The technique of X-ray microanalysis has existed for considerably longer than is, perhaps, generally realised, having been invented in 1951 and first applied to biological material at the beginning of the 1960s (Hall, 1986). Much very useful pioneering work was carried out using wavelength-dispersive detectors. However, there can be no doubt that the introduction of the energy-dispersive detector based on the Si(Li) detector crystal, with its capability of multi-element analysis, and ease of operation led to the technique becoming more widely accepted. For the detection of the majority of elements the ease of use of the EDS detector offsets its lower resolution, and both the continued improvements in the Si(Li) detectors and the introduction of germanium detectors, will help to improve the sensitivity of the technique.

Biological specimens are many and varied with their own special problems for quantification, and perhaps the second most important development in X-ray microanalysis for biologists has been the commercial availability of software for carrying out quantitative routines. The procedure most commonly used for quantification is the continuum normalisation method of Hall (Hall, 1971). With the advent of the microcomputer in the early 1980s the first software for applying the Hall technique to biological specimens became available, and now the importance of quantification of spectra from biological specimens is recognised with all major detector manufacturers offering this option with their software.

Electron probe X-ray microanalysis (EPXRMA) is only one of a large number of techniques that are available for investigating the chemical composition of cells and tissues. These techniques use a wide variety of physical and chemical principles, and each has its own advantages and disadvantages (see Table B1). Some are non-microscopical and, at best, give information only at a rather coarse level of spatial resolution, although their chemical sensitivity may be very high. Among the microscopical techniques, many should be regarded as highly specialised, in the sense that the equipment is not easily available, and in some cases is not produced commercially. Other methods, such as histochemistry and autoradiography, may be regarded as routine, but do not, on the other hand, yield the same type of information as XRMA and other techniques. Indeed, while all the techniques listed in Table B1 can be used to study the chemical composition of cells or tissues, it seems that they are complementary rather than competing. For example, histochemical procedures can identify compounds, such as enzymes, which are not generally accessible by the other methods. As well as autoradiography, techniques such as SIMS (Thellier, Ripoll & Berry, 1991) and LAMMA, which can identify isotopes, are well adapted for the study of dynamic processes using tracers. Of the techniques considered in the following two chapters, particle-induced X-ray emission (PIXE) is most similar to EPXRMA, yet nevertheless shows important differences in specimen penetration, resolution and sensitivity.

Introduction

In 1949, Castaing and Guinier combined for the first time the technique of electron optics and X-ray analysis for the production of the first electron probe analyser (Castaing & Guinier, 1949). Developments continued during the years from a static beam instrument into a scanning device. However, all the analysis was initially limited to thick specimens and it was only at a later date that the first prototype microanalyser was specifically developed for thin specimens (Duncumb, 1962). The detection and counting of X-rays was done by using conventional wavelength dispersive crystal spectrometers. A major breakthrough in X-ray analysis came in 1968 with the development of the solid state energy dispersive X-ray detector (Fitzgerald, Keil & Heinrich, 1968). This type of detector provided higher collection efficiency than the wavelength dispersive spectrometer as well as more reproducible results because mechanical movements were no longer necessary.

Nowadays X-ray microanalysis is a well used technique for analysis of biological material in combination with the transmission and scanning electron microscopes. In this chapter new developments on the energy dispersive detectors will be discussed. Special attention will be given to the advantage of using a high purity germanium crystal instead of Si(Li) crystal in the EDS detector.

X-ray generation

When a focused electron beam hits a specimen several processes occur at the surface and in the specimen. In Fig. 1.1 a schematic presentation of the different interactions is presented.

Introduction

X-ray microanalysis (XRMA) is a young science, but in the past 20 years it has produced major advances in the material and biological sciences. This is particularly the case for the study of the uptake of pollutant heavy metals by animals and plants.

Before the arrival of XRMA it had been established that marine animals were capable of taking-up metals in large amounts and concentrating them in particular tissues. These effects were investigated by atomic absorption spectrophotometry and to a limited extent by histochemistry. However, structural resolution was limited and most specimen preparation procedures either removed the metals from the tissues or produced some degree of translocation. These restrictions did not assist investigation into the cytology of metal metabolism. This situation changed dramatically with the arrival of XRMA which coupled the structural resolution of the electron microscope with the element identification of the energy dispersive X-ray detector. Early applications to marine organisms were described by Nott & Parkes (1975) and Walker et al. (1975a,b).

The metals that are found in tissues can be grouped biologically into those essential for metabolism and those which are non-essential. Most metals are toxic when they occur in excess amounts. Chemically the metals can be grouped into the hard acid type (for example Na, Mg, K and Ca) which bind electrostatically with a preference for O > N > S donating ligands, and the soft acid type (for example Cu, Cd, Hg and Ag) which bind covalently with a preference for S > N > O.

Introduction

Electron probe microanalysis is a sensitive tool to localise elements in biological cells and tissues. In comparison to most specimens studied by this method in materials sciences, cells are not static objects, but vary in their elemental composition, depending on their functional state. Therefore, intracellular components, in particular diffusible ions, have to be immobilised in the defined functional state to be investigated. Rapid freezing, also called cryofixation, is the most promising approach to meet this requirement. This chapter points to the crucial role of appropriate freezing techniques for biological electron probe microanalysis. This is in particular important for X-ray microanalytical studies of the following biological features: (1) intracellular element compartmentation, (2) cell viability and membrane damage, (3) ion transport systems and (4) ion shifts related to dynamic processes in cells.

Preparation paths for electron probe microanalysis

Cells and tissues to be studied in an electron microscope have to be converted into a solid state specimen which is compatible with vacuum. Chemical preparation methods established for morphological investigations are based on fixation with aldehydes, staining by heavy metal salts, dehydration by alcohol and embedding in resin. Thereby, diffusible substances such as electrolyte ions are re-distributed and washed out (Zierold & Schäfer, 1988). As an alternative, low temperature preparation protocols as sketched in Fig. 7.1 were developed. They all start with cryofixation or with specimen sampling which means the appropriate handling of the specimen before rapid freezing to ensure the arrest of the functional state of interest.

Specimen preparation is the cornerstone of successful biological X-ray microanalysis. A damaged or compromised specimen cannot be ‘recovered’ by sophisticated analytical hardware or software, or even by a skilled operator. But what constitutes a ‘good specimen’ or an ‘acceptable degree of damage’? Is there such a thing as ‘the ideal preparative procedure’? Answers to these fundamentally important questions are explicitly or implicitly offered in the chapters in this section of the book. However, two general observations should be made at this juncture.

First, there is no single ideal preparative procedure that meets all the requirements of the biological electron probe X-ray microanalyst (even if we confine our considerations to thin specimens of soft tissues). The starting point must be the nature of the biological problems that are to be tackled. Once this has been clearly defined then a suitable preparation procedure may be adopted, but even then the question (and thus the analytical objectives) may need to be modified in the light of the anticipated level of preparative damage. Measuring electrolyte concentrations in a small cohort of distinctive cells lying some distance beneath the surface of a complex tissue is a good example of a ‘real microanalytical problem’, as distinct from the analysis of well-chosen model systems. Electrolyte analysis requires cryofixation to limit redistribution, but the quality of freezing will not be high in deep cells. So the prospect of analysing compartments finer than ‘nucleus’ and ‘cytoplasm’ in such preparations may be unrealistic.

Introduction

Perspective: directly deposited microdroplets

Electron probe X-ray microanalysis (EPXMA) is capable of yielding quantitative data from excited specimen volumes < 1 μm3. Thus, the analysis of fluid microvolumes in the range 10−11 to 10−10 litres is comfortably within the scope of this technique. For this reason, EPXMA was advocated at an early stage in its technological evolution for the analysis of nanolitre fluid volumes (Ingram & Hogben, 1967). Subsequently a number of laboratories have independently pursued physiological questions by the analyses of diverse biological fluids. This family of related microdroplet methods is described fully elsewhere in this volume by Roinel and Rouffignac (Chapter 11). However, in order that the sprayed microdroplet technique be seen in a realistic perspective a brief resumé of the more conventional ‘direct deposition’ microdroplet methods is provided here.

In general, two major variants of the microdroplet method are practised (Fig. 10.1).

1. 1. Microdroplets of known volume. Microdroplets of known volume are dispensed from calibrated micropipettes, mounted on polished solid (bulk) support materials, then dehydrated by freeze-drying to produce the all-important microcrystallites that minimise X-ray absorption. They are then individually analysed by wavelength-dispersive spectrometry (WDS).

2. 2. Microdroplets of unknown volume. Microdroplets of samples and standards of indeterminate volume are dispensed from constant-volume micropipettes, mounted on thin-film supports, then dehydrated by flash evaporation and subsequently analysed by energy-dispersive spectrometry (EDS). A selected list of the combinations of available options adopted by some of the major laboratories in this specialised field is presented in Table 10.1.

Although not ideal for making a major subdivision of bacteria, Gram's method of staining is convenient because everyone uses it (in one of its many modifications) when characterizing a bacterium. There is an unfortunate tendency to omit this step, especially when dealing with cultures isolated on selective media which, by virtue of the inhibitory agents they contain, may in theory affect the bacterial staining reactions; it is simply assumed that colonies on selective media have the appropriate (or expected) tinctorial and shapely qualities. This may often be tolerated (and sometimes acceptable) in busy routine laboratories where, in the words of W.S. Gilbert, it may be regarded as ‘merely corroborative detail, intended to give artistic verisimilitude to an otherwise bald and unconvincing narrative’. However, our objectives are neither speed nor artistry; we want to know the identity of the organisms and for this we need accurate characterizations. We cannot afford therefore to omit making smears and staining them by Gram's method, even if we do not pursue microscopy any further.

Division into major groups

Before dealing with the Gram-negative bacteria proper, we would remind readers that there are a few organisms that seem to be on the borderline between the Gram-positive and the Gram-negative, for example Gemella, a genus which until recently was thought to consist of Gram-negative cocci. Other bacteria show an unusual phenomenon in that they develop some Gram-positivity as cultures age, in contrast to the Gram-positive bacteria, which usually become Gram-negative as the cells age and degenerate.

It goes without saying that adequate clinical information is essential to enable appropriate culture media to be chosen for primary isolation. We believe that the subsequent investigation of the organism(s) thus isolated should be approached with an open mind; the organism(s) should be regarded as unknown and the process of identification started from the basic (primary) characters.

Theory of identification

In theory the identification of a bacterium consists of a comparison of the unknown with the known, the object being the ability to say that the unknown is like A (one of the known bacteria) and unlike B–Z (all other known bacteria); another and arguably equally important objective is to be able to say that the unknown organism is the same as A and thus to give it a name or identification tag. When we say that it is A we imply that it is different from all the other known bacteria, B–Z. All identification schemes depend on knowing a great deal about the already identified (or known) units, but the human memory can cope with only a small proportion of this knowledge, so memory aids are essential. In practice there are at present two distinct methods of making the identification; but the feasibility of a third method using a computer, first suggested by Payne (1963), was demonstrated by Dybowski & Franklin (1968) and is discussed in Chapter 10.

The first method is familiar to all biologists and uses the dichotomous key.

GENERAL CONSIDERATIONS

The ability of some bacteria to grow on media that are inhibitory to other bacteria gives the former organisms additional special characteristics that may help in their identification. The media themselves, with their selective and differential qualities, are of particular interest in diagnostic bacteriology and for this reason we have sometimes included a few notes on the isolation of certain bacteria and have referred occasionally to enrichment and other media used in culturing organisms from clinical material. We emphasize, however, that this Manual is concerned with how to identify organisms that have been isolated and not with how to isolate them.

In this appendix, media are listed under the following section numbers and headings:

1. A2.1 Basic media

2. A2.2 Enriched and enrichment media

3. A2.3 Differential and selective media

4. A2.4 Media for enhancing pigment production

5. A2.5 Media for carbohydrate studies

6. A2.6 Miscellaneous media

Cleaning and sterilization of glassware

Glassware for media such as Koser's citrate, in which there is a single source of an element such as carbon or nitrogen must be chemically clean to be free from it. A recommended method of ensuring this is to boil all tubes in 20% nitric acid for 5–10 minutes and then wash and rinse well with with glass-distilled water. Tubes are dried in an oven in the inverted position in baskets lined with filter or blotting paper to prevent the mouths of the tubes touching the metal.

For over 25 years now, medical bacteriologists all over the world have turned to ‘Cowan and Steel’ as their first reference book when they encountered an unfamiliar bacterial isolate. A generation of laboratory workers has grown up with it. They turned to it not only because there was clear information on how to examine isolates, with concise details of culture media and test methods that were applicable to the great majority of bacteria of medical importance, but also because of the famous successive tables that led from genera with their minidefinitions to species with their characters. These were combined with practical hints on where one might go wrong, and succinct information on the pathogenic species. The tables contained carefully chosen data in just the right amount for a useful laboratory manual on the identification of medically important bacteria.

In the years since the last edition, test methods and the variety of bacteria of medical interest have both grown considerably. Not only have poorly studied areas like the ‘diphtheroids’ been much clarified, but a number of newly recognized pathogens such as legionellae have become important. Medical and other workers will therefore welcome this new edition, which follows closely the emphasis and style of its predecessors. The contributors and editors are to be congratulated on their labours in bringing a complex field to the concise summary that is contained here, often in the face of difficulties in finding convenient diagnostic tests for the newer taxa.

For scientific purposes, the naming of living organisms requires rules to ensure standardization and consistency in nomenclature and thus in the unequivocal communication of identity. The necessary scientific rules for naming all living organisms are embodied in international codes of nomenclature. Separate codes have been formulated for animals, plants, bacteria and viruses. The rules governing the application and use of names for bacteria were long overdue before they were promulgated in the first Bacteriological Code approved in 1947 (Buchanan, St. John-Brooks & Breed, 1948). This code, which was largely based on the Botanical Code, was both long and complicated. Subsequent attempts to increase its usefulness resulted in further difficulties despite the accompanying explanations given in the revised versions published in 1958 (Buchanan et al., 1958) and in 1966. Such rules are necessarily framed in legalistic language, which is not always easily understood. The Bacteriological Code has two basic guidelines: (i) to avoid translation into different languages, all names must be Latinized so as to be clearly recognized both scientifically and internationally and (ii) the names must have definite positions in the relevant taxonomic hierarchy.

Informal or vernacular names for bacteria such as ‘coliform organisms’ and the ‘tubercle bacillus’ are not governed by the Code although it may help to regulate their use by recommendations for good practice. Informal names such as those for organisms which differ only antigenically are governed by separate rules.

Scientific names are used for species, genera and higher ranks.

In 1968, Dybowski & Franklin suggested various theoretical models for using computers to help with the identification of bacteria, but the first practical systems for computer-assisted identification were not introduced until 1973 by Friedman et al. and Lapage et al.. They depended on a full set of standardized routine tests carried out under strict laboratory control. One such system for strains difficult to identify in diagnostic laboratories has been perpetuated in the Identification Services Laboratory in the National Collection of Type Cultures (NCTC) at Colindale, London. The work of this laboratory was extensively reviewed by Willcox, Lapage & Holmes (1980). Other national culture collections have set up similar services for their own countries. The ranges of the tests applied have increased progressively in order to accommodate new species and genera. Bacterial groups which have been updated recently or examined in this way include the aerobic fermenting and non-fermenting Gram-negative rods (Holmes & Dawson, 1983; Holmes, Pinning & Dawson, 1986), the Gram-positive aerobic cocci (Feltham & Sneath, 1982) and the genus Bacillus (Logan & Berkeley, 1984). Other bacterial groups have been investigated elsewhere.

Commercial kit suppliers have constructed large databases for their particular range of tests and organisms. As with the Tables in this Manual, the test reactions for identification databases or matrices have often been compiled from a variety of sources. The tables or matrices consist of columns usually showing the percentage of positive test reactions, each row representing a cluster or group of similar organisms.

Identicard set for genus identification

Use a card with a single row of holes (Fig. 9.1).

For the master cards, the key to the numbered holes for the characters is shown in Table El, and the plan for notching the holes in Table E2. All positive characters are notched from the hole to the edge of the card. For certain genera two or more cards may be needed to make provision for different reactions (‘dee’ characters) or differences in reading a test (e.g. catalase with Aerococcus). A guide card with the holes numbered or labelled is useful for the probing (sorting) operation.

Notching the unknown. Positive characters of the strain under test are notched from the appropriate hole (Table El) to the edge of the card.

Probing. When notching the card for the unknown strain is complete put it at the front of the pack of master cards, correctly faced, and test each hole in turn with a probe such as a knitting needle.

A probe through the hole of a positive character of the unknown strain will, when lifted, remove master cards that are negative for that character and these can be discounted. The card for the unknown strain and master cards of genera in which that character is positive will drop out of the pack; these cards are collected together and tested again for another character, and so on.

In Section 8.4 we discussed briefly the impact of numerical taxonomy and the newer genetic techniques for measuring bacterial relatedness and evolutionary divergence on traditional classification and nomenclature as used in this Manual. Because of their fundamental significance for taxonomy in the future we reproduce in full, with permission from the International Committee on Systematic Bacteriology, the report of the ad hoc committee set up to consider the various approaches to bacterial systematics and their relationship to each other published in the International Journal of Systematic Bacteriology by Wayne et al. (1987) on behalf of the International Union [formerly Association] of Microbiological Societies.

Report of the Ad Hoc Committee on Reconciliation of Approaches to Bacterial Systematics

L. G. WAYNE, D. J. BRENNER, R. R. COLWELL, P. A. D. GRIMONT, O. KANDLER, M. I. KRICHEVSKY, L. H. MOORE, W. E. C. MOORE, R. G. E. MURRAY, E. STACKEBRANDT, M. P. STARR, AND H. G. TRUPER.

Our ‘Diagnostic Tables for the Common Medical Bacteria’ were originally published in the Journal of Hygiene. The tables seemed to fill a need and the demand for reprints was so great that Cambridge University Press reprinted them in pamphlet form.

Many inquired about the technical methods, and there were constant complaints that the methods were not described and that the text lacked details of the taxonomic problems. We resolved, therefore, to expand the original paper and to prepare a book which would give sufficient detail of media and methods to justify its description as a laboratory manual.

Although designed for medical workers we hope that others will use it.

The value of a laboratory manual was impressed on one of us in 1935 at the British Postgraduate Medical School. Dr A. A. Miles had prepared a loose-leaf mimeographed manual to supplement (and improve on) a popular laboratory handbook. With this example in mind a manual suited to the special needs of the National Collection of Type Cultures was prepared, and contributions were made by other members of the Collection staff, particularly Mrs P. H. Clarke, Miss H. E. Ross, Miss C. Shaw, and Mr C. S. Brindle. The National Collection Manual in turn became the basis for the appendices to the present Manual.

Features such as brightly coloured pigments may be characteristic of a few species as well as a good pointer to the nature of the organism and its ultimate identification. But pigments may mislead, as in the Biblical case of bleeding polenta which is no longer considered to be a miracle but a phenomenon that can be produced experimentally by a bacterium and in nature is commonly produced by a yeast (Merlino, 1924; Gaughran, 1969; Cowan, 1956a, 1910b). As bacteria present few gross diagnostic features they must be looked at closely; the characters sought and the tests applied will depend on knowledge and expertise as well as experience with similar organisms; the approach to identification will be conditioned by professional training and intuitive skill. The observations made and the tests applied are aimed at characterizing the organism so that it can be described (the technical term for the list of characters is a description) and compared with descriptions of other, previously identified and classified organisms.

Bacterial characterization

The difference between characterization for classification and for identification lies not so much in the tests themselves as in the emphasis placed on the results of the tests. Although it is not universally accepted, most taxonomists now support the Adansonian concept that, for classification, equal weight should be given to each character or feature.