3D chemistry speeds up and improves search for industrial enzymes

A needle in a haystack, one in a million, a winning lottery ticket: There are many ways to describe the success of finding an enzyme in nature, and making it work in industry.

​However, the explorer’s task has become easier in recent years with the development of three dimensional (3D) computer modeling, which is helping scientists navigate the vast tracts of data that emerge in the process of optimizing molecules for industry.

“We get a window on the C1chemistry of life: These molecules are made by nature, and their chemical structures are super fine tuned. Looking at that molecule means you are looking at something that has evolved over millions of years,” say Novozymes Science Manager Leonardo De Maria, who works with computerized 3D modeling to try and answer questions as old as life itself.

From backyards to oceans
Based in the company’s protein engineering department, he and his colleagues use cutting-edge technology to optimize enzymes found in nature, so as to deploy them in products that are useful to industry. These enzymes are produced by microorganisms, such as bacteria and fungi, which are found everywhere, from backyard gardens and woodlands, to deserts and oceans. 

“Nature has not optimized these enzymes to work in brewing or baking, nor survive inside a liquid detergent and live on supermarket shelves. This is where Novozymes comes into play,” Leonardo says.

Knowing where and how the enzyme they are seeking will be used helps researchers find the appropriate natural habitats that resemble the industrial conditions where the enzymes will eventually be deployed. For instance, they have dredged freezing Arctic lakes, and tapped hot springs, for enzymes suitable for deployment in cold-water wash, or in the baking industry. What follows is a series of steps to improve, or optimize, an enzyme’s characteristics, so it can be used in an industrial process.

Stringing nature’s beads
Enzymes are proteins, which accelerate chemical processes and are efficient at specific tasks, such as breaking down fats and sugars. Proteins themselves are made up of chains of amino acids, of which just 20 occur in nature. Building a protein is like stringing a necklace using combinations of 20 beads of different colors. Living organisms manufacture proteins in their cells based on instructions written into their DNA, and amino acids are the material those instructions are translated into.
In Novozymes’ factories, enzymes are produced by various bacteria and fungi, which have been given specific instructions about the types of enzymes they must make. While nature has strung its beads in a particular way, these combinations, and the way in which the chains are put together, often need to be altered to create an industrial enzyme. But finding the right combination of amino acids to use in optimizing an enzyme is like picking the winning number in a lottery with an astronomically large pool of tickets.

More fun than video games
At Novozymes, tens of thousands of molecules are screened in the lab to find the optimal amino acid combination that could lead to a product. The search space is enormous, and 3D modeling plays a key role in establishing meaningful paths towards finding the optimal combination. Simulations of enzymes, rendered in 3D on computer screens, help researchers visualize and understand vast amounts of data. Donning chunky 3D glasses - the type used when playing video games - they can see amino acid chains magnified to millions of times their size, and can virtually zoom-in, stretch out, and reorder the structures, without having to use petri dishes, pipettes or cultures.

Amino acid chains can be long and complex. When viewed as 3D structures, these chains appear folded, like a strand of crumpled spaghetti, and help bring closer together items in the chain that would otherwise be far from each other. Computer simulations allow scientists to identify specific regions within the enzyme structure that could hold the key to the property that needs be optimized. This data is translated into DNA instructions fed to organisms to make prototype enzymes, which are then tested in miniaturized versions of industrial processes.

Moreover, computer modeling helps recreate the situations an enzyme might encounter when deployed in an industrial process. For example, a virtual washing environment can examine different factors, such as how a molecule reacts to different fabrics, stains and detergent components, or help predict if a particular enzyme will remove fatty stains at low wash temperatures.

Collaborating for a good result
These processes demand close interplay between the computer modeling, biology and chemistry labs, with all sharing data. And it takes time. Improving the stability of an enzyme, that is, ensuring it lasts under specific conditions, could take several months, or longer, depending on the stability to be achieved. More complex processes, like improving performance, could take even longer.

“If we are working with a customer or partner, we know in advance what we are looking for, as we are given very precise requirements,” says Leonardo. “When we develop products by ourselves, we need to understand what it is that could appeal to a diversity of customers. Along with marketing and sales, we work to translate that into a product.”

Getting quicker at winning the lottery
Computational technologies, including 3D modeling, have improved the odds of finding a winning lottery ticket by reducing the number of false starts and closing in on those parts of a molecule that are likely to produce the desired result when manipulated.

K“We will generate more prototype enzymes, and get better and faster at analyzing them, so the chance of finding the right enzyme is higher,” Leonardo says. The implication is that, over time, there could be faster and better-quality innovation as molecule scanning, 3D modeling and lab-based experimentation improve, and as Novozymes’ in-house knowledge about enzymes and proteins increases.

The potential gains for industry are enormous, given that enzymes can replace traditional chemicals with renewable, biological materials, as well as help cut water and energy use in a variety of industrial processes. This translates into lower carbon dioxide emissions, fewer harmful effluents from industrial processes and reduced consumption of raw materials. The outcomes are cost savings for manufacturers and end consumers, and a more sustainable approach to global economic growth.

Picture captions
Top right: Novozymes Science Manager Leonardo De Maria wears 3D glasses as he studies a virtual model of an enzyme.
Middle left: The picture shows a virtual model of a structure known as SLAC: a small laccase from Streptomyces coelicolor.
Bottom right: The picture shows the virtual model of Candida Antarctica lipase B (CALB) molecule.