Generally it is much more cost-effective to increase the output of a fermentation plant by improving the yield per tank than by building and equipping extra fermentation tanks, although in some low-cost parts of the world the latter procedure is sometimes still practiced.
The product yield can be increased by medium/process improvement and by genetic strain improvement. However, the yield increase possible from medium/process improvement alone is limited by the maximum biosynthetic capacity of the production strain. After this is reached, no significant further increase can be obtained without improving the strain by mutation or other genetic means. In practice, strain improvement goes hand-in-hand with medium and process improvement, and the high titres seen in many fermentation processes today are the result of both being employed synergistically.
Screening for improved titre mutants of industrial production strains is a specialised process because these mutants are comparatively rare, but many developments in screening methodology over the past few decades have significantly improved the chances of finding improved strains. Although the rational screening procedures described in my related article allow the screening of much higher numbers of mutants than is possible using random screening, it must be remembered that all rational screens are pre-screens and any mutants picked up by them still have to be confirmed in conventional screens. Also, it must be pointed out that our knowledge of the biosynthetic pathways for many products, and especially their regulation, is inevitably limited and so we cannot predict every possible type of mutant that could lead to improved yield, or even be sure where the rate-limiting steps in our current strains may be. On the other hand, random screening allows ALL types of yield improvement mutants to be found, predicted or not.
Optimisation of Mutagenesis
Choice of Mutagen: There are many factors which may influence the choice of mutagen for strain improvement. Many people classify mutagens in terms of the kind of DNA damage that they produce. However, efficient DNA repair mechanisms exist which accurately repair such damage without causing any changes in the DNA base-pair sequence. Mutation only occurs when the level of damage is so high that the normal cellular repair mechanisms cannot cope with it, and error-prone repair is induced to ensure cell survival. So what is more important than the type of DNA damage a mutagen causes is the type of error-prone DNA repair pathway that it induces. Early workers identified different error-prone repair pathways induced in bacteria (such as the so-called “SOS” repair discovered in Escherichia coli) by for example, UV, Nitrosomethylguanidine (NTG) and Ethyl Methane Sulphonate (EMS). However, when DNA repair is examined across a wide range of prokaryotes and eukaryotes the situation becomes much more complex. In addition to repair pathways, the phenomenon of Mutagen Specificity has to be taken into account. Several studies have shown that any given combination of strain, mutagen and plating medium will only produce a specific, limited spectrum of all the possible mutations that exist. Consequently the only way to ensure that the maximum possible variety of mutations is obtained is to use a range of mutation procedures and plating media.
Optimum Dose: Many investigators assume that mutation treatments that give a high degree of lethality (90%; 99%) are efficiently producing mutants. However, this idea is based on earlier work with classical fungi and bacteria whose genetics was fairly well understood. Many fermentation products today are produced by less well understood species of organisms obtained from unusual environments. Killing is not mutagenesis. Induction of cell death does not guarantee that mutagenesis is occurring, and indeed many “standard” mutation treatments may not be effective at all on some strains. In order to ensure that mutagenesis is occurring, and indeed to find the optimum dose for the treatment, it is necessary to measure the increase in the frequency of some sort of genetic marker among the surviving population. The easiest marker to use is alteration in colony morphology, where this occurs. Other markers that can be used are resistance markers (resistance to toxic anti-metabolites) or auxotrophic markers. Titre may itself also be used. While increased-titre mutations are too rare for this purpose, decreased-titre mutations are only too common and can be used to estimate the efficiency of mutagenesis.
Despite having generated a wide variety of titre-improvement mutations in the most efficient way possible, they will not be detected unless the screen employed allows efficient levels of expression.
Monocultures: One of the most frequent reasons for failure to detect improved mutants, or subsequently losing them, is that the colonies are mixed. That is, they contain a mixed population consisting of cells of the improved mutant plus non-mutated cells, or worse cells of a reduced titre mutant. This can arise in many ways – mutation of mycelial fragments or clumps of spores rather than uninucleate material, improperly separated colonies due to bad spreading technique or crowded plates, or cross-contamination between colonies (a particular problem with sporulating strains).
Physiological Constraints: Another often overlooked fact is that the screening conditions must be capable of allowing increased titre mutations to be expressed. For example, when using surface culture bioassay screens, it needs to be remembered that many standard plate media may not support adequate production of the metabolite in question. Additionally, even with shake flask screens, the medium and other conditions need to be carefully designed so that they do not artificially limit the expression of improved mutants. I have discussed medium design in a related article on this site.
Types of Screens
Various types of screens are in use today, ranging from agar plate screens, microtitre well-plate screens, test tubes and shake flask screens using flasks of various types and sizes. In my experience, the best correlation with the large scale stirred tank conditions (i.e. terms of scaling up the improved mutants) is found with the 250 ml and 500 ml Erlenmeyer flasks, and all other types of screens must be regarded as pre-screens to be subsequently confirmed in these.
Another consideration is whether one is looking for rate mutants, with an increased production rate but which may become exhausted early on, or stamina mutants with a normal (or increased) production rate but which are able to continue producing for a longer period. This will obviously affect the time of harvest used in the screen.
It is rare indeed that the improved mutants from a single-level screen can be accurately identified and directly scaled up. Because of the errors in measurement of titre, and the fact that the conditions of a given screen may not accurately reflect those of the final production vessel, it is more usual to use a Multi-Level Screen. This means that the isolates pass through a series of screens in which the numbers are progressively reduced and the degree of replication progressively increased, and the screening conditions (media, fermentation time) may also be varied. Any improved mutants which survive this gauntlet then pass to the small-scale stirred tanks in the pilot plant for final scaling up.
Pooling and Recycling
The accuracy of our screens is much less than we would like, and often it is too low to allow small increases in titre to be detected. Mutations to large increases in titre are rare and mutations to small increases may be more common, although if the magnitude of these is smaller than the testing error of the screen, they cannot be reliably detected with a single pass screen. The technique of Rapid Recycling, or Pooling and Recycling, was devised to overcome this problem. In Pooling and Recycling, the top, say, ten percent of improved isolates from a primary screen are not taken for further confirmation, but are pooled, re-mutated and re-screened. Any strains which appeared in the top 10% by error will tend to be lost in subsequent cycles, while genuinely improved strains will tend to be enriched in the population and come to dominate it. Additionally, as the number of cycles progresses, new mutations will occur which will follow the same enrichment process. After a sufficient number of cycles, the titre of the entire population will have increased via the accumulation of small mutations to such a level that it is clearly improved compared to the starting strain. At this point, the best isolates can be taken from it for further confirmation and testing.
Miniaturisation and Automation
The desire to screen large numbers of isolates in order to increase the chances of detecting an improved strain has led people to develop many ingenious techniques. Miniaturisation is one obvious route to take, and by using microtitre well-plates this enables such miniaturised screens to be automated. The problem with such an approach is that the fermentation conditions inside these wells are very different from those of the commercial scale stirred tank! In fact I have often likened the degree of difference between a microtitre well and a shake flask as being as great as that between a shake flask and a stirred tank fermenter. This inevitably leads to the selection for mutations which overcome the limitations of the well-plate environment but are not relevant in the stirred tank i.e. it leads to the selection of a large number of mutants that cannot be scaled up, even to shake flasks. The problem here is one of accurately measuring QUANTITATIVE differences between strains, which are subject to a large number of influences and errors. In my experience, microtitre screens are much more efficient at detecting all-or-nothing QUALITATIVE differences, such as the presence or absence of a particular enzyme activity, rather than detecting improved titre mutants.
Process and Medium Development
Once we have confirmed an improved mutant in our shake flask screen, it has to be scaled up to be of any commercial benefit. This requires an understanding that an improved mutant is a NEW strain. It cannot be expected to demonstrate its maximum potential under the same process conditions that were used for the old strain. At the very least it will require a rebalanced medium with increased concentrations of precursors and other ingredients, and it may also have an altered morphology or other growth differences. Getting the best performance from a new mutant requires experimentation and close collaboration between the pilot plant team and the strain improvement team who identified and characterised it. Often, a new mutant will show only a small improvement in its initial stirred tank trials, but by re-optimising the fermentation process a much larger improvement can be obtained, often greater than it originally demonstrated in the shake flask screen.
It can be seen from the above article that there are many ways in which classical “random” screening can be made more efficient. Most improved mutants have been and still are obtained using these methods, and they should be part of any strain improvement laboratory’s armoury just as much as rational screening and genetic engineering techniques.