Georges Belfort Discusses Breakthroughs in mRNA Purification

November 4, 2025

What makes mRNA purification different from other bioprocessing methods? In this episode of Bioprocessing Unfiltered, host Alois Jungbauer, professor and head of biotechnology at the Institute of Bioprocess Science and Engineering at BOKU University, speaks with Georges Belfort, institute professor of chemical and biological engineering at Rensselaer Polytechnic Institute about his research in mRNA purification and his long, illustrious career in the industry, including the impact of his discoveries, the challenges encountered, and the lessons learned along the way. He also shares how he translates his research into products, what his lab is working on, and navigating competition from other companies.

GUEST BIO

Georges Belfort, Ph.D., Institute Professor, Chemical & Biological Engineering, Rensselaer Polytechnic Institute
Dr. Georges Belfort, endowed institute chaired professor, received his bachelor’s degree in CHME at the University of Cape Town and Ph.D. in engineering from UC Irvine. He was chairman of the managing board of the Society of Biological Engineers (AIChE) for the past 10 years.  He has broad research interests, including mass transfer and membrane filtration, protein misfolding and kinetics, single molecule force spectroscopy, RNA purification, and bioseparations. He has received several major awards in the U.S. on separations, including ACS (1995), AIChE (2000), the North American Membrane Society (2014), and the ACS Murphree Award in Industrial and Engineering Chemistry (2008). He was also one of the 100 Chemical Engineers of the Modern Era as part of the AIChE Centennial Celebration in 2008.

He was elected a member of the U.S. National Academy of Engineering (February 2003) and foreign member of the Bologna Academy of Science (Italy 2012). He was awarded (with Steven Cramer) the $500,000, NAE Bernard M. Gordon Prize (April 2025). He was awarded an honorary Ph.D. (Engr) degree from the University of Cape Town (2019). He has published over 260 peer-reviewed publications with over 20,000 citations, 25 book chapters, and has 18 assigned patents in separations science, biotechnology, health sciences, and transport phenomena. He collaborates with his wife, Dr. Marlene Belfort, distinguished professor at UAlbany, an RNA Lifetime Achievement award winner, and discoverer of introns in prokaryotes.

HOST BIO

Alois Jungbauer, Ph.D., Professor & Head, Biotechnology, Institute of Bioprocess Science and Engineering, BOKU University
Professor Alois Jungbauer received his Ph.D. in food technology and biotechnology from the University of Natural Resources and Life Sciences in Vienna, Austria in 1986. Since then, he has served as a professor at the department of biotechnology. He teaches protein technology and downstream processing and bioprocess engineering. He also acts as area head and deputy director of research in the Austrian Centre of Industrial Biotechnology. He is currently working in the field of bioengineering of proteins, plasmids, and viruses with special focus on expression, downstream processing, and characterization of large biomolecules. For more than 10 years, he has been working on continuous manufacturing of biopharmaceuticals. As a proliferate researcher, he has more than 340 publications on recombinant protein production and bioseparation, 17 patents, and 12 book contributions and recently a monograph entitled “Protein Chromatography, Process Development and Scale Up”. He is executive editor and co-founder of Biotechnology Journal, and a member of multiple editorial boards from numerous journals in biochemical engineering. He acts also the vice president of research of the European Society of Biochemical Engineering Science.

TRANSCRIPT

So I'm Alos Jungbauer, and it's a pleasure for me to do the podcast bioprocessing unfiltered with a very renowned person in the field, Professor George Belford. He has been uh around for uh several decades, yeah, an outstanding expert in the field of bioprocess engineering. But I leave uh uh to him to to uh actually uh talk about uh his career and then later on we will actually elaborate on uh the latest problems in bioprocess engineering in respect to advanced therapeutics and biopharmaceuticals.

George, please uh thank you very much for inviting me and for having me join you. Um it's my pleasure to be here. Um I'm a professor at Rensselaer Polytechnic Institute for 47 years. Um Rensselaer is uh celebrating its 200th anniversary last year. It's the oldest English-speaking technical engineering university outside the military in English, have to be careful. And um I've worked on different fields, uh, but mostly my specialty, if I have one, is uh how to purify materials, whether it's uh desalination, proteins, and now RNA. Those are the three areas we've been in for working on and focusing on in my lab.

Yeah, so you mentioned uh uh RNA, mRNA, yeah. Uh I think uh this became very popular the past uh five years upon the uh uh pandemic where they have successfully produced uh uh mRNA based vaccine. Yeah. And I think since then you have uh realized that there are shortcomings or deficiencies in the bioprocessing in purification of mRNA. Uh could you explain us a little bit how you came across uh why people use uh let's say uh concepts which are not fully adequate for mRNA purification and why mRNA purification is different to proteins or uh let's say other biomolecules?

Yes, I can try. So um so after after the during the past twenty or thirty years, uh people have focused on uh the single most important uh pharmaceutical drug produced by biotechnology, which is monoclonal antibodies, which were discovered at Cambridge University in England and uh developed uh immediately in the United States. Uh and the question there was how can you build a process that everybody could use? And the um the name for that is a platform. Yeah. And so platforms came up, and generally when you purify proteins, and I'll come back to the RNA in a minute, I just want to lay the groundwork for because all of us came from the protein field, and then suddenly we started seeing this change and how um RNA has saved some 20 over 20 million lives, probably including mine and yours, and that uh we take these uh vaccines and they're extremely effective. So the protein development over the these many years was very simple. They decided to grow, you can have make proteins either in cells, and um, and more recently we've been looking at making proteins without cells in cell-free media, together with a group uh headed by Jim Schwartz at Stanford. But we would do this the the purification, and other people do the upstream. That is upstream meaning the synthesis of the material. Um, and so hybridomas were developed, originally discovered in England. Uh these are spleen cells that have been combined with cancer cells. The cancer cells have infinite life, spleen cells don't, and they make these antibodies. And so this was a major breakthrough in the 19th, I forget exactly when was it? No, 19.

80s, 80s, 80s, 80s, early 80s, yeah.

And then after that, um they started wanting to manufacture it in large scale. I think over about 70% of all major drugs today are involved with uh monoclonal antibodies. Anyway, the idea in uh uh producing them, uh so they are generally produced in our body uh at quite high levels. And what happens is we wanted to emulate that. So we built a hollow fiber reactor in which we put the cells on the outside, these hybridomas, and then we added the food or the nutrients in the feed along the bore, and we were able to get very high levels of antibody produced by these cells. I remember distinctly um people coming from MIT and from Carv from Cornell to visit us to see how we did that. And then uh I think people decided that was too slow or the scale up was limited, and so they started making them in bioreactors, big steel reactors. And I don't want to go through the whole story, but what you do is you take that those cells, you first spin them down, remove them either by centrifuge or a membrane, and then you pass it through the most important step, that is protein A. And protein A is a molecule that was discovered um about that same time that basically binds to the FC portion, that is the leg of the antibody. The antibody is a Y, and at the leg of the antibody is called the FC portion. And the end of the Y, the end part, are three loops on each, and those loops are the binding domains. And so, anyway, these antibodies are about 155 kilodaltons. So to answer your question, um, and so when the pandemic oh, sorry, I forgot so they were captured by protein A, and then there were other post-translational uh tests they made through iron exchange, they clean it up, and they have this uh product, um, both with uh chromatography and with um membranes. Now, chromatography is the prime separation mechanism and has always been with proteins. Membranes are the cleanup version. They clean up at the end or they clean up at the beginning, but they aren't the prime uh uh uh step. So now comes the pandemic, and suddenly um uh the two companies, Moderna and BioNTech in Germany, um started uh got got this, they got the sequence from China uh in January, I think it was January 10th. January 10th, and then it was 41 days that Moderna sent the first vial to NIH to be checked of a uh vaccine material that they felt could be uh used as a future vaccine. It was basically purified RNA. In that very short time, the question was how do you purify this material? How do you first synthesize it now you're pure? So I told you that monoclonal antibody, which is my control, was made out of cells made them, hybridoma cells. But the easiest way to make RNA, messenger RNA, these are linear single strands of RNA, like a piece of string, and it has segments in it that are folded into bulbs and whatever, and it has some double-stranded component, but mostly RNA has lots of um hydrogen bonding components from the building blocks, the four base pairs. Proteins are built out of twenty building blocks, so they're more complicated. RNA is built from four building blocks and less complicated. However, this molecule is gargantic, it's ten, nine to ten times bigger than a monoclonal antibody. And that's not trivial because if it's very big, it doesn't have a diffusion coefficient. That means it can't diffuse well into beads. Remember I said beads are the dominant separating media. So what they did was they drilled holes through beads, so they're porous beads, or they made beads with bigger pores to try to get these materials in. But when you do that, you lose surface area, and there are other problems. When this material, these very large RNAs get into a bead, what happens is they could basically occlude or inhibit more coming in, because they could be at the entrance, and there are lots of problems.

Um material would really block block the poor. That's correct.

It's big enough because they would would would block they could block it, so so you don't use all this area to to to bind. Like you can with proteins, which is much easier, they're much smaller and they can get in. So the question was, what did they do? Well, they had no choice, they had to do it really quickly, and so they looked in the lab and they saw, ha, there are there are two methods that we could use. We could put, so I have to go back and tell you what RNA looks like. Messenger RNA has several components to it. Remember, the reason we need it is to clone a protein inside your body. So when you put the messenger RNA inside, you actually it gets red uh and and it can read off the protein that you want to buy. And so what we what generally we're doing with this is to clone the spike protein on the outside of the COVID-19 virus. That spike protein you want to teach your body to build up resistance to so that when the real virus comes, you know all about it. You can fight it. So I think the idea was we better use high performance liquid chromatography, high pressure. It's very high pressure, it works in the lab really well, but it's really complicated and it's really expensive to make very large-scale systems. So the crew or the the the real uh comment the real breakthrough was that the US government gave over hundreds of millions of dollars, maybe billions of dollars, to these two companies so that they didn't care which technique to use. They used the only the ones that worked. They used scaled up ones, even though they were expensive, and they used that one, but they used the other one, the oligo DT also.

But it was actually uh uh on uh uh they're more say rush. Yeah. That's uh if they would have uh time for a year or two.

Well then they could have done slowly, yes.

If they could have done slowly, they had to take what what was available. Yeah. Just to to to not to blame them, no, no, no, but uh but I think they had to do that to save lives.

To save lives. Because the people were dying all over the world. So you're right, they had to do this. They took whatever they could. Yep. So the thing, the two things they took was they took an affinity uh bead with a pause in it, and they put oligo DT in it. T they grafted it on, and T binds to the tail of uh the messenger RNA. Messenger RNA is like a dog. It has a tail, and the tail is made from one of those base pairs, A. And it's like fishing when you want to throw it in and you catch the fish you want, and it's that tail that is caught selectively by the oligo DA, DT, on these beads. And so it worked beautifully, except it's only about one to five percent efficient. One to five percent efficient. So that's very expensive. As one person once said to me, we're spending millions, if not billions, on these beads. Yeah, and um so that's the story of where I come in. So when when you ask me, so how does it work? So we decided, I I've always tried to move orthogonally to where everybody else is going. If everybody's working on this problem, I said, that's fine. Uh, there's too many people, I want to go in this direction. So I tried to go in 90 degrees from them. And I thought, you know, I'll write a short grant. Um actually I didn't even know about it. My colleague Todd Probitsian walked in my office and said, George, the US government is going to give money for uh working on the purification of RNA. Why don't you write a grant? So I said, All right, what did I write on? So I said, I write it on uh using membranes instead of beads. And it turns out it was a very good move.

Yeah, but it's clear because you had uh uh many, many years uh of uh uh experience uh on membranes and you know with proteins. We use them between proteins, but you know how membranes uh uh work, yeah. And what's wrong with them and what's wrong with members and the challenge. And how to chemically modify them and so on and so forth.

Yeah, so I think uh So that's what we did for 20 years when we're modifying membranes. So we thought we could easily modify them and put the oligo T in a membrane and push this very large molecule through this porous media and not allow it to diffuse. So you just push it through, capture it, and see if it works. So there was an interesting problem. I wrote the grant, it was four pages, I'll never forget it. And I think three I forgot about it. And about three months later I got an email to say you've got the money. And I couldn't believe it. There were several problems. I never had messenger RNA, I never had the right membranes, I never had the oligo DT, and I never had a team, and we had no RNA lab. So we had to solve all those problems. I didn't sleep for a few nights because I wanted the money, but I didn't know how to get all of this. So I immediately started phoning CEOs of companies, the top person, not the second and the people I know. I didn't even know some of these people. And they all said, we'll help you. Everyone. It was unbelievable. And Milipore, uh at least, excuse me, IDT was the best. They said, we'll give you all the oligos you need, we'll give them you with FAM on them. FAM is a fluorescent tag, so that then you know you can see the tag with what you're binding and how you bind. So we worked with them for many years now. They've actually they still, we have an agreement that just ended about a month ago, where they gave us everything free for five years, four years. Now they want us to pay. But anyway, it's been wonderful and they've been wonderful. So I think I then phoned up the membrane company, CITEVA, and they said they have a group in London, and they'll give us what, what, what, what are they, Wattman? Yeah, Wattman membranes to sell. They are cellulose membranes, yeah, and they are excellent. Then I phoned up and I got in touch, I was a consultant some years ago with um a company in Japan, and um they make uh uh hollow fiber modules for virus separation. So this is asahi. Asahi. I was a consultant to them in America, and so they sent me, they sent me membranes too. So suddenly we had the membranes, we had the oligos, and now we had the but but it's not trivial to build an RNA lab. And so the question is, what am I gonna do?

But what is the what is the challenge to maybe for uh what is the challenge for the uh RNA lab? Yeah, what is different to the RNA? RNA does not live long in the environment.

On our fingers, and everywhere we are, are are Pac-Men that eat RNA. So we have them all alone, they're all over. So you can't just start like proteins, put it in a glass and start running. You have to wash it with special water, get rid of the RNAs, you have to do everything on the table, you have to make it, it's a special lab. However, I was very lucky. My wife is a famous RNA person. So she said to me, I will arrange for you, come across, bring your team, and we went across 16 miles to the other university where she works at U Albany. It's just across the road in Albany, and we're in Troy. So it was easy. We drove across, and there's an RNA institute there. And the director, Andy Berglund, is a wonderful person. And he said, George, we'll help you do everything. We'll tell you what to buy on your grant, we'll teach you how to run a lab, you bring your students here, we'll do it all. So they did it, and within two months we were running with a complete RNA lab.

So the your your network, your connections actually uh accelerated the project.

I I I don't know, I was lucky. They were nice people. There were lots of very nice people. They could have said, I'm sorry, find somebody else. Actually, one company did. When I asked them for RNA, they said you have to ask somebody else. So we did. Okay. So but we got we got everything, yeah. But it was just lucky too, you know. I think a young person would hesitate to find a CEO that they don't know in a company. I mean, the what did what happened with IDT was um they're in Iowa. They're the single major producers of uh oligopeptides in the United States. So I asked my wife, who's the top person? She said, I don't know anybody, but they're the people you want to go to. So I phoned up the CEO. I said, I don't know you, but I've looked up your background. You're a biochemist. Now you're in the business world. I need your help. And he said, I'll help you.

Great. That was it. Yeah.

So it was wonderful. They were wonderful.

No, so I think what uh we have seen also with your presentations, your students' posters here at the bioprocessing summit, yeah. You have uh uh really progressed a lot, yeah. Uh you are not yet fully there, yeah, but it's really impressive what you have uh achieved in the uh uh past uh couple of years. Yeah. Uh could you uh tell us what uh uh challenges, what open scientific questions you would now uh uh uh uh tackle uh in this uh let's say next period. Yeah? Because you successfully uh developed a membrane with oligotity for affinity purification. We have seen you have uh peptides selective for messenger, kept messenger arena, and so on and so forth. So what is left?

Uh so there's a lot left. And by the way, none of it was done by me. It was done by my students and postdoc and my collaborator. So I have to give them credit. So my collaborator, without him, we couldn't do the peptides. His name is Pankaj Garande, and his ex-student, who now is a postdoc in our group, with him, he's also here in our group, and he's he and she, her name is uh Ming Young or Claudia Hu. And uh so she did all this discovery. And what we did was we took her peptides, grafted them on, and the student that did that is Riddie Bunnock, and then once we've got it on, the membrane, we now had to test it, and that is Thomas Newman. However, that wasn't enough. We also wanted to know to prove that if we if we produced or discovered a peptide or ligand, like I call a hook, if this hook can catch the right thing, we need to find out. So we in my group have done for 20 years, actually 1989, we published our first paper on single molecule force measurements in PNAS. So that was to look at uh at lysozyme, see how it actually sits on a surface. Now we can do it very easily in the lab. The student's name is Surya Kala, and he actually can measure forces between a a um RNA with a cap or without a cap and a small peptide or an oligo DT, any hair monula, and we can measure the forces. So that was one of the posters we showed. Many people are very interested in it because it's a direct measurement. And you can also change the, if you want to know how it not only how the forces are, but how you can get it off elude it. Because we need to elude it to catch the product. You can change the buffers and find the optimal buffer.

My question is now, what is translate your findings, your research into a product or into, let's say, into the bioprocessing industry?

Well, I think there's several things that could be done. Um, first of all, I think the big product would be um, we sort of think I was I was approached today about starting a company um by a British scientist who said to me he started a company, why don't you? So I said, Well, I'm gonna speak to Bob Langer about that and see if we could start a company. Um But really, I I'm not so interested in money. I'm interested in technology and and discovery and how to do research. Um So I think the first thing is we need to make these membranes commercially available. If they're not commercially available, it's not good enough to have it in a university. And I don't know if the companies want to are interested, if they are or they're not, we will see. So we also have a method, a new method, with Thomas and Riddy, in which we don't use a hook. It's got no hook, but we can still catch the fish. And we do it by by actually using the difference in the properties between the unwanted immunogenic, that means causing an immune effect in the body, double-stranded versus a single stranded, which you want. That's the vaccine or therapeutics. So I think the future is we need to have a membrane. Now the second thing is we need a good module. The modules need to be designed. Module is a box in which you put the membranes in. It turns out these modules that are built today have quite a lot of dispersion in them. And dispersion is a fluid mechanics mass transfer problem in that you get mixing. And you don't want to get mixing because for affinity, you want it to go through even. You want to have the black fluid. That's right. Yeah. It has to go through the pores. Even you can go mix, have some going through, some not going through. So we we have designed and patented a new small test cell, both for holo fiber. It's the first hollow fiber replacement. You can put a second olive fiber and you don't throw it away. Currently, you have to throw them away because there's uh the companies like um uh the Japanese company, yeah, Asahi, they sell you holo fiber, but then you have to throw it away. We can take the olive fiber and replace it. So we have a pattern on that and we built it, and it works.

But I'm I'm uh interested. Uh companies have approached you uh already about your findings, discoveries, and so on, or not? Yes, yes, yes.

Some have. I can't tell you who, but they don't want me to say it. No, no, it's fine. That's clear. Some have asked us, you know, what are you doing? Can you work with us? And um so that there are other possibilities, by the way. I've told you oligotea is a hook, peptides can be hooked, and our peptides we can make a hook against any fish we want, any molecule. And thirdly, the non-ligand method. But there is a fourth way that people are talking about, and that is to use organic chemistry to make small organic hooks. It may be cheaper, it has a much bigger monomer library, 500 instead of 20 peptides or four in in so it's possible. And there's a group in England that uh approached us. There are other interesting things.

Um I think I I I think your peptides already very are very promising. Yeah, this what I see. This is a really promising uh yeah uh uh affinity line. Yeah, that's what it is. Because but there is competition. There is some competition. Uh competition with from other groups or other companies.

One other, well, there's a company in San Francisco that does that. But I don't think they do what we do. We don't only look for the peptide. We do a full analysis of the simulations of the molecules binding. So we want to look for binding, and then we start to see that half the peptide is not binding. So then we change the amino acids to make it bind correctly.

So we do that binding efficiency that improved the because this would reduce the costs of that.

And also reduce how much you need.

Yeah, it's clear.

Yeah, yeah. Yeah. But there is one endemic problem with these peptides. So I and also OligoDT and all of these uh binding systems, and I just want to tell you what it is. When you have a very large molecule with a tail, like a very large fish, let's say it's a whale. You're catching a whale, it's got a tail, and you can catch it by the tail. So you put you and you want to catch it on a surface, say on the carpet. What you do is you cover the whole carpet with these ligands, and they will catch the tail. But only two or three of them will catch the tail. The other ninety-nine percent is in the shadow underneath. It doesn't do anything. You've wasted your time, your money. But the question is, and this is what we're looking at now, can you minimize the amount of binding and can you make less of it? And then you can actually bind only and not pay for so much and still have statistical binding. So we're we're looking at that very carefully now. How can you optimize and minimize the cost? But I do want to tell you there are two other opportunities that. This whole field that's coming up. One is circular RNA.

Yeah.

So there is a new company develop uh called RNA Vait in Cambridge University, next to Cambridge University, started by the Nobel Prize winner Venki Ramakrishnan, who actually is a colleague of my my wife. So we've been in touch with him to talk about how to purify circular RNA. There are a lot of opportunities there, and it's not easy. Second, when you make circular, sometimes it gets nicked. And the nicked molecule is identical to the circular. How do you separate it? Oh, that's really that that's so I think there are ways to do it. Um I don't want to give away all the ideas, but there is a way to do it using a physical separation. And thirdly, and by the way, you could also bind stuff to the two ends because the one has ends and the doesn't, and if you react something to it, you may be able to separate it that way. But we're looking at this problem, and then there's self-amplifying RNA.

Yes.

I have a big worry about that molecule. I'll tell you why. It's a molecule that you put in and it has a segment, it's much bigger than the RNA we're talking about. Instead of being 4,000, it's 10,000 base pairs. So it's another ten times larger, it's massive, five times larger. The problem with this is not so much its idea, but the concept is you really want to put a molecule in your body that never that keeps self-amplifying and keeps going. Maybe FDA won't be so pleased with us to go. So it's been approved in India, but not yet in the United States.

I think uh my impression after this uh discussion with you or what you have uh uh shown us, uh you have really developed uh enabling technology to reduce uh the cost uh of goods and manufacturing costs of uh mRNA molecules, which are uh have a great future and also uh may really benefit uh uh uh mankind, yeah, not only as a COVID or influenza vaccine, uh cancer vaccine, uh, and so forth. So uh uh I thank you uh for actually thank you very much for inviting me, and thank you for it. I hope I gave you some ideas of no, I think these were brilliant ideas uh about mRNA purification and Alcing process.

Stay Connected