed in 1.5% KMn04 in veronal-acetate buffer (0.3 mM sodium acetate; 0.3 mM sodium barbital, pH 7.6) for 20 min at room temperature. The specimens were then washed three times and dehydrated in increasing concentrations of ethanol wash. This was followed by 100% propylene oxide twice. The cells were then infiltrated with a 50:50 mix of propylene oxide and the POLY/BED 812 (Polysciences, Inc., Warrington, PA) for two days. The preparations were dried by a vacuum overnight, infiltrated with 100% POLY/BED with accelerator 2,4,6-Tri(dimethylaminomethyl) phenol (DMP-30®, Polysciences, Inc.) for another two days, and then incubated in an oven overnight at 60°C. The resulting samples were then mounted on the blocks and the blocks were then sectioned (by D. Player, Department of Anatomy and Cell Biology, University of Florida College of Medicine) and prepared for examination on a JEOL 100CX II transmission electron microscope.

PAGE 56

CHAPTER 3 CHARACTERIZATION OF PICHIA PASTORIS MUTANTS DEFECTIVE IN GLUCOSE-INDUCED MICROAUTOPHAGY Introduction Methylotrophic yeast P. pastoris can utilize methanol as a sole carbon source by synthesizing those peroxisomal and cytosolic enzymes necessary to assimilate methanol. When P. pastoris cells were grown on methanol, their peroxisomes were induced in large size and numbers. During the subsequent adaptation to glucose or ethanol, their peroxisomes were rapidly degraded (Tuttle et al., 1993). P. pastoris has been an ideal model to investigate the molecular events of peroxisome biogenesis (Subramani, 1996) as well as peroxisome degradation (Tuttle and Dunn, 1995). Our lab has identified two distinct pathways for the degradation of peroxisomes in P pastoris (Tuttle and Dunn, 1995) namely, microautophagy and macroautophagy. These pathways are independently regulated, but share at least one common event, vacuolar degradation. In order to better define the molecular aspects of peroxisome turnover in P. pastoris, glucose-induced selective autophagy deficient (gsa) mutants have been generated from the 43

PAGE 57

44 Steps for the Identification of those Proteins Required for Microautophagy Mutagenesis of Parental Strain Isolation and Selection of Mutants I Backcross and Complementation Analysis \ Biochemical and Morphologic Characterization of Mutants I Rescue of Mutants with Genomic DNA i Determination of Genomic DNA Sequence i Further Studies Figure 3-1 . A flow chart of mutants characterization. A mutant was selected for its inability to degrade AOX during glucose adaptation. Once its biochemical and morphological profile had been identified, it went through backcrossing and complementation analysis. The mutant was then transformed with a P. pastoris genomic DNA library to identify the responsible gene. Once the gene was recovered and verified, the gene was sequenced and, according the sequencing result, the subsequent studies were determined and carried out.

PAGE 58

parental strain GS1 15 by chemical mutagenesis (Tuttle and Dunn, 1995) for their inability to degrade peroxisomal alcohol oxidase (AOX). The putative mutants that appeared purple in an AOX direct colony assay were selected. They were then subjected to liquid media AOX and FDH assays for further verification. The mutant strains were then backcrossed with parental GS190 (an arginine auxotroph) and then mated with another mutant in a histidine auxotrophic background (see below), to identify their complementation groups. The selected mutants were also backcrossed with parental GS190 and GS1 15 two to four times to eliminate background mutations that were generated in the chemical mutagenesis unrelated to gsa phenotype. A flow chart of mutant generation, identification and characterization steps has been included in Fig. 31. In this chapter, I will detail the genetic, biochemical and morphological characterization of gsa mutants. Complementation Group Identification The purpose of the complementation analysis of mutant strains is to provide an indication of the genetic relationship of the mutations involved and thus to further characterize the gsa mutants. Complementation groups were identified by backcrossing different gsa mutants with an arginine auxotroph GS190. The resulting arginine mutants were then mated with other gsa mutants (histidine auxotrophs) as well as parental GS1 15. Their progeny were analyzed

PAGE 59

46 for their peroxisomal AOX degradation ability. If the progeny (diploid cells) rescued the defect of both mutant strains (regaining AOX degradation ability), then these two strains represent two different complementation groups and two unique genes in each of the mutants is defective. If they complemented each other, they belong to the same complementation group with possible different mutated alleles. A sample complementation study graph of strain WDY4, WDY8 and WDY10 is shown in Fig. 3-2. AOX activities of six hour glucose adaptation were presented as a percentage of zero hour activities. As shown in the graph, the progeny of WDY8 and WDY4 could degrade AOX and little AOX activity remained after 6 hour glucose adaptation, and thus they complemented each other. WDY4 and WDY8 did not complemented WDY10. The same was true for WDY1 0 and GS11 5. The result of AOX activity assays of these diploids clearly showed that WDY4, WDY8 and WDY10 are three separate gsa groups. The fact that WDY10 was not complemented by GS1 15 made it clear that WDY10 has a dominant trait. Currently, I have identified eight complementation groups. All gsa mutants except gsaS showed recessive traits while gsaS showed a dominant trait. WDY2, WDY3 and WDY8 belong to the same gsa1 mutant group and are thus named as gsa1-1, gsa1-2 and gsa1-3 respectively. A list of gsa mutants is shown in Table 3-1 . Some of the mutants showed variable AOX degradation results, such as WDY9, 11, 12, 14 and WDY15. This variability may be due to several mutations in one single mutant strain and thus this strain might need to be backcrossed before complementation analysis. Since there

PAGE 60

47 gsa Mutants Complementation Analyses Matings Figure 3-2. Identification of gsa mutants complementation groups. The figure is used to illustrate how complementation groups were identified. WDY4, WDY8 and WDY10 were screened from frozen stocks and identified as gsa mutants. They were then subjected to backcrossing. The resultant WDY8a1 , WDY1 Oa1 , WDY4a1 (Arg auxotroph, generation one) were then mated with GS1 15 cells and themselves. The progenies of mating were subjected to AOX assay. Values of AOX activities after 6 hour glucose adaptation were represented as a percentage of the 0-hour activity. They were measured at least three times in the separate experiments and were shown on the graph as average ± sd. The result showed that WDY4, WDY8 and WDY10 belong to three new complementation groups. However, WDY10 showed a dominant trait since the prototroph of WDY10 and other mutants as well as GS11 5 showed no rescue effect. WDY4 and WDY10 were subsequently named as gsa4 and gsaS. WDY8 was subsequently found to be in the same gsa group as gsa1-1 (data not shown) and was thus named as gsa1-3 (WDY3 has been identified as gsa1-2).

PAGE 61

48 are still more potential mutants and their complementation groups have not been identified, more work needs to be done on the remaining mutants. Although the identified eight gsa complementation groups cover only part of the regulatory molecules that control the process of autophagy, further characterization of these mutants now is possible because of the establishment of these complementation groups. Future studies of the remaining mutants will yield information to better cover all the molecules participating in the peroxisome degradation process as well as to better understand the autophagy mechanism. Screening for asa Mutants At the time of complementation analysis of gsa mutants, the gsa mutants were also going through backcrossing cycles (see Materials and Methods for backcrossing steps). The purpose of backcrossing was to prepare gsa mutants for further genetic study. Because the mutant strains were generated by random chemical mutation, the mutations occurred at various sites in the genome. By backcrossing two to four times back & forth from histidine auxotroph background to arginine auxotroph background, most background mutations that were not related to gsa phenotype were eliminated. In eight gsa complementation groups {gsa1 to gsaS) (see table 2-1) that have been identified, gsa1-1 and gsa2 have been successfully backcrossed four times. Gsa3, gsa5, gsaJ and gsa8 have been backcrossed two times and gsa4

PAGE 62

49 has been backcrossed one time, but the phenotype of this strain is not stable in its arglnine auxotroph background. Gsa6 has a stable and strong phenotype biochemically. However, we have failed in three separate attempts to backcross it. Once it was placed into the arginine auxotroph background, the phenotype was lost. It is possible that the phenotype effect of gsa6 is an accumulated one of several defective genes. Backcrossing that leads to the loss of any one of them shows no phenotype in gsa6. Gsa1-1 and gsa7 were later chosen for further genetic analysis. Glucose Induced Microautophaqv is Defective in asa Mutants Previously, we have shown that parental strain GS115 undergoes microautophagy during glucose adaptation. The loss of AOX activity, the degradation of AOX protein and the degradation of peroxisomes were compatible (Tuttle and Dunn, 1995). We selected gsa mutants on the basis that the mutant could not degrade AOX and peroxisomes. Of the collection of possible mutants, I have identified eight gsa complementation groups. All mutants were unable to degrade peroxisomal AOX during glucose adaptation while their ethanol-induced macroautophagy proceeded normally (Fig. 3-3 and Fig. 3-4). The biogenesis of peroxisomes observed during methanol induction was intact in all these mutants as determined by light microscopy. The degradation of AOX and FDH during glucose adaptation was rapid in parental

PAGE 63

50 GS1 1 5 cells (also see Fig. 1 -2). Less than 20% of the activities of AOX and FDH was left in GS115 after six hour glucose adaptation. However, in mutants presented in these two graphs, they lacked the ability to degrade AOX that has been induced under methanol growth condition. Compare to parental GS115 cells, AOX degradation in response to glucose adaptation in all mutants was inhibited to various extents. The degradation of FDH, a cytosolic protein that required for methanol utilization, was also shut down in most gsa mutants. However, in gsa4 and gsaS, FDH degradation proceeds nearly normal when compared with parental GS115. Since FDH is a cytosolic protein, one possibility is that a mutation block an early event of peroxisome microautophagy pathway in this mutant but the early cytosolic protein degradation pathway may be controlled differently from that of peroxisome degradation and thus the FDH degradation process is intact. Interestingly, in the complementation analysis done by Mr. Kendal (Kendal, unpublished data, 1993), gsa4 strain did not complement either gsa1-1 or gsa2 cells. This may be either due to more than one mutant gene in it since gsa4 was not backcrossed at the time of experiment or to some unknown reasons. As for gsa8 cells, I found it is a dominant mutant in our complementation analysis. We still need to define the relationship of dominant gsa8 and FDH degradation. Ethanol induced peroxisome degradation in all mutants proceeded normally (see Table 3-1 ). This confirmed that our gsa mutants only affected early events of microautophagy of peroxisomes since vacuolar degradation of

PAGE 64

AOX and FDH Activities in gsa Mutants during Glucose Adaptation GS115 gsal-3 gsa4 gsaS gsa6 gsaS Figure 3-3. AOX and FDH activities of gsa mutants under glucose adaptation. P. pastoris (GS1 1 5) and gsa mutants were cultured in methanol induction medium until stationary, at which time glucose was added to begin adaptation (time zero). Cell-free extracts were prepared at 0and 6-hour time point and alcohol oxidase (AOX) and formate dehydrogenase (FDH) activities were measured. GS1 1 5, gsa1-3, gsa4, gsa5, gsa6 and gsaS are included in this figure. Values are presented as a percentage of the 0-hour activity. All assays were measured at least three times and the values shown on the graph represent the average ± sd.

PAGE 65

AOX and FDH Activities in gsa Mutants during Glucose Adaptation GS115 gsal-1 gsa2 gsaS gsa7 Figure 3-4. AOX and FDH activities of gsa mutants under glucose adaptation. P. pastoris (GS115) and gsa1-1, 2, 3 and gsa 7 were cultured in methanol Induction medium until stationary, at which time glucose was added to begin adaptation (time zero). Cell-free extracts were prepared at 0and 6-hour time point and alcohol oxidase (AOX) and formate dehydrogenase (FDH) activities were measured according to Materials and Methods. Values are presented as a percentage of the 0-hour activity. All assays were measured at least three times and the values shown on the graph represent the average ± sd.

PAGE 66

53 Table 3-1 Characterization of gsa Mutants strains gsa group Genotype Phenotype Glucose ETOH AOX FDH AOX FDH GS115 Parental strain his4 + + + GS190 Parental strain arg4 + + + PPF1 Parental strain his4, arg4 + + + WDY1 gsa2 his4, gsa2 + WDY2 gsa 1-1 his4, gsa 1-1 + WDY3 gsa1-2 his4, gsa 1-2 + WDY4 gsa4 his4, gsa4 + + WDY5 gsa 5 his4, gsa 5 + WDY6 gsa 6 his4, gsa 6 + WDY7 gsa? his4, gsa? + WDY8 gsa 1-3 his4, gsa 1-3 + WDY10 gsa 8 his4, gsa 8 + + WDY13 qsa3 his4, gsa3 +

PAGE 67

54 peroxisomes was intact in ethanol induced macroautophagy whicli is shared with microautophagy. Although my work focused on glucose-induced microautophagy, the studies of ethanol-induced macroautophagy are interesting and important. This is because most of autophagy mutants in S. cerevisiae and H. polymorpha are macroautophagy mutants. Only P. pastoris can undergo both degradation pathways during glucose and ethanol adaptation. A list of complete biochemical profiles of gsa mutants and parental strains are compiled in Table 3-1 . GS115 is the parental strain used for chemical mutagenesis. GS190 is used for backcrossing. PPF1 is a double auxotroph used later in the experiment. Their biochemical profiles are provided. All identified gsa mutants are included in the table. Some of them are not listed here due to the variable results of their biochemical studies such as WDY9, 11 , 1 2, 1 4 and WDY1 5. While I was studying gsa1-1, I also backcrossed WDY1 (gsa2) four times to prepare for further genetic analysis. After I finished biochemical and morphological studies of gsa2, I transformed Sfsa2 cells with a genomic DNA library to identify GSA2 gene. The colonies appeared white in an AOX direct colony assay regained their AOX degradation ability and thus should bear rescue plasmids with GSA2 gene. Much to my surprise, about 1 0% of the colonies appeared white on the AOX direct colony assay and thus were putative rescue clones. The positive and negative control strains tested in the direct colony assay worked fine which meant the detection of white rescue colonies was not due to technical errors. I eliminated some of the false positive rescue

PAGE 68

55 colonies by using liquid media AOX assay. However, I still have too many promising colonies. Twelve plasmids were recovered from 12 of these colonies. Each plasmid appeared different based on restriction enzyme analysis. Some of them had no genomic DNA insert but still rescued grsa2. Although I could not exclude the possibility that there are GSA2 or GSA2 suppressor genes in my collections of "rescue colonies", I could not identify them. I tested different growth conditions in order to seek optimal conditions to grow and to characterize the rescue clones. One surprising phenomenon was that the gsa2 mutant was rescued by addition of 80 pi of 10mg/mL histidine at the beginning of glucose adaptation during liquid media AOX assay. Subsequent studies verified that this was a true phenomenon and the doubling of histidine amount at the beginning of methanol induction did not affect this result. This fact promoted me to re-consider the failed gsa2 rescue. In the rescue study, only mutants bearing pYM8 containing plasmids and expressing histidine could survive and grow. Since gsa2 can be complemented by histidine, the expression of histidine in the complemented strains partially or completely offset the phenotype of gsa2 and thus resulting in a lot of rescue colonies. The expression level of histidine might be quite different in each clone and thus not all strains bearing pYMS could rescue gsa2. We do not know why histidine rescues gsa2 when it was added upon glucose adaptation. A review of published references did not help us either. One possibility is the addition of histidine at the time of glucose adaptation kicks in a nonspecific

PAGE 69

GS115 gsal-1 gsa2 gsa4 gsaS gsaS P o GS115 gsal-2 gsal-3 gsa3 gsa6 gsa7 Figure 3-5. AOX degradation and histidine addition during glucose adaptation. The test was used to identify gsa mutants that are not affected by histidine addition The experiment was done as a liquid media AOX assay described in Materials and Methods except that at the zero time point of glucose addition, glucose or glucose/histidine were added to each gsa mutant to compare the effect of histidine addition. All assays were measured at least three times and the values shown on the graph represent the average ± sd.

PAGE 70

57 protein degradation pathway. It is also possible that histidine may directly activate macroautophagy. To further probe this problem, I first screened all the mutants that I have and test their ability to degrade AOX when histidine was added at the time of glucose adaptation. It was not surprising to find that gsa1 was not rescued by histidine. Using non-backcrossed gsa2 strain for the same experiment, I verified that the histidine rescue effect exists and concluded that the phenomenon was not due to improper selection during the backcrossing. All other gsa mutants were tested for the effect of histidine addition at the beginning of glucose adaptation and Fig. 3-5 showed the AOX activities remaining after a six-hour glucose adaptation in these gsa mutants. The effect of histidine addition was compared with control groups grown in the same condition except no histidine addition at the onset of glucose adaptation. As shown in the upper panel, histidine restored AOX degradation ability (or partially restored) in gsa2, 4, 5 and gsa8 while in gsa1, gsa3, gsa6 and gsa 7, histidine has no effect on their AOX degradation. However, the rescue effect varied in these three strains which may be a reflection of the expression level of histidine. Gsa1, gsa3, gsa6 and gsaZ were chosen for further genetic studies. Morphological Studies of gsa Mutants The ultrastructural analysis was done by the potassium permanganate fixation protocol (Veenhuis, et al., 1983). The basic phenomena in all these

PAGE 71

58 mutants were their inability to degrade peroxisomes in a prompt manner albeit these eight mutants showed distinct presentations (Fig. 3-6 and Fig. 3-7). In gsa1-1 mutant, the vacuole appeared round and small. It lacked the finger-like extensions or invaginations present in glucose-adapting GS115 cells (see Fig. 1-3). The morphological presentation of the vacuole did not change much during glucose adaptation. The vacuole was not active since there were few autophagosomes and little cell debris inside the vacuole when compared to GS115 cells. All this suggested that gsa1-1 is required for the initiation of microautophagy and the molecule mutated in gsa1-1 may control an early step of microautophagy, that is, the initiation of autophagy. The limited degradation of AOX and FDH probably is due to inefficient degradation of peroxisome via macroautophagy under glucose adaptation conditions. In gsa4, the peroxisomes were identified outside of the yeast vacuole after 3 hour's glucose adaptation while autophagosomes were seen associated with the yeast vacuole. Although AOX degradation was defective in gsa4 cells (see Fig. 3-7), the extent of the defect, when judged by AOX degradation, was much less when compared with other mutants (Fig. 3-3). This was in agreement with the morphological data. We knew in gsa4, cytosolic protein FDH's degradation proceeds normally. It is possible that the defect in gsa4 was mild and although degradation of AOX ability was inhibited, the cytosolic protein FDH degradation still could proceed via impaired microautophagy or inefficient macroautophagy. It is also possible that the mutant has a recognition defect

PAGE 72

59 and this defect disrupted the microautophagy of peroxisomes while its cytosolic protein degradation in gsa4 cells was intact. It should be interesting to pursue gsa4 to identify the defective gene of this mutant. The genetics data showed that gsaS has a dominant trait. When examined under electron microscopy, it bore similarity to gsa4 cells albeit to less extent than to the active level of the yeast vacuole. During glucose adaptation, autophagosomes with cell debris or cytosolic contents were seen around and inside the yeast vacuole (Fig. 3-6). However, large peroxisomes outside of the yeast vacuole were easily identified, and we did not observe vacuolar degradation of peroxisomes. Our biochemical data showed that the AOX degradation was defective. However, FDH degradation proceeded normally. I suspected that the slow progress of macroautophagy leads to a much greater loss of cytosolic protein FDH while the defect in microautophagy keeps AOX level relatively higher in gsa8. Gsa4 and gsa8 should be in the same group and their defective gene products control the recognition step in the process of peroxisome microautophagy downstream of Gsa1p. In gsa2 cells, finger-like extensions of the vacuole were found surrounding the peroxisomes after 3 hour glucose adaptation and this was not seen at the time of methanol induction. However, it appeared that fusion of vacuolar membrane did not occur, and thus, sequestration of peroxisomes was not complete (Tuttle and Dunn, 1995). Although the morphology suggested that microautophagy was shut down in gsa2 cells, we did observe some small

PAGE 73

60 autophagosomes were in contact with the yeast vacuole. The biochemical data showing that AOX degradation was defective corroborated this observation. However, slow degradation of AOX and FDH did exist and it appeared that the gsa2 cells might be able to adapt to their deficiency and the environmental situation by turning on the macroautophagy pathway albeit much less efficiently. I suspected a fusion step of vacuolar membrane during the engulfment of peroxisomes is defective in gsa2 mutants and Gsa2p should act downstream of Gsa1p. Morphologically, gsa3, gsa6 and gsa7 showed similar characteristics as gsa2. The initiation of autophagy was turned on during the glucose adaptation. The vacuole has formed protrusions that surround peroxisomes. However, we did not observe peroxisome degradation in the yeast vacuole. The biochemical data of AOX and FDH activities during glucose adaptation also supported that peroxisome degradation was impaired in all these three mutants. It appears these mutants are defective in an event at or before the fusion of protrusions of vacuolar membrane, a homotypic vacuolar membrane fusion event. I speculated the molecules that are defective in these mutants may either coordinate the fusion steps in a complex or interact in a step by step manner. The proteins may also reside at different compartments such as peroxisomal and vacuolar membrane or in cytosol. In conclusion, gsa2, gsa3, gsa6 and gsa7 are classified in a mutant group with defects in homotypic vacuolar membrane fusion.

PAGE 74

61 In gsa5 cells, during methanol induction, we observed peroxisomes, mitochondrial and cell debris were wrapped with several high density layers of membrane, a hallmark of macroautophagy. This means that macroautophagy has been turned on in this particular mutant during methanol induction, which is unusual. Some of these autophagosomes were in contact with the yeast vacuole. When the cells were examined at 3 hour of glucose adaptation, we observed the same phenomena as in methanol induction. The macroautophagy was apparently still functioning since multiple membrane layers of cell contents could be easily identified. The yeast vacuole has formed membranous protrusions and the sequestration of peroxisomes by vacuolar membrane protrusions was initiated. However, a subsequent step was defective since no peroxisome degradation was observed. The biochemical data showed that AOX and FDH degradation was defective in gsa5 and supported our conclusion that gsa5 was a very distinct mutant defective in microautophagy. I speculated that the defective gene product may inversely control a common event of microautophagy and macroautophagy. Defects in this gene product turned on macroautophagy and interfered with microautophagy process during glucose adaptation. However, this is only a simplified explanation. Although I am aware of this mutants is unique in its phenotypic presentation, I put gsa5 in the same group of vacuolar membrane fusion defect mutants. Further genetic studies of this mutant will reveal the function as well the role of the defective gene product that control the microautophagy and macroautophagy process.

PAGE 75

Figure 3-6. Ultrastructural studies of gsa1-1, gsa2, gsa3, gsa6, gsa7 and gsaS mutants under glucose adaptation. Gsa mutants were grown on methanol medium and then adapted to glucose for 3 hours. Cells were prepared by a potassium permanganate protocol for ultrastructural analysis (Tuttle and Dunn, 1 995). The representative morphology of mutants at 3 hour adaptation is shown here. The detailed description is in the text. N, nucleus; P, peroxisome; V, vacuole and M, mitochondria. Bar: 0.9 jjM.

PAGE 76

63 Figure 3-7. Ultrastructural studies of gsa4 and gsa5 mutants under glucose adaptation. Gsa4 and gsa5 cells were grown on methanol and then adapted to glucose for 3 hours. Cells of Oh and 3h adaptation were prepared by a potassium permanganate protocol for ultrastructural analysis (Tuttle and Dunn, 1995). The representative morphology of gsa4 and gsaS is shown here. The detailed description is in the text. N, nucleus; P, peroxisome; V, vacuole and M, mitochondria. Bar: 0.7|jM.

PAGE 77

64 In conclusion, The morphological and biochemical characteristics in all eight gsa mutants clearly suggested that the gene products defective in these gsa mutants regulate different steps in peroxisome microautophagy. Further genetic studies will reveal their identities and help us to understand the mechanism of autophagy. Chapter Summary In this chapter, I have characterized most of the available mutants in our stocks and eight gsa complementation groups {gsa1 to gsa8) have been identified. Strain gsa1-1, 2, 3, 5, 7, and gsa8 have been backcrossed two to four times and are ready for further genetic studies. Gsa1-1 to gsa8 have also been characterized biochemically and morphologically. All gsa mutants lacked the ability to degrade peroxisomes during glucose adaptation. However, their abilities to degrade peroxisome during ethanol adaptation were not affected. I have divided these mutants into four groups. The first group includes gsa1, which is defective in the initiation of microautophagy. The second group includes gsa4 and gsaS that have selective degradation defect in AOX but not FDH in peroxisome autophagy. A recognition step might be defective in these two mutants. The third group includes gsa2, gsa3, gsa6 and gsa 7. Our data suggested that they are homotypic vacuolar membrane fusion mutants

PAGE 78

65 P. pastoris Methanol Peroxisomes induction i gsal Glucose Signaling i gsa4,gsa8 * Peroxisome Sequestration i gsa2,gsaS,gsa5, ^ Vacuolar Membrane gsa6, gsa7 Membrane Fusion i pra,prb Vacuolar Degradation of Peroxisomes Figure 3-8. Microautophagy of peroxisomes and gsa mutants in P. pastoris. The process of microautophagy of peroxisomes is depicted here based on the biochemical and morphological evidence we had (see text). Mutations in specific GSA genes acted at the specific stages of microautophagy are also indicated. All GSA genes seems act at a step before vacuolar degradation. The vacuolar degradation mutants pra and prb are also shown in the diagram to complete the microautophagy pathway.

PAGE 79

downstream of Gsa1 p. The final group includes gsa5. In this group, macroautophagy is abnormally turned on under both methanol and glucose growth condition while a fusion step of vacuolar membrane protrusions of microautophagy is shut down as in group 3. As shown in the diagram, all mutants are defective at a step before vacuolar degradation. The representative vacuolar protein degradation mutant pra and prb are also included in the diagram. Although all defects in gsa mutants are specific for glucose-induced microautophagy, the other vacuolar degradation pathway(s) might share a common degradation step or, a common regulatory molecules with the microautophagy pathway. An example of this is microautophagy and macroautophagy at least share a common step of vacuolar degradation and in gsa5, the mutated gene product regulates a step in microautophagy buy might also control a step in macroautophagy. The distinct morphological and biochemical phenotypes described in this chapter hint that the defective genes in these gsa mutants may regulate different steps of microautophagy of peroxisomes (Fig. 3-8). The next step after mutant characterization is to identify the defective genes in the mutants and to study the functions of their gene products in the microautophagy process. Some of the mutants in our original possession such as WDY9, 11, 12, 14 and WDY15 showed variable biochemical results and thus we chose not to study these mutants further at this time. Gsa6 cannot be backcrossed to arginine auxotrophs. However, since it is a stable mutant,

PAGE 80

we can rescue it without backcrossing. Gsa8 has a dominant trait and thus classic rescue study is not applicable for it. A clone bank could be made from this strain and the gene can be identified by screens for autophagy defects after transformation. In the following chapters, I will discuss the genetic studies conducted in gsa1-1 and gsa7 mutants that we believed control the steps of initiation and vacuolar membrane fusion events of microautophagy of peroxisome.

PAGE 81

CHAPTER 4 GSA1 PROTEIN IS PFK1 PROTEIN Introduction Our studies presented in chapter 3 showed WDY2 (gsa1-1) is a genetically stable and a morphologically distinct mutant. It probably signals the initiation of peroxisome microautophagy. We chose it as the first mutant for further genetic studies. As mentioned in the chapter 3, we followed the experimental steps (see Fig. 3-1 ) by first backcrossing WDY2 four times. This eliminated the multiple mutations generated during chemical mutagenesis that were not related to gsa1-1 phenotype. We then complemented gsa1-1 with a genomic DNA library to recover the gene that corrected the defect in gsa1-1. In the following two chapters, I will detail my pursuit of Gsa1p. Gsa1-1 Siqnalino an Upstream Event of Microautophacv Microautophaov of Peroxisomes is Defective in asa1-1 As mentioned in chapter 2, I used peroxisome alcohol oxidase (AOX) and cytosolic formate dehydrogenase (FDH) as a quantitative index to monitor 68

PAGE 82

69 CO Pi O K VO h 120 100 80 < H U < m N z m o 60 40 GS115 WDY2 GLUCOSE GS115 WDY2 ETHANOL Figure 4-1 Glucose and ethanol adaptation in parental GS11 5 and WDY2 (gsa1-1) mutant. The production of peroxisomes was induced first by culturing the cells in methanol induction medium. After 24-36 hours, glucose or ethanol was added to the medium at final concentrations of 2% and 0.5%, respectively. Cell free extracts were prepared at 0 and 6 hours of adaptation. Alcohol oxidase (AOX) and formate dehydrogenase (FDH) activities were assayed as described in Materials and Methods and presented as a percentage of the activity measured at 0 hour. The values represent the mean ± s.d. of three or more determinations.

PAGE 83

70 peroxisome microautophagy oi gsa1-1 and GS1 15 during glucose adaptation. As shown in Fig. 4-1, the glucose-induced degradation of both AOX and FDH was dramatically reduced in gsa1-1 when compared with that of parental GS115. Meanwhile, the loss of AOX in gsa^-^ cells during six hours of ethanol adaptation was comparable to that observed in GS1 15 cells. FDH degradation was not observed under ethanol adaptation as in GS115. The data suggested that the mutation in gsa1-1 affected only the glucose-mediated degradation pathway, while its ethanol-mediated degradation pathway proceeded normally. In addition, vacuolar function appeared normal in the gsa1-1 cells since under the ethanol-mediated degradation pathway, the vacuolar degradation of peroxisome is intact. Our previous studies have shown that both degradation pathways shared a common vacuolar degradation step. In conclusion, our biochemical data supported that gsa1-1 is a glucose-mediated peroxisome autophagy mutant with an early event defect. Morphological Study Revealed That a Step Before Peroxisome Sequestration Is Blocked in asa1-1 We next examined the ultrastructure of gsa1-1 cells during glucose adaptation. Large peroxisomes were induced and the yeast vacuole was small and round in parental GS1 15 cells at zero hour of adaptation. When GS1 15 cells were adapted to glucose, the peroxisomes were few in number when compared with 0 hour glucose adaptation. They were found to be associated with the vacuole. The vacuole contained cellular structures such as

PAGE 84

71 peroxisomes and cell debris (Fig. 1-3). In addition, finger-like extensions of the vacuole surrounding peroxisomes were common. In gsa1-1 cells at 0 hour of adaptation, gsa1-1 cells had large peroxisomes that have been induced by methanol while the yeast vacuole was small and round. After three hour of glucose adaption, the peroxisomes were found outside the vacuole in gsa1-1. The vacuole appeared "inactive", that is, it lacked cellular contents, membrane invaginations, and finger-like extensions. Also shown in Fig. 4-2 is the WDY1 (gsa2) mutant. When comparing the morphology of these two mutants, it supported the hypothesis that gsa2 is different from gsa1-1 and Gsa1p should act before Gsa2p. The ultrastructural data and the biochemical data of gsa1-1 and gsa2 suggested that both gene products GSA1 and GSA2 are required for the delivery of peroxisomes to the vacuole for degradation during glucose adaptation. I suggested that microautophagy was mainly shut down in both mutants in early steps of microautophagy and that the limited autophagy observed in our biochemical analysis occurred by inefficient macroautophagy under glucose adaptation condition. The different morphological presentations ofgsa1-1 and gsa2 showed that their gene products control different events of peroxisome microautophagy. My observations supported the conclusion that Gsa1p controls the initiation of microautophagy while Gsa2p controls a downstream event of microautophagy after Gsa1p such as the fusion step of vacuolar membrane during the engulfment of peroxisomes by yeast vacuole.

PAGE 85

72 Figure 4-2. Morphology of gsa1-1 and gsa2 under glucose adaptation at Oh and 3h time points. The ultrastructural analysis of gsa1-1 and gsa2 was done by a potassium permanganate fixation protocol of yeast cells at Oh and 3h of glucose adaptation. At 3h of glucose adaptation, gsa1-1 showed no morphological change when compared to Oh. However, in gsa2, yeast vacuole formed protrusions to wrap the peroxisomes. This is an indication that the defective gene in gsa2 probably controls a step downstream of gsa1-1 in the microautophagy process. M: mitochondrion; N: nucleus; P: peroxisome; V: vacuole. Bar: 0.9 |jM.

PAGE 86

PDLT1 Complemented asa1-1 Phenotvpe 73 The biochemical and morphological studies have suggested that Gsa1p control the initiation step in microautophagy. I chose gsa1-1 for further genetic analysis. The main purpose of genetic study is to recover the gene that is defective in a specific gsa mutant. A P. pastoris genomic DNA library (a generous gift of Dr. J. M. Cregg, Oregon Graduate Institute) was made and put into an episomal shuttle vector pYM8. WDY2 {his4, gsa1-1) was transformed with the genomic library and plated on transformation medium lacking histidine. The resulting transformants were screened by an AOX direct colony assay to test their ability to degrade AOX in gsa1-1 cells (Tuttle and Dunn, 1995). If the cell acquired the GSA1 gene and the gene were transcribed, it should rescue the defect of gsa1-1 and thus regain the ability to degrade AOX. When examined on nitrocellulose by direct colony assay, the rescued colonies should appear white while non-rescued should appear purple. Of the fourteen putative rescued colonies that were isolated by Dr. Dan Tuttle, only two maintained the wild type ability to degrade AOX when assayed in liquid media. The plasmid DNA isolated from these two clones was amplified in E. coli DH5a cells. Both vectors contained a 5.7 kb genomic DNA insert that displayed identical restriction maps. I next tried to verify the ability of this plasmid, pDLT1 , to complement the mutant gsa1-1 cells by reintroducing it back to gsa1-1. A 3.8 kb BamH I fragment was also subcloned into pYM8 (pDLT-BB) and used to

PAGE 87

74 transform gsa1-1. Since the instability of the episomai vector, pYM8, contributed to inconsistent measurements of AOX degradation, we induced stable integration of these constructs prior to measuring their ability to complement the gsa1-1 phenotype through YPD-YND-YPD non-selection and selection media growth procedure. Those gsa1-1 cells stably integrated with pDLT1 and pDLT-BB efficiently degraded AOX during glucose adaptation (Fig. 4-3 and 4-4) in a manner comparable to GS115. This result showed that 5.7 kb insert containing pDLT1 could restore AOX degradation ability of gsa1-1 cells. Rescue Study of Different Fragments in the Insert of pDLT1 We next tried to further define regions of the 5.7 kb DNA insert in pDLT-1 that was responsible for complementing the defect of gsa1-1 (Fig. 4-3 and Fig. 4-4). Restriction fragments of the 5.7 kb genomic DNA were subcloned into pYM8 and pYM4 (Fig. 2-1 ). Using these constructs, stable transformants of gsa1-1 were isolated and their ability to degrade AOX during glucose adaptation examined. Stable integration of the pDLT1 constructs would presumably occur at the locus homologous to the genomic DNA insert, since pYM8 contained only 164 nucleotides of DNA homologous to P. pastoris. A 3.8 kb BamH I fragment (pDLT-BB) rescued the gsa1-1 phenotype, while the DNA insert without this 3.8 kb fragment (pDLT-AB) did not complement the mutation. I also examined whether smaller fragments from within this 3.8 kb region could

PAGE 88

75 rescue gsa1-1 or not. My results showed that all fragments tested such as the 2.3 kb Hind III / BamH I fragment (pDLT-HB) did not rescue gsa1-1 and the minimum rescue fragment is 3.8 kb BamH I fragment. Since pDLT-HB and pDLT-XB did not rescue gsa 1-1, the critical rescue part probably lies between BamH 1(644) to Hind III (2260) sites. Once the smallest rescue DNA fragment was identified, the next step was to sequence this 3.8 kb fragment. The result of sequence is shown in next section. Sequencing and Sequence Analysis of pDLT1 Insert The 3.8 kb BamH I fragment was subcloned into pBluescript and sequenced in both directions by nested deletion methods. The sequence revealed one long open reading frame without a start codon and five short open reading frames encoding proteins of 70 1 15 amino acids (Fig. 4-4). The short open reading frames did not match any amino acid sequences in protein databases accessible to NCBI, but the large open reading frame showed close amino acid homology to the a-subunit of phosphofructokinase (Pfk1p) of S. cerevisiae and Kluyveromyces lactis. We next sequenced the initial 644 nucleotides of the genomic DNA insert of pDLT-1 . The start codon of PFK1 was identified in addition to 161 base pairs of non-coding 5' upstream DNA. No additional open reading frames were revealed.

PAGE 89

76 Figure 4-3. pDLT1 rescues the gsa1-1 phenotype. pDLT1 was recovered by Dr. Dan Tuttle by transforming a genomic DNA library of P. pastoris and then screened the rescue colonies that appeared white on the direct colony assay. The recovered plasmid was identified as pDLT1 and it rescued gsa1-1 by reintroducing it into gsa1-1 cells. A 3.8kb BamH I fragment cut from pDLTI's 5.7 kb insert also rescued the mutant phenotype in vector pYM8 (pDLT-BB) but not pYM4 (pWP-BB). The rescue result of a full length PFK1 gene (pWP-PFK) is also shown here. A Hind III BamH I fragment (see next figure) did not rescue gsa1-1. AOX activities were tested in a six-hour glucose adaptation. The values of AOX activities were represented as a percentage of 0 hour values. All assays were measured at least three times and the values shown on the graph represent the average ± sd.

PAGE 90

77 Xhol (3191) BamH I Spe I Hind III Spe I (644) (1598) (2260) (3206) BamH I (4512) ^ < (162) PFK 1 (3134) 5.7 kb Vector pDLT-1 pDLT-AB pDLT-BB pDLT-HB pDLT-AS pDLT-SS pDLT-XB Rescues WDY2 Yes No Yes No No No No pWP-PFK pWP-BB pWP-HB Yes No No Figure 4-4. Identification of the genomic DNA that rescues gsa1-1 during glucose adaptation. Genomic DNA fragments were subcloned into pYM8 (pDLT) or pYM4 (pWP) vectors and used to stably transform gsa1-1. GS115, gsa1-1 and transformed gsa1-1 cells were grown on methanol for 24-36 hours then adapted to 2% glucose medium. Cell free extracts were prepared at 0 and 6 hours of adaptation and assayed for alcohol oxidase (AOX) activities. Glucose-induce degradation of AOX in those gsa1-1 cells that had been rescued by complementing DNA was rapid and comparable to parental GS115 cells (e.g., pDLT-BB). However, when gsa1-1 cells were transformed with noncomplementing DNA, AOX degradation was slow and comparable to nontransformed gsa1-1 mutants (e.g., pDLT-HB).

PAGE 91

78 The smallest complementing DNA had five short putative open reading frames with start codons. Three of these open reading frames resided within pDLT-HB while two resided in pDLT-SS. Neither construct was capable of rescuing gsa1-1 (Fig. 4-3). Since we were stably integrating the genomic DNA prior to testing its ability to complement gsa1-1, the smallest complementing DNA does not necessarily represent the entire open reading frame of the rescue gene in gsa1-1. Indeed, we suggest that the 3.8 kb BamH I fragment complemented gsa1-1 by correcting the mutation upon homologous recombination into the PFK1 locus. In a parallel experiment, I also integrated the complementing DNA fragments into the HIS4 locus o^gsa1-1 and tested their ability to rescue gsa11. This was done by constructing the 2.3 kb Hind III / BamH I fragment, the 3.8 kb BamH I fragment and the complete PFK1 gene including 161 nucleotides of 5' upstream DNA into pYM4, which contains a P. pastoris HIS4 gene. Prior to transformation these vectors were linearized by a single cut within the HIS4 gene to direct their integration to the HIS4 locus. Under these conditions, neither pWP-HB nor pWP-BB rescued gsa1-1 (Fig. 4-4). We propose that the inability of pWP-BB to complement was due to the absence of the PFK1 start codon when the vector integrated into the HIS4 locus. Indeed, when the start codon was present, PFK1 (pWP-PFK) complemented the gsa1-1 phenotype (Fig. 4-4). This experiment showed that PFK is responsible for the rescue effect of gsa1-1. Our morphological studies of mutants and their transformants

PAGE 92

79 showed that PFK restored the ability of gsa1-1 cells to degrade peroxisome (Fig 5-5 and Fig 5-6). The results that Pfk1 p might be Gsa1 p was a surprise for me since PFKp has identified function in glycolysis. Before I could reach the conclusion that Pfkl p is Gsa1 p, I needed more data about gsa1-1. Northern Blot Analysis of Qsa1-1 Mutants At the time of identifying genomic DNA that rescues gsa1-1, I also tried to use Northern analysis to 1) identify possible open reading frames that might transcribe in this 3.8 kb region and 2) observe the message RNA (mRNA) level of PFK1 in different yeast strains. A 2.5 kb BamH I / Xho I DNA fragment inside the pDLT insert was used as a probe to identify on Northern blots those mRNA's of GS1 15, WDY1 (gsa2) and WDY2 {gsa1-1) that were being transcribed in situ. One predominant mRNA at 3.2 kb and two minor mRNA's at 2.7 and 1 .5 kb were identified in GS1 1 5 cells (Fig. 4-5). The 3.2 kb mRNA matched the predicted size for PFK1 mRNA. Neither 2.7 kb nor 1 .5 kb mRNA corresponded with any predicted open reading frames within the PFK1 locus (Fig. 4-4). We believed they bound to the membrane nonspecifically. The 3.2 kb mRNA was also present in Sfsa2, but was greatly diminished in gsa1-1 cells. These results could not be attributed to sample loading since the amounts of 18S and 28S rRNA loaded onto the lanes were comparable. It is possible that the 3.2 kb mRNA was neither transcribed nor stable in gsa1-1 cells. A mutation

PAGE 93

80 within the promoter region may suppress transcription, while a premature stop codon would terminate translation causing a destabilization of the mRNA. Our studies suggested that the mutation was not within the upstream 5' region but within the PFK1 encoding region, since pDLT-BB which lacked bases -161 through +483 of the PFK1 gene was able to rescue gsa1-1 when stably integrated (see Fig. 4-4). This experiment showed that PFK1 is the only transcribing gene in this region and PFK mRNA was greatly reduced in gsa1-1 cells and the further evidence that PFK1 gene is related to gsa1-1 mutation. Phosphofructokinase Activity Assay in asa1 The absence of the 3.2 kb PFK1 mRNA in gsa1-1 led me to measure the phosphofructokinase activities in these cells (see Materials and Methods). We found that WDY2 {gsa1-1) cells had less than 15% of the phosphofructokinase activity present in parental GS1 15 cells whereas WDY1 {gsa2) cells had comparable PFK activity level to that of GS1 15 (Fig. 4-6). The data were consistent with GSA1 encodes Pfkip and gsa1-1 is a pfk1 mutant. Phosphofructokinase is a heterooctamer composed of four a-subunits encoded by PFK1 and four (3-subunits encoded by PFK2 in yeast cells (Berger and Evans, 1992, Arvanitidis and Heinisch, 1994). The P. pastoris PFK1 gene is 2.97 kb long encoding a protein of 989 amino acids with a predicted molecular mass of 108.7 kDa (Fig. 4-7). The PFK protein catalyzes the

PAGE 94

Figure 4-5. Northern blot analyses of GS1 15, WDY1 (gsa2), and WDY2 {gsa1-1) cells. Cells were grown in YPD medium, harvested and total RNA prepared. The RNA was separated on an agarose gel and transferred to a nylon membrane. After staining with methylene blue to identify rRNA (lower panel), the membrane was probed with a ^^P-labeled 2.5 kb BamH I / Xho I fragment from the original 5.7 l
PAGE 95

Phosphofructokinase Activity in gsa Mutants 500 1 400 00 > *••-» Q < 200 u B >> N S W 100 0 GS115 gsa2 gsal-1 gsal-2 gsal-3 Figure 4-6. PFK activity is greatly reduced in WDY2 {gsa1-1). Since the sequence data and Northern blot showed that PFK1 might be GSA1, we measured the activity of PFK protein. The parental GS115 and gsa2 showed normal PFK activity while the activity of PFK protein in gsa1-1 reduced greatly (less than 10%). This clearly showed that we recover the right gene. The PFK activities of gsa1-2 and gsa1-3 are also included. Interestingly, the activity levels in these three gsa1 mutants are totally different. All assays are measured at least three times and the values shown in the graph represent the average ± sd.

PAGE 96

83 conversion of fructose 6-phosphate to fructose 1 ,6-bisphosphate. The enzyme controls one of the rate limiting steps in glycolysis and is highly regulated by many metabolites including fructose 2,6-bisphosphate, AMP and citrate. The catalytic and regulatory sites of the Pfk1 protein are highly conserved between P. pastoris, S. cerevisiae and H. sapiens (Fig. 4-8). For example, there exists 60-85% amino acid identity within the fructose 6-phosphate binding site, the ATP binding site, the fructose 2,6-bisphosphate binding site, the AMP binding site, and the citrate binding site (Arvanitidis and Heinisch, 1994). Interestingly WDY3 {gsa1-2) and WDY8 {gsa1-3) previously identified as the same gsa group of WDY2 (gsaf-"/) also showed decrease of PFK activity. This also verified that they belong to the same gsa1 complementation group. However, gsa1-2 showed only a two-thirds decrease of the activity when compared with GS115 cells and in gsa1-3, about 20% activity remained. I speculate that different activity levels of PFK reflected different missense or nonsense mutation sites in these mutants. For example, in gsa1-2 and gsa1-3, the mutation may well be missense but at important positions such as a substrate binding site or an ATP binding sites. If PFK protein in P. pastoris is also composed of aand P-subunits as in S. cerevisiae, the mutations in gsa1-2 and gsa1-3 could also be in PFK2 gene. These strains are very useful for future studies. Based on the ability of genomic DNA fragments to rescue the gsa1-1 cells and the absence of PFK1 mRNA, we speculate that the gsa1-1 mutant has a premature stop codon near the N-terminus of the PFK1 gene. This would

PAGE 97

84 result in the synthesis of a truncated mRNA which may lead to the instability of PFK1 mRNA as we observed on our PFK mRNA study. The translation could also be prematurely stopped. Carbohydrate metabolism is complicated in yeast and some of the pathways are still not fully understood because different yeast strains show different pathways and abilities to metabolize sugars (Wills, 1990, Wills, 1996). However, evidence has accumulated that S. cerevisiae can grow on glucose in pf/c mutants (Lobo and Maitra, 1983, Schmitt, et al, 1984). Glucose is metabolized via the pentose phosphate pathway in pfk mutants (Schmitt, et al, 1984, Heinisch and Zimmermann, 1985, Jacoby et al, 1993, Boles, et al, 1993). This pathway then joins the glycolysis at glyceraldehyde-3-phosphate level (Wills, 1990). This may also be the case in P. pastoris. Sibirny et al. (1987) found that when Pichia pinus cells were grown on methanol, the enzymes for methanol utilization were greatly induced. The subsequent glucose induced catabolite repression of transcription of methanol metabolism enzymes is controlled by the glucose catabolite repression gene 1 {GCR1). By using glycolysis enzyme assay, they found that the gcr1 mutant has low level of PFK activity. They have evidence that in gcr1 mutant, the loss of repression was not due to the damage of the glycolysis pathway. However, no follow-up studies of this gcr1 mutant can be found. This was an indirect evidence that Pfkip might participate in glucose repression of AOX protein via autophagy.

PAGE 98

85 IGl GGATCTTTCTTCTCTTGCTATAAATCAMCATTATTCATACAGAGTTTAATCGA 107 TTCAACAAAGCATAACATTGATTGCAATTGGTTCCGTCTTGAACTGCTAAGGAG 63 AACAATCATAAATTAGTATTTTGTTTGCTTGTTAGACTCAAATCGAATTACAGA 1 " 2 TGCCAGAACCATCTATAAGTGCACTTTCCTTCACTTCGTTTGTCACTAATGATG 2 PEPSISALSFTSFVTNDD 56 ACAAACTGTTTGAAGAGACTTTCAATTTTTACACGAAGTTGGGCTTCCACGCAA 20 KLFEETFNFYTKLGFHAT 110 CACGCTCATATGTTAAAGACAACCGGTCAGACTTTGAATTGACGGGGATTTCCA 3B RSYVKDNRSDFELTGIST 164 CGGATTCAATCAAGGAAATCTGGCTGGAAAGTTTCCCACTATCTGAAGTGGTCG 56 DSIKEIWLESFPLSEVVE 218 AAACGTCAGCTGGTAGAGAGTTGAGAAAACCACTGCAAGAATCTGTGGGCTACC 74 TSAGRELRKPLQESVGYQ 272 AATCTGAAGCTCTTCTGGGATATTCTCCCTACCAGAGTGACGGTGTTGTTATAA 92 SEALLGYSPYQSDGVVIK 326 AATTAAGGTTATCAAATCATGACCTTCAGAAAAACAAAGACTTGCCCGGTGAAG 110 LRLSNHDLQKNKDLPGEV 380 TTACGTTTTTCACCGCTAGTATCGACAAATTAAGAGCTAAACTCATTGAAATTG 128 TFFTASIDKLRAKLIEIG 434 GTGCTGAGATAATTCCCTCAGAAATAGACCTTGTTGAATTTTCAACCAAGGATC 146 AEIIPSEIDLVEFSTKDP 488 CTATGGGCGACGTCATTAGCTTTTCTTCTTATCCCTCTTTGAGTTCCAAGAAGA 164 MGDVISFSSYPSLSSKKK 542 AGATTACCTCTCCAGACTTTTTCCTCCACCCTAAGAAGGAAGTACGCTCCCAAG 182 ITSPDFFLHPKKEVRSQE 596 AATCAATAGTTGAGCAGGTTAAATCTGAAGAAGGTAAGAAGAAGATTGCCATCA 200 SIVEQVKSEEGKKKIAII 650 TAACTTCAGGTGGAGACGCACCGGGAATGAATGCTGCAGTAAGGGCTGTGACAA 218 TSGGDAPGMNAAVRAVTR 704 GAGCCGGTATTTTCTATGGCTGTAAAGTTTACGCTTGTTATGAAGGTTACACTG 236 AGIFYGCKVYACYEGYTG 768 GACTGGTTAAGGGTGGTGATATGTTAAAGGAACTGCAGTGGCAAGATGTCCGTG 264 LVKGGDMLKELOWODVRG 812 GTTTACTTTCCATTGGTGGTACCATAATTGGTACTGCAAGAAGTAAGGAATTCA 272 LLSIGGTIIGTARSKEFR 866 GAGAACGATGGGGCCGTCTTCAAGCTTGCTACAATATGGTCAGCAATGGTATTG 290 ERMGRLQACYNMVSNGID 920 ATGCGTTAGTTGTTTGTGGAGGTGACGGATCTCTTACAGGTGCCGATCTATTTC 308 ALVVCGGOGSLTGADLFR 974 GAAATGAATGGCCTGAACTGATAAAGGAACTTTTGGGTGAGGGCAAAATTACAA 326 NEHPELIKELLGEGKITK 1028 AAGAACAATATGAAACACACAGAAACTTGACAATCGTAGGTCTCGTTGGTTCTA 344 EQYETHRNLTIVGLVGSI 1082 TCGATAACGATATGTGCGGAACTGATTCCACAATTGGTGCTTATTCCTCATTGG 362 DNOMCGTDSTIGAYSSLE 1 136 AGAGAATCATAGAGCTGGTAGACTACATCGATGCTACTGCCGCCTCCCATTCAC 388 RIIELVDYIDATAASHSR 1190 GAGCCTTCGTGGTGGAAGTCATGGGTAGACATTGTGGATGGTTAGGTTTAATGT 398 AFVVEVMGRHCGWLGLMS 1244 CCGGAATTGCTACTGGAGCTGATTACATTTTCATCCCTGAAAGACCTCCAAGTG 416 GIATGADYIFIPERPPSE 1298 AAACAAACTGGAAGGACGACTTGAAGAAAGTCTGTTTGAGACATAGAGAGAAAG 434 TNWKDOLKKVCLRHREKG 1352 GACGCAGGAAGACCACCGTTATTGTTGCTGAAGGTGCTATTGATGACCAACTGA 452 RRKTTVIVAEGAIDDQLN 1406 ACCCTATCACTTCTGAAGAGGTGAAAGATGTACTAGTGGAGATTGGTTTGGACA 470 PITSEEVKDVLVEIGLDT 1460 CTCGTATTACCCGTCTAGGACATGTCCAAAGAGGTGGAGCTCCGTGTGCTTTTG 488 RITRLGHVQRGGAPCAFD 1514 ATAGATTCTTGGCCACTGTTCAAGGTGTTGATGCTGTTAGGGCTGTTTTAGAAA 606 RFLATVQGVDAVRAVLES 1568 GCACCCCAGCAATTCCTTCTCCTGTCATCAGCATTTTGGAGAACAAAATTGTTC 524 TPAIPSPVISILENKIVR 1622 GCCAGCCGTTGGTGGAATCTGTTGCTCAAACAAAGACTGTCAGTGATGCTATCG 542 QPLVESVAQTKTVSDAIE 1676 AGGCCAAGGATTTCGATAAAGCTTTGAAATTAAGAGACCAAGAGTTTGCCACAT 560 AKDFOKALKLROQEFATS 1730 CATATGAGAGCTTCCTGTCCGTTTCCAAGTATGACGATGGATCATATCTAGTAC 578 YESFLSVSKYDDGSYLVP 1784 CAGAGAGCTCAAGATTAAATATTGCCATCATCCATGTGGGAGCTCCAACATCTG 596 ESSRLNIAIIHVGAPTSA 1838 CGTTGAATCCTGCCACAAGAGTTGCTACTTTGAACTCGTTGGCAAAAGGACACA 614 LNPATRVATLNSLAKGHR 1892 GAGTTTTTGCTATTCGAAACGGATTTGCAGGATTAATTCGCCACGGCGCTGTAC 632 VFAIRNGFAGLIRHGAVR 1946 GAGAGCTCAACTGGATTGATGTTGAGGACTGGCACAACACAGGTGGGTCGGAGA 660 ELNWIDVEDWHNTGGSEI 2000 TTGGCACCAACAGAAGTCTTCCTAGTGATGATATGGGCACTGCGGCGTACTACT 668 GTNRSLPSDDMGTAAYYF 2054 TCCAGCAATACAAGTTTGATGGTCTTATTATTATCGGNGGATTTGAAGCTTTCA 684 QQYKFDGLIIIGGFEAFT 2108 CAGCTCTGTACCAGCTGGACGCAGCTCGCGCTCAGCATCCTATCTTCAATATTC 704 ALYQLDAARAQHPIFNIP 2162 CAATGTGTTGNCTTCCAGCTACTGTTTCTAATAACGTTCCTGGTACCGAGTATT 722 MCXLPATVSNNVPGTEYS 2216 CCTTAGGGTCTGACACATGTCTAAACACCTTGTCTGGATACTGTGATGCTGTGA 740 LGSDTCLNTLSGYCDAVK 2270 AACAATCTGCTTCTGCTAGTAGAAGAAGAACATTTGTTGTGGAAGTTCAAGGTG 768 QSASASRRRTFVVEVQGG 2324 GATACTCAGGATATCTTGCCAGCTACGCTGGTCTGATCACAGGAGCTTTGGCTG 776 YSGYLASYAGLITGALAV 2378 TTTATACTCCTGAAAACCCAATCAACCTTCAAACAGTGCAGGAAGACATTGAAT 794 YTPENPINLQTVQEDIEL 2432 TGTTGACTCGAACATACGAGGAAGACGATGGTAAGAACAGATCGGGTAAAATCT 812 LTRTYEEDDGKNRSGKIF 2486 TTATTCATAATGAAAAGGCTTCAAAGGTTTACACCACGGATCTGATTGCTGCTA 830 IHNEKAS KVYTTDLIAAI 2640 TCATAGGTGAAGCTGGAAAGGGTACGTTTGAGAGCCGTACTGCCGTGCCTGGTC 848 IGEAGKGTFESRTAVPGH 2594 ATGTACAACAGGGTAAATCTCCCTCATCTATTGACCGGGTTAATGCCTGCAGAC 866 VQOGKSPSSIDRVNACRL 2648 TGGCTATCAAATGTTGTAACTTCATCGAGGACGCCAATTTCCAGGTGAAACACA 884 AIKCCNFIEDANFOVKHN 2702 ATGCCAATTTGAGCGCCGACGAACGTCATTTGAGATTCTTTTGCGATGACGGAG 902 ANLSADERHLRFFCDDGV 2756 TTAAGACATCTGCAGTGAGCGGCAAATCTTCCGTGATAGATGATAACACGTCAG 920 KTSAVSGKSSVIODNTSV 2810 TGGTCATTGGAATCCAAGGTTCCGAGGTTACATTCACTCCTGTAAAACAGCTAT 938 VIGIQGSEVTFTPVKQLW 2864 GGGAGAAGGAAACTCATCATAAGTGGCGAAAGGGTAAGAACGTTCATTGGGAGC 956 EKETHHKHRKGKNVHUEO 2918 AGTTGAACATTGTCTCTGACCTCTTGAGTGGTCGTTTGTCTATTCGTACCACGT 974 LNIVSDLLSGRLSIRTT* 2972 AAAAGACGGATCAAATCGGTTGTTTGGGTACTAAAGACAATCCATTTTTTTTCT 3026 TNCTCTCGAGCTGGATGAAACTAGTGCATGTACGAATCCGCGTGTAATCTACTG Figure 4-7. Nucleotide and predicted amino acid sequences of P. pastoris PFK1 gene. The 5' upstream region in the original 5.7 kb clone included 161 bases. The Pfk1 protein is predicted to be 989 amino acids with a molecular mass of 108.7 kD. The sequence of PFK1 has a GenBank accession number of U73376.

PAGE 99

PP MPEPSISALSFTSFV. .TNDDKLFEETFNFYTKLGFHATRSYVKDNRSDFELTGIST. .D 56 SC -QSQDSCYGVAFRSIITNDEALFKKTIHFYH-LGFATV
PAGE 100

Glycolysis Pathway and asa Mutants 87 Glucose signals the onset of peroxisome degradation in the glucose adaptation pathway. It may either directly send a signal to initiate the onset of microautophagy of peroxisomal and cytosolic proteins or exert that effect through a subsequent glycolysis enzyme or product. The discoyery that GSA1 is PFK1 prompted me to probe this question in gsa1-1 as well other gsa mutants. I want to know If glucose is the direct signaling molecule and if other carbon sources could signal the initiation of peroxisome degradation. We asked ourselyes which entry point in the glycolysis pathway is related to the autophagy signaling process and if this is related to Pfk1 function in the glycolysis pathway. By proyiding different carbon sources for yeast to grow, we found that P. pastoris could utilize fructose, mannose, glycerol and pyruyate as the sole carbon sources for growth. Howeyer, when these carbon sugars were used as adaptation carbon source, we obserye pyruyate could not initiate peroxisomal AOX degradation in parental GS1 15 strain while fructose, mannose and glycerol could do the job as efficiently as glucose in GS115 (Fig 4-9 and Fig 4-10). Pyruyate, fructose and mannose could not initiate microautophagy in gsa1-1 and gsa2 mutants while glycerol did initiate the degradation process in GS1 15 cells as well as in gsa1 and gsa2 mutants. In a PFK1 knockout, a null mutant of gsa1-1 1 isolated later in the experiment also could undergo glycerol induced autophagy (see chapter 5).

PAGE 101

88 These experiment data suggested that 1 ) Glucose is not the molecule to directly initiate microautophagy since other carbon sources could also do so; 2) A signal was sent out to initiate microautophagy during glycolysis before pyruvate step; 3) The signal might be sent out between glycolysis steps such as fructose-1 -phosphate to glyceraldehyde-3-phosphate. Pfk1p is in the middle of these steps and it may well be the initiation signal. However, I speculate that Pfk1p may direct either signal the microautophagy or initiate the microautophagy process through a downstream molecule (not necessarily in the glycolysis pathway). In Fig. 4-10, AOX could be degraded in all mutants when glycerol was used as an adaptation carbon source. It is possible that glycerol addition initiates another yet to be identified peroxisome autophagy pathway. This pathway may share some of the steps that have been defined previously such as vacuolar degradation. In Sibirny's study (1987), they found carbon sources such as glucose, fructose, mannose and galactose repressed the synthesis of alcohol oxidase in the wild type yeast P. pinas. The repression with these sugars was impaired in a gcr1 mutant which was subsequently identified containing a defective Pfk enzyme. However, glycerol repression proceeded normally in this mutant. These findings corroborated our view that glucose is not the direct signal molecule of microautophagy and the carbon repression is related to Pfk protein. However, a downstream enzyme or glycolysis product is still possible to be responsible for the initiation of autophagy process. Further studies are needed to reach a conclusion.

PAGE 102

89 Carbon Source Entry Points in Glycolysis Pathway Glucose GK, HK, ZWF, PGI G-6-P > Gluconate-6-P PGI ^ FructOSe-6-P < Mannose PFK I FructOSe-l,6-P Fructose I ALD, TIM I T >' Glyceraldehyde-3-P < Glycerol GDH, PGK, PGM 2-Phosphoglycerate ENO, PK ^ Pyruvate < Pyruvate ETOH TCA cycle Figure 4-9. Glycolysis pathway and the entry point of carbon metabolites. The flow chart of glycolysis is shown here and the carbon source entry points are also included. The results of induction of peroxisome degradation by these carbon sources are shown in the next figure.

PAGE 103

90 Adaptation after MIM with Different Carbon Sources Glucoce Mannose H Fructose ^Glycerol B Pyruvate osiis Different Strains Figure 4-1 0. Metabolites in the glycolysis pathway and their induction of the degradation of AOX. GS1 1 5 cells and mutants were grown on methanol induction media for 24 to 36 hours and then were subjected to adaptation sources such as glucose, glycerol, mannose, fructose and pyruvate. Cell extracts were collected at zero hour and six hours after adaptation. A comparison of AOX activities before and after 6h glucose adaptation was presented here. Values were presented as an average of percentage of the zero hour activity ± sd.

PAGE 104

Glycolysis and psa Mutants Since GSA1 has been identified as PFK1, the question raised by this finding is if other gsa mutants are also defective in PFK or other glycolysis enzymes. To approach this questions, I screened all my existing mutants with a PFK activity assay. WDY3 which has been previously identified as gsa1-2 showed a 70% decrease of its PFK activity (see Fig. 4-6). Another mutant WDY8 also showed 80% decrease of its PFK activity. All other mutants had comparable PFK activity level as parental GS1 15. Next I tested the activities of some other glycolysis enzymes in some of the existing gsa mutants (see Materials and Methods). The assay included other two main regulatory step enzymes, hexose kinase and pyruvate kinase. The results are shown in Fig. 4-1 1 . When these mutants were examined with available enzyme assays, except for gsa1-1, no significant difference was seen in these enzyme activity levels when compared with parental GS1 1 5. In gsa11 mutant, PFK activity was greatly reduced by 90% as expected. However, the activity of phosphoglycerate mutase (PGM) of gsa1-1 increased three fold when compared with parental GS115 and other mutants. The pentose shunt pathway has been identified as an alternative pathway for glycolysis to continue in pfk mutants (Heinisch and Zimmermann, 1985, Jacoby, et al., 1993). The cells circumvented the PFK action step in the glycolysis pathway via the pentose shunt pathway to enter glycolysis via glyceraldehyde-3-

PAGE 105

92 Glycolysis Enzyme Assay of gsa Mutants Glucose GK, HK, PGI 1 Frucose-6-P PFK i Fructose-l,6-P ALD, TIM 1 Glyceraldehyde-3-P GDH, PGK, PGM 1 2-Phosphoglycerate ENO, PK Pyruvate GS115 gsal gsaS gsa6 gsa7 gsaS GK* 387 358 307 358 261 282 HK 153 126 149 119 160 141 PGI 1440 1679 1433 1666 1174 1134 PFK 404 16.S 433 434 421 351 ALD 602 1204 853 942 947 735 TIM 20 16 11 14 18 16 PGM 1129 4420 1175 1238 920 735 ENO 116 99 67 113 49 75 PK 645 593 417 479 517 489 I * All enzyme acssy valnes are represented as mU/mg TCA cycle Figure 4-1 1 Glycolysis enzyme activities and gsa mutants. Glycolysis enzyme assays were applied to test possible glycolysis mutants in our possession. The principle of assays is based on a coupled enzyme assay resulting in the oxidation or reduction of NADH/NAD or NADPH/NADP (see Materials and Methods). The enzymes tested are GK: glucokinase; HK: hexokinase; PGI: phosphoglucose isomerase; PFK: phosphofructokinase; ALD: aldolase; TIM: Triosphosphate isomerase; PGM: phosphoglycerate mutase; ENO: enolase; PK: pyruvate kinase. The values were an average of three measurements.

PAGE 106

93 phosphate. This happens two steps upstream of PGM where it acts. Since glycolysis is altered in gsa1-1, I suspected that the change of a particular glycolysis product concentration level may send an up-regulation signal for the transcription of PGM. Gsa1-1 has been backcrossed and it is unlikely that a mutation in PGM gene promoter region up-regulates the expression of PGM. PGM has been found to undergo rapid dephosphorylation during trichocyst exocytosis in Paramecium tetraurelia (Treptau et al.,1995). A protein phosphatase/kinase system was suspected to be involved in the modification of PGM and thus the regulation of exocytosis in P. teraurelia cells (Kissmehl, et al., 1997). However, the role of PGM in this exocytosis is still under investigation and debate. Nevertheless, it is surely worth investigating the role of PGM in the autophagy process. The study of glycolysis enzyme assays in these gsa mutants showed clearly that the tested gsa mutants except for gsa1 did not involve the tested glycolysis steps. Future genetic studies of the remaining gsa mutants will yield useful information about the glucose induced microautophagy. Chapter Summary WDY2 (gsa1-1) strain was used for the biochemical, morphological as well as genetic studies. My result showed gsa1-1 is defective in peroxisome degradation during glucose adaptation after methanol induction. Its

PAGE 107

peroxisome degradation proceeds normally during ethanoi adaptation. The morphological evidence supported that gsa1-1 is defective in the initiation of microautophagy of peroxisomes. By complementing gsa1-1 with a genomic DNA library, we were able to recover the plasmid pDLT1 that rescued the gsa1-1 mutant phenotype. The minimum fragment that rescued gsa1-1 is a 3.8kb BamH I fragment (BB) within pDLTl The reintroduction of pDLT1 and pDLT-BB into gsa1-1 confirmed its rescue effect. The subsequent sequencing showed that the main open reading frame codes the PFK1 gene. Phosphofructokinase activities as well as PFK1 mRNA levels were greatly reduced in gsa1-1 cells. This verified that our cloning of pDLT1 is correct and gsa1-1 is pfk1. However, we knew that Pfk1p is a glycolysis pathway enzyme with defined function. It is a heterooctameric protein composed of 4 a-subunits and 4 P-subunits, a key regulatory enzyme in the glycolytic pathway. The glycolysis enzyme assays of mutants showed other gsa mutants are not directly linked to glycolysis. The carbon source adaptation experiments I finished showed the microautophagy signal is sent out between fructose-1phosphate and glyceraldehyde-3-phosphate. In order to better define the role of Pfk1 p in the peroxisome microautophagy process, I need to test if Pfk1 p is necessary for the initiation of microautophagy upon glucose addition. I also need to test if the effect of Pfk1 p is a direct one or an indirect effect by simply restricting the conversion of fructose-1 -phosphate to fructose-1 ,6-bisphosphate or by up or down regulating other molecules that are specific to peroxisome

PAGE 108

microautophagy. If Pfk1p has a direct role in microautophagy, I also need to identify what the functional role of Pfk1 p is in peroxisome autophagy. To approach these questions, I first tested if PFK1 is necessary for the peroxisome autophagy by generating a PFK1 knockout strain Apfk1 and comparing its phenotype with that of gsa1-1. I also examined the function of Pfk1p in glycolysis and its relationship to peroxisome microautophagy. I will present the results of these studies in the following chapter.

PAGE 109

CHAPTER 5 PFK1P IS REQUIRED FOR THE INITIATION OF PEROXISOME MICROAUTOPHAGY Introduction In the previous chapter, I have shown that Pfk1p is Gsa1p. However, the question remained unanswered as to what the role of Pfk1p is in the peroxisome microautophagy. It is possible that Pfk1p simply assists in the conversion of glucose to a downstream true effector in glycolysis. It is also possible that PFK protein directly regulates the microautophagy of peroxisomes. Our previous results left these two possibilities still open. As a first step to probe this problem, I tried to verify that the PFK1 gene is essential for the microautophagy of peroxisomes. I used gene disruption method to knockout the PFK1 in Pichia pastoris and to compare its biochemical and morphologic profiles with that of gsa1-1. Then I mutated the catalytic site of the Pfk1 protein and tested the ability of this mutant pfk1 gene to complement Apfk1 cells. I tried to define the relationship between PFK activity and its role in glycolysis with its possible role in peroxisome microautophagy via this experiment. The following sections are the results of these studies. 96

PAGE 110

97 Degradation of Peroxisomes by Microautophagy Requires PFK1 PFK1 Gene Disruption and the Verification A 5.6 kb PFK1 "knockout" fragment was constructed by replacing an 1 .2 kb Hind III fragment in PFK1 with a 3.0 kb S. cerevisiae ARG4 (Fig 5-1 , upper and middle panel). PPF1 {his4, arg4) was transformed with this knockout fragment and stable transformants were selected on a transformation plate with histidine supplement but lacking arginine. We expected to find a 5.6 kb DNA fragment on Southern blots if pfk1::ARG4 was inserted into the PFK1 locus. In the parental strains and the gsa mutants, we found a 3.8 kb BamH I fragment that hybridized to our PFK1 probe. However, the 3.8 kb BamH I fragment in the "knockout" WDK01 {his4, pfk1::ARG4) had been replaced by the expected 5.6 kb BamH I "knockout" fragment (Fig. 5-1, lower panel). Interestingly, another transformant PPF1 1-1 had both 3.8 kb and 5.6 kb DNA locuses suggesting that the pfk1::ARG4 did not incorporate into the PFK1 site but instead into an unknown second site in this particular transformant. The 5.6 kb fragments were detected using a S. cerevisiae ARG4 probe (data not shown). WDK01 was thus named as Apfk1 or Agsa1. The PFK1 Knockout Has the Same Phenotype as asa1-1 The phosphofructokinase activity in WDK01 was 15% of that measured in parental PPF1 (Fig. 5-1 ). This activity was comparable to that measured in gsa1-1 mutant. The presence of normal PFK1 in PPF11-1 was substantiated

PAGE 111

98 Bam HI (644) Hindlll (1048) Hindlll (2206) BamHI (4512) 3.8kb PFKl S. cerevisiae A RG 4 gene Bam HI (644) HindUI (1048) HindUI (4387) BamHI (6288) I I o Q «^ r! p 1 o ^ ^ 5.6kb pmi::ARG4 5.6kb pflil::ARG4 i 3.8kb PFKl 300 58 44 522 660 36 565 PFK activities (mU/mg) Figure 5-1 . Generation of the PFK1 knockout. The knockout fragment (5.6 kb) was prepared according to Materials and Methods. PPF1 was transformed with this fragment and stable Arg* transformants were isolated. The genomic DNA of the knockout (clones 1 & 2) and other cells were isolated and digested with BamH I for Southern blot. The expected 5.6 kb DNA band was present in WDK01 and PPF11-1 cells. However, PPF11-1 cells also had the normal 3.8 kb DNA band. Cell extracts were prepared and phosphofructokinase activities measured in these strains. The values represent averages of three or more determinations and are expressed in units per mg of protein.

PAGE 112

99 Figure 5-2. Glucose-induced degradation of AOX and FDH in PPF1 and WDK01. PPF1 and WDK01 were grown in methanol induction media then adapted to glucose. At 0 and 6 hours of adaptation, cell extracts were prepared and AOX and FDH activities measured. WDK01 cells were transformed with pWP-PFK containing normal PFK1 gene and pWP-pfk containing a mutated pfk1 gene (see Materials and Methods). Stable transformants were isolated and their ability to degrade AOX and FDH during glucose adaptation determined. The values are presented as a percentage of the activities measured at 0 hour and represent the mean ± s.d. of three or more determinations.

PAGE 113

100 by its high levels of phosphofructokinase activity comparable to PPF1 (Fig. 51 ). We next compared the ability of PPF1 and WDK01 to degrade AOX and FDH during glucose adaptation (Fig. 5-2). PPF1 and WDK01 were grown in methanol and adapted to glucose medium for six hours. Over 80% of the AOX activity and 70% of the FDH activity were lost in PPF1 . This rate of degradation was comparable to that observed in GS1 1 5. The degradation of AOX and FDH was significantly impaired in WDK01 . The amount of these enzymes remaining at six hours of adaptation was comparable to that seen for WDY2. The losses of both AOX and FDH in PPF1 1-1 that contained one wild type and one mutant copy of PFK1, were similar to PPF1 (data not shown). The result of this experiment showed that the WDK01 mutants have the same phenotype as gsa1-1. Pfkl p is essential for the microautophagy of peroxisome. Verification bv Site Directed Mutagenesis of P. pastoris PFK1 Gene of the Distinction between Rescue Function and PFK Activity Normal PFK1 and Catalvticallv-inactive pfkl Complement Aofkl The PFK1 knockout data suggested that Pfkl protein is required for the onset of glucose-induced microautophagy. Since the identified primary role of PFK protein is to metabolize glucose, it is possible that Pfkip is needed to produce a "signal" and initiates microautophagy. In order to evaluate whether

PAGE 114

101 Partial Amino Acid Sequence Alignment of Phoshofructokinase sc pp HU KKIAVt TSGGESP KKIAi: SMNAAVRAV^RTGIHFGCDVFAV ;mnaavrav7rk T S GG CjAPpMN AAVRAVfr RP|G 1[ F YGCKVYAq YEG^ TSGGCAQ G] YVGAKVYF: YEGY YEGY SC pp HU EG 1 LLR3GKYLKKMAWEDVRGWL3I GGH GLVKGGDMLKELQWQCVRGIIS dGMVDGGlSN I AEADWE SMS S IlLDM ::gg'1 ] IGT AF IG SME AI SKE FI lER idgAFico^Fr FI.KR llRj SC pp HU E GR :^QAAGNLI SC GI DALWC GGDGSLTGAE LFF E V\ GR LQACYNMVS^ GI DALWC GGDGSLTGAE E GR LKAACNLLQF GI TNLCVI ^GGDGSLTGA| ^ LFF^ LFFF EUPSL 326 EV\PEL 331 EV\SGL 145 SC pp HU VEELVAES^FT ^EEVAPYKMS] VG I VGSIDNE M£ GTE ST IKELLGE3
PAGE 115

102 HmdIII(1048) BamHI (644) Clal(1082) Bsu36I(2378) Hindlll (2260) EcoRV (2496) BamHI (45 12) CM (4057) hoi (3191) 4517 5'-GTT GOT TCT ATC GAT AAC GAT ATG TGC-3' DNA sequence VGSIDNDMC Amino acid sequence C GTT GGT TCT ATC AGT AAC GAT ATG TGPrimer sequences A CCA AGA TAG TCA TTG CTA TAC ACG CAsp362Ser Amino acid exchange Figure 5-4. Site directed mutagenesis of P. pastoris PFK1 gene. The upper panel of the figure is a restriction map of PFK1 gene and its franking region. Important restriction enzyme sites are also included. The middle panel illustrates the site of mutation in the PFK1 gene and the primer sequences constructed for site-directed mutagenesis. The Cla I (position 1 082 from the beginning of the PFK1) was abolished due to the mutation and this was used as the first screen criteria for the right construction. Verification of mutation was done by sequencing from both directions. Pfk1 was inserted into Xho I Not I site in the pWP4 and this pWP-pfk1 was used to transform WDK01 (Apfkl) and gsa1-1 cells. As a control, pWP-PFK1 was constructed and transformed into the mutant cells. *lndicates site of mutation.

PAGE 116

103 the degradation of peroxisomes required catalytically active phosphofructokinase, I transformed WDK01 with either PFK1 or pfk1 gene with its fructose 6-phosphate binding site inactivated (Berger and Evans, 1992, Arvanitidis and Heinisch, 1994). The full length of PFK1 with 161 bases of its 5' flanking region was constructed into pYM4 (pWP-PFK). In a second construct, the aspartic acid at position 362 was changed to a serine by site directed mutagenesis as described in Materials and Methods (pWP-pfk) (Fig. 5-3 and 5-4). Both constructs were introduced into WDK01 and gsa1-1 cells. The integration was induced by linearization of the constructs at the His4 gene site. Stable transformants were then subjected to AOX and FDH assays under glucose adaptation condition and PFK assay as described in Materials and Methods. Phosphofructokinase activity was increased twofold in WDK01 cells stably transformed with pWP-PFK. However, this activity was still significantly lower than that measured in parental PPF1 (Fig. 5-1). Since PFK1 was directed to its HIS4 gene locus and the 5' non-transcribing flanking region of PFK1 includes only 161 nucleotides, it may not include the complete promoter and thus only allow low level transcription of PFK1. WDK01 that had been stably transformed with pWP-pfk had only 50% of the phosphofructokinase activity measured in WDK01. Thus, cells not expressing PFK1 {pfk1::ARG4) have more phosphofructokinase activity than cells expressing a mutant pfkl protein {pfk1:.ARG4, pfk1).

PAGE 117

104 Restoration of AOX Degradation Ability of Apfkl by PFK1 We next examined the ability of PFK1 and pfk1 to restore the degradative capacity of WDK01 (Apfkl). The glucose-induced degradation of AOX and FDH was dramatically slower in WDK01 cells than in parental PPF1 cells. Substantial levels of both AOX and FDH remained in WDK01 cells at 6 hours of adaptation (Fig. 5-2). When WDK01 cells were transformed with either PFK1 or pfk1, the degradation of AOX and FDH was comparable to PPF1 cells (Fig. 5-2). The data indicate that PFK1 is required for glucoseinduced microautophagy. Howeyer, this degradatiye event proceeds independent of its glycolytic activity. We propose Pfk1p is a bifunctional protein independently regulating glucose metabolism and glucose-mediate microautophagy. Nonconventional functions have been attributed to many glycolytic enzymes (Smalheiser, 1996). In addition to their abilities to bind DNA and actin (Clarke, et al., 1985, Ronai, 1993), several enzymes have been shown to have nonenzymatic functions. For example, glucophosphoisomerase stimulates growth of neuronal tissues (Mizrachi, 1989). Interestingly, the trophic effects are found only when this enzyme is in the enzymatically inactive monomeric form. Neuron-specific enolase enhances neuronal survival and can be found within the nuclei of damaged nerves (Angelov, et al., 1994, Takei, et al., 1991 ). The redistribution of cytosolic enolase to the nucleus during neuronal repair was transient suggesting a role for this protein in

PAGE 118

105 neuron growth. Robbins and coworkers reported that a single base mutation in glyceraldehyde 3-phosphate dehydrogenase alters endocytosis in CHO cells (Robbins, et al., 1995). In addition, this enzyme plays a role in neuronal apoptosis (Ishitani, et al., 1996). Cellular levels of glyceraldehyde 3phosphate dehydrogenase have been shown to increase in dying neurons. However, if this increase is suppressed by the addition of antisense oligonucleotides, the neurons continue to survive. The mechanism of glyceraldehyde 3-phosphate dehydrogenase in these cellular processes is unknown, but appears to be unrelated to its glycolytic activity. Treptau et al. (1995) found phosphoglucomutase (PGM) underwent rapid dephosphorylation during trichocyst exocytosis in Paramecium tetraurelia. A protein phosphatase/kinase system was suspected to be involved in the modification of PGM and thus the regulation of exocytosis in P. teraurelia cells (Kissmehl, et al., 1997). The glycolytic enzymes have been long known to bind reversibly to the cytoskeleton (Arnold and Pette, 1968, Roberts and Somero, 1989). This binding leads to the activation of the enzymes and their close proximity to each other when bound to cytoskeleton provides a more efficient glycolytic flux (Clarke, et al., 1985). However, the effects of binding on actin structure and function have not been investigated. It will be interesting to investigate the possibilities that some of glycolytic enzymes or glycolysis products do have non-conventional functions that signal or regulate cell dynamic change under stress via processes such as endocytosis, exocytosis or autophagy.

PAGE 119

106 Morphological Characterization of Aofkl We have shown that suppressed degradation of peroxisomes in gsa1-1 cells was coincident with peroxisomes accumulating outside the yeast vacuole (Fig. 5-5). Thus, we propose that GSA1 is required for an event upstream of vacuolar degradation, possibly the initiation of autophagy. If GSA1 were PFK1 as we expect, the morphology of WDK01 (Apfkl) during glucose adaptation should be indistinguishable from that observed for gsa1-1. This was indeed the case. In both Apfkl and gsa1-1 cells after 3 hour glucose adaptation, the peroxisomes appeared normal and were found outside the vacuole. The vacuole contained minimal cellular debris and was noticeably round with no invaginations or protrusions as seen in GS1 15. The indistinguishable morphological presentation o\gsa1-1 and Apfkl, the similar biochemical presentations in these two strains and, the absence of PFK mRNA in Apfkl (data not shown) suggested that gsa1-1 is a null mutant of PFK1. The Apfkl cells were also transformed with pWP-PFK containing a full length PFK1 and pWP-pfk with the PFK active site mutagenized. The morphologic observation was also carried out in these transformants. As shown in Fig. 5-5, In Apfkl cells that had adapted to glucose for 3 hours, cells maintained large peroxisomes. The yeast vacuole was round and little cell debris presented. However, in Apfkl and gsa1 transformed with either pWPPFK or pWP-pfk (Fig. 5-6), after three hour glucose adaptation, peroxisomes

PAGE 120

107 Figure 5-5. Morphology of GS115, gsa1-1 and Apfk1 during glucose adaptation. GS115, gsa1-1 and Apfk1 were grown on methanol and then adapted to glucose for 3 hours. Cells were prepared by a potassium permanganate fixation protocol for ultrastructural analysis (Tuttle et al.,1993). The vacuoles in gsa1-1 and Apfk1 cells contained little cellular debris and finger-like protrusions seen in GS1 15 (arrows) were absent. Peroxisomes in gsa1-1 and Apfk1 were found outside the vacuole. Occasionally, a peroxisome was seen to be surrounded by multiple membranes (arrowhead). Nu, nucleus; P, peroxisome; and M, mitochondria. Bars, 0.2 |jM in upper right panel and 0.5 for all other panels.

PAGE 121

108 Figure 5-6. Morphology of WDK01 {Apfk1), WDY2 {gsa1) and its transformants during glucose adaptation. Gsa1-1 cells and Apfk1 transformed with pWP-pfk and pWP-PFK were grown on methanol induction medium and then adapted to glucose for 3 hours. Cells were harvested, fixed in potassium permanganate, and prepared for electron microscopy study. In cells transformed with pWP-PFK or pWP-pfk, the yeast vacuole was active in wrapping the peroxisomes by its protrusions of vacuole membrane. In the left lower panel o\ Apfk1 cells transformed with site-directed mutagenized pfk1, a peroxisome is inside the vacuole and its membrane has been lost. In the upper right panel o\gsa1-1 cells transformed with normal PFK1, a peroxisome has been partially degraded. N, nucleus; P, peroxisome; and M, mitochondria; V, vacuole. Bar, 0.7 pM.

PAGE 122

109 were wrapped by the protrusions of the yeast vacuole and some of peroxisomes were being degraded inside the vacuole in both transformants. No evidence of peroxisome macroautophagy was observed since no multiple membrane layers were seen around these peroxisomes which is the hallmark of macroautophagy. The biochemical data (Fig. 5-2) also support that the degradation of peroxisome proceeds normally in these mutants transformed with either PFK1 or pfk1 containing gene vectors. Our combined results suggested that PFK modulates microautophagy by a mechanism independent of its glycolytic activity. Chapter Summary To verify that Pfk1 protein is required for microautophagy, I generated a knockout strain Apfk1 with disrupted PFK1 gene through homologous recombination of a 5.6 kb knockout fragment into the PPF1 genome. I examined the ability of Apfk1 to degrade AOX in response to glucose adaptation. As observed in gsa1-1, Apfk1 was unable to degrade AOX and FDH during glucose adaptation. In addition, the morphology of glucoseadapting Apfk1 cells was indistinguishable from that observed in gsa1-1 cells. This verifies that Pfk1p is needed for the microautophagy. I also mutated the catalytic site of the Pfk1 protein and tested its ability to complement Apfk1 cells. I showed that despite its mutated catalytic site and the loss of PFK

PAGE 123

110 activity in the transformant, pfk1 complemented Apfk1 when analyzed by its AOX degradation ability. The morphological studies of gsa1-1 and Apfk1 transformed with PFK1 gene and pfk1 gene also supported this view. In conclusion, I have shown that Pfk1p is required for the onset of glucosemediated microautophagy of peroxisomes independent of its glycolytic activity. The questions remaining are 1) Does PFK2 has a role in the initiation of microautophagy? 2) What is the signal pathway of microautophagy and how does Pfk1 p fit in there? 3) What other regulatory molecules are present in this process? I suggested that PFK modulates microautophagy by a mechanism independent of its glycolytic activity. It may either directly modulate autophagy via another yet to be identified functional domain or it may use its kinase domain to modified a substrate which in turn regulates the autophagy process. Further pursuit of gsa1 as well as other gsa mutants will address these questions.

PAGE 124

CHAPTER 6 CHARACTERIZATION OF gsa7 MUTANTS Introduction A comparison study of glucose induced AOX and FDH degradation after methanol induction in gsa7 and other mutants has shown in Fig. 3-4 in chapter 3. Gsa7 showed impaired ability to degrade AOX and FDH in a six-hour glucose adaptation when compared with that of parental GS1 15 cells. The peroxisomal AOX and cytosolic FDH degradation in the ethanol adaptation pathway was not affected in gssT (Table 3-1). In order to identify possible glycolysis mutants, I have tested some important glycolysis enzyme activities in selected gsa mutants and found gsaJ was not involved in glycolysis steps (Fig. 4-1 1 ). Gsa? was chosen for subsequent morphologic and genetic studies. I will discuss the results in the following sections. Morphological Studies of psa? Ultrastructure studies were also carried out in gsa? mutants. Several representative images of gsa3 and gsaJ mutants at zeroand three-hour 111

PAGE 125

112 glucose adaptation are presented in Fig. 6-1 . At zero hour of adaptation, large peroxisomes were induced due to the need for methanol metabolism enzymes in both gsa3 and gsa 7. The yeast vacuole was relatively round and small. No vacuolar membrane invagination and protrusion was observed. After three hours' glucose adaptation, gsa3 and gsa7 showed the projections of the vacuolar membrane, which sometimes wrapped the peroxisomes. However, no vacuolar membrane fusion was observed in gsa3 and gsaZ which bore similarity to that of the gsa2 mutant (see morphological description in Chapter 4). We did not observe peroxisome degradation in the yeast vacuole in gsa3 and gsa7. This was corroborated by our biochemical studies in these two strains that degradation of peroxisomal AOX was impaired. In fact, we believed that morphologically, gsa2, gsaS and gsa7 belong to the same mutant group defective in the homotypic fusion event of microautophagy (see Chapter 3) and we believe that Gsa2p, Gsa3p and Gsa7p act downstream of Gsa1p in the glucose induced peroxisome autophagy pathway. Recovery of GSA7 Gene and its Verification PYWP7-4 Rescues Qsa7 Phenotvpe Gsa7 cells were transformed by electroporation with a genomic DNA library (see Materials and Methods and also Chapter 4). The putative rescue colonies of gsa7 mutants that appeared white (AOX was degraded) on the filter

PAGE 126

113 Figure 6-1 Morphology of gsa3 and gsa7 mutants during glucose adaptation. Gsa3 and gsa7 mutants were grown on methanol media for 24-36 hours and then were switched to glucose for adaptation. The cells were collected at 0 and 3 hours of adaptation and prepared by a potassium permanganate fixation protocol for ultrastructural analysis (see Materials and Methods). Cells at 0 hour showed large peroxisomes were induced. They were round and with little vacuolar activity. At 3 hours of glucose adaptation, the vacuolar protrusions of gsa3 and grsa 7 wrapped around the peroxisomes but no vacuolar membrane occurred. Bar = 0.9 |jm.

PAGE 127

114 paper in a direct colony assay were collected. A total of 30,000 colonies were screened by this method. Thirty clones were collected from the direct colony assay screening for liquid media AOX assay verification. Only clone 1 , clone 4 and clone 5 showed reproducible rescue ability by both assays. The three plasmids were recovered from these clones and were named pYWP7-1 , pYWP7-4 and pYWP7-5 accordingly. The first two plasmids contain the overlapping genomic DNA fragment identified by restriction enzyme digestion. However, the insert in pYWP7-1 is 0.5 kb longer at the 5' end of the insert than that of pYWP7-4 which has about a 4 kb insert. pYWP7-5 has a 10 kb insert according to the estimation of restriction enzyme digestion. PGR of pYWP7-1 , pYWP7-4 and pYWP7-5 with two primers generated from sequence inside pYWP7-4 insert yielded identical bands of predicted DNA fragment size (data not shown). Therefore, we concluded that all inserts in these three plasmids overlap. When these three plasmids were reintroduced into gsa7 by transformation, all could rescue gsa7 mutants. Only the rescue results of pYWP7-1 and pYWP7-4 were shown in Fig. 6-2. Identification of GSA7 Once the rescue function of pYWP7-1 and pYWP7-4 and pYWP7-5 were verified, the 4.0 kb insert in pYWP7-4 was sequenced from both directions by DNA Sequencing Core of University of Florida. By searching the NCBI sequence database with the insert sequence, three open reading frames that code putative proteins were returned. A 1 04 amino acid of C terminal protein

PAGE 128

115 NcoI(349) Bgllin979) Clal(292) ^ Mfel(1735) \ Mfel(2257) Xbal (141) Seal (3828) YGK24SC YHRmw (GSA7) pYWP7-l pYWP7-4 pWP-GSA7 4073 SWB Rescue + + + pYM-SWI3 Figure 6-2 Verification of GSA7. The restriction map of pYWP7-1 insert is provided here. One full length ORF and two partial ORFs were found to be included in the insert. The pYWP7-4 and pYWP7-1 could rescue gsaZ when reintroduced into gsaJ. A DNA fragment containing complete putative GSA7 gene showed rescue effect while the DNA fragment containing N-terminal half of the Sw/3 gene could not rescue gsaZ. This further suggested that the full length ORF encodes our Gsa7p.

PAGE 129

116 sequence that is homologous to S. cerevisiae Ygr245cp is at the beginning of the insert (see Fig. 6-2). A 340 amino acid of N terminal protein sequence that is homologous to Swi3 protein of S. cerevisiae is at the end of insert. In between is a full length putative protein that is homologous to S. cerevisiae Yhr171wp. SwiSp has been identified as a global transcription factor in many species while the functions of Ygr245cp and Yhr171wp have not been identified. In order to verify that the putative P. pastoris homolog of YHR171w gene is our GSA7, I subcloned P. pastoris DNA fragments containing SWI3 and YI-iR1 71 w genes into pYM8 (an episomal vector) or pYM4 (a non-episomal vector) respectively to study which gene was responsible for the rescue effect. The results of rescue study of these fragments were shown in Fig. 6-2. The DNA fragment in pYM4 containing YlHR171w could complements gsa7 but not the pYM8 containing SWI3. Sequence Analysis of GSA7 The DNA and amino acid sequences of GSA7 are shown in Fig. 6-3. GSA7 gene encodes a putative 654 amino acid protein with a molecular weight of 71 kDa. A sequence comparison of Gsa7p with its homologs in yeast and human is shown in Fig. 6-4. As we can see, the Gsa7p and its yeast homologs in S. cerevisiae and S. pombe show greatest homology in the C terminal half of the proteins. It also showed high homology to a partial human sequence identified by searching EST database. Conceivably, this C-terminal region

PAGE 130

117 1 AAAAAGAAAGATGCGGGGATGCGCAATAAGCTTTTCATTCACTGATGATAGGGTTGCGGATCTTCTTCCCAGAAGCCTGTTTTTTTTTTC 91 GCAAAAATCTTGAAAGACGTTTCCACTGGTAGGCTGTTGATATCTTCTAAACAGGCCAGATATCCCCCTTTTTTAGTACGCTGAAAGCCA 181 AGCAGTTTACAGGAATGGCGTGGTCGTGTCAGAATCAAATGTCGCGTTAGGCCCCATCGAGCCTTCCCTATTTGTCTAACCTTTAGAAGG 271 TTCCTTCCCCCACACAGTAACCATGACAAGCATGGATATCCCCTATTCGCAAATAAGTTCATTTGTGAACTCGTCATTCTTCCAGAAGGT MTSMDIPYSQISSFVNSSFFQKV 361 GTCCCAATTGAAGCTCAACAAATATAGACTGGACGACACGGACAAGGCGATAGTGGGCAGTGTCGATTTTAAGTTCATTGGGAAGAACCA 121 SQLKLNKYRLDDTDKAIVGSVDFKFIGKNO 451 GCCAACAAGTTTATCTGTAGATGAGTCAAGCTTCAATGATAATATAACTTACACGCATGCACAGTTCCCGGTCAAAGGAATACTGAAGAA 151 PTSLSVDESSFNDNITYTHAQFPVKGILKN 541 CTTAAACACGGTTGAGGATTTCAGAAAGGTGGACAAGAATGAATTTCTTCAATCACAGGGTTTGGTGGTGCATAAATCAATTCAAGACCG 181 LNTVEDFRKVDKNEFLQSQGLVVHKSIQDR 631 TTCGTGTCTAAAAGATCTCTCAAAGTTGACCCAATTTTTCATTCTGTCCTTTAGCGACTTGAAAGGGTTCAAATTCATCTATTGGTTTGG 211 SCLKDLSKLTQFFILSFSDLKGFKFIYWFG 721 CTTCCCATCATTAGTGAGTAGATGGAAAGTAAACAAACTGAGTGGTCTAACTGAATCGCAGATAGAGCCGTATGAAAGTAAGCTTAACGA 241 FPSLVSRWKVNKLSGLTESQIEPYESKLNE 811 GTGGCTAAATGCTCGTTTGCCCATTGAGCAAAAGCAAGCTTTCATTATTGACAACCTTGAATTCAAACCGTTTGAACAATTGTCAAGTTT 271 ULNARLPIEQKQAFIIDNLEFKPFEQLSSF 901 TTCACCTGACGATCAGCTTAACATTGGGTTTATTGACACCAGCAGCATCCTCAACAAATGTTCCACCCAGTTGAGAAATATTCTGTACAT 301 SPDOQLNIGFIDTSSILNKCSTQLRNILYM 991 GCTGGCTTATTATGGCTTTGAGAACATCAAAGTATACAATTTCAGATTCAACAATACCACATCCTTTACATTAGACATCACTCTTGCTGA 331 LAYYGFENIKVYNFRFNNTTSFTLDITLAE 1081 GCCTCTTACTTCCGAGCCAAAAACAACAGGGTGGGAGAGAACTGCTCAAGGTAAGTTGGGCCCCAAACTGGCCGATATAGGTGCTTTGGT 361 PLTSEPKTTGWERTAQGKLGPKLADIGALV 1171 TGACCCTGCCCGCTTGGCTGACCAATCAGTTGATCTAAATTTAAAGCTGATGAAATGGAGAGTCATGCCTGAACTTGATCTGGATATCAT 391 DPARLADQSVDLNLKLMKWRVMPELDLDII 1261 AAAGAATAGTAAGGTTCTACTTCTCGGTGCTGGAACACTGGGAAGTTATGTCTCGAGAGTATTGCTAGGATATGGAGTTCGACACATTAC 421 KNSKVLLLGAGTLGSYVSRVLLGYGVRHIT 1351 GTTTGTTGATAATGGTAAAGTTTCATTCTCTAACCCTGTTAGACAACCGTTGTTCAATTTTACAGATTGTTTGGAGGGAGGTGCTCCGAA 451 FVONGKVSFSNPVRQPLFNFTDCLEGGAPK 1441 AGCCGAAACTGCAGCCAAAGCATTGAAATTAATTTTTCCGTTAATAACAAGCCAAGGATATAACTTGGAAGTACCTATGGCTGGACACCC 481 AETAAKALKLIFPLITSQGYNLEVPMAGHP 1531 GGTTACCGATGAAAAAAGACAGTATGAAGACTATCAAAGGTTAGTGACCCTAATAAAGGAACATGATGTAGTTTTCCTTCTAATGGATTC 511 VTDEKRQYEDYQRLVTLIKEHDVVFLLMDS 1621 AAGGGAAACGAGGTGGCTGCCTACCGTGCTTTGCAACGTTTTTGATAAAATTTGCATCACTGCTGCATTGGGATTTGATTCATACCTGGT 541 RETRWLPTVLCNVFOKICITAALGFDSYLV 1 71 1 AATGAGACATGGAAACTTGTTTAATACCGAGCACATAGAAGCGGAAGAGAACTCTCACAGGCTGGGATGCTATTTCTGCAACGATATCAT 571 MRHGNLFNTEHIEAEENSHRLGCYFCNDIl 1801 TGCTCCGAAAGATAGCACAACTGATCGAACTTTGGATCAAATGTGCACAGTGACCAGACCAGGGGTAGCCTTACTTGCTAGTTCTTTGGC 601 APKDSTTDRTLDQMCTVTRPGVALLASSLA 1891 TGCGGAACTATTTGTTTCCATTCTTCAACACCCTCTAAAGAGTCATGCGCCTGCATCACTCCATGATAATGCCACAGTTCTTGGATGTTT 631 AELFVSILQHPLKSHAPASLHDNATVLGCL 1981 ACCACAACAACTCCGAGGGTTTCTTCACAATTTCGAAACCTCCAAACTTGAAGCAAATAACTACGAATACTGCTCTGCATGTTCGATACA 661 PQQLRGFLHNFETSKLEANNYEYCSACSIQ 2071 GGTATTAAACGAATACAAATCCAGAACTTGGGATTTTGTCAAAGATGCCCTGAATGAAAACAATTATCTTGAGGATTTGACAGGCCTTAC 691 VLNEYKSRTWDFVKDALNENNYLEDLTGLT 2161 TAAGGTCAAGCAAGAATCTGAGATAGCCGAGAAGAAGTTTCAAGAGTTTGAAAACGGTTTAGAGTTTAGTGATGAAGATTCAGAATGGAT 721 KVKQESEIAEKKFQEFENGLEFSDEDSEWl 2251 AAACTAATATACAAAAACTAATCCATGGGGTTCTGGAAAATTAAGAAAAAATATATTTAATGACCCTTCTTTTTTTGCTTATCGATATGA 751 N * 2341 GCTCGTAGAACCTTTTGCTTATCACGAAGAGCCATTTTCTGTTTGCTTTGCACACCTCTCTTGTGGATGGTCATCAAGAAGTTCTTCTTA 2431 CGAGCCTTCTCTTTATTAGTGGTAGAATGCTCTCCTTCTCTCTTTCCTCTTCTAGAACCATGAGCATCTTCATCACGACCAGCCTTGACA Figure 6-3 Complete nucleotide and amino acid sequence of GSA7. GSA7 gene encoded a 654 amino acid, 71 .2 kD putative novel protein. No known protein motifs were recognized by our motif search.

PAGE 131

Gsa7 Yhr171w S.ponibe Gsa7 Yhr171w S.pombe Gsa7 Yhr171w S.pombe Gsa7 Yhr171w S.pombe Gsa7 Yhr171w S.pombe Gsa7 Yhr171w S.pombe Gsa7 Yhr171w S.pombe Gsa7 Yhr171w S.pombe Gsa7 Yhr17lH HGsa7 S.pombe Gsa7 Yhr171w HGsa7 S.pombe Gsa7 Yhr171w HGsa7 S.pombe Gsa7 Yhr171w HGsa7 HTSMDIPYSQIS..SFVNSSFF QKVSQLKLNKYRLDDTDKAI VGSVDFKFI 49 -s-ervlsyapafk--ldt---el-r---dvlk--s-cqpltvnl-lhn51 -fvgkalqf-sfhs-idatfHhqlsnykvek--ldasp-tihglcfn-ysrgni-iv-gea 60 G
PAGE 132

119 should be the main functional domain of the protein and thus the focus of our future study. Interestingly, this is also the region that has homology to ubiquitinactivation (UBA1 ) enzyme of S. cerevisiae . A protein motif search of Gsa7p and their other yeast counterparts using Prosite program failed to yield any active site or signature sequence that has been identified in UBA1 in yeast or human. AqsaJ Generation and its Phenotype Studies I have shown that P. pastohs YHR171w complement gsa7. I next constructed a gene knockout fragment of GSA7 gene and used it to knockout the GSA7 in a double auxotroph parental strain PPF1. A Hind III Bgl II fragment in GSA7 were replaced by a S. cerevisiae Arg4 gene (see Fig. 6-5, upper panel). A double auxotroph PPF1 (his4, arg4) were transformed with the knockout fragment (5.1 kb, cut out by Apa I and Sea I) by electroporation. The histidine auxotrophs grown on the transformation plates were chosen and direct colony assay of AOX were applied as a screening method. The colonies appearing purple is defective in degradation of AOX and thus the possible knockout mutant. Three clones in this assay showed purple color and were selected. The genomic DNAs of two clones along with GS115, gsa7 and two other non-rescue transformants C3 and C4, were extracted for PGR screening. The primers used in PGR are depicted in Fig. 6-5. The 5' primer is inside the knockout fragment and precedes 350 bases of the start code of GS/A7 while 3' primer is outside the knockout fragment and just behind the stop codon of GSA7 gene. In a true knockout strain, the PGR fragment should be 4.9 kb while in control strains, the

PAGE 133

120 fragment should be 2.5 kb. The lower panel of Fig. 6-5 showed the PGR results of these strains. Agsa7-1 and Agsa7-2 have the predicted 4.9 kb fragment size but without the 2.5 kb fragment and thus the true knockout. The AgsaYs were then subjected to AOX liquid media assay during glucose adaptation. As expected, the AgsaYs lacked the ability to degrade AOX thus mimic the phenotype of gsaZ while all other strains tested show normal ability to degrade AOX. This experiment showed that the gene we cloned is necessary for the microautophagy of peroxisomes and thus is Gsa7p. A HA Tagged Gsa7p Rescues P. pastohs To study the Gsa7 protein function of P. pastohs, 1 also needed to verify that Gsa7p is expressed in P. pastohs. I also wanted to identify the localization of the protein in the cells. I designed two primers flanking the GSA7 gene for PGR purpose. The 5' primer is 350 bases upstream of the start codon to include its own promoter while the 3' primer is at the end of GSA7 gene containing a HA epitope tag sequence. The GSA7-HA gene was amplified by PGR constructed into a pYM4 vector (see Materials and Methods). The HA tagged YHR171w gene (a homolog of GSA7 in S. cerevisiae) was also amplified by PGR of S. cerevisiae genomic DNA (provided by Ms P. Wu) using the same strategy as in P. pastohs. The PGR products were then constructed into pYM4. Both vectors were introduced into gsa7 and stable transformants were isolated. The

PAGE 134

121 Apal (1488) Ncol (349) \ Hind 111(1776) BglII(I979) Seal (3828) PpGSA7 SWI3 2.5 kb GSA7 Apa I (1488) Ncol (349) Hind 111(1776) Bglll (4754) Seal (6603) ScARG4 4.9 kb gsa7::ARG4 00
PAGE 135

122 expression of both gene products in P. pastoris was confirmed by Western blot with an anti-HA epitope tag antibodies (Fig. 6-6). Clones P6B, P3B and P3G were gsaJ transformed with a HA tagged GSA7 gene of P. pastoris while clones S1 , S3, S4 were gsa7 transformed with a HA tagged YHR171w gene of S. cerevisiae. As shown in Fig. 6-6, no HA tagged protein was detected in done P6B, S1 and gsaJ. The HA tagged Gsa7ps were detected by the HA antibodies in all other clones at the predicted size of 71 .1 kDa. When these clones were tested for their ability to degrade AOX activity by glucose adaptation assay, I found clone P6B and S1 were defective in degradation of AOX during a six hour glucose adaptation while all other clones regained their ability to degrade AOX during glucose adaptation. This is well correlated with the data that Gsa7p or Yhr171wp were expressed in these rescue clones. Interestingly, our preliminary result showed that, when clone P3B, P3g and S3 were grown on methanol and then adapted to glucose, the protein levels in these three clones did not change significantly at 0, 1 and 6 hour time point (data not shown). Our preliminary data suggested that Gsa7p is a cytosolic one. However, the location of the Gsa7p still need to be confirmed by immunofluorescence study. Chapter Summary In this chapter, I described the characterization of gsa? mutant as well as the identification of the G SAT gene. GsaJ is defective in peroxisome

PAGE 136

123 Figure 6-6 Detection of a HA epitope tagged Gsa7p and Yhr1 71 wp in transformed gsaJ. HA epitope tagged GSA7 and YHR171wv/ere constructed into pYM4 with their endogenous promoter and introduced into gfsa7. The transformed gsa7 cells were grown on YPD overnight and cells were prepared for Western blot analysis (the upper panel). Both Gsa7p and Yhr171wp have a predicted molecular mass of 71 kD. However, Gsa7p is higher than Yhr171wp on the Western blot with a molecular mass difference of 2-4 kD to that of Yhrl 71 wp. These strains were also tested their ability to degrade alcohol oxidase (AOX) after a six hour glucose adaptation. The values presented are the percentage of AOX activity remaining after a 6 hour glucose adaptation. They were measured at least three times and presented as a mean of the total measurements. ''Gsa7p; *Yhr171wp.

PAGE 137

124 microautophagy. However, the ethanol induced macroautophagy pathway is intact. Our ultrastructural analysis showed that gsa7 is defective in the an event of homotypic fusion of yeast vacuolar membrane in the autophagy process similar to gsa2 and gsa3 (see chapter 3). The functional complementation study of gsa7 mutants with a genomic DNA library yielded a plasmid with an insert of 4.0 kb that rescued the mutant phenotype. Sequencing of this insert revealed a possible complete ORF of GSA7 coded a 71 .1 KD protein that is homologous to Yhr171wp of S. cerevisiae. Rescue study with different fragments of the insert verified that this putative GSA7 in the insert could rescue gsa7. The GSA7 gene knockout experiment showed that GSA7 is essential for the peroxisome microautophagy during glucose adaptation. The HA epitope tag sequence was attached to the end of GSA7 and YHR171w gene and introduced into gsa7 mutants. The Western blot confirmed the expression of HA tagged Gsa7p and Yhr171wp in the transformed gsa7 cells and these cells regained their ability to degrade AOX. This verified that the function of Gsa7p is conserved in P. pastoris and S. cerevisiae. Since microautophagy has not been observed in S. cerevisiae while macroautophagy has been identified in S. cerevisiae and our GSA7 gene appeared to control a step of homotypic vacuolar membrane fusion in microautophagy, it will be interesting to knockout YI-IR171w gene in S. cerevisiae and to observe its phenotype and compare with that of gsa7. Further study of this protein in these two yeast strains will yield more information about the functional role of Gsa7p.

PAGE 138

125 The discovery of Gsa7p and its verification give us an opportunity to examine for tlie first time a specific molecule that controls one of the important steps in the peroxisome autophagy process. We can do a whole array of experiments that previously not possible. GSA7 gene and its S. cerevisiae counterpart YHR171w gene with a HA epitope tag have been put into pHWOlO vector behind a GAP promoter and the pPIC3 vector with an AOX gene promoter (Both vectors were provided by Dr. J. M. Cregg). Important information such as the relationship of vacuolar fusion genes can be derived if overexpression of Gsa7p in these plasmid could rescue gsa2 and gsa3, gsaS and the other fusion mutant gsa5. We can also construct GSA7 gene into a GAPDH promoter driven plasmid or an AOX prompter driven plasmid that we have and observe the overexpression of this protein in both wild type GS1 15 and mutant gsa7. We can also test if the expression or overexpression of the protein could rescue gsa7 or other gsa mutants in the same fusion defective group such as gsa2 and gsa3. Such experiments could probe the action sequence of these genes in the homotypic fusion event. Since no significant protein motifs were recognized by our gene database search, we could also use random PGR mutagenesis to mutate the GSA7 gene. The products can then be introduced into gsa7 and use AOX direct colony assay to identify those clones that could not rescue gsa7. Such clones should contain plasmid with mutated GSA7 ai critical protein functional sites. Sequence analysis of the plasmid in such clones will give us information as to which domain might be important to the protein function. The

PAGE 139

126 C-terminal half sequence of the human GSA7 homolog has been sequenced and we obtained the clones possibly containing the whole human GS>A7gene (purchased from ATCC). We knew that this C-terminal half bears the highest homology throughout species and has sequence homology to ubiquitinactivating enzyme in S. cerevisiae's. Sequence analysis of this human GSA7 gene, further study of this gene expressing in P. pastoris mutants or vice versa and the functional study of this gene product in mammalian cell system will yield more information as to how this Gsa7p works in the autophagy process in both yeast and possibly mammalian cells system. Drs Oshumi and Klionsky's groups recently found that S. cerevisiae's Gsa7p might act with other novel ubiquitin-like pathway enzymes and responsible for the defect in their autophagy mutants (personal communication). It is important to test the possible role of our GsaZp's in this alternative pathway and if possible, to identify other players in this process.

PAGE 140

CHAPTER 7 CONCLUSIONS AND PROSPECTS Introduction Protein degradation in eukaryotic cells occurs in all subcellular compartments. It is complicated in that there are many pathways for degradation of intracellular proteins (Olson et al., 1992). One reason for this diversity is the variety of intracellular compartments in which proteins are degraded. In addition, various proteins need to be degraded at different rates depending on the physiological state of the cell. The lysosomal protein degradation pathway is one of such pathways in mammalian cells. The functional equivalent of lysosome in yeast is its vacuole. Proteins or cell organelles are delivered to the yeast vacuole for degradation in a similar manner as in mammalian ceils (Takeshige, etal, 1992, Dunn, 1994, Chiang, etal., 1996). Our lab has characterized a system in methylotrophic yeast P. pastoris to investigate protein degradation by observing the process of peroxisome degradation in the yeast vacuole (Tuttle, et al., 1993). The general pathways of peroxisome induction and degradation under different carbon sources in P. pastoris are depicted in Fig. 7-1 . When P. pastoris cells are grown on glucose, 127

PAGE 141

128 they maintains a few small peroxisomes. However, when cells are grown on methanol, large peroxisomes are induced. During the subsequent growth carbon source switch such as glucose or ethanol, peroxisomes underwent microautophagy and macroautophagy respectively (Tuttle and Dunn, 1995). A simplified illustration of the concept of microautophagy and macroautophagy is provided in Fig. 7-2. In yeast, microautophagy is a process that peroxisomes or other cell components are engulfed by vacuolar membrane protrusions and the contents get degraded thereafter. Macroautophagy refers to peroxisomes or cell components are surrounded first by several layers of membrane to form autophagosomes and the autophagosomes then fuse with the yeast vacuole to release the contents of autophagic bodies for degradation. Based on previous work in this lab, we proposed the microautophagy process includes glucose signaling, peroxisome sequestration, vacuolar membrane homotypic fusion and vacuolar degradation of peroxisomes (see Fig. 7-4) and specific regulatory molecules needed to be synthesized to modulate or coordinate this autophagy process (Tuttle and Dunn, 1995). In order to investigate the mechanism of microautophagy, we have generated glucose-induced selective autophagy (gsa) mutants that are unable to degrade peroxisomal AOX during glucose adaptation. By investigating the biochemical and morphological phenotypes as well as genetic characteristics of gsa mutants, we tried to understand the regulatory mechanism of glucose induced microautophagy in this yeast. My studies consists of three parts. The first part is the characterization of gsa mutants.

PAGE 142

129 P. pastoris Grow on glucose Grow on methanol \ Maintain a few Large peroxisomes small peroxisomes are induced Glucose adaptation Ethanol adaptation Microautophagy of Macroautophagy of peroxisomes peroxisomes Figure 7-1 Pathway of peroxisome induction and degradation under glucose and ethanol adaptation in methylotrophic yeast Pichia pastoris. Under glucose growth condition, P. pastoris maintains minimum peroxisomes. However, under methanol growth condition, large peroxisomes are induced. During the subsequent adaption with either glucose or ethanol, cells undergo microautophagy and macroautophagy respectively to degrade peroxisomes.

PAGE 143

130 Glucose Induced Microautophagy £thanol Induced Macroautophagy Figure 7-2 Microautophagy and macroautophagy of peroxisomes in P. pastoris. The simplified figure is used to illustrate the concept of microautophagy and macroautophagy of peroxisomes during nutritional adaptation. P. pastoris is grown on methanol till stationary and glucose or ethanol is added to begin adaptation. Upon glucose addition, peroxisomes undergo microautophagy while during ethanol adaptation, peroxisomes undergo macroautophagy in P. pastoris.

PAGE 144

The second is the genetic study of gsa1-1 mutant. The third is the genetic study of gsa7 mutants. The studies I have finished provide insights on the poorly understood phenomenon of yeast microautophagy process. Characterization of gsa Mutants The characterization of gsa mutants included biochemical analysis of alcohol oxidase and formate dehydrogenase and the morphological characterization. Well-characterized mutants were then subjected to backcrossing and complementation analysis and were used for further genetic studies. I have characterized most of the available mutants in our stocks, and eight gsa complementation groups {gsa1 to gsa8) have been identified. All gsa mutants lacked the ability to degrade peroxisomes during glucose adaptation. However, their ability to degrade peroxisomes during ethanol adaptation was not affected. Morphological studies of gsa mutants revealed that they have distinct phenotypes. Gsa1 (pfk1) is defective in the initiation of microautophagy (see Fig. 7-4). Gsa4 and gsa8 are defective in AOX and peroxisome degradation. However, their cytosolic protein degradation pathway seems intact. I believe they are defective at a step of peroxisome recognition in the microautophagy. Gsa2, gsa3, gsa6 and gsa? are defective in a homotypic fusion step of the vacuolar membrane. In gsa5, macroautophagy was turned on during methanol induction and glucose adaptation, however, the degradation of peroxisome

PAGE 145

132 under glucose adaptation via microautophagy was defective. I believe it also belongs to the group of homotypic fusion mutants but gsaS may inversely control a different step in the event of vacuolar membrane fusion than gsa2, gsa3, gsa6 and gsa7. Recently Subramani's lab (Sakai et al., 1998) also generated peroxisome degradation mutants using the same method we developed (Tuttle and Dunn, 1995). However, they used a parental strain containing a constitutively expressed GFP-SKL protein to help in visualizing yeast peroxisomes. They have identified 6 peroxisome microautophagy (pag) mutant complementation groups. It will be interesting to see if any of pag mutants overlaps with our gsa mutants. The biochemical and morphological phenotypes revealed in these studies indicate that the defective genes in these gsa mutants regulate different microautophagy steps that I have defined in the introduction of this chapter (also see Fig. 7-4). These simple but powerful tools can be used to further characterize the remaining gsa mutants as well as the functional roles of the corresponding gene products and thus the mechanism of microautophagy. Reoulation of Sianalino of Microautophaov bv Gsalp The gsa1-1 mutant is the first mutant we chose for further genetic study. A complete PFK1 sequence was recovered from plasmid that rescued gsa1-1. PFK1 mRNA was greatly reduced in gsa1-1 and so was its PFKp activity. This

PAGE 146

133 showed that gsa1-1 is pfk1 mutant. Our PFK1 knockout experiment confirmed that PFK1 is essential for the onset of autophagy. Our site-directed mutagenesis of PFK1 active site experiment suggested that PFKp activity is not needed for the initiation of autophagy. PFKp is a highly regulated protein with many conserved domains capable of binding ATP, GTP, AMP, citrate, fructose 1 ,6-bisphosphate as well as actin. This protein is the major regulatory checkpoint allowing the cell to respond to changes in glucose concentrations. Our data suggested that the regulation of glycolysis and microautophagy by glucose is modulated by the asubunit of phosphofructokinase. At this time, we have yet to define the role for this protein in microautophagy, and we have not observed any role for the 3subunit of phosphofructokinase in this process. Therefore, we do not know if microautophagy is mediated by the heterooctomeric phosphofructokinase complex composed of both a and p subunits or by monomeric a-subunit. We can only speculate that phosphofructokinase or its a-subunit may directly interact with either a vacuolar or a peroxisomal membrane protein thereby initiating the sequestration event. As an alternative hypothesis, Pfk1 protein might either add a phosphate to a substrate other than fructose-6-phosphate, which in turn regulates the microautophagy process or it may modify a sugar on a glycoprotein and this modified protein subsequently regulate the homotypic fusion of vacuolar membrane in the autophagy process. We can mutagenize the ATP/ADP binding site (Fig 7-3) that is adjacent to fructose-6-phosphate

PAGE 147

134 Substrate and regulatory sites in the a subunit of phosphofructokinase Putative actin-binding site © N ^ / AMP \ F6P F2.6P \ ATP/GTP CKrate ADP/ATP Figure 7-3 Putative substrate binding and regulatory sites in Pfk1 p. This map showed important substrate and regulatory sites in the alpha subunit of phosphofructokinase. These sites are conserved across species. The putative Ser/Thr phosphorylation sites are also included.

PAGE 148

135 binding site to see if the association of ATP to Pfk1 p is important for autophagy. If the elimination of ATP binding site in PFK1 prevented the autophagy, the kinase activity of PFK protein might be important for the autophagy. We can also generate a PFK2 knockout and observe if the microautophagy is normal in Apfk2. if microautophagy were defective in Apfk2 mutant, then PFK protein is important in this process. We should compare PFK mRNA levels in all three gsa1 strains and if possible, Identify the mutation sites in gsa1 mutants. Another way is to knockout the GAPDH gene downstream of PFKp action site in glycolysis (see Fig 4-9). If blockage of this glycolysis step in P. pastoris has no effect on autophagy, then Pfk1p is more likely to directly regulate the autophagy process. We can also generate random mutations in PFK1 by PGR and then construct the PGR products Into pYM8 and, transform gsa1 with the plasmld. We can use direct colony assay to identify plasmidcontaining colonies that do not rescue gsa1. Sequencing these plasmids may reveal sites or motifs in Pfk1p that regulate the autophagy process. Regulation of a Homotvpic Vacuolar Fusion Event of MIcroautophaav bv Gsa7D Gsa7 is defective in peroxisome degradation. It is a mutant which we believe is defective in homotypic fusion of vacuolar membrane. The morphological studies of gsa7 supported that a step in the fusion event of vacuolar membrane is defective during the engulfment of peroxisomes. When

PAGE 149

136 gsaZ cells were transformed with a Pichia pastoris genomic library, three overlapping inserts in pYM8 were recovered. Their complementing effect was verified by reintroducing them into gsa7 cells. The sequence reveals one complete open reading frame and two partial ORFs. I have confirmed that only the DNA fragment with complete ORF rescue gsa7. The putative Gsa7p (encoded by the complete ORF) is a 654 amino acid protein with no identified function. The protein has homologs in S. cerevisiae, S. pombe, C. elegans and H. sapiens and its C terminal half showed homology to ubiquitin-activating enzyme. The GSA7 gene knockout studies verified that GSA7 is essential for the microautophagy process. By transforming gsa7 cells with a HA epitope tagged full length GSA7 gene or its yeast counterpart, YHR171ww\ih their endogenous promoter, I showed that Gsa7p and Yhr171wp were expressed in gsa7 and could rescue gsa7. We knew Gsa7p is important for the homotypic fusion event in the autophagy. How Gsa7p does that? One speculation is that it participates in a novel ubiquitin-like protein degradation pathway (see chapters) and the degradation of a particular protein is important for the homotypic fusion of vacuole membrane. It is also possible that the ubiquitination of a regulatory protein leads to targeted vacuolar degradation of peroxisomes. Further studies of Gsa7p such as to mutagenize the putative ubiquitin binding cysteine, will yield information as to how Gsa7p acts in the cells, who else are its partners and where the protein functional domain of Gsa7p locates.

PAGE 150

Prospects and Conclusion 137 In recent years, great progress has been made in understanding the mechanisms by which proteins and organelles are delivered to the vacuole from the cytoplasm. Various independent studies and approaches have been conducted and applied in the studies of protein and cell organelles transport from the cytoplasm to the vacuole (Klionsky, 1997). However, many questions remained unanswered. How are the signals for autophagy transduced? How is membrane binding or engulfment of a substrate initiated and carried out? Are vand t-SNAREs and rab proteins involved in the targeting and delivering process? How do cells sense the stop signals of autophagy and transduce them to the related cell organelles? Although my study did not resolve all the questions, significant progress has been made due to the isolation of two GSA genes that play a role in the microautophagy process. As stated in the introduction chapter, peroxisomes are induced by growing cells on methanol. In the subsequent glucose adaptation, cells undergo rapid degradation of peroxisomal and cytosolic proteins. This process is similar to microautophagy that has been defined in mammalian cell system. It also needs molecules synthesis to regulate and coordinate this process (Tuttle and Dunn, 1995). The microautophagy process of peroxisomes generally consists of glucose signaling, vacuole recognition and peroxisome sequestration, homotypic vacuolar membrane fusion and finally, the vacuolar degradation of peroxisomes

PAGE 151

138 (Fig. 7-4). Gsa mutants have been generated to identify the regulatory molecules In the peroxisome microautophagy. The glucose induced peroxisome microautophagy and it relationship with gsa mutants is shown in Fig. 7-4. Currently, I have identified eight gsa complementation groups with distinct characteristics. A gene that is important for the autophagy process is defective in each of my gsa mutants. Identification of these genes is critical to understand how cell sense that peroxisomes need to be degraded, how the microautophagy process is initiated and how peroxisomes are targeted for vacuolar degradation. Gsa1p has been identified as Pfk1p. It is needed for the initiation of microautophagy. In gsa1-1, a step in an initiation step is defective as shown in Fig. 7-4 and this is the only mutant that in this signaling mutant group. Further test of our hypothesis to probe how PFK1 acts in microautophagy process as well as possible participants is crucial for the understanding of microautophagy process. Gsa7 have been identified in a homotypic vacuolar membrane fusion mutant group. GSA7 gene has been recovered and it encode a novel protein with no known function in yeast and other species. We propose that it is required for a step at or upstream of homotypic fusion event of vacuolar membrane. We hypothesized that functional role of Gsa7p is to regulate the vacuolar membrane fusion event in microautophagy. The hypothesis I proposed in the first half of this chapter and the possible research experiments will give us an answer for that.

PAGE 152

139 Glucose Addition gsal (pfkl) Vacuole gsa4, gsaS gsa2, gsa3, gsaS, gsa6, gsa7 pra, prb tophagy Signaling Peroxisome Peroxisome Recognition and Sequestration i Homotypic Vacuolar Membrane Fusion i Vacuolar Degradation of Peroxisomes Figure 7-4 Pathway of glucose induced peroxisome microautophagy and it relationship with gsa mutants. The microautophagy pathway of peroxisomes during glucose adaptation is depicted here. Gsalp (Pfk1p) assists signaling the onset of microautophagy while Gsa7p acts at a fusion step of the yeast vacuolar membrane. Other mutants that act in different steps of microautophagy are also included in the figure.

PAGE 153

140 Our data of gsa mutants characterization (also see chapter 3) suggest that this genetic model will yield the identities of more unique GSA genes. Furthermore, there exists more than eight complementation groups of gsa mutants. Further screen and biochemical studies of gsa mutants are needed to cover all the regulatory steps and molecules in the microautophagy process and thus more details can be added to the diagram presented in Fig. 7-4. The molecular characterization of microautophagy mutants in P. pastoris I have finished has clearly shown that this autophagy study model will aid us to know better about the poorly understood autophagy mechanism. The studies I have done and those of others should make it obvious that this is one of the most dynamic and fascinating study area in the eukaryotic cells.

PAGE 154

REFERENCES Ahlberg, J., and H. Glaumann. 1985. Uptake-microautophagy-and degradation of exogenous proteins by isolated rat liver lysosomes. Effects of pH, ATP, and inhibitors of proteolysis. Exp Mol Pathol 42: 78-88. Angelov, D. N., W. F. Neiss, A. Gunkel, 0. Guntinas Lichius, and E. Stennert. 1 994. Axotomy induces intranuclear immunolocalization of neuronspecific enolase in facial and hypoglossal neurons of the rat. J Neurocytol 23: 218-33. Arnold, H., and D. Pette. 1968. Binding of glycolytic enzymes to structure proteins of the muscle. Eur J Biochem 6: 163-71. Arvanitidis, A., and J. J. Heinisch. 1994. Studies on the function of yeast phosphofructokinase subunits by in vitro mutagenesis. J Biol Chem 269: 8911-8. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, J. A., and Struhl, K. (1988). Current protocols in molecular biology. K. Janssen, ed. (New York: John Wiley & Sons, Inc). Baba, M., K. takeshige, H. Baba, and Y. Ohsumi. 1994. Ultrastructural analysis of the autophagic process in yeast: Detection of autophagosomes and their characterization. J Cell Biol 124: 903-913. Berger, S. A., and P. R. Evans. 1992. Site-directed mutagenesis identifies catalytic residues in the active site of Escherichia coli phosphofructokinase. Biochemistry 31 : 9237-42. Blangy, D., H. Buc, and J. Monod. 1968. Kinetics of the allosteric interactions of phosphofructokinase from Escherichia coli. J Mol Biol 31 : 1 3-35. Boles, E., J. Heinisch, and F. K. Zimmermann. 1993. Different signals control the activation of glycolysis in the yeast Saccharomyces cerevisiae Yeast 9: 761-70. 141

PAGE 155

142 Bradford, M. M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72: 248-54. Breitenbach-Schmitt, I., J. Heinisch, H. D. Schmitt, and F. K. Zimmermann. 1984a. Yeast mutants without phosphofructokinase activity can still perform glycolysis and alcoholic fermentation. Mol Gen Genet : 530535. Breitenbach-Schmitt, L., H. D. Schmitt, J. Heinisch, and F. K. Zimmermann. 1984b. Genetic and physiological evidence for the existence of a second glycolytic pathway in yeast parallel to the phosphofructokinasealdolase reaction sequence. Mol Gen Genet : 536-540. Chen Zion, M., G. Lilling, and R. Beitner. 1993. The dual effects of Ca2+ on binding of the glycolytic enzymes, phosphofructokinase and aldolase, to muscle cytoskeleton. Biochem Med Metab Biol 49: 173-81. Chiang, H. L., and R. Schekman. 1991. Regulated import and degradation of a cytosolic protein in the yeast vacuole. Nature 350: 313-8. Chiang, H. L., R. Schekman, and S. Hamamoto. 1996. Selective uptake of cytosolic, peroxisomal, and plasma membrane proteins into the yeast lysosome for degradation. J Biol Chem 271 : 9934-41 . Chiang, H. L., and H. L. Shieh. 1997. Identification of novel vesicles in the cytosol to lysosome targeting and degradation pathway. Page 1781. 37th American Society for Cell Biology Annual Meeting. Molecular Biology of the Cell, Supplement, Washington, DC. Clarke, F., P. Stephan, D. Morton, and J. Weidemann. 1985. Glycolytic enzyme organization via the cytoskeleton and its role in metabolic regulation. CRC Press, Boca Raton. Crane, D. I., J. E. Kalish, and S. J. Gould. 1994. The Pichia pastoris PAS4 gene encodes a ubiquitin-conjugating enzyme required for peroxisome assembly. J Biol Chem 269: 21835-44. Cregg, J. M., K. J. Barringer, A. Y. Hessler, and K. R. Madden. 1985. Pichia pastoris as a host system for transformations. Mol Cell Biol 5: 3376-85.

PAGE 156

143 Cregg, J. M., I. J. van der Klei, G. J. Suiter, M. Veenhuis, and H. Harder. 1990. Peroxisome-deficient mutants of Hansenula polymorpha. Yeast 6: 8797. Cregg, J. M., T. S. Vedvick, and W. C. Raschke. 1993. Recent advances in the expression of foreign genes in Pichia pastoris. Biotechnology NY^^: 905-10. de Duve, C. 1965. Functions of microbodies (peroxisomes). J Cell Biol. 27: 25A-26A. Dodt, G., N. Braverman, D. Valle, and S. J. Gould. 1996. From expressed sequence tags to peroxisome biogenesis disorder genes. Ann NY Acad Sci 804: 516-23. Dodt, G., and S. J. Gould. 1996. Multiple PEX genes are required for proper subcellular distribution and stability of Pex5p, the PTS1 receptor: evidence that PTS1 protein import is mediated by a cycling receptor. J Cell Biol 135: 1763-74. Dunn, W. A., Jr. 1990a. Studies on the mechanisms of autophagy: formation of the autophagic vacuole. J Cell Biol 110: 1923-33. Dunn, W. A., Jr. 1990b. Studies on the mechanisms of autophagy: maturation of the autophagic vacuole. J Cell Biol 110: 1935-45. Dunn, W. A. J. 1994. Autophagy and related mechanisms of lysosomemediated protein degradation. Trends in Cell Biology 4: 139-143. Dunn, W. A., Jr., M. K. Raizada, E. S. Vogt, and E. A. Brown. 1994. Growth factor-induced neurite growth in primary neuronal cultures of dogs with neuronal ceroid lipofuscinosis. Int J Dev Neurosci 12: 185-96. Dyson, R. D., J. M. Cardenas, and R. J. Barsotti. 1975. The reversibility of skeletal muscle pyruvate kinase and an assessment of its capacity to support glyconeogenesis. J Biol Chem 250:3316-21 . Egner, R., M. Thumm, M. Straub, A. Simeon, H. J. Schuller, and D. H. Wolf. 1993. Tracing intracellular proteolytic pathways. Proteolysis of fatty acid synthase and other cytoplasmic proteins in the yeast Saccharomyces cerevisiae. J Biol Chem 268: 27269-76.

PAGE 157

144 Gleeson, M. A., and P. E. Sudbery. 1988. The methylotrophic yeasts. Yeast 3: 1-15. Goldfischer, S. 1996. Peroxisomal diseases: a microscopist looks through the retrospectroscope. Ann NY Acad Sci 804: 424-6. Harding, T. M., A. Hefner Gravink, M. Thumm, and D. J. Klionsky. 1996. Genetic and phenotypic overlap between autophagy and the cytoplasm to vacuole protein targeting pathway. J Biol Chem 271 : 17621-4. Harding, T. M., K. A. Morano, S. V. Scott, and D. J. Klionsky. 1995. Isolation and characterization of yeast mutants in the cytoplasm to vacuole protein targeting pathway. J Cell Biol 131: 591-602. Hauser, K., R. Kissmehl, J. LInder, J. E. Schultz, F. Lottspeich, and H. Plattner. 1997. Identification of isoforms of the exocytosis-sensitive phosphoprotein PP63/parafusin in Paramecium tetraurelia and demonstration of phosphoglucomutase activity. Biochem J 323: 289-96. Heikoop, J. C, M. van den Berg, A. Strijiand, P. J. Weijers, W. W. Just, A. J. Meijer, and J. M. Tager. 1992. Turnover of peroxisomal vesicles by autophagic proteolysis in cultured fibroblasts from Zellweger patients. EwrJCe// 6/0/ 57: 165-71. Heinisch, J., and F. K. Zimmermann. 1985. Is the phosphofructokinasereaction obligatory for glucose fermentation by Saccharomyces cerevisiae? Yeast 1: 173-5. Hill, D. J., A. C. Hann, and D. Lloyd. 1985. Degradative inactivation of the peroxisomal enzyme, alcohol oxidase, during adaptation of methanolgrown Candida boidinii to ethanol. Biochem J 232: 743-50. Hoffman, M., and H. L. Chiang. 1996. Isolation of degradation-deficient mutants defective in the targeting of fructose-1 ,6-bisphosphatase into the vacuole for degradation in Saccharomyces cerevisiae. Genetics 143: 1555-66. Huang, P. H., and H. L. Chiang. 1997. Identification of novel vesicles in the cytosol to vacuole protein degradation pathway. J Cell Biol 13680310.

PAGE 158

145 Ishitani, R., K. Sunaga, A. Hirano, P. Saunders, N. Katsube, and D. M. Chuang. 1996. Evidence tliat glyceraldehyde-3-phosphate dehydrogenase is involved in age-induced apoptosis in mature cerebellar neurons in culture. J Neurochem 66: 928-35. Jacoby, J., C. P. Hollenberg, and J. J. Heinisch. 1993. Transaldolase mutants in the yeast Kluyveromyces lactis provide evidence that glucose can be metabolized through the pentose phosphate pathway. Mol Microbiol 10: 867-76. Kaiser, C, S. Michaelis, and A. Mitchell. 1994. Mettiods in Yeast Genetics. Cold Spring Harbor Laboratory Press, Plainville. Kalish. J. E., G. A. Keller, J. C. Morrell, S. J. Mihalik, B. Smith, J. M. Cregg, and S. J. Gould. 1996. Characterization of a novel component of the peroxisomal protein import apparatus using fluorescent peroxisomal proteins. EMSO J 15:3275-85. Kato, N. 1990. Formate dehydrogenase from methylotrophic yeasts. Methods Enzymol 188:459-62. Klionsky, D. J. 1997. Protein transport from the cytoplasm into the vacuole J MembrBiol 157: 105-15. Kopitz, J., G. 0. Kisen, P. B. Gordon, P. Bohley, and P. 0. Seglen. 1990. Nonselective autophagy of cytosolic enzymes by isolated rat hepatocytes. J Cell Biol 1 1 1 : 941 -53. Lardeux, B. R., and G. E. Mortimore. 1987. Amino acid and hormonal control of macromolecular turnover in perfused rat liver. Evidence for selective autophagy. J Biol Chem 262: 14514-9. Lazarow, P. B., and Y. Fujiki. 1985. Biogenesis of peroxisomes. Ann Rev Cell Biol 1: 480-530. Liu, H., X. Tan, K. A. Russell, M. Veenhuis, and J. M. Cregg. 1995. PER3, a gene required for peroxisome biogenesis in Pichia pastoris, encodes a peroxisomal membrane protein involved in protein import J Biol Chem 270: 10940-51. Liu, H., X. Tan, M. Veenhuis, D. McCollum, and J. M. Cregg. 1992. An efficient screen for peroxisome-deficient mutants of Pichia pastoris J Bacterid 174: 4943-51.

PAGE 159

146 Lobo, Z., and P. K. Maitra. 1983. Phosphofructokinase mutants of yeast. Biochemistry and genetics. J Biol Chem 258: 1444-9. Luiken, J. J., M. van den Berg, J. C. Heikoop, and A. J. Meijer. 1992. Autophagic degradation of peroxisomes in isolated rat hepatocytes. FEBS Lett 304: 93-7. Matsuura, A., M. Tsukada, Y. Wada, and Y. Ohsumi. 1997. Apglp, a novel protein kinase required for the autophagic process in Saccharomyces cerevisiae. Gene 192:245-50. McCollum, D., E. Monosov, and S. Subramani. 1993. The pas8 mutant of Pichia pastoris exhibits the peroxisomal protein import deficiencies of Zellweger syndrome cells~the PAS8 protein binds to the COOHterminal tripeptide peroxisomal targeting signal, and is a member of the TPR protein family. J Cell Biol 121 : 761 -74. Mizrachi, Y. 1989. Neurotrophic activity of monomeric glucophosphoisomerase was blocked by human immunodeficiency virus (HIV-1) and peptides from HIV-1 envelope glycoprotein. J Neurosci Res 23: 217-24. Mortimore, G. E., A. R. Poso, and B. R. Lardeux. 1989. Mechanism and regulation of protein degradation in liver. Diabetes Metab Rev 5: 4970. Moser, H. W., and A. B. Moser. 1996. Peroxisomal disorders: Overview. Anna Rev Cell Biol 9: 427-441 . Noda, T., and Y. Ohsumi. 1998. Tor, a phosphatidylinositol kinase homologue, controls autophagy in yeast. J Biol Chem 273:3963-6. Nuttley, W. M., R. K. Szilard, J. J. Smith, M. Veenhuis, and R. A. Rachubinski. 1995. The PAH2 gene is required for peroxisome assembly in the methylotrophic yeast Hansenula polymorpha and encodes a member of the tetratricopeptide repeat family of proteins. Gene 160: 33-9. Olson, T. S., S. R. Terlecky, and J. F. Dice. 1992. Pathways of intracellular protein degradation in eukaryotic cells. Pages 89-1 14 in T. J. Ahern and M. C. Manning, eds. Stability of protein pharmaceuticals. Part 2: In vivo pathways of degradation and strategies for protein stabilization. Plenum Press, New York.

PAGE 160

147 Robbins, A. R., R. D. Ward, and C. Oliver. 1995. A mutation in glyceraldehyde 3-phosphate dehydrogenase alters endocytosis in CHO cells. J Cell Biol 130: 1093-104. Roberts, S. J., and G. N. Somero. 1989. Properties of the interaction between phosphofructokinase and actin. Arch Biochem Biophys 269: 284-94. Ronai, Z. 1 993. Glycolytic enzymes as DNA binding proteins. Int J Biochem 25: 1073-6. Sakai, Y., A. Koller, L. K. Rangell, G. A. Keller, and S. Subramani. 1998. Peroxisome Degradation by Microautophagy in Pichia pastoris: Identification of Specific Steps and Morphological Intermediates. J Cell Biol 141: 625-636. Scott, S. v., A. Hefner Gravink, K. A. Morano, T. Noda, Y. Ohsumi, and D. J. Klionsky. 1996. Cytoplasm-to-vacuole targeting and autophagy employ the same machinery to deliver proteins to the yeast vacuole. Proc Natl Acad Sci USA 93: 12304-8. Shimozawa, N., T. Tsukamoto, Y. Suzuki, T. Orii, Y. Shirayoshi, T. Mori, and Y. Fujiki. 1992. A human gene responsible for Zellweger syndrome that affects peroxisome assembly. Science 255: 1132-4. Shirahama, K., T. Noda, and Y. Ohsumi. 1997. Mutational analysis of Csc1A/ps4p: involvement of endosome in regulation of autophagy in yeast. Cell Struct Fund 22:501 -9. Sibirny, A. A., V. I. Titorenko, and B. D. Efremov. 1987. Multiplicity of mechanisms of carbon catabolite repression involved in the synthesis of alcohol oxidase in the methylotrophic yeast Pichia pastoris Yeast 3233-241 . Smalheiser, N. R. 1996. Proteins in unexpected locations. Mol Biol Cell ? 1003-14. Sreekrishna, K., R. G. Brankamp, K. E. Kropp, D. T. Blankenship, J. T. Tsay, P. L. Smith, J. D. Wierschke, A. Subramani, and L. A. Birkenberger. 1997. Strategies for optimal synthesis and secretion of heterologous proteins in the methylotrophic yeast Pichia pastoris. Gene 190: 55-62. Subramani, S. 1993. Protein import into peroxisomes and biogenesis of the organelle. Annu Rev Cell Biol 9: 445-478.

PAGE 161

148 Subramani, S. 1996. Convergence of model systems for peroxisome biogenesis. CurrOpin Cell Biol 8: 513-8. Takei, N., J. Kondo, K. Nagaike, K. Ohsawa, K. Kato, and S. Kohsaka. 1991. Neuronal survival factor from bovine brain is identical to neuronspecific enolase. J Neurochem 57: 1178-84. Takeshige, K., M. Baba, S. Tsuboi, T. Noda, and Y. Ohsumi. 1992. Autophagy in yeast demonstrated with proteinase-deficient mutants and conditions for its induction. J Cell Biol 1 1 9: 301 -1 1 . Terlecky, S. R., W. M. Nuttley, D. McCollum, E. Sock, and S. Subramani. 1995. The Picliia pastoris peroxisomal protein PAS8p is the receptor for the C-terminal tripeptide peroxisomal targeting signal. EMBO J 14: 362734. Terlecky, S. R., E. A. C. Wiemer, W. M. Nuttley, P. A. Walton, and S. Subramani. 1996. Signals, receptors, and cytosolic factors involved in peroxisomal protein import. Ann NY Acad Sci 804: 11-20. Thumm, M., R. Egner, B. Koch, M. Schlumpberger, M. Straub, M. Veenhuis, and D. H. Wolf. 1994. Isolation of autophagocytosis mutants of Saccharomyces cerevisiae. FEBS Lett 349: 275-80. Titorenko, V. I., I. Keizer, W. Harder, and M. Veenhuis. 1995. Isolation and characterization of mutants impaired in the selective degradation of peroxisomes in the yeast Hansenula polymorpha. J Bacteriol 177: 35763. Tolbert, N. E. 1981. Metabolic pathways in peroxisomes and glyoxysomes. Ann Rev Bioctiem 50: 1 33-1 57. Tomlison, E. 1992. Direct colony assay for alcohol oxidase. Biotechniques 13: 710. Treptau, T., R. Kissmehl, J. D. Wissmann, and H. Plattner. 1995. A 63 kDa phosphoprotein undergoing rapid dephosphorylation during exocytosis in Paramecium cells shares biochemical characteristics with phosphoglucomutase. Biochem J 309: 557-67. Tsukada, M., and Y. Ohsumi. 1993. Isolation and characterization of autophagy-defective mutants of Saccharomyces cerevisiae. FEBS Lett 333: 169-74.

PAGE 162

149 Tuttle, D. L., and W. A. Dunn, Jr. 1995. Divergent modes of autophagy in the methylotrophic yeast Pichia pastoris. J Cell Sci 108: 25-35. Tuttle, D. L., A. S. Lewin, and W. A. Dunn, Jr. 1993. Selective autophagy of peroxisomes in methylotrophic yeasts. Eur J Cell Biol 60: 283-90. van der Klei, I. J., W. Harder, and M. Veenhuis. 1991a. Biosynthesis and assembly of alcohol oxidase, a peroxisomal matrix protein in methylotrophic yeasts: a review. Yeast 7: 195-209. van der Klei, I. J., W. Harder, and M. Veenhuis. 1991b. Methanol metabolism in a peroxisome-deficient mutant of Hansenula polymorpha: A physiological study. Arch Microbiol 156: 15-23. van der Klei, I. J., and M. Veenhuis. 1996. Peroxisome biogenesis in the yeast Hansenula polymorpha: a structural and functional analysis. Ann NY Acad Sci 804: 47-59. Veenhuis, M., A. Douma, W. Harder, and M. Osumi. 1983. Degradation and turnover of peroxisomes in the yeast Hansenula polymorpha induced by selective inactivation of peroxisomal enzymes. Arch Microbiol 134193-203. Veenhuis, M., W. Harder, J. P. van Dijken, and F. Mayer. 1981. Substructure of crystalline peroxisomes in methanol-grown Hansenula polymorpha: evidence for an in vivo crystal of alcohol oxidase. Mol Cell Biol 1 94957. Veenhuis, M., J. P. van Dijken, S. A. Pilon, and W. Harder. 1978. Development of crystalline peroxisomes in methanol-grown cells of the yeast Hansenula polymorpha and its relation to environmental conditions. Arch Microbiol 117:1 53-63. Waterham, H. R., Y. de Vries, K. A. Russel, W. Xie, M. Veenhuis, and J. M. Cregg. 1 996a. The Pichia pastoris PER6 gene product is a peroxisomal integral membrane protein essential for peroxisome biogenesis and has sequence similarity to the Zellweger syndrome protein PAF-1 Mol Cell Biol 16: 2527-36. Waterham, H. R., Y. De Vries, K. A. Russell, W. Xie, M. Veenhuis, and J. M. Cregg. 1996b. The Pichia pastoris PER6 gene encodes the yeast homologue of human PAF-1. Ann NY Acad Sci 804: 773-4.

PAGE 163

150 Waterham, H. R., M. E. Digan, P. J. Koutz, S. V. Lair, and J. M. Gregg. 1997 Isolation of the Pichia pastoris glyceraldehyde-3-phosphate dehydrogenase gene and regulation and use of its promoter Gene 186:37-44. Wills, C. 1990. Regulation of sugar and ethanol metabolism in Saccharomyces cerevisiae. Crit Rev Biochem Mol Biol 25: 245-80. Wills, C. 1996. Some puzzles about carbon catabolite repression in yeast Res MicrobioMAJ: 566-72. Yahraus, T., N. Braverman. G. Dodt, J. E. Kalish, J. C. Morrell, H W Moser D. Valle, and S. J. Gould. 1996. The peroxisome biogenesis disorder group 4 gene, PXAAA1 , encodes a cytoplasmic ATPase required for stability of the PTS1 receptor. EMBO J 15: 2914-23

PAGE 164

BIOGRAPHICAL SKETCH I was born in Shanghai, P. R. China in 1964. My father, Gangjun Yuan, is a head master of an elementary school and my mother, Zhengjuan Xu, is a pediatrician. I graduated from the distinguished high school, Shanghai Middle School, in 1981 and was enrolled in the Second Military Medical University majoring medicine. I graduated in 1 986 after four-years of study and one year's internship. I did a three-year residency at a hospital in Shannxi, China, and got back to Shanghai in 1 989 as an assistant editor of the Journal of Reproduction and Contraception, at Shanghai Institute of Planned Parenthood Research. In August 1993, I entered the Department of Anatomy and Cell Biology in the College of Medicine at the University of Florida for a Ph.D. majoring cell and developmental biology. My research focuses on the regulatory mechanism of protein and cell organelle microautophagy in yeast P. pastoris. 151

PAGE 165

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of PhilosQfihy. William A. Dunn, Jr., Chair Associate Professor of Anatomy and Cell Biology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of PhilosophN Johyi P. Aris Assistant Professor of Anatomy and Cell Biology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Gudrun S. Bennett Research Professor of Anatomy and Cell Biology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosoc Alfred S//Lewin Professor of Molecular Genetics and Microbiology This dissertation was submitted to the Graduate Faculty of the College of Medicine and the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. December 1 998 ^t^^-^^^Qj-^J^^ Dean, College of Medicine Dean, Graduate School