Rapid Isoelectric Point Determination in a ... - ACS Publications

Rapid Isoelectric Point Determination in a Miniaturized Preparative Separation Using Jet-Dispensed Optical pH Sensors and Micro Free-Flow Electrophore...
7 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

Rapid Isoelectric Point Determination in a Miniaturized Preparative Separation Using Jet-Dispensed Optical pH Sensors and Micro FreeFlow Electrophoresis Christin Herzog,† Erik Beckert,‡ and Stefan Nagl*,† †

Institut für Analytische Chemie, Universität Leipzig, Linnéstrasse 3, 04103 Leipzig, Germany Fraunhofer-Institut für Angewandte Optik und Feinmechanik (IOF), Albert-Einstein-Strasse 7, 07745 Jena, Germany

‡

S Supporting Information *

ABSTRACT: Herein, the fabrication, characterization, calibration, and application of integrated microfluidic platforms for fast isoelectric point (pI) determinations via free-flow electrophoresis with integrated inkjet-printed fluorescent pH sensor microstructures are presented. These devices allow one to determine the pI of a biomolecule from a sample mixture with moderately good precision and without addition of markers in typically less than 10 s total separation and analysis time. Polyhydroxyethyl methacrylate (pHEMA) hydrogels were covalently coupled with fluorescein and hydroxypyrene trisulfonic acid (HPTS)-based pH probes. These were piezoelectrically jet-dispensed onto acrylate-modified glass as pH sensor microarrays with a diameter of 300−600 μm and thicknesses of 0.4−2.4 μm with high spatial accuracy. Microchip fabrication and integration of these pH sensor arrays was realized by multistep liquid-phase photolithography from oligoethylene glycol precursors resulting in glass-based microfluidic free-flow isoelectric focusing (μFFIEF) chips with integrated pH observation capabilities. The microchips were characterized with regard to pH sensitivity, response times, photo-, and flow stability. Depending on the sensor matrix, they allowed IEF within a pH range of roughly 5.5−10.5 with good sensitivity and fast response times. These microchips were used for FFIEF of small molecule markers and several protein mixtures with simultaneous monitoring of local pH. This allowed the determination of their pI via multispectral imaging of protein and pH sensor fluorescence without addition of external markers. Obtained pI’s were generally in good agreement with known data, demonstrating the applicability of the method for pI determination in micropreparative procedures within a time frame of a few seconds only.

T

values for relatively simple molecules such as small peptides without a pronounced tertiary structure. The predominant way to determine the pI is via electrokinetic techniques such as electrophoresis, electroosmosis, electroacoustics, or other measurement modalities.1 Among those, electrophoretic procedures are particularly widespread and routinely applied in many laboratories. In isoelectric (IEF) focusing, analytes are separated by their isoelectric point (pI) in a pH gradient, typically established by ampholytes. IEF may be applied in capillary, gel, or free-flow electrophoresis and is very often used as one dimension in multidimensional separations, for example, in 2D gel or capillary electrophoresis for proteomics applications.4−6 In combination with mass spectrometry (MS), powerful biomolecular analyte separation, quantitation, and identification platforms can be employed.7

he isoelectric point (pI, also IEP) is the point of no net charge of a molecule, coordination compound, aggregate, surface, or any other assembly. To understand the behavior of compounds on a molecular scale, knowledge of their surface charge at different pH and therefore their isoelectric point is essential. Among many other fields, this is of particular importance in biochemistry and molecular biology, where charges present in the building blocks of life such as amino acids, peptides, proteins, enzyme cofactors, nucleotides, their subunits, and polymers determine essential functionality, such as solubility and reactivity. The pI may be determined, in the simplest case, by measurement of the natural pH of solutions or dispersions of the compound,1 but this requires highly pure substances and defined conditions that are often difficult to achieve, particular for biological composites. Because the pI of a molecule is determined by the acid and base constants (pKa and pKb) of their constituents, therefore calculation of the isoelectric point has been done and known for a very long time.2 In recent years, automated computational methods and tools have been made available.3 However, so far this approach yields only accurate © 2014 American Chemical Society

Received: May 5, 2014 Accepted: September 11, 2014 Published: September 11, 2014 9533

dx.doi.org/10.1021/ac501783r | Anal. Chem. 2014, 86, 9533−9539

Analytical Chemistry

Article

incompatible with respect to solvents and viscosity, or the fabrication process is incompatible with the microchip production process. Thete et al. presented photopolymerized hydrogel-based microarrays fabricated by a piezoelectric drop on demand technique and integrated these into glass-based microfluidic platforms for sensing liquid analytes for pH and solvent polarity,27 while Abe et al. integrated inkjet-printed chemical sensing into a filter-paper microfluidic channel structure for fast pH, glucose, and protein assays.28 In previous work,29 we presented the direct observation of the pH gradient in the separation bed of a μFFIEF platform by an integrated photopolymerized pH sensor layer. This approach circumvented indirect subsequent offline analytical methods and allowed direct monitoring of separation performance and isoelectric point and therefore potentially analyte identification. Yet the approach necessitated the use of photopolymer-based sensor matrixes with a limited set of usable pH probes and generated relatively thick sensor layers of several tens of micrometers resulting in only moderate response times of around 30 s that limited their use for fast imaging of separation parameters and isoelectric points. Herein, we present a novel straightforward approach for FFIEF chips with integrated fast-response pH sensors with different working pH ranges and apply these for monitoring of isoelectric points in model compounds and proteins over a wide pH range. Polyhydroxyethyl methacrylate (pHEMA) hydrogels were covalently modified with fluorescent pH probes and used for automated piezoelectric printing of microsized pH-sensitive structures on glass. FFIEF microchip fabrication using these substrates was carried out via a two-step liquidphase photopolymerization method in a modification of earlier procedures29−31 to integrate a jet-dispensed pH sensor field inside the separation bed of a μFFIEF chip. The microchip was used for separation of proteins and small molecule markers with integrated online monitoring of the development, respectively, the stability of the established pH gradient. Therefore, precise online quantification of separation parameters and isoelectric points throughout the microfluidic separation procedure could be realized.

These separation procedures were miniaturized onto microchip-based platforms for faster analysis, smaller sample amounts, better resource efficiency, and more system integration using microfluidic chips that subsequently perform several analytical or preparative steps.8,9 Microfluidic free-flow electrophoresis (μFFE) is a method for continuous separation and preparation of minute amounts of biomolecules under mild conditions,10 and miniaturized free-flow isoelectric focusing (μFFIEF) platforms were realized in various formats and for several applications,11−17 with a focus on the separation of proteins and other (poly)-amino acids. In principle, IEF is able to both separate compounds from mixtures and determine their pI, but in practice this analysis is mostly done via addition of markers of defined pI’s that allow the setup of a calibration curve. While being a successful approach for many multidimensional separations, this procedure is limited particularly in miniaturized systems in that it is invasive, and it adds external compounds to the analyte mixture, which is a particular restriction for preparative or semipreparative procedures such as free-flow electrophoresis. It is also subject to experimental deviations because differences in temperature, solvent composition, or other environmental factors that exert an influence on the pI may not reflect on these marker molecules in the same way as for analyte compounds.18 Miniaturized free-flow electrophoretic platforms are particularly attractive because of their mild separation conditions enabling one to retain proteins or other compounds in their native conformation and because of their continuous operational mode that is attractive for seamless integration into other procedures such as flow chemical transformations or biological response screenings. It is, however, difficult to realize narrow and baseline separated bands with a chromatographic resolution >1.5, because several band broadening effects influence the separation quality negatively: the width of the injected sample stream, hydrodynamic broadening, electrodynamic broadening, and electromigrational dispersion. In the current state of development, μFFIEF platforms are reported to yield a resolution of 0.2−0.4 pH units depending on experimental conditions.13,15 Optical chemical sensors offer fast, sensitive, selective analysis and real-time monitoring of different analytes and parameters. In particular, fluorescent pH sensors have proven vital for pH monitoring in bioprocesses and biotechnology, cell biology, medical, and environmental research fields.19 Fluorescent pH sensor matrixes have been integrated into miniaturized microfluidic devices in various forms, as monolayers on channel surfaces,20 via photopolymerization,21−23 as bead arrays in microchannels24 or beads steered by optical traps,25,26 or monolayers and used primarily for cell culture,27 and environmental water analysis.28 Inkjet printing has been known for almost 150 years and nowadays may be regarded as a family of several printing techniques that propel a droplet from a reservoir onto a surface, typically in a noncontact manner. Often piezoelectric elements are used for automated high speed droplet generation in several modes. Inkjet printers are common in many industrial and consumer applications but have rarely been used for microfluidic optical sensor fabrication, although these types of chemical sensors are typically produced using a coating solution that is applied on a solid substrate through various methods such as blade, spin, or spray coating. Yet for the majority of inkjet techniques, these formulations or dispersions are largely

■

MATERIALS AND METHODS Details on chemicals and materials and experimental protocols for the preparation of pH-probe conjugated polymers, protein labeling, fluorescence microscopy, characterization of pH sensors, and image analysis are available in the Supporting Information. Inkjet Printing of Probe-Doped Polymers. Microscope glass slides (Roth) were cleaned, dried, and treated for 5 min with a 0.16% (v/v) solution of (3-methacryloyloxypropyl)trichlorosilane (TPM) in n-heptane and trichloromethane (4:1, v/v). Finally, the slides were rinsed with n-heptane. Fluorescent pH-sensitive polymer matrixes were applied on a TPM-coated glass slide at 70 °C. The print assembly consisted of a valve MV 100 with needle adapter and driver DCON 100 using needles with an inner diameter of 100 μm and an outer diameter of 200 μm (Picodostec, Germering, Germany) arranged to desktop robot DR2203N (Nordson EFD, Pforzheim, Germany). For the generation of spot array at each point, the valve opened for 2 ms and the robot stop for 50 ms, and for rows a robot speed of 20 mm s−1 and a valve opening time of 2 ms were used. The height and width of sensor structures were determined via 9534

dx.doi.org/10.1021/ac501783r | Anal. Chem. 2014, 86, 9533−9539

Analytical Chemistry

Article

contact surface profilometry with a Dektak 3030 apparatus (Veeco, Mannheim, Germany). Photopolymerization of the Microfluidic Assembly. μFFE chips with inkjet-printed sensors were produced via photopolymerization of oligoethylene glycol diacrylate (OEGDA) using modifications of a procedure previously described for microfluidic chip and μFFE chip production.29−31 Cavities were generated with 50 μm granularity using a Point II powder blaster (Barth, Königsbach-Stein, Germany) into microscope glass slides to prepare cover plates. 70 μL of OEG-DA with 1% (w/w) 2,2-dimethoxy-2-phenylacetophenone was applied on a nonsilanized glass slide as a bottom plate, and a TPM-coated cover glass slide that contained powder-blasted access holes was put on top of the polymer while avoiding air bubble formation between the slides. A photomask containing the μFFE structure on polyethylenetherephthalate foil was produced via offset print at 3600 dpi (DTP-System-Studio, Leipzig, Germany) and applied on top of the assembly. Photopolymerization was carried out with a 4 in. FE5 Flood Exposure (Süss MicroTec, Munich, Germany) equipped with a Hg lamp (14 mW cm−2 at 365 nm) at an illumination time of 1.2 s. The nonpolymerized prepolymer was removed in a partial vacuum. The bottom plate was replaced by a silanized glass slide containing the sensor assembly. For bonding of this slide with the microfluidic structure, soft pressure was manually applied to the glass slides, and they were illuminated again at an exposure time of 1.2 s using the above Flood Exposure instrument. Characterization of pH Sensors. The microchip was positioned on an IX 70 epi-fluorescence microscope (Olympus, Hamburg, Germany) equipped with an Hg arc lamp and a ProEM 512 CCD camera (Princeton Instruments, Warstein, Germany) for fluorescence detection. The pH dependency of integrated sensor layer was measured via flushing the microchip with britton robinson buffers (BRB) 40 mM of different pH (2−12) at a flow rate of 50 μL min−1 and detection of the fluorescence intensity. Response times were determined by flushing the microchip at the same conditions periodically with pH 3 and pH 10 BRBs. For the flow stability test, the microchip was flushed with pH 10 BRB and the fluorescence intensity was observed. A permanent illumination at λexc 420−480 nm using the Hg lamp was employed to investigate the photostability of the sensor layer. Microfluidic Free-Flow Isoelectric Focusing (μFFIEF). To generate flow in the separation area, neMESYS syringe pumps (cetoni, Korbussen, Germany) were applied. The generation of liquid flow in the electrode channels was done by a syringe pump PHD 22/2000 (Harvard Apparatus, Holliston, MA). The electric field was induced by a HCL 356500 power supply (FuG Elektronik, Rosenheim, Germany) in the negative mode. The microscopic setup was as described in pH sensor characterization. For FFIEF, the sample solutions were injected into the chip through the center inlet and focused by flanked ampholyte flows, respectively, sheath flows, and electrode channels (Figure 1). The cathodic flow consisted of 40 mM BRB pH 10 (8.2 mS cm−1) and the anionic flow 40 mM BRB pH 3 (2.3 mS cm−1). Analytes were separated in an electrolyte containing 2% ampholytes and 0.1% Tween 20 in water (590 μS cm−1). The small molecule IEF markers were diluted 1:10 with water to obtain a concentration of 200 mg L−1. Proteins were used undiluted in the concentrations described in the Supporting Information. The parameters for the separation of IEF markers were 1.13 and 1.93 mm s−1 linear velocity and 70 and 170 V

Figure 1. Illustration of experimental setup for μFFIEF separation. Seven inlets for electrolytes, sheath flow, ampholytes, and analytes.

resulting in a current of 40 and 140 μA, respectively. For the separation of protein mixtures, a linear velocity between 1.30 and 3.56 mm s−1 and a voltage of 580−1700 V with an observed current between 280 and 540 μA were established. The corresponding values for each separation are depicted in the respective figure caption.

■

RESULTS AND DISCUSSION Synthesis of pHEMA-Derivatives with Covalently Bound Fluorescent pH Probes. In our work, we have focused on polyhydroxyethyl methacrylate (pHEMA)-based pH sensor matrixes for microfluidic chip integration, because of their high proton permeability, good mechanical stability, and strong adhesion to glass substrates. However, physical immobilization of pH probes into this matrix, in the form of nanoparticles or water-insoluble probes, did not yield satisfactory results for the application because of low signal intensities or insufficient retention of probes in the polymer matrix. Therefore, we attached the pH-sensitive fluorochromes fluorescein and hydroxypyrene trisulfonic acid (HPTS) covalently to pHEMA using a previously described (FITCpHEMA)32 or a combination of earlier described synthetic procedures (HPTS-Amino-pHEMA)33,34 for coupling of functional groups to pHEMA (see the Supporting Information). Fluorescein was attached to pHEMA via its reactive isothiocyanate derivative (FITC) in N,N-dimethylacetamide (Supporting Information, Scheme 1a). HPTS was activated to its 1,4,6-trisulfonyl chloride and bound in basic environment to an amino-modified pHEMA (Amino-pHEMA) that was produced via copolymerization of 2-hydroxyethyl methacrylate (HEMA) with N-(3-aminopropyl)-2-methylacrylamide (Supporting Information, Scheme 1b). Inkjet Printing of pH Sensors on Glass. The jet dispensing technique uses valve reservoirs that are piezoelectrically actuated and allow droplet generation in defined volumes on a wide array of solutions and substrates. The system works with different dosing needles and can be applied for production of various shapes with diverse proportions. Herein, we evaluated both sensor spot arrays and line-shaped sensor structures for pH determination in FFIEF. The displayed exemplary rows were 7 mm in length, 388 ± 58 μm in width, and 782 ± 461 nm in height (Figure 2a). We also realized pH sensor spots with a diameter of 553 ± 83 μm and a height of 2.4 ± 1.0 μm in a spot array of 11 × 3 functional elements in an area of 18.75 mm2 (Figure 2b). This allowed monitoring of the pH at defined points in the separation bed. The size of the sensor structures shows a strong dependency on polymer matrix and the viscosity of the printed solution. Therefore, several polymer concentrations from 1% to 5% were printed. For this concentration range, we found that the higher is the polymer amount, the better is the printing performance, possibly because of the higher viscosity of the solution. A 9535

dx.doi.org/10.1021/ac501783r | Anal. Chem. 2014, 86, 9533−9539

Analytical Chemistry

Article

structure (Figure 3e) was fixed on the cover slide with the microfluidic structure (Figure 3f). In the last step, the assembly was bonded together irreversibly using pressure and another UV illumination (Figure 3g) applying a light-impermeable mask covering the separation bed to prevent bleaching of the sensor matrix during photopolymerization. Via this procedure that does not expose the sensor matrix to light or reagents of the photopolymerization, we could integrate these sensors into a moderately complex microfluidic structure for free-flow isoelectric focusing (μFFIEF) in a relatively straightforward way. Characterization of Microchip-Integrated InkjetPrinted pH Sensors. The characteristics of integrated pH sensors were investigated using 40 mM BRBs. Both sensors showed a pronounced pH dependency switching from almost no fluorescence in the acidic range to strong green fluorescence in the alkaline range. The pH area covered with these sensors was different as sigmoidal fits determined a pKa 6.92 ± 0.04 for FITC-pHEMA and 9.15 ± 0.06 for HPTS-Amino-pHEMA (Figure 4a). Because these matrixes can be applied with good

Figure 2. Representative false-colored fluorescence images of pH sensor microstructures in BRB pH 10 with blue light excitation: (a) pH sensor rows and (b) pH sensor array (probes covalently bound to pHEMA on glass).

concentration of 5% probe-coupled polymer in an ethanol/ water mixture (9:1 w/w) produced the best homogeneity, shape, and reproducibility. Fabrication of Microfluidic FFE Chips with Integrated Inkjet-Printed pH Sensors. For fabrication of microfluidic free-flow electrophoresis structures and integration of pH sensor matrixes therein, we used a multistep photopolymerization procedure that is schematically shown in Figure 3. During

Figure 4. Characterization results of FITC-pHEMA (left) and HPTSAmino-pHEMA (right) are presented at λexc 420−480 nm, λem 510− 560 nm. (a) pH dependence of the fluorescence intensity of integrated sensor structures over a pH range from 2 to 12: FITC-pHEMA, determined pKa 6.9 and HPTS-Amino-pHEMA, pKa 9.2. (b) pH response: blue arrows, change to pH 10; green arrows, change to pH 3 (HPTS-Amino-pHEMA: S-G smoothed with 600 points).

Figure 3. Fabrication of μFFIEF chips with integrated inkjet-printed pH sensors, schematic illustration. (a) Deposition of OEG-DA including photoinitiator between silanized cover glass plate and nonTPM coated bottom glass slide. (b) UV exposure using a photomask containing the microfluidic structure. (c) Removal of nonpolymerized OEG-DA. (d) Removal of the bottom plate. (e) Attachment of a spotted and TPM-coated glass slide containing the inkjet-printed sensor structure to the assembly. (f) UV exposure using a lightimpermeable mask. (g) A real-color fluorescence image with UV excitation of a μFFIEF microchip with a fluorescent pH sensor row.

sensitivity in the range of roughly pKa ± 1.5, these sensors allowed determining IEPs from around 5.5 to 10.7. Calibration curves were very reproducible over time and fabrication charge. No significant aging effect and differences of manufacturing charges could be noticed over a time span of several weeks when the microchips were stored dry in the dark (Supporting Information Figure S-1). The covered pH range could be extended by coupling other pH-sensitive chromophores to pHEMA using the procedures outlined above. Response times of sensor structures are critical to allow fast monitoring and highly dependent upon ion permeability of the sensor matrixes. Using the above hydrogel sensors with covalently linked fluorophores, we obtained very fast response times in both directions for FITC-pHEMA and HPTS-AminopHEMA (t95 acidic → alkaline 1.6 s, 4.4 s, respectively, t95 alkaline → acidic 0.3 s, 1.8 s; Figure 4b). Response in the direction from alkaline to acidic conditions was faster than that in the other direction. FITC-pHEMA showed a faster response, presumably due to better proton permeability within the unmodified uncharged pHEMA. In comparison to our previous work using a photopolymerized fluorescein-dextran pH sensor

the first part of the fabrication, a microfluidic structure was prepared between a nonsilanized bottom glass plate and an acrylic silanized cover glass slide containing powder-blasted access holes. Oligoethylene glycol diacrylate (OEG-DA) was applied as a thin layer on the silanized slide, and the second slide was aligned on the bottom of the former slide (Figure 3a). After introduction of a photomask containing the chip layout and UV photopolymerization (Figure 3b), nonpolymerized liquid residues were removed under reduced pressure resulting in a microfluidic FFE structure with a height of around 20 μm (Figure 3c).29,31 Herein, this structure was only incompletely cured at 1.2 s illumination time to enable the abstraction of the nonsilanized bottom plate that did not covalently bind to the acrylic prepolymer. For integration of jet-dispensed pH sensors on glass, this slide was manually removed (Figure 3d), and a TPM-silanized glass slide containing an printed pH sensor 9536

dx.doi.org/10.1021/ac501783r | Anal. Chem. 2014, 86, 9533−9539

Analytical Chemistry

Article

Table 1. Characterization of the Two Microchip Integrated pH Sensor Systems: Photostability, Retention, Response Time, pKa, and Approximate Working Range

FITC-pHEMA HPTS-AminopHEMA

photobleaching (% s−1)

leaching of sensor matrix (% s−1)

t95 pH 3→pH 10 (s)

t95 pH 10→pH 3 (s)

pKa

approximate working range

0.028 0.005

0.000 0.001

1.6 4.4

0.3 1.8

6.92 ± 0.04 9.15 ± 0.06

5.5−8.5 7.7−10.7

layer for IEF,29 these inkjet-printed microchip pH sensors showed a response that was about 1 order of magnitude faster, presumably mostly due to a much lower layer thickness on the order of 1 μm. This nicely demonstrates the potential and applicability of inkjet-printed pH sensors in IEF and other applications. Photo- and flow stability of the sensor structures in microchips were investigated and are summarized along with other parameters in Table 1. These chip-integrated pH sensors were very stable against a hydrodynamic flow at a flow rate of 1 mm s−1, and washing out of sensor matrixes could not be detected at a significant rate (Supporting Information Figure S2a). Photostability was examined using blue illumination with a 2.5× objective and is summarized in Table 1. FITC-pHEMA showed only moderate photostability with a decrease of 0.028% s−1, which indicated that the matrixes may only be used less than an hour under constant illumination. However, drifts in the relative response to pH were not observed within this time frame. HPTS-Amino-pHEMA showed a much better photostability and could be used for several hours without a strong decrease in fluorescence intensity (Suppporting Information Figure S-2b). However, because the matrixes only need to be used for a time frame of typically less than a minute for a single IEP determination, both matrixes showed acceptable performance. μFFIEF with FITC-pHEMA for pI Determination in the Physiological pH Range. The combined microchips were used for isoelectric point determinations, and their performance was investigated first using small molecule markers of known pI. Separation of IEF markers could be successfully realized with printed FITC-pHEMA sensor rows and is shown in Figure 5a. The markers could be separated with a good chromatographic resolution R (Supporting Information) between 0.6− 1.4 at a separation time of 7.5 s and were observed with mercury lamp UV excitation. The IEF of pI markers showed a performance comparable to previously described μFFIEF platforms.13,15 The integrated pH sensor was observed simultaneously in an individual spectral channel. Fluorescence intensity showed the pH gradient in the separation bed and a stable IEF at the conditions depicted below. The pI’s calculated from sensor intensities show a good accordance to the expected pI’s, especially in the most sensitive range with pI of the markers at 5.5, 7.2, and 7.6 deviating 0.0−0.5 pH units. The deviation of 0.5 of the pI marker 7.2 could be caused by its low fluorescence intensity and therefore complex peak center determination, which is important for the reliable calculation of pI. The markers of pI 4.0 and 9.0 that were outside of this region showed a higher discrepancy of 1.0 and 0.7 pH units, respectively (Table 2). pH gradient calibration using pI markers and the herein introduced method using immobilized pH sensor system layers were compared and plotted (Figure 5b). It is immediately visible that an integrated sensor layer is able to give a much more detailed view of the pH gradient within its dynamic range

Figure 5. IEF of small molecule markers and protein (fluorescently labeled with P503) mixtures with online pH monitoring via FITCpHEMA sensor rows (a,d) and array (c). Top: False-colored fluorescence image of the analyte channel. Bottom: Electropherogram and corresponding pH readout. (a) Small molecule markers (linear velocity, 1.13 mm s−1; electric field, 170 V, 140 μA), (b) comparison of obtained pH via immobilized pH sensor and pI markers (extract from (a)), (c) the proteins bovine serum albumin (BSA), conalbumin, and chymotrypsin (linear velocity, 3.49 mm s−1; electric field, 1700 V, 280 μA; electropherogram, 2 points baseline correction), and (d) proteins bovine serum albumin (BSA), conalbumin, and chymotrypsin (linear velocity, 3.49 mm s−1; electric field, 1700 V, 280 μA; electropherogram, 2 points baseline correction).

Table 2. pI Determination Results of μFFIEF Separated Analytes with FITC-pHEMA analyte IEF marker 4.0 IEF marker 5.5 IEF marker 7.2 IEF marker 7.6 IEF marker 9.0 α-lactalbumin type III (bovine milk) bovine serum albumin conalbumin type I (chicken egg white) myoglobin (horse heart) α-chymotrypsin type II (bovine pancreas)

literature pI

pI (μFFIEF)

deviation to lit.

4.0 5.5 7.2 7.6 9.0 4.5−4.836

5.02 5.75 6.71 7.56 8.32 4.83

± ± ± ± ± ±

0.23 0.17 0.14 0.16 0.13 0.04

+1.0 +0.3 −0.5 +0.0 −0.7 +0.0

4.7−4.937,38 6.0−6.339

5.16 ± 0.34 5.99 ± 0.33

+0.2 +0.0

6.5−7.237 8.4−8.837,39

7.04 ± 0.06 7.62 ± 0.33

+0.0 −0.8

in the separation bed. Good analytical performance can also be demonstrated quantitatively using the results at the location of the pI markers 5.5, 7.2, and 7.6. These pI markers are within the dynamic range of the FITC-pHEMA sensor layer. The rootmean-square sum of pH deviations (n = 4) showed a value of 9537

dx.doi.org/10.1021/ac501783r | Anal. Chem. 2014, 86, 9533−9539

Analytical Chemistry

Article

Alkaline pI protein separations could be also realized with real time identification. Myoglobin, horse cytochrome C, and bovine cytochrome C were separated and quantified. The separation could be realized within 6.5 s. The separation of cytochrome C molecules shows only a chromatographic resolution of 0.06 caused by their rather similar pI in a matrix with a pH gradient from 3 to 10. However, the cytochrome C proteins were inside the working range of the sensor and showed a good agreement of calculated pI based on the AminopHEMA linked HPTS intensity with the literature reported pI’s (Figure 6b, Table 3).

0.3 pH units using the calibration based on the sensor layer and 0.6 pH units for a fit by the pI markers only. These multifunctional microchips were employed for separation as well as pI determination of proteins labeled with the fluorescent pyrylium-based dye P50335 and on-chip observation of the pH gradient. Bovine serum albumin (BSA), conalbumin, and chymotrypsin (Figure 5b) were separated using a sensor spot array for pI determination and lactalbumin and myoglobin using a microchip sensor row (Figure 5c). These proteins were separated within 2.4 s with a chromatographic resolution of 0.7, 0.8, respectively, 1.0. Both sensor structures worked well, but sensor rows proved to be better suited for the analysis as they allowed extraction of the pH over the whole width of the separation bed, only limited by the image resolution of CCD camera, while the array was limited to its number of independent spots. Nevertheless, whatever sensor shape was present, with a sigmoidal fit over the separation bed the pH could be calculated and made a quantification of analytes possible with only small deviations of no more than 0.2 pH units in pI quantification. An exception was the protein αchymotrypsin type II from bovine pancreas that showed a deviation of 0.8 pH units from literature values (Table 2). The sensor spot array is somewhat limited as the spot size of around 500 μm led to a lower resolution as compared to the sensor rows and therefore lesser accuracy in pI determination. μFFIEF with HPTS-Amino-pHEMA for pI Determination in the Alkaline Range. The online observation of pH could be also realized using the HPTS-Amino-pHEMA sensor matrix and was tested first with small molecule markers (Figure 6a). These molecules are beneficial for testing workwise

Table 3. Summary of pI Determinations via μFFIEF of Separated Analytes with HPTS-Amino-pHEMA analyte

literature pI

pI (μFFIEF)

deviation to lit.

IEF marker 9.0 cytochrome C (horse) cytochrome C (bovine)

9.0 9.0−9.437 10.4−10.837

9.51 ± 0.05 9.48 ± 0.02 10.35 ± 0.01

+0.5 +0.1 +0.0

In prior work on macro- and micro-FFIEF, a linear pH gradient of the separation bed was often presumed. Our multifunctional microchip enabled the separation and pI identification less dependent on the shape of the pH gradient, because it enabled the observation of the spatially resolved pH.

■

CONCLUSION Rapid isoelectric point determination within a few seconds is enabled by integration of fluorescent pH sensor matrixes into miniaturized electrophoresis platforms. In this work, the feasibility and applicability of jet-dispensed fluorescent pH sensors in microfluidic free-flow electrophoresis chips is demonstrated. The sensor structures were produced by a piezoelectric microprinter, which enabled accurate and spatially resolved deposition of sensor areas that were integrated into functional microfluidic chips using multiple photopolymerization steps. The fabrication of these multifunctional microchips is straightforward and could be realized in a few hours in a regular laboratory environment. The microchips were used for isoelectric focusing of model small molecule and protein compounds with different sensor matrixes and demonstrated good applicability for isoelectric point determinations within a time scale of only a few seconds. An even less invasive separation of bioanalytes on the microscale could be accomplished by integration of switchable pH actuators for pH gradient generation in μFFIEF instead of ampholyte mixtures.40 Further research will be aimed at using these methods in pI determinations in proteomics and other areas. Also, integrated pH sensors could be used to enhance the capabilities of analytical separation methods such as (miniaturized) capillary IEF and gel IEF. Extending the working range and improving the accuracy of these sensors could be achieved by integrating multiple sensors in a single microchip, ratiometric sensor readout, and combination of this method with μFFE assemblies that can perform the labeling step within a microchip,31b or label-free detection of biomolecules using deep UV fluorescence.41 Performing all of these analytical steps on a single microchip with integrated sensors promises to enable isoelectric point determinations of unknown samples within less than a minute of total preparation and analysis time.

Figure 6. IEF of small molecule markers and proteins (fluorescently labeled with P503) with online pH observation via HPTS-AminopHEMA sensor rows. Top: False-colored fluorescence image of the protein channel. Bottom: Electropherogram and corresponding pH readout. (a) Small molecule markers (linear velocity, 1.93 mm s−1; electric field, 70 V, 40 μA; electropherogram, 7 points baseline correction, S-G smoothed with 20 points) and (b) proteins myoglobin, cytochrome C from horse, and cytochrome C from bovine (linear velocity, 1.30 mm s−1; electrical field, 580 V, 540 μA; electropherogram, 4 points baseline correction, S-G smoothed with 20 points).

performance because of fast osmotic mobility due to their small size and therefore better separation than the larger proteins. The markers could also be separated well in HPTSbased pH sensor FFIEF microchips within 4.4 s with a chromatographic resolution of about 1.0 with the exception of 0.7 for IEF markers 7.2 and 7.6. However, only the marker with pI 9.0 could be determined by this sensor with a deviation of 0.5. Yet because this isoelectric point is on the edge of the range of the former sensor system, this experiment demonstrates the possibility of adjusting the working range by choice of the sensor matrix. 9538

dx.doi.org/10.1021/ac501783r | Anal. Chem. 2014, 86, 9533−9539

Analytical Chemistry

■

Article

Sens. Actuators, B 2006, 114, 984−994. (d) Baldini, F.; Giannetti, A.; Mencaglia, A. A. J. Biomed. Opt. 2007, 12, 024024. (e) Ray, A.; Koo Lee, Y.-E.; Epstein, T.; Kim, G.; Kopelman, R. Analyst 2011, 136, 3616−3622. (f) Vallejos, J. R.; Uplekar, S.; da Silva, J. F.; Brorson, K. A.; Moreira, A. R.; Rao, G. Biotechnol. Bioeng. 2012, 109, 2790−2797. (20) Sebra, R. P.; Kasko, A. M.; Anseth, K. S.; Bowman, C. N. Sens. Actuators, B 2006, 119, 127−134. (21) (a) Lee, S.; They, B. L.; Cote, G. L.; Pishko, M. V. Sens. Actuators, B 2008, 128, 388−398. (b) Lee, K. S.; Boccazzi, P.; Sinskey, A. J.; Ram, R. J. Lab Chip 2011, 11, 1730−1739. (22) Czugala, M.; Gorkin, R., III; Phelan, T.; Gaughran, J.; Curto, V. F.; Ducrée, J.; Diamond, D.; Benito-Lopez, F. Lab Chip 2012, 12, 5069−5078. (23) (a) Brigo, L.; Carofiglio, T.; Fregonese, C.; Meneguzzi, F.; Mistura, G.; Natali, M.; Tonellato, U. Sens. Actuators, B 2008, 130, 477−482. (b) Natali, M.; Begolo, S.; Carofiglio, T.; Mistura, G. Lab Chip 2008, 8, 492−494. (24) Klauke, N.; Monaghan, P.; Sinclair, G.; Padgett, M.; Cooper, J. Lab Chip 2006, 6, 788−793. (25) Maruyama, H.; Arai, F.; Fukuda, T. Lab Chip 2008, 8, 346−351. (26) Mela, P.; Onclin, S.; Goedbloed, M. H.; Levi, S.; Garcia-Parajo, M. F.; van Hulst, N. F.; Ravoo, B. J.; Reinhoudt, D. N.; van den Berg, A. Lab Chip 2005, 5, 163−170. (27) Thete, A.; Gross, G.; Henkel, T.; Koehler, J. Chem. Eng. J. 2008, 327−332. (28) Abe, K.; Suzuki, K.; Citterio, D. Anal. Chem. 2008, 80, 6928− 6934. (29) Jezierski, S.; Belder, D.; Nagl, S. Chem. Commun. 2013, 49, 904− 906. (30) Beebe, D. J.; Moore, J. S.; Yu, Q.; Liu, R. H.; Kraft, M. L.; Jo, B.H.; Devadoss, C. Proc. Natl. Acad. Sci. U.S.A. 2000, 9, 13488−13493. (31) (a) Jezierski, S.; Gitlin, L.; Nagl, S.; Belder, D. Anal. Bioanal. Chem. 2011, 401, 2651−2656. (b) Jezierski, S.; Tehsmer, V.; Nagl, S.; Belder, D. Chem. Commun. 2013, 49, 11644−11646. (32) Parker, J. W.; Laksin, O.; Yu, C.; Lau, M.-L.; Klima, S.; Fisher, R.; Scott, I.; W. Atwater, B. Anal. Chem. 1993, 65, 2329−2334. (33) (a) Badey, B.; Boullanger, P.; Domardd, A.; Crosb, P.; Delairb, T.; Pichot, C. Macromol. Chem. Phys. 1997, 945−957. (b) Tian, Y.; Su, F.; Weber, W.; Nandakumar, V.; Shumway, B. R.; Jin, Y.; Zhou, X.; Holl, M. R.; Johnson, R. H.; Meldrum, D. R. Biomaterials 2010, 31, 7411−7422. (c) Moad, G.; Moad, C. L.; Krstina, J.; Rizzardo, E.; Berge, C. T.; Darling, T. R. Polymer synthesis. International patent WO 9615157A1. (34) (a) Aller, R.; Zhu, Q. Optical pH sensor. U.S. patent US2006088722 (A1). (b) Offenbacher, H.; Wolfbeis, O. S.; Fürlinger, E. Sens. Actuators 1986, 9, 73−84. (35) Wetzl, B. K.; Yarmoluk, S. M.; Craig, D. B.; Wolfbeis, O. S. Angew. Chem., Int. Ed. 2004, 43, 5400−5402. (36) Phillips, G. O.; Williams, P. A. Handbook of Hydrocolloids, 2nd ed.; Woodhead: Cambridge, 2009; p 313. (37) Righetti, P. G.; Caravaggio, T. J. Chromatogr., A 1976, 127, 1− 28. (38) Ang, W.; Eilimelech, M. J. Membr. Sci. 2007, 296, 83−92. (39) Righetti, P. G.; Tudor, G.; Ek, K. J. Chromatogr., A 1981, 220, 115−194. (40) Cheng, L.-J.; Chang, H.-C. Lab Chip 2014, 14, 979−987. (41) Koehler, S.; Nagl, S.; Fritzsche, S.; Belder, D. Lab Chip 2012, 12, 458−463.

ASSOCIATED CONTENT

S Supporting Information *

Materials, experimental setup and procedures, and additional data on pH probe synthesis, sensor characterization, and electrophoretic separations. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Tel.: +49 341 97 36066. Fax: +49 341 97 36115. E-mail: [email protected]. Website: http://www.uni-leipzig. de/~nagl/. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work is supported by the Deutsche Forschungsgemeinschaft (DFG, NA 947/1-1, 1-2). The assistance of Prof. M. Lorenz and G. Ramm from the Physics Department, University of Leipzig, for surface profiling is gratefully acknowledged. We thank Prof. O. S. Wolfbeis (Universität Regensburg) for donation of P503 label.

■

REFERENCES

(1) Kosmulski, M. Surface Charging and Points of Zero Charge; CRC Press: Boca Raton, FL, 2009. (2) Levene, P.; Simms, H. S. J. Biol. Chem. 1923, 55, 801−813. (3) (a) ExPASy Bioinformatics Resource Portal, Compute pI/Mw tool, http://web.expasy.org/compute_pi/, accessed on 2014/04/20. (b) Koslowski, L. P. Calculation of protein isoelectric point, http:// isoelectric.ovh.org/, accessed on 2014/04/20. (4) Righetti, P. R. Electrophoresis 2006, 27, 923−938. (5) Shimura, K. Electrophoresis 2009, 30, 11−28. (6) Rabilloud, T.; Chevallet, M.; Luche, S.; Lelong, C. J. Proteomics 2010, 73, 2064−2077. (7) Zhu, G.; Sun, L.; Keithley, R. B.; Dovichi, N. J. Anal. Chem. 2013, 85, 7221−7229. (8) Sommer, G. J.; Hatch, A. V. Electrophoresis 2009, 30, 742−757. (9) Chen, H.; Fan, Z. H. Electrophoresis 2009, 30, 758−7656. (10) (a) Eijkel, J. C. T.; van den Berg, A.; Schasfoort, R. B. M. Electrophoresis 2008, 29, 977−993. (b) Turgeon, R. T.; Bowser, M. T. Anal. Bioanal. Chem. 2009, 394, 187−198. (11) Xu, Y.; Zhang, C. X.; Janasek, D.; Manz, A. Lab Chip 2003, 3, 224−227. (12) Lu, H.; Gaudet, S.; Schmidt, M. A.; Jensen, K. F. Anal. Chem. 2004, 76, 5705−5712. (13) (a) Kohlheyer, D.; Besselink, G. A. J.; Schlautmann, S.; Schasfoort, R. B. M. Lab Chip 2006, 6, 374−380. (b) Kohlheyer, D.; Eijkel, J. C. T.; Schlautmann, S.; van den Berg, A.; Schasfoort, R. B. M. Anal. Chem. 2007, 79, 8190−8198. (14) Song, Y. A.; Hsu, S.; Stevens, A. L.; Han, J. Anal. Chem. 2006, 78, 3528−3536. (15) (a) Albrecht, J. W.; Jensen, K. F. Electrophoresis 2006, 27, 4960− 4969. (b) Albrecht, J. W.; El-Ali, J.; Jensen, K. F. Anal. Chem. 2007, 79, 9364−9371. (c) Wen, J.; Wilker, E. W.; Yaffe, M. B.; Jensen, K. F. Anal. Chem. 2010, 82, 1253−1260. (16) Song, Y.-A.; Chan, M.; Celio, C.; Tannenbaum, S. R.; Wishnok, J. S.; Han, J. Anal. Chem. 2010, 82, 2317−2325. (17) Walowski, B.; Hüttner, W.; Wackerbarth, H. Anal. Bioanal. Chem. 2011, 401, 2465−2471. (18) Righetti, P. R. J. Chromatogr., A 2004, 1037, 491−499. (19) (a) Zanzotto, A.; Szita, N.; Boccazzi, P.; Lessard, P.; Sinskey, A. J.; Jensen, K. F. Biotechnol. Bioeng. 2004, 87, 243−254. (b) Stahl, H.; Glud, A.; Schroeder, C. R.; Klimant, I.; Tengberg, A.; Glud, R. N. Limnol. Oceanogr. Methods 2006, 336−345. (c) Gerlach, J.; Pohn, B.; Karl, W.; Scheideler, M.; Uray, M.; Bischof, H.; Schwab, H.; Klimant, I. 9539

dx.doi.org/10.1021/ac501783r | Anal. Chem. 2014, 86, 9533−9539