Nano-High-Performance Liquid Chromatography with Online

Feb 1, 2013 - acetic acid (DOTA).14−16 DOTA is available as a bifunctional. Received: ..... Ph.D. Thesis, Humboldt University, Berlin,. 2009. Analyt...
1 downloads 0 Views 1MB Size
Technical Note pubs.acs.org/ac

Nano-High-Performance Liquid Chromatography with Online Precleaning Coupled to Inductively Coupled Plasma Mass Spectrometry for the Analysis of Lanthanide-Labeled Peptides in Tryptic Protein Digests Angela Holste,†,‡ Andreas Tholey,‡ Chien-Wen Hung,‡ and Dirk Schaumlöffel*,† †

Université de Pau et des Pays de l’Adour/CNRS UMR 5254, Laboratoire de Chimie Analytique Bio-Inorganique et Environnement/IPREM, 64053 Pau, France ‡ Institute for Experimental Medicine−Div. Systematic Proteome Research, Christian-Albrechts-Universität, 24105 Kiel, Germany S Supporting Information *

ABSTRACT: Low background signals are an indispensable prerequisite for accurate quantification in bioanalytics. This poses a special challenge when using derivatized samples, where excess reagent concentrations are increasing the background signal. Precleaning steps often are time-consuming and usually lead to analyte losses. In this study, a set of labeled model peptides and a protein digest was analyzed using inductively coupled plasma mass spectrometry (ICPMS), coupled to nano ion pairing reversed-phase high-performance liquid chromatography (nano-IP-RP-HPLC). In addition, matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS) was used for peptide identification. Peptides were labeled with lanthanide metals using bifunctional DOTA-based (1,4,7,10-tetraazacyclododecane-1,4,7,10tetraacetic acid) reagents. The resulting metal excess was removed online during nano-HPLC, by trapping the labeled peptides on a C18-precolumn and washing them prior to their elution to the analytical column. Different ion pairing reagents like TFA (trifluoroacetic acid) and HFBA (heptafluorobutyric acid) were used in the study to enhance interactions of the different peptide species with the C18 material of the precolumn. HFBA even allowed the detection of a highly hydrophilic peptide that was not retained using TFA. It was shown that for the mixture of labeled model peptides, even a short 3 min washing step already enhanced the removal of the excess reagents significantly, whereas peptide losses were observable starting with a 10 min washing time. A 6 min washing time was determined to be the best parameter for lowering the lanthanide metal background while maintaining maximum peptide recovery. Alternative precleaning setups using EDTA to enhance the removal of free metal or an offline approach using solid phase extraction did not show promising results. The application of the optimized method to labeled peptides in a lysozyme digest showed results comparable to those obtained with model peptides.

M

impossible to quantify less-abundant proteins due to the low ionization efficiency of the regarded elements, polyatomic interferences, and high background signals. This can be overcome by the introduction of an element into the analyte which is easier to be ionized, such as metals.10,11 Especially lanthanides, which need very low ionization energies, and thus are ionized quantitatively in ICPMS, have virtually no background in biological samples. One way to connect those rare earth metals to a protein is to conduct a derivatization, using a compound that adds a chelating function to the analyte. The metal-loaded complex afterward functions as a label for quantification in ICPMS. Examples for labeling reagents used with lanthanides are diethylenetriamine pentaacetic acid (DTPA)12,13 and 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA).14−16 DOTA is available as a bifunctional

olecular mass spectrometry (MS), mainly applying soft ionization methods, such as matrix-assisted laser desorption ionization (MALDI) and electrospray ionization (ESI), are the most commonly used methods for the identification of peptides and proteins; further, they play a major role in quantitative proteomics, ever since the introduction of relative protein quantification with ICAT1 and the many other approaches such as SILAC and iTRAQ that have since followed.2−4 Inductively coupled plasma mass spectrometry (ICPMS) has been shown to be a powerful tool for absolute quantification of elements in environmental and life sciences.5 Unfortunately it is not fit as a stand-alone technique in the field of quantitative proteomics, since molecular information is lost during the ionization process. In combination with molecular MS,6 ICPMS can be applied (e.g., for absolute protein quantification using heteroelements like S, P and Se that are naturally occurring in proteins).7−9 The sensitivity for most of these heteroelements in ICPMS is limited though, making it © 2013 American Chemical Society

Received: December 20, 2012 Accepted: February 1, 2013 Published: February 1, 2013 3064

dx.doi.org/10.1021/ac303618v | Anal. Chem. 2013, 85, 3064−3070

Analytical Chemistry

Technical Note

lanthanide(III) chloride hexahydrate salts (holmium, thulium, lutetium, erbium), praseodymium(III) nitrate hexahydrate, lysozyme from chicken egg white, Millipore ZipTip C18, Glu1-fibrinopeptide B and α-cyano-4-hydroxycinnamic acid (CHCA) were purchased from Sigma-Aldrich (Taufkirchen, Germany). Sequencing grade modified trypsin was obtained from Promega (Madison, WI). Deionized water (18.2 MΩ cm) was prepared with a Millipore Milli-Q water purification system Advantage A10 (Merck, Molsheim, France). Sample Preparation. The derivatization of peptides and protein digests with two different DOTA compounds, targeting either the amino- or the sulfhydryl functions present in the molecules, was performed according to Gregorius et al.15 The products were characterized by MALDI MS/MS. Protein Digestion. Five hundred micrograms of lysozyme was reduced with 1.25 μmol TCEP in 50 mM HEPES (pH 7.8) at 60 °C for 1 h. Digestion with 10 μg trypsin took place overnight at 37 °C. Peptide Labeling. Reduction of all peptides was conducted at 60 °C for 1 h, using a 2-fold molar excess of TCEP per cysteine residue. Derivatization of the peptides as well as the digest with Mal-DOTA was performed using a 20-fold molar excess of reagent regarding cysteine residues, 2 h at 37 °C in 50 mM HEPES (pH 7). NHS-DOTA was solved to a 500 mM solution in water-free DMSO. Free cysteine residues were alkylated with a 6-fold molar excess of MMTS regarding SH groups, reaction of free amino groups was carried out using a 100-fold molar excess of NHS-DOTA for 1 h at room temperature in 75% ACN + 25% 100 mM HEPES (pH 8). For the digest, an approximation was done for the free amino groups based on a tryptic in silico digest, excluding missed cleavages.24 The lanthanide salts were dissolved in 100 mM TEAA buffer (pH 5) to 1 M stock solutions. For metal complexation, a 10fold molar excess of lanthanide ions to the respective DOTA reagent was used. Complexation was completed after vortexing. The final concentrations were 15 μM model peptide mixtures and 0.2 μM digests. Apparatus. The mixture of the model peptides was separated on a ULTIMATE System (Dionex, Sunnyvale, CA), which was either coupled to an Agilent 7500ce ICPMS (Agilent Technologies, Tokyo, Japan) for specific detection of lanthanides or to a Probot microfraction collector (Dionex, Sunnyvale, CA), to be spotted on Opti-TOF 384-well plates (Applied Biosystems, Darmstadt, Germany) for consecutive peptide identification by MALDI-MS. Peptide Separation by Nano-IP-RP-HPLC. Volumes of 2 μL of a 5 μM peptide solution were loaded over a 5 μL sample loop onto an Acclaim PepMap C18 trap column (5 μm, 0.3 × 10 mm; Dionex, Sunnyvale, CA) with 0.1% aqueous TFA, 3% ACN, and a flow rate of 30 μL/min. The trap column was flushed for 3, 6, 10, or 20 min. Peptides were eluted onto an Acclaim PepMap100 C18 separation column (5 μm, 75 μm × 150 mm; Dionex, Sunnyvale, CA). Separation was performed with a flow rate of 0.3 μL/min and the following gradient for model peptides (after a 6 min flush): 6−26 min, 5%−70% B; 26−27 min, 70%−95% B; 27−37 min, 95% B; 37−38 min, 95%−5% B; 38−59 min, 5% B. For digests: 6−46 min, 5%−70% B; 46−47 min, 70%−95% B; 47−57 min, 95% B; 57−58 min, 95%−5% B; 58−76 min, 5% B. Eluent A was 0.05% aqueous TFA and eluent B 80% ACN, 0.04% TFA, 20% deionized water (v/v/v). Wash runs (35 min) were

chelating agent with different functional groups (e.g., NHSactivated carboxyl groups targeting free amino groups and maleimido-monoamide or iodoacetamide17 for cysteine residues). The latest approaches reported on DOTA/lanthanidelabeled proteins and peptides were comprised of capillary electrophoresis (CE)-ICPMS,18 laser ablation (LA)-ICPMS applied on Western blot immunoassays with labeled antibodies,19 flow injection analysis (FIA)-ICPMS of dissolved twodimensional (2D)-gel electrophoresis spots,20 and 2D-LCICPMS based on strong cation exchange (SCX) and reversedphase (RP) chromatography.21 Most of those ICPMS configurations are combined with ESI-MS as a complementary molecular MS technique. The necessary excess of reagents and lanthanide salts for quantitative derivatization interferes with analysis, both in molecular and elemental MS. A high concentration of salts can impair the ionization of the analytes in ESI and MALDI or hamper proper crystallization of the matrix in the later method. In ICPMS, the lanthanide excess can interfere with the detection of the labeled peptides by elevating distinctly the metal background signal. Sample preparation commonly includes cleaning steps, such as desalting and concentration of the sample to enhance detection in molecular MS but unfortunately does not rule out losses in peptide or protein recovery. Approaches include a large variety of RP resin-based products (ZipTips, magnetic C18 beads, spin columns), on target washing of MALDI spots,22 and liquid chromatography setups. While LC-MALDI MS is an offline setup with fraction collection, in LC-ICPMS online couplings23 metals passing through the analytical column will inevitably reach the detector. As a consequence, an extensive tailing peak caused by the excess of metal in the beginning of the chromatogram with an elevated background throughout the length of the chromatography was reported.11−14 For sensitive quantification of labeled peptides in LC-ICPMS, it is a necessity to get a low baseline for the metal to be monitored. The main focus of this study was set on the implementation of an online precleaning step for labeled peptide samples using a C18 precolumn in nano-LC prior to the detection with ICPMS. A set of model peptides and a digest of lysozyme were therefore labeled with 165Ho, 141Pr, 169Tm, or 175Lu in the form of NHS- and maleimido (Mal-) DOTA complexes. In additional experiments, the possible benefits of employing solid-phase extraction using ZipTips, different ion pairing reagents, and EDTA were monitored. The best precleaning parameters were applied on a tryptic digest of lysozyme labeled with both DOTA reagents and then analyzed with ICPMS and MALDI MS/MS.



EXPERIMENTAL SECTION Chemicals and Materials. DOTA-NHS-ester was purchased from CheMatech (Dijon, France), maleimido-monoamide-DOTA from Macrocyclics (Dallas, TX). Model peptides were synthesized by standard Fmoc solid-phase synthesis (R. Pipkorn, DKFZ, Heidelberg, Germany); sequences included peptide A: EGHIARNCRA; peptide B: LRRACLG; peptide C: GACLLPK; peptide Hy: ESLSSSEE. Tris(2-carboxyethyl)phosphine (TCEP), S-methylmethanethiosulfonate (MMTS), triethylammonium acetate buffer (TEAA), HEPES, trifluoroacetic acid (TFA), heptafluorobutyric acid (HFBA), ethylenediaminetetraacetic acid tetrasodium salt (EDTA), dimethyl sulfoxide (DMSO), acetonitrile (ACN, E Chromasolv), the 3065

dx.doi.org/10.1021/ac303618v | Anal. Chem. 2013, 85, 3064−3070

Analytical Chemistry

Technical Note

Figure 1. Nano-LC-ICPMS chromatograms of mixtures of 165Ho labeled peptides A, B, C: (a) Labeled with Mal-DOTA analyzed with 3, 6 , 10, and 20 min washing times, and the NHS-DOTA-labeled peptide mixture analyzed with a 6 min washing time. Peptide A is marked for the Mal-DOTA runs to demonstrate the peptide loss after 20 min washing time. [The dotted frame in (a) corresponds to the detailed view in (b).] (b) Raw signals for the four Mal-DOTA mix samples, after the start of elution from the trap column. The arrow marks the moment when the valve switches on column, and the loading buffer from the trap column reaches the ICPMS. The double-faced arrow marks the excess metal background used for calculation of the background mean values. Erbium added to the eluent allowed monitoring of the nebulization stability.

model peptides was used. Then, the optimized method was applied to a tryptic digest of lysozyme. Optimization of the Online Precleaning Step. The labeled peptides were loaded and trapped on the C18 precolumn and then washed with mobile phase before they were eluted to the analytical column. Different times for this washing step were applied; metal background and peptide recoveries were monitored in ICPMS. Additionally, the same mixture was analyzed via MALDI MS to confirm the identity of the labeled peptides and to determine the degree of the derivatization. For this purpose, the nano-LC was coupled to a fraction collector that directly mixes the fractions of the chromatographic run with the matrix and spots them on a MALDI target. Washing time of the trap column is referring to the time period during the nano-LC run when the trap column is flushed with a standard loading buffer (3% ACN and 0.1% TFA) after peptide injection. This time period ends when the valve switches to the analytical column, starting the gradient, and thereby elution of the peptides from the trap. The 20-fold molar excess of Mal-DOTA for the labeling of cysteine residues, together with the 10-fold molar excess of lanthanide ions regarding Mal-DOTA, caused a significant amount of nonpeptide-bound metal. In nano-LC-ICPMS, already for a 3 min washing step, no peak of excess metal was observable (Figure 1a) but instead a slightly elevated but steady metal background. Figure 1a shows a set of nano-LCICPMS runs with washing times ranging from 3 to 20 min for 165 Ho Mal-DOTA-labeled peptides A, B and C, plus the same mixture of peptides labeled with 165Ho NHS-DOTA with a 6 min wash time. The chromatograms of the Mal-DOTA-labeled sample show three peaks corresponding to the three model peptides, because they each contain just one cysteine whose sulfhydryl group is targeted by the Mal-DOTA reagent. Backgrounds from metal excess for all Mal-DOTA runs are very low in comparison with the respective peptide peak heights. Due to the higher excess used for derivatization with NHSDOTA (100-fold over free amino groups), the background caused by the excess metal in this case was significantly higher. Also, there were more free amino groups to be targeted than

performed between sample injections. UV-detection was performed at 214 nm. ZipTip. 40 μL sample, containing a total of 2 μg peptide were adjusted to 0.1% HFBA, aspired into the C18 resin and treated according to protocol. MALDI-MS. For MALDI-MS, eluting peptides were directly mixed with matrix solution (3 mg/mL CHCA in 70% ACN, 0.1% TFA, 5 nM Glu1-fibrinopeptide B) in a ratio of 1:3 (v/v) and spotted in 15 s intervals from minute 4 to minute 65 via the Probot microfraction collector. Measurements took place on an AB SCIEX TOF/TOF 5800 (AB Sciex, Darmstadt, Germany) that was used for analysis. Details on operating parameters can be found in the Supporting Information. ICPMS. Instrumental settings (nebulizer gas flow, rf power, ion lens voltages) were tuned daily by monitoring the signals of erbium that was added in form of a standard solution to both eluents A and B. Final erbium concentration in the two eluents was at 40 μg/L. The nano-HPLC system was coupled with the ICPMS instrument via a previously developed nanonebulizer.25,26



RESULTS AND DISCUSSION The objective of this study was the development of a nano-LC online precleaning step for derivatized peptide samples prior to analysis via ICPMS and MALDI-MS and the application of the developed method to more complex samples such as protein digests. This precleaning aimed for a more reliable peptide analysis through lower metal background signals. The reduction of metal load transferred to the MS also contributes to the robustness of the analytical setup (e.g., by preventing unnecessary instrument contamination). In general, for achieving a quantitative derivatization of all peptides in a mixture, the use of excess reagents is inevitable. Different approaches were followed in this study in order to remove the excess lanthanide salts and unbound DOTA complexes from the samples. For this purpose, nano-LC with online analyte concentration/purification on a C18-precolumn (coupled to ICPMS) was applied. The main focus was to find an optimal ion pairing reagent for efficient binding of the analytes on the trap column and to optimize the trap column washing time. For method optimization, a simple peptide mix of three labeled 3066

dx.doi.org/10.1021/ac303618v | Anal. Chem. 2013, 85, 3064−3070

Analytical Chemistry

Technical Note

Figure 2. Nano-LC-ICPMS analysis of DOTA-labeled peptide mixtures. (a) Tm NHS-DOTA-labeled peptide Hy analyzed with HFBA (orange) and with TFA (green) in the loading buffer. (b) Ho NHS-DOTA mixture peptides A, B, C, and Hy (blue) and Tm NHS-DOTA peptide Hy (orange), duplex analysis with an HFBA loading buffer directly after the sample preparation. (c) Reinjection of sample shown in (c) after two hours, showing lowered recovery for Hy. (d). Ho Mal-DOTA-labeled peptides A, B, and C, analyzed with the TFA loading buffer (green), HFBA loading buffer (blue), and peptide mixture, purified with ZipTip and analyzed with the HFBA loading buffer (red).

more than 75% of peptide A and 25% of peptide B were lost compared to the 3 min wash run, while peptide C was still mostly unaffected with a loss of only 6% (see Figure S-1a of the Supporting Information). After washing for 3 min, the holmium background was 1.24% of the peak height of peptide C; 6 min equals a 0.44% background and 10 min a 0.13%, whereas after 20 min the background was lowered to 0.08% (Figure S-1b of the Supporting Information). As a consequence, the 6 min trap wash length was identified as the best parameter for removal of metal excess while maintaining maximum recovery for the different peptides. In previous studies, a broad and intense elution profile caused by a metal excess was commonly observed for HPLC, nanoLC setups with different column materials (C8 and C18), and different types of reagents, such as lanthanide complexes of DTPA or Mal-DOTA and ferrocene-based compounds.11,13,14 A high peak at the beginning makes it impossible to analyze peptides that elute at the same time, whereas the tailing impairs the accurate quantification of the later peaks. In an earlier work,12 Lu DTPA-labeled peptides were analyzed by nano-LCICPMS without a trap precolumn. The peak caused by the excess reagents exceeds the peak height of the highest peptide peak by approximately 15 times. In our study, almost the entire excess reagent was removed by a trap column setup with an optimized purification step; the

cysteines, so that the chromatographic profile of the NHSDOTA-labeled model peptides varied from those of MalDOTA because of partial double derivatizations (Figure 1a). Figure 1b shows an overlay of the raw signals for the excess metal backgrounds of the same set of chromatograms. Erbium was added to the eluents for the gradient, to monitor the stability of the nebulization. Due to the fact that there was no erbium added to the loading buffer, the respective 166Er signal decreased for the time-equivalent of a trap column’s eluent volume, after the valve switches to the analytical column. With the starting gradient, the leftover excess reagents which were not removed by the washing step were finally washed off the trap column continuously. As expected, the metal background was considerably lowered with each further elongation of the washing step prior to separation. Best results regarding minimal background levels were obtained with 20 min of washing time, where the level could be decreased by 93%, compared to the 3 min flush. Simultaneously, losses in peptide recovery were observed in the form of smaller peak areas. In particular, hydrophilic peptides A and B were affected from the decreased recovery (Figure 1a). After 6 min, peptide peak heights remained virtually the same as after 3 min. Losses were already observable for peptide A after 10 min washing time, while the other two peptides were still almost completely recovered. After 20 min, 3067

dx.doi.org/10.1021/ac303618v | Anal. Chem. 2013, 85, 3064−3070

Analytical Chemistry

Technical Note

achieved for the 6 min washing time in the section above. Direct addition of an equivalent amount of EDTA into the sample vial was tested as an alternative procedure leading again to an elevated metal background. In conclusion, no satisfactory reduction of metal background could be achieved using EDTA; therefore, this approach was not further followed in this study. Offline Purification with ZipTip C18. A frequently applied method for offline concentration or desalting of peptide samples is solid phase extraction using, for example, ZipTips. In principle, only the peptides should bind to the C18 residues and impurities should be washed out during the cleaning process. The ZipTip C18 standard protocol was applied using 0.1% HFBA instead of 0.1% TFA, as HFBA is supposed to strengthen the interaction of hydrophilic peptides to the C18 material.27 Figure 2d shows the same sample (peptides A, B, C, Hy), injected with and without ZipTip cleanup. The metal background, but also signal intensities for all peptides in the cleaned sample “HFBA ZipTip”, were significantly reduced. The overall peptide recovery for the three standard peptides was only around 4%. MALDI-MS for Peak Identification. Due to the different capillary lengths and the different backpressures caused by the nano-LC coupling either to the nebulizer for ICPMS or the MALDI spotter, retention times of nano-LC-ICPMS and nanoLC-MALDI-MS runs, did not match. In order to align these runs, UV traces were used, which are supposed to show the same chromatographic profile for both hyphenations (shown for lysozyme digest in Figure S-2a and b of the Supporting Information). After matching the two UV-chromatograms, the delay between ICPMS and MALDI-MS detection was determined, consecutively allowing the calculation of the proper alignment timings for both techniques. Figure 3 shows a representative overlay for an experiment with three maleimido-DOTA-labeled model peptides A, B, and C, analyzed with the 6 min standard protocol using 0.1% TFA as an ion-pairing reagent. Labeling as observable for peptide A is not complete. This represents a common problem for labeling reactions.15 Moreover, multiple peaks for a single

residual amount of excess metal did not cause a significant signal in the chromatograms when starting the elution on the analytical column (Figure 1). Use of HFBA for the Detection of Hydrophilic Peptide Species. Very hydrophilic peptides are characterized by poor retention on RP stationary phases27 in particular when TFA is used as ion pairing reagent and, hence, can undergo significant sample loss in precolumn approaches. Therefore, we tested the replacement of TFA by the more apolar heptafluorobutyric acid (HFBA) in the loading buffer; the eluents for the following analytical separation still contained TFA. In order to trace the signal of the hydrophilic peptide, Hy, in the chromatogram during the analysis of the four-peptide mix by nano-LC-ICPMS, a 165Ho NHS-DOTA-labeled mixture of the peptides A, B, C, and Hy was spiked with a separately 169 Tm NHS-DOTA-derivatized single peptide Hy and analyzed as a multiplex. Due to this, the Hy peak was easily recognizable in the chromatogram for the peptide mixture. It was shown that Hy can be trapped on the C18 precolumn when applying HFBA (Figure 2a). As a side effect, it was observed that the height of the Hy peak decreased with reinjections, up to the point when it could not be detected anymore. This was to be seen in the mixture as well as for the single Hy peptide and did not affect the other peptides in the mixture. Figure 2 (panels b and c) show the analysis of the 165Ho-labeled mixture together with the 169Tmlabeled single peptide, after the first injection and after reinjection after 2 h. After two hours, peak recovery was only 11% in the peak area and height for the 169Tm-labeled peptide, whereas after 5 h, it was not detectable anymore (not shown). Apart from this observation, it was found that the addition of HFBA to the loading buffer caused a retention time shift and impaired peptide separation, when using HFBA instead of TFA before elution of the peptides from the trap to the column. This is shown in Figure 2d, for the mixture of the peptides A, B, and C analyzed with 0.1% TFA and with 0.1% HFBA, respectively; presumably a small amount of HFBA was still transferred onto the analytical column, due to too short equilibration times with the starting of the analytical gradient buffer. The retention times (after identification of the peptides by MALDI-MS) showed that all the peptides, both with maleimido-DOTA(Figure 2d) and NHS-DOTA-label (not shown) were separated completely when using TFA, whereas the application of 0.1% HFBA lead to an extensive overlapping in their retention. EDTA as Additive. A second online precleaning approach tested in this study was based on the same instrumental setup, but using a modified loading buffer, containing EDTA. Being a far less efficient chelator than DOTA (e.g., for Lu3+: log K(EDTA)= 19.8; log K(DOTA) = 25.4),28 EDTA was expected to aid the excess metal removal on the trap, thus enabling a reduction of the wash time in the following step, without rivaling the already existing DOTA complexes. The applied amount of EDTA was based on the flow rate of the loading pump and the injected amount of metal, so that finally the amount of added EDTA equaled the amount of unbound metal in the sample. The peptides were loaded onto the trap column with the standard loading buffer for 1 min, washed with the modified loading buffer (3% ACN with 1 mM EDTA, pH 7) for 2 min, and lastly, washed again with the standard loading buffer for 3 min, to make sure that no EDTA reached the analytical column. All samples analyzed after the use of EDTA showed an elevated background in comparison with the results

Figure 3. Nano-LC-ICPMS/nano-LC-MALDI-MS overlay for the model peptides A, B, and C labeled with Ho Mal-DOTA: ICPMS chromatogram of 165Ho (above) and extracted ion chromatograms for the respective masses of the labeled peptides in MALDI-MS (below). 3068

dx.doi.org/10.1021/ac303618v | Anal. Chem. 2013, 85, 3064−3070

Analytical Chemistry



ACKNOWLEDGMENTS This work has been funded by the Clusters of Excellence “Inflammation@Interfaces” (A.T.) within the Excellence Initiative of the German Federal Government, by the German-French PROCOPE program of the DAAD (A.H.) and CampusFrance (D.S.), the BMBF-project “SweeOmics” (C.-W.H., A.H.) and the BQR project of the University of Pau (D.S.).

peptide can appear after labeling with Mal-DOTA, corresponding to stereoisomers as reported before by other groups.16,29,30 Application on a Lysozyme Digest. A tryptic digest of lysozyme was labeled using 169Tm NHS-DOTA and 141Pr MalDOTA. The labeled digests, combined after derivatization, were analyzed by nano-LC-ICPMS and in parallel by nano-LCMALDI-MS/MS. Due to the optimized online precleaning procedure encompassing the 6 min wash steps described above, we did not observe a high peak with long tailing in nano-LCICPMS for either of the two labeled digests. The baseline was elevated accordingly, but this did not interfere with the interpretability of the peaks. All further details concerning the digest samples can be found in Figure S-2 of the Supporting Information.



CONCLUSION It was shown that excess reagents, in particular metal ions, can be removed online by application of a precolumn, which enables the trapping and cleaning of labeled peptides. The optimized parameters allowed for an improvement of peptide recovery, while lowering the background by more than 60% in comparison to a shorter cleaning step. The choice of the ionpairing reagent is a critical parameter. With the use of TFA, best results were achieved. For hydrophilic peptides, HFBA was chosen for the trapping step; however, the observed unsatisfactory separation of the peptides in the following TFA-based separation is clearly an artifact caused by carryover of HFBA onto the analytical column. In contrast to the online precleaning procedure, application of EDTA and ZipTips did not give promising results. An implementation of this precolumn/precleaning step in other chromatographic setups is straightforward and might enhance the performance for analysis with ICPMS also in other fields of analytics (e.g., environmental analyses). The parallel use of nano-LC-ICPMS and nano-LC-MALDI-MS/MS was established for the analysis of lanthanide-labeled peptides and is a first step toward application of combined element/molecular MS for the analysis of more complex protein mixtures in quantitative proteomics. Nevertheless, to reach this goal, further developments will be necessary with improved derivatization yields, improved separation schemes, and the development of bioinformatic tools to align ICP and molecular MS information and for the proper readout of quantitative information in chromatographically nonresolved peaks. ASSOCIATED CONTENT

S Supporting Information *

This material is available free of charge via the Internet at http://pubs.acs.org.



REFERENCES

(1) Gygi, S. P.; Rist, B.; Gerber, S. A.; Turecek, F.; Gelb, M. H.; Aebersold, R. Nat. Biotechnol. 1999, 17, 994−999. (2) Ong, S. E.; Blagoev, B.; Kratchmarova, I.; Kristensen, D. B.; Steen, H.; Pandey, A.; Mann, M. Mol. Cell. Proteomics 2002, 1, 376− 386. (3) Ross, P. L.; Huang, Y. N.; Marchese, J. N.; Williamson, B.; Parker, K.; Hattan, S.; Khainovski, N.; Pillai, S.; Dey, S.; Daniels, S.; Purkayastha, S.; Juhasz, P.; Martin, S.; Bartlet-Jones, M.; He, F.; Jacobson, A.; Pappin, D. J. Mol. Cell. Proteomics 2004, 3, 1154−1169. (4) Gerber, S. A.; Rush, J.; Stemman, O.; Kirschner, M. W.; Gygi, S. P. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 6940−6945. (5) Pröfrock, D.; Prange, A. Appl. Spectrosc. 2012, 66, 843−868. (6) Tholey, A.; Schaumlöffel, D. TrAC, Trends Anal. Chem. 2010, 29, 399−408. (7) Rappel, C.; Schaumlöffel, D. Anal. Bioanal. Chem. 2008, 390, 605−615. (8) Becker, J. S.; Boulyga, S. F.; Pickhardt, C.; Becker, J.; Buddrus, S.; Przybylski, M. Anal. Bioanal. Chem. 2003, 375, 561−566. (9) Tastet, L.; Schaumlöffel, D.; Bouyssiere, B.; Lobinski, R. Talanta 2008, 75, 1140−1145. (10) Kutscher, D. J.; Bettmer, J. Anal. Chem. 2009, 81, 9172−9177. (11) Bräutigam, A.; Bomke, S.; Pfeifer, T.; Karst, U.; Krauss, G. J.; Wesenberg, D. Metallomics 2010, 2, 565−570. (12) Rappel, C.; Schaumlöffel, D. Anal. Chem. 2009, 81, 385−393. (13) Patel, P.; Jones, P.; Handy, R.; Harrington, C.; Marshall, P.; Evans, E. H. Anal. Bioanal. Chem. 2008, 390, 61−65. (14) Yan, X.; Xu, M.; Yang, L.; Wang, Q. Anal. Chem. 2010, 82, 1261−1269. (15) Gregorius, B.; Schaumlöffel, D.; Hildebrandt, A.; Tholey, A. Rapid Commun. Mass Spectrom. 2010, 24, 3279−3289. (16) Esteban-Fernández, D.; Scheler, C.; Linscheid, M. W. Anal. Bioanal. Chem. 2011, 401, 657−666. (17) Schwarz, G.; Beck, S.; Weller, M. G.; Linscheid, M. W. Anal. Bioanal. Chem. 2011, 401, 1203−1209. (18) Yang, M. W.; Wang, Z. W.; Fang, L.; Zheng, J. P.; Xu, L. J.; Fu, F. F. J. Anal. At. Spectrom. 2012, 27, 946−951. (19) Waentig, L.; Jakubowski, N.; Hardt, S.; Scheler, C.; Roos, P. H.; Linscheid, M. W. J. Anal. At. Spectrom. 2012, 27, 1311−1320. (20) Bergmann, U.; Ahrends, R.; Neumann, B.; Scheler, C.; Linscheid, M. W. Anal. Chem. 2012, 84, 5268−5275. (21) Esteban-Fernández, D.; Ahrends, R.; Linscheid, M. W. J. Mass Spectrom. 2012, 47, 760−768. (22) Šalplachta, J.; Ř ehulka, P.; Chmelík, J. J. Mass Spectrom. 2004, 39, 1395−1401. (23) Schaumlöffel, D. Anal. Bioanal. Chem. 2004, 379, 351−354. (24) ExPASy PeptideMass web site. http://web.expasy.org/peptide_ mass/ (accessed Oct 31, 2012). (25) Giusti, P.; Lobinski, R.; Szpunar, J.; Schaumlöffel, D. Anal. Chem. 2006, 78, 965−971. (26) Rappel, C.; Schaumlöffel, D. J. Anal. At. Spectrom. 2010, 25, 1963−1968. (27) Tholey, A.; Toll, H.; Huber, C. G. Anal. Chem. 2005, 77, 4618− 4625. (28) Byegård, J.; Skarnemark, G.; Skålberg, M. J. Radioanal. Nucl. Chem. 1999, 241, 281−290. (29) Ahrends, R. MeCAT - Neue Wege in der Peptid- und Proteinquantifizierung. Ph.D. Thesis, Humboldt University, Berlin, 2009.





Technical Note

AUTHOR INFORMATION

Corresponding Author

*Dirk Schaumlöffel, Université de Pau et des Pays de l’Adour/ CNRS UMR 5254, Laboratoire de Chimie Analytique BioInorganique et Environnement/IPREM, Hélioparc, 2 av. du Prés ident Angot, 64053 Pau, France. E-mail: dirk. schaumloeff[email protected]. Tel: +33-559-407760 Fax: +33559-407674. Notes

The authors declare no competing financial interest. 3069

dx.doi.org/10.1021/ac303618v | Anal. Chem. 2013, 85, 3064−3070

Analytical Chemistry

Technical Note

(30) Pieper, S. Metallchelatkomplexe für die element-massenspektrometrische Quantifizierung von Peptiden und Proteinen. Ph.D. Thesis, Humboldt University, Berlin, 2008.

3070

dx.doi.org/10.1021/ac303618v | Anal. Chem. 2013, 85, 3064−3070