Application of N-halogeno-N-sodiobenzenesulfonamide reagents to the selective detection of 5-methylcytosine in DNA sequences

Article abstract of DOI:10.1021/ja311229n

To surmount the challenges of the locus determination and accurate quantification of 5-methyl-2′-deoxycytidine (^5MedC) in DNA fragments that contain multiple ^5MedC residues, we designed and synthesized two N-halogeno-N-sodiobenzenesu

Full text of DOI:10.1021/ja311229n

Communication
pubs.acs.org/JACS
Application of N‑Halogeno‑N‑sodiobenzenesulfonamide Reagents to
the Selective Detection of 5‑Methylcytosine in DNA Sequences
†
†
Tianlu Wang, Tingting Hong, Tun Tang, Qianqian Zhai, Xiwen Xing, Wuxiang Mao, Xiaolong Zheng,
Liang Xu, Jinjun Wu, Xiaocheng Weng, Shaoru Wang, Tian Tian, Bifeng Yuan, Bing Huang, Lin Zhuang,
and Xiang Zhou*
College of Chemistry and Molecular Sciences, Key Laboratory of Biomedical Polymers of the Ministry of Education, Wuhan
University, Wuhan, Hubei 430072, P. R. China
*
S Supporting Information
technique cannot address situations involving the oxidation of
thymidines. Other combinations, such as the V O /LiBr or
ABSTRACT: To surmount the challenges of the locus
determination and accurate quantification of 5-methyl-2′-
2
5
NaIO /LiBr pairings, have partially solved the aforementioned
4
5
Me
deoxycytidine ( dC) in DNA fragments that contain
multiple 5 dC residues, we designed and synthesized two
N-halogeno-N-sodiobenzenesulfonamide reagents that
detection difficulties, but the sensitivity of these assays requires
Me
13
additional improvement. Even the single molecule real time
sequencing (SMRT) technique, which can directly detect DNA
5
Me
provide a new chemical method for probing
dC in
methylation without bisulfite conversion, possesses the problem
DNA sequences. When the strategy we provided was
14,15
of a high error rate.
In addition, traditional detection
5
Me
combined with β-glucosyltransferase,
dC could be
16,17
methods such as bisulfite sequencing
are time-consuming
distinguished from 5-hydroxymethyl-2′-deoxycytidine
and cumbersome. The other reported methods for detecting
5
hm
(
dC) and deoxycytidine (dC) through the introduction
5Me
5
hm
dC are either instrument-intensive or limited by their
of a glucose moiety to the hydroxyl group of dC.
sensitivity and accuracy. In view of the aforementioned
problems with traditional sequencing methods, it is vitally
important to develop a rapid and sensitive chemical method for
the accurate quantification of the cytosine methylation status of
genomic DNA.
5
-Methylcytosine (5mC) is an essential epigenetic modification
that frequently appears in CpG sequences and acts as an
1
important factor in the silencing of genes. In recent years, with
The apparent differences between 5MedC and dC are the
nucleophilicity of the double bond in the pyrimidine rings and
the steric hindrance that is caused by the methyl group of
the development of epigenetics, high importance has been
attached to the selective detection of 5mC in genes, given the
strong correlation of this genetic modification with various
aspects of gene control, such as gene regulation, genomic
5Me
dC; this hindrance led us to identify and characterize a new
2
imprinting, and X chromosome inactivation, among other
specific base-sensitive compound for the quantitative con-
5
Me
effects. It has been reported that a high level of 5mC at CpG
islands within promoters and the global hypomethylation of
genomic DNA, which induces gene instability, can produce the
activation of oncogenes and the high occurrence of various
version of
dC or dC. Sharpless and co-workers have
18−20
reported that chloramine-T (TsN−ClNa)
can serve as an
excellent nucleophile for attacking the double bond of olefins
and their derivatives, producing outstanding yields, particularly
for substrates with electron-donating groups. The similarity of
dC and 5 dC residues to olefin derivatives prompted us to
design and synthesize various chloramine-T derivatives to
assess their activities toward the specific substrates dC and
3
−5
diseases.
Thus, identification of the 5mC level and status in
Me
genes is important for the early detection and treatment of
many tumors. However, because of the subtle differences
between deoxycytidine (dC) and 5-methyl-2′-deoxycytidine
5
Me
5Me
5Me
(
dC), distinguishing
dC from dC is a difficult and
dC.
challenging task.
Here we report on the results obtained for these derivative
compounds, N-sodio-N-chloro-p-nitrobenzenesulfonamide (1)
and N-sodio-N-bromo-m-nitrobenzenesulfonamide (2) (Figure
To date, many initiatives have attempted to address thi6s
challenge. Various methods based on restriction enzymes,
7
8
methylation-specific PCR amplification, photooxidation, and
1
). By combining the results obtained using both of these
9
DNA photoligation have been developed for the detection of
compounds, we can accurately identify the number and loci of
5
mC. Meanwhile, other chemical methods involving the
5Me
dC residues in DNA sequences. This novel indirect
5
Me
conversion of either dC or dC residues in DNA have been
detection method is capable of overcoming the problems
discussed above and can detect multiple 5 dC residues and
loci at a time, satisfying the need for the reliable detection of
proposed. For instance, the Maxam−Gilbert chemical mod-
Me
5
Me
ification was applied to identify the dC residues indirectly by
means of the interference of the methyl group in the reaction
5Me
1
0
dC.
with hydrazine. Moreover, OsO has been assessed as a
4
5
Me
potential reagent for differentiating between
dC and dC
because OsO reacts differently to the distinct nucleophilicities
Received: November 16, 2012
Published: January 9, 2013
4
5
Me
11,12
of the double bonds in dC and dC.
However, the OsO4
©
2013 American Chemical Society
1240
dx.doi.org/10.1021/ja311229n | J. Am. Chem. Soc. 2013, 135, 1240−1243

Journal of the American Chemical Society
Communication
would be converted into two products after the treatments with
1
or 2. After desalting, the obtained DNA products were
treated with hot piperidine (90 °C, 30 min) to induce cleavage
at damaged pyrimidine bases, and the calculations of bond
energies in the pyrimidine rings were performed [see the
Supporting Information (SI)] to study the possible mechanism
(
Table S1 in the SI). The final results were analyzed through
denaturing polyacrylamide gel electrophoresis (PAGE); the
electrophoresis results are depicted in Figure 2b. In lane 1 of
this figure, the results of the reaction with 1 can be observed;
this compound presented excellent selectivity, producing an
obvious cleavage at the dC site of the tested oligodeoxynucleo-
tide. By contrast, in the reaction with 2 (Figure 2b, lane 2), two
Figure 1. Structures of the two compounds we designed and
synthesized.
At first, we prepared oligodeoxynucleotide 1 (ODN 1), a
short sequence of VHL tumor-suppressor gene containing both
dC and 5 dC residues. ODN 1 was subjected to two
different treatments, each of which was expected to target
different base residues selectively, allowing us either directly or
Me
21
5Me
products generated by cleavage at the dC and dC sites of the
target DNA sequence can clearly be observed. Impressively,
both compounds demonstrated high selectivity, with negligible
5
Me
13,22
indirectly to distinguish between dC and
treatments employed a mixture of 100 mM Tris-HCl buffer
pH 5.0), 2 mM 1 or 2, and 50% acetonitrile. The difference
dC. Both
reaction at the vulnerable dT and dG sites
of the DNA.
Moreover, by subtracting the lane 1 results from the lane 2
results, we could indirectly obtain an accurate quantification of
the number and loci of the 5 dC residues in the tested DNA;
(
Me
between the two treatments was that the reaction with 1 was
5
Me
performed with 100 μM I and incubated at 60 °C for 1 h,
thus, this method possesses the potential to detect
dC
2
whereas the other reaction was performed under milder
conditions, proceeding at 50 °C for 10 min without any
catalyst. As shown in Figure 2a, it was expected that ODNs
residues in a convenient and precise way.
The sensitivity and accuracy of this method were tested later
by applying this method to longer sequences containing
multiple dC and 5 dC residues. All of these prepared ODNs
were short sequences of VHL tumor-suppressor gene.
Me
5
Me
Denaturing PAGE analyses of ODN 4 containing five
dC
residues after treatment with 1 and 2 using the protocol
described above are depicted in Figure 3a, which clearly
indicates the number and exact loci of the 5 dC residues in the
tested DNA sequence without producing any additional
cleavage at the dT and dG sites. To ensure the selective
Me
5
Me
cleavage of dC and
eight
dC a longer sequence (50 mer) with
5
Me
dC residues (ODN 5; Figure 3b), a lower
concentration of 2 (250 μM) and a lower pH (3.0) were
adopted. PAGE analyses of other strands with different
numbers of 5 dC residues after cleavage by 1 and 2 and hot
piperidine treatment are presented in Figure S1 in the SI. When
the strategy was performed directly on the double-stranded
DNA (dsDNA), we still observed the same piperidine-sensitive
cleavage sites compared with single-stranded DNA (ssDNA).
Thus, this method can also be used to identify the amount and
loci of 5 dC in dsDNA (Figure S2).
Me
Me
The methylation-selective methods that have previously been
reported in the literature lack the ability to discriminate
23
between 5-hydroxymethylcytosine (5hmC) and 5mC, which
limits their applications with respect to the detection of
methylation. To confirm that our approach demonstrates a
preference for 5 dC, synthetic 24-mer single-stranded ODN 3′
Me
containing two dC, two 5 dC, and one
Me
5hm
dC was designed.
Another single-stranded DNA, ODN 3, which had the same
5
hm
sequence as ODN 3′ except that the
dC residue was
replaced with another 5 dC residue, was chosen as a control.
Applying the method provided, we observed a relatively weaker
band at the site of 5 dC from this comparative examination.
Me
Figure 2. (a) Proposed products of reactions of ODNs 1 (blue) and 2
hm
(
green). (b) Sequence of ODN 1 and PAGE analysis of this ODN
5
Me
after relevant treatments. Lane 1: labeled DNA was incubated at 60 °C
However, the difference in steric hindrance due to
dC and
5hm
5m
for 1 h in a mixture of 100 mM Tris-HCl buffer (pH 5.0), 100 μM I , 2
2
dC is not sufficient for the discrimination of dC and
dC using this method. Thus, we introduced a glucose moiety
to the hydroxyl group of dC by using β-glucosyltransferase
to increase the steric hindrance of
As shown in Figure 4, the band corresponding to
mM 1, and 50% acetonitrile, giving a clearly visible cleavage at the dC
site. Lane 2: two cleavage products at the dC and 5MedC sites were
observed when labeled DNA was incubated in a solution containing 2
mM 2, 100 mM Tris-HCl buffer (pH 5.0), and 50% acetonitrile at 50
5hm
5hm
1
5,24
5hm
(
β-GT)
dC further.
5
hm
°
C for 10 min. Lanes 3−5: Maxam−Gilbert G, A + G, and C
dC
sequencing lanes.
completely disappeared in lane 4, indicating that the double
1
241
dx.doi.org/10.1021/ja311229n | J. Am. Chem. Soc. 2013, 135, 1240−1243

Journal of the American Chemical Society
Communication
Figure 4. Sequences of single strands ODN 3, ODN 3′ and ODN 3″
and polyacrylamide gel electrophoresis (PAGE) analysis of these
sequences. Lane 1, ODN 3′ was treated with 1, indicating the locations
of dC. Lane 2, ODN 3 was treated with 2 under the condition
described above, demonstrating the amount and locations of dC and
5
Me
dC, which was used as a comparation to lane 3 and lane 4. Lane 3,
5hm
ODN 3′ that contain
condition as lane 2, demonstrating a relative weaker band
corresponding to 5hmC. Lane 4, 5hmC was catalyzed to form
dC residue was performed at the same
5
gmC by β-glucosyltransferase and then the corresponding DNA 3″
was reacted with 2 under the same condition. No visible cleavage was
observed at the site of 5gmC.
mechanisms we have proposed in Scheme S1 in the SI. In
particular, we suggest that in the reaction with 1, iodine reacts
+
with 1 to produce sodium iodide and an I source, which then
reacts with cytosine to form an iodonium cation intermedi-
1
8,25
+
ate.
Because of the large atomic radius of iodine, I has a
preference for dC because of the steric hindrance created by the
methyl group in 5 dC. Negligible cleavage of dT residues by 1
was observed for the same reason. However, the reaction with 2
Figure 3. (a) Sequence of ODN 4 and PAGE analysis of the reaction.
Lanes 1 and 2: DNA was treated with 1 and 2, respectively, under the
same reaction conditions described in Figure 2. Lanes 3−5: Maxam−
Gilbert G, A + G, and C sequencing lanes for ODN 4. (b) Sequence of
ODN 5 and PAGE analysis of this ODN after the treatment described
above. Lanes 1−3: Maxam−Gilbert G, A + G, and C sequencing lanes.
Lane 4: ODN 5 was treated with 1 under the same reaction conditions
as described above. Lane 5: ODN 5 was treated with a solution
containing 250 μM 2, 100 mM Tris-HCl buffer (pH 3.0), and 50%
acetonitrile at 50 °C for 10 min.
Me
+
can generate a Br source through an uncatalyzed hydrolysis.
Because of the different radii of iodine and bromine atoms,
both 5 dC and dC can readily form a bromonium cation
Me
5gm
intermediate, but it is hard for dC to form this intermediate
+
+
because of its larger steric hindrance. Both I and Br ions
barely attack dT residues because of their electron-deficient
C5−C6 double bonds and the steric hindrance of the methyl
group at their C5 sites.
bond of 5hmdC is protected from forming bromonium cation
intermediate as a result of the increased steric effect.
Initially, various other Br and I sources were assessed for
+
+
5
Me
selectivity at the C5−C6 double bonds of dC and dC. Like
Otherwise, excitingly, 1 demonstrates the particular ability to
cleave dC quantitatively (lanes 1 in Figure 3a,b) without
significant cleavage at any other residues. Therefore, this
compound can offer another alternative for preparing C lanes.
Furthermore, because of the toxicity and explosiveness of
hydrazine, the Maxam−Gilbert method is a much more
dangerous and cumbersome way to create C ladders than our
one-step strategy.
1, chloramine-T also possesses the ability to generate iodonium
cation intermediates in a reaction catalyzed by I . However,
2
because of the electron-donor group at the para position on the
benzene ring of chloramine-T, the cleavage of dC by
chloramine-T is relatively minimal, as observed in PAGE
analyses that demonstrated a lower yield for the reaction with
chloramine-T than the reaction with 1 (Figure S3). A variety of
other catalysts (e.g., CuCl , CuCl, and KOsO ) were tested, but
2
4
5
Me
To provide further confirmation of the mechanism of the
selective transformation of dC and 5MedC, we analyzed the
MALDI-TOF mass spectrometry data for an ODN with the
same sequence as ODN 1 but without the fluorophore (Figure
S11) after it had first been treated with 1 or 2 and then purified
by HPLC (Figure S12). These data indicate that the possible
products from the reaction are substitutions of hydrogen in dC
no reactions could be observed at the sites of
dC or dC
residues (Figure S4). Thus, it appears that the reaction of 1 at
+
dC sites can be attributed to the formation of I . Meanwhile, a
variety of substituted bromoamine-T derivatives were designed
and screened with the goal of selecting the most sensitive one
among the examined compounds. Several of the tested
compounds, such as bromamine-B (PhSO −NBrNa), which
2
5Me
or dC residues; these substitutions could occur through the
contains no substituent on its benzene ring, also demonstrated
1
242
dx.doi.org/10.1021/ja311229n | J. Am. Chem. Soc. 2013, 135, 1240−1243

Journal of the American Chemical Society
Communication
the ability to react with 5MedC and dC; however, some cleavage
bands were relatively weaker when compared with 2.
Furthermore, the same results were obtained when the
ODNs were treated with either CH PhSO −NBrNa or p-
Bourget, J. A.; Stewart, R.; Ruotti, V.; Millar, A. H.; Thomson, J. A.;
Ren, B.; Ecker, J. R. Nature 2009, 462, 315.
(
(
(
(
2) Bird, A. Genes Dev. 2002, 16, 6.
3) Counts, J. L.; Goodman, J. I. Cell 1995, 83, 13.
4) Beck, S.; Rakyan, V. K. Cell 2008, 24, 231.
5) Keshet, I.; Schlesinger, Y.; Farkash, S.; Rand, E.; Hecht, M.; Segal,
3
2
NO PhSO −NBrNa (Figure S5). In general, the results
2
2
obtained above demonstrate that all of the bromamine-T
derivatives basically have the ability to produce piperidine-
E.; Pikarski, E.; Young, R. A.; Niveleau, A.; Cedar, H.; Simon, I. Nat.
Genet. 2006, 38, 149.
5
Me
sensitive cleavage sites at dC and
dC residues upon
(6) (a) Issa, J. P.; Ottaviano, Y. L.; Celano, P.; Hamilton, S. R.;
treatment under the optimum conditions. However, because
of the weaker cleavage bands at some sites when these other
derivatives were used, we chose the relatively better one,
compound 2.
Davidson, N. E.; Baylin, S. B. Nat. Genet. 1994, 7, 536. (b) Xiong, Z.
G.; Laird, P. W. Nucleic Acids Res. 1997, 25, 2532.
(7) Herman, J. G.; Graff, J. R.; Myohanen, S.; Nelkin, B. D.; Baylin, S.
B. Proc. Natl. Acid. Sci. U.S.A. 1996, 93, 9821.
(
8) Tanabe, K.; Yamada, H.; Nishimoto, S. J. Am. Chem. Soc. 2007,
29, 8034.
9) (a) Ami, T.; Ogino, M.; Taya, Y.; Takemura, Y.; Fijimoto, K.
Meanwhile, the influences of the pH and reaction temper-
ature were also tested. The results show that the pH has little
influence on the reaction with 1 but greatly influences the result
of the reaction with 2 because of the hydrolytic process. In
addition, the reaction temperature impacts both reactions.
Lower temperature may lead to lower yields, while higher
temperature can decrease the selectivity, producing unwanted
cleavage bands at the sites of dT (Figures S6 and S7). Besides,
the acetonitrile solvent was added to the solution to increase
1
(
Nucleic Acids Symp. Ser. 2009, 53, 203. (b) Ogino, M.; Taya, Y.;
Fujimoto, K. Org. Biomol. Chem. 2009, 7, 3163.
(10) Maxam, A. M.; Gilbert, W. Proc. Natl. Acad. Sci. U.S.A. 1977, 74,
5
60.
(11) Okamoto, A.; Tainaka, K.; Kamei, T. Org. Biomol. Chem. 2006,
4, 1638.
(12) Tanaka, K.; Tainaka, K.; Kamei, T.; Okamoto, A. J. Am. Chem.
Soc. 2007, 129, 5612.
19
the nucleophilicity of the reaction.
(13) Bareyt, S.; Carell, T. Angew. Chem., Int. Ed. 2008, 47, 181.
In summary, we have developed a novel indirect chemical
method for distinguishing between cytosine and 5mC in DNA
sequences. In addition, this method allows us to identify the
(14) Booth, M. J.; Branco, M. R.; Ficz, G.; Oxley, D.; Krueger, F.;
Reik, W.; Balasubramanian, S. B. Science 2012, 336, 934.
15) Song, C. X.; Szulwach, K. E.; Fu, Y.; Dai, Q.; Yi, C. Q.; Li, X. K.;
(
5
Me
number and loci of the
dC residues both in ssDNA and
Chen, C. H.; Zhang, W.; Jian, X.; Wang, J.; Looney, T.; Zhang, B.;
Godley, L. A.; Hicks, L. M.; Lahn, B. T.; Jin, P.; He, C. Nat. Biotechnol.
dsDNA accurately and efficiently by combining results obtained
using the two compounds examined in this study. Furthermore,
2
011, 29, 68.
5
Me
5hm
this strategy also enabled us to distinguish
dC from
dC
(16) Cokus, S. J.; Feng, S. H.; Zhang, X. Y.; Chen, Z. G.; Merriman,
B.; Haudenschild, C. D.; Pradhan, S.; Nelson, S. F.; Pellegrini, M.;
Jacobsen, S. E. Nature 2008, 452, 215.
by increasing the steric hindrance of 5hmC using β-
glucosyltransferase.
(17) Frommer, M.; Mcdonald, L. E.; Millar, D. S.; Collis, C. M.;
Watt, F.; Grigg, G. W.; Molloy, P. L.; Paul, C. L. Proc. Natl. Acad. Sci.
U.S.A. 1992, 89, 1827.
ASSOCIATED CONTENT
Supporting Information
■
*
S
(
18) Jeong, J. U.; Tao, B.; Sagasser, I.; Henniges, H.; Sharpless, K. B.
J. Am. Chem. Soc. 1998, 120, 6844.
19) Ando, T.; Kano, D.; Minakata, S.; Ryu, I.; Komatsu, M.
General methods; synthesis and characterization data; exper-
imental details of Maxam−Gilbert methods for G, A + G, and C
sequencing lanes; other PAGE analyses; ab initio calculations of
the energies of CC bonds in the pyrimidine rings; HPLC and
MALDI-TOF MS data; and detailed proposed mechanisms of
the reactions with 1 and 2. This material is available free of
charge via the Internet at http://pubs.acs.org.
(
Tetrahedron 1998, 54, 13485.
(20) Albone, D. P.; Aujla, P. S.; Taylor, P. C. J. Org. Chem. 1998, 63,
9569.
(21) Kaelin, W. G.; Maher, E. R. Trends Genet. 1998, 14, 423.
(22) Stemmler, A. J.; Burrows, C. J. J. Biol. Inorg. Chem. 2001, 6, 100.
(23) Jin, S. G.; Kadam, S.; Pfeifer, G. P. Nucleic Acid Res. 2010, 38,
No. e125.
AUTHOR INFORMATION
Corresponding Author
xzhou@whu.edu.cn
■
(24) Pastor, W. A.; Pape, U. J.; Huang, Y.; Henderson, H. R.; Lister,
R.; Ko, M.; McLoughlin, E. M.; Brudno, Y.; Mahapatra, S.; Kapranov,
P.; Tahiliani, M.; Daley, G. Q.; Liu, X. S.; Ecker, J. R.; Milos, P. M.;
Agarwal, S.; Rao, A. Nature 2011, 473, 394.
Author Contributions
T.W. and T.H. contributed equally.
†
(25) Itoh, N.; Izumi, Y.; Yamada, H. Biochemistry 1987, 26, 282.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
X. Zhou thanks the National Basic Research Program of China
973 Program) (2012CB720600 and 2012CB720603), the
■
(
National Natural Science Foundation of China (91213302), the
National Grand Program on Key Infectious Disease
(
2012ZX10003002-014), and the Program for Changjiang
Scholars and Innovative Research Team in University
IRT1030).
(
REFERENCES
■
(
1) Lister, R.; Pelizzola, M.; Dowen, R. H.; Hawkins, R. D.; Hon, G.;
Filippini, J. T.; Nery, J. R.; Lee, L.; Ye, Z.; Ngo, Q. M.; Edsall, L.;
1
243
dx.doi.org/10.1021/ja311229n | J. Am. Chem. Soc. 2013, 135, 1240−1243