Article pubs.acs.org/ac
Ultrasensitive Detection of Bacteria by Microchip Electrophoresis Based on Multiple-Concentration Approaches Combining Chitosan Sweeping, Field-Amplified Sample Stacking, and Reversed-Field Stacking Zhi-Fang Wang, Shuang Cheng, Shu-Li Ge, Huan Wang, Qing-Jiang Wang,* Pin-Gang He, and Yu-Zhi Fang* Department of Chemistry, East China Normal University, Shanghai 200062, People's Republic of China S Supporting Information *
ABSTRACT: In this paper we describe an on-chip multiple-concentration method combining chitosan (CS) sweeping, reversed-field stacking, and fieldamplified sample stacking for highly efficient detection of bacteria. Escherichia coli was selected as a model bacterium to investigate the efficiency of this multiple-concentration method. CS was the most suitable sweeping agent for microchip electrophoresis, replacing the usually used cetyltrimethylammonium bromide for capillary electrophoresis. The additive taurine had a synergistic effect by enhancing the interaction between CS and the surface of the bacteria, thus improving the analysis sensitivity. All steps of the concentration method and related mechanisms are described and discussed in detail. A concentration enhancement factor of approximately 6000 was obtained using this concentration method under optimal conditions as compared to using no concentration step, and the detection limit of E. coli was 145 CFU/mL. The multiple-concentration methodology was also applied for the quantification of bacteria in surface water, and satisfactory results were achieved. The application of this methodology showed that the concentration enhancement of bacteria clearly conferred advantageous sensitivity, speed, and sample volume compared to established methods.
M
additionally, it is impossible to definitively identify unknown bacteria with a single test. During the past decade, the analysis of bacteria by capillary electrophoresis (CE) has generated great interest, and CE has been explored as an alternative approach for its high speed, relative simplicity, and low cost. Hjerten et al. reported for the first time the possibility of transporting microorganisms through a capillary by applying an electrical field under appropriate experimental conditions.3 Since then, there have been several reports of CE-based methods to analyze microorganisms, including the separation of different bacterial species,4,5 the CE behavior of microorganisms,6,7 and the reproducibility of CE separations of microorganisms.8,9 Because the infectious dose of a pathogen in water, food, and clinical samples can be very low, considerable efforts have been directed to the development of methods to rapidly detect low concentrations of pathogens in actual samples. A variety of electrophoretic concentration techniques for the analysis of bacteria have been developed to improve the sensitivity of CE, including isotachophoretic focusing,10 large-volume sample stacking with polarity switching,11 capillary isoelectric focusing,4,12 and pH-induced stacking and normal stacking.13
ethods to detect and identify pathogenic microorganisms are crucial to ensure consumer and patient safety, maintain product quality, and meet regulatory requirements. Therefore, there is an increasing demand for rapid, sensitive, and broadly applicable methods to assess the sterility of products. Currently, several methods are employed to test bacteria for potentially dangerous microbial agents. The simplest and most widely used approach is the direct inoculation method.1 However, the direct inoculation method and its derivatives require a few days to several weeks to complete, and great care must be taken to prevent any contamination during the cultivation of the bacteria. Furthermore, growth medium-dependent methods can only detect microorganisms that are capable of growing in a particular medium and under certain experimental conditions. To overcome the shortcomings of standard microbiological methods, other approaches that are rapid alternatives to traditional culture-based methods have been developed. These alternative methods provide equally accurate results and include impedance measurements and enzyme immunoassays, as well as molecular techniques, such as nucleic acid hybridization, the polymerase chain reaction, and flow cytometry.2 Although these rapid tests are promising, there are drawbacks associated with them, including a high cost and relatively complex procedures that require professional training; © 2012 American Chemical Society
Received: November 10, 2011 Accepted: January 12, 2012 Published: January 12, 2012 1687
dx.doi.org/10.1021/ac202991u | Anal. Chem. 2012, 84, 1687−1694
Analytical Chemistry
Article
attractive reagent for use in flocculation and different binding applications. Bacteria and fungi are negatively charged because of their carboxylate and phosphate groups at the outer membrane; attractive electrostatic interactions may lead to flocculation. Hence, the high adsorbing ability of CS makes it useful for bacterial concentration and as a sweeping carrier that retains cell viability. CS sweeping allows the detection and quantification of bacteria in a single peak with a high efficiency without the formation of multiple peaks attributable to irregular clusters and aggregates of bacterial cells. This sweeping method functions regardless of the electrophoretic heterogeneity or the microbial composition of the sample. In addition to CS sweeping, field-amplified sample stacking and reversed-electric-field stacking were also integrated into the multiple-concentration strategy. Conditions were optimized for a five-step procedure with a gated injection that resulted in an enrichment factor of approximately 6600 for the preconcentration of E. coli. When this method was applied to determine the content of E. coli in groundwater, the results were satisfactory, showing that this method was rapid and highly sensitive and required only a small sample volume.
Recently, Armstrong and co-workers analyzed a microbial contamination in the laboratory using a three-injection CE technique with cetyltrimethylammonium bromide (CTAB) sweeping, and they attained a 50-cell detection limit for the microbial contamination.14−16 The surfactant additive CTAB swept all cells from the sample into the blocking agent plug, producing a single peak for all cells. Subsequently, Armstrong et al. adapted this approach for ionic liquid-supporting electrolytes for the detection of microbial contaminations, especially Salmonella spp. in mixed samples, by utilizing a CE− fluorescence in situ hybridization (FISH) method.16 Furthermore, Armstrong et al. applied this coupled FISH labeling approach for the detection of Candida albicans in blood samples and achieved a detection limit lower than 2.0 × 103 CFU/mL by increasing the blood sample injection volume.17 On the basis of the CE detection of bacteria, some researchers have explored the possibility of analyzing bacteria using microchip electrophoresis (MCE), a rapid analysis method that reduces sample consumption as well as integrates and automates multiple analytical processes. For example, Shintani et al. demonstrated the separation of lactic acid bacteria and Saccharomyces cerevisiae using quartz microchip capillary electrophoresis.18 Additionally, Li et al. described the identification of enteropathogenic Escherichia coli using MCE.19 Though MCE has been used to analyze bacteria, it has an inadequate detection limit, and the detection of microbial contamination is more challenging because of the characteristically low sample volume, short separation distance, and low concentration of MCE. One way of improving the sensitivity is to integrate an online sample stacking method. Therefore, more powerful sample concentration methods are highly desired for MCE analysis of low-abundance bacteria. To overcome the low sensitivity of MCE, various online electromigration-based sample preconcentration techniques have been studied over the past two decades. These on-chip stacking methods include field-amplified sample injection (FASI),20 field-amplified sample stacking (FASS),21 largevolume sample stacking (LASS),22 isotachophoresis (ITP),23 affinity-based sample sweeping,24 solid-phase extraction,25 isoelectric focusing,26 and temperature gradient focusing.27 However, a single-step stacking method for MCE cannot meet the low detection limit requirement for low-abundance analytes. A combination of two sample stacking methods (FASI and sweeping), referred to as cation-selective exhaustive injection and sweeping, achieved nearly 1 million-fold signal enhancement.28 Qiong et al. recently reported a highly effective on-chip preconcentration method combining FASI and bovine serum albumin sweeping for ultrasensitive detection of green fluorescent proteins on a simple cross-channel microchip device.29 These multiple-concentration techniques with high concentration factors inspired us to investigate the possibility of applying a multiple-concentration method to achieve a sterile testing at lower concentrations of sample with the MCE system. The aim of this study was to find a suitable combination of concentration approaches for the detection of low-abundance microorganisms by MCE. After a series of experiments, we combined field-amplified sample stacking, reversed-electric-field stacking, and chitosan (CS) sweeping in a multiple-concentration approach for MCE detection of bacteria. CTAB has been used as a sweeping reagent in CE for sterile detection; however, CTAB can lyse cells, resulting in a poor sensitivity. CS is positively charged in solution and is therefore a very
■
EXPERIMENTAL SECTION Materials and Apparatus. Tris(hydroxymethyl)aminomethane (Tris), boric acid, and ethylenediaminetetraacetate (EDTA) were obtained from Aldrich (Milwaukee, WI). Chitosan, taurine, and L-cysteine were obtained from the Shanghai Chemical Plant (Shanghai, China). E. coli, Lactobacillus casei, and Streptococcus thermophilus were purchased from the Microorganism Germplasm Resource Bank (Guangdong, China). SYTO 62 red fluorescent nucleic acid stain (5 mM) dissolved in DMSO was purchased from Molecular Probes (Eugene, OR). All solutions were mixed with ultrapure water from Milli-Q filtration systems (Millipore, Bedford, MA). All MCE experiments were conducted using a homemade microfluidic system coupled to an laser-induced fluorescence (LIF) detection device. The MCE system was made jointly by Shanghai Spectrum Ltd. Co., Zhejiang University, and our research group.30 Briefly, a diode laser (5 mW) was used to generate an excitation beam at 635 nm. The fluorescence signal was spectrally isolated using an edge filter and was subsequently collimated with an achromatic lens before being focused onto the photomultiplier tube (PMT). The high voltage power unit variable in the range 0 ± 6 kV was used for on-chip sample injection and zone electrophoretic separation. The amplified current was transferred directly through a 10 kΩ resistor to a 24 bit A/D interface at 10 Hz (Borwin, JMBS Developments, Le Fontanil, France) and stored in a personal computer. Preparation of the Bacterial Suspension. E. coli was cultured in Luria−Bertani (LB) broth at 30 °C for 13 h, and L. casei and S. thermophilus were cultured in deMan Rogosa Sharpe (MRS) broth at 30 °C for 13 h. Bacterial cell pellets were harvested by centrifugation at 3400 rpm for 5 min, and the pellets were washed five times with sterile water to remove the residual medium and to wash the bacterial cells. The pellet was suspended in the sample buffer, which had been filtered through a sterile 0.45 μm filter (Millex GV, Millipore) to remove any bacterial contamination. Before MCE analysis, the bacterial cell concentration of the culture was estimated by measuring the optical density (OD) at 600 nm with a spectrophotometer. Red fluorescent nucleic acid SYTO 62 was added to the bacterial suspension to stain the bacterial cells. A total of 1.5 1688
dx.doi.org/10.1021/ac202991u | Anal. Chem. 2012, 84, 1687−1694
Analytical Chemistry
Article
Figure 1. Schematic diagram of the sample loading, online concentration, and separation of bacteria: (A) preloading the sample, (B) loading the sample, (C) field-amplified sample stacking and sweeping, (D) reversed-voltage stacking, and (E) separation. The black zone represents the area of the concentrated bacterial sample, the gray zone represents the sample matrix, the clear zone represents the running buffer containing CS, and the arrow indicates the direction of the electroosmotic flow. The voltages (kV) applied to the reservoirs at each step are indicated.
mL of the bacterial suspension (approximately 1.0 × 107 CFU/ mL) was mixed with 3.0 μL of 5 mmol/L SYTO 62 to obtain a final concentration of 10 μmol/L SYTO 62. After 30 min, the bacterial suspension was centrifuged for 5 min at 3400 rpm. The colored supernatant was then removed, and the pellet was washed repeatedly with sterile water until the supernatant was colorless. Finally, the stained bacterial cells were suspended and diluted in TBE buffer containing taurine to obtain a cell concentration of approximately 1.0 × 107 CFU/mL. Prior to use, the freshly prepared cell suspension was vortexed for 60− 90 s to prevent cell aggregation. Quantification of E. coli in Surface Water with the Plate Count Agar (PCA) Method. To confirm the practical application of the multiple-concentration method, the inoculation experiments were analyzed with the PCA method. The number of microorganisms in the surface water samples was controlled by serial dilutions and by plating 100 μL of each dilution on LB agar. After incubation of the plates at 30 °C for 13 h, the colonies were counted. The concentration of E. coli was 65 CFU/mL in surface water. MCE Conditions. The glass microchip design used in these experiments consisted of a simple cross-channel. Microchannels were etched to a depth of 25 μm and a width of 100 μm. The separation channel was 50 mm in length, and all other channels had a length of 10 mm measured from the channel intersection. Platinum electrodes were inserted into the reservoirs, providing an electrical contact between the power supply and the electrolyte solution. Before a new microchip was first used, it was washed with 98% H2SO4 and deionized water for 10 min. Subsequently, the channels were flushed with 1 M NaOH for 20 min, deionized water for 10 min, and the MCE running buffer for 10 min. Before each injection, the microchannels were washed with 1 M NaOH for 2 min, deionized water for 1 min, and 22.5 mM TBE (Tris/borate/EDTA, 22.5 mM Tris, 22.5 mM boric acid, and 0.5 mM EDTA) running buffer containing CS for 2 min. To clean and sterilize the microchip after each run, the microchip was conditioned with ethanol for 3 min. Therefore, a total reconditioning time of 8 min elapsed before a new run was started. After the washing steps, a precoating process was conducted with a separation buffer containing cationic CS. A TBE buffer solution containing CS was added to the buffer reservoir (BR), the buffer waste reservoir (BW), the sample reservoir (SR), and the sample waste reservoir (SW). To avoid hydrodynamic flow,
the liquid levels were identical in all reservoirs. The solution in the SR was replaced by the sample solution after the baseline was stable under a voltage of 2.10 kV. The voltage was applied at the BW in reversed polarity and at the SR, BR, and SW in the grounded state. A stable, reversed electroosmotic flow (EOF) baseline indicated that the dynamic coating process was at equilibrium (i.e., cationic CS in the separation buffer had been adsorbed to the silica channel walls). Using the Multiple-Concentration Method To Detect Bacteria. For field-amplified bacterial stacking, the bacterial cell suspension should be prepared with an appropriate buffer solution with a much lower conductivity compared with the MCE running buffer. Due to the fact that taurine can focus the bacterial cells and thus improve the sensitivity, a small amount of taurine was added into the suspension. Therefore, a TBE buffer containing 20 mM taurine and 0.2 mg/mL CS was selected as the running buffer, and the bacterial suspensions were prepared in a 10-fold-diluted TBE buffer solution containing 10 mM taurine. The higher concentration of taurine in the running buffer blocked the swept bacterial cells and further lowered or eliminated their electrophoretic velocity. A five-step MCE procedure was designed as shown in Figure 1. The voltage configuration of each step was responsible for the preconcentration and stacking in the MCE method. For subsequent experiments, E. coli was selected as the model bacterium to investigate the multiple-concentration phenomenon.
■
RESULTS AND DISCUSSION Bacterial Cell Loading. For sterility tests of the actual sample, the number of bacterial or fungal cells may be very low. Therefore, to ensure that the sample injected into the separation channel contains a sufficient number of microbial cells, a large sample injection volume is necessary. We selected a gated injection method to introduce a sample plug into the separation channel for subsequent concentration and separation with reversed polarity. In addition, a gated injection prevented the contamination of the sample buffer due to the pushback of the separation buffer during the separation step. At the preinjection stage, the voltage of each of the reservoirs was accurately adjusted to ensure that the sample solution flowed from the sample reservoir to the sample waste reservoir and that the running buffer solution flowed from the buffer
1689
dx.doi.org/10.1021/ac202991u | Anal. Chem. 2012, 84, 1687−1694
Analytical Chemistry
Article
Figure 4. Optimization of the CS concentration. The electropherogram conditions were the following: E. coli sample with a concentration of 4.0 × 105 CFU/mL prepared in a 10-fold-diluted TBE buffer containing 10 mM taurine. The running buffer consisted of 22.5 mM TBE with 20 mM taurine and 0.01, 0.05, 0.1, 0.2, 0.5, or 1 mg/mL CS.
Figure 2. Schematic diagram of CS sweeping and field-amplified stacking. See the text for details.
reservoir to both waste reservoirs (Figure 1A). The voltage configuration was then changed to the gated injection state. To avoid CS migration from the BR channel to the intersection, causing a lateral sweeping of the sample plug, the applied voltage of the BR reservoir was set to a lower value than that of the injection cross during the injection. Under these conditions, the running buffer solution in the BR channel moved away from the injection cross, and the bacterial cells and the sample buffer were electrokinetically pumped into the separation channel without lateral sweeping at the injection cross. The bacterial cells also entered the BR channel because of the lower voltage of the BR reservoir (Figure 1B). Meanwhile, fieldamplified stacking occurred in the SW and BR channels because of the difference in conductivity between the sample solution buffer and the running buffer (the black zone in Figure 1B). The signal of the microchip analysis depended on both the sample volume and the concentration factor. A longer injection time introduced more sample solution for the subsequent sweeping step. However, the injection time should be optimized because the limited total length of the separation channel and the retention factor of the bacteria in the micelles are insufficient for concentrating large-volume samples. After careful optimization of the conditions, 10 s was selected as an optimal injection time that allowed complete stacking before the sample reached the detection point. CS Sweeping and Field-Amplified Stacking. After 10 s of gated injection of bacterial cells, the voltage configuration
was changed (Figure 1C). The purpose of the new configuration was to produce a much larger concentration enhancement by providing additional time for stronger interactions between the bacterial cells and the CS micelles. Because of the difference in the ionic strength between the sample buffer and the running buffer, field-amplified sample stacking occurred simultaneously with CS sweeping. During this process, the sample zone in the BR channel entered the separation channel, and two sample plugs formed at the front and the back boundaries of the sample matrix. However, because of the rapid movement of charged species in the lowconductivity matrix, the back sample band had a higher net velocity than the front sample band, and the bands ultimately merged. Field-amplified sample stacking is based on the uneven distribution of the electric field in a continuous buffer system, leading to the rapid movement of charged species in a lowconductivity matrix; this movement is immediately retarded at the interface between the two buffers with different ionic strengths. As shown in Figure 2A, when the running buffer in the BR channel flowed into the analysis channel, the sample plug was positioned between the two running buffers. Bacterial cells were further stacked at the front boundary of the sample plug because of the different electric field strengths of the
Figure 3. Signal enhancement of the multiple concentration of E. coli.: (A) signal intensity without concentration, (B) signal intensity with reversedfield preconcentration, (C) signal intensity with field-amplified preconcentration and sweeping, (D) signal intensity with a combination of fieldamplified stacking, reversed-field stacking, and CS sweeping. The E. coli sample (4.5 × 106 CFU/mL) was prepared in a 10-fold-diluted running buffer consisting of 22.5 mM Tris/borate, 0.5 mM EDTA, 20 mM taurine, and 0.2 mg/mL CS (CS was only present during the sweeping concentration step). 1690
dx.doi.org/10.1021/ac202991u | Anal. Chem. 2012, 84, 1687−1694
Analytical Chemistry
Article
Figure 5. (A) Electropherograms of L. casei at a concentration of 1.30 × 106 CFU/mL processed with only sweeping and field-amplified stacking. (B) Electropherograms of the mixed sample containing E. coli, L. casei, and S. thermophilus (the concentration of each bacterium was 1.30 × 105 CFU/mL) obtained with the multiple-concentration method. The sample buffer and the running buffer were 10-fold-diluted TBE with 10 mM taurine and undiluted TBE with 20 mM taurine and 0.2 mg/mL CS, respectively.
sample and the running buffer zones in the separation channel. In contrast, the positively charged CS micelles migrated against the EOF and swept the bacterial cells from the sample zone. At the initiation of the separation step, the anionic bacteria migrated with the EOF toward the anode. Their movement was immediately retarded at the interface between the two buffers, further stacking the bacterial cells (Figure 2B). When the CS molecules in the front of the sample plug (the anodic side) migrated toward the cathode and passed through the sample zone, the CS molecules were adsorbed on the surface of the bacterial cells, thus giving the bacteria a positive charge. Therefore, the direction of the movement of the microorganisms changed from the anodic side to the cathodic side. These bacterial cells then traversed the sample zone and arrived at the separation buffer zone (on the cathodic side). At the separation buffer zone, cellular aggregation occurred, and large macroparticles or bacterial aggregates formed. The
electrophoretic mobility of these macroparticles was quickly lost because of the large size of the aggregates and the higher concentration of the blocking agent (taurine) in the separation buffer (Figure 2C). Theoretically, the concentration enhancement depends on the conductivity ratio of the two buffers, and a high ratio should be used to maximize the signal enhancement.31 However, an increase in the ionic strength produces a higher current and Joule heating, which are detrimental for the MCE separation. Therefore, the ionic strength increase must be balanced against the negative effect of Joule heating. Furthermore, the conductivity ratio of the sample buffer and the separation buffer affects the distribution of the electric field strength across the channel, influencing the sweeping result. On the basis of these limitations, the ionic strength ratio of the two buffers was optimized on the basis of the observed stacked peaks. Finally, a 22.5 mM TBE solution (pH 8.5) was selected as the running 1691
dx.doi.org/10.1021/ac202991u | Anal. Chem. 2012, 84, 1687−1694
Analytical Chemistry
Article
problem, an additional step, termed reversed-field stacking, was performed in the analysis channel (Figure 1D). Under reversed-field conditions, large macroparticles with low electrophoretic mobilities moved toward the injection cross with the velocity of the EOF. In contrast, the swept zone in the BR channel migrated against the EOF because of its electrophoretic mobility. This resulted in the merging of the two swept zones, circumventing the previously observed tailing of the peaks (Figure 1D). Simultaneously, most of the lowconductivity buffer was pumped from the separation channel into the SW channel. In addition, microchip-based separation systems have short separation distances that are 10−100 times shorter than the separation distances of standard capillary-based systems, limiting the concentration enhancement. The reversed polarity step was maintained for 20 s, similar to the sample time in the front steps, so that the countermigrating band of the swept analytes could move to the injection cross within a short distance, thus lengthening the effective separation. All three steps of the multiple-concentration method were important to achieve the highest bacterial concentration enhancement. Figure 3 displays the electrophoretic profiles of 4.5 × 106 CFU/mL E. coli for the different concentration steps. Without concentration (Figure 3A), only a small peak was observed. For the reversed-field sample stacking (Figure 3B), the bacterial cells were enriched and the signal was clearly enhanced. For the combined CS sweeping and field-amplified stacking (Figure 3C), the signal was greatly enhanced and an additional small peak appeared. For the combination of fieldamplified sample stacking, reversed-field stacking, and CS sweeping (Figure 3D), the cells were completely focused and the signal increased considerably. In the multiple-concentration method, CS was important for the concentration of the bacteria. CS coated the bacterial outer membrane and rendered it positively charged; the bacteria could then migrate from the sample plug toward the anode and were thus concentrated. In addition, the high adsorbing ability of CS induced an aggregation of bacterial cells that further enhanced the concentration of the bacteria. Therefore, the combination of sweeping and aggregation may have synergistically enhanced the concentration of the bacteria. These results demonstrated that the concentration steps allowed the complete focusing of bacterial cells during the electrophoretic process and that the steps were crucial for the detection of low-abundance analytes. Notably, even if the bacteria were not completely focused, they did not appear as multiple small peaks, usually attributable to irregular clusters and aggregates of bacterial cells.9,11 This observation may be due to the low concentration of the bacteria in the sample. Effects of Taurine in the Buffer Solution. A number of different CE approaches have been reported to control the process of aggregation of bacteria and thus improve the reproducibility, selectivity, and sensitivity of microbial analysis.14,32 Some small zwitterionic molecules, including sarcosine, taurine, betaine, and trimethylamine N-oxide, which belong to a class of compounds known as osmolytes (or reverse denaturants), are involved in the stabilization and folding of tertiary and quaternary structures of proteins and therefore significantly affect the charge of proteins on the surface of microorganisms.33 Sakaguchi reported that the adsorption of bacteria to CS was improved when cysteine was used as a blocking agent.34 We tested four small zwitterionic molecules, including L-cysteine, and found that these additives improved the sensitivity of microbial analysis. However, only taurine
Figure 6. Quantification of E. coli suspensions using the multipleconcentration method. The experimental conditions were the following: [E. coli] = (A) 6.0 × 102 CFU/mL, (B) 6.0 × 103 CFU/ mL, (C) 6.0 × 104 CFU/mL, (D) 6.0 × 105 CFU/mL, (E) 6.0 × 106 CFU/mL, and (F) 6.0 × 107 CFU/mL. The samples were prepared in a 10-fold-diluted running buffer. The running buffer consisted of 22.5 mM Tris/borate, 0.5 mM EDTA, 20 mM taurine, and 0.2 mg/mL CS.
Figure 7. Quantification of E. coli in surface water using the multipleconcentration method. The buffers were identical to those listed in Figure 5.
buffer, and a sample buffer concentration of approximately 1/ 10 of that of the running buffer was selected. The lower ionic strength of the sample buffer allowed CS to efficiently coat the surface of the bacteria and therefore effectively sweep the bacterial cells from the sample zone. A sweeping time of 15 s was selected to accompany the field-amplified concentration process. This combination resulted in an expected synergistic effect between the stacking and the sweeping to enhance the peak intensity. After the CS sweeping and the field-amplified sample stacking, the bacterial cells were concentrated in the running buffer matrix instead of in the low-conductivity sample matrix. Reversed-Field Stacking. Because the voltage of the BR reservoir was lower than that of the injection cross, the sample solution entered the BR channel from the SR channel (during the sample injection), where the sample solution was swept by the micelles (Figure 1B). The part of the swept sample zone that existed in the BR channel was also pushed into the separation channel during the field-amplified sample stacking and sweeping steps (Figure 1C). This resulted in the tailing of peaks or the formation of another sample zone. To solve this 1692
dx.doi.org/10.1021/ac202991u | Anal. Chem. 2012, 84, 1687−1694
Analytical Chemistry
Article
close enough that the back sample zone merged into the front sample plug and leaked from the concentration zone during the separation process. However, L. casei could be completely focused when all the concentration steps were conducted. Even when three bacteria were mixed in one sample and then subjected to this method, only a single peak was observed (Figure 5B). Validation of the Multiple-Concentration Method. To evaluate the merits of the multiple-concentration method, the linearity, the limit of detection (LOD), and the reproducibility were determined. Figure 6 displays the calibration curve obtained for the E. coli suspension (prepared with a 10-folddiluted running buffer) using the described multiple-concentration steps. The linearity data were established by injecting a series of bacterial suspensions at different concentrations ranging from 6.0 × 102 to 6.0 × 107 CFU/mL. The slope of the calibration curve was 1.118 with a correlation coefficient of R2 = 0.996. An enhancement factor of approximately 6600-fold was obtained compared to when no concentration was performed (a peak area of 67.3 for 4.5 × 106 CFU/mL without any concentration steps and a peak area of 60.3 for 6.0 × 102 CFU/mL with multiple-concentration steps). The calibration curve equation is log A = 1.118 + 0.260 log C. The correlation between the peak area and the injected amount of cells gradually deviated from linearity, likely because of fluorescence quenching for large numbers of cells (which were injected and focused into a highly concentrated entity). If many types of microbes are present in one sample, this deviation from linearity may be more apparent because of the different aggregation states (of different geometries) and refractive indexes. The aggregation states of different types of bacteria are complex and will be investigated in future studies. The limit of detection, defined as 3 times the background noise, obtained by this electrophoretic method is approximately 145 CFU/mL for E. coli. The reproducibility was determined by repeating the analysis five times with an E. coli suspension at a concentration of 6.0 × 103 CFU/mL. The calculated relative standard deviation (RSD) of the resulting peak area was
