VI. Applications of S-layers

25

Transcript of VI. Applications of S-layers

VI. Applications of S-layers

1

Uwe B. Sleytr2;a, Hagan Bayley

2;b, Margit Saèra

3;a, Andreas Breitwieser

a,

Seta Kuëpcuëa, Christoph Mader

a, Stefan Weigert

a, Frank M. Unger

a,

Paul Messnera, Beatrice Jahn-Schmid

a, Bernhard Schuster

a, Dietmar Pum

a,

Kenneth Douglasc, Noel A. Clark

c, Jon T. Moore

c, Thomas A. Winningham

c,

Samuel Levyd, Ivar Frithsen

c, Jacques Pankovc

e, Paul Beale

c, Harry P. Gillis

f,

Dmitri A. Choutovg, Kevin P. Martin

g

aZentrum fuër Ultrastrukturforschung und Ludwig Boltzmann-Institut fuër Molekulare Nanotechnologie, Universitaët fuër Bodenkultur,

A-1180 Vienna, Austria

bDepartment of Medical Biochemistry and Genetics, Texas ApM Health Science Center, 440 Reynolds Medical Building,

College Station,

TX 77843-1114, USA

cDepartment of Physics, University of Colorado, Boulder, CO 80309, USA

dDepartment of Molecular, Cellular, and Developmental Biology, University of Colorado, Boulder, CO 80309, USA

eAstralux, Inc., 2386 Vassar Drive, Boulder, CO 80303, USA

fDepartment of Chemistry and Biochemistry, UCLA, Los Angeles, CA 90095, USA

gMicroelectronics Research Center, Georgia Institue of Technology, Atlanta, GA 30332, USA

Abstract

The wealth of information existing on the general principle of S-layers has revealed a broad application potential. The most

relevant features exploited in applied S-layer research are: (i) pores passing through S-layers show identical size and

morphology and are in the range of ultrafiltration membranes; (ii) functional groups on the surface and in the pores are aligned

in well-defined positions and orientations and accessible for binding functional molecules in very precise fashion; (iii) isolated

S-layer subunits from many organisms are capable of recrystallizing as closed monolayers onto solid supports at the air-water

interface, on lipid monolayers or onto the surface of liposomes. Particularly their repetitive physicochemical properties down to

the subnanometer scale make S-layers unique structures for functionalization of surfaces and interfaces down to the ultimate

resolution limit. The following review focuses on selected applications in biotechnology, diagnostics, vaccine development,

biomimetic membranes, supramolecular engineering and nanotechnology. Despite progress in the characterization of S-layers

and the exploitation of S-layers for the applications described in this chapter, it is clear that the field lags behind others (e.g.

0168-6445 / 97 / $32.00 ß 1997 Federation of European Microbiological Societies. Published by Elsevier Science B.V.

PII S 0 1 6 8 - 6 4 4 5 ( 9 7 ) 0 0 0 4 4 - 2

FEMSRE 586 28-10-97

1This review is part of a series of reviews dealing with different aspects of bacterial S-layers; all these reviews appeared in Volume

20/1-2 (June 1997) of FEMS Microbiology Reviews, thematic issue devoted to bacterial S-layers.

2Guest Editor.

3Corresponding authors : Prof. Dr. Margit Saèra : Tel. : +43 (1) 476 54 ext. 2208; Fax: +43 (1) 478 91 12; E-mail :

[email protected]. Dr. Frank M. Unger: Tel. : +43 (1) 476 54 ext. 6052; Fax: +43 (1) 478 91 12; E-mail: [email protected].

Dr. Bernhard Schuster: Tel. : +43 (1) 476 54-2203; Fax: +43 (1) 478 91 12; E-mail: [email protected]. Prof. Dr. Kenneth

Douglas: Tel. : +1 (303) 492 1515; Fax: +1 (303) 492 2998; E-mail: [email protected]. Prof. Dr. Dietmar Pum: Tel.: +43 (1)

476 54 ext. 2205; Fax: +43 (1) 478 91 12; E-mail: [email protected]

FEMS Microbiology Reviews 20 (1997) 151^175

enzyme engineering) in applying recent advances in protein engineering. Genetic modification and targeted chemical

modification would allow several possibilities including the manipulation of pore permeation properties, the introduction of

switches to open and close the pores, and the covalent attachment to surfaces or other macromolecules through defined sites on

the S-layer protein. The application of protein engineering to S-layers will require the development of straightforward

expression systems, the development of simple assays for assembly and function that are suitable for the rapid screening of

numerous mutants and the acquisition of structural information at atomic resolution. Attention should be given to these areas

in the coming years.

Keywords: Ultra¢ltration membrane; Immunoassay; A¤nity technique; Dipstick; S-layer; Patch pipette; Black lipid membrane; Solid

supported membrane; Filtration; Nanotechnology; Immobilization; Patterning; Immunotherapy of type 1 allergies; Polysaccharide vaccine

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152

2. Exploitation of the ¢ltration and immobilization potential of S-layers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153

2.1. Continuous cultivation of S-layer-carrying organisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153

2.2. S-layer ultra¢ltration membranes (SUMs) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154

2.3. S-layers as matrices for covalent binding of biologically active macromolecules . . . . . . . . . . . . . . . . . . . . . 154

2.4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

3. Vaccine applications of crystalline bacterial surface layer proteins (S-layers) . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

3.1. Approaches to immunotherapy of cancers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

3.2. Antibacterial vaccines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

3.3. Approaches to immunotherapy of type 1 allergies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

3.4. Conclusions and perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

3.4.1. Cancer immunotherapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

3.4.2. Antibacterial vaccines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

3.4.3. Immunotherapy of type 1 allergies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

4. Planar supported lipid membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159

4.1. Lipid membranes generated on pipettes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159

4.2. Black lipid membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159

4.3. Solid supported lipid membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160

4.4. Conclusion and perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161

5. Parallel nanofabrication using microbial S-layers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162

5.1. Nanopatterned metal oxide/S-layer composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162

5.2. Pattern transfer using metal oxide/S-layer masks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

5.3. In situ recrystallization to develop S-layer `designer patterns' . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166

6. Supramolecular engineering with S-layer membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167

6.1. S-layers as immobilization matrices for a geometrically de¢ned binding of biologically functional molecules 168

6.2. Recrystallization of S-layer proteins on solid supports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168

6.3. Lithographic patterning S-layers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168

6.4. S-layers as patterning structures in biomineralization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171

6.5. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171

Acknowledgements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171

References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172

1. Introduction

This review deals with the application of S-layers

from bacteria. S-layers (surface layers) are envelope

structures exterior to the cell wall proper or replacing

other cell wall structures. Most common among

these surface components are regularly arranged

S-layers composed of protein or glycoprotein sub-

FEMSRE 586 28-10-97

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175152

units which form crystalline monomolecular assem-

blies. They form porous lattices completely covering

the cell surface. An important feature of these struc-

tures are their repetitive physicochemical properties.

Evidently, these layers play important biological roles

which are described in other chapters of this issue.

There are several emerging applications of S-layers

for di¡erent purposes which will be discussed here.

2. Exploitation of the ¢ltration and immobilization

potential of S-layers

Margit Saèra

3, Andreas Breitwieser, Seta Kuëpcuë,

Christoph Mader, Stefan Weigert, Uwe B. Sleytr

Crystalline bacterial cell surface layers (S-layers)

possess pores identical in size and morphology and

functional groups aligned in well-de¢ned lattice po-

sitions [1^5]. These speci¢c features make S-layers

unique as biopolymers with repetitive surface proper-

ties down to the subnanometer scale. S-layer proteins

also reveal the ability to self-assemble into two-di-

mensional crystalline arrays in suspension, on solid

supports, at the air/water interface and on lipid ¢lms

[6]. Most importantly, S-layer proteins can be pro-

duced in large amounts by continuous cultivation of

S-layer-carrying organisms. Thus, S-layer proteins

can be considered biopolymers with properties

ideally tailored by nature for many biotechnological

applications such as the production of isoporous ul-

tra¢ltration membranes [2], as matrices for the im-

mobilization of a variety of materials including bio-

logically active macromolecules [3^5] and for

functionalization of inanimate materials [6].

The properties of S-layers are in clear contrast to

those of most polymers used for biotechnological

applications. Due to their amorphous structure, con-

ventional polymers have a wide pore size distribution

and a random arrangement of functional groups.

This is one of the major reasons why ultra¢ltration

membranes produced from amorphous polymers by

the phase inversion process do not possess sharp

molecular mass cut-o¡s [7]. Even using di¡erent

polymer mixtures, the introduction of functional

groups by grafting or the chemical modi¢cation of

preformed ultra¢ltration membranes this technolog-

ically relevant problem has not yet been solved [8].

Thus, the application of ultra¢ltration membranes in

biotechnology is still limited to desalting or concen-

trating of protein solutions.

In comparison to amorphous polymers, S-layers

reveal a high density of functional groups on the

outermost surface [9,10]. When used as a matrix

for the immobilization of macromolecules, closed

monolayers of covalently bound foreign proteins

are formed which frequently re£ect the symmetry

of the underlying S-layer lattice [11^15]. Since amor-

phous polymers have a spongy structure and a lower

density of functional groups, it is not possible to

build up monolayers of covalently bound macromo-

lecules. Usually, more than 90% of the macromole-

cules are immobilized inside the three-dimensional

polymeric network leading to di¡usion-limited reac-

tions and unspeci¢c adsorption [16].

2.1. Continuous cultivation of S-layer-carrying

organisms

For producing large amounts of cell wall frag-

ments and S-layer protein of de¢ned quality,

S-layer-carrying organisms are grown in continuous

culture under steady state conditions. One of the

most common organisms used for producing

S-layer-carrying cell wall fragments and S-layer pro-

tein for biotechnological applications is Bacillus

stearothermophilus PV72, for which a synthetic

growth medium was developed [17] by applying the

pulse and shift technique in continuous culture [18].

In comparison to complex medium, the use of syn-

thetic medium is advantageous with respect to cost

and to limit the organisms to better de¢ned metabol-

ic pathways. By changing the speci¢c growth rate

during continuous cultivation, both S-layer protein

production and the activity of the cells' autolysins

could be controlled, which was of major importance

for the formation of an inner S-layer during the cell

wall preparation procedure [17]. Feeding of an ami-

no acid mixture composed of Ala, Gly, Leu, Ile, Val,

Glu, Asp, Asn and Gln to the continuous culture

signi¢cantly stimulated S-layer protein production

while growth in the presence of excess glucose led

to reversible reduction of S-layer protein synthesis

[19]. In the case of B. stearothermophilus PV72, ex-

pression of a second S-layer gene could be triggered

by increasing the dissolved oxygen concentration in

FEMSRE 586 28-10-97

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175 153

the culture [20]. For example, at a DO of 20^30%,

the hexagonally ordered S-layer lattice of the wild-

type strain was produced, which was composed of

the SbsA protein. This S-layer has been used for

the production of S-layer ultra¢ltration membranes

(SUMs) [21]. At higher oxygen concentration

(DO=50%) synthesis of SbsA was replaced by that

of the SbsB protein which assembled into an oblique

S-layer lattice composed of subunits with a molecu-

lar mass of 97 000 [20,21]. SbsB was able to recrys-

tallize into closed monolayers on a great variety of

surfaces such as poly-L-lysine-coated EM grids,

Langmuir-Blodgett lipid ¢lms or unilamellar lipo-

somes. This unique feature has made SbsB particu-

larly suitable for applications in molecular nanotech-

nology [6]. Both S-layer proteins have been cloned in

Escherichia coli where they retained the ability to

assemble into sheet-like structures in the cytoplasm

of the host cells [22].

2.2. S-layer ultra¢ltration membranes (SUMs)

Analysis of the mass distribution and permeability

properties of isolated S-layer lattices of various Ba-

cillaceae revealed that they function as isoporous

molecular sieves with a pore size of 4^5 nm corre-

sponding to a molecular mass cut-o¡ in the range of

40 000 [1^3]. SUMs are produced by depositing

S-layer-carrying cell wall fragments (with a complete

outer and inner S-layer covering the peptidoglycan-

containing layer) or S-layer self-assembly products

on micro¢ltration membranes, crosslinking the

S-layer protein with glutaraldehyde under a pressure

of 2U105Pa and reducing Schi¡ bases with sodium

borohydride [2,21]. Presently, SUMs are produced in

a microprocessor-controlled apparatus with a size of

30U60 cm. Since crosslinking of the S-layer subunits

with glutaraldehyde blocks the amino groups, SUMs

show a net negative charge on the surface and inside

the pores. The negative surface charge density was

determined for the square S-layer lattice of Bacillus

sphaericus CCM 2120 [9,10]. It could be demon-

strated that 61 carboxyl groups were available per

S-layer subunit which corresponded to 1.6 carboxyl

groups per nm2S-layer lattice [9,10].

Adsorption studies and contact angle measure-

ments con¢rmed that the net negatively charged

standard SUMs are basically hydrophobic [9], which

could be changed by covalently binding appropriate

low molecular mass nucleophiles to the carbodi-

imide-activated carboxyl groups of the S-layer pro-

tein [9,10,23,24]. By applying selected modi¢cation

reactions, 7.0 carboxyl or amino groups or other

functional residues such as hexadecylamine or glu-

cosamine could be introduced per nm2SUM [9,10].

The use of such modi¢ed SUMs for adsorption stud-

ies with proteins allowed a determination of correla-

tions between the physicochemical properties of the

SUM surface, the molecular characteristics (dimen-

sion, net charge) of the adsorbed proteins and the

£ux losses caused by protein adsorption [8,10]. Fur-

thermore, it could be demonstrated that adsorption

of a single layer of protein molecules inside the pores

of the S-layer lattice can cause £ux losses of up to

80% from the initial water £ux [9]. In contrast, pro-

tein molecules that were too large to enter the pores

and adsorbed to surface-located S-layer protein do-

mains led to £ux losses of only 20%. Due to the wide

pore size distribution observed for synthetic ultra¢l-

tration membranes, similar studies on the correlation

between the molecular size of the adsorbed proteins

and the £ux losses as a consequence of fouling can-

not not be readily interpreted.

2.3. S-layers as matrices for covalent binding of

biologically active macromolecules

For immobilization of biologically active macro-

molecules such as enzymes (invertase, glucose oxi-

dase, glucuronidase, L-glucosidase, naringinase, per-

oxidase), ligands (protein A, streptavidin) or mono-

and polyclonal antibodies, S-layer-carrying cell wall

fragments possessing a complete outer and inner

S-layer or SUMs were used [3^5]. In both materials,

the S-layer lattice was crosslinked with glutaralde-

hyde before the carboxyl groups of the S-layer pro-

tein or the hydroxyl groups of the carbohydrate

chains in S-layer glycoproteins were activated for

covalent binding of foreign macromolecules [11,13^

16,25]. S-layer-carrying cell wall fragments with im-

mobilized protein A could be applied as escort par-

ticles in a¤nity cross-£ow ¢ltration for isolation and

puri¢cation of human IgG from serum or of mono-

clonal antibodies from hybridoma cell culture super-

natants [15,16]. Both the protein A, which was co-

valently bound to the carbodiimide-activated

FEMSRE 586 28-10-97

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175154

carboxyl groups of the S-layer protein, and the sub-

sequently adsorbed IgG formed a monolayer on the

S-layer surface. The a¤nity microparticles were

highly resistant against shear forces under cross-

£ow conditions. No protein A or S-layer protein

leakage could be observed [16]. S-layer-carrying cell

wall fragments with covalently bound IgG were also

applicable as microparticles in immunoassays [25].

The IgG was linked either to the carboxyl groups

of the protein moiety, or to the cyanogen bromide-

activated or periodate-oxidized carbohydrate chains

of S-layer glycoproteins [25]. Since de¢ned amounts

of IgG could be immobilized on the S-layer lattice,

the use of S-layer microparticles in immunoassays

led to highly reproducible absorption curves. Fur-

ther, no unspeci¢c adsorption of test substances or

enzyme-antibody conjugates applied during the im-

munoassay was observed [25].

SUMs with immobilized monoclonal antibodies

were also used as reaction zones for dipstick-style

immunoassays [14]. In the course of dipstick devel-

opment, the methods for immobilizing antibodies

had to be optimized. For this purpose, human IgG

was either directly coupled to the carbodiimide-acti-

vated carboxyl groups of the S-layer protein (Fig.

1a), or it was adsorbed onto a SUM with covalently

bound protein A (Fig. 1b). Alternatively, human IgG

was biotinylated and bound to a SUM onto which

streptavidin was immobilized in a monomolecular

layer (Fig. 1c) [14]. Comparison of the three di¡erent

methods showed that in case of the protein A-SUM,

700 ng human IgG was bound per cm2membrane

area, which exactly corresponded to a monolayer of

uniformly oriented IgG molecules with a compact

state of the Fab region (Fig. 1b). When IgG was

covalently bound to the carbodiimide-activated car-

boxyl groups of the S-layer protein, 375 ng could be

immobilized per cm2

SUM, corresponding to a

monolayer of randomly oriented IgG molecules

(Fig. 1a). In the case of biotinylated IgG, 150 ng

was adsorbed per cm2

streptavidin-SUM, corre-

sponding to a maximum 60% coverage of the S-layer

surface with IgG (Fig. 1c). To investigate the suit-

ability of SUMs with human IgG immobilized by the

three di¡erent methods as reaction zone for immuno-

assays, increasing concentrations of rabbit anti-hu-

man IgG were applied. The bound anti-human IgG

was ¢nally quantitated with an anti-rabbit alkaline

phosphatase conjugate using p-nitrophenyl phos-

phate as substrate. As shown in Fig. 2, the IgG im-

mobilized via protein A gave the highest response

whereas the two con¢gurations led to comparable

absorption values in the immunoassay. Since the

amount of the directly coupled IgG was twice that

of the bound biotinylated IgG, the results indicate

that the biotinylated human IgG had retained a

higher biological activity, most probably due to the

presence of a spacer in the biotin molecules [14].

FEMSRE 586 28-10-97

Fig. 1. Schematic drawing illustrating the immobilization of IgG

to carbodiimide-activated carboxyl groups of the S-layer protein

of SUMs (a), to protein A covalently bound to the S-layer lattice

(b) and after biotinylation to a streptavidin-modi¢ed SUM (c).

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175 155

Nevertheless, considering the costs of protein A or

streptavidin and the a¤nity of protein A to the dif-

ferent types of antibodies, the direct immobilization

of IgG is the most economical method for prepara-

tion of SUM-based dipsticks. However, the compa-

rative immobilization studies demonstrated that

S-layers are well suited for determining the binding

density of immobilized foreign macromolecules [14].

Moreover, the formation of a monolayer of the di-

rectly coupled IgG on the SUM surface prevented

di¡usion-controlled reactions and unspeci¢c adsorp-

tion [14].

For developing the dipstick hardware, small SUM

disks (6 mm in diameter) with the immobilized anti-

body were punched out of larger SUM samples and

aligned between two pieces of te£on foil. The upper

piece had an opening of about 3 mm diameter where

the SUM surface with the immobilized antibody was

exposed (Fig. 3). After gluing the pieces of foil to-

gether the dipsticks were ready for use. The align-

ment of the SUM disks between the pieces of foil

was necessary for completely covering the micro¢l-

tration membrane. By this production method, back-

ground reactions due to unspeci¢c adsorption of the

micro¢ltration membrane could be prevented.

So far, two di¡erent types of dipstick-style immu-

noassays have been developed: one for determining

birch pollen allergen-speci¢c IgE in human serum for

diagnosis of allergies, which represent a serious prob-

lem in Europe (about 15% of the population are

a¡ected by this disease [26]) and the other for mon-

itoring tissue plasminogen activator (t-PA) in human

blood or plasma during t-PA therapy [27]. In both

cases, a speci¢c monoclonal mouse antibody was co-

valently bound to the carbodiimide-activated carbox-

yl groups on the S-layer surface in SUMs. After

binding of the recombinant major birch pollen aller-

gen Bet v 1a to the monoclonal mouse antibody BIP

1 the dipstick was ready for use. The bound r Bet v

1a was recognized by IgE. After washing with water

the dipstick was incubated with an anti-IgE alkaline

phosphatase conjugate. The substrate used for the

enzyme was 5-bromo-4-chloro-3-indolylphosphate/

nitroblue tetrazolium (BCIP/NBT), which in case of

a positive reaction formed a violet precipitate on the

SUM surface. Semi-quantitative determination or ex-

act quantitation of the bound IgE was possible either

by comparing the intensity of the precipitate with a

color card or by measuring it with a re£ectometer.

For determination of t-PA, the dipstick with the im-

mobilized monoclonal mouse antibody 3-VPA was

incubated in human plasma together with an anti-t-

PA peroxidase conjugate for 15 min. As substrate for

peroxidase, 3-amino-9-ethylcarbazole was chosen,

which in case of a positive reaction formed a red

precipitate on the SUM surface. The concentration

of t-PA which can be determined with the SUM-

based dipsticks is between 0 and 200 ng/ml plasma.

The whole procedure takes about 20 min and is

therefore well suited for t-PA monitoring during

t-PA therapy after myocardial infarcts.

FEMSRE 586 28-10-97

Fig. 2. Absorbance values at 405 nm for di¡erent concentrations

of anti-human IgG applied to SUMs on which human IgG was

either covalently bound to the S-layer protein, immobilized via

protein A, or after biotinylation linked to a streptavidin-modi¢ed

SUM.

Fig. 3. Schematic drawing showing the preparation of a SUM-

based dipstick. The monoclonal antibodies (MAB) were linked to

carbodiimide-activated carboxyl groups of the S-layer protein.

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175156

2.4. Conclusions

Due to the presence of pores identical in size and

morphology and the uniform and dense arrangement

of functional groups on the outermost surface,

S-layers are particularly suitable for the production

of isoporous ultra¢ltration membranes with sharp

molecular mass cut-o¡s and as matrices for the con-

trolled immobilization of macromolecules. For pro-

ducing a¤nity microparticles, antibodies or ligands

such as protein A were covalently bound to the

S-layer lattice of S-layer-carrying cell wall fragments

showing a complete outer and inner S-layer. Such

a¤nity microparticles were applied as escort particles

in a¤nity cross-£ow ¢ltration or as microparticles in

immunoassays. SUMs with immobilized monoclonal

antibodies proved to be well suited as novel reaction

zones for dipstick-style immunoassays. Since anti-

bodies or ligands can be immobilized as monolayers

on the S-layer lattice, problems arising with di¡u-

sion-limited reactions and unspeci¢c adsorption

were prevented.

3. Vaccine applications of crystalline bacterial surface

layer proteins (S-layers)

Frank M. Unger3, Paul Messner,

Beatrice Jahn-Schmid, Uwe B. Sleytr

The experimental use of crystalline bacterial sur-

face layer proteins (S-layers) as combined carrier/ad-

juvants for vaccination and immunotherapy has pro-

gressed since 1987 in three areas of application:

immunotherapy of cancers, antibacterial vaccines,

and antiallergic immunotherapy. Work performed

from 1987 to 1991 in the cancer immunotherapy

and antibacterial vaccine areas has been reviewed re-

cently [28]. Here, these applications are only brie£y

summarized. Since 1994, investigations have focused

on the immunological properties of S-layer conju-

gates with Bet v 1, the main allergen of birch pollen.

The goal of these studies is the suppression of the IgE-

mediated manifestations of type 1 hypersensitivity.

3.1. Approaches to immunotherapy of cancers

The suitability of S-layers as combined carrier/ad-

juvants for therapeutic cancer vaccines was ¢rst sug-

gested by Smith et al. [29]. These investigators found

that S-layer conjugates of the tumor-associated gly-

cans LGal1,3KGalNAc (T antigen) and KFuc1,2L-

Gal1,4[KFuc1,3]LGlcNAc (Lewis y antigen) primed

BALB/c mice for strong, hapten-speci¢c delayed-

type hypersensitivity responses (DTH). Signi¢cantly,

the DTH responses were achieved without the use of

an extraneous adjuvant. When administered alone,

conjugates of the same haptens with bovine serum

albumin (BSA) elicited only weak DTH responses.

However, the DTH responses elicited by the BSA

conjugates in conjunction with aluminum oxide

(alum) as adjuvant were similar to those observed

with hapten-S-layer conjugates. An interesting fea-

ture of the arti¢cial antigens formed by conjugating

tumor-associated glycans to S-layers is the balance of

humoral and cellular immune responses elicited in

mice. In particular, those antigenic preparations con-

taining cross-linked S-layers elicited very low titers of

antibody in relation to the magnitude of the DTH

responses. Adoptive transfer experiments indicated

that the observed DTH responses are mediated by

T-helper cells [29]. From these ¢ndings, the conclu-

sion was drawn that the natural propensity of

S-layer protomers to assemble into large, two-dimen-

sional arrays endows them with intrinsic adjuvant

properties. DTH responses speci¢c for tumor-associ-

ated glycans were also observed when the hapten-

S-layer conjugates were administered to mice by

the oral/nasal route [30]. These responses were at

least as strong as those following intramuscular ap-

plication of the S-layer conjugates.

3.2. Antibacterial vaccines

Conjugates of S-layers with oligosaccharides

derived from the capsular polysaccharide of Strep-

tococcus pneumoniae type 8 elicited immunopro-

tective antibodies in mice [31] as shown in a serum

killing assay. Sera from mice immunized with Strep-

tococcus pneumoniae type 8-oligosaccharide-

S-layer conjugates reduced S. pneumoniae type 8

colony forming units by 99% on blood agar plates,

whereas sera from mice that had been immunized

with S. pneumoniae type 8 capsular polysaccharide

had no e¡ect. The S. pneumoniae type 8-S-layer

conjugates also elicited DTH in mice to an ex-

FEMSRE 586 28-10-97

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175 157

tent comparable with the e¡ect of heat-killed bacte-

ria.

3.3. Approaches to immunotherapy of type 1 allergies

In allergic individuals, the production of IgE anti-

bodies is mediated by the TH2 helper lymphocyte

subset; by contrast, non-allergic individuals produce

low levels of speci¢c IgG to allergens, a response

mediated by TH0/TH1 cells [32]. Successful induction

of tolerance to allergens in atopic patients was found

to be associated with a shift from a TH2-type to a

TH0/TH1-like cytokine pattern (decreased IL-4 pro-

duction and increased IFN-Q production) [33^35].

Based on these experiences, it has been suggested

that redirecting of regulatory T lymphocytes toward

a TH0/TH1-like cytokine secretion pattern might con-

stitute a promising strategy for therapy and possibly

prophylaxis of type 1 allergy [36]. Previous results of

immunization experiments with tumor antigen-

S-layer conjugates had suggested that immune re-

sponses in animals can be modulated toward a

TH1- or a TH2-directed response through the choice

and construction of the respective S-layer conjugates

[28,29]. Therefore, since 1994, the immunological

properties have been explored of conjugates formed

from S-layers and Bet v 1, the main allergen of birch

pollen [37]. These conjugates contain intact B-cell

epitopes, as demonstrated in inhibition experiments

using human Bet v 1-speci¢c IgE. Also, the S-layer-

Bet v 1 conjugates were shown to be immunogenic

in mice. The peptides created by antigen processing

of Bet v 1-S-layer conjugates appear to be similar

to those derived from natural allergen, as indi-

cated by the proliferation of Bet v 1-speci¢c T-cell

clones. When human, allergen-speci¢c TH2 lym-

phocytes were stimulated with Bet v 1-S-layer con-

jugates, a modulation of the cytokine produc-

tion from a TH2- to a TH0/TH1-like pattern was

observed [38]. These ¢ndings are taken to indicate

the potential use of S-layers as carrier/adjuvants for

immunotherapeutic vaccines in Type 1 hypersensi-

tivity.

3.4. Conclusions and perspectives

3.4.1. Cancer immunotherapy

The recent discoveries of two new pathways of

antigen processing ^ the CD-1 and the vacuolar

pathway ^ in addition to the well-known MHC class

I and MHC class II mechanisms suggest new ap-

proaches to the stimulation of e¡ective immune re-

sponses against tumor-associated antigens [39]. In

this context, it will be necessary to identify immuno-

logical mechanisms ^ humoral or cellular ^ capable

of activating therapeutically useful antitumor e¡ec-

tor functions [40]. It is expected that the recently

developed S-layer-covered liposomes [41] will consti-

tute a valuable tool in this exploratory research,

serving as a vehicle for hapten-S-layer conjugates

in combination with a variety of protein, lipid or

glycolipid immunomodulators. To provide a repre-

sentative selection of target antigens, a program of

chemical synthesis of tumor-associated glycans has

been initiated.

3.4.2. Antibacterial vaccines

Conjugates composed of S-layers and oligosac-

charides corresponding to fragments of bacterial

capsular polysaccharides have been demonstrated

to elicit protective antibody titers in animals [31].

The absence of measurable toxicity and the availabil-

ity of a multitude of S-layer preparations [42], many

of which are immunologically not cross-reactive,

make S-layers an interesting choice of carrier/adju-

vant for experimental conjugate vaccines against

encapsulated pathogens that occur in a variety of

serotypes (e.g. S. pneumoniae) [43]. At present, devel-

opment of such experimental vaccines suitable for

tolerance and e¤cacy studies in humans is under

way.

3.4.3. Immunotherapy of type 1 allergies

The ¢nding that TH0/TH1-like mediator secretion

can be induced in allergen-speci¢c, TH2 helper lym-

phocyte clones in vitro has suggested further experi-

ments with peripheral blood cells from individuals

allergic to birch pollen. When T-helper clones were

established from these cells, the use of Bet v 1-

S-layer conjugates as the primary stimulating antigen

resulted in a much higher proportion of Bet v

1-speci¢c TH0/TH1 clones than did the use of un-

conjugated Bet v 1 [44]. Development is in prog-

ress of Bet v 1-S-layer conjugates suitable for ex-

perimental immunotherapy of pollen allergy in

humans.

FEMSRE 586 28-10-97

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175158

4. Planar supported lipid membranes

Bernhard Schuster

3, Dietmar Pum, Uwe B. Sleytr

Membrane-associated and integrated proteins are

a class of biological macromolecules important in

medicine and biotechnology, particularly in the de-

velopment of biosensors [6,45]. Investigations of re-

lationships between structure and function of pro-

teins and cell membranes are commonly performed

with arti¢cial lipid layers as matrices [46^48]. These

model systems replicate the physiological environ-

ment of ionophores [48] without the many compli-

cating cell-speci¢c factors and e¡ects on ion channels

formed in classical patch clamp investigations

[49,50]. The electrophysical and geometrical simplic-

ity of planar lipid membranes provides an environ-

ment conducive to sensitive assay for ionophores

[51]. However, the poor stability of plain lipid mem-

branes and the sometimes denaturing environment

for transmembrane proteins require support of the

lipid membranes by biocompatible substrates [52].

A useful approach is the application of recrystallized

bacterial cell surface-layers (S-layers) as supporting

layers [6].

Currently, there are several strategies for the for-

mation of planar lipid membranes [51,53,54], includ-

ing bilayers covalently coupled to a supporting sub-

strate [55,56]. Due to the broad interest in this ¢eld

of research, many multidisciplinary approaches have

been suggested. This chapter summarizes the most

common strategies for the formation of plain and

supported lipid membranes and discusses ideas

which methods might be used for speci¢c problems.

4.1. Lipid membranes generated on pipettes

One approach to generate lipid membranes span-

ning an ori¢ce is the use of patch pipettes made of

glass [54,57,58]. There are two methods to generate

planar lipid bilayers on the tip of the pipette. One

procedure, known as the `tip-dip' method, is to pass

the pipette twice through a lipid monolayer at the

air/water interface [54]. The lipid monolayer can be

generated either by the Langmuir-Blodgett (LB)

technique [59] or by adding vesicle suspensions into

the subphase [60]. On the other hand, the pipette can

be punched through a much larger, preformed lipid

bilayer [57,58]. Both methods lead to the generation

of a small (6 40 Wm diameter) clamped membrane

(Fig. 4A). The remarkably tight lipid-glass interac-

tion causes a seal resistance of several tens of giga-

ohms [50]. It was demonstrated that membranes gen-

erated by the LB technique can be supported by the

recrystallization of isolated S-layer protein subunits

on the lipid/water interface [6,61,62]. This is per-

formed simply by injecting a solution of monomers

of bacterial S-layers into the subphase. After a cer-

tain time, large coherent arrays of recrystallized

S-layer lattices closely attached to the lipid mem-

brane can be observed by electron microscopic tech-

niques [6,61,62].

Solvent free membranes generated on pipettes pro-

vide low background noise and a short current set-

tling time after an applied voltage step [63]. General

experimental limitations of this method are restricted

access to the side of the bilayer bathed by the pipette

solution and that the membrane patch may be too

small to allow reasonable rates of direct channel in-

corporation [53]. Preliminary studies indicate that

S-layer-supported lipid membranes generated on pi-

pettes (see Fig. 4B) gain in stability in terms of life

time especially when the membranes are functional-

ized with incorporated ionophores (B. Schuster, D.

Pum and U.B. Sleytr, submitted). Nevertheless, the

generation of lipid membranes on pipettes requires

considerable experience and technical support [63].

4.2. Black lipid membranes

This method involves the generation of a lipid bi-

layer over a septum with an ori¢ce of 40^800 Wm

[47,51]. For this purpose either a small drop of lipid

dissolved in alkane is placed on the opening of the

septum [46] or the membrane is formed from two

lipid monolayers at an air/water interface by the ap-

position of their hydrocarbon chains through an

aperture made in a hydrophobic partition which sep-

arates the two monolayers (Fig. 5A) [51,64]. How-

ever, both methods require a preconditioning of the

ori¢ce with hexadecane or decane solved in a volatile

solvent [64]. The success of the membrane formation

depends strongly on the shape of the ori¢ce [53].

Thus, much e¡ort has been put into the design of

sophisticated apertures [53,65]. In analogy to lipid

membranes generated on pipettes, aperture-spanning

FEMSRE 586 28-10-97

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175 159

lipid membranes can also be supported with S-layers

(Fig. 5B). In comparison to plain black lipid mem-

branes they reveal a decreased tendency to disinte-

grate especially when membrane proteins are as-

sembled and incorporated in the lipid membranes

([66] ; B. Schuster, D. Pum, H. Bayley and U.B.

Sleytr, in preparation).

The advantage of black lipid membranes is their

easy handling and the ability to add ionophores or

change the electrolyte on both sides of the membrane

[47]. S-layer-supported lipid membranes have the po-

tential advantage of a greater additional stability.

However, there is solvent left within the lipid bilayer

leading to an inhomogeneous hydrophobic core in

the lipid membrane. The application of black lipid

membranes is limited especially for single channel

measurements when relatively high activation poten-

tials (vVv 300 mV) are necessary [58]. Another dis-

advantage is the poor signal resolution of black lipid

membranes compared to membranes generated on

pipettes [58].

4.3. Solid supported lipid membranes

Solid supported lipid membranes are most com-

monly generated on surfaces such as metals

[55,67,68], glass [56], silicon, mica [69] or on polym-

ers [52,70]. Di¡erent methods have been developed.

The lipid bilayer can be generated by applying a

drop of a lipid dissolved in hexadecane on a freshly

cut tip of a te£on-coated wire (Fig. 6) [67]. Another

method is to bind long chain mercaptans to a gold

FEMSRE 586 28-10-97

Fig. 5. (A) Schematic illustration of a black lipid membrane generated by the method of Montal and Muëller [51]. (B) An S-layer-

supported black lipid membrane.

Fig. 4. (A) Schematic illustration of a lipid bilayer, generated on the tip of a patch clamp pipette. (B) The lipid bilayer on the pipette is

supported by a crystalline S-layer.

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175160

surface [68] and subsequently transfer a second lipid

layer either by LB techniques [59,68] or by vesicle

spreading [69] (Fig. 7A). The lipid membrane may

also be separated from the metal surface by ultrathin

water containing polymer ¢lms (Fig. 7B) [52,70] or

S-layers (Fig. 7C) [6,66]. The ultrathin layer of water

covering the metal surface is thought to provide a

natural environment for the lipid membrane and for

associated or integrated molecules.

The advantage of these methods is that large areas

of lipid membranes can be generated [56] and that

the membrane gains mechanical stability [55,67]. But,

especially with the method involving the cutting of a

te£on-coated wire within a drop of lipid dissolved in

hexadecane, the roughness of the metal surface in-

duces great variations in the thickness of the lipid

membrane and thus the membrane experiences a

strong inhomogeneous electric ¢eld [67]. One way

to overcome this problem is the binding of long

chain mercaptans on £at gold surfaces. Unfortu-

nately, light, especially at short wavelengths, leads

to photoinduced electrical currents as radiation in-

teracts with the gold surface [68]. Separating the sol-

id surface and the lipid membrane by water-contain-

ing, polymer cushions of a few nanometer in

thickness [52] or S-layers [6,66] provides a more nat-

ural environment. The lipid membrane gains in

stability and transmembrane proteins protruding

from the lipid membrane are separated from the

metal surface [52]. In comparison to polymer cush-

ions, S-layer lattices represent a structurally much

better de¢ned matrix which can be speci¢cally func-

tionalized [6]. A general disadvantage of solid sup-

ported membranes is that they provide access to only

one side of the membrane.

4.4. Conclusion and perspectives

The increasing interest in and application of elec-

trophysical techniques has led to the development of

di¡erent approaches to generate lipid matrices. Im-

portant requirements for detailed investigations of

membrane-associated or membrane-integrated bio-

molecules are lipid ¢lms with an adequate £uidity

combined with su¤cient stability [45]. Membranes

generated on pipettes are most appropriate for inves-

tigations of single ion channels [49,63], providing a

low background noise and short current settling time

after an applied voltage step [63]. Black lipid mem-

branes, which are also appropriate for measurements

on single ion channels, and supported lipid mem-

branes generated on te£on-coated metal wires are

relatively simple techniques, suitable for measure-

ments with larger electrical signals [51,64]. These lip-

id membranes contain variable concentrations of

hexadecane or decane within their hydrophobic

FEMSRE 586 28-10-97

Fig. 7. (A) Schematic illustrations of a lipid bilayer built up by a monolayer of long chain mercaptans covalently bound to a gold surface

and an attached phospholipid monolayer. (B) The solid supported phospholipid bilayers are separated from the substrate by an ultrathin

polymer or (C) by a crystalline S-layer.

Fig. 6. Schematic illustration of a solid supported lipid bilayer

generated on a te£on-coated metal wire.

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175 161

core. Due to the increased thickness of the hydro-

phobic core, some transmembrane biomolecules lose

their biological activity as they are not able to span

the lipid membrane [54]. Solid supported lipid mem-

branes reveal a higher mechanical stability although

the membranes are in some cases rather inhomoge-

neous [67]. Another concept is to modify the solid

surface either by covalently linked lipid monolayers

[68] or by polymer cushions to get a more homoge-

neous membrane with a biocompatible environment

[52,70].

A new approach to study membrane-associated

and integrated biomolecules is the application of

S-layer-supported lipid membranes [6,66]. With

S-layers large scale homogeneous lipid membranes

with an enhanced stability especially in the presence

of ionophores can be generated ([63] ; B. Schuster, D.

Pum, H. Bayley and U.B. Sleytr, in preparation).

When these structures are spread over apertures,

functional studies on membrane-associated or inte-

grated molecules can be combined with structural

studies by transmission electron microscopy and

scanning tunnelling or atomic force microscopy

[71]. In addition S-layer-supported lipid membranes

generated on apertures made of silicon have the ad-

vantage of being accessible from both sides and the

electrolyte can be easily changed. Composite S-layer/

lipid membranes could provide ideal biomimetic

structures for studying functions of transmembrane

proteins [6,71]. Thus, composite S-layer/lipid mem-

branes represent a new promising tool for studying

functions of many biomacromolecules at meso- and

macroscopic scale [6,66,71].

5. Parallel nanofabrication using microbial S-layers

Kenneth Douglas3, Noel A. Clark, Jon T. Moore,

Thomas A. Winningham, Samuel Levy,

Ivar Frithsen, Jacques Pankovec, Paul Beale,

Harry B. Gillis, Dmitri, A. Choutov,

Kevin P. Martin

Biomimetic approaches to the fabrication of ad-

vanced materials is an interdisciplinary undertaking

which has proliferated in recent years [72]. One of

the many facets of this electric ¢eld is the use of two-

dimensionally organized organic surfaces as biomi-

metic templates for material deposition and fabrica-

tion. Mesoscopic crystalline monolayers are particu-

larly attractive for nanometer scale fabrication

because their periodicity provides a length scale ame-

nable to structural study by a variety of established

methods which represent a technology base for the

development of new nanometer fabrication techni-

ques. This work is an extension of the metal deco-

ration and shadowing methods employed in bio-

membrane structural investigations [73,74].

Additionally, the structural redundancy of periodic

monolayer arrays o¡ers a signi¢cant advantage in

that a single preparation yields many examples of

the same process, enabling £uctuation e¡ects, which

will become of increasing importance as device size is

reduced, to be e¡ectively probed.

5.1. Nanopatterned metal oxide/S-layer composites

It has been reported that titanium oxide-coated,

two-dimensional microbial S-layers can be employed

as masks for the parallel nanostructuring of the

underlying substrate [75]. In those experiments, ion

beam milling of the metal oxide/protein crystal com-

posite was used to create a screen containing peri-

odic arrays of nanometer scale holes which possess

the same 22 nm periodicity as the underlying protein

crystal lattice [76]. It was then shown that this metal

oxide screen can act as a mask for the transfer of this

array of holes to the underlying substrate.

Details of the experimental procedures used to ob-

tain the S-layers and to utilize them in biologically

derived nanofabrication have been described previ-

ously [72]. Brie£y, the two-dimensional protein crys-

tals form the surface layer (S-layer) of the bacteria

Sulfolobus acidocaldarius. S-layers are isolated by a

modi¢ed version of the procedure used by Michel et

al. [77]. These protein crystals (V1 Wm in diameter)

are deposited from an aqueous suspension onto sub-

strates (e.g. (100) crystalline silicon) and coated with

titanium at an oblique angle of incidence (40 from

normal incidence) by electron beam deposition (Fig.

8). In experiments designed to produce a metal

screen with periodic nanodimensional holes (nano-

screen), the average titanium thickness is measured

in vacuo by a quartz crystal monitor to be 1.2 nm.

The titanium ¢lm subsequently oxidizes when ex-

posed to air. Metal oxide thickness measured by

FEMSRE 586 28-10-97

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175162

atomic force microscopy (AFM) and independently

con¢rmed by spectroscopic ellipsometry is 3.6 nm.

After ion milling at normal incidence with 2 keV

argon ions and a beam current density of 7 WA/

cm

2for 12 min, the titanium oxide/S-layer composite

assumes the form of a nanoscreen as shown in Fig.

9b. In experiments designed to produce a periodic

array of metal oxide dots (nanodot array), the aver-

age titanium thickness as deposited is 0.6 nm (which

oxidizes to 1.8 nm prior to ion milling). After ion

milling with 2 keV argon ions at a current of 7 WA/

cm

2for 12 min, the titanium oxide/S-layer composite

assumes the form shown in Fig. 9c.

In order to explain the titanium oxide particle re-

organization on the S-layer which leads to these pat-

terns, a statistical mechanical model similar to the

Ising model was created (manuscript in preparation).

This model is a solid-on-solid model with titanium

atoms stacked in a simple hexagonal lattice on top of

a protein and silicon surface which simulates the

S-layer structure. The model is based by the binding

energies between two Ti atoms, between a Ti atom

and the silicon substrate, and between a Ti atom and

the S-layer protein. The motion of the Ti atoms is

governed by a Metropolis Monte Carlo process. The

system tends to evolve towards its lowest free energy

state (from di¡erent initial conditions) in both the

case of nanoscreen and nanodot array formation.

Fig. 10a shows the time evolution for the case of

the nanodot array as simulated by the model, and

Fig. 10b shows AFM images of the experimentally

realized nanodot array.

5.2. Pattern transfer using metal oxide/S-layer masks

As discussed earlier, argon ion milling has been

used to transfer the S-layer pattern to a substrate

surface in the form of a periodic array of etch pits

[75,78]. More recently, the dry etching technique of

low energy electron enhanced etching (LE4) has also

FEMSRE 586 28-10-97

Fig. 8. Processing steps to make nanostructures by S-layer lithography; (a) deposition of S-layers onto substrate; (b) shadow metallization

of S-layers by electron beam vaporization of titanium; (c) ion milling to reorganize the metal coating and transfer the resulting pattern to

the substrate.

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175 163

been employed to transfer an ordered nanopattern

onto a silicon surface while in£icting minimal etch

damage to the silicon lattice. (Similar studies have

been done using gallium arsenide as the substrate.)

This lack of damage will be of particular importance

for fabrication on the nanometer length scale since

the depth of the damage layers produced by more

standard ion-enhanced techniques approaches the

size of the fabricated structures. Moreover, the peri-

odic alteration of the silicon surface chemistry nucle-

ates the growth of a metal nanocluster array from a

metal ¢lm evaporated onto the surface after the met-

al oxide/S-layer mask has been removed (manuscript

in preparation).

Previous reports have demonstrated that, as a

method of pattern transfer, LE4 is competitive with

standard dry etching techniques in yield, rate, aniso-

tropy, and surface smoothness; however, LE4 has

the added advantage of doing minimal damage

[79^81]. Because LE4 etches by delivering electrons

with kinetic energies of 1^15 eV to the surface along

with a reactive species, the momentum imparted to

the sample is negligible, resulting in minimal dam-

age. Metal oxide/protein crystal nanoscreens on a

FEMSRE 586 28-10-97

Fig. 9. Atomic force microscope (AFM) images of (a) titanium-coated S-layer prior to ion milling; (b) ion-milled titanium-coated S-layer

formed into a nanoscreen; (c) ion-milled titanium-coated S-layer formed into a nanodot array.

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175164

FEMSRE 586 28-10-97

Fig. 10. (a) Computer simulation of the creation of a nanodot array upon ion milling of an S-layer/titanium composite; the number of

steps in the Monte Carlo modeling simulation is indicated beneath each panel. (b) AFM images of the time evolution of a nanodot array

formed from ion milling of an S-layer/titanium nanocomposite ¢lm.

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175 165

crystalline silicon surface were etched by LE4 in a dc

plasma con¢guration in 100 mTorr of H2 at room

temperature. Samples were then thinned by tripod

(HRXTEM). Fig. 11 shows a low magni¢cation

view of an area of a sample exposed to LE4. The

periodic nanometer scale pattern of the protein crys-

tal mask has been etched into the Si lattice to a

depth of 10 nm. Because the cross-section occurs at

an arbitrary orientation with respect to the rows of

etched holes, inhomogeneities in the pattern and

even missing holes in the image may arise as repre-

sented schematically in the inset to Fig. 11. Etched

features appear fairly isotropic. However, in previous

experiments, Si(100) which has been LE4 patterned

on a micron length scale with metal and with dielec-

tric masks has shown etch directionality of various

degrees, from nearly vertical sidewalls to classical

isotropic etching [80]. The extent of undercut in-

creases with hydrogen partial pressure in all cases.

Substantial improvement in the results presented

here should be possible through process optimiza-

tion.

Following pattern transfer by LE4, the metal ox-

ide/S-layer mask was stripped o¡ with a 1:1 solution

of H2SO4 :H2O at 130³C. the sample was then

subjected to an oxygen plasma for 30 s at 1 keV

and V8 mA. 12 A

î

of titanium was then deposited

by electron beam evaporation at normal incidence.

Upon AFM examination, the sample revealed or-

dered arrays of nanoclusters displaying the same

symmetry and lattice constant as the S-layer pattern-

ing mask. The mechanism for the nanocluster array

formation is currently under investigation.

It is anticipated that such two-dimensionally or-

dered metal nanocluster arrays may themselves be

used as LE4 etch masks in order to pattern the sub-

strate material into an array of nanodots. Such small

groups of atoms are sometimes referred to as quan-

tum dots or quantum boxes.

5.3. In situ recrystallization to develop S-layer

`designer patterns'

While there are a wide variety of potential appli-

cations for surfaces which can be nanostructured us-

ing randomly deposited, naturally occurring S-layers

as masks, it is clear that the more interesting appli-

cations will require growth of crystals in selected

places. Therefore, it is desirable to work toward

the development of methods whereby low resolution

(submicron) patterns can serve as templates for mo-

lecular self-assembly of high resolution (10 nm)

structures, which, in turn, are templates for nano-

meter scale fabrication. The goal is to develop `de-

signer' patterns by replacing the naturally occurring

S-layer crystals with recrystallized S-layers grown in

situ in a selected geometry on a chosen area of a

surface which is to be nanostructured.

FEMSRE 586 28-10-97

Fig. 11. High resolution cross-sectional transmission electron micrograph (HRXTEM) of a portion of a crystalline silicon sample which

has been nanopatterned using an S-layer/titanium mask and the dry etching technique of low energy electron enhanced etching (LE4).

The inset shows how the arbitrary orientation of the cross-section in£uences the appearance of the pattern.

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175166

One means to this goal is to use low resolution

templating by means of patterned deep ultraviolet

(UV) lithography of organosilane self-assembled

monolayer (SAM) ¢lms to create partitioning of

the substrate into regions of hydrophilic and hydro-

phobic a¤nities. The recrystallization of the solubi-

lized S-layer monomeric units will then be nucleated

on this patterned area and following this one can

perform parallel nanofabrication (as described previ-

ously) using these `designer patterns'. The use of

deep UV photochemistry and patterning of SAM

¢lms to achieve the low resolution templating is

based on the work of Calvert et al. [82]. One ap-

proach is to simply use the silanol functionality

(Si-OH) which is exposed after deep UV photocleav-

age of the organic groups from the SAM ¢lm (Fig.

12). The silanol functionality can be used to bind

glycosylated protein crystal monomers by hydrogen

bonding. In order to provide a high disparity in

binding a¤nity between the substrate areas which

were designated to form crystals and those which

were not, one would use an organosilane whose

functional group is hydrophobic, e.g. trimethoxysi-

lane. Thus, after UV patterning, the exposed polar

regions have a markedly higher probability for

chemisorption of the glycoprotein than the unex-

posed (hydrophobic) regions. Alternatively, one

could employ the silanol sites resulting from deep

UV irradiation as reaction sites for the chemisorp-

tion of a second silane ¢lm with a functionality

known to show preferential of the S-layer protein

monomers.

Another route to preferential recrystallization

would be mimetic of that of Dressick et al. who

have used selective electroless (EL) metallization in

patterning experiments [83]. They reported an appli-

cation of the patterning technique described above to

chemisorbed ligand-bearing organosilanes such

as 2-[2-(trimethoxysilyl)ethyl]pyridine] (PYR). Pat-

terned deep UV irradiation resulted in photodesorp-

tion of the pyridyl chromophore from the surface of

the exposed areas. In the unexposed areas, intact

pyridyl ligands can interact with Pd(II) solution spe-

cies to generate catalyzed surfaces amenable to EL

metallization processes. It is important to the pat-

terning that Pd(II) deposition is selective, i.e. it

does not occur in the absence of the PYR nitrogen

group. Consequently, only the areas unexposed to

the UV and thus still possessing the surface-bound

Pd species will catalyze EL metallization. Regarding

site-selective S-layer recrystallization experiments,

once the PYR surfaces are coated with Pd the photo-

lysed (Si-OH) surfaces will have a preferential a¤n-

ity for chemisorption of glycoproteins to nucleate

recrystallization (Fig. 13). Another possibility would

be to perform one more chemisorption reaction to

deposit a second organosilane ¢lm which has a de-

termined a¤nity for the glycoprotein subunits.

6. Supramolecular engineering with S-layer

membranes

Dietmar Pum

3, Uwe B. Sleytr

S-layers have been shown to be excellent pattern-

ing structures in molecular nanotechnology due to

their high molecular order, high binding capacity

and ability to recrystallize with perfect uniformity

on solid surfaces, at the air/water interface and on

lipid ¢lms [6]. In particular, the recrystallization of

S-layer subunits on substrates such as silicon, gal-

FEMSRE 586 28-10-97

Fig. 12. (a) Chemisorbed ligand-bearing organosilanes on silicon substrate. (b) After patterned deep UV irradiation. (c) Selective electro-

less metallization with Pd coating unexposed areas. (d) Protein adsorption at hydrophilic (silanol) sites.

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175 167

lium arsenide, gold or glass allows their application

as immobilization matrices in the ¢eld of supramo-

lecular engineering [6].

6.1. S-layers as immobilization matrices for a

geometrically de¢ned binding of biologically

functional molecules

The controlled immobilization of functional mol-

ecules on surfaces is a useful tool for the develop-

ment of supramolecular structures. In conventional

carriers the location, local density and orientation of

functional groups and the porosity and pore size are

only de¢ned approximately. In S-layer lattices, the

properties of a single constituent protein unit are

replicated with the periodicity of the lattice and

thus de¢ne the characteristics of the two-dimensional

immobilization matrix. Electron microscopic studies

have shown that macromolecules may be immobi-

lized on S-layers as densely packed crystalline arrays

[11,84,85]. Molecules may be bound to S-layer latti-

ces either by non-covalent interactions (e.g. ionic

bonds, hydrogen bonds or hydrophobic interactions)

or by covalent attachment after activation of func-

tional groups (e.g. carbodiimide activation of car-

boxyl groups) on the S-layer or the molecules. The

pattern of bound molecules frequently re£ects the

lattice type, the size of the morphological unit and

the distribution of physicochemical properties on the

array. So far, S-layer lattices have been used as im-

mobilization matrices for a broad spectrum of bio-

logically active proteins (e.g. enzymes, antibodies,

ligands). Based on these structures, a broad range

of amperometric or optical biosensors and solid

phase assays have been developed [4,66,86^88].

6.2. Recrystallization of S-layer proteins on solid

supports

S-layer proteins isolated from numerous organ-

isms can be recrystallized on solid surfaces such as

silicon, glass, carbon or synthetic polymers [89]. The

orientation of the protein array on the support is

determined by the physicochemical anisotropy of

the surface properties of S-layer lattices. Electron

and scanning force microscopic studies have revealed

that recrystallized S-layers are oriented with their

charge-neutral, more hydrophobic outer face against

hydrophobic solid surfaces. It has been demon-

strated that S-layer proteins usually do not assemble

into coherent monolayers on hydrophilic substrates.

The determination of the orientation of a recrystal-

lized S-layer is particularly easy for oblique lattice

symmetry while higher spacegroup symmetries re-

quire advanced image processing methods.

6.3. Lithographic patterning S-layers

The basic requirements for manufacturing supra-

molecular devices are a spatial control of the immo-

bilization matrix and a geometrically de¢ned binding

of functional molecules. Patterning S-layers may be

achieved by exposure to deep ultraviolet radiation

(ArF; 193 nm) [90,91]. Prior to irradiation the re-

crystallized S-layer must be dried (e.g. in a stream

of high purity nitrogen gas) to remove excess water

FEMSRE 586 28-10-97

Fig. 13. (a) Chemisorbed ligand-bearing organosilanes on silicon substrate. (b) After patterned deep UV irradiation. (c) Selective electro-

less metallization with Pd coating unexposed areas. (d) Protein adsorption at hydrophilic (silanol) sites.

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175168

not required for maintaining the structural integrity

of the protein lattice. Test patterns could be trans-

ferred onto the S-layer by a mask which was brought

in direct contact with the wafer (Fig. 14). The pat-

tern on the mask (100 nm thick chromium coating

on synthetic quartz glass) consisted of lines and

squares (feature sizes ranging from 200 nm to

1000 nm) with di¡erent line-and-space ratios. The

experiments have shown that the S-layer lattice was

completely removed by ArF radiation with a dosage

of V200 mJ/cm2which was supplied in two pulses

of 100 mJ/cm2each (Fig. 15). Exposure to KrF ra-

diation (248 nm) showed that the S-layer was only

carbonized but not removed even after increasing the

dosage up to 10 laser pulses of V350 mJ/cm2each

[90]. In this context it is important that the unex-

posed regions of the S-layer maintained its integrity

as immobilization matrix for functional macromole-

cules (e.g. enzymes, antibodies) or as supporting

structures for functional lipid membranes.

FEMSRE 586 28-10-97

Fig. 14. (a) Schematic drawing of patterning a recrystallized S-layer on a silicon wafer by deep ultraviolet radiation. (b) The S-layer is ab-

lated in the exposed regions. The integrity of the protein matrix is retained in the unexposed areas and may be used either as (c) an im-

mobilization matrix for binding functional molecules, or (d) as natural nanoresist enhanced by refractive clusters or electroless metalliza-

tion for subsequent reactive ion etching.

Fig. 15. Scanning force microscopic image of a patterned S-layer

on a silicon wafer. Bar, 2000 nm, z-range 10 nm.

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175 169

As an alternative to their application as immobi-

lization matrices, S-layers might also make useful

natural nanoresists (Fig. 14d). Since S-layers are

only 5^10 nm thick, which is much thinner than

conventional resists, considerable improvement in

edge resolution in the fabrication of submicron

structures can be expected. In this approach the

S-layer regions which have not been removed by ra-

diation could be either used as a matrix for selective

coating with refractory clusters or ampli¢ed by elec-

troless metallization as already demonstrated for

monolayer ¢lms [92]. In our latest results, we have

succeeded in using the S-layer as the top layer in a

two-layer resist system (Fig. 16) [91]. As the bottom

layer resist a spin-coated novolak resist was used. It

is well known that this resist material can be ablated

by exposure to KrF excimer laser radiation. The

patterning was performed in two steps by starting

with a transfer of the pattern onto the S-layer by

exposure with ArF radiation (193 nm) (Fig. 16a).

Subsequently the wavelength was changed to 248

nm (KrF radiation) and the novolak resist at the

bottom ablated by a blank exposure using the pat-

terned top S-layer as mask (Fig. 16b). This technique

yielded very steep sidewalls in the resist material

(Fig. 16c, Fig. 17).

S-layers have also been used in the fabrication of

FEMSRE 586 28-10-97

Fig. 17. Scanning electron micrograph of a novolak resist which

was structured by a blank exposure using a patterned S-layer on

top as mask (see text for details).

Fig. 16. Schematic drawing of the two-layer resist technology. The recrystallized S-layer is the top layer and a spin coated novolak resist

the bottom layer. (a) First, the S-layer is patterned by ArF radiation (= 193 nm). (b) Subsequently the wavelength is changed to 248 nm

(KrF radiation) and the novolak resist at the bottom is ablated by a blank exposure using the patterned top S-layer as mask. (c) This

technique yields very steep sidewalls in the resist material.

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175170

nanometric metallic templates [75,93]. In this ap-

proach a nanometer thick metal ¢lm (Ta/W or Ti)

was deposited onto an hexagonal S-layer and

thinned by ion milling. In the course of this proce-

dure a perforated metal mask was obtained where

the periodically arranged holes (V15 nm in diame-

ter) resembled the underlying S-layer structure.

6.4. S-layers as patterning structures in

biomineralization

Biomineralization is becoming increasingly impor-

tant in the synthesis of inorganic materials exhibiting

uniform particle size, morphology, oriented nuclea-

tion and assembly [94]. S-layers may be used in this

¢eld as a geometrically precisely de¢ned surface with

accurately positioned nucleation sites for biomineral-

ization. This has already been demonstrated in the

precipitation of CdS on S-layer protein monolayers

on hydrophobic solid supports. In this study, the

corrugated and net negatively charged inner face of

an S-layer was exposed to the reagent. Electron mi-

croscopic studies in combination with image process-

ing techniques have shown that the localized nega-

tive charge and periodic surface topography favor

the formation of nanoparticles with approximately

5 nm diameter (Fig. 18). This was not surprising

since recent studies on a cyanobacterial S-layer

clearly demonstrated that a crystalline surface layer

can be involved in the production of ¢ne grain min-

erals in lakes [95,96].

6.5. Conclusion

Studies on the ultrastructure, chemistry, genetics,

morphogenesis and function of S-layers have re-

vealed a very broad application potential for two-

dimensional (glyco)protein crystals as patterning

structures in molecular nanotechnology and supra-

molecular engineering. In particular, S-layers which

have recrystallized on solid supports represent

unique matrices for geometrically de¢ned binding

of functional molecules and as natural nanoresist

for structuring surfaces. In the future, the possibility

of generating metallic point patterns on S-layer lat-

tices could lead to the development of nanostruc-

tures with novel electronic or optical properties.

Acknowledgments

The work of U.B.S. and coworkers on the exploi-

tation of S-layers was supported in part by grants

from the Austrian Science Foundation, Project

FEMSRE 586 28-10-97

Fig. 18. (a) Transmission electron micrographs of a nanometric point pattern of CdS particles obtained by biomineralization on an S-

layer with oblique lattice symmetry. Protein appears white, CdS particles dark. Bar, 60 nm. (b) Corresponding computer image recon-

struction to (A). Bar, 10 nm.

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175 171

S7204-MOB, the Federal Ministry of Science and

Transportation, and the Austrian National Bank,

Project 5525. The S-layer-covered liposomes are

being developed by Drs. S. Kuëpcuë and C. Mader.

F.M.U., P.M., B.J. and U.B.S. gratefully acknowl-

edge the contribution to the cancer immunotherapy

and antibacterial vaccine sections by Drs. R.H.

Smith and A.W. Malcom and their associates at

Chembiomed Ltd., Edmonton, Alberta, Canada.

Drs. W. Schmid, Irmgard Wenzel and R. Prenner

of the University of Vienna have expertly synthesized

tumor-associated carbohydrate antigens. For con-

struction, formatting and puri¢cation of hapten-

S-layer conjugates we thank Drs. C. Schaë¡er and

A. Zenker.

The authors thank Drs. D. Kraft, O. Scheiner and

C. Ebner for valuable suggestions and for the hospi-

tality of their facilities at the General Hospital

(AKH) in Vienna. Financial support was provided

by the Austrian Science Foundation, Project S7206-

MOB (to P.M.), and the Federal Ministry of Science

and Transportation, and the Austrian National

Bank, Projects 5525 (to U.B.S.) and 6011 (to

F.M.U.).

B.S., D.P. and U.B.S. would like to thank B. Wet-

zer for critical reading of the manuscript and S. Die-

luweit for helpful discussions. Their work was

supported by grants from the Austrian Science

Foundation, Projects S7204 and S7205, the Austrian

Federal Ministry of Science and Transportation, and

the Austrian National Bank, Project 5525 (to

U.B.S.).

The research of the American laboratories was

funded in part by the Army Research O¤ce, the

Air Force O¤ce of Scienti¢c Research, the Colorado

Advanced Technology Institute, and the Colorado

Advanced Materials Institute.

The work of U.B.S. on supramolecular engineer-

ing was supported by the Austrian Science Founda-

tion, Grants S7204 and S7205, and the Federal Min-

istry of Science and Transportation, and the

Austrian National Bank, Project 5525.

References

[1] Saèra, M. and Sleytr, U.B. (1987) Molecular-sieving through

S-layers of Bacillus stearothermophilus strains. J. Bacteriol.

169, 4092^4098.

[2] Saèra, M. and Sleytr, U.B. (1987) Production and character-

istics of ultra¢ltration membranes with uniform pores from

two-dimensional arrays of proteins. J. Membrane Sci. 33,

27^49.

[3] Saèra, M., Kuëpcuë , S. and Sleytr, U.B. (1996) Biotechnological

applications of S-layers. In: Crystalline Bacterial Cell Surface

Proteins (Sleytr, U.B., Messner, P., Pum, D. and Saèra, M.,

Eds.), pp. 133^159. R.G. Landes/Academic Press, Austin, TX.

[4] Saèra, M. and Sleytr, U.B. (1996) Biotechnology and biomi-

metic with crystalline bacterial cell surface layers (S-layers).

Micron 27, 141^156.

[5] Sleytr, U.B. and Saèra, M. (1997) Bacterial and archaeal

S-layer proteins: structure-function relationship and their bio-

technological applications. Trends Biotechnol. 15, 20^26.

[6] Pum, D. and Sleytr, U.B. (1996) Molecular nanotechnology

and biomimetics with S-layers. In: Crystalline Bacterial Cell

Surface Proteins (Sleytr, U.B., Messner, P., Pum, D. and Saèra,

M., Eds.), pp. 175^209. R.G. Landes/Academic Press, Austin,

TX.

[7] Nakao, S. (1994) Determination of pore size and pore size

distribution. 3. Filtration membranes. J. Membrane Sci. 96,

131^165.

[8] Stengaard, F.F. (1988) Characteristics and performance of

new types of ultra¢ltration membranes with chemically modi-

¢ed surfaces. Desalination 70, 207^224.

[9] Weigert, S. and Saèra, M. (1995) Surface modi¢cation of an

ultra¢ltration membrane with crystalline structure and studies

on interactions with selected protein molecules. J. Membrane

Sci. 106, 147^159.

[10] Weigert, S. and Saèra, M. (1996) Ultra¢ltration membranes

prepared from crystalline bacterial cell surface layers as model

systems for studying the in£uence of surface properties on

protein adsorption. J. Membrane Sci. 121, 185^196.

[11] Saèra, M. and Sleytr, U.B. (1989) Use of regularly structured

bacterial cell surface layers as matrix for immobilizing macro-

molecules. Appl. Microbiol. Biotechnol. 30, 184^189.

[12] Saèra, M., Kuëpcuë , S., Weiner, C., Weigert, S. and Sleytr,

U.B. (1993) S-layers as immobilization and a¤nity matrices.

In: Advances in Paracrystalline Bacterial Surface Layers

(Beveridge, T.J. and Koval, S.F., Eds.), pp. 195^204. Plenum,

New York.

[13] Kuëpcuë , S., Mader, C. and Saèra, M. (1995) The crystalline cell

surface layer from Thermoanaerobacter thermohydrosulfuricus

L111-69 as an immobilization matrix: in£uence of the mor-

phological properties and the pore size of the matrix on loss of

activity of covalently bound enzymes. Biotechnol. Appl. Bio-

chem. 21, 275^286.

[14] Breitwieser, A., Kuëpcuë , S., Howorka, S., Weigert, S., Langer,

C., Ho¡mann-Sommergruber, K., Scheiner, O., Sleytr, U.B.

and Saèra, M. (1996) 2-D protein crystals as an immobilization

matrix for producing reaction zones in dipstick-style immuno-

assays. BioTechniques 21, 918^925.

[15] Weiner, C., Saèra, M. and Sleytr, U.B. (1994) Novel protein A

a¤nity matrix prepared from two-dimensional protein crys-

tals. Biotechnol. Bioeng. 43, 321^330.

FEMSRE 586 28-10-97

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175172

[16] Weiner, C., Saèra, M., Dasgupta, G. and Sleytr, U.B. (1994)

A¤nity cross-£ow ¢ltration: puri¢cation of IgG with a novel

Protein A a¤nity matrix prepared from two-dimensional pro-

tein crystals. Biotechnol. Bioeng. 44, 55^65.

[17] Schuster, K.C., Mayer, H.F., Kieweg, R., Hampel, W.A. and

Saèra, M. (1995) A synthetic medium for continuous culture of

the S-layer carrying Bacillus stearothermophilus PV72 and

studies on the in£uence of growth conditions on cell wall

properties. Biotechnol. Bioeng. 48, 66^77.

[18] Kuhn, H., Friederich, U. and Fiechter, A. (1979) De¢ned

minimal medium for a thermophilic Bacillus sp. developed

by a chemostat pulse and shift technique. Appl. Microbiol.

Biotechnol. 6, 341^349.

[19] Pink, T., Langer, K., Hotzy, C. and Saèra, M. (1996) Regula-

tion of S-layer protein synthesis of Bacillus stearothermophilus

PV72 through variation of continuous cultivation conditions.

J. Biotechnol. 50, 189^200.

[20] Saèra, M., Kuen, B., Mayer, H.F., Mandl, F., Schuster, K.C.

and Sleytr, U.B. (1996) Dynamics in oxygen-induced changes

in S-layer protein synthesis from Bacillus stearothermophilus

PV72 and the S-layer de¢cient variant T5 in continuous cul-

ture and studies on the cell wall composition. J. Bacteriol. 178,

2108^2127.

[21] Sleytr, U.B. and Saèra, M. (1989) Structure with membranes

having continuous pores. US Pat. No. 4,886,604.

[22] Kuen, B., Saèra, M. and Lubitz, W. (1996) Heterologous ex-

pression and self-assembly of the S-layer protein SbsA of Ba-

cillus stearothermophilus in Escherichia coli. Mol. Microbiol.

19, 495^503.

[23] Kuëpcuë, S., Saèra, M. and Sleytr, U.B. (1991) Chemical mod-

i¢cation of crystalline ultra¢ltration membranes and immobi-

lization of macromolecules. J. Membrane Sci. 61, 165^175.

[24] Kuëpcuë, S., Saèra, M. and Sleytr, U.B. (1993) In£uence of co-

valent attachment of low molecular weight substances on the

rejection and adsorption characteristics of crystalline protein-

aceous ultra¢ltration membranes. Desalination 90, 65^76.

[25] Kuëpcuë, S., Sleytr, U.B. and Saèra, M. (1996) Two-dimensional

paracrystalline glycoprotein S-layers as a novel matrix for the

immobilization of human IgG and their use as microparticles

in immunoassays. J. Immunol. Methods 196, 73^84.

[26] Scheiner, O. and Kraft, D. (1995) Basic and practical aspects

of recombinant allergens. Allergy 50, 384^391.

[27] Binder, B.R. (1995) Physiology and pathophysiology of the

¢brinolytic system. Fibrinolysis 9, 3^8.

[28] Jahn-Schmid, B., Messner, P., Unger, F.M., Sleytr, U.B.,

Scheiner, O. and Kraft, D. (1996) Toward selective elicitation

of TH1-controlled vaccination responses: vaccine applications

of bacterial surface layer proteins. J. Biotechnol. 44, 225^

231.

[29] Smith, R.H., Messner, P., Lamontagne, L.R., Sleytr, U.B. and

Unger, F.M. (1993) Induction of T-cell immunity to oligosac-

charide antigens immobilized on crystalline bacterial surface

layers (S-layers). Vaccine 11, 919^924.

[30] Smith, R.H., Babiuk, L.H. and Stockdale, P.H. (1981) Intra-

nasal immunization of mice against Pasteurella multocida. In-

fect. Immun. 31, 129^135.

[31] Malcolm, A.J., Best, M.W., Szarka, R.J., Mosleh, Z., Unger,

F.M., Messner, P. and Sleytr, U.B. (1993) Surface layers from

Bacillus alvei as a carrier for a Streptococcus pneumoniae con-

jugate vaccine. In: Advances in Paracrystalline Surface Layers

(Beveridge, T.J. and Koval, S.F., Eds.), pp. 219^233. Plenum,

New York.

[32] Ebner, C., Schenk, S., Naja¢an, N., Siemann, U., Steiner, R.,

Fischer, G.W., Ho¡mann, K., Szeèpfalusi, Z., Scheiner, O. and

Kraft, D. (1995) Non-allergic individuals recognize the same

T-cell epitopes of Bet v 1, the major birch pollen allergen, as

atopic patients. J. Immunol. 154, 1932^1940.

[33] Secrist, H., Chelen, C.J., Wen, Y., Marshall, J.D. and Umet-

su, D.T. (1993) Allergen immunotherapy decreases interleukin

4 production in CD4

�T-cells from allergic individuals. J. Exp.

Med. 178, 2123^2130.

[34] Varney, V.A., Hamid, Q.A., Gaga, M., Ying, S., Jacobson,

M., Frew, A.J., Kay, A.B. and Durham, S.R. (1993) In£uence

of grass pollen immunotherapy on cellular in¢ltration and

cytokine mRNA expression during allergen-induced late-phase

cutaneous responses. J. Clin. Invest. 92, 644^651.

[35] Jutel, M., Pichler, W.J., Skrbic, D., Urwyler, A., Dahinden, C.

and Mueller, U.R. (1995) Bee venom immunotherapy results

in decrease of IL-4 and IL-5 and increase of IFN-gamma

secretion in speci¢c allergen-stimulated T-cell cultures. J. Im-

munol. 154, 4187^4194.

[36] Holt, P.G. (1994) A potential vaccine strategy for asthma and

allied atopic diseases during early childhood. Lancet 344, 456^

458.

[37] Ferreira, F.D., Ho¡mann-Sommergruber, K., Breiteneder, H.,

Pettenburger, K., Ebner, C., Sommergruber, W., Steiner, R.,

Bohle, B., Sperr, W.R., Valent, P., Kungl, A.J., Breitenbach,

M., Kraft, D. and Scheiner, O. (1993) Puri¢cation and char-

acterization of recombinant Bet v 1, the major birch pollen

allergen. J. Biol. Chem. 268, 19574^19580.

[38] Jahn-Schmid, B., Graninger, M., Glozik, M., Kuëpcuë, S., Eb-

ner, C., Unger, F.M., Sleytr, U.B. and Messner, P. (1996)

Immunoreactivity of allergen (Bet v 1) conjugated to crystal-

line bacterial cell surface layers (S-layers). Immunotechnology

2, 103^113.

[39] Ojcius, D.M., Gachelin, G. and Dautry-Varsat, A. (1996) Pre-

sentation of antigens from microorganisms residing in host-

cell vacuoles. Trends Microbiol. 4, 53^59.

[40] Livingston, P.O.L. (1995) Approaches to augmenting the im-

munogenicity of melanoma gangliosides: from whole melano-

ma cells to ganglioside-KLH-conjugate vaccines. Immunol.

Rev. 145, 147^166.

[41] Kuëpcuë , S., Saèra, M. and Sleytr, U.B. (1995) Liposomes coated

with crystalline bacterial cell surface protein (S-layer) as im-

mobilization structures for macromolecules. Biochim. Bio-

phys. Acta 1235, 263^269.

[42] Sleytr, U.B., Messner, P., Pum, D. and Saèra, M. (1996) Crys-

talline Bacterial Cell Surface Proteins. Appendix, pp. 211^225.

R.G. Landes/Academic Press, Austin, TX.

[43] Anonymous (1996) FDA and industry see complications in

combination vaccines. ASM News 62, 10^11.

[44] Jahn-Schmid, B., Siemann, U., Messner, P., Unger, F.M.,

Sleytr, U.B., Kraft, D. and Ebner, C. (1995) Crystalline bac-

terial surface layers (S-layers) as carrier-adjuvants for selective

FEMSRE 586 28-10-97

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175 173

TH1 vaccination responses. Applicability in the immunother-

apy of allergies? Immunobiology 194, 222.

[45] Scheller, F., Schubert, F., Pfei¡er, D., Wollenberg, U., Ren-

neberg, R., Hintsche, R. and Kuëhn, M. (1992) Biosensors:

fundamentals, technologies and applications. In: Fifteen

Years of Biosensor Research in Berlin-Buch (Scheller, F.

and Schmid, R.D., Eds.), Vol. 17, pp. 3^10. VCH, Weinheim.

[46] Hanke, W. and Schlue, W.-R. (1993) Physical properties of

biological membranes and planar lipid bilayers. In: Biological

Techniques Series (Sattelle, D.B., Ed.), pp. 9^22. Academic

Press, London.

[47] Schindler, H. (1989) Planar lipid-protein membranes: strat-

egies of formation and of detecting dependencies of ion trans-

port functions on membrane conditions. Methods Enzymol.

171, 225^253.

[48] Kagan, B.L. and Sokolov, Y. (1994) Use of lipid bilayer mem-

branes to detect pore formation by toxins. Methods Enzymol.

235, 691^705.

[49] Marty, A. and Neher, E. (1995) Tight-seal whole-cell record-

ing. In: Single-Channel Recording (Sakmann, B. and Neher,

E., Eds.), pp. 31^52. Plenum, New York.

[50] Milton, R.L. and Cardwell, J.H. (1990) How do patch clamp

seals form? P£uëgers Arch. 416, 758^765.

[51] Montal, M. and Muëller, P. (1972) Formation of bimolecular

membranes from lipid monolayers and a study of their elec-

trical properties. Proc. Natl. Acad. Sci. USA 69, 3561^

3577.

[52] Sackmann, E. (1996) Supported membranes: scienti¢c and

practical applications. Science 271, 43^48.

[53] Wonderlin, W.F., Finkel, A. and French, R.J. (1990) Optimiz-

ing planar lipid bilayer single-channel recordings for high res-

olution with rapid voltage steps. Biophys. J. 58, 289^297.

[54] Coronado, R. and Latorre, R. (1983) Phospholipid bilayers

made from monolayers on patch-clamp pipettes. Biophys. J.

43, 231^236.

[55] Zivman, M. and Tien, H.T. (1991) Formation of a bilayer

lipid membrane on rigid supports: an approach to BLM-

based biosensors. Biosens. Bioelectron. 6, 37^42.

[56] Tamm, L.K. and McConnell, H.M. (1985) Supported phos-

pholipid bilayers. Biophys. J. 47, 105^113.

[57] Sigworth, F.J., Urry, D.W. and Prasad, K.U. (1987) High-

resolution recordings show rapid current £uctuations in gra-

micidin A and four chemical analogues. Biophys. J. 52, 1055^

1064.

[58] Andersen, O.S. (1983) Ion movement trough gramicidin A

channels. Biophys. J. 41, 119^133.

[59] Zasadzinski, J.A., Viswanathan, R., Madsen, L., Garnaes, J.

and Schwartz, D.K. (1994) Langmuir-Blodgett ¢lms. Science

263, 1726^1733.

[60] Labarca, P. and Latorre, R. (1992) Insertion of ion channels

into planar lipid bilayers by vesicle fusion. Methods Enzymol.

207, 447^470.

[61] Pum, D., Weinhandel, M., Hoëdl, C. and Sleytr, U.B. (1993)

Large-scale recrystallization of the S-Layer of Bacillus coagu-

lans E38-66 at the air/water interface and on lipid ¢lms.

J. Bacteriol. 175, 2762^2766.

[62] Pum, D. and Sleytr, U.B. (1994) Large-scale reconstitution of

crystalline bacterial surface layer proteins at the air-water in-

terface and on lipid ¢lms. Thin Solid Films 244, 882^886.

[63] Hamill, O.P., Marty, A., Neher, E., Sakmann, B. and Sig-

worth, F.J. (1981) Improved patch-clamp techniques for

high-resolution current recording from cells and cell-free

membrane patches. P£uëgers Arch. 391, 85^100.

[64] Alvarez, O. (1986) How to set up a bilayer system. In: Ion

Channel Reconstitution (Miller, C., Ed.), pp. 115^129. Ple-

num, New York.

[65] Eray, A.R., Numan, D.S., Liu, L., Koch, A.R., Mo¡ett, D.F.,

Silber, M. and Van Wie, B.J. (1994) Highly stable bilayer lipid

membranes (BLMs) formed on microfabricated polyimide

apertures. Biosens. Bioelectron. 9, 343^351.

[66] Saèra, M. and Sleytr, U.B. (1996) Crystalline bacterial cell sur-

face layers (S-layers) : from cell structure to biomimetics.

Progr. Biophys. Mol. Biol. 65, 83^111.

[67] Hianik, T., Passechnik, V.I., Sargent, D.F., Dlugopolsky, J.

and Sokolikova, L. (1995) Surface potentials and solvent re-

distribution may explain the dependence of electrical and me-

chanical properties of supported lipid bilayers on applied po-

tential and bilayer history. Bioelectrochem. Bioenerg. 37, 61^

68.

[68] Seifert, K., Fendler, K. and Bamberg, E. (1993) Charge trans-

port by ion translocating membrane proteins on solid sup-

ported membranes. Biophys. J. 64, 384^391.

[69] Contino, P.B., Hasselbacher, C.A., Ross, A.J.B. and Nemer-

son, Y. (1994) Use of an oriented transmembrane protein to

probe the assembly of a supported phospholipid bilayer. Bio-

phys. J. 67, 1113^1116.

[70] Spinke, J., Yang, J., Wolf, H., Liley, M., Ringsdorf, H. and

Knoll, W. (1992) Polymer-supported bilayer on a solid sub-

strate. Biophys. J. 63, 1667^1671.

[71] Sleytr, U.B., Saèra, M., Messner, P. and Pum, D. (1994) Two-

dimensional protein crystals (S-layers) : fundamentals and ap-

plications. J. Cell. Biochem. 56, 171^176.

[72] Douglas, K. (1996) Biomimetic approaches to nanostructural

fabrication. In: Biomimetic Materials Chemistry (Mann, S.,

Ed.), pp. 117^142. VCH, New York.

[73] Rash, J.E. and Hudson, C.S. (1979) Freeze-Fracture: Meth-

ods, Artifacts, and Interpretations. Raven, New York.

[74] Nermut, M.V. (1983) The cell monolayer technique: an appli-

cation of solid phase biochemical and ultrastructural research.

Trends Biochem. Sci. 8, 303^306.

[75] Douglas, K., Devaud, G. and Clark, N.A. (1992) Transfer of

biologically derived nanometer-scale patterns to smooth sub-

strates. Science 257, 642^644.

[76] Douglas, K., Clark, N.A. and Rothschild, K.J. (1986) Nano-

meter molecular lithography. Appl. Phys. Lett. 48, 676^679.

[77] Michel, H., Neugebauer, D.-Ch. and Oesterhelt, D. (1980)

The 2-D crystalline cell wall of Sulfolobus acidocaldarius :

structure, solubilization, and reassembly. In: Electron Micro-

scopy at Molecular Dimensions (Baumeister, W. and Vogell,

W., Eds.), pp. 27^35. Springer Verlag, New York.

[78] Holland, B.W., Douglas, K. and Clark, N.A. (1994) Biolog-

ically derived nanometer-scale patterning on chemically modi-

¢ed silicon surfaces. Mat. Res. Soc. Symp. Proc. 330, 121^123.

[79] Gillis, H.P., Choutov, D.A., Martin, K.P. and Song L. (1996)

FEMSRE 586 28-10-97

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175174

Low energy electron-enhanced etching of GaAs(100) in a

chlorine/hydrogen DC plasma. Appl. Phys. Lett. 68, 2255^

2257.

[80] Gillis, H.P., Choutov, D.A., Steiner IV, P.A., Piper, J.D.,

Crouch, J.H., Dove, P.M. and Martin, K.P. (1995) Low en-

ergy electron-enhanced etching of Si(100) in hydrogen/helium

direct-current plasma. Appl. Phys. Lett. 66, 2475^2477.

[81] Gillis, H.P., Clemons, J.L. and Chamberlain, J.P. (1992) Low-

energy electron beam enhanced etching of Si(100)-(2U1) by

molecular hydrogen. J. Vac. Sci. Technol. B10, 2729^2732.

[82] Calvert, J.M., Georger, J.H., Peckerar, M.C., Phersson, P.E.,

Schnur, J.M. and Schoen, P.E. (1992) Deep UV photochem-

istry and patterning of self-assembled monolayer ¢lms. Thin

Solid Films 210/211, 359^362.

[83] Dressick, W.J., Dulcey, C.S., Georger Jr., J.H. and Calvert,

J.M. (1993) Photopatterning and selective electroless metalli-

zation of surface-attached ligands. Chem. Mat. 5, 148^150.

[84] Saèra, M., Kuëpcuë , S., Weiner, C., Weigert, S. and Sleytr, U.B.

(1993) Crystalline protein layers as isoporous molecular sieves

and immobilization and a¤nity matrices In: Immobilised

Macromolecules: Application Potential (Sleytr, U.B., Mess-

ner, P., Pum, D. and Saèra, M., Eds.), pp. 71^86. Springer

Verlag, London.

[85] Pum, D. Saèra, M. and Sleytr, U.B. (1993) Two-dimensional

(glyco)protein crystals as patterning elements and immobilisa-

tion matrices for the development of biosensors. In: Immobi-

lised Macromolecules: Application Potential (Sleytr, U.B.,

Messner, P., Pum, D. and Saèra, M., Eds.), pp. 141^160.

Springer Verlag, London.

[86] Neubauer, A., Pum, D. and Sleytr, U.B. (1993) An ampero-

metric glucose sensor based on isoporous crystalline protein

membranes as immobilization matrix. Anal. Lett. 26, 1347^

1360.

[87] Neubauer, A., Hoëdl, C., Pum, D. and Sleytr, U.B. (1994) A

multistep enzyme sensor for sucrose based on S-layer micro-

particles as immobilization matrix. Anal. Lett. 27, 849^865.

[88] Neubauer, A., Pum, D., Sleytr, U.B., Klimant, I. and Wolf-

beis, O.S. (1996) Fiber-optic glucose biosensor using enzyme

membranes with 2-D crystalline structure. Biosens. Bioelec-

tron. 11, 315^323.

[89] Pum, D. and Sleytr, U.B. (1996) Monomolecular reassembly

of a crystalline bacterial cell surface layer (S-layer) on un-

treated and modi¢ed silicon surfaces. Supramol. Sci. 2, 193^

197.

[90] Pum, D., Stangl, G., Sponer, C., Fallmann, W. and Sleytr,

U.B. (1997) Deep ultraviolet patterning of monolayers of crys-

talline S-layer protein on silicon surfaces. Coll. Surf. B Bio-

interfaces (in press).

[91] Pum, D., Stangl, G., Sponer, C., Riedling, K., Hudek, P.,

Fallmann, W. and Sleytr, U.B. (1997) Patterning of mono-

layers of crystalline S-layer proteins on a silicon surface by

deep ultraviolet radiation. Micro- Nanoeng. (in press).

[92] Calvert, J.M., Chen, M.S., Dulcey, C.S., Georger, J.H.,

Peckerar, M.C., Schnur, J.M. and Schoen, P.E. (1991) Deep

ultraviolet patterning of monolayer ¢lms for high resolution

lithography. J. Vac. Sci. Technol. B 9, 3447^3450.

[93] Douglas, K. and Clark, N.A. (1986) Nanometer molecular

lithography. Appl. Phys. Lett. 48, 676^678.

[94] Mann, S. (1995) Biomineralization and biomimetic materials

chemistry. J. Mat. Chem. 5, 935^946.

[95] Schultze-Lam, S., Harauz, S. and Beveridge, T.J. (1992) Par-

ticipation of a cyanobacterial S-layer in ¢ne grain mineral

formation. J. Bacteriol. 174, 7971^7981.

[96] Schultze-Lam, S. and Beveridge, T.J. (1994) Nucleation of

celestite and strontianite on a cyanobacterial S-layer. Appl.

Environ. Microbiol. 60, 447^453.

FEMSRE 586 28-10-97

U.B. Sleytr et al. / FEMS Microbiology Reviews 20 (1997) 151^175 175