From Plant Proteins and Metabolites to Protein Networks and Metabolic Pathways
Proteomics and metabolomics are two of the “omics” technologies that are still underrepresented in plant biology despite their well-recognized value to crop science. With the help of examples, this talk will show how using these approaches contribute to advancing our understanding of plant coping strategies and defense mechanisms when they are under stress.
icon search Searchable Transcript
Toggle between list and paragraph view.
[00:00:00.800]The following presentation
[00:00:02.260]is part of the agronomy and horticulture seminar series
[00:00:05.840]at the University of Nebraska-Lincoln.
[00:00:09.040]Good afternoon, everyone and welcome
[00:00:10.900]to the UNL Department of Agronomy and Horticulture
[00:00:15.040]I am MarK E Bull.
[00:00:16.100]I am an associate professor in the department.
[00:00:20.180]So I'm quite happy to introduce Sophie today.
[00:00:23.110]I know her for quite a long time now
[00:00:24.810]for over 15 years almost now.
[00:00:29.010]And Sophie joined UNL in 2015,
[00:00:31.810]as a director of Proteomics and Metabolomics Facility,
[00:00:35.310]at the Nebraska Center for Biotechnology.
[00:00:38.080]And she's also a research associate professor
[00:00:40.120]in the Department of Agronomy and Horticulture.
[00:00:43.600]So Sophie earned her PhD in 2004
[00:00:46.560]from the University of Lille 1 in France in plant biology.
[00:00:50.660]After she get her PhD she moved to the US
[00:00:53.980]where she completed two different projectile fellowships.
[00:00:56.600]One with Daniel Trackman, at the Danforth Center
[00:01:00.010]and one with Qi Hu Chen at the University of Florida.
[00:01:03.620]In both position Sophie gained experience in proteomics
[00:01:07.117]and metabolomics approaches and she conducts
[00:01:09.820]those experiments on different plant species
[00:01:12.400]like maize, peas, most of Arabidopsis
[00:01:15.560]and in response to values biotic and abiotic stresses.
[00:01:19.940]So, Sophie can gained really outstanding expertise
[00:01:22.440]in metabolomics and proteomics.
[00:01:25.610]And as a reflection of that in 2007,
[00:01:29.530]she becomes a manager of the proteomics
[00:01:31.507]and mass spectrometry facility at the Danforth center.
[00:01:34.600]And there she was running this facility for eight years
[00:01:38.000]applying mass spectrometry based omics tools
[00:01:41.020]to various different types of samples, not just plants,
[00:01:43.915]but also RJ, animals, human samples.
[00:01:48.000]As a consequence, she has a pretty outstanding
[00:01:53.490]number of paper publications 57,
[00:01:56.360]including searching as a first auto.
[00:01:58.700]In 2010, she received the outstanding
[00:02:00.770]Scientist Technologist Award
[00:02:02.410]from the Association of Biomolecular Resource Facilities.
[00:02:06.380]So Sophie is really an expert
[00:02:07.600]in mass spectrometry technology.
[00:02:09.700]She's operating here the core facilities in the data center,
[00:02:13.200]and she provides services such protein,
[00:02:15.180]small molecules tools on to our community.
[00:02:17.930]And today she's going to talk to us
[00:02:19.630]about the study of proteins and metabolites,
[00:02:22.200]and how it's critical to understand biology
[00:02:24.460]at the system level.
[00:02:27.150]Well, thanks, Mark for the introduction.
[00:02:29.980]And thanks for everyone who's attending to this seminar.
[00:02:34.080]My name is Sophie Alvarez, and I'm the Director
[00:02:37.280]of the Proteomics and Metabolomics Facility.
[00:02:40.940]Today, my talk is gonna be describing the methodologies
[00:02:46.120]and platform that we've established in the lab
[00:02:49.140]over the last five years.
[00:02:51.317]And those platforms are used to study proteins
[00:02:57.360]And I'm gonna go over a few examples,
[00:03:01.040]where we successfully apply these methodologies
[00:03:05.880]to address plant related biological questions
[00:03:11.090]Anyway, so my first slide is to introduce our team
[00:03:15.100]and the core facility.
[00:03:17.550]Currently, we are five members including myself,
[00:03:21.580](mumbles) here, the assistant director
[00:03:26.640]is mostly responsible for the proteomics platform,
[00:03:31.660]and is our lab manager.
[00:03:33.630]And we have two research technicians, Felicia and Laurie,
[00:03:39.170]mostly working on the metabolomics platforms
[00:03:42.710]and Felicia also involved in the proteomics platform.
[00:03:46.500]So the proteomics and metabolomics facility
[00:03:50.900]is part of the Nebraska Center for Biotechnology.
[00:03:54.060]We are located at the Beadle Center.
[00:03:57.117]So we serve the needs for biological
[00:04:00.590]mass spec-related research across all Nebraska campuses,
[00:04:05.960]but also do external work from institution
[00:04:10.770]and private companies.
[00:04:12.890]So the way the core facility can operate
[00:04:17.280]is based on a fee-for-service business model.
[00:04:22.060]So what does that mean, is for all the samples
[00:04:24.750]that we run through the different platforms,
[00:04:28.210]we charge the user for a cost.
[00:04:32.560]And for internal users, that cost is actually subsidized.
[00:04:37.500]It's not the true cost of the actual experiment.
[00:04:40.460]And this subsidy comes from funding
[00:04:46.860]from the Nebraska Research Initiative.
[00:04:49.180]So the NRI has been really helpful to us
[00:04:54.680]with the establishment of the core facility
[00:04:56.760]and over the years.
[00:04:59.390]The other source of funding that operates the facility
[00:05:05.970]is from grant, grant funding through collaborations.
[00:05:13.120]So, I was just gonna give a few numbers
[00:05:17.590]with regards to the facility because maybe some of you know
[00:05:23.080]but maybe some of you don't actually know
[00:05:25.100]that this has been only our fifth year
[00:05:28.540]running this core facility.
[00:05:30.400]I consider myself that we are still
[00:05:32.210]quite a new core facility after five years.
[00:05:36.100]We've started this lab from scratch by installing equipment
[00:05:42.610]that was funded by NRI and by establishing
[00:05:46.000]many different types of approaches and methodologies
[00:05:52.810]to run our services, our current services.
[00:05:55.440]So we have a GC-MS instrument, a standalone HPLC
[00:06:01.810]and now we have total of three mass spec platforms,
[00:06:10.150]The third one was just recently added earlier this year.
[00:06:14.950]So just to give you an idea on how many samples
[00:06:19.040]we are able to process every year.
[00:06:22.130]So obviously, that number went up over the years
[00:06:27.010]as we were establishing ourselves at UNL,
[00:06:31.780]establishing our methodologies.
[00:06:33.960]So we last year ran over 3500 samples,
[00:06:39.870]those samples came from about 100 different users.
[00:06:44.590]And that user base has also obviously been increasing
[00:06:47.740]over the years.
[00:06:48.940]What we noticed here, with this plateau and this decrease
[00:06:53.100]is not really a concern since the number of samples
[00:06:55.810]has been still going up.
[00:06:57.820]And this is explained by the fact that 42%
[00:07:01.030]of our user last year, were actually repeat users.
[00:07:04.910]So what that means is that they've been submitting sample
[00:07:07.990]more than once.
[00:07:10.130]So you can also see in this graph that we got
[00:07:14.860]internal users, but also external users from academic
[00:07:20.850]And amongst our users, UNL users, last year 82%
[00:07:26.460]were from INR and College of Arts and Science.
[00:07:29.581]So we serve quite a large community,
[00:07:32.990]internally and externally.
[00:07:38.520]So this slide might be a little bit too busy,
[00:07:42.505]I don't need to go through it,
[00:07:45.650]really the take home message here is that,
[00:07:48.720]with all the knowledge that we have now, we know that
[00:07:54.570]the phenotype is not directly linked to the genotype.
[00:08:00.630]So the central dogma of molecular life,
[00:08:06.120]which starts with the transcription of the DNA into RNA,
[00:08:11.490]and then translated into proteins,
[00:08:15.820]proteins that have a catalytic activity
[00:08:18.480]which also convert metabolites,
[00:08:21.350]well this dogma is actually highly regulated.
[00:08:26.360]It's highly regulated and there are lots of
[00:08:30.722]feedback mechanisms that are operating
[00:08:33.710]at each omics layer.
[00:08:36.240]So looking at the genomic expression
[00:08:40.830]and looking at levels of transcript or proteins
[00:08:45.420]or metabolites, if you were looking at all of them,
[00:08:48.810]don't expect them to be correlated or associated
[00:08:53.010]because of all that fine regulation going on.
[00:08:56.300]So one would really think that to be able
[00:09:00.400]to understand fully a phenotype to actually study
[00:09:06.973]what's happening are those are multiple layers,
[00:09:10.290]so multiple layers of that dogma.
[00:09:17.195]And this is where we can help.
[00:09:21.070]So if anyone needs to study proteomics or metabolomics,
[00:09:29.310]we are here to run the project.
[00:09:34.180]So, this slide is just sort of describing the main workflows
[00:09:41.150]that we have for our proteomics platform.
[00:09:45.320]So mainly what we do are listed here.
[00:09:50.090]So the quantitative proteomics,
[00:09:52.830]which is really looking at protein abundance in samples
[00:09:58.110]and looking at the changes between different samples,
[00:10:03.190]either comparing a wild type to a mutant,
[00:10:07.670]or comparing a control condition to a treated sample.
[00:10:15.140]Those studies, we can actually do it two different ways.
[00:10:20.510]They are called label-free approach or labeling approaches.
[00:10:25.360]And this is depicted here by this workflow
[00:10:28.710]where we would actually get from the user the actual sample.
[00:10:33.840]So we do a lot of the sample prep.
[00:10:36.560]So we're not just mass spectrometers,
[00:10:39.750]we actually do a lot of sample preparation
[00:10:42.850]and data analysis.
[00:10:45.480]So what we would do depending on what the sample is,
[00:10:50.840]we would extract the protein.
[00:10:52.670]So we have different types of extraction
[00:10:54.380]depending on the sample.
[00:10:55.950]So we would do an extraction and then digest the protein
[00:11:03.100]because what we do in the lab
[00:11:04.690]is bottom up approach proteomics and I'm gonna talk
[00:11:07.250]a little bit more about that in my next slide.
[00:11:10.420]And then we run the proteins or the peptides here
[00:11:14.100]into the mass spec.
[00:11:16.040]And once it's run on to the mass spec, the data is analyzed
[00:11:19.630]and typically for proteomics,
[00:11:22.360]the proteins are first identified and then we quantify them.
[00:11:26.080]The quantification that we get here
[00:11:28.260]for quantitative proteomics project
[00:11:30.330]is relative quantification.
[00:11:32.410]So again, what that means is we not getting concentration,
[00:11:38.000]can give you constant actual concentration of proteins
[00:11:40.820]in your sample, but we can tell you
[00:11:42.600]if by comparing different sample, if this protein abundance
[00:11:46.410]is going up or going down.
[00:11:49.810]So, if you go through this workflow,
[00:11:53.330]this is the workflow for level three.
[00:11:56.260]But what we can do as well is take a labeling approach
[00:12:01.000]where after the proteolysis, we use TMT labeling.
[00:12:05.390]So it's chemistry that we use,
[00:12:07.400]specific chemistry that we use to label the sample
[00:12:10.730]that will be typically followed by fractionation
[00:12:13.220]before it goes on to the mass spec platform.
[00:12:16.380]And I'm gonna hopefully be able to, if I have time,
[00:12:19.784]explain a little bit more about that.
[00:12:22.838]The other main type of proteomics that we do
[00:12:30.640]is interactome studies.
[00:12:33.520]So those interactome studies, what that means
[00:12:35.930]is a lot of the samples that we actually receive
[00:12:39.520]at the facility are not the original sample
[00:12:43.970]but processed sample.
[00:12:45.810]Processed sample through a pull down or Co-IP experiments.
[00:12:50.860]So please note that those pull downs and Co-IP experiments
[00:12:54.560]are not done in our lab, but are actually optimized
[00:12:58.330]by the user.
[00:13:00.010]But where we intervene, where we start
[00:13:03.100]is after the pull down is done, we get the bins,
[00:13:09.280]or we take the early weight and then we would
[00:13:12.780]clean the sample typically before digest
[00:13:16.600]and then run on the mass spec.
[00:13:21.130]I'm sure, hopefully, everybody knows
[00:13:24.140]what I mean by interactome study
[00:13:26.150]is trying to identify protein interrupters
[00:13:29.840]of a protein, specific protein or trying to identify
[00:13:34.630]the components of a protein complex.
[00:13:38.610]This third type of proteomics that we do
[00:13:43.270]So now, what we do is not just look at protein abundance
[00:13:46.330]in the sample, but we're looking at the abundance
[00:13:48.940]of phospho-proteins, phosphorylated proteins.
[00:13:53.610]And probably know that phosphorylation
[00:13:56.750]is an important modification that can regulate
[00:14:03.070]So, phosphoproteomics project, what we typically recommend
[00:14:08.940]is to do a phospho-enrichment before we actually run
[00:14:12.800]the samples onto the mass spectrometers.
[00:14:15.140]And hopefully, I will be able to give you
[00:14:17.780]a little bit more details later on.
[00:14:19.990]But because of the type of equipment that we have,
[00:14:25.310]it doesn't stop there,
[00:14:26.820]we can help with different type of project
[00:14:30.000]that can be done using a mass spectrometer,
[00:14:32.690]and I'm not mentioning this too much,
[00:14:35.760]but we obviously is not just about phosphorylation as a PTM
[00:14:40.180]but we can do other type of PTM as well.
[00:14:45.200]So, just to note that those type of proteomics experiment
[00:14:50.780]are for discovery.
[00:14:52.900]So they're usually used for discovery type studies.
[00:14:58.820]Okay, so this is really to explain you the fundamental
[00:15:04.525]or the principle of the proteomics platform that we use.
[00:15:10.290]Because by understanding this,
[00:15:13.770]you'll be able to better understand the data you get
[00:15:16.900]if you're gonna be using us in the future
[00:15:19.860]and also understand some of the challenges that we face
[00:15:24.400]when we do proteomics, because proteomics
[00:15:26.940]is not always perfect.
[00:15:29.060]So, this is just at the top showing the different steps
[00:15:37.354]of the actual experiment that we would go through.
[00:15:41.040]So, like I said, the type of mass spectrometry
[00:15:45.250]based proteomics approach that we use
[00:15:47.980]is a bottom-up approach,
[00:15:49.700]as opposed to what is also called top down.
[00:15:52.890]And I'm not gonna go through top down but a bottom-up
[00:15:57.790]means that we are taking proteins,
[00:16:00.490]we are digesting them into peptides
[00:16:03.140]and then we identify the peptides and from that
[00:16:06.470]we reassembled into proteins,
[00:16:08.850]which is why it's called bottom up.
[00:16:11.330]So this is what's showing in this slide.
[00:16:14.510]So we have our protein.
[00:16:16.440]Here is a very simple example where there is only one,
[00:16:19.700]but often our samples are not just a single protein
[00:16:22.850]but multiple complex protein sample.
[00:16:26.150]So we take the proteins, we are gonna digest it,
[00:16:29.730]like I just previously said,
[00:16:32.930]the proteins that we use for digestion is trypsin,
[00:16:37.350]most of the time or by default typically
[00:16:39.870]unless it's not appropriate.
[00:16:41.710]But there are some cases it might not be appropriate.
[00:16:46.733]Because a trypsin digest on the C-terminal of arginine
[00:16:54.060]So the proteins have to have arginine and lysine
[00:16:56.560]so that we can digest them into peptides.
[00:16:59.410]So you can see here already one challenge is that
[00:17:02.980]if the proteins don't have arginine and lysines,
[00:17:08.460]that the trypsin won't leave them and we won't be able
[00:17:10.640]to identify the protein.
[00:17:13.040]So once they are digested, we are gonna run them
[00:17:17.730]on the mass spectrometer, the LC mass spec platform.
[00:17:21.490]And this is very simplified but basically,
[00:17:24.860]we go through two rounds of acquisition of information.
[00:17:28.740]So every round here is acquiring
[00:17:32.700]different types of information that's gonna be used
[00:17:34.940]for the identification.
[00:17:36.680]So in the first round, all the peptides
[00:17:39.930]go through the different components of the mass spec
[00:17:43.690]to the detector and the detector will be able to give us
[00:17:49.221]first information, which is the intact mass of the peptide.
[00:17:54.620]So we'll get that first information and then what we'll do
[00:17:57.990]is then send each single peptide through the collision cell
[00:18:02.730]for fragmentation, for further fragmentation.
[00:18:04.950]So what we're trying to do is now fragment
[00:18:08.150]so that we've got information at the amino acid level.
[00:18:12.400]So we're gonna fragment do that for all the peptides
[00:18:15.020]and get a second type of mass spectrum, which is here,
[00:18:21.900]So those two type of information are gonna be used
[00:18:26.180]for the identification.
[00:18:27.960]And this is how it works,
[00:18:29.240]because, now we collect so much information
[00:18:33.800]in one single run that we don't do
[00:18:37.150]a manual de novo sequencing anymore
[00:18:39.880]trying to identify the amino acid sequence of the peptides.
[00:18:44.580]So it's all done through softwares, certain giant databases.
[00:18:51.060]And this is how we actually sequence a protein,
[00:18:54.307]we're not really sequencing a protein
[00:18:56.960]or sequencing a peptide per se, we're just inferring
[00:19:00.930]the sequence by matching our data to a database.
[00:19:06.540]So and this is how it works.
[00:19:07.830]Once we have our experimental data,
[00:19:11.170]this will be submitted to the software
[00:19:13.080]which is gonna convert all those fragments,
[00:19:16.040]all those ions, parent ions and ion fragments into picklist.
[00:19:20.650]And then what we're gonna do is give some information
[00:19:23.430]to the software so that it will be able to match the best
[00:19:29.900]to the proteins.
[00:19:31.730]That information is very important
[00:19:33.880]because if you don't put enough information,
[00:19:36.780]you're most likely not gonna see your proteins.
[00:19:40.080]So the first one is the database, sequence database.
[00:19:43.450]So the way it works is you need a protein database
[00:19:47.810]to be able to identify your protein.
[00:19:49.920]So again, if the database we use doesn't contain
[00:19:54.270]the proteins that you are working with or interested
[00:19:58.610]we will not identify the protein.
[00:20:01.890]So the sequence database we select is very important.
[00:20:04.730]So that's why you have to give us as much information
[00:20:08.691]We will, depending on the species of the sample
[00:20:13.590]that you work with, we'll select the best database for that,
[00:20:17.920]then we'll tell the software what protease we use.
[00:20:21.340]So here we use trypsin, so that the software can take
[00:20:25.410]all the protein sequences in that database,
[00:20:27.970]and do a in silico digestion to produce this list
[00:20:31.910]of protein peptide.
[00:20:35.540]And then what we're gonna tell the software as well
[00:20:38.090]is give some information on the instrumentation
[00:20:40.200]that was used and the way the MS2 was produced.
[00:20:45.200]Because there is a little bit of difference
[00:20:47.830]depending on the instrument.
[00:20:49.860]And from each peptide, it will then generate,
[00:20:53.470]do a in silico fragmentation to generate MS/MS scans,
[00:21:00.130]So you're gonna have theoretical MS scan
[00:21:02.320]and theoretical MS/MS scan, and the software is gonna try
[00:21:06.020]to match your observed data with whatever is present
[00:21:10.440]in the database.
[00:21:11.600]So the other thing that's important to indicate
[00:21:14.560]during that search at the beginning of the search
[00:21:16.620]is if you have post translational modification
[00:21:19.610]of some of the amino acid.
[00:21:22.110]Obviously, if you do not ...
[00:21:23.960]So here, we are measuring mass with a mass spectrometer.
[00:21:27.250]So if you have a modification on the amino acid,
[00:21:31.510]that modification means that it's gonna shift the total mass
[00:21:36.670]of that peptide.
[00:21:38.650]So if we don't tell the software that there is possibility
[00:21:42.230]that this peptide is modified, that this amino acid
[00:21:46.170]is modified that it won't produce the correct size peptide,
[00:21:50.800]and therefore won't be able to match it
[00:21:52.860]to our observed data.
[00:21:53.940]So that will stay unidentified unfortunately.
[00:21:57.500]So as much information that can be given to us
[00:22:01.700]the best quality data you will get out of it.
[00:22:05.380]So once you've got peptide spectrum matches,
[00:22:08.540]then the software will assemble that into proteins.
[00:22:14.480]So now I'm gonna just quickly describe
[00:22:19.200]the two main quantitative proteomic approaches
[00:22:22.940]that we run in the lab.
[00:22:24.050]So those two methodologies we routinely run that in the lab.
[00:22:28.990]So one is the label-free approach and the other one
[00:22:32.150]is the labeled approach using TMT.
[00:22:34.660]So, just for you to understand what's the difference
[00:22:41.120]between the two, and why use one over the other.
[00:22:45.160]So, as you can see here, typically,
[00:22:48.890]you would have your sample, we would extract the protein,
[00:22:51.900]maybe do some further purification,
[00:22:54.900]once you have the proteins we will digest them into peptide,
[00:22:57.960]like I said, most likely using trypsin.
[00:23:01.310]And then after that, if we take a label-free approach,
[00:23:05.357]then each individual sample will then be run separately
[00:23:11.120]into the mass spec platform.
[00:23:14.140]And then we will compare the data
[00:23:16.348]during the data analysis process.
[00:23:19.790]If we take a labeling approach, we will use a TMT chemistry
[00:23:27.300]to label each sample with a different tag,
[00:23:30.970]and then we'll combine those samples into one
[00:23:35.070]and run the sample only once on the mass spectrometer.
[00:23:38.980]So, one of the reason TMT labeling his advantages
[00:23:48.350]is because it will reduce experimental variation
[00:23:53.100]only some of it because obviously,
[00:23:55.640]you still have some experimental variation introduced
[00:23:58.500]before the labeling, but it will remove anything after that.
[00:24:02.990]All the variation would be removed.
[00:24:06.060]The capability of this TMT right now,
[00:24:13.450]this TMT chemistry is that you can actually
[00:24:16.330]combine up to 16 different samples into one run right now.
[00:24:22.950]Obviously, one of the main drawback is that those reagents
[00:24:30.250]obviously costly and they gonna add to the budget.
[00:24:35.560]The other drawback as well is that,
[00:24:40.070]it may or may not be a huge problem,
[00:24:42.700]but we will need a certain amount of proteins
[00:24:46.290]to be able to do a TMT labeling.
[00:24:48.800]So, typically 25 to 50 micrograms are required.
[00:24:53.590]When we do a label-free, it's better if we get
[00:24:59.640]more than five microgram per sample,
[00:25:01.810]but we have done experiment where we had less than five,
[00:25:06.679]less than five microgram.
[00:25:09.910]So yeah, we can do using less than five
[00:25:13.410]but if you have more is always better,
[00:25:17.130]but ultimately what is actually loaded on the instrument
[00:25:20.890]is only one microgram of the sample,
[00:25:23.230]one microgram of protein.
[00:25:26.410]So the difference between the two is how the quantification
[00:25:32.593]Here the quantification is done at the MS1 scan,
[00:25:35.470]so had I shown you when that Ms one scan is acquired
[00:25:42.370]is when the intact peptide level for the labeled approach,
[00:25:49.670]the quant is done at MS2.
[00:25:52.510]So, after fragmentation, and here I did MS3
[00:25:56.580]because with the new instrument that I was talking about
[00:26:00.640]that we just got installed earlier this year,
[00:26:03.310]we are gonna be able to actually do
[00:26:09.330]this type of labeling approach using MS3
[00:26:12.915]for the quantification.
[00:26:16.270]But really, if you are gonna be doing
[00:26:19.740]a labeling TMT experiment, the main advantage of doing it
[00:26:27.620]is that you're gonna be able to mine more deeply
[00:26:32.320]into the protein.
[00:26:34.390]So this is the protein coverage that we get
[00:26:39.120]from a label-free.
[00:26:41.124]So between 2000 3000, obviously it depends on the sample.
[00:26:45.560]But with a TMT labeling, we're gonna be able
[00:26:47.890]to increase that coverage by double.
[00:26:52.920]Right now, some examples, we can go up to 1000 a go
[00:26:57.830]when we get the new instrument running and optimized
[00:27:01.930]to be able to identify 10,000 or more proteins.
[00:27:07.990]And this is only feasible because of a fractionation step
[00:27:13.090]that is included right after the labeling,
[00:27:17.210]where instead of now running only a two-hour run,
[00:27:21.810]we're actually gonna run the sample for 24 hours.
[00:27:24.630]So that's why we get this increase in protein coverage.
[00:27:31.630]Okay, so now I'm just gonna briefly show you some work
[00:27:37.580]that we've done with collaborators
[00:27:40.270]and show you which one of those methodology
[00:27:44.960]we've been using.
[00:27:47.130]So, we did some work with Meter Nusinow,
[00:27:51.439]who is a Pi at the Danforth Center in St. Louis.
[00:27:54.380]And back a few years ago,
[00:27:57.150]it was interesting in the Evening Complex in Arabidopsis.
[00:28:02.880]The Evening Complex, regulates the circadian clock
[00:28:14.352]So the way it's interacting with other pathway,
[00:28:19.520]some information was known, but not necessarily
[00:28:22.810]confirmed information and certainly
[00:28:25.030]not confirmed interactors.
[00:28:28.521]So, this Evening Complex is actually a multiprotein complex
[00:28:32.930]made of ELF3, ELF4, and LUX.
[00:28:38.463]So the way they designed their experiment
[00:28:41.440]was quite sophisticated.
[00:28:44.240]They used affinity purification,
[00:28:47.680]this is an interactome study.
[00:28:50.274]So, they used affinity purification to try to identify
[00:28:53.720]their proteins, the proteins interactor.
[00:28:56.810]So they had the protein bait that was tagged with FLAG
[00:29:07.040]and His, so it was doubly tagged,
[00:29:09.730]because they decided to do a tandem affinity purification.
[00:29:13.730]So it's not just one single purification but a double,
[00:29:17.820]and really that helps a lot with interactomes studies
[00:29:21.460]to remove some of the nonspecific binding that happens a lot
[00:29:26.000]with this type of experiment.
[00:29:27.780]So they had the (mumbles) and then what they also had
[00:29:34.820]included in their study design is in addition
[00:29:38.160]to the wild tag, they had mutants.
[00:29:42.340]A mutant for ELF3 and a mutant for the phytochrome B.
[00:29:47.400]And by comparing the results from the the mass spec data
[00:29:54.240]that we got from the different samples,
[00:29:57.000]they came up with new, I mean, already known interactors,
[00:30:02.930]but also new interactors of this Evening Complex.
[00:30:07.890]So obviously, it was not just about affinity purification,
[00:30:15.000]the power of this experiment was also from the genetic tools
[00:30:19.157]that they were using as well.
[00:30:20.550]And then what they did once they got this data
[00:30:23.230]is that they did additional further experiment
[00:30:25.820]of two-yeast hybrid to confirm
[00:30:29.130]some of the direct interactions
[00:30:33.233]from protein interactors that were identified.
[00:30:36.410]So they published this interactome of the Evening Complex,
[00:30:41.640]and they basically showed that it was connected
[00:30:45.750]to the light sensing and photomorphogenesis pathway
[00:30:50.590]And what was really also interesting about this study
[00:30:53.230]and they were not expecting that
[00:30:54.580]is to find those new interactors.
[00:30:57.330]Obviously, not direct interactors,
[00:30:59.440]but secondary interactors.
[00:31:05.020]Those MLK Mut9-Like kinases, that they sure serve as well
[00:31:13.070]as key regulators in linking the environmental inputs
[00:31:17.170]from light and temperature to developmental output.
[00:31:20.730]So, being the flowering and growth that are controlled
[00:31:23.710]by the security of circadian clock.
[00:31:26.250]So they were able to correlate this to the EC,
[00:31:31.080]the Evening Complex.
[00:31:33.550]And they also added some more work
[00:31:37.309]that was just recently published actually,
[00:31:41.190]trying to characterize mutants to be able to understand
[00:31:46.390]what the role of MLK is, by actually doing
[00:31:49.370]a forceful proteomics and proteomic study.
[00:31:52.120]So we did not actually, that was not done with us
[00:31:55.420]but that really, was quite a good example
[00:31:59.530]of what can be done using mass spectrometry.
[00:32:04.870]Another example of work that we've done here
[00:32:09.320]with David Holding in agronomy.
[00:32:13.600]So he was working with this mutant of maize RDM4,
[00:32:19.510]and he was trying to just characterize the mutants
[00:32:22.557]and he had quite a lot of information already
[00:32:26.460]with regards to the phenotype
[00:32:28.460]and did further experimentation on the storage proteins
[00:32:33.830]to understand what was happening in the kernel.
[00:32:38.455]So he's been able to separate the storage protein
[00:32:42.620]from the non storage protein and showed that
[00:32:45.530]in the mutant RDM4, there was a decrease of synthesis
[00:32:52.700]And that there was probably some changes in the non-zein,
[00:32:59.720]non-zein proteins from the kernel.
[00:33:02.640]They did some complimentary experiments
[00:33:06.870]using amino acids analysis that was done
[00:33:10.150]by another collaborator to show that the lysine content
[00:33:15.860]was also increased in the mutant.
[00:33:18.340]And those three observable really showed that,
[00:33:26.330]they suspected there was a protein rebalancing
[00:33:29.030]in the RDM4 mutants.
[00:33:32.302]So, they had all of this information already,
[00:33:35.690]all of this data, but they wanted to ...
[00:33:39.720]They asked us basically to do so a proteomic study
[00:33:43.380]in addition so that they can confirm what they were seeing
[00:33:47.610]were also the non-zein proteins,
[00:33:51.360]and if there was really a change in the protein content.
[00:33:56.980]So for this experiment, we used TMT labeling approach.
[00:34:02.390]So here we use the TMT 10-plex kit or chemical reagents
[00:34:10.290]because the experimental design, the study design,
[00:34:13.500]only included for biology pro-replicates of the control
[00:34:16.640]and the mutant, we used only eight labels out of the 10.
[00:34:22.716]So I will try to go quickly through the next slide.
[00:34:28.740]Basically, I wanted to sort of try to explain
[00:34:32.890]what is TMT labeling and how it works,
[00:34:35.240]because I'm sure a lot of people have heard of it
[00:34:37.990]but don't really understand it,
[00:34:39.780]and it's actually quite easy to understand.
[00:34:42.750]So I'm gonna try and go through this slide.
[00:34:45.410]Basically, you have one chemical reagent
[00:34:49.150]that has this structure,
[00:34:51.510]which includes a group that's gonna react to the peptide,
[00:34:55.565]what is called balance group and the actual reporter ions.
[00:35:00.016]This is the reporter ion.
[00:35:01.279]So this level is gonna react to the n terminal
[00:35:06.110]of the peptides.
[00:35:07.820]And once we run on to the mass spec, after MS/MS,
[00:35:13.340]we're gonna have a cleavage here that's gonna
[00:35:17.400]give the individual report ion from each sample.
[00:35:22.660]And we'll be able to measure the abundance from each sample.
[00:35:28.200]So the strategy and the way it works is that
[00:35:32.080]you have this exact reagent in 10 different version,
[00:35:36.660]but all of them have the same size,
[00:35:38.540]they're all the same mass.
[00:35:40.570]What changes is that some of the carbon and nitrogen
[00:35:45.740]are labeled with stable isotope, which means that,
[00:35:50.024]and they all have the same number of stable isotope,
[00:35:54.420]all of them.
[00:35:55.660]So that means that the overall mass is not gonna change
[00:35:58.270]because they have all the same number,
[00:36:00.980]but the position of them is different.
[00:36:03.970]So just here for information going from a C 12 to a C 13,
[00:36:09.810]it's gonna add this much to the size of the compound,
[00:36:17.180]when you go from N 14 to N 15, this is the difference.
[00:36:21.640]And they are gonna play with the fact that
[00:36:24.070]it's actually not giving the same difference
[00:36:26.860]to give this TMT 10-plex because the first TMT
[00:36:32.140]commercially released was actually only a TMT 6-plex.
[00:36:37.340]And this is where it comes from,
[00:36:38.580]because you have some of the carbon and nitrogen
[00:36:45.437]on the balance group and only one here on the reporter ion.
[00:36:49.810]And as you go up here, you go from a size of 126 to 127
[00:36:56.248]to 128 to 129 to 130, you're adding a stable isotope
[00:37:00.470]on the reporter ion group and removing it from the balance
[00:37:04.080]so that even though the overall mass is the same one
[00:37:07.330]we're gonna cleave here,
[00:37:09.140]the reporter ion mass is gonna be different.
[00:37:12.780]And here is the difference between 127N and 127C
[00:37:19.390]is that the stable isotope here is on the nitrogen,
[00:37:24.780]but here it's on the carbon.
[00:37:26.180]So that means those two reagents are gonna have a difference
[00:37:35.120]And only that small difference actually
[00:37:37.970]can still be detected by the mass spectrometer,
[00:37:41.190]only if you have a high mass resolution mass spectrometer.
[00:37:45.270]So people need to appreciate that even if you have
[00:37:49.920]a mass spec, someone with a mass spectrometer in their lab,
[00:37:53.590]that doesn't mean that the mass spectrometer
[00:37:56.950]can do everything,
[00:37:59.670]proteomics, metabolomics, and targeted TMT, it can't.
[00:38:08.250]Depending on the instrument you have you can run,
[00:38:11.300]it's optimized for running a certain type of experiment.
[00:38:14.890]And in this case, to run this type of experiment,
[00:38:17.050]you need a high resolution mass spectrometer.
[00:38:19.370]So this is just some of the spectrum we get to explain
[00:38:22.270]to you and to show you that we can resolve
[00:38:25.070]the different reporter ions.
[00:38:28.166]I've not taken data from David Holding experiment
[00:38:32.950]because it was a 8-plex.
[00:38:34.090]So I actually took from a different experiment
[00:38:36.490]where we actually use the 10 TMT plex.
[00:38:39.900]So when you label the peptides,
[00:38:42.730]all the peptide is not gonna change in mass,
[00:38:45.060]it's all gonna be the same, so as shown here.
[00:38:48.837]So in this pick, you actually have the data
[00:38:50.850]from 10 different samples.
[00:38:53.000]And when you do your MS2 scan, so after fragmentation,
[00:38:57.340]you're gonna see the reporter ion here at the lower range
[00:39:02.150]of the spectrum.
[00:39:04.170]And if I zoom in in this area, now you can see six picks.
[00:39:11.290]But in the middle ones here, for middle ones,
[00:39:15.010]you actually have two picks but they are so close
[00:39:17.500]to each other because remember, they're only a difference
[00:39:19.840]of 0.006 that it's hard to tell, but if you zoom in,
[00:39:24.570]so I'm zooming on this 127 if you zoom in on the 127,
[00:39:29.740]you will be able to see the two piece of resolve,
[00:39:33.880]one being the 127N and one of being the 127C.
[00:39:39.180]And this is how the samples are gonna be quantified
[00:39:44.900]at the MS2 level by looking at the abundance
[00:39:48.320]of this reporter ion for each sample.
[00:39:51.810]So after we ran the experiment,
[00:39:54.970]we identified over 7600 protein.
[00:40:00.076]A PC analysis showed that indeed,
[00:40:04.820]the protein content was actually very different
[00:40:07.910]between the wild type and and the mutant.
[00:40:13.264]But we should confirm some of the previous observations
[00:40:15.970]that were made and that with this changing protein content,
[00:40:21.750]what we saw is that the ribosome biogenesis pathway
[00:40:26.170]was increased while, the plant hormones transduction
[00:40:30.840]was actually decreased in RDM4.
[00:40:35.780]Alright, so now this example,
[00:40:38.430]and that will be my last example for our proteomics platform
[00:40:42.110]is to show when we actually apply for proteomics
[00:40:49.400]for this project.
[00:40:50.630]So this is a project from Joe Louis in Entomology
[00:40:54.750]and Gautam Sarath, from USDA.
[00:40:58.060]So they were working on greenbug infestation on switchgrass.
[00:41:03.380]They already had acquired data from transcriptomic
[00:41:08.800]and they wanted to confirm some of what they were seeing
[00:41:12.170]from the transcriptomic data.
[00:41:13.720]But in addition, they also wanted specifically
[00:41:16.610]to look at phosphorylation levels to see
[00:41:20.050]if they could pinpoint specific transcription factors
[00:41:25.700]that were regulated during the greenbug infestation.
[00:41:29.500]So for doing this experiment,
[00:41:31.688]we applied our phosphoproteomics platform,
[00:41:35.566]we did a phospho enrichment and this is sort of
[00:41:39.410]to illustrate why we have to do a phospho enrichment
[00:41:43.250]when you wanna look at phosphorylation,
[00:41:47.150]is because when you look at all the proteins
[00:41:50.210]that we can identify in a sample,
[00:41:52.430]this is how much we can actually identify,
[00:41:55.510]this is what you can quantify.
[00:41:57.500]And the phosphorylation amount is just down here.
[00:42:00.440]So it's really low abundant, really low stoichiometry,
[00:42:03.480]highly dynamic modification and it's also a labile.
[00:42:08.588]Which means that if you are gonna be doing phosphoproteomics
[00:42:12.270]you need to think of using phosphatase inhibitor, I'm sorry,
[00:42:17.340]and you need to make sure you keep your samples
[00:42:19.890]at minus 80 as well,
[00:42:22.540]to keep those phosphate groups attached to the protein.
[00:42:25.990]So, this is the methodology that we use for phosphorylation.
[00:42:30.850]We use sort of homemade stage tips
[00:42:34.520]with the titanium dioxide resin.
[00:42:38.030]So, the resin is placed into this tip
[00:42:41.560]and we will be pipetting the sample into the tip
[00:42:45.800]and the phosphopeptides will be binding
[00:42:47.850]to the titanium dioxide
[00:42:49.740]while all the non-phosphorylated peptide will be washed off.
[00:42:54.260]And then the last step is to use ammonia
[00:42:59.010]to release the phosphopeptide from the resin.
[00:43:02.870]So with this project ...
[00:43:06.235]So this is the study design for the project,
[00:43:09.420]so they have the switchgrass, they use aphids
[00:43:14.201]for infecting the switchgrass.
[00:43:18.660]It was at day 10 that they compared control samples
[00:43:24.070]to infected samples.
[00:43:24.903]We did a total protein and phosphoprotein study.
[00:43:29.561]And what is really interesting, I think to me,
[00:43:33.120]most interesting of this study was that really
[00:43:37.320]it showed them the amount of correlation that you can find
[00:43:40.920]between transcriptomic data and proteomics data
[00:43:44.080]and also between proteomics data and phosphoproteomics data.
[00:43:47.946]You can see in this first graph here.
[00:43:51.950]This is comparing the proteomics to transcriptomics
[00:43:57.310]that you can see that a lot of the proteins
[00:44:01.730]that were identified in the study were not even identified
[00:44:04.960]at the transcriptomic level and the reverse is also true.
[00:44:10.460]And then you have green shoeing,
[00:44:13.340]what was actually identified by both in the right direction,
[00:44:18.170]but also sometimes in the opposite direction.
[00:44:21.250]So, you can see that both type of study can actually
[00:44:25.050]give you more information and hints
[00:44:27.620]into how some proteins are regulated.
[00:44:32.030]And this final graph here is showing the comparison
[00:44:35.930]of the protein level with the phosphorylation level.
[00:44:41.890]And again, what you see is that a lot of those proteins
[00:44:46.590]were actually only showing a difference
[00:44:53.780]in the phosphorylation level and not necessarily
[00:44:56.280]in the abundance level.
[00:44:59.853]And that was the case, for example,
[00:45:01.840]for me to transcription factor.
[00:45:05.090]That was shown here that was actually also regulated
[00:45:09.130]by reversible phosphorylation.
[00:45:12.110]And it's also already known as the regulator
[00:45:14.620]of jasmonic acid responses in rise
[00:45:17.750]during planned difference.
[00:45:19.210]But the other information here is that
[00:45:22.010]by doing this type of phospho enrichment,
[00:45:24.360]we are able to enrich for nuclear proteins and specifically,
[00:45:29.430]transcription factors because transcription factors
[00:45:32.690]are very low abundant and when we do complex protein studies
[00:45:37.950]well, we don't tend to identify a lot of them
[00:45:40.250]but by doing phosphoproteomics,
[00:45:42.532]because a lot of them are phosphorylated,
[00:45:44.020]we are able to enrich for them.
[00:45:47.300]Okay, so now I'm gonna spend the last 15 minutes
[00:45:51.740]on our metabolomics platform.
[00:45:55.299]Metabolomics platform, we've been spending the last four
[00:46:00.520]or five years really working more
[00:46:03.280]on targeted metabolomics data and targeted metabolomics.
[00:46:06.761]And the limitation for that was because we were sort of
[00:46:11.210]limited by instrumentation time,
[00:46:14.430]only having one high resolution instrument in the lab.
[00:46:18.040]So the targeted metabolomics, is we have several platform
[00:46:24.260]basically that can be used the GC-MS, the HPLC,
[00:46:29.280]with a DD fluorescence detector and LSD,
[00:46:34.340]or we can use LC-MS, we also use LC-MS for targeted work.
[00:46:40.790]So, the targeted work also we typically also
[00:46:45.020]start from samples.
[00:46:46.170]So we will extract the metabolite from the sample,
[00:46:50.090]we can do extractions for more hydrophobic compounds
[00:46:53.840]or for more hydrophilic compounds.
[00:46:56.410]If we do a targeted approach, typically for the analysis,
[00:47:02.158]there is no identification step
[00:47:04.210]because we are targeting compounds,
[00:47:07.534]so we already know what they are.
[00:47:08.367]So the only data analysis we do is quantification
[00:47:11.770]and here because it's standards using standards.
[00:47:15.490]Typically, we have authentic standard for those compounds,
[00:47:18.040]we can run external standard curves
[00:47:20.150]for absolute quantification.
[00:47:23.380]We can also do targeting on the GC-MS as well,
[00:47:27.430]even though it's not really shown in this diagram.
[00:47:30.870]For the untargeted metabolomics.
[00:47:32.550]So, we've been trying to do some untargeted metabolomics
[00:47:36.161]on the GC-MS, but that requires typically derivatisation
[00:47:40.330]of the samples first.
[00:47:42.390]Obviously, we can do volatile studies on the GC-MS,
[00:47:48.150]but what was missing was being able to do
[00:47:51.023]also LC-MS untargeted work.
[00:47:54.823]And because of the fact that there are some limitations
[00:47:59.590]of only using GC-MS for untargeted,
[00:48:03.714]there was something really missing in the lab.
[00:48:05.490]So, we started doing some experiments
[00:48:08.060]on our high resolution instrument that was used,
[00:48:11.360]was used mainly for proteomics at the time.
[00:48:14.730]So I was trying to start establishing some methods
[00:48:18.760]and now with the arrival of this new instrument,
[00:48:22.920]we are gonna be able to basically
[00:48:26.010]do untargeted metabolomics profiling using the QE-HF
[00:48:32.190]which was the last instrument we were using for proteomics
[00:48:36.200]and the new instrument is gonna be used
[00:48:38.180]for more characterization of the compounds
[00:48:42.300]when we are not able to identify the compounds.
[00:48:45.530]And you will see that identification is the challenge
[00:48:48.510]number one, with untargeted metabolomics.
[00:48:52.580]And this is shown here in the data analysis
[00:48:57.641]of untargeted data is that the first thing we usually do
[00:49:04.480]is quantification, relative quantification.
[00:49:07.030]So this is discovery type, similar to proteomics.
[00:49:11.660]And then we try to identify the picks
[00:49:15.100]using my spectral libraries using authentic standards.
[00:49:20.420]Just as a note, because maybe some people don't know
[00:49:24.690]what's the range of a small molecule,
[00:49:27.590]typically we're talking about a 50 to a 1500 Dalton.
[00:49:31.460]And when I talk about metabolomics,
[00:49:34.250]I do not include lipidomics.
[00:49:38.000]Lipidomics is another field another type of expertise
[00:49:42.060]that we don't have.
[00:49:45.230]So, when to use HPLC, when to use GC-MS, or LC-MS.
[00:49:51.510]It really depends on three main factors,
[00:49:54.660]the actual compound that you're interested in,
[00:49:57.320]the class of compound you're interested in
[00:49:58.970]because unfortunately, there is not one universal extraction
[00:50:04.870]for looking at all the compounds
[00:50:06.780]and there is not one approach to look at all the compounds.
[00:50:12.410]So somehow, when you do targeted work,
[00:50:16.700]that's fine because you are targeting
[00:50:18.190]a specific class of compound.
[00:50:19.570]But when you wanna do an untargeted study,
[00:50:22.600]then you still have to make a choice as to what compound
[00:50:25.690]you wanna target.
[00:50:27.300]So GC, usually can do quite a wide range of compounds,
[00:50:32.100]obviously best for volatile compound
[00:50:34.330]but with derivatisation.
[00:50:37.350]With a derivatisation step we can study some polar compound
[00:50:43.780]and some non-polar compounds.
[00:50:47.320]When it comes to using HPLC or LC-MS,
[00:50:52.510]we need to know what sort of compound,
[00:50:54.500]class of compound you're interested in
[00:50:55.980]because we will have to use a different
[00:50:58.090]type of chromatography.
[00:51:00.220]The other factors that we also take into account
[00:51:04.270]is the abundance of the compounds in the sample.
[00:51:06.890]If it's very abundant, and it's something
[00:51:09.330]that you've extracted and purified
[00:51:11.320]to more simple compound type,
[00:51:15.090]then HPLC might just be enough, as long as obviously GG
[00:51:19.116]or for instance, detector can work.
[00:51:21.400]If it's very low, we might have to use mass spectrometry
[00:51:30.050]either GC-MS or LC-MS depending again
[00:51:32.360]on the class of compound.
[00:51:34.930]For specificity, the best is obviously to use LC-MS
[00:51:40.030]for best specificity.
[00:51:41.730]And I'm gonna talk about what I mean by that
[00:51:46.010]in the next few slides.
[00:51:48.381]So I'm not gonna necessarily talk any more about GC-MS
[00:51:50.897]and HPLC, I'm just gonna concentrate on deals
[00:51:53.460]with the LC-MS platforms that we have.
[00:51:55.770]And this is the targeted platform,
[00:51:57.530]which uses a triple quadrupole that allows you
[00:52:01.250]to do what is called MRM scan
[00:52:04.820]for multiple reaction monitoring.
[00:52:07.020]So the way it works is that because it's targeted,
[00:52:10.510]we know what we're looking for,
[00:52:12.470]we typically would have standards for that.
[00:52:14.900]We have information that we can use to make
[00:52:19.098]the method more specific for quantification.
[00:52:23.340]So we can tell the instrument when it receives
[00:52:28.173]all those ions to select for a specific mass.
[00:52:31.240]And we're gonna use fragmentation information as well
[00:52:34.520]to specifically tell the instrument to only look
[00:52:37.930]at this fragment ion, which is only its signature ion
[00:52:43.100]for this compound.
[00:52:44.130]So this is an example, for example, of gibberellic acid.
[00:52:47.700]So that means we really increase the specificity
[00:52:50.670]and we also increase sensitivity
[00:52:52.400]by using this type of scanner.
[00:52:54.420]And this is sort of illustrating what I'm trying to say
[00:52:57.950]between, if you require specificity
[00:53:01.440]and what's the difference
[00:53:02.940]between looking at HPLC chromatogram or MRM chromatogram
[00:53:07.060]is that if it was just an HPLC chromatogram,
[00:53:10.430]you would have...
[00:53:11.360]So this is showing two different compounds eluting
[00:53:14.660]at the same time, where you would have only one peak,
[00:53:17.680]one signal for both compounds.
[00:53:19.530]And you will not know how much of it comes from one compound
[00:53:23.670]and how much comes from the other compound.
[00:53:26.270]When you do mass spectrometry,
[00:53:27.720]you'll be able because of the fact that
[00:53:29.620]they have different mass.
[00:53:31.470]So those two compounds here co-eluting gibberellic acid
[00:53:34.930]and jasmonic acid isoleucine,
[00:53:36.570]because they have different size different mass,
[00:53:40.130]we can identify them separately
[00:53:43.010]and quantify them separately.
[00:53:46.000]So on this targeted instrument,
[00:53:51.240]we've been over the last few years, in the last four years,
[00:53:54.770]we've been optimizing for a set of classes of compounds,
[00:54:00.750]that was based on the projects we were getting,
[00:54:04.340]that was based on those projects we were getting.
[00:54:06.240]So this is a list of them with the number of compounds
[00:54:11.230]that are included.
[00:54:12.350]Our top one acid that we run the most
[00:54:14.940]is phytohormone acid.
[00:54:17.650]And the phytohormone and amino acids were initiated first
[00:54:23.545]by a collaboration with Daniel Schachtman in agronomy,
[00:54:28.080]who was trying to identify the compounds present
[00:54:31.860]in root exudates and in soil.
[00:54:35.923]And that was work done on maize, to trying to understand
[00:54:41.600]the root metabolism and its influence on interaction
[00:54:44.470]with soil microbes.
[00:54:45.980]So we now have been processing lots of different samples
[00:54:49.240]using this phytohormone acid.
[00:54:51.710]Lots of different species but also type of samples,
[00:54:56.290]including microbial products and also glands from wasps.
[00:55:03.320]So untargeted LC-MS, and wanted to speed up here.
[00:55:08.530]But really the take home message that you need to understand
[00:55:12.740]is really you need a high resolution mass spectrometer.
[00:55:16.500]You need a high resolution so that you can get
[00:55:19.130]all those numbers after the decimal place.
[00:55:23.100]And also, not that only but also the MS/MS fragmentation
[00:55:28.500]of the compounds to be able to identify.
[00:55:30.890]I'm not able to go through the slides but basically,
[00:55:34.240]if you only had one type of information it won't be enough
[00:55:38.040]to actually confidently identify the compound.
[00:55:41.350]Just by having MS/MS and the fact that the information
[00:55:44.870]is available in the library,
[00:55:46.544]you're able to come to identify confidently this compound.
[00:55:51.160]And this is just to show that mass spec is really powerful.
[00:55:56.600]But without LC HPLC at the front of the mass spec,
[00:56:00.458]it wouldn't be as powerful, because in this case
[00:56:03.890]where a zeatin exists as a trans and cis sterile isomer
[00:56:09.160]if we were not able to separate them chromatographically,
[00:56:12.150]we would not be able to know which one is present
[00:56:15.180]in the sample.
[00:56:16.920]So basically, this is just showing you the different steps
[00:56:21.290]for the data processing of untargeted data.
[00:56:25.830]And as you will see, quantification is done first
[00:56:30.020]and then we try to identify the compounds.
[00:56:33.490]So it's either we get a positive match,
[00:56:36.160]because we have a reference standard or you can get
[00:56:40.390]a putative match from a library search.
[00:56:44.475]And those are the databases that are used
[00:56:48.160]for this type of identification.
[00:56:50.550]But if you still don't get a match,
[00:56:52.490]there are additional tools that can be used to try
[00:56:55.150]and help you in predicting what is this unknown compound.
[00:57:01.050]And like I said, we will have also an instrument
[00:57:03.800]that will be able to help with characterizing
[00:57:06.680]also those compounds.
[00:57:08.720]So just to finish up this presentation,
[00:57:12.340]just to remind people that we have, well, not this past year
[00:57:16.810]because of the pandemic, but we've been doing workshops.
[00:57:21.720]We do this spring course every year and we also offer
[00:57:28.960]one-on-one training for self-service users.
[00:57:32.210]So just wanna thank our staff members, our user base,
[00:57:37.920]and some people here for advising me over the years
[00:57:43.180]and of course funding support.
[00:57:50.550]So maybe waiting for questions to arrive,
[00:57:53.490]I have the first question for you Sophie.
[00:57:56.870]In your presentation, you mentioned that you can work
[00:57:59.330]with cells to establish the podiums.
[00:58:03.300]Could you provide more information
[00:58:04.860]about cells that you already manage in your facility?
[00:58:08.720]How many of them do you need?
[00:58:10.140]What could be the expectation in terms of
[00:58:12.100]the number of protein you can detect?
[00:58:15.199]Okay, well, over the last four five years
[00:58:20.040]we've been processing quite a lot of a number of samples,
[00:58:23.880]I don't have an actual exhaustive list of them,
[00:58:27.900]I've not being keeping track of them but really we do
[00:58:30.960]lots of different type of cells.
[00:58:32.400]We've been doing, it could be just cell culture
[00:58:35.810]from plants Arabidopsis, or we're talking here
[00:58:39.840]about more of cell culture from mammalian cells,
[00:58:44.480]from human cells.
[00:58:45.760]So we do actually a lot of that for human and mammalian.
[00:58:50.567]When it comes to the cell type sample,
[00:58:53.810]more than we've done for plants but we can also easily do it
[00:58:57.801]for a plant sample.
[00:58:59.640]I mean, the number of proteins that we can identify
[00:59:02.310]really also depend on what is the type of cell
[00:59:05.840]but for mammalian and human, we can easily identify
[00:59:09.770]2000 proteins, two to 3000 proteins
[00:59:15.470]by using just our label-free approach.
[00:59:19.810]So do you expect to have even more sensitivity
[00:59:22.310]I guess, yes, with the new mass spectrophotometer?
[00:59:25.990]So yes, with the additional mass spectrometer,
[00:59:30.120]we are hoping that because of the increased sensitivity
[00:59:33.790]and speed that we'll be able to increase that number,
[00:59:36.860]but obviously the TMT labeling approach
[00:59:39.130]is still the best approach to get
[00:59:43.940]an increased protein coverage.
[00:59:49.273]Thanks for the questions.
[00:59:54.560]I'm going to ask another question.
[00:59:56.470]I also noticed that you mentioned this idea
[00:59:58.950]of special proteomics, is it directly related
[01:00:03.190]to this proteome interactome or are we basically moving
[01:00:08.100]into the trend that we can localized
[01:00:12.010]maybe hundreds of different proteins
[01:00:14.480]at a different kind of cell relocalization?
[01:00:18.660]Yes, so some people have been doing that.
[01:00:22.848]I mean, this is where we're really not involved
[01:00:25.730]with this type of experiment,
[01:00:27.110]'cause this is something that has to be optimized
[01:00:31.600]by the user in their lab.
[01:00:33.420]But basically, where we are involved is once they actually
[01:00:38.462]have been able to purify those specific protein
[01:00:43.648]to a location so that we can then identify them.
[01:00:48.550]But yeah, our level of expertise only comes
[01:00:55.990]after they actually ran this type of interactome study
[01:01:01.330]or special proteomic study, where we come and try
[01:01:05.550]and identify these proteins.
[01:01:07.584]But that sometimes means consulting with them
[01:01:11.580]to make sure then while they are preparing their samples,
[01:01:14.440]they're gonna be able to not use things that might interfere
[01:01:18.970]with the mass spectrometer,
[01:01:20.950]which is why we often run samples from interactome studies
[01:01:26.500]by SDS page, so that we can clean up
[01:01:29.850]the presence of detergent in the sample.
[01:01:32.600]So it's good even if we're not necessarily involved
[01:01:36.480]with this type of experiment at the beginning,
[01:01:39.810]it's nice to be able to consult as well with us
[01:01:42.100]to see if there are any alternative ways of preparing
[01:01:46.430]Because some people want us to sometimes do the digestion
[01:01:52.370]from the beads instead of running it onto a gel.
[01:01:56.760]But we can only do that if they've been using
[01:01:59.520]a very high grade of reagent that are not gonna be full
[01:02:06.040]They're gonna just swamp off our mass spectrometer.
[01:02:15.348]So thank you, Sophie.
[01:02:16.490]Thank you very much for giving this talk.
Log in to post comments