So far, we are familiar with only a fraction of global biodiversity, and even less has been scientifically categorised and examined. If we bear in mind that less than one per cent of all known organisms can be cultured in the laboratory, this highlights even more clearly the gigantic untapped potential that resides in the biosphere.

The purpose and art of harnessing natural resources at BRAIN consists in continually discovering and optimising new active biomolecules in order to make them available for nature-based applications. But biodiversity is found not only in the soil samples that BRAIN scientists are forever taking. Many other people have long been taking full advantage of nature’s cornucopia in quite a different way – in silico, i.e. digitally by means of computers.

BRAIN has developed outstanding expertise in this area over the past few years. One of its specialists is Dr Alexander Pelzer, Research Scientist & Project Manager at BRAIN AG. Since joining BRAIN in 2014, he has constantly been on the look-out for new enzymes. These are proteins popularly known as “biocatalysts”, which accelerate (bio)chemical reactions and may be of interest for various BRAIN projects. Enzymes consist of 20 different amino acids lined up in a defined sequence, like a string of pearls. These acids interact and give rise to pleated three-dimensional structures.

But Pelzer does not only go through flora and fauna. Instead, he systematically sifts through and analyses “big data” in ever-growing databases containing published metagenome data on organisms that thrive in extremely varied places and surfaces on or in the ground. Using state-of-the-art software and his rich store of experience, he compares the genetic information of these organisms with the DNA of other described microbes that are known to produce particularly useful enzymes. Things get really interesting when computer-based and 3D analyses find similar sequences in promising organisms.

The properties and functions of a large number of enzymes on the market have already been described in detail, and this market is constantly growing. Enzymes that are required for very special applications rather than the classical enzyme markets are of particular interest. The intention is to provide novel and effective enzymes for this speciality market.

Identifying new enzyme candidates

In a model test to discover new enzymes in silico that break complex substrates down into small fragments, Pelzer started by looking at the amino acid sequence of three well-known representatives of their class. He then compared them with over 29,000 similar sequences from metagenome databases. In a series of studies, he compared all the molecules to establish whether they have similar sequences. He also made use of computer-based CLANS analyses that can be used to compare the position of the amino acid sequences and to represent them in a spatial model.

A further selection criterium in this stage of work involved observations based on over 20 years of BRAIN experience with its in-house BioArchive and research into biodiversity. This evaluation basically entails comparison with nature. The properties of the enzyme candidates’ original organisms are examined in detail and discussed in the team. Questions like “Which organism does the enzyme come from?” or “Which temperature range can the organism tolerate?” offer valuable indications whether the discovered enzyme might also work in the new application environment targeted by BRAIN.

Only around 20 highly interesting enzymes were left over at the end of the series of tests using the specific in silico model to discover new enzymes. These are candidates whose quality and activity makes them appear particularly suited for speciality markets.

Optimising enzymes via protein engineering?

When it comes to development Pelzer’s colleagues in BRAIN’s molecular biology labs take over. Dr Klaus Liebeton, Research Scientist & Platform Coordinator Microbial Expression at BRAIN, works in close conjunction with Pelzer. Every time Pelzer has identified promising enzyme candidates in metagenome databases, Liebeton’s task is to find a suitable expression strain in BRAIN’s repertoire, such as a bacterium, fungus or yeast. This strain must be capable of incorporating the basic DNA of the desired biocatalyst and reliably translating it into an enzyme.

Here too, the key is preselecting the right gene sequence for the biocatalyst out of a trillion possibilities. For years, Liebeton has devoted himself to gaining an understanding of the connections between the DNA sequence and the relevant protein yield.

But the story doesn’t end there. Once we have examined what quantity of the new enzyme the expression strain produces, and have also characterised the enzyme’s properties, we go back and forth between the molecular biology lab and protein engineering.” In protein engineering, scientists work on the DNA that codes for the target protein and modify the DNA at precisely those points where the need for change was identified. When the DNA is translated into the enzyme, this change is transferred to the enzyme’s amino acid sequence. This makes it possible to modify specific enzyme properties and enhance desirable characteristics.

“Nature didn’t develop enzymes for use in industrial biotechnology, but protein engineering enables us to use promising enzyme candidates from biodiversity as a starting point and to improve them so they are tailored to the conditions that prevail in a wide variety of industrial processes.

If I am looking for an enzyme with a special property, such as increased heat resistance or activity at a lower pH value, I can use 3D models to judge which changes in the amino acid sequence might improve the enzyme properties.”

The first step is to analyse the enzyme models on the computer, especially those regions that define a specific enzyme property, and then to modify them as necessary. State-of-the-art modelling techniques help to customise the proteins step by step towards maximum effect. The pleating of the enzymes offers entry points for this digital comparison. A software programme shows Pelzer and his team three-dimensional full-colour representations of the preselected enzymes on the computer monitor (Fig. 2). It takes a great deal of background knowledge and experience to interpret the images correctly. Each optimisation step on the computer is followed by repeated DNA modification and examination of the enzyme properties in the lab.

Based on in silico biodiversity and using digital sequence comparisons and protein engineering, BRAIN has come a long way in expanding access to “nature’s toolbox”.