INTRODUCTION
Mucus is a hydrated gel network that coats the luminal surfaces of epithelia throughout the body. It varies greatly in composition, quantity and thickness depending on its localisation and function. The common feature of these gels is the presence of one or more molecules we now term mucins. These molecules are large glycoproteins, their most characteristic features being the presence of tandemly repeated peptide domains rich in serine and/or threonine and proline. Typically more than 60% of their mass is accounted for by O-linked oligosaccharides. This is a growing family of gene products that has some 10 members at present.
Currently we distinguish between molecules which remain tethered to the cell surface and contribute to
the glycocaelix and those which are secreted to form the gel matrix of mucus. It is now clear that the
rheological properties of mucus are largely due to the physical properties of these latter mucin molecules
and their interactions. They are particularly characterised by their large mass (5-40x10^6), physical size
(500-5000nm) and polydispersity.
This polydispersity appears to arise naturally due to the linear, disulphide bond mediated assembly of the
molecules from a variable number of subunits (2-3x10^6). These subunits are distinctive in having heavily
glycosylated domains interspersed with partially glycosylated or unglycosylated domains.
The mechanisms of assembly are not completely clear yet but the evidence suggests a strong similarity with
the molecule called von Willebrand Factor (vWF) which dimerises with a C-C terminal reaction and
subsequently tetramerises with an N-N terminal association.
A number of gel forming mucins have been found to have strong homology in their N and C termini to the
so-called D-domains of vWF which have been shown to be responsible for its assembly. The different
glycosylation of mucin core proteins can give rise to many different molecules and currently it is a
central theme in mucin research to understand the nature and biological functions of mucins in mucus gels
in terms of the mucin gene products present, their glycosylation and state of oligomerisation.
This problem is particularly relevant in the gastrointestinal tract where there is a diverse array of
mucins localised in specific cells that are under defined secretory control.
ELECTRON MICROSCOPY
Electron microscopy is a valuble tool for the physical characterisation of large biomolecules such as mucins. Using this method we definitively demonstrated that mucin subunit assembly was based upon a linear architecture. Subsequently we have used various techniques of high resolution microscopy to study the number and weight average size distribution of these molecules, the number and organisation of their subunits, the presence of glycosylated and "naked" protein domains and to identify the presence of specific epitopes on subpopulations of the mucins that might be present in a mixture.
It is also possible by quantitative dark field microscopy to study their mass/length. We have employed two
major methods in our work
a) monolayer spreading techniques that have been adapted directly from the DNA field
b) replica shadowing methods that are commonly used for high resolution imaging of all kinds of
biomolecules.
MONOLAYER SPREADING METHOD
Four steps underlie the application of the monolayer method. They are :
Preparation of mucins
This is an extensive topic, too large to be dealt with here in detail. Briefly we prefer the application
of isopycnic density centrifugation as it yields the molecules free of lipids, proteins and nucleic acids
at the 1% level. It is gentle and can be performed with strict control of protease activity, which is
important for ascertaining the native size distribution of the molecules.
Preparation of carbon coated grids
Thin carbon films (2-5nm) are produced by the indirect evaporation of carbon onto 2cm square blocks of
freshly cleaved mica surfaces. The mica is left in a humid environment for approximately 1 hour and the
film is subsequently removed by floating it off on the surface of clean water in a petri dish.
A rubber O-ring of 2.5cm diameter is gently lowered onto the floating carbon film and allows the intact
film to be steered on the water surface over the EM grids (15-25 400-600 mesh) that have previously been
placed at the bottom of the dish on a wire mesh.
The water level is lowered by suction and the carbon film deposited on the grids.
The grids are dryed in an oven at 60° for 2 hours.
Deposition of mucins on grids
The spreading method was originally developed by Kleinschmidt using cytochrome C as the spreading agent.
This method was modified by Lang et al and Koller et al for the improved imaging and analysis of DNA.
The basis of the method is the creation of a protein or lipid bilayer in which the long filamentous
molecules are entrapped and thereafter can be removed onto the surface of the grid.
We prefer the diffusion/adsorption based geometry as it requires little material and use the spreading
agent benzyldimethylalkylammonium chIoride (BAC) as introduced in ref 4. We have perfomed this procedure
in a wide variety of solution conditions including 6M guanidinium chloride and find it very tolerant,
though the presence of extraneous detergents or lipids or large amounts of small proteins can be a problem
as they interfere with the BAC monolayer.
Staining and/or shadowing
The grids are touched to the surface film and thereafter washed for a few seconds in 95% ethanol.
Mucins are typically polyanions rich in sialic acid and/or sulphate and thus can be positively stained by
washing the grids in ethanolic uranyle acetate. For generating high molecular contrast the molecules may
also be rotary shadowed with heavy metals such as platinum or tungsten.
Advantages and disadvantages
This is a simple, rapid, salt tolerant and sensitive technique requiring very little material.
The molecules are actually present on the grid unlike replica methods and thus it is possible to stain
the grids with other reagents such as antibodies before they are contrasted and, in this way, highlight
specific domains on the molecule.
However the method is dependent on the surface properties of the grids and is demanding in the cleanliness
of water and reagents. Thus the results may on occasion become non-reproducible due to changes in the
experimental milieu that are not easy to detect or analyze.
b) A micro version of the same procedure can be arranged by adding the BAC to the mucin in solution and
transferring 40µl drops to a teflon surface. Within minutes the BAC forms a monolayer on the solution
surface in which the mucin molecules become trapped. The surface film is touched to the carbon coated grid
as above.
a) A solution of mucins (typically about 10 ml) at a concentration of 0.01-0.1 µg/ml in any aqueous
solvent is poured into a teflon trough or small petri dish. A drop (1µl) of a solution of BAC (100 µg/ml)
is touched to the surface and the solution left for 5-15 minutes. In this time the mucins diffuse to the
surface and become entrapped in the BAC monolayer. A carbon coated electron microscope grid (400-600 mesh)
is touched to the surface and thereafter washed in 95% ethanol, dried and rotary shadowed
REPLICA METHOD
This method is also commonly called "rotary shadowing". However this procedure is not the essential principle of the method. There are many variants of this method that are current in different laboratories. We typically use a modification described by Mould et al because we find that it minimises fragmentation of very large molecules that can take place using the more common drop nebulisation method. A drop of solution (20µl) containing the mucins at concentrations from 1µg/ml to .01µg/ml depending on size are put on a 2cm square of freshly cleaved mica. The two clean surfaces are again joined and left together for a few minutes. The mica sandwich is placed in a beaker of 0.2M ammonium acetate and the two sheets separated under the solution and left to wash for 1-10 minutes. The two mica sheets are removed from the beaker, plunged into liquid nitrogen, allowed to cool and then placed face up on a copper block previously cooled in liquid nitrogen. The block is placed in an evaporation unit which is pumped down, essentially freeze drying the molecules on the mica. When dry, the mica is rotary evaporated with platinum at an angle of 5°-10° and thereafter a thin layer of carbon 10nm is evaporated onto the platinum surface. The mica is stored overnight in a dessicator containing 0.1M acetic acid and the carbon replica is removed the next day onto a clean water surface and transferred onto grids as described above.
Advantages and disadvantages
The method is reliable, efficient, reproducible in experienced hands, yields high resolution data and is
tolerant of a variety of other molecules and salts in the starting solution. However preparation is rather
a lengthy process and it can take some time to search the appropriate concentration conditions for
studying your molecules.
SUMMARY AND PERSPECTIVES
Electron microscopy gives a unique view of the architecture and microstructure of the mucins and will in the future be employed to study the organisation of oligosaccharides on the protein core and their interactions with specific proteins in the mucus gel. The method can yield powerfuI, quantitative and qualitative information about the heterogeneity of mucin species present, their number and weight size distribution, mass per unit length and interactions with other proteins. To date these procedures have been employed in specialist biochemical laboratories but in principle the methods are not complicated and could be disseminated more generally. The improvement in the rapidity and simplicity of biochemical operations would now make it possible to assay by electron microscopy the mucins present in biopsy quantities of material taken for diagnosis. Thus in the future these methods of molecular analysis may have a role to play in digestive pathology.