In 2010, a team of researchers at NMBU discovered one of the world’s most powerful biomass‑degrading enzymes, later named lytic polysaccharide monooxygenases (LPMOs). Published in Science, the breakthrough reshaped industrial biotechnology.
Today, these enzymes help unlock energy and valuable products from non‑edible biomass such as straw, forest residues and shrimp shells—and ongoing discoveries from the NMBU team continue to reveal how LPMOs work and how they can be used most effectively.
Here, professor Vincent Eijsink outlines the LPMO-story from the NMBU perspective:
Enzymatic depolymerization of crystalline polysaccharides such as chitin and cellulose is important for development of a more circular economy, where biomass is exploited in better and more sustainable manners. Such enzymatic depolymerization is very challenging because the enzymes doing this, chitinases, cellulases, need to bind to a single polymer chain into their active sites.
“Extracting” such a single chain from a crystalline fiber (or surface) poses a major energetic barrier; a very strong network of hydrogen bond interactions needs to be broken:
Solving the challenges
As early as in 1950, Elwyn Reese, a famous researcher in the cellulase field, and co-workers proposed that depolymerization of cellulose by hydrolytic enzymes (cellulases) also involves a “decrystallizing” factor called “C1”. Such a factor would be needed to make the insoluble and often crystalline polysaccharide more accessible to hydrolases, solving the challenges described above. This is Reese’s famous C1-Cx hypothesis.
Since 1950, some researchers had been chasing C1, without success. Of note, there were some indications that some carbohydrate-binding modules (CBMs – protein domains with a substrate binding function only) perhaps could decrystallize cellulose.
We have been studying the enzymatic depolymerization of these hard-to-degrade insoluble polysaccharides since the late 1990ies, initially working on chitin. If one grows a bacterium called S. marcescens on chitin, one sees a set of distinct proteins being overexpressed, as nicely illustrated in Fig. 2 of Suzuki et al., 1998 (see figure). These bands represent chitinases, but there is one protein there, CBP21, for chitin-binding protein of 21 kDa, that is a lonely CBM of the CBM33 family, and that at the time was known as chitin-binding protein only. Similar proteins had been described in a few other bacteria.
Addressing new issues
In the period 2001 - 2005, we were studying these chitinases addressing issues such as how they cleave (exo or endo), the structural determinants and impact of processivity, the ability to use these enzymes to valorize chitin, and synergy between the enzymes. After having done chitinase work in the first phase of his work as a PhD student, Gustav Vaaje-Kolstad started working on this CBP21 protein. In a first 2005 paper, we described the crystal structure of the protein, which revealed some remarkable features. In particular, the many aromatic residues in this “CBM”, which are typically associated with substrate-binding (so, with CBM function), turned out to be on the inside of the protein and not on the substrate-binding surface (Nb. today we know why that is).
In a later study that same year (2005), we then showed that this CBP21 protein drastically boosts the activity of chitinases. This was the first ever, crystal clear proof of the existence of something like the C1 factor proposed by Reese. There was a lot we did not yet understand at the time, and we were lucky that we somehow were using reaction conditions in which CBP21 was active. We did not yet know that CBP21 was an enzyme. We had no idea about how it worked, but, in accordance with general thinking at the time, we were speculating about it using binding affinity to act as some sort of wedge. At the time, we did not realize that these proteins were enzymes; the title of the paper in which we described the synergy of CBP21 with chitinases even includes the term “non-catalytic”!
Subsequently, others, in particular researchers at Novozymes, described similar effects for proteins acting on cellulose that at the time were erroneously classified as family GH61 glycoside hydrolases, but the underlying mechanisms remained unknown. Partly by coincidence, we discovered that CBP21 needed an electron donor to exert its boosting effect on chitinases, which suggested that CBP21 in fact catalyzes a chemical reaction.
This prompted us to see if we could find chitin fragments generated from chitin by CBP21, profiting from the analytical competence of Bjørge Westereng who had then joined the group. Indeed, using mass spectrometry, we could detect fragments, and these fragments turned out to be oxidized. In fact, we found that we could degrade chitin quite a bit, by just treating it with CBP21 and a reductant (see picture below). Then, we knew that we had found a completely new type of enzyme reaction that entails oxidative cleavage of glycosidic bonds. We showed that this reaction depended on the presence of oxygen and a reductant (= electron donor).
In a 2010 paper published in Science, we then showed that these “C1-proteins” in fact are enzymes that oxidatively break glycosidic bonds and that, by doing so, boost the activity of classical hydrolytic enzymes. These enzymes, today known as Lytic Polysaccharide Monooxygenases (LPMOs), act on the surfaces of chitin and cellulose fibrils, rather than on single polysaccharide chains, as illustrated by, for example, the experiment depicted in Fig. 2E,F of the 2010 paper.
The picture below, an artist impression, illustrates the difference between LPMOs (red, acting on a surface) and cellulases (yellow, acting on a single chain):
It was immediately obvious from our 2010 results that GH61 proteins likely had the same activity as CBP21, but on cellulose, and that other members of the CBM33 family (to which CBP21 belongs) likely would act on cellulose. This was indeed shown in several studies published in 2011, by different groups.
In 2010, we did not yet see that LPMOs use copper as a co-factor, partly because experiments done by us and by others gave confusing results that put us on the wrong track. Two other groups pointed at copper in 2011 and since then it has been clear that LPMOs are mono-copper enzymes, which, notably, makes them unique in terms of their catalytic mechanism and power. It is worth noting, as we already did in 2005, that these proteins are widespread in Nature, suggesting that they may have multiple functions. It is also worth noting that LPMOs are small and in a way “simple” proteins, while at the same time catalyzing a very challenging reaction.
In 2010, again, we were lucky because with chitin, and using these reaction conditions, the synergistic effects between CBP21 and chitinases were exceptionally huge. In 2023 we published a paper providing a rational explanation for our luck (Østby et al., 2023). This paper provides important hints for further optimization of the use of LPMOs in biorefining.
Discovering new LPMOs
In the years that followed, many new LPMOs were discovered, with different substrate specificities. Importantly, many fungi have tenths of LPMO genes and one would thus expect functional diversity. Such diversity could, for example, relate to the complexity of the plant cell wall; one can envisage that different LPMOs are needed to degrade different substructures in the plant cell wall. There is data supporting this. While they are generally associated with degradation of biomass such as plant cell walls (= lignocellulosic biomass) and chitin-rich materials, they have other functions that largely remain unclear. Today, it is clear that LPMOs have roles in plant pathogen interactions, microbial development and microbial virulence. The precise nature of these roles remains unclear; there are large discoveries waiting to be made.
As to the use in biorefineries and the catalytic mechanism, progress was made, albeit slowly. LPMOs are easy to produce but difficult to work with, and high-quality kinetic data did not become available before about 2018 (and are still scarce). LPMOs are unstable if not used “correctly”; they damage themselves. Most importantly, collectively, the field, including us, did not understand how LPMO reactions actually take place.
In the laboratory LPMO reactions were (and are still) typically done as follows: substrate + LPMO + reductant in anaerobic conditions (so with oxygen). We realize now (see below) that, when using such conditions, hydrogen peroxide is generated in the reaction mixtures and that it is the generation of this co-substrate that limits the reaction. This has “fooled” the field, including us.
The name LPMOs indicates that these enzymes use molecular oxygen as a co-substrate, as we and others had shown. Driven by a brilliant post-doc, Bastien Bissaro, and following years of “seeing things that we did not understand”, we discovered in 2016 that the rate-limiting step in typical LPMO reactions is the (slow) generation of hydrogen peroxide, which is the true co-substrate of these enzymes. Hence these enzymes are not oxygen-using monooxygenases but hydrogen peroxide-using peroxygenases. In 2016 (BioRxiv) / 2017 (Nature Chemical Biology), we claimed (and showed) that LPMOs are in fact peroxygenases that use hydrogen peroxide rather than oxygen. Among many things, we pointed out that this had been overlooked because in commonly used conditions for LPMO reactions hydrogen peroxide will be formed. Our findings were controversial for some, and still are, for a very few, although semantics play a role here. Today, it is clear that LPMOs primarily are peroxygenases. In this 2016/2017 study, we also showed how and why LPMOs are susceptible to self-inactivation if not “used correctly”. These findings have major implications for bioprocess development.
It is worth mentioning that reaction rates originally reported for LPMOs were typically in the 1 per minute range, so very, very slow (for an enzyme). When running LPMO reactions under “peroxygenase conditions” (so with supply of hydrogen peroxide), the enzymes show “normal” reaction rates on the order of 1 to 100 per second.
Our work since 2017, including our applied work, has been much focused on understanding what is really happening in LPMO reactions and on how such reactions can be optimized (e.g. by keeping the LPMO undamaged and by assessing different ways of supplying/controlling access to hydrogen peroxide). We have spent quite some energy on showing that our 2016/2017 findings were correct and we have been doing the first thorough kinetics work in collaboration with (and often led by) the groups of Roland Ludwig (Vienna) and Priit Väljamäe (Tartu). We believe that we have a pretty good understanding of things today, and an ability to predict/optimize/control reactivity and enzyme stability, however, only in clean systems (see below).
We have also done structure-function studies that have gradually disclosed structural features that determine LPMO functionality. Despite major efforts, such as multiple studies of bacterial LPMOs by one of our leading LPMO researchers, Zarah Forsberg, several questions remain. As an example, the structural determinants of the substrate specificity of LPMOs are only partly known. More recently, we have been heavily engaged in studying how the wider copper environment (the “second sphere”) affects LPMO function.
The team of Roland Ludwig has been very important for us and we are grateful for a highly productive collaboration over many years. Recently, Ludwig et al. developed hydrogen peroxide sensors that, since 2023, allow doing accurate and extensive kinetic measurements with LPMOs. This is helping us gathering a much deeper understanding of LPMO catalysis.
Despite all these developments and discoveries, optimizing LPMO function in industrial processes with “messy” substrates (like pretreated lignocellulosic biomass) remains a major challenge, because these substrates contain large amounts of redox active material (lignin). Lignin reacts with oxygen (-> producing hydrogen peroxide) and with hydrogen peroxide (consuming the LPMO co-substrate); lignin can act as reductant. Because of this, reactions with lignin are very hard to control and even light may play a role. One of the things that is becoming clear is that the “state of the lignin”, and, thus, the method used to pretreat the biomass, has important effects on how LPMOs perform in subsequent enzymatic saccharification of such biomass.
LPMOs today
LPMOs are today an important component of commercial cellulolytic enzyme cocktails for processing of lignocellulosic biomass, and their impact on biomass processing efficiency is evident from multiple studies. Ongoing research in the field is focusing on optimizing the interplay between redox-active LPMOs, reducing power, molecular oxygen, hydrogen peroxide, and regular cellulases, an interplay that also depends on how the lignocellulosic biomass is pre-treated. Importantly, while LPMOs are widely used and while their biotechnological importance is undisputed, we would claim that their full potential has not yet been harnessed in bioprocessing, due to a lack of full understanding and control of the factors that govern their activity, stability and interplay with (multiple types of) hydrolytic enzymes. In the past decade, novel LPMO activities have been discovered, such as activity on cellulose-bound hemicelluloses, and implementation of these activities in enzyme cocktails for biomass processing may lead to further improvements. Chitin-rich biomass is another abundant polysaccharide biomass, the biological processing of which is less well developed compared to lignocellulosic biomass. LPMOs will be central in developing efficient enzyme technologies for processing and valorization of chitin.
While originally studied in the context of biomass conversion, it is now clear that LPMOs are ubiquitous in Nature, occurring in a wide variety of bacterial and fungal communities, viruses, plants and insects. This suggests that LPMOs have a wide range of biological functions, for example related to cellular development and microbial pathogenicity, processes that may also entail the degradation of recalcitrant polysaccharides. Thus, the impact of the discovery of LPMOs will extend beyond the impact on enzymatic processing of polysaccharides in biomass.
Next to being biocatalysts with great biotechnological potential, LPMOs are of major interest because of their unprecedented mono-copper catalytic center and mechanism. The unique chemistry of LPMOs is a subject of much ongoing fundamental research by biotechnologists and chemists. Moreover, LPMOs are inspiring biotechnologists and chemists to design and evolve novel enzymes or synthetic catalysts. LPMOs have properties that make them very suitable for catalyst development: they are small, easy to produce, soluble, and evolvable (the latter shown by huge natural variation in sequences). Moreover, in contrast to many other redox enzymes, they use a simple, cheap co-factor, copper.
We currently have a several projects that include the use of LPMOs in processing of chitin-rich biomass, such as fungal cell walls remaining after the industrial production of citric acid (Valuable, ChitoVal and Mycocircle). A larger body of our current LPMO work, mainly funded by an ERC-Synergy grant, is less biomass focused and more focused on catalyst development in general. Major current goals are:
- Fully understand how LPMOs work.
- Can LPMO features be emulated in synthetic catalysts?
- Engineer LPMOs or de novo design LPMO-inspired proteins to catalyze novel reactions: (a) selective oxidation (C-H bond activation) of organic molecules and (b) oxidation of plastics.
For more information and contact details:
research group