ABSTRACT
The microreaction volumes of PCR chips (a microfabricated silicon chip bonded to
a piece of flat glass to form a PCR reaction chamber) create a relatively high
surface to volume ratio that increases the significance of the surface chemistry in the polymerase chain reaction (PCR). We investigated several surface passivations in an attempt to identify `PCR friendly' surfaces and used those
surfaces to obtain amplifications comparable with those obtained in
conventional PCR amplification systems using polyethylene tubes. Surface
passivations by a silanization procedure followed by a coating of a selected
protein or polynucleotide and the deposition of a nitride or oxide layer onto
the silicon surface were investigated. Native silicon was found to be an
inhibitor of PCR and amplification in an untreated PCR chip (i.e. native
silicon) had a high failure rate. A silicon nitride (Si
3
N
4
) reaction surface also resulted in consistent inhibition of PCR. Passivating
the PCR chip using a silanizing agent followed by a polymer treatment resulted
in good amplification. However, amplification yields were inconsistent and were
not always comparable with PCR in a conventional tube. An oxidized silicon (SiO
2
) surface gave consistent amplifications comparable with reactions performed in
a conventional PCR tube.
As the demands of diagnostic laboratories continue to increase as a result of
rising health costs, the necessity for rapid and inexpensive medical tests has
become increasingly apparent (
1
-
4
). A reduction in tissue or fluid sample volumes, a reduction in test reagent
volumes, high through put of samples, the reduction of contamination and an
increased ease of use through automation are all necessary elements to both
reduce the cost of diagnostic tests and increase the speed in obtaining
results.
The implementation of microfabricated devices in the research laboratory setting
has begun to address these issues in all areas of laboratory testing, from
sample preparation (
5
) to sample reaction, separation and detection (
6
-
9
). By miniaturizing the sample preparation procedure, less of the potentially
infectious specimen is necessary, translating into smaller specimens taken from
patients. By miniaturizing the reaction chamber, a faster reaction (
10
) is possible using only nanoliter to femtoliter volumes of reagents. By
miniaturizing the detection apparatus, less of the reacted sample is necessary
for measurement. The combination of these three microsystems on one
microfabricated device will create a single micro total analysis system ([mu]TAS) performing the tasks of several large instruments (
11
,
11
).
Microfabricated devices are currently being used to perform a variety of
chemical (
13
-
14
) and enzymatic reactions, such as the glucose oxidase reaction (
15
,
16
), ligase chain reaction (LCR) (
17
) and polymerase chain reaction (PCR) (
18
,
19
). These devices can also be used to directly amplify human genomic DNA from
lymphocytes introduced directly into the microchambers (
20
). However, in order for the [mu]TAS to compete with other systems the analytical performance of these
devices must be comparable with or superior to conventional reaction systems. Surface chemistry plays an
especially significant role in any reaction performed inside the
microfabricated devices. The surface to volume ratio in a conventional PCR
reaction tube (e.g. Perkin-Elmer MicroAmpT Reaction Tube) is ~1.5 mm
2
/[mu]l. As the geometry and volume of the reaction chamber is altered this ratio increases ~5-fold in the case of a capillary tube and 13-fold for a PCR chip. No studies that deal specifically with
the passivation (any chemical or physical treatments that render a surface
inert) of surfaces in microfabricated structures and their use with PCR have
been reported. In this paper we examine the effects of silicon and treated
silicon surfaces on PCR and the effectiveness of silicon-glass microfabricated devices used to perform PCR. We further examine
possible surface treatments to passivate the silicon-glass chips in an attempt to identify an inert surface compatible with
PCR.
Silicon powder 325 mesh (Aldrich Chemical Co. Inc. Milwaukee, WI) was used to
test the effect of silicon on PCR. It was treated with one of the following
silanizing agents: SurfaT, AquaSilT, dimethyldichlorosilane (DMDCS), dimethylchlorosilane (DMCS), trimethylchlorosilane (TMCS) (Pierce, Rockford, IL) or SigmaCoteT (Sigma Chemical Co., St Louis, MO). Several hundred milligrams of the silicon powder were incubated at room temperature in an
excess of silanizing agent in a test tube overnight. The excess silanizing
agent was then removed and the silanized silicon pellets were dried overnight in an oven at 70oC. The pellets were then washed 10 times with double distilled, deionized
water and again oven dried overnight. The silanized silicon powder was
aliquoted into separate tubes in batches of 20 mg each and 1 ml polymer
solution was added. The following polymer solutions (10 mg/ml) were prepared in
0.1 M Tris buffer, pH 8.6: poly-[alpha]-alanine, poly-L-aspartic acid, polyglycine, poly-L-leucine, poly-DL-phenylalanine, poly-DL-tryptophan,
poly-L-lysine, polyvinylpyrrolidone, polyadenylic acid, polymalei- mide or maleimide (Sigma Chemical Co.). The silicon was
incubated with the polymer solution at room temperature overnight. The
supernatant was removed and the treated silicon powder was then dried overnight
in an oven at 45oC.
The PCR reactions were performed using a Perkin-Elmer's GeneAmpr PCR Reagent Kit with AmpliTaqr DNA polymerase and run in the Perkin-Elmer GeneAmp PCR System 9600 (Norwalk, CT). A 100 [mu]l reaction mixture contained 72 [mu]l double distilled, deionized, autoclaved water, 10 [mu]l 10* reaction buffer, 200 [mu]M each dNTP, 0.3 [mu]M each primer, 2 ng [lambda] phage control template and 10
U AmpliTaqr DNA polymerase. The sequences of the primers were: primer 1, 5'- GATGAGGTCGTGTCCGTACAACTGG-3'; primer 2, 5'-GGTTATCGAAATCAGCCACAGCGCC-3'. The thermal conditions of
the System 9600 were as follows: one cycle at 94oC for 1 min; 35 cycles at 94oC for 15 s, 60oC for 15 s and 72oC for 1 min; one cycle at 72oC for 5 min. Approximately 4.6 mg of each type of
treated silicon powder was measured into separate reaction tubes. The PCR
reaction mixture (28 [mu]l) was added to each tube and then all were cycled in the System 9600 using
the above conditions.
The amplified products were detected using a 2% agarose gel (N930-2774; Perkin-Elmer) in 1* TBE buffer, pH 8.3. The 1* TBE buffer contained 100 mM Tris, 90 mM boric acid and
1.0 mM EDTA (Life Technologies, Grand Island, NY). The gel was stained with 1 [mu]g ethidium bromide (Sigma Diagnostics, St Louis, MO) per 10 ml gel. The
samples were electrophoresed at 100 V for ~30 min.
Silicon chips were fabricated by the Alberta Microelectronic Center (Edmonton,
Alberta, Canada) using standard photolithographic procedures (
21
). Each 14 * 17 mm chip was etched to a depth of 115 [mu]m. The surface-polished Pyrextm glass cover (14 * 17 mm; Bullen Ultrasonics Inc., Eaton, OH) was
essential to guarantee good anodic bonding and to minimize non-specific adsorption. The silicon chips were soaked in a H
2
SO
4
/H
2
O
2
(2:1 v/v) bath at 120oC and then washed several times with >1 l deionized, distilled water. The
silicon chips were placed on an aluminum plate that was heated to 500oC on an insulated hot plate (PC-300; Corning, Corning, NY). The temperature was monitored using a
surface thermometer (Hallcrest, Glenview, IL). Pyrex glass covers were placed
on top of each silicon chip. The silicon and glass were anodically bonded by
applying 1000 V (with a current of <1 mA) throught the aluminum plate and glass cover. A d.c. Kepco power pack (APH
1000M; Kepco Inc., Flushing, NY) was used to apply the necessary voltage.
Thermal cycling of the PCR chips was performed using a custom fabricated device from Faulkner Instruments (Pitman, NJ). The device was capable of thermal cycling four PCR chips simultaneously. This
device incorporated a Peltier heater/cooler (9500/071/040; ITI Ferrotec,
Chelmsford, MA) centrally located under an oxygen-free copper block (40 * 40 mm) containing a 10 k[Omega] thermistor (YSI 44016; Yellow Springs Instruments, Yellow
Springs, OH). It was necessary to keep a constant air flow of ~40 l/min under the thermal cycling device to dissipate heat. The air flow
was monitored by a flowmeter (Gilmont F-400; Gilmont Instruments Inc., Barrington, IL). The heater/cooler and
thermistor were connected to a Modular Laser Diode Controller (LDC-3900; ILX Lightwave, Boceman, MT) through an RS232 interface. The Laser
Diode Controller was connected to a 486 PC through a GPIB interface, where it
was controlled using a virtual instrument built on LabVIEW for Windows (National Instruments, Austin, TX). The virtual instrument automated thermal cycling, giving cycle
times of ~3 min.
The powdered silicon surface treatments which qualitatively showed the greatest amount of amplification were then applied to the PCR chips. The
silanizing agent was pipetted into the PCR chip and allowed to incubate at room
temperature for 15 min. It was then removed by applying a negative pressure to
the exit port. This vacuum was applied overnight to remove all of the
silanizing agent and to dry the PCR chip. The chips were then washed with 1 ml
autoclaved, double distilled, deionized water. The chips were again emptied and
dried by applying a vacuum to the exit port. Each PCR chip was then filled with
one of the selected polymer solutions and incubated at room temperature for at
least 1 h. The polymer solution was removed using a negative pressure applied
to the exit port. The PCR chip was then rewashed with 0.5 ml autoclaved, double distilled, deionized water. The chip was emptied by again
applying a negative pressure to the exit port of the chip.
Wafers used to fabricate the PCR chips were also treated with 1000 Å thick layers of silicon oxide (SiO
2
) or silicon nitride (Si
3
N
4
) using standard deposition techniques (
22
,
23
) (Alberta Microelectronics Center). The silicon and glass were then anodically bonded as
described above. The glass covers of these oxide- and nitride-coated PCR chips were not treated.
A 100 [mu]l reaction mixture contained 74.5 [mu]l double distilled, deionized, autoclaved water, 10 [mu]l 10* reaction buffer, 200 [mu]M each dNTP, 0.3 [mu]M each primer, 1 ng template (either [lambda] phage DNA or
Campylobacter jejuni
bacterial DNA) and 2.5 U AmpliTaqr DNA polymerase. The primers used for the amplification of [lambda] phage DNA were as described above. The sequences for amplification of
C.jejuni
bacterial DNA were: primer 1, 5'-CTTCAGGGATGGCGATAGCA GATAG-3'; primer 2, 5'-GCACTGAACCAATGTCGGCTCTGAT-3'. Approximately 10 [mu]l of reaction mixture
were pipetted directly into the entry port of the PCR chips. The chips were
then positioned on the thermal cycling device and sealed with silicon rubber
gaskets. The chips were held in place by spring-activated clamps incorporated in the thermal cycling device. Thermal
cycling conditions for amplifing [lambda] phage DNA were as follows: one cycle at 94oC for 1 min; 35 cycles at 94oC for 15 s, 60oC for 15 s and 72oC for 1 min; one cycle at 72oC for 5 min. Thermal cycling conditions for
amplifing
C.jejuni
bacterial DNA were as follows: one cycle at 94oC for 1 min; 35 cycles at 94oC for 30 s, 60oC for 30 s and 72oC for 1 min; one cycle at 72oC for 5 min.
Positive controls of the PCR were run in parallel to the PCR chips in the Perkin-Elmer GeneAmp PCR System 9600 in order to ensure direct comparability with
results obtained from the PCR chips. A 10 [mu]l reaction volume was used for the positive control. Additionally, the
GeneAmp PCR System 9600 was programed to mimic the thermal profile (identical
ramp and hold times) of the PCR chip thermal cycler.
After cycling, the reaction mixtures were removed from the PCR chips using a
custom fabricated device (Faulkner Instruments). This device clamped the PCR chips in place while a positive pressure was
applied to the entry port of the chip, causing the reaction mixture to be
ejected through a polyethylene tube (Clay Adams, Parsippany, NJ). The amplified
mixtures were collected in polypropylene microcentrifuge tubes. Amplified
products were detected using a 2% agarose gel (N930-2774; Perkin-Elmer) in a 100 mM Tris, 90 mM boric acid, 1.0 mM EDTA buffer, pH
8.3 (Life Technologies). The gel was stained with 1 [mu]g ethidium bromide (Sigma Diagnostics) per 10 ml gel. The samples were run
at 120 V for ~30 min.
Initial tests using silicon powder indicated that untreated silicon is an
inhibitor of PCR (Fig.
1
, lane 2). In Figure
1
, lane 1 represents the positive control, where no silicon was present in the
reaction mixture, and lane 2 represents the reaction that included ~4.6 mg powdered silicon in the reaction mixture. Subsequent lanes (lanes 3-13) show signals from reactions conducted with silicon powder coated
with the silanizing agent SurfaSiltm followed by coating with polymer solution. PCR reactions conducted with
silicon powder coated with the silanizing agent SigmaCotetm followed by polymer solution showed poorer amplification compared with
the use of SurfaSiltm alone (data not shown).
Use of silanization and polymer coatings.
Surface treatments identified as `PCR friendly' were then applied to the PCR
chips under the assumption that PCR would react to surface treatments in the
same manner on silicon chips as they did on silicon powder. We selected the
following surface treatments: SurfaSiltm followed by treatments with poly-[alpha]-alanine, polyadenylic acid or polyvinylpyrrolidone;
SigmaCotetm followed by treatments with polyglycine, poly-L-leucine or polyadenylic acid. The yields of products from
reactions run in the treated PCR chips paralleled the yields from reactions run
with treated silicon powder in reaction tubes in the System 9600 (data not
shown). For example, where silicon powder coated with SurfaSiltm and polyadenylic acid gave higher yields of amplified product than
silicon powder coated with SurfaSiltm and polyglycine, PCR chips coated with SurfaSiltm and polyadenylic acid gave higher yields of amplified product
than PCR chips coated with SurfaSiltm and polyglycine. The same was true in selecting a silanizing agent for
the PCR chips. PCR chips treated with SurfaSiltm and a polymer gave much stronger amplification signals than those
treated with SigmaCotetm.
Figure
2
shows results from four PCR chips cycled together. The results indicate that an
untreated PCR chip (native silicon) (Fig.
2
, lane 2, and Fig.
3
, lane 2) displays some degree of inhibition of Taq DNA polymerase. PCR chips
coated with `PCR friendly' reagents such as SurfaSiltm followed by polyadenylic acid or polyvinylpyrrolidone (Fig.
2
, lanes 3 and 4 respectively) produce amplifications that are equivalent to
those run in conventional polyethylene tubes (Fig.
2
, lane 1). Lane 5 displays the results from an amplification in a PCR chip
treated with SurfaSiltm without any polymer treatment. Although signals like this were common,
amplifications using silane-treated PCR chips were inconsistent, giving yields that were not
quantitatively uniform and variable from run to run. However, occassional
signals were obtained that were even greater than those obtained from the
conventional system. These irregular results were most likely due to a lack of
uniformity in the treatments of the silicon chips. There were often visible
variations in the surfaces of these chips, most likely the result of the manual
methods used to coat them. It was also possible that the surface treatments may
have degraded during thermal cycling. As a result, we investigated procedures
that were standardized and more easily quality controlled.
The authors would like to thank the Perkin-Elmer Corp. for the use of their GeneAmpr PCR System 9600. We would also like to thank Dr Irving Nachamkin and Miss H.Ung
(Department of Pathology and Laboratory Medicine, University of Pennsylvania)
for providing the
C.jejuni
chromosomal DNA used in these studies. 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 PW and LJK and the University of Pennsylvania, with full endorsement by the Conflicts
of Interest Committee of the University of Pennsylvania. PW and LJK hold
minority stock in ChemCore Corp.
REFERENCES
Return