Application of Metal-Coded Affinity Tags (MeCAT): Absolute Protein

The metal-coded affinity tagging (MeCAT) approach has been devised as a .... Total protein of E. coli was reduced with 0.05 mM TCEP, then 0.25 mM EDTA...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

Application of Metal-Coded Affinity Tags (MeCAT): Absolute Protein Quantification with Top-Down and Bottom-Up Workflows by MetalCoded Tagging U. Bergmann,†,‡ R. Ahrends,† B. Neumann,‡ C. Scheler,‡ and M. W. Linscheid*,† †

Department of Chemistry, Humboldt-Universitaet zu Berlin, Brook-Taylor-Strasse 2, 12489 Berlin, Germany Proteome Factory AG, Magnusstrasse 11, 12489 Berlin, Germany

‡

S Supporting Information *

ABSTRACT: As the quantification of peptides and proteins extends from comparative analyses to the determination of actual amounts, methodologies for absolute protein quantification are desirable. Metal-coded affinity tags (MeCAT) are chemical labels for peptides and proteins with a lanthanide-bearing chelator as a core. This modification of analytes with non-naturally occurring heteroelements adds the analytical possibilities of inductively coupled plasma mass spectrometry (ICPMS) to quantitative proteomics. We here present the absolute quantification of recombinantly expressed aprotinin out of its host cell protein background using two independent MeCAT methodologies. A bottom-up strategy employs labeling of primary amino groups on peptide level. Synthetic peptides with a MeCAT label which are externally quantified by flow injection analysis (FIA)-ICPMS serve as internal standard in nanoHPLC−ESI-MS/MS. In the top-down approach, protein is labeled on cysteine residues and separated by two-dimensional gel electrophoresis. Flow injection analysis of dissolved gel spots by ICPMS yields the individual protein amount via its lanthanide label content. The enzymatic determination of the fusion protein via its β-galactosidase activity found 8.3 and 9.8 ng/μg (nanogram fusion protein per microgram sample) for batches 1 and 2, respectively. Using MeCAT values of 4.0 and 5.4 ng/μg are obtained for top-down analysis, while 14.5 and 15.9 ng/μg were found in the bottom-up analysis.

L

determination of therapeutically reasonable doses, their pharmacokinetic behavior, and their administration to patients requires quantitative monitoring with robust and reproducible methods.4−6 As therapeutic proteins may undergo alterations upon storage and after administration which in turn give rise to different protein species, methodologies which resolve such protein species and therefore allow for their parallel and selective quantification are highly interesting. The quantitative comparison of different analytes, however, requires a common basis for calibration. Protein quantification using proteomic techniques usually requires labeling techniques and/or internal standards such as isotopically labeled peptides. In turn, common strategies for protein quantification cannot be performed untargeted, and calibration must always be done with custom-made standards.

ife sciences and biotechnology nowadays widely make use of comparative quantitative proteomics for the characterization of biological systems. Most studies are aimed at the discovery of differentially regulated proteins to elucidate pathways, to identify biomarkers, or to monitor expression of protein products. Differential quantification techniques are employed to address these questions, and the results of these methodologies are abundance ratios of peptides or proteins, respectively. They have been greatly fuelled by isotope tagging techniques in combination with mass spectrometric detection. As the current focus of quantification in proteomic analyses is shifting from only intersample comparison to determination of actual contents for many applications, the integration of analytical techniques dedicated for quantitative work is highly desirable.1 Protein and peptide pharmaceuticals are important drugs and therapeuticals for a variety of indications, e.g., peptide hormones like insulin, erythropoetin, or antibodies, which are used as vaccines.2,3 Their production from raw material or from recombinant fermentation, the stability upon storage, the © 2012 American Chemical Society

Received: January 18, 2012 Accepted: May 10, 2012 Published: May 31, 2012 5268

dx.doi.org/10.1021/ac203460b | Anal. Chem. 2012, 84, 5268−5275

Analytical Chemistry

Article

gene was conjugated with β-galactosidase, and the resulting fusion gene for aprotinin::β-galactosidase was transformed into an expression system with a heat-inducible promoter. Cells were grown in a lab scale fermenter first at 30 °C to allow growth up to the desired cell density, and then the temperature was raised to 40 °C for induction of recombinant target protein synthesis (Figure 2). Samples were collected just before and

Element mass spectrometry is an interesting technique for both assessment of standards and direct quantification of protein analytes due to its high analytical sensitivity and selectivity.7 It has been employed previously for the direct identification and quantification of protein analytes carrying naturally occurring intrinsic heteroatoms like phosphorus, sulfur, selenium, and metal ions.8−11 Chemical labeling of proteins with tags containing elements which are not naturally present in biological systems allows this orthogonal detection technique to be exploited for virtually all proteins.12 The metal-coded affinity tagging (MeCAT) approach has been devised as a chemical labeling strategy for proteomic analysis that imposes a mass offset between different peptides or proteins by coding them with lanthanide ions which are captured in a macrocyclic complex based on DOTA (1,4,7,10tetraazacyclododecane-N,N′,N″,N‴-tetraacetic acid). Complete derivatization of cysteine thiol groups using maleimide chemistry with these tags allows for sensitive and accurate quantification of peptides and proteins by inductively coupled plasma mass spectrometry (ICPMS) using an external calibration.13 The MeCAT strategy is fully compatible with contemporary proteomic workflows, and as a key feature the direct quantification of metal-tagged proteins and peptides by inductively coupled plasma mass spectrometry (ICPMS) is made accessible for proteomics thanks to very low limits of detection for the rare earth elements.14 The quantification with ICPMS also yields actual sample amounts rather than only abundance ratios like comparative analysis schemes do. Here we demonstrate the MeCAT technology applied to the targeted quantification of a selected protein within a complex sample using two independent strategies. A bottom up strategy applies LC−MS-based proteomics: proteolysis of total protein, labeling and coding of the complex mixture, subsequent multiplex analysis of the samples, and externally quantifying differentially labeled test peptides. This workflow uses the amine-reactive labeling of proteolytic peptides with DOTA on their N-termini and lysine side chains, the coding with a lanthanide, multiplexing with other samples or internal standards, and subsequent LC−MS-based proteomics analyses. Other than cysteine-reactive chemistry, all peptides are derivatized, which results in theoretical coverage of the whole accessible proteome. As true HPLC separation is required for direct quantification analytes, but not feasible for peptides in complex mixtures, the quantification of element-tagged species on the peptide level has been limited to samples of fairly low complexity until now.15 We therefore propose the use of synthetic peptides of the target protein which are metal-labeled and externally assessed with ICPMS to serve as spiked internal standards in nanoHPLC−ESI-MS/MS. The second workflow uses the following top-down approach (2-DE/FIA-ICPMS): intact MeCAT labeled proteins in gel spots from two-dimensional electrophoresis are excised and subsequently dissolved for infusion to ICPMS as shown previously.13 As a biotechnological model system, which represents common production methods of recombinant proteins for therapeutic or diagnostic purposes, we investigated a high cell density cultivation of Escherichia coli designed for the overexpression of a fusion protein. The polypeptide aprotinin is a clotting actuator which has been employed in surgery for minimization of blood loss.16 It was originally produced from bovine lung tissue, but there have been attempts to move production to recombinant technology. For fermentation, its

Figure 1. Schematic comparison of the different workflows presented in this study: (A) absolute quantification by spiking an ICPMSquantified MeCAT labeled standard peptide into the labeled MeCAT sample, (B) absolute quantification of MeCAT labeled proteins; proteins are labeled on the protein level and quantified using an external metal salt standard, (C) enzymatic assay to determine protein amount via its activity. The ◀ indicates when the standard or sample is quantified by ICPMS using an external metal salt standard.

after heat induction during the course of product formation to compare the influence of heat shock conditions and product formation to the proteome of the expression host. This study aimed for the absolute quantification of the aprotinin::β-galactosidase fusion protein at different time points of the fermentation process by means of chemical labeling of the total proteome with metal tags and subsequent analysis by common proteomic workflows combined with ICPMS.

■

EXPERIMENTAL SECTION Cultivation of bacteria, protein extraction, and proteolysis is described in the Supporting Information (Supplementary Text 3). Labeling of tryptic peptides was carried out with a 100-fold molar excess of MeCAT-NHS (assuming an amino group content of 825 g/mol). It was critical to maintain the peptide concentration above 1 μg/μL, to have 50% acetonitrile in the reaction mixture, and to stabilize the pH at 8.5. Accordingly, peptides were buffered with 100 mM TEAB and the reagent was redissolved in 75% ACN before addition to the peptides. The pH was adjusted by the further addition of 100 mM triethylamine in acetonitrile immediately after the addition of MeCAT-NHS to ensure the reaction proceeds quantitatively. After 1 h, the solution was dried, redissolved in 100 mM triethylammonium acetate, and labeled with a 1.25-fold molar excess of the respective lanthanide chloride (dissolved to 100 mM in water). The mixture was then lyophilized, redissolved in 0.5% formic acid, and purified by solid phase extraction (SPE) on C18 reverse phase and strong cation exchange material using 5269

dx.doi.org/10.1021/ac203460b | Anal. Chem. 2012, 84, 5268−5275

Analytical Chemistry

Article

Figure 2. Time course of parameters during fermentation of recombinant Escherichia coli designed for the overexpression of the clotting factor aprotinin.

For flow injection analysis, gel spots were dissolved as described in the D-SDS-PAGE manual (Proteome Factory, Berlin, Germany). These liquid samples were analyzed using a setup consisting of an HPLC pump (model K-1001, Knauer, Berlin, Germany) operating at 200 μL/min and a Famos autosampler (Dionex, Idstein, Germany) interfaced with an Element XR mass spectrometer via an HPLC nebulizer (gas flow 0.87 L/min) and a desolvation chamber (Elemental Scientific, Omaha) as described earlier.13 Ion traces for relevant elements were recorded in the low-resolution mode, and peak areas of the resulting signals were calculated for further analysis.

UltraMicroSpin Tips (Nest Group, Southborough, MA) successively. Synthetic peptides were obtained from Schafer-N (Copenhagen, Denmark) in >97% purity and used as supplied. Stock solutions (5 mg/mL) were prepared by reconstitution in 20% ACN supplemented with 5 mM TCEP where appropriate. Cysteine-containing peptides were carbamidomethylated prior to further use. MeCAT labeling and coding was performed with 1 mM peptide in the reaction mixture as described above, followed by purification and desalting. Sample Preparation for Top-Down Analysis. Total protein of E. coli was reduced with 0.05 mM TCEP, then 0.25 mM EDTA, 2.5 mM sodium acetate, and 0.36 mM MeCATDBM reagent were added. The mixture was supplemented with 10−20% acetone and buffered to pH 7. The reaction was allowed to proceed at room temperature with shaking for 12 h. Residual reagent was then quenched with DTT. Two-dimensional gel electrophoresis was carried out according to Klose and Kobalz with minor modifications.17 In the second dimension, a dissolvable gel matrix was used (DSDS-PAGE, Proteome Factory, Berlin, Germany). For silver staining, the FireSilver staining kit (Proteome Factory, Berlin, Germany) was used. Separation and Detection. Analysis of peptides was carried out with nanoHPLC−ESI-MS/MS as described earlier.14 Element-selective mass spectrometry was performed with inductively coupled plasma mass spectrometry using an Element XR mass spectrometer (ThermoFisher Scientific, Bremen, Germany), which was operated in low-resolution mode. For the quantification of metal-tagged peptide standards, a MicroMist nebulizer (Glass Expansion, West Melbourne, Australia) in the self-aspirating mode was used for continuous sample infusion. Purified MeCAT-labeled peptides were diluted by a factor of 10 000 with 2% nitric acid and analyzed to determine the absolute peptide concentration. Ion intensities were recorded for 1 min and averaged for further calculations. External calibration was performed with a thulium standard in a concentration range from 10 to 500 000 ppt.

■

RESULTS AND DISCUSSION The aim of our study was to demonstrate that the MeCAT technology delivers accurate absolute quantification results on the peptide or protein level independent of the applied workflow (Figure 1). Here we focused on the determination of aprotinin to demonstrate the suitability of both methods for quantitative assessment of active biopharmaceutical ingredients (bio APIs). In a bottom-up strategy, samples are hydrolyzed with trypsin and MeCAT-labeled. A target peptide is synthesized, metal-labeled, externally quantified by ICPMS, and mixed with the sample for use as an internal standard in a duplex assay using electrospray MS. The top-down approach is comprised of metal-tagging of proteins and quantification of 2DE-separated proteins by FIA-ICPMS after dissolution of gel matrix. The well-known enzymatic determination of βgalactosidase activity served as a reference method to estimate the accuracy of MeCAT techniques. Establishing of a Peptide Level MeCAT Strategy for the Quantification of Aprotinin. For the absolute quantification using MeCAT on the peptide level, we established a labeling strategy which targets primary amino groups. The general suitability of this method is demonstrated in Supplementary Text 1 in the Supporting Information: collision induced-dissociation yields spectra suitable for database search (Figure S-1 in the Supporting Information), complete labeling and coding of analytes is shown (Figure S2 in the Supporting Information), differentially labeled peptides 5270

dx.doi.org/10.1021/ac203460b | Anal. Chem. 2012, 84, 5268−5275

Analytical Chemistry

Article

Figure 3. Overlay of extracted ion chromatograms of expected m/z ranges and corresponding mass spectra for four tryptic, differentially MeCAT labeled peptides of the fusion protein from a duplex analysis are displayed. Samples were drawn prior and after temperature induction, digested, differentially labeled, and combined for analysis. Panel A shows respective traces of the sample before the temperature shift (MeCAT-Lu) while panel B shows corresponding data obtained from sample drawn after heat induction (MeCAT-Ho). Relative abundance has been normalized to the signal intensity of the Ho-labeled peptide for each corresponding peptide pair. Panels C−F show the mass spectra in the expected m/z range for the respective peptide pairs. For the LC−MS analysis, 4 μg of total E. coli cell lysate were injected.

Table 1. Chemical and Physical Properties of the Identified Peptides That Were Chosen for Quantification Are Listeda

a

sequence

length

m [Da]

M + MeCAT [Da]

m/z, charge

RT [min]

no. tags

FNDDFSR AGLCQTFVYGGCR LAAHPPFASWR SVDPSRPVQYEGGGADTTATDIICPMYAR

7 13 11 29

899.37 1487.66 1251.65 3126.44

1451.47 2039.76 1803.74 3678.53

726.74 (2+) 1020.89 (2+) 902.88 (2+) 1227.18 (3+)

95.5 246.0 246.9 337.1

1 1 1 1

Peptides were custom-synthesised and subjected to MeCAT-labelling for utilization as internal standards.

check for their absence in the lutetium-labeled sample and their presence in the holmium-labeled sample (Figure 3). The analysis of the chromatographic data shows that the fusion protein is not present before the temperature shift due to absence of Lu-labeled peptides, and after 6 h of overexpression its respective peaks can be clearly identified via the Ho-labeled signals (Figure 3). This data demonstrates that the thermoinducible aprotinin-fusion protein is translated after temperature induction, but relative quantification is impossible since the protein was not detected before temperature induction because it has no basal expression at normal growth temperatures, and therefore it is not possible to attribute a fold-change for the onset of expression of aprotinin. To achieve quantitative information, an absolute quantification approach must be established. Absolute Quantification on Peptide Level. The four most abundant peptides identified from the fusion protein were chosen as standard peptides (Table 1). Labeled synthetic peptides were then absolutely quantified by ICPMS (Figure S11 in the Supporting Information). Our internal standard strategy relies on target peptides that are sourced as synthetic material for differential metal tagging. To accurately determine the concentration of the metal-coded internal standard peptides, the standards were purified from

coelute in RP-HPLC (Figure S-3 in the Supporting Information), and duplex analysis is feasible (Figure S-4 in the Supporting Information). The linearity of the ESI response was shown in the relevant concentration range (Figure S-5 in the Supporting Information). The quantitative assessment of the target protein has further requirements peculiar to aprotinin that need to be met: the protein of interest must be detectable via proteolytic peptides with appropriate elution profiles which are detectable without spectral interferences. To check these constraints, samples were drawn at different time points from aprotinin fermentation, MeCAT-labeled, and analyzed by LC−ESI-MS/MS. In addition, it had to be shown that no aprotinin was present before heat induction. Therefore, two samples were cleaved with trypsin, and completeness of proteolysis was checked by SDS-PAGE and silver staining. Proteolytic peptides were then differentially MeCAT labeled: the sample drawn before the temperature shift was lutetium labeled, while the sample taken 6 h after the temperature shift was holmium labeled. Samples were combined for duplex analysis with nanoHPLC−ESI-MS/ MS, and resulting data was submitted to database searches to obtain peptides from the fusion protein. Extracted ion chromatograms of selected identified peptides were created to 5271

dx.doi.org/10.1021/ac203460b | Anal. Chem. 2012, 84, 5268−5275

Analytical Chemistry

Article

Figure 4. Analysis of differentially labeled peptides from the E. coli fermentation by nanoHPLC−ESI-MS/MS. Sample from batch 2 collected 6 h after the temperature shift (Ho) and standard peptides (Tm) were differentially labeled and pooled for analysis. Extracted ion chromatograms (EIC) of four selected differentially metal-tagged peptides (Ho, black traces; Tm, blue traces) from the fusion protein are displayed. NL indicates the normalization level of each peptide pair with the peak area ratio listed below. A red chromatogram indicates the total ion chromatogram (TIC) of the analysis.

excess reagent and lanthanide salt by successive solid phase extraction on C18 reverse phase and strong cation exchange material. Subsequently we performed FIA/ICP-MS with dilutions of the individual peptide solutions (Figure S-11 in the Supporting Information). Elemental mass spectrometry is employed here for the quantitative assessment of the peptide standards using an external calibration with a well-defined inorganic lanthanide chloride solution. Despite its analytical power in terms of linear detection range and high sensitivity, the design of the presented workflow rendered it necessary to employ just ICPMS for simple infusion analyses of the synthetic peptide standards to establish their concentration. Using this information, it was possible to recalculate the concentration of each peptide in a mixture of nominally equimolar synthetic peptides, so an absolute quantification of endogenous peptides that had been analyzed by nanoHPLC− ESI-MS/MS was feasible. For the absolute quantification of aprotinin within the samples of the E. coli model system, we chose three individual samples, each collected at different physiological states of the cells from two cultivation batches. Two samples were drawn from each fermentation batch: one sample just before the heat induction (t = 0) and another sample 6 h after the temperature shift for heat induction, respectively. For each sample, three peptides were selected for the quantification. Tryptic peptides were differentially labeled with Lu (batch 1, t = 0) or Ho (batch 1, after 6 h) while the internal standards were Tm-labeled. The FIA-ICPMS quantified internal standards were then spiked into the labeled sample, and the mixture was analyzed with nanoHPLC−ESI-MS/MS. Despite the general rule of adding internal standard as early as possible, here internal peptide standards were added after proteolysis, as favored by Kuzyk et

al. where their addition prior to tryptic hydrolysis resulted in an overestimation of the actual sample peptide concentrations.18 The obvious loss of internal standard has been attributed to kinetic effects that do not apply to the sample peptides that emerge during the course of incubation with the protease. For the manual evaluation of ion traces, peak areas of the selected peptides were determined by their extracted ion chromatograms with an isolation width of ±10 ppm (Figure 4). The peak area ratios of differentially labeled peptides and the known concentrations of the internal standards supplied the absolute concentration of sample peptide in the analysis. As an estimation of the method precision, we considered three or four peptides to determine the concentration of the fusion protein. As with all techniques that determine information about protein concentration from derived peptides, the analytical precision of the observed analytes determines the coefficient of variation of the overall protein. With the shift of quantitative analyses to MS-based strategies on the peptide level, especially with many published large scale experiments using isobaric tagging, this aspect moved into the focus of our research.19 We used four internal standard peptides for the quantification of the fusion protein (Table 2). The analytical precision of the method can thus be probed by comparison of the individual amounts of peptide determined. Frequently, targeted approaches rely on only one or two internal standards per analyte protein which compromises analytical precision. We were able to determine the absolute amount of the fusion protein in total protein with 14.5 ± 3.5 for batch 1 and 15.9 ± 7.2 ng/μg for batch 2. Next we asked if we would obtain the same results by applying the MeCAT technology on the protein level. 5272

dx.doi.org/10.1021/ac203460b | Anal. Chem. 2012, 84, 5268−5275

Analytical Chemistry

Article

demonstrated previously. In an earlier study, proteins labeled with thulium and lutetium were absolutely quantified. The labeling with different element tags on proteins was tested (La to Lu). Because of the ionic radii which affect the migration on SDS-PAGE (Figure S-7 in the Supporting Information), the quantification approach on the protein level is limited to six labels (Tm, 171Yb, 172Yb, 174Yb, 176Yb, and Lu; Figure S-8 in the Supporting Information). After the staining and 2-DE separation protocol optimization, the absolute quantification of the fermentation product within the complex E. coli host cell protein background was approached. A sample was drawn 2 h after heat induction (Figure 2) and divided in two aliquots of 25 μg of protein and differentially labeled with MeCAT tags carrying isotopically pure 172Yb and 176Yb ions, respectively. After separation by 2-DE and silver staining, the identity of the protein spots was confirmed by nanoHPLC−ESI-MS/MS and a database search. Spots identified as aprotinin::β-galactosidase were excised from a duplicate gel, dissolved, and analyzed by FIA-ICPMS in triplicate (Figure 5). Using an external calibration with 174Yb, a total content of 5.4 ± 0.2 ng/μg (for 172 Yb) and 4.0 ± 0.1 ng/μg (for 176Yb) of fusion protein in total protein was determined with a technical standard deviation of ±3.4%. The sample-to-sample difference was 21%. Absolute Quantification by an Enzymatic Assay. For comparison with an established β-galactosidase quantification method, the fusion protein amount was also determined in an enzymatic assay by measurement of the β-galactosidase activity over the time of induction (Figure S-12 in the Supporting Information). A fusion protein content of 8.3 ng/μg (batch no. 1) and 9.8 ng/μg (batch no. 2) in total protein was calculated.22 The results of the three presented methods and samples are summarized in Table 3. All values are in the same analytical range, which underscores the analytical robustness and reproducibility of metal tag based absolute protein quantification. The general suitability of MeCAT for relative quantification out of complex mixtures has been shown in both top-down and bottom-up style experiments. Two different analysis workflows, one targeted at cysteine residues and the other aimed at peptide N-termini and lysine side chains have been successfully established for this. Absolute quantification of the target protein was demonstrated, and the results lie within the same order of magnitude as the known enzymatic assay that served as the reference method. This is a very good agreement given the

Table 2. Protein Concentrations As Determined from the Four Different Peptides of the Fusion Protein in Two Different Biological Samplesa content of fusion protein in total protein as calculated from peptide amount [ng/μg]

batch no. 1

batch no. 2

AGLCQTFVYGGCR SVDPSRPVQYEGGGADTTATDIICPMYAR FNDDFSR LAAHPPFASWR

18.4 12.0 11.1 16.6

5.2 18.7 18.5 21.2

a

The protein concentration is calculated from these values with the table deviation of the peptides, giving the analytical precision of the assay.

Absolute Quantification with the Top-Down Methodology. For the absolute quantification of aprotinin on the protein level, we applied MeCAT-maleimide chemistry targeting the thiol groups as presented earlier14 to a sample from batch 2. To quantify the aprotinin-β-Galactosidase fusion protein in an absolute fashion, an optimization of the multiplexing ability (Figures S-7 and S-8 in the Supporting Information), staining (Figure S-10 in the Supporting Information), and separation (2-DE) of labeled complex protein mixtures (Figures S-6 and S-9 in the Supporting Information) was performed (Supplementary Text 2 in the Supporting Information). These optimizations allowed us to subsequently label, separate, and detect the fusion protein with the best 2-DE resolution and sensitivity. Separation and detection of intact protein analytes offers a significant advantage that bottom-up approaches cannot: information that would be lost on the peptide level is conserved. In addition, sample complexity does not increase by orders of magnitude during sample processing. Assuming an average protein mass in a proteome of 40 kDa with 5% lysine and arginine residues, every protein might yield about 30 peptides each.20 Two-dimensional gel electrophoresis is still undoubtedly the benchmark technique for the separation of complex protein samples in a top-down fashion, due to its unparalleled capability of resolving up to thousands of proteins and/or protein species.21 For the parallel analysis of different samples on a single gel, differentially labeled samples must be combined for a multiplex analysis. Accurate quantification requires a protein to migrate in electrophoresis irrespective of the incorporated metal ion. The general applicability of MeCAT with 2-DE has been

Figure 5. Fusion protein spots after 2-DE separation of the E. coli proteome from batch 2. Arrows show position of aprotinin::β-galactosidase fusion protein. Panels A and C show the gel region from a sample before heat induction without and upon MeCAT labeling, respectively. In panels B and D, sample drawn 2 h after the temperature shift are shown with the fusion protein present, in panel B without and in panel D after MeCAT labeling. The calibration curve with inorganic standard is depicted in panel E. Panel F gives the determined ion intensities (in cps) of the dissolved spot via FIA-ICPMS. 5273

dx.doi.org/10.1021/ac203460b | Anal. Chem. 2012, 84, 5268−5275

Analytical Chemistry

Article

Table 3. Concentrations and Precision of the Fusion Protein As Determined with Three Techniques in Two Different Samples

content of aprotinin [ng/μg total protein]

batch no. 1, peptide level

batch no. 1, enzymatic assay

batch no. 2, peptide level

batch no. 2, 2DE/FIAICPMS (172Yb)

batch no. 2, 2DE/FIAICPMS (176Yb)

batch no. 2, enzymatic assay

14.5 ± 3.5

8.3

15.9 ± 7.2

5.4 ± 0.2

4.0 ± 0.1

9.8

biological variation that has to be considered in protein quantification and the fact that indeed no default value is known. To assess method accuracy delivered by the two MeCAT methods, the results have to be compared to the results of the enzymatic assay which serves as a reference method which is assumed to be accurate. The peptide-based MeCAT method yield 174% and 198% of these values, while with the proteinbased workflow 55% and 40% of the reference’s results are obtained. In all methodologies, sample may be lost for analysis: successive clean up steps on peptide level and transfer from first to second gel dimension for protein level may result in sample loss, while protein denaturation during cell lysis may occur before the enzymatic assay. The extent of sample loss in the peptide workflow may be probed by adding the internal standard immediately after proteolysis to ensure it experiences the same procedures as sample peptides. The recovery of metaltagged proteins during 2-DE can be analyzed by running a selection of HPLC-purified, ICPMS-quantified proteins on 2DE, dissolving their gel spots and analyzing their metal content again by ICPMS. Both MeCAT strategies yield results that fall into the same analytical range as those obtained by the reference, i.e., they differ by a factor of about 2. Given that no real default values are known and as there is no reference material available, this is a reasonable result in terms of accuracy. Method precision for the peptide workflow is given via the different quantities found for different target peptides. These values vary with 24% (batch 1) and 45% (batch 2) around each other. The reason for this may be different recovery rates that peptides exhibit during analysis workflow due to different physical and chemical properties. For the protein level strategy, both analyses have been performed using the same sample which allows giving an estimate about method precision. The standard deviation for the ICPMS analysis is ±3.4% while the standard deviation of the samples for the whole workflow is 21%. This data is for both methods in good agreement with data obtained from the enzymatic assay which is assumed to be correct. In the bottom-up approach, which covers all available proteolytic peptides, metal-coded standard peptides are employed to cope with high sample complexity. For a direct absolute quantification without a need for custom-made standards, the thiol-reactive MeCAT reagent was used in combination with two-dimensional electrophoresis and FIAICPMS of selected spots, allowing for the first time the direct assessment of individual protein amounts from a complex sample. This straightforward methodology is only limited to proteins containing at least one cysteine residue which still covers about 80% of all proteins in the E. coli proteome.23 Both methodologies are capable of absolute protein quantification by calibration with simple inorganic standards. In addition to the absolute quantification of individual analytes, this allows for the comparison of different analytes, too. The top-down approach even allows discerning different protein species which arise upon storage or due to metabolic processing and which may significantly differ in their activity or even toxicity. This is not possible with peptide-centric strategies.24

The absolute quantification of proteins is an important leap for the pharmaceutical industry as it gives the actual amount of an active bio API. Enzymatic assays yield indirectly measured activities rather than actual amounts, and immunological assays rely on antibodies, which are difficult to assess for specificity and activity. Metal coded tagging may therefore open an alternative approach to protein quantification when antibodybased approaches reach their limits, either due to unwanted cross-reactivity when the resolution of very similar protein species is demanded or when limited numbers of antibodies are available for insect cells. Peptide quantification methods using isotopically labeled peptides must rely on the stability of the supplied standards, which degrade over time. Additionally, peptide quantification by organic mass spectrometry misses the comparability of different analytes due to different ionization behavior. This lack of interanalyte comparability and traceability to well-defined standards can be overcome by metalcoded tagging of analytes, which employs simple inorganic standards for calibration.

■

CONCLUSIONS

We have developed and applied two independent techniques for the absolute quantification of proteolytic peptides and intact proteins from a complex biological system. The bottom-up strategy is based on the differential metal-coded tagging of sample peptides and their respective internal standards. The top-down approach employs the metal-tagging of the proteome, followed by common two-dimensional electrophoresis and flow injection ICPMS. Thus, it is even possible to discern protein species that may arise from protein modifications due to pharmacokinetic behavior or decay upon storage. These effects are easily overlooked by immunological techniques but move into the focus of regulatory measures and therefore will be necessary to assess. Metal-coding offers the possibility to directly compare different peptides, proteins, and protein species and therefore may become a valuable tool to address such questions. The approaches presented in this contribution employ the robust and powerful quantification tool ICPMS with simple infusion or flow injection analyses to limit further technical complexity of the analysis workflow. Element mass spectrometry therefore serves as an easy extension to proteomic workflows. It allows straightforward workflows yielding actual sample amounts instead of abundance ratios. Looking further, the total percentage of a selected protein within a proteome may be accessed. This additional information can be obtained with HPLC−ICPMS even if samples are too complex for baseline separation of species. By using the total peak area covered by sample peptides and knowledge of the statistic composition of the investigated proteome, the quota of the target proteins within the whole proteome can be assigned. This information is usually obtained by using different methods for the total protein amount and the target protein as it is difficult to access with a single method. The total percentage of, e.g., host cell proteins within a pharmacological product may be determined in this fashion. 5274

dx.doi.org/10.1021/ac203460b | Anal. Chem. 2012, 84, 5268−5275

Analytical Chemistry

■

Article

(24) Schluter, H.; Apweiler, R.; Holzhutter, H. G.; Jungblut, P. R. Chem. Cent. J. 2009, 3, 11.

ASSOCIATED CONTENT

S Supporting Information *

Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Address: Laboratory of Analytical and Environmental Chemistry, Department of Chemistry, Humboldt-Universitaet zu Berlin, Brook-Taylor-Str. 2, 12489 Berlin, Germany. Phone: +493020937575. Fax: +493020936985. E-mail: m.linscheid@ chemie.hu-berlin.de. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Prof. Dr. M. Popovic, Beuth-Hochschule für Technik, Berlin, for the generous donation of the E. coli samples. We thank Kyle Kovary for careful reading of the manuscript during his spare time. U.B. and R.A. contributed equally to this work.

■

REFERENCES

(1) Wang, M.; Feng, W. Y.; Zhao, Y. L.; Chai, Z. F. Mass Spectrom. Rev. 2010, 29, 326−348. (2) Xuereb, F.; Chaignepain, S.; Breilh, D.; Godde, F.; Saux, M. C.; Lenz, C.; Glueckmann, M.; Schmitter, J. M. Anal. Bioanal. Chem. 2011, 400, 2073−2084. (3) Rappel, C.; Schaumloffel, D. Anal. Chem. 2009, 81, 385−393. (4) Ezan, E.; Dubois, M.; Becher, F. Analyst 2009, 134, 825−834. (5) Chao, T. C.; Hansmeier, N.; Halden, R. U. J. Proteomics 2010, 73, 1641−1646. (6) Ohtsuki, S.; Uchida, Y.; Kubo, Y.; Terasaki, T. J. Pharm. Sci. 2011, 100, 3547−3559. (7) Sanz-Medel, A. Anal. Bioanal. Chem. 2005, 381, 1−2. (8) Wind, M.; Feldmann, I.; Jakubowski, N.; Lehmann, W. D. Electrophoresis 2003, 24, 1276−1280. (9) Grebe, M.; Profrock, D.; Kakuschke, A.; Broekaert, J. A.; Prange, A. Metallomics 2011, 3, 176−185. (10) Rappel, C.; Schaumloffel, D. Anal. Bioanal. Chem. 2008, 390, 605−615. (11) Giusti, P.; Schaumloffel, D.; Encinar, J. R.; Szpunar, J. J. Anal. At. Spectrom. 2005, 20, 1101−1107. (12) Prange, A.; Proefrock, D. J. Anal. At. Spectrom. 2008, 23, 432− 459. (13) Ahrends, R.; Pieper, S.; Kuhn, A.; Weisshoff, H.; Hamester, M.; Lindemann, T.; Scheler, C.; Lehmann, K.; Taubner, K.; Linscheid, M. W. Mol. Cell. Proteomics 2007, 6, 1907−1916. (14) Ahrends, R.; Pieper, S.; Neumann, B.; Scheler, C.; Linscheid, M. W. Anal. Chem. 2009, 81, 2176−2184. (15) Esteban-Fernandez, D.; Scheler, C.; Linscheid, M. W. Anal. Bioanal. Chem. 2011, 401, 657−666. (16) Sivasubramaniam, S.; Murali, B.; Smallpiece, C. J.; Smith, B.; Kalkat, M. Ann. Card. Anaesth. 2009, 12, 84−86. (17) Klose, J.; Kobalz, U. Electrophoresis 1995, 16, 1034−1059. (18) Kuzyk, M. A.; Smith, D.; Yang, J. C.; Cross, T. J.; Jackson, A. M.; Hardie, D. B.; Anderson, N. L.; Borchers, C. H. Mol. Cell. Proteomics 2009, 8, 1860−1877. (19) Karp, N. A.; Huber, W.; Sadowski, P. G.; Charles, P. D.; Hester, S. V.; Lilley, K. S. Mol. Cell. Proteomics 2010, 9, 1885−1897. (20) Brownridge, P.; Beynon, R. J. Methods 2011, 54, 351−360. (21) Jungblut, P. R.; Holzhutter, H. G.; Apweiler, R.; Schluter, H. Chem. Cent. J. 2008, 2, 16. (22) Putnam, S. L.; Koch, A. L. Anal. Biochem. 1975, 63, 350−360. (23) Patton, W. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2002, 771, 3−31. 5275

dx.doi.org/10.1021/ac203460b | Anal. Chem. 2012, 84, 5268−5275