In situ determination of enzyme monolayer activity
J.J. Ramsden (Biozentrum, Basel) and M.G. Cacace (CNR, Siena)

The fabrication of plates coated with enzymes is considered to be of great importance, e.g. for biosensors in which a monolayer of receptor proteins is attached to a transducer able to convert the presence of captured ligand into an electrical or optical signal. Although the number of enzymes per unit area can nowadays be determined with good accuracy, very little information on the activities of adsorbed enzymes is generally available. This contribution describes a method for determining adsorbed enzyme activity (turnover number)

Urease is adsorbed to a metal oxide plate and exposed to its substrate urea in the presence of a pH indicator. The plate is a planar optical waveguide, and optical waveguide lightmode spectroscopy (OWLS) is used to precisely monitor the buildup of the adsorbed protein monolayer. The lightmode spectrum depends sensitively on the refractive index profile in the vicinity of the the waveguide surface, i.e.\ within the evanescent field created by the guided light. By choosing an indicator whose color change lies in the vicinity of the wavelength l of the guided light, the change in the so-called ``anomalous" or selective dispersion at l can be very much greater than the corresponding change in optical density. Since the system is well defined both optically and hydrodynamically, the enzyme activity can be calculated from the measured shift in the lightmode spectrum during substrate decomposition.

The method can be used to determine the activity of any adsorbed or surface-immobilized enzyme whose substrate yields a pH change upon decomposition. The use of tunable wavelength sources would further generalize the approach, enabling it to encompass any substrate transformation involving an appropriate change of color.

A surface engineering approach towards the development of cell-based biochips
S. Makohliso
(Centre for Gene Therapy \& Division of Surgical Research, CHUV, Lausanne)

For investigating the electrical activity of neurones, the patch clamp is highly sensitive, but it is invasive and can only address one cell at a time. A biochip is noninvasive and can address many cells. The biochip strategy is to pattern a smooth planar substrate (Si) with teflon and gold. Selected amino acid sequences grafted onto the gold promote cell adhesion, whereas the teflon inhibits it. The electrical circuit is completed by placing a gold ring around the neurone. The main problem is that such biochips are typically one thousand times less sensitive than the patch clamp, probably due to poor adhesion between the neurone and the gold substrate. The origin of the observed signals is ascribed to capacitance (C) changes. Nevertheless, the link between C and (for example) the opening of Ca$^{2+}$ channels by glucose is not clear at present, and may be obscured by the possibility that cells adapt to their new, immobilized, environment by changing the expression of membrane channel receptors.

Functionalized surfaces for biomicroanalytical assays
G.A. Urban
(Institut fur Mikrosystemtechnik, Albert-Ludwigs-Universitát, Freiburg-i.-Br.)

The `lab on a chip' promises new tools for metabolic investigations and the determination of physical parameters characterizing the health of a patient. In contrast to genomics (knowledge of the DNA) and proteomics (knowledge of the proteins), function (what reactions are facilitated and how different enzymes interrelate) is emphasized here.

First steps towards realizing this goal are the development of multisensor extracellular probes. As well as intrinsic sensor performance, the performance of the system as a whole and its interface with humans have to be considered. The effective operation of sophisticated biomolecule receptor arrays often depends crucially on proper water management in the environment of the enzyme.

An online toxicological sensor based on the optical waveguide technique
J. Vörös
(Laboratory for Surface Science and Technology, ETH Zurich)}

Initially spherical osteoblasts are deposited on smooth planar titania surfaces. They adhere and spread within 1--2 h, as quantified by confocal laser scanning microscopy (CLSM), which requires that the cell membranes are fluorescently labelled, and by in situ optical waveguide lightmode spectroscopy (OWLS), as described by Ramsden, J.J., Li, S.-Y., Heinzle, E. & Prenosil, J.E., An optical method for the measurement of number and shape of attached cells in real time, Cytometry {\bf 19} (1995) 97--102. A model toxin (sodium hypochlorite) causes the cells to bead up and become detached.

Adsorption of serum proteins on modified metal oxide surfaces measured by optical waveguide techniques
G. Kenausis
(Laboratory for Surface Science and Technology, ETH Zurich)}

Waveguide surfaces can be readily modified by block copolymers or grafted dendromers. This is a robust and extremely versatile method of changing the balance of repulsive and attractive forces at the protein--surface interface, and hence increasing or decreasing the amount of protein adsorption. OWLS is a powerful and sensitive technique for quantifying the thickness of the polymer coating, and the amount of subsequent protein adsorption.

Multilayer protein assemblies at solid/liquid interfaces
E. Brynda and M. Houska
(Institute of Macromolecular Chemistry, Prague)}

One problem with the classic biosensor format comprising a monolayer  of receptors coating a transducer is that the change in whatever   attribute is being transduced (mass, polarizability etc.) is proportional  to analyte size, and in consequence small analytes have much higher  detection limits than large ones. A strategy to overcome this limitation is to  coat the transducer with receptor multilayers. Small analytes can  readily penetrate the outer layers and the total change equals monolayer  response multiplied by the number of layers.

To fabricate such structures the first protein (receptor) layer is adsorbed at a pH such that its electrostatic charge is opposite from that of the substrate. Next, a polyion of opposite charge (i.e.\ the same as that of the original substrate) is adsorbed, followed by a second protein layer, and so on. Finally the proteins are crosslinked with glutaraldehyde, and by simply raising or lowering the pH the polyion `scaffold' is expelled, leaving a porous protein multilayer. The pore cavity size can be controlled via the size of the polyion.

The interaction of myelin basic protein with synaptic membranes
D. Mikeladze
(I.S. Beritashvili Institute of Physiology, Tbilisi)

The creation of the myelin sheath around the neurones is the final stage in the development of the central nervous system (CNS). It ensures the high resistance and low capacitance necessary for efficient transmission of the nervous impulse. Myelin plays a very important r\^ole in cell-cell interactions of the CNS, and most neural inflammation and disease involves some demyelination. Myelin basic protein (MBP) comprises 50\% of the myelin. Hence the detailed biophysical characterization of MBP and its interactions should offer the key to understanding many neural disorders.

Mapping the radial distribution function of adsorbed particles: results and biological significance
Ph. Lavalle
(INSERM, Strasbourg and Biozentrum, Basel)

Particles deposited on a surface are characterized principally by:
(i) fluctuations in the number of adsorbed particles);
(ii) the radial distribution function (rdf) of the particles;
(iii) their jamming limit; and
(iv) the kinetics of particle deposition.

This contribution describes experimental work in which monolayers of negatively charged latex particles of different radii were deposited onto positively charged glass slides. The deposits were photographed and the images quantified, enabling the rdf to be determined.

The particles' spatial distributions could be quantitatively modeled using the extended random sequential adsorption (ERSA) model:
(i) a position for attempting adsorption of a new particle is chosen at random;
(ii) if the new particle would not overlap with any previous ones, it is deposited and fixed in place.
(iii) if it does overlap, up to $n_s$ further deposition attempts are made, along the line joining the centers of the new and the overlapped particles, up to a maximum center-to-center distance $d_s$.

The model was fitted to the experimental data to determine $d_s$ and $n_s$. The variation of these parameters with particle radius $r$ revealed that with increasing $r$, $d_s$ decreased and $n_s$ increased (approaching the ballistic deposition (BD) model), and with decreasing $r$, $d_s$ increased and $n_s$ decreased (approaching the random sequential adsorption (RSA) model).

The investigation of particulate layers using the film balance
Z. Horvolgyi
(Department of Physical Chemistry, Budapest Technical University)

Partially wettable particles can be spread on the surface of an aqueous subphase to form a floating monolayer. If these particles are (weakly) cohesive, their behavior upon compression and expansion is much more complex than that of the more familiar repulsive particles. It could be shown that a number of features of the compression-expansion isotherms are related to the advancing or receding contact angles between the particle material with the subphase liquid.

Particles were imaged on the monolayer using Brewster angle microscopy (BAM),
or transferred to solid substrates using the Langmuir-Blodgett (LB) technique and examined using a variety of optical methods.

The application of percolation theory and fractal concepts in tablet manufacture
H. Leuenberger
(Department of Pharmacy, Basel University)

The percolation threshold (i.e.\ the particle density at which a spanning cluster joining all particles in the system appears) is an important parameter controlling the properties of particulate aggregates. Whereas in two dimensions there is a single threshold, in three dimensions there are two and in the transition region between them both sublattices (those of the particles and of the material in which they are embedded) percolate, i.e.\ are continuous.

In the vicinity of a threshold $p_c$ many important physical parameters $x$ (e.g.\ mean cluster size) follow a power law $x \sim \mid p - p_c \mid^\gamma$  where $p$ is the lattice occupancy (density). Many of the exponents $\gamma$ are universal, and do not depend on the details of the system.

The Cayley tree (Bethe lattice) has the advantage that it can be manipulated analytically relatively easily, and many exact results are known. A lattice coordination number $z$ appropriate to the system under examination can be estimated from the measured porosity of a particulate aggregate, and the corresponding Bethe lattice then used to predict features of the real system.

Modification of biological surfaces via adsorption of multifunctional polyelectrolytes
J.A. Hubbell
(Institute for Biomedical Engineering and Department of Materials,
ETH Zurich and University of Zurich)

Tissue engineering, of which wound healing is a part, is steeped in the notion of communication. 
One may identify:
(i) silence---blocking both diffusible and immobilized elements;
(ii) touching---immobilized bioactive elements can interact;
(iii) speaking---diffusible elements interact; and
(iv) listening---responding to a cell-derived signal.

Different problems demand different approaches. For example, the arterial healing cascade must be blocked in the case of therapeutic injury. A useful blocking material is lactic acid (LA)-PEG copolymer. This is introduced as PEG-LA-Ac units into the blood (Ac = photopolymerizable acrylate). In the presence of dye and a sensitizer, polymerization is initiated by light to produce a knotted polymer. The LA later degrades, and the PEG is washed out.
In situ polymerization is necessary to ensure that the (gel-like) polymer is conformal with uneven tissue surfaces and can interpenetrate with the extracellular matrix.

Other sophisticated materials are coated with oligopeptides (typically 10 fmol/cm^2---about 0.1\% of a monolayer) recognized by cell surface receptors. This allows one to selectively control adhesion to different cell types.

Pathogenesis of implant-associated infections
W. Zimmerli (Abt.\ f\"ur Infektiologie, Kantonsspital Basel)

Implants are highly susceptible to infection by bacteria, which are firmly adherent and resistant to antibiotics. The rest of the body is unaffected, however, and the infection rapidly clears once the implant is removed. The bacteria are not antibiotic resistant if removed and cultured, and any bacteria escaping into the blood are cleared by the host's defence mechanisms. The implant gives the appearance of promoting a local immunodeficiency. The phenomenon is currently not understood.

The surface of biomolecules:\ how their topology accounts for their biological role and biophysical properties
M.G. Cacace
(Centre for the Study of Germinal Cells, CNR, Siena)

A mesoscopic approach is emphasized. Proteins may be placed on a solvatochromic scale, which measures to what degree their surfaces compete for hydration water. This provides a coherent basis with which many recognition phenomena can be understood (see also M.G. Cacace, E.M. Landau \& J.J. Ramsden, The Hofmeister series:\ salt and solvent effects on interfacial phenomena}, Q. Rev. Biophys. 30 (1997) 241--278).

Measurement of nanomechanical parameters at interfaces using atomic force microscopy
N. Rozlosnik (Department of Biological Physics EOtvös University, Budapest)

Proteins are typically soft (compliant) amd adhesive, unlike the classical materials of engineering. Penetration and retraction of the tip of an atomic force microscope into a protein multilayer allows the mechanical properties of the protein to be characterized.
Initially the tip moves as in a viscous liquid, and then upon further penetration an elastic force is experienced. Upon retraction, viscous damping occurs.

On a possible mechanism of crystallization in lipid mesophases
P. Nollert (Biozentrum, Basel)

Lipidic cubic phases are infinite periodic minimal surfaces, characterized typically by two principal curvatures. They have shown remarkable promise in facilitating the crystallization of integral membrane proteins. In order to extend and generalize the method, effort is being focused onto the effect of the components (eight!) present in a protein-lipid mixture on the phase behavior, the identification of the phases present during crystallization, and the role of the lipid.

It is proposed that the cubic phases promote crystallization by firstly concentrating the protein monomer via salt-induced contraction, sufficiently to initiate nucleation. The rest of the lattice then provides an ample reservoir of protein for crystal growth.

Summing up
J.J. Ramsden (Biozentrum, Basel)

There has been an enormous diversity of specific topics covered in this conference, and it has been truly enriching to follow the unfolding of a rich web of cross connexions and mutual illumination during these past two days, during which, to an even greater extent than anticipated, most of the themes touched on most of others.

One of the aims was simply to disseminate novel experimental techniques which have been developed by specialists in one field, but remain unknown to specialists in others, although they have the potential to greatly enhance the armoury of methods of bio-physico-chemical characterization! This aim has been accomplished, and the results are largely self-evident, so I shall not expound further on them here, but rather confine myself to the conceptual strands.

One such strand starts with the molecular basis (Cacace) of the interactions between proteins, cells and synthetic materials (Hubbell, Kenausis, Nollert). Of no less importance for materials destined for use in the body are their mechanical properties (Rozlosnik). Understanding these interactions should have far reaching implications for the pathology of the central nervous system (Mikeladze), and the infection of artificial implants (Zimmerli). The sensors developed for investigating the response of cells to their environment (Mahkoliso, Vörös) will also be exceedingly useful for elucidating these interactions.

Another strand starts with the basis of how assemblies of particles interact with one another based on their purely geometrical properties (Lavalle, Leuenberger). Only when the geometrical aspects are understood can one unravel the parts of observed behaviour due to specific chemical interactions (Cacace). The importance of the latter has been unambiguously demonstrated by Hórvolgyi, and the fabrication of composite multilayers (Brynda, Houska) makes use of both geometrical and chemical properties.

These two strands become intertwined, and an important take-home message is that sometimes it is best to consider proteins and cells  particles with purely geometric properties, whereas under other circumstances their specific chemical and biological functionality dominates their behavior.