Episode 14 of the Modern Chemistry podcast is a deep discussion with Elisa Fadda about all things glycosylation, and in particular, how glycosylation of the SARS-CoV-2 spike protein plays an essential role in the infectivity of the COVID-19 causing virus. Since 2014, Elisa has been a lecturer in the Department of Chemistry at Maynooth University.
Elisa Fadda obtained her PhD in 2004 from the Department of Chemistry at the Université de Montréal under Professor Dennis R. Salahub. From May 2004 to May 2008, she worked as a post-doctoral fellow in Dr Régis Pomès group in Molecular Structure and Function at the Hospital for Sick Children (Sickkids) Research Institute in Toronto. From June 2008 until May 2013, Elisa worked as a research associate and honorary research lecturer in Prof Robert J. Woods group in the School of Chemistry at NUI Galway. In 2013 she was awarded a Post-Graduate Certificate in Teaching and Learning in Higher Education from the Centre for Learning and Teaching (CELT) at NUI Galway. In August 2013, Elisa became an Assistant Lecturer in the Department of Chemistry at Maynooth University, taking on a Lecturer position since 2014.
You will hear the following terms used during the interview. I've included some descriptions here.
The publication we refer to early on in the discussion is available at https://www.sciencedirect.com/science/article/pii/B9780128194751000560?via%3Dihub. A full list of Elisa’s publications is available at her group website.
Elisa is contactable on social media, and you can find her on
LinkedIn https://www.linkedin.com/in/elisa-fadda-a012b194/ (although, Elisa admits, she's rarely on LinkedIn)
On Twitter, search @ElisaTelisa
The group website is https://efadda73.wixsite.com/elisafadda
Our theme music is "Wholesome" by Kevin MacLeod (https://incompetech.com)
Music from https://filmmusic.io
License: CC BY (http://creativecommons.org/licenses/by/4.0/)
Connect with me (Paul) at https://www.linkedin.com/in/paulorange/
H.E.L. group can be found at
online,
on LinkedIn at
https://www.linkedin.com/company/hel-group/
on Twitter, we're @hel_group,
or search for us on Facebook
Elisa Fadda (00:01):
I'm a chemist by training but I, let's say, abandoned chemistry as a... If you want to consider it as a pure science, what chemistry is supposed to be if you ask a synthetic chemist. 20 years ago? It's just I've been doing mostly biophysics, so since then, quantum chemistry first and then biophysics.
Paul Orange (00:34):
But in the paper you sent me, the very first part of the abstract I thought was just a really fascinating place to start. What it says is, and I'm sorry to quote your own work back at you but, "The chemical nature and heterogeneity of most complex carbohydrates makes their structural characterization very difficult if not possible through experimental biology." I just thought that's such a bold and impressive place to start from where it says, "Chemistry or experimentation in the lab will not tell us enough-
Elisa Fadda (01:07):
No.
Paul Orange (01:07):
... for what we need to know." What does the computational side allow you then, that lab techniques won't allow you to do?
Elisa Fadda (01:15):
Yeah. Fundamentally, there are limitations that for, let's say, experimental characterization of these glycans or complex carbohydrates in general, the longer the worse. Due to the fact that chemically they're extremely flexible molecules, they're all single-bond linked. Some of those bonds are more mobile than others, so if you consider that on top of that, carbohydrates unlike proteins or nucleic acid are branched polymers. They don't go linearly, they can go any direction. They have different hydroxyl groups that can be functionalized with other branches. All of them can move independently relatively to their own, let's say, structure.
Elisa Fadda (02:00):
This allows them to move within a time scale that is so fast that experimental methods are not able to capture them, to take a picture of them. [inaudible 00:02:13] protein is actually stable, especially in cryogenic environments. You can crystallize it and you can actually-
Paul Orange (02:22):
Sure.
Elisa Fadda (02:22):
... leave it there and image it through X-ray crystallography or cryo-EM. Meanwhile, the glycans retain a lot of dynamics within also cryogenic environments so you can see them. It's like taking a picture of a moving object, let's say a running dog. What you see is just a blur if you don't know that is a dog because it just moves faster than the shutter speed of your optical camera. You just can't tell, so what you have is just a blur.
Elisa Fadda (02:55):
What computational methods allow you to do is to fit a structure, like an [inaudible 00:03:02] structure within this blur and capture the dynamics of these molecules within that real-time scale, let's say nanoseconds to microseconds. Microseconds pertaining to glycans that undergo particularly rare conformational transitions or changes [inaudible 00:03:22]. No experimental methods per NMR can give an image or give you information at such a rapid time scale. That is actually the domain of molecular dynamics or classic simulation methods, so methods that are based on micromechanics or classical physics.
Paul Orange (03:48):
When I think back to when I did my degree, the statement was, "Oh. Proteins, yeah. Structure is very important for function and you've got the primary through the quaternary structure. Oh, and they might have a sugar stuck on somewhere as well." That was almost like an asterisk sub-note in the text. But looking at the work you've been doing... Or maybe we should just say this is on viruses but obviously very relevant at the moment is the spike protein on the coronavirus. It's amazing how much of that structure actually is made up of these glycans and the effect it has on the function.
Paul Orange (04:24):
I think I understood what was in your paper, but people don't want to hear me explain it. Maybe you should explain it and make sure that it is right. But, fascinating.
Elisa Fadda (04:35):
Thank you. A whole lot of proteins in general that undergo the circulatory pathway are glycosylated. Something that hasn't... Well, it's been understood for years and years, however not really properly appreciated by, especially by the structural biology community, for different reasons. Because fundamentally, let's say, up to now where cryo-EM is all the hype, rightfully. I'm a big cryo-EM fan. I want a cryo-EM tee shirt.
Elisa Fadda (05:10):
Most of the CBD structures are from X-ray crystallography. In order to crystallize a protein the first thing you do is shave off the glycans because otherwise you won't get protein crystals whatsoever at all unless the glycan is bound to the protein so it's fixed to the protein, but that's a different process. Because in the PDB you could not see these glycans that were magically not there. It is an issue that stems, I think, from this process but it was not intentional.
Elisa Fadda (05:46):
However, glycans are... Two to three percent of the human genome is dedicated to glycosylated proteins. That is a process that it has an enormous complexity to it itself. The variety of glycans that you can have on the proteins are enormous. Proteins that are external to the cell, on the cell surface, inside the cell are glycosylated, they can be glycosylated. These glycans do stuff. Obviously, the cell would not invest such an amount of energy and evolution in order to get stuff that is a mere decoration.
Paul Orange (06:29):
Sure.
Elisa Fadda (06:30):
What has been discovered through the years through tireless work of different type of scientists because glycans are not template-driven. There is no information in the DNA that tells you, "That glycan is this type of glycan there." So have revealed that glycosylation of a protein through the ER and Golgi allows a protein to fold, preserve its structure and trafficking of the protein outside the cell. This is intrinsic properties of the glycosylated protein.
Elisa Fadda (07:06):
Then after the protein is either embedded within the cell membrane because it's a member of protein or is secreted outside, the glycans can perform a lot of functions, an enormous type of functions that are, let's say mediating protein and protein interaction. They actually represent the first port of entry for infection, so toxins, bacteria and viral infection. Okay?
Paul Orange (07:35):
Yeah.
Elisa Fadda (07:35):
And we had COVID. Well, did we get to viruses in general, enveloped viruses? They are viruses that are actually wrapped with a membrane that is called an envelope because they're [inaudible 00:07:53] and they present outside these fusion proteins that are called spikes. They are called spikes because they come out like this. This is not just the coronavirus. That is all enveloped viruses. The spike proteins are all different. They have different mechanisms that allow the virus to enter the cell, so that's why they have to be external, because those are the trigger for the virus to get in.
Elisa Fadda (08:22):
One thing that all of these spike proteins have in common, aside from the fold, they look, a lot of them, like mushrooms. If you see the HIV gp120, gp41, it looks like a stubbier mushroom than-
Paul Orange (08:38):
Right.
Elisa Fadda (08:39):
... a spike protein does, that's in the coronavirus. It's that they are extremely heavily glycosylated. As you said, these proteins are covered literally in glycans. The coronavirus is actually less covered than the HIV, like gp120 and gp41.
Paul Orange (08:58):
Right.
Elisa Fadda (09:00):
You can barely see the protein in those [inaudible 00:09:04] proteins. They are completely fur, covered in this glyco fur. That is evolutionary advantageous for the virus because it camouflages it from the immune system. Being outside, being exposed to the immune system as a virus enters, if it didn't have a cloak or what is called a glycan shield, immediately the immune system will recognize it as a foreign thing. Meanwhile, the virus, what does it do? During replication uses our own cell machinery, or whatever host cell machinery to glycosylate, to use the host glycans and it puts them on its own proteins so that the immune system doesn't recognize the virus because it's actually dressed up as a human. Right?
Paul Orange (10:01):
Right. Right.
Elisa Fadda (10:02):
Because it's covered by stuff that the immune system sees all the time. So it's just like, "Well, this is cool. Let it go. That's fine." What the immune system recognizes is protein epitopes or glycan protein epitopes, so it can see a bit of the proteins that do not belong. At that point, there is an immunity response.
Paul Orange (10:26):
Sorry, I was just going to say, when these viruses replicate and they put this glycan coat around themselves, it's not that the immune system is no good at recognizing anything with a glycan, it's that it's seen this before and it says, "Oh, no. This looks like one of our own cells so we should just leave it alone."
Elisa Fadda (10:44):
Yeah, yeah. The coronavirus glycans are very human. They're very human like. Right?
Paul Orange (10:52):
Yeah.
Elisa Fadda (10:53):
Meanwhile, other viruses that have a much denser glycosylation, they have a slightly different glycosylation in terms of the type of glycans. They're still made by human cells but because they are so dense, they cannot be really elaborated much so they would represent something that, for example in humans, is not that common.
Elisa Fadda (11:17):
A type of glycan that is not that common because human proteins are not so densely glycosylated. But the immunity response, they're very low in immunogenic, in that sense. It's a very good disguise, unfortunately. Depending on how much of the protein is covered, it's hidden. It makes it really difficult to build an immunity response and design a vaccine for it because [crosstalk 00:11:47].
Paul Orange (11:47):
I know you've been talking about this and I think there was a phrase you used in your paper which was carbohydrate force field. As a science fiction nerd, that really resonated with me. Am I right in thinking in what you said in the publication as well, that the glycosylation coat isn't just about the virus being able to stealthily infect new cells, there are actually-
Elisa Fadda (12:10):
Mm-hmm (affirmative).
Paul Orange (12:10):
... important functional roles that it plays in terms of enabling the spike protein to bind to the other host cell?
Elisa Fadda (12:18):
Yeah. This functional role of the glycan shield is something that we discovered recently in collaboration with Rommie Amaro's group with USCSD and it's unique to this particular virus and entry. What does this spike do? What the coronavirus spike has to do is open in order to initiate the process of entering the whole cell by binding a receptor that is a primary receptor, it is called ACE2. This opening is actually a complete, a very complex conformational change of the actual protein architecture of it. Because of its actually protein architecture, the glycan shield, what does it do in this case? It actually compounds it, it stabilizes it in the open confirmation like putty. You know?
Paul Orange (13:21):
Yeah.
Elisa Fadda (13:23):
Because otherwise there would be holes everywhere and nature, especially at that level of small, tiny, tiny things at that level or scale of dimensions, as soon as there's a vacuum or a hole that is created, water floods in and it breaks everything apart. What we discovered through multi-microsecond molecular dynamic simulation that required extremely large computational resources to do is that the glycans actually support this open spike in a way that depends... And this is episode two that we've just [inaudible 00:14:08].
Elisa Fadda (14:08):
That it depends on the type of glycosylation. So not all glycans are able to do this. This type of glycosylation actually affects the type of support. It's like a crutch. If you think that the glycan is propping up this part of the protein that needs to bind the receptor... If you think about it like a good crutch that fits your armpit would be fine but if the crutch is up to your hip, it's not going to do much [inaudible 00:14:41]. Or if there's no crutch, there's no opening. You'd just fall down.
Elisa Fadda (14:46):
What we looked at is how the different types of glycans or different types of glycosylation can affect that, and that's as repercussional and therapeutic interventions. As I said earlier, there's a lot of enzymes that are dedicated to remodeling these glycans in our own cells. Take them down, put them up, decorate them variously so [inaudible 00:15:17] actually. Selective inhibition of different pathways. You can actually create a type of glycosylation that is not functional at all. That's the work that we've been doing so far in order to understand what this glycan shield does within the context of the coronavirus.
Paul Orange (15:40):
Taking the coronavirus as an example then, if it's got the type of glycosylation on it that it currently has, it means that the virus circulating around our bodies has a good chance of infecting a cell and causing the disease. But then, if you change the glycosylation or I guess mutate it in some way, actually although the virus may be going around in your bloodstream, it doesn't have a chance to bind to your cell and give you the disease and at some point your body will just eliminate that virus.
Elisa Fadda (16:12):
Yeah because it's useless. It would be not as effective. It would not create havoc. [inaudible 00:16:20] not replicating effectively so the number of viruses circulating would be much lower. It would be a mild infection. As you said, your body gets rid of it. How? I don't know. I'm not an immunologist or a medical doctor. What the virus does fundamentally, in my understanding that's based on viral evolution, is to make itself more and more effective and replicate. That's its job. Killing the host is not the primary. Actually, no, it's a side effect because a dead person or a dead carrier is not a good carrier for replication but to allow for maximum [inaudible 00:17:08] that's in replication. If you start meddling with it, it would not do that. I guess that would be a rubbish virus.
Paul Orange (17:21):
Although obviously everybody is very focused on the coronavirus at the moment for very obvious reasons. It's not limited just to that virus, right? I think you've shown in this paper, was it HIV and a flu virus as well? Yeah.
Elisa Fadda (17:37):
Yeah. In that review I was looking at three cases which have been tackled extensively with computational studies and where computational studies have provided the missing [inaudible 00:17:57] information, especially the dynamics of the glycan shield that is important to understand where do we look for therapeutic strategies or for vaccine designs, to optimize vaccine designs and so on. It is a characteristic that all envelope viruses share, like herpes virus, Lassa virus. They're all envelope viruses so they all have these very glycosylated spikes.
Elisa Fadda (18:35):
What we found is just actually something that it's important also to underline that everybody talks about, mutations of the protein. Well, I think if there's something that influenza teaches us is that the glycan shield mutates quite effectively. As you know, the glycan shield can change the epitope specificity, can change affinity and this is actually... The evolution of the glycan shield is something that is happening in SARS viruses. There are differences, for example, fundamental differences that we looked at between the coronavirus 1... Well, the coronavirus because we thought there was not going to be the episode two, in 2003 that is very similar to the coronavirus that we have now that actually have conferred more potency to this virus. Possibly a more infectious virus. There are probably several other types of consideration to be made.
Elisa Fadda (19:49):
There have been several changes in the glycan shield that actually optimized the function of this, the activity of the spike. We have to look at the whole evolution, just not about the protein evolution itself. What makes it more active? What makes it more contagious? Yeah, in the whole scheme of things.
Paul Orange (20:19):
Yeah. I think what you were saying is really interesting in terms of the multidisciplinary approach to look at the virus. I know in the UK there's been a lot of media about the fact that we've got all these great sequencing capabilities and we're sequencing variants. That's highly useful to do but if that's not telling you how the thing looks and operates because of the glycan shield then changing as well and you can't... I don't know how much of that is detectable. You need to bring all those additional factors into the equation.
Elisa Fadda (20:54):
The large majority of the glycan shield is made of a particular type of glycosylation that's called N-glycans. N-glycans do not occur everywhere in the sequence of a protein. They occur only in correspondence of what is called a sequon, so there has to be a sequence of three amino acids in a row. Any [inaudible 00:21:22] glycans because they link to an [inaudible 00:21:25] that is in protein lingo as N. So N is to be followed by whatever amino acid, so we'll call it X, [inaudible 00:21:33]. Okay?
Paul Orange (21:34):
Mm-hmm (affirmative).
Elisa Fadda (21:34):
And either a [inaudible 00:21:36]. From sequencing you can tell what sequons are there in every protein that you can sequence. Some of the sequons are occupied, so they might be glycosylated, other sequons are not. It depends on where they are in the structure. From sequencing a protein you get a string of amino acids. If you don't have a structure, you don't know where these things are. They can be inside a protein, outside a protein. All of this glycan shield is outside because it needs to be reached by proteins and enzymes in order to put them there.
Elisa Fadda (22:22):
But you can tell if you have a protein that, if a sequence appears or disappears. So in a mutant you would tell if there is a new glycan or an old glycan and some positions might suppress other positions. The glycan shield, as we described in our paper in collaboration with Ben Schulz at the University of Queensland, it's undergoing extensive mutation. The positions are appearing and disappearing and some important positions that disappeared recently, especially between CoV-1 and CoV-2, have allowed the virus possibly to actually use even a different mechanism to be even more infectious.
Paul Orange (23:11):
Right.
Elisa Fadda (23:11):
This is also, again, glycan-based. The cell is covered by glycans. Not only the glycans that I described as N-glycans before but a lot of different other glycans are very large [inaudible 00:23:28]. They're [inaudible 00:23:28] glycans they are called. If you imagine that all the cells are super furry, super furry. A virus needs to actually like fur in order to get in there. It needs to interact positively or proactively with this fur and because of some mutation, we think, that have changed the glycan shield within this virus. The coronavirus 2 has acquired an uncanny potential in interacting without this fur.
Paul Orange (24:08):
Right.
Elisa Fadda (24:09):
What happens is, well, the virus goal is to reach the receptor. There is a cell-bound receptor. The cell-bound receptor is within this fur, so as soon as the virus approaching really likes the fur, binds to the fur and eventually is going to find the receptor. A virus that is not too keen on the fur will find the receptor less.
Elisa Fadda (24:32):
There's a method that a lot of viruses use for localization, so they'll localize themselves onto the cell. It's an evil machinery but it's an evil machinery that can be understood and should be understood and all these mechanisms and... I don't know. I don't want to say [inaudible 00:24:55] it is. I'm not on the virus's side, but it is extremely clever. It's all based on glycoscience, like the glycobiology and glycans interactions between the virus and the cell. If you do not take that into consideration, you're just missing the whole plot. You're missing the story completely.
Elisa Fadda (25:19):
That's why during this year, one of the main missions that myself and many other glycoscientists had was just to promote as much as possible the importance of glycans in biology and their fundamental role in viral pathogenesis that is sadly disregarded for by the mainstream science or biophysics in particular, where I belong till now.
Paul Orange (25:54):
But there's been quite a lot of press interest you said since you've published this work, right? And obviously with what's going on in the world.
Elisa Fadda (26:03):
Yeah, there has been because I think it's just what we showed is actually the protein that has been on everybody's mouths. People that knew that were in the business and people that were not in the business, it doesn't look anything like what has been shown before. It is completely covered in glycans. It was a missing part that nobody has highlighted before. Like, "Okay. Just actually look at the murderer." You know?
Paul Orange (26:36):
Yeah.
Elisa Fadda (26:36):
"This [crosstalk 00:26:36], not that guy. It's the other guy." It revealed pretty much clearly, very clearly and obviously that you just needed to use that model, a fully glycosylated protein in order to understand how to win this war because if you don't, you're just looking at a wrong murderer, wrong [inaudible 00:27:06]. You're just actually probably diverting a lot of energy and effort into something that is not even going to be as successful if successful in the first place.
Elisa Fadda (27:18):
The second part, that actually was probably received more from more in-the-know people as that this was the first time the glycan shield was revealed as functional in the work of a protein. Usually, as I said, glycans might be either directly involved in the binding... So the protein function, let's say that it has to encounter another protein in order to do something, so they might mediate this encounter as this is an immune response. For example, glycans are really important for that. Glycans and glycan interactions, all that sort.
Elisa Fadda (28:00):
Or for the protein to exist in the first place as a folded protein because they facilitate folding. Or trafficking so the protein finally gets out where it belongs. But they've never been, like in viral protein, implicated in the actual mechanism of infection. The glycans are as guilty as the protein in making this machinery perfect as it is because well, a non-glycosylated protein A, does not exist. I'm sorry for all the people that actually worked on all the non-glycosylated protein from a functional point of view and the structural point of view because they cannot see them, so that's what it is.
Elisa Fadda (28:49):
It doesn't exist because A, it wouldn't fold or it wouldn't look like that all. B, it just wouldn't work. It just would not work. To answer again. Sorry, long-winded answer. I think just the press interest was because we imaged what the actual protein looks like in all its beauty and in all its dangers.
Paul Orange (29:17):
Yeah. Given where we are at the moment, people are hearing a lot about infectivity, our numbers. The popular press is reporting all of that. I suppose you've shown the cellular and molecular mechanisms that are actually making that happen down at the very sharp end when a virus gets into a cell. I really applaud the work that you've done in being able to show that to people, that it isn't just about like, "Did somebody cough on me or not?" Or, "Did I wash my hands properly?" That's an entry mechanism. It's also how the viral coat is structurally designed that's supporting the function of getting-
Elisa Fadda (30:04):
Yes.
Paul Orange (30:04):
... into the cells and causing people to be ill. So yeah, I think that's a really beautiful part of the work that you've done here.
Elisa Fadda (30:12):
[inaudible 00:30:12].
Paul Orange (30:11):
It's great.
Elisa Fadda (30:13):
We're very proud of it. It's been a great collaboration that is ongoing in so many directions still. One of the positive things with this year, year and a half at this point, within the science community is that everybody pulled in together. Especially in my experience with Rommie's group, with Jason McLaren's group, where authors of this work and with Ben Schulz with the work on the other part of the work and contacts through conferences, through talks, through chats and whatnot. Everybody pulled in together.
Elisa Fadda (31:00):
Our own, let's say, glory or our own aims were completely disregarded. It was just like, "What can I do to help? How can I help by working with somebody else that has more expertise in that particular field?" Which was essential. I don't think that any of us singularly could've done this work to the completeness that it was done at. This is a lot scientists have all worked together. We abandoned everything we were doing in March last year.
Elisa Fadda (31:35):
Everything else was just put on the side and probably funding agencies are going to be like, "What?" But it's just we dedicated all our resources to this, 100%. This also goes to a testament to the dedication of the students who were always at risk literally because they have a project, it's funded. They were like, "Okay. So should I shelve it and then work on the coronavirus? What's going to be for my PhD or my degree?" That went really well.
Paul Orange (32:15):
I completely agree with I think everybody I've spoken to on the podcasts this year. That's something that's come through is exactly like you say that they'd stopped what they were doing, "How can I help?" In all sorts of ways. I think that comes through, that when we pull together we can do amazing things.
Elisa Fadda (32:35):
[inaudible 00:32:35]. That's a lesson that I hope people will remember. I definitely will. It doesn't exist to do things alone anymore. There's no reason or no point. What I can do I can do well but it's limited. So if I join forces with somebody else that has expertise in such-and-such, we can do much greater things and give really valuable insight instead of just the salami slice of the story and say, "Oh, yeah. I haven't done this because I'm not able to [inaudible 00:33:13] ask somebody else." You know?
Paul Orange (33:13):
Mm-hmm (affirmative).
Elisa Fadda (33:13):
It's just more [inaudible 00:33:18]. A comprehensive science story requires a lot of input from people that come from different directions. That's been one of the highlights or silver linings of this year, I guess.
Paul Orange (33:35):
Yeah. I completely agree, completely agree. I'm just wondering if I could just change direction ever so slightly because you touched on something that I was interested in asking. You mentioned that it takes a lot of computing power to do these simulations. I was going to ask is it just from your [inaudible 00:33:52] laptop? I'm guessing not. Could you give us an idea of what are we talking about in terms of processing power needed to do these kind of simulations?
Elisa Fadda (34:03):
Yeah. To give you an idea, the full spike embedded in the membrane with all this water and [inaudible 00:34:15] environment will count at 1.7 million atoms, so 1.7 million centers that you need to run calculations and the calculation scales with the number of atoms. Okay? The calculations are [inaudible 00:34:27].
Paul Orange (34:28):
So it's a power function, right? It's something to the power of 1.7 million. Or-
Elisa Fadda (34:34):
It's just N would be 1.7 million. [inaudible 00:34:38] with that number. The more atoms you have, the more calculations you have to do. To do those simulations, like the simulations that Rommie Amaro's group ran were run on one of the fastest supercomputers in the world that is in Texas. They were run, if I'm not mistaken and off the top of my head, 280 nodes in parallel. That is an-
Paul Orange (35:14):
Wow.
Elisa Fadda (35:15):
... enormous amount. Those are CPUs. Once we had figured out that you actually don't need the whole spike, you just need the top of the spike in order to properly address the mechanism of opening, we reduced obviously in the number of atoms but not by a super great deal. It just was probably be less than half in some things. What we used is GPUs. GPUs are much faster. We used between two and four GPUs at once in parallel and those are SOAR simulations but in our last experiment there was... I don't remember if it was 12 models or even more than that and multi, multi microsecond of each of them was run for 10 months.
Elisa Fadda (36:15):
Those experiments were 10 months of experiments run on European computational resources from Price. We got 15.8 million core hours, so that's enormous. We used them for 10 months and ran on a computer in Italy that's called MARCONI 100. You need very potent, some very powerful supercomputers to run these things.
Paul Orange (36:44):
Yeah. Maybe for people who aren't familiar with this, something so small is so complex to model that it requires, as you say, some of the most powerful computing in the world to do. It's insane but again, it's a testament to everybody throwing in and putting that effort in and making those resources available.
Elisa Fadda (37:08):
Yeah. Just for example, Price in May of last year... Usually they have recurring calls for these allocations of very large resources. They had a special fast-track of a world of resources only for COVID projects. That was instrumental because usually the review of these things takes months. Six months, seven months. That took one month of review. Everybody was just pulling in and all the reviewers were like, "Speed up. Get that on a fast-track," and so we were able to start in June and we finished in December.
Paul Orange (37:52):
Wow. Okay. There was something I made a note of to ask you which was, working in simulations and a lot of people have been home working, I was just wondering, how much of this were you able to do remotely? I was also going to ask about a little bit of a mix of the physical science and you've already answered, a big fan of cryo-EM, for instance. There's clearly a connection, right? It doesn't all just happen on a screen for you. There's work done with collaborators or with members of your team on physical studies as well. How has that been working through this process as well? Have you still been able to do the physical science part?
Elisa Fadda (38:39):
Yeah. Like me, I don't do any of the physical science part. We have been working remotely quite successfully. I haven't been in the lab since April last year. I haven't seen my students in 3D since April of last year. We would have to connect to those big resources or those computing centers, high-performance computer centers as they're called, remotely anyway. We have a big resources in Ireland as well but the computer is far away from here, we are just always, so connecting from home or connecting from work is identical. For our work, it's been fine.
Elisa Fadda (39:27):
In terms of the experiments that people have done in direct support of our work, there were people that were allowed to be in the lab. For the first part of the work, Jason McLaren's group, Jason's lab is one of the first that resolved the spike proteins coronavirus 2 structure. Nowadays it is world-renowned. It is the reason why we have the vaccine, I'll have to say. His work has been incredible for the advancements that we have at this point in every... To combat the virus full stop. Again, he was in the lab when he did the experiments on our mutants. Our [inaudible 00:40:26] mutants that was last summer.
Elisa Fadda (40:29):
For other, in terms of building these structures or reconstructing the glycan shield, what we need essentially that is vital is glycoanalytics. What we need is to know what glycans goes where. What type of these very complex sugars do we need to rebuild and analyze precisely? This has been again, the tireless work and another one of the, in this case the glyco-hero of this year, that is Max Crispin at the University of Southampton, who has published... He has a preprint first in February of last year, the complete profile of the glycosylation of the spike and then followed up so many other different studies that complimented this. Then finally, this work was published in Science in June.
Elisa Fadda (41:20):
Again, I think another fundamental contribution to this effort, and I hope this... I think it's pretty clear, the role of preprints and preprinting in science advancement. We would be months behind if we were not able to get these preprints, these data on preprint and then be ourselves, the peer reviewer. I remember getting those papers. It was like, "Okay. We get this," and you read the paper and then you assess as much as you can if it's good or not and if you can use it or not and then it goes into peer review.
Elisa Fadda (42:10):
If we had to wait for peer review, for example for Max Crispin's work, we wouldn't have been able to start our simulations until six months after and we would've lost six months. It's fundamental that the science is communicated through this avenue and as by our archive rights there is a note, "This is not to be peer reviewed. Watch out." You know? It's a warning.
Paul Orange (42:39):
Yeah.
Elisa Fadda (42:40):
However, a peer reviewed article is usually reviewed by two people and not necessarily... It's two, three people if you're lucky. It's never guaranteed that these two, three people are actually competent. If you put out a preprint, everybody sees it. To your own demise. That's why I think that overall there are some flukes, obviously. Everywhere there's flukes. But the overall quality of the preprints and how useful, how fundamental some of them have been through this progress is that, I think, one of those things that we should cherish after all this nightmare is over.
Paul Orange (43:24):
Yeah. Yeah, I agree with you. Yes, the debate about the peer review process is one I don't think that's ever going to stop, right?
Elisa Fadda (43:35):
[inaudible 00:43:35] it's not peer reviewed, I'm just going to... Good thing that I'm home. I cannot [inaudible 00:43:41]. But it's just ludicrous. It's a ludicrous statement. There is so much absolute garbage that's peer reviewed that circulates out. I always see the tireless work of Elizabeth Beck, finding out what's figures that have been meddled with or Photoshopped. This is all peer reviewed. It's something that needs to be, and I guess we're in the process of reevaluating slowly but surely. Don't get me wrong. I value peer reviewed. I'm a reviewer. I think it's a great step through progress to have all your work examined and critically examined. However, for different reasons, especially because of academic institutions, how it is and how you're judged, peer review is just so biased. It's not foolproof, let's put it that way and I'll be nice.
Paul Orange (44:46):
Okay. Fair enough. Fair enough. I'm aware of time. I think maybe the last thing I'd like to just ask is then, what's next for you? I think I saw that you've got a funded sabbatical coming up.
Elisa Fadda (45:04):
Yeah. I've got an approved sabbatical, so where I'm going to be for six months, so that's very exciting, at the Institute for Glycomics in Gold Coast in Australia. The Institute for Glycomics is actually a whole institute who are dedicated to glycoscience and understanding glycoscience, understanding carbohydrates in infections, in cancer, in every... Biology in general and it's actually led and founded by Professor Mark von Itzstein who discovered Relenza, which is one of the two now. He's going to kill me. I think one of the only two therapeutics that are active against influenza [inaudible 00:46:04].
Paul Orange (46:06):
Okay. Physically, you'll be going to Australia for that, yeah?
Elisa Fadda (46:14):
Yeah.
Paul Orange (46:14):
Yeah? Okay.
Elisa Fadda (46:14):
If they let me in.
Paul Orange (46:16):
Yep. Okay.
Elisa Fadda (46:17):
I'm hoping to get vaccinated by that time and it will be February 2022.
Paul Orange (46:22):
Oh, I'm sure. I'm sure you'll be vaccinated by then. I'm just [crosstalk 00:46:29]-
Elisa Fadda (46:28):
We are slow but not that slow. After that, I hope to be a free citizen again.
Paul Orange (46:38):
I guess, is there anything else that I should've asked you or anything that you want to make a comment on or talk about?
Elisa Fadda (46:47):
No. Other than thank you so much for having me here and for again, use your time as a glyco-evangelist. Spread my word of the glyco world, the word around. It's just that is so important to consider all sides of biology, not just one. It's a very important protein, [inaudible 00:47:16] proteins but it's only one side of the picture.
Paul Orange (47:20):
No. First of all, it's me who should be thanking you. You are an evangelist. You come across incredibly passionately about what you do. The glyco biome, I guess, is incredibly important in lots of biological processes and as we've learned over the past couple of years disease processes in particular. I encourage anybody who's glyco-curious to get into it. It's probably one of the toughest areas of biochemistry or biology to look into, right? Because as you say, these things are very difficult to examine experimentally, [crosstalk 00:47:55] so we need new tools for this.
Elisa Fadda (47:57):
They are difficult to look at. What a colleague of mine once asked is like, "Why is there not that many people [inaudible 00:48:04] that actually look at glycans from a computational biophysics point of view?" It's because they're extremely difficult to look at. They require a background in order to understand them. In chemistry, you need to understand the chemistry of a carbohydrate in order to actually even look at it. I've been working with carbohydrates for a very long time at this point and I still have trouble distinguishing different monosaccharides if I don't draw them or I don't look at them. I cannot, at a glance say, "Bang. That's this and that." Like linkages. Especially when they're in 3D. When they're in 2D, it's fine.
Elisa Fadda (48:51):
They are difficult to look at. They're extremely flexible, so they're impossible to look at experimentally. They are not template-driven, so you cannot sequence nothing. And also, the biology of carbohydrates or the glycans is enormously complex. You need a biology background or you need the will to go for it like I did. I have a chemistry degree, a theoretical chemistry degree in the end, so to me actually it was a discovery that I had to do by myself and work myself together with supervisors and colleagues and learn. I'm still learning loads. Sometimes I'm like, "Oh, my god. I didn't know that." It happens every day, so it makes it interesting.