Large fragment Bst DNA polymerase for whole genome amplification of DNA from formalin-fixed paraffin-embedded tissues.

Abstract

Formalin-fixed paraffin-embedded (FFPE) tissues represent the largest source of archival biological material available for genomic studies of human cancer. Therefore, it is desirable to develop methods that enable whole genome amplification (WGA) using DNA extracted from FFPE tissues. Multiple-strand Displacement Amplification (MDA) is an isothermal method for WGA that uses the large fragment of Bst DNA polymerase. To date, MDA has been feasible only for genomic DNA isolated from fresh or snap-frozen tissue, and yields a representational distortion of less than threefold.
We amplified genomic DNA of five FFPE samples of normal human lung tissue with the large fragment of Bst DNA polymerase. Using quantitative PCR, the copy number of 7 genes was evaluated in both amplified and original DNA samples. Four neuroblastoma xenograft samples derived from cell lines with known N-myc gene copy number were also evaluated, as were 7 samples of non-small cell lung cancer (NSCLC) tumors with known Skp2 gene amplification. In addition, we compared the array comparative genomic hybridization (CGH)-based genome profiles of two NSCLC samples before and after Bst MDA. A median 990-fold amplification of DNA was achieved. The DNA amplification products had a very high molecular weight (> 23 Kb). When the gene content of the amplified samples was compared to that of the original samples, the representational distortion was limited to threefold. Array CGH genome profiles of amplified and non-amplified FFPE DNA were similar.
Large fragment Bst DNA polymerase is suitable for WGA of DNA extracted from FFPE tissues, with an expected maximal representational distortion of threefold. Amplified DNA may be used for the detection of gene copy number changes by quantitative realtime PCR and genome profiling by array CGH.

Full text

Page 1

BioMed Central
Page 1 of 10
(page number not for citation purposes)
BMC Genomics
Open Access
Methodology article
Large fragment Bst DNA polymerase for whole genome
amplification of DNA from formalin-fixed paraffin-embedded
tissues
Sarit Aviel-Ronen1,2, Chang Qi Zhu2, Bradley P Coe3, Ni Liu2,
Spencer K Watson3, Wan L Lam3 and Ming Sound Tsao*1,2,4
Address: 1Department of Pathology, University Health Network, 200 Elisabeth Street, Toronto, ON M5G 2C4, Canada, 2Division of Applied
Molecular Oncology, Ontario Cancer Institute/Princess Margaret Hospital, University Health Network, 610 University Avenue, Toronto, ON M5G
2M9, Canada, 3Department of Cancer Genetics and Developmental Biology, British Columbia Cancer Research Centre, 675 West 10th Avenue,
Vancouver, BC V5Z 1L3, Canada and 4Department of Laboratory Medicine and Pathobiology, Faculty of Medicine, University of Toronto, 100
College Street, Toronto ON M5G 1L5, Canada
Email: Sarit Aviel-Ronen - saronen@uhnres.utoronto.ca; Chang Qi Zhu - czhu@uhnres.utoronto.ca; Bradley P Coe - bcoe@bccrc.ca;
Ni Liu - niliu@uhnres.utoronto.ca; Spencer K Watson - swatson@bccrc.ca; Wan L Lam - wanlam@bccrc.ca;
Ming Sound Tsao* - ming.tsao@uhn.on.ca
* Corresponding author
Abstract
Background: Formalin-fixed paraffin-embedded (FFPE) tissues represent the largest source of
archival biological material available for genomic studies of human cancer. Therefore, it is desirable
to develop methods that enable whole genome amplification (WGA) using DNA extracted from
FFPE tissues. Multiple-strand Displacement Amplification (MDA) is an isothermal method for WGA
that uses the large fragment of Bst DNA polymerase. To date, MDA has been feasible only for
genomic DNA isolated from fresh or snap-frozen tissue, and yields a representational distortion of
less than threefold.
Results: We amplified genomic DNA of five FFPE samples of normal human lung tissue with the
large fragment of Bst DNA polymerase. Using quantitative PCR, the copy number of 7 genes was
evaluated in both amplified and original DNA samples. Four neuroblastoma xenograft samples
derived from cell lines with known N-myc gene copy number were also evaluated, as were 7
samples of non-small cell lung cancer (NSCLC) tumors with known Skp2 gene amplification. In
addition, we compared the array comparative genomic hybridization (CGH)-based genome profiles
of two NSCLC samples before and after Bst MDA. A median 990-fold amplification of DNA was
achieved. The DNA amplification products had a very high molecular weight (> 23 Kb). When the
gene content of the amplified samples was compared to that of the original samples, the
representational distortion was limited to threefold. Array CGH genome profiles of amplified and
non-amplified FFPE DNA were similar.
Conclusion: Large fragment Bst DNA polymerase is suitable for WGA of DNA extracted from
FFPE tissues, with an expected maximal representational distortion of threefold. Amplified DNA
may be used for the detection of gene copy number changes by quantitative realtime PCR and
genome profiling by array CGH.
Published: 12 December 2006
BMC Genomics 2006, 7:312 doi:10.1186/1471-2164-7-312
Received: 10 August 2006
Accepted: 12 December 2006
This article is available from: http://www.biomedcentral.com/1471-2164/7/312
© 2006 Aviel-Ronen et al; licensee BioMed Central Ltd.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0),
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Page 2

BMC Genomics 2006, 7:312http://www.biomedcentral.com/1471-2164/7/312
Page 2 of 10
(page number not for citation purposes)
Background
With growing interest in the genomic characteristics of
various human tumors and a steep increase in the availa-
bility of genomic tests for both clinical and research pur-
poses, the amount of genomic DNA available from
biological samples may limit the practicality of genomic
analysis. Having been used for decades, formalin-fixed
paraffin-embedded (FFPE) tissues comprise the most
common form of human tissue samples archives. There-
fore, it is desirable to establish a whole genome amplifica-
tion (WGA) method specifically for DNA extracted from
FFPE tissues. Two main approaches for WGA have been
developed: thermocycling protocols and isothermal
amplification methods.
Several thermocycling protocols have been used, includ-
ing the degenerate oligonucleotide primed-polymerase
chain reaction (DOP-PCR) [1-4], primer extension
preamplification (PEP) [5-7], tagged-PCR (T-PCR) [8],
and single cell comparative genomic hybridization
(SCOMP), also known as linker adaptor-PCR [9,10].
What common in all these protocols are the PCR principle
of temperature-dependent cyclic amplification, and the
use of primers with a random sequence to allow for mul-
tiple binding sites. They differ in primer design and the
sequence of temperature changes. Their amplification
magnitude is a few hundredfold and the size of their DNA
product ranges from 200–3000 bases. Each technique has
its advantages and limitations, varying from incomplete
genomic coverage to preferences for certain DNA length
(e.g. shorter alleles in DOP-PCR [4]), and inconsistency in
the magnitude of amplification and elaborated protocol
(SCOMP).
Isothermal amplification methods refer to Hyper-
branched Strand Displacement Amplification (HSDA),
which is also known as Multiple-strand Displacement
Amplification (MDA) [11-13]. MDA is based on two prin-
ciples [14-16]: (1) the ability of the polymerase to cause
strand-displacement, and (2) random initiation points
using random primers. The 5' end of each strand is dis-
placed by another upstream strand that is growing in the
same direction. Displaced single strands are targeted by
new random priming events. As more DNA is generated
by strand displacement, an increasing number of random
priming events occur, forming a network of hyper-
branched DNA structures of high molecular weight. As the
reaction proceeds, thousands or even millions of copies of
the original DNA are generated. Two enzymes are capable
of catalyzing MDA: Φ29 DNA polymerase and the large
fragment of Bacillus stearothermophilus (Bst DNA polymer-
ase, large fragment). Previous work has shown that MDA
using Bst DNA polymerase on intact DNA (e.g. DNA iso-
lated from fresh or snap-frozen tissue) gives rise to robust
amplification with a representational distortion of less
than threefold [14-16]. We have investigated the feasibil-
ity of MDA on DNA from FFPE tissue using the Bst DNA
polymerase, and evaluated the magnitude of representa-
tional distortion using quantitative realtime PCR (QPCR)
and whole genome tiling array comparative genomic
hybridization (CGH) representing complete coverage of
the human genome.
Results
The Bst DNA polymerase yielded a median 990-fold
(range 613–1618) of DNA amplification (Figure 1). The
reaction efficiency for commercial DNA and DNA
extracted from snap-frozen samples was comparable and
achieved a median amplification of 803-fold (range 613 –
1043), whereas the FFPE derived DNA was amplified
slightly more, with a median amplification of 1035-fold
(range 839 – 1618). The amplification of the negative
control also generated DNA product, which was consist-
ent in amount to the other samples. Amplification of
DNA from the human pancreatic ductal epithelium
(HPDE) cell line yielded 1422 ± 310- and 1560 ± 144-fold
changes without and with DNA shearing, respectively.
DNA replication products were of very high molecular
weight as they were larger than 23 Kb (Figure 2).
QPCR analysis revealed similar findings in all samples,
whether of normal tissue or tumoral nature, carrying
known gene copy number abnormalities. All tested genes
were found in all FFPE and Bst amplified samples, and
their relative gene copy number was within 3-fold range of
Mean amplification of DNA by Bst polymerase
Figure 1
Mean amplification of DNA by Bst polymerase. All
reactions started with 10 ng of target DNA. FFPE samples:
Lung 1–5, neuroblastoma xenografts (LAN-5, NUB-7, SK-N-
BE(2), NBL-S) and NSCLC 1–7. Intact DNA samples: FL
(Frozen Lung) 1–4 and Positive C. (Control). Negative C.
(Control) contained water in lieu of target DNA. For each
sample the mean and SD of 2–6 independent experiments is
shown.

Page 3

BMC Genomics 2006, 7:312http://www.biomedcentral.com/1471-2164/7/312
Page 3 of 10
(page number not for citation purposes)
non-amplified samples (Figure 3 to 5). In normal lung
samples, the expected copy number ratio of any given
gene to GAPDH was 1. The values shown in Figure 3 are
the average ratios of the five samples tested before and
after Bst amplification. The average ratios of the non-
amplified samples were close to 1 (range 1.08–1.26),
while those following amplification were somewhat
higher (range 1.20–2.00) but within 3-fold (Figure 3). For
neuroblastoma xenografts where the N-myc gene is highly
amplified, the representational distortion introduced by
Bst amplification was negligible relative to the magnitude
of gene amplification (Figure 4). For genes with low
amplification levels such as the Skp2 gene in NSCLC
(Skp2/PIK3R1 ≤ 6), an increase in gene copy was detected
following Bst amplification with a bias of up to 3-fold
(Figure 5). It should be noted that in two of the non-
amplified samples (NSCLC no. 6 and 7), the Skp2/PIK3R1
ratio was lower than 3, and therefore within the bias
range. Nevertheless, it was detectable after Bst amplifica-
tion.
In the negative control, the amplification reaction pro-
duced substantial amounts of DNA. However, no genes
were ever detected by QPCR, indicating that the measured
product was the result of a spurious amplification of the
primers.
Array CGH genome profiles of the Bst-amplified DNA
from two NSCLC tumours were similar to profiles
obtained using their respective non-amplified DNA (Fig-
ure 6A). Hybridization of Bst-amplified samples against
Bst-amplified reference DNA allowed for the identifica-
tion of genomic changes that were below 3-fold change.
For each of the array CGH clones, the ratio of sample to
reference signal defines the changes in gene content of a
given tumour. This ratio should not change for each clone
when an optimal WGA method is used. Figure 6B illus-
trates that the correlation of such ratios was near ideal
(1:1) between the un-amplified and Bst-amplified DNA
from NSCLC 8; similar finding was found in NSCLC 9.
The four CGH arrays had variable quality; therefore a dif-
ferent number of human bacterial artificial chromosome
(hBAC) clones was evaluated for each pair of arrays
hybridized to non-amplified and amplified DNA (Table
2). Both NSCLC samples had normal gene content in
more than half of the hBAC clones, and detected changes
were of low gene-dosage. Analysis using the aCGH-
Smooth software showed that 58.3–84.3% of the clones
had a matching call (normal gene content/amplification/
Amplification of genes in formalin-fixed paraffin embedded normal lung tissue
Figure 3
Amplification of genes in formalin-fixed paraffin
embedded normal lung tissue. Mean values of relative
amount of each gene to GAPDH in five samples before and
after Bst polymerase amplification are shown. Error bars for
± 2 SD of mean values of Bst amplified samples are drawn. X3
Mean Not Amp is the calculated expected 3-fold representa-
tional distortion range. Gene copy numbers following Bst
amplification resembled respective values in non-amplified
samples and were within 3-fold change.
Gel electrophoresis of Bst DNA polymerase amplification products
Figure 2
Gel electrophoresis of Bst DNA polymerase amplifi-
cation products. From left to right: (1) Lambda DNA-Hind
III digested ladder; FFPE samples: (2, 3) normal lung 3 & 4; (4,
5) neuroblastoma xenografts LAN-5 & SK-N-BE (2) and (6,
7) NSCLC 3 & 4; (8) Commercial DNA; (9) Negative con-
trol. Samples were analyzed in 0.5% agarose gel, stained with
SYBR-green II. 10% by volume of the amplification product
was used for the gel electrophoresis.

Page 4

BMC Genomics 2006, 7:312http://www.biomedcentral.com/1471-2164/7/312
Page 4 of 10
(page number not for citation purposes)
deletion) following Bst amplification. The percentage of
matching clones correlated with the level of genomic
changes: the higher the un-amplified sample/reference
ratio was (especially > 3), the more likely it was to be
detected correctly following Bst amplification. NSCLC 8,
which had higher level of genomic gains compared to
NSCLC 9 (as reflected by the highest ratio of un-amplified
sample/reference of 2.9 vs. 2.1), also exhibited a higher
percentage of hBAC clones with matching profiles (84.3%
vs. 58.3% respectively). The level of disagreement
between the paired arrays was expressed by the number of
clones that changed after Bst amplification. Both NSCLC
samples had more amplified than deleted clones. Like-
wise, more amplified than deleted clones were undetected
after Bst amplification. However, based on the percentage
of originally amplified/deleted clones, gene deletions
were more prone to escape detection following Bst MDA
(95.83–100%) compared to gene amplifications (67.22–
92.31%).
In contrast to the successful Bst MDA of FFPE DNA,
repeated attempts to amplify FFPE DNA using Φ29 DNA
polymerase failed. Although the latter yielded 568 ± 342-
fold DNA amplification and the reaction product was vis-
ible on a gel, QPCR of genes successfully validated on Bst
MDA products consistently failed on Φ29 DNA polymer-
ase WGA products. To rule out the inadequacy of FFPE
DNA or of the Φ29 DNA polymerase reaction, we
repeated the QPCR reactions on non-amplified FFPE DNA
and DNA from frozen tissue before and after Φ29 DNA
polymerase amplification and were able to detect all
genes.
Discussion
Lage and Dean et al [14,15] reported that MDA demon-
strates high-amplification potential and excellent loci rep-
resentation with less than 3-fold bias. Our study showed
for the first time that Bst MDA is feasible and reliable for
WGA even on FFPE DNA. We have demonstrated that in
three groups of FFPE samples (normal lung tissue, neu-
roblastoma xenografts and NSCLC), the relative content
of different genes was maintained following the amplifi-
cation. If a bias existed, it was limited to a 3-fold change.
We deliberately monitored the content of genes that are
located on separate, unrelated regions of the genome to
provide a good estimation of the overall amplification
process. Array CGH data further supports the adequacy of
Bst MDA on FFPE DNA. Hybridization against Bst-ampli-
fied reference DNA allows detection of genomic changes
that are even below 3-fold change.
In our study, median amplification ranged from 803- to
1035-fold, and was higher for FFPE samples than for
intact DNA isolated from snap-frozen tissue or commer-
cial DNA. This is more than the 250-fold reported previ-
ously [14]. The discrepancy might be attributed to the
method used for quantitation of the template and prod-
Detection of Skp2 amplification in NSCLC samples following whole genome amplification by Bst DNA polymerase
Figure 5
Detection of Skp2 amplification in NSCLC samples
following whole genome amplification by Bst DNA
polymerase. The ratios of Skp2 to PIK3R1 gene were main-
tained in Bst amplified vs. non-amplified NSCLC samples.
Error bars represent SD.
N-myc gene content in Bst amplified vs. non-amplified neu-roblastoma xenografts
Figure 4
N-myc gene content in Bst amplified vs. non-amplified
neuroblastoma xenografts. For neuroblastoma
xenografts, where N-myc gene is highly amplified, relative
gene content in Bst amplified samples was comparable to the
respective values in non-amplified samples and the represen-
tational distortion was negligible. Note: NBL-S is a neuroblas-
toma cell line that lacks N-myc amplification and
appropriately the calculated copy numbers were 1.12 ± 0.03
for non-amplified DNA and 1.14 ± 0.35 for Bst amplified
DNA. Error bars represent SD.

Page 5

BMC Genomics 2006, 7:312http://www.biomedcentral.com/1471-2164/7/312
Page 5 of 10
(page number not for citation purposes)
Array CGH of NSCLC before and after Bst amplification
Figure 6
Array CGH of NSCLC before and after Bst amplification. (A) Data for chromosomes 1, 5 & 8 are displayed as a karyo-
type diagram with values corresponding to log2 ratio of Cy5/Cy3 spot signal (SeeGH v1.6). The genome profile following Bst
amplification was similar to the profile of the original sample. Clones with log2 ratio <0.5 at 1p escaped detection following Bst
amplification. (B) Scatter diagram comparing ratio of Cy5/Cy3 spot signal of NSCLC 8 before and after Bst amplification. Solid
line: expected 3 fold representational distortion. Dashed line: desired (1:1) ratio of ideal WGA devoid of representational dis-
tortion. Comparison of the signal ratio for NSCLC 8 before and after Bst amplification shows it is near ideal (1:1) ratio.

Page 6

BMC Genomics 2006, 7:312http://www.biomedcentral.com/1471-2164/7/312
Page 6 of 10
(page number not for citation purposes)
ucts (PicoGreen DNA quantitation vs. NanoDrop spectro-
photometry). The wide range in amplification yield may
be due to variability in DNA quality and tissue fixation. A
possible explanation for the higher yield in FFPE com-
pared to intact DNA could be the preferential amplifica-
tion of shorter DNA fragments by the Bst polymerase
[17,18]. However, the partial shearing of intact genomic
DNA did not result in a significant change of the amplifi-
cation yield. Yet, our results contradict previous mathe-
matically-based predictions of a lower yield with sheared
DNA [14].
Our QPCR analysis of gene copy content of FFPE samples
following Bst amplification demonstrated up to a 3-fold
change with respect to the non-amplified samples, fitting
earlier reports of up to 3-fold representational bias
[14,15]. To compare the bias resulting from various WGA
methods is challenging, as the reported characteristics of
each method depend on the initial amount of DNA tem-
plate used, as well as on the application under investiga-
tion. The representational bias can be implied from the
reported range of efficiency rates for amplification of DNA
sequences of several microsatellite loci. When performed
in single cells, DOP-PCR efficiency rate ranged from 0–
10% and was inferior to PEP-PCR and improved PEP-PCR
(I-PEP-PCR) that ranged from 0–20% and 20–50%,
respectively [6]. Dean et al [15] specifically compared the
representational bias of three WGA methods and reported
a 103–106 representational bias with DOP-PCR, 102–104
bias with PEP-PCR, and less than a 3-fold bias with MDA,
which remained almost constant between 100- to
100,000-fold amplification.
It is known that MDA by either Φ29 or Bst DNA polymer-
ases gives products even in the absence of DNA template.
This is thought to result from spurious amplification of
the primers. Lage et al [14] reported that background DNA
synthesis was completely suppressed when modified
primers with two 5'-nitroindole (universal base) residues
were used. We were unable to eliminate primer amplifica-
tion reactions despite the use of modified primers. How-
ever, this spurious primer amplification appears to be
Table 2: Comparison of array CGH genome profiles before and after Bst amplification
Number of clones Normal gene contentGene amplificationGene deletion
NSCLC 8 (2.9*)
Before Bst MDA16277 12779 345048
Comparing after to before MDA Matching clones
Clones that changed
% of abnormal clones before MDA
13725 (84.3%)
2552 (15.7%)
12592 (91.75%)
187 (7.3%)
1131 (8.24%)
2319 (90.9%)
67.22%
2 (<0.01%)
46 (1.8%)
95.83%
NSCLC 9 (2.1*)
Before Bst MDA23942135389483922
Comparing after to before MDA Matching clones
Clones that changed
% of abnormal clones before MDA
13960 (58.3%)
9982 (41.7%)
13231 (94.8%)
307 (3.1%)
729 (5.2%)
8754 (87.7%)
92.31%
0 (0%)
922 (9.2%)
100%
Data from each of the four array CGH experiments (NSCLC 8 & 9 before and after Bst amplification) was normalized and replicate data points with
standard deviation of log2 ratio > 0.075 were excluded. Following, data was analyzed with aCGH-Smooth software, which identifies breakpoints and
areas of gene amplification and deletion. The numbers presented refer to clones that were evaluable both before and after Bst amplification for each
of the tumors. * Highest ratio of un-amplified sample/reference.
Table 1: Primers sequences
GeneAmplicon Forward primerReverse primer
GAPD
NMYC
SS18L2
GHR
PIK3R1
COPS5
LATS2
SKP2
125 bp
111 bp
127 bp
120 bp
144 bp
61 bp
101 bp
102 bp
5'-GGTAAGGAGATGCTGCATTCG-3'
5'-CGCAAAAGCCACCTCTCATTA-3'
5'GTAGGGATGAGGTCTCCCTTTGT-3'
5'-GACTGGCCACTTAGCTGTCTTTG-3'
5'-TCATTTGTGGGATGACTTAGATTTG-3'
5'-TCGACATGCACCTTGTTTGG-3'
5'-GAGTCAGGGAACCTGGCTTTAA-3'
5'-GGGTACCATCTGGCACGATT-3'
5'-CGCCCAATACGACCAAATCTAA-3'
5'-TCCAGCAGATGCCACATAAGG-3'
5'-GAAATGCGGAGCTGGTGTG-3'
5'-GGAGTCCTTTGAGTAGCAGCAACT-3'
5'-AAAGTTGACAGTCCTGAATATTTTTAATATATAAA-3'
5'-TGAAAACAGCTGCAATCCCC-3'
5'-ATATGACTCTTCGGCAGCTGC-3'
5'-GATACTGCTATTCTGAAAGTCTTTTTCTTC-3'

Page 7

BMC Genomics 2006, 7:312http://www.biomedcentral.com/1471-2164/7/312
Page 7 of 10
(page number not for citation purposes)
tolerable, since tested genes were consistently detected by
QPCR in all the Bst-amplified samples but not in the
amplified negative control.
The two enzymes used for MDA, Φ29 DNA polymerase
and the large fragment of Bacillus stearothermophilus, have
distinct qualities. Where as Bst DNA polymerase is devoid
of the 3'→5' exonuclease, Φ29 DNA polymerase holds
this proofreading activity. Therefore, Φ29 DNA polymer-
ase has a lower error rate, more efficient amplification
reactions and appears to be more suitable for additional
sequencing studies. However, Bst DNA polymerase
seemed to demonstrate greater fidelity for copy number
than Φ29 DNA polymerase. One reason is the signifi-
cantly reduced activity of Φ29 DNA polymerase at the tel-
omeres. In contrast, Bst DNA polymerase can switch
templates and is therefore less affected by the proximity of
genes to the telomere [14]. Thus, Bst MDA may be a better
choice for WGA in array-CGH studies. It was suggested
that MDA (by either Φ29 or by large fragment Bst DNA
polymerase) is appropriate for array-CGH from the point
of view of both sequence fidelity and representation dis-
tortion with the following guidelines. First, amplification
should be < 1000-fold to keep the representation distor-
tion less than threefold [14-16]; second, only gene copy
number changes that are minimally threefold can be reli-
ably detected [14], and finally, array-CGH study design
should include amplification of the sample of interest and
reference genomic DNA under identical conditions to
minimize biases [14]. Based on our results, we also rec-
ommend that these guidelines should be adopted when-
ever Bst MDA is followed by QPCR for gene copy number
evaluation. Copy number changes detected following Bst
amplification are reliable only if they are higher than the
3-fold representational distortion range. Thus, it is
expected that high copy number changes (e.g. N-myc in
neuroblastoma xenografts) will be easily detected, since
even with 3-fold amplification bias, the change in copy
number is conspicuous compared to normal gene con-
tent. Although low copy number changes (e.g. Skp2 in
NSCLC) are also detectable, the difference in copy
number relative to normal gene content may be attenu-
ated with 3-fold representation distortion.
Our array CGH results using FFPE DNA are remarkably
similar to results previously obtained on intact DNA. With
1000-cell experiments, Lage et al reported a concordance
between the amplified and non-amplified DNA array
CGH of 53.6–83.3% [14]. We found concordance of
58.3–84.3% when Bst amplification was applied to 10 ng
of DNA starting material. Lage stated that altered loci with
relatively high gene-dosage alterations were detected with
high reproducibility. Likewise, we observed that gene
amplifications were consistently detected, even when
smaller than 3-fold, yet the detection sensitivity correlated
with the level of genomic changes. We noted that gene
deletions were prone to be missed following Bst amplifi-
cation and similar findings can be seen in the data pre-
sented by Lage et al. Array properties, the quality of the
amplified DNA, the length of deleted areas and the low
level of gene content change in deletions, which is
strongly affected by the amplification representation bias
introduced by MDA, all contribute to the reduced detec-
tion ability of deletions following Bst amplification. Alto-
gether, we found that Bst MDA on FFPE DNA is reliable
for following genome profiling by array CGH, in particu-
lar for the detection of gene amplification.
Prior to 2003, there were very few studies on the use of
MDA compared to the widely reported PCR-based WGA
[19]. However, there has recently been a growing interest
in the method, as reflected by the increasing number of
papers published during the past two years. MDA was
reported to succeed in a variety of applications, including
sequencing [20,21], microsatellite marker analysis
[22,23], SNP analysis [24,25], genotyping [26] and array-
CGH [14,27]. All these studies used the Φ29 DNA
polymerase, and all but one [28] used DNA isolated from
fresh or snap-frozen tissue samples. Like others [26], we
have failed in our attempts to amplify FFPE samples with
the Φ29 polymerase. Although the reaction yielded DNA
that was 500-fold greater than the initial amount, no
genes could be detected by QPCR and results were consist-
ent with a spurious amplification of the primers. On the
other hand, Wang et al [28] reported a successful Φ29
amplification of FFPE-derived DNA after the addition of a
preliminary restriction enzyme fragmentation step. The
modified protocol was named Restriction and Circulariza-
tion Aided Rolling Circle Amplification (RCA-RCA). Our
work is the first to describe the use of the large fragment
of Bst DNA polymerase for WGA of FFPE DNA.
Aside from PCR-based techniques and MDA, a novel
approach to WGA is the T7-based linear amplification of
DNA (TLAD). TLAD appears free of sequence and length-
dependent biases, and is thus applicable to FFPE DNA.
However, this technique requires purification following
each step and is therefore laborious and vulnerable to
sample loss [19,29].
Conclusion
We have shown that the large fragment of Bst DNA
polymerase is suitable for WGA of DNA extracted from
FFPE tissues, with an expected representational distortion
of up to threefold. Amplified DNA may be used for the
detection of gene copy number changes by QPCR and
genome profiling by array CGH.
The expected application, magnitude of findings and the
limits of the method are factors that need to be considered

Page 8

BMC Genomics 2006, 7:312http://www.biomedcentral.com/1471-2164/7/312
Page 8 of 10
(page number not for citation purposes)
when WGA is chosen. The virtues of Bst DNA polymerase
use for WGA should be emphasized: the method is effi-
cient, technically easy, gives high yield and is suitable for
FFPE DNA. We believe that in time, Bst DNA amplifica-
tion will become part of routine molecular laboratory
work for research and clinical purposes.
Methods
Tissue materials and genes
The University Health Network Research Ethics Board has
approved this study protocol. Tissues were obtained from
NSCLC patients who underwent tumour resection at the
University Health Network. Snap-frozen tissues were
banked within 30 min after resection. Archival paraffin
embedded tissues were 4 to10 years old (from 1994–
2000); they were fixed in 10% buffered formalin and
processed according to routine pathology departmental
protocols. DNA was extracted from four snap-frozen nor-
mal lung tissue samples and five FFPE normal lung tis-
sues. Commercial human male DNA (Novagen, Madison,
WI) served as a positive control, and water in lieu of target
DNA served as the negative control. Gene copy number
both prior to and following Bst amplification was assayed
by QPCR for GAPDH (12p13), N-myc (2p24.1), SS18L2
(3p21), GHR (5p12-13), PIK3R1 (5q13.1), COPS5
(8q13.1) and LATS2 (13q11-12). Four xenografts of neu-
roblastoma cell lines with known N-myc gene copy
number including LAN-5 [30,31], NUB-7 [32,33], SK-N-
BE (2) [34,35] and NBL-S [36] were also studied. Seven
non-small cell lung cancer (NSCLC) samples with known
Skp2 gene amplification [37] were also similarly evalu-
ated. Tumor cells from the xenograft and NSCLC samples
were enriched by manual microdissection from sections
stained by toluidine blue. To compare the amplification
yield in intact and fragmented DNA, we used DNA from
three clones of the normal human pancreatic ductal epi-
thelium (HPDE) cell line. DNA was sheared by sonication
to strands of less than 4 Kb.
DNA extraction from FFPE tissue
Tissue sections 5 μm thick were de-waxed in toluene with
vibration for 10 seconds, followed by 2 ethanol washes
(with 100% and 75% ethanol) after brief vortexing.
Digestion with 0.5 mg/ml Proteinase K (Roche, Laval, QC,
Canada) in digestion buffer (50 mM KCl, 10 mM Tris-HCl
(pH 8.3), 0.1 mg/mL gelatin, 0.45% Igepal CA-630 and
0.45% Tween) was carried out over-night, and was inacti-
vated in 100°C for 10 minutes. The samples were purified
by Phase Lock Gel (Eppendorf, Westbury, NY) with Phe-
nol:Chloroform:Isoamyl Alcohol 25:24:1, and precipi-
tated in 3 volumes of 0.1 M NaOAc in 95% ethanol
overnight at -20°C. DNA was re-suspended in water at
37°C for 1 hour.
Bst DNA polymerase and Φ29 DNA polymerase MDA
DNA was amplified by the large fragment of Bst DNA
polymerase according to Lage et al [14]. Briefly, 10 ng of
DNA were mixed with 1.5 μL of 10× ThermoPol buffer
(New England Biolabs, Ipswich, MA) and primers (ran-
dom 7-mers with an additional two nitroindole residues
at the 5' end and a phosphorothiate linkage at the 3' end)
at a concentration of 100 μM in 15 μL. DNA was denatur-
ated at 96°C for 2 minutes, cooled at room temperature
for 10 minutes and then placed on ice. The reaction mix-
ture was then brought up to 50 μL with 400 μM dNTPs in
1× ThermoPol buffer, 0.35 units/μL Bst DNA polymerase,
large fragment (New England Biolabs, Ipswich, MA) and
4% final concentration of DMSO. T4 gene 32 protein
(Amersham Biosciences, Piscataway, NJ) was added to the
reaction in final concentration of 30 ng/μL. The reaction
was carried out at 50°C for 6 hours and inactivated at
80°C for 15 minutes. Amplified samples were purified
using Phase Lock Gel (Eppendorf, Westbury, NY) with
Phenol:Chloroform:Isoamyl Alcohol 25:24:1 and precip-
itated in 3 volumes of 0.1 M NaOAc in 95% ethanol over-
night at -20°C. Amplified DNA was re-suspended in water
at 37°C for 1 hour and the concentration was measured
by NanoDrop (NanoDrop Technologies, Wilmington,
DE). 10% by volume of the amplification products were
analyzed on 0.5% alkaline agarose gels stained with SYBR-
green II (Molecular Probes, Eugene, OR). For Φ29 DNA
polymerase, we used the GenomiPhi DNA amplification
kit (GE Healthcare, Piscataway, NJ) according to the man-
ufacturer's protocol.
Gene copy number evaluation based on quantitative
realtime PCR
QPCR was performed using the SYBR Green technique in
a Mx3000P® QPCR System (Stratagene, Cedar Creek, TX).
Primers were designed using Primer Express software v1.5
[38] (Applied Biosystems, Foster City, CA) and tested for
their specificity by alignment using the BLASTN program,
followed by dissociation curve and primer efficiency tests.
The target amplicons were 61-144 bp in intron 1 of each
gene. The sequences of the genomic primers used for sub-
sequent QPCR assays are listed in Table 1. Five ng of non-
amplified target DNA was used in each QPCR reaction,
compared to 50 ng of Bst-amplified DNA. Gene copy
number was normalized to that of GAPDH, and pooled
DNA from five FFPE normal lung samples was used as a
calibrator. Relative gene copy number was calculated
using the formula: 2-ΔΔCt [38,39]. In NSCLC samples, Skp2
was normalized against PIK3R1 (instead of GAPDH) since
PIK3R1 was previously reported to show no amplification
in NSCLC, unlike GAPDH [37,40].
Array CGH of non-amplified and Bst amplified samples
Paired un-amplified and Bst-amplified DNA from two
NSCLC tumours were studied and compared for genomic

Page 9

BMC Genomics 2006, 7:312 http://www.biomedcentral.com/1471-2164/7/312
Page 9 of 10
(page number not for citation purposes)
profile changes using whole genome array CGH. We used
the "27 K" high-density hBAC Sub Megabase Resolution
Tiling (SMRT) set array CGH (BCCRC, Vancouver, BC),
which contains two replicates for each clone and has a res-
olution of 70–80 Kb. The experiments were performed as
previously described [41,42]. Briefly, 400 ng of both sam-
ple and reference male genomic DNA (Novagen, Missis-
sauga, ON, Canada) were labeled with Cyanine-5 and
Cyanine-3 dCTPs (PerkinElmer, Woodbridge, ON, Can-
ada), respectively. Bst-amplified samples were hybridized
against Bst-amplified reference DNA. All MDA DNA was
sheared by sonication into 3–4 Kb prior to labeling. Fol-
lowing hybridization, arrays images were captured by the
charge-coupled device (CCD) scanner system (Perk-
inElmer, Wellesley, MA, USA) and analyzed with Soft-
WoRx Tracker (Applied Precision, Issaquah, WA, USA).
Data were normalized using a three-step normalization
framework [43]; replicate data points with a log2 ratio that
exceeded a standard deviation of 0.075 were excluded.
Data were analyzed with SeeGH v1.6 [44] and aCGH-
Smooth [45] software with the Lambda and breakpoint
per chromosome settings set to 6.75 and 100, respectively.
This analysis defines chromosomal breakpoints and iden-
tifies chromosomal sections with abnormal gene content,
namely areas of gene amplification or deletion. As the
chromosomal location of each hBAC clone is known, the
breakpoint information can be presented at the clone
level where each clone can be defined as having either
amplification, normal content or deletion. Each of the
arrays was independently analyzed and evaluable clones
before and after Bst amplification were compared. The
concordance between paired arrays was indicated by the
number and percentage of clones with matching profiles,
while discordance was shown as the number and percent-
age of clones that changed after Bst amplification (Table
2). The concordance between paired arrays was also dem-
onstrated by the correlation of the sample to reference sig-
nal ratio between un-amplified and Bst-amplified DNA
(Figure 6B).
Abbreviations
CGH – comparative genomic hybridization, FFPE – for-
malin-fixed paraffin-embedded, hBAC – human bacterial
artificial chromosome, MDA – multiple-strand displace-
ment amplification, NSCLC – non-small cell lung cancer,
PCR – polymerase chain reaction, QPCR – quantitative
realtime PCR, WGA – whole genome amplification.
Authors' contributions
SAR carried out the molecular genetic studies, performed
the statistical analysis, contributed to the conception of
the study and its design and drafted the manuscript. CQZ
participated in the molecular genetic studies and contrib-
uted to the statistical analysis. BPC participated in statisti-
cal analysis. NL and SKW participated in molecular
genetic studies. WLL contributed to the conception and
design of the study. MST conceived of the study, partici-
pated in its design and helped to draft the manuscript. All
authors read and approved the final manuscript.
Acknowledgements
This work was supported by the Canadian Cancer Society and National
Cancer Institute of Canada (NCIC) grant #015184. Dr. Aviel-Ronen is a fel-
low of the Canadian Institutes of Health Research
(CIHR) Training Program for Clinician Scientists in Molecular Oncologic
Pathology (STP-53912) and a recipient of the Ontario Cancer Institute
Knudson Research Fellowship and NCIC Terry Fox Foundation Clinical
Research Fellowship. Dr. Tsao is the M. Qasim Choksi Chair in Lung Can-
cer Translational Research.
The authors thank Dr. Herman Yeger for his generosity in providing the
paraffin blocks of neuroblastoma xenograft tumors, Dr. Roya Navab for her
help with the gel electrophoresis, Ms. Suzanne K. Lau for her comments
and Dr. Jeremy Squire for advice.
References
1. Telenius H, Carter NP, Bebb CE, Nordenskjold M, Ponder BA, Tun-
nacliffe A: Degenerate oligonucleotide-primed PCR: General
amplification of target DNA by a single degenerate primer.
Genomics 1992, 13:718-725.
2.Cheung VG, Nelson SF: Whole genome amplification using a
degenerate oligonucleotide primer allows hundreds of geno-
types to be performed on less than one nanogram of
genomic DNA. Proc Natl Acad Sci USA 1996, 93:14676-14679.
3.Harada T, Okita K, Shiraishi K, Kusano N, Furuya T, Oga A, Kawauchi
S, Kondoh S, Sasaki K: Detection of genetic alterations in pan-
creatic cancers by comparative genomic hybridization cou-
pled with tissue microdissection
oligonucleotide primed polymerase chain reaction. Oncology
2002, 62:251-258.
4.Grant SF, Steinlicht S, Nentwich U, Kern R, Burwinkel B, Tolle R:
SNP genotyping on a genome-wide amplified DOP-PCR
template. Nucleic Acids Res 2002, 30:e125.
5.Zhang L, Cui X, Schmitt K, Hubert R, Navidi W, Arnheim N: Whole
genome amplification from a single cell: Implications for
genetic analysis. Proc Natl Acad Sci USA 1992, 89:5847-5851.
6.Dietmaier W, Hartmann A, Wallinger S, Heinmoller E, Kerner T, Endl
E, Jauch KW, Hofstadter F, Ruschoff J: Multiple mutation analyses
in single tumor cells with improved whole genome amplifica-
tion. Am J Pathol 1999, 154:83-95.
7.Wang VW, Bell DA, Berkowitz RS, Mok SC: Whole genome
amplification and high-throughput allelotyping identified five
distinct deletion regions on chromosomes 5 and 6 in micro-
dissected early-stage ovarian tumors. Cancer Res 2001,
61:4169-4174.
8.Grothues D, Cantor CR, Smith CL: PCR amplification of mega-
base DNA with tagged random primers (T-PCR). Nucleic Acids
Res 1993, 21:1321-1322.
9. Klein CA, Schmidt-Kittler O, Schardt JA, Pantel K, Speicher MR,
Riethmuller G: Comparative genomic hybridization, loss of
heterozygosity, and DNA sequence analysis of single cells.
Proc Natl Acad Sci USA 1999, 96:4494-4499.
10.Stoecklein NH, Erbersdobler A, Schmidt-Kittler O, Diebold J, Schardt
JA, Izbicki JR, Klein CA: SCOMP is superior to degenerated oli-
gonucleotide primed-polymerase chain reaction for global
amplification of minute amounts of DNA from microdis-
sected archival tissue samples. Am J Pathol 2002, 161:43-51.
11.Lizardi PM, Huang X, Zhu Z, Bray-Ward P, Thomas DC, Ward DC:
Mutation detection and single-molecule counting using iso-
thermal rolling-circle amplification. Nat Genet 1998,
19:225-232.
12.Dean FB, Nelson JR, Giesler TL, Lasken RS: Rapid amplification of
plasmid and phage DNA using phi 29 DNA polymerase and
and degenerate

Page 10

Publish with BioMed Central and every
scientist can read your work free of charge
"BioMed Central will be the most significant development for
disseminating the results of biomedical research in our lifetime."
Sir Paul Nurse, Cancer Research UK
Your research papers will be:
available free of charge to the entire biomedical community
peer reviewed and published immediately upon acceptance
cited in PubMed and archived on PubMed Central
yours — you keep the copyright
Submit your manuscript here:
http://www.biomedcentral.com/info/publishing_adv.asp
BioMedcentral
BMC Genomics 2006, 7:312http://www.biomedcentral.com/1471-2164/7/312
Page 10 of 10
(page number not for citation purposes)
multiply-primed rolling circle amplification. Genome Res 2001,
11:1095-1099.
Lizardi PM: Multiple Displacement Amplification Connecticut, United
States; 2000.
Lage JM, Leamon JH, Pejovic T, Hamann S, Lacey M, Dillon D, Seg-
raves R, Vossbrinck B, Gonzalez A, Pinkel D, Albertson DG, Costa J,
Lizardi PM: Whole genome analysis of genetic alterations in
small DNA samples using hyperbranched strand displace-
ment amplification and array-CGH. Genome Res 2003,
13:294-307.
Dean FB, Hosono S, Fang L, Wu X, Faruqi AF, Bray-Ward P, Sun Z,
Zong Q, Du Y, Du J, Driscoll M, Song W, Kingsmore SF, Egholm M,
Lasken RS: Comprehensive human genome amplification
using multiple displacement amplification. Proc Natl Acad Sci
USA 2002, 99:5261-5266.
Rook MS, Delach SM, Deyneko G, Worlock A, Wolfe JL: Whole
genome amplification of DNA from laser capture-microdis-
sected tissue for high-throughput single nucleotide polymor-
phism and short tandem repeat genotyping. Am J Pathol 2004,
164:23-33.
Srinivasan M, Sedmak D, Jewell S: Effect of fixatives and tissue
processing on the content and integrity of nucleic acids. Am
J Pathol 2002, 161:1961-1971.
Siwoski A, Ishkanian A, Garnis C, Zhang L, Rosin M, Lam WL: An effi-
cient method for the assessment of DNA quality of archival
microdissected specimens. Mod Pathol 2002, 15:889-892.
Hughes S, Arneson N, Done S, Squire J: The use of whole genome
amplification in the study of human disease. Prog Biophys Mol
Biol 2005, 88:173-189.
Monstein HJ, Olsson C, Nilsson I, Grahn N, Benoni C, Ahrne S: Mul-
tiple displacement amplification of DNA from human colon
and rectum biopsies: Bacterial profiling and identification of
helicobacter pylori-DNA by means of 16S rDNA-based
TTGE and pyrosequencing analysis. J Microbiol Methods 2005,
63:239-247.
Mai M, Hoyer JD, McClure RF: Use of multiple displacement
amplification to amplify genomic DNA before sequencing of
the alpha and beta haemoglobin genes. J Clin Pathol 2004,
57:637-640.
Hosono S, Faruqi AF, Dean FB, Du Y, Sun Z, Wu X, Du J, Kingsmore
SF, Egholm M, Lasken RS: Unbiased whole-genome amplifica-
tion directly from clinical samples. Genome Res 2003,
13:954-964.
Holbrook JF, Stabley D, Sol-Church K: Exploring whole genome
amplification as a DNA recovery tool for molecular genetic
studies. J Biomol Tech 2005, 16:125-133.
Tzvetkov MV, Becker C, Kulle B, Nurnberg P, Brockmoller J,
Wojnowski L: Genome-wide single-nucleotide polymorphism
arrays demonstrate high fidelity of multiple displacement-
based whole-genome amplification. Electrophoresis 2005,
26:710-715.
Pask R, Rance HE, Barratt BJ, Nutland S, Smyth DJ, Sebastian M,
Twells RC, Smith A, Lam AC, Smink LJ, Walker NM, Todd JA: Inves-
tigating the utility of combining phi29 whole genome ampli-
fication and highly multiplexed
polymorphism BeadArray genotyping. BMC Biotechnol 2004,
4:15.
Sjoholm MI, Hoffmann G, Lindgren S, Dillner J, Carlson J: Compari-
son of archival plasma and formalin-fixed paraffin-embedded
tissue for genotyping in hepatocellular carcinoma. Cancer Epi-
demiol Biomarkers Prev 2005, 14:251-255.
Hughes S, Lim G, Beheshti B, Bayani J, Marrano P, Huang A, Squire JA:
Use of whole genome amplification and comparative
genomic hybridisation to detect chromosomal copy number
alterations in cell line material and tumour tissue. Cytogenet
Genome Res 2004, 105:18-24.
Wang G, Maher E, Brennan C, Chin L, Leo C, Kaur M, Zhu P, Rook
M, Wolfe JL, Makrigiorgos GM: DNA amplification method tol-
erant to sample degradation. Genome Res 2004, 14:2357-2366.
Liu CL, Schreiber SL, Bernstein BE: Development and validation
of a T7 based linear amplification for genomic DNA. BMC
Genomics 2003, 4:19.
Seeger RC, Danon YL, Rayner SA, Hoover F: Definition of a thy-1
determinant on human neuroblastoma, glioma, sarcoma,
and teratoma cells with a monoclonal antibody. J Immunol
1982, 128:983-989.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
single nucleotide
26.
27.
28.
29.
30.
31.Hamano S, Ohira M, Isogai E, Nakada K, Nakagawara A: Identifica-
tion of novel human neuronal leucine-rich repeat (hNLRR)
family genes and inverse association of expression of
Nbla10449/hNLRR-1 and Nbla10677/hNLRR-3 with the
prognosis of primary neuroblastomas. Int J Oncol 2004,
24:1457-1466.
Dimitroulakos J, Squire J, Pawlin G, Yeger H: NUB-7: A stable I-
type human neuroblastoma cell line inducible along N- and
S-type cell lineages. Cell Growth Differ 1994, 5:373-384.
Torkin R, Lavoie JF, Kaplan DR, Yeger H: Induction of caspase-
dependent, p53-mediated apoptosis by apigenin in human
neuroblastoma. Mol Cancer Ther 2005, 4:1-11.
Squire JA, Thorner PS, Weitzman S, Maggi JD, Dirks P, Doyle J, Hale
M, Godbout R: Co-amplification of MYCN and a DEAD box
gene (DDX1) in primary neuroblastoma. Oncogene 1995,
10:1417-1422.
Smith AG, Popov N, Imreh M, Axelson H, Henriksson M: Expres-
sion and DNA-binding activity of MYCN/Max and Mnt/Max
during induced differentiation of human neuroblastoma
cells. J Cell Biochem 2004, 92:1282-1295.
Cohn SL, Salwen H, Quasney MW, Ikegaki N, Cowan JM, Herst CV,
Kennett RH, Rosen ST, DiGiuseppe JA, Brodeur GM: Prolonged N-
myc protein half-life in a neuroblastoma cell line lacking N-
myc amplification. Oncogene 1990, 5:1821-1827.
Zhu CQ, Blackhall FH, Pintilie M, Iyengar P, Liu N, Ho J, Chomiak T,
Lau D, Winton T, Shepherd FA, Tsao MS: Skp2 gene copy number
aberrations are common in non-small cell lung carcinoma,
and its overexpression in tumors with ras mutation is a poor
prognostic marker. Clin Cancer Res 2004, 10:1984-1991.
PE Applied Biosystems: User Bulletin #2: Relative Quantitation of Gene
Expression Foster City, CA: Applied Biosystems; 1997.
Livak KJ, Schmittgen TD: Analysis of relative gene expression
data using real-time quantitative PCR and the 2(-delta delta
C(T)) method. Methods 2001, 25:402-408.
Massion PP, Kuo WL, Stokoe D, Olshen AB, Treseler PA, Chin K,
Chen C, Polikoff D, Jain AN, Pinkel D, Albertson DG, Jablons DM,
Gray JW: Genomic copy number analysis of non-small cell
lung cancer using array comparative genomic hybridization:
Implications of the phosphatidylinositol 3-kinase pathway.
Cancer Res 2002, 62:3636-3640.
Ishkanian AS, Malloff CA, Watson SK, DeLeeuw RJ, Chi B, Coe BP,
Snijders A, Albertson DG, Pinkel D, Marra MA, Ling V, MacAulay C,
Lam WL: A tiling resolution DNA microarray with complete
coverage of the human genome. Nat Genet 2004, 36:299-303.
Coe BP, Lockwood WW, Girard L, Chari R, Macaulay C, Lam S,
Gazdar AF, Minna JD, Lam WL: Differential disruption of cell
cycle pathways in small cell and non-small cell lung cancer.
Br J Cancer 2006.
Khojasteh M, Lam WL, Ward RK, MacAulay C: A stepwise frame-
work for the normalization of array CGH data. BMC Bioinfor-
matics 2005, 6:274.
Chi B, DeLeeuw RJ, Coe BP, MacAulay C, Lam WL: SeeGH – a soft-
ware tool for visualization of whole genome array compara-
tive genomic hybridization data. BMC Bioinformatics 2004, 5:13.
Jong K, Marchiori E, Meijer G, Vaart AV, Ylstra B: Breakpoint iden-
tification and smoothing of array comparative genomic
hybridization data. Bioinformatics 2004, 20:3636-3637.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.