 Professor of enzyme technology, head of the protein chemistry and enzyme technology section at Department of Biotechnology and Biomedicine, Technical University of Denmark. The section comprises eight professor research groups in total 75 persons, including 20 PhD students. She is a group leader of enzyme technology in the section. Her research interests are many among them. Applied enzyme technology, including enzyme enzymatic biorefining of biomass, agroindustrial side streams, starch, pectin, and seaweeds for production of bioactives and functional food compounds. Dr. Mayer. I'm just getting connected here. Well, thanks a lot for this arrangement. Thanks a lot for inviting me to talk today, the last speaker. I'm very pleased to be here, and I'm grateful to both the Biotechnology Institute and Amano, of course. So I'm obviously from Denmark. As you can see, we have this little red flag. And I'm going to talk about the new proof processes and ingredients with enzymes. So now I'll see how this works. So in general, I'd like just to introduce one slide with what we actually do regarding enzyme technology as a research discipline. So we actually have enzymes as a major disciplinary research area at the university. And we generally treat, if you look at the bottom here, many different types of raw materials, very much like what we've heard today, both residues. And as you can see, maybe on the left, seaweeds as well. That's one of my crazy PhD students showing his misery, how large his seaweeds are. I'm going to talk a little bit about that today. Then we're having some processing residues in our part of the world. It's, for example, a sugar beet. And recently, we're even working on CO2 conversion as a carbon substrate and even plastics, as you can see at the bottom here. Then our work line is that we do some kind of bioinformatics to actually find various enzymes, discover new enzymes, or understand the diversity of the enzymology. Maybe try the first homology structure. Then we very often produce the enzymes ourselves. We purchase a synthetic gene. And then we do kinetics, which is the middle part, the lower part, which is actually my main research area. And then we very often do engineering pretty much as what has been told today. And then we finally validate with the analytics and very often have another structural go at it, even with an ordinary structure, or at least an alpha fold with some docking, as I'll talk about. And then this continues forever. So we gain new knowledge about these enzymes, how they work. Then we get new ideas, getting new conversion concepts on board, very much in contact with industry. I come from a very small country, so we're pretty much, how can I say, connected not so much to the enzyme industry only, but certainly to a lot of processors in Europe. So I've decided to consider some of the major challenges confronting the modern food science field today, as we've heard amply already. And I'm going to pick, for examples, where we have some results that are in line with the title of the talk, and I think the whole vibe today that we need to do some sustainable processes. We need to work to solve some large challenges. But at the same time, there has to be some kind of a business model. So the first is a little bit about corn bran. Then I'm going to talk about enzymatic synthesis of human milk oligosaccharides. I'm going to introduce those. Enzymatic seaweed biorefining as a quite new field we are tackling. And then since I know this is a kind of functionality interest here, we have a little bit about a functional starch modification that's a new field for me, but we have a little bit on that. So one of the first things, as we've also talked about, that's like using the byproducts. So obviously, corn bran is a huge sidestream from corn starch processing here in the US. I believe, or we have tried to calculate, that it's about 55 to 60 million tons per year. Currently, a low value byproduct. It's very resistant to normal enzymatic degradation of the backbone because you can see, as I've tried to draw here, a little bit exaggerated structural model with the cross-linking, but the problem is it's very, very substituted. The Arabino-Silent constitutes 20% to 40%. We know that solubilized Arabino-Silents, they can also be from other types of cereal, of course. They are particularly substituted when it comes to corn bran. So the corn bran Arabino-Silent is usually insoluble because of the cross-links. But we're interested in seeing if we could solubilize and get some of these potentially healthy effects out of it. And I'd like to say now that in climatic solubilization with classical endo-acting endo-Silentases that are amply available commercially and that are used commercially in wheat processing, for example, and in bread making, they don't work on this corn bran. So we decided to look for one. And we looked into corn pathogens and did bioinformatics. We've detailed that in a couple of papers here and found an enzyme that we call GH30. That's a different type of enzyme family. And it comes from a bacterium called Dekaya chrysanthimii, which was formerly called something else. It attacks other plant crops as well. So the idea here is not to completely chop the Silent down to monomeric constituents, but try to dissolve some oligomeric components without being too sensitive to the enzymatic treatment time and dose. And we actually see that we can release some oligomers that, of course, over time they do decrease a little bit, but it's not like the usual thing which happens with endo-acting enzymes where you get the oligomers into solution and then they are chopped up very, very quickly. Here we kind of retain the oligomeric structures. And the first thing we looked at that is, so what is this enzyme doing? Well, to the very right you will see a green circle. And it does actually chop quite close to a methylated glucuronic acid, the type that Emma Master talked about earlier. So I can offer you some here, Emma. And then we're able to cross link it again by a lac case so we can actually functionalize this material, which is partly cross link still, but we can use it as a new type of hydrogel in foods, for example. The chemistry I've shown to the right that is that we actually do show that the ferulic acid does decrease and the dipherulates increase. So we know when we add a lac case two different types, they'll act a little bit slightly differently, then we get the cross links. In general, we are then very, very interested in these ferulic acid cross links that in many ways hinder further processing of the original corn bran. So we decided to look a little bit at these and I can reveal that maybe for the last 10 years we've been looking at if there were enzymes that would actually release these dipherulates by actually then cleaving near the ester bond, just like normal ferulic acid esterases do, but then actually then open up the cross links, as we would call it. And recently I've had a very, very good PhD student or PhD fellow who worked on this and he used the solubilized GH30 oligomers to test if we could actually find dipherulic acid releasing enzymes by testing a range of various ferulic acid esterases. And at first I'd like to say we did spend quite some time developing a robust LC-MS analysis to actually detect these dipherulates. And yes, indeed, some fungal and a few bacterial, two bacterial we found, they're actually able to catalyze this dipherulic acid release, but mainly from the solubilized oligomers. So that's also interesting. That's probably how it works, that you need to get these dipherulates in solution, so to speak. So we do actually infer that what we do release with this GH30, that small sandwich compounds of cross linked complexly substituted ferulate. Then we've tried to, and I can say at this time of the day, we've tried to understand, so what is the structure function, why do they, why do some release these diphes? And I tried to dock it. We've tried to do some docking experiments, but to be perfectly honest, we're not 100% sure of why some of the fungal ferulol esterases can do it and others cannot. There might be some loop and lid function that I've shown to the right, but we're currently pursuing that further. So the scientific part is, of course, to understand how the enzymes exactly do this. They will all release the ferulic acid as well, but only some will actually attack these tight dipher structures. Then the question is, can we actually confirm that once the ferulic acid dipherulates are present in a cross linked fashion, is that good in the gut, even if they are solubilized? We know that when they are insoluble, they are not prebiotic at least. So we did very recently, this is just an in vitro fermentation over 24 hours, although with genuine human feces. So these results have to be taken with a pinch of salt. But we did then treat the original oligomers with alkaline, so you remove the dipherulates. And then we also cross-linked it again with lacquer. So we got an exaggerated cross-linking, and then we tested how that affected the 24 hour fermentation in vitro and used inulin as a control, as I hope you can see. And we did see some significant changes. Of course, this is only a small fraction that I have time to show. These results are sometimes extremely complicated to actually interpret objectively. But we did see that certain Bifidobacteria and bloutia, they did increase in the relative abundance. We even could nail those down to a prebotelacopry and something called Anorostepis heterus. I don't know if you can see that at the bottom. And then we had the actual abundance and did a number of other assessments of the microbiology. But what we do believe is that actually these oligomers that have some degree of cross-linking, they do work because they are in fact soluble. But if we cross-link too much, they become insoluble, and that somehow does not improve their immediate prebiotic activity. But on the other hand, when we remove the dipheriolates, they have lesser, at least this beneficial bacterial growth promoting activity. So there's more to be done in this field. But I feel this could be good news for urban oocylin and use of this material for a value-added proposition. Next subject, enzymatic synthesis of human milk oligosaccharides. Human milk oligosaccharides, I think you may know that that's quite an abundant constituent of human breast milk. It has many beneficial effects for the breastfed baby. And there is now a big impetus for actually biosynthetic production, not to avoid breastfeeding, but to actually improve the infant formula that are even sold after the baby has been breastfed. We also know now that some of these, they even have a beneficial effect in adults and in the elderly. So we have embarked on this for a while now. And the interesting thing is there's a lot of different structures built on a basic blueprint structure with only five monomers that I've tried to sketch here, mainly lactose in the reducing end, and then decorations or elongations with innastuclucosamine, galactose, and then decorations with frucose and cialic acid. So we have all these as our playground for synthesis. And we're actually trying to now enhance the number of different types very recently. And it happens to be right in our backyard in Denmark, actually, where they have started this factory production in E. coli of the small, very few, three, four are particularly abundant. It's the two prime fucosil lactose. It's something called six prime cialil lactose, which is less prevalent at present than then. It's something called LNNT that has like this. And the new idea is that we'll actually try to then decorate with cialil or fucosil, as I'll show you. It's particularly difficult for E. coli to actually produce and secrete the cialilated. So we have, for a while, worked on the, we call it the non-virulent twin, or cousin, from the original troponosomal crudzi producing, or the troponosomal crudzi cialidase, which is tied to the virulence of that troponosoma disease. And we ended up finding an original Argentinian paper where they had, by five, six, very few amino acid substitutions in the active site. They had changed the enzyme to be, to work in the reverse, to transglycosylate, or transcialilate, and that's what we have been then expanding from there. One other point I'd like to emphasize that is to use another sidestream from dairy production, and it turns out we all know what happens when the cheese clots. But remember, there's also a glycoprotein released, and that glycoprotein from the kappa casein turns out to be cialilated, and we have used that as a cialid donor for this transglycosylation reaction. And some of my colleagues here that do this, they are pictured here, Brigitte and Jesper. And the next is, of course, we have engineered this with something we call the loop engineering, exactly to make the transglycosylation work better, and to make a very, very long story short. We're now up to having 15 mutations near the active site and in the loops, carefully developed from analysis, and we now have something called now TR15 and TR16 with these mutations that involve both those in the active site, and some around the loop to actually see if we can direct that cialil in the reverse reaction. So the next application step is, of course, okay, does this work in milk? Instead of producing only the ingredients, could you do this in milk prior to the drying, prior to the infant formula preparation, or even as a type of active beverage product? And then in Denmark, we have a system where you can become an industrial postdoc or an industrial PhD, and this is actually Valentina. She looks like she just got out of high school. She's the mother of two children and a very bright young postdoc with me, and she worked on this project where she showed with these newly developed trans-cialil acting enzymes that you could do the exact same reaction in milk. You could be concerned that actually in milk, there's both the lipids and the proteins, but the reactions run readily, and she was also forced to actually determine the activation energy, and she found that this TR15 was actually far better, even had a lower activation energy than the original Tryponosomacruzis true trans-cialidase that you see at the bottom here, but that worked as well. Then she got a little bit more ambitious and asked if she could try to do simultaneous silulation and fucrosilation in the milk, where the idea is to actually expand that palette of different types of human milk, all the saccharides you would have, and you can easily do that now with some of the components available. We actually had to add that silylated glycoprotein. It couldn't actually work well enough on the original Kappa-Casin, but by adding this commercial component, the two prime fucrosilactose, then you can use that as a donor onto the LNNT. We have another fucrosidase that we have shown in a different study. It's called TF that I've pictured here to the upper right. That's from a bacterium called Tanorella facinthia. It's one that actually causes oral disease, but the enzyme is taken out of the organism and works well to transglycosylate without even being engineered, and that also works perfectly well. It's not the highest yields we have recorded, but we could actually produce some of these more complex DSLNT in the milk and in the buffer as well in one reaction. Then we got curious about if there was a system among these glycosylhydrolase family 29 if we could predict this transglycosylation ability. There is a lot of hydrolytic data available in literature, and we have also done this a lot. It's easy enough to see which kind of compound that is hydrolyzed. I've tried to map it here. It's a little bit small. We have done kind of a new bioinformatic analysis where we try to group these to understand how these work better, and we have a bioinformatician, a Christian who's pictured to the very right there who's championing this. Then when we add the actual trans, well, first there is a complex substrate specificity. We don't really get anything from that. Then when we try to add the transglycosylations, that's these black spots that I hope you can see. The question is, is there way or are there certain groups of these enzymes that actually transglycosylate better? And the answer is no. We're not able to identify these groupings according to the sequence and structure. So we have to conclude that actually this transglycosylation ability, which is quite interesting from a biological point of view, it seems to be broadly distributed across this family and we cannot predict this at all from the motive sequences that we use like this. Then I'm gonna run very quickly through some of the brown seaweeds. Brown seaweeds, the kelp type, they are now actually grown even in the North Atlantic now because the hydrocolloid market seems endless, alginate you use as a hydrocolloid, it used as I'll show on the next slide, and we now have even fuccoidin and laminarin that can also be used. We'll not eat the raw seaweed anymore in the Stone Age in Denmark, this was eaten, but it's not considered a good palatable food source today. So we try to see, can we actually modify the alginate of some of these cultivated seaweed, which doesn't necessarily have a good gelling alginate in itself, and you can do that by enhancing the guluronic acid components, and you can do that by removing the manuronic acid parts and there's plenty of calcium there. So that's actually what we have done. We have considered various strategies, but we have at first tried to see if we can simply remove some of the manuronic acid enzymatically. And here are some other guys from my lab, Bowen Casper, and they furthermore wanted to see if we could find, because of the novelty, if we could find these enzymes in fungi, which they did, and this is found in a particular group here that I have circulated in green, and then they went ahead, as I've tried to sketch, because they always think biotechnology and what can we do. So we produced some of these enzymes in our lab, where we have a Piccia pastoris production facility, I can say, with a couple of laptop fermenters, and true enough, they did find in this Salina, P. Salina fungus, actually new types of alginate liases that they have characterized up and down, and the original idea was to understand how this fungus actually adapted to grow on this brown seaweed. We thought there was a big kind of revelation in understanding that it would have a completely different set of caseins in order to thrive there, but no, not at all. It turned out that it was only a small set of alginate liase, alginate degrading enzymes that actually distinguished it from other fungi. So that means that we could probably easily find more, and just to show, then we found one that was a little bit faster on the poly-M, poly-manoronic acid, and went ahead and considered that, and it can actually modify alginate so you get a more poly-gluoronic acid content. Then another one is, he discovered, both discovered a new exo-polysaccharide liase family, eight, the first fungal, in that family that can do the same. It selectively removes manoronic acid, and maybe that has a little bit a better prospect for modifying the alginate. We're currently trying to understand if we can use these products for something, and we also want to understand how this enzyme functions. He had to move on to another project, so we're currently struggling a little bit to continue this type of work. Then I can see for the sake of time, I'll only spend a few minutes on some of our favorite enzymes. That's the fuccoidin modifying enzymes. We started out by looking at the fuccoidin, which is a very highly valued bioactive polysaccharide in brown seaweeds, and again, very similar to what Pam Ismail explained, we're trying to extract using enzymes in order to get the structures out intact. They come as sulfated compounds. They have interesting backbones, very rich in fuccos, of course, that's why it's called fuccoidin. Then they have alternating 1, 3, 1, 4, alpha 1, 4, 1, 3 bonds, and various patterns of the sulfate substitutions, and we know that they have a range of biological activities that appear tied to the sulfation pattern, and of course the backbone pattern that we don't quite understand yet. Then we want to make sure that the fuccoidins can be used, and that's very important to get some kind of homogenous products because the fuccoidin does change depending on the origin, and of course depending on the type of extraction where you can destroy it pretty well. So we have a large program where we discover these enzymes that keep giving, as we say. There's a new family as well. They're organized roughly in two families, some that mainly catalyze the hydrolysis of the alpha 1, 4 bond in the backbone, that's the GH107, then some that generally mainly catalyze the alpha 1, 3 in the backbone, that's GH168. Of course, we have then managed to find some that don't fit into this pattern. There are very few, so it's easy enough to do this by alignment, finding other GH107, but some we have found they even cleave the alpha 1, 3. Then another problem, these enzymes, they would usually come, strangely enough, from marine derived microbes, and they're very, very unstable when we try to express them in E. coli, as in you cannot even do the Western blood before it's been decayed. And then Maria, to the right here at the bottom, she decided to actually chop them up and remove these C-terminal parts that are actually called lamg constructs that are kind of the arms on which the enzyme is tied when it's tied to the microbe, and then we get these severely C-terminally truncated enzymes that work extremely well. They also express well, and they chop down the fuccoid and extremely well, and we have written quite some articles about that because that's a very interesting trick to actually stabilize these enzymes. We are now aware that apart from, first, we spent more than a year trying to remove the sulfate groups. We thought they were in the way, but it turns out that they are actually not in the way some of these substitutions on the fuccoidin that vary in different seaweeds, and we're particularly interested in something called sagrina latissima, which is grown in the cold northern hemisphere, that they actually use these substitutions as kind of recognition tags for the enzymes. We're not quite clear about how that works yet, but we're finding constantly new types of oligomers that we then test with medical doctors that we have a collaboration with in Germany, that they have all kinds of strange biomedical functions, and that's why I'd like to just, in the interest, I know the Amano is also interested in the kind of pharmaceutical era. So here's an area that's completely free. This is not tied to production or anything that we get the enzymes from any companies. Some of these fuccoidin products, exactly as explained, they're, of course, more worth than the enzymes themselves, but they work in various biomedical fashions, and one of the most interesting that is they have some kind of bone healing function. We have worked with people who actually break the bones in these sheep, then stitch them together with some of this fuccoidin as well. Another type of the structure, and exactly those that we have enzymatically extracted, they'll help work on the eye disease, it seems, or at least the processes that leads to this AMD, which is called the H-related macular degeneration, which is actually increasing all over the world, including in this country, and there's no cure, except if you want to go to your doctor and have an injection directly into the eye once a month. Then just briefly, I can see them a little bit over time, but I have three slides about the functional starch modification because there's a very interesting aspect that I added when some of the lectures today that actually the potato starch industry, quite big in Europe, they're also looking for clean label products. Most of the potato starch is used in foods in everything from gummy bears to mayonnaise and dressings and artificial cheese, but they use chemical modifications. So now their vision is to use enzymatic modification. One of the first things we have looked at, that's the alpha glucanotransferases, the same type that do the cyclo-dextrins, that the previous speaker talked about, but here we have located some that actually elongate the chains. As you can see, I have drawn here in this pitiful sketch, but we actually decided to look into this. They are intracellular enzymes. They don't add new chains, they elongate chains. And to make a very long story short about, they have different modular make-ups. They are distributed across a wide range of different types of bacterial species. There is some, which is patented by Avive in Holland from a thermothemophilus, which is highly thermostable. We don't necessarily need a highly thermostable enzyme here, but we use that as our benchmark. It's the one called TT in the around five o'clock. And then we test it, and that's just how we normally do it. We test their optoma in a special statistical setup, and we define the optoma. And as you can see, true enough, the upper right TT has a very high thermostability. Whereas the others, they have a pH temperature optimum medium range. Then we have a new technique where we actually measure the chain elongation, including the chain elongation change. And as you can maybe see, they all, when you go from the yellow to the more purple, they pick up or remove the shorter that become slightly longer. But the patterns are relatively similar. We had hoped to find some that stood out, and we're still looking a little bit at that. The main difference is now in their activity. And we don't completely, we simply don't completely understand why we know that this, the TT that I've encircled here, which is the second to the right on the upper left diagram has a more condensed structure, but a more open crevice as it's active site. It's much, much better at generating the type of chain structure if you want in the starch that forms these gels at very, very low enzyme dosage. So we're working on this still. It's very, very interesting that an enzymatic treatment will cause the starch to gel. So applied enzyme technology research has a broad impact. As you all know, and we have both discovered new enzymes, we've done a bit of new protein engineering, and at the same time created some new scientific insight, which is never a complete, but at least we're working on it. And we're having new processes, even new molecular products, including these fuccoidins and human milk oligos for sure. And then we also have a little bit better by resource utilization by using, for example, the corn brand, Erbino Cylon. And I'd like to address, especially to the Protein Innovation Center here, we've also recently embarked on something called bio-catholic protein mobilization of plant foods. That's, of course, to remove the things around the protein so they can become a little bit more accessible. And that's also on peas, exactly the type of peas which were shown, and we call it give peas a chance in the project. And we have just initiated this, but I'd be very pleased to collaborate more on this. So that's all for today. This is the sponsors, Innovation Fund Denmark. They're the ones in Denmark that actually support our industry university research. We, of course, get significant funding from the Novo Nordisk Foundation. We've also been lucky to have some Chinese scholarship, PhD students, as well as the European and something called the KMC. That's the Danish Potato Factory Union. Thank you very much for your attention. I'm thinking about the Erbino Cylon, your first project and the ferulic acid dimers. You said, if I understood well, you can remove these with alkali treatment, but then there is some reduced bioactivity. Have you thought about controlling, but then too much is also a problem. So controlling the cross-linking through lackeys, dehydrogenase coupling system. My understanding is the free radical can be used as an electron acceptor for the dehydrogenases. Thanks for the question. Well, that's what we thought we did. But those of you, like you have been to my office, so before we embarked on this experiment, we did try to envisage up and down the wall. So what will happen? Because we believe the solubility has a significant effect. In this type of test that we did, that's just a mere kind of potential prebiotic if they can support the growth of some beneficial organisms, especially the B-phetobacteria that we will know that that will be overall considered as a step in the right direction at least. It's not a good enough experiment to just do a batch fermentation over 24 hours. But once you begin to cross-link more than what we have just released, they become insoluble and they actually form these gels. And somehow, I think it's merely that physical insolubility that does reduce the access of the enzymes from the microbes. We've seen some of the supporting results in this. We of course have the original equivalent kind of insoluble material. And these dipherulic acid-releasing enzymes, they just don't work as well on that. So I think there's something there related to the solubility. And we simply cannot distinguish if there is an effect of the type of dipherulic acid. We cannot control that with that case. It just polymerizes. I have a quick question. Okay. You showed a phylogenetic tree of the transglycosylation enzymes and said you couldn't predict which ones would have that activity. Was that based on primary amino acid sequence? And have you tried alpha-fold and looking at three-dimensional or, sorry, tertiary or secondary structure? Thanks for the question. We did select a range of enzymes around that first kind of exactly a dendrogram built on the alignment. It's not a phylogenetic tree as such. And then of course, because the assay to assess the transglycosylation is, it's certainly not a high throughput. So we first decided to say, okay, what's the hydrolytic capability? And then we use that same series of enzymes for the transglycosylation. And the technique we use for the bioinformatics, it's not that full chain alignment. It's an amino acid kind of signature comparison method where we have eight amino acids in a row and then kind of in silico chop up the protein to see if these patterns of small signatures, as we call them, if there is some kind of explanation. We've seen in other instances that the resolution that you get by this in carbohydrate processing enzymes does give a functional annotation better than the case the family will do because within each fold, there can be plenty of different functionalities just as there is. But unfortunately, at first, we didn't see any congruence between any kind of pattern. We saw this transglycosylation as spread. But we would normally do that, yes. Fold it up, maybe even dock it. You maybe even get the structure, yeah. Thank you. I have a question. First, thank you for your presentation. Very nice, impressive work. I have a question related to the synthesis of the human oligosaccharides. So, say it again, please. The synthesis of the human oligo, human milk oligosaccharides. So, is there any regulatory prohibition? What about labeling and such? Great question. Fortunately, we collaborate with the industry here, actually one particular company. And even though these are exact structural identical components, but there's a particularly large country east of here, far east begins with the sea, where this is not allowed. But they are having some tricks where they kind of evolve the bacteria to produce these enzymes and produce these components. But right now, our work here is for in vitro work and it's then for other markets, but where the regulatory access is easier. Several places in Europe, for example. Yeah. Very good, thank you. So, I have the honor to do the closing remarks and share the closing remarks with.