INFORMATION TO USERS

This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.

The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.

In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6” x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

University Microfilms International A Bell & Howell information Company 300 North Zeeb Road. Ann Arbor. Ml 48106-1346 USA 313 761-4700 800 521-0600 Order N u m b e r 9238146

Membrane assembly studies of subunit H, leader peptidase and M13 procoat proteins Escherichiain colt

Cheng, Shiyuan, Ph.D.

The Ohio State University, 1992

UMI 300 N. Zeeb Rd. Ann Aibor, MI4H106 MEMBRANE ASSEMBLY STUDIES OF SUBUNIT H, LEADER PEPTIDASE AND M13 PROCOAT PROTEINS IN ESCHERICHIA CPU

DISSERTATION

Presented in Partial Fulfillment of the Requirement for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

By

SHIYUAN CHENG. B. S.

The Ohio State University

1992

Dissertation Committee: Approved by ^

* " * e “ i w Pappachan E. Kolattukudy Advisor

James A. Cowan Ohio State Biochemistry Program*—•" To my wife, Bo Hu and my parents, for their continuous love, understanding and support. To my lovely daughter, Sarah (Ruirui), for bringing me wonderful joy. ACKNOWLEDGEMENTS

I would like to express my sincere appreciation to my mentor Dr. Ross E. Dalbey for his kind understanding, his great encouragement at the turning point of my graduate career, and for his generous moral and material support as well as for his thoughtful advice and guidance throughout the research.

I would also give thanks to Dr. Samuel Kaplan for instructing me how to express the subunit H protein. My sincere appreciation also goes to my colleagues, Heng-Yi, Bill, Liming, Meesook, Lee, Guoqing, Xintong, Ruby, Songho and people in laboratories of Drs. Ming-Daw Tsai and James A. Cowan, for their friendship, helpful discussions and sharing instruments. Special thanks also to Drs. Andreas Kuhn and Gunnar von Heijne for their cooperation.

Finally to my family, especially to my lovely wife, whose love, patience, understanding, encouragement, support and priceless gift of my beautiful daughter make me forever indebted. To Sarah (Ruirui) for being the world’s most wonderful and beautiful daughter. Thanks also to my parents and siblings for their continuous love, encouragement and support throughout my life .

i i i VITA

July 1, 1957 ...... Born in Wuhan, P. R. China

January, 1982 ...... B.S. in Biochemistry, Wuhan University Wuhan, P. R. China

September, 1985-Present ...... Graduate Research Associate or Teaching Associate in The Ohio State Biochemistry Program, The Ohio State University.

PUBLICATIONS

Shen, L. M., Lee, J. I., Cheng, S. Y., Jutte, H., Kuhn, A., and Dalbey, R. E. (1991) "Site-directed Mutagenesis to Define the Limits of Sequence Variation Tolerated for Processing of the M13 Procoat Protein by the Escherichia coli Leader Peptidase", Biochemistry, 30, 11775- 11781.

i v Farooqui, A. A., Rammohan, K. W., Cheng, S. V., and Horrocks, L. A (1988) "Membrane Bound Diacylglycerol Lipase in Bovine Brain: Purification, Characterization, and Regulation" Frontiers of Chemistry: Vol. 1, Biotechnology. 75-89. Published by Chemical Abstract Services, Columbus, Ohio.

Field of Study:

Major Field: Biochemistry Emphasis in Molecular and Cell Biology, Professor Ross E. Dalbey, Advisor.

v TABLE OF CONTENTS

PAGE DEDICATION i i

ACKNOWLEDGEMENTS i i i

V ITA i v

LIST of TABLES...... ix

LIST of FIGURES...... x

LIST of ABBREVIATIONS...... x i v

CHAPTER PAGE

I INTRODUCTION ...... 1

1.1. The Role of Charged Amino Acid Residues in Membrane Protein Topology ...... 1 1.2. Subunit H Protein of the Photosynthetic Reaction Center from Rhodobacter sphaeroides ...... 11 1.3. Protein Export in £. coli ...... 15 1.4. £. coll Leader Peptidase ...... 2 2

11. MATERIALS AND METHODS...... 2 7

2.1. Materials...... 2 7 2.2. Bacterial Strains, Media and Plasmids ...... 2 9 2.3. Special Growth of the Cells Carrying Subunit H Protein Gene of the Photosynthetic Reaction Center from Rhodobacter sphaeroides ...... 3 0 2.4. Preparation of E. coli Cell Membrane Fractions. 31

v i 2.5. Alkaline Phosphatase Activity Assay ...... 3 3 2.6. Reverse Protease Mapping on Subunit H Protein Expressed in £. soli ...... 3 4 2.7. Alkali Fractionation ...... 34 2.8. SDS-PAGE Gel Electrophoresis and Fluorography 3 5 2.9. Agarose Gel Electrophoresis ...... 41 2.10. Polyacrylamide Gel Electrophoresis for DNA Sequencing Analysis ...... 41 2.11. Preparation of M13 RF (Replicative Form) and Plasmid DNA ...... 4 2 2.12. Oligonucleotide-Directed Mutagenesis ...... 4 5 2.13. DNA Sequencing ...... 4 9 2.14. Subcloning ...... 5 0 2.15. Transformation and Transfection ...... 53 2.16. Screening for Positive Clones with Different Methods ...... 54 2.17. Preparation of Spheroplasts ...... 5 7 2.18. Protease Accessibility Assay ...... 59 2.19. Immunoprecipitation ...... 6 0 2.20. Immunoblotting ...... 61

III. THE USE OF GENE FUSION TO DETERMINE WHICH REGIONS OF A PROTEIN ARE IMPORTANT FOR THE MEMBRANE ORIENTATION OF SUBUNIT H PROTEIN OF THE PHOTOSYNTHETIC REACTION CENTER FROM RHODOBACTER SPHAEROIDES ...... 63

3.1. Introduction ...... 6 3 3.2. Construction of Subunit H-PhoA Gene Fusions .. 6 7 3.3. Expression in £. coli of Subunit H Protein of the Photosynthetic Reaction Center from Rhodobacter sphaeroirifis ...... 7 6 3.4. The Insertion of Subunit H Protein into the Cytoplasmic Membrane of £. coli ...... 7 7 3.5. Expression of the Subunit H-PhoA Fusion Proteins ...... 8 5 3.6. The Negatively Charged Amino Acid Clusters Are Not Determinants of Membrane Orientation of Subunit H Protein ...... 8 6 3.7. Conclusion ...... 91

v i i IV. SECB-DEPENDENT AND SECB-INDEPENDENT INSERTIONS OF£. COLI LEADER PEPTIDASE 9 6

4.1. Introduction ...... 9 6 4.2. SecB Is Required for the Complete Translocation of Altered Leader Peptidases .... 101 4.3. Conclusion ...... 113

V. REQUIREMENT OF A MEMBRANE ANCHOR DEPENDS ON WHETHER OR NOT A PROTEIN IS SEC-DEPENDENT FOR MEMBRANE ASSEMBLY ...... 1 1 8

5.1. Introduction ...... 118 5.2. Inverted Leader Peptidase Requires Its Second Hydrophobic Domain for Membrane Assembly .. 124 5.3. Other Substitutions with Arginines in H2 Region Did Not Affect the Kinetics of Inverted Leader Peptidase Membrane Insertion ...... 12 7 5.4. The Membrane Anchor Region Is Not Essential for Sec-Dependent Procoat Membrane Insertion 129 5.5. Insertion of Inverted Leader Peptidase Mutants That Is Not Affected by a Positive Charge Is SecA-independent ...... 134 5.6. Conclusion ...... 139

LIST OF REFERENCES ...... 143

v i i i LIST OF TABLES

TABLE PAGE

2.1. Preparation of SDS-PAGE gel (with urea) ...... 37

2.2. Preparation of SDS-PAGE gel (with urea, for procoat protein) 3 8

2.3. Preparation of SDS-PAGE gel (without urea) ...... 3 9

2.4. Preparation of SDS-PAGE gel (for gradient gel without urea) ...... 40

3.1. Alkaline phosphatase activities of Subunit H-PhoA fusion proteins 9 0

4.1 Chaperones involved in E. coli protein export ...... 100

i x LIST OF FIGURES

FIGURES PAGE

1.1. Subceliular organelles of eukaryotic cells and protein traffic pathways within the cell ...... 2

1.2. Principal locations of bacterial membrane proteins . 3

1.3. Schematic illustration of integral and peripheral membrane proteins ...... 5

1.4. Model of the membrane orientations of three proteins of the plasma membrane of £. coli ...... 7

1.5. Model to explain why positively charged amino acid residues are determinants for membrane protein topology ...... 9

1.6. Three dimensional x-ray crystallographic structure of the photosynthetic reaction center from Rhodobacter sphaeroides ...... 1 3

1.7. pro-OmpA protein translocation via the export machinery ...... 1 9

1.8. Membrane integration of M13 phage procoat protein via Sec-independent pathway ...... 2 1

1.9. Current model of membrane topology of leader peptidase ...... 24

1.10. Membrane topology of inverted leader peptidase ...... 2 6 2.1. Screening for subcloned inverted leader peptidase mutants with restriction enzyme digestion after isolation of the plasmid DNA ...... 56

2.2. Screening for the subcloned procoat-OmpA A21-33 and A21-39 mutants with immunoprecipitation ...... 58

3.1. Membrane topology of subunit H protein within the photosynthetic reaction center from Rhodobacter sphaeroides ...... 65

3.2. Use of alkaline phosphatase gene fusions to analyze the membrane protein topology ...... 6 8

3.3 Construction of subunit H fpuhAVphoA gene fusion 7 0

3.4 In-frame fusion of subunit H-PhoA protein by oligonucleotide-directed gene deletion ...... 71

3.5. Constructs of subunit H-PhoA fusion proteins or subunit H mutants ...... 7 3

3.6 Expression of subunit H protein of the photosynthetic reaction center in E. coli strain HJM114 ...... 7 8

3.7. Alkali fractionation of membranes derived from £.. coli HJM114 cells which express the subunit H protein ...... 81

3.8. Reverse protease mapping of subunit H protein in the membrane vesicles isolated from E. coli HJM114 cells ...... 8 3

3.9. Expression of the subunit H-PhoA fusion protein in E. coli CCl 18 strain fPhoA~) ...... 8 7

3.10. Summary of the predicted membrane topology of the mutated subunit H-PhoA fusion proteins in this study ...... 9 2 x i 4.1 A working model of leader peptidase insertion into the plasma membrane of £. coli ...... 9 7

4.2. Gene constructs of leader peptidase used in this study and their membrane assembly properties ...... 102

4.3. The principal of protease accessibility assay ...... 104

4.4 Translocation of the large carboxyl terminal domain of wild-type leader peptidase is SecB-independent 106

4.5. SecB-dependent membrane insertion of leader peptidase A222-323 (pRD9) ...... 109

4.6. Membrane topology of pre-MBP, pro-OmpA and leader peptidase after the proteins are translocated across the plasma membrane ...... 111

4.7. Leader peptidase A4-50 is SecB-dependent for complete translocation of its large carboxyl-terminal domain ...... 1 1 2

5.1. Three classes of integral membrane proteins ...... 119

5.2. Signal hypothesis model of protein translocation ,. 121

5.3. Gene constructs of the wild-type inverted leader peptidase and its mutants, R64, R67, R69, R70, R71, R72 and A70-74 ...... 125

5.4. Substitution with arginine at position 67 and 69 (R67 and R69) or deletion 70 to 74 (A70-74) blocks the insertion of the inverted leader peptidase ...... 126

5.5. Substitution with arginine at other position of the second hydrophobic segment of the inverted leader peptidase does not block the export of P1 region .... 128

5.6. Kinetic study of the inverted leader peptidase mutant R72 that translocates across the membrane. 130

x i i 5.7. Gene constructs of procoat-OmpA 828 wild-type, OM30R, A21-33 and A21-39 ...... 132

5.8. The Membrane anchor region in procoat 828 and its mutants is not critical for the cleveage ...... 133

5.9. Deletion of the second transmembrane domain wihtin procoat mutants does not impair the assemblyacross the plasma membrane ...... 135

5.10. Inverted leader peptidase mutants R70 and R72 are not stable in SecA,s (CJ105) and SecY,s {CJ107) £- strains ...... 13 7

5.11. Instability of R70 and R72 in SecAts and SecYts strains and Sec-independent export of the inverted leader peptidase mutant R72 ...... 138

x i i i LIST OF ABBREVIATIONS

A pr ampicillin resistance ATP adenosine 5'-triphosphate Bis N, N'-methylene-bis- a cryla m ide bp base pair(s) BCIP bromochloroindolyl phosphate dATP 2'-deoxyadenosine 5’-triphosphosphate dATPaS 2’-deoxyadenosine 5'-[a-thio]-triphosphate dCTP 2'-deoxycytosine 5'-triphosphate dGTP 2’-deoxyguanosine 5'-triphosphate dTTP 2’-deoxythymindine 5'-triphosphate DTT dithiothreitol DMF dimethylformamide EDTA ethylenediamine tetraacetic acid EtBr ethidium bromide IPTG isopropyl-13-D- thiogalactoside kb 1000-base pairs KD 1000 Daltons LB Luria broth LPase Leader peptidase MBP maltose binding protein NBT nitroblue tetrazolium chloroide OmpA outer membrane protein A PAGE polyacrylamide gel electrophoresis

x i v PGG polyethyleneglycol PMSF phenylmethylsulfonyl fluoride RF replicative form SB sample buffer SDS sodium dodecyl sulfate Sec secretion TCA trichloride acetate Tet tetracycline TLCK L-1-tosolyamido 2-lysine chloromethyl ketone TPCK L-1-tosolyamido 2-phenyl ethyl chloromethyl ketone Tris 2-a m ino-2-( hydroxy me thy I)-1,3 propanediol UTP uridine 5'-triphosphate

xv CHAPTER I

INTRODUCTION

1.1 The Role of Charged Amino Acid Residues in Membrane Protein Topology. Cells, both prokaryote and eukaryote, are composed of different subcellular organelles. In eukaryotic cells, these organelles are the nucleus, mitochondria, chloroplast (in plant cells), endoplasmic reticulum, peroxisome, lysosome, Golgi apparatus and cytoplasmic membrane (Fig. 1.1). In prokaryotic Gram-negative organisms, the subcellular structures are the inner membrane (plasma membrane), periplasm and outer membrane; in Gram-positive organisms, there is only a cytoplasmic membrane (Fig. 1.2.).

Among the subcellular organelles, the cytoplasmic membrane has the function of not only confining a cell to a certain size and shape, but also being a barrier or exchanger between the interior of the cell and its environment. In order to carry out these functions, the membrane usually has a unique composition of lipids and proteins. The lipids form a fluid bilayer which comprise a major part of the membrane and exist 1 2

Endocytosis

7 Peroxisome Secretory© & Vesicle

/ Outer memb Secretory Lysosome Nucleus Granule

Golgi Complex memo matrix (ntermemb. Mitochondria space. Outer memb stroma ntermemb thylokoid space inner lumen Rough thylokoid memb Endoplasmic memb. Reticulum Chloroplast

Figure 1.1. Subcellular organelles of eukaryotic cells and the protein traffic pathways within the cell. • nucleus-encoded proteins and ■ mitochondrial or chloroplast-encoded proteins. 3

Integral outer membrane protein

Outer = . membrane * peripheral membrane protein Integral plasma membrane f ^ 0 ytoplSmic protein Cytosolic protein /M e m S n e

**-— Periplasmip, ^ ^ - space

Figure 1.2. Principal locations of bacterial membrane proteins. The diagram is representive of a gram-negative bacterium. In the case of gram-positive bacteria, there is no definite outer membrane lipid bilayer. 4 in a taii-to-tail fashion with their head groups facing the outside of the bilayer {Fig. 1.3.). To date, many of the biological functions of the lipids remain unknown. Membrane proteins, synthesized in the cytoplasm, are localized in the membrane either as integral or peripheral forms (Fig. 1.3.). Although peripheral proteins associate with the membrane in a loose and dynamic fashion, integral proteins assume their unique asymmetric topologies embeded within the membrane. This phenomenon raises several questions. How do membrane proteins insert into the membrane during or after they are synthesized in the cytoplasm? Where is the information within the polypeptide that determines its functional topology? The former question can be described as protein export which will be discussed in one of the following sections and the latter is the focus of this section.

In recent years, many lines of evidence have suggested that the information required for proper topology of membrane proteins resides with the distribution of charged amino acid residues that lie on either side of transmembrane segments of the proteins (Dalbey, 1990; Model and Russell, 1990; Boyd and Beckwith, 1990). After analysis of over a hundred membrane proteins from prokaryotes and eukaryotes, von Heijne proposed a "positive inside rule" which states that hydrophobic transmembrane segments orient themselves in the membrane with the positively charged amino acids facing the cytoplasm Integral proteins

Peripheral proteins

Figure 1.3. Schematic illustration of integral and peripheral membrane proteins. Peripheral membrane proteins may be removed under certain conditions without disrupting the membrane structure. Integral membrane proteins, in contrast, can be released only if the membrane is disrupted. 6

(von Heijne, 1986; von Heijne and Gavel, 1988). While this rule is not absolute, exceptions do exist especially in eukaryotes, experimental and statistical analysis of membrane proteins with known or predicted structure indicates that it represents a good generalization. Figure 1.4. shows some membrane proteins depicting their positively charged amino acid residues flanking the hydrophobic segments (Fig. 1.4.).

The central role of positively charged amino acid in controlling the topology of membrane proteins has been demonstrated for many proteins. All prokaryotic and most eukaryotic leader peptides which function to initiate export contain positively charged residues at their amino terminus. Alteration of these amino acids indicated that the basic residues are necessary for efficient protein translocation (Inouye fiia l, 1982; Vlasuk filaL, 1983; Lino fiia i, 1987; Puziss Alai., 1989). The net charge at the amino terminus of the mature sequence of bacterial exported proteins is also important for their translocation. Addition of positively charged residues to the N-terminus of the mature region of several proteins blocked protein export (Li a ia i , 1988; Yamane and Mizushima, 1988; Laws and Dalbey, 1989; Summer et al.. 1989; Zhu filai., manuscript in preparation). Gene fusion studies with alkaline phosphatase in £. coli have revealed that the basic amino acids are important in anchoring cytoplasmic 7

Periplasm

Inner Membrane

Cytoplasm MALF Procoat Leader Peptidase

Figure 1.4. Model of the membrane orientations of three proteins in the plasma membrane of £. coil* 8

domains of membrane proteins in their normal orientation (Boyd and Beckwith, 1989). More dramatically, the critical role of positively charged residues has been shown by "flipping" the orientation of an individual (von Heijne si at, 1988, Haeuptle si ai-» 1989; Parks siai-, 1989; San Miilan si si., 1989; Szczena- Skorupa and Kemper, 1989; Sato slai-. 1990) or pairs (von Heijne, 1989; Nilsson and von Heijne, 1990) of transmembrane segments. This was achieved by moving positively charged amino acid residues from one end of the transmembrane region to the other or by exchanging polar domains flanking the hydrophobic regions.

Even though how positively charged amino acid residues determine membrane topology remains unknown, several possible mechanisms have been suggested (Fig. 1.5.). For example, Gallusser and Kuhn proposed that the positive charges at the amino- and carboxyl-terminal regions of the M13 procoat protein electrostatically bind to the head groups of the acidic lipids within the membrane (Gallusser and Kuhn, 1990). In this way, they would anchor both regions of the protein on the cytoplasmic surface (Fig. 1.5.B). Furthermore, recent results raised the possibility that the basic residues in the leader peptide as well as at the amino terminus of mature protein interact with acidic residues within the components of the export machinery (Akita M ai. 1990; Zhu fiiai, manuscript in 9

Periplasm

Plasma membrane

+ + Cytoplasm

Periplasm

Plasma membrane

Cytoplasm

Figure 1.5. Model to explain why positively charged amino acid residues are determinants for membrane protein topology. (A) Positively charged residues in a typical exported protein. (B) Altered exported protein with positive charges on both sides of N- and C- terminal region interact with the head groups of lipid within the membrane and anchor both regions in cytoplasm. (C) High pKa of basic amino acids prevents insertion of the hydrophobic region. (D) Alteration of positive charge distribution results in an inverted orientation. 1 0

preparation). On the other hand, the high pKa of the positively charged residues, especially arginine, would make their translocation energetically more costly if they had to be neutralized before the export can be initiated (Fig. 1.5.C; Summer fiia l, 1990). The last obstacle for positively charged residues penetrating the membrane is the unfavorable distribution of electrical charge surrounding the membrane. For instance, the electrochemical potential across the membrane is negative inside which makes it favorable for basic residues to stay in the cytoplasm. As expected, the addition of positive charges to the amino terminus of the mature region blocks the normal membrane translocation (Fig. 1.5.D; Li siai., 1988; Yamane and Mizushima, 1988; Zhu fiiai-, manuscript in preparation).

While in prokaryotic systems the "positive-inside rule" fits quite well, in eukaryotes, it is still an open question as to what is the determinant for the orientation of the transmembrane segments of membrane proteins. However, several lines of evidence suggest that positively charged residues may play the role here as well. In proteins with single transmembrane regions, the influence of positively charged residues have been revealed by either exchanging the amino or carboxyl terminus of the hydrophobic signal sequences of a secretory protein or by inserting additional basic residues at the 11 amino-terminal side which inverts the orientation of the transmembrane segment (Haeuptle siaL, 1989; Szczesna- Skorupa filaL, 1988). For multispanning eukaryotic membrane proteins, it has been suggested that the charges flanking the first transmembrane region play a critical role in the membrane protein topogenesis and are the deciding factors for the topology of other transmembrane segments since they insert sequentially following the lead of the first one (Hartmann filai-, 1989). This proposal does not explain, however, why von Heijne sees the same charge distribution bordering transmembrane segments in most eukaryotic membrane proteins (von Heijne and Gavel, 1988).

1.2. Protein Subunits of the Photosynthetic Reaction Center from Rhodobacter sbhaeroides Rhodobacter sphaeroides is a typical gram-negative bacterium and it is grown chemoheterotrophically in the presence of oxygen. When the oxygen tension is reduced below threshold levels, the bacterial cell responds by synthesizing an extensive intracytoplasmic membrane system (Drews and Oleze, 1981). This new membrane houses the numerous components of the photosynthetic apparatus including the key complex in this photosynthetic bacterium, the photosynthetic reaction center. The synthesis of the components in this reaction center is modulated by light intensity. The membrane biosynthesis and 1 2 photosynthesis have been studied at the molecular level (Kaplan, 1978; Kaplan and Arntzen, 1982). The reaction center uses light energy in the formation of high-energy electrons, and it is these electrons which are the driving force of photosynthetic electron transport. The photosynthetic reaction center is an integral membrane protein complex, which contains a number of cofactors that mediate the primary photochemistry. There are three protein subunits in the reaction center. There are L, M and H having 281, 306 and 260 amino acid residues respectively (Williams filai., 1983; Williams fiiaL, 1984; Williams aial-, 1986). The cofactors are four bacteriochlorophylls, two bacteriopheophytins, two quinones, and one iron. Following the solution of the x-ray crystallographic structure of the photosynthetic reaction center from Rhodopseudomonas viridis (Deisenhofer a ia l, 1985), the three dimensional structure of the reaction center from Rhodobacter sphaeroides was also solved (Allen ai al,, 1987a; Allen aial-. 1987b; Yeates alal-. 1987; Feher aial-. 1989; Fig. 1.6.). The structures of the reaction center from both species are similar with only minor differences (Feher aial*. 1989). The cofactors of the reaction center from Rhodobacter sphaeroides are arranged along two branches which are approximately related to each other by a two-fold symmetry axis (Allen aial-. 1987a). The protein subunits L, M and H are organized as LM complex and H subunit. The L and M proteins both have five transmembrane segments Figure 1.6. Three dimensional x-ray crystallographic structure of the photosynthetic reaction center from Rhodobacter sphaeroides. 1 4 forming the complex with two-fold rotational symmetry axis that approximately relates the two subunits. The secondary structure of the H subunit is very different from that of the L or M subunits. The H subunit has only one transmembrane helix that starts on the periplasmic side near the amino terminus. The bulk of the H subunit lies on the cytoplasmic side which forms a globular domain with one nontransmembrane helix. Extensive contacts of the H subunit with LM complex stabilize the reaction center structure (Allen sial-, 1987b; Fig 1.6). Integral membrane complex LM binds to the cofactors while H does not interact with them. Comparison of the reaction center primary structures from different Rhodobacterial species revealed that the similarity between the three-dimensional structures of the reaction center is consistent with the similarity of the amino acid sequences (Allen fiia l, 1987b). Besides the detailed structural studies described here, the molecular biology aspect of the reaction center from Rhodobacter sphaeroides has been carried out extensively by Kaplan and colleagues (Kiley and Kaplan, 1988). The cDNAs that code for the three protein subunits have been cloned and expressed in both Rhodobacteria and E. coli cells (Williams fila!-, 1983, Williams fiia l, 1984, Williams fiia l, 1986; Donohue filai-, 1986; Youvan filai. 1984). The use of gene manipulation techniques including site-directed mutagenesis have revealed how the interactions occur between the protein subunits and which part of thees proteins is critical 1 5 for the photosynthetic function (Hoger and Kaplan, 1985; Sockett fiia l, 1989). But how the H subunit assume its unique membrane orientation (i.e. only one transmembrane helix crosses the membrane with very small fragment facing periplasm while the bulk part of the protein stays in the cytosol) remains unknown.

1.3. Protein Export in £ . coli Most export or secretory proteins in both eukaryotes and prokaryotes reach the cell membrane or outside medium via the export pathway. In eukaryotic cells, proteins enter the secretory pathway at the rough endoplasmic reticulum membrane and travel through the Golgi apparatus to the cell surface or to the secretory (storage) granules, the lysosomes, or the vacuoles (Fig. 1.1.). In prokaryotes, the translocation of the proteins typically occurs with the help of the export machinery (see below) and ends up in either the inner or outer membrane, the periplasmic space, or secreted out of the cells (Fig. 1.2.). More strikingly, most of these proteins are synthesized in a higher molecular weight precursor form with a cleavable leader (or signal) peptide. The necessity of the leader peptide for protein export was first observed by Milstein and colleagues (Milstein fiia l. 1972), who showed that the secreted immunoglobin light chains are shorter than the precursor proteins synthesized in vitro. The shorter protein was found to be a processed mature polypeptide and the leader peptide was 1 6

removed by a leader or signal peptidase in the cell (Blobel and Sabatini, 1970, 1971; Blobel and Dobberstein, 1970, 1971). The function of leader peptides was then shown to interact with the export machinery in both eukaryotes (Blobel 1980; Walter fiia l., 1984, 1986) and prokaryotes (Wickner fit fii.,1991). Although there is no sequence homology among the leader peptides of known exported or secretory protein precursors, there are several common features within the leader peptides: (1) all cleavable leader peptides are located exclusively in the N- terminus of the precursors; (2) they range in length from 13 to 36 amino acid residues; (3) there is at least one positively charged amino acid at their N-terminus; (4) a unique hydrophobic central core region is composed of 9 to 15 amino acids that have a strong tendency to form a helix across the membrane; and (5) there are small non-charged amino acid residues at -1 and -3 relative to the cleavage site.

In most cases, the leader peptide is necessary but not sufficient for protein export. By altering the amino acid residues in the N-terminus of the mature region of the exported protein, Knowles and colleagues demonstrated that a certain feature of the mature region is also required for efficient translocation. Substitution of a positively charged amino acids within the first 14 amino acids immediately following the leader peptide abolishes the protein export (Summers and 1 7

Knowles, 1989; Summers fiial*. 1989). The primary structural requirement of the mature region for export is still not clear. However, this region may be involved in the binding of a cytosolic protein, SecB which is a component of export machinery (Randall fiial*, 1990; Collier and Bassford, 1989; Collier fiial*, 1990).

Besides the requirement for a leader peptide and a compatible mature region, a special protein export machinery is often involved in translocation. The export apparatus contains cytosolic proteins as well as membrane components, In the past decade, the secretory apparatus and its components have been characterized in both higher and lower eukaryotes (Rapoport, 1986; Walter and Lingappa, 1986; Pfanner and Neupert, 1987; Pfaller fiial., 1989) and they are beyond the scope of the discussion in this section. On the other hand, with the help of newly developed in vitro transcription, translation and translocation systems, the export apparatus in the prokaryote, £. coli. has been elucidated in much more detail than its eukaryotic counterpart (Wickner and Lodish, 1985; Dalbey fiiai*. 1987; Model and Russell, 1990; Wickner a ia l. 1991). Two bacterial exported proteins, maltose-binding protein (MBP) and outer membrane protein A (OmpA) have been chosen as model proteins for these export studies. Early studies done by Randall and colleagues found that protein translocation often occurs after 1 8

the newly synthesized precursor reaches "critical molecular weight" and a loosely folded conformation of the preprotein (i.e., unfolded) is required for the export (Randall, 1983; Randall and Hardy, 1986). More recently, based on work with the pro-OmpA protein, Wickner and colleagues proposed a model of protein export in E coli. which is catalyzed by two proteins, a soluble chaperone and a membrane-bound translocase (Wickner fiial., 1991; Fig. 1.7.). SecB, the major chaperone for export, interacts with mature region of preproteins to form a complex for translocation (Collier fiia l, 1988). The translocase consists of functionally linked peripheral and integral membrane protein domains. The SecA protein is the peripheral membrane subunit domain of the translocase (Oliver and Beckwith, 1982; Cabelli fii ai., 1988), and is the primary receptor for the SecB/preprotein complex (Cunningham and Wickner, 1989). SecA hydrolyzes ATP, promoting cycles of translocation, protein release, Ap (transmembrane electrochemical potential)-dependent translocation, and rebinding of the preprotein to SecA (Lill fil ai., 1989, 1990; Hartl filai-, 1990; Schiebel filai-, 1991). The membrane-embedded domain of the translocase is the SecE/Y protein with three subunits, SecE, SecY and protein I polypeptides. SecE/Y protein stabilizes and activates SecA and participates in translocation (Brundage filai-, 1990; Hartl filai-, 1990). It is noteworthy that there is a requirement for energy in the process of export (Chen and Tai, 1985; Yamada filai-, ^ 'T y Leader + Peptidase

Figure 1.7. pro-OmpA protein translocation via the export machinery. Hatched region represents the leader peptide in pro-OmpA. 20

1989; Yamane filai-, 1987). There are two consecutive energetically distinct steps in the translocation (Geller and Green, 1989). ATP is involved in the early stage by providing the energy source to initiate the insertion of the amino terminus of the preprotein and exposing the leader peptidase cleavage site to the periplasmic space. The second step, which completes the transfer of the protein across the membrane, is triggered by either ATP hydrolysis or the transmembrane electrochemical potential (Wickner fiial-, 1991; Schiebel filai-, 1991). Although there is some speculation for the role of the electrochemical potential in the export, the details of how it promotes the preprotein translocation still remain unknown.

Although conditionally lethal mutants in the sec genes affects the secretion of most bacterial proteins, certain small proteins such as M13 procoat (Wotfe filai-. 1985), Pf3 coat protein (Wickner filai*. 1991) and honeybee prepromelittin (Cobet filai-. 1989) do not require the Sec proteins for translocation. The best example of these is M13 procoat protein, the first preprotein with a typical leader peptide found not to require Sec proteins for membrane insertion (Wolfe filai*. 1985). The cascade of Sec-independent membrane assembly of M13 procoat protein is illustrated in figure 1.8. It was speculated that interaction between the two apolar regions of procoat and the three charged segments of the protein may function together 21

T T T T T l INNER MEMBRAN ' m.

Hydrophobic ArfmUfll _ . porNfi*o C'tOvo,. location L„^^ EttcFroifofa Ending Phag« o«*(nb(^

Figure 1.8. Membrane integration of M13 phage procoat protein via Sec-inedpendent pathway. 22

to shield the apolar surface and stabilize a insertion competent conformation prior to membrane interactions (Wickner aial., 1991). But the common requirement of a membrane electrochemical potential in both Sec-dependent and Sec- independent export pathways makes the theme more complicated.

1.4. E. coli Leader Peptidase The other central enzyme in the protein export machinery of coli is leader peptidase. The function of leader peptidase is to release the translocated or preproteins which are bound to the outer surface of inner membrane by removing the leader peptides. The mature proteins can then reach their destination in the periplasm, outer membrane or outside the cell (Dalbey and Wickner, 1985; Fikes and Bassford, 1987). There are two major leader peptidases in coli. lipoprotein leader peptidase that cleaves only lipoproteins (Tokunaga filai-, 1982) and leader peptidase that acts on a wide variety of preproteins (Watts fil a t, 1983).

The leader peptidase in £.. coli was first found to be involved in the processing of M13 procoat protein to coat protein (Chang filai-, 1978). Since then the gene was isolated (Date and Wickner, 1981), cloned and sequenced (Wolfe filai-. 1983). The protein was also purified to homogeneity from an overproducing 23

bacterial strain (Wolfe fiia l, 1982). Leader peptidase consists of single polypeptide chain with a molecular mass of 37,000 daltons. The membrane topology of leader peptidase is shown in figure 1.9. The first two hydrophobic domains (H1 and H2) of the leader peptidase span the inner membrane twice with the third hydrophobic segment (H3) and a long carboxyl terminus facing the periplasm (Wolfe filai., 1983; Moore and Miura, 1987; San Millan fiia l, 1989). Its membrane assembly has been studied extensively. The mechanisms of the insertion of the two transmembrane domains differ from each other (Dalbey and Wickner, 1986; Dalbey fiia l, 1987). H2 is an internal uncleavable signal peptide which is required for the insertion of the large carboxyl terminal domain (Dalbey and Wickner, 1987) via Sec protein- and membrane electrochemical potential- dependent mechanism (Wolfe and Wickner, 1984; Wolfe fiial-, 1985; Lee and Dalbey, unpublished data). More recently, Dalbey and colleagues demonstrated that the insertion of H1 may occur spontaneously without requiring either the Sec proteins or the transmembrane electrochemical potential (Lee fiia l, 1992).

Positively charged amino acid residues also play an important role in determining the membrane topology of the leader peptidase. The charge distribution along the polypeptide obeys the "positive inside" rule (von Heijne 1986). A highly positively charged region (residues 30-52) was defined as a Periplasm

Inner Membrane

Cytoplasm

Leader Peptidase

Figure 1.9. Current model of membrane topology of leader peptidase. 25

"translocation poison" domain because this region prevents the translocation of the large carboxyl-terminal domain (von Heijne fiia l, 1988). In addition, the orientation of H1 can be reversed by having positively charged amino acids in front of it (Law and Dalbey, 1989; San Millan fiial-. 1989). More interestingly, when this "translocation poison" domain is substituted by a neutral spacer and a positive charge cluster is inserted in front of H1, the membrane orientation of the leader peptidase is reversed {von Heijne, 1989). Four lysines added to N-terminus of the protein promotes an Njn-Cjn topology and a lysine added between H1 and H2 endorses Nou(-Cout topology (Nilsson and von Heijne,

1990; Figure 1.10.). The successful construction of inverted leader peptidase also raises the possibility to study a mechanism similar to that of the M13 procoat protein because of the similarity of the topology between the procoat and the inverted leader peptidase. 26

P2

Periplasm

H1 I_I2 Cytoplasmic H1 H2 Membrane

Cytoplasm P1

Wild Type Inverted

Figure 1.10. Membrane topology of inverted leader peptidase. CHAPTER II

MATERIALS AND METHODS 2.1. Materials For cell growth media, bacto-tryptone, bacto-agar, yeast extract and casamino acids were purchased from Difco Laboratories. All salts, arabinose, thymine (Vitamin B) and all 20 amino acids were from Sigma Chemical Company. Fructose and glucose were also from Sigma.

For SDS-polyacrylamide gel electrophoresis and fluorography, urea, sodium sulfate (reagent grade) and B-mecaptoethonal were obtained from Aldrich Chemical Company. Sodium dodecyl sulfate (SDS), acrylamide, N,N'- methylene-bis-acrylamide, sodium persulfate (APS), ultra pure TEMED, bromophenol blue and Coomassie brilliant blue G-250 were from Bio-Rad Laboratories. 2-amino-2-(hydroxymethyl)- 1,3-propanedio! (Tris) and glycine were also from Sigma.

The nucleotides, ATP, UTP, dATP, dGTP, dCTP, and dTTP were purchased from Pharmacia. Dithiothreitol (DTT), phenylmethytsulfonyl fluoride (PMSF), ampicilin (Amp) and tetracycline (Tet) were from Sigma. Pansorbin ceils (StaphA) 27 28 for immunoprecipitation was from CalBiochem. Isopropyl-B-D- thiogalactopyranoside (IPTG) was purchased free of dioxane from Research Organics, Inc.. Sucrose was from ICN Biochemicals, p-nitrophenyl phosphate (Sigma 104), L-1- tosylamido 2-phenyl ethyl chloromethyl ketone (TPCK) and L-1- tosylamido 2-lysine chloromethyl ketone (TLCK) were from Sigma Chemical Inc. Trans [35S]-label, a mixture of 85% [35Sj methionine and 15% [3SS] cysteine, 1000 Ci/mmol, and [35S] dATP (5000Ci/mmol) were from ICN K&K Laboratories Inc. and Du Pont-New England Nuclear, respectively.

Trypsin (L-1 -tosylamido 2-phenyl ethyl chloromehtyl ketone treated) and soybean trypsin inhibitor were obtained from Worthington. Hen egg white lysozyme (54,000 units/mg) was from Sigma. DNA polymerase I (Klenow fragment) and alkaline phosphatase (bovine intestine) were from Boehringer Mannheim. All DNA restriction enzymes, T4 DNA kinase and T4 DNA ligase were from Bethesda Research Laboratories, Promega and New England Biolabs. DNA Sequenase 2.0 kit and its components were purchased from United State Biochemicals. All oligonucleotides used for site-directed mutagenesis were synthesized in the Biochemical Instrument Center at The Ohio State University. 29

For Western blot, the Protoblot kit (Rabbit anti-IgG conjugated with alkaline phosphatase) was purchased from Promega. Nitroblue tetrazolium chloride (NBT), bromochloroindoly! phosphate (BCIP), Tween-20 and dimethylformamide (DMF) were from Sigma.

For desalting and partial purification of oligonucleotides, and isolation of restriction fragments after digesting plasmid DNA with a restriction enzyme, Sephadex G-50 and G-25 were obtained from Pharmacia.

2.2. Bacterial Strains, Media and Plasmids E. coli strains JM101 ( (lacpro) thi/P, supE, traD36, proAB, laclq ZM15), MC1061 { lacX74, araD139, (ara, leu)7697, galU, galK, hsr, hsm, strA), HJM114 {( lac-pro)/F (lac-pro)) and RZ1032 (ung-, dut) are from our collection. CJ105 (HJM114, SecAts, leu82::Tn10) and CJ107 (HJM114, SecYts) were obtained from Dr. William Wickner in the University of California, Los Angeles. CK1953 (F", lacU169, araD139, rbsR, rpsL, thiA, relA, motA, zhe::Tn10, malTc, SecB::Tn5) was obtained from Dr, Carol Kumamoto in Tufts University. CC118 (araD139(ara, leu) 7697 lacX74, galK, thi, rpsE, rpoB, argE, phoA20, galE, recA1) used for expression and assays of alkaline phosphatase was from Dr. Robert Gennis in the University of Illinois, Urbana. 30

For expressing plNG-1 encoded proteins in £.. coli. M9 minimal salt media was prepared as described in Miller (1972) and was supplemented with 0.5% fructose, 50 pg/ml of each amino acid except methionine and 0.5% arabinose. For expressing of subunit H protein of the photosynthetic reaction center from Rhodobacter sphaeroides in E. coli. the M9 media was supplemented with 0.4% succinic acid and 0.06% casamino acids (Dr. Samuel Kaplan, personal communication). For assays of alkaline phosphatase activity, aerobic rich medium or LB broth (Maniatis filai. 1982) was used as described (Chepuri and Gennis, 1990)

The pRBHL19 plasmid is a derivative of pUCl9 and carries the full length cDNA of subunit H protein of the photosynthetic reaction center (Sockett filai-, 1989). This plasmid was a gift from Dr. Samuel Kaplan (University of Texas, Health Science Center at Houston). The pING plasmid (Johnston fila i , 1985) which contains the arabinose regulatory elements and the arabinose promoter was obtained from Dr. Gary Wilcox (Ingene,

Inc.).

2.3. Special Growth of the Cells Carrying Subunit H Protein of the Photosynthetic Reaction Center from Rhodobacter sphaeroides 31

£. coli cells bearing pRBHL19 plasmid encoding subunit H protein of the photosynthetic reaction center was inoculated onto 10 to 20 ml of M9 medium supplemented with 0.4% succinic acid, 0.06% casamino acids and 50 pg/ml ampicilin and the culture was growing in a 37°C waterbath overnight. The saturated culture was back diluted to fresh medium to 0.05 absorption units at a wavelength of 600 nm. The expression of subunit H protein was induced by adding 0.2 pM IPTG. The cells was grown at 37°C either overnight or until the density of the culture reached about 1.45 O.D.6Q0. The cells then was harvested and used for preparation of cell membranes.

2.4. Preparation of E. coli Cell Membrane Fractions The overnight grown cells were chilled quickly in an ice/water bath. Protease inhibitors, TLCK and TPCK, were added to a final concentration of 0.5 mM. After chilling, cells were harvested by centrifuging in a pre-cooled (4 to 6°C) rotor at 10,000 rpm for 10 minutes. The pellets from 500 ml cell culture were washed in 100 ml of ice-cold 10 mM Tris, 1 mM Na4EDTA, pH 8.0, containing 0.5 mM TPCK and TLCK. The cells were spun down and resuspended in 10 ml of buffer A (20% sucrose (weight/volume), 50 mM Tris, 5 mM Na4EDTA, pH 8.0, 0.5 mM TPCK and TLCK). Fresh lysozyme was added to final concentration of 0.2 mg/ml and the cells was stirred on a 32 ice/water bath for about 30 minutes. The treated cells were centrifuged in a pre-cooled rotor at 10,000 rpm for 10 minutes to pellet newly formed spheroplasts. The spheroplasts were washed with ice-cold buffer A and resuspended in buffer B (10 mM Tris, 1 mM Na4EDTA, pH 8.0, 0.5 mM TPCK and TLCK). Then the spheroplasts were sonicated three times for 5 minutes in ice bath at the setting 6 to 8 in a sonicator. Care should be taken at this step to keep the sonicated spheroplasts cold by cooling them at certain time intervals. The broken cells were pelleted by centrifuging them 10 minutes at 10,000 rpm in a pre-cold rotor. The cell membrane fraction then was washed with TE, pH 8.0 once in the presence of TPCK and TLCK and the unbroken cells were removed by centrifuging at 8,000 rpm for 5 minutes. The fraction was then resuspended in TE, pH 8.0 and spun at 40,000 rpm, 4°C for 2 hours. The pelleted membrane fraction was resuspended in 500 pi to 2 ml TE, pH 8.0. Protein concentration of the membrane fraction was measured by a modified method (Peterson, 1977) and the expression of the subunit H protein of the photosynthetic center was analyzed by immunoblotting (see section 2.20.) after SDS-polyacylamide gel electrophoresis. 33

2.5. Alkaline Phosphatase Activity Assay Alkaline phosphatase activity of PhoA-puhA (subunit H gene) fusion proteins (see Chapter 111) was determined as described by Brickman and Beckwith (Brickman and Beckwith, 1975) with the following modifications (Chepuri and Gennis, 1990). An overnight culture of E. coli strain CC118 containing subunit H-PhoA fusion protein or controls was used to inoculate 10 ml of LB medium and grown to the early exponential phase. 1 ml of cells was centrifuged, washed, and resuspended in 1 ml of 1 M Tris, pH 8.0. One drop of 0.1% SDS and 2 to 3 drops of chloroform were added to the cell suspension, which was then incubated at 37°C in a waterbath for 5 minutes to permeabilize the cells. After incubation, p-nitrophenyl phosphate (substrate for alkaline phophatase, Sigma 104) was added (final concentration of 0.4 mg/ml), and the conversion of p-nitrophenyl phosphate to p-nitrophenol was monitored as a function of time in a spectrophotometer by measuring the absorbance change at 420 nm. The alkaline phophatase specific activity was determined by the change in absorbance at 420 nm. It is expressed in units of nm/min/optical density of the cell suspension measured at 600 nm. This result is then multiplied by a factor of 1000. 34

2.6. Reverse Protease Mapping of Subunit H Protein Expressed in coli. The protease digestion of subunit H protein expressed in E coli was done by the following method (Samuel Kaplan, personal communication). After isolation of the membrane from HJM114 cells expressing the subunit H protein (see section 2.4.), the membrane vesicles were formed by washing with ice-cold 20% sucrose, 50 mM Tris, 5 mM Na4EDTA, pH 8.0 once and the spheroplasts were collected by spinning in an Eppendorf microcentrifuge and resuspended in the same solution. The proteases, trypsin (TPCK treated) and chymotrypsin (TPCK and TLCK treated) were added to a final concentration of 1% (weight/weight) and the mixture was incubated on ice for 2 and 10 minutes, respectively. A control was also done as described above on sonicated samples which were sonicated for 5 minutes three times as described before with periods of cooling. The treated samples were then analyzed on a 8%-25% gradient SDS- polyacrylamide gel electrophoresis.

2.7. Alkali Fractionation To determine the membrane location, i.e. whether the protein is located in a peripheral or a transmembrane topology, alkali fractionation analysis was done as described by Davis ei al (1985). HJM114 cells containing pRBHL19 were grown at 37°C 35 in M9 medium supplemented with 0.4% succinic acid and 0.06% casamino acids to a density of about 1.45 optical density at 600 nm wavelength. After isolation of membrane fraction, 1 ml ice- cold 0.1 M NaOH was added into the tube and samples were then subsequently vortexed vigorously and centrifuged at 4°C in an Eppendorf microfuge for 15 minutes. The supernatant was removed and precipitated with 0.15 ml of ice-cold 100% trichloroacetic acid. 1 ml of ice-cold 5% trichloroacetic acid was added to rinse the pellet. Both of the trichloroacetic acid treated fractions were incubated on ice for 10 minutes. Precipitates were spun down at 4°C for 10 minutes in an Eppendorf microfuge. Pellets were then washed with 1 ml of ice-cold acetone, and centrifuged at room temperature (25°C) for 2 minutes . After removing the acetone by aspirating, protein pellets were dissolved by heating at 100°C for 5 minutes in 150 pi of TE buffer and analyzed on a gradient SDS polyacrylamide gel electrophoresis (see section 2.8.).

2.8. SDS-PAGE Gel Electrophoresis and Fluorography SDS-PAGE gel electrophoresis was used to analyze leader peptidase, OmpA, procoat and their fusion proteins as described by Ito fiia l (1980). Radioactive polypeptides were visualized by fluorography (Chamberlin, 1979). Prior to electrophoresis, samples were suspended in 30 pi of sample buffer (SB, 10 mM 36

Tris-HCI, pH 8.0, 1 mM EDTA, 2% SDS, 5% B-mecaptoethanol, and 0.001% (w/v) bromophophenol blue) and heated at 95°C for 5 minutes to ensure complete denaturation of proteins. One half of each samples (15 pi) was loaded on a pre-prepared SDS polyacrylamide gel (19% for leader peptidase, OmpA, or their fusion proteins; 23% for procoat protein only, Table 2.1., 2.2.) and run at 30-40 mA for 2 to 3 hours in a running buffer (0.19 M glycine, 0.024 M Tris, pH 7.5 and 0.0034 M SDS). After electrophoresis, the gel was fixed in a fixing solution (50% methanol, 7% acetic acid, v/v) for 45 minutes and then treated with a sodium salicylate solution (0.9 M salicylic acid and 0.9 M NaOH) for 15 minutes. The gel was dried in a vacuum dryer and exposed on a piece of Kodak x-ray film (for fluorography).

SDS-PAGE analysis for subunit H protein and its respective fusion proteins of subunit H and alkaline phophatase was done according to the modified method of Kaplan (Kaplan, personal communication). Protein samples were treated with sample buffer (0.17 mM Tris-HCI, pH 8.0, 200 pg/ml glycerol, 160 mg/ml LiDS and 0.1 mg/ml bromophenol blue) either at room temperature for 30 minutes and at 37°C for 10 minutes. The SDS polyacrylamide gels were prepared as described in Table 2.3. and 2.4. Electrophoresis was performed at 4°C in the running buffer listed above at 20 to 40 mA for 3 to 6 hours. Then the gel was 37

Table 2.1. Preparation of SDS-PAGE Gel (With Urea)

Separating Gel (19%) Acrylamide (60%) 4.8 m I Bis (1%) 1.1 m I 1.0 M Tris-HCI (pH 8.8) 5.0 m l

Urea 5.4 9 10% SDS 50 p i 10% APS 50 p i TEMED 5 P I

Stacking Gel (5%, X2) Acrylamide(30%) 1.67 m I Bis (1%) 1.3 m I 1 M Tris-HCI (pH 6.8) 0.625 ml Urea 3.6 g h 2o 3.68 m I

10% SDS 100 p l 10% APS 50 p i TEMED 10 P i 38

Table 2.2. Preparation of SDS-PAGE Gel (With Urea, for procoat protein)

Separating Gel (23%) Acrylamide (60%) 7.65 m I Bis (2%) 0.88 m l 2.0 M Tris-HCI (pH 8.8) 4.0 m I 1 M NaCI 1.8 m I

Urea 7.2 9 10% SDS 200 p l 10% APS 55 P i TEMED 5.5 p l 39

Table 2.3. Preparation of SDS-PAGE Gel (Without Urea)

Separating Gel (12%) Acrylamide/Bis (37.5:1) 4.5 m l 1.5 M Tris-HCI (pH 8.8) 3.75 m I

h2o 6.52 m I

10% SDS 150 Ml 10% APS 75 Ml TEMED 10 Ml

Stacking Gel (3%) Acrylamide/Bis (37.5:1) 0.75 m I 0.5 M Tris-HCI (pH 6.8) 2.5 m I h 2o 6.3 m I

10% SDS 100 Ml 10% APS 33 Ml TEMED 10 Ml 40

Tabel 2.4. Preparation for SDS-PAGE Gel (For Gradient Gel without Urea)

Separating Gel (8%) Acrylamide/Bis (37.5:1) 4.6 m I 3 M Tris-HCI (pH 8.8) 5.62 m I h 2o 12.2 m I

10% SDS 2 2 5 P I 10% APS 1 1 3 p l TEMED 15 p l

Separating Gel (25%) Acrylamide/Bis (37.5:1) 13.5 m I 3 M Tris-HCI (pH 8.8) 5.63 m I h 2o 3.0 m I

10% SDS 2 2 5 p l 10% APS 1 1 3 p l TEMED 15 P l 4 1

subjected to Coomassie brilliant blue staining according to published protocols or protein transblot for immunoblotting (section 2.20.).

2.9. Agarose Gel Electrophoresis 0.7%, 1% or 1.2% agarose gels were prepared for DNA analysis by dissolving appropriate amounts of ultra pure agarose in 1X TBE buffer in a microwave oven. After the solution was cooled down to about 65°C, ethidium bromide was added to a final concentration of 0.5 pg/ml. A gel was casted in a tray with a comb. Electrophoresis was performed at constant current of 100 to 150 mA for more than one hour. The DNA bands were visualized by UV light and the results were recorded by photography.

2.10. Polyacrylamide Gel Electrophoresis for DNA Sequencing Analysis 8% polyacrylamide gel electrophoresis was carried out for separating the DNA fragments after the sequencing reaction was done. To prepare the gel, 36.04 g of urea was dissolved in 15 ml 5X TBE buffer, 15 ml of acrylamide/bisacrylamide (40:1.33) solution and water was added up to a final volume of 75 ml. After filtration and degassing, the polymerization was initiated by adding 200 pl of 10% APS and 30 pl TEMED. The mixture was 42 poured immediately into a gel casting tray and polymerized for two hours to overnight. The gel was then pre-run at 1500-2000 volts for about 30 minutes. Then the DNA sequencing samples were applied to the gel and electrophoresed at 1500-2000 volts for the appropriate time. The gel was carefully removed from the electrophoresis apparatus and the urea within the gel was released by submerging the gel in 10% acetic acid and 5% methanol for 10 minutes. Then the gel was dried under the vacuum and exposed to a piece of x-ray film to detect the radioactive signals showing the actual sequences of the DNA.

2.11. Preparation of M13 RF (Replicative Form) DNA and Plasmid DNA For M13 RF DNA preparation, 50 pl of M13 phage and 100 pl of £. coli strain JM101 overnight-grown cells were inoculated into 35 ml LB broth medium and grown at 37°C for overnight. On the following day, the phage in the supernatant was collected at 4°C by centrifuging the overnight culture at 6000 rpm for 5 minutes. 20 ml JM101 overnight-grown cells were added into 1 liter fresh LB broth medium and the culture was shaken at 37°C for 1.5 to 2 hours until the optical density of cells suspension reached 0.5 measured at 600 nm (wavelength). 25 ml of the collected M13 phage was added into the culture and shaken at 37°C for another 2 hours. For preparation of plasmid DNA, the 43 coli cells containing plasmid was inoculated into 1 liter LB broth medium and grown at 37°C for more than 20 hours. The cells were harvested at 4°C by centrifugation at 8000 rpm for 10 minute using a JA10 rotor. The pelleted cells were resuspended in 13.5 ml of 15% sucrose, 50 ml Tris-HCI, pH 7.5, 50 mM EDTA and 1 mg/ml freshly prepared lysozyme. After incubation at room temperature for 20 minutes, 14.5 ml of Triton X-100 solution (0.1% Triton X-100, 50 mM Tris-HCI, pH 8.0 and 50 mM EDTA) was added and incubation was continued for another 20 minutes (if lysis did not occur, the incubation would be carried out at 37°C for 30 minutes). After lysis, the cell debris was pelleted at 4°C by centrifuging at 19,000 rpm for 1 hour in a JA20 rotor. 27.5 ml of supernatant was mixed with 26.13 g CsCI and 0.5 ml of 10 mg/ml ethidium bromide (refractive index should be between 1.39 and 1.396 or at density of 1.59 g/ml). Ultracentrifugation was carried out at 20°C at 40.000 to 50,000 rpm for more than 20 hours in a Ti70 rotor. The covalently closed supercoiled DNA (lower band) was visualized by long-wavelength UV illumination and removed carefully with a syringe. Ethidium bromide in the solution was removed by extracting with 5 ml TE-saturated isobutanol three times. The bottom (water) phase was mixed with 15 ml TE, pH 8.0 buffer, 2 ml of 3 M NaOAc and 20 ml of isopropanol and incubated in -70°C for more than 20 minutes (or at -20°C for 2 44

hours to overnight) to precipitate DNA. The DNA was pelleted at 4°C by centrifuging at 17,000 rpm for 50 minutes in a JA20 rotor, then dissolved in 0.5 ml of TE, pH 8.0 buffer and transferred to a 1.7 ml Eppendorf microcentrifuge tube. The DNA solution was mixed with 50 pl of 3 M NaOAc and 0.5 ml of isopropanol, and incubated at -70°C for 20 minutes to precipitate the isolated DNA. The DNA was pelleted again at 4°C by centrifugation at 15,000 rpm in an Eppendorf microcentrifuge for 10 minutes. The DNA then was washed with cold 70% ethanol, dried under vacuum in a speed vacuum condenser and finally dissolved in 0.2 to 1 ml TE, pH 8.0 buffer.

To prepare M13 RF DNA rapidly on a small scale, a single M13 phage plaque was picked up from a LB agar plate, inoculated in 1 ml LB broth and shaken at 37°C for overnight. On the second day, 50 pl of JM101 overnight cells was inoculated into 2 ml of LB broth (usually do 6 identical tubes for one sample) and shaken at 37°C for 1.5 to 2 hour (optical density of the cell resuspension was about 0.5 measured at a wavelength of 600 nm). The overnight-grown phage-infected cells were collected at room temperature by centrifugation in an Eppendorf microcentrifuge at 15,000 rpm for 2 minutes. 10 pl phage (supernatant) was added to the cells (Ab.60Q = 0.5) and the cultures were shaken for another 2 hours. For plasmid 45

preparation, a single £. coli cell colony containing the plasmid was inoculated into 2 ml LB broth medium and grown at 37°C for overnight. In both cases, the overnight grown cells were pelleted by centrifugation and resuspended in 150 pl of GTE buffer (50 mM glucose, 25 mM Tris-HCI, pH 8.0, 10 mM EDTA). After incubation at room temperature for 5 minutes, the cells were lysed by adding 200 pl of freshly-made alkali-SDS mix (0.2 M NaOH and 1% SDS) and further incubated on ice for 5 minutes. Then 150 pl of acetate solution (5 M potassium 3 M acetate, pH 4.8) was mixed with the sample and incubated on ice for another 5 minutes. The sample was spun down at 4°C for 5 minutes and the supernatant was transferred to a new microfuge tube containing 50 pl 3 M NaOAc and 1 ml cold 100% ethanol. DNA was precipitated by incubating the sample at -70°C for 20 minutes and centrifuged at 4°C for 10 minutes. The DNA pellet was washed with 1 ml cold 70% ethanol and dried with a Speedvac. The isolated DNA was dissolved in 20 to 60 pl of TE, pH 8.0 buffer.

2.12. Oligonucleotide-Directed Mutagenesis Oligodirected-mutagenesis was performed to generate in- frame deletions and point mutations in subunit H-alkaline phosphatase fusion gene, inverted leader peptidase, and procoat 828 genes (Zoller and Smith, 1983; Kunkel, 1986; Fig. 3.4.). 46

2.12.1. Preparation of Oligonucleotides Oligonucleotides were synthesized on an Applied Biosystem Model 380B DNA Synthesizer at the Biochemical Instrument Center at The Ohio State University. The oligonucleotides in fresh 100% ammonium bicarbonate were deblocked by incubating the sample at 55°C for 9 to 12 hours. A portion of the oligonucleotides (0.5 ml out of about 3 ml total) were spun in a speed vacuum condenser to evaporate the solution. The dried DNA was resuspended in 0.5 ml water and desalted by Sephadex G-25 column chromatography with 50 mM triethanolamine bicarbonate (TEAB) buffer, pH 8.5 which was pre-equilibrated with the same buffer. A total of 6 fractions were usually collected in 1 ml of volume each and the concentration of oligonucleotide was determined by measuring O.D. ^ (40 pg/ml = 1 O. D. unit). 2 6 0

2.12.2. Isolation off the Single-Stranded Uridine- Containing M13 DNA Template The single-stranded uridine containing M13 DNA template was prepared from M13 phage grown on E. coli strain RZ1032 (ung'. dut) as described (Kunkel, 1985). An overnight grown RZ1032 cell culture was inoculated into 20 ml LB broth and shaken at 37°C. At O.D.^„„ of about 0.2, 5 ml of the cells were 600 added into 1 liter of LB broth containing 0.25 pg/ml uridine 47

(12.5 pl of 5 mg/ml stock), 12.5 pg/ml tetracycline (1 ml of 12.5 mg/ml stock). 50 pl of M13 phage (1011pfu/ml) was added into the culture to infect the cells. The culture was vigorously shaken at 37°C for overnight. At the following day, the cells were pelleted at 4°C by centrifuging in 2 X 500 ml bottles at 10,000 rpm for 10 minutes in a JA10 rotor. The supernatant was mixed with 250 ml of 20% PEG-2.5 M NaCI and incubated on ice for 1 hour to precipitate the phage. The phage was pelleted at 4°C by centrifuging at 10,000 rpm for 10 minutes in the JA10 rotor. The pellet was resuspended in 15 ml of LB media and 5 ml of 20% PEG-2.5 M NaCI was added to precipitate the phage. After incubation on ice for 30 minutes, the phage was pelleted again by centrifugation. The pellet was resuspended in 4.5 ml of TE, pH 8.0 buffer and the cell debris was removed at 4°C by spinning at 8,000 rpm for 5 minutes. The phage supernatant was transferred into a 40 ml centrifuge tube containing 6 ml of CsCI (optical density is 1.593) and mixed. The mixture was then transferred into a 13 ml quick-seal polypropylene ultracentrifuge tube and the unfilled volume was adjusted by adding mineral oil. The tube was sealed and centrifuged at 18°C in an ultracentrifuge at 40,000 rpm for 24 hours in a Ti50 rotor. The phage band (usually near the top of the CsCI gradient phase) was removed with a syringe. The purified phage was extracted in TE-saturated phenol twice to remove M13 coat proteins. 48

After extracting the traces of phenol with diethyl ether twice, the aqueous phase was transferred into a new microfuge tube and the single-stranded M13 DNA was precipitated by ethanol with one tenth volume of 3 M NaOAc. The DNA was pelleted at 4°C by centrifuging at 15,000 rpm in an Eppendorf microfuge for 10 minutes, washed with 70% ethanol and dried in a Speedvac. The DNA was dissolved in TE, pH 8.0 buffer and the concentration of DNA was determined by measuring the O.D. „ units. 7 260

2.12.3. Oligonucleotide Mutagenesis-Annealing, Elongation and Ligation Both the oligonucleotide and the M13 universal primer (a 17 base oligonucleotide) were phosphorylated by T4 DNA kinase in the presence of 1 pl of both 10 mM ATP and kinase buffer (0.5 M Tris-HCI, pH7.6, 0.1 M MgCI2> 50 mM DTT, 1 mM spermidine and

1 mM EDTA). The reaction was carried out at 37°C for 30 minutes. Annealing was performed by mixing 5 pl of single­ stranded uridine-containing DNA template (1 mg/ml), 5 pl of the kinased oligonucleotide (40 pg/ml), 1 pl of the kinased universal primer (40 pg/ml), and 1 pl of the annealing buffer (0.2 M Tris-HCI, pH 7.5, 0.1 M MgCI2, 0.5 M NaCI, 0.01 M DTT) in

12 pl total volume. The mixture was incubated at 65°C for 7 minutes and the annealing reaction was allowed to occur by cooling down the mix slowly at room temperature for over 1 49

hour. The elongation and ligation reactions were performed by mixing the following: 12 pl of the above annealed DNA, 2 pl of ligation buffer (700 mM Tris-HCI, pH 7.5, 70 mM MgCI2, 7 mM ATP, 100 mM DTT), 1 pl each of 10 mM dATP, 10 mM dCTP, 10 mM dGTP, 10 mM dTTP, 2 pl of 5 unit/pl Klenow enzyme, and 2 pl of 1 unit/pl T4 DNA ligase. The reactions were carried out simultaneously by incubating the mixture at 14°C for overnight.

After the mutagenic reaction was done, the DNA mixture of the above was then transfected into coli JM101 (ung\ section 2.15.). The mutated M13 single-stranded DNAs were isolated from single plaques and the mutation was identified by dideoxynucleotide sequencing (Sanger, feial, 1977, also see section 2.13.). The gene fragment carrying the mutation was then subcloned into a expressing plasmid (section 2.14.) and the DNA inserts and plasmid encoded protein were screened by the following analysis (section 2.16.).

2.13. DNA Sequencing The dideoxy method was applied to sequence the single­ stranded M13 DNA template (Sanger fetal. 1977) and the DNA Sequenase 2.0 kit from United State Biochemical, Inc. was used according to procedures suggested by the manufacturer. DNA template and sequencing primer was annealed by mixing 7 pl of 50 isolated single-stranded DNA, 1 pl of primer (O.D.= 0.05 at 600 nm), and 2 pl of annealing buffer. The mixture was incubated at 65°C for 2 minutes, then annealing was allowed by cooling the mix slowly at room temperature for over 1 hour. The annealed mixture was then mixed with 1 pl of 0.1 M DTT, 2 pl of labeling mix (diluted 1:5 by water), 1 pl of [a-35S] dATP (10 pCi/ml) and 1 pl of Sequenase 2.0 (diluted 1:8 by TE, pH 8.0). The labeling reaction was incubated at room temperature for 3 to 5 minutes. The termination reaction was initiated by transferring 3.5 pl out of total 16 pl mix into a microfuge tube containing 2.5 pl of ddATP termination mix (prewarmed at 37°C for more than 2 minutes and labeled as A). Similar transfers were done to the tubes containing ddGTP, ddCTP and ddTTP termination mixes (labeled as G, C and T), respectively. After mixing, the reaction was incubated at 37°C for 5 minutes and quenched by adding 5 pl of stop solution. The samples were heated at 90°C for 2 minutes, loaded on a sequencing gel and analyzed by polyacrylamide gel electrophoresis (section 2.10.).

2.14. Subcloning The M13 RF and plasmid DNA used for subcloning were prepared as described (section 2.11.). Restriction digestion was performed to cut the gene fragment out of the M13 or plasmid vector by mixing 30 pl to 60 pl of vector (plasmid) DNA or M13 51

RF DNA, 25 pl of 10 X restriction buffer, 5 pl each of two restriction enzymes and water to make up to 250 pl of the final volume. After incubation at 37°C for at least 2 hours, the reaction was terminated by adding 25 pl of 10 X loading buffer. The samples were analyzed by agarose gel electrophoresis (section 2.9.) and DNA bands were visualized under long wavelength UV light. Pre-cut and pre-soaked DEAE-cellulose papers (in TE, pH 8.0) were placed right in front of the DNA bands of interest and the DNA was electrophoresed onto the paper by running the gel for another 10 to 15 minutes. The filter was washed with 0.5 ml of low salt buffer (0.15 M NaCI, 0.1 mM EDTA and 20 mM Tris-HCI, pH 8.0), and the DNA was extracted with high salt buffer (1.5 M NaCI, 0.1 mM EDTA and 20 mM Tris-HCI, pH 8.0) at 65°C for 1 hour by vortexing a couple of times during the hour. The eluted DNA was extracted twice by 0.5 ml TE- saturated phenol. After spinning at 15,000 rpm in an Eppendorf microcentrifuge, the aqueous phase was transferred into a new microfuge tube and the trace of phenol was removed by extracting twice with 1 ml of diethyl ether. The DNA fragment was precipitated by adding 50 pl of 3 M NaOAc and 1 ml of 100% ethanol followed by incubating the mixture at -70°C for 20 minutes. The DNA pellet was collected at 4°C by centrifugation for 10 minutes and dried in a Speedvac. The isolated DNA was dissolved in 80 pl TE, pH 8.0 buffer. During this time, a spin 52 column was prepared by pinching a tiny hole at the bottom of a 0.5 ml Eppendorf microcentrifuge tube with a 26 X 5/8 Gauge needle and filled with TE, pH 8.0 pre-soaked Sephadex G-25. The column was spun to remove the solution, additional Sephadex G- 25 was added and spinning was resumed. This process was repeated for several times until the dried Sephadex G-25 filled the tube up. The DNA solution was applied at the top of the spin column and centrifuged a few times until no wet Sephadex G-25 was observed within the column. The flow-through salt-free DNA fragment was collected and run in a agarose gel (section 2.9.) to check the yield.

For ligation of the isolated DNA fragment into plasmid vector, the DNA insert and linearized vector were added into a microfuge tube in 1:1 molar ratio according the yield in the agarose gel. 5 pl of 5 X ligation buffer, 1 pl of 1 unit/pl T4 DNA ligase and water were added to the tube to 25 pl final volume. The mixture was then incubated at 14°C overnight to ligate the DNA.

Figures 3.3. is the demonstration of the subcloning procedures of subunit H gene fpuhAi from pBRHL19 into pUI310 containing the gene encoding alkaline phosphatase or the in­ frame fused ohoA-puhA gene after deletion by oligonucleotide- 53 directed mutagenesis into pUC19 vector, the leader peptidase gene (lep) and procoat-OmpA fused gene from M13mp8 into pING- 1 vector.

2.15. Transformation and Transfection Competent £. coli cells used for transformation and transfection were prepared as follow. Overnight-grown coli cells were inoculated into 50 m! to 80 ml LB broth and shaken at 37°C until the cell culture reached the mid-log phase (O.D.6Q0 was about 0.5 unit). The cells were pelleted at 4°C by centrifugation at 6,000 rpm for 5 minutes and resuspended in 10 ml of ice-cold 50 mM CaCI2- After incubation on ice for 15 minutes, the cells were spun down under the same conditions again and the pellet was resuspended in 2 to 3 ml of the 50 mM CaCL. The incubation was continued for another 15 minutes and 2 the cells were ready for the next step.

For transformation, 300 pl of the competent cells were mixed with ligation mixture or appropriate amount plasmid DNA and incubated on ice for at least 5 minutes. The mixture was heat treated at 42°C for 45 seconds. 1 ml of LB broth was added and mixed by immediately inverting the tube several times , and the culture was incubated at 37°C for 30 minutes. The transformed cells were collected by spinning the culture briefly in an Eppendorf microcentrifuge, removing most of supernatant, 54

and the remaining portion was resuspended and spread on the top of a LB agar plate containing 50 pg/ml ampicillin. The plate was then incubated at 37°C in a dry incubator for overnight.

For transfection, 300 pl of JM101 competent cells and 1 pl to 5 pl of M13 RF DNA or the mutagenic mixture (see section 2.12.3.) were mixed and incubated on ice for at least 5 minutes. The mix was heated at 37°C for two and half minutes, then 1 ml LB broth was added to the tube. After mixing the culture by inversion, a portion or the entire contents was quickly added into a microcentrifuge tube containing 3 to 4 ml of melted LB top agar (47°C) and 0.5 ml of overnight-grown JM101 cells. The whole mix was then poured onto the top of a LB agar plate and incubated at 37°C for 6 hours to overnight.

2.16. Screening for Positive Clones with Different M ethods Two different methods were used for screening of positive clones. One method was to cut the isolated plasmid DNA with certain restriction enzymes and then examine the newly produced DNA inserts to identify the clones. After subcloning the DNA gene fragment into the plasmid vector and transforming the ligated DNA into the coli host, the individual colonies were picked up with a sterilized tooth-pick followed by 55 isolating the plasmid DNA (section 2.11.)- On© tenth (usually 2 pl out of 20 pl total volume) of the prepared plasmid DNA was digested by mixing the DNA, 1 pl each of two restriction enzymes, 1 pl of heat-treated 0.1 mg/ml RNase, 2 pl of 10X restriction buffer and water to a final volume of 20 pl. The mixture was incubated at 37°C for at least 1 hour and analyzed by agarose gel electrophoresis (section 2.9.). The cut-out inserts were visualized by UV light and the results were recorded by photography (Fig. 2.1.).

The other approach was immunoscreening. After observing the clones containing the appropriate DNA inserts by the restriction digestion method (see above, if no restriction enzyme site is available, the subclones can be screened directl), the cells containing the interested plasmid were grown overnight in M9 media supplemented with 0.5% fructose, 50 pg/ml of 19 amino acids except methionine, and 50 pg/ml ampicitlin. The following day, the overnight cells were diluted in a 1:50 ratio into 1 ml fresh M9 media and grown for 2 hours. After inducing expression of plasmid-encoded proteins with 10 pl of 20% arabinose for 30 minutes to 1 hour, the ceils were labeled with p 5S]-methionine, TCA precipitated and immunoprecipitated with the appropriate antibody. The expression of plasmid encoded protein was analyzed by SDS- 56

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Lep-lnv.

Lep-lnv.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Fig. 2.1. Screening for the subcloned inverted leader peptidase mutants with restriction enzyme digestion after isolation of the plasmid DNA. The plasmid DNA was cut by restriction enzymes Sal I and Sma I at 37°C for 1 hour, run on a 0.7% agarose gel, and stained with ethidium bromide. Lanes 1 to 6 were the mutants E68. Lanes 7 to 12 were R69. Lanes 13 to 18 were R67. Lanes 19 to 24 were R64 and lane 25 was DNA molecular weight markers. 57

PAGE electrophoresis and fluorography (Fig. 2.2.).

For screening of positive clones of the subunit H-alkaline phosphatase fusion proteins, immunoblotting analysis was performed as described (section 2.20., Fig. 3.6.).

2.17. Preparation of Spheroptasts Spheroplasts were prepared from £. coli cells as described (Randall and Hardy, 1986). E. coli strains bearing the plasmid encoding leader peptidase, procoat-OmpA, or their respective mutants were inoculated into M9 media supplemented with 0.5% fructose, 19 amino acids except methionine and 50 pg/ml ampicillin and grown for overnight. On the following day, the saturated cells were diluted 1 to 50 into fresh M9 media and shaken for 2 hours. The expression of proteins encoded by the plasmids was induced by adding arabinose at the mid-log phase of the growing culture. The cells then were either pulse-labeled or pulse-chased with [35S]-methionine and chilled quickly with 0.5 ml ice-cold water. The cells were collected at 4°C by centrifugation for 40 second and resuspended in 250 pl of 100 mM Tris-acetate, pH 8.2, 50 mM sucrose and 5 mM EDTA. To remove the outer membrane of the cells, 20 pl of 2 mg/ml freshly made lysozyme was added and immediately mixed with 250 pl of ice-cold water. After 5 minutes incubation on ice, 50 58

Procoat-OmpA Mutants A21-33 A21-39

+ + + 1 2 3 4 5 7 8 9 10 11 12

Figure 2.2. Screening for the subcloned procoat-OmpA

A 21-33 and A21-39 mutants with immunoprecipitation. Cells were grown at 37°C to their midlog phase and incubated with

(+) or without (-) 0.2 % arabinose for 30 minutes. The proteins were pulse-labeled with 10 pCi [35S]-methionine for 1 minute. The samples were subjected to immunoprecipitation, SDS-PAGE and fluorography analysis. Lanes 1 to 6 were the mutants of A21-33 and lanes 7 to 12 were the mutants of

A21 -39. 59 pi of 0.2 M MgS04 was added to stabilize the newly-formed

spheroplasts. The spheroplasts were pelleted by centrifugation for 30 second and resuspended carefully in 500 pi of 50 mM Tris-acetate, pH 8.2, 0.25 M sucrose and 10 mM MgS04. Then the spheroplasts were ready for protease accessibility studies.

2.18. Protease Accessibility Assay The translocation of interested proteins across the inner membrane of £. coli cells was monitored by a protease accessibility technique (Randall and Hardy, 1986, Fig. 4.3.). To do this, 150 pi of spheroplasts was aliquoted into three different microcentrifuge tubes containing 15 pi of 1 mg/ml trypsin, 15 pi each of trypsin and 20% triton X-100, or none of the above, respectively. After mixing, the protease digestion was carried out by incubating the tubes on ice for 1 hour. The reaction was quenched by adding 40 pi of soybean protease inhibitor (10 mg/ml), incubating on ice for 5 minutes followed by the addition of 5 pi of 0.1 M PMSF in 100% ethanol, and further incubating on ice for another 5 minutes. The treated spheroplasts were lysed by adding 250 pi of ice-cold 20% trichloroacetic acid (TCA) and incubated on ice for 30 minutes to precipitate the proteins as well. 60

2.19. Immunoprecipitation The process of immunoprecipitation was carried out as described (Wolfe el a!-, 1982). After precipitation of the proteins with 20% TCA, the samples were centrifuged at 4°C for 5 minutes. The supernatant was aspirated away with a Pasteur pipet. The precipitates were washed by adding 1 ml of ice-cold acetone with vigorous vortexing for 30 seconds. The pellet was collected at 4°C by centrifugation for another 5 minutes and the supernatant was again aspirated away. The pellet was dried at 95°C for 5 minutes at which the 140 pi of 10 mM of Tris-HCI, pH 8.0, 2% SDS was added into the tube. Heating was continued for 15 to 20 minutes with intermittent vortexing to completely dissolve the pellet. The debris was removed by centrifuging the sample at room temperature for 5 minutes and 90% of the supernatant was transferred into a new tube containing 1 ml of ice-cold immunobuffer (10 mM Tris-HCI, pH 8.0, 5 mM EDTA, 150 mM NaCI and 2.5% Triton X-100) and 25 pi of StaphA followed by incubation for 15 minutes on ice for pre-absorbing. StaphA was then removed by brief centrifugation (15 seconds). The supernatant was again transferred into a new tube containing 5 pi or 10 pi of antibody to bind the protein of interest. After 30 minutes incubation on ice, 25 pi of StaphA was added to the sample and incubation was continued for another 30 minutes. The immuno-complex was then spun down 61 and washed with immunobuffer twice by successive suspension and centrifugation. The washed sample was resuspended in 30 pi of 2X Sample buffer (100 mM Tris-HCI, pH 6.8, 20% glycerol, 1.5% 3-mercaptoethanol, 0.02% bromophenol blue and 4% SDS) and heated at 95°C for 5 minutes prior to loading on a SDS-PAGE gel for electrophoresis.

2.20. Immunoblotting Immunoblotting analysis was performed according to the method of Towbin el ai (1979) using the Protoblot kit from Promega, Inc. The proteins separated by SDS-PAGE were transferred onto a nitrocellulose paper as follows. The gel was carefully cut according to the dimension of the protein containing area of the gel and soaked in the transfer buffer (20 mM Tris, 150 mM glycine with or without methanol). The corresponding sized nitrocellulose paper and 3 MM paper were cut out and soaked in the same solution. The above items were assembled in a cassette as a sandwich with 3 MM paper on the outside and the nitrocellulose paper on the side of the anode. The sandwich cassette was put in a transfer apparatus and electrophoresis was carried out at 4°C at 150 mA for 4 hours to overnight. After transfer, the nitrocellulose paper was rinsed with distilled water and agitated in 1 X TBS buffer (0.15 M NaCI, 0.01 M Tris-HCI, pH 7.5 and 0.5 mg/ml Tween 20 with 10 mg/ml 62

BSA ) at room temperature for 10 minutes. The paper was rinsed with water and incubated in same solution in the presence of antibody for 1 hour. After agitating twice the paper in 1 X TBS plus BSA solution for 10 minutes each, the same solution with the secondary antibody conjugated with alkaline phophatase (rabbit IgG, 1:10,000 dilution as recommended by the manufacturer) was added and incubation was continued for another hour. The paper was again washed twice with 1 X TBS (plus BSA) and rinsed with water. A substrate solution for alkaline phosphatase (AP buffer, 100 mM Tris-HCI, pH 9.5, 100 mM NaCI and 5 mM MgCl2) was then prepared containing 66 pi of

50 mg/ml nitroblue tetrazolium chloride (NBT) in 70% dimethylformamide (DMF) and 33 pi of 50 mg/ml bromochloroindoly! phosphate (BCIP) in 100% DMF. The paper was agitated in the AP buffer for 2 to 10 minutes or until the protein band can be seen (Fig. 3.6.). CHAPTER III

THE USE OF GENE FUSIONS TO DETERMINE WHICH REGIONS OF A PROTEIN ARE IMPORTANT FOR THE MEMBRANE ORIENTATION OF SUBUNIT H PROTEIN OF THE PHOTOSYNTHETIC REACTION CENTER FROM RHODOBACTER SPHAEROIDES

3.1. Introduction The photosynthetic reaction center in Rhodobacter sphaeroides is comprised of two pigment-binding subunit proteins, L and M and one non-pigment-binding subunit protein H. The L and M proteins both have five transmembrane segments and forms a complex that binds to pigments while the H protein is anchored in the membrane by a single hydrophobic stretch near the N-terminus (Allen fila l, 1987b). The subunit H protein appears to be vital for the correct and stable assembly of a functional reaction center in the membrane of Rhodobacter sphaeroides (Sockett a ia l, 1989). X-ray crystallographic studies have also revealed that the extensive contacts of the subunit H protein with the LM complex which may stabilize the reaction center structure (Allen filai., 1987b).

63 64

In contrast to L and M, which both have multi-spanning membrane regions, the H protein spans the membrane only once. It only has one helix crossing the membrane with a small segment (residue 1 to 12) exposed in the periplasm and most of the protein faces the cytoplasm. More interestingly, the distribution of charged amino acid residues flanking the transmembrane segment does not follow von Heijne's "positive inside rule" which states that hydrophobic transmembrane segments orient themselves in the membrane with the most positively charged amino acid residues facing the cytoplasm. Instead, a highly negatively charged region (residues 34 to 46) is present immediately after the helix in the cytoplasm (Fig. 3.1.). Thus, the intriguing question here is what is the topogenic determinant for this membrane orientation.

In recent years a number of genetic techniques have been developed to follow and dissect the process of protein localization (Michaelis and Beckwith, 1982; Oliver, 1985; Benson fiiai-, 1985; Silhavy and Beckwith, 1985; Manoil fii a i, 1988). One such technique involves the fusion of genes encoding reporter molecules such as LacZ (Bassford aiaL, 1978; Beckwith and Silhavy, 1983; Silhavy and Beckwith, 1985), and PhoA (Manoil and Beckwith, 1985; Hoffman and Wright, 1985; Manoil and Beckwith, 1986; Boyd a ia l, 1987; San Millan aial-, 1989; Peripalsm

Plasma Membrane

Cytoplasm

Figure 3.1. Membrane topology of subunit H protein of the photosynthetic reaction center from Rhodobacter sphaeroides. 66

Verga and Kaptan, 1989; Chepuri and Gennis, 1990) to the gene{s) of interest to monitor the regulation of expression and localization of a particular protein. Specific fusions to a known region of a gene may then allow genetic dissection of a protein into its structural and functional domains. This will reveal those regions of a protein that are required for specific targeting to one of the cellular membrane systems or compartments (Boyd and Beckwith, 1989; San Millan eiai-. 1989; McGovern and Beckwith, 1991).

Among these gene fusion techniques, the alkaline phosphatase approach is most commonly used in £. coli. Alkaline phosphatase is encoded by phoA gene and the DNA sequence of the PhoA gene has been determined (Chang fiiai-. 1986). After protein synthesis, alkaline phosphatase is translocated across the cytoplasmic membrane and localized in the periplasm of £. coli (Chang £lai-, 1983; Inouye and Beckwith, 1977). Alkaline phosphatase is a homodimer enzyme. It is enzymatically active in its normal location, the periplasm, but it is inactive in the cytoplasm (Schlesinger filai-, 1969; Michaelis filal-, 1983; Hoffman and Wright, 1985; Kim and Wyckoff, 1991). Furthermore, when the enzyme is fused at its amino terminus to a periplasmic domain of an inner membrane protein, it is active; when it is fused to a cytoplasmic domain of such a protein, it is 67 inactive. Thus, the properties of such fusions allow the topogenic determinants of the plasma membrane proteins to be analyzed (Boyd and Beckwith, 1989; Fig. 3.2.).

In order to identify the factor that determines the membrane topology of subunit H protein in Rhodobacter sphaeroides. we have chosen the phoA gene fusion technique and combined it with site-directed mutagenesis. Two subunit H- PhoA in-frame gene fusions were constructed, in which one retains the normal charge distribution after the membrane spanning region and one has the entire negatively charge cluster deleted (H34 and H46, Fig. 3.5.). Four other mutants were also constructed to demonstrate that the PhoA gene fusion approach was appropriate (2LysH34, 2LysH46, H-ln and H-Del, Fig. 3.5.). Our results showed that the negatively charged cluster in the subunit H protein from Rhodobacter sphaeroides is not the topogenic determinant for its membrane orientation. The possible reason for its orientation across the cytoplasmic membrane will be discussed.

3.2. Construction of Subunit H-PhoA Gene Fusions To obtain gene fusions in which PhoA was attached to the beginning of the cytoplasmic domain of subunit H protein, we have used oligonucleotide-directed loop-out mutagenesis 68

1

Periplasm

Plasma Membrane

Cytoplasm

Periplasm

Plasma Membrane

Cytoplasm

Acittve (PhoA +) Inacitive (PhoA -) Alkaline Phosphatase

Figure 3.2. Use of alkaline phosphatase gene fusion to analyze the membrane protein topology. 69

deletions as described previously (San Millan slaL, 1989; Boyd alai-, 1987; Egthedarzadeh and Henikoff, 1986). The overall strategy of the construction is illustrated in figure 3.4. Two key plasmids, both derived from pUC19, were kindly provided by Dr. Samuel Kaplan at the University of Texas, Medical Center in Houston. The first plasmid, pRHBL19 has a 1.32 kb BamHI fragment insert containing the subunit H protein gene from Rhodobacter sphaeroides /Sockett et al.. 1989). The second plasmid, pUI310, which was constructed to facilitate the PhoA fusion analysis, contains a 1.47-kb Pstl-BstE 11 PhoA gene encoding alkaline phosphatase lacking its signal sequence (Verga and Kaplan, 1989). The 1.32 kb BamHI fragment containing the subunit H gene was isolated after digesting the pRHBL19 with the restriction enzyme BamHI. The purified subunit H gene fragment then was subcloned into the BamHI site in pUI310 which places the gene in front of phoA gene (Fig. 3.3.). The correct orientation of the subcloned subunit H gene was confirmed by restriction enzyme mapping. The ligated subunit H-phoA gene fragment was digested by the restriction enzymes Xbal/Kpnl and a 2.79 kb (subunit H 1.32kb + PhoA 1.47kb) fragment was separated from the linear pUC19 vector (2.86kb) on a 1% agarose gel that was run slowly overnight. The larger fragment was then isolated from the gel and subcloned into Xbal/Kpnl cut M13mp18 and mp19 RF DNAs. The M13 RF DNAs 70

ouhA ptoA BamHI BamHI BamHI BstEII

PRHBL19 pUI310

BamHI, BamHI, Phosphatase Isolate the BstEII fragment phoA

1.32kb

puhA pUI310

K)

T4 BamHI Ligase BamHI phoA

BstEII

Subject to in-frame fusion by oligonucleotide-directed deletion {see figure 3.4).

Figure 3.3. Construction of subunit H fpuhA)-phoA gene fusion. 71

CUM BamHI

BamHI

phoA

BstEII

Anneal with H34 or H46 Oligonucleotide (48-mer)

PuhA BamHI

BamHI phoA A primer

BstEII

H34-PhoA or H46-PhoA

BamHI Mutagenic Deletion BstEII

Figure 3.4. In-frame fusion of subunit H-PhoA protein by oligonucleotide-directed gene deletion. 72 were used to transfect Ca2+-treated E. coli JM101 cells and plated out on a LB agar plates supplemented with IPTG and X-gal. The white plaques containing the insert were picked and grown overnight in LB broth at 37°C. Single-stranded DNAs were then isolated and subjected to dideoxy-chain termination DNA sequencing (Sanger e ia i, 1977). The resulting sequences of M13mp18 and M13mp19 clones corresponded to the 3’ and 5’ sequences of subunit H and PhoA genes, respectively. For site- directed mutagenesis, only the M13mp18 clone was chosen to prepare uridine-containing template. Two 48-mer oligonucleotides were synthesized (Biochemical Instrument Center of The Ohio State Univrsity) that correspond to the sequences from both sides in the fusion joint of subunit H-phoA gene. This allowed us to make a deletion from the fusion joint in subunit H gene (nucleotide sequences corresponding to amino acid residue 34 or 46) to the first nucleotide of phoA gene which lacks its signal sequences (Fig. 3.4 and 3.5.). The third oligonucleotide was also synthesized that has a sequence further upstream of the above two oligonucleotides for the purpose of sequencing. Oligonucleotide-directed mutagenesis/deletion was performed as described in chapter II. Deletions were confirmed by sequencing through deletion regions on the M13 derivative with a specific sequencing primer (the third oligonucleotide mentioned above). The two new in-frame 73

Wild type + PhoA euM 44 44 PhoA

H34 I ivwwwwvvvwvvv^vvj puhA 34 phoA

H46 PUhA 46 ohoA

2LysH34 PuhA 34 ohoA

2LysH46 Ift :±~ LWVWWVVWV^VVVM PuhA 48 PhoA

10 A.A. ol Pf3 H-ln ■ ^ - -. puhA

H-Del r ~ / \ '*■ — puhA

Figure 3.5. Constructs of subunit H-PhoA fusion proteins or subunit H mutants. fused mutants were isolated after digestion by Xbal/Kpnl and subcloned back into plasmid pUC19. The positive pUCl9 clones were then selected by examining which plasmid produced a 2.1 kb DNA inserts upon digesting with the appropriate restriction enzymes. Next we detected the expression of the fusion proteins by immunoblotting procedure using an antibody against alkaline phosphatase (Fig. 3.9.). The fusion proteins were then analyzed for protein translocation by assaying the alkaline phosphatase activities of both fusion proteins in vivo.

Previously it has been demonstrated that the introduction of positively charged amino acid residues immediately before transmembrane segment can flip its orientation (von Heijne ai, 1988; San Millan filai-, 1989; Nilsson and von Heijne, 1990). Therefore, we inserted two lysine residues right in front of the only transmembrane segment of subunit H-PhoA fusion protein (Fig. 3.5.) by oligonucleotide-directed mutagenesis (section 2.12.). The mutated subunit H-PhoA fusion proteins should reverse their membrane orientation and expose the cytoplasmic segment of the subunit H protein as well as the alkaline phosphatase to the periplasm. Thus, a dramatic increase in the alkaline phosphatase activity should be observed for these two m utants. 75

Since subunit H protein has only 12 amino acid residues facing the periplasmic side of the membrane, it may be very difficult to detect the insertion of this amino terminus across the membraneby protease-mapping. Only a very small shift in the molecular weight on a SDS-PAGE analysis is expected after digesting with protease from the periplasmic side of the membrane. It also may be that even if subunit H protein inserts across the membrane and is digested by protease, we still can not detect the shift in molecular weight on a SDS-PAGE gel. Thus two mutants were constructed to see this is a problem. The first mutant had a deletion from amino acid residues 2 to 11 of the subunit H protein deleted (H-Del, Fig. 3.5) and the second mutant was constructed as follow. Based on the successful attempt on monitoring the export of the H1 segment of coli leader peptidase across the inner membrane (Lee alai-. 1992), we inserted 10 amino acid residues from the very amino terminus of Pf3 phage coat protein in front of the membrane anchoring domain of subunit H protein (H-in, Fig. 3.5.). We hoped that by constructing these two mutants, with one which its periplasmic domain decreased and another with its periplasmic segment increased, we could distinguish the molecular weight differences among the wild type, deletion and insertion mutants of subunit H protein by SDS-PAGE gel electrophoresis. Therfore, protease-mapping should be a useful assay to detect the 76 insertion of the short periplasmic domain of subunit H protein by protease K mapping (Lee alal-, 1992).

3.3. Expression In £. coli of Subunit H Protein of the Photosynthetic Reaction Center from Rhodobacter sphaeroides The subunit H protein is normally expressed in Rhodobacter sphaeroides. a typical gram-negative bacterium. Since E. coli is different from its normal host, we first wanted to check its expression level in this bacterium. Unfortunately, the expression of the subunit H could not be detected by immunoprecipitation. We then used the well-established immunoblot technique to detect the expression. The pBRHL19 plasmid was first transformed into E. coli strain HJM114. 1 Klett unit (0.014 O.D-600) overnight cells were inoculated into

100 ml M9 medium supplemented with 0.4% succinic acid and 0.06% casamino acids and grown at 37°C overnight with IPTG (2 pM of final concentration) which induces expression of these plasmid endcoded proteins . After isolation of the membrane fraction, the protein profile (total amount 100 pg) was analyzed on a 16% SDS-PAGE gel. The proteins were then transferred onto a nitrocellulose membrane in Tris-glycine buffer without methanol. The blotted proteins were probed by an antibody against subunit H protein and were then monitored by a second 77

antibody conjugated with alkaline phophatase. The reason we used alkaline phosphatase conjugated antibody was that this system is more sensitive than horse radish peroxidase conjugated antibody and is commonly used in immunoblot detection. Figure 3.6. shows that even though the expression of the subunit H protein is low, its expression is clearly detectable.

3.4. The Insertion of the Subunit H Protein into the Cytoplasmic Membrane of £. coli. Since the subunit H protein is a foreign protein in coli. the next question we asked was if the protein has inserted into the plasma membrane of coli. As there is only a small periplasmic membrane fragment of the subunit H protein (total 12 amino acid residues, Fig. 3.1.), two different methods were used to approach this question.

The first aspect we looked at was whether the subunit H protein has been inserted into the membrane of £. coli. This was tested by using an alkali fractionation method (Davis fila i., 1985) which can distinguish between a peripheral and a membrane-embedded protein. By mild alkali treatment of the membrane fraction, a peripheral protein, which is loosely associated with the membrane, will dissociate and enter the supernatant. In contrast, membrane embedded proteins will Figure 3.6. Expression of subunit H of the photosynthetic reaction center in E. coli strain HJM114. The experimental procedure is described in the text. Lane 1 are proteins from a concentrated membrane fraction from Rhodobacter sphaeroides strain 2.4.1 which overexpresses subunit H. Lane 2 are proteins from a membrane fraction from a Rhodobacter sphaeroides strain 601 in which the subunit H gene is deleted. Lane 3 are proteins from a membrane fraction from £. fiflli strain HJM114 bearing pBRHL19 {subunit H gene). Lane 4 are proteins from a membrane fraction from HJM114 bearing pUC19 (the parental plasmid of PBRHL19).

78 79

110KD 84 KD

47 KD 33 KD

24 KD

16 KD

Figure 3.6. Expression of subunit H of the photosynthetic reaction center in E. coli strain HJM114. 80 remain in the membrane. Therefore, after centrifugation, the membrane fraction will pellet at the bottom of the tube, and separate from the supernatant. The proteins in these two different fractions then can be detected by either immunoblotting or immunoprecipitation. Figure 3.7 shows that more than 70% of subunit H protein was pelleted and, thus inserted into the membrane after it was expressed in £_. coli (Fig. 3.7, lane 3, look at the band at 28 KD). This percentage is quite promising for membrane-embeded subunit H, considering the subunit H is a foreign protein in £.. coli and its complete insertion into the cytoplasmic membrane may need extra cellular components or it may require interaction with the other two subunits of the photosynthetic reaction center, L and M, in Rhodobacter sphaeroides (Allen et al.. 1987b),

The second approach was to do reverse protease mapping (Kaplan, personal communication) on subunit H protein. As shown in figure 3.8, after the inner membrane fraction of HJM114 cells (bearing the pBHRL19 plasmid) was isolated, the membranes were subjected to trypsin digestion with or without sonication for different times. If the inner membrane of cells were right-side-out, there would be no change of subunit H protein in the presence or absence of the protease. The full length protein of 28 KD is unchanged (Fig. 3-8, lanes 1, 2 and 3). Figure 3.7. Alkali fractionation of membrane derived from E. coli HJM114 cells which express the subunit H protein of the photosynthetic reaction center from Rhodobacter sphaeroides. A total amount of 100 pg protein of the isolated membrane fraction was diluted with TE pH 8.0 to a total volume of 200 pi. 1 ml ice-cold 0.1 M NaOH was added into the solution and the mix was vortexed vigorously. The mix then was centrifuged at 4°C at 15,000 rpm in an Eppendorf microcentrifuge for 15 min. Then, the supernatant was transferred to a fresh tube. After neutralizing the alkali treated sample, 15 pi of 100% TCA and 1 ml of ice-cold 5% TCA were added to the supernatant and pellet of membranes tubes, respectively. The TCA mixes were incubated on ice for 10 min. The precipitated proteins were spun down at 4°C for 10 min. and washed with 1 mt ice-cold acetone. After dissolving the protein in 150 pi 10 mM Tris-HCI, pH 8.0, 2% SDS, the proteins were analyzed on 8% to 25% gradient SDS-PAGE gel and subjected to immunoblotting. Lane 1 are the proteins from the supernatant. Lane 2 are the proteins from the pellet. Lane 3 are the proteins from membrane fraction without treatment, and lane 4 are the proteins from Rhodobacter sphaeroides strain 2.4.1 (see Fig. 3.6.).

81 82

110KDI 84KD»

47 33 KD^ i: I '■ 24 KD

Figure 3.7. Alkali fractionation of membrane derived from E. coli HJM114 cells which express the subunit H protein of the photosynthetic reaction center from Rhodobacter sphaeroides. Figure 3.8. Reverse protease mapping of the subunit H protein in the membrane fraction isolated from £. coli HJM114 cells expressing the protein. The isolated membrane fraction was first washed with 20% sucrose. One half of the sample was sonicated three times with a Fisher membrane sonicator for 5 min each shock. The sonicated and non-sonicated samples were then digested on ice with chymotrypsin and trypsin mix (1% of weight/weight final concentration) for either 2 or 10 min.

The treated samples were precipitated with 20% TCA, washed with acetone and dissolved in sample buffer. The proteins then were analyzed on a 8% to 25% SDS-PAGE gradient gel and further analyzed by immunoblotting probed with an antibody against the subunit H protein. Lane 1 are the proteins from non-sonicated and undigested membrane fraction. Lane 2 are the proteins from non- sonicated fraction but digested with protease for 2 min. Lane 3 is the same sample of lane 2 but the digestion time was 10 min. Lane 4 are the proteins from the sonicated sample and digested for 2 min. Finally, lane 5 is the same sample of lane 4 but the digestion time was 10 min.

83 84

110 K D

84 K D

47 K D

3 3 K D

l!* '

24 K D

1 2 3 4 5

Figure 3.8. Reverse protease mapping of the subunit H protein in the membrane fraction isolated from coli HJM114 cells expressing the protein. 85

However, if the inner membrane were broken by sonication in the presence of protease, then the 28 KD protein band will decrease. Figure 3.8. shows that up to 85% of 28 KD band was reduced and two new protein bands, 21 KD and 19KD, appeared at the early protease treatment point. These two bands correspond to the proteolytic fragments of the bulk cytoplasmic domain of the subunit H protein (Kaplan, personal communication). When the membrane is broken, this domain is exposed to the protease and partially digested by the enzymes (Fig. 3.8, lanes 4 and 5). These results demonstrated that the major part of the subunit H protein is indeed located on cytoplasmic side (otherwise the 21 KD and 19 KD protein fragments would not be observed). Thus the reverse protease mapping and alkali fractionation data show that the subunit H protein is inserted into the inner membrane with the majority of the protein in the cytosol.

3.5. Expression of the Subunit H-PhoA Fusion Proteins The plasmids encoding H34, H46, 2LysH34, 2LysH46, H-Del and H-ln genes were transformed into an alkaline phosphatase deficient E. coli strain, CC118 (PhoA-). The positive clones of CC118 cells which express the subunit H-PhoA fused genes produced blue colonies on LB plates supplemented with XP (BCIP). An overnight culture of these CC118 cells was used to inoculate 50 or 100 ml of aerobic rich medium supplemented 86 with 0.3% DL-sodium lactic acid and grown to early exponential phase. The cells were then spun down and the membrane fraction of the cells was isolated as described in chapter II. A 100 pg of total protein in the membrane fraction was then subjected to SDS-PAGE and immunoblot analysis using an antibody against alkaline phosphatase. The results in figure 3.9 show that the expression of the subunit H-PhoA fusion proteins was substantially higher than that of subunit H alone and each of the fusion proteins ran at its corresponding molecular weight position. The two protein bands that show up in pUI310, which encodes the phoA gene lacking the signal sequence (Verga and Kaplan, 1989) can be explained by the fact that alkaline phophatase runs in two different forms on SDS-PAGE analysis (Verga and Kaplan, peraonal communication). It should be noted that the expression of the other two mutated subunit H proteins, H-ln and H-Del, was unsuccessful. Possibly these two mutants are very unstable under the conditions we used.

3.6. The Negatively Charged Amino Acid Cluster Are Not Determinants of Membrane Orientation of the Subunit H Protein The effect of the negatively charged residues that immediately follow the transmembrane segment of the subunit H Figure 3.9. Expression of the subunit H-PhoA fusion protein in E. coli CC118, a PhoA* strain. An overnight culture of CC118 cells was grown to the early exponential phase. The cells were harvested and the membrane fraction was isolated as described in text. For each sample, 100 pg of total proteins was analyzed on a 12% SDS-PAGE gel and the expression of the fusion protein was detected by immunoblotting (with an antibody against coli alkaline phosphatase). Lane 1 is purified £.. coli alkaline phosphatase. Lane 2 are the proteins from CC118 cells bearing the pUCl9 plasmid (the parental plasmid of H34 or H46). Lane 3 are the proteins from CC118 cells without any plasmid. Lane 4, 5, 6 are the proteins from three individual clones of 2LysH46. Lane 7 is 2LysH34. Lane 8 is pUI310 that contains ohoA gene lacking the encoded signal peptide. Lane 9 is H46 and lane 10 is H34.

87 88

205 KD #

116.5KD

77 KD

Figure 3.9. Expression of the subunit H-PhoA fusion protein in E. coli CC118, a PhoA* strain. 89 protein was determined by using the well-established alkaline phosphatase fusion approach. In this assay, the enzymatic activity of alkaline phosphatase reflects its location relative to the cytoplasmic membrane. If the enzyme is located inside the cytoplasm, the alkaline phosphatase activity is very low. If the enzyme is located on the periplasmic side of the cytoplasmic membrane, the alkaline phosphatase will dimerize and shows its full activity (Kim and Wyckoff, 1991). Alkaline phosphatase activity was determined as described in section 2.5. (Brickman and Beckwith, 1975; Chepuri and Gennis,1990). Table 3.1. summarizes the alkaline phosphatase activities for each of the subunit H-PhoA fusion proteins studied here. In order to ascertain that the assay worked properly, three controls were performed. We measured the alkaline phosphatase activity of cells bearing pUCl9 (parental plasmid of the subunit H-PhoA fusion proteins), pVPE172 (PhoA was fused at cytoplasmic side of cytochrome B ) and pVPS142 (PhoA was fused at periplasmic side of cytochrome D, Chepuri and Gennis, 1990). As expected, the three controls produced the enzyme activities as they were reported previously (Chepuri and Gennis, 1990). There was no detectable alkaline phosphatase activity in the cells bearing pUC19, which is the parental plasmid for the subunit H-PhoA fusion proteins. In addition, very high enzyme activity was observed in the cells containing pVPS142 while the cells with 90

Table 3.1.*1 Alkaline Phosphatase Activities in Subunit H-PhoA Fusion Proteins

Fusion Sequence at Fusion Fusion Specific Plasmid Junction S ite Activity

H34 CTC CAG ACC 2CCT GTT CTG 3 4 /6 3 774 H46 AAC GAG GAC CCT GTT CTG 4 6 /6 141 2LysH34 AAC TTC AAG AAG CTG GCG 3 4 /6 5,846 2LysH46 AAC TTC AAG AAG CTG GCG 4 6 /6 6,293 PUC19 Parental plasmid N/D pVPS142s CytD-PhoA 15 pVPE1725 CytB-PhoA 5,320

* Note: 1. For all proteins, 4 to 6 individual assays were performed. 2. In H34 and H46, the underlined bases indicate the subunit H sequence and non-underlined bases are for the phoA sequence. In 2LysH34 and 2 LysH46, the underlined sequences represent the nucleotides that code for two lysine residues inserted between amino acid residues 11 and 13 with the replacement of amino acid residue 12 (aspartic acid). 3. The first number indicates the last amino acid residue of subunit H and the second number represents the first amino acid residue of alkaline phosphatase. 4. Specific activity is expressed as alkaline phosphatase units/units of optical density of whole cells at 600 nm. 5. See reference: Chepuri and Gennis, 1990. 91 pVPE172 only produced low enzyme activity, The results shown in Table 3.1. suggests the topologies illustrated in figure 3.10. (Fig. 3.10.). Both H34 and H46 fusion proteins expressed in £. coli give rise to low alkaline phosphatase activity. This shows that their alkaline phosphatase moiety is in the cytoplasm. On the other hand, the 2LysH34 and 2LysH46 fusion proteins result in high alkaline phosphatase activity. This strongly suggest that the two lysine residues that were inserted in front of the transmembrane segment of H34 and H46, respectively, reverses membrane orientation of the hydrophobic domain and exposes the PhoA moiety to the periplasm. Since the phoA fusion protein H34, which lacks the negatively charged cluster, has its PhoA moiety in the cytosol, it shows that the negatively charged cluster is not required to prevent the carboxyl-terminal region from translocating across the membrane.

3.7. Conclusion In this work, we tested whether the membrane orientation of the subunit H protein of the photosynthetic reaction center from Rhodobacter sphaeroides (Fig. 1.8.) is determined by the acidic cluster immediately after the transmembrane segment of the protein (Fig. 3.1.). Our current data shows that subunit H protein from Rhodobacter sohaeroides can be expressed and inserted into the cytoplasmic membrane of coli. With N Periplasm

Plasma Membrane

Cytoplasm j+ N N

2LysH34 H46-PhoA. H34-PhoA. 2LysH46 -PhoA. High Wild type Low alkaline Low alkaline -PhoA. High alkaline subunit phosphatase phosphatase alkaline phosphatase H protein actitvity. activity. phosphatase activity. activity.

Figure 3.10. Summary of the predicted membrane topoplogy of the subunit H-PhoA fusion proteins in this stduy. (g^depicts the alkaline phosphatase moiety. 93 oligonucleotide-directed mutagenesis and PhoA gene fusion techniques, four different subunit H-PhoA fusion proteins were constructed and expressed in £.. coli. The membrane topology of each of the constructed fusion proteins were determined by measuring the alkaline phophatase activity in the cell. The results are shown in figure 3.10. H34 and H46 fusion proteins had low phosphatase activity suggesting that they oriented themselves with their PhoA domain in the cytoplasm. We believe , although without proof, that the amino terminal region can traverse the membrane in both cases despite the lack of negatively charged cluster for the H34 fusion proteins. The other two fusion proteins, 2LysH34 and 2LysH46 which had two positively charged residues introduced before the transmembrane segment had their alkaline phosphatase portion localized in the periplasm. Thus the transmembrane segment reverses its normal orientation (Fig. 3.10.). Our results suggested that the negatively charged cluster is not required to prevent the large carboxyl domain of subunit H from translocating to periplasm. It is possible that the determinants that affect the membrane topology of subunit H may reside in the conformation of the cytosolic globular domain.

Statistical studies by von Heijne (von Heijne 1986; von Heijne and Gavel, 1988) have shown that there is a correlation 94 between positively charged residues and the membrane topology. Positively charged amino acid residues tends to face the cytoplasmic side of the membrane while the region with more negatively charged residues are found with equal abundance on either side of the membrane (Fig. 1.4.). There are several reasons why positively charged residues, which remains in the cytosol, are determinants of protein topology (Boyd and Beckwith, 1990). First, basic residues may interact with the negatively charged lipid head groups of the membrane and prevent the carboxyl terminal domain following it from partitioning into the apolar fatty acid core of the membrane. Indeed, the positively charged amino and carboxyl-terminal regions of M13 procoat protein bind electrostatically to the negatively charged membrane surface (Gallusser and Kuhn, 1990). Second, the positive charges may make it energetically unfavorable to insert across the membrane because the membrane potential has a net positive charge at its outer surface (Zhu fiiai-, manuscript in preparation). Third, the positively charged residues of the polypeptide may not enter the membrane because it is difficult to bury these residues within the membrane. In contrast, introducing a charge cluster of four negatively charged amino acid residues after the leader peptide at the beginning of the mature region of OmpA had no effect on its export. Instead, these acidic residues promote the pro-OmpA 95 translocation (Zhu a ia l, manuscript in preparation). Subunit H protein, is unusual in that it has an acidic cluster after transmembrane segment in the cytosol. We have tested whether this acidic region functions to prevent the large carboxyl- terminal domain from transferring across the membrane. Our results show that this is not the case. It is possible that the inhibition of translocation may occur because this domain may quickly fold soon after the protein is synthesized.

In summary, the unusual position of the negatively charged cluster in the subunit H protein plays no role in preventing the large carboxyl-terminal domain from transferring across the membrane. The membrane orientation of this protein may be due to the rapid formation of the huge globular domain that follows the negatively charged cluster. CHAPTER IV

SECB-DEPENDENT AND SECB-INDEPENDENT INSERTION OF THE E. COLI LEADER PEPTIDASE

4.1. Introduction Leader peptidase is one of the central enzymes in the protein export machinery of coil. The physiological role of leader peptidase is to release the translocated preproteins from the periplasmic side of the plasma membrane. (Dalbey and Wickner, 1985; Fikes and Bassford, 1987). This protein has been chosen as a model system to study membrane protein topogensis in E. coli (Dalbey, 1990).

Leader peptidase is an integral transmembrane protein located in the cytoplasmic membrane of coli. It is comprised of two hydrophobic transmembrane segments H1 (5-22), and H2 (62-76), one cytoplasmic domain P1 (23-61), with the third hydrophobic segment H3 (83-98) and a long carboxyl-terminal domain (99-323) facing the periplasm (Fig. 4.1.; Wolfe e ia i-, 1983; Moore and Miura, 1987). When the first transmembrane segment H1 is deleted, the insertion of the carboxyl-terminal domain of leader peptidase is blocked (von Heijne filal-, 1988), 96 97

Periplasm

Plasma ec A/Y Membrane

Cytoplasm

Figure 4.1. A working model of leader peptidase insertion into the plasma membrane of coli. 98 while the mutant missing H1 plus a short polar region ( 4-50) inserts across the membrane with slightly delayed kinetics (Dalbey siai., 1987). The second hydrophobic segment is required for translocation of the large carboxyl-treminal domain (Dalbey and Wickner, 1987). It constitutes an uncleavable leader peptide (Dalbey M ai-, 1987). In addition, since removal of two of third of the carboxyl terminus blocks export there is presumably a minimal size requirement for translocation (Dalbey and Wickner, 1986). Like most exported proteins of £.. coli. leader peptidase requires the component of the protein export machinery in £.. coli. including SecA and SecY proteins (Wolfe fila l, 1985), SecE protein (Lee and Dalbey, unpublished data), and the proton motive force for its membrane assembly (Wolfe and Wickner, 1984). Finally, the conformation of the leader peptidase undergoes a dramatic change during the export. It is first synthesized in a loosely-folded conformation in the cytoplasm (Wolfe fiiai-, 1985). When the translocation starts, the protein undergoes a conformational change to an export- competent structure and assembles across the membrane.

Compelling lines of evidence have also suggested that many other exported proteins such as protein precursors of maltose binding protein (pre-MBP) and outer membrane protein A (pro-OmpA) are also synthesized in a loosely-folded 99 conformation before the export starts (Randall, 1983; Randall and Hardy, 1986; Collier fiia i.,1988). Besides the involvement of SecA, SecY, SecE and the proton motive force, the other important member of the protein export machinery, SecB, is often required for the translocation (Kumamoto and Beckwith, 1983; Kumamoto and Nault, 1988; Weiss fiiai-, 1990). SecB protein, together with groEL (Bochkareva fiiai, 1988, Martin fii a l, 1991) and DnaK (Phillips and Silhavy, 1990), has been defined as molecular chaperones because they help to promote preprotein export (Table 4.1.). SecB, a cytosolic protein that forms a tetramer with each identical subunit having a molecular weight of 17 kilodaltons, functions in at least two capacities in protein translocation. First, it binds to exported proteins, stabilizes the conformation of newly synthesized preprotein, thus prevents them from misfolding (Bankaitis and Bassford, 1984; Collier fiiai., 1988; Liu fiiai-, 1989; Lecker fiiai-, 1989). Second, SecB plays a role in targeting exported preproteins to the membrane by its interaction with membrane-bound SecA. This has been shown for several exported proteins, including OmpA (Hartl fiia i, 1990) and LamB (Swidersky fiiai-, 1990). It has been proposed that SecB behaves analogous to the eukaryotic signal recognition particle and binds to the leader peptide of precursor protein (Watanabe and Blobel, 1989). However, several laboratories (Randall fiiai-, 1990; Weiss and Bassford, 100

Table 4.1. Chaperones Involved in £. coli Protein Export

Chaperone Exported protein

SecB (16 KD-Tetramer) pre-MBP, AP, pro-OmpA GroE (L & S-Decatetramer) 6-lactamase DnaK (70 KD) LamB-LacZ hybrid protein 101

1990; Gunnon fiiai-, 1985; and Altman fiiai-, 1990) have recently reported that SecB binds to the mature region of pre-proteins.

Previous studies have examined the requirement for SecB protein in the export only for secretory proteins (pre-MBP and pro-OmpA). There are some similarities in the protein export between the well-studied pre-MBP, pro-OmpA and leader peptidase. They are all synthesized in a loosely-folded conformation before translocation starts and their export requires the Sec-components as well as proton motive force. Thus we asked, does leader peptidase require SecB protein for its carboxyl terminus insertion? In this project, we have investigated this possibility with the wild type leader peptidase as well as its mutants in SecB+ and SecB' £.. coli strains. We find that SecB protein is necessary for membrane assembly only when the native structure of leader peptidase is disturbed.

4.2. SecB Is Required for the Complete Translocation of Altered Leader Peptidase The gene constructs of leader peptidase used in this study are shown in figure 4.2. Two of them are altered leader peptidase, namely A 222-323 and A 4-50. A 222-323 lacks half of the periplamisc carboxyl-terminal domain of leader peptidase while A 4-50 is missing H1 plus a short polar region (P1). The 102

A. P2 H3 Periplasm

Plasma H1 H2 Membrane

Cytoplasm

Membrane B. Assembly W T. N-C 3—C - C SecB-lndependent Leader Peptidase

4222-323 N-C SecB-Dependent Leader Peptidase

44-50 C SecB-Dependent Leader Peptidase'

Figure 4.2. Gene constructs of leader peptidase used in

this study and their properties of membrane assembly. 103 membrane assembly properties of leader peptidase A222-323 (Dalbey and Wickner, 1986) and A4-50 (Dalbey fiiai-. 1986; Dalbey and Wickner, 1987) have been previously reported in a SecB+ strain. Each of these leader peptidases were cloned into the plNG-1 vector, a pBR322 derived plasmid carrying the arabinose regulatory element and araB promotor (Johnson fiiai-, 1985). The bacterial strains were HJM114 (F\ lac pro/lac pro), a SecB+ strain from our collection, and CK1953, a SecB" strain from Dr. Kumamoto (Tufts University). Translocation of the carboxyl terminus of leader peptidase was monitored by protease accessibility assay (Dalbey and Wickner, 1986). After induction with arabinose to express the protein in vivo, the bacterial cells were harvested, converted to spheroplasts by lysozyme treatment, and treated with trypsin to assay for translocation (Randal! and Hardy, 1986). The principle of this technique is illustrated in figure 4.3. In this assay, any protein exported across the plasma membrane is sensitive to the external added protease while the protein remaining inside the cells shows resistance.

We first tested the requirement of SecB for the wild type leader peptidase. CK1953 (SecB') strain was transformed with a plasmid pRD8, encoding wild type leader peptidase and grown at 37°C to the early log phase in M9 media containing 0.5% 104

inner membrane outer _ membrane

periplasm

Figure 4.3. The principle of protease accessibility assay. 1 05

fructose and 50 pg/ml of each amino acids except methionine. Arabinose (0.2%) was added to the media to induce the synthesis of the leader peptidase. After one hour, the cells were pulse- labeled with 100 jiCi of [35S]-methionine and subjected to the protease-accessibility assay. Following the treatment with protease, a polyclonal antibody to the leader peptidase was used for immunoprecipitation. The samples were then assayed by sodium dodecyl sulfate-poiyacrylamide gel electrophoresis (SDS-PAGE) and followed by fluorography (Ito fiiai., 1980). Figure 4.4 shows that the leader peptidase (Fig. 4.4., left panel) is digested by trypsin, indicating that the large carboxyl- terminal domain translocates across the plasma membrane even in the absence of SecB. As a control, we confirmed that the SecB-dependent protein outer membrane protein A (pro-OmpA) accumulates (Fig. 4.4., right panel) in the cytosol and is resistant to protease. In contrast, the small amount of mature OmpA, which runs as a doublet in this SDS-polyacrylamide gel, is digested by protease since this outer membrane protein is sensitive to external protease from the periplasmic surface. Ribulokinase (Fig. 4.4, left panel), a cytosolic protein, is not digested by protease, showing that the cytoplasmic membrane is still intact within spheroplasts.

We next tested leader peptidase in which the native Figure 4.4. Transtocation of the large carboxyl- terminal domain of leader peptidase is SecB-independent. coli strain CK1953 (SecB-) expressing wild-type leader peptidase was grown, labeled and converted to spheroplasts as described in the text. Aliquots of these spheroplasts were incubated on ice with either no addition, with trypsin (1 mg/m!) for 60 minutes, or with trypsin after the cells were disrupted by Triton X- 100. After addition of trypsin inhibitor (5 mg/ml) and PMSF (1 mM), samples were immunoprecipitated with antiserum to OmpA (right panel), and with antiserum generated against leader peptidase and ribulokinase (left panel) and then analyzed by SDS-PAGE and fluorography.

106 107

Pro-OmpA Leader - \ Peptidase :~-'1' Omp A r

Ribolukinase

Digestion Digestion Time 0 60 60 Time 0 60 60 (min) - + +fysis (min) - + +lysis

Figure 4.4. Translocation of the large carboxyl- terminal domain of leader peptidase is SecB- 1 08 structure of the protein was disturbed by a large deletion in its carboxyl-terminal domain. pRD9, which encodes leader peptidase A222-323, was transformed into a SecB' strain, CK1953, grown at 37°C, and induced with arabinose to express the mutant leader peptidase. Cells were then pulse-labeled with 100jiCi of [35S]-methionine for one minute and either directly quenched in ice-cold water (1 min pulse) or chased first with 100 pi of 5 mg/ml cold methionine for five minute (5 min chase). Samples were analyzed for translocation as described above and the amount of leader peptidase inserted was estimated by quantification of the bands by densitometry (Shen fila l-, 1991). In an 1 min pulse, leader peptidase A222-323 is completely inaccessible to protease in a SecB' stain (Fig. 4.5,A, left panel). Translocation is completely blocked even after a 5 min chase point (Fig. 4.5.B, right panel). This is in sharp contrast to A222-323 in the SecB+ strain, where most of the protein is accessible to protease at the 1 min pulse point (Fig. 4.5.B, left panel; also see Dalbey and Wickner, 1986).

The last question we examined was why the wild type leader peptidase, which requires SecA/SecY (Wolfe fiiai., 1985) and SecE (Lee and Dalbey, unpublished data), does not require SecB, whereas a number of other SecA/SecY/SecE dependent exported proteins, such as pro-OmpA (Lecker fiiai-, 1989), pre- LamB (Swidersky fiiai-, 1990) and pre-MBP (Kumamoto and 1 09

A. Sec B"

1 min pulse 5 mtn chase

LPase A222-323 - LPase A222-323

pro-O m pA^ pro-OmpA - OmpA - OmpA I Digestion Time 0 60 60 Digestion Time 0 60 60 (min) - + +lysis (min) + +lysis

B. Sec B+

1 mki pulse 5 min chase

LPase A222-323 - M B LPase A222-323

RibuloKinase Ribokrtinase - i + m Digestion Time 0 60 60 Digestion Time 0 60 60 (min) - + +lysis (min) * +lysis

Figure 4.5. SecB-dependent membrane insertion of leader peptidase A222-323 (pRD9). E. coli CK1953 (SecB-) and MC1061 (SecB+) bearing pRD9 was grown, labeled, and analyzed for assembly as described in te x t. 1 1 0

Beckwith, 1983 and 1985) do require SecB for the membrane translocation. Figure 4.6. shows a comparison between the membrane topology after these pre-proteins and leader peptidase translocate across the plasma membrane. For pre-MBP and pro-OmpA, after they translocate across the plasma membrane, both pre-proteins have their amino-terminal leader peptide, anchored to the membrane with the remaining mature region in the periplasm. After the leader peptide is cleaved off by leader peptidase, the mature proteins will be released from

the plasma membrane (Fig. 4.6.). In the case o f leader peptidase, in addition to its internal, uncleavable signal peptide (amino acid residue 50-83), it has another hydrophobic region H1 (4- 22), that may help target the protein to the plasma membrane and, therefore, prevents the protein from misfolding in the cytosol. Thus, this targeting role of H1 would make it unnecessary to need SecB. To test this idea, we asked whether leader peptidase A4-50, which lacks H1 (4-22) as well as a short polar domain (23-50), requires SecB for its translocation. A4-50 was transformed into the SecB' strain, CK1953, grown at 37°C, and the translocation of its carboxyl- terminal domain were analyzed, in an one minute pulse, 70% of leader peptidase A4-50 inserted across the membrane in a SecB' strain (CK1953, Fig.4.7.A, left panel), the same extent as with a SecB+ strain (MC1061, Fig. 4.7.B, right panel). However, while 111

Uncleavable / Leader \ Leader Peptide Peptide

Plasma Membrane

Cytoplasm O Pro-OmpA Pre-MBP Leader Peptidase

Figure 4.6. Membrane topology of pre-MBP, pro-OmpA and leader peptidase after the carboxyl-terminal domain translocated across the plasma membrane. 1 1 2

A. Sec B

1 min pulse 5 min chase LPase A4-50- LPase A4-50- ■■■“

Ributokinase- ^ — Ribulokinase- S t Digestion Digestion Time 0 60 60 Time 0 60 60 (min) - + +iysis (min) • + +lysis

B. Sec B+

1 min pulse 5 min chase

LPase A4-50- • LPase A4-50-

Ribulokinase Ributokinase- .«> Digestion Digestion Time 0 60 60 Time 60 60 (min) + +lysis (min) + +lysis

Figure 4.7. Leader peptidase A4-50 is SecB-dependent for complete translocation of its large carboxyl-terminal domain. EE. coli CK1953 (SecB-) and MC1061 (SecB+) bearing A4-50 was grown, labeled, and analyzed for membrane assembly as described in Fig. 4.5. 1 1 3 over 95% of A4-50 inserts across in the SecB+ strain in the 5 min chase point {Fig. 4.7.B, right panel; also see Dalbey and Wickner, 1987), still only 70% translocates in the SecB strain (Fig. 4.7.A, right panel). We confirmed this SecB* result with A4-50 in four independent studies, of which the average is 70% translocated at the 5 min chase point (panel B, right side). From the data in figure 4.7., we also believe that a small portion of leader peptidase is capable of translocating at the early stage of export in the absence of SecB because leader peptidase can insert co-translationally. In contrast, the post-translationally translocated leader peptidase does require SecB.

4.3. Conclusion In this project, we have analyzed the membrane assembly properties of the wild-type, A222-323 and A4-50 leader peptidase proteins in a SecB* strain. Our data show that wild- type leader peptidase can bypass the SecB requirement for insertion (Fig. 4.4.). However, leader peptidase switches from SecB-independent to SecB-dependent translocation when a large deletion is made in the carboxyl-terminal domain or when H1 plus a short polar domain is deleted (Figs. 4.5. and 4.6.). We suspect that the wild-type leader peptidase does not require the membrane targetting SecB (Hartl filal-, 1990) because it gets to the membrane soon after initiation of protein synthesis and thus 1 1 4 does not require SecB to keep it in a "loosely folded" conformation. Alternatively, other chaperones such as GroEL (Bochkareva fiiaJ.-, 1988) or DnaK (Phillips and Silhavy, 1990) have exchangeable functions as SecB and can promote insertion of the wild-type leader peptidase in a SecB' strain.

The requirement for SecB for A222-322 is absolute (Fig. 4.5.). Possibly this is due to the protein misfolding in the cytosol or folding too rapidly, such that the protein is incapable of inserting across the membrane. In contrast, the requirement for SecB for A4-50 is not absolute, as is the case for pro-OmpA (Kumamoto and Nault, 1989) and pre-MBP (Collier and Bassford, 1988). Even in the absence of SecB, over half of A4-50 can insert across the membrane in the 1 min pulse (Fig. 4.7.). We suspect that SecB is required to stabilize leader peptidase that assembles across the membrane post-translationally. This idea is currently being tested.

The dual functions of SecB, namely binding to exported pre-proteins to prevent them from misfolding and targeting them to the plasma membrane, make it belong to the family of the molecular chaperones. As it shows in Table 4.1., the other two members of the chaperone family in £.. coli are DnaK and GroEL proteins. In the case of GroEL, 14 subunits of GroEL form 115 a complex, which are arranged in two stacked seven-subunits rings with ATPase activities (Chandrasekhar eiai., 1986). The GroEL protein functionally cooperates with the GroES protein, a ring-shaped complex with seven subunits, to inhibit the ATPase activity of GroEL (Chandrasekhar at a)., 1986). The GroE proteins are required for lambda phage head assembly and pre-G- lactamase by transiently associating with newly synthesized "unfolded" proteins to prevent them from completely folding (Sternberg filal, 1973, Bochkareva fiiai-, 1988). The GroEL protein stabilizes the polypeptides in a conformation resembling the "molten globule" state which occurs at the surface of the GroEL protein complexes (Martin fiia l, 1991). More recently, an in vitro study was performed on the chaperonin-dependent folding of two proteins, reduced carboxy-methylated a- lactalbumin and rhodanese. This study revealed that DnaK, a heat-shock protein homologue of £. coli. recognizes the folding polypeptide as an extended chain and cooperates with another heat shock protein, DnaJ. This cooperation in turn stabilize an intermediate conformational state lacking ordered tertiary structure. In a dependent manner, GrpE (Echol, 1990) and ATP hydrolysis, the protein is then transferred to GroEL which acts catalytically in the production of the native state (Langer a ia i-, 1992) 1 1 6

For SecB protein, it can form stable, stoichiometric complexes in vitro with many pre-secretory proteins such as pro-OmpA, pre-PhoE, pre-MBP, and pre-LamB (Bochkareva s ia l, 1988; Crooke gjtal, 1988a, 1988b; Lecker filai-. 1989; Liu e i a l, 1989; Swidersky filai-- 1990). SecB forms a complex with the mature domain of precursor proteins (Bankaitis and Bassford, 1984; Collier 1988; Liu fiial-. 1989; Gonnon al ai , 1989; Lecker atai-, 1989), which modulates their folding and prevents aggregation. Furthermore, the antifolding role of SecB protein was demonstrated that alterations in the leader peptide as well as in the mature region of MBP slow the precursor protein folding totally eliminated the requirement of SecB protein in MBP export (Collier and Bassford, 1989).

In summary, the domains of leader peptidase located either in the long carboxyl-terminal domain or within the H1 plus a short polar region might function analogously to the antifolding factor such as SecB to prevent formation of a compact folded conformation of the remaining region of leader peptidase. Once these antifolding factors were deleted, then protein translocation across the cytoplasmic membrane became SecB- dependent, i.e. the leader peptidase switches the export pathway from SecB-independent to SecB-dependent because of the purturbation in its structure. Our current working model is that 1 1 7

SecB binds to the altered leader peptidase (here are A222-323 and A4-50), keeping them in a translocation-competent conformation, analogous to how it functions with other precursor proteins (Wickner fiial-, 1991). CHAPTER V

REQUIREMENT OF A MEMBRANE ANCHOR DEPENDS ON WHETHER OR NOT A PROTEIN IS SEC-DEPENDENT FOR MEMBRANE ASSEMBLY

5.1. Introduction In bacteria, there are three classes integral membrane proteins based on their orientation across the membrane (Wickner and Lodish, 1985; Fig. 5.1.). The first class, named bitopic, consists of proteins which span the membrane once with their carboxyl termini exposed to either cytoplasmic (type I) or periplasmic (type II) surface of the membrane. In the second, called oligotopic, proteins span the membrane twice with both amino and carboxyl termini on the same side. Finally, termed polytopic, proteins span the membrane multiple times and have more complex structures (Dalbey, 1990).

According to the signal hypothesis proposed by Blobel (Blobel, 1980), proteins generate their complex membrane topologies by a series of independent signals and stop-transfer sequences that start and stop translocation of the polypeptide 11 8 Bitopic Oligotopic I----- i r

Periplasm

i Plajma i i memprane i a a

Cytoplasm

N N

Figure 5.1. Three classes of integral membrane proteins based on their orientation. 1 20 chain, respectively (Fig. 5.2.). In this model, the stop-transfer segment is a passive topogenic element that plays no role in promoting translocation. Rather, it stops the translocation of the protein that was initiated by a proceeding signal sequence. Although proposed over a decade ago, the salient point of this model that there are independent integration segment have held up over the years for a large number of membrane proteins (von Heijne, 1986; von Heijne and Gavel, 1988; Boyd and Beckwith, 1990; von Heijne and Manoil, 1990).

In 1986, it became clear that insertion of all proteins do not necessarily follow this mechanism and for some proteins the idea of independent signals and stop-transfer segments do not hold. This was first shown by Kuhn and his colleagues by studying the precursor of M13 phage coat protein (Kuhn fiial., 1986). M i3 coat protein, a typical Type I bitopic membrane protein (Fig. 5.1.), is a small (50-residue) viral protein which spans the plasma membrane of infected cells prior to assembling to extruding virus particles (Wickner, 1976; Ohkawa and Webster, 1981). It is made as a precursor, termed procoat, with two hydrophobic domains, one a typical 23-residue amino terminal leader peptide and one a membrane anchor region in the mature part of the protein (Koning fiial-, 1975; Sugimoto fiial-, 1977). Surprisingly, both leader peptide and the membrane 121

Uncleaved Cleaveage Signal Stop site transfer

Periplasm

Plasma Membrane

Cytoplasm

Leader peptide Insertion domain

Figure 5.2. Signal hypothesis model of protein translocation. 1 22

anchor domains are required to initiate membrane insertion. Thus the second hydrophobic domain is not a simple stop- transfer segment. It was termed a hydrophobic helper because this segment along with the cleavable leader peptide promote insertion {Kuhn fiial-, 1986). Based on this and other work (Kuhn, 1987), it was proposed that the procoat protein inserts across the membrane as a loop, with both hydrophobic domains promoting insertion (Kuhn fiial, 1986, 1986b; Kuhn, 1987; Fig. 1.10.).

To a certain degree, the M13 phage coat protein work has been down played in the membrane research field because this protein was regarded as a special case. For instance, the M13 coat protein does not require SecA and SecY for its membrane assembly (Wolfe fiia l, 1985; Kuhn fiial-. 1987) and has been shown to insert spontaneously into membrane vesicles (Ohno- Iwashita and Wickner, 1983; Geller and Wickner, 1985). This unusual membrane assembly mechanism was commonly viewed to be due to the small size of the procoat protein (total 72 amino acid residues). However, it has been shown that it is not the overall size of the protein that determines the Sec- dependency (Kuhn, 1987; Kuhn fiial-. 1990); either increasing the length of the protein at the amino or carboxyl terminus does not alter its membrane assembly behavior. In contrast, when the 123 central region is extended by inserting additional amino acids between the amino terminal leader peptide and the membrane anchor then the protein becomes dependent on SecA and SecV for its insertion {Kuhn, 1988).

In this project, we have examined the requirement for a membrane anchor region for another protein that assembles in a Sec-independent manner. Recently, von Heijne (1989) has shown that leader peptidase, with an inverted topology, is a Sec- independent protein. This mutated leader peptidase is synthesized with two hydrophobic domains that are very close together (Fig. 1.10.). We find, like the procoat protein, that the second hydrophobic domain of the inverted leader peptidase is required for assembly, showing that the requirement for H2 of the native procoat is not an exceptional case. We next analyzed a M13 procoat protein (procoat 828) which has the distance between the two hydrophobic regions moved further apart and requires SecA/Y proteins for assembly (Kuhn, 1988). In contrast to the native procoat, this Sec-dependent procoat 828 does not require the second hydrophobic domain. This shows that the requirement of the membrane anchor depends on the Sec- dependency of insertion. 1 24

5.2. Inverted Leader Peptidase Requires Its Second Hydrophobic Domain for Membrane Assembly Previously it was shown that leader peptidase, with an inverted topology, inserts across the membrane independent of the Sec proteins (von Heijne, 1989). The inverted leader peptidase has two transmembrane segments which are separated by a short polar region (P1) comprised of 20 residues that faces the periplasm. The second hydrophobic domain is followed by a large polar carboxyl-terminal domain (P2) in the cytosol (Fig. 1.10.). To evaluate whether the second hydrophobic domain is required for membrane insertion, the following mutations were made (Fig. 5.3.). Phenylalanine 67 was replaced with an arginine, valine 69 was replaced with arginine, or residues 70-74 were deleted. Each of these mutations were made using oligonucleotide-directed mutagenesis, as described in section 2.12. Insertion of the short P1 domain across the membrane was assayed by the production of a shorter protease-resistance fragment upon treating with protease (see section 2.18.). Cells expressing leader peptidase R67 (Lep-inv. R67) were pulse- labeled for 1 min, chased for either 5 sec or 5 min, and then converted to spheroplasts. Aliquots were then treated without trypsin, with trypsin for 60 min or with trypsin after the cell membranes were disrupted with detergent. Figure 5.4. shows that Lep-inv R67 was resistant to protease, indicating that the 125

Inverted Leader Membrane Peptidase Assembly 4 22 61 76 1,W ildtype F- z z “ I C +

4 ?? 61 76 2. R64 N------1 F- ■Tr 1------c +

4 2 2 61 76 3 R67 F o ~ 1 C -

4 2 2 61 76 4. R69 isU^I __._F -an. I C -

4 2 2 61 76 5. R70 F - r z R 1------c +

4 2 2 61 76 6. R71 F R I------c +

4 2 2 61 76 7, R72 N———1 F -EZ R I------C +

4 2 2 61 * 76

8. A70-74 ------F -EZ z Y t j c -

Figure 5.3. Gene constructs of wild-type inverted leader peptidase and its mutants, R64, R67, R69, R70, R71, R72 and A 7 0 -7 4 . 1 26

Lep-inv. R67

O m pA

Lep-inv. R69

Lep— i f 3SdS:

O m pA - ^ t

Lep-inv. A70-74

Chase Time 5" 5' Trypsin + + + + Triton X-100 - + - - +

Figure 5.4. Substitution with arginine at positions 67, 69 (R67 and R69) or deletion of 70 to 74 (A70-74) blocks the membrane translocation of the inverted leader peptidase. coli strain MC1061 cells expressing the inverted leader peptidase mutants were grown, labeled, and analyzed for membrane assembly as described in the text. 127

P1 domain did not cross the membrane bilayer. Under these experimental conditions, outer membrane protein A (OmpA), was accessible to trypsin, indicating that the cells were converted to spheroplasts since OmpA, can only be digested from the periplasmic surface of the membrane. Leader peptidase with an arginine at position 69 was not accessible to protease. Similarly Lep-inv A70-74 does not assemble across the membrane {Fig. 5.4.). Therefore, the inverted leader peptidase requires H2 for membrane assembly, like the native M13 phage procoat protein.

5.3. Other Substitutions with Arginine in H2 Region Did Not Affect the Kinetics of the Inverted Leader Peptidase Membrane Insertion In contrast, other arginine mutations into the H2 region did not effect the kinetics of insertion of P1. Lep-inv with an arginine substituted for alanine 64 inserted across the membrane with the same kinetics as Lep-inv wild-type and was converted by trypsin to a lower molecular weight species (Fig. 5.5., compare panel B with A). Lep-inv with an arginine introduced at position 70 inserts across the membrane in a 1 min pulse, as a protected fragment is observed . However, no protected fragment was detected at the 5 min point (Fig. 5.5., panel C). It is not clear why a protected band is not seen at the 1 28

Lep-inv. W.T. Lep

B. Lep-inv. R64 L e p - ^

c. Lep-inv. R70 L e p -

D. Lep-inv. R71 Lep SB S

Lep-inv. R72

Lep- m — — • in f .t f * * ----

Chase Time S’ Trypsin + + Triton X-100 +

Figure 5.5. Substitution with arginine at other positions of the second hydrophobic segment of the inverted leader peptidase do not block the export of the P1 region. For experimental procedure, see Fig. 5.4. 1 29 later chase point. The inverted leader peptidase may be unstable and be degraded by proteases in the celt after insertion. Similar results were observed with Lep-inv. R71 and Lep-inv. R72 (Fig. 5.5., panel D and E). While a protected fragment is observed at the early chase point, it is not seen at the later chase point.

The kinetics of membrane assembly of lep-inv R72 confirmed that the arginine had little or no effect on insertion. As can be seen in Figure 5.6., the protected, lower molecular weight fragment shows up at the early chase point (5 second) and remains at the 2 minutes chase with non-radioactive methionine to allow the completely export of P1 (Fig. 5.6.). The fact that the protected band is not observed in the 5 min chase point (Figure 5.5) suggest that the large P2 domain is getting translocated to the periplasm. Possibly the membrane anchor domain has been destabilized by the introduced positively charged amino acid residue. This notion is currently being tested. Most likely the large P2 domain is translocated across the membrane in a Sec-dependent manner.

5.4. The Membrane Anchor Region Is Not Essential for Sec-Dependent Procoat Protein Export These studies with the inverted leader peptidase are consistent with the idea that the membrane anchor promotes 1 30

Lep - ■■ I I I

OmpA*

Chase Time 30“ 1* 2' Trypsin + + + + + + Triton X-100 + +

Figure 5.6. Kinetic study of the inverted leader peptidase mutant R72 that translocates across the membrane. 131 insertion when it is nearby the first hydrophobic domain and the insertion mechanism is Sec-independent. Possibly the membrane anchor is necessary for assembly only when insertion is Sec-independent. To test this idea, we asked whether the M13 procoat protein requires the second hydrophobic domain (H2) when its insertion mechanism is converted from Sec- independent to Sec-dependent. Kuhn (1988) showed that procoat inserts across the membrane in a Sec-dependent fashion when an OmpA fragment comprised of 174 amino acids is inserted into the central region between the membrane anchor and the leader peptide. This hybrid protein construct was termed procoat 828. Specifically, three oligonucleotide directed mutations were made in this procoat mutant: arginine was substituted for residue 30 in the center of H2 (procoat 828 OM30R) or residues 21-33 or 21-39 were deleted which removes most or all of H2 (procoat 828 A21-33 and A21-39, respectively, Fig. 5.7.). Cells expressing the procoat 828 and its mutants were pulse-labeled with [35S]-methionine for 1 min and chased with non-radioactive methionine for 5 sec or 5 min and quenched in trichloroacetic acid. As can be seen in Fig. 5.8, 50% of procoat 828 is cleaved to coat in the 5 min chase. Similarly procoat 828 OM30R, 828 A21- 33, and 828 A21-39 is cleaved to coat with approximately the same kinetics, showing that the membrane anchor region is not critical for cleavage (Fig. 5.8 ). LEADER COAT 1

-23 -1 +1 +50 ++ + . - — + + ++ + MKKSLVLK lASV AV ATLVPMLSF AI AEGDDP AK A AFNSLQ AS ATElYIGYAyAMW IVGATIGlI KLFKKFTSK AS Procoat -23 - 1+1 +50 ++ + ---- + + ++ + 1 Wild-type N------

174 A, A. of OmpA +50 -23 ++ ++ 2. Procoat 828

174 A. A. ot OmoA +50 -23 3. Procoat 828 ++ ++ OM30R

174 A. A. ot OmpA +50 -23 4. Procoat 828 ++ A21-33

174 A.A, of OmpA

■23 5. Procoat 828 ++ + ^ - — + - I ------A21 -39

Figure 5.7. Gene constructs of procoat-OmpA (procoat 828) wild-type, OM30R, A21 -33 and A21 -39 used in this study. 1 33

Procoat 828 Procoat 828 Procoat 828 Procoat 828 W ild-type OM30R A 21-33 A 21-39

'W t m s. 5BBH5 ^(55^

5’ 5" 5’ 5’ 5'

Figure 5.8. The membrane anchor region in procoat 828 and its mutants is not critical for the cleavage. 1 34

Protease accessibility was also used as an assay to determine whether the central region had crossed the membrane (see section 2.18.). Cells expressing procoat 828 A21-33 was pulse-labeled with [35S]-methionine for 1 min and chased with non-radioactive methionine for 5 sec or 5 min. Cells were subsequently converted to spheroplasts and procoat 828 was analyzed for protein translocation, as described before. Figure 5.9. shows that trypsin digests the processed procoat whereas the unprocessed molecules were resistant to proteolysis. Almost all of procoat is processed and translocated after 5 min of chase. Similar results were observed with procoat 828 A21- 39 in which the entire membrane anchor domain was deleted showing that it is not required for membrane assembly. Thus, in contrast to the native procoat protein that inserts in a Sec- independent manner, the membrane anchor region is dispensible for Sec-dependent procoat.

5.5. Insertion of the Inverted leader Peptidase Mutants Which Is Not Affected by Positive Charge Is SecA- Independent Previously, the wild type inverted leader peptidase was shown to insert across the membrane independent of the Sec proteins (von Heijne, 1989). To ascertain this Sec-independent property still hold on for the mutants with the positive charge 1 35

Pro 828 coat A21 -33 ——* V * ! . ■ 828 coat A21-33 - r * I 1 Chase Time 5" S' Trypsin + + + + Triton X-100 + - +

Pro 828 coat A21-39 - § - *"•* III* 828 Coat A21 -39 - L_ _ d Chase Time 5" 5* Trypsin + + + + Triton X-100 + +

Figure 5.9. Deletion on the second transmembrane domain of procoat-OmpA 828 dose not impair their export across the plasma membrane. 136 in H2 region, two mutants were selected for following study. They were Lep-inv. R70 and R72. The plasmids bearing these two mutated Lep-inv. genes were transformed into two sec gene temperature sensitive E. coli strains, CJ105 and CJ107 respectively. CJ105 is a SecA temperature sensitive strain and CJ107 is a SecY temperature sensitive strain. At non- permissive temperature (42°C), the expression of the Sec proteins will be shut down, so we can assay the effect of the Sec proteins in the insertion of translocated protein. The cells expressing Lep-inv. R70 and R72 were grown at 30°C for two hours to reach the mid log phase and then shifted to 42°C for four hours to express the proteins. The protein insertion was then assayed as described before. Unfortunately, the mutants were not stable in these two Sec temperature sensitive strains after two minutes chase point. But the insertion of P1 region of these mutants still can be observed at 1 minute pulse-label point (Figs. 5.10. and 5.11.). To overcome this problem, we then took advantage of the observation that the expression of SecA protein can be completely knocked out in SecA wild type strain, MC1061, by briefly treating the cells with sodium azide (Oliver filai-. 1990). Thus the MC1061 cells bearing Lep-inv. R72 were grown and induced at 37°C as described earlier, then sodium azide was added to the final concentration of 0.2 mM. After one minute growing, the cells were pulse-labeled for 1 minute with 137

A. CJ105 Lep mv. R7C Lep— ”

pro-OmA — m m - - O pm A ^ Lep-inv. R72 Lep —

pro-OmA — —— OpmA'"^

B. CJ107 Lep-inv. R70 Lep —

pro-OmA — Opm A

Lep-inv. R72 Lep —

pro-OmA M m m OpmA-^ Chase Time 5 5*

T rypsin *■ + + + lysis tlysis

Figure 5.10. Inverted leader peptidase mutants R70 and R72 are not stable in SecA15 (CJ105) and SecY,s (CJ107) E.. coli strains. For the experimental procedure, see the text. 1 38

A. CJ105 Lep-inv. R70 Lep-inv. R72

Lep~ ft ~ ^ flA -* I * • • — 0 m »

CJ107 Lep— *■ - .,, M tk « ** -t—

Chase Time 30" 1* 21 5* 30" V 2' 5"

MC1061+NaAz treatment

Lep-inv. R72 Lep — ■ 1 i - pro-OmA _ Q p m A ^ ; f g Chase Time 5" 5' Trypsin - + + + + + lysis 4-lysis

Figure 5.11. Instability of R70 and R72 in SecAls and SecY1s strains (A) and Sec-independent export of the inverted leader peptidase mutant R72 (B). 1 39

[35S]-methionine and chased with non-radioactive methionine for 5 minutes. The insertion of the protein was monitored in the same fashion as before. Figure 5.11. shows the effect of sodium azide treatment can be observed in the blockade of OmpA protein export since the translocation of this protein is totally dependent on the SecA (Cabelli fit at, 1988; Cunningham a la t, 1989). The accumulation of the pro-OmpA can be seen at 5 second chase point and the translocated mature OmpA is digested by external added trypsin. After 5 minute chase, all the pro-OmpA are converted to the mature OmpA and exported (Fig. 5.10., panel B). For Lep-inv. R72 mutant, the insertion of P1 domain can be observed clearly at both 5 second and 5 minutes chase points (Fig. 5.10., panel B). Thus, the non-effect insertion of Lep-inv. mutant behaved in the same fashion as its wild type counterpart did under the SecA deficient condition.

5.6. Conclusion Bitopic and poiytopic membrane proteins have been proposed to contain start (Nir|-Cout) and stop-transfer elements (N -C ) that function to start and stop the translocation of the v out in' r polypeptide chains across the membrane (Blobel, 1980). This model, to a large extent, has been supported by the discovery of topogenic sequences. There are two kinds of topogenic sequences that start carboxyl terminus transiocation, namely 140 leader peptides which is cleaved after the protein precursor is exported (Blobel, 1980; Wickner and Lodish, 1985) and uncleavable signal peptides which also mediate protein translocation but remains in the mature protein serving as a transmembrane anchor (Blobel, 1980; Bos elai-, 1984; Dalbey and Wickner, 1987); two types that start amino terminus translocation, termed reverse signal (von Heijne, 1988 and 1990) and a hydrophobic helper (Kuhn siai-, 1986; Kuhn a la l, 1986b); and one type that stops the transfer of the polypeptide chain by anchoring the protein to the membrane, namely stop- transfer (Davis and Model, 1985). Their important properties and functions are now being examined by using molecular genetics.

What determines whether a membrane anchor domain (Nout- C|n) functions as a stop-transfer that halts translocation or a hydrophobic helper element that helps promote translocation? Our data here suggest that the membrane anchor domain functions as a hydrophobic helper domain for Sec-independent proteins and a stop-transfer segment for Sec-dependent proteins. For example, the Sec-independent leader peptidase, with an inverted topology, requires the membrane anchor domain for insertion (Fig. 5.4.), as was observed previously for the Sec- independent procoat Kuhn, 1988). However, when procoat 141 becomes dependent on the Sec proteins for assembly, then the second hydrophobic domain, which is essential for Sec- independent insertion, is not required (Figs. 5.8. and 5.9.). In this case, the membrane anchor is a stop transfer domain.

Our working model of Sec-dependent and Sec-independent membrane insertion for procoat and inverted leader peptidase is following. When the two hydrophobic domains are close together then the protein inserts in a Sec-independent manner presumably because the protein cannot engage the translocation machinery. When the two hydrophobic segments are far apart then the protein engages the Sec-machinery in an unknown fashion. One possible reason why a Sec-dependent protein can initiate translocation of a hydrophilic region with only a signal peptide is because translocation is assisted by the Sec-components. For example, the SecA protein has been shown to target preproteins to the membrane where upon the membrane protein can engage the translocase to catalyze the transfer of the hydrophilic regions of the membrane protein across the membrane (Schiebel Alai-, 1991). Nevertheless, when the protein does not interact with the Sec components, then it must spontaneously insert directly across the membrane. It is energetically unfavorable for a polar region to cross the apolar lipid bilayer {Wickner al ai., 1991). Under these set of conditions, the signal peptide, by 1 42 itself, does not provide enough driving force to translocate the polar region across the membrane. However, it is energetically favorable for the signal peptide, together with the membrane anchor segment, to insert the hydrophilic region across the membrane.

It is noteworthy that the second hydrophobic domain is required for assembly in leader peptidase both in its normal (Wolfe alai-, 1985) and inverted topology. While in the inverted topology it functions as a hydrophobic helper, it acts as uncleaved signal in its normal orientation. Thus, in either orientation, this hydrophobic domain plays an active role in membrane insertion. However, there are some interesting differences that we observe. As a hydrophobic helper, an arginine blocked insertion at residue 67 or 69, whereas these mutations had no or a slight effect when H2 was acting as a signal peptide (Zhu and Dalbey, 1988). In addition, an arginine at position 70, 71 and 72 had very little effect on the inverted leader peptidase; this mutation within the native leader peptidase, on the other hand, blocked translocation of the P2 domain. REFERENCES

Allen, J., Feher, G., Yeates, T., Komiya, H., and Rees, D., (1987a) Proc. Natl. Acad. Sci. USA 84, 5730-5734.

Allen, J., Feher, G., Yeates, T., and Rees, D., (1987b) Proc. Natl. Acad. Sci. USA, 84, 6162-6166.

Altman, E., Emr, S. D. and Kumamoto, C. A. (1990) J. Biot. Chem. 265, 18154-18160.

Andersson, H. and von Heijne, G. (1991) Proc. Natl. Acad. Sci. USA, 88, 9751-9756.

Andersson, H, Bakker, E. and von Heijne, G. (1992) J. Biol. Chem. 267, 1491-1495.

Beckwith, J., and Stlhavy, T. J. (1983) Methods Enzymol. 97, 3-11.

Bankaitis, V. A., Ryan, J. P.. Rasmussen, B. A. and Bassford, P. J. Jr. (1985) Curr. Top. Memb. Transp. 54, 105-150.

Bassford, P. J., Silhavy, T. J. and Beckwith, J. R. (1979) J. Bacteriol. 139, 19-31.

Bedwell, D. M., Klonsky, D. J., and Emr, S. D. (1987) Mol. Cell. Biol. 7, 4038-4047.

143 144

Bedouelle, H., Bassford, P. J. Jr., Fowler, A. V., Zabin, I. Beckwith, J. and Hofnung, M. (1980) Nature, 285, 78-81.

Benson, S. A., Hall, M. N., and Silhavy, T. J. (1985) Annu. Rev. Biochem. 54, 101-134.

Blobel, G. and Dobberstein, B. (1975) J. Cell Biol. 67, 835-851.

Blobel, G. and Dobberstein, B. (1975) J. Cell Biol, 67, 852-862.

Blobel, G. and Sabatini D. D. (1970) J. Cell Biol. 67, 835-851.

Blobel, G. and Sabatini D. D. (1971) Biomembranes 2, Plenum, New York, 193-195.

Blobel, G. (1980) Proc. Natl. Acad. Sci. USA, 77, 1496-1500.

Bochkareva, E. S., Lissin, N. M. and Girshovich, A. S. (1988) Nature, 336, 254-257.

Boeke, J. F., Russel, M., and Model, P., (1980) J. Mol. Biol., 144, 103-1 16.

Boquet, P. M., Manoil, C. and Beckwith, J. (1987) J. Bacteriol. 169, 1 6 6 3 -1 6 6 9 .

Bos, T. J., Davis, A. R., and Nayak, D. P. (1984) Proc. Natl. Acad. Sci. USA, 81, 2327-2331.

Boyd, D., Manoil, C. and Beckwith, J. (1987) Proc. Natl. Acad. Sci. USA, 84, 8525-8529. 145

Boyd, D. and Beckwith, J. (1989) Proc. Natl. Acad. Sci. USA. 86, 9 4 6 6 -9 4 5 0 .

Brickman, E. and Beckwith, J. (1975) J. Mol. Biol. 96, 307-316.

Briggs, M. S. and Gierasch, L. M. (1986) Adv. Prot. Chem. 38, 109- ISO.

Brundage, L., Hendrick, J. P., Schiebel, E,, Driessen, A. J. M., and Wickner, W.(1990) Cell, 62, 649-659.

Brusilow, W. S. A., Gunsalus, R. P., and Simoni, R.D. (1983) Methods Enzymol. 97, 188-195.

Carlson, M., and Botstein, D. (1983) Cell 28, 145-154.

Chamberlain, J. P. (1979) Analyt. Biochem. 98, 132-135.

Chang, C. N., Blobel, G. and Model, P. (1978) Proc. Natl. Acad. Sci. USA, 75, 361-365.

Chang, C. N., Inouye, H., Model, P. and Beckwith, J. (1980) J. Bacteriol., 142, 726-728.

Chang, C. N., Kuang, W.-J., and Chen, E. Y. (1986) Gene (Amst) 44, 1 2 1 -1 2 5 .

Chen, L., and Tai, P. C. (1985) Proc, Natl. Acad. Sci. USA, 82, 4 3 8 4 -4 3 8 8 .

Clement, J. M., and Hofnung, M. (1981) Cell, 27, 507-514. 1 46

Cobet, W., Moll ay, C., Mullar, G., and Zimmarma nn, R. (1981) J. Biol. Chem. 264, 10169-10176.

Cohen, S. N., Chang, A. C. Y., Boyer, H. W, and Heilling, R. B. (1973) Proc. Natl. Acad. Sci. USA, 70, 3240-3244.

Collier, D. N., Bankaitis, V. A., Weiss, J. B., and Bassford, P. J. (1988) Cell, 53, 273-283.

Crooke, E , Guthrie, B., Lecker, S., Lill, R. and Wickner, W. (1988) Cell, 54, 1003-1011.

Dalbey, R. E., and Wickner, W. (1985) J. Biol. Chem. 260, 15925- 15931.

Dalbey, R. E. and Wickner, W. (1986) J. Biol. Chem. 261, 13844- 13849.

Dalbey, R. E. and Wickner, W. (1987) Science, 235, 783-787.

Dalbey, R. E., Kuhn, K., Kreil, G.. Rice, M. C. and Wickner, W. (1987) Chemica Scripts, 27B, 245-248.

Dalbey. R. E., Kuhn, A., and Wickner, W. (1987) J. Biol. Chem. 262, 1 3241-13245.

Dalbey, R. E. and Wickner, W. (1988) J. Biol. Chem. 263, 404-408.

Dalbey, R. E. (1990) TIBS, 15, 253-257.

Dallas, W. S., and Falkow, S. (1980) Nature, 288, 499-501. 147

Date, T., Goodman, J. M. and Wickner, W. (1980) Proc. Natl. Acad. Sci. USA,77, 4669-4673.

Davis, N. G. and Model, P. (1985) Cell, 41, 607-614.

Deisenhofer, J., Epp, O., Miki, K„ Huber, R., and Michel, H., (1985) Nature, 318, 618-624.

Deshaies, R. J., Koch. B. D.p Werner-Washburne, M. Craig, E. and Schekman, R. (1988) Nature, 332, 800-805.

Drews, G. and Oleze, J. (1981) Cell Function and Differentiation. Edited by Akoyuroglou, G. 247-256. Alan R. Liss, Inc., New York.

Donohue, T., Hoger, J., and Kaplan, S., (1986) J. Bacteriol. 168, 9 5 3 -9 6 1 .

Duplay, P., Bedouelle, H., Fowler, A., Zabin, I., Saurin, W., and Hofnung, M. (1984) J. Biol. Chem. 259, 10606-10613.

Echols, H. (1990) J. Biol. Chem., 265. 14697-14700.

Ehamann, M., Boyd, D. and Beckwith, J. (1990) Proc. Natl. Acad. Sci. USA. 87, 7574-7578.

Eisenberg, D., Schwarz, E., Komaromy, M. and Wall, R. (1984) J. Mol. Biol. 179, 125-142.

Emr, S. D., Schwartz, M., and Silhavy, T. J. (1978) Proc. Natl. Acad. Sci. USA, 75, 5802-5806.

Emr, S. D., Hedgpeth, J., Clement, J. M., Silhavy, T. J., and Hofnung, M. (1980) Nature, 285, 82-85. 148

Emr, S. D., Hanley-Way, S. and Silhavy, T. J. (1981) Cell, 23, 79- 88.

Emr, S. D., and Bassford, P. J. Jr. (1982) J. Biol. Chem. 257 5 8 5 2 -5 8 6 0 .

Emr, S. D. and Silhavy, T. J. (1983) Proc. Natl. Acad. Sci. USA, 80, 4 5 9 9 -4 6 0 3 .

Feher, G., Allen, J., Okamura, M., and Rees, D., (1989) Nature, 339, 1 1 1 -1 1 6 .

Ferenci, T. and Sihavy, T. (1987) J. Bacteriol. 169, 5339-5342.

Galluser, A. and Kuhn, A. (1990) EMBO J., 9, 2723-2729.

Geller, B. L. and Wickner, W. (1985) J. Biol. Chem. 260, 13281- 13285.

Geller, B. L., Movva, N. R., and Wickner, W. (1986) Proc. Natl. Acad. Sci. USA, 83. 4219-4222.

Geller, B. L., and Green, H. M. (1989) J. Biol. Chem. 264, 16465- 16469.

Groarke, J. M., Mahoney, W. C,, Hope, J. N., Furlong, C. E., Robb, F. T., Zalkin, H., and Hermodson, M. A. (1983) J. Biol. Chem. 258, 12952- 12956,

Guan, C., Wanner, B. and Inouye, H. (1983) J. Bacteriol., 156, 710- 717. 1 49

Haeuptle, M. T., Flint, N., Gough, N. M. and Dobberstein, B. (1989) J. Cell. Biol. 108, 1227-1236.

Hartmann, E., Rapoport, T. A. and Lodish, H. F. (1989) Proc. Natl. Acad. Sci. USA, 86, 5786-5790.

Hartl, F.-U., Lecker, S., Schiebel, E., Hendrick, J. P. and Wickner, W. (1990) Cell, 63, 269-279.

Hartl, F.-U., and Neupert, W. (1990) Science, 247, 930-938.

Hackett, J., and Reeves, P. (1983) Nucleic Acids Res. 11, 6487- 6495.

Hemmingsen, S. M., Woolford, C., van der Vies, S. M., Tilly, K. Dennis, D. T., Georgopoulos, C. P., Hendrix, R. and Ellis, R. J. (1988) Nature, 333, 330-334.

Hoffman, C. and Wright, A. (1985) Proc. Natl. Acad. Sci. USA, 82, 5 1 0 7 -5 1 1 1 .

Hoger, J. H. and Kaplan, S. (1985) J. Biol. Chem. 260, 6932-6937..

Hoger, J., Chory, J., and Kaplan, S., (1986) J. Bacteriol., 165, 942- 9 5 0

Inouye, H. and Beckwith, J. (1977) Proc. Natl. Acad. Sci. USA, 74, 1 4 4 0 -1 4 4 4 .

Inouye, M., and Halegoua, S. (1980) CRC Crit. Rev. Biochem. 7, 3 3 9 -3 7 1 . 1 50

Inouye, H., Barnes, W. and Beckwith, J, (1983) J. Bacteriol. 149, 4 3 4 -4 3 9 .

Inokuchi, K., Mutoh, N., Matsuyama, S., and Mizushima, S. (1982) Nucleic Acids Res. 10, 6957-6968.

Ito, K., Date, T. and Wickner, W. (1980) J. Biol. Chem. 255, 2123- 2130.

lino, T., Takahashi, M. and Sako, T. (1987) J. Biol. Chem. 262, 7 4 1 2 -7 4 1 7 .

Johnson, S., Lee, J. H. and Ray, D. S. (1985) Gene (Amst) 34, 137- 146

Kaplan, S. (1978) The Photosynthetic Bacteria. Edited by Clayton, R. and Sistron, W. S. 809 - 835. Plenum Publishing Corp., New York.

Kaplan, S. and Arntzen, C. J. (1982) Photosynthetic Energy Convertion by Plants and Bacteria. Edited by Govindjee, D. Academic Press, Inc., New York.

Kaiser, C. A., Preuss, D., Grisafi, P., and Botstein, D. (1987) Science, 235, 312-317.

Kendall, D. A., Bock, S. C. and Kaiser, E. T. (1986) Nature, 321, 7 0 6 -7 0 8 .

Kiley, P. J. and Kaplan, S. (1988) Microbiol. Rev. 52, 50-69.

Kim, E., and Wyckoff, H., (1991) J. Mol. Biol. 218, 449-464. 151

Konings, R. N. H, Hulsebos, T. and Hondel, C. A. C. (1975) J. Virol. 15, 570-578.

Kuhn, A. and Wickner, W. (1985) J. Biol. Chem., 260, 15914- 15918.

Kuhn, A., Kreil, G. and Wickner, W., (1986) EMBO J. 5, 3681-3685

Kuhn, A., Wickner, W. and Kreil, G , (1986) Nature, 322, 335-339

Kuhn, A. (1987) Science, 228, 1413-1415.

Kuhn, A., Kreil, G. and Wickner, W. (1987) EMBO J. 6, 501-505.

Kuhn, A. (1988) Eur. J. Biochem., 177, 267-271.

Kuhn, A., Zhu, H. Y., and Dalbey, R. E. (1990) EMBO J., 9, 2385- 2389.

Kumamoto, C.A. and Gannon, P.M. (1988) J. Biol. Chem., 263, 11554-11558.

Kunkel, T. A. (1985) Proc. Natl. Acad. Sci. USA, 82, 488-492.

Kusukawa.N., Yura.T. Ueguchi, C., Akiyama.Y. and lto,K. (1989) EMBO J., 8, 3517-3521.

Langer, T., Lu, C., Echols. H., Flanagan, J. Hayer, M. K. and Hartl, F. U. (1992) Nature, 356, 683-689.

Laws, J. K. and Dalbey, R. E. (1989) EMBO J., 8, 2095-2099. 1 52

Lecker, S.H., Driessen, A.J.M. and Wickner, W. (1990) EMBO J. 9, 2 3 0 9 -2 3 1 4 .

Li, P., Beckwith, J. and Inouye, H. (1988) Proc. Natl. Acad. Sci. USA, 85, 7685-7689.

Lill, R., Cunningham, K.t Brundage, L. A., Ito, K., Oliver, D., and Wickner, W. (1989) EMBO J. 8, 961-966.

Lill, R., Dowhan, W., and Wickner, W., (1990) Cell, 60, 271-280.

Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning, A Laboratory Manual, Cold Spring Laboratory, Cold Spring Harbor, NY

Martin, J., Langer, T., Boteva, R., Schramel, A., Horwich, A. and Hartl, F.-U. (1991) Nature, 36-42.

Manoil, C. and Beckwith, J. (1985) Proc. Natl. Acad. Sci. USA, 82, 8 1 2 9 -8 1 3 3 .

Manoil, C. and Beckwith, J. (1986) Science, 233, 1403-1408

Manoil, C., Boyd, D. and Beckwith, J. (1988) Trends Genet. Sci., 4, 2 2 3 -2 2 6 .

Manoil, C. (1990) J. Bacteriol. 172, 1035-1042.

McGovern, K. and Beckwith, J. (1991) J. Biol. Chem. 266, 20870- 20876.

McGovern, K., Ehrmann, M. and Beckwith, J. (1991) EMBO J., 10, 2 7 7 3 -2 7 8 2 . 153

Michaelis, S. and Beckwith, J. (1982) Science, 233, 1403-1408.

Michaelis, S. and Beckwith, J. (1982) Ann. Rev. Microbiol. 36, 4 3 5 -4 6 5 .

Michaelis, S., Inouye, H., Oliver, D. and Beckwith, J. (1983) J. Bacteriol. 154, 366-374

San Millan, J. L. , Boyd, D., Dalbey, R., Wickener, W. and Beckwith, J. (1989) J. Bacteriol. 171, 5536-5541.

Miller, J. H. (1972) Experiments in Molecular Genetics, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY.

Milstein, C., Brownlee, G.G., Harrison, T. M. and Matthews, M. B. (1972) Nature, 239, 117-120.

Mizuno, T., Chou, M.-Y., and Inouye, M. (1983) FEBS Lett. 151, 159- 164.

Moore, K. E., and Miura, S. (1987) J. Biol. Chem. 262, 8806-8813.

Movva, N.R., Nakamura, K., and Inouye, M. (1980) J. Mol. Biol. 143, 3 1 7 -3 2 8 .

Meyer, D.l. (1988) Trends Biochem. Sci., 13, 471-474.

Nakamura, K., and Inouye, M. (1979) Cell, 18, 1109-1117.

Nilsson, I. and von Heijne, G. (1990) Cell, 62, 1135-1141. 1 54

Nilsson, I. and von Heijne, G. (1991) J. Biol. Chem. 266, 3408- 3410.

Ohkawa, I. and Webster, R. E. (1981) J. Biol. Chem. 256, 9951- 9958.

Ohno-lwashita, Y. and Wicker, W. (1983) J. Biol. Chem. 258, 1 8 9 5 -1 9 0 0 .

Oliver, D. (1985) Ann. Rev. Microbiol. 39, 615-648.

Oliver, D,,Cabelli, R. J., Dolan, K. M. and Jarosik, G. P. (1990) Proc. Natl. Acad. Sci. USA, 87, 8227-8231.

Overbeeke, N., Bergmans, H., van Mansfeld, F., and Lugtenberg, B. (1983) J. Mol. Biol. 163, 513-532.

Peterson, G. (1977) Analyt. Biochem. 83, 346-356.

Pfanner, N. and Neupert, W. (1987) J. Biol. Chem. 262, 7528.

Pfaller, R., Pfanner, N., and Neupert, W. (1989) J. Biol. Chem. 264, 34 -38.

Park, S., Liu, G., Topping, T. B., Cover, W. H. and Randall, L. L. Science, 239. 1033-1036.

Perlman, D., and Halvorson, H. O. (1983) J. Mol. Biol. 167, 391- 409.

Pugsley, A. P. (1989) Protein Targeting. Academic Press Inc., San Diego, CA. 155

Puziss, J.-L., Fikes, J. D., and Bassford, P. J. (1989) J. Bacteriol. 171, 2303-2311.

Randall, L. L. (1983) Cell, 33, 321-340.

Randall, L. L. and Hardy, S. J. S. (1986) Cell, 46, 921-928.

Randall, L. L., Hardy, S. J. S., and Thom, J. R. (1987) Ann. Rev. Microbiol. 41, 507-541.

Randall, L. L., Topping, T. B., and Hardy, S. J. S. (1990) Science, 248, 860-863

Rapoport, T.A. (1986) CRC Crit. Rev. Biochem. 20, 73-137. Rohrer, J. and Kuhn, A. (1990) Science, 250, 1418-1421,

Ryan, J. P., Duncan, M. C., Bankaitis, V. A. and Bassford, P. J., Jr. (1986) J. Biol. Chem. 261, 3389-3395.

Sander, G., Marsh, R. C., Voigt, J. and Parmeggiani, A. (1975) Biochem istry, 14, 1805-1814.

Sanger, F., Nicken, S., Coulson, A. R. (1977) Proc. Natl. Acad. Sci. USA, 74, 5463-5467.

Sasaki, S., Matsuyama, S.I., and Mizushima, S. (1990) J. Biol. Chem. 265, 4358-4363.

Schlesinger, M. J., Reynolds, J. A. and Schlesinger, S. (1969) Ann. N. Y. Acad. Sci. 166, 368-379.

Schnaitman, C. A. (1970) J. Bacteriol. 104, 890-901. 1 56

Shen, L. M., Lee, J. I., Cheng, S. Y., Jutte, A., Kuhn, A. and Dalbey, R. E. (1991) Biochemistry, 30, 11775-11781.

Silhavy, T. J., and Beckwith, J. (1985) Microbiol. Rev. 49, 398- 418..

Silhavy, T. J., Shuman, H. A., Beckwith, J. and Schwartz, M. (1977) Proc. Natl. Acad. Sci. USA, 74, 5411-5415.

Sivasubramanian, N. and Nayak, D. P. (1987) Proc. Natl. Acad. Sci. USA, 84, 1-5.

Sockett, R., Donohue, T., Varga, A., and Kaplan, S., (1989) J. Bacteriol. 171, 436-446.

Sollner, T., Griffith, R., Pfaller, N. and Neupert, W. (1989) Cell, 59, 1061 Spicer, E. K. and Nobel, J.A. (1982) J. Bio. Chem. 257, 5716-5721.

Spiess, M. and Lodish, H. F. (1986) Cell, 44, 177-185.

Spiess, M. and Handschin, C. (1987) EMBO J., 6, 2683-2691.

Sugimoto, K., Sugisaki, H., Okamoto, T. and Takanami, M. (1977) J. Mol. Biol. 110, 487-509.

Summers, R. G. and Knowles, J. R. (1989) J. Biol. Chem. 264, 20074-20081.

Summers, R. G., Harris, C. R. and Knowles, J. R. (1989) J. Biol. Chem. 264, 20082-20088.

Sutcliffe, J. G. (1978) Proc. Natl. Acad. Sci. USA, 75, 3737-3741. 1 57

Swidersky, U. E., Hoffschulte, H. K. and Muller, M. (1990) EMBO J. 9, 1777-1 785.

Thom, J. R. and Randall, L. L. (1988) J. Bacteriol. 170, 5654-5661.

Tokunaga, M., Loranger, J. M., Wolfe, P. nd Wu, H. C. (1982) J. Biol. Chem. 257, 9922-9925.

Verga, A. R. and Kaplan, S. (1989) J. Bacteriol. 171, 5830-5839

Vlasuk, G. P., Inouye, S., Ito, H., Itakura, K. and Inouye, M. (1983) J. Biol. Chem. 258, 7141-7148. von Heijne, G. and Blomberg, C. (1979) Eur. J. Biochem. 97, 175- 181. von Heijne, G. (1983) Eur. J. Biochem. 133, 17-21. von Heijne, G. (1984) EMBO J. 3, 2315-2318. von Heijne, G. (1985) J. Mol. Biol. 184, 99-105. von Heijne, G. (1986) J. Mol. Biol. 192, 287-290. von Heijne, G., Wickner, W. and Dalbey, R. E. (1988) Proc. Natl. Acad. Sci. USA, 85, 3363-3366. von Heijne, G. and Gavel, Y. (1988) Eur. J. Biochem. 174, 671-678. von Heijne, G. (1989) Nature, 341, 456-458. von Heijne, G. and Manoil, C. (1990) Protein Enginerr. 4, 109-112. 1 58

Walter, P., & Lingappa, V. (1986) Annu. Rev. Cell Biol. 2, 499-516.

Watts, C., Silver, P. and Wickner, W. (1981) Cell, 25, 347-353.

Watts, C.,.Wickner, W. and Zimmermann, R. (1983) Proc. Natl. Acad. Sci. USA, 80, 2809-2813.

Weiss, J. B., Ray, P. H. and Bassford, P. J. Jr. (1988) Proc. Natl. Acad. Sci. USA, 85, 8978-8982.

Wickner, W. (1976) Proc. Natl. Acad. Sci. USA, 73, 1159-1163.

Wickner, W. (1979) Ann. Rev. Biochem. 48, 23-45.

Wickner, W. and Lodish, H. (1985) Science, 230, 400-407.

Wickner, W. (1988) Biochemistry, 27, 1081-1086.

Wickner, W., (1989) Trends Biochem. Sci. 14, 280-283.

Wickner, W., Driessen, A. and Hartl, F-U (1991) Ann. Rev. Biochem., 60, 101-124.

Wiech, H., Sagstetter, M., Muller, G. and Zimmermann, R. (1987) EMBO J., 6, 1011-1016.

Wiedman, M., Kurzchalia, T V., Hartmann, E., & Rapoport, T.A. (1987) Nature, 328, 830-833.

Willians, J. C., Steiner, L. A., Ogden, R. C., Simon, M. I. and Feher, G. (1983) Proc. Natl. Acad. Sci. USA, 80, 6505-6509. 1 59

William, J. C., Steiner, L. A., Simon, M. I. and Feher, G. (1984a) Proc. Natl. Acad. Sci. USA, 81, 7327 - 7332.

Wiltians, J. C., Steiner, L. A., Geher, G., and Simon, M. I. (1984b) Proc. Natl. Acad. Sci. USA, 81, 7303-7307.

William, J. C., Steiner, L. A., Ogden, R. C., Simon, M. I. and Feher, G. (1985) Proc. Natl. Acad. Sci. USA, 80, 6505 - 6509.

William, J. C., Steiner, L. A. and Feher, G. (1986) Proteins, 1, 3 1 2 -3 2 5 .

Wolfe, P. B., Silver, P., and Wickner, W. (1982) J. Biol. Chem. 257, 7 8 9 8 -7 9 0 2 .

Wolfe, P. B., Wickner, W., and Goodman, J. M. (1983) J. Biol. Chem. 258, 12073-12080.

Wolfe, P. B. and Wickner, W. (1984) Cell, 36, 1067-1072.

Wolfe, P. B., Rice, M. and Wickner, W. (1985) J. Biol. Chem. 260, 1 8 3 6 -1 8 4 1 .

Yamane, K. and Mizushima, S. (1988) J. Biol. Chem. 263, 19690- 19696.

Yamane, K., Akiyama, Y., Ito, K., and Mizushima, S., (1990) J. Biol. Chem. 265, 21166-21201.

Yeates, T., Komiya, H., Rees, D., Allen, J., and Feher, G., (1987) Proc. Natl. Acad. Sic. USA, 84, 6438-6442. 1 60

Youvan, D. C., Bylina, E. J,, Alberti, M., Begusch, H. and Hearst, J. Y. (1984) Cell, 37, 945-957.

Zerial, M., Melancon, P., Schneider, C. and Garoff, H. (1986) EMBO J., 5, 1543-1550.

Zimmermann, R., Sagstetter, M , Lewis, M.J. and Pelham, H.R.B. (1988) EMBO J „ 7, 2875-2880.

Zhu, H. Y. and Dalbey, R. E. (1989) J. Biol. Chem., 264, 11833- 11838.

Zhu, H. Y., Cheng, S. Y., Kuhn, A., Geller, B. and Dalbey, R. E. (1992) Manuscript in preparation.