Article Cite This: Anal. Chem. 2017, 89, 13559−13566
pubs.acs.org/ac
Plasma Protein Turnover Rates in Rats Using Stable Isotope Labeling, Global Proteomics, and Activity-Based Protein Profiling Jordan Ned Smith,* Kimberly J. Tyrrell, Joshua R. Hansen, Dennis G. Thomas, Taylor A. Murphree, Anil Shukla, Teresa Luders, James M. Madden, Yunying Li, Aaron T. Wright, and Paul D. Piehowski Pacific Northwest National Laboratory, Richland, Washington 99354, United States S Supporting Information *
ABSTRACT: Protein turnover is important for general health on cellular and organism scales providing a strategy to replace old, damaged, or dysfunctional proteins. Protein turnover also informs of biomarker kinetics, as a better understanding of synthesis and degradation of proteins increases the clinical utility of biomarkers. Here, turnover rates of plasma proteins in rats were measured in vivo using a pulse−chase stable isotope labeling experiment. During the pulse, rats (n = 5) were fed 13C6-labeled lysine (“heavy”) feed for 23 days to label proteins. During the chase, feed was changed to an unlabeled equivalent feed (“light”), and blood was repeatedly sampled from rats over 10 time points for 28 days. Plasma samples were digested with trypsin and analyzed with liquid chromatography−tandem mass spectrometry (LC−MS/ MS). MaxQuant was used to identify peptides and proteins and quantify heavy/light lysine ratios. A system of ordinary differential equations was used to calculate protein turnover rates. Using this approach, 273 proteins were identified, and turnover rates were quantified for 157 plasma proteins with half-lives ranging 0.3−103 days. For the ∼70 most abundant proteins, variability in turnover rates among rats was low (median coefficient of variation: 0.09). Activity-based protein profiling was applied to pooled plasma samples to enrich serine hydrolases using a fluorophosphonate (FP2) activity-based probe. This enrichment resulted in turnover rates for an additional 17 proteins. This study is the first to measure global plasma protein turnover rates in rats in vivo, measure variability of protein turnover rates in any animal model, and utilize activity-based protein profiling for enhancing turnover measurements of targeted, low-abundant proteins, such as those commonly used as biomarkers. Measured protein turnover rates will be important for understanding of the role of protein turnover in cellular and organism health as well as increasing the utility of protein biomarkers through better understanding of processes governing biomarker kinetics.
P
degradation may significantly contribute to the aging process.7−9 Protein turnover also plays a role in diseases not related to aging. For example, cystic fibrosis is caused by increased levels of cystic fibrosis transmembrane conductance regulator (CFTR) degradation.2,10 Chronic obstructive pulmonary disease (COPD) has been associated with higher levels of whole-body protein turnover,11 much like acute and chronic liver disease.12 These diseases are among many in which protein turnover dysfunction plays an important role. Protein biomarkers, especially those found in plasma, are commonly used to diagnose or monitor disease state, disease susceptibility, normal biological processes, or toxicological response. For example, the Food and Drug Administration (FDA) has approved protein biomarkers for clinically diagnosing and monitoring treatment of cancers, including CA125 (mucin 16) for ovarian cancer, prostate-specific antigen
roteins in living cells undergo constant synthesis and degradation. Protein turnover is the balance of these processes and is important for maintaining cellular homeostasis.1 Protein turnover rates are variable among different proteins ranging from minutes to years.2,3 It is hypothesized that the ability of a protein to respond to environmental stimuli is determined by the protein’s turnover rate.4 As such, proteins of high abundance and low turnover rates generally execute more housekeeping functions,5 while smaller proteins with rapid turnover rates are generally more involved with rapid cellular response.4 On an organism scale, protein turnover is important for general health, as a number of physiological abnormalities have been linked with protein turnover dysfunction. Routine protein turnover is a key cellular strategy for combating aging, as old, damaged, or dysfunctional proteins are routinely replaced.6 Loss of protein turnover function plays a major role in aging and many age-related diseases (e.g., Parkinson’s and Alzheimer’s diseases), and it has been hypothesized that damage to long-lived intracellular proteins that elude protein © 2017 American Chemical Society
Received: September 28, 2017 Accepted: November 22, 2017 Published: November 22, 2017 13559
DOI: 10.1021/acs.analchem.7b03984 Anal. Chem. 2017, 89, 13559−13566
Article
Analytical Chemistry
of protein turnover in cellular and organism health as well as processes governing biomarker kinetics.
(PSA) for prostate cancer, and CA19-9 for pancreatic cancer.13,14 Due to expanding proteomic sample preparation techniques as well as separation and mass spectrometry capabilities, proteins are frequently targeted for biomarker discovery. However, translation of protein biomarker candidates to clinical use has been increasingly unsuccessful.13−16 A better understanding of biomarker kinetics has been advocated to increase protein biomarker clinical relevance.17,18 Mechanisms of protein turnover are responsible for determining levels of protein biomarkers. While studies evaluating protein biomarker kinetics have been increasing,17−19 many researchers are still impeded by a lack of understanding of quantitative protein turnover processes, and better measurements are needed, especially for endogenous proteins. Recently, technological advances has allowed for global measurements of protein turnover rates. Protein turnover rates have been measured in various systems using isotopic labeling techniques coupled with proteomic measurements using mass spectrometry. In these techniques, stable isotope amino acids (e.g., 13C6 lysine, 13C6 arginine, or 15N labeling of all amino acids) are used to label proteins by adapting classic pulse−chase or chronic labeling strategies.2,4,20,21 Mass spectrometry can be used to determine ratios of labeled versus unlabeled proteins, and kinetic analyses of these ratios can be used to determine turnover rates. Using this general approach, protein turnover rates have been measured in cell culture systems,22 insects,23 zebrafish,24 tissues and blood of mice,2,25−27 and sperm of mice.28 One particular challenge of using global proteomics approaches to measure turnover rates in plasma is the large dynamic range of protein abundances, which span 10 orders of magnitude.15,29 Considering that the 12 most abundant proteins comprise 95% of total plasma protein, detection and quantification of the remaining less-abundant proteins can be challenging.15,29 One strategy that has been applied to enrich different classes of proteins based on active-site reactions is activity-based protein profiling.30−32 Activity-based protein profiling has been used to successfully enrich a number of biomarkers in complex biological samples, such as serine hydrolases.33,34 Serine hydrolases are a diverse group of enzymes that have been studied for their role in various biological functions like blood coagulation, inflammation, digestion, and nervous cell signaling.35 Serine hydrolases have been used as biomarkers for carbamate and organophosphate exposures34,36 as well as various cancers.37 Due to the importance of protein turnover for cellular and organism health and for determining biomarker levels and kinetics, the goal of this work was to accurately measure plasma protein rates in rats. Protein turnover has not been measured in rats, and biological variability of protein turnover has not been robustly assessed on a global scale in any animal model. Thus, because of the ability to take repeated blood measurements from the same animal, the rat was utilized for global measurements of plasma protein turnover after labeling with isotopically labeled lysine feed. A second focus of this study was to understand protein turnover rates of biomarkers, using serine hydrolases as a model class. An activity-based protein profiling strategy was used to enrich serine hydrolases for turnover rate measurements allowing us to target this enzyme family of interest that could not be easily measured with global proteomic techniques due to the high dynamic range of proteins in plasma. Results of this study provide reliable protein turnover rates that will improve our understanding of the role
■
MATERIALS AND METHODS Chemicals. All chemicals and biological reagents were purchased from Sigma unless otherwise noted. Animals. Rats were used as an animal model to measure plasma protein turnover rates. Adult male Sprague−Dawley rats (n = 5, 200 g, 7 weeks old) were purchased from Charles River Laboratories Inc. (Wilmington, MA, U.S.A.). Rats were housed individually in solid bottom cages using α cellulose bedding (Shepherd’s ALPHA-dri, Animal Specialties, Inc., Hubbard, OR, U.S.A.) under standard laboratory conditions. Water was available ad libitum. All procedures involving animals were in accordance with protocols established in the NIH/NRC Guide and Use of Laboratory Animals (NIH/NRC) and were reviewed by the Institutional Animal Care and Use Committee of Battelle, Pacific Northwest Division. In Vivo Methods. Rat plasma proteins were labeled in vivo using a stable isotope amino acid pulse−chase labeling strategy (Figure 1). Upon arrival, rats were fed 28 g/day of 13C6-labeled (“heavy”) lysine (99%) mouse chow (MouseExpress L-lysine, Cambridge Isotopes Laboratories, Inc., Tewksbury, MA, U.S.A.) for 23 days. In a previous study, this level of feed consumption was 88% of ad libitum.38 After 23 days consuming 13 C6-labeled lysine mouse chow, rats were fed an identical feed with unlabeled (“light”) lysine (day 0) administered at 28 g/ day. Blood (20−50 μL) was sampled from each rat on days 0, 1, 2, 3, 4, 7, 10, 14, and 21 from either the lateral tail vein or saphenous vein using a Goldenrod animal lancet (Braintree Scientific, Inc., Braintree, MA) and Ram Scientific Safe-T-Fill capillary blood collection systems with EDTA as an anticoagulant (Fisher Scientific Co. LLC, Pittsburgh, PA). During blood collections, rats were chemically restrained using isoflurane (Baxter Healthcare Corp., Deerfield, IL, U.S.A.) mixed with oxygen administered using an inhalation anesthesia machine (VetEquip Inc., Pleasanton, CA, U.S.A.). On day 28, rats were euthanized using CO2 gas as an asphyxiant followed by cervical dislocation, and blood was collected by intracardiac puncture. Blood samples were centrifuged for 10 min at 1600 rcm, and plasma was separated from the packed red blood cell fraction. Both fractions were frozen until further analysis. Global Proteomic Sample Preparation. Plasma samples were prepared for liquid chromatography−tandem mass spectrometry (LC−MS/MS) analysis as previously described.39 Briefly, an aliquot (5 μL) of rat plasma was taken from each sample and transferred to a 1 mL 96 well plate for digestion. Samples were denatured and reduced by adding 200 μL of 8 M urea with 10 mM dithiothreitol in 50 mM ammmonium bicarbonate at pH 8 and incubating for 30 min at 37 °C. Cysteine alkylation was achieved by adding iodoacetamide to a final concentration of 40 mM followed by incubation at 37 °C in the dark for 30 min. Samples were diluted 4-fold in 50 mM ammonium bicarbonate (pH 8). Tryptic digestion was carried out by the addition of sequencing grade trypsin (Promega) at an enzyme/substrate ratio of 1:50 followed by incubation for 6 h at 37 °C. The resulting peptides were desalted using C18 solid-phase extraction in a 96 well plate format (Phenomenex). Peptide concentration was determined by BCA assay (Pierce), and samples were diluted to a final concentration of 0.2 μg/μL. Activity-Based Protein Profiling Sample Preparation. Activity-based protein profiling was used to enrich serine hydrolases for turnover rate measurement. Remaining plasma 13560
DOI: 10.1021/acs.analchem.7b03984 Anal. Chem. 2017, 89, 13559−13566
Article
Analytical Chemistry
Probe-labeled proteins were enriched with streptavidin agarose resin (Thermo Fisher Scientific, Rockford, IL; 1−3 mg of biotinylated BSA protein/mL of resin). The resin was placed in a Bio-Spin chromatography column (Bio-Rad, Hercules, CA) on a vacuum manifold and was washed 2 × 1 mL with 0.5% SDS in PBS, 2 × 1 mL 6 M urea in 25 mM ammonium bicarbonate (NH4HCO3), and 4 × 1 mL PBS. The resin was transferred to 4 mL cryovials using two 1 mL aliquots of PBS. An additional 0.5 mL of PBS was added to each tube followed by 1600 μg of protein (in 1.2% SDS in PBS). Normalized protein loading onto streptavidin results in an analysis that evaluates protein function on a per unit mass basis. The total volume of each tube was set to 3.0 mL, giving a final SDS concentration of 0.2%. Tubes were rotated end over end for 4 h at room temperature. Following streptavidin capture of FP2-labeled proteins, the solution was transferred into the BioSpin columns, and the solution was removed. The resin was washed with 0.5% SDS in PBS (1 mL, repeat 2×), 6 M urea in 25 mM NH4HCO3 (1 mL, repeat 2×), Milli-Q water (1 mL, repeat 2×), PBS (1 mL, repeat 8×), and 25 mM NH4HCO3 (1 mL, repeat 4×). The resin was transferred to sealed 1.5 mL tubes using two 0.5 mL aliquots of 25 mM NH4HCO3. Samples were centrifuged at 10 500g, and the supernatant was discarded. Enriched probe-labeled proteins were prepared for LC−MS/ MS analysis. NH4HCO3 (200 μL) was added to the resin for each sample, along with trypsin solution. Resin solutions were placed on the thermal mixer at 37 °C set at 1200 rpm for 3 h. Following trypsin digestion, tubes were centrifuged at 10 500g for 4 min, and the supernatant was carefully collected. To the resin was added NH4HCO3 (150 μL), and tubes were placed on a thermal mixer at 37 °C set at 1200 rpm for 10 min. Again, the tubes were centrifuged at 10 500g for 4 min, and the supernatant was carefully collected and added to the prior collection. Volatiles were then removed from the combined tryptic peptide supernatant using a speed vacuum. The dried peptides were reconstituted in NH4HCO3 (40 μL) and heated for 10 min at 37 °C with mild agitation. To remove any solid particulates, samples were centrifuged at 53 000g for 20 min at 4 °C. From each ultracentrifuge vial was removed 25 μL for subsequent MS analysis. Samples were stored at −20 °C until analysis. LC−MS/MS Analysis. Ratios of light to heavy lysine in rat plasma proteins were measured using LC−MS/MS. Peptide samples were separated using a Waters nano-Aquity UPLC system (Waters) equipped with a homemade 75 μm i.d. × 70 cm length C18 column packed with 3 μm particles (Phenomenex). A 100 min gradient of 100% mobile phase A [0.1% (v/v) formic acid in water] to 60% (v/v) mobile phase B [0.1% (v/v) formic acid in acetonitrile] was used. This system was coupled to a Thermo Q-Exactive mass spectrometer for MS/MS analysis. MS Spectra were collected from 300 to 1800 m/z at a mass resolution setting of 35 000. A top 12 method was used with an isolation width of 2 m/z with +1 charges excluded and an MS2 mass resolution of 17 500. For the activity probe enriched samples, a Thermo Velos Orbitrap Elite mass spectrometer was used with the same mass resolution as Q-Exactive analyses and using a top 10 method. Protein Identification and Quantification. LC−MS/MS data was used to identify plasma proteins and determine ratios of heavy and light lysine within proteins. Thermo.raw files were processed using MaxQuant (ver. 1.5.3.30). Spectra were searched against the Uniprot database for Rattus norvegicus downloaded in May 2016. Searches were carried out using
Figure 1. Experimental workflow to measure protein turnover rates in rats using a pulse−chase stable isotope labeling experiment. During the pulse, rats (n = 5) were fed 13C6-labeled lysine (“heavy”) feed for 23 days to label proteins (1). During the chase, feed was changed to an unlabeled equivalent feed (“light”) (2). Blood was repeatedly sampled from rats over 10 time points for 28 days (3). Plasma samples from individual rates were digested with trypsin for global proteomic analysis (4A). Additionally, activity-based protein profiling was applied to pooled plasma samples to enrich serine hydrolases using a fluorophosphonate (FP2) activity-based probe (4B). Samples were analyzed with liquid chromatography−tandem mass spectrometry (LC−MS/MS) (5). MaxQuant was used to identify peptides and proteins and quantify heavy/light lysine ratios (6). A system of ordinary differential equations was used to calculate protein turnover rates (7).
from 5 rats was pooled within each time point to ensure enough sample for enrichment. For each time point, a 500 μL aliquot of pooled rat plasma, diluted to 4 mg/mL protein in phosphate-buffered saline (PBS), was incubated (90 min, 37 °C) with 100 μM fluorophosphonate (FP2) probe.40 Afterward, samples were incubated (90 min, 37 °C) with TEV− biotin−azide (60 μM), sodium ascorbate (10 mM, prepared fresh in water), tris(3-hydroxypropyltriazolylmethyl)amine (THPTA) (4 mM), and CuSO4 (8 mM, prepared in water). Proteins were precipitated from solution using ice-cold methanol and centrifugation (14 000g for 4 min at 4 °C). The supernatant was discarded, and the pellet was dissolved in PBS with 1.2% sodium dodecyl sulfate (SDS) with the aid of probe sonication and thermal mixing (95 °C). Samples were centrifuged (14 000g for 4 min at 4 °C), and protein concentrations were measured using the BCA assay. 13561
DOI: 10.1021/acs.analchem.7b03984 Anal. Chem. 2017, 89, 13559−13566
Article
Analytical Chemistry
3.2.3 (R Foundation for Statistical Computing, Vienna, Austria). Optimizations of model parameters were achieved using a maximum log likelihood objective and a Nelder−Mead algorithm.
Lys(6) as the heavy label type using a maximum number of labeled amino acids per peptide of 3. Oxidation of methionine and N-terminal acetylation were allowed as dynamic modifications, and carbamidomethylation of cysteine was set as a fixed modification. The “match between runs” feature was enabled using default parameters. All other parameters were MaxQuant defaults for Orbitrap analyses including peptide and protein false discovery rate (FDR) of 0.01. All further data analysis was carried out using the proteinGroups output table, requiring a minimum of two unique peptides for confident protein identification and using the Ratio H/L column for quantitative analyses. Protein Turnover Rate Modeling. A model describing the kinetics of proteins with heavy and light lysine was used to calculate protein turnover rates. Proteins are undergoing constant synthesis and degradation, and these processes have been hypothesized and modeled as first-order kinetic processes.4,21,25,26 To calculate protein turnover rates, we described the disposition of proteins amounts with light and heavy lysine over time using a system of ordinary differential equations. Protein synthesis is usually modeled as a first-order process dependent on the size of an amino acid pool.25 Since we do not have information on the size of various amino acid pools, including light or heavy lysine, we assumed a zero-order process. Over short periods of time, zero-order processes are good estimates of the overall rate of protein synthesis until the size of amino acid pool changes significantly. As such, the change in the amount of protein with light (APl, eq 1) or heavy (APh, eq 2) lysines was modeled as a function of time (t) with a zero-order formation rate constant for each lysine isotope (kf) and a first-order degradation rate constant (kd) for each protein, independent of lysine isotope. The overall change in the amount of protein (APt) as a function of time (eq 3) was the sum of the light and heavy lysine protein dispositions. dAPl = k fl − kd(AP) l dt
(1)
dAPh = k fh − kd(APh ) dt
(2)
dAPt = k fl + k fh − kd(APl + APh ) dt
(3)
■
RESULTS No overt toxicity was observed in rats consuming heavy lysine feed. For most days, rats did not consume all of the provided feed, suggesting that 28 g feed/day was approximately ad libitum consumption for these rats. Using global proteomic techniques, 2292 unique peptides were discovered and used to identify 273 unique proteins in rat plasma (Supporting Information Table 1). Trypsin digestion hydrolyzed proteins at lysine and proline amino acids. Since a 13 C6-labeled lysine strategy was employed, approximately 64% of the resulting peptides and 99% of proteins contained at least one lysine amino acid, allowing ratios of heavy to light lysine to be globally measured for the majority of confidently identified proteins. Fraction of heavy to light lysine in proteins were highest in early time points and decreased over time for the majority of proteins (Figure 2), as expected. Within individual proteins, labeled ratios were highly consistent among the five rats.
Time course ratios of light/heavy lysine measured by mass spectrometry were used to parametrize the protein turnover model. Ratios of light/heavy lysine were converted to fractions of light or heavy lysine, and the total amount of protein was assumed to be 1 over the duration of the experiment. Three parameters (kfl, kfh, and kd) along with the initial amounts of proteins with heavy and light lysine were optimized to the time course data of fractions of proteins with light or heavy lysine. Two different data sets were used to optimize parameters within individual proteins: (1) fits to data from all rats and (2) fits to data from each rat as individuals. The model was fit to data that had valid fraction measurements. Bayesian information criterion (BIC) was used to select confident fits. BICs of
