- 1. Initial ramp to 95° C. and maintain for 1 minute (1.0° C./s to 75° C., 0.1° C./sec to 95° C.).
- 2. Three step thermocycling for 40 cycles (92° C. for 30 sec., 54° C. for 30 sec., and 72° C. for 1 min) or;
- 3. Two step thermocycling for 40 cycles (92° C. for 30 seconds and 60° C. for 60 sec.)

MicroAmp tubes (Applied Biosystems, Foster City, Calif.) with the remaining reaction mixture, designated Macro TaqMan reactions to distinguish them from reactions performed in the microfluidic device, were placed in the GeneAmp PCR System 9700 (Applied Biosystems, Foster City, Calif.) and thermocycled in the 9600 mode. The Macro TaqMan reactions served as macroscopic controls for the reactions performed in the micro fluidic device. The thermocycling protocol was set to match that of the microfluidic device, except that the initial ramp rate was not controlled for the Macro TaqMan reactions.

Once thermocycling was completed, the control and guard lines were depressurized and the chip was transferred onto a glass slide (VWR, West Chester, Pa.). The chip was then placed into an Array WoRx Scanner (Applied Precision, Issaquah, Wash.) with a modified carrier. The fluorescence intensity was measured for three different excitation/emission wavelengths: 475/510 nm (FAM), 510/560 nm (VIC), and 580/640 nm (Passive Reference 1 (PR1)). The Array Works Software was used to image the fluorescence in the micro fluidic device and to measure the signal and background intensities of each 1 nl well. The results were then analyzed using a Microsoft Excel file to calculate the FAM/PR1 ratio for β-actin TaqMan reactions. For conventional Macro TaqMan, positive samples for target DNA were determined using calculations described in the protocol provided by the manufacturer (TaqMan PCR Reagent Kit Protocol). The signal strength was calculated by dividing the FAM/PR1 ratio of the samples by the FAM/PR1 ratio of the controls. A successful reaction was defined as a sample ratio above the 99% confidence threshold level.

III. Results

Initially, AmpliTaq Gold (Applied Biosystems, Foster City, Calif.) was used in TaqMan reactions and FAM/PR1/Control ratios of 1.5-2.0 were produced, compared to Macro TaqMan reaction ratios of 5.0-14.0. Although results were positive, increased signal strength was desired. Therefore, the AmpliTaq Gold polymerase was substituted with DyNAzyme polymerase due to its increased thermostability, proofreading, and resistance to impurities. The standard Macro TaqMan DyNAzyme concentration of 0.025 U/ul was used in the microfluidic experiments. This polymerase change to DyNAzyme produced FAM/ROX/Control ratios of 3.5-5.8. The signal strength was improved, but it was difficult to achieve consistent results. Because it is know that some proteins stick to PDMS, the concentration of the polymerase was increased and surface modifying additives were included. Two increased concentrations of DyNAzyme were tested, 8×(0.2 U/ul) and 4×(0.1 U/ul) the standard concentration for Macro TaqMan, with 100 pg or 10 pg of genomic DNA per nl in the micro fluidic device. Gelatin, Glycerol, and 0.5% Triton-x-100 were added to prevent the polymerase from attaching to the PDMS. The results of the reactions in the micro fluidic device (chip) and the Macro TaqMan controls are shown inFIG. 8.

The microfluidic TaqMan reaction ratios range from 4.9-8.3, while the Macro TaqMan reactions range from 7.7-9.7. Therefore, the signal strength of the TaqMan reactions in chip is up to 87% of the Macro TaqMan reactions. There was no significant difference between 4× or 8×DyNAzyme. The results demonstrate that PCR reactions can be done with greater than 50% signal strength, when compared to the Macro TaqMan reactions, in the microfluidic devices. The results have been consistent through at least four attempts.

EXAMPLE 2Spotting Reagents

I. Introduction

The purpose of the experiment was to demonstrate successful spotted PCR reactions in a microfluidic device. The term “spotted” in this context, refers to the placement of small droplets of reagents (spots) on a substrate that is then assembled to become part of a microfluidic device. The spotted reagents are generally a subset of the reagent mixture required for performing PCR.

II. Procedure

A. Spotting of Reagents

Routine spotting of reagents was performed via a contact printing process. Reagents were picked up from a set of source wells on metal pins, and deposited by contacting the pins to a target substrate. This printing process is further outlined inFIG. 9. As shown, reagents were picked up from a source (e.g., microtiter plates), and then printed by bringing the loaded pin into contact with the substrate. The wash step consists of agitation in deionized water followed by vacuum drying. The system used to print the reagent spots is a Cartesian Technologies MicroSys 5100 (Irvine, Calif.), employing TeleChem “ChipMaker” brand pins, although other systems can be used as described supra.

Pins employed areTelechem ChipMaker 4 pins, which incorporate an electro-milled slot (seeFIG. 9) to increase the uptake volume (and hence the number of printable spots). Under the operating conditions employed (typically, 75% relative humidity and temperature approximately 25° C.), in excess of one hundred spots were printed per pin, per loading cycle. Under the conditions above, the volume of reagents spotted onto the PDMS substrate is on the order of 0.1 nL.

The dimensions of the pin tip are 125×125 μm. The final spot of dried reagent is substantially smaller than this (as small as 7 μm in diameter), yet the pin size defines a lower limit to the readily achievable spot spacing. The achievable spacing determines the smallest well-to-well pitch in the final device. Using such a device and the foregoing methods, arrays with spacings of 180 μm have been achieved. Arrays built into working chips tend to have spacings from 600 to 1300 microns.

Spotting was done using only one pin at a time. The system in use, however, has a pin head which can accommodate up to 32 pins. Printing a standard-size chip (array dimensions oforder 20×25 mm) takes under 5 minutes.

B. Assembly of Spotted Chips

The flow and control layers of the PCR devices are assembled according to the normal MSL process described above. The microfluidic device design is the same as the one described in Example 1. In parallel, a substrate layer composed of 150 μm-thick PDMS with component ratio A:B of 30:1 is formed via spin-coating a blank silicon wafer, and then cured for 90 minutes at 80° C.

The cured blank substrate layer of PDMS (sealing/substrate layer724 ofFIG. 7A) serves as the target for reagent spotting. Patterns of spots are printed onto the substrate, which is still on the blank wafer. The reagents spotted for PCR reactions were primers and probes, specific to the particular gene to be amplified. The spotted reagent included a 1:1:1 volume ratio of 300 nM β-actin forward primer (FP), 300 nM β-actin reverse primer (RP), and 200 nM β-actin probe (Prb). In some cases, it is useful to further tune the chemistry via concentrating the spotted mixture. It has been found that adjusting the concentrations such that primer and probe concentrations are equal to, or slightly higher than, the normal macroscopic recipe value yields consistently good results. Therefore, the spotted reagent is concentrated to be 3 times and 4 times the concentration of the macro reaction. Concentration of the reagents is performed in a Centrivap heated and evacuated centrifuge and does not alter relative FP:RP:Prb ratios. The increased spot concentration results in the correct final concentration when the reagents are resuspended in a 1 nL reaction volume. Spotted reagents need not be limited to primers and probes; nor must all three (FP, RP and Prb) be spotted. Applications where only the probe, or even one of the primers, is spotted can be performed. Experiments have been conducted in which the sample primer/probe sets spotted were TaqMan β-actin and TaqMan RNAse-P.

Following the spotting process onto the substrate layer, the combined flow and control layers (i.e., layers 720 and 722 ofFIG. 7A) were aligned with the spot pattern and brought into contact. A further bake at 80° C., for 60-90 minutes, was used to bond the substrate to the rest of the chip. After the chip has been assembled, the remaining components of the PCR reaction (described in Example 1) are injected into the flow channels of the chip and the chip is thermocycled as described in Example 1.

III. Results

PCR reactions have been successfully and repeatably performed using devices where primer (forward and reverse primers) and probe molecules are spotted. An example of data from a chip in which a reaction has been successfully performed is shown inFIG. 10. The spotted reagents have resulted in successful PCR reactions as defined in Example 1. Successful reactions have been performed using 2-stage and 3-stage thermocycling protocols.

EXAMPLE 3Genotyping

I. Introduction

The purpose of the following experiments was to demonstrate that genotyping experiments can be conducted utilizing a microfluidic device or chip such as described herein. Specifically, these experiments were designed to determine if reactions conducted in the device have sufficient sensitivity and to ensure that other primer/probe sets, besides β-actin, can be performed in the microfluidic device.

II. Methods/Results

A. RNase P Experiment

RNase P TaqMan reactions (Applied Biosystems; Foster City, Calif.) were performed in a microfluidic device as described in Example 1 to demonstrate that other primer/probe sets produce detectable results. RNaseP reactions also require a higher level of sensitivity because the RNaseP primer/probe set detects a single copy gene (2 copies/genome) in contrast to the β-actin primer/probe set. The β-actin set detects a single copy β-actin gene and several pseudogenes, which collectively total approximately 17 copies per genome. The RNase P reactions were run with the same components as described in Example 1, with the exception that the β-actin primer/probe set was replaced with the RNase P primer/probe set. Further, the RnaseP primer/probe set was used at 4× the manufacturer's recommended value to enhance the fluorescence signal. The VIC dye was conjugated to the probe for RNase P and the analysis focused on VIC/PR1 ratios. The results of one of four experiments are shown inFIG. 11.

The VIC/PR1/Control ratios for the Macro TaqMan reactions are 1.23. The corresponding ratios for the TaqMan reactions in the microfluidic device are 1.11 and 1.21. The ratios of the genomic DNA samples in the microfluidic device are above the 99% confidence threshold level. Further, the signal strength of the TaqMan reactions in the microfluidic device is 50% and 93.7% of the Macro TaqMan reactions. The control TaqMan reactions in the microfluidic device have standard deviations of 0.006 and 0.012, demonstrating consistency in the reactions across the micro fluidic device. Therefore, it is determined that the TaqMan reactions in the chip are sensitive enough to detect 2 copies per genome.

B. DNA Dilution Experiment

To further determine the sensitivity of TaqMan reactions in the microfluidic device, dilutions of genomic DNA were tested using the β-actin primer/probe set. Reaction compositions were generally composed as described in Example 1 using 4×DyNAzyme and dilutions of genomic DNA. The genomic DNA was diluted down to 0.25 pg/nl, which corresponds to approximately 1 copy per nl. The result of one dilution series is shown inFIG. 12.

According to a Poisson distribution, 37% of the total number of wells should be negative if the average target number is one.

Well numbers

5, 6 and 7 are below the calculated threshold and, therefore, negative. This suggests that the β-actin TaqMan reactions in micro fluidic chip can detect an average of one copy per nl. Therefore, the sensitivity of the reactions in the microfluidic device is sufficient to perform genotyping experiments.

C. Genotyping Experiment

Because TaqMan in the microfluidic device is capable of detecting low target numbers, preliminary testing of SNP (Single Nucleotide Polymorphism) genotyping was performed using the Predetermined Allelic Discrimination kit (Applied Biosystems; Foster City, Calif.) against the CYP2D6 P450 cytochrome gene. The kit contains one primer set and two probes; FAM labeled for the wildtype or reference allele, CYP2D6*1, and VIC labeled for the CYP2D6*3 mutant or variant allele. Positive controls, PCR products, for each allele along with genomic DNA were tested in the device using the same conditions as described in Example 1. The results from one experiment are shown inFIGS. 13 and 14. The experiment has been repeated at least three times to validate the results and to demonstrate reliability.

As shown inFIG. 13, the Al-1 (Allele 1, CYP2D6*1 wild type allele) and genomic DNA (100 pg/nl) produced an average VIC/PR1/Control ratio of 3.5 and 2.2, respectively, indicating that the genomic DNA was positive for the CYP2D6*1, wild type allele. These values are above the threshold limit for the reactions. The signal strength of the TaqMan reactions in the microfluidic device is 59% and 40% of the Macro TaqMan controls, respectively. Al-2 (Allele 2, CYP2D6*3 mutant or variant allele), which should be negative in the VIC channel, showed some signal over control (1.5), possibly due to FAM fluorescence leaking into the VIC channel of the detector. The leakage can be minimized with an improved detection process.

The Al-2 positive control gave an average FAM/PR1/Control ratio of 3.0, which was 37% of the Macro TaqMan signal and above the calculated threshold limit (seeFIG. 14). The genomic samples were negative for the CYP2D6*3 mutant allele, an expected result since the frequency of the CYP2D6*3 allele is low. Again, it appears that there is some leakage of the Al-1, VIC probe into the FAM channel of the detector. Overall, the SNP detection reactions were successful in the microfluidic device.

EXAMPLE 4Verification of PCR by Gel Electrophoresis

I. Introduction

As an alternative method to prove amplification of DNA was occurring in the microfluidic device, an experiment to detect PCR product by gel electrophoresis was performed. PCR reactions compositions were as described in Example 1, except the TaqMan probe was omitted and the β-actin forward primer was conjugated to FAM.

II. Procedure

A. Microfluidic Device

A three layer microfluidic device, fabricated using the MSL process, was designed and fabricated for conducting the experiments described in this example;FIG. 15 shows a schematic view of the design. Thedevice1500 generally consists of a sample region1502 and a control region1504. Sample region1502 contains three hundred and forty-one 1nl reaction sites1508 represented by the rectangles arrayed alongflow channel1506, which includes inlet via1510 and outlet via1512. Control region1504 contains threecontrol flow channels1514 each containing ten 1nl reaction sites1518, also represented by the rectangles and an inlet via1516. A network ofcontrol lines1522 isolate each

reaction site

1508,1518 when sufficient pressure is applied to inlet via1524. A series ofguard channels1520 are included to prevent liquid from evaporating out of the

reaction sites

1508,1518. The device is a three-layer device as described in Example 1 (seeFIG. 7A). The entire chip is placed onto a coverslip.

B. Experimental Setup

Microfluidic device

1500 was loaded and thermocycled using the 3 temperature profile described in Example 1. The remaining reaction sample was thermocycled in the GeneAmp 9700 with the same thermocycling profile as formicrofluidic device1500. The reaction products were recovered after thermocycling was completed. To recover the amplified DNA, 3 μl of water was injected into sample input via1506 and 3-4 μl of product were removed from outlet via1512. The reaction products fromdevice1500 and the Macro reaction were treated with 2 μl of ExoSAP-IT (USB, Cleveland, Ohio), which is composed ofDNA Exonuclease 1 and Shrimp Alkaline Phosphatase, to remove excess nucleotides and primers. The Macro product was diluted from 1:10 to 1:106. The product fromdevice1500 was dehydrated and resuspended in 4 μl of formamide.

III. Results

Both products, along with negative controls were analyzed, on a polyacrylamide gel.FIG. 15 shows the gel electrophoresis results. The appropriate size DNA band of 294 base pairs in length is observed inFIG. 16.

The products from the Macro reactions are shown on the left hand side of the gel and correspond to about 294 base pairs, the expected size of the β-actin PCR product. The negative controls lack the PCR product. Similarly, the product derived from the device gave the expected β-actin PCR product. Therefore, target DNA was amplified in the micro fluidic device.

EXAMPLE 5Massive Partitioning

The polymerase chain reaction (PCR) has become an essential tool in molecular biology. Its combination of sensitivity (amplification of single molecules of DNA), specificity (distinguishing single base mismatches) and dynamic range (10⁵with realtime instrumentation) make it one of the most powerful analytical tools in existence. We demonstrate here that PCR performance improves as the reaction volume is reduced: we have performed 21,000 simultaneous PCR reactions in a single microfluidic chip, in a volume of 90 pL per reaction and with single template molecule sensitivity.

FIGS. 17a-17ddepict a single bank and dual bank partitioning microfluidic device. where multilayer soft lithography (MSL) (1), was used to create elastomeric microfluidic chips which use active valves to massively partition each of several liquid samples into a multitude of isolated reaction volumes. After injection of the samples intoinlet1703 which is in communication with branchedpartitioning channel system1705 of microfluidic device1701 (FIG. 17b), 2400 90pL volumes1709 of each sample are isolated by closingvalves1707 spaced along (FIG. 17d) simple microfluidic channels. The chip device is then thermocycled on a flat plate thermocycler and imaged in a commercially available fluorescence reader.

We assessed the performance of PCR in the chips by varying the concentration of template DNA and measuring the number of wells that gave a positive Taqman™ signal. We found that a digital amplification is observed when the average number of copies per well is low (FIGS. 18aand18b). A mixture of robust positive and clearly negative signals is observed even when the average number of copies per well is below 1; this implies that even a single copy of target can give good amplification. The number of positive wells was consistent with the number of wells calculated to have ≧1 copy of target by the Poisson distribution (FIG. 19.). This result validates that this system gives amplification consistently even from a single copy of target. Fluorescent signal strengths from microfluidic Taqman™ PCR were comparable to macroscopic PCR reactions with the same DNA concentration—even though the macroscopic reactions contained >10⁴more template copies per reaction.

We believe that the primary source of this remarkable fidelity is the effective concentration of the target: a single molecule in a 90 pL volume is 55,000 times more concentrated than a single molecule in a 5 uL volume. Since the number of molecules of target, n_t, does not change (i.e. n_t=1) and the number of molecules that can produce side reactions, n_s, (i.e. primer-dimers and non-complementary DNA sequences in the sample) is linearly proportional to volume (i.e. n_s∝V), the ratio of target to side reactions is inversely proportional to volume: n_t/n_s∝1/V. Since side reactions are a primary cause of PCR failure (4), the advantage to reducing the volume of the reaction is clear.

PCR amplification from single copies of template has been previously reported (5). However, current methods that achieve reliable amplification from single copies in a macroscopic volumes often require altered thermocycling protocols (e.g. long extension times, many cycles), precautions against mispriming and non-specific amplification (e.g. “hot start” PCR (thermal activation of the polymerase), “booster” PCR, additives to reduce nonspecific hybridization, etc), and are almost always done with two rounds of PCR, where an aliquot of the first PCR is used as template in the second reaction. In contrast, this system achieves reliable amplification from single copies using standard conditions—off-the-shelf primers and probes and a single-round, standard thermocycling protocol. Being completely enclosed, it is also nearly invulnerable to environmental contamination. The ability to do massive numbers of PCR reactions simultaneously provides definite logistical, cost and time advantages compared to macroscopic volumes (1 chip with 21,000 reactions vs. 219 separate 96 well plates, and the associated time, equipment, and tracking infrastructure).

This principle of massive partitioning with a digital PCR readout may be used for absolute quantification of the concentration of target in a sample. It can be used, for example, to genotype a pooled sample of genomic DNA simply by counting the numbers of wells that give a positive for a particular allele, or plurality of alleles as described above. Due to the enhanced resistance to side reactions, it should also be useful in quantifying mutants in a background of wild-type DNA—a problem relevant in cancer detection. The general principle of concentration by partitioning may also be useful in other reactions where detection of single molecules, bacteria, viruses or cells is of interest (e.g. ELISA reactions for protein detection). Digital PCR is described by Brown, et al., U.S. Pat. No. 6,143,496, titled “Method of sampling, amplifying and quantifying segement of nucleic acid, polymerase chain reaction assembly having nanoliter-sized chambers and methods of filling chambers”, and by Vogelstein, et al, U.S. Pat. No. 6,446,706, titled “Digital PCR”, both of which are hereby incorporated by reference in their entirety. The small volumes achievable using microfluidics allow both a massive degree of parallelization and very high target-to-background concentration ratios. High target-to-background ratios allow single-molecule amplification fidelity. These factors suggest that for PCR, smaller really is better.

The invention provides for methods and devices for conducting digital PCR in a microfluidic environment comprising the steps of: providing a microfluidic device having a fluid channel therein, said fluid channel having two or more valves associated therewith, the valves, when actuated, being capable of partitioning the fluid channel into two or more reaction sites or chambers; introducing a sample containing at least one target nucleic acid polymer, actuating the valves to partition the fluid sample into two or more portions, wherein at least one portion contains a target nucleic acid polymer and another portion does not contain a target nucleic acid polymer, amplifying the target nucleic acid polymer, and, determining the number of portions of the fluid channel that contained the target molecule. In preferred embodiments, the microfluidic device comprises an elastomeric material, and more preferably, comprises at least one layer comprising an elastomeric material. In certain preferred embodiments, the microfluidic device further comprises a deflectable membrane wherein the deflectable membrane is deflectable into and out of the fluid channel to control fluid flow within the fluid channel and/or to partition one portion of the fluid channel from another, preferably wherein the deflectable membrane is integral to a layer of the microfluidic device having a channel or recess formed therein, and preferably wherein the deflectable membrane is formed where a first channel in a first layer is overlapped by a second channel in a second layer of the microfluidic device. In some embodiments, the sample fluid contains all of the components needed for conducting an amplification reaction, while in other embodiments, the microfluidic device contains at least one component of an amplification reaction prior to the introduction of the sample fluid. In some embodiments, the microfluidic device further comprises a detection reagent, preferably one or more nucleic acid polymers complimentary to a least a portion of the target nucleic acid polymer, preferably a plurality of different nucleic acid polymers spatially arrayed within a reaction site or chamber of the microfluidic device.

Amplification may be achieved by thermocycling reactions such as PCR, or by isothermal reactions, such as described by Van Ness et al., in U.S. patent application Ser. No. 10/196,740 which has publishes as U.S. 2003/0138800 A1, which is herein incorporated by reference in its entirety for the purpose of teaching an isothermic amplification scheme.

The invention further provides for a protein microcalorimetry assay using a fluorescent dye, for example SYBER green (TM), to measure the conformational changes of a protein, such as denaturation, especially if a protein's denaturation temperature changes when the protein interacts with another moiety such as a ligand or compound or other protein. An additional benefit of using SYBR Green (TM) is that it us used at lower wavelengths than other UV range dyes thus reducing background problems typically associated with many plastic materials.

REFERENCES

1. Unger et al,Science288, 113-116 (2000).
2. The sample channels and control lines are loaded by “blind filling”—PDMS is sufficiently gas permeable that liquid pressurized at a few psi drives the gas out of the channels, leaving them completely filled with liquid. See Hansen et al,PNAS99, 16531-16536 (2002)
3. A 294 bp segment of the human β-actin gene was amplified using a 5′-exonuclease assay (Taqman). Forward and reverse primer sequences were 5′-TCACCCACACTGTGCCCATCTACGA3′ and 5′-CAGCGGAACCGCTCATTGCCAATGG3′, respectively.FIG. 1bwas taken with a TAMRA-based FRET probe,sequence 5′-(FAM)ATGCCC-X(TAMRA)-CCCCCATGCCATCCTGCGTp-3′. The data inFIG. 1cwas taken with a dark-quencher based probe, as large numbers of these primer-probe sets are becoming commercially available. Reactions contained 1×Taqman buffer A (50 mM KCl, 10 mM Tris-HCl, 0.01 M EDTA, 60nM Passive Reference 1, pH 8.3), 4 mM MgCl₂, 200 nM dATP, dCTP, dTTP, 400 nM dUTP, 300 nM forward primer, 300 nM reverse primer, 200 nM probe, 0.01 U/uL Amperase UNG (all from Applied Biosystems, Foster City, Calif.), 0.2 U/uL DyNAzyme (Finnzyme, Espoo, Finland), 0.5% Triton-x-100, 0.8 ug/ul Gelatin (Calbiochem, San Diego, Calif.), 5.0% Glycerol, deionized H₂O and human male genomic DNA (Promega).
4. Quantitative PCR Technology, Chapter on “Gene Quantification”, L J McBride, K Livak, M Lucero, et al, Editor, Francois Ferre, Birkauser, Boston, Mass. p 97-110, 1998.
See E. T. Lagally, I. Medintz, R. A. Mathies,Anal Chem73(3), 565-570 (2001), as well as B. Vogelstein, K. W. Kinzler,PNAS96, 9236-9241 (1999)

It is understood that the examples and embodiments described herein are for illustrative purposes only and that various modifications or changes in light thereof will be suggested to persons skilled in the art and are to be included within the spirit and purview of this application and scope of the appended claims. All publications, patents, and patent applications cited herein are incorporated by reference in their entirety for all purposes to the same extent as if each individual publication, patent or patent application were specifically and individually indicated to be so incorporated by reference.