Activity-Based Protein Profiling for Drug Discovery | ABPP Webinar

[00:00:00]hello everyone i'd like to welcome you to today's webinar titled activity-based protein profiling for drug discovery over the next hour our panelists will discuss how proteins and enzymes in particular play a pivotal role in human physiological and pathological processes they'll explore how activity-based protein profiling or abpp methods exploit the power of chemistry to elucidate enzyme activities and identify novel therapeutic targets at this time i'd like to introduce you to our panelists dr matt boisjo is a professor of pathology at stanford medical school his interests are focused on the use of chemistry to study the role of proteases in human disease in particular in tumor genesis and the life cycle of pathogens dr ben cravat is the gilula chair of chemical biology and professor in the department of chemistry at the scripps research institute his research group develops and applies chemical proteomic technologies for protein and drug discovery on a global scale and has particular interest in studying biochemical pathways in the nervous system and cancer ben is also a co-founder of several biotech companies and also today our session will be moderated by dr brad bacchus he is an associate professor at the ucsf school of medicine and began his career in the pharmaceutical industry spending nearly 10 years with novartis and abbott laboratories brad is also the co-founder of client therapeutics and has contributed founding intellectual property to several other biotech companies before i turn this over to brad i do want to remind all of the attendees that you can post questions during the webinar using this q a panel and your zoom toolbar we will reserve an hour at the end to pose questions to the panelists all right brad over to you welcome all and thanks to cdd for hosting this event and thank to thank to uh all of you for joining activity-based protein profiling is a collection of methods and technology it's had a profound impact on both basic research and drug discovery and development whereas genomic strategies and proteomic approaches are often limited in that they can only depict protein abundance in a sample activity-based protein profiling can characterize protein activity that is members of a particular proteome that are actually turned on that are free of endogenous inhibitors the advantage that this affords for understanding proteins and their relevance in some biology or disease state is obvious now at a high level the cartoon you see here depicts a typical approach you have a complex proteome and a specific member of this proteome or a particular class of targets is targeted with the reactive activity site probe uh that allows for visualization like like a dye is attached or enrichment like a biotin is attached now following labeling proteins can be visualized by in-gel fluorescent scanning or after enrichment downstream lcms analysis can be performed to tell you what's in your sample now abpp probes have been developed for numerous different enzyme classes including hydrolases metalloproteases oxo reductases glutathon s transferases and and many more and these technologies have been expanded broadly today we're really lucky to have both matt and ben with us two scientists who pioneered this technology way back in the 90s and who have really led its innovation to its maturation today uh and all the powerful tools that we have at our disposal now we're going to start out by start out by asking about the genesis of these technologies in their labs and how these evolved to uh durable applications that we're familiar with so matt maybe i can hand it to you tell us about the genesis brad thanks a lot um i really appreciate the opportunity to speak today and cdd for putting together this webinar i think it's a fun topic for for me in particular um and i i really want to say i'm particularly excited to be here with ben cravat who i really think of as a close friend and colleague who i've known now over 20 years i think our relationship and interactions have been a great example of how you can work in a field and sort of synergize with one another rather than sort of competing in a specific field i mean i often would get asked about who are my competitors when i was first starting out in this field and i i would kind of scratch my head and think well the mo the closest person has really been and he's he's a great friend and collaborator so i don't think of it as competition at all um so that kind of brings us to the beginning of this and actually i met first my friend in the late 90s when he was basically a young assistant professor at scripps and i was just getting my lab started at ucsf as a faculty fellow there and he came through um and he had just published his paper on um his fp based probes that you see here on the right and the slide and the reason why he had synthesized these probes was to study a single enzyme called pha fatty acid amide hydrolyze and he gave this amazing talk um at ucsf and this is probably one of the first times i had you know i was in the seminar circuit where the speakers would come and talk and so he came to my office and luckily was on my schedule and we had this amazing conversation um and and realized that we had very similar interests on the left you see is what i was working on actually as a graduate student even before i started at ucsf um was was making these kinds of covalent labels of the proteasome when i was a graduate student and hit a plus lab we were very interested in sort of understanding how that multi-component protease complex worked and there were just a number of things you couldn't do with classical substrates which is what people use to study proteases because there are all these different active sites if you throw a substrate at it you didn't know who was cleaving it which subunit um and so i first synthesized this compound that you see here which was a vinyl cell phone and back in the day we used radioactive labels when it was uh still in vogue to use iodine 125 slop that stuff around the lab and i remember distinctly i tell the story is like one of the aha moments for me as a graduate student was the gel image you see there in the middle labeled autoradiogram i had added i had added the probe on the left you can see the silver stain to the total protein or to the total cellular extract and then when i was pulling this film out i thought the experiment didn't work it was blank blank blank blank and then all of a sudden boom you saw those bands that you see there and to me it was amazing that i could label uh a single subunit in that case of a multi-protease complex in the context of the entire proteome and it looked just as if it was a purified protein which is what you see in those two lanes and then you see a 2d gel image of that so it sort of opened up the door for me this idea that like we really need to use these kinds of covalent labels as a way to track enzymes sort of in real time um and that that really sort of was the direction i took um and i know ben followed along that track as well um if you advance to the next slide maybe charlie um i think i have another really kind of old slide in here yeah this is great this one paper from back in 2002 when we started playing around further and going after other targets so moving from the proteasome we started working on cathepsins as targets um and those are a relatively small family but highly related protease proteases and and we had started with a slightly different type of electrophile here you can see the epoxide based electrophile which is based on a natural product called e64 and this slide is just the purpose of showing you this was that we soon realized that there are all these different modalities for detecting these probes once they're bound and obviously we had done work with i-125 but that was sub-optimal um it gave a very sensitive signal but there was not much you could do with that biotin was another one obviously we could use that to pull out targets but the real step forward um and i think ben and i both started kind of doing this around the same time was putting a fluorophore onto the molecule and then we realized um that if we had a flatbed scanner um or some way of scanning the gel for fluorescence we could see the bands and the funny part was back then there weren't really good commercial flatbed scanners for floor force and so the top image a was actually taken using a dna sequencing setup where i basically labeled individual purified cathepsins with different colored probes and then uh ran those protein samples through the very thin dna type gels on on the machine and the machine was looking for fluorescence and that's how it read out um the dna fragments and so we could actually get a protein gel like image which you see there so that for us was like wow this is really the direction we need to go and then and then eventually commercially you could get flatbed scanners which were a lot easier you could just take your gel and not even remove it from the glass you could literally put your gel down onto the scanner scan and get the kind of image you see you see at the bottom and it was nice because it showed you could get very specific labeling here these are these are labeling of four cathepsin targets in the context of a whole cell and you can see labeled with all the different strategies so that um that was kind of the the early days of activity-based profiling and i think you know where ben and i diverged a bit was that i always had this interest in hydrolytic enzymes so proteases hydrolases esterases um you know enzymes that add water across a bond and ben i think you know made the the really sort of brilliant leap towards just broadening the reactivity and broadening the classes of enzymes that you could see with this and as you know he's now moved even into further into sort of reactivity profiling so um yeah that's kind of my my two cents about the history of the field and i'll let um you know ben add some stuff to here thanks matt it's uh it's also great to be with you here today it's been really fun to reflect on the last you know 20 plus years or more time working together and in this emerging field and um i think as you as you alluded to and pointed out it is important to note i think to the audience that these sorts of technologies you know really emerged from specific problems that both you and i were trying to address in our in our um biological systems of interest whether it be understanding proteasome subunit function or cathepsin activity or in our case the specific functionalities of hydrolases then vault and endocannabinoid metabolism and um you know i've always adhered to the belief that that if you can solve a specific problem in your lab with a new technology or approach oftentimes it will generalize right and and but the converse can be challenging right if you start with the divine concept of trying to create a general technology you don't have a specific problem to work on i think it can be very difficult to know um where to apply that technology and so i i think in large part to the extent that activity profiling had this much broader impact across the field i think it's because it originated with very specific problems that we were trying to solve and those problems end up being problems that many of the labs are trying to solve on their specific proteins of interest as well and um as we'll talk about a little bit later you know the generalized tools that emerged proved versatile not only for our specific proteins or enzymes of interest but for in some cases entire several hundred member protein families so um so yeah so that's all i'll add right now but maybe i'll let you kind of explain how you've taken this into really remarkable areas of imaging and then we can return to the broader you know chemical proteomic approaches that are after that so yeah sure yeah i mean i think for for us the genesis was that you know once we started to realize these things could be labeled with fluorescent tags we wanted to ask the question could you go into more complex systems obviously not just cellular systems but whole organisms and that's sort of the direction that my lab has taken certainly since i started at stanford back in 2003 i think on the next slide we have some images yeah so this just shows some of the exact same probes that we were working back in the early 2000s on cathepsins turned out to be very good for labeling tumors it turns out cathepsins are very highly expressed in the tumor microenvironment and this just shows where we've injected a mouse with one of our activatable fluorescent probes so i would say that that was another big step forward for us was realizing that if we could make them fluorescent we could also make them activatable fluorescent um so we came up with this concept of quenched activity-based probes where um now as the probe covalently labels the target it actually releases a floor quencher for the fluorophore and so you get a signal and this allows you to now do very rapid imaging because you don't have to wait for the for the unbound probe to circulate out of the system and we've been over the last now uh almost 20 years that's coming up on 20 years we've been playing around with this and in vivo systems i'm happy to report that actually one of our probes is now entering or entered into clinical trials in australia and patients with lung cancer and is starting a clinical study in the united states sometime later this year so it's really exciting to actually see this kind of data that you're seeing here in a mouse but actually in a human patient and i think it's an exciting time for contrast agents like this for imaging and i think they're going to have a very big impact i'm guessing in the next three to five years you're going to be seeing most types of tumor surgeries involving a contrast agent where it allows the surgeon to be able to see where the cancer is you can see here on the left yeah go ahead that's really exciting matt that it's going into humans do they do they do this as an ex vivo application or actually uh uh use this uh directly on the patient it's actually uh systemically administered yeah so that's been the hurdle for these kinds of agents getting to the clinic is that um it has to go through kind of the same kind of regulatory steps that a drug would um and that took some effort but now there's a number of companies that have contrast agents that have been through phase two and now some one just went through phase three and has been approved um so that's exciting uh and that that's this is an area that we've been really excited about um pursuing and you can see here sort of on the left like this is the problem during cancer surgery you can see the tumor fluorescent it's very large but the on the right in the yellow region that's the actual tumor bed if you just look at that in the white light you think oh we got all the tumor but when you see the fluorescence and do the pathology you realize we left quite a bit of cancer cell behind and so this is what we're we're gunning for um and and we've been spending a lot of time trying to make our probes more selective basically which is kind of interesting because it's kind of the opposite of ben making things more broad so you can profile more on a proteomic sense we've been we really want selectivity because we don't want to hit you know targets that aren't present outside the tumor um and so that's that's again where i think we've diverged and in a very good way it's allowed us to sort of broaden the field for sure i think there's one more slide that just shows um if you just go to the next one yeah it just shows again like these kinds of probes how selective they can be in vivo i mean this this is akin to the moment of seeing that gel coming out of the processor where there's a single band there here it was like here's a whole animal this is one of the mice injected with our cathepsin activatable probe and you can see these are the mammary fat pad tumors the two of them and you see some other signal elsewhere including lymph nodes but it's really really a bright signal specifically in the tumor so yeah it's been it's been a fun ride along that path with probes and you know another area we've really been interested in is using them as a way to identify new protease and particular new protease targets if we have a system where we know a protease is involved we don't know what it is if we have a compound that's a covalent inhibitor of that activity we can then use that as a probe form to pull out the target and identify it so we've done a lot of that kind of functional biological studies uh where we have a specific phenotype that's really interesting i think that between the two of you you know between targeting specific enzyme and broad classes uh it's very complementary ben do you want to talk more about uh broad yeah coverage of the proteome classes sure maybe go to the next slide i think um as a template yeah so um you know when matt and i were beginning these efforts it happened to coincide with when genome sequences began to appear in their complete form and so that got protein scientists quite interested in initially just kind of cataloging to the extent that we could at that time you know memberships in different large protein families say in the human genome or your genome of of choice um and then that that's where it became clear that at least for some of the activity probes that were already being developed such as those targeting the hydrolases shown here um there was an incredible membership of that type of protein family in humans two to three hundred enzymes of which many of them remain well most of them remain certainly without selective inhibitors and even a substantial fraction remained unannotated as relates to their endogenous functions you know substrates and products and so i do think that one of the more durable applications of activity profiling especially with a probe like this fluorophosphate which can react with hundreds of members of a given enzyme family it allows one to have a universal assay so to speak to look at changes in the activity state of those proteins and then pretty much any biological system of interest and then as we'll discuss in a moment also allow you to use it as a tool to discover inhibitors for for those proteins and so um you know in that regard you know i don't think we necessarily saw foresaw that that broad application until the the scope of the of the proteome was appreciated um but i think that that once it was recognized that there was such a large fraction of unannotated proteins um and uh and these probes such as broad reactivity with with memberships of large families that then it became pretty clear how one could apply them and then i think you know what we also learned during that time was or gained benefit from i should say is advances in mass spectrometry based technologies right so if you go to the next slide you know we begin to ask some simple questions like you know why do we need to restrict ourselves to pockets of enzyme active sites for for chemical reactivity analysis if the mass spec can read out tens of thousands of peptides maybe we could push this technology to look at the reactivity and functionality of say just amino acids in the proteome that are intrinsically nucleophilic like cysteines right and that's led to this hybrid activity slash reactivity based pro profiling concept um this is some recent work published in the last couple years where what's neat is when you start applying that type of approach to look at the ligand ability if you will not to mention the reactivity and functionality of cysteines across the broader proteome you can begin to get the evidence of how many sites of small molecule engagement there exist in the proteome that are far outside what you might call classically druggable space i.e outside the active sites of enzymes right so this is sort of an example you take a compound library and add it to a biological system right and if one of those compounds happens to interact with the site that uh that an activity probe reacts with they compete with one another and you can read that out in a variety of formats you know if you're looking at obviously 10 000 or 20 000 sites in parallel it's going to be a mass spectrometry based readout if you have a more selective probe as matt's described you could look at it by imaging you can look at it by gel based analysis and in very convenient formats and i think this has ended up being a pretty valuable and versatile application of activity profiling because it puts the sort of uh chemistry discovery part very far in front of the biological investigations you know we actually have examples in our lab where we've discovered you know useful tool compounds for proteins and enzymes where the initially the only known biochem biological activity is the probe reacts if the protein reacts to activity probe right and then you can use that that selective tool compound that's been discovered in this way to annotate the functions of the proteins in in rather complex biological systems and so so i think you know when the general learnings from from this type of strategy is you know not only can we expand like i would say like the tool set of of selected chemical probes within druggable space we actually i think realize at least with covalent chemistry that one can radically expand the ligandable or druggable space outside of conventional druggables active sites so um yeah this is just to to point out that like matt mentioned with his imaging tools there are now multiple drug candidates in clinical development but i think from the very moment they were discovered as hits used activity profiling as a guiding tool to optimize their specificity these are actually both covalent inhibitors that are remarkably selective they're for their respective um hydrolytic enzymes found maglipase respectively um and i think that these covalent temperatures you know end up being potentially highly useful drugs in part because they do have such specificity for their targets and i'm not sure that type of information could have been acquired for an enzyme family that's got two or three hundred members without this type of chemical proteomic strategy underpinning the evaluation of of compounds that that were discovered um you know maybe by more conventional screening methods and then optimized using activity profiling as a guiding guiding tool so you know when i began uh you know covalent inhibitors were taboo uh maybe like 20 years ago and what impact do you think abtp has had on changing people's perceptions of alcohol i'll give it energy give a couple thoughts on that and certainly would be interested in maths perspective as well but i think you are correct brad that probably both matt and i kind of got laughed at maybe 15 years ago when we would propose purposefully or intentionally developing covalent chemical probes if not drugs for proteins because it was just thought that they would be you know um indiscriminate labelers of proteins throughout the proteome and you couldn't harness and control their their reactivity um but i think activity profiling as well as as i think some um you know ambitious efforts in oncology initially around covalent kinase inhibitors and the willingness to bring those into the clinic um has i think uh flipped that narrative or that's completely and i think now you'll see many companies i think that would would prefer a covalent inhibitor or complaint ligand if they could develop one assuming that it can achieve the level of potency and selectivity of a drug candidate because you can start separating out pharmacodynamics from pharmacokinetic effects especially for proteins that have reasonably long half-lives and that can allow you to in some ways get away with with compounds that are less perfect on the pharmacokinetic side in exchange for for benefiting from you know living off the half-life of the protein for your pharmacological activity um and of course it also allows you to find pockets and proteins that are hit that are pretty challenging to drug in a reversible way i think kayvon just wrote a really nice review about the perspective on this comparing g12c k ras inhibitors versus the emerging efforts to go after other g-12 variants of mutant variants of k-ras that now use reversible ligands and the challenges are very different right in terms of what you can get away with with a more shallow or difficult drug pocket if covalency can be used as a as a as a way to to harness that that that that um interaction so you know i i think yeah your question i guess getting back to it is is would that maybe phrase another way you know would the enthusiasm for covalent drug development be where it is today without activity profiling and chemical proteomics as an underpinning technology and i actually think it probably wouldn't be because i think there would have been drugs that would have been taken forward that would have failed in the clinic for gross toxicity and you know in fact there's one fawn inhibitor that fell in that category right that wasn't really vetted by chemical proteomics in an aggressive and an intense way and that compound you know ended up having very severe side effects that have nothing to do with fall right many fall inhibitors have gone in and if anything you know font hammers are too safe in some ways from a clinical development perspective in their historical activity so i i do think that that maybe if if chemical proteomics wasn't around to support the optimization of those types of compounds there would be more examples of that type of failure of covalent chemistry from a safety perspective and and that in turn may have may have um sort of reinforced the traditional biases against this type of modality yeah i totally agree with that i mean for me it was a a very real experience because when i was at ucsf as a fellow after i left there i moved to slayer genomics for a few years and did actual some i was involved in several other drug discovery programs including some of their catheps inhibitors that were partnered with merck um but i remember distinctly at the time like if we had any kind of meetings my group was really responsible for trying to develop these kind of covalent active side probes as a way to sort of look at target engagement of their lead molecules but if i ever talked about any of our molecules they would always have to be called tool compounds or or probes but never like a drug lead and ironically at the time stilero was moving into the kinase space and they wanted probes specifically to look at kinase targets and this was back in early 2000 2001 i think um so my my group at the time made um you know took took some of the took a known kinase inhibitor and added electrophilic uh warheads on to it and in fact one of those compounds which was a tool compound as it was called its lara ended up getting sold off and became imbrutinib it was purchased by pharmacyclix um and eventually was the molecule that now is i think one of the the key molecules that pushed the the sort of momentum towards covalent drugs um and and you know we saw that when we were at solari when you had a compound that was a covalent compound you know it made your life way easier you didn't care so much about the pharmacokinetic profiles you cared about how much what was the c-max how much drug got in and then how much target got engaged so it became a pharmacodynamic analysis and that was also much easier to do to be able to figure out how much target did you hit where did you hit it um and so it became clear that this was a you know a good strategy it just i think it took a while getting over the hump that this this idea that if you have a covalent modifier you're going to induce an immune response and and certainly certainly there are um examples of that um so anyway i think it's been fun to watch and uh i mean i i think it's it's been a nice transition towards covalent um type drugs and i think we're going to see a continued expansion of that into the pharmaceutical space yeah i actually think matt following up on that that the you know the g12c ligands are a particular case of interest in my mind because they're to my knowledge arguably the first example of of a covalent drug that was not repurposed from a reversible ligand right that that was designed with you know was sort of a chemist like a covalent chemistry first approach from the beginning right purposeful right like if the covalent county's numbers largely were retroactively modified reversible inhibitors um so you so one could if one want to be a cynic or a skeptic one could say oh yeah well you know that's a a low bar right to be able to take a tight binding reversal inhibitor and then retrofit it with a covalent electrophile to to to achieve your outcomes but but i think the gtlc ligands from amgen and emirati and wellspring and others are a different example right and if that can be generalized i think that that really bears um some thought about how how broadly that could be applicable across the podium for breaking down traditional boundaries of drug ability yeah definitely good conversation on that i think uh maybe maybe we could switch gears just a little bit and ask each of you you know is there a particular direction that this technology is heading or needs to head or a particular direction that you're excited about at this time um sure i mean i can start quickly i i mean i think one of the things that i'm excited to see is that because it's become more um sort of in vogue to work with covalent modifiers you see the expansion in the space in the chemistry space of the electrophilic sort of warheads and traps that are being used now which is just a nice paper that came out sort of profiling the reactivity of all these different electrophiles and then um sort of doing the deep proteomics to ask what gets modified how selective are they what because that's a big question you know how much how much reactivity do you want in a warhead and i've always talked about this you know for me the vinyl cell phone was my first sort of electrophile that i really got excited about and that molecule you know when you talk about regular vinyl cell phones with like an alkyl vinyl sulfone that'll react with the free thiol no problem right it's a microacceptor but it turns out if you put it in the context of these peptide scaffolds so you have an alpha nitrogen that becomes very deactivated and what i was really thought was cool was that here is a molecule that um really just does not react with free thighs and lets you heat it at very high concentrations but then you put it into the cell and boom it forms this covalent bond and it was even more interesting to me because it hit the proteasome i remember you know the first molecule that i started with was a peptide aldehyde called mg132 it was a leucine leucine leucine aldehyde which was one of the only published or known inhibitors of the proteasome at the time other than lactosystem which was a natural product um and i saw that and said well that's one step um the horner wadsworth emmett's reaction away from a final cell phone so let's make the minus the phone version and see if it works and all my chemist friends are like oh that's stupid you know that's that's a michael acceptor michael acceptors react with dials not with hydroxyls and sure enough i made it and tested it and it reacted so here you have an example of a michael acceptor type reaction with a hydroxyl nucleophile so it tells you like the reactivity it really depends a lot on how well you can get things to bind right and and what where you get close to that nucleophile so you know i think it's going to take some calibration to understand you know what is it we're looking for and and i think it also depends on what scaffold you're going to put that electrophile on if you're scaffold something very small like a peptide or even smaller like a drug like you know less than 500 molecular weight molecule you're going to need a lot more reactivity out of your electrophile which is going to potentially give you cross-reactivity with other targets as you get to bigger molecules you probably can get away with much much less reactive electrophiles the problem there is then you get things that don't look as drug-like and maybe don't get into cells and so i think this is the exciting part of the field is we're starting to see expansion of the type of electrophiles people are reporting on more testing of those and i think in the future we'll have a better sort of um i would guess menu of potential things to to build into probes definitely i think you know the ability to target different residues as you said to move beyond sistines uh definitely opens a lot of doors how about you ben yeah um those are great comments by matt and i i certainly support those areas as well but i i might highlight a slightly complimentary one which is you know as we've begun to get outside the active sites of enzymes and you're finding high quality druggable pockets you know these are fundamentally binding first assays right and so we are then sort of uh encumbered with the opportunity or challenge depending on your perspective of relating those binding interactions to function and um we've begun to uncover and publish work on you know some really remarkable allosteric sites that are being discovered uh through this type of binding first approach um even as well as sites that are at the interface of rna dna binding with proteins and such and so it it really then starts to raise some provocative questions about um you know how what what the fraction of these druggable pockets are that are agonistic antagonistic silent contextually silent contextually active against the their proteins of interest and how do those events then relate to say the genetic loss of function of that protein right and and like to me you know to the and i have this you know maybe ridiculous but interesting vision of a provocative vision of ligand discovery and maybe 10 years out from now where you know these sorts of um allosteric and non-active site-based ligand events will provide much more contextualized ways to regulate protein function than just the sledgehammer or complete deletion of a protein genetically maybe even in the context of you know proteoform specific ligand ability right where you can start not only drugging you know a kinase but drugging a kinase in its phosphorylated state or its protein protein interaction state again to get more refined you know could there be a day when chemistry is more refined than genetics in terms of its ability to manipulate and modif and control protein function in cells i think that would be you know quite exciting and then finally i'll mention that you know i think as our lab's certainly probably one of the most liberal in terms of of using the terminology activity profiling across a whole variety of chemical proteomic approaches um and so one could argue you know if you're finding allosteric sites in the podium of drug ability that have reactive cysteines in them are you just finding sort of neo-functional outcomes that are pharmacological or are you sort of rediscovering sites in the proteome that the cell is naturally using it was an own endogenous small molecule electrophilic or otherwise repertoire to regulate protein function in other words all these allosteric stereotypes also themselves activity reflective and i think in many cases they probably are it's just a challenge for us to deduce you know what the endogenous regulators are right i i look back at something like and you know brad you'll be familiar with this right the um the amazing work out of gnf many many years ago to discover the mercilation pocket as being a druggable site in bcr able right or able specifically um that's a good example right of a druggable pocket that that actually is in a site of endogenous regulation outside the active site of the kinase itself and i suspect there's a lot of that going on and maybe chemical proteomics can help find these pockets first with synthetic ligands right then allowing the biological community to go back and figure out what the endogenous ligand might be and then you know ultimately maybe this concept of activity versus reactivity profiling just kind of goes away right and any side of drug ability in the proteome may end up being uh potentially a way of that the cell regulates function of proteins as well because the uh targeting of allosteric sites is certainly an amazing vista to uh to address uh charlie did you want to head to uh your uh your uh uh quiz for the the group yeah hi thanks so uh we're going to launch a poll here um and we would like to give the attendees a chance um to vote here on what you think is most relevant for abp i'm not going to read the choices here hopefully you can read these and make your choices appropriately and they're off to the races everyone i see the the scores coming in people are still voting so we'll give them a few more seconds to to wrap all of this up all right everyone i'm going to end the poll in three two one now thank you very much so here are the results if you guys can all see this so the 38 of the attendees feel that drug discovery leads is the most relevant but follow very closely by target validation so maybe not a clear winner if you guys have any comments on those i think this is what i would hope to see right it can do awesome right excellent all right guys i'm gonna stop sharing this and uh ben back to you or brad yeah how about we uh you know we've had a really uh terrific conversation here uh we have a good amount of time for questions and i noticed there are a number of questions uh can we move to uh questions charlie yeah absolutely so we will uh turn that part over to abe to see which questions we had coming in from the attendees abe all yours thank you sure i'll read a few questions from our audience um the first one is can you use covalent reversible warheads to make an abpt or do you need an irreversible one yeah i think that it i i think you can certainly profile covalent reversible or even reversible ligands for that matter for their interactions with proteins by competitive pro activity profiling but i don't think that those reversible uh electrophiles will be stable to many of the not all but but many of the applications of activity profiling which you know would involve uh steps to denature proteins and they might fall off if there's ways to maybe making comment on that from an imaging perspective maybe there's ways to use them in a manner that doesn't require denaturation and then you're probably okay right but i think a step that they will fall off if once proteins are are denatured that's part of the protocols yeah yeah i think that that's what ben says is absolutely correct but um in the case of like for proteases there are there are definitely warheads that have slow enough off rates that allow you to see them i mean i think that it's challenging to get them to stay on under denaturing sds page conditions but um we we think they're they're useful from the perspective of imaging as long as the molecules stay stay bound long enough for their contrast to be generated um and and there are definitely ones like that boronia is a great example right it reacts with hydroxyls but some of these bordeates have off rates of like 24 hours or longer which is essentially like a covalent molecule so yeah i mean i think you can do it you just have to figure out what your applications are and then understand the on off rates your molecules the question too i always wondered about on off rates for instance for uh say lead discovery you know are there situations where you're missing things because the you know the lingan is moving in and out and the uh covalent uh warhead has an opportunity to sneak in there it looks like dan dan erlinson who we we know putting a question into the panel into the chat which he wasn't supposed to but i'll ask i'll read his question he asks like there are all these warheads being explored um he's asking whether based on our experience there are ones that you would definitely avoid um that's an interesting question i think it depends on the applications i mean someone to keep an eye on is uh near london who's at the weizmann who's been doing a lot of fragment based stuff and it's been fun chatting with him about this because he's looking at covalent fragments um and you know when you ask him about it there are certain warheads that just depending on how big your fragments are will work or won't work i think that there are warheads that are going to be just too reactive um but that could be used in the context of fragment screening for example and then could be detuned a bit um and so i think that there aren't any warheads that i would say absolutely avoid it can't be used at any cost but because he even uses chloroacetomides and other things that are quite reactive um but he kind of needs that to be able to get the initial binding of a very small fragment uh i think what you said it depends on the experiment of interest right if it's looking for a ligandable pocket if it's doing a cell-based experiment or if you want to make a drug right and there those are different requirements for the quality of the electrophile and its stability and such but you know i think it's also important to point out that even with the same electrophile you're beginning to see some guidelines emerging not from i think academic labs as much but from the drug discovery industry about how to attune the stability stability of those electrophiles right so if you look at the g12 ck raster inhibitors or some of the covalent kinetic inhibitors many of those are just you know wrote acrylamides but you start looking at how they're modified as they become optimized drugs and you'll start seeing that if it's a you know a papyridine or the pyrazine acrylamide you'll start seeing these little magic methyl groups decorated around the papers you know the pyridine ring and that's almost certainly to mitigate gst immediate metabolism of those of those drugs so as we get a better sar of the main routes of metabolism that can be used uh as a way to optimize compounds from tools to drugs that improve their their half-lives right so because ultimately if you have a if you have you're not covalent drugs aren't going to have an advantage for short half-life proteins unless you can keep the drugs around right so if you have a target that's one of the perplexing aspects of of the breadth of the approach right if you want to extend this to targets that can't be drug without covalent chemistry right because they don't have a whatever a pocket that's amenable to reversible chemistry as easily but they also have a short half-life that's going to require your covalent ligand to have a pk properties that resemble a reversible drug and i think the industry is starting to do a really good job of figuring out how to make that happen and make that work so um so i wouldn't even say that all electrophiles are even the same electrolysis created equally in the different scaffolds that may be found in right in terms of of its uh durability well here's another question that popped up here can post translational modif uh modifications affect reactivity yeah we just yeah we published a paper on this uh initial forays in this area right looking at uh how phosphorylation events across the proteome can impact proximal cysteine reactivity events and it's pretty interesting you can see phosphorylation events that will stimulate or increase i should say the reactivity of cysteines on a protein and those that will decrease so this gets back to the comment i made earlier about whether one might eventually be able to create proteoform restricted covalent ligands that take advantage of um for instance a phosphorylation or a ptm's effects on um on the reactivity of a protein right and i think it's very early days right that paper was was the first foray into trying to understand this but i think the data are encouraging and certainly warrant further investigation um into considering how the different ways that proteins are modified in cells can impact their um their reactivity with small molecules here's another question uh what's your take on targeting protein protein interaction sites certainly always a very difficult uh challenge yeah i think that's a good one i mean my lab's starting to work a little on that um as we've developed some of these kind of phage display technologies to kind of get at more complex ligands that are used like in this case cyclic peptides or bicyclic peptides that contain the reactive electrophile piece we're starting to look at those kind of interactions because i think they do require these greater protein-like interfaces to be disrupted i think the idea of using covalent and something that has the ability to bind almost like a small protein has real potential for disrupting those things it's been traditionally very challenging but adding the the covalent piece to it i think could really help a lot i don't know then what do you think yeah i i totally agree i we have had some pretty good luck actually with protein protein modulation again with as matt pointed out the the caveat or the not the caveat but the context that the ligand recognition event is probably going to be dictated by not a fragment in that case but more of a elaborated small molecule that has sp3 kind of character to it to allow it to interact with protein protein interfaces that may be a bit flatter than a and and less shaped as a pocket than say the an hp binding pocket of a kinase but um no i'm a big believer that the appropriately designed covalent libraries can can certainly have a big impact on on modulating protein protein interactions here's another question from uh pat fitzgerald who says can you comment on the state of abpp in the context of antibacterial drug discovery notes that both of you have published in the area recently yeah go ahead matt go ahead matt yeah i know that's definitely a big area for us we're really interested in infectious disease um not only for the drug development side of it but also for the imaging side we think this is a big hole right now for you talk to infectious disease doctors and the problem is they can tell you you have an infection through a blood culture but they don't know where it is nor do they know how it's responding to treatment so we're we've been developing and using profiling with actually a lot of benz probes shifting over towards the fluorophosphonates because of their broader reactivity than a lot of our probes and we've identified all sorts of interesting both syrian proteases and syrian hydrolases for bacterial targets we're looking in the gut commensals now and i've identified really interesting targets there as well so i think um yeah there's real potential here i believe that you know abpp could sort of help to better characterize a lot of these poorly characterized bacterial genomes in particular um and then once those are characterized there's the potential to throw whatever molecules are out there at it and i think you know these numbers of like for example even just syrian reactive molecules that ben's word on and abide therapeutics which is now lungbeck has um i think those all could be basically used against these targets so i think for antibiotics it would be quite interesting yeah i mean one of the areas map that i think is under explored with activity profiling but but is just beginning to emerge as a thematic area of interest um again always drawing to you know think about where can chemical biology solve problems that genetics has had trouble with some of these bacteria like tuberculosis they have they have like you know many replicate members of a given sub clamp right and so as a genetic analysis single gene knockouts don't reveal any evidence of functional functionality for those paralogous parallelogress enzymes because they can all compensate for one another but if you can create a tailored polypharmacology chemical probe through activity profiling those might become the most relevant drugs right when all sudden four members of a related clan are being drugged collectively in a way that you know the genetic experiments would have told you each individual enzyme is irrelevant but all four inhibited together become lethal to the to the growth of the bacteria and certainly in enzymes like tb that's the case right where they just have this massive expansion of like hydrolytic enzymes and i suspect that several of them are performing overlapping and complementary functions so that a single genetic knockout is not sufficient to re to record their relevance to the organism's you know growth and viability yeah that that brings up that's a great point then because i mean we've we've been working with some of these kind of um well one was the natural product in in plasmodium falciparum which is the causative agent of malaria and there we uncovered a whole host of syrian hydrolases that we think are lipid processing enzymes and you have to hit multiple of them to really kill the parasite but what's more interesting about that is it becomes very hard for the parasite to become resistant when you hit multiple targets so we're finding very low resistance in the vagina which is really cool the other thing that we're finding is if you take a drug like for the proteasome and you have a reversible binder against the malaria proteasome and an irreversible binder the the reversible binder tends to lead to induction of a resistance phenomena in the in the parasite whereas the irreversible one doesn't and we think it's because what happens is you can make mutations around the binding side of the molecule so you're so you don't bind quite as well and that has a big impact for a reversible binding yeah yeah you can't you take the nucleophile or the enzyme's dead right even if it's slower over time you're still accumulating dead enzyme so uh we think it's really cool from a perspective like that for infectious disease participant asks how about an application of these technologies the gpcr is certainly a a large drug class yeah it's interesting question i mean i do think that there are you know there are some gpcr drugs in the market that actually adventitiously work through disulfide bonding the the rare cysteines that exist on the outside of cells in a free form but one has to be sober about that right and recognize that you know gpcr sequences tend to be have a lot of sequence outside the cell or within the membrane and this may be where one needs to consider alternative amino acids beyond cysteine right because most of the cystine is not all of them but most the cysteines outside the cell are disulfide bonded pretty stably right which would make them inert to electrophilic interactions with small molecules so you know i don't think that there's been a lot of progress in this area yet but over time one can imagine as the um electrophilic chemistries for additional nucleophilic amino acids begin to mature um there might be opportunities in for covalent chemistry in the gpcr space more purposefully than those that have been discovered so far yeah i honestly believe if you get if you get the right electrophiles and you have enough diversity of fairly complex ligands attached there's really no limit to what you could go after you know there are limits to where those molecules can get to if they're very big and they require so much binding energy to deliver the electrophile but i think there's there is definitely the potential to go after non-catalytic and not even very nucleophilic residues if you can get your molecule spine in the right spot another question may be targeted towards matt a participant asked how challenging is it to find and apply appropriate warheads with phage display maybe a little bit more detail on the phage display methods yeah i mean that's what we're kind of playing around with now i mean we've tested some of our known warheads that we've worked that we know work with like proteolytic targets that was our first paper and now we're looking at uh electrophiles that we um can attach and we're going for the much more deactivated electrophiles because they're we think that the the phage approach allows you to get to these sort of billions of complex sequences so now you can sort of make up for the low reactivity of the electrophile with higher binding affinity um so it's just something i think you have to you have to test um and we're in the sort of early days of that great maybe uh down to our last questions uh another participant asked for more detail on the prospects or just the the future of non-cysteine nucleophile targeting uh maybe more on the chemistry there yeah i mean i think that that um there's prospects there but one again has to be a little bit sober about recognizing cysteine's intrinsic nucleophilicity at neutral ph is is dialed for covalent reactivity and the reason that there's so many covalent drug candidates and drugs for syrian hydrolases is that's also what happens with the catalytic nucleophile the serine right it gets it gets driven from a pka of 14 or 15 down to seven or eight and it's also hypernucleophilic in in the context of neutral ph i think when you think about other amino acids like lysine or tyrosine i think you have two choices right uh one which is less attractive to me is to just play in the sandbox of the very small subset of those residues that also happen to have for whatever reason um perturb pkas to fall within the range of hypernucleophilicity at neutral ph that's going to be a small sandbox in my view based on the data we've already generated in our lab and others aren't there aren't that many lysines and tyrosines that are pka perturbed to the proteome the more attractive approach and i think matt was getting to this earlier would be to to strike a slightly better balance of recognition and reactivity right that if one can improve on the reversible binding elements of a small molecule um such that this molecule is is got a high effective molarity in proximity to an otherwise very attenuated nucleophilic residue like a pka unperturbed lysine i think you have a chance to get covalent chemistry towards your advantage in that context um because you'll be able to have the effect of molarity drive the reaction even on a residue that is less overall nucleophilic um and do so with a less reactive electrophile but um those these are early days still and i think that's something that we're going to have to um work through in a case-by-case basis right because that that puts a lot of demands on the reversible binding affinity of the initial ligands that are used as templates for covalent chemistry yeah thank you uh we're coming up against the hour i i'll take this uh time to thank ben and matt for the stimulating uh discussion and thanks to all the participants for all the questions uh i'll hand it back to charlie for uh final comments yeah thank you brad and matt and ben for this discussion today uh as a reminder i would like to let you know that collaborative drug discoveries cdd vault platform offers a comprehensive solution for the storage management and analysis of chemical and biological data if you're interested in more information you can always navigate to our website collaborativedrug.com or if you're interested in finding us in the real world we're back out there participating in these upcoming conferences and finally we'd like to announce that the next webinar that cdd will be hosting is the covid moonshot collaboration webinar on june 9th where we'll have scientists from the covet moonshot project discuss their ongoing collaborations to address the global pandemic so thank you very much again to our panelists and moderator and to all of you attendees for joining us today we hope you have a wonderful rest of your week