https://orcid.org/0000-0001-9397-2914
ABSTRACT
Professor Richard (Rick) Morimoto is the Bill and Gayle Cook Professor of Biology and Director of the Rice Institute for Biomedical Research at Northwestern University. He has made foundational contributions to our understanding of how cells respond to various stresses, and the role played in those responses by chaperones. Working across a variety of experimental models, from C. elegans to human neuronal cells, he has identified a number of important molecular components that sense and respond to stress, and he has dissected how stress alters cellular and organismal physiology. Together with colleagues, Professor Morimoto has coined the term “proteostasis” to signify the homeostatic control of protein expression and function, and in recent years he has been one of the leaders of a consortium trying to understand proteostasis in healthy and disease states. I took the opportunity to talk with Professor Morimoto about proteostasis in general, the aims of the consortium, and how autophagy is playing an important role in their research effort.
Disclaimer
As a service to authors and researchers we are providing this version of an accepted manuscript (AM). Copyediting, typesetting, and review of the resulting proofs will be undertaken on this manuscript before final publication of the Version of Record (VoR). During production and pre-press, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal relate to these versions also.Conversation with Richard I. Morimoto (slightly edited for clarity and brevity)
Thank you very much, Rick, for agreeing to do this interview, and I think that the readers of Autophagy will also appreciate it. As a start, could you say a few words on the concept of proteostasis.
It came about from scientific interactions and extensive discussions between Jeff Kelly (a chemical biologist), Andy Dillin (a geneticist), Bill Balch (a cell biologist) and myself (a molecular biologist). We realized that a complete understanding of the events that lead from protein synthesis, all the way to folding, function, translocation and degradation would be valuable not only for understanding cellular function, cellular health and longevity but it would also provide a foundation for a better understanding of the many hundreds of diseases related to protein conformation. The key was that scientists from very different backgrounds came together to understand this question in its totality. If you look at the 2008 Science paper where we put down for the first time the word and the concept [Citation1] it lays out very clearly the idea that to understand protein homeostasis required a new way of thinking to bring everything together. I think we all appreciate that our respective fields, whether it is autophagy, translation, protein folding, or the regulatory cell stress responses, are themselves quite substantial. What we proposed at that time, and I think that the field has embraced very nicely, is that there is even more accomplished when people come together because what we are studying is a systems network – maybe one of the oldest systems networks.
And did you have an inkling that all of these different pathways would be somehow integrated?
Yes! But let me note that in the very beginning as we had our first meeting in Cold Spring Harbor in 1982 we had everybody there. The point is that, even at the beginning, there was a recognition that there are many events that control the health of proteins starting from their synthesis, and that there were all type of stresses involved. So, the beginning of the field was quite integrative even though we certainly did not understand it very well at that point. Then what happened in the 2000s and the 2010s was that each field had its own group of meetings (entire meetings on autophagy, on the ubiquitin proteasome, on chaperones, on protein transport, translation and cell stress responses). Those meeting are important, but as a complement to them I think that people are starting to appreciate that autophagy, for example, does not “live” alone, translation does not “live” alone, and that there is feedback all the way from autophagy to translation, and of course that is why the stress responses are so integrated.
And this is one reason why this conversation may be interesting for people working on autophagy who haven’t thought about all of this. OK, let me then move forward. You coined the term proteostasis in 2008-2009, but when did you actually realize that in order to tease the system apart you would need to work as a consortium, and I am talking about the proteostasis consortium that you have started together with some other scientists (Editor’s note: this is the link of the consortium: https://www.proteostasisconsortium.com/).
Although it took a little time for the term proteostasis to be adopted by the field once we published the paper, to our surprise and pleasure it then took off very quickly and became an integrative term that has been adopted widely. At the same time, Jeff Kelly, Andy Dillin and I founded a company, Proteostasis Therapeutics, which was the first company focused on protein folding. Jeff Kelly had previously formed a company with Susan Lindquist, FoldRx Pharmaceuticals, which mostly focused on tafamidis and pharmacological chaperones for amyloid polyneuropathy. Their idea was that if you understand folding in detail and can identify a small molecule that can lock in a desired conformation that would prevent misfolding. And although this led to some impressive early successes, Jeff himself recognized that this strategy may be challenging to apply in general and that a complementary approach to understand the cellular protein quality control machinery in detail was necessary. And even as we founded the company, we struggled with the fact that everything is integrated. In those early days people thought for example that Alzheimer disease was simply a disease of amyloid and nobody thought that it is actually the cellular machinery that is changing even though people realized that those diseases were all age-associated. In those early years, the concepts of proteostasis collapse in ageing were a few decades away from being discovered. So, what happened for us after these realizations was the recognition that we needed to create functional teams, outside of the company and purely in the academic world. We then created the proteostasis consortium about a decade ago centered around the effort to get a large NIH grant. These program project grants allow investigators, either within an institution or across the country, to come together around a big concept. We used that as an organizational principle, and, because it is a grant, the number of people involved cannot be large and there has to be a clear focus. So, we proposed that we were going to study Alzheimer and Alzheimer-related diseases through the lens of proteostasis: why, how and when does it fail? To accomplish this required an expert on the proteasome so we invited Dan Finley; we needed people working on autophagy and so we had Steven Finkbeiner and Jeff Kelly; Steven Finkbeiner has also developed some of the best single-cell longitudinal imaging platforms for neurons; we invited Judith Frydman for molecular chaperones, and we invited Evan Powers, a computational biologist who previously developed the E. coli folding network. Once we brought everybody together, we needed to demonstrate that we could function as a team. As you know, scientists are very good at working independently on their own projects but it can be difficult to manage productive teams of researchers in labs across the country. To see how well we can work together, we decided on two team projects that would benefit ourselves and the field. One was the annotation of the human proteostasis network which took us two and a half years and involved 50 people. This was then made publicly available because our mantra has been to share and have everything we do available as open access.
Can you talk a little regarding how you went about the annotation of the proteostasis network?
Our current total is 2989 genes for the human proteostasis network. We broke the network down into anabolic and catabolic control. We first focused on the anabolic aspect: the chaperones, the compartments, the stress signaling processes, everything involved in making a protein, allowing it to be folded and be transported to the right compartment. The second part was on the catabolic aspect: the ubiquitin proteasome and autophagy (Editor’s note: the two papers describing this work are now in bioRxiv; links are available on the consortium website). For each of those aspects, we also engaged experts in the field. For example, we talked to Randy Nixon and Ivan Dikic about autophagy and similarly for other pathways. We wanted to make sure that our lists and annotations were reasonable. We are not trying to do something new but to organize things better. For example, for autophagy or chaperones for that matter, it is hard to find a complete list of all genes regulating all processes.
I guess during this process you kept the original nucleus of the PIs the same but you added collaborators as needed. What about the funding? Did you stay with NIH grants?
We migrated to other sources of funding because our approach has been to work with many different funding agencies. For example, we are in the process of establishing a partnership with ALTOS (the Bay area branch). We were also invited by the Hevolution Foundation, a new funding agency interested to support basic and translational research on healthy ageing. They knew that proteostasis is a very fundamental mechanism for protein health in every cell and tissue. They asked us to look at ageing through the lens of proteostasis which was very logical because it meant that we would focus on cells and tissues that are healthy, for example fibroblasts from healthy people as they age. We also realised that it was important to partner with the Intramural Program at the National Institute on Ageing because they have a project called the “GESTALT study” [Citation2]. In this study, they collect healthy cells and tissues from 900 volunteers that have been screened with a number of physiological and genetic assays to ensure their healthy status. What is important to note here is that, although all of us in our fields study how proteostasis fails in ageing and disease, we don’t really have a good understanding for any mechanism how does it work in normal health: what controls and optimizes a process to make it robust in health. For us this is all very exciting because if we understand the process in healthy cells and tissues, it will help us understand what happens in the hundreds of folding diseases, etc.
This is actually a very nice place for me to segue and ask you about the Proteostasis Index of Healthy Ageing that you are trying to develop. I guess this fits into the next phase of the work of the consortium?
We are very excited about this idea. It is more of a vision rather than something which is obvious. We are taking one of the most complex networks of biology – corresponding in terms of genes to 15% of the human genome. In terms of proteins it is much more and it corresponds to 20-25% of the protein mass of every cell – with neurons and secreting tissues having even more. The idea is that, if we can take all the information including the omics data and functional measurements in basal and in stressful conditions – such as heat shock stress or oxidative stress – but also in response to molecular perturbants such as small molecule modulators, to determine how the proteostasis network responds in cells and tissues of different chronological age. If there is an autophagy response for example, what other changes and over what period of time. Moreover, is the change different for a young healthy 25-year-old fibroblast versus a healthy 89-year-old fibroblast? We are very intrigued by the cross-talk in these pathways especially because very few people pay attention to the crosstalk as opposed to looking at their pathway, be it autophagy or something else. The goal is to collect all of the information and then, working together with colleagues in computation and AI to use integrative reductionism to formulate and test a hypothesis. In other words, the idea is to generate network information and then experimentally test which components are at risk for failure and to identify those components of the proteostasis network that promote robustness. What happens if we remove network components with CRISPR or with siRNA. What we are hoping will come out of this work is an understanding of what changes and when, well before you see that the system fails.
Is this ultimately going to be the protein world’s answer to the methylation clock?
I am not going to say that it is an “answer to” but certainly we are inspired by what Steven Horvath and many others have done to generate a clock, and if somebody asks “is protein quality control linear”, none of us knows. Are there times when there is a dramatic change? And, of course, a difference between the methylation clock and the Proteostasis Index is that ours is all functionally based. We know at the outset the components, and, working with our colleagues, we will have an understanding in mechanistic detail. It is complementary and yet different.
And so, this is now the future plan for the consortium?
Yes, this is one of our goals. Wouldn’t it be wonderful if, instead of having dementia as the evidence of neurological failure, which of course is useful but perhaps a bit too late, we could quantify specific changes to the proteostasis network well in advance that predict a later failure. And the way we will know this is that we will benchmark our data with monogenic versions of these diseases, because obviously there are genetic determinants for Alzheimer or Alzheimer-related diseases, for familial tauopathy and Huntington disease, etc. And although these are ultimately folding diseases, long before there is phenotypic evidence for this one can vary the load of MAPT/Tau or HTT (huntingtin) and ask the question at what point is the system not capable to deal with it? The idea is that maybe things look normal from a distance but if you challenge the cells with a proteotoxic stress you may see a defect much earlier.
It would be very interesting to see where this works takes you. I was listening recently to Randy Schekman describing the effort he is heading tackling Parkinson disease, and although theirs is a single disease focus the idea of doing the basic science work in an organized way and eventually let it inform the clinical aspect is very similar.
Yes, I was also at a meeting with Randy describing the Parkinson consortium and there are common threads. But for us, the main focus is not on a single disease but on the basic underpinnings of the entire proteostasis network, and I would argue that, whether it is a muscle-wasting disease, a metabolic disease or Alzheimer disease, from a proteostasis network perspective, the hope is that we can identify either common or distinct features for all of these diseases at a network level from work in both healthy and disease cells and tissues.
Just as a note, our Institute is also navigating the tricky space between healthy ageing and diseases that are connected to ageing, so I can completely understand the focus of your work. Can we go back and talk about some things that you have already accomplished, such as the annotation of the network. You have a nice figure on the proteostasis website which shows all of the components of the network and how they connect with each other. Can I ask how did you annotate the network?
We decided, because a lot of the analysis on the GO terms is often driven by paper literature but is not done by experts in the field, to follow a different route. We started with key reviews from domain experts and with intrinsic knowledge of the members of the proteostasis consortium. So, for example, Judith Frydman and myself have a pretty strong understanding of the entire world of chaperones and in this way, we had a basis for this part of the network. Some of our information was of course dated without the benefit of recent bioinformatic tools so we had to modify it. As an example, I will describe some work that we did with chaperones. We started out with a list: HSPA/HSP70, HSP90 and so on. And then, because chaperones are highly conserved in terms of sequence and domains, we used this information to go to the literature and pull out everything known about them. Most of it was confirmatory, but it allowed us to be more confident of what was there. And then we made the decision, very much contributed to by Evan Powers at Scripps Research and Suzanne Elsasser, a senior scientist with Dan Finley, to create an organizational structure in terms of class, group, type and subtype annotation to give increasingly specific descriptions of a component’s role in proteostasis [Editor’s Note, link to a preprint of this work is on the Proteostasis consortium website]. This allows people to track the components through the various pathways, etc. I should note that people at the EBI in Hinxton with whom I talked recently have expressed interest to try to integrate this information into their bioinformatics databases. Interestingly, as we did all of this, we realized that there are no examples of how to think about components of a pathway (transcription factors, DNA repair, chaperones, etc.) in such an organizational way, and the way we have done this has become an added bonus of our work to the community.
Once you had this initial phase done, what were the next steps in the annotation work?
Once we validated the list of the components we looked deeper into the families, and I know this because I participated in the work myself. For example, for E3 ligases there are many gene families that have domains and so we used sequence identity, domain structure and identity, function based on biochemical work and genetics, to classify things. Our system was meant to be inclusive – when we were unsure but the sequence or the structural relatedness offered a clue we included this molecule in the list to let other people in the field look at it and discover whether, for example, this particular gene is in fact an E3 ligase.
And I guess we should make the point that all of this is available to everybody.
It is all online, freely accessible. Everything we have done follows the mantra that our efforts are for the community (and this includes academics, biotech and pharma) to benefit from a better resource. Let me take a moment to describe the other project that we have done which is of equal value, which has to do with small molecule proteostasis regulators. If you into the literature and look for molecules that affect folding or autophagy or the proteasome you find many hundreds of molecules. Which ones should you use, and how and why? Because we have Jason Gestwicki as part of the newer group together with Jeff Kelly from the original consortium who are both – together with their teams – fantastic medicinal chemists and chemical biologists, we decided as a contribution to the field to go through the entire literature and look at every molecule implicated in the regulation of proteostasis. We would read the literature, come up with a minimal set of the most reliable molecules, resynthesize them, put them on 96-well plates and make them available to the community. We have already done this and the information is on the website. In order to gain access to the plates, you have to apply by indicating what you are planning to do, what is the assay you will use, and what is the readout that you expect. This information will ensure that the assays are robust. The only agreement the applicants will need to accept is that they will submit their result to an internal intranet so that the next person who wants to use the compounds will benefit from knowing about the data obtained before. Of course, they would have to agree not to publish previous data without the permission of the people who obtained them. By the way, none of these compounds are proprietary drugs, just chemical tools.
We can say that these are compounds that would allow us to manipulate the proteostasis pathway in a very specific way?
Exactly. For example, people who are still using MG132 to inhibit the proteasome need to be aware that this compound affects other pathways in cells. A scientist outside of the field who goes to the literature to find an inhibitor for a pathway may not know which are the best molecules.
Since we are on this topic, what is actually the best compound to inhibit the proteasome?
We would recommend bortezomid.
I was going to ask the next question at the end but I may as well do it now. Obviously, the long-term goal of this work is the generation of some pharmaceutical that would extend lifespan; do you think there is reasonable hope that something like that will happen in the next 10-20 years?
I am very hopeful, but I think that it will come out from a strategy that is different. Our purpose is to understand the basic mechanisms of healthy proteostasis and this is very different from the approaches that everybody else is taking on what fails in a specific disease. It is not a criticism since it makes sense to also have those other approaches. For example, all this effort to enhance autophagy is reasonable given the fact that many diseases tie in to autophagy. However, has anybody asked what are the long-term consequences of enhancing autophagy on the other arms of proteostasis? So, what we are doing is to moderate the discussion in an effective way. What we are saying is let’s understand what is the threshold of autophagy in normal healthy ageing, when does it change, and when it changes does it do so in concert with, for example, a change in translation or a compensatory change in the proteasome. These are the questions that will become important because then when you think about a molecule you have multiple biomarkers that you can score simultaneously. And although with these biomarkers we will not have a mechanistic understanding of the response, we will have a perfect correlation with the response be it in autophagy, the unfolded protein response or the heat shock response.
Going back to autophagy a little, and although this may not be a fair question to ask you since you are not heading the autophagy part of the consortium, but nevertheless, has your work uncovered something interesting or unexpected related to autophagy when you annotated the genes or when doing the other types of work? Have you noticed things that are missing?
For example, with autophagy, one thing we did was to talk to a lot of people in the field; some feedback was provided by Randy Nixon and others as well. Also, because the list has always been open access to the community we are constantly getting feedback from our colleagues. So, what we have now is what I would call a form of scientific crowd sourcing: you don’t have to be a collaborator, you can simply use the resource and then nominate genes or provide additional information for genes already included in the list. I should also mention that our list is a useful resource for people working with omic technologies. Let’s say for example that you are working on small molecules enhancing autophagy. Now you have a list of genes, about 1100 or them, and this allows you to ask which components change in response to your small molecule. Even now, as people publish papers talking about changes in protein homeostasis they are mostly looking at a smaller subset of genes and they use GO term analysis that sometimes connects things in a very indirect way. For example, we recently saw a paper that used GO term analysis showing that chaperones are regulators of the proteasome, which is of course very indirect and likely not a regulatory interaction.
In addition to such more theoretical aspects, have you uncovered something more experimental related to autophagy?
I could mention that work from Jeff Kelly which is in bioRxiv describes MTOR-independent regulators of autophagy.
This is one of the holy grails in the field as you know. In my experience, every time we think we had an MTOR-independent regulator of general autophagy it turned out to be a regulator of a specific selective autophagy response, so it would be very interesting to look at this work.
Yes, this is one of the results of our approach of team collaboration and I will mention a few more details. When Jeff Kelly did a very large molecule screen at Scripps Research, he made the decision, instead of doing all of the biological tests in-house, to contact leading laboratories around the world and offer them the molecules to test. He would give them the test compound and an enantiomer or something equivalent and let them work with it. So, for example, Ivan Dikic’s laboratory at Frankfurt and Steve Haggarty’s laboratory at Harvard has done an exhaustive testing of some of these molecules. This was done with other colleagues, and if you look at the bioRxiv paper it has a long list of collaborators. The bigger point to make is that when we work together in a new way of collaboration we can accomplish much more. Especially in drug discovery where what we do in a single lab is seldom sufficiently robust to take forward.
This is a good place to ask you, what are the mechanisms for people getting in touch with the consortium. You said for example, to use the annotations and if you find something interesting let us know. How do people do that?
We are building into the website a user-friendly portal where you can come in with a set of genes and ask “whether, and how does the proteostasis network interact with this set?” Or one can ask if the gene that they are studying is a new component of the proteostasis network. This is possible because our datasets also contain all of the protein-protein interaction studies, all the Pearson correlations, protein and RNA expression, the yeast two-hybrid results etc. We are about ready to launch this next phase of our portal, which will be even more useful, but for right now we have some information on how people can get in touch by email.
One comment that comes to mind is that for such a huge task that you have given to yourselves you actually have a fairly small core group.
Let me share with you the reason for this. I felt that it was critical to form a highly efficient, well integrated and productive core team, and that this core team is just that: the nucleus of the effort. This core team can then interact with a large number of investigators as the opportunity arises to achieve the bigger goals. In some ways, what we tried to do is very much like forming a biotech company with very well-defined goals.
I remember some time ago when Al Gilman created the Alliance for Cellular Signaling with very high-level goals, but also in some ways with too many people being involved – I guess you have benefitted from knowing about that experience?
Exactly. In my opinion, it is difficult to a large group of people coming together while keeping the effort focused, efficient and productive. What you can do is to have a core that is highly integrated and productive which at the same time benefits openly in partnering with collaborators in every possible way. This is why we decided, even as we started generating data, that we will publish our units of information open access as bioRxiv papers so that everybody can look at what we have as we continue on. That is another approach to making our work interactive and useful to many colleagues and it avoids having to run a large army!
In some way, I think that Al also realized that as well as the Alliance moved on, although he wanted to be very inclusive at the beginning. I can only wish you good luck with your own effort, and I am glad to see that you have the funding to be able to accomplish your goals.
Thank you! We are very excited about our work and I should also mention that there is so much to be done that it is not a question of competition but of collaboration. On our side, we want to stay as a small group with well-defined and big goals, and we will always welcome working cooperatively with others.
Thank you, Rick, for this very useful conversation. From my side, I hope that colleagues in autophagy reading this interview will get to know the proteostasis consortium, how to engage with it and what it is trying to accomplish.
Yes, and as an aside I want to mention that although meetings of individual components of proteostasis are fantastic, there is also the proteostasis meeting at Cold Spring Harbor that brings all of the strands of this network together.