Analytical Chemistry - ACS Publications - American Chemical Society


Analytical Chemistry - ACS Publications - American Chemical Societyhttps://pubs.acs.org/doi/abs/10.1021/ac034819hCachedS...

0 downloads 88 Views 125KB Size

Anal. Chem. 2003, 75, 6837-6842

Development of a Sheathless Interface between Reversed-Phase Capillary HPLC and ICPMS via a Microflow Total Consumption Nebulizer for Selenopeptide Mapping Dirk Schaumlo 1 ffel,† Jorge Ruiz Encinar,† and Ryszard Łobin´ski*,†,‡

Group of Bio-Inorganic Analytical Chemistry, CNRS UMR 5034, He´ lioparc, 2, Av. Pr. Angot, F-64053 Pau, France, and Department of Analytical Chemistry, Warsaw University of Technology, 00-664 Warsaw, Poland

A sheathless interface based on a total consumption micronebulizer operating at flow rates in the range 0.57.5 µL min-1 was developed between capillary HPLC and ICPMS. It allowed the efficient nebulization and transport into the plasma of mobile phases containing up to 100% organic solvent without either cooling the spray chamber or oxygen addition. The coupled system was applied to selenopeptide mapping in a protein fraction isolated from a selenized yeast extract. The detection limits were 150 (80Se) and 200 fg (82Se) for a quadrupole instrument with and without a collision cell, respectively, which is a factor 100-150 less than that reported elsewhere for HPLCICPMS. The minimal peak broadening (∼5 s at the halfheight) allowed baseline resolution of a mixture containing more than 30 selenopeptides, many of which could not be separated using the conventional HPLC-ICPMS coupling. Research toward understanding the mechanisms of interactions of trace elements and metal probes with protein ligands in living organisms requires analytical techniques that are able to provide information on the identity and concentrations of element species occurring in biological tissues at the picogram and lower levels.1 The acquisition of this type of data is possible with hyphenated (coupled) techniques that combine a highresolution separation technique with sensitive element or moleculespecific detection.2 The coupling of HPLC employing different separation mechanisms with inductively coupled plasma mass spectrometry (ICPMS) has been the most widely used.1,2 The need for analysis of smaller samples, such as, for example, digests of spots in 2D electrophoresis, compartments of individual cells, or human biopsy extracts, has spurred in the past decade the development of nanoflow separation techniques in proteomics research.3 Despite the increasing popularity of electrochromato* To whom correspondence should be addressed. Tel.: +33-559-80-6884. Fax: +33-559-80-1292. E-mail address: [email protected] † Group of Bio-Inorganic Analytical Chemistry. ‡ Warsaw University of Technology. (1) Szpunar, J.; Lobinski, R.; Prange, A. Appl. Spectrosc. 2003, 57, 102A-112A. (2) Szpunar, J. Analyst 2000, 125, 963-988. (3) Shen, Y.; Smith, R. D. Electrophoresis 2002, 23, 3106-3124. 10.1021/ac034819h CCC: $25.00 Published on Web 11/12/2003

© 2003 American Chemical Society

graphic techniques, such as capillary electrophoresis4-6 and capillary electrochromatography,4,5 the position of capillary HPLC is still dominant.7,8 Indeed, HPLC is a robust, reliable, and reproducible separation technique for various types of samples, provides high resolution, especially when used in gradient mode, and can be easily implemented to various sample types and application areas. It can be scaled down to nanoflow dimensions and used in single or multidimensional separations.3 However, the capillary HPLC techniques are a priori incompatible with ICPMS because of the flow rates being 100-1000 times lower than those (∼0.7-1 mL min-1) required by conventional nebulizers. Also, the large dead volume (40-100 cm3) of the most commonly used double-pass Scott spray chamber results in long washout times and peak broadening. Furthermore, methanol or acetonitrile added to the mobile phase in reversed-phase HPLC negatively affects the ICP stability9 and results in a decrease in signal intensity and in carbon deposition on the cones. Oxygen addition to the plasma gas10,11 and, consequently, the use of platinum cones are required. The removal of the solvent vapor using a membrane desolvator12,13 or a cooled spray chamber,10,14 commonly used to couple HPLC to ICPMS, will result at lowmicroliter per minute flow rates in peak broadening and thus in loss of chromatographic resolution. Therefore, it is necessary to develop a dedicated interface between capillary HPLC and ICPMS based on a high-efficiency low-sample-consumption zero-dead volume sample introduction system. (4) Altria, K. D. J Chromatogr., A 1999, 856, 443-463. (5) Kasicka, V. Electrophoresis 2001, 22, 4139-4162. (6) Manabe, T. Electrophoresis 1999, 20, 3116-3121. (7) Boyes, B. E.; Gratzfeld-Husgen, A.; Weber, R. Chimia 2001, 55, 48-49. (8) Shen, Y.; Zhao, R.; Belov, M. E.; Conrads, T. P.; Anderson, G. A.; Tang, K.; Pasa-Tolic, L.; Veenstra, T. D.; Lipton, M. S.; Udseth, H. R.; Smith, R. D. Anal. Chem. 2001, 73, 1766-1775. (9) Ferrarello, C. N.; Ruiz Encinar, J.; Centineo, G.; Garcia Alonso, J. I.; Fernandez de la Campa, M. R.; Sanz-Medel, A. J. Anal. At. Spectrom. 2002, 17, 1024-1029. (10) Ruiz Encinar, J.; Ouerdane, L.; Buchmann, W.; Tortajada, J.; Lobinski, R.; Szpunar, J. Anal. Chem. 2003, 75, 3765-3774. (11) Kahen, K.; Strubinger, A.; Chirinos, J. R.; Montaser, A. Spectrochim. Acta, Part B 2003, 58B, 397-413. (12) Cairns, W. R. L.; Ebdon, L.; Hill, S. J. Fresenius J. Anal. Chem. 1996, 355, 202-208. (13) Lustig, S.; Michalke, B.; Beck, W.; Schramel, P. Fresenius J. Anal. Chem. 1998, 360, 18-25. (14) Makarov, A.; Szpunar, J. J. Anal. At. Spectrom. 1999, 14, 1323-1327.

Analytical Chemistry, Vol. 75, No. 24, December 15, 2003 6837

Micronebulizers, such as a Micromist nebulizer,15,16 highefficiency nebulizer,17 microconcentric nebulizer (MCN),16 direct injection nebulizer,18,19 or direct injection high-efficiency nebulizer (DIHEN),20 which work at typical sample uptake rates 10 times lower than conventional nebulizers, reportedly allowed efficient interfacing of microbore (1-mm i.d.) HPLC operated at flow rates between 40 and 100 µL min-1.15,17,19,20 In combination with the Micromist nebulizer, a low-dead-volume cyclonic spray chamber was reported.15 The direct injection nebulizers do not require a spray chamber and have small internal dead volumes. On the other hand, the primary aerosol produced results in high oxide ion rates.21 The only reported capillary HPLC-ICPMS coupling22-25 was optimized for 31P detection and used for the analysis of phosphopeptides23 in a tryptic digest of β-casein22 and the identification of the phosphorylation sites in phosphoproteins.24,25 Three different micronebulizers were tested: the MCN 6000,22 the PFA 100,22-25 and the DIHEN.23 The disadvantages of the former two included the sample uptake rates of 10-100 µL, which were still higher than those demanded by capillary HPLC (∼4 µL), and the large dead volume of the desolvation system of the MCN 6000. Even with the modified DIHEN, the peaks were relatively broad (∼15-s peak width at half-height). Recently, the combination of a total consumption micronebulizer with a single-pass low-volume spray chamber (5 cm3) was shown to allow the complete transport (no drain) of flow rates of ∼6 µL.26 This concept was adopted and enhanced by combining the spray chamber and torch into an integrated sample introduction system.27 The objective of this work was to investigate this concept for the purpose of coupling capillary (300-µm i.d.) HPLC with ICPMS. The interface to be developed should allow sensitive element-specific detection in organic-rich mobile phases and be operated without draining, cooling, makeup liquid, or addition of oxygen to the plasma. The performance of the interface developed was demonstrated by mapping of selenium peptides in a tryptic digest of a selenium-containing protein at the low- and subpicogram level. EXPERIMENTAL SECTION Apparatus. Capillary HPLC System. Capillary HPLC separations were performed using a dual syringe solvent delivery system (15) Polec, K.; Garcia-Arribas, O.; Perez-Calvo, M.; Szpunar, J.; Ribas-Ozonas, B.; Lobinski, R. J. Anal. At. Spectrom. 2000, 15, 1363-1368. (16) Ackley, K. L.; Sutton, K. L.; Caruso, J. A. J. Anal. At. Spectrom. 2000, 15, 1069-1073. (17) Pergantis, S. A.; Heithmar, E. M.; Hinners, T. A. Anal. Chem. 1995, 67, 4530-4535. (18) Chassaigne, H.; Szpunar, J. Analusis 1998, 26, M48-M51. (19) Gammelgaard, B.; Bendahl, L.; Sidenius, U.; Jons, O. J. Anal. At. Spectrom. 2002, 17, 570-575. (20) Acon, B. W.; McLean, J. A.; Montaser, A. J. Anal. At. Spectrom. 2001, 16, 852-857. (21) Becker, J. S.; Dietze, H.-J.; McLean, J. A.; Montaser, A. Anal. Chem. 1999, 71, 3077-3084. (22) Wind, M.; Edler, M.; Jakubowski, N.; Linscheid, M.; Wesch, H.; Lehmann, W. D. Anal. Chem. 2001, 73, 29-35. (23) Wind, M.; Eisenmenger, A.; Lehmann, W. D. J. Anal. At. Spectrom. 2002, 17, 21-26. (24) Wind, M.; Kelm, O.; Nigg, E. A.; Lehmann, W. D. Proteomics 2002, 2, 15161523. (25) Wind, M.; Wesch, H.; Lehmann, W. D. Anal. Chem. 2001, 73, 3006-3010. (26) Schaumlo¨ffel, D.; Prange, A. Fresenius J. Anal. Chem. 1999, 364, 452456. (27) Todoli, J.-L.; Mermet, J.-M. J. Anal. At. Spectrom. 2002, 17, 345-351.

6838

Analytical Chemistry, Vol. 75, No. 24, December 15, 2003

(model 140C, Applied Biosystems, Foster City, CA). A stable flow rate of less than 5 µL min-1 demanded by capillary HPLC columns (i.d. 75-300 µm) is normally achieved by accurate splitting of a higher flow rate gradient (∼100 µL min-1) delivered by the HPLC pump.23 In this work, the flow rate of 106 µL min-1 at the exit of the pump was split 1:26.5 to achieve a flow of 4 µL min-1 at the inlet of the HPLC column using a model IC-100-VAR Accurate splitter (LC Packings, San Francisco, CA). A piece of 0.3-mm-i.d. PEEK tubing was used as a calibrator for the microflow. A µ-Guard column MGU 30 C18 (LC Packings) was used as a precolumn, and the reversed-phase capillary column was a Hypersil C18 BDS (300 µm i.d. × 15 cm, 3 µm, LC Packings). Injections were made using a model 7520 microinjector fitted with a 200-nL Vespel rotor (Rheodyne, Cotati, CA). The connections from the pump to the splitter and from the splitter to the injection valve were made of PEEK tubing. To reduce the dispersion of the sample, the connections from the injection valve to the precolumn and from the capillary column to the interface were made of fused-silica capillaries (75-µm i.d., 280-µm o.d., LC Packings). The guard column was connected to the capillary column by means of a lowdead-volume PEEK fitting. For flow injection experiments, the column and the guard column were replaced by a 30-cm fusedsilica capillary (75-µm i.d., LC Packings). For direct sample introduction experiments, solutions were fed by means of a Harvard apparatus model 22 syringe pump. Capillary HPLC-ICPMS Interface. An ICPMS instrument, ELAN 6000 (Perkin-Elmer Sciex, Thornhill, ON, Canada), was used and compared in terms of detection limits with an ICP collision cell MS (Agilent 7500c, Yokogawa Analytical Systems, Tokyo, Japan). The column eluate was introduced into the ICP via a microflow total consumption nebulizer fitted with a low-deadvolume (8 cm3) spray chamber without drain. The small dimensions of the nebulizer capillary (50-µm i.d., 180-µm o.d., × 41 mm) resulted in a high flow resistance, which allowed a stable flow in the microliter range (0.5-11 µL min-1) without addition of a sheath flow. The nebulizer capillary tip was centered in a sapphire orifice of 254-µm i.d., allowing the generation of a fine aerosol, which could be optimized by adjusting the position of the capillary tip in the nebulizer orifice. The outlet of the HPLC column was connected to the nebulizer capillary by means of a fused-silica capillary (75-µm i.d.). The whole instrumental setup with an enlargement of the interface is shown in Figure 1. Reagents, Solutions, and Materials. Analytical reagent grade chemicals purchased from Sigma-Aldrich (Saint-Quentin Fallavier, France) and water (18 MΩ cm) obtained with a Milli-Q system (Millipore, Bedford, MA) were used throughout unless stated otherwise. Solutions of trypsin (1.33 mg mL-1) and dithiothreitol (36 mg mL-1) were prepared by dissolving the corresponding reagent in 0.1 mol L-1 Tris buffer (pH 7.8). A commercial preparation of selenized yeast (Alltech, Lexington, KY) having a total selenium content of 2.1 ( 0.1 mg g-1 was used as the sample. Microcentrifuge filters for ultrafiltration with a cutoff of 100 kDa were purchased from Millipore. Procedures. Preparation of Samples. The selenoproteincontaining fraction was isolated by preparative size-exclusion chromatography of a water extract of selenized yeast and lyophilized as described elsewhere.10 The lyophilizate was redissolved

ICPMS Conditions. ICPMS measurements conditions (nebulizer gas flow, rf power, and lens voltage) were optimized daily for the highest intensity of the 103Rh signal. For this purpose 25 µg L-1 Rh(III) was added as Rh(NO3)3 to the mobile phases A and B. When the ICP collision cell MS was used, the 80Se isotope was monitored as well. To remove Ar2+ interference, the H2 flow rate in the collision cell was optimized to 3.6 mL min-1, the value for the octopole bias was set to -13.8 V, and the quadrupole bias to -12.5 V. Capillary HPLC-ICPMS Interface Conditions. The nebulization process was optimized by adjusting the position of the nebulizer capillary tip in the nebulizer orifice while monitoring the 103Rh signal by ICPMS. The signal intensity, precision, and stability over time were considered. To minimize dead volume and reduce the peak width, the connections between the injection valve and the nebulizer were optimized. Zero-dead-volume connections were achieved by an exact flat cut of the fused-silica capillaries, which allowed a direct abutting of the capillaries in the fittings. The peak shape was monitored by injection of a 200nL sample aliquot containing 100 µg L-1 selenium (as selenomethionine) in the flow injection mode.

Figure 1. (a) Instrumental setup of the capillary HPLC-ICPMS system: (1) microflow HPLC pump; (2) splitter; (3) calibrator; (4) injector; (5) guard column; (6) capillary column; (7) microflow nebulizer; (8) spray chamber; (9) ICP torch. (b) Detailed view of the interface: (10) fused-silica capillary 75-µm i.d.; (11) fitting; (12) screw for adjustment of the nebulizer tip; (13) O-rings; (14) nebulizer gas (argon) inlet; (15) nebulizer capillary; (16) nebulizer orifice; (17) zerodead-volume connection of the capillary column outlet (fused-silica capillary) to the nebulizer capillary.

in 120 µL of 0.1 mol L-1 Tris buffer and mixed with 15 µL of a dithiothreitol solution. After 1 h, trypsin solution (50 µL) was added and allowed to react for 17 h at 37 °C. The pH of the mixture was adjusted to 5 by adding 2 µL of acetic acid in order to stop the reaction. The mixture was filtered through a 100-kDa cutoff ultrafilter by centrifugation at 3000g for 15 min. The filtrate containing selenopeptides was analyzed by reversed-phase capillary HPLC-ICPMS. Chromatographic Conditions. The mobile phases A and B were 0.1% trifluoroacetic acid (TFA) in water and in methanol, respectively. The solvents were degassed by purging with helium prior to addition of TFA. Selenomethionine was eluted isocratically with 30% of B for the determination of the detection limit as well as for the calibration of the system. The tryptic digest solution was diluted 1+4 with the starting mobile phase (10% B) prior to injection. A 200-nL aliquot was injected. Selenium-containing peptides were separated using a stepwise gradient: 0-5 min 1025% B linear; 5-20 min 25-32% B linear; 20-25 min 32-37% B; 25-35 min 37% B isocratic; 35-55 min 37-42% B linear, and 5560 min 42-100% B linear. The selenium isotopes 77Se, 78Se, and 82Se were monitored on-line by ICPMS in order to detect selenomethionine and selenium-containing peptides.

RESULTS AND DISCUSSION Performance of the Capillary HPLC-ICPMS Interface. The working range of the microflow nebulizer was investigated in terms of signal intensity and precision and the formation rates of oxides and doubly charged ions. In contrast to an earlier work describing a capillary electrophoresis-ICPMS interface,26 where the nebulizer was working in the self-aspiration mode, in this study the sample solution was introduced by a pump without addition of a makeup liquid. Furthermore, the effect of organic solvents typically used in HPLC (methanol, acetonitrile, 2-propanol) on the plasma stability and ionization was studied. Flow Rate, Intensity, and Precision. Low and precise flow rates between 0.5 and 11 µL min-1 of a test solution containing Rh, Ba, Ce (each 25 µg L-1), and Se (250 µg L-1) could be achieved with a syringe pump. Figure 2a shows a linear (r2 ) 0.995) increase in intensity as a function of increasing flow rate between 0.5 and 7.0 µL min-1, to reach a plateau beyond this value. The relative standard deviation of the 82Se intensity measurement decreases with increasing flow rate to arrive at a value of ∼4.5%. From 0.5 µL min-1 on, a stable nebulization could be achieved. At flow rates higher than 7.5 µL min-1, the formation of a thin water film, and above 9.5 µL min-1, a noticeable condensation in the spray chamber, were observed. From these data it can be concluded that the investigated microflow nebulizer showed an analyte transport efficiency of nearly 100% when operated at flow rates below 7.5 µL min-1 and can be regarded as a total consumption nebulizer. The nebulizer behaves similarly to the high-efficiency cross-flow micronebulizer (HECFMN) described elsewhere, for which 100% transport efficiency was measured for a flow rate of 5 µL min-1.28 Oxide Ions and Double Charged Ions. The formation of oxide and doubly charged ions in the ICP may lead to possible interferences in the mass spectrometer. Among several factors influencing the generation of oxide ions, the aerosol droplet (28) Li, J.; Umemura, T.; Odake, T.; Tsunoda, K. Anal. Chem. 2001, 73, 14161424.

Analytical Chemistry, Vol. 75, No. 24, December 15, 2003

6839

Figure 3. Effect of different organic solvents on the selenium signal intensity in ICPMS.

Figure 2. (a) Intensity and relative standard deviation (RSD) of the 82Se signal versus the nebulizer sample uptake rate; (b) formation of oxide ions and doubly charged ions as a function of the nebulizer flow rate.

distribution is directly related to the nebulizer performance. The presence of larger droplets is responsible for the occurrence of cold regions locally in the plasma where oxide ions cannot decompose completely.29,30 In this study, the intensity of the CeO+ ion was determined to study the CeO+/Ce+ oxide ion formation rate as a function of the nebulizer flow rate. Figure 2b shows a slight increase (from 0.3 to 1%) in the oxide formation rate between flow rates of 0.5 and 11 µL min-1. This is very low not only in comparison with conventional cross-flow nebulizers (oxide ion formation rate of 5% at the uptake rate of ∼1 mL min-1) but also with microflow nebulizers such as DIHEN31 and HECFMN.28 Indeed, the latter show a CeO+/Ce+ ratio of 16 and 2.5%, respectively, at a sample uptake rate of 10 µL min-1. These results indicate that the microflow nebulizer investigated in this study generated a very fine aerosol. At flow rates lower than 3 µL min-1, the formation of double charged ions increased, which was observed by measuring the Ba2+/Ba+ ratio, whereas between 3 and 11 µL min-1, this ratio was stable at ∼1.5%. Nebulization of Organic Solvents. The nebulization of organic solvents typically used in HPLC, such as methanol, acetonitrile, and 2-propanol, was investigated. All three solvents could be introduced into the ICP at flow rates of up to 25 µL min-1 without condensation in the spray chamber and without extinction of the (29) Liu, H.; Clifford, R. H.; Dolan, S. P.; Montaser, A. Spectrochim. Acta, Part B 1996, 51B, 27-40. (30) Tanner, S. D. J. Anal. At. Spectrom. 1993, 8, 891-897. (31) McLean, J. A.; Zhang, H.; Montaser, A. Anal. Chem. 1998, 70, 1012-1020.

6840 Analytical Chemistry, Vol. 75, No. 24, December 15, 2003

plasma even in the absence of oxygen. This opens the way to use solvent gradients of up to 100% in capillary HPLC-ICPMS that has hardly been possible otherwise.9,11-14 Since an organic solvent modifies the plasma ionization conditions, its concentration has a significant effect on the signal intensity in ICPMS. For elements with high first ionization energies, such as selenium, the enhancement of the sensitivity in ICPMS by organic modifiers has been reported.32 An addition of 3% v/v methanol to the aqueous chromatographic mobile phase was found to be optimal for Se detection in ICPMS using a conventional (1 mL min-1) nebulizer and a double-pass spray chamber.33 This phenomenon was suggested to involve a charge-transfer reaction from ionized carbon in the plasma to incompletely ionized selenium.33 To investigate the effect of organic solvents on the 82Se signal intensity and on the formation of oxide and double charged ions in this work, aqueous solutions containing up to 100% methanol, acetonitrile, and 2-propanol and Ba, Ce (each 25 µg L-1), and Se (250 µg L-1) were introduced into the plasma at a flow rate of 4 µL min-1 using a syringe pump. Figure 3 shows the intensities measured as a function of the organic solvent concentration and normalized to those measured in the absence of the organic solvent. The maximum intensities were achieved at concentrations of 40% methanol, 30% acetonitrile and 35% 2-propanol, respectively. At the maximum, the 82Se intensity increased almost 5 times for any of the three solvents in comparison with the measurement in pure aqueous solution. The formation of oxide ions and doubly charged ions (results not shown) decreased with increasing organic solvent concentration. For 10% 2-propanol, an exceptionally high Ba2+/Ba+ ratio of 7.5% was observed. Although the introduction of 100% organic solvent was possible without extinguishing the plasma, the selenium ionization process was compromised at high concentrations of the organic phase. At concentrations exceeding 80% methanol or 60% acetonitrile, no intensity for 82Se could be measured. It is interesting to note that the organic solvent effect on the selenium intensity was different when another instrument (shieldtorch plasma equipped with a collision cell) was used. The gain in signal intensity was less pronounced; at the maximum, the 80Se intensity increased only 2 times the intensity at a methanol concentration of 30%, but at 100% methanol, two-thirds of the 80Se intensity could still be “recovered” in comparison with the (32) Larsen, E. H. Spectrochim. Acta, Part B 1998, 53B, 253-265. (33) Larsen, E. H.; Stuerup, S. J. Anal. At. Spectrom. 1994, 9, 1099-1105.

Table 1. Absolute Detection Limits for Selenium in HPLC-, GC-, and CE-ICPMS Couplings technique

column diam, mm

mass analyzer

species

detection limit for Se, pg

ref

capillary HPLC-ICPMS micro HPLC-ICPMS HPLC-ICPMS HPLC-ICPMS HPLC-ICPMS HPLC-ICPMS HPLC-ICPMS GC-ICPMS CE-ICPMS CE-ICPMS

0.3 1 3 4 4.6 4.6 4.6 5 0.075 0.075

collision cell, quadrupole quadrupole reaction cell, quadrupole quadrupole quadrupole quadrupole quadrupole quadrupole quadrupole sector field

selenomethionine selenomethionine selenoamino acids Se(IV) trimethylselenonium ion Se(IV) selenomethionine Me2Se selenomethionine Se(IV)

0.15 3.0 2.6-4.4 28 22 19 59 2.5 28 0.174

this work 19 41 35 37 36 36 38 42 39

Figure 4. Effect of the instrumental setup on the peak shape obtained after injection of a 200-nL sample of selenomethionine (100 µg L min-1 as Se): (a) flow injection mode using a fused-silica capillary (75-µm i.d.); (b) chromatography in the capillary column (300µm i.d.), isocratic elution (30% B).

measurement in water. It was also reported elsewhere that the effect of organic solvents on the plasma, and therefore on the signal intensity, strongly depended on the type of the ICPMS instrument.9 Analytical Performance of the Capillary HPLC-ICPMS Coupling. Peak Shape. After optimization of the connection between the injection valve and the nebulizer to minimize dead volume, the dispersion of the injected analyte was studied by measuring the peak width in the flow injection mode (without an HPLC column). The injection of a 200-nL sample aliquot containing 100 µg L-1 Se (as selenomethionine) resulted in a 3.6- and a 12-s peak width measured at the half-height and at the base, respectively (Figure 4a). The dotted line in Figure 4a shows an ideal Gaussian peak shape for the purpose of comparison. As can be seen, the tailing of the measured peak due to dispersion of the sample and the dead volumes of the spray chamber and the connectors is minimal. When a guard column and an analytical column were present, injection of a 200-nL sample of selenomethionine (100 µg L-1 as Se) in conditions of its nonretention produced a peak 5.6 s wide at half-height and 15 s at the base (Figure 4b) demonstrating that the influence of the column on the dispersion of the sample and on the peak shape was minimal as well. Calibration and Detection Limits. The capillary HPLC-ICPMS system was calibrated by injections of selenomethionine solutions at conditions of weak retention (evaluated for 3.2 min). Figure 5 shows the peaks for injected standard solutions between 10 and

Figure 5. Calibration of the capillary HPLC-ICPMS system demonstrated by double injections of selenomethionine standard solutions between 10 and 250 µg L-1 Se, isocratic elution (30% B). Inset: background intensity of 82Se and the signal of a 2 µg L-1 Se (400 fg) injection.

250 µg L-1 Se, where the double injection of each standard demonstrates the repeatability of the measurement. The regression r2 of the calibration graph was 0.9991, representing good precision of the measurements and linearity. The detection limits for 77Se and 82Se in a 30% (v/v) methanol solution were determined by injection of four blank solutions and four subsequent injections of 2 µg L-1 Se (400 fg). An example of the chromatographic peak obtained under these conditions and its baseline vicinity is shown in the inset to Figure 5. The noise intensity was measured for the blank and on the basis of a signal-to-noise ratio of 3 the detection limit for 82Se was calculated to be 1 µg L-1 and 200 fg, respectively. For 77Se detection, a slightly higher detection limit of 1.7 µg L-1 and 340 fg, respectively, was found. To improve the detection limits for selenium, an ICP collision cell MS (Agilent 7500c instrument) was used. Under optimized collision cell conditions, the Ar2+ ions were removed, enabling an interference-free detection of the most abundant 80Se isotope. This could be demonstrated by measuring the 82Se/80Se isotope ratio of 0.1831 (theoretical, 0.1851) and the 77Se/80Se ratio of 0.1502 (theoretical, 0.1529). The observed differences can be considered negligible in view of the measurement’s standard deviation of 0.0040. When selenomethionine was eluted isocratically with a mobile phase containing 30% methanol, the detection limit for 80Se was determined to be 0.75 µg L-1 and 150 fg. The gain in detection limit (factor 1.3) in comparison with the 82Se detection by ELAN 6000 is much smaller than expected. Indeed, in a comparison study between the normal mode (82Se Analytical Chemistry, Vol. 75, No. 24, December 15, 2003

6841

They are also a factor of 10 lower than those reported for microbore HPLC-ICPMS19 and GC-ICPMS38 (3.0 and 2.5 pg, respectively). Only one study reported comparable detection limits for a capillary electrophoresis-ICPMS coupling using a sectorfield instrument.39 Capillary-HPLC-ICPMS for Selenopeptide Mapping. The developed capillary HPLC-ICPMS coupling was applied for selenopeptide mapping in a tryptic digest of a selenium-containing protein isolated from selenized yeast as described elsewhere.10 Figure 6 compares the chromatograms obtained using a normal 4.6-mm-i.d. C18 column with a conventional interface40 and using a 300-µm capillary C18 column with the interface developed. The gain in resolution is spectacular for both the hydrophilic and hydrophobic peptides. About twice as many peaks could be resolved by capillary HPLC chromatography. Taking into account that 750 times less sample was injected on the capillary column than on the normal column, the capillary HPLC-ICPMS was almost 100 times more sensitive than the conventional HPLCICPMS. Five peaks could be assigned by retention time to selenopeptides, which were previously isolated and identified as described elsewhere.10 Peaks in the chromatogram at retention times between 45 and 58 min correspond to larger selenopeptide (2-3 kDa) fragments of the investigated selenium-containing proteins.

Figure 6. Comparison of conventional HPLC-ICPMS and capillary HPLC-ICPMS for selenopeptide mapping in solutions after the tryptic digestion of selenoprotein. Assignment of peaks (peptide sequence obtained from ref 10): (1) XGHDQSGTK, (2) XNAGR, (3) TYENXK, (4) Ac-SNXMNK, and (5) Ac-SNXXNK (X denotes selenomethionine). (a) 30-µL sample injected on a ODS2 C18 column (4.6 mm i.d. × 25 cm, 5 µm); (b) 0.2 µL of the 1+4 diluted sample injected on a Hypersil C18 BDS column (300 µm i.d. × 15 cm, 3 µm).

detection) and the collision cell mode (80Se detection) reported elsewhere, an improvement in the detection limits for selenium in aqueous solution of a factor 5 was expected.34 This discrepancy can be explained by the different plasma ionization conditions of the two instruments (ELAN 6000 and Agilent 7500c) when organic solvents are introduced. As described above (cf. Figure 3), the presence of 30% (v/v) methanol in the analyzed solution resulted in a 5-fold increase in sensitivity for the ELAN 6000 instrument and only a 2-fold increase in sensitivity for the Agilent 7500c instrument (in comparison with pure aqueous solutions). The detection limits obtained in this work are 100-150 times lower than those reported elsewhere for selenium ICPMS detection in conventional (4.6-mm i.d.) HPLC (Table 1), where the lowest absolute detection limits were between 19 and 59 pg.35-37 (34) Hinojosa Reyes, L.; Marchante Gayon, J. M.; Garcia Alonso, J. I.; Sanz-Medel, A. J. Anal. At. Spectrom. 2003, 18, 11-16. (35) Jakubowski, N.; Thomas, C.; Stuewer, D.; Dettlaff, I.; Schram, J. J. Anal. At. Spectrom. 1996, 11, 1023-1029.

6842 Analytical Chemistry, Vol. 75, No. 24, December 15, 2003

CONCLUSIONS The new sheathless interface developed for the coupling of capillary HPLC with ICPMS offers a significantly higher chromatographic resolution, peak capacity, and detection sensitivity than any of the HPLC-ICPMS couplings ever reported. High concentrations of organic solvents (up to 100%) can be tolerated by the plasma without addition of oxygen or any adverse effects, especially when a shield-torch collision cell instrument is used. The technique is attractive for detection of heteroelement (Se, S, P)-containing compounds in complex biological mixtures and is complementary to electrospray MS. The very small injection volumes at the nanoliter level allow a smaller sample intake for biological elemental speciation analysis and open the way to investigation of biochemical processes on a cellular level. The element-specific, compound-independent signal in ICPMS detection may also be the basis for selenopeptide and -protein quantification by on-line isotope dilution analysis. Received for review July 18, 2003. Accepted October 7, 2003. AC034819H (36) Gonzalez LaFuente, J. M.; Dlaska, M.; Fernandez Sanchez, M. L.; SanzMedel, A. J. Anal. At. Spectrom. 1998, 13, 423-429. (37) Yang, K.-L.; Jiang, S.-J. Anal. Chim. Acta 1995, 307, 109-115. (38) Pe´cheyran, C.; Quetel, C. R.; Lecuyer, F. M. M.; Donard, O. F. X. Anal. Chem. 1998, 70, 2639-2645. (39) Prange, A.; Schaumlo ¨ffel, D. J. Anal. At. Spectrom. 1999, 14, 1329-1332. (40) Ruiz Encinar, J.; Sliwka-Kaszynska, M.; Polatajko, A.; Vacchina, V.; Szpunar, J. Anal. Chim. Acta, in press (DOI 10.1016/S0003-2670(1003)00754-00752). (41) Sloth, J. J.; Larsen, E. H. J. Anal. At. Spectrom. 2000, 15, 669-672. (42) Casiot, C.; Donard, O. F. X.; Potin-Gautier, M. Spectrochim. Acta, Part B 2002, 57B, 173-187.