Article pubs.acs.org/ac
Point-of-Care Assay Platform for Quantifying Active Enzymes to Femtomolar Levels Using Measurements of Time as the Readout Gregory G. Lewis, Jessica S. Robbins, and Scott T. Phillips* Department of Chemistry, The Pennsylvania State University, University Park, Pennsylvania 16802, United States S Supporting Information *
ABSTRACT: This Article describes a strategy for quantifying active enzyme analytes in a paper-based device by measuring the time for a reference region in the paper to turn green relative to an assay region. The assay requires a single step by the user, yet accounts for variations in sample volume, assay temperature, humidity, and contaminants in a sample that would otherwise prevent a quantitative measurement. The assay is capable of measuring enzymes in the low to mid femtomolar range with measurement times that range from ∼30 s to ∼15 min (lower measurement times correspond to lower quantities of the analyte). Different targets can be selected in the assay by changing a small molecule reagent within the paper-based device, and the sensitivity and dynamic range of the assays can be tuned easily by changing the composition and quantity of a signal amplification reagent or by modifying the configuration of the paper-based microfluidic device. By tuning these parameters, limits-of-detection for assays can be adjusted over an analyte concentration range of low femtomolar to low nanomolar, with dynamic ranges for the assays of at least 1 order of magnitude. Furthermore, the assay strategy is compatible with complex fluids such as serum. Measurements based on “time” are the least developed of these unconventional readouts, yet “time” is a readout that can be measured in a number of ways, even without the use of electronic devices. Consequently, herein we describe the development of a new approach for quantifying enzyme analytes by measuring the time that is required for one region on a paper-based device to turn color after a first region turns color (see the Abstract image). This strategy is demonstrated in assays for active enzymes.35 The intensity of the color in either region for this assay is not indicative of the quantity of the analyte; rather, the quantity of the analyte is directly related to the relative time required for the color to appear. This relative measurement enables assays that are internally calibrated for effects of temperature, humidity, and sample viscosity on sample distribution,36−39 and the overall approach requires only that a user begin a measurement once the assay region turns color, thus leading to short measurement times (seconds to minutes), with overall assay times (from application of the sample to the completed test) requiring only 15−30 min. This new assay strategy also offers a remarkable level of sensitivity (femtomolar),3 is selective, is inexpensive (the device is made from paper and microgram quantities of reagents), and is easy to use (the user need only add the sample to the device and then time how long it takes for one region to turn color after the first turns color). Thus,
W
hile qualitative point-of-care (POC) assays are available in the form of dipsticks and lateral-flow tests, quantitative assays pose practical challenges that have been difficult to overcome in an inexpensive and convenient way.1,2 The ideal quantitative POC assay, particularly for use in extremely resource-limited environments such as remote villages in the developing world,3−6 not only should be inexpensive, straightforward to operate, and provide rapid and reproducible quantitative results but also should do so without the use of an external “reader”.7−14 This goal of “reader-less” quantitative POC assays represents a formidable scientific and technical challenge. A key question to address toward this end relates to the type of readout that should be produced by an assay so that the readout is easy to quantify without using electronic devices. Standard spectroscopic and electrochemical analyses typically require an external device to obtain a quantitative result,9,11,13−16 but more recent studies have focused on measurements based on “distance”, “time”, or “the number of regions that turn color” on a device. All three of these outputs can be quantified, in theory, by counting, which represents a first step toward achieving quantitative point-of-care assays that do not require electronics. Representative examples of quantitative and semiquantitative assays (not all are POC assays) based on counting include time-based assays,17−19 semiquantitative20−23 and quantitative24 assays that require counting colored regions, as well as distance-based measurements in the context of nanomotor,25 microfluidic,26−28 spot tests,29 and immunochromatographic and lateral-flow assays.30−34 © XXXX American Chemical Society
Received: August 1, 2013 Accepted: September 27, 2013
A
dx.doi.org/10.1021/ac402415v | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
Figure 1. Detailed depiction of the assay device as well as the reagents used in the assays. (a) Photograph of a three-dimensional paper-based microfluidic device19,24,36,37,44−46 for quantifying active enzyme analytes by measuring the relative time required for a sample to turn a control region green (right-hand region) relative to when an assay region (left-hand region) turns green. The device is made from stacked layers of wax-patterned paper44 that are held together using spray adhesive47 and then laminated. The dimensions of the paper portion of the device are 20 mm × 10 mm × 1.8 mm; the black regions are hydrophobic wax, and the white regions are hydrophilic paper. The white dotted line shows the location of the crosssection depicted in (b). The left channel in (b) is the assay region and the right channel is the control region. (c) Specific substrate reagents are incorporated into the device to provide selective detection of the target enzyme analyte. (d) A hydrophobic oligomer is used to amplify signal for the detection event. Amplification arises from head-to-tail depolymerization19,48−52 in response to hydrogen peroxide that is generated during the detection event (c).19,24,53,54
as well as (ii) the cofactor MgCl2, which is needed by certain enzyme analytes (other cofactors could be added for different enzyme targets). In the left-hand channel (the assay region) in Figure 1b, the sample redissolves a substrate for a target enzyme analyte (compound 1, Figure 1c),55 beginning in layer 3. If the target enzyme is in the sample, it reacts with this substrate and causes release of one molecule of glucose per enzymatic reaction. Once the sample continues through layers 3 and 4 and into layer 5, it encounters bead-bound glucose oxidase (dark blue in Figure 1b,c), which remains immobilized in the fibers of the paper.13 The glucose oxidase (GOX) oxidizes the released
the strategy offers a step forward toward the goal of conducting quantitative point-of-care assays without using auxiliary instruments or electronics.
■
EXPERIMENTAL DESIGN An example configuration of a paper-based microfluidic device40−44 that we designed for this assay is depicted in Figure 1. This device has an entry point for addition of the sample, and hydrophilic channels of paper that split the sample into two equal directions (i.e., layer 2, Figure 1b). Layer 2 also includes (i) buffer salts that are redissolved by the sample to control the pH of the fluid as it distributes through the device, B
dx.doi.org/10.1021/ac402415v | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
Experimental details, device fabrication procedures, and tabulated data are available in the Supporting Information.
glucose and generates hydrogen peroxide as the sample travels laterally in layer 5 of the device. Once the sample reaches the vertical conduit on the far left-hand side of the device in Figure 1b, it encounters an oligomer (compound 2, Figure 1d) that is hydrophobic and thus alters the wetting properties of the paper from hydrophilic to hydrophobic.19,24 In the absence of hydrogen peroxide, the sample travels slowly through this hydrophobic region, but in the presence of hydrogen peroxide, reagent 2 converts to hydrophilic products through a cascade depolymerization reaction,19,48,50,56,57 thus switching the wetting properties of the paper from hydrophobic back to hydrophilic. This switching reaction amplifies the effects of hydrogen peroxide on the flow rate through layers 4 and 3 by converting a large hydrophobic oligomer into hydrophilic products.19,48 This switching reaction also allows the sample to pass through the layers containing 2 with a rate that depends on the concentration of hydrogen peroxide in the sample, which ultimately reflects the concentration of the target enzyme analyte. Once the sample passes through the layers containing 2, it continues to travel in the vertical direction until it redissolves dried green food coloring and carries the highly colored solution to the top layer where the bright green color becomes visible. The control region (right-hand channel in the cross-section in Figure 1b) contains the same reagents in the same order as the assay region, with the exception of bead-bound glucose oxidase. In this control region, the enzyme analyte (if present) will react with substrate 1 deposited into the channel and generate glucose,55 but hydrogen peroxide will not be generated; therefore, hydrogen peroxide will not be present to react with oligomer 2. Hence, the time required for the sample to pass through this control region (and carry the green color to the top of the device) depends on the temperature and humidity under which the assay is conducted, as well as on the viscosity of the sample.37−39,58 These factors will affect sample distribution rates in the assay region as well (the left-hand channel); therefore, this control region normalizes the output of the assay for the effects of these variables on sample distribution. This normalization is implemented by measuring the time required for the control region to turn green relative to when the assay region (the left-hand region) turns green (i.e., Tmeasurement). This measurement time is different than the total time (Ttotal) for the sample to pass from the entrance of the device to the end of the control region (the right-hand region in Figure 1). The measurement time (Tmeasurement) also is different than the assay time (Tassay), which is the time required for the sample to pass from the entrance of the device to the end of the assay region (the left-hand region in Figure 1). The relationship between Tmeasurement, Ttotal, and Tassay is depicted in eq 1. Tmeasurement = Ttotal − Tassay
■
RESULTS AND DISCUSSION Demonstrating Selectivity in the Assays. The performance of the assay is revealed by the calibration curves shown in Figure 2 for the model enzyme analytes alkaline phosphatase (a
Figure 2. Calibration curves for (a) alkaline phosphatase and (b) β-Dgalactosidase, both of which are model enzymes to demonstrate the quantitative assay. The calibration curves were obtained at 19 °C and 20% relative humidity using 2d as the phase-switching reagent and compound 1a for alkaline phosphatase and 1b for β-D-galactosidase. The data points are the average of three measurements, and the error bars reflect the standard deviations of these averages. The insets for (a) and (b) provide an expanded view of the regions that are bracketed using a dotted line. The equation for the line in (a) is y = 0.591x + 0.349, and the equation for the line in (b) is y = 0.122x + 0.436. The data for alkaline phosphatase is black, catalase is green, and β-Dgalactosidase is blue. Values for all of the enzyme assays, in U/L, are available in the Supporting Information.
marker in blood that is indicative of liver function)59−61 and β62 D-galactosidase (a general marker of fecal coliforms in water). In Figure 2a, the analyte is alkaline phosphatase, which uses 1a as the small molecule substrate in the device and 2d as the phase-switching reagent. The calibration curve was generated by depositing samples of alkaline phosphatase in 40 mM HEPES buffer (pH 8.0) to the top of the device and measuring Tmeasurement for the control region to turn green after the assay region turns green. The limit-of-detection for this assay is 320 pM (0.355 U/L) alkaline phosphatase, with a dynamic range of 320 pM (0.355 U/L) to 14.8 nM (16.3 U/L).63 This level of sensitivity far exceeds the sensitivity of colorimetric assays that use camera-equipped cellular phones for quantifying assays in
(1)
Ttotal and Tassay are not measured during an assay and vary depending on the conditions for the assay. Typically Ttotal and Tassay range from 15 to 30 min for Ttotal and from 2 to 15 min for Tassay. The only reason to discuss either of these two times is because Tassay is the period of time in which a user must watch for the appearance of green color in the assay region (the lefthand region in Figure 1) so that Tmeasurement can be made once the assay region turns green. The actual measurement time (Tmeasurement) typically is seconds for low concentrations of the target analyte up to 15 min for high concentrations. C
dx.doi.org/10.1021/ac402415v | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
paper microfluidic devices59 and is even more sensitive than comparable activity-based assays that use a glucose meter to obtain the quantitative readout.55 Perhaps more importantly, the measurement time (Tmeasurement) is proportional to the concentration of the analyte, which is a feature that enables rapid trace-level detection, even if only a qualitative result is desired. Moreover, the assay is selective for alkaline phosphatase, as revealed by the negligible response when catalase or β-Dgalactosidase are added to the device instead of alkaline phosphatase (green and blue data in Figure 2a, respectively). Catalase was used for comparison to alkaline phosphatase since it decomposes hydrogen peroxide rapidly to water and oxygen64 and thus should not provide a measurable response if the mechanism of the quantitative assay relies on analyteinduced production of hydrogen peroxide. β-D-Galactosidase was chosen because it belongs to a different enzyme family than alkaline phosphatase and therefore demonstrates that selectivity between classes can be achieved. If the substrate in the device (i.e., 1) is switched to detect an enzyme other than alkaline phosphatase, then the selectivity switches as well (Figure 2b). The calibration curve in Figure 2b is for the enzyme β-D-galactosidase, which uses 1b (Figure 1c) as the substrate in the device. The limit-of-detection for this assay is comparable to the alkaline phosphatase assay (i.e., 1.94 nM; 693 U/L), with a similar dynamic range (1.94 nM (693 U/ L) to 43 nM (15 360 U/L)).63 More importantly, this second calibration curve demonstrates that the assay can be reconfigured easily by changing the activity-based detection reagent (i.e., 1) to target a variety of enzymes.55 Demonstrating Tolerance to Sample Volume. An important feature of the assay is its ability to provide quantitative results without requiring precise measurements of sample volume (Figure 3). The patterned hydrophilic paper in
has a negligible effect on the results of the quantitative assay (blue data in Figure 4a), as anticipated based on inclusion of
Figure 4. Effect of humidity and temperature on the accuracy of measuring Tmeasurement when using 11 nM (12 U/L) alkaline phosphatase as the model enzyme analyte for the humidity assay and 4.0 nM (4.0 U/L) for the temperature study. The assays were conducted using devices that contained 1a and 2d. The data points are the average of three measurements, and the error bars reflect the standard deviations from these averages. The inset in (b) depicts the linear region of the “temperature vs. Tmeasurement” graph. The equation for the line is y = 0.3079x − 3.6831.
the control region. Variations in temperature, however, do affect the time required for a sample to flow through the device. Temperature-induced changes in sample viscosity affect sample distribution rates37−39,58 but are likely accounted for by the control region, whereas temperature effects on enzymatic activity are not. Instead, as depicted in the black data in Figure 4a, Tmeasurement tracks with the anticipated activity of the target enzyme (alkaline phosphatase in this case).66 The effects of assay temperature on enzymatic activity (and, ultimately assay time) can be accounted for easily, however, since there are three temperature regimes (within the ranges that we tested) that affect the assay: (i) 33 °C. The enzymatic activity is equal within the first temperature range, so no adjustment is needed to the calibration curve in Figure 2a with the exception of adding 1 min to Tmeasurement so that the measurement times match the calibration curve (Figure 2a), which was generated at a warmer temperature (i.e., 19 °C) than this cold temperature range (
