XDENTIFYING DOMAINS OF SHIGA-LIm TOXIN 1 THAT
ARE RESPONSIBLE FOR ITS MEMBRANE TRANSLOCATION
by
Mazen T. Saleh
A Thesis Submitted in Conformity with the Requirements for the Degre of Doctor of Philosophy, Graduate Department of Medical Biophysics, in the University of Toronto
Ocopyright by Mazen T. Saleh. 1997 National Library Bibliothéque nationale I*I of Canada du Canada Acquisitions and Acquisitions et Bibliographie Services services bibliographiques 395 Wellington Street 395, nie Wellington Ottawa ON KIA ON4 OttawaON K1AOW canada Canada
The author has granted a non- L'auteur a accordé une licence non exclusive licence allowing the exclusive permettaot à la National Library of Canada to Bibliothèque nationale du Canada de reproduce, loan, distribute or sell reproduire, prêter, distribuer ou copies of this thesis in microfom, vendre des copies de cette thèse sous paper or electronic formats. la forme de microfiche/nlm, de reproduction sur papier ou sur format électronique.
The author retains ownership of the L'auteur conserve la propriété du copyright in this thesis. Neither the droit d'auteur qui protège cette thèse. thesis nor substantial extracts fiom it Ni la thèse ni des extraits substantiels may be printed or othexwise de celle-ci ne doivent être imprimés reproduced without the author's ou autrement reproduits sans son permission. autorisation. ABSTRACT
Identifyùig domains of shiga-like toxin 1 that are responsible for its membrane translocation
Mazen T. Saleh
Docmr of Philosophy.1997, Graduate Department
of Medical Biophysics, in the
University of Toronto
Shiga-like toxin 1 (SLT-1).produced by pathogrnic strains of Escherchia coli. is a representative mernber of a superfamily of multimeric protein toxios known as the A/B family of toxins. SLT-1 is intemalized by cells via receptor-mediated endocytosis and localizes in various intracellular compartments. However, targets of these toxins are invariably Iocated in the cytoplasm. thus internalized SLT-I rnust translocate across a vesicular membrane in order to deliver its cytotoxic function. Earlier studies have revealed that toxins may exploit conditions in intracellular environments to facilitate translocation. For example, acidic pH of endocytic vesicles might dlow exposure of hydrophobic domains that associate with membranes, a necessary initial step in membrane translocation. SLT-1 was used to determine if this proposed mechanism was a general feature of all members of this family of toxins in tems of their mode of membrane insertion and uansIocation. Fluorescence spectroscopy has revealed that Ttp-34 in SLT-1 B subunit is perturbed at pH lcvels nomally found in endocytic and lysosomal vesicles. This effect was further characterized using circular dichroism. solid-phase receptor binding assays and gel permeation chromatography. The results support the conclusion that the çonformauond change observed with fluorescence studies is a local stmctunl effect and that the secondaiy and tertiary structures of the B subunit remain unaltered during a pH transition between pH
7 and pH 4. Moreover, studies carried out with the holotoxin indicated that the smcture of the toxh rernains largely unaffected by acidic pH in relation to neuual pH. The effects of other intraceUular microenvironments on SLT-1 translocation were investigated. Hydropathy profües of the A subunits of SLT-I revealed the presence of a hydrophobic domain between residues 220 and 246 flanked by positively charged amino acids that resembles the composition of signal sequences and trammembrane domains of integral membrane proteins. Studies with synthetic vesicles containing acidic lipids, such as
phosphatidyl glycerol. have shown that synthrtic peptides corresponding to this region of
the A 1 chain, can spontaneously insert into membrane vesicles. Our hypothesis is that this
region may function in a similar way within the intact Al chah once localized in the ER.
It is therefore concluded that acidic pH alone is not capable of eliciting a structurai change in internalized SLT-1. Rather. other intracellular processes, such as proteolytic processing and disulfide reduction, may be involved in exposing the hydrophobic C- terminus of the Al chain. The exposed domain may then insert into the lumenal side of the endoplasmic reticulum membrane as an initial step in membrane translocation.
Furthexmore. factors endogenous to the ER. such as those involved in signal-sequence processing, protein folding and degradation, may play a role in the translocation of the A chah of SLT-1 across the ER membrane.
... lll ACKNO WLEDGMENTS This body of work is the result of the collective efforts of many individuals. I am grateful for the efforts and contributions of my supervisor, Dr. Jean Gariepy. The present frame of
my scient& thinking was moulded by his relentless pursuit of excellence in reseuch. His patience provided me with the independence in the laboratory that 1 needed and his guidance ailowed me to successfully complete of my work
1 also wish to thank members of my student supervisory cornmittee, Dr. David Rose and Dr. Cheryl Arrowsmith for their concise encouragement, support and helpful discussions; my pst and present partners in the laboratory for making my stay stimulating and pleasant both at work and during evening discussions in bars. Without their efforts, 1 would have probably saved more money and my liver would be in a better state of health. They are :
Beth Boyd, Bruce Carpick, Jim Ferguson, Francesca Bahr, Rajiv Ayra, Deming Liu, Kate Sheldon. Eric Lacasse, Mark Bray, Devender Singh. Reza Kiarasch, Wai-May Lim, Stuart Bisland, Kim Kawarnura, Ray Reilly and Nancy Stokoe, and a special thanks to John Simard.
1 also iake this opportunity to thank our collaborators, Dr. Ioan Boggs and Godha Rangaraj (University of Toronto), for the EPR work presented in chapter 4; Dr. Robert Hodges and Mr. Lome Burke (Department of Biochemistry, University of Alberta) for perfoming rnass spectrometry on the purified peptides and, Dr. Avi Chakrabartty for his help and access to his circular dichroism spectrometer and fluorometer. Finally. I would like to acknowledge
Ms. Tanuja Chitnis and Dr. David Isenman for their help in the early stages of the work presented in chapter 1 as well as Drs. Penelope Stein and Randall Read for providing us with the crystûl coordinates for the B subunit of Shiga-like toxin 1. TABLE OF CONTENTS
List of Tables List of Figures List of Abbreviations CHAPTER 1
Introduction ...... 1.1 The Role of Shiga-like Toxins in Disease ......
1.2 The Structure of Shiga-like Toxin 1 ...... 1.2.1 The Smcture and Function of Shiga -1ike Toxin 1 A subunit ......
1.2.2 The Structure of Shiga-like Toxin 1 B subunit ...... 1.3 The Receptor of Shiga-like Toxin 1 ......
1.4 Receptor-Mediated Endocytosis of Bound Toxin ......
1.5 Stnictural and Functional Homologies with other Toxins ......
1.6 Goals and Objectives ......
CHAPTER 2
Experimzntal Methods 2.1 Biochernicd Methods Purification of SLT-1 ...... Purification of SLT-1 B subunit ......
Purification of the synthetic peptides ...... EIectroelution of SLT-1 A subunit ...... Radiolabeling of SLT-1 and SLT-1 B subunit ......
Gb3 solid phase binding assay ...... 2.1.7 Spin-labeling of ShTA(220-246) ...... 2.1.8 Preparation of lipid vesicles ...... 2.1.9 Binding of radiolabeled peptide to lipid vesicles ...... 2.1.10 Total Phosphorus Assay ...... 2.1.1 1 Amino acid analysis ......
2.2 Characterization of Purified SLT-1 and its B subunit 2.2.1 Polyacrylamide gel electrophoresis ...... 2.2.2 Western blot analysis ...... 2.2.3 Cytotoxicity assay ...... 2.2.4 Cell-free protein synthesis inhibition assay ......
2.2.5 Gel permeaiion c hromatography ......
2.2.6 Calculation of the molar absorption coefficient of SLT-I ......
2.3 Immunological Methods 2.3.1 Generation of rabbit Anti-SLT-1 A subunit antisera ...... 2.3.2 Detection of rabbit anti-SLT-1 antibodies by ELISA ...... 2.4 Biophysical Methods 2.4.1 Fluorescence Spec troscopy ...... 2.4.2 Circular Dichroism ...... *...... 2.4.3 EPR measurements ......
CHAPTER 3
Local Conformational Change in the B subunit of Shiga-like Toxin I at Endosomal pH 3.1 Introduction ...... 55
3.2 ResuIts and Discussion ...... 59 3.2-1 The fluorescence of uyptophan-34 is perturbed at endosomal pH ...... -...... -..-.-.... 3.2.2 The structure of the B-subunit is graduaily altered by pH conditions ranging from pH 2 to 4.5 ...... 3.2.3 The tertiary and quatemary structures of the B subunit pentamer are stable between pH 2 and 10 ......
CHAPTER 4
Effect of Endosomal pH on the Conformation of the Holotoxin
4.1 Introduction ...... ,...... , ....., ...... ---.--.--.-...... --- .-.
4.2 Results and Discussion
4.2.1 Acidic pH alters the microenvironment of tryptophan(s)......
4.2.2 Acidic pH does not alter the exposure of tryptophans in
SLT-1 ...... ,...... --S.-...-...... 4.2.3 pH does not alter the secondary stmcture of SLT-1 ......
CHAPTER 5
Insertion and Orientation of a Synthetic Peptide Representing the Shiga Toxin Ai C-terminus into Phospholipid Membranes
5.1 Introduction ... .,...... ,...... -. ..-... .-. -... .-. .. .-.....* ......
5.2 Objectives ...... *......
5.3 Results and Discussion
5.3.1 ShTA(220-246) adopts a partially helical structure in the presence of negatively charged lipid vesicles ...... 5.3.2 Changes in the fluorescence spectrum of W232A242ShT(220-246)
in the presence of lipid vesicles suggest that this hydrophobie region of the Al domain partitions readily into a lipid bilayer ...... 90
5.3.3 Binding of radiolabeled Wz32A242ShT(220-246)
to lipid vesicles ...... - ...... 94 5.3 -4 Monitoring the interaction of ShTA(220-246) with membranes by EPR ...... -...-...... 94 5.3.5 S hTA(220-246) causes leakage of calcein trapped inside DMPG vesicles ...... --...... 100 5.3.6 A mode1 summaizing the interaction of ShTA(220-246) with a negatively charged lipid bilayer ...... 102
CHAPTER 6
Concluding Remarks
6.1 Does the B subunit play a role in A chain translocation? ......
6.2 Domains andogous to ShTA(220-246) are present on other proteins ......
6.3 Does ShTA(220-246) and the Al chain interact with component(s)
of a protein transporter located in the ER membrane? ......
6.4 The involvement of the ubiquitin-proteosorne pathway ......
6.5 A hypothetical mode1 describing membrane insertion and
translocation of the SLT-1 Ai fragment ......
6.6 Future Directions ......
REFERENCES ...... -...... -. . . - .. . . . -...... LIST OF TABLES
Table 1.1 Relationship between mem bers of the family of Shiga toxins ...... 5
Table 2.1 Amino acid analysis of ShTA(220-246) and W232A242ShT(220-246) ...... 36 Table 2.2 Amino acid analysis of SLT-I and its B subunit ...... 48
Table 4.1 Effect of endosomal pH on the gel permeation properties of SLT-1 and its B subunit ...... 80
Table 5.1 EPR parameten denved from the spectra of spin-labelled
ShTA(220-246) ...... ,.,, ...... 97 LIST OF FIGURES
CHAPTER 1
Figure 1.1 Distribution of 12.9 million deaths among children under five years old in al1 developing countries ......
Figure 1.2 Sequence alignment of A subunits from SLT-1 and related toxins ...
Figure 1.3 Proposed mechanism of the enzymatic A subunit of SLT-1 ......
Figure 1.4 Structure of Shiga-like toxin 1 B subunit ......
Figure 1.5 Various representations of the X-ray crysrallography structure of Shiga toxin ......
Figure 1.6 A mode1 of the structures of Gb3 and Gb4 ......
Figure 1.7 Schematic representation of the receptor-mediated
endocytosis of bound SLT-1 ......
CHAPTER 2
Figure 2.1 Typical chromatognphy profiles observed during the purification of SLT-1 ......
Figure 2.2 Typicÿl chromatognphy protiles observed during the
purification of SLT-1 B subunit ......
Figure 2.3 Reverse-phase purification profde of ShTA(220-246) ......
Figure 2.4 Reverse-phase purification profile of
W232A242S hTA(2 20-246) ......
Figure 2.5 Mass spectnim of ShTA(220-246) ......
Figure 2.6 Mass spectrum of W232A242S hTA(220- 246) ......
Figure 2.7 Electrophoretic mobility of purified SLT-I ...... Figure 2.8 Electrophoretic mobility of purified SLT-1 B subunit ......
Figure 2.9 A typical curve obtained in the Vero ceii cytotoxicity assay ......
Figure 2.10 A typical curve obtained in the cell-free protein synthesis inhibition assay ...... -.-... Fipe2.1 1 Calibntion of the gel permeation column ...... Figure 2.12 The absorption gradient of SLT-I ......
CHAPTER 3
Figure 3.1 Effect of pH on the tryptophan fluorescence intensity of the B subunit of Shiga-like toxin 1 ...... Figure 3.2 Stereoprojections of the B subunit peniamer of Shiga-like toxin 1 ......
Figure 3.3 Effect of pH on the circular dichroism spectrum of the B subunit of Shiga-like toxin I ...... Fi yre3.4 Effect of pH on ellipticity values at 222 nm for the SLT-1 B subunit ...... Figure 3.5 Histognm depicting the effect of pH on the specific binding of radioiodinated Shiga-like toxin 1 B subunit to its natural recepior Gb3 ......
CHAPTER 4
Figure 4.1 Various representations of the structure of SLT-1 highlighting the location of tryptophans ...... -......
Figure 4.2 Tryptophan iluorescence cmission spectra of SLT-1 and its B subunit ...... ,,,..,., ...... 74 Figure 4.3 Change in tryptophan fluorescence of SLT-I as a function of pH ...... 75 Figure 4.4 Acrylamide quenching of Trp fluorescence in SLT-I and its B subunit at neutral and at acidic pH ...... 77 Figure 4.5 CD spectra of SLT-I in various pH conditions ...... 79
CHAPTER 5
Figure 5.1 Hydrophobicity profile of the A subunit of Shiga toxin ...... 86
Figure 5.2 Circular dichroism spectrurn of the synthetic peptide S hTA(220-246) ...... 88 Figure 5.3 Circular dichroisrn spectra of the peptide ShTA(220-246)
in the presence of small unilamellar vesicies composed of DMPG lipids ...... 89
Figure 5.4 Circular dichroism spectra of the peptide ShTA(220-246)
in the presence of smdl unilameilar vesicles composed of DMPC lipids ...... 91
Figure 5.5 Effect of lipid concentration and composition on the tryptophan fluorescence emission spectra of the peptide
W232Az42ShTA(220-246) ...... -...... 93
Figure 5.6 Effect of DMPG vesicles and ascorbic acid on the EPR spectnim of proxyl-labeled ShTA(220-246) ...... 96
Figure 5.7 Reduction of only the immobilized EPR spectral component hy entrapped ascorbic acid ...... 98 Figure 5.8 Peptide-induced release of calcein from DMPG vesicles as a function of time and peptide-to-lipid ratio ...... 101
Figure 5.9 A schematic mode1 depicting possible modes of interaction of peptide ShTA(220-246) with the Lipid bilayer of
DMPG vesicles ...... ,...... 103
CHAPTER 6
Figure 6.1 A schematic diagram depicting the proposed mechanisrn of
membrane translocation by SLT-1 ...... 1 1 1 LIST OF ABBREVIATIONS
AR1 Acute Respiratory Infection
Bip/GRP94/ASP47 Members of the family of heat shock proteins
CD Circular dic hroism spec troscopy
CT Cholem Toxin
DEAE-sepharose Diethyl aminoethyl-sepharose. anion exchange min
DMPG
DMPC Dimyristoyl-L-a-phosphatidyl choline
EDTA Ethylene diamine tekaxetic acid, cation chrlating agent
EHEC Enterohemorhagic Eschenchia coli
EIEC Enteroinvasive Fccherichia coli
EPEC Enteropathogenic Escherichia coli
EPR Electron paramagnetic resonance spectroscopy
ETEC Entero toxigenic Esche richia coli
ER Endoplasmic Reticulum
Gai Galactose
G b3 Globotnasyl ceramide
G b4 Globotetraasyl ceramide
Glc Glucose
GO Golgi Apparatus
HC Hemorrhagic Colitis
xiv HPLC High performance liquid chromatography
HUS Hernomhagic Uremic Syndrome Ks .(a Effective Stern-Volrner quenching coefficient
LT E. coli heat-labile enterotoxin
MRE Mean residue ellipticiy
PMSF Phenyl methyl sulfonyl fluoride. prote% inhibitor
PVDF Polyvinylidine difluoride
RT Ricin Toxin
SDS-PAGE Sodium dodecyl sulfate-polyacrylamide gel electrophoresis
Sec6lp Subunit of the ER protein translocation channel
ShT Shiga Toxin produced by Shigella dysenterae type 1
ShTA(220-246) Residues 220-246 of the A subunit of Shiga toxin
SLT Shiga-like Toxin(s) produced by E. coli
SUV Small unilamellar vesicles
Tm ,X Hyperfine splitting parameter
It, Motional panmeter
TFE Trifl uoroethanol
VT Verotoxin. synonornous with SLT
VTEC Verotoxin-producing Escherichia coli
W232A242ShTA(220-246) Residues 220-246 of the A subunit of Shiga toxin with
substitutions, I232W and C242A
WH O World Health Organization INTRODUCTION CHAPTER 1
INTRODUCTION
The World Health Organization WO)has recently reported that 25% of infant deaths in third world countries is a consequence of diarrhea (Fig. 1.1; WHO, 1995).
Diarrheagenic strains of E. coli represent a major culprit in the incidence of such diseases. These strains cm be classified into four main groups; enterotoxigenic, enteroinvasive, enteropathogenic and enterohernorrhagic (Karmali et al., 1989). Enterotoxigenic E. coli
(ETEC)strains represent the major cause of diarrhea, including travellers' diarrhea, in third world countries. Members of this group produce one or both of a cholera toxin-like heat- labile enterotoxin (LT) and a heat-stable enterotoxin (ST). Enteroinvasive E. coli (EEC) strains cause severe diarrhea that resembles dysentery caused by Shigella dysentery type 1.
Enteropathogenic E. coli (EPEC)strains hclude specific E. coli serotypes that are associated with infantile diarrhea and can be shown to adhere to Hep-2 cells in vitro (Cravioto et al.,
1979). a property not seen with non-pathogenic E. coli or other classes of pathogenic E. coli
(Kmali et al., 1989). Enterohemorrhagic E. coli (EHEC) strains includtl verotoxin- producing E. coli (VTEC) and is charactznzed by the production of one or more verotoxins or Shiga-like toxins. EHEC strains are associated with episodes of hernorrhagic colitis and the production of charactenstic intestinal mucosal lesions (Tzipori et al.. 1988; Karmali et al., 1989; Gyles, C. L.. 1992).
Shiga-iike toxin 1 (SLT-1) represents one example of a large family of bacterial and plant toxins that share cornmon stnictural and functional properties, including the targeting of specific surface receptors of eukaryotic cells. Members of this farnily of toxins are multi- subunit polypeptides ranging in size from 70 kDa to 170 kDa. Cytotoxicity occurs as a result of the intemalization of the receptor-bound toxin and the subsequent translocation of ARI
Diarrhea 23%
Figure 1.1 Distribution of 12.9 million deaths among children under five years old in all developing countnes. ARI; acute respiratory infection. Reproduced from the World Health
Organization (WHO)statistics annual. 1996. the catalytic subunit of the toxin to the cytosol. Mj. objective was to investigate the mechanisms by which SLT-I enters into ceIls and translocates across intracellular membranes
in order to inactivate its intracellular mget, the ribosome. located in the cytoplasm.
. * ... 1.1 The role ot Shimike tomin di-
Patients that developed hemorrhagic colitis during outbreaks of this disease in the early
1980's in North America were found to have consumed contaminated beef patties or meat products from fast food restaurants. The disease was terrned "hamburger disease" (Riley et d., 1983; Wells et al.. 1983). Clinical isolates collected from these patients revealed the presence of a rare toxin-producing E. coli serotype known as 0 lS?:H7 (Riley et al., 1983:
Wells et al.. 1983). Various labontories later reported the isolation of related E. coli serotypes that produce a farnily of closely-related toxins (Table 1.1).
Patients infected with toxin-producing Shigellu or E. coli strains may exhibit lik threatening symptoms such as hemonhagic colitis (HC) and hemolytic urernic syndrome (HUS) resulting from darnage to the colon and kidney, respectively (Riley et al., 1983;
Karmali et al., 1983; for a review see Karmali, M. A., 1989). Toxin-producing E. coli strains normally express more than one type of shiga-like toxins (SLT-1 and SLT-II). The majority of patients exhibit immunological responses (prirnarily production of antibodies of class IgA) against SLT-1 and not against SLT-II (Ashkenazi et al., 1988). Although a mechanism by which active toxin gains entry into the blood circulation has not ken identified. the involvement of the toxins in such affiictions is supported by the presence of neutralizing antibodies in infected patients (Karmali et al., 1985). Tesh et al. (1993) have recently demonstrated that a 50% lethal dose (LDSo) for SLT-II in rnicr was 400 times less than that for SLT-1 when injected intravenously or intraprritonrally. The authors did noi detect any differences in the enzyrnatic activity of SLT-1 versus SLT-II and proposed bat the difference in lethality could be due to differences in toxin stability andor receptor affiity.
Table 1.1 me relationship between members of the famiiy of Shiga toxins. The data was compiled and presented according to the proposal of O'Brien et al. (1994) for rationaiizing the nomendature of tbe E. coli cytotoxins. The toxin ciassification is based on (a) serological specificity (b) host organism and (c)biological activity (qtotoxicity). As shown in the table. SLT-I and SLT-II are serologicalty different from other ~lated toxins. SLT-IIc and SLT-iie differ in terms of receptor binding (SLT-IIe binds to Gb4 instead of Gb3) and in the nature of the host organism (O'Brien et al., 1994; Jackson, M. P., 1990; Jackson & O'Brien, 1994).
Neutralization by Cytotoxicity Antisera Toxin Source6train Host ag ainst Vero Hela ShT SLT-II ShT S. dysen tenue hurnan + + + -
+ SLT-II E. coli/C600 human + + - (933W)
E. coli Pig SLT-Ue S1191. 412
The involvement of toxin in disease was further supported by the fmdings of Fontaine et al. (L988), where infection of macaque monkeys with a toxin-deficient strain of
Shigelh was compared with that of a wild type Shigella suain. The study showed that toxin production was ûssociated with bloody stools. colonic vascular darnage and intestinai inflammatory response. Other evidence indicate that the pnmary target of the toxins is the vascular endothelium (Kavi et al.. 1987; Obrig et al.. 1988) and colonic epithelial cells (Moyer et al.. 1987). Later studies (Wadolkowski et al.. 1990a; Wadolkowski et al., 1990b) have demonstrateci the presence of toxin receptors in mouse kidneys and that renal darnage is
the ultimate cause of death in rnice challenged with purified toxin. Toxin receptors have now
been identified in several tissues including endothelid celis lining blood vessels and red
blood ceiis (Richardson et al., 1988; Bitzan et. al., 1994). In rabbits, a challenge with purified SLT-1 resulted in anorexia. lethargia and limp paralysis with no effect on renal function (Zoja et al., 1992). The authors have shown that the lesions observed comelated with the tissue distribution ofGb3. a glycolipid identified as the recrptor for ShT and SLT-1.
There was no Gb3 present in the hem, liver or the kidney of nbbiu. Tissues having the highest Gbs content were the tissues of the central nervous system. gruirointcsinal tract. lung and spleen.
Once in the circulatory system. the toxin may target cells of the immune system.
Activated B-cells, in particular IgG-cornmitted subsets, are sensitive to SLT-1 in vitro
(Cohen et al.. 1990). The deletion of B cells subsets may explain the relative lack of anti- toxin IgG responses in patients infected with toxin-producing E. coli strains ( Lingwood et ai., 1994). The role of SLT-I in disease was also investigated in rabbits using the pathogenic
E. coli strain, RDEC- 1. An isogenic form of this strain called RDEC-H19A. was produced by infecting RDEC- 1 with an SLT-1-converting bacteriophage. The resulting strain produces a disease in rabbits with symptoms similar to those seen in humans infected with mxin- producing E. coli strains (Sjogren et al.. 1994). The intravenous injection of puditid SLT-
XIe into pigs cm reproduce the symptoms of pig edema disrase, an illness arising as ü result of infections with SLT-ne-producing E. coli strains (Macleod et al., 199 1). Investigation of toxicity in pigs caused by SLT-IIe reveded darnage to the stomach, the eyelids, the cerebellum and the colon (Gyles. C. L., 1992). SLT-producing E. coli strains have also been isolated from cats with aggressive diarrhea (Abaas et al., 1989). 1.2 stnict-
Shiga and Shiga-Like toxins (verotoxins) are closely reiated oligomeric proiein toxins.
Shiga toxin, produced by Shigelln dysenreriae (Shiga et al., 1897) has been known for a century. In the late seventies. a new toxin was isolated from pathogenic svains of E. coli that exhibited cross reactivîty with antisera raised against Shiga toxin (O'Brien et al., 1978). This new toxin was termed Shiga-like toxin (SLT). Independent studies by Konowalchuk and his coworkers also identifïed a cytotoxin produced by certain strains of E. coli which was able to kill Vero cells (Konowalchuk et al., 1977). This toxin was accordingly cdled verotoxin 0.It was later show that SLT and VT were the same toxin. The terms SLT and VT were subsequently used interchangeably.
With the discovery of new members of this family, SLT was named SLT-1 (VT- 1) to distinguish it from the new toxins isolated from E. coli. Two variants of SLT-1 were later identified; SLT-Ik, fiom E. coli isolates recovered tiom humans (Scotland et al.. 1985;
Strockbine et al., 1988) and SLT-IIe. from E. coli isolates from pigs (Marques et al..
1987). The inter-relationship of these toxins is illustrated in Table 1.1.
The primary structure of SLT-I is identical to that of the ShT with the exception of a serine to threonine substitution at position 45 in the A chain. SLT-1 subunit stmcture is a common design found in a larger group of bacterial and plant toxins now referred to as the A/B family of protein toxins (for reviews see, Memtt & Hol, 1995; Bumette, W. N.. 1994; and Read & Stein, 1993). This family of toxins includes cholen toxin (CT),pertussis toxin (PT), diphthena toxin (DT) and the plant toxin. ricin (EU'). Members of this family of toxins are made up of an enzymatic A subunit and a receptor binding B subunit.
SLT-1 is made up of one A subunit and five identical copies of a B subunit (Olsnrs ci al.. 198 1). The gene for the toxin was cloned and sequenced from E. coli O 157:H7 (Takao et al.. 1988) and from bactenophage H-19B (De Grandis et al.. 1987), isolated from the E. coli isolate 026:H19 (Smith et al., 1983) . The association between the A and B subunits involves non covalent interactions. The complex is very stable. as harsh conditions (8M urea. 0.1% propionic acid) are required to dissociate the subunits from each other (Yutsudo et al.. 1987; Head et al., 1991).
13.1 me Structure and Function of Shia-like Toxin 1 A subunit The A subunit is made up of 293 amino acids (Fig. 1.2) and has an isoelectric point of 8.2 (Yutsudo et al.. 1987). whereas the isoelecvic point of the holotoxin is 6.72 to 7.02 (O'Brien & LaVeck, 1983; Petric et al.. 1987). These PI'S are different from those of 5.2 (Downes et al.. 1988) for SLT-IIc and 9.0 (MacLeod et al.. 1991) foi SLT-IIe. The A subunit possesses N-glycosidase activity which cleaves an adenine located at position 4324 in 28s rRNA (Fig. 1.3; Endo et al.. 1988) and resulting in the inhibition of protein synthesis. In the crystal structure of Shiga loxin (Fraser et al.. 1994). the A suhunit was show to be lying on one face of the pentamer (referred to as the A-face; see Fig. 1.4). opposite to the face containing the receptor binding site(s), the R-face. The C-terminus of the A subunit (residues 282-287) adopts a partially helical conformation and is inserted into the central pore created by the pentamer. The C-terminus is relatively hydrophobic, a feature compatible with the hydrophobic pore of the pentamer. The two cysteines at positions 24 1 and 26 1 in the A chin form a disuiphide bond.
The resulting Ioop contains residues 242 to 260 (Fig. 1.2) and is the region of the A chain most susceptible to proteolytic attack Interestingly, the side chain of Met-260 is positioned into the pocket corresponding to the putative N-glycosidase active site of the A chain (Fig.
1.3). The location of this residue may explain why the cleaved toxin is more active than its uncleaved counterpart (Garred et al.. 1995a; Burgess & Roberts, 1993). sincc clzavage in the protease sensitive loop could potentially move this residue away from. thus incrrasc: the SLT-IA AI PI&??- LMID SLT-IIA $":I%!g EI $-PLEHIS QGTTSVSVINI SLT-IIvA ...... VSSLNSI AIS~PLEHISQGATSVSVIN~ RTA IFPKQYP ZLT$ADVRH EIPVLPNRVG
SLT-IA SLT-IIA ATNTFYRF SLT-IIvA TTNTFYRF RTA ..... NSAYFFHP
101 SLT-TA AD ...... TFPGTSAV TLS SLT-IIA SD ..... SVPGVTTV SMT SLT-IIvA SD.. ... SLPGVTTI SMT 9TA FTDVQNRY TFA
SLT-IP. EIMJRFVTVT .SLT-iIA IRAVLRFVTVT PjED RQIQ REF ~~T-IIVA RAVLRFVTVT RE$J~E: 2TA FIICIQMI YIE GE
201 250 SLT-IA S GRS Y VMTAE DYHG QD d~...... SLT-IIA TAPVYTMTPG EYRG ED+...... SLT-IIvA TAPVYTMTPE EYRG SA@...... RTA Y NRRS P.P DP S ESNQ GAFASP IQLQ .. ,.
SLT-IA SLT-IIA SLT-IIvA RT A
SLT-IA SLT-IIA SLT-IIvA "nA
Figure 1.2 Sequence alignment of A subunits from SLT-1 and related toxins. Conserved domains are boxed in green lines and residues in the active site are boxed in red. RTA, A chain of ricin. accessibility to, the active site. The majority of the contacts between the A chain and the B pentamer are due to the A2 fragment (minimal domain; residues 262-293) and a srnaIl contribution from a fbhairpin in the Al fragment (residues 2 18-226).
Trp-203 in the Al fragment has been identified as an active site residue by mutagenesis (Yamasaki et al., 1991; Jackson et al., 1990). Other residues implicated in the catalytic action of the toxin include Glu-167, Tyr-77, Tyr4 14 and Arg-170 (Hovde et al., 1988; Schlossman et al.. 1989: Deresiewicz et al.. 1992; Yarnasaki et al., 1991; see Fig.
1.3). Using deletion analysis. Haddad and coworkers (1993) defined a C-terminal boundary of the enzymatic Ai fragment at Arg-268. In that study. deletion mutants werr scrcened using an in vitro protein synthesis inhibition assay. N-terminal deletions wcre also uxd to map the minimal N-terminus site to Asn-75 (Haddad et al., 1993). The A chain is thus typically processed into an Ai fragment (27kDa; N-terminus) and an A2 domain (5kDa; C- terminus) while associated with the B pentamer.
1.2.2 a Structure of SJ,T-1B subunit The B subunit of SLT-I is made up of 69 arnino acids (Fig. 1.4) and has an isoelectric point of 5.8. as determined by polyacrylarnide gel isoelectric focusing (Yutsudo et ai., 1987). nie sequence contains cysteine residues at positions 4 and 57 involved in a disulfide bond in the functional B subunit and a single tryptophan residue at position 34 (Newland et al.. 1985; Huang et al.. 1986). The crystal structure of SLT-1 B suhunii revealed structural homology with the E. coli heat-labile enterotoxin (Stein al., 1992).The
B subunit exists as a pentamer with each monomer cornposed of 2 three-stranded antiparallel
P-sheets and a 10-residue long a-helix (Stein et al., 1992). The monomers assemble together to form a ring-like structure with a central pore. The diameter of this orifice ranges in values Figure 1.3 Action of the enzymatic A subunit of SLT-1. (A) Clravage site in the adenosine nucleotide (position 4324) that results in the liberation of the adenine moirry (Olsnes et al.,
1990). (B) The conserved nucleotide loop of 28s rRNA recognized by ricin and SLT's. IL mediates the binding of tRNA's during polypeptide synthesis. (C) Location of the nucleotide loop within the ribosome. It is located within the cleft in which peptide bond formation and polypeptide elongation takes place. Rernoval of adenine 4235 inhibits the binding of tRNA's and thereby disabling protein synthesis and ultimately causing cell death. from 11A on the A-face down to 8A on the R-face (Fig. 1.4 and 1.5). The 5 helices line the pore and each contains four leucine residues which creates an internal surface harbouring an overail hydrophobic character. The pentamer of the E. coli LT toxin, on the other hand. displays a hydrophillic cenual pore.
The E. coli LT B subunit (103 arnino acids) and the SLT-I B subunit (69 amino acids) share little sequence identity but have a surprisingly similar xcondary and teniary structures. A comparative study by Sixma and coworkers (1993) have shown that both subunits cm be superimposed with seven out of eight secondary structure eiements retained in both subunits. Using sequence alignment with other members of the SLT family of toxins, the authors also attempted to locate the Gb3-binding site. The aost like!~site was identified in a cleft between monomers and contains residues Asp- 18 and Phe-30 (Stein et al., 1992; Nyholm et al.. 1996). Mutagenesis studies have also identified Phe-30 as an important residue for receptor binding (Bast et ai., 1994). This cleft also includes the side chahs of Asn- 15, Thr- 19 and Thr-2 1. al1 of which are capable of hydrogen bonding (Stein et al.. 1992; Nyholm et al.. 1996). Lys-53 has indirectly been implicated in receptor binding. since the biotinylation of its side chain was shown to reduce the binding or the B subunit to
Gbs (Khine & Lingwood. 1991). The central heiix in the B suhunit is çapped al the N- terminus end by Trp-34. which places this rrsidue in the sarnr Fax of the pcntamer as the putative Gb3 binding domain (referred to as the R-face or the pentamer; Fig. 1.4). Forrnylation of its indole ring has been shown to inhibit the binding of SLT-1 B subunit to Gbs (Boyd. B.. 1991).
Finally, the reduced quenching of the B subunit tryptophan fluorescence by acrylamide in the presence of DMPC/Gb3 vesicles suggesü the proximiy of the Trp-34 to the receptor (Surewicz et al., 1989). The monomeric B subunit dues not bind to Gb3, since the monomers are not observed in solutions and dissociation of the pentamer Figure 1.4 Structure of Shiga-like toxin 1 B subunit (A) Sequence alignment of SLT-I B subunit with SLT-II and SLT-IIv B subunits. Identical residues are boxed in green lines and residues implicated in the binding of SLT-1 to Gb3 are boxed in red. (B) Space-fing mode1 of the X-ray structure of the B subunit pentamer (Stein et al., 1992). The red colour shows the location of the side chahs of the residues involved in receptor binding (Asp-18 & Phe- 30). This top face of the pentamer is the R-face (as defined in the text) where the single tryptophan of the B subunit (purple colour) is located. A SLT-IB SLT-IIB SLT-IIvB
SLT-IB SLT-IIB SLT-IIvB Figure 1.5 Vanous representations of the X-ny crystallography structure of Shiga toxin
(ShT) as determined by Fraser et al. (1994). (A) A stick mode1 of the structure of ShT highlighting the Al hgment (pllow). a hydrophobic domain betwern residuts 220 and 246 in the A subunit (purple). the A2 fragment (magenta) and the active site residues (black bal1 and stick models) including Tyr-77, Tyr-1 14, Glu-167 and Arg-170. The red dots in this region represent the oxygens of Glu 167. The B subunit pentarner is shown in green. (B)
The sarne colour scheme as in (A) Dut in a space fiilling model to show the level of exposure of the domains to the solvent. (C) The back face of the model shown in (B). requires denaniring conditions and the presence of reducing agents. This is also mefor other toxins containing homopentarneric B subunit such as CT and LT. In contrast, the separated monomers in pemissis toxin, which consists of a heteropentarnenc B subunit, (S 1 and S2) are properly folded and can interact . in vitro, with receptor moieties (Saukkonen et al.. 1992). The pnmq structure of the B subunit has been shown to be similar (up to 40% sequence similarity) to the N-terminal50 residues of the hurnan interferon y (IFN- y) receptor
(Lingwood & Yiu, 1992) and the N-terminus extracellular domain of CD19 (Maloney &
Lingwood, 1994). These findings have implications for the role that Gb3 may play in IFN-y- and CD19-mediated signalling and suggests a role for the toxin as a signalling mimic by which the bacterium evades the hosts' immune system to gain an advantage in colonizing the hosts' digestive tract
1.3 The Receptor of Shiea-like Toxin 1
SLT-1 binds to cells expressing the receptor globotriosyi cerarnide (Gb3) on their surface membrane (Fig. 1.6). Gb3. also known as CD77. is a human B ce11 differentiation antigrn and can he found on the surface of epithelial crlls lining the intestine and colon. on rend endothelid cells. as well as on red blood cells and hurnan platelets (Tedder & Isaacs.
1989; Obrig et al., 1993; Lingwood. C. A., 1993; Moyer et al.. 1987). Binding specificity to Gb3. and to a lesser extent to pentosyl cerarnide and galabiosyl ceramide (Cohen et al.,
1987; Lindberg et al., 1987) was described in several labontories using membrane lipid extracts resolved by thin-layer chromatognphy (TU3)and overlay detection methods based on the use of radiolabeled toxins and antitoxin antibodies (Jacewicz et al., 1986; Lindberg et al., 1987; Lingwood et al., 1987) . Al1 of those glycolipids have terminal Gaial-4Gal group and a cenmide lipid tail anchonng ii in the outer leaflet of the membrane. Free saccharide Figure 1.6 A mode1 of the stnictures of Gb3 (A) and Gb4 (B). The terminal disaccharide in
Gbs is interna1 in Gb4 but lectins that bind to Gb3 do not recognize Gbr. The ceramide moiety of the molecule is normaliy heterogeneous showing differenr chah lengths and variable unsaturation of the fany acid chahs (Kiarash et al.. 1994). shows very weak interaction with the toxin (St Hilaire et al.. 1994), which irnplicates the cerarnide in the interaction between the toxin and the receptor. even though it is not directiy involved in the binding . There is one more line of evidence availsble that suggests a role for the ceramide tail in binding; digalactosyl diglyceride, a plant glycolipid that has the same carbohydrate sequence but has glycerol instead of cerarnide, is not recognized by SLT-1. The ceramide group may play a role in maintaining the carbohydrate in a specific orientation by hydrogen bonding between the sphingosine amide and the oxygen linking the fust glucose molecule to the other lipid tail (Stromberg et al.. 199 1; Karlsson. K.-A., 1989). Certain ce11 lines. such as A431, express Gbs on their ce11 surface but are not sensitive to SLT-1 (Sandvig et al.. 1992). On the odier hand. sensitiviy :o die toxin cm be induced in some resistant ce11 lines by ueatment with agents that increase the expression of Gbs on the ceII surface, snch as hydrocortisone or butync acid (Mobassaleh et al.. 1994;
Sandvig et al., 1991). The reason for the resisunce of Gb3-positive cells is not clear. Thest: cells are equally resistant to clraved toxin. elirninating the possibility that resistant cells lack the ability to cleave and activate the toxin (Sandvig et al., 1992). It has been hypothesized that the resistance could be due to different intracellular routing of intemalized toxin molecules, such that membrane translocation is not possible (Sandvig et al., 1994). It is not clear how butyric acid sensitizes cells but it appears to be related to changes in the structure of Gb3. It has been shown that the lipid composition of glycolipids is important for the positioning of such molecules in the ce11 membrane and for the conformation of the carbohydrate moiety ( Pellizari et al., 1992; Stromberg et al., 1991; Karlsson, K.-A.. 1989; and Crook et al.. 1986). The physiological role of Gb3 is not well understood. It rnay represent a second messenger involved in the regulation of ceIl cycle and cellular differentiation. For example. tumor necrosis factor-a (TNF-a)activates a neutral sphingomyelinase in leukemia HL-60 cells (Bielawska et al., 1992; Dressler et al., 1992). The activation of this enzyme leads to an increase in free ceramide which in tum regulates a membrane-associated protein kinase that has been shown to mediate the action of TNF-a.
Using Burkitt's lymphoma ce11 lines. cytotoxicity by SLT-I was shown to be due not only to inhibition of protein synthesis but also due to apoptosis (Manpney et al.. 1993). a regulatory process that can be mediated by ceramide.
1.4 -or-M-ted Endocvtosis of Bound Tod
Following Gbî binding, SLT-I is internalized via receptor-mediated endocytosis from coated pits (Sandvig et al.. 1989). The formation of coated pits was observed in electron micrographs and the intemalization of the toxin could be inhibited by incubating cells at 4 OC or by treating them with dansyl cadavenne, an inhibitor of cytoplasmic trans- glutaminase (Khine & Lingwood, 1993; Davies et al., 1980). The intemalization of the toxin via clathrin-coated pits was further supported by the inhibition of toxin uptake as a result of acidifying the cytosol. which blocks the formation of coated pits (Sandvig et al.. 1987). Clathnn-coated pits can also ba inhibited by potassium daplation from the cytosol. which has been shown to protrct cells from the toxin (Sandvig et al.. 1994). Cells cün also be protected from lysis by the toxin by treating cells with 3-methyladenine and cyclohexirnide (Sandvig & van Deurs, 1992). Other factors that may influence cell sensitivity to SLT-1 include Cal+ transport and certain divalent cations (Sandvig and Brown. 1987).
Vero cells were found to be most sensitive in the presence of CaCh and SC12 and were resistant to the toxin in the presence of MnQ, BaC12. CoCI2 and inhibitors of Ca2+ transport (Sandvig & Brown, 1987). The intracelluiar trafficking of SLT-1 (Fig. 1.7) proceeds via endosornes to the Golgi (GO) and through retrograde transport to the endoplasmic reticulum (ER) and even the nuclear membrane (Sandvig et al., 1992;1994).
Brefeldin A, a dmg that disrupts the GO apparatus. protects cells from the toxin (Sandvig et Figure 1.7 Schematic representation of the receptor-mediated endocytosis of bound SLT-1.
The toxin is composed of an Ai fragment (red rectangle), an Az fragment (yellow rectangle) and B subunits (blue rectangles). Gb3 is represented on the ce11 surface by the green symbols. The initial clathrin (brown-green high1ight)-coated vesicle is formed from the ce11 membrane and localizes the toxin in the Golgi, as well as other compartments but are not shown here for clarity. The toxin is then transported via a retrograde rnechanism to the endoplasmic reticulum from which, the Al fragment could potentially cross the membrane and enzymatically rnodify the 28s rRNA loop located on the ribosomes (shown as irregular purple globules). Toxin & Extracellular Matrix
w etrograde Lysosome Transport + Ribosomes al., 1991). It has been estimated that -4% of bound toxh per hour reach vesicular compartments beyond endosomes (Sandvig et al., 1991). However. before the toxin can reach and inactivate ribosomes, it is required to translocate across vesicular membranes. regardless of which cornpartment it occupies. Passing through the various compartments of the cell, the intemalized toxin is exposed to environments with varying pH. In the endosome, where acidification of the endocytic vesicles begins before hision with lysosomes, pH drops to -5 (Tycko & Maxfield, 1982; Ohkwna et al., 1982). The lumen of the lysosomes is more acidic in nature, reaching a pH of 4.6 (Tycko & Maxfield; Geisow &
Evens, 1984). In the Golgi network, pH is maintained slightly acidic at pH 6.4 (Kim et al., 1996) and goes back up to neutrality (7 to 7.2) in the ER (Bartido et al.. 1995).
The blockage of retrograde transport with brefeldin A protects cells sensitive to the action of ShT and suggests that the toxin must reach the lumenal cornpartments of the Golgi network and the ER before its Al subunit cm be effectively translocated near ribosomd subunits (Garred et al.. 1995b). Since the target sire for the N-glycosidax activity of the A 1 chain is a single adenine base of 28s rRNA, the Ai subunit must possess a mechanism to translocate to the cytoplasmic side of the ER membrane (Endo et al.. 1988; Saxena et al., 1989). In the case of diphthena toxin, there is evidence that the toxin exploits the acidification of the endosome to undergo a conformational change and exposes a hydrophobic domain that enables it to translocate across membranes (Donovan et al., 198 1; Hu et al., 1984; Blewitt et al.. 1985; and Rolf et al., 1993). A similar effect has been observed in the case of Pseudomonas exotoxin A where low pH alters its structure to reveal a membrane binding domain (Farahbakhsh & Wisnieski, 1989).
1.5 Structural and Funçtisnai Hornolo~iesWith Other Toxins
The structure and cellular mechanism of the A/B family of toxins has bmn the focus of several reviews (Middlebrook et al.. 1984; Olsnes et al., 1990; Read & Stein, 1993; and Menia & Hol, 1995). Members of this family of toxins are characterized by coding the enzymatic activity and the receptor binding activity on separate polypeptide chahs, which subsequently associate to form biologically active toxin molecules. The A subunit of SLT-1 contains distinct homologous regions with ricin A subunit and it has ben shown that both
Ca traces are supenmposable ( Sixma et al.. 1993; Read & Stein, 1993). Residucs in the putative active site are distinctly conserved. The B subunit shows sequence homology only to members of the SLT toxin family. However, its crystd structure revealed a surprising homology in terms of its folding and assembly with the E. coli LT toxin, cholera toxin and
Staphylococcus aureus nuclease (Amone et al., 197 1).
In addition to structural similarities, memkrs of the AA3 family of toxins dso share functionai similarities. Although the cellular targets of these toxins are different, they are all located in the cytoplasmic pool of target cells. For this reason, the toxins must overcome the limiting membrane of vesicular compartments during endocytosis. Early work on diphthena toxin using planar iipid bilayers and lipid vesicles showed that the toxin interacts with membranes, and in some cases, forrns transmembrane channels (Donovan et al., 198 1; Hu
& Holmes, 1984; and Zalman & Wisnieski, 1984). Funhermore, those studics showed that membrane interactions of diphtheria toxin were pH drpndent, being highrsi at pH 4.0 ~opH
5.0, comparable to those conditions found in late endosornes and lysosomes. The increased lipid interaction properties of a toxin at endosornal pH requires that the protein undergoes a confornational change to expose hydrophobic domains. Indeed, endosomal pH was found to induce a change in the conformation of diphtheria toxin (Rolf & Eidels, 1993; Blewitt et al., 1985; and Zhao & London, 1988). Dissection of the structure of DT led to the identification of a membrane translocation domain located between residues 202-378 (Zhan et al., 1995; Montich et al., 1995) and was recently show to insert into mode1 membranes (Zhan et al., 1995; Montich et al., 1995). Similar features were also identified in Pseudomonas exotoxin A (PE). The toxin molecule contains a distinct enzymatic, receptor
binding and membrane translocation domains (Ailured et al., 1986; Collier, R. J.. 1988; and
Hwang et al., 1987). As it was observed in the case of diphtheria toxin, PE required
proteolytic cleavage for cytotoxicity (Jinno et al., 1989; Ogata et al., 1990). One additional requirement for PE cytotoxicity is the teuapeptide motif' KDEL or REDL(K) at its C- terminus. a necessary signal for the proper invacellular trafficking of this toxin (Chaudhary et al., 1990).
In contnst to these toxins. however, SLTs and the closely related plant toxin. ricin. do not contain distinct translocation domains nor do they contain the KDEL motif. in spite of this missing feature. SLT-1 is tnnsponed to the GO and the ER and CR routc. is sxposcd to a variety of proteases including. among possible others. trypsin. funn and calpain (Gamd et al., 199%; 199%). The primary cleavage region in SLT-1 is Iocated in the disulfide loop. residues 242-261, in the A subunit (Fig. 1.2; residues 265 to 284 in the alignment nomenclature). Cleavage in this region has been shown to increase the potency of the toxin, both in vitro and in vivo (Garred et al.. 1995a; 1995b). Although results vary between different ceIl lines and incubation periods. deletions in the trypsin and furin sensitive .motif (RXXR) resulted in reduced cytotoxicity by as much as 100fold in T47D cells following a 1 hour incubation period (Garred et al., 1995a). Longer incubation periods seem to reduce the effects of the deletions. One explanalion for this latter observation is thac longer incubation periods result in the toxin being cleaved and activated by RXXR-independent protease(s). At the present time, no one has been able to demonstrate SLT-1 cytotoxicity without intracellular processing. In the case of ricin, the A and B subunits are dready separated from each other in the native toxin, but are associated through a disulfide bond and therefore. cleavage is not required. Interestingly, the proteolytic processing of intemalized ricin has ken demonstrated experimentally ( Blum et al., 1991). . . 1.6 -erttves
In an effort to understand intracellular mechanisms that may alter the structure and localization of SLT toxins. 1 have used purified SLT-I and its individual subunits. The objectives of my thesis were as follows :
1) Revious observations have suggested that the B subunit may be able to unfold in acidified compartments causing the release of the A2 fragment andlor contribute to the translocation of the Ai fragment. 1 decided to investigate the effects of pH on the stmcture of the SLT-1 B subunit (chapter 3). It was found that endosomal pH causes a local conformational change in the B subunit The observed change is not large enough in scale and would not occur in the nght compartment to support a mechanism for the translocation of the enzymatic subunit across the ER membrane. This change in conformation. however. may contribute to other events that occur in the endosorne. such as the proteolytic processing of the toxin A chain en route ro the ER. which is required for the translocation to occur.
2) As observed in the case of diphtheria toxin and exotoxin A, both the A and B subunits may participate in membrane insertion and translocation of the enzymatic fragment to the cytosol. To test for the occurrence of similar effects in SLT-1, the intluence of pH on the structure of the holotoxin has been investigated (chapter 4). It was observed that endosomal pH induces local conformational changes in the holotoxin but does not alter its secondary or tertiary structures. As mentioned in point (1) above. such local changes in conformation rnay facilitate proteolytic processing of the A subunit and disulfide reduction prior to its translocation in the ER compartment.
3) The reduction of the single disulfide bond of the A chah of SLT-1 and its limited proteolysis may facilitate the release of the Al fragment from the holotoxin andor expose its hydrophobic C-terminus. These cellular events may favour its interaction with membranes and suggests a mechanism for its translocation across the ER membrane. To test this hypothesis. we have investigated the interaction of the hydrophobic domain of the Al fragment with membranes (chapter 5). It was found that a synthetic peptide reprwnting this region associates with membranes in a pH-dependent manner. At neutrai pH, as in the lumen of the ER, this peptide inserts across the vesicle membrane. In the context of the translocation of the Al fragment, this peptide may mediate the insertion of the Ai fragment into the inner leaflet of the ER membrane and rnay constitute the fust step in its translocation.
My thesis work has produced two publications in Biochemistry (see chapters 3 and
5). In addition, 1 have contributed to two other bodies of work not discussed in the thesis.
One project was led by Dr. Eric Lacasse and published in BIood under the title "Shiga-like toxin purges human lymphoma from bone rnarrow of severe combined immunodeficient mice. Vol. 88, pp1561-1567". The other piece of work is being conducted by Dr. A. Menikh in coIIabontion with Dr. I. Boggs at the Hospitd for Sick Children. This latrcr work focuses on the use of ATR-FTIR to investigate the interaction between a synthetic peptide
(W232A242ShTA (220-246), see chapter 5) and lipid membranes. The manuscript that describes this work is presentiy king prepared. CHAPTER 2
EXPERIMENTAL METHODS 2.1.1-of-L Several LB agar plates containing carbeniciilh (100 ~g/ml)were streaked using a steel loop with an aliquo t of a culture of the bac terial strain JB 28 previously stored at -70'C. The plates were incubated at 37 % ovemight A single colony from such plates is then used to inoculate a lOOml LB broth containing 100pg/ml carbenicillin and
0.1 % (WN)glucose. The cultures were grown ovemight with mixing (250 rpm) at 37%.
Up to 6L of LB broth were prepared in the same marner and each 6L Eûsk (containing up CO 2L of broth) was inoculated with 5ml of the stock culture grown ovemight. After an incubation period of 18h. the culture was centrifuged at 10,000 rpm for 15 minutes and the pellet was either used irnmediately for subsequent toxin purification steps or stored at -WC for future processing. Protein extraction was performed by suspending the pellet in 400 ml of PBS containing 0.1 mg/ml polymyxin B -0.4 rng/ml EDTA and 0.1 mg/ml PMSF and rnixed vigorously using a blender. The bacterial suspension was subsequently incubated at
37'C with continuous stimng for lh. Cellular debris were removed by centrifugation
(10,000 rpm for 15 min.) and the supematant was collected and kept on ice for subsequent steps. The pellet was reextracted two more tirnes in the same way and the supematmts from the three extractions were pooled. The final supernatant was filtered through a glass-fibre filter paper (0.45 pm pore size) and concentnted using an Arnicon uluafiltration ceIl (500 ml capacity) connected to a niuogen gas cylinder (1; 60 psi). Ultrafiltration membrane (Amicon
YM IO) with a 10.0ûûMW cut off was used. The supematant was concenuated to 80- 100 ml at 4% and dialyzed (using dialysis tubing with a 35ûûMW cut of0 against 2L of 10 mM sodium phosphate buffer, pH 7.4 ovemight followed by another dialysis in the sarne buffer to bring the final NaCl concentration below 5 mM. aThe dialysate was loaded ont0 a hydroxyapatite column (2.5 cm x 30 cm) pre-equilibnted into the dialysis buffer. The column was then washed with diaiysis buffer until the absorbante of eluant was c 0.1 (280 nm). The bound protein was eluted in a batch method with two column volumes (300 ml) of
100 mM sodium phosphate buffer. pH 7.4 . The column flow rate was 1 mi/min. and 4 ml fractions were collected. The eluate was monitored spectrophotomevically at 280 nm. Fractions 36-50 were pooled and dialysed (using the sarne dialysis tubing described above) against 2L of 25 mM imidazole-HCl buffer, pH 7.4 ovemight at 4'C. The pooled fractions were loaded ont0 an anion exchange column (2.5 cm x 20 cm; polybuffer exchanger, Pharmacia Biotech) pre-equilibrated with the imidazoie didysis buffer. The column was then washed with 2-3 column volumes and the pH gradient deveioped (at a flow rate of 1 mllmin.) with Polybuffer 74 adjusted to pH 5.0 . nie elution profile was monitored at 280 nm and fractions (33-41) containing a single peak and centred at pH 6.9 were pooled and dialyzed (10,ûûûMW cut-off) against IO mM sodium phosphate buffer. pH 7.4 ovemight. QtJThe pooled fractions were subsequrntly loaded onto a Cibachron blue F3GA column (2.5 cm x 10 cm; Affi-Gel blur gel. Bio-Rad Laboratones) pre-zquilibra~edwith dialysis buffer followed by washing with 60ml of the same dialysis buffer. Bound toxin was eluted (flow rate of 1.8 ml/min.) using 200 ml of the dialysis buffer containing 0.5M NaCl.
The purified toxin ( fractions 18-21) was diaiysed against 4L of distilled water and lyophilized.
Purification of SLT-1 B subu& Clones expressing the SLT-1 B subunit (pJB 122) were grown in a total culture volume of 6L and the extractions were perfonned as described in section 2.1.1. The final supernatant was concenmted to 80- lûûml at 4'C and dia9scd
(using dialysis tubing with a 3,500 MW cut off) against 2L of 10 mM Tris-HCl buffer. pH Figure 2.1 Typical chromatography profiles observed dunng a purification of SLT-1. (A)
Hydroxyapatite chromatography of a concentrated extract of the bacterial svain JB28 treated with PBS containing polymyxin B. (B) Chromatofocusing of hydroxyapatite fractions containing the toxin using a pH gradient of 7.4-5.0. (C) Elution of bound SLT-1 from Cibacron blue gel using 0.5M NaCl in phosphate buffer, pH 7.4. Toxin-containing fractions were identified using anti-A subunit antisera. Solid bar indicates the fractions that were pooled and prepared for the next pufication step.
Figure 2.2 Typical chromatography profiles observed during the puritication of SLT-1 B subunit. (A) Ion exchange @EAE-Sephacel) chromatography of a concentnted extract of the bacterial strain pJB 122 treated with PBS containing polymyxin B. (B) Chromatofocusing of
DEAE fractions containing the B subunit The column was developed with a pH gradient going from pH 7.4 to pH 4.0. (C) Gel filtration of the B subunit collected from the chromatofucusing step using Sephadex G-50 (Phmacia Biotech.). Fractions containing the B subunit were identified using anti-B subunit antisera prepared in Our laboratory by Beth Boyd (1991). Solid bar indicates the fractions that were pooled and prepared for the next purification step. O 20 40 60 80 120 Fraction number
C 4 8
Fraction number
Fraction number 7.4 ovemight followed by another dialysis in the sarne buffer to bnng the fiai NaCl concentration below 5 rnM. The dialysate was loaded ont0 a DEAE-Sephacel column
(2.5 cm x 30 cm) preequilibrated with the dialysis buffer. The column was then washed with dialysis buffer until the absorbance of eluant was < 0.1 (280 nm). The bound protein was then eluted with a linear gradient (total volume of 600 ml) going from O to 0.6 M NaCl prepared in dialysis buffer. The column flow rate was 1 mumin. and 5 ml fractions were collected. Fractions 73-87 were pooled and dialyzed (using the same dialysis tubing stated above) against 2L of 25 mM imidazole-HC1 buffer. pH 7.4 overnight at 4'C. The pooled fractions were loaded onto an anion exchange column (2.5 cm x 20 cm; polybuffer exchanger, Phamacia Biotech) pre-equilibrated with dialysis buffer. The column was then washed with 2-3 column volumes and the pH gradient developed with Polybuffer 74 adjusted to pH 5.0 . The column flow rate was 1 mumin. Fractions (105-120) containing a single peak and centred at pH 5.8 were pooled and dialysed against 50 mM ammonium bicarbonate buffer. pH 7.4 overnight. The pooled fractions were concentratrd by ultrafiltration to 5 ml or less and loaded onto a G5O gel filtration column (7.5 cm x 50 cm:
Pharmacia Biotech). The column tlow rate was 1.4 mumin. and 4 ml fractions were collected. The purified B-subunit (contained in fractions 36-47) was dialysed against 4L of distilled water and lyophilized.
2.1.3 Peptide Svnthesis and Purification The peptides were assembled by solid phase synthesis on an Applied Biosystems 43 1 peptide synthesizer using FMOC chemistry. The following peptides were synthesized: ShTA(220-246), (14C)Ac-RVGRISFGSMAILGS-
VALILNCHHHA. which represents residues 220 to 246 in the native toxin and
W232A242ShTA(220-246), (I4C)Ac-RVGRISFGSINAWLGSVALILNAHHHA, which contains two substitutions, namely I232W and C242A. The N-terminus of each peptide was acetylated by reacting 100 mg of peptide-resin suspended in dichloromethane with 250 pCi of [1-Wlacetic anhydride ovemight at room temperature. The acetylation step was cornpleted by further treating the sample with 10% (vlv) cold acetic anhydride in dichlorornethane for 30 minutes pnor to cleaving the peptides from the resin. The peptides were purified by reverse-phase HPLC on a C18 column (Bechan ODS; 1 cm O.D. x 25 cm; column was pre-equiiibrated in water: 0.1% (vlv) TFA) and eluted with a Iinear gradient of acetonitrile (AcN): 0.08% (v/v) TFA. The steepness of the elution gradient was 2%
AcN/min. Both peptides typically elutzd from Ihc: column when the composition of the mobile phase reached 60% acetonitrile. The column flow rate was lml/min. The composition of each purifïed peptide was confirmed by amino acid andysis and mass spectrometry. The molar absorption coefficient of W232A242ShTA(220-246) at 280 nm was calculated to be
5700 M-km-1 .The concentration of the melittin stock solution was calculated using the Trp molar absorption coefficient of 5570 M-1 cm-1at 280 nm (Wetlaufer, 1962). The specific activity of the 1°C-labelled peptides was calculated to be 8.6 x IO2 cpm/pg for
W232A242ShTA(220-246) and 9.8 x 102 cpmlpg for ShTA(220-246). The HPLC elution profiles of both peptides are shown in Figures 2.3 and 2.4. The purified peptides were characterized by Mass Spectrometry (Figures 2.5 and 2.6) and amino acid analysis (Table 2.1).
2.1.4 Electroelution of SLT-I A subunit. For this procedure, normally TndGlycine gels (16 cm x 16 cm) were prepared and sample wells either contained 30pl (for the molecular weight markers and controls) or 400~1(for the purified SLT-1). The gels were run as mentioned in section 2.1.4. Immediately following the mn, the gel lane containing the toxin was cut into 2 Figure 2.3 Purifica~ionof ShTA(220-246). Panel A Reverse-phase HPLC profile on a semi-prep C 18 column. Arrow marks the peak containhg the peptide. -el B Reverse- phase HPLC profile of the peak collected from the semi-prep elution on an analytical Cl 8 column. Figure 2.4 Purification of W232A242ShTA(220-246). Panel Reverse-phase HPLC profile on a semi-prep Ci 8 column. Arrow marks the peak containing the peptide. -el B Reverse- phase HPLC profile of the peak collected from the semi-prep elution on an analytical Ci8 column. Figure 2.5 Mass spectrum of ShTA(220-246). System: ion-spray (SCIEX).Each of the
peaks shown represents a detected peptide fragment and is expressed as the ratio of
masskharge (rn/z). For example, the peak identified in the spectnirn by MA (at m/z of
725.4) reflects the presence of a peptide in the sample that has four positive charges (4 H+).
ïherefore, the mass of the peptide can be calculated from the foilowing formula : peptide mus = [(rn/z x z) - z]
From the peak at 725.4. the mass of the peptide is : [(725.4 x 4) - 41, or 2897.6 daltons. The expected mass of this peptide is 2898 daltons. Figure 2.6 Mass spectmm of W232A242ShTA(220-246). System: ion-spray (SCEX). The mass of the peptide cm be extracted from the data as showin Figure 2.5. mm sections. The section containing the A subunit, as identifid by staining a comsponding control lane, was mercut into smder sections (2 mm x 10 mm) and the protein with. was eluted out using mode1 422 Electro-eluter (Bio-Rad laboratones, Hercules, CA). The eluted A subunit was used to generate rabbit anti sera as describecl below in section 2.3.1.
Table 2.1 Arnino acid analysis of p-ed synthetic peptides.
. --- sh~~(220-246) W232A242S hTA AminoAcid Number Eqemü Number Present Number Expected Number Present Ala 3 3.2 + 0.3 4 4-4 -1: O. 1 A% 2 1.9 I0.6 2 1.7 IO.l Asx 2 2.2 f 0.1 2 2.3 i O. 1
G~Y 3 3.3 + O. 1 3 3.2 + O. 1 His 3 3.7 + 0.4 3 4.1 f 0.1 Leu 3 3.4 f 0.4 3 3.6 f 0.1 Re 4 4.0 I0.2 3 2.9 3- O. 1
Phe 1 1.1 t 0.2 1 0.9 +- 0.0 Trp -- -- 1 ND Ser 3 2.0 i O. 1 3 2.0 + 0.0 Val 2 2.2 =f= 0.3 2 2.0 I 0.0 CYS 1 ND - 1 -- 1.5 &dd&&g of SLT-1 and SLT-1 B dunitA mixture containing 180 pl of O. 1 M
sodium phosphate, pH 7.4 , 1 mCi of NalzsI and two iodobeads (Pierce Chernical Co.) was
deposited into a polypropylene tube and incubated at room temperature for 5 minutes. A 20
pl aliquot of SLT-1 or SLT-1 B subunit (1 mdml in water) was added to the reaction tube
and the resulting mixture was further incubated at room temperature for 10 minutes. The
iiquid was removed from the reaction via1 and loaded ont0 a Sephadex G25 çolurnn
(1 cm x 20 cm) previously equilibrated in 10 mM sodium phosphate. pH 7.4 . The column fIow rate was 0.4 rnVmin and fractions of 1.0 ml were collected and monitored by counting
the radioactivity in a y counter. The labelled protein eluted in the void volume and was well
resolved from the rernaining free 125L The preparations typically had a specific activity of
approximately 3x 106 cprn/pg of protein.
2.1.6 Gh?Solid Phase Rindin~Assav. The binding assay was performed following the
method of Rarnotar et al. (1990). Briefly, wells of flexible polyvinylchloride 96-well plates
were coated with 10 ng of Gb3 (Gb3: PC:Cholrstrrol. 2: 105) in methanol:chlorofom (9:1 )
and the solvents were allowrd to rvaporate ovemight at room temperature. Wells were thrn blocked with 200 pl of 2% (wfv) BSA in 10 mM PBS for 2 hours at room temperature.
washed 6 times with PBS (at the given pH) More 1251-labelled B subunit (3 x 104 cpmlwell) was added either in the presence (50 pg; 1000-fold excess) or absence of unlabeled B-subunit. After incubation at room temperature for 2 hours, the wells were
washed 6 times with PBS (at the given pH), cut and the bound radioactivity quantïfied in a y-
counter. To confirm that the denatured protein did not bind to its receptor in this assay, a sample of the B subunit was boiled for 10 minutes and the assay subsequently performed at neutral pH. Error bars represent the standard deviation associated with expenments perfomed in tnpïicate.
7 Preayl-la- of -0-7.46) Pm.Purified peptide (1.0 mg) dissolved in 160 pL of 8M urea and 40 pL of 3-maleimido proxyl (10 mglm1 in methanol) were mixed together to give a 5 molar excess of the maleimido-proxyl gmup over the reactive thiol side chah (Cys-242) of ShTA(220-246). Fifty pL of 0.25M MES buffer. pH 6.2 was then added to the mixture and the reaction left to proceed ovemight at room temperature. Excess proxyl label was removed by reverse phase HPLC using a semi-preparativr Cl8 column
(Bechan ODS; 1 cm O.D. X 25cm). nie column was drveloped with a linex gradittn~[rom
O to 100% acetonitrile/0.08% TFA. The steepness of the gradient was 1% AcN per minute and the proxyl-labelled peptide eluted frorn the column when the composition of the mobile phase reached 60% acetonitrïle. The column flow rate was 1mUmin. The concentration of proxyl-ShTA(220-246) samples was detennined by integrating the EPR spectra and comparing the spectrai signals to lhat of a free maleimido proxyl standard dissolved in 5 mM phosphate, pH 7.0. From this analysis. the molar ratio of proxyl group coupled to peptide was determined to be 0.83.
2.1.8 Preoaration of Linid Vesicles. DMPG or DMPC were dispensed from a lipid stock solution (500 pL of a 20 mg/ml solution) prepared in ch1oroform:rnethanol (1: 1 ) into a 16 x 125 mm borosilicate test tube and drkd under a Stream of nitrogen. Residual solvents wcrc removed under vacuum for three hours. The lipid films were then rauspended and hydralrd in 2 mL of either 5 mM phosphate. pH 7.0 or in 5 mM acetate. pH 5.0 using 10 cycles of fast freeze-thaw (Mayer et al., 1985). The large multilamellar vesicles formed were then subjected to sonication in a bath sonicator to visual clarïty. The small unilarnellar vesicles (SUV) produced by this procedure were used directly in the experiments. For calcein leakage experiments, the vesicles were fonned in the presence of 20 mM calcein in either phosphate or citrate buffer. Free calcein was removed by gel filtration on a Sephadex G 15 column (1.5 cm x 40 cm) and eluted using the same buffer in which the vesicles were prepared.
. . 1.9 Rindirg of Radiolabeled Peptide to 1. ipid Vesick W23 2A242S hTA(220-246) samples
(42 pg dissolved in either 5 mM phosphate, pH 7.0 or in 5 mM acetate, pH 5.0; swc activity. 8.6~102 cprn/pg) were dispensed into centrifuge tubes (13 x 5 1 mm; Beckrnan. polyallomer). Increasing amounts of preformed DMPG or DMPC vesicles suspended in the same buffer were then added to the tubes to reach a final volume of 4.0 ml. The tubes were then centrifuged at 175,000 x g for 5 hrs and the radioactivity of the supernatant was measured (Wang et d.,1993). Lipid concenuations were determined by the phosphorus assay (Rouxr et al.. 1975). The fraction of peptide bound to vesicles was calculated as 1 -
[SiipidStotaLJ,where Slipidrepresents the radioactivity (in cpm) ÿssociatrd with rhr supernatant recovered from centrifuging sarnples containing the K-labelled peptide in the presence of lipid vesicles at a peptide-to-lipid rnolar ratio of 150. The term StOta1represents the radioactivity (in cpm) recorded for control supernatants (radiolabeled peptide in the absence of lipid vesicles).
2.1.10 Total Phosphorus AuThe amounts of lipids used in the experiments were determined by quantifying the amount of inorganic phosphorus produced from complete hydrolysis of the phospholipids according to the method of Rouser st al. (1970). Following preparation of lipid vesicles (see section 2.1.8). aliquots from the vesicles stock solution were dispensed into glas tubes (8 mm 0.d. x 120 mm) and dried by lyophilization. Lipids in the tubes were then extracted with a chloroform:methanol (2: 1. v/v) mixture and dispensed into a new set of tubes (same type as above). Lipids were hydrolyzed by addition of 0.65 ml perchloric acid per tube and heated at 150' C in a heated metal block for 30 minutes. Approximately 2.3 of the tubes extended outside of the surface of the block which served as a condenser. Tubes were not allowed to dry at al1 times. After cooling, to each tube were added : 3.0 ml water, 0.5 ml of 2.5% ammonium molybdate and 0.5 ml of 10% ascorbic acid. Colour was developed by heating in a boiling water bath for 5 minutes and Lhr intensity measured at 797 nm. A standard curve was constructed from the sarne lipids used in the experïments.
2.1.1 1 Amino Acid An- Protein and peptide samples were subjected to amino acid analysis to establish their composition and to calculate the concentrations of stock solutions
(Henrikson & Meredith. 1984). Sarnples (containing approximately 50 nmol of the most common amino acid in that particulv sample) dissolved in water were mixed with 30 nmol of norleucine (acting as an intemal standard). and dried in acid-washed tubes (5 mm x50 mm). The tubes were placed in a via1 containing 200~1of 1% (vh) phenol in 6N constant boiling HCI. The vials were tlushed with nitrogen and sealed undrr vacuum. Vapour-phase acid hydrolysis was camrd out for 48-72h ai 108'C at 100 rnTorr. The vials werr then allowed to cool to room temperature and dned under vacuum. The samples were dissolved in
10 pl of a solution of ethanol:H20:triethanolamine, 212: 1 (v/v/v), vortexed and then redried under vacuum. The samples were resuspended in 20 p1 of denvatization solution
(ethanol:H20:triethanolamine:PITC, 7: 1: 1 : 1 (v/v/v/v)) and ailowed to stand at RT for 20 minutes. The samples were then dned under vacuum. The resulting phenylthio- carbamyl derivatives were separated and quantitated by reversephase HPLC (Bechan System Gold,
Beckrnan Instruments Inc., San Ramon, CA). Samples were dissolved in 100 pl of sample buffer (5 mM sodium phosphate, pH 7.2 containing 5% (vlv) acetonitrile). 20 pi aiiquots of
each sample were injected onto a silica-based C 18 column equilibrated in buffer A (19 g sodium acetate trihydrate, 0Sml triethanolamine and 0.06% acet0nit.de in 1L of H20,pH
6.4 adjusted with acetic acid). The column was developed with a linear gradient (0%)
acetonît.de in water. The elution profile was recorded at 254 nm and each amino acid was quantitated by integrating peak areas using Beckman System Gold software. A correction factor for each amino acid was obtained from reference chromatograrns of a standard mixture of amino acids (standard H, Pierce Chernical Co.) containing 30 nmol of norleucine to account for variation in the degrce of derivatization with PlTC and ihr differcntial stabilitirs
of amino acids and peptide bonds during the hydrolysis process.
2.2 Characterization of puritied SLT-I and its B subunit.
P-e Gel E-. Tris-tricine gels for discontinuous SDS-PAGE analysis of low molecular weight proteins were prepared as described by Schigger and von Jagow (1987). In some cases, ready-made tricine gels were used (Bio-Rad labontories,
Hercules,CA.; catdog # 16 1-0922). Protein sarn ples (1- 1Opg) were dissolved in Laernmli
sample buffer (Laemmli, U. K., 1970) consisting of 50mM Tris. pH 6.8 and containing 2%
(wlv) SDS. 4% (v/v) P-rnercaptoethanol and 5% (vlv) glycerol and heaied for 3 min. ar
1WC. Separating gels contained 10- 158 (wlv) acrylamide. 3% (wlv) bis-acrylamide, 0.3M
Tris-HCI, pH 8.45, 10% (vlv) glycerol, 0.1% (wlv) SDS, 0.05% (w/v) ammonium persulfate and 0.05% (vlv) TEMED. Stacking gels (km)contained 5% (wlv) acrylamide, 3% (wIv) bis-acrylamide, 0.1% (wlv) SDS, 0.05% (w/v) ammonium persulfate and 0.1% (v/v) TEMED.Proteins were separated under a constant current of 30mA/gel and visualized 42 by staining in 0.2% (w/v) Coomassie Brilliant Blue in 50% (v/v) ethanol and 10% (v/v) acetic acid. The mobility of SLT-1 and its subunits are show in Figures 2.7 and 2.8.
FoAnalvsis.1owing SDS-PAGE. proteins were eletrophoretically transferred to polyvinylidene difluonde (PVDF) membranes (Immobilon; Millipore Corp., Bedford. MA) using a semi-dry tmnsfer system (Amencan B ionetics, Hayward, CA), based on the method of Towbin et al. (1979). The transfer time was 15 min. at 250 mA, The membranes were then incubated in a blocking buffer [2% (wlv) Carnation powdered milk in
0.1M Tris-HC1.0.15MNaCl. pH 7.41 for 2h at room temperature (RT) or overnight at 4'C to minimize non-specific binding. The membranes were then washcd in Tris-buffered saline (nS) sevenl times before incubating them in Tl3S containing 0.26 (wlv) powdered milk and dilutions of either anti-B or anti-A subunit rabbit misera or rnurint: anti-B subunit mAb.
Following incubation at RT for 2h. the membranes wwere washed extensively with TBS and incubated in TBS containing horse ndish peroxidase (HRP)conjugated anti-rabbit or anti- mouse antibodies at a dilution of 1:50ûû at RT for 1.5h. After washing the membranes with TBS. bound antibodies were detected by the method of Young (1989). Bnefiy, 30 mg of' 3.3'-diaminobenUdine tetrahydrochloride and 10 mg of Qchloro- 1-naphthol were dissolved in 5 ml of methmol and combined with 40 ml of PBS and lOpl of 304 (vlv) hydrogen peroxide. Colour development was stopped by washing the membranes in distilled water.
. . 3 Cy- ASSIW,
The cytotoxicity assay was performed using Vero cells essentially as described by
Gentry and Dalrymple (1980) for HeLa crlls. Vero calls wcre susprnded in rissut: culturc microwell plates at a density of 6000 to 80cells per well in a total volume of 100pl. The Fi yre 2.7 Electrophoretic mobility of SLT-1. Fractions collected during the purification protocol were resoived on SDS-PAGE (Tricine. 16.5%C) and stained with coomassie blue.
Lane 1. crude 1828 exaact; lane 2, pooled hydroxyapatite fractions; lane 3. pooled chromatofocusing fractions; lane 4, punfied SLT-1 eluted from Cibachron blue gel and dialysed against water. Figure 2.8 (A) Electrophoretic mobility of purified SLT-1 B subunit. SDS-PAGE (Tricinr.
16.58 C) of purified SLT-1 B subunit stained with coomassir blue. Lanr 1. low molrcular weight markers (Pharmacia Biotech.); lane 2, purified B-subunit (6 pg); lanr 3. 12 pg of B subunit. (B) Western blot of SLT-1 and its B subunit The proteins were separated on SDS-
PAGE and the gel processed as described in section 2.1.5. Lanes 1 and 4; Kaleidoscope molecular weight markers (Bio-Rad laboratones. Mississauga, Ont.). Lane 2; purified B subunit as identified by anti-B subunit anUsera (Boyd et al.. 1991). Lane 3 ; purified SLT-1 as identified by anti-A subunit antisera (see section 2.2.1). 45
plates were incubated in a Caincubator ovemight at 37'C. The following day. SLT-1 was
added (in a volume of lOpl of media) to each well at various dilutioiis. The plates were
incubated at 37'C for a further period of 18 hours. The media from the plates was removed
by aspiration and the wells were rinsed once with fresh media Finaily, lOOpl aliquots of
media were added to each well followed by 10pl of the ceil proliferation reagent WST
-1 (Boehringer Mannheim. Germany). The plates were retumed to the 37'C incubator for 1.5
to 2 hours and the colour developed by the reagent was rneasured at 450 nrn and 690 nm. Results are expressed as a ratio of the two wavelengths and are convened to percent survival by taking the value of the colour intensity in the control wells (cells with no toxin added) as
100% survival. A representative assay is shown in Figure 2.10.
2.2.4 Cell-free Protein Svnthesis Inhibition as sa^ The protein synthesis inhibition assay
was performed using the TNT@ coupled reticulocyte lysate systern (Prornega Corp.. Madison WI) as described by the manufacturer in the technical bulletin number 126. Briefly,
a mixture of TNT reaction mix (24~1).TNT SP6 polymerase (24~1)~arnino acid mix (24~1.
lacking Leu and Met), Rnasin (12~1)and SP6 (12~1)was prepared. In a typical assay, 4pl of
this mixture was added to 12.5~1rabbit reticulocyte lysate in a reaction tube. The toxin. in
PBS, was added to this reaction mixture to give a final reaction volume of 25~1.Tubes were
incubated at 30'C for 1.5 hours. Subsequenrly. 5pl fractions of a 1Ox dilution of each reaction mixture was used to measure the luciferase activity in a luminometer. A typiçal result of this assay is shown in Figure 2.10. 0.000 1 0.001 0.0 1 O. 1 I Toxh concentration (ng/ml)
- Figure 2.9 A typical titration curve obrained in the Vero ceil cytotoxicity assay. Toxin concentration @,M)
Figure 2.10 A cypical curve obtained in the cell-free protein synthesis inhibition assay. Table 2.2 Amino acid analysis of pdedSLT-1 and its B subunit on Ch- The column (superdex 75. pharmacia biotech.) was calibrated at room temperature and the values of Vao$ for SLT-I and its B subunit were determined under the sarne conditions. Experùnents were run eicher in phosphate-buffered saline, pH 7.2 or in citrate-buffered saline, pH 5.0. Standards used were as follows : Blue dextran (2 ma), Bovine serum albumin (66 kh). Carbonic anhydrase (29 Da).
Ribonuclease A (18.8 kDa) and Myoglobin (14.5 Da). The retention time of blue dextran was used for the value of V,.
Figure 2.11 Calibration of the gel pemeation column.
t V, is the void volume of the column as determined by the retention time of blue dexuan. V, is the elution volume of standard proteins. of The Mohr Ab- Three stock solutions of
SLT-1 (that differed in concentrations) were prepared such that their absorbance at 280 nm was between O. 1 and 0.7 . Fractions of each solution were used for quantitative amino acid analysis as described in section 2.1.1 1. page 40. Based on the number of moles of Phe (see Table 4.2)- the concentrations of the stock soIution were cdculated and the vaiues were used
L- 0.5 - Slope = 84000 (UmoLcm) - - 0.4 - - - 0-3 - - - - 0-2 - -
O 1 2 3 4 5 6 7 SLT-I concentration (mole&) x10-~ Figure 2.12 ïhr absorption gradient of SLT-1.
Accordhg to Beer-Lambert law for absorption. where A is absorbance. c is the concentration. 1 is the path length (1.0 cm) and E (dope of the curve in Fig. 2.12) is the absorption coefficient It was determined to be 8.4~104L Mol-1 cm-1. eneration of Rabb t Antiser~The antisera were generated in New Zealand white rabbits using a primary injection and two booster injections. The pnmary injection was prepared by emulsifying 1.0 ml of SLT-1 A subunit (0.1 mg/ml in PBS) with 1.0 ml of Freunds complete adjuvant. The 2.0 mls of this mixture were used to inject two rabbits (1.0 ml each). The booster injections were prepared by emulsifying 1.O ml of SLT-I A subunit (0.05 mgml in PBS) with 1.0 ml of Freunds incomplete adjuvant. The first booster injection was deiivered to the rabbiü four weeks after the pnmary injection foilowed by the second booster injection four weeks after. The rabbits were bled 10 days &ter the second booster injection and the blood was crntrifuged al 3000 rpm for 10 minutes. Thr anri sera (supernatant) was collected in fractions of 0.2 ml ach and irnmediately stored at -70 'C.
U.2 Detection of Rabbit Anti-SLT-I A subunit Antibodies by ELISA. Purified SLT-I was dissolved in PBS at a concenu?ition of 10 pg/ml. The toxin was coated on 96-well ELISA plates (1 pg/well) at RT for 1.5 hours. The uncoated surface in the wells was then blocked by incubation in PBS containing 1% (wiv) powdered skim mik at RT for 1.5 hours. Serial dilutions of the antisera were made in PBS/ 0.05% TWEEN 20 and added to the wells, in triplicate. in a volume of 100 pVwell. Following an incubation period of 1.5 hrs at RT, the plates were washed several times with PBS and peroxidase-conjugated goat anti-rabbit antibody (at a dilution of 1:2000) was addrd and incubacd at RT for a further 1.5 hours. Finally, plates were washed and bound mtibodies were detected with the peroxidase substrate 2,2'-azino-bis(3-ethylbenzthiazoline-6-sulf0nic acid) diammonium salt (ABTS) as descnbed by the manufacturer (Sigma Chernical Co., ST. Louis, MO). The bound 0.01 0.00 1 0-0001 Antisera Dilution
Figure 2.13 Detection of immobilized SLT-I with ELISA. Four rabbit antisera were generated: STAl and STA2 (from two rabbits injected with the electroeluird SLT-1 A subunit) and STA3, STA4 (from two rabbits injected with pre-boiled elecuoeluted SLT-1 A subunit). Other details are as descnbed in section 2-32. page 5 1. 53 peroxidase converts the ELISA reagent ABTS to a product that absorbs at 405 m. This absorbance is thus a merisure of the amount of bound anti-A subunit antibodies. A srpicd ELISA assay is show below in Figure 2.14. The antisera were used consistently to identify
SLT-I and its A subunit fragments on dot blots and western blots following SDS-PAGE. nie antisera dilution used in these protocols was usually 1/1000.
2.4 Biophysical Methods
24.1 Fluorescence Spectroscopy. Fluorescence spectra were recorded on either a Perkin Elmer LS-3 fluorescence spectrometer (cell path length, 1.0 cm) or on a Photon Technology International fluororneter. The B subunit was dissolved in 0.15 M NaCl to a concentration of 80 pg/ml. Titrations were performed by adding known volumes of dilute HCl and NaOH solutions to the cuvette containing the protein smple. The measurements were obtained at 25'C. The excitation wavelength was set at 298 nm (tryptophan indole group) with the maximal emission signal occurring at 350 nm. Fluorescence emission spectra were recorded from 320 to 410 nm. In experirnents with peptide
W232Az42ShTA(220-246) and lipid vesicles. titrations were performed by adding 15 pL aliquots of preformed vesicles (5 mg/rnl) suspended in eirher 5 rnM phosphate. pH 7.0 or 5 mM aceüite. pH 5.0 to a 1 cm pathlength cuvette containing the peptide W232A242Sh-
TA(220-246) (3 FM) dissolved in 2.8 ml of the sarne buffer. The resulting solution was continuously stirred while spectra were acquired. The excitation wavelength was set at 280 nrn and tryptophan emission spectra were recorded from 300 to 420 nm. In the case of experiments involving calcein leakap from lipid vesicles. the excitation wavelength used was 490 nrn and the emission signal was monitored at 520 nm. In acrylamide quenching expenments, SLT-1 (0.18 mglml) or its B subunit (0.137 mg/ml) were prepared in either 20 mM phosphate buffer, pH 7.4 or 20 mM acetic acid, pH 3.0 (adjusted with NaOH). The samples (3.0 ml) in the cuvette (1.0 cm x 1.0 cm) were under continuous mixing at RT.
Acrylamide, at a concentration of 15 M, was added to the cuvette in 10 pl aliquots and me spectnim was recorded after each addition. The relative change in fluorescence was expressed in some figures as the ratio FOE, were Fo represents the initial sarnple fluorescence in the absence of either lipids of acrylamide and F represents the fluorescence following the addition of aliquots of lipid vesicles or ac~ylamide.
Circw Dichroism. Circular dichroism spectra were recorded at 23'C on either a
JASCO spectropolarimeter Mode1 1-720 (cell path length. 0.1 cm) or a AVIV CD spectrometer mode1 62A DS. The B subunit concentration was 140- 160 &ml dissolved in 0.15 M NaCl. Individual samples were prepared for each pH condition tested and the protein concentration was spectroscopically detemined pnor to and &ter spectrum collection using the extinction coefficient value at 278 nm for the B subunit of 9,500 M-km-1 (monomer; Boyd. B., 1992). The pH tiaation expenments were performed in triplicate. Each spectnim represents the average of 12 accumulated spectra. The B subunit has an isoelecvic point of 5.7 (Boyd, B., 1992). No aggregation was obsemed at pH levels near this value as confimed by the tecording of identical CD spectra for ten-fold diluted samples of the B subunit In experiments with peptides and lipid vesicles, a 15 pL aliquot of a 2-4 mM stock solution of either peptide dissolved in 50% (vh) TFE was diluted to a final volume of 2.8 mL in either 5 mM phosphate. pH 7.0 or 5 mM acetate. pH 5.0 and disponsed into a 1 cm pathlength cuveta. The final peptide concentration was iypically betwwn 10 and
20 pM. Spectra were collected with an avenging tirne period of 10 seconds per point. Each point represents the averaged ellipticity value recorded over a 0.5 nrn spectral interval. When spectra were recorded in the presence of lipids, the peptide was added directly to a cuvette containhg lipid vesicles suspended in a constantly stirring solution. Estimates of the helical fraction of the peptide were calculated by the mediod of Greenfield and Fasman (1969) using
the foilowing equation :
where the value of [û]222helix for a 27-amino acid long peptide was calculated to be -35.740 deg.cm2.drnol (Chen et al.. 1974). and the value for coi1 was determined from a peptide solution prepared in 6M guanidine.HC1.
2.4.3 EPR measurements, Spectra were collected using a Varian E-104B EPR spectrometer operateci in continuous wave mode and equipped with a Varian variable temperature module.
Samples were containcd in a sealed 50 PLglass micropipette. The peptide concentrations used were typically 20-30 pM and the peptide-CO-lipidmolar ratio was maintained constant ai
1: 120. Ascorbic acid causes the reduction of the spin label free radical thus abolishing its
EPR spectnim. In the case of ascorbate reduction experiments. 2 pL of 1.0 M sodium ascorbate prepared in either 5 mM phosphate, pH 7.0 or 5 mM acetate, pH 5.0 buffers, was added to 120 pL of the peptide/vesicle mixture imrnediately before data collection. In some reduction experiments. ascorbate ions were entnpped into SWs during vesicle formation and the loaded vesicles pded from free ascorbate ions by gel filtration chromatography as described above for calcein. Peptide solutions were thcn mixed with loaded SUVs just prior to data acquisition. For fast. nearly isotropie motion. the motional parameter su was obtained as described elsewhere (Eletr and Keith, 1972). In the case of spectra ~tlecting slower motion. relative changes in motion were determined from the maximum hyperfine splitting, Tm,. However, due to the low amplitude of the spectra and broadness of the outer high field peak, this value was measured as the separation of the outer low field hefrom the midpoint of the centre peak where it crossed the basefine, rather than as one half the sepantion of the outer low field fine from the outer high field line. LOCAL CONFORMATIONAL CHANGE IN THE B SUBUNIT OF
SHIGA-LIKE TOXIN AT ENDOSOMAL pH
This chapter represents an edited version of a publishrd paper rzprinted with permission from : " Saleh, T. M. and J. Gariépy (1993) Local conformational change in the B subunit of
Shiga-Like toxin at endosomal pH. Biochemistry 32:9 18-922." copyright 1993 American Chernical Society. 3.1 Introduction
Understanding the mechanism of entry of protein toxins into cells is of fundamental importance in the development of prophylactic approaches and of strategies to deiver dmgs. antigens and bioactive agents to ceilular compartments. A panicularly successfui toxin design is the "wu-subunit assembly " where one subunit (A subunit) is an enzyme that specifically modifies a regdatory or essential component of the host cellular machinery while one or more copies of another subunit (B subunit) is (are) involved in the recognition and binding of the toxin to surface receptors located on target cells. Exarnples of this structural design include diphtheria toxin, cholera toxin, the E. coli heat-labile enterotoxins and the
Shiga (ShT) and Shiga-like (SLT) toxins. The B subunits of Shiga and related toxins
(Seidah et al., 1986; Calderwood et al.. 1987; DeGrandis et al., 1987; Jackson et ai..
1987a.b; and Strockbine et al., 1988) reprexnt the smallest subunits (- 70 amino acids) known to participate in this type of structural motif. These B subunits spontanrously assemble into pentamers in solution (Ramotar et al.. 1990; Head et al., 199 1; and Saleh & Gariépy, 1993) and bind to unique ce11 surface glycolipids (Lindberg et al.. 1987; Lingwood et al., 1987). The pentamer non-covalently associates with a single A chah and is endocytosed through coated pits (Sandvig et al.. 1989). The A chah must traverse vesicular compartments and reach the cytoplasm for the toxin to be effective. In this work, results are presented which suggest that in addition to the three functions already canied out by this small protein (pentmerization. A subunit and glycolipid-binding properties). the B subunit may play a role in the mechanisrn of translocation of the A chain.
Earlier work on the isolated A subunit of SLT-1 had shown that it can interact with
DMPC vesicles (change in tryptophan tluorescence; Surewiçz et ai.. 1989). Binding of' the A chah to lipid vesicles may have been due to the hydrophobic nature of its A2 domain. In fact, the crystal structure of the holotoxin has reveaied that the C-terminal residues of the A chah (A2 domain) are inserted into the hydrophobic central pore of the pentamer (Fraser et al.. 1994). The A2 domain is strongly associated with the B pentamer and it is highly unlikely that this domain would be available to interact with membranes following the intemalization of the toxin. One possible rnechanism by which the A2 dornain could be released would be through the unfolding of the B subunit as a result of acidification of intracellular vesicles containing the toxin. In addition to releasing the A2 fragment, unfolding of the B subunit may prornote its interaction with membranes and facilitate the release of the enzymatic Ai fragment to the cytosol.
The objective of this study was to probe for the existence of one or more conformational states for the B subunit of Shiga-like toxin 1 at acidic pH conditions that would suggest its involvement in the translocation mechanism of the A chain across the endosomal cornpartment. These states were contrasted with those arising from protein denaturation events based on the sensitivity of its receptor-binding property to protein denaturation.
3.2.1 The fIuorescence of tryptophan-34 is perturbed ut endosornul pH. The presence of a single tryptophan residue at position 34 (Fig. 1.4) in the sequence of the B subunit allows us to monitor local pH-sensitive perturbations in the structure of the pentamer induced by pH changes using fluorescence spectroscopy. As shown in Figure 3.1, a sharp and reversible transition in fluorescence intensity occurs at pH 4.5 . No shift in wavelength maxima (350 nm) was observed under such conditions. From the crystal structure of the B subunit Figure 3.1 Effect of pH on rhe tryptophan fluorescence intrnsity of the B subunît of Shiga- like toxin 1. Samples were initially prepared in dupiicatt: in wlitrr conÿiining 150 mM NaCl.
One sample was tiuated to low pH and one to high pH conditions ( O ) then both were titrated back to neuual pH ( ). Ae,,=298 nm;he,.=350 nm. pentamer, the tryptophan side chah in each monomer is located at the N-terminus of the single a-helix (residues 36 to 46; Figure 2A; Stein et al.. 1992) foming a "doorwaytfto the central channel of the pentamer. The loss in fluorescence intensity below pH 5 could be interpreted as the result of an increased exposure of the indole ring to a polar environment. The crystal structure solved at neuiral pH provides a clear indication that tryptophan-
34 is exposed in all five subunits of the pentamer (Stein et al., 1992). Aspartic or glutarnic acid side chains generally have pKa values between 4.4 and 4.6 in the context of proteins (Cantor & Schirnmel. 1980) suggesting that the ionization srate of one or more of these side chains may influence the environment of Trp-34. Examination of the crysral structure of the pentamer does not place any such side chains in the vicinity of the indole ring. However, in three of the five subunits of the pentamer, the guanidinium group of arginine-33 is located les than 2.8 A from side chah oxygens of either Asp 16 or Asp 18 present on a loop region of an adjacent subunit (Figure 3.2). This spatial proximity suggests the existence of a salt bridge that may be disrupted by the protonation of such aspartic acid side chains. The presence of a free guanidinium group at low pH rnay then influence the polarity of the tryptophan side chah environment or destabilize the N-terminus of the a-helix through an
unfavorable interaction with the helix dipole (Ho1 et al.. 1978; 198 1;Shoemaker et al.. 1987).
3.2.2 7he structure of the B subunit is gradually altered by pH conditions ranging from pH
2 to 4.5. The secondary structure of the B subunit of Shiga-like toxin 1 remains unaffected by pH conditions between 4.5 and pH 10 as identical spectra are recorded by circular dichroism in the far U.V. region (195 to 250 nm). The spectrum of the B subunit is altered as the pH of the sample drops below 5 (Figure 3.3A). These pH-sensitive spectral changes are reversible until pH 2 is reached. The pentamer is significantly denatured at pH conditions below 2 and above 10 (Figures 3.38 and 3.4) as the spectnirn of the pentamer at neutral pH cannot be regenented following the back titration of such samples. The circular Figure 3.2 Stercoprojcctions of the B subunit pentamer of Shiga-like toxin 1 (adapted from
Stein ct al.. 1992). (A) Rihbon-like representation of the pzntamer (Insigh~II soîlware:
Biosym Technologies. San Diego. CA). The single a-helix in each monomer gives rise to a closc arrangement of 5 helices in the pentamer. The indole side chin of uyptophan-34 is located at thc top of each hclix. The B subunits have identical sequences and are composed of 69 amino acids. Sccondary structure composition of the B subunit (Stein et al.. 1992):
Pl. rcsiducs 3-8; P2. rcsidues 9-14; P3. residues 20-24; P4, residues 27-31; a-helix. rcsiducs 36-46; 05, rcsidues 49-53; 06, residues 65-68. The positions of Asp-16, Asp-17,
Asp- 18 and Arg 33 an: illustrated in relation to Trp-34. (B) Identical representation as in A) without thc rihhons. Thrcc possiblc salt bridges are highlighted ( A ) where the distance scparating Lhc aspartic acid sidc chain oxygens of -either Asp- 16 and Asp- 18 and the guanidinium group ou Arg-33 is less than 2.8 Â (sec text). The subuniis of the pentamcr are dphahcticdly labelled from A io E and the putative salt bridges would represent intenubunit con tacts.
63 dichroism spectra of SLT-1 B between pH 2 and 10 have negative maxima at 208 and 222 nm supporting the presence of a-helices (Figure 3.3A; Brahms & Brahms. 1980). At neutral pH. residues 36 to 46 adopt an a-helical conformation which represents 16% of the secondary structure elements of the molecule (Stein et al.. 1992). A graph of eilipticity values at 222 nm as a function of pH (Figure 3.4) suggests a change in a-helical content starhg at pH conditions lower than 4.5, with ellipticity rneasurements at 222 nm incxasing from -8,100 deg.cm2.dmol-1 at neutral pH to -9,MO deg.cm2.drnol-1 at pH 2 (Fig. 2.4).
The gradua1 change in conformation observed between pH 2 and 4.5 (Figures 3.3A and 3.4) is not concomitant with the sharp local perturbation observed for Trp-34 (Figure
3.1) and would suggest the existence of at least two conformational stages for the B subunit pentamer at low pH. As discussed below, both states involve stable conformations rather than denatured states of the B subunit since the pentarner retains its ability to bind the glycolipid Gb3 under these conditions. The conformational ~ansitionoccumng below pH
4.5 may be shifted to a higher pH (closer to endosomal pH for example) in the context of the B subunit being bound to a Gb3-containing membrane or in association with the A subunit.
Indeed, the pH-induced conformational change in the A chah of diphtheria toxin is sensitive to its environment, with the transition pH between hydrophillic and hydrophobic states king shifted from pH 3.5 to 5 in the presence of smail unilarnellar vesicles (Zhao & London, 1988). Its isolated B subünit also undergoes a pH-induced exposure of a hydrophobic domain (Boquet et al., 1977; and review by London, E., 1992).
3.2.3 The tertiary and quatemary structure of the B subunit pentamer are stable between pH 2 and 10. The crystal structure of SLT-1 B subunit has recently been derived from the
X-ny diffraction analysis of crystals grown at neutral pH (pH 6.8-7.0; Stein et al.. 1992). The protein exists as a pentamer. The peptide backbone adopts a structure reminiscent of the 200 210 220 230 240 250 200 210 220 230 240 250 Wavelength (nm) Wavelength (nm)
Figure 3.3 Effect of pH on the circula dichroisrn spectrum of the B subunit of Shiga-like toxin 1 (195-250 nm region). (A) Spectra were recorded under neutrai and acidic pH conditions: pH 7.0, pH 4.0. pH 3.0. and pH 2.0 (see legend in figure). (B) Spectra of denatured forms of the B subunit at pH 1.3 and pH 12.0 are contrasted with the reference spectmm of the B subunit recorded at pH 7.0 . Each spectmm represenü the average of 12 accumulated spectra (refer to experirnentat methods in chapter 2, page 54). Figure 3.4 Effect of pH on eliipticity values ai 2?2 nm. Ellipticity measurrrnents ([el222nm) were derived from spectra recorded as descnbed in Figure 2.3. The graph was consvucted using spectra collected from three independent titrations. The pH and protein concentration of each sarnple were detemined prior to and after specuum acquisition. The ellipticity data was extracted from individual spectra representing the average of 12 accumulated spectra (refer to experimental methods in chapter 2. page 54). larger B subunit of the E. coli heat-labile enterotoxin (Sixrna et al., 1991; Stein et al. . 1992;
Sixma et al.. 1992). Each B subunit is composed of 6 B-stmds and one a-helix Linked together by a series of shon tums. Pentamerization occurs as a result of intersubunit contacts between two Pstrands (P2 and P-6) on adjacent B subunits (Figure 3.2A; Stein et ai.,
1992). The SLT-1 B subunit spontaneously assembles into pentamers in solution, with the presence of monomers or other rnultimers not king observed by gel fütration chromatography or by sedimentation equilibnum experiments (Ramotar et al., 1990; Saleh & Garîépy, 1993). Monitoring the oligomeric state of the B subunit by high performance gel filtration chromatography at pH 5, thus approximating the pH of the endnsomal or lysosomal compartments (Maxfield. 1982; Geisow. 1984; Geisow & Evans. 1984). revealed that the structure of the oligomer remains intact eluting with the expected retention time and showing no sign of disaggregation (results not shown).
A more important criteria proving the integrity of the B subunit structure was to monitor the ability of the pentamer to bind to its natunl receptor. the glycolipid globotriaosylceramide (Gb3) (Lingwood et al.. 1987). Indeed, denaturation events such as the reduction of the single disulfide bond in SLT-1 B foliowed by carboxyarnidomethylation, lead to the dissociation of the pentamer and result in a loss of receptor binding property (Boyd, 1991). In this assay, Gb3 was immobilized ont0 solid phase and radioiodinated
SLT-1 B was allowed to associate with the bound receptor (Rarnotar et al., 1990). The radiolabeled subunit was incubated in a range of pH conditions, pnor to and during the binding step. The level of specific binding was assessed by performing the binding step in the presence and absence of an excess of unlabeled B subunit. From the histogram presented in Figure 3.5, one can conclude that the specific binding of the toxin to its recaptor is not affected by acidic pH as low as 3.5. A significant level of specitic binding still occurs even after exposing SLT-I B to pH levels below 2. A similar result was observed when the radiolabeled B subunit was aliowed to bind initially to Gbs at neutral pH. and the complex later exposed to lower pH conditions (pH 3 and pH 5; results not shown). The stability of the tertiary or quatemary structure was more sensitive to alkaline conditions panicularly pH levels above 10. Boiling the radiolabeled B subunit for 10 minutes pnor to performing the binding assay resulted in a total loss of binding to Gb3 (Figure 3.5).
In the changes obsewed in the tryptophan fluorescence and circular dichroism experiments under acidic pH conditions do not parallel a loss in tertiary or quaternary structure. These changes suggest the existence of a discrete conformational transition at endosomal pH's. A similar effect was not observed at alkaline conditions where denaturation as monitored by 105s of receptor binding property correlates with large changes in ellipticity values at 222 nm (Figure 3.5). Figure 3.5 Histograrn depicting the effect of pH on the specific binding of radioiodinated
Shiga-like toxin 1 B subunit to its naturd receptor Gb3. Open bars. toul binding; closed bars. non-specifïc binding (presence of excess unlabeled B subunit). B. denanired B subunit as a result of boiling the protein sarnple pnor to performing the binding assay. Error bars represent the standard deviation associated with experiments performed in triplicate (refer to expenmental methods in chapter 2. page 37). CHAPTER- 4
EFFECT OF pH ON THE CONFORMATION OF THE
HOLOTOXIN
This chapter represents a series of unpublished results and observations relating to the effect of pK on the structure of the holotoxin. CHAPTER 4
4.1 In traduction
Exposure of intemdized bacterial toxins to reduced pH levels may resdt in conformational changes in one or several parts of the molecule. Work with diphtheria toxin (DT) provided evidence that the A fragment may be responsible for membrane penetration and translocation, a function previously suspected to be encoded by other subunits. Zhao and coworkers (1988) have demonstrated that diphtheria toxin A subunit undergoes a reversible conformational change at pH below 5, which results in the exposure of tryptophanyl residues. These conformational changes also result in the exposure of a hydrophobic domain. as judged by an increase in toxin binding to detergent micelles.
Proteolysis protection experiments with Pseudornonus urruginusu exotoxin A (PE) and diphthena toxin implicates that both the A and B subunits rnay function together in a concerted manner during membrane insertion and translocation. Incubation of cells with 1 251- labelled diphtheria toxin followed by treatrnent of cells with pronase showed that two fragments of 21 kDa and 25 kDa were protected from proteolysis (Moskang et al.. 1988).
The 2 1 kDa fragment was identified as the A fragment whereas the 25 kDa represents the B fragment. Exotoxin A, on the other hand. is a single polypeptide chah of 613 amino acids
(Grey et al., 1984) cornprising three functional domains (1. II and In). In expenments involving exotoxin A and lipid vesicles. it was found that the toxin induces vesicles permeabilization at acidic pH by formation of ion-conductive channels (Menestrina et al.. 1991). The toxin dornain interacting with lipids was hypothesized to be the highly helical domain II of the toxin and includes residues 253-30 (Allured et al.. 1986). This region falls between domain 1 (enzymatic domain) and domain III (receptor binding domain). It is therefore possible that in some protein toxins. domains may fûnction as one unit in membrane penetration. In light of previous results observed with DT and PE, we decided to test whether or not SLT-1 may have membrane active properties and to rnonitor the effect of pH on the structure of the holotoxin. Since most of the hydrophobie regions of SLT-I are buried in the mature structure, the toxin will have to alter its conformation. We hypothesize that the acidification of its environment may trigger a confonnationai rearrangement. Unlike in the case of diphtheria toxin and exotoxin A, acidîc pH does not induce SLT-1 to translocate across surface membranes of cells. Proteolysis protection experirnents would thus not be revealing. Fluorescence and circular dichroism spectroscopies provide sensitive tools to monitor small changes in the structure of the holotoxin. The B subunit of SLT-1 does undergo a small conformational change at low pH levels but such a change may not be directly involved in membrane translocation of the A subunit (see chapter 3). The A subunit of SLT-I may mediate its own translocation across intracellular membranes. as was observed in diphtheria toxin. Alternatively. the site of membrane translocation by SLT-1 has been hypothesized to br in the lumen of the endopiasmic reticulum (ER) of sensitive ceIls (Gamed et al.. 1995a). where the pH is nrar neutrality (Bartido et al.. 1995). It cmbe therefore argued that Shiga and related toxins rnay not need to change structure at acidic pH since the translocation would occur at neutral pH. Nonetheless, the toxin may require acidic pH to change its conformation in order to alter its interaction with endogenous factors (proteases. Lipids or protein translocation factors) en route to the ER. Such interactions would be crucial for subsequent interactions with and translocation across the ER membrane. In this section. we report the effect of acidic pH on the conformation of SLT-1, as determined by fluorescence spectroscopy and CD, and discuss the results in the context of membrane insertion and translocation. Figure 4.1 Various representations of the structure of SLT-1 highlighting the locations of the tryptophans. (A) The A subunit (yeilow) has two Trp residues (stick models in magenta) at positions 203 and 277. Each of the B chains (green colour) has one Trp residue (stick model in magenta) at position 34. (33) The same colour scheme as in (A) in a space-filling mode1 to show the exposure of the Trp side chains to the solvent It can be seen that the tryptophans in the A subunit are not exposed to the solvent (C) The same model shown in (B) but the structure is rotated such that the view is at the receptor binding face of the B subunit pentamer. Trp-34 in the B subunit is shown in magenta and dethose of the A subunit, the B subunit Trp is clearly exposed to the solvent The C-terminal helix of the A subunit is shown (yeilow), as well as a portion of Trp-277 (magenta) through the cenaal pore of the pentamer.
4.2.1 Acidic pH Alters the Microenvironment of Tryptophun(s). SLT-I contains a total of seven Trp residues; one in each of the five B chahs and two in the A subunit (Fig. 4.1). These hainsic probes allowed us to monitor the conformation of the toxin at different regions in the toxin. The crystal structure of the free SLT-1 B subunit and that of the holotoxin show that the five Trp residues (Trp-34) in the B subunit pentamer are clustered and exposed to the solvent (Fig. 4.1). This feature is reflected in the emission fluorescence spectrum of the B subunit (Fig. 4.2). which shows a maximum at about 353 nm. The Trp fluorescence emission spectrum of SLT-1 is not syrnmetrical . indicative of heterogencity in the environment of Trp residues (Fig. 4.2). The specuurn of SLT-1 also shows a distinct plateau at about 335 nm. This blue shift in maxima retlrcts the shielded nature of Trp in the
A subunit (Fig. 4.1 ).
The pH titration of SLT-I shows a complex tiuation curve that reflects the heterogenous Trp populations within the toxin. Association of the A subunit with the B pentamer occurs through the central pore and through the A-face of the pentarner (Fig. 4.1) and does not affect Trp-34 in the B subunit. Thus. It would be reasonable to assume that pH will alter the microenvironment of Trp-34 to a similar extent as in the free B subunit. Titration of Trp-34 with pH shows a sharp and reversibie quenching of the fluorescence centred at pH 4.5 (see chapter 3). This is not observed in the pH titration of the holotoxin. Instead. the titration curve of SLT-1 (Fig. 4.3) shows three distinct regions: a region of small quenching between pH 7 and pH 4. a region of sharp quenching ktwren pH 4 and pH 3 followed by another slow quenching region below pH 3. The changes observed brtween pH 7 and pH 3 are reversible whereas those observed below pH 3 are not reversible. It will not be possible to assign these regions to specific Trp residues. However, it shows that the Wavelength (nm)
Figure 4.2 Tryptophan fluorescence emission spectra of SLT-I (solid line) and SLT-1 B subunit (dashed line) in PBS. The protein concentrations used were 0.18 mg/ml (18 phd - Trp) for SLT-I and 0.137 mgml (18 pM Trp) for SLT-1 B subunit (refer to experimenlal methods in chapter 2. page 53). Figure 4.3 Change in uyptophan fluorescence of SLT-1 as a function of pH. The pH of the
solution (150 mM NaCl in water) containing the.toxin was reduced by the addition of IO@
aliquots of 60 mM HCI to the 1.0 cm cuvette, containing a total of 3.0 ml of sample. The solution was continuously mixed. The toxin concentration was 0.18 mg/ml (refer to expenmental methods in chapter 2, page 53). 76 microenvironment of more than one Trp in the holotoxin may be perturbed by acidic pH.
42.2 Acidic pH Does Not Alter nie Exposure of Tryptophans in SLT-1. Acrylarnide quenching was used to monitor the exposure of tryptophans in SLT-1. Quenching of Trp-34 fluorescence in the free B pentarner with acrylarnide clearly demonstrates the exposure of the indole ring and correlates with the emission maximum at 353 nm observed in Figure 4.2. The quenching effect can be fitted by the Stem-Volrner equation (Birks, I. B.. 1970):
which gives a value of 8.1 for K,,(eff) at pH 7.4, a value comparable to numbers reported for other proteins with exposed tryptophan (Eftink & Ghiron, 1976; 198 1; Macek et al., 1995). The sarne equation cm be used to determine K,,(eff) from the initial rate of quenching of SLT-1 (Fig. 4.4) and a value of 2.4 was obtained at pH 7.4. This value agrees with other values obtained from proteins containing shielded Trp residues ( Eftink & Ghiron, 1976; 1981; Macek et al., 1995) and further supports the results observed in the previous section. The quenching of Trp emission at acidic pH is achieved at a similar rate to that observed at neutnl pH (Fig. 4.4). as determinsd from the initial rate of acrylmidc quenching. At higher acry lamidc concen tntions, the quenching curvrs deviate frorn cach other, particularly in the case of the holotoxin. The deviation is not large and is primarily due to experirnental enors. It can be concluded from this section that the decrease in Trp fluorescence at acidic pH in the holotoxin is not due to exposure of shielded tryptophans but rather due to changes in the polar@ of the Trp microenvironment as a result of changes in the ionic state of charged side chains in the vicinity of Trp. [Acrylamide] , Molar
Figure 4.4 Acrylarnide quenching of Trp fluorescence in SLT-1 and its B subunit at neutral and at acidic pH. The quenching was performed in either 20mM phosphate buffer (pH 7.4) or 20 mM acetic acid (adjusted ro pH 3.0 with NaOH). Concentrations of proteins werr as specified in Fig. 4.2 (refer to expenmental methods in chapter 2. page 53). 4.2.3 pH hes not akr the secondmy strucntre of SLT-1. Monitoring the secondary structure of SLT-1 pmvides information about global changes in its conformation that may not be apparent by fluorescence spectroscopy. The CD spectnim of SLT-1 in the far UV
(Fig. 4.5) shows negative maxima at 222 nm and at 208 nm. which retlect the presence of a- helices. This result was expected since the X-ray structure of the holotoxin shows chat the A subunit is made up of a large number of a-helices (Fraser et al.. 1994). The CD specti-um of SLT-1 at neutral pH gives an ellipticity value of - -12.000 deg.cm2.dmol-1 at 222nm (Fig.
4.5). This value translates to an a-helical content of approximately 30%. This value agrees with the calculated a-helical content derived from the X-ray structure of SLT-I (Fraser et al.. 1994). where the A subunit is about 30% a-helical and the B subunit is approxirnately 15% a-helical. The CD spectrum of SLT-1 at pH 7.4 and pH 5. representing a range of pH levels expected to be present in intracellular cornpartments. are identical (Fig. 4.5). This result indicates that the secondary structure of SLT-1 is not sensitive to mild acidic conditions. Both fluorescence and CD obsemations were contïrmed by size exclusion chromatognphy (Table 4.1). Changes in the folding of a protein results in changes io its volume. which in turn changes its mobility in a high-pressure size exclusion experiment
(Corbett and Roche. 1984). As shown in Table 4.1. pH 5.0 does not affect the folding of SLT-1 or its B subunit The differences in retention times seen in Table 4.1 between neutral and acidic pH fall within expimental error. This technique is limited in sensitivity and wili not detect small changes in conformation. However, given the scale of conformational changes that would be required to mediate membrane insertion of the holotoxin or its enzymatic fragment, it is clear that endosomal pH alone does not produce such effects on the holotoxin or its B subunit In contrast to these results. endosomal pH have significant effects on the conformation of DT and PE. Employing similar techniques. Blewitt et al. (1985) have shown that endosomal pH induces a conformational change in DT leading CO rhe exposure of al,b 15 Ili, IlL, IF,, - SI,, - - -
------
C* - -- - -15 11111,1, 1,~~,1,, 111, 200 210 220 230 240 250 Wavelength
Figure 4.5 CD specw of SLT-I in various pH conditions. The specua of the toxin (0.12mg/ml final concentration) were collected in 20mM phosphate. pH 7.4; 20mM citrate, pH 5.0 and in 20mM acetic acid, pH 3.0. The pathlength was 0.1 cm and the experiments were perfonned at room temperature (refer to experimental methods in chapter 2, page 54). Trp residues. Associated with those changes was the exposure of hydrophobic surfaces that lead to increase in the membrane interaction of DT. SimiIar effects were observed in PE, where low pH was found to induce a change in the structure of the toxin that were suggestive of partial unfolding and increased hydrophobicity (Farahbakhsh et al., 1987; Jiang and London. 1990).
Table 4.1
Effect of endosornai pH on the gel pemeation properties of SLT-1 and its B subunit
SLT-I 1.36 1.33
* See chapter 2, page 49 for details
In conclusion, it was observed that SLT-1 does not contain a pH-sensitive rnolecular "switch", which could potentially lead to the exposure of a previously-shielded hydrophobic domain segments and thus, allow the toxin to insert into the membrane in the appropnate cornpartment and mlocate the Ai hgment to the cytosol. It is therefore necessary to re- evaluate the search for a pH-triggered switch in the holotoxin and search for another mechanism that members of this family of toxins may utihto access their target in the 8 1 cytosol. As was observed in the case of the isolated SLT-1 B subunit (chapter 3)- acidic pH induces local conformational changes in die structure of the holotoxin. During intracelMar trafficking, such effects may be important for the proper cleavage of the toxin andlor the reduction of the disulfide bond holding the Al and the A2 fragments. Improper inuacellular processing could be detrimental to the delivery of the enzymatic fragment to the cytosol. CHAPTER 3
INSERTION AND ORIENTATION OF A SYNTHETK PEPTIDE
REPRESENTING THE C-TERMINUS OF THE Ai DOMAIN OF
SHIGA TOXIN INTO PHOSPHOLIPID MEMBRANES
This chapter represents an edited version of a published papa reprintrd with permission from : " Saleh. T. M. , J. Ferguson. J. M. Boggs and J. Gariépy (1996) Insertion and orientation of a synthetic peptide representing the C-terminus of the Ai domain of Shiga toxin into phospholipid membranes. Biochemistry 3593254334." copyright 1996 American Chernical Society. CHAPTER 5
Since the target site for the N-glycosidase activity of the Ai chah is a single adenine base of 28s rRNA (Endo et al., 1988; Saxena et al., 1989). the Al subunit must possess a mechanism to translocate to the cytoplasmic side of the ER membrane. In the case of diphthena toxin. there is evidence that the toxin exploits the acidification of the endosomal cornpartment to undergo a conformational change and exposes a hydrophobic domain that enables it to translocate across membranes (Donovan et al., 198 1; Hu et al., 1984; Blewitt et al., 1985; and Rolf et al., 1993). A similar effect has been observed in the case of
Pseudomonas exotoxin A where low pH alters its structure to reveal a membrane binding domain (Farahbakhsh & Wisnieski, 1989). Other toxins such as ShT and ricin appear to lack a response to endosomal pH and may have devrloped an altemate mechanism for relocating their enzymatic domain to the cytosol.
One possible mechanism is the exposure of a hydrophobic domain as a result of intracellular processing events enroute to the ER. Members of the ShT farnily of toxins contain a 9-amino acid long segment (residues 246-254) which include the well-defined furin-sensitive motif RXXR (for review. see Gordon & Leppla, 1994). Roteolytic cleavage at multiple sites within this nonapeptide segment takes place dunng the intravesicular trafficking of the toxin (Takao et al.. 1988; BIum et al., 199 1; and Garred et al.. 1995,b).
An inspection of the A-chah sequence of ShT and related roxins reveals the presence of a hydrophobic signal-sequence-like domain between residues 220 and 246 (Fig. 4.1 ). This hydrophobic region contains residues usually found in signal peptides and transmem brme protein segments (Perlman & Havorson, 1983; Briggs & Gierash. 1986). The rrgion immediately precedes the start of the nonapeptide cleavage sites (residues 246 to 254) located in the A-chah protease-sensitive loop (Cys-242 to Cys-261) and would represent the C- terminus of the Al fragment The X-ray structure of Shiga toxh shows that ihis hydrophobic segment of the A subunit is not initially exposed to solvent in the native toxin but rather hidden at the interface between the A subunit and the B subunit pentamer (Fnser et al., 1994). In the case of ricin, the isolated A subunir shows higher avidity for mzmbmes than the holotoxin, indicating the exposure of a hydrophobic domain upon dissociation from the B subunit (Ishida et al., 1983; Utsumi et al., 1984; Yarnasaki et al., 1988; Utsumi et al., 1989; Bilge et al., 1995). We therefore hypothesize that the reduction of the single disulphide bond of the A chah of ShT and its limited proteolysis may facilitate the release of the Al fragment from the holotoxin andor expose its hydrophobic C-terminus thus favouring its interaction with membranes.
5.2 Objectiva
In an effort to study the potential of the C-terminus tail of the Al domain to interact with membranes, peptides corresponding to residues 220 CO 246 of the ShT A subunit were synthesized and their interaction with srnall unilamrllar vesiclcs (SUV) was çharx~nzcd.
Spectroscopie results demonsirate that Lhis region of the Ai domain inserts into nrgacivdy charged membranes and assumes an a-helical structure with a large fraction of the bound peptide probably spanning the width of a negatively charged lipid bilayer at neutral pH.
5.3 ResuIts and Discussion
The Ai chah of Shiga and related toxins must transit from the lurnenal side of the endoplasmic reticulum to the cytoplasmic side to be able to effectively interact with ribosomal subunits and block protein synthesis. This translocation event may be initiated by a specific domain of the toxin as in the case of other bacterial toxins (Pseudonwnas exotoxin A, diphtheria toxin). Such a domain has not ken identified in bacterial toxins such as Shiga and Shiga-like toxins. We propose that the C-terminus created as a result of proteolysis and reduction of the A chain in this toxin family rnay represent a functional domain tnggering membrane translocation. Recently, a proline mutation (Pro250Ala) in a 12-residue hydrophobic segment located near the C-teminus of the A-chain of ricin has been shown to affect the ability of ricin to intoxicate Vero cells but did not affect its N-glycosidase activity
(Simpson et al.. 1995). This finding was interpreted as suggesting that this hydrophobic domain may be involved in membrane translocation.
A represlntation of the A chain of Shiga toxin in Figure 5.1 (pancls -4 and B). highlights the presence of a proteolytic loop (Cys-242 to Cys-261) flanked on its amino terminus by a 27-amino acid long signal-sequence like domain (residues 220-246). This domain represents the C-terminus of the Ai subunit created as the result of transcytosis modifications. The hydrophobicity profile of the A chain (Fig. 5.1A; derived using the rnethod of Engelman et al., 1986) further suggests that this domain (darkened area) rnay represent a transmembrane region. Sequence alignments of Shiga and related toxins (Fig.
5.K) with a segment of the A chah of ricin implicated in membrane translocation (Simpson et al., 1995) illustrates the existence of a common hydrophobic region within these molecules. Ricin and Shiga toxins have A chains with homologous functions. The crystal stnicture of Shiga toxin (Figure 1.5) has indicated the existence of only a limited nurnbrr of interactions betwwn ics A 1 domain and the B subunit pentamer. in particular. residurs 226 to 242 are buned under most of the Al chain, in close proximity to the B pentamer. but making no contact with the pentarner (Fraser et al., 1994). Thus, most of the C-terminus of the Ai chain is not initially accessible to solvent as part of the intact toxin. We hypothesize that this domain becornes exposed after cleaving the protease-sensitive loop at one or more - Figure 5.1 (a) Hydrophobicity profile of the A subunit of Shiga toxin (ShT) using the
Goldman. Engelman and Steitz (GES)scale (Enplman et al.. 1986). The A chai.of Shiga toxin is proteolyticaily cleaved hto two domains, namely a catalytic A 1 domain and a short
A2 domain associated with the binding subunit of the toxin. The shaded area in the hydrophobicity profile identifies the sequence from residues 225 to 242 as representing the most dominant hydrophobic region of the Ai chah
(b) Schematic representation of ShT A subunit organization showing the relative location of the hydrophobic region in relation to the known protease sensitive loop (Cys 242 to Cys
26 1). Established (A) and putative (cathepsins B and D; 1 ) cleavage sites within the loop are highlighted on the amino acid sequence (residues 220 to 246).
(c) Sequence alignment of the C-terminus of ShT Al with regions of other ribosome- inhibiting toxins. Abbreviations and sequences listed in this panel are associated with the following toxins: SLT-IIv. residues 212 to 25 1 of the A subunit of Shiga-like toxin II variant
; RTA, residues 225-268 of the A chain of ricin; AbA, residues 217 to 249 of the A subunit of abrin; DT. residues 341 to 380 of diphtheria toxin. Several protein toxins possess a stretch of hydrophobic amino acids (boxed sequences) flanked by positively charged amino acids at neutral pH (R, Arg) as well as histidines (H)which possess a positive charge at endosomal pH. ShTA SLïïivA RTA AbA DT 87 sites and reducing the disulphide bridge between Cys-242 and Cys-26 1. The newly created signal sequence-like C-terminus rnay then possess membrane active properties that would faditate the Ai chah insertion from the lumenal side of the Golgi or ER membranes.
Peptides were thus synthesized to investigate the ability of this region of the Al domain to interact and insert into phospholipid membranes. Two peptides rep~senùng residues 220 to 246 of the Ai domain, were assernbled by solid-phase peptide synthesis.
Their amino terminus was acetylated with acetic anhydride CO insure the absence of an unnatural positive charge. The peptide ShTA(220-246) includes a naturdly occumng cysteine residue at position 242 which provided a unique site for the subsequent incorporation of a paramagnetic probe (proxyl group) to monitor die insereiûfi and orientation of the peptide into unilamellar vesicles. The peptide analog, W232A242Sh-TA(220-246), included two mutations, namely Cys242Ala and Ile232T1-p to monitor its interaction with membranes by fluorescence spectroscopy. Specaoscopic experiments were performed at two pH values, namely 5.0 and 7.0 to mimic representative pH conditions observed in endosomal and ER comparûnents respectively.
5.3.1 ShTA(220-246) adoprs a partially helicd structure in rhe presence of negariveiy charged lipid vesicles. The circular dichroism spectmm of ShTA(220-246) dissolved in 5 mM phosphate. pH 7.0, suggests that îhis peptide lacks structure (random coil) in aqueous solutions (Fig. 5.2). The mean residue ellipticity (MRE) at 222 nm was expenmentally determined to be -5,000 I600 deg.cmz.dmol-1. The propensity of ShTA(220-246) to adopt a partially a-helical conformation in hydrophobic environments was demonstrated in the presence of SDS micelles (MRE of - 18,Oûû 2 900 deg.cm2.dmol-1 ai 222 nm). This finding was further supported by recording the specuum of the peptide in 50% (v/v) TFE, where the ellipticity maxima at 222 nrn reached a value of -22,000 + 1,100 deg.cm2.dmol-1. The l 200 21 0 220 230 240 250 Wavelength (nrn)
Figure 5.2 Circular dichroism spectra of the peptide ShTA(220-246) dissolved in either
5 mM phosphate, pH 7.0 (--), in 10% (w/v) SDS (- -) .or in 50% (vlv) TFE (-). The
peptide concentration was 15 pM (refer to expenmennl rnethods in chapter 2. page 54). 200 210 220 230 240 250 Wavelength (nm)
Figure 5.3 Circular dichroism spectra of the peptide ShTA(220-246) in the presence of
small unilarnellar vesicles composed of DMPG lipids. The spectrum of the peptide dissolved
in 5 mM phosphate buffer. pH 7.0 (---) is compared to that of Ihe peptide exposed to DMPG
vesicles at pH 7.0 (-) and pH 5.0 (- -). The peptide concenuaiion was 15 PM (rcfer io expenmental methods in chapter 2. page 54). peptide also assumed a partiaily a-helical structure in the presence of vesicles construc ted with negatively charged lipids (DMPG; Fig. 5.3). Sirnilar CD specua were recorded at pH
5.0 and pH 7.0, with a negative maxima at 222 nm of 16,750 I1,200 deg.cm2.dmol-1 at a
peptide-to-lipid ratio of 150. This ellipticity value translates to an approximate helicai content
of 4540%. Since about half of the added peptide was bound to the lipid (see below). this fmding indicates that the membrane-bound peptide is mostiy a-helical in structure at pH 7.0 and to a lesser extent at pH 5.0. In the presence of vesicles composed of neutrai Lipids such as DMPC. the peptide appears to possess significantly les a-helical character with a MRE
222 nm of - 11,000 I850 deg.cm~.dmol-1at pH 7.0 and half that value at pH 5.0 (Fig. 5.4).
A broad negative maxima centred at 218 nm was observed in the specinim of ShTA(220-
246) exposed to DMPC vesicles at pH 7.0 and suggests the existence of one or more peptide confornations that would include B-sheet containing structures. ShTA(220-246) tends to slowly aggregate in aqueous buffen (hours) giving rise to a circular dichroisrn specvum showing a minima between 213 and 218 nrn, a characteristic feature of B-sheet structure
(data not shown). Aggregation may be nucleated by the presence of DMPC vcsicles at pH
7.0 .
5-12Changes in the fluorescence spectrum of WJ2A242ShTA(220-246)in the presence of lipid vesicles suggest that this hydrophobie region of the AI dornain partitions readily into a lipid biloyer. Substituting isoleucine with tryptophan at position 232 provided us with an intrinsic probe to monitor the environment of the hydrophobic core of the peptide. The Trp emission spectra of the peptide in solution showed a maximum at 354 nm, indicating that the tryptophan side chah was exposed to the aqueous environment As DMPG vesicles were added to the peptide solution at pH 7.0, a blue shift (354 nm to 335 nm) in the emission maximum of the peptide was observed (Fig. 5.5A) indicating that the Trp side chain had Waveiength (nm)
Figure 5.4 Circular dichroism specm of the peptide ShTA(220-246) in the presencr of
small unilarnellar vesicles composed of DMPC lipids. The spectrum of the peptide dissolved
in 5 mM phosphate buffer, pH 7.0 (---) is compared to that of the peptide exposed to DMPC
vesicles at pH 7.0 (-) and pH 5.0 (- -). The peptide concentration was 15 j.M (refer to
experimental methods in chapter 2, page 54). migrated to a hydrophobic environment (Lakowicz. 1983). Changes in the fluorescence intensity at 335 nm (FEo) were also monitored as a function of lipid concentration (Fig.
5.5B). The increase in fluorescence intensity correlated with the obsemed spectral shift to 335 nm (Fig. 5.5A). Both measurements support the concept that the indole ring has relocated to a hydrophobic environment (Wharton et al., 1988). Similar spectral characteristics were observed at pH 5.0 in the presence of DMPG vesicles. Changes in the spectra of W232A242ShTA(220-246) in the presence of DMPC vesicles at pH 7.0, were comparable to those observed in the presence of DMPG vesicles. However, these changes occurred at higher lipid-to-peptide ratios. These results were not observed for the peptide mixed with DMPC vesicles at pH 5.0. Mixtures of acidic and zwitterionic lipids were also used to mimic physiological conditions expected on the inner lediet of die ER membrane.
The composition of the inner leaflet of rat liver rough endoplasmic reticulurn is dominated by zwittenonic (neutral) lipids such as phosphatidylcholine and phosphatidylethanolamine (Bolien & Higgins, 1980; Zachowski, 1993). However. acidic lipids such as phosphatidylsenne and phosphatidylinositol account for 7% up to 254 of the rough ER inner leaflet composition (Bollen & Higgins. 1980). This lipid leaflet is thus negatively charged. The association of W232A242ShTA(220-246) with such membranes was assessed using DMPC vesicles containing a level of DMPG (20530 ratio,
DMPG:DMPC) representative of the proportion of negatively charged Lipid present in the rat liver ER inner lipid leaflet. The effects on both the fluorescence intensity and shift in wavelength emission maxima of W232A242ShTA(220-246) were similar to those observed for the peptide interacting with vesicles composed of DMPG lipids only (Fig. 5.5A,B). although requiring more lipids. Whan vesicles were constnicted with a 595 ratio of DMPG to DMPC. the changes observed were intemediate to those recorded when
W232A242ShTA(220-246) was added to either pure DMPG or DMPC vesicles alone (data Lipid:peptide (mole ratio)
O 10 20 30 40 50 60 70 Lipidlpeptide (mole ratio) not shown). The influence of mixed lipid vesicles on the fluorescence spectral properties of the peptide suggests that the presence of low levels of negatively charged lipids (such as
DMPG)in membranes can strongly dic tate the association of W 232A242S hTA(220-246) with lipid vesicles.
53.3 Binding of rodiolabeled W32A242ShTA(220-246)tu lipid vesicles. A 1"C-acetyl group was initially introduced at the N-terminus of W*32A242ShTA(220-246) during synthesis.
The radiolabelrd peptide was shown by centrifugation to intrraçt directly with srndl unilamellar vesicles. The fraction of bound peptide was determined to be 0.48 f 0.09 at pH
7.0 and 0.51 + 0.07 at pH 5.0 in the presence of DMPG vesicles while values of 0.31 f
0.07 at pH 7.0 and 0.1 I0.05 at pH 5.0 respectively, were caiculated for the fraction of peptide bound to DMPC vesicles. These resulis suggest that W232A242ShTA(220-246) partitions more favourably into a membrane containing negatively charged lipids such as DMPG lipids than in neutral lipids exemplified by DMPC. In addition. the reduced fraction of peptide bound to vesicles containing a zwitterionic lipid (DMPC)at pH 5.0, suggests that the protonation of the three C-terminal histidines of W232A242ShTA(220-246) may accentuate its partitioning into the aqueous phase. Since the helical content of the peptide in the presence of DMPG vesicles at both pH 5.0 and 7.0 was estimated to be approximatdy 50% and since the CD spectrum of the peptide free in solution (Fig. 5.2-5.4) lacks helical character, one can conclude bat the population of membrane-bound peptide must adopt a mostly a-helical structure.
5.3.4 Monitoring the in teraction of Sh TA(220-246) with membranes by EPR. The naturally occurrîng cysteine residue (Cys-242; see Figure 5.1) located near the C-terminus of ShTA(220-246) provided a useful site to insen a paramagnetic proxyl group in the peptide. As monitored by EPR spectroscopy, the covalently attached proxyl probe displayed its expected nearly isotropie rnobility when the peptide was dissolved in an aqueous environment The spectnim of the proxyl-peptide conjugate highiighn the anticipated triplet of narrow spectral lines (Figure 5.6A). The value of the proxyl's motional parameter was significantly increased nble5.1; 0.30 ns) when compared to the free spin label (0.02 ns), reflecting the fact that the probe is less mobile when bound to the peptide. When the labelled peptide was mixed with vesicles of DMPC, there was little effect on the spectrum of the probe except for a small increase in z, values (Table 5.1). suggesting char there wulittlr interaction of the peptide with DMPC vesicles at pH 7.0. The circular dichroism spzctrum of ShTA(220-246) observed in the presence of DMPC vesicles at pH 7.0 (Figure 5.4). suggests the existence of peptide conformers containing P-sheet structure and indirectly supports the presence of microaggregates of the peptide on the surface of DMPC vesicles. Such an event would explain Our binding results where 3 1% of the peptide population was found associated with DMPC vesicles at a peptide-to-lipid ratio of 150.
In the presence of negatively charged DMPG vesicles, the spectrum at 26.5 'C contained two spectral cornponents (Figure 5.6B). One component (mobile component). representing the free peptide. was defined by the set of sharp lines previously observed in the spectrum of the proxyl-peptide in solution (Figure 5.6A). The other spectral component
(immobilized component) was broader with greater hyperfine splitting and retlected the more restricted motion of the spin-labelled peptide bound to DMPG vesicles. The arnount of immobilized component was greater at pH 5 than at pH 7 (Figure 5.7A and 5.7B). The mobility of the spin-labelied peptide bound to DMPG rose with increasing temperature, as Figure 5.6 Effect of DMPG vesicles and ascorbic acid on the EPR specmm of proxyl-
labeiled ShTA(220-246). The EPR spectnirn of the labelled peptide was recorded for the peptide dissolved in 5 mM phosphate buffer. pH 7.0 (specmm A). in the presence of DMPG vesicles suspended in the same buffer (specuum B). and after the exremal addition of 12 mM (final concentration) ascorbic acid to the peptide/vesicles mixture (spectnm C). For details. refer to experimental methods in chapter 2, page 55.
Table 5.1
EPR parameters derived from spectra of spin-labelled ShTA(220-246)
in different environrnents a
Free spin label 7.0 26.5 0.02
Spin-labelled peptide 7.0 in buffer
Spin-labelled peptide 7.0 + DMPC vesicles
Spin-labelleci peptide 7.0 + DMPG vesicles
a Experimental conditions and the denvation of T ,, and 7, values are described in the section detailing experimental procedures. Figure 5.7 Reduction of only the immobilized EPR spectral component by entrapped ascorbic acid. Superimposed EPR spectra of the proxyl-labelled peptide in the presence of DMPG vesicles at pH 7.0 (panel A. solid line spectrum) and of the labelled peptide interacting with DMPG vesicles loaded with ascorbic acid. at pH 7.0 (panel A. dotted line specmm). Superirnposed EPR specwa of the proxyl-labelled peptide in the prrszncr: ol' DMPG vesicles at pH 5.0 (panel B. solid line spectrum) and of the labelled peptide interacting with DMPG vesicles loaded with ascorbic acid. at pH 5.0 (panel B. dotted line spectmm). For details, refer to expenmental methods in chapter 2, page 55. indicated by a decrease in the value of the maximum hyperfine splitting parameter, Tm,
(Table 5.1). The Lipid exists in the gel phase at 6.5 'C, and in the liquid crystalline phase at
37.5 'C. There were no significant changes in line shape when the temperature of the peptide in solution was varied, indicating that the observed changes in the specûa with temperature in the presence of DMPG vesicles were not due to changes in the conformation of the free peptide, but rather due to its binding to the lipid bilayer. The change in rnobility as a function of temperature was greater at pH 7 than at pH 5. suggesting that the spin label on the peptide was more sensitive to the DMPG gel to liquid crystalline phase transition at pH 7 than at pH 5 (Table 5.1). The preceding results demonstrate that the C-terminus of the peptide associates with
DMPG membranes. This conclusion is supported by the motional restriction of the proxyl group and its sensitivity to the lipid phase transition. However, motional restriction could occur as a result of either peptide binding to the bilayer surface or peptide insertion into the bilayer. Use of ascorbic acid to reduce the N-O* group to N-OH. thus übolishing its EPR signal. enabled us to determine the directionality of membrane insertion. Addition of ûscorbic acid (10-20 mM) to the outside of DMPG vesicles almost completely reducrd the mobile spectrai component (sharp lines) as shown in figure MC. Interestingly, only a partial reduction (at most 50%) of the immobilized component was observed at pH 7.0 (Figure 5.6C). indicating that only a fraction of the bound peptide C-terminus is accessible to ascorbic acid from the outside of vesicles. At pH 5, - 70% of the immobilized component was reduced by extemally added ascorbic acid (data not shown) suggesting a greater access of ascorbic acid to the probe. Ascorbic acid completely reduced the spectmm of the labelled peptide in the presence of DMPC vesicles, indicating that the spin label is completely accessible to the aqueous phase in the presence of this neutnl Iipid.
To monitor if the C-terminus of the peptide could penetrate the bilayrr of DMPG vesicles, the spin-labelled peptide was added to ascorbic acid-loaded vesicles at pH 7.0. In contrast to the effect of ascorbic acid added to the outside of the vesicles, the ascorbic acid (12 mM) trapped inside the vesicles caused an almost complete reduction of only the immobilized component of the spectnim leaving the mobile component unaffected (Figure 5.7A). The addition of the spin-labelled peptide to vesicles containing ascorbic acid at pH
5.0, did not alter any components of the spin label specvum (Figure 5.7B). These results suggest thar at pH 7. the C-terminus of the bound peptide is accessible to intnvesicular ascorbic acid and may thus insert through the lipid bilayer. At pH 5, the proxyl group must reside mostly near the surface of the SUV outer leaflet (Figure 5.7B and preceding paragnph) where it is more accessible to ascorbate added outside the vesicles. The partial or complete protonation of the three C-terminal histidines (positions 243-245) at pH 5.0 may contribute to the lack of penetration of the peptide into the lipid bilayer, thus favounng the orientation of ShTA(220-246) dong the surface of the membrane at endosomal/lysosomal pHs (pH 4.5-6.0).
5.3.5 ShTA(220-246) causes leakuge of calcein trapped inside DMPG vesicles.
The C-terminus of the Al domain may possess membrane-active properties. The addition of ShTA(220-246) to DMPG SUVs at pH 7.0 containing self-qurnching concentrations of calcein resulted in a rapid increase in fluorescence intensity at 520 nm. indicating the release of the calcein fluorophore. Similar results were recorded at pH 5.0
(data not shown). The release of calcein was observed at peptide-to-lipid molar ratios of 1:2 and 1:4 (1:4 ratio is shown in Figure 4.8). No significant leakage was observed at lower peptide-to-lipid ratios (1: 17 or 150; Figure 5.8). Melittin. a pore-forming peptide used as a positive control, required peptide-to-lipid ratios of the order of 1:20 to achieve sirnilar effects on vesicles loaded with calcein as those observed for ShTA(220-246) (Figure 5.8). When lime (seconds)
Figure 5.8 Peptide-induced leakage of calcein from DMPG vesicles as a function of time and peptide-to-Iipid ratio. The peptide ShTA(220-246) was added to a cuvette containing 5 mM phosphate buffer. pH 7.0 and calcein-loaded DMPG vesicles. The fluorescence excitation wavelength was set at 490 nm and the increase in the emission signal due to the release of cdcein was monitored at 520 nm. The molar ratio of ShTA(220-246) to DMPG
Lipid was either 1:4 (4). 1: 17 (0).or 150 (0). As a positive control, the peptide mrlittin was used at a rnelittin to lipid molar ratio of 1:17 (A). A curve (O) reprrxnting die background leakage of calcein from DMPG vesicles observed in the absence of peptide is also shown. The fluorescence signal corresponding to 100% release of calcein from DMPG vesicles, was detemined by adding Triton-XI00 (1% (v/v) final concentration) to each peptide:Iipid mixture (refer to experimental methods in chapter 2. page 53). DMPG vesicles were added to clear solutions of ShTA(220-246). the resulting mixtures became ~rbidat high peptide-to-lipid ratios. suggesting that the vesicles were staning tu aggregate and fuse under such conditions. If die lipid vesicles were added to the peptide solution such that the peptide-to-lipid ratio was low (150). no turbidity was observed. Consistent with Our calcein leakage experiments, this result indicates that leakage of calcein seen at a high peptide-to-lipid ratio may be the result of membrane insertion and disruption (detergent-like property) rather than channel formation.
5.3.6 A mode1 summarizing the interaction of ShTA(220-246) with a negatively charged lipid bilnyer. The results in this study indicate that the peptide representing die C-terminus of the
Ai fragment of ShT inserts into DMPG vesicles. niree possible types of interaction of ShTA(220-246) with a DMPG Lipid bilayer are presented in Figure 5.9. which explain the restricted motion of the bound peptide and its sensiiivity to the lipid phwtransition. At pH
7. the N-terminus of ShT(220-246) is positively charged (Arg-220 and Arg-224) whilr its
C-terminus remains mostly electrically neutral. Interactions of the peptide N-terminus with lipid head groups present on the outer leaflet of the vesicles are thus favoured. The remaining segment of the peptide (residues 225-246) may either insert into the bilayer in an orientation parallel to the surface of the membrane (Figure 5.9; conformer A) or rnay traverse the lipid bilayer (Figure 5.9; conformer B). Results from fluorescence and circular dichroisrn spectroscopies cannot distinguish between confomers A and B since both membrane conformations would lead to the burying of the tryptophan side chah (fluorescence spectroscopy) and to the observation of a peptide adopting a mostly a-helicd structure (circular dichroism). Only the EPR results presentrd in this study allow us to differrntiatr betwren such confomers. Most of the peptide population musi traverse thc DMPG hilaycr with the C-terminus facing the intenor of vesicles since the C-terminus proxyl group (1)cm Figure 5.9 A schematic mode1 depicting possible modes of interaction of peptide
ShTA(220-246) with the lipid bilayer of DMPG vesicles. At pH 5. the dominani inwraciion is depicteci by conformer A. where the peptide adopts a partially a-helical confomaiion and partitions within the bilayer with the helix axis being parallel to the normal axis of the membrane. At pH 7, a population of the peptide inserts and traverses the lipid bilayer
(conformer £3). Populations of both conformers A and B may exist at pH 7. The open star represents the position of the proxyl probe within the peptide while the letter N indicates the location of its amino terminus. Open rectangles depict the length of the putative a-helical region within the peptide. The dotted Ihe as part of the rectangle in conformer B represenü an extension of the helical region beyond residue 242 (the site of incorporation of the proxyl label). be readily reduced by ascorbate entrapped inside vesicles and (2) is reduced to a much greater extent under such conditions than as a result of ascorbate being present on the
outside the vesicles. The putative transmembrane segmenr could include rrsidurs 225 LO 246 (starhg after Arg-224), suggesting that the resulting helix could incorponte up to 22 mino acids, a length sufficient to span most lipid bilayers (Persson and Argos, 1994; Rost et al.,
1995). This hypothesis would imply that up to 81 % (22 of 27 amino acids) of the peptide midues would have to adopt an a-helical conformation for the proxyl probe (Cys-242) to be located near the surface of the inner leafiet of a bilayer. Circular dichroism results and binding measurements of peptide to lipid vesicles predict that most of the residues in
ShTA(220-246) bound to DMPG vesicles assume an a-helical geomeüy. More precisely, approximately half of the peptide is free in solution in equilibrium with the remaining peptide population existing as one or more a-helical membrane-bound conformers (as demonsuated by binding experiments, circular dichroism and EPR results: Figure 5.3 and 5.6). In summary. the dominant ShTA(220-246) conformer hound to DMPG vssiclcs ar pH 7.0 would be conformer B (Figure 5.9). At pH 5.0, both the N- and C-termini are positively charged (Arg-220 and -224; His-243. -244 and -245). Reduction experiments with ascorbic acid indicate that the C- terminus of ShTA(220-246) is only accessible from the outer surface of the Lipid vesicles. suggesting that the peptide does not traverse the bilayer at pH 5.0. The association of ShTA(220-246) with lipid vesicles at pH 5.0, was confimed by circular dichroism, fluorescence, and bindhg experirnenu with radiolabeled peptide. Thus, the hydrophobic cote of the peptide probably inserts into the lipid bilayer in an orientation best depicted by conformer A (paralle1 to the surface of the bilayer) with its charged termini bound to negatively charged lipid head groups (Figure 5.9). Since the toxin must travel through xidic compartments (pH 5-6) prior to reaching the ER lumen (neutral pH). the insertion of the Al C-terminus into negatively charged membranes may be resuicted to a particular orientation during the transit stage. However, either mode of insertion would be favourable since it means bringing the Al fragment in contact with the membrane; a required sep for membrane tram location. Findly, onented CD measurements of magainins, a family of 23-residues frog peptides with antimicrobial properties, demonstrawl than they adopt an a-hehcal conformation with two distinct orientations in relation to a Lipid membrane ( either inserted parallei or perpendicular to the lipid bilayer) (Luddte et al.. 1994). This report supports our results that membrane-bound ShTA (220-246) can insert into a lipid bilayer in the two proposed orientations. CONCLUDING REMARKS 6.1 Dues the B subunit play a role in A chuin translocation ? The observation of a local conformational change in the B subunit pentamer at acidic pHs predicted for comparunents associated with endocytosis, suggested that the B subunit may encode a rnechanism for
promoting or actuating the procrssing. release andor translocation of the A çhain to the cytoplasm. The potential role of the B oligomer remains unclear but could involvc the tive a-helices. This site would be favoured by analogy with the E. coli heat-labile enterotoxin where most of the A chain contacts with its B subunit pentarner occur through the central helices (Sixma et al., 1991; Sima et al.. 1992). The change in tryptophan fluorescence intensity (Figure 2.1) between pH 5 and 7 would be related to a local conformationd change occurring near the NH2 terminus of these helices. The crystal structure of the B subunit indicates that the tryptophans in the pentamer are already mostly exposed to the solvent at neutral pH. Thus, at pH levels below 5, the indole side chains must either be Iocated in proxirnity of charged side chains that are revealed within this pH range or that the tryptophans and poten tially the a-helices are partially reoriented to a more solvent-expoxd environment. Since the B subunit selectively binds the glycolipid Gb3. ir suggests that thc receptor-binding domain of the B subunit maintains a rigid tertiary or quatemaiy structure and that the receptor-binding function is not influenced by acidic pH levels (Figure 3.5).
As in the case of the influenza virus hearnagglutinin and diphtheria toxin, the ShT B subunit would have to undergo a change in conformation large enough in scale to directly mediate the membrane insertion and translocation of the Al chain. Given the small size of the B chah (69 amino acids) and the several functions it is known to perfom (binding ro rhc
receptor, binding to the A subunit and oligornerization) it does not appthat it cm undergo
large scale conformational changes without affecting any of those functions. Endosomal pH
does not alter the oligomerization of the B subunit or the holotoxin. as determined by gel permeation chromatography (Table 4.1). The B subunit also maintains its receptor binding propenies at those pH levels (Figure 3.5). This aiiowed us to conclude that the B subunit may not be direcdy involved in the translocation of the Al fragment to the cytosol. Instead, it may be required to orient the A subunit for proper proteolytic processing and for the loc~zationof the toxin in the ER
6.2 Domains annlogous to ShTA(2ZO-246) are presenr on other proteins. In xarching For a domain in the holotoxin that may bt: involved in the Ai chain translocation. WC tiiund a signal sequence-like domain at the C-terminus of this chain (residurs 220-246) which may be the sought-after functional domain (chapter 5). Amino acid sequences similar to this peptide sequence can be found in a number of proteins. Of particular interest is the human asialoglycopro tein (ASGP) recep tor H 1. Protein fusion expenments have shown that this sequence (residues 38-65) was sufficient to result in membrane insertion and translocation of the C-terminal domain of rat a-tubulin (Spiess & Lodish. 1986). The presence of such a hydrophobic segment is also seen in the C-terminus of polyomavirus middle-T antigen (residues 394-421) which is involved in membrane binding (Markland et al. 1986). In diphtheria toxin, the hydrophobic sequence shown in Figure 5.L is located within the putative membrane translocation domain (Moskaug et al.. 199 1 ; Choe ei al.. 1992). Point mutation studias on this domain have rrvealed that substiiuting glutarnic acid ai position 362 with lysine allowed the toxin to translocate at a higher pH than the wild typa (Falnes et al., 1992). More recently. spin-labelled analogs of the isolated transmembrane domain of diphtheria toxin have ken show to insert into and permeabilize large unilamellar vesicles (Zhan et al., 1995).
6.3 Does ShTA(220-246) and the A 1 chuin interact with component(s) of a protein maruporter located in the ER membrane? The import of nascent protein chahs associated with ribosomes into the ER lumen is thought to involve the recognition of signal sequences on newly synthesized proteins by a signal sequence binding protein (ilseif an ER membrane protein) in associarion with other known components of the protein translocation machinery to the ER (signal recognition particle (SRP). SRP receptor, ribosome and transmembrane channel protein). One hypothetical scenario rnay be that the C-terminus of the ShT Al domain inserts into the ER membrane and allows the catalytic domain to diffuse dong the plane of the membrane until it is recognized by a signal sequence binding protein. Subunits of the transmembrane channel protein or recently assembled channel proteins may then be recruited or be already positioned in proximity to the complex of the ShT A 1 domain and the signal sequence binding protein. favouring the ultimate translocation of the Al domain across the ER membrane. As a result of interactions with the ER membrane and associatrd protein components, the population of bound Ai chain may cxist as a series of panly unfoldcd intermediates. which may thrn be reçognizrd as appropriate substraiss for cxpon (çytosd) through the uansmembrane protein channel. The involvement of various ER protein factors linked to the folding of proteins such as the protein disulfide isomerase, p88 or calnexin, HSP47, GRP94 and Bip (Gething and Sambrook, 1992; Gaut and Hendershot. 1993; Simon, S., 1993) rnay be crucial for the export process to occur.
6.4 The involvement of the ubiquitin-protemorne pathway. Altematively. the hydrophobic C- terminus of the Ai fragment would be exposed and would contain a free cysteine (Cys-242) 110 available for reversible disulfide bond formation with resident proteins of the ER cornpartment These features have been proposed to mediate protein retention and degradation in the ER (Schmitz et al., 1995; Reddy et al., 1996). One possible scenario is that upon an-ival of SLT-1 in the pre-GolgiIER comparunents. the exposed hydrophobic region in the C-terminus of the Al fragment is recognized by an ER resident protein (such as
Bip in eukaryotes or Der1 in yeast) and is then vanslocated to the cytoplasmic side of the ER to be degraded by the ubiquitin-proteasorne pathway. This pathway. which is initiated at the cytosolic side of the ER (Biederer et al., 1996). is the major pathway for degradation of several misfolded and mutated proteins including free MHC class 1 molecules (Raposo et al., 1995). mutant forms of the cystic fibrosis uansmembrane conducmce regulator (CFTR) associated with cystic fibrosis (Ward et al.. 1995) and Sec6lp cornplex, a key component of the ER protein translocation machinery (Biederer et al., 1996). nie interaction of any of those ER proteins rnentioned above with SLT-1 remains to be investigated.
6.5 A hyporhetical modrl describing membrune inserrion and rranslocarion of the SLT-I AI frogrnent. Intoxication of cells by Shiga and Shiga-like toxins occurs through dire distinct stages; The binding of the B subunit to its receptor, the intracellular trafficking and proteolytic processing (primarily of the A subunit) of the toxh foilowed by the membrane translocation of its Al domain leading to the catalytic inactivation of the ribosome. Membrane translocation of the enzymatic fragment requires that it cornes in close proximity to, and perhaps making contact with. the vesicular membrane. The results of this work and from the work of other Iaboratories (Surewicz et al.. 1989) indicate that the holotoxin possesses little membrane active properties. Processed toxin however. may expose the hydrophobic domain (see chapter 4) which we have shown to have strong membrane active properties. This feature would satisfy the requirement of bringing the enzymatic fragment in close proximity &$!p- - - receptor, Gb3 4- --- - B subunit ------A2 fragment- SLT-1 - hydrophobic C-terminus of Al ALchai11 1
polypeptide foldinghfolding chaperones endogenous to the ER
protein translocation channel
Figure 5.1 A schematic diagram depicting the proposed mechanism of membrane translocation by SLT-1. Panel A This is the initial step in the translocation. The Ai chain
interacts with che ER membrane through its hydrophobic C-terminus either direcdy (1) or via transient association with an endogenous factorkhaperone of the ER (2). Panel B The next stage would involve unfolding of the Al chah by ER chaperones followed by the
translocation step. The translocated Al chain may either remain associared with the ER
membrane and gain access to ribosomes docked ont0 the membrane (3) or be released from the membrane to inactivate free ribosomes (4). Panel C An altemate translocation pathway could involve pro tein-translocaring channels. The lareral diffusion of the membrane inserted
Al may bring this region in close proximity to the translocon (the protein translocation
channel) either after unfolding (5) or before unfolding (6) takes place. Because of the
homology berween the A 1 C-terminus and signal sequences. the translocon may recognize
this region of the Al and subsequently facilitate its translocation. A Cytoplasm to the vesicular membrane (Figure 5.1; panel A, step 1). The next step could follow one of two routes; the inserted C-terminus could allow the A 1 fragment to spontaneously escape the
ER cornpartment, or it could result in the lateral diffusion of the A 1 fragment to bring it in close proximity to specialized componenu of the ER protein translocation complex and aiiow for the assisted translocation of the catalytic fragment (Figure 6.1). For example, major histocompatibility complex (MHC) molecules targeted to the inner side of the ER membrane require the binding of small peptides for subsequent processing and targeting to the cell surface. In the absence of such peptides. MHC molecules are unstable and are targeted for proteolysis. It is becorning evident that this is accomplished by unfolding and translocation (through specialized protein chmnels) to the cytosol. where they are recycled tiy proteolysis in the proteasorne (Raposo et al., 1995). The proposed mechanism for the translocation of the SLT-1 Ai fragment is attractive in that it implies that the ER protein translocation complex could function in both directions; the normal direction being the translocation of nascent chains from the ribosome on the cytosolic side to the inside of the ER. Although the proposed mechanism is highly speculative, its appeal resides in the facts that A) a machinery provided by the host ce11 at the level of the ER membrane can recognize a broad array of sequences similar to ShTA(220-246) and B) cm transport large proteins across this membrane barrier in association with ribosomes which C) are the target site for the action of the cawlytic Al domain.
6.6 Future Directions. The mode1 in Figure 5.1 highlights the potential existence of two distinct mechanistic events, that may act in concert, in delivering the enzymatic Al fragment from the ER to the cytosol. This hypothesis nises two questions; one is whether or not the proposed hydrophobic C-terminus of the Al fragment is actually involved in membrane translocation and the second issue of whether or not any ER resident proteins interact with the toxh in the ER to affect translocation.
The fomer question can be effectively addressed by modifying the character of the this region to either reduce or abolish its contribution (if any) to the translocation process.
This region of the Al has two main features, the uninterrupted hydrophobic core and the positive charges at both temini (Fig. 5.1). Site-directed mutagenesis could be used to alter these features. Certain hydrophobic residues in the core of this region may be converted to bulky charged residues (such as arginines) to intempt the hydrophobicity of the domain. Similarly, the arginines at the N-terminus or the histidines at the C-terminus rnay be converted to neutral residues. The effect of such mutations on the toxins' ability to intoxicate cells may then be detennined. For an effective comparison, however, such mutations should not affect the folding, enzymatic activity or intracellular targeting of the toxin. The latter question concems the potential interaction of both soluble and membrane- associated ER proteins with the toxin molecule. most probably its A subunit Such factors could be identified by CO-immunoprecipitation cxperiments, given the availability of antibodies directed against ER resident factors. Other strategies include cross-linking expenrnents by probing ceIl extract from cells pretreated with toxui followed by western blot analysis using an ti-toxin or an ti-ER proteins an tibodies. REFERENCES
Abaas, S., A. Franklin, 1. Kuhn. F. Orskov. and L Orskov. 1989. Cytotoxin activity on
Vero cells among Escherichia cofi strains associated with diarrhea in cats. Am J. Vet. Res.
50:1294-6.
AUiired. V. S., R. J. Collier, S. F. Carroll, and D. B. McKay. 1986. Structure of exotoxin
A of Pseudomnas aeruginosa at 3.0-Angstrom resolution. froc. Natl. Acad. Sci 83: 1320-
1324.
Amone, A., C. J. Bier. F. A. Cotton. V. W. Day. E. E. Jr Hazen, D. C. Richardson. A.
Yonath. and J. S. Richardson. 197 1. A high resolution structure of an inhibitor cornplex of the extracellular nuclease of Staphylococcus aureus . 1. Experimental procedures and chah tracing. J. BioL Chem. 246:2302-23 16.
Ashkenazi, A. U., T. G. Cleary, E. Lopez. and L. K. Pickering. 1988. Anticytotoxin- neutralizing antibodies in immune globulin preparations: potential use in hernolytic-uremic syndrome. J. Pedian. 113: 1008- 10 14.
Bartido, S. M., S. Diment, and C. S. Reiss. 1995. Processing of a virai glycoprotein in the endoplasmic reticulum for class II presentation. Eur. J. Immun. 2522 1 1-22 19.
Biederer. T., C. Volkwein and T. Sommer. 1996. Degradation of subunits of the Sec6lp cornplex. an integral component of the ER membrane. by the ubiquitin-proteasorne pathway.
EMBO J. 15:2069-2076. 115
Bielawska, A., C. M. Linardic, and Y. A. Hannun. 1992a Modulation of ceIl growth and differentiation by ceramide. FEBS Len. 307:2 1 1-2 14.
Bielawska, A., C. M. Linardic. and Y, A. Hannun. 1992b. Ceramide-mediated biology. Detexmination of structural and stereospecific requirements through the use of Kacyl-phenyl aminoalcohol analogs. J. Bid. Chem 267: 18493-7.
Birks, J. B. 1970. Photophysics of aromatic molecules. New York, NY.. Wiley- Interscience, pp 433-447.
Bitzan, M., S. Richardson, C. Huang, B. Boyd, M. Petric, and M. A. Karmali. 1994. Evidence that verotoxins (Shiga-like toxins) from Escherichia coli bind to P blood group antigens of human erythrocytes in vitro. Infect. Immun. 62:3337-3347.
Blewitt, M. G., L. A. Chung, and E. London. 1985. Effect of pH on the conformation of diphtheria toxin and its implications for membrane penetration. Biochemist~.24545844.
Blum, J. S., M. L. Fiani, and P. D. Stahl. 1991. Proteolytic cleavage of ricin A chain in endosomal vesicles. Evidence for the action of endosornal proteases at both neutral and acidic pH. J. BioL Chem. 266:2209 1-5.
Bollen, 1. C., and J. A. Higgins. 1980. Phospholipid asymrnetry in rough- and smooth- endoplasmic-reticulurn membranes of untreated and phenobarbital-treated rat liver. J.
Biochem 189:475-80. 116
Boquet, P., M. S. Siiveman, and A. M. Pappenheimer Jr. 1977. Interaction of diphtheria
toxin with mammalian cell membranes Prog. Clin. BioL Res. IF501 -509.
Boyd, B., G. Magnusson, 2. Zhiuyan, and C. A. Lingwood. 1994. Lipid modulation of
glycolipid receptor function. Availability of Gal(a1-4)Gal disaccharide For verotoxin
binding in natural and synthetic glycolipids. Eur. J. Biochem 223:873-87 8.
Boyd, B. 1992. MSc. thesis
Boyd. B., S. Richardson, and J. Gariépy. 1991. Serological responses to the B subunit of Shiga-like toxin 1 and its peptide fragments indicate that the B subunit is a vaccine candidate
to counter the action of the toxin. Infect. Immun 59:750-757.
Brahms, S., and I. Brahms. 1980. Determination of protein secondary structure in solution
by vacuum ultraviolet circular dichroisrn. J. Mol. Biol. 138: 149-78.
Briggs, M. S.. and L. M. Gisrasch. 1986. Molecular mechanisms ol' protrin secretion: the
role of the signal sequence. Adv. Prut. Chem. 38: 109-80.
Brinkmann, U., and 1. Pastan. 1994. Immunotoxins against cancer. Biochim. Biophys.
Acta. 1198:27-45.
Brown, J. E., S. W. Rothman, and B. P. Doctor. 1980a Inhibition of protein synthesis in
intact HeLa cells by Shigelfa dysenterine 1 toxin. Infect. I~mmwi.9:98- 107. 117
Brown, J. E., M. A. Ussery, S. H. Leppla, and S. W. Rothman. 1980b. Inhibition of protein synthesis by Shiga toxin: activation of the toxin and inhibition of peptide elongation.
FEBS Letters. 117:84-8.
Brown, J. E., D. E. Griffin, S. W. Rothman, and B. P. Doctor. 1982. Purification and biological characterization of shiga toxin from Shigelh dysen teriae 1. Infect. I~Mu~.36:996
1005.
Brown, J. E.. T. G. Obrig, M. A. Ussery. and T. P. Moran. 1986. Shiga toxin from
Shigellu dysenteriae 1 inhibits protein synthesis in reiiculocyte lysatrs hy inactivation ut' arninoacyl-@NA binding. Microbial Pah 1:325-34.
Burgess. B. J., and L. M. Roberts. 1993. Proteolytic cleavage at arginine residues within the hydrophillic disulphide loop of the ficherichia coli Shiga-like toxin 1 A subunit is not essential for cytotoxicity. Mol. Microbiol. 10: 17 1-9.
Burks. C.. and D. M. Engelman. 1981. Cholessryl myristate conformation in liquid crystalline mesophases deterrnined by neutron scattering. Proc. Natl. Acad Sci U.S.A.
78:6863-7.
Bumette. W. N. 1994. AB5 ADP-ribosylating toxins: comparative anatorny and physiology.
[Review]. Srruct. 2: 15 1-8.
Calderwood, S. B., F. Auclair, A. Donohue-Rolfe, G. T. Keusch, and J. J. Mekalanos.
1987a Nucleotide sequence of the Shiga-like toxin genes of Escherichia coli. Proc. Natl. Acad Sei. USA. 84:4364-4368.
Calderwood, S. B.. and J. J. Mekalanos. 1987b. Iron regulation of Shiga-Iike toxin
expression in Escherichia coli is mediated by the fur locus. J. Bacteriol. 169:4759-64.2 1.
Calderwood, S. B., D. W. Acheson, M. B. Goldberg, S. A. Boyko, and A. Donohue-
Rolfe. 1990. A system for production and rapid purifcation of large amounts of the Shiga toxinlshiga-like toxin 1B subunit. Infect. Immun. 58:2977-32.
Cantey, J. R. 1985. Infectious diarrhea. Pathogenesis and risk factors. Am J. Med 78:65-
75.
Cantey, J. R. 1985. Shiga toxin--an expanding role in the pathogrnesis of inkctious diseases. J. Infect. Dis. 151:766-7 1.
Cantor, C. R., and P. R. Schirnmel. 1980. In biophysical chernistry, part 1, pp49. W. H. Freeman and Company. San Francisco, CA.
Chen, Y.-H., J. T. Yang, and K. H. Chau. 1974. Determination of the helix and P forni of proteins in aqueous solution by circular dichroism. Biochernisq 13:3350-3359.
Cohen, A., G. E. Hannigan. B. R. G. Williams. and C. A. Lingwood. 1987. Roles of globotriosyl- and galabiosylceramidr in verotoxin binding and high affinity interferon receptor. J. Biol. Chem. 262: 1 7088- 1709 1. 119
Cohen, A., V. Madrio-Marina, Z Esuov. M. H. Freedman. C. A. Lingwood, and KM. Dosch. 1990. Expression of glycolipid receptors to Shiga-like toxin on human B lymphocytes: a mechanism for the failure of long-lived antibody response to dysenterie disease. Int. Irnmwz. 2: 1-8.
Collier, R. J. 1988. Structure-activity relationships in diphtheria toxin and Pseudumonas aerughosa exotoxin A. [Review]. Cancer Treat. Res. 37:25-35.
Corbett, R. J. T.. and R. S. Roche. 1984. Use of high-speed size exclusion chromatography for the study of protein folding and s~bility.Biochernistq 23: 1888- 1894.
Cravioto, A, R., R, J. Gross, S. M. Scotland, and B. Rowe. 1979. An adhesive factor found in strains of Escherzkhia coli belonging to the traditional infantile enteropathogenic serotypes. Curr. Microbiol. 3:95-99.
Crook. S. J.. J. M. Boggs. A. 1. Vistnes, and K. M. Koshy. 1986. Faciors ii1Tecting surface expression of glycolipids: influence of lipid rnvironmrn~and çrnimidc composition on antibody recognition of cerebroside sulfate in liposomes. Biochemistq 257488-94.
Davies. P. I., M. M. Comwell. J. D. Johnson. A. Reggianni, M. Myers. and M. P. Murtaugh. 1984. Studies on the effects of dansylcadavenne and related compounds on receptor-mediated endocytosis in cultured cells. [Review] Diabetes C'are. 7 Suppl 1:35-4 1.
Davies. P. J., D. R. Davies. A. Levitzki, F. R. Maxfield. P. Milhaud, M. C. Willingharn, and 1. H. Pastan. 1980. Transglutaminase is essential in receptor-mediated endocytosis of alpha 2-rnacroglobuiin and polypeptide hormones. Namre. 283: 162- 167.
Dawidowicz, E. A. 1993. Lipid sorting and trafficking in cells. Curr. Oph. Srntct. Biol.
3:495-498.
De Grandis, S., J. Ginsberg, M. Toone, S. Climie, J. Friesen, and J. Brunton. 1987.
Nucleotide sequence and promoter mapping of the Escherichia coli Shiga-like toxin operon of bacteriophage H- 1%. J. BacterioZ. 169:43 13-9.
Deresiewicz, R. L., S. B. Calderwood, J. D. Robertus, and R. J. Collier. t 992. Mutarions affecting the activity of the Shiga-like toxin 1 A-chain. Biochemistry. 31:3272-3280.
Donovan, J. J.. M. 1. Simon, R. K. Draper. and M. Montal. 1981. Diphtheria toxin forms transmembrane channels in planar iipid bilayers. Proc. Natl. Acad Sci. U.S.A. 78: 172-6.
Downes, F. P., T. J. Barret, J. H. Green, C. H. Aloisio, J. S. Spika, N. A. Srrockbine, and 1. K. Wachsmuth. 1988. Affmity purification and characterization of shiga-like toxin II and production of toxin-specific monoclonal antibodies. Infect. Irnmim. 56: 1926- 1933.
Dressler, K. A., S. Mathias, and R. N. Kolesnick. 1992. Tumor necrosis factor-alpha activates the sphingomyelin signal transduction pathway in a cell-free systrm. Science.
255: 17 15-17 18.
Edelman, R., and N. F. Pierce. 1984. From the National Institute of Allergy and Infectious
Diseases. Surnmary of the 19th United States-Japan Joint Cholera Conference. J. Infect. Dis. Eftink, M. R., and C. A. Ghiron. 1976. Exposure of tryptophanyl residues in proteins.
Quantitative determination by fluorescence quenching studies. Biochernisny. 15672-680.
Eletr, S.. and A. D. Keith. 1972. Spin-label studies of dynamics of lipid allcyl chains in
biological membranes: role of unsatunted sites. Proc. Nari. Acod Sci. U.S.A. 69: 1353-7.
Endo, Y., K. Tsurugi, T. Yutsudo, Y. Takeda. T. Ogasawara, and K. Igarashi. 1988. Site of action of a Vero toxin (VT2) from Escherichia coli 0157:H7 and of Shiga toxin on eukaryotic ribosomes. RNA N-glycosidase ac tivi ty of the toxins. Eur. 1. Biochem. 171:45-
50.
Engelman, D. M., T. A. Steitz. and A. Goldman. 1986. Identifying nonpolar uansbilayer helices in amino acid sequences of membrane proteins. Ann. Rev. Biophys. Biophys. Chem
15:32 1-53.
Falnes, P. O., 1. H. Madshus, K. Sandvig, and S. Olsnes. 1992. Replacement of negative by positive charges in the presumed membrane-inserted part of diphthena toxin B fragment. Effect on membrane translocation and on formation of cation channels. J. Biol. Chem
267: 12284-90.
Farahbakhsh, 2. T., and B. J. Wisnieski. 1989. The acid-uiggered entry pathway of
Pseudomonas exoroxin A. Biochernisrry. 28:580-5. 122
Farahbakhsh, 2. T., R. L. Baldwin, and B. J. Wisnieski. 1987. Effect of pH on the
conformation of Pseudononas Exotoxin A. J. Biol. Chem. 262:2256-2261.
Fontaine, A., J. Arondel, and P. J. Sansonetti. 1988. Role of shiga toxin in the pathogenesis of bacillary dysentery, studies by using a tox- mutant of shigelh dysenteriae 1. Infect Immun. 56:3099.
Fraser, M. E., M. M. Chemaia, Y. V. Koziov, and M. N. James. 1994. Crystal structure of the holotoxin from Shigella dysenreriae ar 2.5 A resolution. Nature Struct. Biol. 159-64.
Garred, O., E. Dubinina, P. K. Holm, S. Olsnes, B. van Deurs, J. V. Kozlov, and K- Sandvig. 1995a. Role of processing and intracellular transport for optimal toxicity of Shiga toxin and toxin mutants. Eip. Cell Res. 218:39-49.4 1.
Garred. O., B. van Deurs, and K. Sandvig. 1995b. Furin-induced cleavage and activation of
Shiga toxin. J. Biol. Chem 270: 108 17-2 1.
Gaut, J. R., and L. M. Hendershot 1993. The modification and assembly of proteins in the endoplasrnic reticulurn. Cum Opin. Cell Biol. 5589-95.
Geisow. M. J. 1984. Fluorescein conjugatcs as indicators of subcellular pH. A critical evaluation. Exp. Cell Res. 150:29-35.
Geisow, M. J., and W. H. Evans. 1984. pH in the endosorne. Measurernents during pinocytosis and receptor-mediated endocytosis. Erp. Cell Res. 150:36-46. 123
Gemski, P., and D. E. Griffui. 1980. Isolation and characterization of miniceIl-producing
mutants of Shigella spp. Infect. Immun. 30:297-302.
Gentry. M. K. and J. M. Dalrymple. 1980. Quantitative microtiter cytotoxicity assay for
Shigella toxin. 3. Clin. MicrobioL 12:361-366.
Gething, M. J.. and J. Sambrook. 1992. Protein folding in the cell. Nature. 355:33-45.
Ghetie, M.-A., and E. S. Vitetta 1994a. Recent developments in irnrnunotoxin therapy.
Curr. Opin. Immun. 6:707-7 14.
Ghetie, V., and E. Vitetta. 1994b. Immunotoxins in the therapy of cancers: from bench to chic. Phamc. Ther. 63:209-234.
Gordon, V. M., and S. H. Leppla. 1994. Proteolytic activation of bacterial toxins: role of bacterial and host cell proteases. Infect. Immun. 62:333-340.
Gottstein. C.. U. Winkler, H. Bohlen, U. Diehl, and A. Engert. 1994. Immunotoxins: is there a clinical value? Ann. Oncol. 3: S97-S 103.
Gray, G. L., D. H. Smith, J. S. Baldridge, R. N. Harkins, M. L. Vasil, E. Y. Chen, and
H. L. Heyneker. 1984. Cloning, nucleotide sequence, and expression in ficherichia coli of the exotoxin A structural gene of Pseudomonas aeruginosa. Proc. Nad. Acad Sci. 81:2645- 2649. 124
Greenfield, N., and G. D. Fasman. 1969. Computed circuiar dichroism spectra for the evaluation of protein confornation Biochemistry. 8:4 108- 16.
Griffm. D. E., and P. Gemski. 1983. Release of Shiga toxin from Shigella dysenteriae 1 by polyrnyxin B. Infect. Immun, 40:425-8.
Griffi, D. E., M. K. Gentry, and I. E. Brown. 1983. Isolation and charactenzation of monoclonal antibodies to Shiga toxin. Infect. Immun. 41:430-3.
Gyles, C. L., S. A. De Grandis. C. MacKenzie. and J. L. Brunton. 1988. Cloning and nucleotide sequence analysis of the genes determining verocytotoxin production in a porcine edema disease isolate of Escherichia coli. Microbial Pathogen. 5419-26.
Gyles, C. L. 1992. Escherichia coli cytotoxins and enterotoxins. Can J. Microbiol. 38:734-
46.
Haddad, J. E., A, Y. Al-Jaufy and M. P. Jackson. 1993. Minimum domain of the Shiga toxin A subuni t required for enzymatic activity. J. Bacteriol. 1754970-4978.
Hale, T. L., and S. B. Formal. 1980. Cytotoxicity of Shigella dysenteriae 1 for cultured marnmaiian cells. Am J. Clin. Nutri. 33:2485-90.
Hale, T. L., and S. B. Formal. 1981. Protein synthesis in HeLa or Henle 407 celis infected with Shigella dysenteriae 1, Shigella flemerî 2a, or Salmonella typhimurium W 11 8. Infect.
Immun. 32:137-44. 125
Head, S. C., M. A Kmali, and C. A. Lingwood. 1991. Preparation of VTI and VT2 hybrid toxins from their purified dissociated subunits. Evidence for B subunit modulation of a subunit function. 3. Biol. Chem 26636 17-2 1.
Hofmann, S. L. 1993. Southwestern intemal medicine conference: Shiga-like toxins in hemolytic-uremic syndrome and throm bo tic throm bocytopenic purpura. Am J. Med Sci . 306:398-406.
Hol, W. G., P. T. van Duijnen. and H. J. Berendsen. 1978. The alpha-hrlix dipole and the properties of proteins. Nature. 273:443-6.6 1.
Hol, W. G., L. M. Halie, and C. Sander. 1981. Dipoles of the alpha-helix and beta-sheet: their role in protein folding. Nature 294532-6.
Hovde, C. J., S. B. Calderwood, J. J. Mekalanos, and R. J. Collier. 1988. Evidence that glutamic acid 167 is an active site residue of Shiga-iike toxin 1. Proc. Natl. Acad Sci. USA.
85:2568-2572.
Hu. V. W.. and R. K. Holmes. 1984. Evidence for direct insertion of fragments A and B of diphtheria toxin into mode1 membranes. J. Biol. Chem 259: 12226-33.
Huang, A., J. Friesen, and J. L. Brunton. 1987. Charactenation of a bactenophage that canies the gene for production of shiga-like toxin 1 in Escherichia coli J. Bacteriol.
169:4308-43 12. 126
Huang, A, S. De Grandis, J. Friesen, M. A. Karmali, M. Peuic, R. Congi, and J. L. Brunton. 1986. Cloning and expression of the genes specitying Shiga-like toxin production
in Escherichia coli W 19. J. Bacteriol, 166375-379.
Ishida, B.. D. B. Cawley. K. Reue, and B. J. Wisnieski. 1983. Lipid-protein interactions during ricin toxin insertion into membranes. Evidence for A and B chah petration. J. Biol-
Chem 2585933-7.
Jacewicz, M., and G. T. Keusch. 1983. Pathogenesis of Shigella diarrhea. VU. Evidence for a translocation step in the cytotoxic action of Shiga toxin. J. Infect. Dis. 148:844-54.
Jacewicz, M., H. Clausen, E, Nudelman, A, Donohue-Rolfe, and G. T. Keusch. 1986.
Pathogenesis of shigelia diarrhea. XI. Isolation of a shigrlia toxin-binding glycolipid from rabbit jejunum and HeLa cells and its identification as globotnaosylcerarnide. J. Exp. Med
163: 139 1-404.
Jacewicz, M. S., M. Mobassaleh, S. K. Gross, K. A. Balasvbnmanian, P. F. Daniel, S. Raghavan, R. H. McCluer, and G. T. Keusch. 1994. Pathogenesis of Shigella diarrhea: XW. A mammalian ce11 membrane glycolipid. Gbs, is required but not sufficient to confer sensitivity to Shiga toxin. J. Infect. Dis. l69:53 8-546.
Jackson, M. P., J. W. Newland, R. K. Holmes, and A. D. O'Brien. 1987. Nucleotide sequence analysis of the sttuctural genes for Shiga-like toxin 1 encoded by bactenophage
933J from Escherichia coli. Microbial Path. 2: 147-53. 127
Jackson. M. P. 1990a Structure-function analyses of Shiga toxin and the Shiga-like toxins.
Micro biai Path. 8:235-42.
Jackson, M. P., R. L. Deresiewicz, and S. B. Calderwood. 1990b. Mutational analysis of
the Shiga toxin and Shiga-like toxin II enzymatic subunits. I. Bacteriol. 172:3346-50.
Jackson, M. P., E. A. Wadoikowski, D. L. Weinstein, R. K. Holrnes, and A. D. O'Brien. 1990~.Functional analysis of the Shiga toxin and Shiga-like toxin type II variant binding
subunits by using site-directed mutagenesis. J. Bacteriol. 172:653-77 1.
Jiang, I. X., and E. London. 1990. Involvernent of denaturation-like change in Pseudomonas Exotoxin A hydrophobicity and membrane penetration detemined by
characterization of pH and thermal transitions. J. BioL Chem. 265:8636-8641.
Karlsson, K.-A. 1989. Animal glycosphingolipids as membrane attachment sites for
bacteria. Annrr. Rev. Biochem. 58:309-350.
Karmali, M. A. 1989. Infection by verocytotoxin-producing Escherichia coli. Clin.
Microbiol Rev. 2: 15-38.
Karmali, M. A., M. Petric. C. Lim, R. Cheung, and G. S. Arbus. 1985. Sensitive method for detecting low numben of verotoxin-producing Eschenchia coli in mixed cultures by use of colony sweeps and polymyxin extraction of verotoxin. J. Clin. Microbiol. 22:614-9.
Karmali, M. A. 1989. Infection by verocytotoxin-producing Escherichia coli Clin. Micro. Rev. 2: 15-38.
Karmali, M. A., M. Petric, C. Lim, P. C. Fleming, and B. T. Sieele. 1983. Ercherichiu coli cytotoxin. haemolytic-uraemic syndrome, and haemorrhagic colitis Iletter]. hcet.
2: 1299- 1300.
Kavi, J., 1. Chant, M. Maris, and P. E. Rose. 1987. Cytopathic effect of verotoxin on endothelid celis [letter]. iuncet 2: 1035.
Khine, A. A.. and C. A. Lingwood. 1994. Capping and receptor-mediated endocytosis of cell-bound verotoxin (Shiga-like toxin) 1: chernical identification of an arnino acid in the B subunit necessary for efficient receptor giycolipid binding and cellular intemalization. J. CelL
Physiol. 161:3 19-332.
Kiarash, A.. B. Boyd, and C. A. Lingwood. 1994. Glycosphingolipid receptor function is modified by fatty acid content. Verotoxin 1 and verotoxin 2c preferentially recognize different globotriaosyl ceramide fatty acid homologues. J. Biol. Chem 269: 1 1138- 1 1 146.
Kim, J. H., C. A. Lingwood, D. B. Williams. W. Furuya. M. F. Manolson. and S.
Grinstein. 1996. Dynamic rneasurement of the pH of the Golgi cornplex in Living cells using retrograde transport of the verotoxin receptor. J. Cell BioL 134:1 387- 1399.
Konowdchuk, J., J. 1. Speirs, and S. Stavnc. 1977. Vero response to a cytotoxin of
Escherichin coli Infect. Immun. 18:775-779. 129
Laemmli, U. K. 1970. Cleavage of structurai proteins during assembly of the head of bacteriophage T4- Nature. 227:680-685.
Lakowicz, J. R. 1980. Fluorescence spectroscopie investigations of the dynamic properties of proteins, membranes and nucleic acids. J. Biochem Biophys. Meth. 2:9 1 - 119.
Lakowicz, J. R. 1983. Principles of fluorescence spectroscopy. Plenum press, New York.
Lindberg, A. A., J. E. Brown, N. Stromberg, M. Westling-Ryd, J. E. Schultz and K. A.
Karlsson. 1987. Identification of the carbohydrate receptor for Shiga toxin produced by
Shigella dysenteriae type 1. J. Biol. Chem. 262: 1 779-85.
Lingwood, C. A. 1996. Role of vcrotoxin receptors in pathogenesis. [Review] Trenak in
Microbiol. 4:147-53.
Lingwood, C. A. 1994. Verotoxin-binding in human rend sections. Nephron 66:2 1-8.
Lingwood. C. A. 1993. Verotoxins and their glycolipid receptors. meview] Adv. Lipid
Res. 25: 189-2 11.
Lingwood, C. A. and S. K. Yiu. 1992. Glycolipid modification of alpha 2 interferon binding. Sequence similarity between the alpha 2 interferon receptor and verotoxin (Shiga- like toxin) B-subunit. J. Biochem. 283 :25-6.
Lingwood, C. A., H. Law, S. Richardson, M. Petric, J. L. Brunton, S. De Grandis, and 130
M. Karmali. 1987. Glycolipid binding of purifïed and recombinant Escherichia coli produced vero toxin in vitro. J. Bioi. Chem 262: 88 34-9.
London, E. 1992. Diphtheria toxin: membrane interaction and membrane translocation.
Biochim Biophys. Acta. 1113:25-51.
Ludtke, S. J., K. He, Y. Wu, and H. W. Huang. 1994. Cooperative membrane insertion of magainin comlated with its cytolytic activi ty. Biochim Biophys. Acta. 1190: 18 1-4.
Macek, P., M- Zecchini, C. Pederzolli, M. Dalla Serra, and G. Menestrina. 1995. Inuinsic tryptophan fluorescence of equinatoxin II, a pore-forming polypeptide from the sea anemone
Actinia equinn L, monitors its interaction with lipid membranes. Eur. J. Biochem 234:329- 335.
MacLeod. D. L., C. L. Gyles, and B. P. Wilcock. 199la. Reproduction of edema dixase of swine with purified Shiga-Iike toxin-lT variant. Ver. Path. 28:66-73.
MacLeod, D. L., C. L. Gyles, A. Valdivieso-Garcia and R. C. Clarke. 1991b. Physicochemical and biological properties of purified Escherichia coli Shiga-like toxin II variant.leproduction of edema disease of swine with purified Shiga-like toxin-II variant
Infect. Immun. 28:66-73.
Maloney, M. D., and C. A. Lingwood. 1994. CD19 has a potential CD77 (globotriaosyl ceramide)-binding site with sequence similariiy to verotoxin B-subunits: implications of molecular mirnicry for B ce11 adhesion and entemhemorrhagic Escherichia coli pathogenesis. J. Exp. Med 180:191-201.
Mangeney, M., C. A. Lingwood, S. Taga. B. Caillou, T. Tursz, and J. Wiels. 1993.
Apoptosis induced in Burkitt's lymphoma cells via Gb3/CD77, a glycolipid antigen. Cancer
Res. 5353 14-53 19.
Mangeney, M., Y. Richard, D. Coulaud, T. Tursz, and J. Wiels. 1993. CD77: an antigen of
germinal center B celis entering apoptosis. Eur. J. Immun. 21: 1 13 1- 1 140.
Markland, W., S. H. Cheng. B. A. Oostn. and A. E. Smith. 1936. In vitro mutagenesis of the putative membrane-binding domain of polyomavims middle-T antigen. J. Virol 59:82-9.
Marques, L. R. M., J. S. M. Peiris, S. J. Cryz, and A. D. O'Brien. 1987. Escherichia coli strains isolated from pigs with edema disease produce a variant of Shiga-like toxin II. FEMS
Micro biol. Lm. 44:33-3 8.
Maxfield, F. R. 1982. Weak bases and ionophores rapidly and reversibly raise the pH of endocytic vesicles in cuitured mouse fibroblasts. 3. Cell Biol. 95676-8 1.
Mayer, L. D., M. J. Hope, P. R. Cullis, and A. S. Janoff. 1985. Solute distributions and trapping efficiencies observed in freeze-thawed multilarnellar vesicles. Biochim Biophys.
Atm. 817: 193-6.
McCarthy, M. P., J. P. Eamest, E. F. Young, S. Choe. and R. M. Stroud. 1986. The molecular neuro biology of the acetylcholine receptor. Ann. Rev. Neuroscience. 9:353-4 13. 132
Menesuina. G.. C. Pederzolli. S. Forti. and F. Gambale. 1991. Lipid interaction of
Pseudomonas aemginosa exotoxin A. Biophy. J. 60: 1388-1400
Merria, E. A.. and W. G. J. Hol. 1995. AB5 toxins. Curr. Opin. Struct. Biol. 5165-171.
Middlebrook, J. L., and R. B. Dorland. 1984. Bacterial toxins: cellular mechanisms of action. Microbiol. Rev. 48: 199-22 1.
Mobassaleh, M., O. Koul, K. Mishra, R. H. McCluer, and G. T. Keusch. 1994. Developmentally regulated Gb3 galactosyluansferase and alpha-gaiactosidase determine
Shiga toxin receptors in intestine. Am J. Physiol. 267:G6 18-24.
Montich, G., S. Scarlata. S. McLaughlin, R. Lehmann. and J Sel 1995. Thermodynamic characterization of the association of small basic peptides with membranes containhg acidic lipids. Biochirn Biophys Acta 1146: 17-24.
Moskaug, J. O., H. Stenmark, and S. Olsnes. 1991. Insertion of diphtheria toxin B- fragment into the plasma membrane at low pH. Characterization and topology of inserted regions. J. Biol. Chem 2662652-9.
Mosmann, T. 1983. Rapid colorimelric assay for cellular growth and survivd: application to prolifention and c ytotoxicity assays. J. Immun. Meth. 6555-63.
Moyer. M. P.. P. S. Dixon, S. W. Rpthman, and J. E. Brown. 1987. Cyrotoxicity of shiga toxin for human colonic and iled epithelial cells. Infect. Zmmun 55: 1533. 133
Newland, J. W., N. A. Stroçkbine, S. F. Miller, A. D. O'Brien, and R. K-Holmes. 1985.
Cloning of Shiga-like toxin structural genes from a toxin converting phage of Escherichia
coli. Science. 230:179-8 1.
Nyholm, P.-G.,G. Magnusson. 2. Zheng, R. Norel, B. Binnington-Boyd and C. A.
Lingwood. 1996. Two distinct binding sites for globotriaosyl ceramide on vero toxins: identification by molecular modelling and confirmation using deoxy analogues and a new glycolipid receptor for all verotoxins. Chem & Biol. 3:263-275.
O'Brien, A. D., M. E. Chen, R. K. Holmes, J. Kaper, and M. M. Levine. 1984.
Environmental and human isolates of Vibrio cholerae and Vibrio parahaemolyticus produce a
ShigelZa dysenteriae 1 (Shiga)-Wte cytotoxin. Laneet. lJ7-8.9 1.
O'Brien, A. D., G. D. LaVeck, D. E. Griffin, and M. R. Thornpson. 1980. Charactenzation of Shigella dysenteriae 1 (Shiga) toxin purified by anti-Shiga toxin affinity chromatograp h y.
Infecr. Immun. 30: 170-9.
O'Brien, A. D., G. D. LaVeck, M. R. Thompson, and S. B. Formal. 1982. Production of
Shigella dysenteriae type 1-!ike cytotoxin by Escherichia coli. J. Infect. Dis. 146:763-9.
O'Brien, A. D., and G. D. LaVeck. 1983a. Purification and characterization of a Shigella dysentenae 1-lïke toxin produced by Escherichia coli.. Infect. Immun. 40:675-83.
O'Brien, A. D., T. A. Lively, T. W. Chang, and S. L. Gorbach. 1983b. Purification of
Shigella dysenteriae L (Shi@-like toxin from Escherichia coli 0 157:H7 strain associated wi th haemorrhagic colitis Iletter]. hcet. 2:573.
O'Brien, A. D., J. W. Newland, S. F. Miller, R. K. Holmes, H. W. Smith, and S. B.
Fornial. 1984. Shiga-like toxin-converting phages from Escherichia coli strains that cause
hernorrhagic colitis or infantile dianha Science. 226:694-6.
O'Brien, A. D., L. R. Marques, C. F. Kerry, J. W. Newland, and R. K. Holmes. 1989.
Shiga-iike toxin converting phage of rnarohemorrhagic Escherichiu coli sirah 933.
Microbial Path. 6:3 8 1-90.
O'Brien, A. D., A. R, Melton, C. K. Schmitt, M. L. McKee, M. L. Batts, and D. E.
Gnffin. 1993. Profile of Escherichia coli 0 157:H7 pathogen responsible for hamburger- borne outbreak of hemorrhagic colitis and hemolytic uremic syndrome in Washington. L
Clin Microbioi. 3l:2799-80 1.
Obrig, T. G.,P. J. Del Vecchio, J. E. Brown. T. P. Moran. B. M. Rowland, T. K. Judge, and S. W. Rothman. 1988. Direct cytotoxic action of Shiga toxin on human vascular endothelid ceUs. Infecr. Immzcn. 562373-8.
Obrig. T. G., C. B. Louise, C. A. Lingwood, B. Boyd, L. Barley-Maloney, and T. O.
Daniel. 1993. Endothelid heterogeneity in Shiga toxin receptors and responses. J. Biol.
Chem. 268: 15484-8.
Obng, T. G., T. P. Moran, and R. J. Colinas. 1985. Ribonuclease activity associated with the 60s ribosome-inactivating proteins ricin A, phytolaccin and Shiga toxin. Biochem. Biophys. Res. Corn 130:879-84.
Obrig. T. G., T. P. Moran, and J. E. Brown. 1987. The mode of action of Shiga toxin on peptide elongation of eukaryotic protein synthesis. J. Biochem 244287-94.
Ogata. M.. V. K. Chaudhary. 1. Pastan. and D. J. FitzGerald. 1990. Processing of
Pseudomnas exotoxin by a cellular protease results in the generation of a 37,000-Da toxin fngment that is translocated to the cytosol. J. Biol. Chem. 26520678-85.
Ogata, M., V. K. Chaudhary, M. C. Willingham, S. Adhya. D. FitzGerald, and 1. Pastan.
1989. Domain II mutants of Pseudomonas exotoxin defïcient in translocation. J. Biol,
Chem. 264: 15953-9.
Ohkuma, S.. Moriyarna. Y. And T. Takano. 1982. Identification and characterization of a proton pump on lysosomes by fluorescein-isothiocyanate-dextranfluorescence. Proc. Natl.
Acad Sci USA 79:2758-2762.
Olsnes. S., J. O. Moskaug, H. Stenrnark. and K. Sandvig. 1990. Translocation of diphtheria toxin to the cytosol and formation of cation selective channels. [Review]. J.
Physiol. 84: 19 1 -6.
Olsnes. S.. R. Reisbig. and K. Eiklid. 198 1. Subunit structure of Shigella cytotoxin. J. Biol.
Chern. 256:8732-8.
Oostenvijk, E.. A. Kalisiak. J. C. Wakka. D. A. Scheinberg, and L. J. Old. 1991. Monoclonal antibodies against Gala 1-4Galp 1-4GIc (Pk, CD77) produced with a synthetic
glycoconjugate as immunogen: reactivity with carbohydrates. with fresh frozen human
tissues and hematopoietic tumors. Inr. Cancer. 48:848-854.
Pascher, 1.. S. Sundell, K. Harlos, and H. Eibl. 1987. Conformation and packing properties of membrane lipids: the crystal structure of sodium dimynstoylphos-
phatidylglycerol. Biochim Biophys. Acta. 896:77-88.
Pellizzari, A., H. Pang and C. A. Lingwood. 1992. Binding of verocytotoxin 1 to its
receptor is influenced by differences in receptor fatty acid content Biochemis~31: 1363- 1370.
Perlman, D., and H. O. Halvorson. 1983. A putative signal peptidase recognition site and
sequence in eukaryotic and prokaryotic signal peptides. J. Mol. Biol. 167:39 1-409.
Persson. B., and P. Argos. 1994. Prediction of transmernbrane segments in proteins
utilking multiple sequence alignments. Mol. Biol. 237: 182-92.
Petric, M.. M. A. Kmdi, S. Richardson, and R. Cheung. 1987. Purification and biological properties of Escherichia coli verocytotoxin. FEMS Microbiol. Lett 41: 63-68.
Pudymaitis, A., and C. A. Lingwood. 1992. Susceptibility to verotoxin as a function of the cell cycle. J. Cell. Physiol. 150:632-639.
Rarnotar. K., B. Boyd, G. Tyrrell, J. GariBpy, C. Lingwood, and J. Brunton. 1990. Characterization of Shiga-like toxin 1 B-subunit purified from overproducing clones of the
SLT- 1 B cistron. J. Biochem 272:805-8 1 1,
Raposo, G., H. M. Van Santen, R. Leijendekker, H. J. Geuze and H. L. Ploegh. 1995. Misfolded major histocompatibility complex class 1 molecules accumulate in an expanded
ER-Golgi intermediate cornpartment J. Cell Biol. 131: 1403- 14 19.
Read, R, J. And P. E. Stein. 1993. Toxins, Curr, Biol. 3:853-860.
Reddy, P., A. Sparvoli, C. Fagioli, G. Fassina and R. Sitia. 1996. Formation of reversible disulfide bonds with the protein matrix of the endoplasmic reticulum correlates with the retention of unassembled Ig light chains. EMBO J. 152077-2085,
Richardson, S. E., M. A. Karmali, L. E. Becker, and C. R. Smith. 1988. The histopathology of the hemolytic uremic syndrome associated with Verocytotoxin-producing
Escherichia coli infections. Hum Pathol. 19: 1 102- 1108.
Riley. L. W.. R. S. Remis, S. D. Helgerson, H. B. McGee, .J. G. Wells. B. R. Davis. R. J. Hebert, E. S. Olcott, L. M. Johnson, N. T. Hargrett, P. A. Blake, and M. L. Cohen. 1983. Hemorrhagic colitis associated with a rare ficherichia coli serotype. N. EngL J. Med
308:68 1-685.
Robertus, J. 199 1. The structure and action of ricin, a cytotoxic N-glycosidase. Sem Cell
Biol. 2:23-30. 138
Roehm, N. W., G. H. Rodgers, S. M. Hatfield. and A. L. Glasebrook. 1991. An irnproved
colorimevic assay for ce11 proliferation and viability utilizing the tetrazolium salt XIT. 1.
Immun. Meth. 142:257-265.
Rolf, J. M., and L. Eidels. 1993. Structure-function analyses of diphthena toxin by use of
monoclonal an tibodies. Infect. Imm.. 61:994- 1003.
Rost, B.. R. Casadio. P. Fariselli, and C. Sander. 1995. Transmembrme helices predicted at 95% accuracy. Prut. Sei. 4521 -33.
Rouser, G.. S. Fkeischer. and A. Yamamoto. 1970. Two dimensional thin Lqer chrornatographic separation of polar lipids and determination of phospholipids by phosphorus analysis of spots. Lipidr. 5494-6.
Saleh. M. T.. and J. Garikpy. 1993. Local confornational changes in the B subunit of
Shiga-like toxin 1 at endosomal pH. Biochemistry. 32:9 1 8-922.
Saleh, M. T., J. Ferguson. I. M. Boggs, and J. Gariépy. 1996. Insertion and orientation of a synthetic peptide representing the C-terminus of the Al domain of Shiga toxin into phospholipid mern branes. Biochemistry. 3593259334.
Sandvig, K., E. Dubinina. O. Gamd. K. Prydz, J. V. Kozlov. S. H. Hansen. and B. Van
Deurs. 1993. Entry of Shiga toxin into cells. Int. J. Med. Microbiol. Virol. Parasitol. Infect.
Dis. 278: 296-305. 139
Sandvig, K., O. Garred, K. Prydz, J. V. Kozlov. S. H. Hansen, and B. van Deurs. 1992.
Retrograde transport of endocytosed Shiga toxin to the endoplasmic reticulum. Nature.
3585 10-512.
Sandvig, K. And B. Van Deurs. 1992. Toxin-induced ce11 lysis: Protection by 3-
methyladenine and cycloheximide. Erp. Cell Res. 200: 253-262.
Sandvig, K., K. Prydz, M. Ryd, and B. van Deurs. 1991. Endocytosis and intracellular
transport of the glycolipid-binding ligand Shiga toxin in polarized MDCK cells. J. Cell Biol.
1 l3:553-62.
Sandvig, K.. S. Olsnes, J. E. Brown. O. W. Petersen, and B. van Deurs. 1989. Endocytosis from coated pits of Shiga toxin: a glycolipid-binding protein from
dysenteriae 1. J. Ce11 Biol. 108: 133 1-43.12 1.
Sandvig, K and J. E. Brown. 1987. Ionic requirernents for entry of Shiga toxin from
Shigella dysenteriue 1 into cells. Infect. Immun. 55298-303.
Saukkonen, K., W. N. Bsrnette, V. L, Mar, FI. R. Masure, and E. 1. Tuornanen. 1992.
Penussis toxin has eukaryotic-like carbohydrate recognition domains. Proc. Nad. Acod Sci.
USA 89:1 18-22.
Saxena, S. K., A. D. O'Brien, and E. J. Ackerman. 1989. Shiga toxin, Shiga-like toxin II variant, and ricin are al1 single-site RNA N-glycosidases of 28 S RNA when rnicroinjected into Xenopus oocytes. J. Biol. Chem 264596-60 1. 140
Schlossrnan, D., D. Withers, P. Welsh, A. Alexander, J. Robertus, and A, Frankel- 1989. Role of glutamic acid 177 of the ricin toxin A chah in enzyrnatic inactivation of ribosomes.
Mol. Ce11 BioL !MO 12-502 1,
Schmitz, A., M. Maintz, T. Kehle and V. Herzog. 1995. In vivo iodination of a misfolded
proinsulin reveals CO-localizedsignals for Bip binding and for degndation in the ER. EMBO
J. 14: 109 1- 1098.
Scotland, S. M., R, H. Smith, and B. Rowe. 1985. Two distinct toxins active on Vero cells
from Escherichia coli 0 157. Lancer, II:885-886.
Scudiero, D. A., R. H. Shoernaker, K. D. Paull. A. Monks, S. Tierney, T. H. Nofziger. M. J. Cumens, D. Seniff. and M. R. Boyd. 1988. Evaluation of a soluble tetrazoliurn/formazan assay for ce11 growth and drug sensitivity in culture using human and other tumor ce11 lines. Cancer Res. 48:4827-4833.
Seidah, N. G., A. Donohue-Rolfe. C. Lazure, F. Auclair, G. T. Keusch, and M. Chretien.
1986. Cornpiete arnino acid sequrnce of Shigella toxin B-chah A novel polypeptide containing 69 arnino acids and one disulfide bridge. J. Bioi. Chem 261: 13928-3 1.
Sherwood. D., D. R. Snodgrass. and A. D. O'Brien. 1985. Shiga-like toxin production from Fcr herichia coli associated with calf diarrhoea. Vet. Record 1l6:2 17-8.
Shiga. K. 1897. Study on the cause of shigellosis. I (in Japanese). Saikingaku-Zasshi
25: 790-8 1O. 141
Shoemaker, K. R., P. S. Kim, E. J. York, J, M. Stewart, and R. L. Baldwin. 1987. Tests
of the helix dipole mode1 for stabilization of alpha-helices. Nature. 326563-7.
Simon, S. 1993. Translocation of proteins across the endoplasmic reticulum. Curr. Opin.
Cell BioL 5581-8.
Simpson. J. C., C. Dascher, L. M. Roberts. J. M. Lord, and W. E. Balch. 1995. Ricin cytotoxicity is sensitive to recycling between the endoplasmic reticulum and the Golgi
cornplex. 3. Biol. Chem. 270:20078-83.
Sixma, T. K., P. E. Stein, W. G. Ho1 and R. J. Read. 1993. Cornparison of the B-pentamer
of Heat-labile enterotoxin and verotoxin- 1: two structures with remarkable simdarity and dissimilarity. Biochernistry 32: 19 1- 198.
Sixma, T. K., S. E. Pronk, K. H. Kalk, E. S. Wartna, B. A. van Zanten. B. Witholt, and W. G. Hol. 199 1. Crystal structure of a cholera toxin-related heat-labile enterotoxin from E. coIi [see comments]. Nature. 3Sl:37 1-7.13 1.
Sjogren, R., R. Neill, D. Rachmilewitz, D. Fritz, J. Newland, D. Sharpnack, C. Colleton,
J. Fondacaro, P. Gemski, and E. Baedeker. 1994. Role of Shiga-like toxin 1 in bacterial enteri tis: corn parison between isogenic Escherichia coli stnins induced in rabbits [sez commenü]. Gmtroenterology. 106:306- 17.
Smith, H. W., P. Green, and Z. Parsell. 1983. Vero cell toxins in Escherichia coli and related bacteria: transfer by phage and conjugation and toxic action in labontory anirnals, chickens and pigs. J. Gen. MicrobioL 129:3 12 1-3 137.
Spiess, M., and H. F. Lodish. 1986. An internal signal sequence: the asialoglycoprotein
receptor membrane anchor. Ceil. 44: 177-85.
St Hilaire. P. M., M. K. Boyd, and E. J. Toone. 1994. Interaction of the Shiga-like toxin type 1 B-subunit with its carbohydrate receptor. Biochemisny 33: 14452- 14463.
Stein. P. E., A. Boodhoo. G. J. Tyrrell, J. L. Brunton, and R. J. Read. 1992. Crystal
structure of the cell-binding B oligomer of verotoxin-l from E. coli. Namre. 355748-50.
Stirpe, F., L. Barbieri, M. G. BatteUi. M. Soria, and D. A. Lappi. 1992. Ribosome- inactivating proteins from plants: present stanis and future prospects. Biotech. 10:405
Strockbine. N. A.. L. R. Marques. R. K. Holmes, and A. D. O'Brien. 1985a.
Characterization of monoclonal antibodies against Shiga-like toxin from Escherichia di.
Infect. Immun.. 50:695-700.
Strockbine, N. A.. L. R. M. Marques, J. W. Newland, H. W. Smith, R. K. Holmes, and
A. D. O'Brien. 198%. Chmcterization of monoclonal antibodies against Shiga-like toxin from Escherichia coli Infect. Immun. 50: 695-700.
Strockbine, N. A., M. P. Jackson. L. M. Sung, R. K. Holmes, and A. D. O'Brien. 1988.
Cloning and sequencing of the genes for Shiga toxin frorn Shigellu dysenteriue type 1. J-
Bacteriol. . 170: 1 116-22. 143
Stromberg, N., P.-G. Nyholm, 1. Pascher, and S. Nomark. 1991. Saccharide orientation at
the ce11 surface affects glycolipid receptor function. Proc. Nad Acad Sei. USA. 88:93&
9344.
Surewicz, W. K., K. Surewicz, H. H. Mantsch, and F. Auclair, 1989. Interaction of
Shigella toxin with globotriaosyl cerarnide receptor-containing membranes : A fluorescence study. Biochern Biophys. Res. Cornm. 160: 126-132.
Takao, T., T. Tanabe, Y. M. Hong, Y. Shimonishi, H. Kurazono, T. Yutsudo, C. Sasakawa, M. Yoshikawa, and Y. Takeda. 1988. Identity and molecuiar structure of Shiga- like toxin 1 (VTl) from Escherichia coli 0157:H7 with that of Shiga toxin. Microbial
Pathogen. 557-69.
Tedder, T. F., and C. M. Isaacs. 1989. Isolation of cDNAs encoding the CD19 antigen of human and mouse B lymphocytes. A new member of the immunoglobulin superfarnily. J.
Immun. 143:7 12-7.
Tesh, V. L., J. A. Burris, J. W. Owens, V. M. Gordon, E. A. Wadolkowski, A. D- O'Brien, and J. E. Samuel. 1993. Cornparison of the relative toxicities of Shiga-like toxins I and type II for mice. Infect. Immun. 61:3392-3402.
Tycko, B. And F. R. Maxfield. 1982. Rapid acidification of endocytic vesicles containing at-macroglobulin. Cell. 28:643-65 1.
Tzipori, S., K. 1. Wachsmuth, J. Srnithers, and C. Jackson. 1988. Studies in gnotobiotic 144 piglets on non-O157:H7 Escherichia coli serotypes isolated from patients with hemoahagic colitis. Gastroenterology 94:590-597.
Uckun, F. M. 1993. Immunotoxins for the treatment of leukemia Br& J. Haemat. 85435- 438.
Utsumi, T., Y. Aizono, and G. Funatsu. 1984. Interaction of ricin and its constituent polypeptides with dipalmitoyIphosphatidylcho1ine vesicles. Biochim Biophys- Acta. 772:202-8.
Utsumi, T., Ide, A. & Funatsu, G. 1989. EZicin A-chah induces fusion of small unilarnellar vesicles at neutral pH. FEBS Lett. 242:255-258.
Vallera, D. A. 1994. Immunotoxins: will their clinical promise be fulfilled? Blood. 83:309- 317.
Waddel, T., A. Cohen, and C. A. Lingwood. 1990. Induction of verotoxin sensitivity in receptor-deficient ce11 lines using the receptor glycolipid globoaiosylceramide. Proc. Nutl.
Acad Sci. USA. 87:7898-790 1.
Wadolkowski, E. A., L. M. Sung, J. A. Bums, J. E. Samuel, and A. D. O'Brien. 1990a.
Acute rend tubular necrosis and death of mice orally infected with Escherichia coli strains that produce Shiga-like toxin type 1I.Infect. Immun. 58:3959-3965.
Wadokowski, E. A., J. A. Burris. and A. D. O'Brien. 1990b. Mouse modei for 145
colonization and disease caused by enterohemorrhagic Escherichia coli 0157:H7. Infect.
Immun. 58:2438-2445.
Wang, Z., J. D. Jones, J. Rizo, and L. M. Giensch. 1993. Membrane-bound conformation
of a signal peptide: a transferred nuclear Overhauser effect analysis. Biochernistry.
32: 1399 1-9.
Ward. C. L.. S. Omura, R. R. Kopito. 1995. Degradation of CFTR by the ubiquitin- proteasorne pathway. Ceil. 83: 12 1 - 127.
Wells. J. G., B. R. Davis. 1. K. Wacchssmuth, L. W. Riley, R. S. Remis, R. Sokolow, and G. K. Moms. 1983. Labontory investigation of hemorrhagic colitis outbreaks associated with a rare Escherichia serotype. J. Clin. Microbiol. 185 12-520.
Wharton, S. A., S. R. Manin. R. W. Ruigrok. J. J. Skehel. and D. C. Wiley. 1988.
Membrane fusion by peptide analogues of influenza virus haemagglutinin. J. Gen-Virol.
69: 1847-57.
Wu. Y. N., M. Gadina, J. H. Tao-Cheng. and R. J. Youle. 1994. Retinoic acid disrupts the
Golgi apparatus and increases the cytosolic routing of specifc protein toxins. J. Cell Biol.
125:743-753.
Yamasaki, N.. Y. Nagase & G. Funatsu. 1988. Hydrophobicities of Ricin D and its constituent polypeptide chahs. Agric. Biol. Chem 52:102 1- 1026. 146
Yamasaki, S., M. Furutani, K. Ito. K. [garashi, M. Nishibuchi, and Y. Takeda. 1991. Importance of arginine at position 170 of the A subunit of Vero toxin 1 produced by
enterohemorrhagic Eschenchia coli for toxin activity. Microbiol. Pathog. 11: 1-9.
Yazdi, P. T., and R. M. Murphy. 1994. Quantitative analysis of protein synthesis inhibition
by transferrin-toxin conjugates. Cancer Res. 54:6387-6394.
Yutsudo, T.. T. Honda T. Miwatani and Y. Takeda. 1987. Physicochemical characterization of A and B subunits of Shiga toxin and reconstitution of holotoxin from isolated su bunits. Microbiol. Immun. 31: 189- 197.
Zachowski, A. 1993. Phospholipids in animal eukaryotic membranes: transverse asyrnmetry and movement. J. Biochem 294: 1- 14.
Zalman, L. S., and B. J. Wisnieski. 1984. Mechanism of insertion of diphdiena toxin: peptide entry and pore size determination. Proc. Nad. Acad. Sei. USA. 8l:334 1-3345.
Zhan, H., K. I. Oh, Y. K. Shin, W. L. Hubbeil, and R. J. Collier. 1995. interaction of the isolated transrnembrane domain of diphtheria toxin with membranes. Biochemistry.
34:4856-63.
Zhao, J. M., and E. London. 1988. Conformation and mode1 membrane interactions of diphtheria toxin fragment A. J. Biol. Chem 263: 15369-77.
Zoja, C., D. Coma, C. Fuina, G. Sacchi, C. Lingwood, M. P. Doyle, V. V. Padhye, M. Abbate and G. Rernuzzi. 1992. Verotoxin glycolipid receptors determine the localization of rnicroangiopathic process in rabbits given verotoxin- 1. J. Lob. Clin. Med 120:229-238.
Zoban, T. M., P. R Austin, P. E. Jablonski, and C. J. Hovde. 1994, Purification of recombinant Shiga-iike tain type 1 A 1 fragment from Escherichia coli Prof. Eip & Purif.
S:Z9 1-5. IMAGE EVALUATION TEST TARGET (QA-3)
APPLIEO IMAGE. lnc -.-5 1653 East Main Street --- - Rochester. NY 1- USA ------Phone: 716i482-0300 ---- Fa: 716/288-5989
O 1993. Appiied Image. lm.. Aii RWts Resetwd