Depletion of Abundant Plasma Proteins by Poly(N ... - ACS Publications

Article pubs.acs.org/ac

Depletion of Abundant Plasma Proteins by Poly(N‑isopropylacrylamide-acrylic acid) Hydrogel Particles Gerard Such-Sanmartín, Estela Ventura-Espejo, and Ole N. Jensen* Department of Biochemistry and Molecular Biology, University of Southern Denmark, Odense M, DK-5230, Denmark S Supporting Information *

ABSTRACT: Protein and proteome analysis of human blood plasma presents a challenge to current analytical platforms such as mass spectrometry (MS). High abundance plasma proteins interfere with detection of potential protein biomarkers that are often 3−10 orders of magnitude lower in concentration. We report the application of pH-sensitive poly(N-isopropylacrylamide-acrylic acid) hydrogel particles for removal of abundant plasma proteins, prior to proteome analysis by MS. Protein depletion occurs by two separate mechanisms: (1) hydrogel particles incubated with low concentrations of plasma capture abundant proteins at higher efficiency than low abundance proteins, which are enriched in the supernatants, whereas (2) hydrogel particles incubated with high concentrations of plasma capture and irreversibly trap abundant proteins. During the elution step, irreversibly trapped proteins remain captured while low abundance proteins are released and recovered in the eluate. We developed a series of distinct depletion protocols that proved useful for sample depletion and fractionation and facilitated targeted analysis of putative biomarkers such as IGF1-2, IBP2-7, ALS, KLK6-7, ISK5, and PLF4 by selected reaction monitoring (SRM) liquid chromatography (LC)-MS/MS. This novel use of hydrogel particles opens new perspectives for biomarker analysis based on mass spectrometry.

T

been widely studied for its capacity to trap soluble proteins based on electrostatic and hydrophobic interactions.13−20 This negatively charged hydrogel becomes protonated at low pH conditions in turn increasing its hydrophobicity and expelling water, which leads to shrinking of the hydrogel and a concomitant reduction of particle diameter and pore size. This process also occurs during protein absorption to the hydrogel particles: as the cationic proteins bind the anionic acid groups, particles experience a continuous shrinking equivalent to the addition of an acid.20 Hydrogel particles have been explored for many diverse uses such as intelligent delivery systems or biosensors (see reviews21,22). However, the number of applications in the proteomics field is still low. One recent application was based on the utilization of the polymeric mesh for molecular weight “sieving” of proteins, thereby capturing only the low molecular weight fraction while excluding medium and large size proteins.23 We present a novel application of hydrogel particles for sample preparation of human blood plasma samples that enables MS-based detection of low abundance proteins. Synthesized anionic hydrogel particles exhibited a capacity to deplete high abundance plasma proteins in a concentration

he discovery and monitoring of low abundance plasma biomarkers is one of the largest challenges in proteomics.1 This is due to the enormous complexity of the human plasma proteome, which probably contains more than 20 000 different protein species2 at concentrations that span over 10 orders of magnitude.3 The most abundant plasma protein, serum albumin, constitutes approximately 50% of the protein content in plasma, whereas the 22 most abundant proteins represent 99% of the total protein content.4 The large differences in individual protein concentrations in plasma have implications for mass spectrometry (MS) driven proteomics experiments. It leads to matrix effects, undersampling, and suppression during electrospray ionization,5 limiting the accessible concentration range for protein analysis to 3−4 orders of magnitude.6 This challenge can to some extent be overcome by extensive fractionation of plasma samples at the protein or peptide level by using (combinations of) centrifugal ultrafiltration,7 chromatography,8 electrophoresis,9 or precipitation.10 Other methods such as affinity based depletion of abundant proteins up to the top-20 most abundant plasma proteins11 are widely used in plasma proteomics projects although at the cost of increased complexity and limited throughput in the sample preparation protocols. Poly(N-isopropylacrylamide) (pNIPAm) based hydrogel particles are well-characterized water-swellable polymers responsive to temperature, pH, and other external stimuli.12 The family of anionic pH-sensitive hydrogels containing carboxylic groups such as acrylic acid (pNIPAm-AAc) has © 2014 American Chemical Society

Received: September 22, 2013 Accepted: January 8, 2014 Published: January 8, 2014 1543

dx.doi.org/10.1021/ac403749j | Anal. Chem. 2014, 86, 1543−1550

Analytical Chemistry

Article

temperature under axial rotation. Supernatants containing noncaptured proteins were recovered after centrifugation (20 830g, 10 min). Hydrogel particles were washed three times by successive centrifugations as previously described. Protein elution was performed by addition of the corresponding elution buffer of each protocol (volumes and composition indicated in the Results section), followed by incubation for 15 min. In protocols with two elution steps, a second elution buffer was added after a 15 min incubation and incubated for an additional 40 min. Supernatants were transferred to Eppendorf tubes and evaporated in a vacuum centrifuge until dryness for subsequent SDS-PAGE analysis (see SDS-PAGE Analysis section) or concentrated in a 10 kDa 0.5 mL centrifugal ultrafiltration device (Millipore, Hellerup, Denmark) for subsequent in-filter digestion (see below). In-Filter Digestion. In-filter tryptic digestion was performed for buffer exchange and removal of small analytes and nondigested proteins.25,26 Depleted samples were concentrated in a 10 kDa 0.5 mL centrifugal ultrafiltration device (Millipore), followed by the addition of 400 μL of 50 mM ABC at pH 7 and centrifugation (washing step performed two times). Filtrates were discarded. Proteins were reduced with 100 μL of 5 mM dithiothreitol (DTT) in ABC with 1% (w/v) sodium deoxycholate27,28 (SDC) (ABC-SDC) for 30 min at 56 °C. Protein alkylation was followed by addition of 5 μL of 200 mM 2-iodoacetamide (IAM) in ABC-SDC for 30 min in the dark at room temperature. The sample was centrifuged and followed by two washing steps (see previous description). Proteins were digested with 500 ng of porcine trypsin in 100 μL of ABC-SDC and kept at 37 °C for 16 h. The generated peptides were passed through the filter by centrifugation following an additional wash with 100 μL of H2O and were recovered in the filtrate. SDC was removed by precipitation with the addition of formic acid (FA) to a final concentration of 5% (v/v), followed by centrifugation (20 830g, 10 min). Supernatants were transferred to a microcolumn consisting of a 200 μL pipet tip packed with 2 layers of SDB-XC Empore Disk (Sigma-Aldrich), previously pre-equilibrated first with 50 μL of ACN and second with 100 μL of 0.1% (v/v) FA.29 After sample loading, the microcolumn was washed with 100 μL of 0.1% FA, and peptides were eluted with 50 μL of 50% ACN 0.1% FA. The eluted solution was evaporated in a vacuum centrifuge until dryness, resuspended in 20 μL of 1% FA and analyzed in triplicate. SDS-PAGE Analysis. Samples were resuspended in NuPAGE LDS Sample Buffer prior to SDS-PAGE analyses, performed in NuPAGE 4−12% Novex Bis-Tris gels and MOPS SDS running buffer (Invitrogen, Copenhagen, Denmark), using a standard SDS-PAGE protocol.30 Proteome Analysis by LC-MS/MS. LC-MS analyses were performed using an Easy-nLC II (Thermo Fisher Scientific, Odense, Denmark) coupled to a LTQ Orbitrap Velos hybrid instrument (Thermo Fisher Scientific, Bremen, Germany) equipped with a Nanospray Flex Ion Source interface (Thermo Fisher Scientific, Odense, Denmark). Peptides were loaded onto a custom-made emitter of 20 cm × 75 μm packed with ReproSil-Pur 120 AQC18 3 μm reverse phase material from Dr. Maisch GmbH (AmmerbuchEntringen, Germany). An analysis cycle encompassed column equilibration and sample loading with solvent A (0.1% v/v FA) at maximum pressure (280 bar), running the gradient at room temperature and with a flow rate of 200 nL/min from 0% to 34% of solvent B (95% ACN, 0.1% FA) in 147 min, to 100% in 5 min and at 100% for 8 min. The cone voltage was set to 2300 V and 250 °C, operating in

dependent manner. We optimized the hydrogel particles-based depletion method and developed a series of complementary sample processing protocols that were employed for comprehensive proteome analysis, using trypsin digestion for peptide generation and liquid chromatography (LC)-MS/MS for protein detection. We demonstrate the usefulness of this approach for targeted analysis by selected reaction monitoring (SRM) LC-MS/MS analysis for the detection of 13 plasma proteins in the ng/mL concentration range.

■

EXPERIMENTAL SECTION Reagents. All reagents were obtained from Sigma-Aldrich (Copenhagen, Denmark) unless otherwise stated. Milli-Q grade H2O was generated with a Purelab Ultra Elga system (Glostrup, Denmark). Synthetic stable isotope labeled peptides with 13 C/15N-Lys or 13C/15N-Arg in C-terminus were purchased from JPT (Berlin, Germany). These included the sequences APQTGIVDECCFR, GIVEECCFR, EPGCGCCSVCAR, ALAQCAPPPAVCAELVR, THEDLYIIPIPNCDR, AVYLPNCDR, APAVAEENPK, TELLPGDR, DFALQNPSAVPR, LISPQDCTK, LSELIQPLPLER, AVFLTEALER, and AGPHCPTAQLIATLK for the proteins IGF1, IGF2, IBP2, IBP3, IBP4, IBP5, IBP6, IBP7, ALS, KLK7, KLK5, ISK5, and PLF4, respectively. Peptides were resuspended in 20% acetonitrile (ACN)/80% 50 mM ammonium bicarbonate (ABC) at pH 7, and stored at −20 °C. PNIPAM-AAc Hydrogel Particles Synthesis. pNIPAmAAc hydrogel particles were synthesized by free radical precipitation polymerization.24 NIPAm, N-N′-methylenebis(acrylamide), and AAc monomers were dissolved in H2O at concentrations of 86, 7, and 12 mM, respectively, in a final volume of 100 mL and purged with argon for 30 min. The solution was kept at 75 °C for 30 min prior to the addition of potassium persulfate salt to a final concentration of 0.9 mM. After 5 h under mild mixing, the solution was cooled to room temperature and left overnight. The hydrogel particles solution was aliquoted in several 1.5 mL Eppendorf tubes in volumes of 1 mL and washed five times with 1 mL of H2O by resuspending the pellet through vortexing and centrifuging at 20 830g for 10 min, discarding the supernatant. Pellets were further resuspended in 1 mL of H2O, pooled in a 50 mL Falcon tube, and diluted with H2O to a final volume of 45 mL. The solution was stored at 4 °C. The dry weight of the particles indicated an approximate particle concentration of 18 g/L. This stable homogeneous suspension was used as the stock solution for the pNIPAm-AAc hydrogel particles. Mean particle diameter was measured at 23 °C by dynamic light scattering measurements (BI-200SM, Brookhaven Instruments, Vienna, Austria). Laser wavelength was 633 nm, and scattering angle was 90 degrees, measuring for 120 s, giving an average particle diameter of 2.4 μm. Human Plasma Depletion with pNIPAm-AAc Hydrogel Particles. Plasma samples were depleted by using hydrogel particles based on eight different protocols defined by specific incubation, washing, and elution conditions. The incubation step consisted of having different volumes of pooled human blood plasma samples (Innov-research, Handen, Sweden) (61.6 g/L protein concentration, determined by Qubit Fluorometric Quantitation, Invitrogen, Copenhagen, Denmark) diluted to a final volume of 800 μL with H2O and mixed with different volumes of hydrogel particles at specific plasma-to-particle ratios. Initial volumes of plasma and particles are detailed in the Results section. The mixture was incubated for 20 min at room 1544

dx.doi.org/10.1021/ac403749j | Anal. Chem. 2014, 86, 1543−1550

Analytical Chemistry

Article

Figure 1. SDS-PAGE profiles of noncaptured (A) and captured (B) plasma proteins upon incubation with hydrogel particles. Incubation of plasma and hydrogel particles at LOW and HIGH plasma-to-particle ratios enable protein depletion by recovering the noncaptured and the captured proteins, respectively. Each lane corresponds to incubations at increasing volumes of plasma for a fixed volume of 50 μL of the stock hydrogel particles solution. Volumes are indicated at the bottom of the figure. Lane 1: plasma (0.5 μL).

were calculated using a “precursor ions area detector” with mass accuracy of 2 ppm, accounting for the areas of the top 3 most abundant peptides detected.32 The database employed was UniProtKB/Swiss-Prot (2013.01) filtered for Homo sapiens taxonomy. Results were filtered with peptide false discovery rate (FDR) of 5% calculated with Percolator, peptides of rank 1, and protein Mascot score above 10. Proteins are reported as protein groups, i.e., the minimum set of protein sequences that adequately accounts for all observed peptides.

positive ion mode and using argon as a collision gas. MS measurements were performed at a mass resolution setting of 60.000 fwhm (at 400 m/z), acquiring the top 15 CID MSMS spectra triggered at an ion intensity of 5000 using Wideband Activation (normalized collision energy of 35%, 15 ms activation time), 500 ms/1 × 106 and 150 ms/5 × 103 as maximum injection times and signal for the FTMS (MS) and IT (MSn), respectively. Data was acquired using Thermo Xcalibur version 2.1.0.1140 software. Selected Reaction Monitoring (SRM) LC-MS/MS Analysis. Targeted peptide analyses were performed using an Easy-nLC II coupled to a TSQ Vantage triple quadrupole mass spectrometer (Thermo Fisher Scientific, Bremen, Germany). Peptides were loaded onto a custom-made emitter of 15 cm × 75 μm packed with ReproSil-Pur 120 AQC18 3 μm reverse phase material from Dr. Maisch GmbH (AmmerbuchEntringen, Germany). An analysis cycle encompassed column equilibration and sample loading with solvent A (0.1% v/v FA) at maximum pressure (280 bar) running the gradient at room temperature and 250 nL/min from 5% to 20% of solvent B (95% ACN, 0.1% FA) in 15 min, to 40% in 5 min, to 100% in 4 min, and at 100% for 8 min. The cone voltage was set to 2300 V, and the cone temperature was set to 200 °C, operating in positive ion mode and using argon as a collision gas. MS measurements were obtained at a mass resolution of 0.7 Da fwhm (at 400 m/z). Data was acquired using Thermo Xcalibur 2.0.7 software. Collision energy and data analysis was performed using Skyline 1.2 (MacCoss Lab Software31), using peptide ion peak areas for quantification. Protein Identification. Protein identification was performed with Proteome Discoverer v1.4.0.270 (Thermo Fisher Scientific) and Mascot v2.3.02 (Matrix Science, London, UK). Mascot searches were performed using the following settings: maximum of 2 missed cleavages using trypsin/P (cleaves Lys and Arg), 10 ppm and 0.6 Da for precursor and fragment mass tolerance, respectively. Carbamidomethyl was set as a static modification of cysteines, and deamidation (Asn-Gln) and oxidation (Met) were set as dynamic modifications. Peak areas

■

RESULTS Concentration Dependent Interaction between Hydrogel Particles and Blood Plasma Samples. Anionic hydrogel particles are known to exhibit strong protein absorption based on ionic and hydrophobic interactions. We studied how these particles interact with proteins depending on the initial sample concentration, defined here as plasma-toparticle ratios. Synthesized hydrogel particles were incubated with human blood plasma, a very complex mixture of proteins of different properties and sizes, at volumes ranging from 1 to 40 μL in a series of increasing plasma-to-particle ratios. We analyzed both the noncaptured proteins by recovering supernatants and the captured proteins by eluting them from hydrogel particles using acidic conditions. SDS-PAGE analyses of the noncaptured and captured proteins are shown in Figure 1A and 1B, respectively. Exact volumes of plasma and particles used in the incubation are indicated in the figure. The MW range of the proteins observed after incubation with hydrogel particles did not show any significant difference when compared to the MW range initially present in plasma (Figure 1, lane 1). As such, there was an absence of any MW discrimination in the capture process, i.e., hydrogel particles did not capture proteins based on their MW. The SDS-PAGE image of the noncaptured proteins (Figure 1A) revealed different protein profiles depending on the plasma-to-particle ratios, as indicated by two groups marked as LOW (lanes 2−4) and HIGHER (lanes 5−8). These profiles showed a different distribution of the most abundant proteins 1545

dx.doi.org/10.1021/ac403749j | Anal. Chem. 2014, 86, 1543−1550

Analytical Chemistry

Article

in the mixture. When employing low plasma-to-particle ratios, some of the most abundant proteins were removed including, e.g., serum albumin (∼65 kDa) or immunoglobulins (∼150 kDa) (Figure 1). This phenomenon of protein removal was further explored as discussed in the sections below. Inspection of the SDS-PAGE image of the captured proteins (Figure 1B) revealed changes in the protein profiles depending on the plasma-to-particle ratios, as indicated by two groups marked as LOWER (lanes 2−8) and HIGH (lanes 9−15). In this case, plasma proteins that were initially captured were eluted with an acidic solution to disrupt protein−particle interactions. Captured proteins increased in concentration at higher plasma volumes. However, some of these proteins showed a sharp and progressive decrease of concentration for incubations at plasma volumes above a particular threshold, ending up with a complete depletion from the mixture. This is illustrated with serum albumin, which was depleted at the plasma-to-particle ratio corresponding to 12:50 (lane 9). The number of depleted proteins progressively increased for incubations at higher volumes of plasma. This led to mixtures with varying complexity depending on the employed HIGH plasma-to-particle ratio. This strategy for depletion of proteins in plasma samples was further studied. Plasma depletion of some of the most abundant proteins was possible by incubating pNIPAm-AAc hydrogel particles with blood plasma samples at low and at high plasma-to-particle ratios (Figure 2). Both mechanisms yielded complementary

spiked into blood plasma samples, they were also irreversibly captured and depleted from the mixture, but surprisingly, they also accelerated the irreversible capture of abundant plasma proteins of the mixture (data not shown), indicating that this process did not affect proteins in an independent manner. Multiple Protocols for Plasma Depletion Based on Hydrogel Particles. The ability to control the interaction between proteins and hydrogel particles made it feasible to develop multiple protocols based on specific plasma-to-particle incubation ratios and recovering either the noncaptured or the captured proteins. The aim was to define a number of complementary protocols useful for sample fractionation at the protein level, prior to LC-MS/MS proteome analysis. Eight protocols were developed (protocols A to H, Table 1) and initially evaluated by SDS-PAGE (Figure 3). Seven Table 1. Eight Different Hydrogel Particles-Based Depletion Protocolsa elution P

ratio

wash

first step

second step

A B C D E F G H

2:10 3:10 40:10 3:10 3:10 40:10 1:50 1:50

H2O H2O H2O H2O H2O NH3 H2O H2O

1% TFA in H2O 1% TFA in H2O 1% TFA in H2O 5% TFA in ACN 5% TFA in ACN 1% TFA in H2O − (recover supernatant) 1% TFA in H2O

− − − H2O 0.7% NH3 − − −

a

Particles were mixed at a certain plasma-to-particle ratio for 15 min, followed by three washing steps and final elution for 15 min. A two step elution consisted of the addition of a second elution solvent after 15 min of incubation. Protocol H included spiking of ovalbumin in plasma. P: protocol; Ratio: plasma-to-particle ratio (v/v).

protocols were based on the mechanism of irreversible capture of proteins. Different high plasma-to-particle ratios afforded different degrees of depletion with an increasing number of irreversibly captured proteins at higher ratios (Figure 3, protocols A, B, and C at lanes 2, 3, and 4). Using the same high plasma-to-particle ratio, different optimized conditions resulted in distinctive protein profiles. For example, the application of a 2-step elution first with organic and then with aqueous solutions resulted in immunoglobulin-free depleted mixtures (comparing protocols D and E at lanes 5 and 6 with protocol B at lane 3). Likewise, including alkaline rinsing of hydrogel particles resulted in a concentration of a group of proteins characterized by low MW (comparing protocol F at lane 7 with protocol C at lane 4). Using a low plasma-to-particle ratio, the irreversible capture of plasma proteins was accelerated by spiking a single protein (ovalbumin) at a high concentration (protocol H at lane 8). A low plasma-to-particle ratio also enabled plasma depletion by recovering the noncaptured proteins (protocol G at lane 7). This protocol provided the most distinctive profile of all eight protocols, as later observed by large-scale MS-driven proteome analysis (discussed in the next section). Proteome Analysis of Hydrogel Particles-Depleted Human Plasma. SDS-PAGE analysis of proteins recovered by different hydrogel particle protocols demonstrated the utility for depletion of plasma samples. We further studied this by employing all eight protocols (Table 1) for proteome analysis by LC-MS/MS (see Experimental Section). We compared the

Figure 2. Schematics of protocol for human plasma depletion by hydrogel particles.

protein mixtures based on the protein distribution shown by SDS-PAGE (Figure 1A, lanes marked as LOW ratios and Figure 1B, lanes marked as HIGH ratios). At low ratios, the observed mechanism indicated a higher binding affinity for some proteins possibly depending on their ionic and hydrophobic properties. At high ratios, the progressive lower intensity and ensuing depletion of certain proteins suggested an irreversible capture of these proteins inside the hydrogel particles that only occurred above a certain plasma-to-particle ratio. This phenomenon was also observed for pure proteins such as bovine ovalbumin and human transferrin, which were irreversibly captured only when they were incubated at high concentrations (data not shown). When these proteins were 1546

dx.doi.org/10.1021/ac403749j | Anal. Chem. 2014, 86, 1543−1550

Analytical Chemistry

Article

detected, vitronectin was the 24th, and plasma kallikrein was the 54th in raw plasma. Likewise, five protocols (A, B, C, E and H) showed an enrichment of the family of immunoglobulins (i.e., considering all proteins marked as immunoglobulins), while the other three protocols showed less (D) or much less (F and G) abundance of immunoglobulins in the mixture. Four protocols (B, C, F, and G) displaced serum albumin from the top 10 list. The relative abundance of serum albumin decreased from 23.4% protein content in plasma to less than 0.4% in the protocols and to less than 6.5% by all eight protocols. Serotransferrin changed from the 4% in plasma to less than 0.04% in 4 protocols. Haptoglobin, present at 4.4% in plasma, was detected in only 2 protocols. Alpha-1-antitrypsin changed from 2.7% in plasma to less than 0.02% in 5 protocols. These differences were illustrated by representing the different concentrations of the top 10 most abundant proteins in raw plasma for each one of the eight hydrogel particles-depleted plasma protocols (Figure 5).

Figure 3. Protein profiles obtained by different hydrogel particlesbased depletion protocols. SDS-PAGE lane 1: plasma (0.5 μL). Lanes 2−9: protocols A to H, described in Table 1.

list of proteins identified for each protocol. The distribution of the top 10 most abundant proteins (Figure 4) showed differences in which proteins were depleted to a higher degree. For all protocols, most of the top 10 most abundant proteins were not present in the list of raw plasma (see the list next to each protocol in Figure 4). For example, plasminogen was the most abundant protein in protocol D and the fifth in protocol H; lysozyme C was the third in protocol F and the fifth in protocol C. Vitronectin was the ninth in protocol G, and plasma kallikrein was the ninth in protocol E. In comparison, plasminogen occupied position 37, lysozyme C was not

Figure 5. Relative abundances of the top 10 most abundant plasma proteins detected in each hydrogel particles-based depletion protocols. Vertical axis: logarithmic ratio of protein relative areas of depleted and nondepleted plasma. Horizontal axis: plasma and eight hydrogel particles-based depletion protocols.

The complementarity between protocols was further studied by determining how many proteins were exclusively detected in

Figure 4. Distribution of the top 10 most abundant proteins after hydrogel particles-depleted plasma samples. Distribution of the top 10 most abundant proteins in blood plasma, according to relative intensities for each hydrogel particles-based depletion protocols. 1547

dx.doi.org/10.1021/ac403749j | Anal. Chem. 2014, 86, 1543−1550

Analytical Chemistry

Article

kallikrein inhibitor ISK5 (low ng/mL range36) were selected on the basis of their reported roles as potential biomarkers.37−41 Comparison of the developed protocols (Table 1) identified protocol D as the most sensitive on average, based on previous large-scale proteome analysis data (Table 2) and on SRM

each case. We observed a partial overlapping of identifications, indicating that each protocol represented a distinctive alternative as a depletion method. This opened the possibility of combining all protocols to increase the total number of identifications, i.e., using the eight protocols as a sample fractionation strategy. The combination of the eight protocols led to the identification of 448 proteins (FDR 5%, ignoring immunoglobulins and keratins). This number represented a significant improvement when compared to the average 129 identified proteins detected by each individual protocol (see the complete list in Supplementary Table 1, Supporting Information). For the eight protocols used, the most orthogonal protocol was determined by representing the number of unique protein identifications for each case versus the rest of the protocols (Figure 6). Protocol G showed the highest number of unique proteins.

Table 2. Assessment of Depletion Protocols for Selected Proteinsa

IGF1 IGF2 IBP2 IBP3 IBP4 IBP5 IBP6 IBP7 ALS PLF4 ISK5 KLK6 KLK7

protocols where the protein was detected

coverage (%)

Mascot score

D, E, F A, D, E, F, H A, B, C, D, E A, B, C, D, E, F A, B, D, E A, B, C, D, E, F A, D, E, F D, E A, B, G A, B, C, D, E C, D, F n.d. n.d.

21.54 17.22 51.38 46.05 37.98 52.94 31.67 19.50 14.05 36.63 3.48

243 253 980 369 486 412 280 265 352 298 158

a

Table showing hydrogel particles-based depletion protocols for each different targeted protein. The best protocol is underlined, based on the highest Mascot score obtained (shown in the corresponding column). Coverage refers to the identified part of the protein sequence. n.d.: not detected.

analyses (data not shown). Proteins were detected on the basis of proteotypic tryptic peptides in plasma. Synthetic stable isotope labeled peptide homologues were added as internal standards for ion signal correction and unambiguous identification. Correlation of transitions and coelution of the endogenous and synthetic peptides confirmed the identity of the endogenous peptides (Supplementary Figure 1, Supporting Information). Coefficients of variation (CV) showed good repeatability of the hydrogel particles-based depletion method across 12 replicates, with values