1
1
Supplemental Data (Methods, Tables and Figures)
2
3
Methods
4
COLLECTION OF BLOOD AND TISSUE SAMPLES
5
Chorionic villus samples (CVS) were collected during conventional prenatal diagnosis
6
sessions in the first trimester of pregnancy. Placental tissue samples were collected from
7
euploid third-trimester pregnancies after delivery and from euploid and trisomy 21
8
pregnancies after termination of pregnancy. The fetal chromosomal status was confirmed by
9
full karyotyping. Tissue samples were stored at -80 C until extraction. Maternal peripheral
10
blood samples (12 - 20 mL EDTA) were collected from all subjects. An additional 12 mL of
11
blood was collected into EDTA tubes from the third trimester pregnancies 24 hours after
12
delivery. Gestational ages of the first-, second- and third-trimester samples were 12 – 14
13
weeks, 17 – 21 weeks and 38 – 40 weeks, respectively. In this proof-of-concept study, the
14
experimentalists were not blinded with regard to the clinical status of the specimens.
15
16
PROCESSING OF BLOOD AND TISSUE SAMPLES
17
Maternal peripheral blood samples were processed by a double centrifugation protocol as
18
previously described (1). The blood cell portion was recentrifuged at 2 500 g, and any residual
19
plasma was removed. Samples were stored at -80 C until extraction. DNA from the
20
peripheral blood cells and that from maternal plasma was extracted with the blood and body
21
fluid protocol of the QIAamp DNA Blood Mini Kit and the QIAamp DSP DNA Blood Mini
22
Kit, respectively (Qiagen, Hilden, Germany). DNA from the CVS and placentas was extracted
23
with the QIAamp DNA Mini Kit (Qiagen) according to the manufacturer’s tissue protocol.
24
25
SCREENING OF DIFFERENTIALLY-METHYLATED REGIONS ON CHROMOSOME 21
1
2
1
BY COBRA
2
The COBRA procedure is described as follows. Bisulfite conversion was performed on one
3
microgram of each DNA sample using the EZ DNA Methylation Kit (Zymo Research, Orange,
4
CA) according to manufacturer’s instructions. Forty nanograms of bisulfite-converted DNA
5
(calculation based on the original DNA input) were then subjected to PCR amplification with
6
reagents supplied in the HotStar Taq DNA Polymerase Kit (Qiagen, Hilden, Germany).
7
Reagent compositions for each PCR are listed in Table S1 in the online Data Supplement.
8
Typically, PCR was performed in a 20 L reaction with 1x PCR Buffer, MgCl2, 50 M of
9
each dNTP, forward and reverse primers, HotStar Taq polymerase, with or without 2X PCRx
10
Enhancer (Invitrogen, Carlsbad, CA). The thermal profile was 95 °C for 15 min, followed by
11
45-55 cycles of 95 °C for 20 s, appropriate annealing temperature for 30 s, 72 °C for 1.5 min,
12
and a final extension of 72 °C for 3 min. PCR products were then subjected to restriction
13
enzyme digestion. The restriction enzyme to be used for each respective locus was chosen for
14
its ability to distinguish between the methylated and unmethylated sequence after bisulfite
15
conversion. In essence, restriction sites were only present in either the methylated or
16
unmethylated sequence but not both, so that one of the sequences would be digested while the
17
other would remain intact. Restriction enzyme digestions were performed in 20 L reactions
18
with 5 L PCR products, 1x appropriate buffer, and 10 U restriction enzyme (or none for
19
mock digestion) under the manufacturer’s recommended temperatures for 2 h. All enzymes
20
were purchased from New England Biolabs (Ipswich, MA). Digested products were then
21
analyzed by agarose gel electrophoresis.
22
23
CLONING AND BISULFITE SEQUENCING ANALYSIS OF THE
24
DIFFERENTIALLY-METHYLATED LOCUS IDENTIFIED BY COBRA
25
The same PCR products from the COBRA analysis were used for cloning and bisulfite
2
3
1
sequencing. To analyze methylation status at single molecule resolution, the PCR product was
2
TA-cloned into a plasmid vector using the pGEM-T Easy Vector System (Promega, Madison,
3
WI). The inserts from the recombinant clones were PCR amplified using vector primers T7
4
and SP6, and were then analyzed by cycle sequencing using the BigDye Terminator Cycle
5
Sequencing v1.1 kit (Applied Biosystems, Foster City, CA) according to the manufacturer’s
6
instructions. After ethanol precipitation, the samples were resuspended in 10 L of Hi-Di
7
formamide and run on a 3100 DNA Analyzer (Applied Biosystems). The sequencing data
8
were analyzed using the SeqScape software (Applied Biosystems). The completeness of
9
bisulfite conversion was confirmed before scoring. To calculate the methylated site frequency,
10
the total number of methylated CpG sites was divided by the total number of CpG sites in the
11
cloned region (2).
12
13
CONVENTIONAL REAL-TIME QUANTITATIVE PCR ANALYSIS OF THE HLCS
14
LOCUS
15
MSRE digestion. For each placental and maternal blood cell DNA sample, 100 ng DNA was
16
subjected to MSRE digestion. Restriction enzyme digestion was performed in a 50 L
17
reaction with 1x appropriate buffer, 50 U of BstUI (or none for mock digestion) (New
18
England Biolabs) at 60 C for at least 16 h. For each pre- and post-delivery maternal plasma
19
sample, 1.6 mL plasma was used for DNA extraction, and was eluted in 50 L of deionized
20
water, 21 L of which was subjected to restriction enzyme digestion. Enzyme digestion was
21
performed in a 30 L reaction with 1x appropriate buffer, 30 U of BstUI (or none for mock
22
digestion) at 60 C for at least 16 h. Digested products were then analyzed by real-time
23
quantitative PCR. The selected restriction enzyme only digested the unmethylated DNA but
24
not the methylated DNA. Since the data from the COBRA analysis has shown that HLCS is
25
hypermethylated in placental tissues and unmethylated in maternal blood cells, we expected a
3
4
1
proportion of DNA from placental tissues would remain detectable while most of the DNA
2
from maternal blood cells would be digested, thus becoming undetectable after restriction
3
enzyme treatment.
4
Real-time quantitative PCR analysis. A real-time PCR assay was developed using the 7300
5
Real-time PCR Systems (Applied Biosystems) for the quantitative analysis of HLCS genomic
6
DNA with and without restriction enzyme digestion. Two and a half microliters of restriction
7
enzyme treated or mock-digested DNA samples were used in the real time PCR assay in a
8
total reaction volume of 25 L using 96-well reaction plates (MicroAmp Optical 96-Well
9
Reaction Plate, Cat No. N8010560, Applied Biosystems). Multiple water blanks (no template
10
controls) were included in every PCR experiment. Each reaction contained 1x TaqMan
11
Universal PCR Master Mix (Applied Biosystems), 300 nM of each of the forward and reverse
12
primers (Integrated DNA Technologies, Coralville, IA), and 100 nM of the hydrolysis probe
13
(Applied Biosystems). The sequences of the primers and probes are listed in Table 1. The
14
thermal profile was 50 °C for 2 min, 95 °C for 10 min, followed by 50 cycles of 95 °C for 15
15
s, and 60 °C for 1 min. All reactions were run in duplicates, and the mean quantity was taken.
16
The fluorescent signals were collected and analyzed by the SDS v1.3.0 software (Applied
17
Biosystems). As the detectable HLCS molecules after restriction enzyme digestion
18
represented the methylated fraction, we expressed the real-time quantitative PCR as a
19
methylation index. The methylation index of a sample was calculated by dividing the copy
20
number of HLCS after enzyme digestion by that obtained with mock digestion. Maternal
21
plasma DNA analysis was performed to show the postpartum clearance of the
22
hypermethylated HLCS molecules.
23
24
INFORMATION FOR THE REAL-TIME QUANTITATIVE PCR EXPERIMENT OF THE
25
HLCS LOCUS
4
5
1
The primers were analyzed by “UCSC In-Silico PCR” (http://www.genome.ucsc.edu). The
2
full amplicon sequence was BLAT searched (http://www.genome.ucsc.edu) to ensure
3
specificity. Assay specificity was further confirmed by subjecting the real-time PCR products
4
from mock-digested maternal blood cell and placental DNA samples to agarose gel
5
electrophoresis. A single band with a size of 96 bp was obtained. No template control (water
6
blank) was performed with 40 replicates and no positive signal was observed. Serially-diluted
7
human genomic DNA originally quantified by NanoDrop ND-1000 (Thermo Fisher Scientific,
8
Waltham, MA), using the DNA-50 setting, was used as the quantitative standard with a linear
9
dynamic range from three copies per reaction to 10 000 copies per reaction. The limit of
10
detection (LOD) for the assay was three copies per reaction, which was estimated to be
11
equivalent to 45 copies/mL for plasma DNA detection. Twenty duplicated reactions (40
12
replicates) were performed at the LOD concentration and 100% of the wells were positive
13
with a median quantification cycle of 36.8 (range: 35.7-40.5). The mean and standard
14
deviation (SD) of quantification cycles of these 40 replicates was 37.1 and 1, respectively.
15
The intra-assay variation was determined by measuring the concentration of a placental DNA
16
sample diluted to a concentration similar to that of the maternal plasma sample (50 copies per
17
reaction, which was estimated to be equivalent to 890 copies/mL for plasma DNA detection)
18
for 20 duplicates (40 replicates). This resulted in a mean and SD of quantification cycles of
19
32.3 and 0.17, respectively. To express the results in copies/mL, the mean HLCS
20
concentration and the SD detected from the 40 replicates were 780 copies/mL and 89
21
copies/mL, respectively.
22
A total of three real-time quantitative PCR experiments were performed to evaluate the
23
HLCS concentration in the DNA samples. The performance of the standard curve in each
24
experiment is illustrated in Supplemental Table S3. Supplemental Table S4 comprises a
25
checklist for all the information for the real-time quantitative PCR.
5
6
1
2
CHROMOSOME DOSAGE ANALYSIS BY MICROFLUIDICS DIGITAL PCR
3
The specificity of the digital PCR assays was determined with the optimized PCR conditions.
4
Assay specificity on the BstUI-digested genomic DNA samples. DNA samples from the CVS,
5
placenta, and maternal blood cells were subjected to the HLCS, RASSF1A, and ZFX/Y digital
6
PCR analysis after BstUI digestion. After enzyme digestion, the DNA was diluted to a
7
concentration of 1 - 2 ng/L for loading into the reaction mixture for digital PCR analysis.
8
DNA samples from the CVS and placenta consisted predominantly of enzyme
9
digestion-resistant HLCS and RASSF1A molecules, thus signals should be detected after
10
restriction enzyme digestion. On the other hand, we would expect no or low level of detection
11
in the blood cells as they were hypomethylated at these two loci.
12
The ZFX/Y loci did not contain any BstUI enzyme digestion sites, thus restriction
13
enzyme treatment should not confer any effects on the DNA molecules. The ZFY molecules
14
constituted the fetal-derived sequences from a male fetus for the ratio comparison.
15
Fetal-specificity on the BstUI-digested plasma DNA samples. After BstUI digestion on the
16
maternal plasma DNA samples, a four-fold dilution with deionized water was performed. The
17
samples were then mixed with the reaction mixture for loading onto the microfluidics chips
18
for analysis.
19
20
References for Supplemental Data Methods
21
1.
Chiu RWK, Poon LLM, Lau TK, Leung TN, Wong EM, Lo YMD. Effects of
22
blood-processing protocols on fetal and total DNA quantification in maternal plasma.
23
Clin Chem 2001;47:1607-13.
24
25
2.
Chiu RWK, Chim SSC, Wong IH, Wong CSC, Lee WS, To KF, et al.
Hypermethylation of RASSF1A in human and rhesus placentas. Am J Pathol
6
7
1
2
2007;170:941-50.
3.
Lun FMF, Chiu RWK, Chan KCA, Leung TY, Lau TK, Lo YMD. Microfluidics
3
digital PCR reveals a higher than expected fraction of fetal DNA in maternal plasma.
4
Clin Chem 2008;54:1664-72.
5
7
8
1
Tables and Figures
2
Table S1. The list of information regarding the 51 COBRA assays, including the names
3
and locations of the loci, the primer sequences, reaction conditions and results. [Excel
4
spreadsheet.]
5
6
8
9
1
Table S2A. Detection of HLCS, RASSF1A, ZFY and ZFX in the mock- and
2
BstUI-digested DNA samples by digital PCR.
3
Sample
Maternal blood 1_mock
Maternal blood 1_Bst UI
Maternal blood 2_mock
Maternal blood 2_Bst UI
Placenta 1_mock
Placenta 1_Bst UI
Placenta 2_mock
Placenta 2_Bst UI
4
The copy numbers are corrected for the Poisson distribution using the formula shown in the text.
Copy number
RASSF1A
464
0
321
0
287
187
113
81
HLCS
487
6
374
17
295
196
158
106
ZFY
0
0
0
0
158
164
93
90
ZFX
455
459
397
437
171
177
64
95
5
6
Table S2B. Detection of HLCS and RASSF1A in the untreated and BstUI-digested
7
plasma DNA samples by digital PCR.
8
9
Replicate
1
1
2
1
1
2
Sample
Post-delivery plasma 1_untreated
Post-delivery plasma 1_BstUI
Post-delivery plasma 1_BstUI
Post-delivery plasma 2_untreated
Post-delivery plasma 2_BstUI
Post-delivery plasma 2_BstUI
HLCS
172
9
6
106
4
2
Copy number
RASSF1A
109
1
0
66
3
1
The copy numbers are corrected for the Poisson distribution using the formula shown in the text.
10
11
9
10
1
Table S3. Performance of the calibration standard in three experiments.
Experiment
Slope
Y-intercept
PCR efficiency
R2
1
-3.69
38.2
0.87
0.99
2
-3.69
38
0.87
0.99
3
-3.56
37.7
0.91
0.99
2
3
4
Table S4. MIQE checklist. [Excel spreadsheet.]
5
6
10
11
1
Figure S1. Genomic sequences of the four digital PCR assays (HLCS, RASSF1A, ZFY
2
and ZFX). The chromosomal location on the March 2006 human reference sequence (NCBI
3
Build 36.1) of the UCSC Genome Browser is indicated. The underlined nucleotides represent
4
the enzyme recognition sites. The bold nucleotides in capital letters represent the forward and
5
reverse primers. The bold nucleotides in small letters represent the MGB probe sequence.
6
7
8
9
10
11
11
12
1
Figure S2. Gel electrophoresis of COBRA results for (A) HLCS region B2, and (B)
2
C21orf81. Two trisomy 21 placentas (T21 PLN), two first-trimester normal placentas (Normal
3
PLN 1st), two third-trimester normal placentas (Normal PLN 3rd), and two first-trimester
4
maternal blood cells (MBC) were analyzed. PCR products were incubated with (+) or without
5
(-) BstUI enzyme. DNA methylation was detected by the appearance of the smaller size
6
digestion products. One kb ladder (Invitrogen Carlsbad, CA) (M) was used in gel
7
electrophoresis.
8
(A)
Normal PLN 1st Normal PLN 3rd
T21 PLN
BstIUI
+
-
+
-
+
-
+
+
-
-
+
-
MBC
+
-
M
+
-
9
10
517/506 bp
11
220 bp
12
13
14
15
Normal PLN 1st Normal PLN 3rd
T21 PLN
(B)
BstIUI
+
-
+
-
+
-
+
-
+
-
+
-
MBC
+
-
M
+
-
16
17
517/506 bp
18
220 bp
19
20
21
22
12
13
1
Figure S3. Quantification of HLCS DNA from four placental tissues and eight maternal
2
blood cells by real-time quantitative PCR. The methylation index of a sample was
3
calculated by dividing the copy number of HLCS after enzyme digestion by that obtained
4
from mock digestion.
5
n=4
n=8
Placenta
Maternal blood cell
1.2
6
8
9
10
Methylation index
1.0
7
0.8
0.6
0.4
0.2
11
0.0
12
-0.2
13
14
15
13
14
1
Figure S4. Postpartum clearance in plasma detected by real-time quantitative PCR.
2
Detection of HLCS in 10 pairs of pre- and post-delivery maternal plasma samples (A) with (+),
3
or (B) without (-) restriction enzyme digestion. Paired samples from the same subject are
4
depicted by identical symbols connected by a line.
5
(A)
1200
HLCS (copies/mL plasma)
6
7
8
9
10
11
1000
800
600
400
200
0
12
Pre-delivery (+)
Post-delivery (+)
Pre-delivery (-)
Post-delivery (-)
13
15
16
17
18
19
20
(B)
4500
HLCS (copies/mL plasma)
14
4000
3500
3000
2500
2000
1500
1000
21
22
14