Anal Bioanal Chem DOI 10.1007/s00216-014-7781-0

NOTE

Mass spectrometric analysis of nanoscale sample volumes extracted from open microchannels after sample preconcentration applied on amyloid beta peptides Saara Mikkonen & Johan Jacksén & Åsa Emmer

Received: 20 February 2014 / Accepted: 20 March 2014 # Springer-Verlag Berlin Heidelberg 2014

Abstract A new instrumental concept for extraction of nanovolumes from open microchannels (dimensions 150 μm×50 μm, length 10 mm) manufactured on silicon microchips has been used in combination with a previously developed method for preconcentrating proteins and peptides in the open channels through electromigration. The extracted nanovolumes were further analyzed using nanoelectrospray ionization (nESI) or matrix-assisted laser desorption/ ionization mass spectrometry (MALDI-MS) directly or with subsequent enzymatic protein digestion in a nanodroplet prior to the MS analysis. Preconcentration of the samples resulted in a 15-fold sensitivity increase in nESI for a neurotensin solution, and using MALDI-MS, amyloid beta (Aβ) peptides could be detected in concentrations down to 1 nM. The method was also successfully applied for detection of cell culture Aβ. Keywords Mass spectrometry . Microchannel . Preconcentration . Amyloid beta . MALDI . Nano-ESI

Introduction In the field of clinical and bioanalytical chemistry, the available sample volumes, as well as analyte concentrations, are often very low. As a response, the development of miniaturized techniques for handling and analyzing small sample volumes has progressed remarkably during the past few decades. Although downsizing provides advantages including reduced analysis time and sample and reagent consumption, it S. Mikkonen (*) : J. Jacksén : Å. Emmer Department of Chemistry, Division of Applied Physical Chemistry, Analytical Chemistry, School of Chemical Science and Engineering, KTH Royal Institute of Technology, Teknikringen 36, 100 44 Stockholm, Sweden e-mail: [email protected]

also places higher demands on the detection sensitivity. In order to achieve an improved detectability, sample preconcentration can be performed. Numerous—either electrophoretic, such as field-amplified stacking (FAS), or extraction-based, as solid-phase extraction (SPE)—preconcentration techniques have been adapted to and used in microfluidic devices [1]. Additionally, devices with several, integrated sample pretreatment procedures have been presented. For example, Wang et al. managed to successfully perform digestion, SPE, and capillary electrophoresis (CE) separation in a single microfluidic device coupled to ESI-MS [2]. Although integration and automation possibilities are important advantages of micrototal analysis systems, these types of devices often require complicated fabrication and/or experimental procedures, limiting the potential application possibilities for a specific design. Another approach to increase the detectability is to use a more sensitive detection method. MS generally offers both high sensitivity and identification possibilities. Several interfaces between sample pretreatment procedures in microfluidic devices and matrix-assisted laser desorption/ionization (MALDI)- or ESI-MS have been proposed [3]. The coupling to ESI-MS is normally implemented on-line and most often using microchips with integrated ESI emitter tips [3]. In some cases, however, off-line handling is advantageous; an off-line system, for example, provides an additional degree of freedom regarding coupling of different procedures. In the present work, sample preconcentration based on electromigration of proteins and peptides in open microchannels [4] was coupled to nanoelectrospray ionization (nESI)- and MALDI-MS by extraction of the analyte zone during the preconcentration. To achieve this, a new instrumental concept for extracting nanodroplets from the microchannels, utilizing the accessibility of sample from the open system, was developed.

S. Mikkonen et al.

Materials and methods Neurotensin, cytochrome c (CytC) (equine heart), trypsin (bovine pancreas), 2-[N-cyclohexylamino] ethanesulfonic acid (CHES), trifluoroacetic acid (TFA), phosphate-buffered saline (PBS), bovine serum albumin (BSA), and Aβ 1-38 and 1-40 were from Sigma Aldrich (Stockholm, Sweden). Ammonium bicarbonate (NH 4 HCO 3 ) was from BDH Laboratory Supplies (Poole, England), 2,5-dihydroxybenzoic acid (DHB) from Bruker Daltonik GmbH (Bremen, Germany), fluorocarbon liquid (FC-77) from 3M (St. Paul, MN, USA), formic acid from Fluka (Steinheim, Germany), and hydrofluoric acid (HF) (50 %, VLSI Selectipur) and acetonitrile (ACN) were from Merck (Darmstadt, Germany). All water used was deionized water from a Milli-Q Synergy 185 system (Millipore Corp., Bedford, USA). The preconcentration procedure utilizing open microchannel (dimensions 150 μm×50 μm, length 10 mm, Fig. 1) silicon chips is described in detail elsewhere [4]. Shortly, the channel was filled either with a discrete sample volume (2 μL including droplets at the channel ends) or sample solution that was supplied from both channel ends at 20 μL/h. By simultaneously applying voltage (36.4 V for 1– 10 min) across the channel length, analytes formed concentrated zones in the channel through migration according to charge. Neurotensin, CytC, and Aβ 1-38 and 1-40 were used as model substances. For the preconcentration of CytC, the channel was filled with CHES buffer (20 mM, pH 8.6) prior to preconcentration. A sample of supernatants of SH-SY5Y cells overexpressing APP695 cell culture Aβ (preparation according to immunoprecipitation (IP) procedure described in [5] with 6E10 antibody, but with modified washing (three steps PBS/0.1 % BSA, two steps 50 mM NH4HCO3, and one step water) and elution (50 μL 0.5 % formic acid)) steps followed by vacuum centrifugation was also analyzed. Fig. 1 Platform for preconcentration and extraction of nanodroplets from the open microchannels. Insets: photographs of microchip and extraction devices

To couple the preconcentration to nESI-MS, a borosilicate nESI needle (ES380, Thermo Scientific, Stockholm, Sweden) was positioned into the open channel (inset in Fig. 1) and a fraction of the channel content was extracted using vacuum suction. Thereafter, the needle was transferred to the ion trap mass spectrometer (ESQUIRE, Bruker Daltonics, Bremen, Germany) equipped with an off-line nanospray interface, and a previously developed method of performing nESI-MS analysis from discrete nanolitersized sample volumes was utilized [6]. nESI-MS analysis was performed in positive mode, with scan resolution 13,000m/z/s applying 450–550 V. EsquireControl was used for instrument control and DataAnalysis (Bruker Daltonics) for evaluation of data. When the preconcentration was coupled to MALDIMS, a nanodroplet was instead extracted utilizing capillary forces. A short fused silica capillary (CM Scientific, Silsden, UK, id/od 20/150 μm, tip etched to cone shape using HF) was positioned in the channel, and applying pressurized air the extracted sample was deposited on a MALDI-plate (inset in Fig. 1). In some experiments, the plate was precovered with an FC-77 liquid to prevent evaporation of the small droplets, and enzymatic digestion was performed by the addition of digestion buffer (trypsin in ammonium bicarbonate at pH 8.3, with final buffer concentration 10 mM and an enzyme:protein ratio 1:4 for the initial protein concentration). The digestion was terminated by adding MALDI-matrix and removing the FC-77. DHB was used as matrix (10 mg/mL in 2:1 for digestion experiments and 9:1 0.1 % TFA:ACN for the Aβ-samples), and a Bruker Reflex III MALDI-TOF with a SCOUT 384 ion source and Flex control software (Bruker Daltonics) was used. Using a reflectron positive method, 500–1,000 laser pulses were collected with an intensity of 60–85 %.

Mass spectrometric analysis of nanoscale sample volumes

Fig. 2 a nESI-MS spectrum, 500 nM neurotensin, preconcentrated 1 min. b nESI-MS spectrum, 500 nM neurotensin, without preconcentration. c MALDI-MS spectrum of 1 nM CytC digest (2.5 h

in a 7.5 nL droplet) after preconcentration in the open channel and extraction from a position next to the cathode

Results and discussion

same amount of time (Fig. 2b). It could be observed that the preconcentration resulted in an approximately 15 times increased signal-to-noise ratio (S/N). The advantages of having open access to the sample were also demonstrated by performing further sample pretreatment in nanodroplets extracted from the open channels. In the resulting MALDI-spectrum after preconcentration (10 min, pump rate 20 μL/h) of a 1-nM CytC solution (pI∼10, positively charged at pH 7, 3.3 fmol supplied to channel) and digestion under a liquid lid, three peaks could be assigned to CytC (Fig. 2c, the remaining peaks were confirmed to be part of the sample matrix). By using an open system, reagents can be added to the droplets during an extended period of time, it means that the plausible reactions are not limited to fast kinetics reactions, which otherwise often is the case in chipbased systems coupled to MS [6]. In comparison to the nanodroplet digestion, a batch digestion of 1 nM CytC was performed (total volume 100 μL, same digestion buffer but enzyme:protein 1:20), resulting in only one peak

During preconcentration, analytes migrate in the channel based on their net charge. In the previous work, the analyte distribution along the length of the channel during and after preconcentration was studied [4]. Detection was performed either visually (by using colored or fluorescent substances) or by applying MALDI-matrix to the channel after preconcentration and performing MALDI-MS analysis directly within the open channel. In this work, nanodroplets were instead extracted from the channel during preconcentration and thereafter analyzed with nESI- and MALDI-MS. In Fig. 2a, the nESI-spectrum after preconcentration of 500 nM neurotensin (pI∼9.7, positively charged at pH 7) is shown. The sample solution was continuously supplied and voltage applied for 1 min, and a sample droplet was extracted from a position next to the cathode in the channel. In comparison, another emitter was manually filled with the same solution, without preconcentration, and data was collected for the Fig. 3 MALDI-MS spectra of Aβ-peptides. a 1 nM 1-38 and 1-40, preconcentrated 1 min (the remaining peaks were confirmed to belong to Aβ 1-38). b 1 nM 1-38 and 1-40, without preconcentration. c Cell culture Aβ, preconcentrated 3 min. d Cell culture Aβ, without preconcentration

S. Mikkonen et al.

corresponding to CytC (data not shown). Thus, it can be concluded that preconcentrating the protein solution and performing the digestion in a nanodroplet increased the sensitivity. In general, the described procedure is useful when a fast and simple method for controlled preconcentration of small volumes is required. As an example of such an application, immunoprecipitated cell culture samples of Aβ-peptides were analyzed. First, standards of Aβ 1-38 and 1-40 were preconcentrated in concentrations ranging from 1 μM down to 1 nM. The preconcentration was performed in an acidic solution (0.1 % formic acid), resulting in the target peptides (pIs∼5–6) being positively charged and thus becoming concentrated at the cathode. The channel was filled with sample solution, and after 1–3 min of voltage application, an 8.8-nL droplet was extracted from a position next to the cathode and deposited on a MALDI-plate. In the top half of Fig. 3, the MALDI-spectrum obtained after preconcentration of 1 nM Aβ 1-38 and 1-40 is shown in comparison to an equal volume of the same solution without preconcentration, clearly demonstrating the usefulness of the method. Next, the method was applied on a real sample of cell culture Aβ (41.8 ng/mL Aβ 140 before IP). The MALDI results (bottom half of Fig. 3) show that peptides not observed without preconcentration (in particular Aβ 1-38 and 1-40) could easily be detected after preconcentration.

Conclusion The newly developed instrumental concept and the described preconcentration procedure utilizing open channel microchips have shown to improve the detection sensitivity in both MALDI- and nESI-MS analysis of proteins and peptides. The fact that the system is open towards the surroundings combined with the simple design of the chip-based platform results in a broad versatility of application possibilities. Earlier

work has shown that the open channel can be used directly as a MALDI-target [4], and as shown in the present study, the accessibility of sample can be used to extract sample directly from the channel during preconcentration either for direct coupling to nESI- or MALDI-MS or for further sample pretreatment in nanodroplets. The platform can, for example, be used for preconcentrating samples of low concentrations of Aβ-peptides and has shown to be applicable for real sample analysis.

Acknowledgments Patrik Ek is acknowledged for the help with the experimental setup and Thomas Frisk for the chip fabrication. Ute Haußmann, Hans-Wolfgang Klafki, and Jens Wiltfang are acknowledged for the Aβ cell culture sample and IP-protocol. The Swedish Research Council is acknowledged for financial support (grant no. 621-20094095).

References 1. Giordano BC, Burgi DS, Hart SJ, Terray A (2012) On-line sample preconcentration in microfluidic devices: a review. Anal Chim Acta 718: 11–24 2. Wang C, Jemere AB, Harrison DJ (2010) Multifunctional protein processing chip with integrated digestion, solid-phase extraction, separation and electrospray. Electrophoresis 31(22):3703–3710 3. Gao D, Liu H, Jiang Y, Lin J-M (2013) Recent advances in microfluidics combined with mass spectrometry: technologies and applications. Lab Chip 13(17):3309–3322. doi:10.1039/c3lc50449b 4. Mikkonen S, Khihon Rokhas M, Jacksén J, Emmer Å (2012) Sample preconcentration in open microchannels combined with MALDI-MS. Electrophoresis 33(22):3343–3350 5. Schieb H, Weidlich S, Schlechtingen G, Linning P, Jennings G, Gruner M, Wiltfang J, Klafki HW, Knölker HJ (2010) Structural design, solidphase synthesis and activity of membrane-anchored β-secretase inhibitors on Aβ generation from wild-type and swedish-mutant APP. Chem Eur J 16(48):14412–14423 6. Ek P, Stjernström M, Emmer Å, Roeraade J (2010) Electrospray ionization mass spectrometry from discrete nanoliter-sized sample volumes. Rapid Commun Mass Spectrom 24(17):2561–2568