Determination of Pyrimidine and Purine Bases by Reversed-Phase

ARTICLE pubs.acs.org/ac

Determination of Pyrimidine and Purine Bases by Reversed-Phase Capillary Liquid Chromatography with At-Line Surface-Enhanced Raman Spectroscopic Detection Employing a Novel SERS Substrate Based on ZnS/CdSe Silver Quantum Dots Carolina Carrillo-Carri
on,†,‡ Sergio Armenta,† Bartolom
e M. Simonet,‡ Miguel Valc
arcel,‡ and Bernhard Lendl*,† † ‡

Institute of Chemical Technologies and Analytics, Vienna University of Technology, Getreidemarkt 9/164 AC, A-1060 Vienna, Austria Department of Analytical Chemistry, Marie Curie Building (Annex), Campus de Rabanales, University of C
ordoba, E-14071 C
ordoba, Spain

bS Supporting Information ABSTRACT: We have developed a new SERS substrate based on the reduction of silver nitrate in the presence of ZnS-capped CdSe quantum dots. This substrate showed higher sensitivities as compared to a hydroxylamine-reduced silver sol. On the basis of this new substrate, at-line SERS detection was coupled with capillary liquid chromatography (cap-LC) for the separation and selective determination of pyrimidine and purine bases. For this purpose, wells of a dedicated microtiter plate were loaded with 20 μL of the SERS substrate and placed on an automated x,y translation stage. A flow-through microdispenser capable of ejecting 50 pL droplets, at a frequency 100 Hz, was used as the interface to connect the cap-LC system to the wells loaded with SERS substrate. A detailed study of the dependence of both the separation and the surface-enhanced Raman spectra of each base on the pH was performed to optimize the system for maximum sensitivity and selectivity. Highly satisfactory analytical figures of merit were obtained for the six investigated bases (cytosine, xanthine, hypoxanthine, guanine, thymine, and adenine) with detection limits ranging between 0.2 and 0.3 ng injected on the capillary LC column, and the precisions were in the range of 3.0 6.3%.

S

eparation and determination of nucleic acid bases, the constituents of DNA, is an interesting and challenging task. Purine and pyrimidine bases in urine, serum, or plasma show noticeable differences between healthy subjects and individuals with various types of cancer.1,2 The concentration of nucleic acid components in physiological fluids, tissues, and cells are related to the catabolism of nucleic acids, enzymatic degradation of tissues, dietary habits, and various salvage pathways. Therefore, detection of elevated levels of these substances could be indicative of certain diseases or metabolic disorders.3 Aside from the four principal DNA nucleic bases (adenine, guanine, thymine, and cytosine), xanthine and hypoxanthine are two important derivates that are usually also present. Lim et al. recently showed that xanthine and hypoxanthine could be formed from guanine and adenine during DNA hydrolysis or enzymatic degradation.4,5 As such, the quantification of total purines is normally defined as the sum of adenine, guanine, xanthine, and hypoxanthine.6 The most widely used technique for the determination of nucleic bases in biological samples is high performance liquid chromatography (HPLC). The reversed phase is at present the most commonly used liquid chromatographic method for the determination of nucleic acid constituents, using C8 or C18 r 2011 American Chemical Society

columns.7 9 The nucleotides, nucleosides, and bases can be separated by reversed phase, but some of these cannot be resolved simultaneously and require the ion-pairing mode, which has a wider range of parameters that need to be adjusted and optimized for separation.9,10 The difficulty of resolving mixtures containing both purines and pyrimidines results from a phenomenon known as “vertical base stacking”. Brown and Grushka showed that this phenomenon had a significant effect on the retention time and orders of the nucleic acid constituents.11 The majority of nucleotides, nucleosides, and nucleic bases are generally detected by UV spectroscopy. However, the major challenge encountered upon developing separation techniques based on UV detection is the unique identification of the analyte in a certain peak of the chromatogram. Therefore, more information-rich detection techniques including mass spectrometry,12 14 NMR,15 fluorescence,15,16 or electrochemical detection17,18 have also been employed. In this context, vibrational spectroscopy is also of interest due to the molecular specific information contained in IR Received: August 1, 2011 Accepted: November 3, 2011 Published: November 03, 2011 9391

dx.doi.org/10.1021/ac201821q | Anal. Chem. 2011, 83, 9391–9398

Analytical Chemistry and Raman spectra. When it comes to the detection of very small quantities, surface-enhanced Raman scattering is gaining popularity due to the exceptional sensitivity provided by this technique. In surface-enhanced Raman scattering (SERS) spectroscopy, the Raman vibrational modes of a molecule located in close proximity to a rough noble metal surface are strongly enhanced. Enhancement factors in the order of 103 106 are typical for SERS, and in cases where resonance Raman effects also contribute to the signal, in addition to SERS, enhancement factors greater than 1014 can be obtained, which make single molecule detection possible.19 21 The substrates employed in SERS are of key relevance. In an ideal case, they should show large and constant enhancement factors, have a long shelf life, and be simple to make. Among the most popular SERS substrates are silver colloids that, in most cases, are prepared by chemical reduction of silver nitrate. Over the years, different ways of preparing these substrates have been developed, the most popular being those using citric acid (Lee Meisl), boron hydride (Creighton), or hydroxylamine (Leopold Lendl) as reducing agents. In particular, the use of hydroxylamine has gained popularity because this synthesis route involves simple mixing of two reagents and can be performed at room temperature. Nevertheless, there is a constant desire to improve the enhancement factors of the colloids employed for SERS experiments. More than a decade ago it was shown that surface-enhanced Raman scattering (SERS) spectroscopy could be an attractive technique to identify analytes after separation by GC or HPLC.22 27 Online, as well as at-line, coupling procedures have been proposed. Whereas online coupling appears advantageous from the viewpoint of apparent experimental simplicity, there are important practical difficulties related to this approach. Memory effects are often encountered when analytes adhere to solid-state SERS substrates or when the employed SERS sols attach to the tubings of flow system used. Furthermore, in the case of online detection, the adapted chemical conditions, such as the composition of the mobile phase and the chosen pH, are often a compromise between the needs of the chromatographic separation and of the SERS detection, respectively. In this work, we propose a more flexible and robust at-line approach, which allows optimization of the conditions for separation and detection separately and which, by design, avoids unwanted memory effects. This employs a cap-HPLC system, using a flow-through microdispenser as an interface to achieve robust SERS detection employing a novel silver quantum dot (Ag QD) SERS substrate. ZnS/CdSe quantum dots were used, as these QDs are stable and can be synthesized and purified with a very narrow distribution of their diameter. Using this system, separation and sensitive detection of the main purine and pyrimidine bases (adenine, guanine, thymine, and cytosine) and two of their most common degradation products (xanthine and hypoxanthine) were achieved.

’ EXPERIMENTAL SECTION Reagents and Materials. All chemical reagents were of analytical grade and used with no additional purification. Cadmium oxide (99.99%), trioctylphosphine oxide (TOPO, 99%), trioctylphosphine (TOP, 90%), selenium (powder, 100 mesh, 99.99%), diethylzinc solution (ZnEt2, ∼1 M in hexane), bis(trimethylsilyl) sulfide ((TMS)2S), anhydrous methanol, and anhydrous chloroform were purchased from Sigma Aldrich (Madrid, Spain). Hexylphosphonic acid (HPA) was obtained

ARTICLE

from Alfa Aesar (Karlsruhe, Germany). Silver nitrate (99.5%), hydroxylamine hydrochloride (98%), sodium hydroxide, rhodamine 6G, crystal violet chloride, adenine, and 2,6-dinitrotoluene were obtained from Sigma Aldrich (Vienna, Austria). Methanol (HPLC grade) was purchased from Sigma (Schnelldorf, Germany). Adenine, guanine, cytosine, thymine, xanthine, and hypoxanthine were purchased from Sigma (Vienna, Austria). Stock solutions of the standards (1000 mg/L) were prepared by dissolving adenine, cytosine, and thymine in water and guanine, xanthine, and hypoxanthine in 2 M perchloric acid. These solutions were stable for at least 2 weeks when stored at 4 °C. Diluted standards were prepared from this solution by appropriate dilution in the mobile phase (methanol:water, 1:1). Working standard mixture solutions were prepared from the individual stock solutions by appropriate dilution in a phosphate buffer (pH = 7). These diluted working solutions were prepared daily. Instrumentation. Raman Spectroscopy. Raman spectra were acquired with a confocal Raman microscope (LabRAM HR, Jobin Yvon Ltd., Bensheim, Germany) using a 633 nm laser line (17 mW) and a charge-coupled device (CCD) detector with 1024  256 pixels. A grating with 600 grooves/mm, a confocal aperture of 500 μm, an entrance slit of 100 μm, and an acquisition time of 20 s were selected for the experiments. A 100 microscope objective was used to record the spectra from the sample present in microwells. Under these experimental conditions, the volume measured was approximately 1 μm3. Characterization of the SERS colloids was performed using a UV vis diode array spectrometer (HP 8452A, Hewlett-Packard), as well as a transmission electron microscope (FEI TECNAI F20) and a secondary electron microscope (FEI Quanta 200 3D DualBeam-FIB). Capillary Liquid Chromatography and Fractionation Unit. The capillary LC chromatographic system consisted of an Ultimate 3000 Dionex with a 1 μL injection loop and a C18 Acclaim PepMap (300 μm i.d.  15 cm, 3 μm, 100 Å) separation column from Dionex (Dionex Corporation, GmbH, Germany). For the separation, methanol was used as solvent A and deionized water was solvent B. A flow rate of 3 μL min 1 was selected throughout. A variable wavelength UV vis detector (Ultimate UV detector, LC Packings, Dionex) set at 260 nm was placed before the microdispenser interface. The UV detection window was made in the untreated fused silica capillary (i.d. 50 μm, o.d. 363 μm) by burning away a small piece of surrounding polyamide coating on the capillary. An in-house-developed program, running in LabVIEW 8.5 (National Instruments, Austin, TX) was used to control the UV vis detector and register the chromatograms.28 The fractionation unit consisted of a flow-through microdispenser connected to the capillary column of the LC system via a fused silica capillary (i.d. 50 μm, o.d. 364 μm, and 40 cm long) and a computer-controlled x,y stage (Newport THK, Compact Linear Axis) carrying a homemade PTFE microtiter plate to collect aliquots from the eluent by means of the flow-through microdispenser. This microliquid handling device formed by two silicon microstructures is ideally suited for hyphenation with capillary LC systems mainly because of its small internal volume.29 A schematic of the microdispenser used can be found in a previous work.30 The piezoceramic element of the dispenser was driven by a dc power supply (HGL 5630 DLBN) together with a computer-controlled arbitrary waveform generator (Agilent 33120A, Agilent Technologies, Palo Alto, CA) which provided an electronic pulse with defined amplitude (15 V) that allowed ejection of a 50 pL droplet per pulse. The computer controlled x,y stage moved in step sizes of 5 μm and with a 9392

dx.doi.org/10.1021/ac201821q |Anal. Chem. 2011, 83, 9391–9398

Analytical Chemistry

ARTICLE

Figure 1. UV vis spectra of the silver colloids formed as a function of the concentration of ZnS/CdSe quantum dots present in the reaction volume.

maximum distance of 90 mm  40 mm. The PTFE microtiter plate contained a line of wells, spaced by 0.5 mm, each measuring 3 mm in diameter and 3 mm depth. The microdispenser, as well as the x,y stage, were operated with the help of an in-housewritten MS Visual Basic 6.0 based software program. Preparation of the Ag QDs SERS-Active Substrates. First, core shell quantum dots (ZnS-capped CdSe) were synthesized using CdO as precursor following the procedure described by Peng’s group.31 This reaction scheme produces a solution of QD in chloroform with a concentration of 2 mg/mL. From this solution, 100 μL, corresponding to 0.2 mg of CdSe/ZnS QDs, were dispersed in 5 mL of a solution containing hydroxylamine hydrochloride (8.5  10 3 M) to which 0.04 mL of NaOH 2 M were added. Dispersion of the quantum dots in the aqueous phase was achieved through the application of ultrasound (50 W, 60 Hz) for 10 min. During this step, the chloroform evaporated and an aqueous solution containing QDs stabilized by hydroxylamine was obtained. The hydroxylamine QDs solution was rapidly added to 45 mL of an aqueous solution of silver nitrate (1.1  10 3 M) under vigorous stirring. Finally, the reaction mixture was stirred for an additional 10 min. The final concentrations of each reagent were 1  10 3 M of Ag, 8.5  10 4 M of hydroxylamine, and 4 mg/L of CdSe/ZnS QDs. For comparison, hydroxylamine-reduced silver colloids were prepared following the procedure reported by Leopold and Lendl32 where the final reaction solution contained 1.5  10 3 mol/L hydroxylamine hydrochloride and 1  10 3 mol/L silver and no QDs. Optimized Cap-LC SERS Conditions. Separation Conditions. The following multistep gradient was used: 0 min 2% A, 3 min 2% A, 10 min 10% A, 22 min 22% A, 24 min 22% A, 26 min 2% A, 30 min 2% A. The column was kept at 40 °C, and the experiments were performed with 3 μL min 1 eluent flow rate. All of the eluent was transferred to the microdispenser without flow splitting. Optimized Fractionation of the Cap-LC Chromatogram. Two different working conditions of the fractionation system

Figure 2. (i) SEM picture of the silver colloid formed. As compared to the Leopold Lendl colloid (no QDs added), the newly prepared ZnS/CdSe colloids showed the formation of larger clusters while retaining the shape of the individual nanometer-sized colloids to a greater extent. (ii) TEM image of the Ag QD composite at two magnifications (A and B). (C) Amplification of the marked area showing the crystalline planes of a QD within the Ag QD structure. (D) Electron diffraction pattern of QD.

consisting of the microdispenser and the microtiter plate were adapted by selecting appropriate translation speeds of the x,y stage while maintaining a dispensing frequency of 100 Hz and a flow-rate of 3 μL min 1. Continuous System. By setting the translation speed of the stage to 350 μm/s, about 8 s of the chromatogram were collected in one well of the microtiter plate. Therefore, using these settings it was possible to record 3D SERS chromatograms (intensity/ wavenumber/time). Optimized System for Increased Sensitivity. After having optimized the chromatographic separation and having determined the retention time of each analyte, conditions could be set 9393

dx.doi.org/10.1021/ac201821q |Anal. Chem. 2011, 83, 9391–9398

Analytical Chemistry

ARTICLE

Figure 3. SERS spectra of the studied analytes recorded at the different pH values. The selected pH value for each one is marked with an asterisk. All spectra were collected using acquisition times of 20 s.

to collect the maximum amount of each analyte in one well. This was achieved by setting the translation speed to 20 μm per second. Each well was filled with 20 μL of the Ag QD colloidal solution. From the fractionation of the cap-LC chromatogram into different well plates, it was possible to adjust the conditions of the SERS substrate in each well according to the needs of the analytes eluting at the respective retention time. In this particular application, this experimental flexibility offered the opportunity to adjust the pH in each well plate for maximum sensitivity. SERS Measurements. For the recording of SERS spectra from the wells, the microtiter plate was properly positioned under the microscope and the laser focused manually inside the solution of each well. To record SERS spectra from each well sequentially, the position of the microplate was varied by means of a manually controlled x,y stage.

’ RESULTS AND DISCUSSION Colloidal SERS Substrate Based on Silver ZnS/CdSe Quantum Dots (Ag QDs). The reaction sequences that take place

during reduction most likely involve complex formation between the CdSe/ZnS QDs, the hydroxylamine, and the silver ions, in addition to the observed silver reduction. We observed that,

in the presence of QD, the pH of the solution decreased less (final pH 9.5) than in the absence of QDs (pH 7). This minor decrease in pH indicates that fewer protons were released to the environment, which in turn means that fewer hydroxylamine molecules are oxidized and that QDs do also act as reducing agents. This hypothesis is supported by the fact that an excess of hydroxylamine was found in the final Ag QD solution when the concentration of hydroxylamine was set to 1.5 mM. When completely removing hydroxylamine during the reduction step, QDs were not well dispersed and large silver aggregates were formed. Additionally, increasing the concentration of QDs present during the reduction step resulted in an increase of the average size of the particles of the silver sol formed which tend to precipitate quickly. Therefore, the hydroxylamine concentration was set to 0.85 mM and the QD concentration to 4 mg/L to ensure colloidal particles of an appropriate size with sufficient dispersion of the QDs for all subsequent analyses. Characterization of the Ag QD Colloids by UV vis Spectroscopy. The optical characterization of Ag QD colloids was carried out using UV vis spectroscopy. The absorption maximum of the UV vis spectrum of the colloidal solution provides information on the average particle size, whereas its full width at half-maximum can be used to estimate particle dispersion. As can be seen in Figure 1, UV/vis spectra confirm the formation of a 9394

dx.doi.org/10.1021/ac201821q |Anal. Chem. 2011, 83, 9391–9398

Analytical Chemistry

ARTICLE

new class of colloid. For conventional Leopold Lendl silver sols (without the presence of QDs) the plasmon band has a λmax at 405 nm. However, for Ag QD colloids, a red shift and a broadening of the plasmon band was observed as the amount of QDs added increased from 0 to 10 mg/L, suggesting the formation of larger particles. The decrease in absorption for the high concentrations of QD can partly be explained by rapid precipitation of the produced Ag QD colloids. For practicability of the planned SERS measurements, the QD concentration was set to 4 mg/L, which corresponds to a Ag/QD ratio of 27 (w/w). Characterization of the Ag QD Colloids by Electron Microscopy. SEM and TEM were used to study the morphologies of the synthesized Ag QD colloids. As can be seen in SEM images (Figure 2i), the Ag QD composite presents a sponge-like morphology where, however, the shape of individual colloid particles at several tens of nanometers is more preserved than in the case of the Leopold Lendl colloids. Also, it should be noted that larger aggregates are formed in the presence of the quantum dots. Figure 2ii shows TEM images of the colloids. QDs attached on the silver surface or located between silver particles were occasionally observed and were probably not homogenously distributed on the surface, as QDs were observed only in some parts of the colloid. Figure 2ii also shows the typical selective area electron diffraction pattern of a QD inside the Ag QD structure. The recorded electron diffraction pattern matches well with the standard diffraction pattern of a very high crystalline structure, confirming that the QDs attached on the silver surface preserved their crystallinity. SERS Activity of the Ag QD Colloid. Maximum SERS intensities were obtained with the Ag QD colloid containing 4 mg/L of QDs, and these intensities were significantly greater than those obtained from colloids without QDs as shown in the Supporting Information (S1). The higher SERS activity of the Ag QD colloid was attributed to several reasons: (i) The silver aggregates obtained in the presence of QDs are larger in size, and this different degree of aggregation of the silver colloid can be responsible for the SERS enhancement. (ii) Some QDs act as a spacer between silver nanoparticles and thus generate ‘hot spots’ that contribute to an increase in the overall observed enhancement factors. (iii) QDs also act as a coreductor of the reaction together with hydroxylamine, leading to a different silver structure with a particular spongy morphology that is more active in SERS. Analysis of a 2  10 9 M R6G solution showed blinking in the intensities of the characteristic Raman bands. This may be explained by the fact that on average only 1.2 molecules were present in the sampled volume, which was estimated to be 1 μm3 and demonstrates the high sensitivity achievable with this technique. The Ag QD colloids showed stable SERS activity

over 20 days. After a month, some sedimentation of larger particles was observed and the SERS activity began to decrease gradually. After 40 days, the colloid lost half of its activity (ca. 55% of the relative intensity of the rhodamine band at 1510 cm 1). Chromatographic Conditions and Operation of the Fractionation Unit. Selection of Eluent Solvent, pH, and Flow Rate. The selection of the eluent was governed by the following considerations: (i) it should yield well-separated peaks of the analytes under investigation, and (ii) it should not adsorb readily on the SERS-active silver surface and thereby prevent the analyte from gathering there. In several previous works,6,7,30 36 the HPLC separation of this type of compound was performed with a mobile phase containing approximately 20% by volume of acetonitrile. However, AcNcontaining eluent produces a low intensity background spectrum that can make detection of analytes with a low concentration difficult. Additionally AcN also adsorbs onto the silver SERS surface and can interfere with the interaction of the SERS substrate and analyte,24 as can some ions contained in common buffer solutions used as the mobile phase. Therefore, we replaced acetonitrile with methanol as mobile phase although a more complex gradient program had to be used in order to achieve a good separation with the capillary-LC system. Using the simple methanol/water mobile phase, one can be sure that the observed vibrational bands are exclusively due to the analyte.

Figure 4. 3D SERS chromatogram obtained from a 5 mg L 1 mixture standard injected in the cap-LC column. The speed of the stage was fixed at 350 μm s 1. Each analyte was collected on wells filled with 20 μL of the Ag QD solution at different pH values according with the strongest interaction of each analyte (pH = 8 for cytosine, pH = 5 for xanthine, pH = 4 for hypoxanthine, pH = 4 for guanine, pH = 6 for thymine, and pH = 4 for adenine).

Table 1. Analytical Features of the Cap-LC SERS Method for the Six Nucleic Acid Bases cytosine

xanthine

retention time (s)

502

615

Raman intensity

(802 cm 1)

(660 cm 1)

guanine

thymine

868

953

1091

1314

(730 cm 1)

(653 cm 1)

(737 cm 1)

(735 cm 1)

regression equation

y = 2230x 1.0 10

0.9 10

0.7 2.0

0.8 10

0.95 10

0.67 2.5

linearity (R2) LOD (mg L‑1)

0.991 0.30

0.996 0.27

0.999 0.21

0.996 0.24

0.998 0.29

0.998 0.20

LOQ (mg L‑1)

1.0

0.90

0.70

0.80

0.95

0.67

RSD (%)

6.3

4.7

4.1

4.4

3.0

3.5

18

y = 19116x

9395

54

y = 3463x

50

y = 3445x

adenine

lineal range (mg L‑1)

17

y = 3456x

hypoxanthine

586

y = 18208x

61

dx.doi.org/10.1021/ac201821q |Anal. Chem. 2011, 83, 9391–9398

Analytical Chemistry It is well-known that at physiological pH the predominant form of the bases is the keto form. The existence of a tautomer mainly in the amino or keto form rather than in the imino or enol form will cause a decrease both in the extent of stacking and in the retention time. For that reason, the better resolution was achieved at a pH between 7 and 8. At pH values lower than 6, stronger stacking interaction took place between the analytes, resulting in poor separation with only four peaks being observed in the chromatogram. Additionally, at pH values higher than 8, the resolution was worse. Therefore, the pH optimum for the sample injected on the LC column was fixed at 7 by preparing the working solutions in a phosphate buffer. The effect of the mobile phase flow rate affects not only the separation efficiency but also the peak shape. A flow rate of 3 μL/ min was found to be optimal. At higher flow rates, the eluting peaks were clearly resolved whereas at lower flow rates, analysis time was increased unnecessarily. Under the optimum conditions, the separation of the six analytes cytosine, xanthine, hypoxanthine, guanine, thymine, and adenine eluted at 502, 615, 868, 953, 1091, and 1314 s, respectively. A typical UV chromatogram obtained when a standard mixture of 2 mg/L was injected on the capillary column is included in the Supporting Information (S2). SERS Detection. pH Dependence of the SERS Spectra of the Pyrimidine and Purine Bases. The pH of the colloidal substrate is a critical parameter for obtaining high-quality SERS spectra. Adjustment of the pH of the Ag QD SERS-active solution showed that, in order to obtain the highest SERS intensities, the optimum pH was very different for each base studied, as shown in Figure 3. The best pH for the SERS-active solution for each analyte was determined as follows: pH = 8 for cytosine, pH = 5 for xanthine, pH = 4 for hypoxanthine, pH = 4 for guanine, pH = 6 for thymine, and pH = 4 for adenine. These differences can be explained by the pKa values of the studied bases and their tautomer equilibrium. It has been reported that the pKa values of all compounds studied here are lowered substantially in the presence of the SERS substrates,37 meaning that lower pH values than originally thought, according to their pKa, have to be employed to obtain SERS spectra of neutral molecules. For all bases, it was found that the deprotonated forms adsorbs preferably on the Ag QD surface instead of the cationic or protonated forms. However, depending on the studied base (pKa and the tautomeric equilibrium), the pH required to obtain the neutral and more stable form, and therefore the pH dependence on the SERS signal for each base, was different for each base. Although adenine can undergo a change from amine to imino form, the amine form is strongly favored and therefore is the predominant form at any pH; hence, little change was observed in this base by varying the pH between 4 and 8. In stronger acidic media, adenine undergoes protonation on the N(1) position rather on the amino group, and even at pH values lower than 2, adenine is double protonated at N(1) and N(7). The adsorption of these latter protonated forms on the silver surface was not favored. Because of similar purine structure, the SERS spectra of adenine, guanine, xanthine, and hypoxanthine showed a very similar pH dependence in that the strongest spectra were obtained under weak acidic conditions. On the other hand, a different pH dependence was observed for the pyrimidine bases. For pyrimidines, the keto form predominates at neutral pH (between 6 and 8), but at alkaline pH, the tautomeric equilibrium shifts to the enol form. At alkaline pH, the hydrogen N(3) of

ARTICLE

Table 2. Comparison of the Proposed Method with Other Developed Techniques for the Separation and Determination of Purine and Pyrimidine Bases analytical technique cap-LC SERS

analytes adenine

LOD 0.20 mg/L

reference this work

(Ag QD as substrate)

RP-HPLC DADa

guanine

0.24 mg/L

cytosine

0.30 mg/L

thymine xanthine

0.29 mg/L 0.27 mg/L

hypoxanthine

0.21 mg/L

cytosine

0.047 mg/mL

guanine

0.105 mg/mL

thymine

0.057 mg/mL

44

uracil

0.024 mg/mL

ion-pair-RP-HPLC UVb

adenine

1.75 mg/L

10

CZE UVc

guanine adenine

1.17 mg/L 0.8 mg/L

45

(β-cyclodextrin as additive)

CZE UVd

guanine

1.2 mg/L

cytosine

1.5 mg/L

thymine

1.3 mg/L

uracil

1.8 mg/L

adenine

0.8 mg/L

46

(c-MWNTs as additive)

CZE EDe

SiCNP/GCE DPVf

guanine

1.2 mg/L

cytosine

1.5 mg/L

thymine

1.3 mg/L

uracil

1.8 mg/L

adenine

2.47 μM

guanine

1.57 μM

xanthine

4.17 μM

hypoxanthine adenine

7.67 μM 0.015 μM

guanine

0.015 μM

thymine

0.14 μM

cytosine

0.14 μM

47

48

a

RP-HPLC DAD: Reversed-phase high-performance liquid chromatography with diode-array detection. b Ion pair-RP-HPLC UV: Ionpair reversed-phase high-performance liquid chromatography with UV detection. c CZE UV: Capillary zone electrophoresis with UV detection using β-cyclodextrins as additive into the running buffer. d CZE UV: Capillary zone electrophoresis with UV detection using carboxylic multiwalled carbon nanotubes as additive into the running buffer. e CZE ED: Capillary zone electrophoresis with electrical detection using a carbon disk electrode. f SiCNP/GCE DPV: Silicon carbide nanoparticlemodified glassy carbon electrode differential pulse voltammetry.

thymine is removed, indicating the weak basicity of the ring nitrogen. Again, it could be observed that the protonated form was not readily adsorbed on the silver substrate. In summary, it was found that the amine and uncharged form (for adenine and cytosine) and the keto and uncharged form (for guanine, xanthine, hypoxanthine, and thymine) led to the highest SERS signals, and this implies that they were the predominant forms present on the Ag QD surface. The results found here are in agreement with conclusions in previous works.38 43 Figure 4 shows the 3D capillary LC SERS chromatogram obtained using 9396

dx.doi.org/10.1021/ac201821q |Anal. Chem. 2011, 83, 9391–9398

Analytical Chemistry the continuous system while in Supporting Information (S3) we show SERS spectra of the six bases at their optimum pH after their separation and collection with the fractionation system using the optimized system for increased sensitivity, having adjusted the colloidal solution in each well to that most appropriate for the individual bases. Analytical Features. The linearity, sensitivity, limits of detection (LOD) and quantification (LOQ), and precision of the system were obtained and are summarized in Table 1. Different mixture standards of the six bases from 0.1 to 10 mg/L were analyzed following the whole procedure: chromatographic separation, fractionation, and finally SERS detection. Calibration graphs were constructed by plotting the Raman intensity of the selected peak for each analyte versus the concentration injected on the LC column in mg/L. The bands corresponding to the ring breathing mode (735 cm 1 for adenine, 653 cm 1 for guanine, 660 cm 1 for xanthine, 730 cm 1 for hypoxanthine, 802 cm 1 for cytosine, and 737 cm 1 for thymine) were used as the analytical signal, as they are the most intense and characteristic for these analytes. Note that the profile of the SERS spectra of each analyte was similar at different concentration levels. The shape of the Raman spectra was not concentration dependent and only the overall intensity of the bands was affected. The LOD and LOQ were calculated from the standard deviation of five independent analyses at a concentration level of 2.5 mg/L divided by the analytical sensitivity (slope of the calibration equation) and multiplying by 3 or 10 for LOD and LOQ, respectively. The repeatability of the total procedure was tested by repeated analysis of 2.5 mg/L mixture standard (n = 5) with relative standard deviations (RSD) between 3.0 and 6.3%. Note that the detection limits obtained with the proposed methodology are considerably lower than the most of those previously reported for the analysis of purines and pyrimidines bases using different techniques (Table 2). This clearly demonstrates the favorable and efficient coupling of a capillary-LC system with SERS spectroscopy using Ag QD as substrate.

’ CONCLUSION In this contribution, we demonstrated the successful coupling of a capillary-LC system with SERS spectroscopy for the separation, identification, and quantification of six nucleic acid bases. Moreover, we have reported a very efficient SERS substrate with spongelike morphology exploiting the reductor potential of QDs and their integration in silver nanostructure to allow optimal aggregate size, shape, and spacing of the Ag particles. Following the proposed methodology, it is very simple to record the SERS spectra of each analyte at its most appropriate pH value. The Ag QD SERS-active solution can readily be adjusted to a desired pH value in order to obtain the maximum SERS enhancement for each analyte. The strategy was to fill each well of the microtiter plate with the substrate solution previously adjusted to specific pH value(s) in each of the wells according to the elution order of the analytes and their required optimum pH values. This requires prior knowledge of the order of elution of compounds and to which well each analyte will be collected, under a given set of working conditions. Note that the presented methodology has the potential to be further optimized for achieving even higher sensitivities. Considering the small spot size of the confocal Raman microscope, significantly smaller wells could be used to achieve a higher analyte concentration in a single well. Very accurate handling of

ARTICLE

picoliter to nanoliter volumes can be identified as the key for such improved interfaces. In the case where this can be achieved on a routine basis, SERS detection in a capillary or nanoseparation system could become a powerful detection technique, adding vibrational spectroscopic information for improved qualitative as well as quantitative analysis.

’ ASSOCIATED CONTENT

bS

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

’ AUTHOR INFORMATION Corresponding Author

*Tel/fax: +43-1-58801-15140/15199. E-mail: blendl@mail. zserv.tuwien.ac.at.

’ ACKNOWLEDGMENT The authors express their gratitude to the Spanish Ministry of Innovation and Science for project CTQ2007-60426, and to the Junta of Andalusia for their project FQM02300. C.C.-C. also thanks the Ministry for the award of a Research Training Fellowship (Grant AP2006-02351). B.L. acknowledges financial support received from the Austrian FFG within the Research Studio Austria program. ’ REFERENCES (1) Di Pietro, M. C.; Vannoni, D.; Leoncini, R.; Liso, G.; Guerranti, R.; Marinello, E. J. Chromatogr., B: Biomed. Sci. Appl. 2001, 751, 87–92. (2) Sandoval Guerrero, K.; Revilla V
azquez, A.; Segura-Pacheco, B.; Due~ nas-Gonz
alez, D. Electrophoresis 2005, 26, 1057–1062. (3) Liu, L.; Ouyang, J.; Baeyens, W. R. G. J Chromatogr., A 2008, 1193, 104–108. (4) Lou, S. J. Food Sci. 1998, 63, 442–444. (5) Lim, K. S.; Huang, S. H.; Jenner, A.; Wang, H.; Tang, S. Y.; Halliwell, B. Free Radic. Biol. Med. 2006, 40, 1939–1948. (6) Reynal, S. M.; Broderick, G. A. J. Dairy Sci. 2009, 92, 1177–1181. (7) Pi~ neiro-Sotelo, M.; Rodríguez-Bernaldo de Quir
os, A.; L
opezHern
andez, J.; Simal-Lozano, J. Food Chem. 2002, 79, 113–117. (8) Katayama, M.; Matsuda, Y.; Shimokawa, K.; Tanabe, S.; Kaneko, S.; Hara, I.; Sato, H. J. Chromatogr., B 2001, 760, 159–163. (9) Brown, P. R.; Robb, C. S.; Geldart, S. E. J. Chromatogr., A 2002, 965, 163–173. (10) García del Moral, P.; Arín, M. J.; Resines, J. A.; Díez, M. T. J. Chromatogr., B 2005, 826, 257–260. (11) Brown, P. R.; Grushka, E. Anal. Chem. 1980, 52, 1210–1215. (12) Porcelli, B.; Pagani, J.; Lorenzini, L.; De Martino, A.; Catinella, S.; Traldi, P. Rapid Commun. Mass Spectrom. 2005, 8, 443–450. (13) Lemr, K.; Adam, T.; Frycak, P.; Friedecky, D. Adv. Exp. Med. Biol. 2000, 486, 399–403. (14) Frycak, P.; Huskova, R.; Adam, T.; Lemr, K. J. Mass Spectrom. 2002, 37, 1242–1248. (15) Suzkowska, A. Spectroscopy 2002, 16, 379–385. (16) Ramos-Salazar, A.; Baines, A. D. Anal. Biochem. 1985, 145, 9–13. (17) Zen, J.-M.; Chang, M.-R.; Ilangovan, G. Analyst 1999, 124, 679–684. (18) Wang, H.-S.; Ju, H.-X.; Chen, H.-Y. Anal. Chim. Acta 2002, 461, 243–250. (19) Nie, S.; Emory, S. Science 1997, 275, 1102–1106. (20) Kneipp, K.; Wang, Y.; Kneipp, H.; Perelman, L. T.; Itzkan, I.; Dasari, R.; Feld, M. S. Phys. Rev. Lett. 1997, 78, 1667–1670. 9397

dx.doi.org/10.1021/ac201821q |Anal. Chem. 2011, 83, 9391–9398

Analytical Chemistry

ARTICLE

(21) Kneipp, K.; Kneipp, H.; Deinum, G.; Itzkan, I.; Dasari, R. R.; Feld, M. S. Appl. Spectrosc. 1998, 52, 175–178. (22) Sheng, R.; Ni, F.; Cotton, T. M. Anal. Chem. 1991, 63, 437–442. (23) Bell, S. E. J.; Spence, S. J. Analyst 2001, 126, 1–3. (24) S€agm€uller, B.; Schwarze, B.; Brehm, G.; Schneider, S. Analyst 2001, 126, 2066–2071. (25) S€agm€uller, B.; Schwarze, B.; Brehm, G.; Trachta, G.; Schneider, S. J. Mol. Struct. 2003, 661, 279–290. (26) Trachta, G.; Schwarze, B.; S€agmuller, B.; Brehm, G.; Schneider, S. J. Mol. Struct. 2004, 693, 175–185. (27) Heaps, D. A.; Griffiths, P. R. Appl. Spectrosc. 2005, 59, 1305–1309. (28) Wagner, C.; Armenta, S.; Lendl, B. Talanta 2010, 80 1081–1087. (29) Armenta, S.; Lendl, B. Anal. Bioanal. Chem. 2010, 397, 297–308. € (30) Miliotis, T.; Kjellstr€om, S.; Onnerfjord, P.; Nilsson, J.; Laurell, T.; Edholm, L. E.; Marko-Varga, G. J. Chromatogr., A 2000, 886, 99–110. (31) Peng, Z. A.; Peng, X. J. Am. Chem. Soc. 2001, 123, 183–184. (32) Leopold, N.; Lendl, B. J. Phys. Chem. B 2003, 107, 5723–57. (33) Kazoka, H. J Chromatogr., A 2002, 942, 1–10. (34) Makkar, H. P. S.; Becker, K. Br. J. Nutr. 1999, 81, 107–112. (35) Ivandini, T. A.; Honda, K.; Rao, T.; Fujishima, A.; Einaga, Y. Talanta 2007, 71, 648–655. (36) Monser, L. Chromatographia 2004, 59, 455–459. (37) Oh, W. S.; Kim, M. S.; Suh, S. W. J. Raman Spectrosc. 1987, 18, 253–258. (38) Szeghalmi, A. V.; Leopold, L.; Pınzaru, S.; Chis, V.; SilaghiDumitrescu, I.; Schmitt, M.; Popp, J.; Kiefer, W. J. Mol. Struct. 2005, 735 736, 103–113. (39) S
anchez-Cortes, S.; García-Ramos, J. V. Surf. Sci. 2001, 473, 133–142. (40) Sheng, R.; Ni, F.; Cotton, T. M. Anal. Chem. 1991, 63 (5), 437–442. (41) Bell, S. E.; Sirimuthu, N. M. J. Am. Chem. Soc. 2006, 128 (49), 15580–1. (42) Chowdhury, J.; Mukherjee, K. M.; Misra, T. N. J. Raman Spectrosc. 2000, 31, 427–431. (43) Badr, Y.; Mahmoud, M. A. Spectrochim. Acta, Part A 2006, 63, 639–645. (44) Yu, L.; Zhao, J.; Li, S. P.; Fan, H.; Hong, M.; Wang, Y. T.; Zhu, Q. J. J. Sep. Sci. 2006, 29, 953–958. (45) Wang, P.; Ren, J. J. Pharm. Biomed. Anal. 2004, 34, 277–283. (46) Xiong, X.; Ouyang, J.; Baeyens, W. R.; Delanghe, J. R.; Shen, X.; Yang, Y. Electrophoresis 2006, 27, 3243–53. (47) Chen, G.; Chu, Q.; Zhang, L.; Ye, J. Anal. Chim. Acta 2002, 22, 225–233. (48) Ghavami, R.; Salimi, A.; Navaee, A. Biosens. Bioelecton. 2011, 26, 3864–3869.

9398

dx.doi.org/10.1021/ac201821q |Anal. Chem. 2011, 83, 9391–9398