Mechanisms of autophagy in methylotrophic yeasts


Material Information

Mechanisms of autophagy in methylotrophic yeasts
Physical Description:
xiii, 160 leaves : ill. ; 29 cm.
Tuttle, Daniel Lee, 1955-
Publication Date:


Subjects / Keywords:
Research   ( mesh )
Autophagocytosis -- physiology   ( mesh )
Autophagocytosis -- genetics   ( mesh )
Pichia -- anatomy & histology   ( mesh )
Pichia -- physiology   ( mesh )
Pichia -- genetics   ( mesh )
Methanol -- physiology   ( mesh )
Ethanol -- physiology   ( mesh )
Glucose -- physiology   ( mesh )
Nitrogen -- pharmacology   ( mesh )
Endopeptidases   ( mesh )
Microbodies -- physiology   ( mesh )
Vacuoles -- physiology   ( mesh )
Gene Library   ( mesh )
Models, Biological   ( mesh )
Department of Anatomy and Cell Biology thesis Ph.D   ( mesh )
Dissertations, Academic -- College of Medicine -- Department of Anatomy and Cell Biology -- UF   ( mesh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph.D.)--University of Florida, 1995.
Bibliography: leaves 148-159.
Statement of Responsibility:
by Daniel Lee Tuttle.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 002290610
oclc - 49350012
notis - ALP3780
System ID:

This item is only available as the following downloads:

Full Text








I have to thank, first of all, my parents who bestowed on my brothers and

me good genes, lots of love, and a stable home. I hope they are somewhere

where they can know what I have done and that my educational

accomplishments make them proud and give them some repayment for their

sacrifices. Thanks also go to my brother George for his unwavering support and


To the many wise professors and students I have known and my patient

cat Sadie, who have contributed indirectly by example, goes my appreciation.

To my committee members, Al Lewin, Henry Baker, Gudrun Bennett, my

chairman Bill Dunn, and to Jim Cregg and Martin Gleeson, my personal P.

pastors consultants, go all the thanks of someone who could not have done

what he did without them. A special thank you goes to Dave Kendall of the

University of Bath for being a willing and able participant in this project. I also

wish to thank Denny Player for excellent ultramicrotomy and photography work.


ACKNOWLEDGEMENTS ......................................................... ii

LIST O F TA BLES .......................................................... ......................... vi

LIST OF FIGURES ....................................... .................. vii

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

ABSTRACT .............................................. .......................... xi



Methanol Metabolism and Peroxisomes in Yeast ....................... 1
The Yeast Vacuole .......................................... ................. 4
Vacuolar Biogenesis ........................................ .............. 5
The Vacuole Regulates Cytosolic Concentrations of
Substrates ........................................... 9
Protein Degradation in Yeast ..................................... .......... 10
Protein Degradation Within the Endoplasmic Reticulum ........ 11
Protein Degradation in the Cytoplasm ............................... 12
Protein Degradation in the Vacuole .................................... 15
Genetic Analysis in Methylotrophic Yeasts ............................... 30
Classical Genetics ................... ............................................ 30
Molecular Biology ....................................................... 31

2 MATERIALS AND METHODS ........................... ............ 34

Yeast Strains and Media .................... ................. .. 34
Enzyme Assays ............................................. .................... 35
Preparation of Antisera ................................. ..... ............... 36
Electron M icroscopy ......................................... ...................... 39
Protein Synthesis and Degradation .......................... ...... 40
Synthesis of Peroxisomal Proteins .................................... 40
Degradation of Peroxisomal and Mitochondrial Proteins ........ 41

Measurement of Overall Protein Degradation .................... 42
Isolation of Glucose-Induced Selective Autophagy-Deficient (gsa)
Mutants .................................. ..... ...... .................. 43
Mutagenesis and Screening ................................................ 43
G genetic A nalysis................................ ....................... 45

METHYLOTROPHIC YEASTS ............................................... 48

Introduction ................................................................................... 4 8
Loss of Methanol-Induced Enzymes During Glucose Adaptation .. 52
Degradation of Methanol Assimilating Enzymes ................. 52
Loss of Enzyme Activity .................................. ........... 57
Autophagy of Peroxisomes Induced by Glucose ...................... 59
Ultrastructural Examination of Autophagy ........................... 59
Vacuolar Degradation-Deficient Mutants Verify the Site of
Peroxisome Degradation During Glucose Adaptation............... 64
Analysis of PrA/PrB-defective Strains ................................. 65
Peroxisome Degradation in PrA/PrB-Deficient Strains During
Glucose Adaptation .............................................. ........... .. 66
Ultrastructural Observations During Glucose Adaptation ....... 68
Selective Autophagy .................................. ....................... 70
Chapter Sum m ary ........................................ ..................... ... 72

A UTO PHA G Y ............................................. ................... 75

Introduction ................................................. ............................ 75
Mutagenesis, Screening, and Verification ................................. 77
Backcrossing and Complementation ........................................... 81
Biochemical and Ultrastructural Examination of WDY1 and
W DY2............................... .. ............................ 83
Chapter Sum m ary ........................................ ..................... ... 88

PICHIA PASTOR IS ................................................................. 91

Introductio n .................................................................................... 9 1
Ethanol Adaptation ............................................... ............... 92
Characterization of Ethanol Adaptation in Wild Type Cells ..... 92

Investigation of the Site of Inactivation of AOX During
Ethanol A adaptation ..................................... ..................... ... 97

Ethanol Adaptation in gsa Mutants ........................................... 101
Nitrogen Starvation ..................................................................... 103
Comparison of Vacuolar Uptake of Cellular Components ........ 105
Comparison of Selectivity During Nitrogen Starvation and
Glucose and Ethanol Adaptation ......................................... 107

Evaluation of Autophagy Induced by Nitrogen Deprivation
in gsa mutants ..................... ... ................................. 109
Chapter Summary ........................................ .................... 111

6 CONCLUSIONS, MODELS, AND PROSPECTS ......................... 114

Catabolite Inactivation ....................................... 114
Autophagy ................................................... ...................... 118
Microautophagy ........................ ..... ........................ 119
Macroautophagy ........................................ 124
D ivergent Pathw ays ................................................... .............. 127
Glucose Adaptation ............................................................ 128
Ethanol Adaptation ............................................................ 130
Nitrogen Starvation ............................................................ 131
Prospects and Conclusions ..................... ................... ............ 132



REFERENCES ............................................................. 148

BIOGRAPHICAL SKETCH ................................................. 160


Table Page

2-1 Yeast strains ......................................... 34

5-1 Comparison of the characteristics of autophagy induced by various
stim uli ................ .. ........ ..... ... .................. ........ 118


Figure Page

1-1 Methanol metabolism and its compartmentalization in methylotrophic
yeasts ................... ...................... 2

1-2 Processing, sorting, and activation of carboxypeptidase Y, a
representative vacuolar protease ....................................... 7

2-1 Immunoelectron microscopic localization of alcohol oxidase in methanol-
induced P. pastors ................................. ....... ......... ..... 38

3-1 Comparison of the proteins profiles of methanol-induced P. pastors cells
and cells undergoing glucose adaptation ............................ 49

3-2 Turnover of peroxisomal proteins in methanol induction and glucose
adaptation media .................................... ...... ............... 54

3-3 Synthesis of alcohol oxidase by H. polymorpha and P. pastoris during
stationary phase in methanol-induced cultures and in glucose
adapting cultures relative to synthesis during exponential
growth (log phase) ......................... ... ........................ 56

3-4 Degradation of peroxisomal and cytosolic enzymes during glucose
adaptation ............................. .... ..................... ... 58

3-5 Ultrastructure of methylotrophic yeasts grown in glucose and
m ethanol .................. .... .... .... ........................ 60

3-6 Morphological characterization of glucose-induced peroxisomal
degradation in P. pastoris ..................................................... 61

3-7 Morphological characterization of glucose-induced peroxisomal
degradation in H. polymorpha .............................................. 63

3-8 Protein degradation induced by histidine starvation is blocked in the
absence of functional proteinases A and B ......................... 67

3-9 Proteinase mutants are not able to degrade methanol-induced
components during glucose adaptation ............................... 69

3-10 Degradation of peroxisomal and mitochondrial proteins during
glucose-mediated peroxisome removal in H. polymorpha
and P pastoris ................................ ............... ............. 71

4-1 Screening for glucose-induced selective autophagy-deficient
mutants (gsa) by direct colony assay for alcohol oxidase ....... 78

4-2 Comparison of degradation of methanol-induced enzymes in vacuolar
protease-deficient mutants, gsa mutants, and the parental
strain (G S 115) ............................................. ......... .......... 80

4-3 Analysis of specific and nonspecific degradation by gsa mutants ..... 84

4-4 Ultrastructural examination of glucose adaptation in WDY2 ........... 87

5-1 Degradation of peroxisomal and cytosolic enzymes during ethanol
adaptation ....................... .... .................... 93

5-2 The effects of cycloheximide on the degradation of methanol-induced
enzymes and peroxisomes during ethanol and glucose
adaptation .................. .. ............... ...... 95

5-3 Morphological characterization of ethanol-induced peroxisomal
degradation in P. pastors ..................... ......................98

5-4 Proteinase mutants are not able to degrade methanol-induced
components during ethanol adaptation .................................. 100

5-5 Comparison of ethanol adaptation in GS115 and WDY2 ................ 102

5-6 Effects of nitrogen starvation on parental and proteinase mutant
P. pastoris ............... .. ... ....................................... 104

5-7 Effect of nitrogen deprivation on peroxisomal, cytosolic, and
mitochondrial enzymes ................................... .................. 108

5-8 Comparison of autophagy of peroxisomes induced by glucose,
ethanol, and nitrogen deprivation in wild type and gsa mutant
P. pastors strains ..................... .................................. 110

6-1 Proposed steps in microautophagic uptake of peroxisomes by the
vacuole during glucose adaptation in P. pastors .................. 121

6-2 Proposed steps in microautophagic uptake of random cytoplasmic
components during nitrogen deprivation in P. pastors. .......... 123

6-3 Proposed steps in macroautophagic uptake of peroxisomes by the
vacuole during ethanol adaptation in P. pastors ................... 126

6-4 Postulated events mediated by unknown molecules which lead to
the degradation of cytoplasmic components during glucose
and ethanol adaptation and nitrogen starvation...................... 129

A-1 Map of the E. coli P. pastors shuttle vector pYM8 ......................... 141

A-2 Test of the ability of integrated pDLT1 to complement the autophagy
defect in W DY2 ...................... .... ............................... 145

A-3 Proposed restriction map of genomic library insert in pDLT1 .............. 146


AOX: alcohol oxidase
BSA: bovine serum albumin
CCO: cytochrome c oxidase
DHAS: dihydroxyacetone synthase
DSM: diploid selection medium
EA: ethanol adaptation medium
F1p: mitochondrial F1 ATPase, 3 subunit
FBP: fructose-1,6- bisphosphatase
FDH: format dehydrogenase
GA: glucose adaptation medium
gsa: glucose-induced selective autophagy mutant
LSM: low sulfate medium
MIM: methanol induction medium
SA: heat-killed, glutaraldehyde-fixed Staphylococcus aureus cells
SM: sporulation medium
TX100: triton X-100
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




May 1995

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

When the yeasts Hansenula polymorpha and Pichia pastors utilize

methanol as sole carbon and energy source, peroxisomes and cytosolic

enzymes which metabolize it are induced. If the cells are subsequently switched

to media containing glucose or ethanol, the superfluous peroxisomes and

cytosolic methanol assimilation enzymes are degraded. The purpose of the

present project was to examine the process of turnover of cellular components

utilizing biochemical, ultrastructural, and genetic means, thereby advancing the

understanding of the mechanisms of autophagy.

Cellular proteins were radioactively labeled to show that glucose

mediates the selective, enhanced degradation of methanol-induced

peroxisomes. Utilizing mutant strains of P. pastoris, which contain null alleles for

vacuolar proteinase A and proteinase B, rendering the vacuole proteolytically

inactive, it was determined that the vacuole is the site of degradation of

peroxisomes and cytosolic format dehydrogenase. Morphological examination

revealed species- and nutrient-specific modes of delivery of cytoplasmic

components to the vacuole for degradation. Macroautophagic sequestration of

peroxisomes is induced in H. polymorpha by glucose and in P. pastors by

ethanol. In contrast, glucose initiates selective microautophagy in P. pastors.

Nitrogen deprivation also initiates microautophagy in P. pastors, but of a

morphologically distinct type which degrades cytoplasmic components

nonselectively. Microautophagy in P. pastors requires protein synthesis

whereas macroautophagy does not. Models are presented which illustrate

hypothesized steps in the ultrastructure and mechanisms of the various modes

of autophagy.

In order to evaluate the molecular basis of autophagy, mutants defective

in glucose-induced selective autophagy (gsa) were isolated and characterized.

Members of two complementation groups, WDY1 and WDY2, were determined

to have proteolytically active vacuoles but were unable to conduct

microautophagy in response to glucose or nitrogen starvation. Macroautophagy

proceeds normally in these mutant strains in response to ethanol. A genomic

library clone has been isolated which functionally complements the genetic

defect in microautophagy in WDY2.

This analysis of P. pastors provides the first models for microautophagy

in yeast. The models and tools for genetic analysis of autophagy presented

herein will be useful for evaluating the molecular mechanisms of selective and

nonselective autophagy in yeast.


Methanol Metabolism and Peroxisomes in Yeast

Certain species of yeast within the genera Candida, Hansenula, and

Pichia are able to utilize methanol as the sole source of carbon and energy so

are referred to as methylotrophic (Fukui et al., 1975; van Dijken et al., 1975).

Many of the enzymes necessary to assimilate methanol are not present in

detectable amounts in cells grown on other carbon and energy sources and

consequently must be induced during adaptation to methanol-limited media.

Hydrogen peroxide is one of the products of the first enzyme-mediated reaction

of methanol metabolism (Figure 1-1); thus, as with other enzymes which

generate hydrogen peroxide, this enzyme (alcohol oxidase; AOX) and certain

others of this pathway are located within peroxisomes. Alcohol oxidase in its

active form is a homo-octameric flavoprotein containing FAD as a prosthetic

group (van der Klei et al., 1990). AOX has a low affinity for its substrate thus

large quantities of it are necessary for the yeast to subsist on methanol (van

Dijken et al., 1975; 1982; Harder et al., 1987; Gleeson and Sudberry, 1988).

Crystalline inclusions composed of alcohol oxidase (AOX) form in methanol-

induced peroxisomes, occupying most of the lumen of peroxisomes, which have

a diameter of between 0.9 and 1.3 pm. These peroxisomes are surrounded by a

Figure 1-1. Methanol metabolism and its compartmentalization in methylotrophic
yeasts. Methanol is oxidized in the peroxisome by alcohol oxidase (AOX)
yielding formaldehyde and H202 which is rapidly detoxified by catalase (CAT).
Formaldehyde is then either oxidized in the cytosol following interaction with
reduced glutathione (GSH) by formaldehyde dehydrogenase (FAH) and format
dehydrogenase (FDH) or assimilated via the xyulose 5-phosphate cycle. The
enzymes active in the xyulose 5-phosphate cycle include dihydroxyacetone
synthase (DHAS) which catalyzes the reaction of xyulose 5-phosphate (Xu5-P)
with formaldehyde, yielding glyceraldehyde 3-phosphate (GAP) and
dihydroxyacetone (DHA), dihydroxyacetone kinase (DHAK) yielding
dihydroxyacetone phosphate (DHAP), fructose 1,6-bisphosphate aldolase
(FBAse) yielding fructose bisphosphate (FBP), and fructose 1,6-bisphosphatase
(FBPase) yielding fructose 6-phosphate (FeP). By a series of rearrangement
reactions, xyulose 5-phosphate is regenerated to complete the cycle. The
pathway is responsible for the generation of GAP made available for production
of biomass. Underlined and boxed acronyms are enzymes, boxed acronyms are
enzymes which have been used as markers for degradation in the present study.
Adapted from Gleeson and Sudberry (1988).

unit membrane and appear in clusters in the cell, occupying as much as 80% of

cell volume under optimal growth conditions (Gleeson and Sudberry, 1988). The

two peroxisomal proteins AOX and dihydroxyacetone synthase together may

account for > 50% of the protein mass in a fully methanol-induced cell (see Fig.

3-1). The peroxisomes are visible even at the light level and clustered

peroxisomes are the dominant feature in cells viewed by electron microscopy.

In methylotrophic yeasts, methanol is utilized in two ways: a catabolic

pathway to produce energy in the form of NADH2 or an anabolic pathway to

produce substrates for the synthesis of biomass (Fig. 1-1; van Dijken et al.,

1975;1982; Harder et al., 1987; Gleeson and Sudberry, 1988). The first step of

methanol utilization takes place in the peroxisome and entails its oxidation by

AOX (also referred to as methanol oxidase) the products of which reaction are

hydrogen peroxide and formaldehyde. Catalase is diffused in the peroxisomal

matrix and detoxifies the hydrogen peroxide formed, producing water and

molecular oxygen. Formaldehyde is the substrate which can be directed down

either the catabolic or anabolic pathways. Catabolism involves a linear pathway

involving the complete oxidation of the formaldehyde after it diffuses into the

cytoplasm. In the cytosol, formaldehyde reacts with reduced glutathione and is

subsequently oxidized by formaldehyde dehydrogenase and format

dehydrogenase (Fig. 1-1, top half). NAD is reduced to NADH2 in the last two

steps, yielding two reducing equivalents per molecule of methanol.

Alternatively, methanol can be assimilated into cellular mass via the

xyulose 5-phosphate cycle (Fig. 1-1, bottom half). The first step in this cycle is

the reaction between formaldehyde and xyulose 5-phosphate, catalyzed by the

peroxisomal enzyme dihydroxyacetone synthase (DHAS). The products of this

reaction, glyceraldehyde 3-phosphate (GAP) and dihydroxyacetone, diffuse into

the cytosol and undergo a series of rearrangement reactions which serve to

regenerate xyulose 5-phosphate and produce 1/3 molecule of GAP available for

biomass production per molecule of methanol assimilated. The methanol

utilization enzymes AOX, DHAS, and format dehydrogenase (FDH) are

undetectable in cells grown on carbon sources other than methanol, e.g.,

glucose and ethanol.

The Yeast Vacuole

The vacuole serves as the main degradative organelle in Saccharomyces

cerevisiae, the yeast that has been most extensively studied relative to

degradation and vacuolar function. In some ways the vacuole is the functional

equivalent of the mammalian lysosome: 1) Vacuolar component proteins are

synthesized and directed to the vacuole via the secretary pathway and

proteolytically activated in the vacuolar lumen. 2) It is acidic and contains a

variety of hydrolytic enzymes. 3) It is the site of degradation of endocytosed

materials. 4) Recent evidence suggests that it is also the site of degradation of

intracellular components, i.e., the endpoint for materials sequestered by

autophagy. In other ways it is quite different: 1) while in mammals there are

many small lysosomes in each cell, the yeast vacuole is quite large, occupying

10-20% of the cell volume, and usually there is only one, lobulated vacuole per

cell; and 2) the yeast vacuole actively and precisely regulates cytosolic

concentration of many different constituents.

Vacuolar Biogenesis

A discussion of the biogenesis of the yeast vacuole includes biosynthesis,

sorting, targeting, and processing of the hydrolases which reside in the vacuole

(Fig. 1-2; for reviews see Klionsky et al., 1990; Raymond et al., 1992). The

hydrolase which has been most highly studied and has served as a model of a

typical soluble vacuolar protein is carboxypeptidase Y (CPY).

Processing and sorting of vacuolar protease. Temperature-sensitive

mutants in the secretary pathway, sec61and sec62, block endoplasmic reticulum

(ER) translocation at the restrictive temperature, causing the accumulation of the

unglycosylated, signal sequence-containing form of CPY (preproCPY). At the

permissive temperature, the 20 amino acid signal peptide is proteolytically

removed at the time of translocation into the ER. These data suggest that, as is

usual for proteins in the secretary pathway, CPY is synthesized as a precursor

that translocates into the ER at which time the N-terminal signal sequence is

cleaved yielding the zymogen proCPY (Blachly-Dyson and Stevens, 1987;

Johnson et al., 1987).

Dolichol-mediated glycosylation occurs in the ER with the addition of core

oligosaccharides at four sites on proCPY (Trimble et al., 1983). ProCPY is

further modified in the Golgi apparatus with the elongation of the core

oligosaccharides with additional mannose residues yielding the fully

glycosylated Golgi precursor form (see Fig 1-2). Unlike the long mannose outer

chains of secreted yeast proteins, CPY and the other vacuolar proteins undergo

limited elongation. There is strong evidence that the type of glycosylation

modifications is not responsible for sorting of secretary proteins away from

vacuolar proteins. This evidence includes the ability of cells to correctly sort

vacuolar proteins in the presence of tunicamycin, which blocks the addition of N-

linked oligosaccharides to proteins (Klionsky et al., 1988). Furthermore, hybrid

proteins which contain sections of vacuolar proteins fused to the secretary

protein invertase are efficiently delivered to the vacuole despite the fact that the

same carbohydrate elongation takes place as in the secreted wild type invertase

(Johnson et al., 1987; Klionsky et al, 1988). These data suggest that the sorting

signals reside in the polypeptide chains themselves. It was determined, using

fusion proteins and site-directed mutagenesis, that the vacuolar targeting signal

of CPY resides in the N-terminal region of proCPY (Valls et al., 1987). This is in

contrast to the case of mammalian lysosomal proteins, which are diverted from

the secretary path by the interaction of mannose 6-phosphate residues attached

to the proteins and mannose 6-phosphate receptors in the Golgi (reviewed in

Griffiths et al., 1988; Kornfeld and Mellman, 1989).

Transport of vacuolar proteases to the vacuole. Data have recently been

presented which suggests that Golgi-modified vacuolar proteins do not traffic



Figure 1-2. Processing, sorting, and activation of carboxypeptidase Y. a
representative vacuolar protease. Newly synthesized carboxypeptidase Y
(preproCPY) enters the secretary pathway after translation on the rough
endoplasmic reticulum (RER) at which time the signal sequence (pre: denoted
by the black box) is cleaved. N-linked core glycosylation occurs at four sites on
the protein in the RER (vertical lines) which is further processed in the Golgi by
addition of mannose residues (branching diagonal lines). In the late Golgi
vacuolar proteins are sorted from proteins destined to be secreted (e.g.,
invertase = INV). Fully glycosylated CPY (proCPY or pCPY) leaves the Golgi,
presumably in vesicles, which fuse with endocytic vesicles, in this case
containing endocytosed alpha-factor (aF), before finally fusing with the vacuolar
membrane where the pCPY is deposited into the vacuolar lumen. This inactive,
zymogen form of CPY is proteolytically cleaved, in turn, by the endoproteases
proteinase A (PrA) yielding active CPYA and proteinase B (PrB) yielding the
protein of mature size (CPYB).

directly to the vacuole but rather by way of an endosome-like intermediate

(Vida et al., 1993). In these studies, intercompartmental transport was

reconstituted in permeabilized spheroplasts using pulselchase protocols and

subcellular fractionation. After a 5 minute pulse and 30 second chase, 30 to 40%

of the Golgi-modified precursor form of CPY fractionated at a density

intermediate between the vacuole and Golgi. Mutant CPY lacking a functional

sorting signal was detected in Golgi fractions but not the intermediate density

compartment, indicating that the intermediate density compartment is

downstream of the Golgi. In addition, the Golgi-modified CPY co-fractionated

with "S-a-factor internalized at 150C. The authors concluded that CPY is sorted

away from secretary proteins in the late Golgi and proceeds to the vacuole via

an endosome-like intermediate (see Fig. 1-2). Davis et al. (1993) provided

further evidence for convergence of the endocytic and vacuolar biogenesis

pathways based on the fact that REN1, a gene which blocks a-factor receptor

delivery to the vacuole, and VPS2, a gene required to deliver newly synthesized

vacuolar proteases to the vacuole are identical genes.

Vacuolar biogenesis mutants. A large set of mutants has been isolated

which are defective in vacuolar biogenesis (see Rothman et al., 1989 for a

description of the genes involved). At least 49 genes in the PEP, END, VPL,

VPS, and VPT complementation groups are required to target and deliver

soluble enzymes, including CPY, to the vacuole. These genes include a

member of the ras-like GTP binding family with striking similarity to mammalian

rab5 (VPS21; Horazdovasky et al., 1994); a phosphatidylinositol 3-kinase and a

protein kinase acting as part of a hetero-oligomeric protein complex (VPS34 and

VPS15, respectively; Stack et al., 1993); and a peripheral vacuolar membrane

protein (PEP3: Preston et al., 1991; a.k.a. VPS18: Robinson et al., 1991) among

others. Of these genes one of the best characterized is PEP4, encoding the

vacuolar endoprotease proteinase A (PrA), which is required for the processing

of vacuolar zymogens and overall proteolytic activity of the vacuole (Jones,

1984; Ammerer et al., 1986; Hirsch et al., 1992).

Processing of vacuolar zvmogens. A total of seven vacuolar proteases

are known: two endoproteases, PrA and proteinase B (PrB); two

carboxypeptidases, CPY and carboxypeptidase S; two aminopeptidases,

aminopeptidase I (AP-I) and aminopeptidase Co; and dipeptidyl aminopeptidase

B (reviewed in Jones, 1991a; Knop et al., 1993). Of these PrA, PrB, AP-I, and

CPY are transported to the vacuole in inactive, zymogen form while the others

are transported in their active forms. The pathway of activation of these

zymogens has been worked out in detail by genetic and biochemical means

revealing that PrA autocatalyzes its own proteolytic cleavage to active form and

then acts on PrB, AP-I, and CPY to activate them (Ammerer et al., 1986;

Teichert et al., 1987; Hirsch et al., 1992). While the PrA-catalyzed cleavage will

activate these proteases, this cleavage alone is not enough to process the

zymogens to mature size and form; this requires a further clip by PrB, though the

functional significance of this final trimming is not known (see Fig. 1-2).

The Vacuole Regulates Cytosolic Concentrations of Substrates

The vacuolar H" ATPase is the enzyme most responsible for generating

an electrochemical potential difference of protons across the vacuolar

membrane, maintaining the vacuolar lumen at ~ pH 6.0 (Preston et al., 1989).

The proton gradient is the primary force for the transport of amino acids, Ca2, P1,

Zn2', Mg2*, and other ions via proton antiporters (Ohsumi et al., 1981; 1983).

Eight different amino acid/proton antiporters have been identified in the vacuole

of S. cerevisiae, specific for arginine, arginine-lysine, histidine, phenylalanine-

tryptophan, tyrosine, glutamine-asparagine, and leucine-isoleucine (Sato et al.,

1984). While amino acid concentrations in the cytosol remain relatively

constant, those in the vacuole fluctuate widely depending on availability of

nutrients and stage of growth (Kitamoto et al., 1988). Arginine is usually the

major amino acid stored in the vacuole, probably since it is the most nitrogen-

rich amino acid and therefore serves as a nitrogen reserve. In fact, nitrogen

starvation evokes a transfer of much of the vacuolar arginine pool into the


Protein Degradation in Yeast

Proteolysis is an integral part of post-translational control that functions

either by limited clips that regulate the function of a protein or total degradation

to regulate the quantity of a protein or to dispose of defective proteins (see

Teichert et al., 1987; Jones, 1991a; Rendueles and Wolf, 1988; Holzer, 1976).

Proteolytic activities have been found in nearly all compartments of yeast,

including mitochondria, endoplasmic reticulum, cytoplasm, vacuole, and

periplasm. Particular instances of limited proteolysis which activate vacuolar

zymogens have been discussed above so here I will review what is known about

the total degradation pathways which occur in the endoplasmic reticulum,

cytoplasm, and especially the vacuole.

Protein Degradation Within the Endoplasmic Reticulum

In order to show degradation within the endoplasmic reticulum, it must be

shown that lysosomal/vacuolar proteases are not involved and that the protein

does not exit the endoplasmic reticulum. Protein degradation within the

endoplasmic reticulum appears to include elimination of misfolded mutant

proteins as well as regulation of levels of normal proteins. The control of 3-

hydroxy-3-methylglutaryl coenzyme A reductase (HMGR) levels in the

endoplasmic reticulum has been studied in mammalian cells to reveal a highly

regulated system modulating synthesis and degradation depending on cellular

needs (Nakanishi et al., 1988). The half-life of HMGR in cultured cells varies

between 40 minutes and >10 hours depending on the rate of synthesis in the

mevalonate pathway of isoprenoid synthesis. Recent work in S. cerevisiae has

suggested that HMGR levels are regulated in a manner similar to that observed

in mammals and that it is normally degraded in the endoplasmic reticulum

(Hampton and Rine, 1994).

The degradation of abnormal proteins in the endoplasmic reticulum has

also been studied in S. cerevisiae (Finger et al., 1993). In this study, the fate of

mutant forms of the vacuolar enzymes CPY and PrA was determined. The

mutant PrA form exhibited no proteolytic activity in vitro while the mutant CPY

was susceptible to trypsin cleavage in vitro, unlike the wild type form, indicating

an altered structure. No Golgi-specific carbohydrate modifications were evident

on either protein and subcellular fractionation studies revealed an endoplasmic

reticulum localization. The degradation rates of both aberrant enzymes were

similar, suggesting that both were degraded by a similar mechanism within the

endoplasmic reticulum.

Protein Degradation in the Cytoolasm

Proteinase E is the S. cerevisiae equivalent of the mammalian

proteasome (Achstetter et al., 1984a,b), a multifunctional enzyme complex

responsible for stress-induced proteolysis of cytoplasmic components

(Heinemeyer et al., 1991), for degradation of abnormal proteins (Egner et al.,

1993), and probably for degradative catabolite inactivation of at least some

glucose-sensitive enzymes, e.g., phosphoenolpyruvate carboxykinase (Burlini et

al., 1989), malate dehydrogenase (Funagama et al., 1985), and possibly

fructose-1,6-bisphosphatase, though this is controversial (Teichert et al., 1989;

Schork et al., 1994; Chiang and Schekman, 1991). The proteasome has been

highly conserved from yeast to man and is composed of 12 to 14 subunits of

between 20 to 35 kDa comprising an enzyme complex of a total of about 700

kDa forming a hollow cylinder which displays a variety of proteolytic activities.

Heinemeyer et al. (1991) isolated two complementation groups (PRE1; PRE2)

which are defective in the "chymotrypsin-like" activity of the S. cerevisiae

proteasome. The PRE1 gene is essential for life and encodes a 22.6 kDa

protein with significant homology to proteasome subunits of the rat and

Drosophila. Diploids homozygous for the temperature-sensitive allele prel-1 are

defective in sporulation and under nutritional and temperature stress conditions

exhibit decreased protein degradation and accumulate ubiquitin-protein


In order to ensure that the degradation of cytosolic proteins is limited to a

specific subset of those present, the proteins to be degraded must be targeted to

the proteasome in some way. The ubiquitin-linked proteolytic pathway serves

this purpose in organisms as diverse as mammals and yeast (Finley and Chau,

1991). Proteins to be degraded are multi-ubiquitinated on specific lysine

residues (Chau, 1989; Johnson et al., 1992). Similarly to the results for

proteasome mutants discussed above, S. cerevisiae mutants which are unable to

ubiquitinate proteins either due to the inability to express ubiquitin (Finley et al.,

1987) or are unable to express ubiquitin conjugating enzymes (E2s; Seufert and

Jentsch, 1990) are susceptible to stress. The fact that proteasome mutants

exhibit decreased proteolysis rates and accumulate ubiquitin-protein conjugates

under stress conditions (Heinemeyer et al., 1991) and decreased degradation of

proteins which are subject to ubiquitin-mediated degradation in wild type cells

(Seufert and Jentsch, 1992), strongly suggest that the proteasome is part of the

ubiquitin system which is responsible for a substantial portion of cytosolic protein


The glucose effect occurs in yeast cells following addition of glucose to

cultures which were previously growing in media containing an unrelated carbon

source. In these cases glucose catabolites act to transcriptionally repress the

synthesis of certain enzymes (catabolite repression) and also to induce the

inactivation (by degradation or otherwise) of certain enzymes already present

(catabolite inactivation; Holzer 1976; Holzer and Purwin, 1986). As was noted

above, some controversy exists regarding the site of degradation of proteins

undergoing catabolite inactivation as during adaptation to glucose (see Schork

et al., 1994 and response of H.-L. Chiang and R. Schekman following). This

debate is based on the use of vacuolar and proteasomal mutants to determine

the site of degradation. Vacuolar mutants have been used in several cases to

indicate that certain glucose-sensitive enzymes are not degraded in the vacuole

(see above; Burlini et al., 1989; Funagama et al., 1985; Chiang and Schekman,

1991). But, in the case of fructose-1,6-bisphosphatase, one laboratory reports

that a proteasome mutant was unable to degrade it while vacuolar mutants could

(Teichert et al., 1989; Schork et al., 1994). In another laboratory (Chiang and

Schekman, 1991) vacuolar mutants were unable to degrade this protein in

response to glucose. These investigators suggested that the proteasome

mutants used by the former group exhibited a pleiotropic phenotype and

therefore are not reliable, and Chiang and Schekman are doing further work to

settle the matter (response to Schork et al., 1994). The role of the proteasome

in degradation in yeast is not well defined at this point in time due to a relative

lack of research on this system compared to the huge amount of research which

has been conducted on the vacuole.

Protein Degradation in the Vacuole

The majority of the work done regarding the degradative role of the

vacuole to date has involved the investigation of the proteases which are

resident and their proteolytic specificities. This has left the questions of how and

which proteins and organelles enter the vacuole largely unanswered. As with

the mammalian lysosome, the major routes of entry into the vacuole are via the

secretary pathway (biogenesis), via endocytosis, and via autophagy. Biogenesis

has already been discussed; consequently this section shall mainly deal with

endocytosis and autophagy.

Endocvtosis. Internalization of alpha-factor (aF), one of the two peptide

hormones responsible for synchronizing mating between the two cell types of S.

cerevisiae, and its receptor serves as the model system for endocytosis in yeast.

aF receptor is endocytosed constitutively when not bound to its ligand and in a

regulated manner when bound to aF (Davis et al., 1993). Normal rates of aF

receptor internalization is dependent upon clathrin, suggesting internalization via

clathrin coated pits as commonly occurs with this class of G protein-coupled

receptors in higher eukaryotes (Tan et al., 1993). After internalization, aF bound

to its receptor proceeds via vesicular intermediates (Raths et al., 1993; Davis et

al., 1993; Singer and Reizman, 1990; Schimmoller and Reizman, 1993) through

early and late endosome compartments (Davis et al., 1993; Wichman et al.,

1992; Schimmoller and Reizman, 1993). The early endosome may be the site of

convergence of the vacuolar biogenesis and endocytic pathways (see Fig. 1-2;

Vida et al., 1993; Davis et al., 1993) where vacuolar and endocytosed proteins

transiently reside in identical compartments. Transport between endosomal

compartments is dependent on YPT7, a ras-like small guanine nucleotide-

binding protein (Wichmann et al., 1992; Schimmoller and Reizman, 1993).

Finally, aF is delivered to the vacuole where it is degraded (Wichmann et al.,


Autophaav. Eukaryotic cells possess specific systems for degrading their

own proteins and organelles when they become superfluous, or have

deteriorated, or when the cell is under stress and requires substrate for new

protein synthesis and organelle biogenesis. Autophagy is the process whereby

the lysosome/vacuole takes up cellular components either by surrounding

portions of the cytoplasm within invaginations of its membrane and assimilating

and degrading the contents (microautophagy; Ahlberg et al., 1985; Mortimore et

al., 1988) or by fusing with and assimilating the contents of autophagosomes (or

autophagic vacuoles) in which portions of cytoplasm have been sequestered

(macroautophagy; Dunn, 1990a,b; Takeshige et al., 1992; Baba et al., 1994). In

mammalian cells, autophagic vacuoles are formed in response to amino acid

deprivation in perfused rat livers when portions of the rough endoplasmic

reticulum (RER) lose their ribosomes, pinch off from the main body of the RER,

and surround portions of the cytoplasm sequestering the contents from the

remainder of the cytoplasm (Dunn, 1990a). The autophagic vacuoles formed

may contain ribosomes, mitochondria, peroxisomes and soluble proteins. The

newly formed autophagic vacuoles rapidly become acidic, possibly by acquiring

vacuolar H' ATPase complexes by fusion with vesicles containing them derived

from the Golgi apparatus (Dunn, 1990b). The autophagic vacuole then acquires

hydrolytic enzymes by fusing with primary lysosomes and thereby matures into a

degradative autolysosome (Dunn, 1990b). Macroautophagy in yeast has not

been so well characterized but analysis of autophagy mutants should soon yield

more data on this process (Tsukada and Ohsumi, 1993; Thumm et al., 1994) and

will be discussed in the next section.

Nonselective autophaav. Autophagy is considered by many to be mainly

responsible for a nonselective bulk turnover of proteins which occurs at the site

for nonselective degradation, i.e., the lysosome/vacuole, in response to

nutritional or developmental stress (Takeshige, et al., 1992; Simeon et al., 1992;

Mortimore and P6so,1987; Knop et al., 1993; Ahlberg et al., 1985; Jones, 1991a;

Kopitz et al., 1990). Kopitz et al. (1990) used electrodisruption of the plasma

membrane of isolated hepatocytes to deplete cells of their cytosol and soluble

proteins to form "cell corpses" which still contain their organelles, including

autophagosome. In this study seven proteins for which normal half-lives where

calculated and which ranged from 0.9 h to 17.4 h were found to be sequestered

into isolated autophagosomes at the same rate in the presence of inhibitors of

lysosomal proteolysis. The authors concluded that the autophagic-lysosomal

system operates nonselectively.

Evidence also exists for nonselective uptake of cytoplasmic components

by autophagy in S. cerevisiae. Utilizing mutants strains of S. cerevisiae which

contained null mutant copies of the genes for the vacuolar endoproteases PrA

and PrB, it was found that the vacuole is responsible for about 40% of the

proteolysis which occurs during logarithmic growth but this increases to 85%

during starvation for nitrogen (Teichert et al., 1989). After 24 h of nitrogen

starvation, 45% of all cellular proteins had been degraded in the vacuole.

Cytosolic proteins which are normally degraded independently of the vacuole,

e.g., enzymes of the gluconeogenic pathway (fructose-1,6-bisphosphatase and

phosphoenolpyruvate carboxykinase) were found to accumulate in the vacuoles

of nutrient deprived vacuolar mutants (Egner et al., 1993). Coincident with the

accumulation of cytosolic proteins in the vacuole there was an accumulation of

vesicular bodies within the vacuolar lumen, suggesting that cytoplasmic contents

are being taken up into the vacuole by autophagy but are not degraded because

of the lack of protease activity in vacuolar mutant cells (Egner et al., 1993;

Simeon et al., 1992).

Ohsumi and his colleagues (Takeshige et al., 1992; Baba et al., 1994)

have been studying autophagy in S. cerevisiae by rendering the vacuole

proteolytically inactive either by utilizing strains lacking PrA and PrB or by

treating normal cells with the serine protease inhibitor phenylmethylsulfonyl

fluoride (PMSF). As a result of depriving the cells of nitrogen, carbon, or

essential amino acids, "autophagic bodies" were seen to accumulate in the

vacuoles of cells with inactivated vacuoles. The concentrations of three

cytosolic marker enzymes, alcohol dehydrogenase, glucose-6-phosphate

dehydrogenase, and phosphoglycerate kinase, were present in the autophagic

bodies at the same concentration as in the cytosol but were not present in the

vacuolar sap. The autophagic bodies also contained ribosomes, mitochondria,

RER, and glycogen and lipid granules at approximately the same density as in

the cytoplasm. Using a freeze-substitution method of fixation for transmission

electron microscopy which preserves yeast ultrastructure remarkably well,

spherical structures bound by double membranes were observed in the

cytoplasm near the vacuole which contained cytoplasmic components and had a

similar appearance to autophagic vacuoles seen in mammalian systems (Dunn,

1990a,b). These autophagic vacuoles were also observed with their outer

membranes continuous with the vacuolar membrane as if fusing with the

vacuoles and depositing the autophagic bodies into the vacuolar lumen.

Taken together, these results strongly suggest that nutrient deprivation in

S. cerevisiae initiates a process whereby random portions of the cytosol are

surrounded by a double membrane of unknown origin forming autophagic

vacuoles. These autophagic vacuoles then fuse with and deposit their contents

into the vacuole where they are degraded to provide substrates for new

synthesis which will allow the cells to adapt to prevailing nutrient conditions and

remain viable.

Selective autophagy. A review of the mammalian literature reveals

several interesting instances of selective autophagy. When fed clofibrate (an

anti-hyperlipidemia drug), rats strongly induce peroxisomes containing the

enzymes catalase and acyl-CoA oxidase (Luiken et al., 1992). In three separate

experiments, hepatocytes were isolated from rats which had been fed clofibrate

and then incubated in the absence of amino acids for 4 h to induce the

autophagic pathway. The levels of peroxisomal and cytosolic proteins before

and after incubation were determined. The relative levels of both peroxisomal

enzymes were ~60% after incubation when compared to 0 h levels; this was

significantly lower than the cytosolic marker enzymes levels which were all in the

range of 79 to 89% of initial. Co-incubation with 3-methyl-adenine, an inhibitor

of autophagy, or long-chain fatty acids, which are substrates for acyl-CoA

oxidase, blocked the decrease in peroxisomal enzymes while short-chain fatty

acids, not substrates for acyl-CoA oxidase, did not. These data indicate that

peroxisomes are specifically degraded in cases in which the enzymes they

contain are not needed (i.e., when incubated without acyl-CoA substrates).

Patients with cerebro-hepato-renal Zellweger syndrome and cultured

fibroblasts from patients with this genetic defect have few normal peroxisomes;

instead membrane structures referred to as peroxisomal ghosts form. These

contain some peroxisomal membrane proteins but lack normal peroxisomal

matrix contents (Santo et al., 1988; Arias et al., 1985). Apparently, the lesion in

this disease is a defect in peroxisomal biogenesis in which certain peroxisomal

matrix enzymes are mislocalized to the cytosol and rapidly degraded, instead of

transported into the peroxisomes, (Tager et al., 1986). Heikoop et al., (1992)

used double labeling immunoelectron microscopy on Zellweger fibroblasts to

show that most peroxisomal ghosts in fact contain lysosomal hydrolases.

Treatment with the autophagy inhibitor 3-methyl-adenine caused an increase in

the number of morphologically normal peroxisomes, suggesting that Zellweger

peroxisomes are selectively degraded by autophagy.

There has been extensive study of the fate of phenobarbital-induced

smooth endoplasmic reticulum (SER) and the resident detoxifying enzymes

cytochrome P450 and NADPH-cytochrome P450 reductase upon withdrawal of

the drug. Bolender and Weibel (1973) showed morphometrically that there was

a concurrent increase in the total volume (800%) and number (96%) in

autophagic vacuoles and a preferential removal of SER membranes. Masaki et

al. (1984) showed a decrease in the cellular content of SER enzymes and a

concurrent increase in the number autophagic vacuoles after phenobarbital

withdrawal. The same group directly showed, using quantitative immunoblots

and immunoelectron microscopy, that autophagic vacuoles isolated from liver

cells recovering from phenobarbital treatment contain large amounts of the SER

enzymes cytochrome P450 and NADPH-cytochrome P450 reductase (Masaki et

al., 1987).

These data show that in certain nutritional, pathological, and

pharmacological situations selective autophagic degradation can be induced in

higher eukaryotes. As discussed above, several instances of selective,

degradative "catabolite inactivation" have long been recognized as occurring in

S. cerevisiae (Holzer 1976; Holzer and Purwin, 1986). Of the enzymes known to

undergo catabolite inactivation, there is evidence for only one being degraded in

the vacuole (Chiang and Schekman, 1991) and even this is controversial

(Schork et al., 1994). Therefore, little is known about how a soluble protein

might enter the vacuole. It has been suggested that a certain protein(s) which

traverses the secretary pathway may serve as a receptor or channel to allow

soluble protein entry into the vacuole (Chiang and Schekman, 1991).

The methylotrophic yeasts have served as a model system both for

peroxisome biogenesis and degradation due to the ease of peroxisome induction

and detection in this system (see discussion above in "Methanol Metabolism

and Peroxisomes in Yeast"). Bormann and Sahm (1978) were the first to note

that methanol-induced cells of the methylotrophic yeast Candida boidinii,

containing high levels of the peroxisomal enzymes alcohol oxidase and catalase,

exhibited a marked decrease in enzyme activity and number of peroxisomes

when subsequently incubated in a new carbon source (ethanol). It was

observed that the degradation of peroxisomes stimulated by ethanol did not

require protein synthesis. Later experiments by Hill et al. (1985) confirmed the

fact that the loss of peroxisomal enzyme activity was due to degradation of the

protein and not by limited or reversible inactivation. Electron microscopic

examination of cells undergoing ethanol adaptation revealed that peroxisomal

enzyme degradation was due to degradation of the entire peroxisomes and not

to certain enzymes within the peroxisome. Together these data indicate that

methanol-induced peroxisomes in C. boidinii are degraded by autophagy upon

addition of ethanol to the media.

The fate of methanol-induced peroxisomes has been most carefully

studied in Hansenula polymorpha. Veenhuis et al. (1978) added glucose to a

final concentration of 0.4% to stationary methanol-grown cultures and

determined that alcohol oxidase and catalase activities decreased precipitously

over the ensuing 4 h to 22% and 34% of the values at 0 h, respectively.

Morphometric evaluation of transmission electron micrographs of cells before

and after 4 h of glucose adaptation revealed a decrease in the volume occupied

by peroxisomes of 92%. These authors concede that their data do not prove

that the peroxisomes and enzymes disappear due to degradation since they

were measuring enzyme activity and not protein and their system could not take

possible changes in peroxisome synthesis into account. Their morphometric

analyses did not reveal the cause of the disappearance of the peroxisomes and

diluting out among new daughter cells could not be ruled out.

The same laboratory (Bruinenberg et al., 1982) investigated the

phenomenon of glucose-mediated catabolite inactivation in H. polymorpha

further, using HPLC to show that indeed AOX and catalase protein are being

depleted at approximately the same rates as enzyme activity during adaptation

to glucose. Switching methanol-induced H. polymorpha cells to media lacking

any carbon source had minimal effects on AOX and catalase activity and protein

levels, suggesting that the loss of protein is directly mediated by glucose or its

metabolites. Again, synthesis effects were not evaluated so it could not be

determined whether peroxisome degradation was enhanced or whether it

remained the same and peroxisome synthesis had decreased. An interesting

phenomenon was noted when methanol-induced cells were switched to glucose

media which also contained 6 mM KCN; in this case, AOX activity fell very

rapidly (80% loss in 1 h) but AOX protein levels remained the same for at least 4

h. They were able to account for this phenomenon by noting that the time

course for activity loss corresponded almost exactly with decreases in AOX-

bound FAD. It appears that KCN mediates a loss of the FAD prosthetic group

which renders the enzyme largely inactive.

A detailed ultrastructural description of the events that occur in H.

polymorpha when methanol-induced cells are switched to glucose was prepared

by this group (Veenhuis et al., 1983). The earliest change that was noted was

the appearance of a variable number (2 to 12) of layers of electron-dense

membranes surrounding individual peroxisomes within a cluster which appeared

to sequester a given peroxisome from the cytosol into an autophagosome.

Other cytoplasmic components were never observed together with peroxisomes

in an autophagosome as determined by transmission electron microscopic

surveys of glutaraldehyde/Os04 or KMn04 fixed cells or isolated

autophagosomes. The authors suggested that the limiting membranes of the

autophagosomes arose de novo since they saw no evidence for enwrapping of

peroxisomes by existing membrane components. Notwithstanding this fact,

Veenhuis and colleagues had earlier noted that methanol-induced peroxisomes

where almost always seen in close juxtaposition to endoplasmic reticulum

cisternae (Veenhuis et al., 1978). Therefore, it can not be ruled out that in H.

polymorpha, as in higher eukaryotes (Dunn 1990a), organelles may be

sequestered from the cytosol by segments of endoplasmic reticulum.

Similarly to rat liver autophagosomes (Dunn 1990a,b), the early

autophagosomes of H. polymorpha cells undergoing catabolite inactivation

were devoid of vacuolar hydrolases and then acquired them as they matured

(Veenhuis et al., 1983). Hydrolases were observed to be procured by two

different mechanisms: 1) extensions of the limiting membranes of the

autophagosome protruded into a nearby vacuole, surrounding a portion of the

vacuole and incorporating that portion of the vacuole into the thereby newly

formed autolysosome; or 2) autophagosomes were seen to fuse with vacuoles

and deposit their contents (a peroxisome) into the vacuolar lumen where the

peroxisomes were degraded. The authors noted that the molecular mechanisms

behind these actions remain unknown.

One of the events that must occur if peroxisomes are to be selectively

degraded is that they must be recognized as destined for degradation. Evidence

to support this hypothesis is provided by van der Klei et al. (1991) who studied

the fate of AOX in H. polymorpha peroxisome assembly mutants (PER mutants).

One class of these mutants does not grow in methanol but develops very large

AOX crystalloids in the cytosol which are not bounded by peroxisomal

membranes. When glucose is added to methanolic cultures of this mutant class,

they begin to grow but the AOX crystalloids are not degraded or seen to be

taken up into the vacuole, and AOX activity and protein remain unchanged

during a 4 h time course. In the wild type H. polymorpha, as shown in the

previous studies already mentioned, AOX activity and protein and peroxisomes

were > 50% depleted in 4 h of glucose adaptation. This suggests that the

peroxisomal membrane mediates the degradation of AOX: Presumably

molecules in the peroxisomal membrane allow the peroxisomes to be recognized

by the membranes which form the limiting membranes of autophagosomes. In

contrast to glucose, ethanol caused a rapid decrease in AOX activity in PER

mutants which form cytosolic crystalloids, decreasing AOX activity levels by

-50% in 4 h (van der Klei et al., 1991). Immunoblot analysis of AOX during

ethanol adaptation of PER mutants revealed that AOX protein levels did not

diminish during this time, suggesting "modification inactivation" rather than

degradation. This result may be explained by a loss of AOX-bound FAD which

has previously been shown to inactivate the protein (Veenhuis et al., 1983). The

mechanism of inactivation was not determined in this study (van der Klei et al.,

1991) but cytochemical analysis revealed that AOX crystalloids which had lost

activity during ethanol treatment had not been taken up into the vacuole but,

rather, remained in the cytosol.

Autophaqy mutants. While examples of autophagy have been reported in

yeast, still very little is understood about the mechanisms of autophagic

sequestration of specific substrates, and the field has remained largely

descriptive. Only very recently has the power of yeast genetics been brought to

bear on the problem of autophagy, and important details regarding the molecular

mechanisms of autophagy should be forthcoming.

The first autophagy-defective mutants in yeast were produced in vacuolar

proteinase-deficient strains of S. cerevisiae (Tsukada and Ohsumi; 1993). It has

been noted that starvation for nitrogen, essential amino acids, or carbon

produced an increase in nonselective protein degradation mediated by the

vacuoles of normal yeast (Teichert et al., 1989). Correspondingly, starvation

caused an accumulation of autophagic bodies in the vacuoles of cells with

defective PrA and/or PrB, i.e., with proteolytically inactive vacuoles (Takeshige

et al., 1992; Baba et al., 1994). Tsukada and Ohsumi (1993) chemically mutated

a S. cerevisiae strain lacking both PrA and PrB and then looked for vacuolar

accumulation of autophagic bodies by light microscopy. Clones that were

unable to sequester cytoplasmic components into the vacuole were readily

identifiable in this way (no accumulation in the vacuole during starvation) and

were classified APG mutants. Further characterization of a clone designated

apgl-1 revealed that it was sensitive to nitrogen deprivation, i.e., it became

inviable after 2 d in nitrogen-free media as indicated by uptake of phloxine B

(red colonies) when plated onto agar plates containing this dye. Thereafter

nitrogen deprivation sensitivity was used as an initial selection to aid in the

screening for further mutants and 75 recessive apg mutants were assigned to 15

complementation groups. It was determined that the vacuoles of these mutants

were normal for functions other than autophagy, e.g., they were acidic and they

were able to accumulate endocytosed substances (Tsukada and Ohsumi, 1993).

Therefore, it was concluded that the defects being studied in these yeast were

not caused by a generalized defect within the vacuole itself.

Egner et al. (1993) showed that when one of the two subunits of the

cytosolic protein fatty acid synthase is overexpressed in starved protease mutant

S. cerevisiae, then the excess, unassembled subunits accumulate in the

vacuole, along with other cytosolic proteins. In contrast, in fed cells, the

unassembled subunits are degraded in the cytosol by the proteasome. Since

the vacuole is the site of nonselective degradation during starvation, the

diversion of unassembled subunits from their normal site of degradation to the

vacuole was taken as evidence that they are taken up into the vacuole by a

nonselective bulk process. These same investigators took advantage of this

nonselective accumulation as a marker for autophagy to screen for autophagy

mutants (AUT) using a colony screening procedure (Thumm et al., 1994).

Chemically mutagenizing PrA/PrB deficient strains which also lacked a

functional gene for one of the two fatty acid synthase subunits (e.g., a) and then

using a direct colony immunoassay for the other fatty acid synthase subunit

(e.g., 3), they screened for clones which were not able to accumulate the

appropriate fatty acid synthase subunit in the vacuole. This rather time-

consuming process (5 to 9 d for screening) has so far yielded 3

complementation groups.

Analysis of the proteins encoded by the 18 genes found to affect

autophagy which have been isolated to date should yield the first information on

the molecular mechanisms of nonselective autophagy. Autophagy has only

recently been acknowledged to occur in yeast. In that short time, more progress

has been made than in the several decades that has been expended in the

study of mammalian autophagy. Now that the advantages of yeast cell biology

and genetics are being utilized in regards to autophagy, our new knowledge of

the mechanisms of autophagy should enable steady progress both in yeast and

in higher eukaryotes, if homologous genes are found to act in these diverse

organisms, as has proven to be the case for secretion, nuclear transport, protein

sorting and other areas of basic cell function.

Genetic Analysis in Methylotrophic Yeasts

As must be obvious from the foregoing discussions, Saccharomyces

cerevisiae is by far the best studied of the yeasts both in terms of cell and

molecular biology; there are hundreds of mutant strains available to aid one's

studies and straightforward strategies have been worked out for the molecular

genetic manipulation of the S. cerevisiae genome. In most cases this renders S.

cerevisiae the clear choice for mutagenesis and gene cloning strategies. One

case contrary to this rule is that of the study of selective autophagy, a

phenomenon not known to exist in S. cerevisiae but very easy to manipulate in

the methylotrophic yeasts P. pastors and H. polymorpha. Recent developments

have made these yeast more amenable to cell biological and genetic studies.

Classical Genetics

Two of the genera of methylotrophic yeasts, Pichia and Hansenula are

ascomycetous: their life cycles are characterized by genetically defined stages,

i.e., haploid and diploid (Gleeson et al., 1984; Cregg, 1987; Gleeson and

Sudberry, 1988; Gould et al., 1992). The life cycles of P. pastors and H.

polymorpha are very similar to that of S. cerevisiae in that they can be

maintained indefinitely in the haploid vegetative state and by changing the

composition of their growth media they can be forced to mate and become

diploid and remain vegetative in this state if desired. Changing the media again

causes the diploids to undergo meiosis and haploid spores result which can

germinate in favorable media and resume haploid vegetative growth. Also like

S. cerevisiae, these yeasts exist in one of two mating types. Mating type

switching occurs under poor nutritional conditions so these yeasts are

homothallic, and mating can be caused to occur between any two strains.

These features just mentioned allow the routine use of backcrossing and

complementation analysis. These invaluable techniques are a large part of the

reason that species that can be maintained in the haploid state are such

powerful systems for the identification of new genes. A prolonged haploid state

makes the consequences of recessive mutations easily detectable.

Backcrossing makes it possible for mutant strains to be refined so that they carry

only one mutant gene relevant to the phenotype of interest, and

complementation analysis allows the phenotypic identification of

complementation groups or genes.

Molecular Biology

Development of the methylotrophic yeasts P. pastors and H. polymorpha

as host systems for DNA transformations has been taken place over the past

decade (Cregg et al., 1985; Tikhomirova et al., 1986, 1988; Roggenkamp et al.,

1986; Gleeson et al., 1986). As it turns out P. pastors is more amenable to

transformation, yielding transformation rates of -104 transformants/ pg DNA,

similar to that obtainable in S. cerevisiae but 1 to 2 orders of magnitude higher

than has been obtained in H. polymorpha (J.M. Cregg, personal

communication). Due to the fact that the AOX promoter of P. pastors is tightly

regulated, i.e., tremendously induced in the presence of methanol and is almost

completely turned off in its absence, this yeast has been developed as a host

system for the production of foreign proteins (Cregg and Madden, 1988;

Wegner, 1990; Sreekrishna et al., 1988), e.g., the secreted isoform of the

Alzheimer's amyloid beta-protein precursor (Wagner et al., 1992), Bordetella

pertussis pertactin (Romanos et al., 1991), mouse epidermal growth factor

(Clare et al., 1991), bovine lysozyme (Brierly et al. 1990), and human tumor

necrosis factor (Sreekrishna et al., 1989). Reports of as much as 3 g of

protein/liter of culture were caused to be secreted into the media.

A number of genes have been cloned and sequenced in P. pastors,

particularly ones whose products are active in the peroxisome biogenesis

pathway (Koutz et al. 1989; Ellis et al., 1985; Spong and Subramani, 1993;

Crane et al., 1994; McCollum et al., 1993; Gould et al., 1992). To accomplish

this, certain tools or reagents have to be available and have been developed in

at least two separate labs: a set of auxotrophic strains with essentially wild type

genetic background; plasmids that act as E coli-P. pastors shuttle vectors,

(those developed have been based on the plasmid pBR322) containing the

ampicillin resistance gene, an E. coli origin of replication, P. pastors

autonomous replication sequences and an auxotrophic selectable marker, e.g.,

the histidine dehydrogenase gene from either S. cerevisiae or P. pastors; and

genomic DNA libraries for the isolation of P. pastors genes by functional

complementation of mutants or by nucleic acid hybridization (Gould et al., 1992;

Cregg et al., 1985).


With these tools the molecular cell biology of P. pastoris can be probed

and the molecular mechanisms of processes for which methylotrophic yeasts are

good models, e.g., peroxisome biogenesis and degradation, are beginning to be



Yeast Strains and Media

The essentially wild type methylotrophic yeasts used in these studies

were a leucine auxotrophic leucinee) strain of Hansenula polymorpha (A16, a

backcross derivative of strain L1 (Gleeson et al., 1986); a histidine auxotroph

(histidine') of Pichia pastors (GS115) and an arginine auxotroph of P. pastors

(GS190-3). These strains were all the very generous gifts of J.M. Cregg. The

strains of P. pastoris lacking proteinase A (PrA) and/or proteinase B (PrB') are

all histidine auxotrophs: SMD1163 (PrA, PrB'); SMD1165 ( PrB-); and SMD1168

(PrA). These strains were kindly donated by L.V. Benningfield of the Phillips

Petroleum Co. Licensing Office as part of the Pichia Yeast Expression System.

TABLE 2-1. Yeast strains

A16 (H. polymorpha) leu2' Leucine
GS115 (P. pastors) his4 Histidine'
GS190-3 (P. pastors) arg4 Arginine
SMD1163 (P. pastors) his4, pep4, prb1 Histidine, PrA, PrB"
SMD1165 (P. pastors) his4, prb1 Histidine', PrB'
SMD1168 (P. pastors) his4, pep4 Histidine', PrA'
WDY1 (P. pastors his4, gsal-1 Histidine-, GSA1
WDY2 (P. pastors) his4, gsa2-1 Histidine, GSA2
WDY3' (P. pastors) his4, gsa2-2 Histidine', GSA2"
See text for details of nomenclature
b Strains developed during the course of this study

The media utilized for the growth of yeast in this study consist of the following:

MIM (methanol induction medium): 6.7 g/L yeast nitrogen base without amino
acids (Difco), 0.5% methanol, 40 mg/L histidine or arginine, and 40 pg/L

GA (glucose adaptation medium): 6.7 g/L yeast nitrogen base without amino
acids, 2.0% glucose, 40 mg/L histidine or arginine, and 40 pg/L biotin.

EA (ethanol adaptation medium): 6.7 g/L yeast nitrogen base without amino
acids, 0.5% ethanol, 40 mg/L histidine or arginine, and 40 Ig/L biotin

YPD: 1% yeast extract, 2% bactopeptone (both from Difco), 2% dextrose

SM sporulationn medium): 0.5% sodium acetate, 1.0% KCI, 1.0% dextrose

DSM diploidd selection medium): 6.7 g/L yeast nitrogen base without amino
acids, 2.0% glucose, and 40 pg/L biotin.

Transformation plates: 6.7 g/L yeast nitrogen base without amino acids, 2.0%
glucose, 1 M sorbitol, and 40 gg/L biotin.

LSM (low sulfate medium): autoclave stock solution of 10x salts: 6 g/L
MgCl2.6H20, 1.23 g/L K2HPO4, 8.77 g/L KH2PO4, 5 g/L NaCI, 10 g/L
NH4CI; autoclave stock solution of 1 M CaCI2; autoclave stock solution of
100 mg/L FeCI3-6H20; to prepare working solution of LSM mix 0.1 g yeast
extract in 10 mL 10x salts and bring to 100 mL with water, pH to 5.8 and
autoclave; after cooling add 0.9 mL CaCI2 and 0.5 mL FeCi3; supplement
with amino acids and carbon source as needed.

Enzyme Assays

Preparation of cell-free extracts of yeast was accomplished by removing

2mL samples of cultures to culture tubes on ice at appropriate times and

harvesting cells by centrifugation at 1000 x g, 40C for 1 min and aspirating the

supernatant; 1 mL of a buffer containing 20 mM Tris/CI, pH 7.5, 50 mM NaCI, 1

mM EDTA, was added to the cell pellet and phenylmethylsulfonyl fluoride to a

final concentration of 1 mM was added while mixing just prior to preparation of

extract. Approximately 0.5 mL of 0.5 mm diameter acid-washed glass beads

were added and then the tubes were vortexed 2 x 1min, placing tubes on ice 2 1

min between each occasion of vortexing. The tubes were then centrifuged 1000

x g, at 40C for 6 min, and the supernatants carefully aspirated and placed in 1.5

mL microfuge tubes on ice.

Formate dehydrogenase (FDH) was assayed utilizing sodium format as

a substrate and following the reduction of NAD+ at 340nm according to the

procedures of Kato (1990).

Measurement of alcohol oxidase (AOX) activity was performed with

methanol as a substrate, producing hydrogen peroxide, in turn metabolized by

horseradish peroxidase (HRP) to oxidize 2,2'-azino-bis(3-ethylbenz-thiazoline-6-

sulfonic acid) (ABTS). The generation of oxidized ABTS was followed

spectrophotometrically at 410 nm (Sahm and Wagner, 1973).

Proteinase A (PrA) assays were performed according to Jones (1991b)

and direct colony assays for carboxypeptidase Y (CPY) activity were done

according to Jones (1977).

Preparation of Antisera

Formate dehydrogenase purified from Candida boidinii (100 Ig; Sigma

Chemical Co.) was emulsified in Freund's adjuvant and injected intradermally

into a rabbit. A boost injection containing 50 pg FDH emulsified in Freund's

adjuvant was administered intradermally. The immunological properties of FDH

from C. boidiniiand P. pastoris have been shown to be quite similar, even though

the subunit size is somewhat larger in P. pastors (Hou et al., 1982).

Immunological analysis was performed on cell-free extracts which were

electrophoresed on 7.5% SDS-polyacrylamide gels and then transferred to

nitrocellulose. Immunoblotting was performed according to the ECL method of

Amersham (Arlington Heights, IL). On immunoblots, a single band of Mr 45x103

was recognized in methanol-induced cellular extracts of P. pastors (Tuttle and

Dunn, in press). This band is in close agreement with that reported for FDH

subunits of P. pastors (Hou et al., 1982). Preabsorption of the antiserum with

purified FDH greatly diminished the signal on immunoblots. Preabsorption with

the same amount of an unrelated protein (AOX) had little effect on

immunostaining (data not shown).

Purified AOX from P. pastors (104 pg; Sigma Chemical) was emulsified in

Freund's adjuvant and injected into a lymph node of a rabbit. Boost injections

were administered subcutaneously at 6 wk intervals that included 52 pg of

antigen in Freund's adjuvant. On Western blots, the antiserum recognized a

single protein of 74 kDa in total cellular extracts of both P. pastors and H.

polymorpha only when grown under peroxisome-induction conditions (Tuttle and

Dunn, in press). Immunoelectron microscopic analysis of methanol-induced

yeast displayed strong, specific localization limited to the crystalloid core of giant

peroxisomes (see Fig. 2-1).

Figure 2-1. Immunoelectron microscopic localization of alcohol oxidase in
methanol-induced P. pastors. Wild type P. pastors cells (GS115) were grown
to stationary phase then fixed, embedded, and subjected to immunoelectron
microscopic analysis utilizing alcohol oxidase antisera. AOX was localized with
10 nm gold conjugated to goat anti rabbit antibodies. p = peroxisomes m =
mitochondrion; n = nucleus; bar = 0.5 pm.

Electron microscopy

Ultrastructural analysis was performed on potassium permanganate-fixed

cells in which membrane profiles are effectively delineated in methylotrophic

yeasts (Veenhuis et al., 1983). Cells were harvested from a 1 mL of culture

sample by centrifugation for 1 min at 15,000 x g at room temperature, washed in

1 mL water, and fixed in 1 mL 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 dehydrated in increasing concentrations of ethanol, ending

in absolute, followed by 100% propylene oxide then infiltrated with a 50:50

mixture of propylene oxide and (Polysciences, Inc., Warrington, PA) for 2 d. The

samples were pelleted at 15,000 x g and then resuspended in fresh POLY/BED

812 and placed under vacuum overnight then polymerized in a 60C oven for 24

h. The blocks were sectioned on a Reichert UltraCut E microtome (Cambridge

Inst.; Deerfield, IL) by Denny Player and examined on a JEOL 100CX II

transmission electron microscope.

Immunoelectron microscopic localization of alcohol oxidase was

performed on glutaraldehyde-fixed, POLY/BED 812-embedded cells as

described by Clark (1991). To block non-specific binding of antibodies the

sections were incubated in 8% bovine serum albumin (BSA) in TBS then

incubated in 1:100 rabbit anti-alcohol oxidase serum in TBS + 0.1% BSA for 1 h

at room temperature in a moistened chamber. After washing in 0.1% BSA in

TBS, grids were incubated in a 1:30 dilution of protein A conjugated to 10 nm

colloidal gold (Amersham, UK) for 1 h to localize antigen/antibody complexes

followed by washing in TBS.

Protein Synthesis and Degradation

Synthesis of Peroxisomal Proteins

Single colonies from YPD plates were pre-cultured in low sulfate medium

(Daum et al., 1982) containing 30 mg/L leucine or 20 mg/L histidine, and 2%

glucose. Pre-cultures were inoculated into low sulfate medium with 0.5%

methanol and then incubated until the desired growth phase had been attained.

Methanol-induced exponential and stationary phase cells and stationary phase

cells to which glucose had been added were labeled for 15 min with

["S]methionine/cysteine (Tran5S-label, ICN Biomedicals, Irvine, CA, USA).

Cell-free extracts were made by vigorous vortexing of harvested cells in the

presence of glass beads. Total cellular proteins were separated by 7.5%

sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE;

Laemmli, 1970), stained with 1.0% Coomassie blue, and the bands

corresponding to the methanol-induced peroxisomal enzymes dihydroxyacetone

synthase (DHAS) and alcohol oxidase (AOX) were excised from the gels (see

Fig. 3-1). The gel pieces were solubilized in a scintillation cocktail composed of

0.54% PPO, 0.0135% dimethyl POPOP in a solution of 90% toluene and 10%

TS-1 Solubilizer (Research Products International, Inc., Mt. Prospect, IL, USA)

and radioactivity quantified using a Beckman LS5000TD scintillation counter. In

some cases, the labeled proteins were electrophoretically transferred to

nitrocellulose and their relative amounts quantified from an autoradiograph by

laser densitometry.

Degradation of Peroxisomal and Mitochondrial Proteins

Single colonies from YPD plates were precultured in low sulfate medium

containing 2% glucose for 24 h. The cells were then inoculated 1:33 into low

sulfate medium with 0.5% methanol and 20 pCi/mL ["S]methionine/cysteine and

incubated until stationary growth phase was reached. At this time, methanol-

induced exponential and stationary cultures and glucose adapting cultures were

chased with unlabeled methionine and cysteine at final concentrations of 0.1 M

each. At various times of chase, cell extracts were prepared and the

radioactivity present in DHAS and AOX determined as described above. The

data ([3S] cpm/g cellular protein) was expressed relative to non-glucose

adapting stationary cultures.

The B-subunit of mitochondrial F1 ATPase (FilB) was immunoprecipitated

from samples of the extracts of glucose repressed cells using modifications of

the methods of Aris and Blobel (1989). Aliquots of Staphylococcus aureus (10-

15% suspension of heat-killed, glutaraldehyde fixed cells; SA) were washed 3

times in 2 volumes of immunoprecipitation buffer (20 mM Tris/CI, pH 7.8, 150

mM NaCI, 2 mM EDTA, 0.02% NaN3; IP) + 0.5%BSA + 1%Tx100 + 0.2%SDS,

then resuspended in 1 volume of the same buffer. Cell-free extracts were

transferred to a tubes containing one tenth volume of washed and blocked SA

then incubated 1-2 h, 4C to pre-clear the extracts then harvested by

centrifugation for 10 min in a microfuge at full speed. Supernatants were

transferred to fresh tubes and diluted 10-fold in IP + 0.5%BSA + 1% TX100 +

0.2% SDS to solubilize the proteins. Polyclonal antisera raised against Flp

(Lewin and Norman, 1983) were added in amounts sufficient to bind to all

available antigen and incubated overnight at 4C. Washed, blocked SA was

added at a proportion of 5 volumes of SA suspension/volume of antiserum used,

incubated 1-2 h, 4*C, and the SA was pelleted in a microfuge at full speed for 3

min and the supernatant discarded. Precipitated immune complexes were

washed in IP + 1% TX100 + 0.2% SDS and in IP alone. The antigens were

eluted by adding buffer containing 10 mM Tris/CI, pH 8.0 + 1 mM EDTA + 2%

SDS + 1% B1 mercaptoethanol + 20% sucrose to the samples, using the same

volume of buffer as the original volume of SA suspension used. Samples were

heated to 95*C for 2 min to aid dissociation of immune complexes. SA was

removed by centrifugation (5 min) and the supernatant saved for polyacrylamide

gel analysis. The immunoprecipitates were electrophoresed on 7.5%

polyacrylamide gels and transferred to nitrocellulose or dried and quantitated by

phosphorimager analysis (Molecular Dynamics Phosphorimager).

Measurement of Overall Protein Degradation

Histidine auxotrophic yeast strains were precultured in YPD then

inoculated 1:30 into medium containing 6.7 g/L yeast nitrogen base without

amino acids (Difco), 2% glucose, 40mglL histidine, 40 ig/L biotin, and 1 pCi/mL

["C]valine (Amersham). Cultures were incubated ~18h then washed twice in 6.7

g/L yeast nitrogen base without amino acids. Individual cultures were then

divided into two equal portions and incubated with 10mM cold valine chase in

the absence or presence of histidine (40mg/ml) in yeast nitrogen base without

amino acids, 2% glucose and 40 pg/L biotin. Aliquots were collected on ice at 0,

5, and 8 hours of chase with the immediate addition of ice cold trichloroacetic

acid to a final concentration of 20% w/v. The samples were incubated on ice for

at least 1h then centrifuged. The supernatant was aspirated and kept separate

from the pellet which was solubilized in 0.5mL Scintigest (Fisher Scientific Co.)

The radioactivity in the supernatant (acid-soluble counts) and solubilized pellet

(acid-insoluble counts) was counted on a Beckman LS5000TD scintillation

counter. Percentage degradation was calculated as the ratio of TCA-soluble

counts at each time point to acid-insoluble (protein-associated) counts at 0 hour

of chase multiplied by 100 and normalized to one hour of chase.

Isolation of Glucose-Induced Selective Autophavq-Deficient (.sa) Mutants

Mutagenesis and Screening

Pichia pastors cells (GS115) were mutagenized according to the

methods of Cregg (1990). Briefly, single colonies were grown in YPD medium

overnight to an Awo of 0.5 to 1.0 and then washed twice with 0.1M sodium

citrate, pH 5.5. Cells were resuspended in 40 mL sodium citrate plus 100 pg/mL

N-methyl-N'-nitro-N-nitrosoguanidine and incubated at room temperature

without shaking for 1 hour followed by detoxification by mixing with an equal

volume of 5% sodium thiosulfate. Cells were then washed three times with

sterile water and inoculated into 140 mL YPD cultures and incubated for 2 to 6

hours at 300C. The mutagenized yeast were harvested and concentrated in

fresh YPD to 10 Awo units and sterile glycerol was added to a final concentration

of 30%, aliquoted, and allowed to equilibrate at room temperature for 2 hours

and then frozen at -800C for later use. Cells were viable for screening for > 1

year of storage.

Mutagenized cells were screened for the loss of the ability to degrade

alcohol oxidase in response to a shift in carbon and energy source from

methanol to glucose. The screening protocols are modifications of the direct

colony assay procedures of Gleeson et al. (1984) and Tomlinson and Esser

(1992). Aliquots of mutagenized yeast were thawed and diluted in water and

sonicated for 30 s to break up cell clumps without reducing their viability. One

hundred pL of the diluted cells were spread on MIM plates and incubated until

colonies formed (5 to 6 days). These master plates were replicated

to GA plates by placing sterile 85 mm nitrocellulose circles directly onto the

master plates until they were soaked through and then placing onto GA plates.

These plates were incubated at 30C for 12 to 14 h during which time

peroxisomes would be repressed in cells with a normal selective autophagy

pathway. It was then necessary to spheroplast the cells to allow access to

cellular contents for detection of peroxisomal AOX. To accomplish this, the

colony circles were treated as follows: the nitrocellulose circles are placed on

paper filters soaked with 20 mM dithiothreitol in 67 mM potassium phosphate

buffer, pH 7.5, containing 20 mM EDTA and incubated for 5 min at room

temperature. The nitrocellulose circles were then moved to fresh paper filters

soaked in 0.25 mg/mL Zymolyase 20T in 67 mM potassium phosphate buffer, pH

7.5 and incubated at 30C for 1 hour. For colorimetric assay, the nitrocellulose

circles were placed on fresh filters soaked in 0.13% methanol + 3.4 U/mL HRP +

0.56 mg/mL ABTS in 33 mM potassium phosphate, pH 7.5 and incubated at

room temperature until color develops. The originals of the positive replica

colonies were picked from the master plates and grown in YPD and mutant

status was verified as follows: putative mutants were grown in MIM to stationary

phase and the cultures were sampled at 0 h and 6 hours after addition of solid

glucose to 2% final concentration and assaying for AOX and FDH activity.

Measurement of the decrease in methanol-induced enzyme activity after glucose

treatment allowed quantitation of the defect. The amount of activity present at

the end of 6 h glucose treatment was compared to enzyme activity in the

parental strain (see Fig 4-2).

Genetic Analysis

Backcrossing. In order to obtain mutant strains in which the defective

gene of interest is the only mutated gene different from that found in the parental

strain, the haploid mutants were sequentially mated to essentially wild type

strains (GS115 or GS190-3) containing complementary auxotrophic markers and

haploid progeny recovered. Several rounds of this backcrossingg" procedure

result in strains in which the only difference from the parental strain is the mutant

gene of interest. A single round of backcrossing was accomplished as follows:

the mutant strains were streaked on YPD plates in 2 cm X 2 cm patches and the

parental strain of the complementary auxotrophy (i.e., His" x Arg-) was spread on

YPD plates to form a lawn at a density of 1x107 cells/plate and both were

incubated at 300C overnight. The lawn and the patch plates were both replica

plated to a single SM plate and incubated overnight at 300C to induce mating.

These plates were then replicated to DSM plates (lacking amino acids) and

incubated for 2 to 3 d, until diploid (prototrophic) colonies appeared. Colonies

from these plates were streaked onto fresh DSM plates and incubated 1 to 2 d

and then streaked to YPD plates overnight. Cells from this plate were streaked

to SM plates and incubated 4 d at 30C to induce meiosis and sporulation at

which time some of these cells were harvested with an inoculation loop and

etherized to kill any remaining diploid cells. Etherization consists of putting a

loopful of cells from the SM plate into 1 mL sterile water, adding 1 mL

diethylether, vortexing and incubating at room temperature for 20 min. An

aliquot from this mixture was diluted 1:100 and 100 plL spread on YPD plates

which were incubated for 2 d so that the spores germinate to grow vegetatively.

Sterile water was placed on these plates and swirled to loosen some of the cells

and collected; this cell suspension was diluted and plated on MIM to yield 100 to

500 colonies/plate. The resulting colonies were screened by direct colony assay

and then verified for the gsa phenotype as described above. Only gsa mutants

of the opposite auxotrophy were selected, i.e., if the mutant strain was His'

before backcrossing, only Arg mutants were selected to ensure that the original

mutant had not come through the procedure without backcrossing.

Complementation analysis. To determine the number of different mutant

genes which were present in my collection of gsa mutants, the ability of diploids

produced by mating different mutant strains to degrade peroxisomes was

assessed. If the diploids produced by mating two mutant strains is normal then

the strains are said to complement, indicating that the genes mutated in the two

strains are different. Accordingly, if they do not complement, the two strains

contain presumably different mutant alleles of the same gene. To accomplish

this, a His' backcross of one mutant strain was mated to an Arg mutant strain as

described above, diploids were recovered from DSM plates and inoculated into

YPD media and cultured until stationary. The ability to degrade peroxisomes

was assessed as described above.



Certain species of yeast in several genera can grow in media containing

methanol as the sole carbon and energy source, e.g., Candida boidinii,

Hansenula polymorpha, and Pichia pastoris. At the time that H. polymorpha and

P. pastors are fully methanol-induced, i.e., grown up in methanolic media until

the carbon source is depleted and thereby constituting a non-growing or

stationary culture, the peroxisomal enzymes alcohol oxidase (AOX) and

dihydroxyacetone synthase (DHAS) comprise >50% of the cells' total protein

mass. These two proteins are important mediators of methanol utilization (see

Fig. 1-1) and are virtually undetectable in cells grown with glucose as the sole

carbon source (see Fig 3-1; Gleeson and Sudberry, 1988; Gould et al., 1992;

Veenhuis et al., 1983). This being the case, it seems reasonable that as a cell

adapts to new carbon sources it would be advantageous for the cell to rid itself

of these proteins and to reutilize their component amino acids. In this study, I

have chosen the ascomycetous methylotrophic yeasts H. polymorpha and P.

pastors to investigate the fate of enzymes necessary for the utilization of

methanol when the carbon and energy source is changed. One of the initial


Oh lh 8h

Figure 3-1. Comparison of the protein profiles of methanol-induced P. oastoris
and cells undergoing glucose adaptation. Cultures of P. pastors were grown to
stationary phase in methanol then adapted to glucose for 8 h. Samples of equal
volume were removed from cultures at 0, 1, and 8 h after the addition of glucose.
Cell extracts were prepared, solubilized in SDS, and component proteins
separated by SDS-PAGE and identified by Coomassie staining. Alcohol oxidase
(AOX) at 74 kDa and dihydroxyacetone synthase (DHAS) at 78 kDa are
identified (arrows). Five unknown proteins, the levels of which appear to remain
relatively constant during glucose treatment, are identified by closed circles.

aims of the present study was to determine if the peroxisomal proteins were in

fact degraded and by what mechanism. Autophagy has been most extensively

described in mammalian cells, and in particular, rat liver cells (Ahlberg et al.,

1985; Dunn 1990a,b; Kopitz et al., 1990; Mortimore et al., 1988).

Experimentation conducted in these cells has on the whole shown autophagy,

especially that induced by nutritional stress, to be a nonspecific process, i.e.,

soluble and organellar cytoplasmic components are taken up into

autophagosomes or lysosomes without regard for their identities or their

senescence. Similarly, in S. cerevisiae cells in which autophagy has been

induced by nutritional stress, uptake of cytoplasmic components into the vacuole

appears to be nonselective based on the fact that the contents of the vacuole in

vacuolar protease mutant cells was indistinguishable from the cytoplasm itself

(Takeshige et al., 1992; Baba et al., 1994). These results are not surprising

considering that starvation-induced autophagy produces substrates for synthesis

of replacement housekeeping proteins required to allow the cells to remain

viable during poor nutritional conditions.

In the case of adaptation to glucose by cells previously adapted to

methanol, in which the cells are undergoing major, rapid changes in the cellular

content of specific enzymes in order to adapt to a change in carbon and energy

source, the autophagic response might be expected to be more selective. In this

case it may be hypothesized that a more limited subset of cellular proteins is

being degraded and a different set of proteins synthesized. Rather than

degrade cytoplasmic proteins randomly, which in this event would seem to be

not necessary and therefore energetically wasteful, specific proteins and

organelles required to utilize methanol may be replaced by other specific

proteins required to utilize glucose.

While several authors have investigated the disappearance of methanol-

induced peroxisomes in methylotrophic yeasts (Bormann and Sahm, 1978;

Bruinenberg et al., 1982; Hill et al., 1985; Veenhuis et al., 1978, 1983), previous

studies have not directly measured protein degradation. Consequently, it has

not been determined whether the decrease in morphologically identifiable

peroxisomes and associated enzymes observed in response to the nutrient

adaptation effect is due to selective or enhanced degradation or, rather, to a

decline in enzyme/peroxisome synthesis in conjunction with unaltered,

constitutive degradation.

The aims of this part of my study were as follows: 1) to determine whether

the degradation of methanol-induced peroxisomes is brought about by

autophagy; 2) to investigate the means by which methanol-induced peroxisomes

are sequestered and the cellular site of degradation; and 3) to determine

whether the degradation of these components is selectively enhanced during

adaptation to glucose.

Loss of Methanol-Induced Enzymes During Glucose Adaptation

Degradation of Methanol Assimilating Enzymes

Several investigators have measured the loss of enzyme activity and

protein in methanol-induced yeasts in various stages of growth (Bormann and

Sahm, 1978; Bruinenberg et al., 1982; Hill et al., 1985; Veenhuis et al.,

1978,1983). The present work is the first to measure synthesis and degradation

in cells rapidly growing in methanol or grown to stationary phase in methanol.

Degradation cannot be assumed to be equal to the loss in enzyme activity or

protein because the cellular content of a given protein is the combined result of

degradation and synthesis. By metabolically labeling cells during growth in

methanol and adapting to glucose in chase media, I was able to unequivocally

measure methanol-induced enzyme degradation. In addition, I estimated

enzyme synthesis by quantifying incorporation of "S-labeled amino acids into

peroxisomal enzyme proteins during logarithmic and stationary growth in

methanol media.

The peroxisomal enzymes alcohol oxidase (AOX) and dihydroxyacetone

synthase are very prevalent in methanol-induced H. polymorpha and P.

pastors such that the bands containing these proteins can be easily recognized

by Coomassie staining gels in which cell-free extracts have been

electrophoresed ( see Fig 3-1; Douma et al., 1985). To measure degradation of

peroxisomal proteins directly P. pastors and H. polymorpha cells were

metabolically labeled with ["S]methionine/cysteine during the entire time cells

were growing in methanolic media containing low sulfate to stationary phase. At

this point the cells were harvested and resuspended in media containing excess,

unlabeled methionine and cysteine (to chase the radiolabeled amino acids) and

glucose or methanol. Samples were taken at 0 and 3 hours of chase, cell-free

extracts prepared and separated on 7.5% SDS-polyacrylamide gels. The bands

corresponding to DHAS and AOX (see Fig. 3-1) were excised from the gels,

solubilized and the radioactivity quantified in a scintillation counter. In some

cases the gels were transferred to nitrocellulose blots to which film was exposed.

In these cases the bands corresponding to AOX and DHAS on the resulting

autoradiographs were quantified by laser densitometry. These data yielded

essentially the same results as those in which bands were excised from the gels

and counted. The results from both of these procedures were combined and are

shown in Fig. 3-2. It can be observed that the peroxisomal proteins are stable in

methanol but rapidly degraded in the presence of glucose. This indicates that

the degradation of peroxisomal proteins is enhanced when the carbon and

energy source is changed to glucose in both methylotrophic yeasts examined.

I next examined whether, since AOX and DHAS are stable in stationary

cultures in methanolic media, synthesis of these proteins is low. In order to test

this hypothesis the incorporation of radiolabeled methionine and cysteine into

peroxisomal proteins was evaluated. H. polymorpha and P. pastors cells were

grown in low sulfate media containing methanol either to log or stationary growth

phase. At this time cultures were either harvested and resuspended in glucose-

containing media (stationary cultures) or maintained in methanolic media

120 -
120 Chased In Methanol Medium

100 E:] Chasd In Gluose Midhm,,

0 M 'R
M-C 0
0 60

.2 40

0 20


t.E 3) 80
E S 60
0) >I 40
3 20

H. polymorpha P. pastors

Figure 3-2. Turnover of peroxisomal proteins in methanol induction and glucose
adaptation media. Cultures of H. polymorpha and P. pastors were grown to
stationary phase in low sulfur media with methanol as the sole carbon source, in
the presence of ["S]cysteine/methionine. At 0 h, unlabeled cysteine and
methionine were added to chase the label in the absence (solid bars) or
presence (hatched bars) of 2.0% glucose. Replicate samples were removed at 0
and 3 h of chase, cell-free extracts prepared and fractionated by SDS-PAGE.
Radioactivity corresponding to dihydroxyacetone synthase and alcohol oxidase
was quantified (see Chapter 2) and the values shown are the mean + s.e.m. for
3 or 4 determinations in two separate experiments.

(stationary and exponentially growing cultures). [jS]methionine/cysteine was

added to replicates of stationary cultures for 15 minutes every 2 hours during the

first 8 hours after glucose addition. Exponential cultures were pulsed for 15

minutes. After the pulses cultures were harvested and cell-free extracts

prepared and separated by SDS-PAGE and the bands corresponding to AOX

and DHAS quantitated as above. These data show that peroxisomal enzyme

synthesis is lowered by 95% during stationary phase relative to that during

exponential growth (Fig. 3-3). Moreover, synthesis of these proteins is not

significantly decreased after the addition of glucose.

In summary, these results verify that methanol-induced peroxisomal

enzyme degradation is enhanced during adaptation to glucose. Furthermore,

since peroxisomal enzyme synthesis is very low during stationary phase, any

decrease in enzyme protein must be due to degradation under these growth


Further confirmation of the actual loss of proteins that function in the

methanol assimilation pathway during glucose adaptation has been derived by

use of immunoblots employing polyclonal antisera specific for the methanol-

induced enzymes AOX (peroxisomal) and format dehydrogenase (FDH;

cytosolic). Cell-free extracts were prepared from hourly samples of P. pastors

cultures during 6 hours of glucose adaptation and were either assayed for AOX

and FDH activity or the proteins separated on 7.5% SDS polyacrylamide gels

prior to transfer to nitrocellulose. Immunoblot analysis was conducted according


oB p H.polymorpha

0 0 P.pastorls
g 0

0 D


L x

0 -

Stationary Phase Glucose Adaptation

Figure 3-3. Synthesis of alcohol oxidase by H. polvmorpha and P. pastors
during stationary phase in methanol-induced cultures and in glucose adapting
cultures relative to synthesis during exponential growth (log phase). H.
polymorpha and P. pastors cultures were grown to exponential or stationary
phase in low sulfur media containing methanol as the sole carbon and energy
source. The incorporation of ["S]cysteine/ methionine into alcohol oxidase
during exponential phase, and at various points during the first 8 hours of
stationary phase and glucose adaptation was quantitated as described in
Chapter 2. The results represent the mean + s.e.m. of four determinations from
two separate experiments and are expressed as a percentage of the
incorporation of radioactivity into alcohol oxidase in methanol-induced
exponentially growing cultures.

to the ECL protocol of Amersham utilizing polyclonal antibodies monospecific for

either FDH or AOX at concentrations of 1:10,000.

The effects of glucose adaptation are presented in Fig 3-4 and clearly

demonstrate the glucose-induced decrease in enzyme activity and protein in the

case of both peroxisomal (AOX) and cytosolic (FDH) methanol catabolism


Loss of Enzyme Activity

In order to rapidly measure degradation of methanol-induced enzymes

during adaptation glucose, enzyme assays were conducted on samples taken

from cultures adapting to these carbon sources at various time points. Cases in

which measuring enzyme activity during adaptation is a reliable method for

assessing enzyme protein degradation are those in which both the enzyme is

being synthesized at a very low rate or not at all and in which inactivation is

brought about by degradation. This is the circumstance for peroxisomes and

cytosolic FDH during glucose adaptation. Furthermore, inasmuch as the

objective of the present project is to investigate the manner in which

peroxisomes are sequestered, rather than the act of degradation per se, if

peroxisomal sequestration parallels enzyme inactivation then enzymes assays

will be useful to assess the presence of these two proteins. Samples were taken

and assayed by volume rather than by protein concentration on the basis of

enzyme activity per methanol-induced cell volume at 0 hour of glucose

adaptation. The assumption is that in any volumetric sample taken during the


1 0 1 2 3 4 6
1 01 2346
120- M AOX

,- Z- A FDH

X 80 -

> 60- i\

i 40

wu 20

0 1 2 3 4 5 6

Hours after addition of Glucose

Figure 3-4. Degradation of peroxisomal and cytosolic enzymes during glucose
adaptation. Pichia pastors cells (GS115) were cultured in methanol induction
medium until stationary at which time glucose was added to begin adaptation.
Cell-free extracts were prepared at selected times and peroxisomal alcohol
oxidase (AOX) and cytosolic format dehydrogenase (FDH) activities measured.
The values represent the average s.e.m. of six determinations. Protein levels
of AOX and FDH within extracts of glucose adapting cells were analyzed by
immunoblots (inset).

course of adaptation, the same number of cells which were present in a

methanol-induced state at 0 h will be also present in that sample.

Autophaya of Peroxisomes Induced By Glucose

Ultrastructural Examination of Autophagy

P. pastors and H. polymorpha cells were fixed for transmission electron

microscopy utilizing a potassium permanganate protocol which is useful for

detecting membrane events (Veenhuis et al., 1983; Tuttle et al., 1993). When

grown to stationary phase in media originally containing 0.5% methanol as the

sole carbon and energy source, the yeast cells contain many large peroxisomes

(Fig 3-5B; Veenhuis et al., 1983; Gould et al., 1992). The peroxisomes are

observed assembled in clusters in the cytoplasm, often adjacent to endoplasmic

reticulum, rounded (i.e., spherical or lobulated) vacuoles, and other cytoplasmic

organelles. In contrast, cells grown in glucose-containing media, i.e., not

previously induced in methanol, are devoid of identifiable peroxisomes (Fig. 3-

5A; Veenhuis et al., 1983; Gould et al., 1992). The objective of these

ultrastructural studies, then was to investigate how the cells were able to

dispose of the methanol-induced peroxisomes and return to the morphology

typical of glucose-grown cells. In methanol-induced P. pastors drastic changes

in the ultrastructure of the vacuole are observed within 0.5 to 1 hour after the

addition of glucose (Fig 3-6). The earliest change noted is the contortion of the

vacuole to a cup-shaped appearance with a cluster of peroxisomes in the hollow

formed (Fig 3-6A,B).


r a r

n )

Figure 3-5. Ultrastructure of methylotrophic yeasts grown in glucose and
methanol. H. polymorpha (A) and P. pastors (B) cells were incubated in cultures
containing glucose (A) or methanol (B) as the sole carbon and energy sources
then fixed with potassium permanganate and examined by transmission electron
microscopy. p = peroxisome; m = mitochondrion; n = nucleus; v = vacuole; Bar =
0.5 pm.

I'" w--



figure 3-. Morpnological cnaractenzation or alucose-inaucea peroxisomal
degradation in P. pastors. P. pastors was grown in methanol medium until
stationary growth was achieved. Glucose was then added to the cultures to a
final concentration of 2.0% then incubated for one additional hour and potassium
permanganate fixed as described in Chapter 2. p= peroxisome; v = vacuole; m =
mitochondrion; n = nucleus; Bar = 0.5 pm.

, n A


In other micrographs, vacuoles are seen completely encircling a peroxisome

cluster (Fig. 3-6C). Another prevalent configuration observed is a cluster of

peroxisomes entirely within the vacuolar lumen with the members of a

peroxisome cluster observed in various stages of degradation (Fig 3-6D).

These observations suggest a course of events reminiscent of

mammalian microautophagy (see Chapter 1). In some way the vacuole is

triggered to change shape, distending to cup around a cluster of peroxisomes

and continuing to extend arms until the peroxisomes are completely surrounded

but still not taken up into the vacuole (Fig 3-6A-C). Subsequently, the inner

vacuole membrane disappears, perhaps by degradation or fusion with the

peroxisome membranes, and the peroxisomes are thereby incorporated into the

vacuolar lumen were they are degraded by the vacuole-resident hydrolases (Fig


The ultrastructure of H. polymorpha cells during glucose adaptation is

strikingly different. In this yeast, during the first hour of glucose treatment,

individual peroxisomes within a cluster can be seen surrounded by from 2 to 12

extra membrane layers of unknown origin (Fig. 3-7B; Veenhuis et al., 1983). In

some cells peroxisomes are seen singly or in small groups in the lumen of the

vacuole, sometimes apparently partially degraded (Fig. 3-7B). A possible

sequence of events has been described by Veenhuis et al. (1983) and bears a

marked resemblance to mammalian macroautophagy in which cytoplasmic

components are first sequestered by a double membrane body forming an

autophagosome (Dunn, 1990a) which fuses with lysosomes forming a


Figure 3-7. Morphological characterization of glucose-induced peroxisomal
degradation in H. polvmorpha. H. polymorpha was grown in methanol medium
until stationary growth was achieved. Glucose was then added to the cultures to
a final concentration of 2.0% then incubated for one additional hour and
potassium permanganate fixed as described in Chapter 1. Arrow in A points to
outer autophagosome membrane layer. p = peroxisome; v = vacuole; m =
mitochondrion; Bar = 0.5 pm.

degradative autolysosome (Dunn, 1990b). A similar course of events may take

place in the case of H. polymorpha. One possible scenario is that membranes

of unknown origin (possibly endoplasmic reticulum cistemae, if true to the

mammalian process) sequester individual peroxisomes from the remainder of the

cytosol forming an autophagosome. The limiting membranes of the

autophagosome may possess the ability to recognize and fuse with the

degradative vacuole whereby peroxisomes are deposited into the vacuolar

lumen and degraded.

Vacuolar Degradation-Deficient Mutants Verify the Site of Peroxisome
Degradation During Glucose Adaptation

The biogenesis of the vacuole is presently being studied in

Saccharomyces cerevisiae (see Klionsky et al., 1990 for review) and it has been

determined that the vacuolar endoproteases proteinase A (PrA) and proteinase

B (PrB) are responsible for self-activating their zymogen forms as well as for the

zymogen forms of several other major vacuolar proteases such as

carboxypeptidase Y (CPY; see Chapter 1). In the absence of functional PrA and

PrB, the proteolytic capacity of the vacuole is sharply curtailed. The homologs

of PrA and PrB have been identified in P. pastors and strains prepared by

homologous recombination which lack functional forms of either or both of these

proteases (M.A.G. Gleeson, personal communication). Initially these strains

were tested to determine what effect these gene knockouts had on vacuolar


Analysis of PrA/PrB-Defective Strains

Enzyme activity assays were conducted for the presence of active PrA

and CPY according to Jones (1977; 1991b). In SMD1163 (PrA/PrB) and SMD

1168 (PrA), PrA activity was decreased by ~80% compared to the wild type

strain GS115. CPY activity was evaluated utilizing a semi-qualitative

colorimetric colony assay, and indicated minimal CPY activity in both PrA"

strains. Interestingly, a color developed in SMD1165 (PrB) intermediate in

intensity to that of the wild type strain and the PrA strains suggesting a more

stringent requirement for PrA in zymogen activation than for PrB, consistent with

the S. cerevisiae literature (Klionsky et al., 1990).

It has been shown in S. cerevisiae that culturing cells in media lacking

nitrogen, carbon, or essential amino acids causes a large increase in

nonspecific autophagic sequestration of cytoplasmic components and their

degradation in the vacuole (Teichert et al., 1989; Takeshige et al., 1992; Baba et

al., 1994). In S. cerevisiae mutants lacking PrA and PrB the increase in overall

protein degradation rate due to nutrient deprivation is abolished. P. pastoris

strains with and without functional PrA and/or PrB were tested for their ability to

degrade endogenous proteins in response to amino acid deprivation. Cells were

metabolically labeled in low sulfate containing media containing [3S] methionine

and cysteine (see Chapter 2) and glucose until stationary then harvested,

washed and resuspended in chase media containing or lacking histidine (an

amino acid required for growth of these strains). Overall protein degradation

was estimated by quantifying the ratio of trichloroacetic acid-soluble and

insoluble counts in serial samples (Fig 3-8). The results indicate that in the

strain containing functional copies of the genes encoding PrA and PrB (GS115)

depriving the cells of histidine caused an increase in overall protein degradation

(compare open squares to closed squares in Fig. 3-8). This indicates that

during this type of nutritional stress the cell responds by increasing the

degradation of cellular proteins. On the contrary, depriving PrA and/or PrB-

defective strains of histidine caused no increase in protein degradation,

indicating that these proteases are essential for starvation-induced degradation.

These data suggest that the homologs of PrA and PrB in P. pastoris are

essential for starvation-induced degradation. Furthermore, these proteases are

implicated in the activation of vacuolar zymogens suggesting a role in vacuolar

biogenesis similar to their counterparts in S. cerevisiae (see Chapter 1).

Peroxisome Degradation in PrA/PrB-Deficient Strains During Glucose

Glucose adaptation experiments were performed on the PrA/PrB double

knockout strain of P. pastors (SMD1163). The cultures were grown to stationary

phase in methanolic media then glucose was added to the cultures to turn on

autophagy of peroxisomes; as in previous experiments with GS115, samples

were taken at the time of glucose addition and 6 h later and cell-free extracts

prepared. The extracts were assayed for alcohol oxidase and format

dehydrogenase (FDH) activity and the results are represented in Figure 3-9A.

The degradation of AOX and FDH was almost completely blocked in these

2.0 -
D GS115 + HIS
1.5 ASMD1163
z 0.5

I 0.00 0
..-.. ... .....

-0.50 I
1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0


Figure 3-8. Protein degradation induced by histidine starvation is blocked in the
absence of functional proteinases A and B. Parental P. pastors (GS115: his4)
cells were metabolically labeled with [4C]valine and then chased with media
containing cold valine in the absence or presence of histidine (HIS). Proteinase
mutant strains, SMD1163 (his4, pep4, prbl); SMD1165 (his4, prbl); and SMD
1168 (his4, pep4), were labeled with [14C]valine and chased in the absence of
histidine. The release of trichloroacetic acid-soluble radioactivity was measured
and the percent cellular protein degraded calculated as a percentage of total
radioactivity at zero hour.

mutants. These data show that, in the absence of PrA and PrB, these cells are

not able to degrade either AOX or FDH. Based on previous data that PrA and

PrB are required for vacuolar degradation, this suggests that the vacuole is the

site of degradation of both peroxisomes and the cytosolic enzyme FDH.

Ultrastructural Observations During Glucose Adaptation

The effect of glucose adaptation was also studied at the ultrastructural

level in the putative vacuolar mutant strain of P. pastors (SMD1163). Methanol-

induced cells were sampled when grown to stationary phase (0 h) and at 6 h

after the addition of glucose and fixed for morphological examination by the

potassium permanganate procedure to enable visualization of membrane events

(Fig. 3-9B). At 0 h, the mutant cells exhibited clusters of peroxisomes and the

other cytoplasmic organelles with no apparent differences from the parental

strain (not shown). After 6 h of glucose adaptation, the vacuoles of the putative

vacuolar mutants were filled with undegraded peroxisomes (Fig. 3-9B). Other

recognizable organelles were not observed in the vacuole. This result suggests

that the vacuole is unhindered in its ability to sequester the peroxisomes but that

it lacks the proteolytic functionality required to degrade them.

The data demonstrate that the P. pastors homologs of PrA and PrB are

required for vacuole-mediated degradation. Evidence is presented that

suggests that these endoproteases act as vacuolar zymogen activators, a

function they have been shown to serve in S. cerevisiae. Mutants of P. pastors

defective in PrA and PrB activity can be utilized to determine whether a

1 120 A
100 AOX
180 FDH

GS115 SMD1163

I i-w

Figure 3-9. Proteinase mutants are not able to degrade methanol-induced
components during glucose adaptation. Parental GS115 (his) and mutant
SMD 163 (his, pep4, prbl) cells were grown in methanol then adapted to
glucose. Alcohol oxidase (AOX) and format dehydrogenase (FDH) activity were
determined at 0 hours and 6 hours of glucose adaptation (A). The values
represent the average + s.e.m. of at least three determinations done at 6 hours
of adaptation and is reported as a percentage of that measured at zero hour. At
6 hours of glucose adaptation, the mutant cells were fixed and processed for
ultrastructural examination (B). Intact peroxisomes (*) were observed within the
vacuoles (v) of proteinase mutant cells (b). Arrowheads indicate the perimeter of
the yeast vacuole. p = peroxisome; Bar: 0.5 pm.

particular degradation event takes place in the vacuole by evaluating

biochemically whether or not a decrease in degradation has occurred in

themutant strain relative to the wild type (see Fig. 3-8; 3-9B). Additionally,

ultrastructural analysis revealed an accumulation of undegraded peroxisomes in

the vacuole of protease mutants, suggesting the vacuole as the normal site of

degradation (3-9B). By utilizing these mutants in this manner, the vacuole has

been shown to be the normal site of degradation of methanol-induced

peroxisomes and FDH during glucose adaptation (see Fig. 3-9). The vacuole

also plays a major r6le in the degradation of cellular components during

starvation for essential amino acids (see Fig. 3-8).

Selective Autophaqy

To directly measure the degradation of mitochondrial proteins during

glucose adaptation in H. polymorpha and P. pastoris, cells were metabolically

labeled with ["S]cysteine and methionine during growth to stationary phase in

methanolic media. The cultures were then chased in the presence of glucose

and the mitochondrial protein Fi-ATPase, p subunit (F1p) was

immunoprecipitated from cell-free extracts generated at 0, 1, and 3 hours of

chase (Tuttle et al., 1993). The amount of label associated with Flp was

quantitated by phosphor-imaging after separating the extracts by 7.5% SDS-

PAGE. The amount of label incorporated into the peroxisomal enzymes AOX

and DHAS was also quantitated as above (see Fig. 3-2, 3-3). Figure 3-10 shows


S -------------.
0 -100 1

80 AOX
*= \ ADHAS

0 60 i

0 40

20 -

"" 0

0 1 2 3 0 1 2 3
Hours After Addition of Glucose

Figure 3-10. Degradation of peroxisomal and mitochondrial proteins during
glucose-mediated peroxisome removal in H. polvmorpha and P. pastors.
Cultures of yeast were grown to stationary phase in low sulfur media with
methanol as the sole carbon source, in the presence of "S-methionine/cysteine.
At 0 hour solid glucose was added to a final concentration of 2% to begin
adaptation and solid methionine and cysteine were added to prevent reutilization
(chase) of the radioisotope. Duplicate samples were removed at various times,
protein extracts were prepared and analyzed as described in Chapter 2. For F,
ATPase, B-subunit (F1B), protein was immunoprecipitated from the protein
extracts, fractionated by electrophoresis and quantified with a Phosphorimager.
AOX and DHAS were quantified by counting the radioactivity associated with the
respective protein bands excised from Coomassie stained gels. Values shown
are the mean SEM for 3 or 4 determinations in three separate experiments.
AOX = alcohol oxidase; DHAS = dihydroxyacetone synthase.

the results of this analysis which indicate that while peroxisomal proteins are

rapidly degraded, the mitochondrial protein F I is stable.

Ultrastructural evidence also suggests that the autophagic degradation of

peroxisomes during glucose adaptation is selective (see Fig. 3-6, 3-7, 3-9). In

both H. polymorpha and P. pastors, mitochondria or other cytoplasmic

organelles were never seen sequestered along with peroxisomes.

Chapter Summary

Utilizing metabolic labeling/chase protocols in both H. polymorpha and P.

pastors, it was demonstrated that the synthesis of the peroxisomal enzymes

DHAS and AOX was very low during stationary phase in methanol-induced

cultures and therefore the disappearance of DHAS and AOX proteins which

occurs during glucose adaptation must be due to enhanced degradation (see

Fig. 3-3). It was also directly demonstrated that these peroxisomal proteins are

quite stable in stationary methanolic cultures and their degradation is rapidly

enhanced after the addition of glucose (see Fig. 3-2).

In cultures of H. polymorpha and P. pastors grown to stationary phase in

methanolic media, the peroxisomal enzymes AOX and DHAS are synthesized at

low rates and are quite stable (see Fig. 3-2, 3-3). Upon addition of glucose to

these cultures, the peroxisomal enzymes are rapidly degraded (Fig. 3-2, 3-4). In

P. pastors, clusters of peroxisomes are sequestered directly by the vacuole and

degraded there as in microautophagy (see Fig. 3-6). In H. polymorpha,

individual peroxisomes are sequestered by multiple membrane layers and then

fuse with the degradative yeast vacuole and deposit the peroxisomes into the

vacuole where they are degraded as in macroautophagy (see Fig. 3-7; Veenhuis

et al., 1983).

I have provided evidence that PrA and PrB are vacuolar proteases which

act similarly to their counterparts in S. cerevisiae. These proteases are required

for the proteolytic function of the vacuole, due to their innate endoproteolytic

activity and probably by activating zymogen forms of vacuolar proteases (e.g.,

CPY, see above). This evidence consists of 1) the inhibition of degradation in

histidine-starved vacuolar mutant strains (see Fig. 3-8); 2) the absence of CpY

activity in cells lacking functional PrA and PrB; and 3) the accumulation of

undegraded peroxisomes in the vacuole during glucose adaptation (see Fig. 3-

9B. Utilizing these mutants, I have demonstrated that in P. pastors the vacuole

is the site of peroxisomal degradation during adaptation to glucose (see Fig. 3-

9). In addition, the cytosolic methanol catabolic enzyme FDH is also degraded

in the vacuole although the method of vacuolar uptake is unclear (see Fig. 3-9).

Finally, I presented evidence that peroxisome degradation induced by

glucose is selective in that mitochondrial components that are not degraded

simultaneously. This evidence is both biochemical (see Fig. 3-1; 3-10) and

morphological (see Fig. 3-6, 3-7). All the evidence presented indicates that the

degradation of peroxisomes occurs due to autophagy. Morphologic evidence


shows that peroxisomes are taken up into the vacuole as a whole where their

constituent enzymes are degraded nonspecifically.



An important reason that one uses ascomycetous yeasts to study

eukaryotic systems is the ease of identification of genes that function in a

pathway of interest afforded by certain of these yeasts. Since they can be

maintained indefinitely in the haploid state, mutations have only to affect the

single allele of a given gene to affect the phenotype. These mutations can be

brought about by random mutagenesis or by gene-replacement with a cloned

allele and identified by screening for a phenotype which suggests a defect in the

pathway being studied. Subsequently, haploid mutant strains can be mated with

other mutants to identify different complementation groups (affected genes)

present in the mated mutants. Alternatively, haploid mutants can be mated to a

strain with a wild type background backcrossingg) to produce diploids with a

normal phenotype (if the mutation is recessive) which can then be stimulated to

undergo meiosis, yielding haploid progeny, some mutant, some wild type for the

gene of interest. During meiosis in yeast, homologous recombination events are

common so that after several rounds of backcrossing, haploid strains may be

obtained in which the gene of interest will be the only mutated gene (aside from

selectable auxotrophic markers, e.g., HIS4).

Genes may be cloned by transforming haploid mutants with a plasmid-

borne library, assaying for complementation, and then recovering the DNA in

rescued mutants. Sequence analysis leads to the ready identification of the

proteins in the pathway of interest. Further genetic manipulation and study of

the proteins can yield data on protein localization, interactions with other cellular

components, structure, and function.

Much time and effort has been expended on the investigation of the

regulation and mechanism of autophagy in mammalian systems (see Chapter 1).

This work has revealed the existence of regulatory pathways which stimulate

autophagy (e.g., starvation for certain amino acids) and some of the

morphological features which led to the designation of macroautophagy and

microautophagy. Unfortunately, due to the difficulty of performing classical and

molecular genetic analyses in higher eukaryotes, the molecular mechanisms of

these phenomena remain almost entirely unknown.

In order to take full advantage of the model for selective autophagy which

the methylotrophic yeast system provides (see Chapter 3), I undertook to

develop a system for the screening, verification, and characterization of P.

pastors strains which are defective in glucose-induced selective autophagy.

The specific aims of this part of the present study were: 1) to chemically

mutagenize an essentially wild type strain of P. pastors; 2) screen for mutant

colonies which are unable to degrade methanol-induced peroxisomes in

response to glucose; 3) verify and characterize the mutant phenotype; 4) and

backcross the mutants and assign complementation groups.

Mutagenesis, Screening, and Verification

A strain of P. pastoris which is wild type except for the HIS4 selectable

marker for histidine auxotrophy (GS115) was mutagenized using N-methyl-N'-

nitro-N-nitrosoguanidine (which mainly causes base conversions leading to point

mutations) according to the methods of Cregg et al. (1990). After treatment for 1

hour, leading to a 95% kill rate, cells were aliquoted and frozen at -800C until

tested for autophagy deficiency.

In preparation for screening, mutagenized cells were spread on MIM agar

plates containing 0.5% methanol (master plates). The plates were incubated for

4 to 5 days to allow 1 to 2 mm colonies to form, at which time the plates were

replicated onto nitrocellulose circles which were subsequently placed onto GA

agar plates (containing 2.0% dextrose) or fresh MIM plates (positive control for

the assay; see Fig. 4-1A) and incubated for 14 to 16 h. At this time the AOX-

containing peroxisomes in normal cells will have degraded, therefore the

strategy for screening was to identify colonies which still retained AOX activity

after this period of adaptation to glucose. In order to do this, the replica-

containing nitrocellulose circles were submitted to colorimetric direct colony

assays for AOX activity as described in Chapter 2 and colonies retaining AOX

activity (Fig. 4-1 B) were selected from the master plates for verification and

further characterization.



Figure 4-1. Screening for glucose-induced selective autophagy-deficient mutants
(qsa) by direct colony assay for alcohol oxidase. Mutagenized colonies were
incubated on MIM plates for 5 d then replicated to nitrocellulose circles. The
nitrocellulose circles were then placed on fresh MIM plates (A; positive control)
or glucose plates (B) and incubated 14 h to initiate glucose-mediated
peroxisome degradation. The colonies in (A) all show positive reaction product
while in (B) most of the colonies are not visible because they have degraded the
methanol-induced peroxisomes and so remain white. Putative gsa mutants retain
alcohol oxidase activity after glucose treatment (dark colonies in B).

In this way, 48 positive colonies were selected out of approximately

200,000 colonies screened. These 48 colonies underwent a verification

procedure which consisted of the same procedure as described in Chapter 3 to

quantitate the effect of glucose adaptation on methanol-induced proteins.

Briefly, the colonies were grown up in YPD precultures, inoculated 1:30 into 20

mL MIM cultures, incubated -2 d, then glucose was added to a final

concentration of 2% to initiate glucose adaptation. Samples were taken at 0 and

6 h after addition of glucose and peroxisomal AOX and cytosolic FDH activity

assayed to quantitate the putative autophagy defect during glucose adaptation.

Of these 48 positive colonies, 15 were verified to be defective in AOX

and/or FDH degradation to the extent that they retained at least 2-fold more

enzyme activity than the parental GS115 cells after 6 h glucose adaptation (Fig.

4-2). These 15 strains were tentatively designated glucose-induced selective

autophagy-deficient (gsa mutants). Further confirmational assays indicated that

the autophagy deficiency was variable from culture to culture for some of these

strains, suggesting the putative strains were not clonal either 1) because of

contamination with other strains due to the clumping characteristic of P. pastors

cells in culture or 2) because reversion of some of the mutant cells to the

parental (or other) allele caused mixed populations in the cultures. Three of the

original isolates, designated gsa4.3, gsai2.1, and gsa13.1 (see Fig. 4-2), were

selected for backcrossing and complementation analysis by virtue of the severity

of the exhibited autophagy defects and their relative consistency in tests of

glucose adaptation.


2.0 -

Vacuolar Proteinase-
Deficient Mutants

*Alcohol Oxidase
o Formate Dehydrogenas

Il ll ll lll ll




gsa mutants

Figure 4-2. Comparison of degradation of methanol-induced enzymes in
vacuolar protease-deficient mutants, asa mutants, and the parental strain
(GS115). The amount of alcohol oxidase and format dehydrogenase activity
present in samples removed from cultures 6 h after glucose addition relative to
that present at 0 h was compared among vacuolar protease mutants lacking PrA
and/or PrB, the 15 original isolates of the screening procedure for gsa mutants
which were considered verified (designated by numerals on X axis; see text for
details), and GS115. The results shown are normalized to GS115, the mean for
which is set at 1.0 (horizontal line).



I l l I I I I

I1I1I1IA~~lllllllll1411111 II111II




Backcrossing and Complementation

Backcrossing was carried out by replica plating the histidine-requiring gsa

mutants with an arginine-requiring strain with a wild type background (GS190-3)

to SM agar plates, lacking nitrogen, causing them to mate. Diploids were

selected on plates lacking amino acids (DSM) since the auxotrophic mutations

should complement to produce a prototrophic diploid. The diploids where then

forced to undergo meiosis on the same nitrogen-free agar medium which causes

them to mate (SM) and the resulting spores are regenerated to vegetative

growth, then screened for the gsa phenotype by direct colony assay and for

auxotrophy. Importantly, the gsa mutants selected from this backcross were

arginine auxotrophs, ensuring that the resultant gsa backcross could not have

gone through the procedure without mating and the attendant homologous

recombination events which are the main purpose of backcrossing. In other

words, the genotype of the gsa mutant selected from crossing of a gsalhis4

strain with a GSAlarg4 strain would be gsalarg4 (see Table 2-1).

In the next step in the backcrossing regimen, a gsalarg4 strain was

crossed to GSA115 (GSAIhis4) and the end product was a gsalhis4 strain. In

order to obtain a strain that reliably contains only a singly gsa mutation, 4 to 6

rounds of backcrossing are necessary.

Complementation analysis was carried out in the same manner, with the

exception that cells of one histidine auxotrophic gsa clone were mated with cells

of an arginine auxotrophic clone of a different gsa isolate. The resulting diploid

strains were tested for the ability to degrade methanol-induced AOX and FDH in

response to glucose in the usual manner. Strains were considered

complementary, i.e., contained mutations in different genes, if AOX was

degraded at the wild type rate. Accordingly, strains which did not complement,

degraded AOX at a rate significantly slower than the wild type, and were

considered to contain different mutant alleles of the same gene. Controls for

complementation analysis consisted of crossing gsa mutants to one of the wild

type strains (positive control, diploid exhibits normal rate of AOX degradation;

complementation of the gsa defect by the wild type strains also demonstrates the

recessive nature of the mutant allele) and crossing a given gsa clone to a clone

containing the same mutant allele, i.e., a backcrossed strain of the same clone,

but a different auxotrophic marker (negative control in which diploid exhibits the

mutant rate of AOX degradation).

This analysis has revealed two complementation groups designated gsal

and gsa2 and two mutant alleles of gsa2 have been isolated, designated gsa2-1

and gsa2-2. The strains carrying these mutations have been named according

to their gsa alleles: WDY1 (derived from original isolate gsal2.1), WDY2

(derived from gsa4.3), and WDY3 (derived from gsal3.1; see Table 2-1, Fig. 4-

Biochemical and Ultrastructural Examination of WDY1 and WDY2

After backcrossing WDY1 and WDY3 once and WDY2 four times, these

strains were further analyzed for their ability to degrade methanol-induced

peroxisomes and FDH in response to glucose (Fig. 4-3A). This was tested by

the usual 6 h glucose adaptation procedures as described in Chapter 2. These

three strains were all severely defective in the removal of both methanol-induced

peroxisomes (AOX: solid bars in Fig. 4-3A; compare mutants to GS115) and

cytosolic FDH (Fig. 4-3A, hatched).

Using the same cell-free extracts as for the AOX and FDH assays

described in the preceding paragraph, these strains were tested to determine

whether they had lost the selectivity of degradation exhibited by the parental

strain (GS115; see Fig. 3-6, 3-12). This was accomplished by measuring

enzyme activities not acutely regulated by glucose in wild type P. pastors, i.e.,

mitochondrial cytochrome c oxidase (CCO) and cytosolic fructose-1,6-

bisphosphatase (FBP; Fig. 4-3B). After 6 hours of glucose adaptation, the

activities of these two enzymes remain relatively constant in parental GS115,

suggesting that mitochondria and cytosolic components are not randomly

degraded with the methanol-induced peroxisomes and FDH. Similarly, in the

gsa mutants, the levels of these enzymes remain constant or 20 to 50% higher

(Fig 4-3B). These data suggest that selectivity of degradation is not lost in these

mutants since, while peroxisomes and FDH are degraded to some degree, the

degradation of these components is not enhanced relative to the wild type. In

90 A 220 B
0 75


I 15 140-

0 80


P. pastors Strain

Figure 4-3. Analysis of specific and nonspecific degradation by osa mutants.
Enzyme assays were conducted on cell-free extracts prepared from various
mutant and wild type (GS115) strains at 0 and 6 h of glucose adaptation. The
results are expressed as the percentage of enzyme activity present at 6 h
relative to that present at Oh. Activity of the glucose-sensitive proteins alcohol
oxidase (AOX) and format dehydrogenase (FDH; A). Activity of the glucose-
insensitive proteins cytochrome c oxidase (CCO) and fructose-1,6-
bisphosphatase (FBP; B). Bars represent mean + s.e.m.

fact, the levels of CCO and FBP tend to be higher in the gsa mutants,

suggesting that a certain level of constitutive degradation of these cellular

constituents occurring in stationary phase in methanol in the wild type cells is

blocked in the mutants. These data suggest that constitutive degradation of

cytoplasmic components also proceeds by autophagy. This possibility seems

more likely than glucose induction of these enzymes since glucose catabolites

are known to induce the inactivation of FBP in S. cerevisiae (Holzer, 1976;

Chiang and Schekman, 1991; Schork et al., 1994).

In order to rule out the possibility that the phenotype being tested was

due to a generalized defect in vacuolar degradation, two analyses were

undertaken. First, the activities of the important vacuolar proteases PrA and

CPY (see Chapters 1 and 3) were assayed in WDY1 and WDY2 and found to

contain approximately the same levels of these enzymes as the wild type.

Second, ultrastructural examinations of these two strains were carried out. It

was expected that, if the vacuole had been rendered proteolytically inactive

while autophagic sequestration was still operative, then peroxisomes would

accumulate in the vacuole as was shown to be the case in vacuolar proteinase

deficient mutants (see Fig. 3-10). Indeed, based on the standard glucose

adaptation assay, the vacuolar protease mutants are autophagy-deficient (see

Fig. 3-10) due to their inability to degrade sequestered components. However,

as mentioned previously, the purpose of this study was to understand how and

where cellular components are sequestered, not the act of degradation per se

and therefore, if gsa mutant strains were found to be deficient in vacuolar

proteolysis, these strains would not be further studied.

Cells of the strain WDY2 were fixed for electron microscopic examination

using the standard potassium permanganate protocol at 0, 1.5, and 5 h of

glucose adaptation (Fig 4-4). As a check that the cultures being sampled

actually contained autophagy-defective strains, cell-free extracts were prepared

at 0 and 6 hours of glucose adaptation from these same cultures and assayed

for AOX activity. The biochemical results indicated that the cells which were

fixed were mutant (data not shown). Just prior to initiation of glucose adaptation,

the morphology of WDY2 (Fig. 4-4A) was indistinguishable from wild type cells.

These cells contained peroxisomes in clusters as usual, typical mitochondria,

and nuclei.

Instances of autophagic sequestration events were not detected in WDY2

at 1.5 of glucose adaptation (Fig. 4-4B) at a time point when peroxisomes are

commonly seen being taken up by microautophagy in the wild type cells (see

Fig. 3-6). After 5 h of glucose adaptation, microautophagy was still not observed

in WDY2, but examples of macroautophagic sequestration were observed (Fig.

4-4C). In this micrograph, the vacuole appears to have fused with a peroxisome-

containing autophagosome. The vacuolar membrane is continuous with the

outer autophagic vacuolar membrane (see arrow in Fig. 4-4C) in which a

peroxisome, still encased in extra membranes layers is observed. This process

is similar to events described during glucose adaptation in H. polymorpha (see

Fig. 3-7; Veenhuis et al., 1983). Macroautophagy has not been observed in wild

Figure 4-4. Ultrastructural examination of glucose adaptation in WDY2. WDY2
was grown to stationary phase in methanol and then glucose was added to a
final concentration of 2.0% to initiate glucose adaptation. Samples were
removed just prior to glucose addition (A), at 1.5 h after glucose addition (B),
and 5 h after glucose addition (C). Arrow in C points to vacuole membrane which
is continuous with autophagosome membrane surrounding peroxisome. p =
peroxisome; m = mitochondrion; n = nucleus; v = vacuole; Bar = 0.5 pm.

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EBMI1V55Q_EOSWXF INGEST_TIME 2012-09-24T13:53:40Z PACKAGE AA00011831_00001