JMD 2001, Vol. 3, No. 4
Copyright © 2001 American Society for Investigative Pathology & Association for Molecular Pathology
Robot Printing of Reverse Dot Blot Arrays for Human Mutation Detection
Stephen Lappin*,
Jeff Cahlik
and
Bert Gold*
From Quest Diagnostics,
*
Van Nuys, and Beckman/Coulter,
Fullerton, California
 |
Abstract
|
|---|
We report on a generally useful, partially
automated, human mutation detection method based upon printing
moderate density oligonucleotide arrays using a biorobot on
activated nylon membranes. The Beckman Biomek 2000 was adapted to this
task through fabrication of aluminum membrane filter holders and the
development of an addressable Tool Command Language (Tcl)
program, which can be invoked through BioScript. During program
execution, a robot arm is moved along the
x, y, and
z axes to expel liquid, without
dripping, from disposable barrier pipette tips and then to
touch the drops on preactivated membranes. Printed arrays
consist of alternating rows of oligonucleotides containing normal and
mutant sequences. Hybridization of biotin labeled polymerase chain
reaction products derived from human patient genomic DNA samples are
visualized using chemiluminescent or chromogenic indicators. This
technique allows unequivocal genotyping of 32 mutations at the
ß-thalassemia locus (11p15.5) and of 34 mutations and one
polymorphism at the cystic fibrosis transconductance membrane regulator
locus (7p35).
|
Introduction
|
|---|
Reverse dot blot (RDB) or reverse allele specific oligonucleotide
(Reverse ASO) hybridization is an important method for genotyping
common human mutations.1, 2, 3
It is especially commonly used
in providing diagnostic genotypes in a high mutation spectrum, medium
or high frequency disorders such as cystic fibrosis, or in the
diagnosis or carrier status assessment of hemoglobin C (HbC),
hemoglobin E (HbE), hemoglobin S (HbS), and
ß-thalassemias.4, 5, 6, 7, 8, 9, 10, 11, 12
In an RDB procedure, exons (or
other regions of interest) are amplified by the polymerase chain
reaction (PCR) using labeled oligonucleotide primers.2
In
all of our work, we have used a 5' biotin label on our PCR
primers13
since we have found that this does not affect
the integrity of the amplification reaction.14
Amplification products (amplicons) are then denatured and allowed to
hybridize with mutation specific DNA probes, covalently bound to solid
supports. Subsequent to hybridization and washing, bound nucleic acids
are incubated with an enzyme conjugated to streptavidin. The bound
enzyme-conjugated, streptavidin-biotin-nucleic acid complex is then
washed and incubated with a chromogenic or luminogenic substrate, which
allows visualization of hybridized spots. Using this method, we can
rapidly determine patient genotypes through a systematic examination of
signals detected in an array.
We were motivated to automate the production of RDB strips when we observed
that array printing constrains the throughput of reverse ASO analytic
procedures and that manufacture is relatively slow, tedious, and error
prone. Because the position of each probe "dot" in the array is
critical for correct genotype assignment, even a small inadvertent
error caused by mistaken manual spotting can cause inaccurate genotype
assignment. On the other hand, automated printing of the reverse ASO
matrix considerably speeds the genotyping process, allows greater
quality assurance (QA) and quality control (QC) over required reagents
and greatly diminishes the likelihood of human error in diagnostic
genotyping.
In this report, we describe a robot-based method for spotting
oligonucleotide hybridization probes on solid membrane supports. The
spotting is accomplished by programming a Beckman Biomek 2000 robot
capable of x, y, and z axis
translation to deposit approximately one microliter aliquots of
oligonucleotide at each position on a membrane array. The deposit is
followed by hybridization with labeled PCR products that can be
visualized using chemiluminescent or colorimetric secondary indicators.
|
Materials and Methods
|
|---|
Specimen Preparation
Total genomic DNA was extracted from peripheral blood leukocytes
using a proprietary method (Gentra Generation Capture Column Kit,
Minneapolis, MN), according to the manufacturer’s instructions.
Amniotic fluid cells or chorionic villi were extracted using salt
precipitation (Puregene Kit, Gentra Systems, Minneapolis,
MN). After purification, DNA was stored in a buffer containing 10
mmol/L Tris-HCl, 1 mmol/L disodium EDTA, pH 8.0 at 4°C.
Oligonucleotide probes were synthesized using a C6-amino-link
phosphoramidite as the final ligand. As a result, the amino moiety was
present on the 5' end of the product. Subsequent to synthesis,
oligonucleotide probes were deprotected, purified,
spectrophotometrically quantified, and dissolved into a spotting buffer
containing 0.5M sodium bicarbonate pH 8.4, 10 mg/ml phenol red, and
0.1% v/v Tween 20. Oligonucleotide stocks were aliquoted and sealed
into U-bottom microtiter plates (containing 250 µl of each probe
solution) and/or 2 ml deep-well plates. Probes are stored frozen at
-20°C. Human ß-globin and CFTR probe sequences are detailed in
Table
1. An indicator spotwasproduced using 5% v/v black India ink suspension in spotting buffer.
Membrane Washing Procedure
Preactivated, 8.5 x 11-inch, 0.45-µm nylon, Immunodyne ABC
membrane was obtained from Pall Corporation (East Hills, NY) (Part No.
BC045H2721). Subsequent to spotting, membranes were air dried and then
immersed in excess 0.1N NaOH with agitation for twenty minutes. The
membranes were then washed three times for 5 minutes each in reagent
grade water. They were then permitted to air dry before being trimmed
for use. Strips were stored desiccated for up to 6 months at -10°C
before use.
CFTR amplifications took place in 50-µl reactions containing 1 x Stoffel buffer (10 mmol/L KCl, 10 mmol/L Tris-HCl, pH 8.3), 5 mmol/L
MgCl2 (Multiplex I) and 6 mmol/L
MgCl2 (Multiplex II and Multiplex III), 200
µmol/L each of dNTP, 10 pmol of each primer, 0.9 µl Taq
Start antibody and 6.6 units of AmpliTaq DNA polymerase, Stoffel
fragment. PCR primers for the three cystic fibrosis multiplex reactions
are detailed in Table 1
.
Thermocycling conditions were uniform for the three multiplexes and
include an initial denaturation step at 94°C for 7 minutes, followed
by 35 cycles of denaturation at 94°C for 30 seconds, annealing at
57°C for 30 seconds and polymerization at 72°C for 1 minute.
Cycling culminated with a final extension at 72°C for 10 minutes.
ß-Globin gene amplifications took place in 25-µl reactions
containing: 1X PCR buffer B; (15 mmol/L
(NH4)2SO4,
2 mmol/L MgCl2, 60 mmol/L Tris-HCl pH 8.5), 200
µmol/L each dNTP, 0.125 µg each primer, and 0.5 units of
Taq polymerase. PCR primers for the three ß-thalassemia
PCR reactions are detailed in Table 1
. Thermocycling conditions were
identical for the three primer pair sets and included an initial
denaturation step at 94°C for 5 minutes, followed by 35 cycles of
denaturation at 95°C for 15 seconds, annealing at 55°C for 30
seconds, and polymerization at 72°C for 45 seconds. Cycling
culminated with a final extension at 72°C for 10 minutes.
CFTR Genotyping, Hybridization, and Visualization
Pooled amplicons from the three multiplex PCRs were added to
membranes pre-wet with hybridization buffer (2X saline sodium citrate
(SSC), 0.1% sodium dodecyl sulfate (SDS) at 42°C. The amplicons and
membranes were boiled and permitted to hybridize for at least 3 hours
at 42°C. Hybridization solutions were poured off and membranes were
washed twice in an excess of wash buffer (0.75 x SSC, 0.1% SDS).
Membranes were then agitated in a freshly prepared conjugate solution
(1:4000 dilution of streptavidin-horseradish peroxidase
conjugate,Boehringer Mannheim, Part No. 1089153, in 0.5X SSC, 0.1%
SDS) for 30 minutes at room temperature. At the end of the incubation,
membranes were washed three times for 3 minutes each with 0.5X SSC,
0.1% SDS; then twice for three minutes each with 0.1 mol/L sodium
citrate, pH 5.0. The strips were then exposed to 0.01% w/v
tetramethylbenzidine dihydrochloride (TMB) substrate (Sigma T-8768) in
a freshly prepared, very dilute solution of hydrogen peroxide
(0.00225% v/v) in 0.1M sodium citrate pH 5.0 with agitation for 10 to
20 minutes in the dark. Color development was stopped by rinsing the
strips four times with deionized water for 3 minutes each. Results were
recorded photographically while the strips were wet.
ß-Globin Genotyping, Hybridization, and Visualization
Pooled amplicons from the three ß-globin PCR sets were added to
membranes pre-wet with hybridization buffer (2 x SSC, 0.1% SDS)
at 42.3°C. The amplicons and membranes were boiled and permitted to
hybridize for one hour at 42.3°C in a shaking water bath. Amplicon
containing hybridization solution was poured off and membranes were
washed twice for 5 minutes each in an excess of hybridization buffer,
followed by two final washes for 10 minutes each in 0.5 x SSC,
0.1% SDS at 42.3°C in a shaking water bath. Strips were then
agitated in a blocking buffer (17 mmol/L
Na2HPO4, 8 mmol/L
NaH2PO4,125 mmol/L NaCl,
5% SDS, pH 7.2) for 10 minutes at ambient temperature. This was
followed by an incubation with a freshly prepared streptavidin-alkaline
phosphatase conjugate solution (1:2000 dilution of
streptavidin-alkaline phosphatase conjugate [New England Biolabs, CDP
Star Kit] in blocking buffer) for 5 minutes at room temperature. At
the end of the incubation, membranes were washed three times in 100
mmol/L NaCl, 10 mmol/L MgCl2, and 100 mmol/L
Tris-HCl, pH 9.5 for five minutes each. The remainder of luminescent
detection was done in accordance with the CDP Star Kit manufacturer’s
instructions.
The Beckman Biomek 2000 was purchased from Beckman Instruments
Corporation (Fullerton, CA). Included in the purchase were: a P20
single-tip pipette tool, an MP20 eight-tip pipette tool, a P200L
single-tip pipette tool, an MP200 eight-tip pipette tool, and a P1000L
single-tip pipette tool. Also purchased were the left and right side
modules, a systems controller computer, the tip disposal optional
software and Bioworks operating system.
Membrane holders were fabricated from 0.100- and 0.150-inch thick
milled aluminum, as described in the text. Pipette tips were purchased
from Beckman and were prepared for spotting by removing and discarding
seven tips from the ninth row of the tip container. The programs
described in this manuscript have been deposited in the Indiana
University Bioarchive ftp site indexed at
http://iubio.bio.indiana.edu/. The Beckman Biomek was programmed using
Tool Command Language (Tcl)15
containing some
Biomek-specific machine commands.
Image analysis was conducted using 16-bit TIFF files obtained through
high resolution (600 pixels) scanning of photographs of stained
membranes. These were saved as grayscale images and adjusted to 640
pixels horizontal to prepare them for the Scion Image version of NIH
Image. Further information about this freely distributed program is
available at http://rsb.info.nih.gov/nih-image/. Ninety manual and 90
automated spots were imaged and measured using a Look Up Table (LUT)
which ranged from 0 (no color) to 255 (black). Five unhybridized
regions were analyzed and averaged to provide a background for each
blot. Differences between background and LUT values were calculated and
summed to provide the statistics in Table 2
. Size measurements were done by micrometer measurement directly against
spotted, developed nylon strips.
|
Results
|
|---|
Aluminum membrane holders were fabricated with machine holes that
match pinions at specific positions within microwell plate stations on
the deck of the biorobot work surface. The machine hole and pinion
placements allow 8.5 x 11-inch preactivated sheets of Immunodyne
ABC membrane to fit securely on the work surface. A washable rubber mat
lies beneath the activated nylon membrane sheet during printing.
Photographs and schematic diagrams of the fabricated membrane holders
are shown in Figure 1
. In Figure 1A
, the membrane holder originally machined for the 64
ß-thalassemia RDB array is shown. Figure 1B
shows the membrane holder
fabricated to accommodate an array that interrogates 34 CFTR mutations
and one polymorphism. After several weeks of operation, we observed
warping of the relatively thin aluminum used for fabrication of the
ß-thalassemia RDB array (approximately 0.100 inch). This convinced us
that a thicker grating would be required in the fabrication of future
membrane holders. As a consequence, the CF membrane holder was milled
to a thickness of 0.150 inch. This has mitigated the warping problem
observed with the ß-thalassemia strip membrane holder. Detailed
blueprints with dimensions for the membrane holder tools are provided
in Figure 1C
, for the CFTR RDB array; and in Figure 1D
, for the
ß-thalassemia RDB array.
View larger version (71K):
[in this window]
[in a new window]
|
Figure 1. Blueprints and photographs of custom fabricated aluminum membrane
holders. A: The membrane holder for the ß-thalassemia
printing array has three chambers, each with enough area to print eight
arrays, each screening for 32 mutations. In sum, RDB arrays
representing 24 patients are printed per 8.5 x 11-inch sheet of
Immunodyne ABC. B: This is the membrane holder for the
cystic fibrosis assay; note the two chambers, each capable of hosting
10 CF arrays. One 8.5 x 11-inch sheet can therefore supply arrays
sufficient to genotype 20 patients at the CFTR locus. C:
Blueprint of the original ß-thalassemia membrane holder tool,
compatible with the Biomek 2000. This was fabricated from approximately
0.100" aluminum at Beckman Instruments, Inc., in Fullerton, CA. After
a short period of use, we noticed that some warping had occurred as a
consequence of the relatively thin aluminum used; therefore, a new
cystic fibrosis membrane holder (blueprint at
D) was fabricated from thicker
aluminum, milled to 0.150 inch.
|
|
Figure
2provides an overview of the Biomek in the process of printing
ß-thalassemia arrays. The MP200 pipette tool on the robot allows
mounting eight pipette tips simultaneously. After the preactivated
Immunodyne ABC membrane is manually placed within the membrane holder,
the Bioworks program can be initiated. The Tcl program, called from
BioScript, first prints eight equidistant spots; the program then
re-aligns the eight-channel mechanical pipettor so that the subsequent
eight spots are pipetted between the previous set of eight spots. The
program continues as described in the legends to Figures 2
and 3
.
View larger version (125K):
[in this window]
[in a new window]
|
Figure 2. Photograph of the Biomek 2000 printing ß-thalassemia arrays. The
MP200’s largest pipetting tool on the Biomek allows mounting eight
tips simultaneously and pipetting these repetitively. After mounting
the preactivated Immunodyne ABC membrane within the membrane holder
borders, the Tcl program first prints eight equidistant spots; the
program then re-aligns the eight-channel mechanical pipettor. The
subsequent eight spots are pipetted between the previous set of eight
spots. The pipettor is then reset to an incremented initial
orientation, one row further down the array, so that the next
eight-spot row will appear adjacent to, but clearly demarcated from,
the initial eight-spot row. Interspersing of spots between the next row
of eight spots is again achieved as noted above. This procedure is
repeated for two subsequent rows of spots. Each array consists of four
rows, where normal and mutant oligonucleotides in the array
alternate.
|
|
View larger version (17K):
[in this window]
[in a new window]
|
Figure 3. Flow chart for RDB spotting Tcl script. After initialization, the core
of the program consists of three nested loops which are recursively
executed. As illustrated in this diagram, tip pick-up, the move to
probe solution, and the move to the destination array are the essential
steps for printing; these are followed by a reorientation step, an
incremental advance in orientation, and a repeated tip pick-up and
printing procedure. Finally, when array printing is complete, an index
spot, consisting of India ink, is printed, as described in Materials
and Methods. This index spot allows for orientation of the RDBs after
an NaOH wash and air drying.
|
|
The flow chart shown in Figure 3
provides a representation of robot
activity during the spotting procedure. At initialization, the robot
arm is oriented at origin and source location, tip location, tool
location, destination array locations, and pipette locations with the
slopes required to provide pick-up specified. The robotic arm then
mounts the eight-tip tool and sufficient barrier tips to pipette
adequate solution for the spotting of one row of strips in the 8.5 x 11-inch array. The mechanical arm then moves to the destination
coordinates and dispenses a small amount (approximately 1 µl of
solution) onto the tip of the pipette, which is then lowered exactly
onto the membrane and then rapidly raised. The destination of the robot
arm is then directed to the next corral. x-y axis
movement resumes after reloading oligonucleotides. Reorientation to a
new corral occurs twice in the ß-thalassemia procedure, but only once
in the CF printing procedure, consistent with the number of chambers in
each membrane holder tool.
Patient specimens presented for cystic fibrosis genotyping at Quest
Diagnostics (Los Angeles, CA) until mid-2000 were routinely evaluated
for 34 mutations and one polymorphism. From 1996 to 1999, this was done
using three reverse dot blot strips that were manually prepared in a
three-strip array, the schematic of which is shown in Figure 4
(C, schematic of the three manual strips). The PCR reactions are
highly, though not completely, multiplexed so that one patient’s CF
gene can be evaluated in three separate PCR reaction tubes. Our efforts
to completely multiplex the PCR reactions were not successful.
Amplicons are then pooled and denatured before hybridizing to the three
reverse dot blot strips, as shown in Figure 4A
. This panel contains one
patient set of three reverse dot blot strips. This manually printed
reverse dot blot has been determined to be that of a G542X/normal
heterozygote. The comparable machine printed RDB strip is shown in
Figure 4B
.
View larger version (70K):
[in this window]
[in a new window]
|
Figure 4. Photographs and schematics of cystic fibrosis genotyping colorimetric
RDB strips. Specimens undergoing cystic fibrosis genotyping at Quest
Diagnostics are routinely evaluated for 34 mutations and one
polymorphism. In the past, this was accomplished using a three strip
array of manually prepared reverse dot blot strips, the schematic of
which is shown in C. The PCR reactions are highly, though
not completely, multiplexed so that one patient’s CF gene can be
evaluated by three separate PCR reactions. In both the old and new
procedures, amplicons are pooled and denatured before hybridizing to
reverse dot blot strips as described in Materials and Methods.
A contains one patient set of three manually printed reverse
dot blot strips. The patient shown is a G542X/normal heterozygote.
B shows the comparable machine printed RDB strip.
C and D provide layout maps of the array for the
three manually printed strips and for the single, robot printed strip,
respectively.
|
|
Automated printing provided an opportunity to simplify layout of the CF
RDB strips. This involved a rearrangement of the 35 probed loci into
alternating normal and mutant rows respectively. Normal RDBs were not
repetitively spotted on automated, printed strips, so that the
hybridization signal would not be reduced by the presence of different
amounts of normal amplicon as the hybridization proceeded (Figure 4D
is
a schematic of the new RDB array for CF with duplicated normal spots
removed; entirely blackened dots represent regions where no DNA was
spotted). As can be seen from the automated printing of the CF reverse
dot blot strip, in Figure 4B
, as compared with Figure 4A
, spots are now
more uniform. Less volume is used for machine versus manual
spotting, thereby representing an economy of oligonucleotide probe
achieved through use of the automated process. In addition, batch sizes
of strips are larger and more accurate than can be manually produced.
Spot-to-spot variability was measured using two different techniques.
The first was to measure spot diameter, using a micrometer, as this is
proportional to the amount of each nucleotide spotted.16
Diameter measurements resulted in a coefficient of variation of 14%
for the manual spots and 11% for the Biomek spots. The NIH
Image Analysis program was also used to judge differences in spot
intensity between manual and machine manufactured arrays. This method
demonstrated greater consistency in average image density using the
Biomek (Table 2)
.
Biomek barrier tips are prepared for printing by removing the lower
seven tips in the tenth row. This provides a way for the (normally)
eight tip-based MP20 tool, to print the single India ink orientation or
index spot. If the tips were not removed, a tool change to a single tip
tool or a risk of confounding the clean strip preparation with
microwell plate-bound contaminants might exist.
Included in the bicarbonate buffered oligonucleotide spotting mix is a
phenol red indicator. The presence of indicator allows the robot
operator a rapid method to monitor missed or incorrectly oriented
oligonucleotide spots. Missed spots are not a frequent occurrence. When
these do occur, however, they can be manually corrected because the
phenol red marks missed positions in the array. On one occasion, we
observed incorrectly oriented angular arrays with overlapping spots.
These proved to be an indication of wear on one robot cable, which was
constraining free movement of the robot arm. Although an error
correcting feedback loop might have prevented this problem, our
troubleshooting revealed that the error correction system does not
function while running a Tcl script. Thus, the indicator dye is useful
because it allows direct observation of array symmetry and an early
warning of problem print runs.
Each strip requires an index or orientation spot because of its
symmetry. Otherwise, the array could be mistakenly oriented and mutant
miscalls could result. The problem of an index spot has been solved by
using a dilute suspension of India ink in spotting buffer. This allows
printing an index spot which survives subsequent washing and
visualization steps in both the chemiluminescent and the colorimetric
procedures.
A 9-pin serial port connection between the Beckman BioRobot and the
controlling computer allows each step in the spotting process to be
monitored. The robot microprocessor, located on semiconductor panels
within the Biomek, records command completion and issues acknowledgment
to the controller computer, using a 9600-band serial port. Thus, robot
action can be monitored in a terminal emulator program window in the
controller computer screen.
Increasing Amounts of Probe Allow Increases in Signal Intensity
From the outset of engineering the reverse dot blot, two
approaches were conceived to improve signal strength from refractory
hybridization of multiplex PCR products at a given ionic strength. The
first approach, taken in perfecting manually spotted strips, is to
alter the amino-conjugated probe sequence to provide improved
hybridization between these and interrogated, labeled PCR products. The
second approach is to seek improved saturation of probe coating on the
solid-support and thereby use the law of mass action to improve the
signal strength. As can be seen in Figure 5
, an experiment aimed at titrating probe attached to a solid support
against uniform amounts of single amplicon was successful. Amino-link
conjugated oligonucleotides spotted at 1, 3.3, 5, 20, 50, and 100
µmol/L show an increasingly strong signal. This demonstrated the
feasibility of using increasing concentrations of a single probe to
improve signal strength. Experiments were then carried out which were
intended to optimize signal strength for each probe.
As indicated in Figures 2
and 6
, ß-thalassemia RDB matrices are spotted on a large activated
membrane. Three columns of eight membranes (each with 64 probes, for a
total of 32 mutable sites interrogated) are spotted on 8.5 x
11-inch sheets. We print a black index spot to reference the relative
position of each strip. The black spot appears as an unexposed region
when the film is developed after exposure to chemiluminescent
substrate. Sample ß-thalassemia reverse dot blot arrays are shown in
Figure 6
. Because of the complexity of the array, manual spotting,
though achievable, is less accurate than the mechanized method
demonstrated here. Thirty-two mutations are analyzed in this reverse
dot blot method, which provides a 95% worldwide detection rate for
ß-thalassemia. The developed strips in Figure 6
, A and B, are from
the same ivs 2,1/negative heterozygote. The strips in Figures 6A
and 6C
were manually spotted, while those in Figures 6B
and D were printed by
the robot. Figure 6C
and D, strips were each hybridized with
amplification products from the same patient who is affected with
sickle cell disease.
View larger version (34K):
[in this window]
[in a new window]
|
Figure 6. Comparison of manually and robot printed ß-thalassemia reverse dot
blot strips. Four developed strips resulting from chemiluminescent
visualization of ß-thalassemia RDBs are shown. Two of these,
A and C, were printed manually, while
B and D were printed using the Biomek biorobot as
discussed in the text. In addition, E is a schematic map
which provides reference for the locations of probes printed within the
array. A and B resulted from hybridizing
amplicons from an ivs 2,1 G > A/normal heterozygous patient;
C and D resulted from hybridization against
amplifications primed with DNA from an HbS homozygous patient. Note the
unexposed reference spot visible in the bottom right corner of
B and D.
|
|
|
Discussion
|
|---|
Inspiration for automating the printing of RDB arrays was a result
of the requirement to provide definitive genotype assignments for the
patients. These assignments must be accurate, reproducible and
relatively inexpensive. Automating the most tedious and critical part
of the process, the printing of the RDB arrays, appeared to us the most
expeditious way to accomplish the objective of mechanizing genotype
assignments generally.
In the course of refining the process for automated RDB printing, we
encountered several technique simplifications that improved the
readability of our array. Among these are inclusion of a low
concentration of non-ionic detergent in our spotting mixture to improve
oligonucleotide distribution on the membrane; programming the deposit
of 1 µl drop of liquid on the pipette tip before touch-down to spot;
and providing a diluted India ink index spot to orient the membrane or
autoradiograph subsequent to visualization.
A method of dispensing a small amount of liquid onto the pipette tips
followed by touch-down onto the membrane allowed a more regular
distribution of oligonucleotide on the membrane. Initially the program
was designed to bring the tool directly onto the surface of the
membrane, with a minimum tolerance between membrane and pipette tip,
and eject the oligonucleotide-containing buffer directly onto the
membrane. However, this technique gave rise to three problems:
variability in the tautness of the membrane sitting within the holder
yielded a lack of uniformity in spot diameter; occasionally, the
presence of the tip tool caused indentations in the membrane; and we
observed a halo effect in developed reverse dot blots, which we believe
resulted from uneven distribution of oligonucleotide on the membrane.
Each of these problems was ameliorated or completely resolved by a
combination of adding Tween 20 detergent to the oligonucleotide mixture
and re-writing the program so that a small amount of liquid was
extruded and left hanging at the lower end of the tips, followed by a
gentle touch-down and rapid pull away from the membrane by the robot
tip holder. This method was called touch-down pipetting. As a result,
strips with the clarity of those shown in Figure 4B
and 6B
and 6D
are
routinely produced.
We believe that the array printing method described here can be
generally applied to the reproducible detection of any point mutation
or small deletion or insertion in any amplified genomic sample. The
general utility of this RDB printing method might allow mechanized
production of screening strips for aldolase B mutations (which cause
hereditary fructose intolerance),17
non-deletion
-thalassemia,18
or adult onset mitochondria disorders
such as Leber’s hereditary optic neuropathy.19
This assay
is flexible, inexpensive to implement, and uses off-the-shelf
commercially available hardware, reagents, and software. While RDB
assays generally are limited by an inability to detect large or
quantitative deletions and an inability to characterize all but modest
expanded repeat sequences, where ascertaining of individual base pair
changes can lead to molecular diagnosis, these strips can provide a
means of accurate and reproducible genotype assignments. Automated
spotting of reverse dot blot strips allows us to print 80 to 100 strips
in one hour with a minimum of operator intervention. Automation has
also allowed us to make higher density strips more accurately than by
hand. In the case of the cystic fibrosis genotyping test, it has
permitted us to array all of the probes required for a complete test on
a single strip. This single-strip array increases the efficiency of
processing the strips for colorimetric visualization and reporting
results on individual patients.
One recent report which created excitement in the molecular diagnostics
community is the observation that the Biomek can be used for printing
PCR product microarrays onto glass slides.20
However, in
the past, immobilized amplicons have been noted to provide less even
results than RDB arrays.1
This is because immobilized
amplification products are generally longer and more subject to steric
hindrance during hybridization than is an immobilized oligonucleotide
on a reverse dot blot. In addition, amplification product
immobilization efficiency in a forward ASO test depends on
amplification success. A compilation of information from multiple
forward ASOs would be required for equivalent CF or ß-thalassemia
genotype assignments to those presented here. We therefore think it
likely that RDBs will remain a significant tool for human mutation
testing for years to come.
|
Acknowledgments
|
|---|
The authors thank Drs. C.S. Chiang, Jean Amos, Wayne Wecksler, and
Peter Noce for their support. In addition, we wish to acknowledge that
Dr. Mark Rabin, Jessica Novak, and Adalbert Kowalski contributed to
early versions of the work described here. Also, Dave Helphrey and Dr.
John Fawcett of Beckman/Coulter provided the initial ideas and
introductions that made this work possible. A preliminary report on
some of the data contained in this publication was presented at a
poster session of the 1998 American Association of Clinical Chemists
meeting in Chicago, Illinois. The abstract accompanying that
presentation was printed in Clinical Chemistry, vol. 44, no. 6,
supplement.21
|
Footnotes
|
|---|
Address reprint requests to Bert Gold, Laboratory of Genomic Diversity, Human Genetics Section, National Cancer Institute at Frederick, Frederick, MD 21702. E-mail: goldb{at}ncifcrf.gov
Biomek and BioWorks are registered trademarks of
Beckman/Coulter, Inc.
Stephen Lappin’s current address is Amgen, 1 Amgen Center Drive,
Thousand Oaks, CA 91320.
Jeff Cahlik’s current address is ACME-Automation.com, Tustin, CA
92782.
Accepted for publication August 15, 2001.
|
References
|
|---|
-
Zhang Y, Coyne MY, Will SG, Levenson CH, Kawasaki ES: Single-base mutational analysis of cancer and genetic diseases using membrane bound modified oligonucleotides. Nucleic Acids Res 1991, 19:3929-3933[Abstract/Free Full Text]
-
Kawasaki E, Saiki R, Erlich H: Genetic analysis using polymerase chain reaction-amplified DNA and immobilized oligonucleotide probes: reverse dot blot typing. Methods Enzymol 1993, 218:369-381[Medline]
-
Sutcharitchan P, Embury SH: Advances in molecular diagnosis of inherited hemoglobin disorders. Curr Opin Hematol 1996, 3:131-138[Medline]
-
Cuppens H, Buyse I, Baens M, Marynen P, Cassiman JJ: Simultaneous screening for 11 mutations in the cystic fibrosis transmembrane conductance regulator gene by multiplex amplification and reverse dot-blot. Mol Cell Probes 1992, 6:33-39[Medline]
-
Cuppens H, Marynen P, De Boeck C, Cassiman JJ: Detection of 98.5% of the mutations in 200 Belgian cystic fibrosis alleles by reverse dot-blot and sequencing of the complete coding region and exon/intron junctions of the CFTR gene. Genomics 1993, 18:693-697[Medline]
-
Chehab FF, Wall J: Detection of multiple cystic fibrosis mutations by reverse dot blot hybridization: a technology for carrier screening. Hum Genet 1992, 89:163-168[Medline]
-
Sutcharitchan P, Saiki R, Huisman TH, Kutlar A, McKie V, Erlich H, Embury SH: Reverse dot-blot detection of the African-American beta-thalassemia mutations. Blood 1995, 86:1580-1585[Abstract/Free Full Text]
-
Wall J, Cai S, Chehab FF: A 31-mutation assay for cystic fibrosis testing in the clinical molecular diagnostics laboratory. Hum Mutat 1995, 5:333-338[Medline]
-
Grody WW, Dunkel-Schetter C, Tatsugawa ZH, Fox MA, Fang CY, Cantor RM, Novak JM, Bass HN, Crandall BF: PCR-based screening for cystic fibrosis carrier mutations in an ethnically diverse pregnant population. Am J Hum Genet 1997, 60:935-947[Medline]
-
Rady MS, Baffico M, Khalifa AS, Heshmat NM, el-Moselhy S, Sciarratta GV, Hussein IR, Temtamy SA, Romeo G: Identification of Mediterranean beta-thalassemia mutations by reverse dot-blot in Italians and Egyptians. Hemoglobin 1997, 21:59-69[Medline]
-
Villalobos-Torres C, Rojas-Martinez A, Villareal-Castellanos E, Cantu JM, Sanchez-Anzaldo FJ, Saiki RK, Barrera-Saldina HA: Analysis of 16 cystic fibrosis mutations in Mexican patients. Am J Med Genet 1997, 69:380-382[Medline]
-
Winichagoon P, Saechan V, Sripanich R, Nopparanta C, Kanokpongsakdi S, Maggio A, Fucharoen S: Prenatal diagnosis of beta-thalassemia by reverse dot-blot hybridization. Prenat Diagn 1999, 19:428-435[Medline]
-
Syvanen AC, Bengtstron M, Tenhunen J, Soderlund H: Quantification of polymerase chain reaction products by affinity-based hybrid collection. Nucleic Acids Res 1988, 16:11327-11328[Abstract/Free Full Text]
-
Mitchell LG, Merril CR: Affinity generation of single-stranded DNA for dideoxy sequencing following the polymerase chain reaction. Anal Biochem 1989, 178:239-242[Medline]
-
Ousterhout J: K Tcl and the Tk Toolkit. 1994 Addison-Wesley Professional Computing, Boston MA
-
Rose D: Microfluidic Technologies and Instrumentation for Printing DNA Microarrays in Microarray Biochip Technology. Edited by M Schena. Natick MA, Eaton Publishing, 2000, pp 19–38
-
Lau J, Tolan DR: Screening for hereditary fructose intolerance mutations by reverse dot-blot. Mol Cell Probes 1999, 13:35-40[Medline]
-
Chan V, Yam I, Chen FE, Chan TK: A reverse dot-blot method for rapid detection of non-deletion alpha thalassaemia. Br J Haematol 1999, 104:513-515[Medline]
-
Schollen E, Vandenberk P, Cassiman JJ, Matthijs G: Development of reverse dot-blot system for screening of mitochondrial DNA mutations associated with Leber hereditary optic atrophy. Clin Chem 1997, 43:18-23[Abstract/Free Full Text]
-
Macas J, Nouzova M, Galbraith DW: Adapting the Biomek laboratory automation workstation for printing DNA microarrays. Biotechniques 2000, 1998:25:106–110
-
Lappin S, Novak J, Rabin M, Cahlik J, Gold B: Automated printing of reverse dot blot arrays for the detection of human b-globin mutations. Clin Chem 1998, 44(Suppl):A32 (Abstract)
Top
Abstract
Introduction
