Introduction

Development work in the plant is needed for a number of reasons. These may be the need to introduce a new process developed in the laboratory, the need to improve existing plant processes and to correct faults, and the need to deal with problems like infection. Constant attention is needed to hold an existing high standard of working, as results may drift downwards over a long period. There may be times when it is desired to reduce the rate of production, temporarily, and an economical process may be needed to make this possible. The two serious problems that sometimes arise, infection and yield fluctuation, are discussed in Chapter 13, where methods for the study of plant batches are also considered.

This subject has been discussed in a number of papers, for example Queener & Swartz (1979), Küenzi & Auden (1983), and in general by Vandamme (1984). These accounts tend to stress the difficulty and complexity of the subject. This is often the case, especially with highly developed processes. In the earlier stages this is often less difficult. With penicillin, for example, the early processes worked quite well in the plant, though infection and reproducibility presented greater problems. Often the initial difficulty is to get good growth in the inoculum; this difficulty can be overcome after a certain amount of experimentation. As the process is developed, problems such as those described in Chapters 5 and 6 become more apparent. As production increases and the steps in progress become less, the critical factors become more important, and experimental work becomes more difficult to interpret.

The move from development in the laboratory or research pilot plant to development in the main plant involves very considerable changes in methods and ideas. These arise mainly from the size of the equipment, but also from the very different approach that is necessary in commercial operations. In a way, it can be said that the laboratory worker is usually free to think and to experiment and to regard the fermentation process as flexible. Some experiments work and are followed up, others do not do so well and are forgotten. As the work moves to the main plant, the assumption is that the process is now clearly defined, and the objective is to bring it into operation with a minimum of risk, without embarking on a series of experiments. In any case, in large-scale work, the idea of discarding a batch is unthinkable on account of the expense. Thus the introduction of a new process to the plant requires a programme that is well thought out and very reliable, and fully backed up by laboratory work.

In this connection, the main role of the pilot plant is to check laboratory processes as well as possible to establish their suitability for plant work. The larger pilot-plant fermenters may be used for this (1000–50001), as well as for producing kilogram quantities of new substances for test.

The role of the main plant is commercial, i.e. to produce quantities of material for sale at a profit. The firm's profitability, and the availability of cash resources for research and development, depend upon the success of the main plant.

Reviewing the field of antibiotic process development, it can be seen to be a mixture of chance, biochemical and physiological understanding, and the adaptation of a biological process to a physical framework defined by the fermentation equipment.

In wild strains, antibiotic production forms part of the system of differentiation that lies behind the cycle of growth and reproduction; it is usually associated with the end of vegetative growth and the onset of sporulation. Such a rigid system is unsuited to development, and the initial stages of mutation and selection have the effect of loosening this linkage so that antibiotic production starts to extend over a longer period of growth. Thus, at this stage, production often appears to be growth linked and to be associated with a particular phase of growth that must be extended as long as possible. Later on in development, growth and production become relatively independent, and production can be maintained over a long period, giving very high production levels. This often requires that the cells be grown in a particular form of pellets, enabling the productive phase to be held for many days.

There are two obvious limitations to high levels of production, first the supply of oxygen for energy production (and possibly for the maintenance of a favourable aerobic balance in the cells), second the inhibitory effect of the antibiotic and of other products in the fermentation. With wild strains, many substances are inhibitory, such as phosphate, but many of these effects are lost in mutation. Glucose, however, usually remains inhibitory to antibiotic production.

Origins of the development of antibiotic processes

Dr Alexander Fleming first observed penicillin in 1928. A contaminant colony of Penicillium notatum growing on an agar plate covered with bacterial colonies had brought about the lysis of all the colonies around it. In further culture experiments, the disinfecting effect of penicillin was confirmed, but the techniques available at the time were insufficient to enable it to be extracted and tested further. Ten years later, Florey and Chain, with a large team of scientists at Oxford, managed to prepare extracts and demonstrate their efficacy. As a result of this, production began in order to supply the material to the armed forces, and the first use in the field began in 1943 in North Africa. In 1945–6, as supplies increased, penicillin became generally available.

Professor L. P. Garrod, a leading bacteriologist, commented later that no other such trivial observation as that of Fleming has had such momentous consequences. Fifteen years afterwards the mould was being cultivated on a gigantic scale to produce a substance of unexampled power. There was no other chemotherapeutic agent that could be used with such a complete disregard for possible harmful effects. Although its action is selective, it could be used to treat diseases that previously could only be dealt with with difficulty or not at all. Along with the sulphonamides, available in 1935, it dealt with half the emergencies occurring in medicine; it completely transformed the outlook, reducing fatalities to a minimum.

The production of penicillin involved major efforts in process development, and these provide a useful introduction to the present book.

Several antibiotic fermentations have been discussed and illustrated by curves of growth and production, and the impression has probably been given that it is easy to obtain good results. In fact, it is not difficult to obtain good results, but many beginners find it hard at first. This is because it is always essential, if good results are to be obtained, to use optimal growth conditions and to inoculate with a culture in perfect condition for the job. Achieving this condition may be easy or difficult, depending on the strain being used, but it always requires attention. For optimal production it is usually also necessary to meet a number of special conditions during the fermentation; the importance of these conditions may be unexpected.

These subsidiary factors therefore deserve attention. The subject is difficult because the critical factors differ for each fermentation and have to be worked out by experience. This chapter will therefore give a number of illustrative examples that demonstrate the kind of details likely to need consideration.

The fact that it is necessary to have all stages of the inoculum in the right condition suggests that cell differentiation is the key factor (see Chapter 3.6). However, too little is known at present to be certain of the exact way differentiation brings about the observed effects, and an empirical approach is necessary.

Development work usually means testing various fermentation procedures so as to obtain the best results. These tests can only be useful if all conditions are optimal. The basic process itself must be optimised, and all subsidiary factors, such as the inoculum, must be adjusted to give ideal conditions.

Infection and yield variations can cause serious problems in production work. In the early days of the antibiotics industry, infection was often a most-serious problem, sometimes halting production for months at a time. The situation greatly improved after several years' hard work, and plants now work over long periods without trouble. Infection can occasionally break out, and a brief description of the problems it creates is therefore useful.

Outbreaks of groups of low-yielding batches, lasting for periods of time, can also cause trouble. These outbreaks may be of two kinds. Firstly there are relatively small drops in yields (15–25%), probably caused by variations in technique. Secondly, very severe falls in output, to 50%, sometimes less, associated with changed metabolism, for which there is no obvious reason.

These two problems, of infection and yield variation, are difficult to deal with. A team effort between engineers and plant staff is usually needed, a microbiologist being often called in to provide an extra pair of eyes and hands, since it usually happens that the local people find the cure. The two problems, however, are often related, since falls in yield may be due to infection, and it may be necessary to decide which.

The problem of infection

The nature of infection

Infection arises from the contamination of the batches with invading microorganisms, usually bacteria. Infection may be at low or high level. The first may have little or no effect, while with a heavy infection there are serious changes in appearance, and the culture has to be discarded. In most cases, the infection is relatively low, but it is noticeable, has some effect on production, and can be confirmed by testing.

Before we move to a detailed account of the behaviour of antibiotic fermentations and their development, the present chapter is intended to briefly describe the organisms used, methods of culture, mode of growth, and general behaviour, based on a consideration of ordinary microorganisms rather than special high-yielding mutants. This will be illustrated by two student-type experiments. However good the descriptions may be, they can never be as useful as a visit to a laboratory where work is in progress, which will bring out the methods used and the type of expertise required for success.

The isolation of antibiotic-producing microorganisms

General screening techniques

Antibiotic-producing organisms are obtained as a result of search programmes. Fungi and streptomycetes are isolated from soils and other materials by plating samples and picking off suitable colonies. This procedure is outlined later in this chapter. The isolated colonies are transferred to slopes and tested in a suitable manner. A modern isolation and test system has been described by Nolan (1986). Targets of 10 000–100 000 colonies, tested annually, are achieved by growing the isolates in small cups in a stationary block, using a liquid medium. The solutions obtained are then diluted and screened en masse, the tests being conducted in similar blocks. These are then read using a computerised system that not only records the test results but also compares these results with those from previous tests (held in memory) and with a target pattern and picks out any that show the desired pattern. This avoids a personal and painstaking study of the data, which can be time consuming and subject to error.

Introduction

Microbial physiology is the study of the manner in which cells grow and live. It is concerned with their metabolism, growth, reproduction, and response to the environment. In the case of the production of antibiotics, the formation of these substances must be added to these functions. The environment becomes that of submerged fermentation systems, which are affected by the physical conditions used and the effect that the presence of the organism has upon these physical conditions.

Chapter 2 gives a general account of the behaviour of the streptomyces and fungi. In the present chapter a more-detailed account of the behaviour of antibiotic-producing strains is given. The subject is a large one, and it is only possible to give those aspects which are of practical interest. The discussion will also be based on the behaviour of relatively high-yielding strains, stressing features of this behaviour that are important in commercial production of antibiotics and that can be used as a basis for a consideration of advanced development work.

The chemical and biological behaviour of the microorganisms is very complex, involving many reactions, within and without the cells. To attempt to describe such a multi-dimensional system is almost impossible. Workers in the field, therefore, adopt the physiologist's view and consider the cell as a mechanism through which material flows, giving rise to cells and products. The interaction between these flows, the properties of the cells, the conditions used, and the regulatory systems involved can then be considered in terms of systems engineering, without going into too much detail about the biochemical and other mechanisms involved, so as to produce a relatively simple model.

The productivity of antibiotic fermentations depends on the medium and conditions used, and also on the state of the genotype, i.e. the nucleus of the cells. Strain modification by mutation of the nucleus to give high yields has long been used and is illustrated in Fig. 3.9 in the case of oxytetracycline production. Mutation implies random modification of the nucleus. Recombination techniques and gene insertion methods can also be used to modify the nucleus in a more systematic way. Gene insertion appears to be the method of the future.

Since the use of mutation and other methods involve changes in the nucleus, strain development is often seen as a branch of genetics. There is no denying the connection, but, operationally, strain development is more akin to the long-established science of breeding, which has, over very many years, led to important improvements in plants and animals used in farming.

In strain development, the search is for new strains that (1) give better results and more product, (2) are capable of further development, i.e. have a flexible genome, and (3) are convenient in use. In breeding, the classical approach has been based on the careful choice of parental material, optimal conditions of culture for the expression of the desired effect, and detailed recording to aid selection and future planning. These points remain of maximum importance in industrial strain improvement. Selection is an important part of the system. Selection requires a great deal of work, and the choice of the correct system is of key importance.

Mutation produces both good and bad effects on the nucleus. Selection is usually only narrowly based, mainly on productivity.

The antibiotics factory

Although antibiotics process development is concerned with the fermentation stage of manufacture, it forms part of the activities of a group of works units which involve all stages of production. It is therefore important to appreciate the scope of the activities that are included in a typical antibiotics factory. It is likely that such a plant will include the following facilities or work areas.

1. (a) Basic physical inputs: steam raising, electricity, water, and effluent disposal.

2. (b) Laboratories for analysis and for the production of seed cultures for the fermentation plant.

3. (c) The fermentation building and its ancillary equipment.

4. (d) Development facilities, often in the form of a pilot plant.

5. (e) Filtration and extraction buildings.

6. (f) Plant for the preparation of pharmaceutical products for sale.

7. (g) Engineering maintenance facilities of various types.

8. (h) Management and administration.

Works buildings are commonly referred to as “sheds”. This has no derogatory significance, and will be adopted in later descriptions.

All these facilities are of great importance for success, and have their own operational methods and considerations. For example, steam, electricity, and water are used to a very large extent, each costing millions of pounds per year. Water is important for cooling the fermenters, and it is often necessary to provide for the cooling of the water and for re-cycling, especially in warm weather, to make effective use of it. One plant employs a large refrigeration system for this purpose. Many plants have to provide treatment units to bring down the biological oxygen requirement of the effluent before it can be discharged.

Introduction

The type of work required in this field is wide ranging and can vary from the simple to the elaborate types of experimentation. As experimental work is costly, there is a tendency to make many tests on individual batches, especially with stirred culture, and to obtain as much data as possible, often using standard routine methods. This can present a more-complicated picture than is currently needed, but it proves useful for future reference.

In earlier times, equipment was often designed and made by user firms. This was, in some ways, convenient, but it was expensive in time and in design costs. At present, equipment is usually made by specialised firms who supply not only shakers and fermenters but also ancillary equipment; these firms also carry out the installation. There is extensive instrumentation, both on the fermenters and in the analytical laboratory, and built-in computers are used for recording, control, storage of data, and its analysis. Some of these changes have been introduced to economise on staff and avoid over-night supervision.

To bring the subject into focus, Table 7.1 summarises some of the common types of fermentation apparatus and their intended uses. The numbers are not intended to be taken literally, but they give an indication of likely requirements. In general, the smaller sizes of equipment are used for academic work, or for screening and quick tests, where a few clear results are needed. Industrially, for laboratory work on process development, larger fermenters are preferred, for example 20–30-1 capacity, so as to avoid confusion over volume changes and to provide for extraction work.

Fermentation process development is a very wide field, and the variety of subjects and tasks involved, very great. One worker may be glad to have achieved 1 gram/litre of a new product, another, with an old one, is stuck at 30 g/l; yet others wrestle with poor or erratic growth, infections, or with economic problems. How can some sort of balance be struck when writing on such a subject? In the present case, an attempt has been made to present a simple account of the subject, illustrated by a few examples. The result is bound to be speculative and over-simplified, and the omission of so many valuable ideas must not be taken as a lack of respect for them. It is only possible to give a few pointers towards problems and to interesting areas of work, which may suggest further analysis of problems and ways to their solution and, hopefully, provide interest to those unfamiliar with the field.

I have to thank the late Professor J. S. Hough for suggesting the writing of the book, and my dear wife for her patience with the resulting decay of home and garden. I am also glad to acknowledge the help of firms and individuals in providing information: Dr J. P. Herman and Glaxo Research, ICI Plc, Agricultural and Pharmaceuticals Divisions, Dr S. W. Carleysmith and Beecham Pharmaceuticals, Dista Ltd, Liverpool, and Dyckerhoff AG, Gulheim, West Germany. I must also thank Dr I. S. Hunter for providing the stylised graph in Fig. 3.9d, and Dr Elizabeth Wellington for reading the drafts and making many helpful suggestions.

Introduction

The general problem

It is interesting to consider the problems that have to be solved in moving productivity towards high yields, based on a consideration of the material so far presented. While an increase in the specific productivity of the cells is important, many changes are also needed in primary metabolism and growth. Some of these are summarised in Tables 10.1 and 10.2. Table 10.1 lists some of the shifts in behaviour that are needed, while Table 10.2 shows some of the factors that have to be taken into account in both primary and secondary metabolism so as to obtain high yields.

Table 10.1 brings out the great differences between high-yielding and wild strains, especially the independence of secondary metabolism in relation to growth patterns and primary metabolism; the whole process of secondary metabolism becomes smoother and more flexible. Table 10.2 brings out the complexity of primary metabolism and the number of factors connected with it that have to be optimised to give good production. Secondary metabolism is, in comparison, simpler, though it depends on the properties of the cells being favourable to it. For the best results, secondary metabolism must be able to work well while primary metabolism is changing. Productivity is often connected to growth rate (μ). Productivity seems really to be related to the state of the cells. This may imply a particular value of μ, but cases arise where productivity is high, even in the absence of growth.

As mutation and selection take place, both good and bad effects are produced. It is possible that at some times primary metabolism may be optimal for production, and at other times it may be less good.