50

In particular, I can use it to investigate how I can achieve the best possible yield of a

product (the “sinks”, see above) with starting products (the “sources”), for example if I

want to biotechnologically produce citric acid for the kitchen or nanocellulose for trans­

parent displays – to give a well-known and a very modern example. Similarly, I can now

compare all the metabolic possibilities for one organism with various other organisms and

in this way see what peculiarities are present or even what diversions and alternatives one

organism has and the other does not. You can also see in this way that different strains of

bacteria, such as meningococci, use different pathways to achieve the same rate of growth,

allowing both a pathogenic, disease-causing lifestyle and a more benign lifestyle with

greater effort on amino acid synthesis but less aggressiveness against the human host

(Ampattu et al. 2017).

It is important to validate the modelled (and thus only predicted) metabolic differences

by further experimental data. Since individual errors are corrected by the metabolic flux

network model (in a metabolic flux, all enzymes must work together at the same rate), data

such as RT-PCR measurements on the mRNA expression of metabolically active enzymes

can also be used, for example. These mRNA measurements are “indirect” because only the

mRNA is measured and not the protein or enzyme activity; however, this works well in

practice, with only 5–10% error for fluxes from a network of 30–100 enzymes, as con­

firmed by metabolite measurements (Cecil et al. 2011, 2015). Examples of applications

include the changing lifestyle of chlamydiae (bacteria) during infection (as elementary

bodies and subsequently as reticular bodies; Yang et al. 2019) or the mutual metabolic and

regulatory responses to infection events in fungal infections of fungus and host (Srivastava

et al. 2019).

This is particularly interesting if I want to use it for medical purposes, for example to

develop an antibiotic. Then I am interested in the metabolic pathways that as many bacte­

ria as possible have in common, but which are absent in the sick person and can therefore

be blocked by the antibiotic without endangering the sick person, but at the same time

killing all bacteria that have this metabolic pathway.

The flux calculations also open up the possibility of identifying individual enzymes that

are particularly critical for the survival of the bacteria (because the failure of a particular

enzyme affects, for example, all flux modes that provide an essential cofactor for the bac­

terium and not just a few). This may also help in finding new drugs against insidious fun­

gal infections. One can also re-examine the detailed effects of an antibiotic with gene

expression analyses and a calculation of the resulting metabolite fluxes as well as single

metabolite measurements (Cecil et al. 2011; YANAsquare program). This then helps to

find new drugs against multidrug-resistant staphylococci, for example (Cecil et al. 2015).

At present, we also want to link the different modelling levels (Chaps. 1, 2, 3, 4 and 5)

more intensively in order to better protect plants against drought stress and infections, for

example by identifying key enzymes that have an alternative regulatory function (e.g.

aconitase, which, in addition to its metabolic function in the citric acid cycle, also regu­

lates IRE in mRNAs, see Sect. 2.2) and alter regulation favourably for drought stress or

resistance to infection.

4  Modeling Metabolism and Finding New Antibiotics