CLIN. CHEM. 40/9, 1815-1818 (1994)
Oak Ridge Conference
#{149}
PCR in a Silicon Microstructure
Peter Wilding,’
Mann A. Shoffner, and Larry J. Kricka
Devices for performing polymerase chain reactions (PCR)
have been developed for use with photolithographed silicon. Microchambers capable of holding between 5.0 and
10 LL of PCR reagents were constructed by etching specific areas of rectangular silicon chips (17 x 15 mm),
which were then capped with Pyrex glass. These silicon
devices (PCRChips), which were etched to depths of
40-80 m, permitted free flow of fluids in the microchan-
nets and microchambers. Access to the microchambers
was through holes in the silicon. Thermal cycling of the
PCR reagents was achieved by placing the disposable
PCRChip in a small holder containing a computer-controlled Peltier heater-cooler. Successful amplification was
demonstrated by electrophoresis of products in agarose
gel containing ethidium bromide, and the migration of the
product was compared with that obtained in a commercially available thermal cycler. The thermal characteristics
of the silicon, coupled with the high surface area to volume ratio in the new devices, are particularly advantageous features for amplification by PCR.
Indexing Term.: polymerase chain reaction/micromachining/silicon chips/temperature control
The widespread
application
of the polymerase
chain
reaction
(PCR) (1) and allied nucleotide amplification
techniques
(2-4) has resulted in the development of
many devices or systems to provide the thermal cycling
necessary
for these procedures.
These devices (5-7) are
based on a variety of design principles for heat transfer,
including
water baths, air baths, dry blocks (e.g., aluminum), and microdevices
fabricated
from silicon (8).
Widely different
conditions
are used to achieve amplification. Conditions
varied between
these different systems include cycling times, operating temperatures, and
reactant volumes and composition. However, the overall
aim of these developments
has been to reduce cost, time
of reaction, and contamination,
while at the same time
achieving adequate amplification
of the targeted nudeotide sequence for subsequent
detection or analysis.
Recently, we reported results of studies on the control
and manipulation
of the flow of biological fluids in micromachined
silicon structures
(9) as part of our development of microstructures
for various analytical
purposes. Numerous
types of micromachined
devices have
been designed for a diverse range of applications (e.g.,
motors, cantilever
beams, gears). Clinical
laboratory
science currently
has few applications for micromachined devices, but there are strong indications
that
Department of Pathology and Laboratory Medicine, University
of Pennsylvania, Philadelphia, PA 19104.
‘Author for correspondence. Fax 215-662-7529.
Received April 19, 1994; accepted June 21, 1994.
they will play a major role in the future. The pressures
influencing this trend include the need for economy in
manufacture,
consumables, and operation; the movement of the analytical operation
from the laboratory
to
the point-of-care; and the use of much smaller amounts
of body fluid for testing.
Exploitation
of silicon as a substrate for micromachined devices
is well established in the engineering
fields (10-12).
In biomedicine, the list is shorter but
includes
some novel products.
Several companies, in-
cluding Molecular Devices Corp. (13) and i-STAT Corp.
(14), have used silicon as the substrate to develop microdevices for pH measurement
and biosensors,
respectively; others have exploited the thermal conductivity of
silicon to construct thermal cycling devices for use in
PCR (8). In some of these casesthe silicon is subjected to
little or no modification, other than preparation for coating with various reagents and shaping to facilitate use
in a microdevice (14, 15). In others, the silicon is etched
in a specific manner, with use of photolithographic techniques to construct grooves, channels, or chambers
to
facffitate
specimen handling (9). By using thin wafers of
silicon coated with a photoresistant
material
and photographic masks patterned with the structures to be
fabricated,
the silicon can be selectively etched with
etching
solutions such as potassium
tures can be fabricated in a highly
hydroxide.
Strucreproducible form
with submicrometer definition.
We now report the development of a specific micromachined device with an associated
thermal cycler for microscale
PCR.
Materials and Methods
Silicon Chipsand Heater-Cooler Devices
The PCRChips were manufactured by Micrel Semiconductor (San Jose, CA) with standard photolithographic procedures (see Fig. 1) to have one reaction
chamber (4.4 L for a 4O-m etch and 8.9 L for a
80-sm etch). The etched wafers were cut into rectangles
(-17 x 15 mm) and capped with Pyrex glass by anodic
bonding (16). The chips were held in a custom-fabricated device (see Fig. 2) from Faulkner
Instruments
(Pitman, NJ), and located on a thermoelectric
heatercooler assembly incorporating
a single heater-cooler
(TE9501/017/030A
ITI Ferrotec,
Chelmsford, MA) fused
to an oxygen-free
copper block (11.50 x 11.50 mm) containing a 10-kfl thermistor (YSI 44016; Yellow Springs
Instrument,
Yellow Springs,
OH). The heater-cooler
was connected
to a Thermoelectric
Cooler Controller
(Series 3 TC3; Alpha Omega Instruments).
PCR reagents (-15 pL) contained
in a plastic Eppendorf
pipette tip were transferred to the PCRChip by placing
the pipette tip over one of the holes (entry port) in the
CLINICAL CHEMISTRY, Vol. 40, No. 9, 1994
1815
The final concentrations of the reactants
able PCRChip.
were: nucleotides,
200 nnol/L
each;
Taq polymerase,
1.0
U/10 p1; primers,
0.875 moI/L
each; DNA template,
0.2 ngIlO p1. The thermal cycling (usually 35 cycles)
was performed automatically with use of the computercontrolled heater-cooler/controller
unit.
Fig. 1. Single-chamber PCRChip etchedto a depth of 80 m.
chamber volume, -9
tL
overall sIze, 17 x 15 mm.
silicon chip and then applying a slight negative
pressure to the other hole (exit port). During cycling, the
entry and exit ports of the PCRChips were sealed with
small washers mounted on the edge of the chip holder.
Cycling
times were limited
with this system, giving
cycle times of -3 mm (see Fig. 3). After completion
of
the PCR reaction, the PCRChip was emptied by applying a positive pressure to displace the chamber contents
through
a disposable
piece of 30-gauge
polyethylene
tubing
(Clay Adams, Div. of Becton Dickinson,
Parsippany, NJ) into a microcentrifuge
tube. From the centrifuge tube the sample was transferred
to 2% agarose gel
and subjected to electrophoresis
with a Horizon 58 Electrophoresis
system
(GIBCO
BRldLife
Technology,
Gaithersburg, MD) Standard markers
(50 to 1000 bp;
ProMega,
Madison,
WI) were also run for comparison
purposes.
PCR
Before locating the PCRChip on the chip holder, we
mixed the PCR reactants (nucleotides, AmpliTaq DNA
polymerase,
primer, and the bacteriophage
A DNA sample) in tubes and transferred
the mixture to the dispos-
Results
The anodic bonding of the silicon to the glass cover
was effective, and no failures or leaks were encountered
even with fluids heated to >90#{176}C.
The filled PCRChips
were evsiniined with a microscope (Aristomet;
Wild Leitz, Heerbrugg,
Switzerland), and monitored during
thermal cycling with a black and white video camera
(Dage-MTI, Michigan
City, IN) and a video-cassette
recorder (PVM-122;
Sony, Teaneck, NJ). Air bubbles were
rarely noted and when present in small volumes did not
interfere with the PCR reaction. DNA amplified by PCR
in small volumes (<5 p1) were detected
by ethidium
staining.
bromide
Studies with the PCRChip (Fig. 1) included variations
of reactant
composition and temperatures.
During the
latter studies, the temperature in the chambers
was
checked with small temperature-indicating
devices
made of paper coated with specific liquid crystals (Hallcrest, Glenview, IL). Characteristic
color changes
of
these liquid crystal devices (e.g., green to blue) provided
a simple indication (Fig. 4) of the temperature
of the
chip (± 1.0#{176}C).
Evidence
for successful
amplification of
was demon(Fig- 5). The
model DNA target used during the development
stage
was primarily
bacteriophage
A DNA (a 500-bp segment
was amplified),
but amplification of other targets was
also demonstrated
(e.g., F 508 cystic fibrosis mutation).
specific
strated
DNA
sequences
in the PCRChip
by conventional electrophoresis
DIscussIon
The successful demonstration of controlled fluidics in
microchannels
and the development of microfiltration
Spring Activated
ChipClamp
P
Washer
Holes
2 Leads from
Thennlstor
Oxygen Free
Copper Block
2 Leads from TEC
Pedestal which
holds
syrInge durlngN
Thermal Electric
Cooler (TEC)
Pedestal which
holds tubing
during ejection.
To 9-Pin
Serial Port
1816 CUNICAL CHEMISTRY, Vol. 40, No. 9, 1994
Fig. 2. Schematic of holder for silicon
PCRChip incorporatinga heater-cooler
assembly and luer fittings to which syringeswould attach for removal of samples.
12345
Temperature (Degrees C)
70
0
30 60 90 120 150 180 210
Time s
330
390 420 45)
Fig. 3. Typical temperature cycles used during operation of the
siliconPCRChip (one completecycletakes 3 mm).
devices (9, 17) are strong indicators that micromachined
chambers
that allow sample preparation, thermal cycling, and detection of amplified nucleic acids to be incorporated in a single device will be developed for PCR.
Such devices will reduce the cost of DNA amplification
and bring about the development of convenient systems
for basic research and clinical application. The thermal
characteristics
of silicon, coupled with the high surface
area to volume ratios (>100 mm2/10 p1), are particularly advantageous features of the micromachined devices.
Reports by Wittwer et al. (18, 19) indicate that the
total thermal cycling time for PCR can be reduced if the
reactant
volumes are small and the appropriate
sample
temperature cycling is achieved. Our experience supports the expectation that reactant volumes can be reduced to <10 p1, and our current development of thermal cycling is directed at reducing total cycling times to
well under 45 min. Use of appropriate controller and
heater-cooler devices will ensure this.
The intensity of the ethidium bromide-stained bands
in the gel illustrated
in Fig. 5 shows that the PCR
reactions performed in PCRChips were not as efficient
FIg. 5. Electrophoresls of PCR reactants In agarose gel containing
ethidiumbromide.
Lanes 1 and5contain standard markers (1000,750,500,300,150,50 bp);
lane 2 shows evidence
of amplificationof a segment of bacteriophage lambda
DNA (500 bp) in a Perkin-Elmer Model 9600 thermal cycler. Cycle times and
temperaturesin the Perkin-Elmercycler were95#{176}C
for 15 s; 55#{176}C
for 15 s;and
72#{176}C
for I mm, with a 50-mm denaturatlon
penod at 95#{176}C
before cycling and
a 5-mm elongationperiod at 72#{176}C
after cycling. Lanes 3 and 4 show amplification performed in the PCRChIp with identical reagents arid comparable
cyclingtimes and temperatures as used In lane2.
as those carried out in the commercial
cycler (i.e., Perkin-Elmer Model 9600). However, modifications to the
insulation of the heater-cooler device, coupled with appropriate optimization of the reactant
concentrations,
cycling temperatures, and cycling times, should further
improve the amplification achieved with the new device.
The use of microscopy and video-filming to monitor the
filling of the PCRChips was useful and will allow optimization of chamber architecture
in future chip designs.
The detection of amplicate in this work was based on
electrophoresis in agarose gel containing ethidium bromide. However, use of other possible detection systems
(20,21) may eliminate the need to remove the axnplicate
from the PCRChip.
In conclusion, we have demonstrated PCR in a microfabricated device holding <10 .tL of reaction mixture.
We did not achieve amplification equivalent to that
obtained in a widely used thermal cycler. However, the
results clearly indicate the feasibility of microscale PCR
in a device that should facilitate low-cost PCR and automation. Further work is directed at system optimization and testing clinically relevant DNAs (e.g., human
iminunodeficiency virus).
This work is the subject of patents and patent applications assigned to the University of Pennsylvania
and licensed to ChemCore Corp. (Malvern, PA). Some of the work was performed under
a Sponsored Research Agreement from ChemCore Corp to P.W.
and L.J.K. and the University of Pennsylvania,
with full endorsement by the Conflicts of Interest Committee of the University of
Pennsylvania.
P.W. and L.J.K hold minority stock in ChemCore
Corp.
Fig.4. System used to #{243}heck
temperatures in the PCRChip: Three
liquidcrystaldetectorstrips(on thin paper) change color at the
temperaturesindicated(55, 75, and 95#{176}C).
References
1. Templeton MS. The polymerase chain reaction. History, methods and applications. Diag Mol Pathol 1993;1:58-72.
2. Lizardi PM, Guerra CE, Lomeli H, Tussie-Luna I, Kramer FR.
Exponential
amplification
of recombinant
RNA hybridization
probes. Biotechnology 1988;6:1197-202.
CLINICALCHEMISTRY, Vol. 40, No. 9, 1994
1817
3. Backman KC, Wang CNJ. Detecting target nucleic acid in a
sample. Eur patent 1989; 320 308.
4. Panaccio M, Georgesz M, Lew AM. FoLT PCR: a simple PCR
protocol for amplifying DNA directly from whole blood. BioTechniques 1993;14:238-43.
5 HaffL, Atwood JG, DiCesare J, Katz E, Picozza E, Willisims JF,
Woudenberg T. A high performance system for the automation of
the polymerase chain reaction. BioTechniques
1991;1O:102-12.
6. Findlay JB, Atwood SM, Bergmeyer L, Chemeffi J, Christy K,
Cummins T, et al. Automated closed-vessel system for in vitro
diagnostics based on polymerase chain reaction. Clin Chem 1993;
39:1927-33.
7. Wittwer CT, Filinore GG, Hillyard DR. Automated polymerase
chain reaction in capillary tubes with hot air. Nucl Acids Res
1989;17:4353-7.
8. Northrup MA, Ching MT, White RM, Watson RT. DNA amplification with a microfabricated
reaction chamber. In: Digest of
technical papers: transducers
1993 (Proc. 7th mt. conf. on solidstate sensors and actuators). New York Institute of Electrical and
Electronic Engineers, 1993:924-6.
9 Wilding P, Pfahler J, Bau HH, Zemel JN, Kricka U. Manipulation and flow of biological fluids in straight channels micromachined in sfficon. Clin Chem 1994;40:43-7.
10 Wise KD, Najafl K Microfabrication
techniques for integrated
sensors and microsystems. Science 199 1;254:1335-42.
11. Angell JB, Terry SC, Barth PW. Silicon micromechanical
devices. Sci Am 1983;248:44-55.
12. Kricka LI, Nozaki 0, Wilding P. Micromechanics
and
1818 CUNICAL CHEMISTRY, Vol. 40, No. 9, 1994
technology-implications
and applications
in the clinical laboratory. J Int Fed Clin Chem 1994;6:54-9.
13. McConnell HM, Owicki JC, Parce JW, Miller DL, Baxter GT,
Wada HG, et al. The Cytosensor microphysiometer: biological
applications in silicon technology. Science 1992;257:1906-12.
14. Erickson KA, Wilding P. Evaluation of a novel point-of-care
system: thei-STATportable clinical analyzer. Clin Chem 1993;39:
283-7.
15. Tsay YG, Lin CI, Lee J, Gustavson EK, Appelquist H, Magginetti P, et al. Optical biosensor assay (OBA’#{176}).
Clin Chem 1991;
37:1502-5.
16. Wallis
glass-metal
17. Kricka
G. Direct-current
polarization
during field-assisted
sealing. J Am Ceram Soc 1970;53:563-7.
LI, Nozaki 0, Heyner S, Garside WT, Wilding P.
Applications of a microfabricated
device for evaluating
sperm
function. Cliii Chem 1993;39:1944-7.
18. Wittwer CT, Fillmore GC, Garling DJ. Minimizing the time
for DNA amplification by efficient heat transfer to small samples.
Anal Chem 1990;186:328-31.
19. Wittwer CT, Marshall BC, Reed GH, Cherry JL. Rapid cycle
allele-specific amplification: studies with the cystic fibrosis AF
locus. Cliii Chem 1993;39:804-9.
20. Higuchi R, Dollunger G, Walsh PS, Griffith R. Simultaneous
amplification and detection of specific DNA sequences. Biotechnol-
ogy 1992;10:413-7.
21. Livache T, Fouque B, Teoule R. Detection of HN1 DNA in
biological samples by an homogeneous assay: fluorescence measurement of double-stranded
RNA synthesized from amplified
DNA. Anal Biochem L994;217:248-54.