Site-selective depurination by a periodate-dependent deoxyribozyme

Article abstract of DOI:10.1039/b704507g

A deoxyribozyme is identified that mediates the site-selective depurination of its 5′-terminal guanosine nucleotide using periodate (IO ₄^-) as an obligatory cofactor. The Royal Society of Chemistry.

Full text of DOI:10.1039/b704507g

View Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Site-selective depurination by a periodate-dependent deoxyribozyme{
Claudia Ho¨bartner, P. I. Pradeepkumar and Scott K. Silverman*
Received (in Cambridge, UK) 26th March 2007, Accepted 3rd May 2007
First published as an Advance Article on the web 10th May 2007
DOI: 10.1039/b704507g
The 112-mer 10FN10 deoxyribozyme was incubated with
40 mM NaIO4 in the absence of UDP-GlcNAc in 50 mM
NaOAc (pH 5.2) at 37 uC for 2 h. After precipitation to remove
periodate, the DNA was subsequently treated with a 30-mer
39-NH2 oligonucleotide and 10 mM NaCNBH3. PAGE revealed
that ca. 10–20% of the DNA was converted to a slower-migrating
band (Fig. 1B). This is consistent with periodate-induced
formation of an amine-reactive functional group on the DNA
and subsequent reductive amination with the added 39-NH2
oligonucleotide. Treatment with NaCNBH3 during the latter step
was required to observe the product.§
A deoxyribozyme is identified that mediates the site-selective
depurination of its 59-terminal guanosine nucleotide using
periodate (IO42) as an obligatory cofactor.
Nucleic acid enzymes (ribozymes and deoxyribozymes) that use
small-molecule compounds as cofactors are interesting from both
fundamental and applied viewpoints. The first report of a cofactor-
dependent nucleic acid enzyme was of a deoxyribozyme that uses
histidine to facilitate RNA cleavage.1 Oxidative cleavage of DNA
by a deoxyribozyme can be achieved using Cu2+ either with or
without ascorbate,2–4 and alcohol dehydrogenation or aldehyde
reduction can be performed using NAD+ or NADH, respec-
tively.5,6 A DNA aptamer for the fluorophore dihydrotetramethyl-
rosamine has weak redox activity.7 The discovery that the natural
RNA-cleaving glmS ribozyme requires glucosamine 6-phosphate
(GlcN6P) as an obligatory coenzyme8,9 suggests that the use of
cofactors by nucleic acid enzymes could have practical value. This
is particularly important as efforts expand to apply RNA and
DNA in many contexts, such as sensor applications.10,11
We used strategically 29-ribo-modified versions of 10FN10 to
narrow the DNA reaction site by using internal transesterification
at phosphorus (‘‘alkaline hydrolysis’’) to follow the location of a
32P-radiolabel. Inclusion of a 29-hydroxyl group at G16 of 10FN10
did not affect the reductive amination yield (not shown). The
reductive amination product from rG16-10FN10 with a 15-mer
39-NH2 oligonucleotide was 59-32P-radiolabeled, and subsequent
cleavage by alkaline hydrolysis led to the expected product
(Fig. 2A, rG16 lanes). Inclusion of 29-hydroxyl groups at both G2
and G16 decreased the reductive amination yield by about 4-fold,
but the rG2,rG16-10FN10 reductive amination product was still
readily isolated by PAGE and 59-32P-radiolabeled. Cleavage of this
product by alkaline hydrolysis showed the band expected if
attachment of the 39-NH2 oligonucleotide occurred at either the
first or second nucleotide (G1 or G2; Fig. 2A, rG2,rG16 lanes).
Importantly, attachment at the third nucleotide (A3) or further
towards the 39-end of 10FN10 should lead to a larger cleavage
band, which was not observed. Therefore, in the reductive
As part of our overall efforts to investigate the catalytic abilities
of DNA, we designed a new in vitro selection experiment that was
intended to use uridine 59-diphospho-N-acetylglucosamine (UDP-
GlcNAc) as an electrophilic small-molecule substrate for attack by
the DNA 59-hydroxyl group as nucleophile, thereby attaching
GlcNAc to the DNA. Separation of catalytically active deoxyri-
bozymes during the selection procedure was to be achieved via a
three-step procedure: (i) NaIO4 oxidation of the vicinal diol at
positions 3 and 4 of the newly appended GlcNAc sugar ring; (ii)
reductive amination with NaCNBH3 to attach
a 39-NH2
oligonucleotide; and (iii) PAGE separation of the deoxyribozyme
sequences that were increased in length due to attachment of the
39-NH2 oligonucleotide.{ After 10 selection rounds, we were
surprised to find that the N70 DNA pool had attained substantial
catalytic activity (65-fold above background; Fig. 1A) that was
independent of UDP-GlcNAc but required NaIO4. In particular,
omission of periodate led to undetectable activity (,0.5%).
Furthermore, 59-phosphorylation of the DNA pool suppressed
the activity by at least 10-fold.§ Apparently, periodate leads to
formation of an amine-reactive moiety on the DNA, likely near
the 59-terminus, without first requiring attachment of GlcNAc that
would be oxidized by the periodate. One individual deoxyribo-
zyme, 10FN10, was chosen for further characterization of its
periodate-dependent reaction.
Fig. 1 Periodate-induced formation of an amine-reactive moiety by the
10FN10 deoxyribozyme. (A) Selection progression, shown as activity of
the uncloned selection pool versus round number. A positive control
substrate that bears a covalently attached GlcNAc moiety§ gave 26–44%
yield when tested in parallel during each round. (B) Reaction of the
periodate-treated round 10 DNA pool or the 10FN10 deoxyribozyme with
a 30-mer 39-NH2 oligonucleotide and NaCNBH3 leads to 10–20% of the
reductive amination product. The unreacted pool is shown as a standard.
The slight gel shift between the standard and the other two lanes is from
inclusion of a DNA splint that holds together the deoxyribozyme and the
39-NH2 oligonucleotide (see supplementary information for reaction
details).
Department of Chemistry, University of Illinois at Urbana-Champaign,
600 South Mathews Avenue, Urbana, Illinois 61801, USA.
E-mail: scott@scs.uiuc.edu; Fax: 217 244 8024; Tel: 217 244 4489
{ Electronic supplementary information (ESI) available: Details of all
experimental procedures. See DOI: 10.1039/b704507g
This journal is ß The Royal Society of Chemistry 2007
Chem. Commun., 2007, 2255–2257 | 2255

View Online
Fig. 2 Narrowing the site of reactivity to either G1 or G2 of the 10FN10
deoxyribozyme. (A) Gel image showing cleavage by alkaline hydrolysis (+)
of the 59-32P-radiolabeled reductive amination product. 29-Hydroxyl
groups were incorporated into 10FN10 either at G16 (rG16) or at both
G2 and G16 (rG2,rG16). Std = 15-mer or 30-mer 39-NH2 oligonucleotide.
(B) Schematic depictions of the expected rG16-10FN10 and rG2,rG16-
10FN10 cleavage products if the 15-mer 39-NH2 oligonucleotide (zigzag
line) becomes attached at nucleotide G1, G2, or A3. For attachment at
any of the first 16 positions, the control rG16-10FN10 cleavage product
would be slightly larger than the 30-mer standard (i.e., 31-mer plus cyclic
phosphate), as observed. Only in the case of attachment at G1 or G2
would the rG2,rG16-10FN10 cleavage product be slightly larger than the
15-mer standard (i.e., 17-mer plus cyclic phosphate), as observed. For
attachment at A3 or any other position after G2, the rG2,rG16-10FN10
cleavage product would be a 29-mer, which is not observed (,0.5%).
Scheme 1 Structural possibilities 1–3 for the product of reaction at
nucleotide G1 or G2 of the 10FN10 deoxyribozyme before reductive
amination. Compounds 1 and 2 represent depurination at G1 and G2,
respectively, whereas 3 depicts oxidation of the 59-terminal hydroxyl
group. Aldehyde functional groups are marked with dashed arrows, and
guanine nucleobases are shown as G.
NaBH4 are inconsistent with the deoxyribonolactone structure,
which should not require NaCNBH3 to form a stable adduct with
a 39-NH2 oligonucleotide and should not be reduced by NaBH4.
To determine more specifically which of 1–3 corresponds to the
periodate oxidation product, we performed the oxidation and
reductive amination sequence using the 29-ribo-G16 version of the
10FN10 deoxyribozyme (rG16-10FN10; cf. Fig. 2) and a 15-mer
39-NH2 oligonucleotide. The product was PAGE-purified, and
alkaline hydrolysis (i.e., cleavage at rG16) was used to remove the
39-terminal portion of the deoxyribozyme (this provides a smaller
product and therefore more accurate mass determinations).
Without further purification, analysis of the cleaved product by
MALDI-MS revealed that the fragment containing both the
39-NH2 oligonucleotide and the 59-terminal portion of the
deoxyribozyme has m/z 9648 ¡ 10 (Fig. 3)." The calculated m/z
values for the reductive amination products of 1a, 1b, 2, and 3 are
9691, 9643, 9691, and 9822, respectively (note that the NaIO4
oxidation 1a A1b removes two heavy atoms, as shown in Scheme 1,
and the product from 3would have much higher mass because it has
not been depurinated). Of 1–3, the observed m/z is consistent only
with 1b as the structure of the product before reductive amination.
The selectivity for formation of 1b over alternatives such as 2 or 3
can be conservatively estimated as ¢ 90%, based on the lack of
peaks at either higher expected m/z value. Therefore, we conclude
that NaIO4 induces site-selective depurination of the 10FN10
deoxyribozyme at nucleotide G1, leading to 1a. This subsequently
forms 1b after periodate oxidation of the vicinal diol.
amination product the 39-NH2 oligonucleotide is attached at either
G1 or G2.
We considered three likely candidates on the 10FN10
deoxyribozyme for the attachment site of the 39-NH2 oligonucleo-
tide by reductive amination (Scheme 1). First, the guanine
nucleobase of G1 could be lost (i.e., depurination), followed by
reductive amination of the 39-NH2 oligonucleotide with the
resulting abasic site (1). The initially unmasked aldehyde (1a) is
also a 1,2-diol; therefore, 1a could be oxidized to a dialdehyde (1b).
Second, the guanine nucleobase of G2 could be lost, leading to an
abasic site that cannot be oxidized to a dialdehyde (2). Third, the
59-hydroxyl group on nucleotide G1 of the DNA could be oxidized
to a 59-aldehyde (3). Sodium periodate was reported to oxidize
benzylic alcohols under relatively harsh conditions,12 providing
some precedent for this type of reaction (which would clearly
require a catalyst to occur under mild conditions).
All three structures 1–3 are compatible with the outcomes of
two additional experiments.§ First, reduction of the NaIO4-treated
product with NaBH4 before reductive amination entirely sup-
pressed formation of the reductive amination product, consistent
with direct reduction of an amine-reactive aldehyde functional
group. Second, heating in 90 mM NaOH at 95 uC for 5 min after
NaIO4 treatment entirely suppressed formation of the reductive
amination product, consistent with strand scission (and therefore
loss of the aldehyde functional group) induced by removal of a
proton a to the carbonyl group.13 This latter experiment is also
consistent with a fourth possibility for the oxidation product, a
deoxyribonolactone,14 which could react with a 39-NH2 oligonu-
cleotide to form an amide. However, both the NaCNBH3
requirement and the suppression of product formation by
The mechanism of periodate-induced depurination at G1 is
uncertain. Several additional experiments provide some clues.§
KIO4 is as effective a cofactor as NaIO4, but activity is
undetectable (,0.5%) if the nucleotide at position 1 is mutated
2256 | Chem. Commun., 2007, 2255–2257
This journal is ß The Royal Society of Chemistry 2007

View Online
guanosine nucleotide. The depurination reaction requires periodate
as cofactor and does not involve divalent metal ions or H2O2.
Cofactor-independent DNA depurination (deglycosylation) reac-
tions have been reported,18,19 but the requirement observed here
for periodate indicates that 10FN10 leads to a different type of
depurination reaction. The DNA-catalyzed reaction is site-selective
because we do not detect any products arising from depurination
at sites other than G1 (e.g., G2). In contrast, the Cu2+-dependent
oxidative self-cleavage of several deoxyribozymes2–4 is not highly
site-selective; indeed, the original report described those self-
cleavage reactions as merely ‘‘region-specific’’.4 Therefore, in
several characteristics the 10FN10 deoxyribozyme provides a
novel example of a cofactor-dependent DNA-catalyzed reaction.
This work was supported by the US National Institutes of
Health (GM-65966 to S.K.S.) and an Erwin Schro¨dinger
Fellowship from the Austrian Science Fund (to C.H.). We thank
Marc Greenberg for comments on the manuscript and Shana
Kelley for discussion of potential depurination mechanisms.
Fig. 3 MALDI mass spectrometry data showing that NaIO4 induces
reaction of the 10FN10 deoxyribozyme by depurination of G1 to form 1a
and subsequent oxidation to form 1b. (A) Data trace obtained after
alkaline cleavage of the reductive amination product from the rG16-
10FN10 deoxyribozyme with the 15-mer 39-NH2 oligonucleotide. The
filled circle marks the peak location expected if 1b is the final product of
NaIO4 treatment; the open circle marks the location expected for either 1a
or 2; and the X marks the location expected for 3. (B) Assigned structure
of the observed reductive amination product from 1b.
Notes and references
{ The selection was performed using procedures similar to those in our
previous report.20 See supplementary information for additional details of
the selection procedure.
§ See supplementary information for experimental details.
" See supplementary information for additional details of MALDI sample
preparation and analysis.
I Analysis of 10FN10 using the mfold secondary structure prediction
algorithm did not suggest any strong secondary structure elements other
than those involving the 59-terminal nucleotides (see supplementary
information). We have not performed any systematic experiments to
investigate the secondary structure of 10FN10 in its 59-terminal region.
to anything other than guanosine.I Modest activities (y5–6-fold
lower than with NaIO4) were observed with 40 mM NaClO4 or
1 mM KMnO4; this has not been investigated further. Oxidants
such as H2O2 or ammonium persulfate led either to no observed
reaction or to nonspecific DNA degradation, depending on
1 A. Roth and R. R. Breaker, Proc. Natl. Acad. Sci. U. S. A., 1998, 95,
6027–6031.
2 N. Carmi, S. R. Balkhi and R. R. Breaker, Proc. Natl. Acad. Sci.
U. S. A., 1998, 95, 2233–2237.
concentration. Divalent metal ions such as Mg2+, Ca2+, Zn2+
,
Co2+, and Cu2+ (which were present during the original selection
procedure) are not required for the reaction, which occurs with the
same yield even in the presence of 5 mM of the chelator EDTA.
Our data indicate that periodate is obligatory for the initial
conversion of the 10FN10 DNA to 1a and not required solely for
1a A 1b, which is a straightforward oxidation of a vicinal diol. If
periodate were dispensable for formation of 1a, then omission of
periodate would lead to formation of the reductive amination
product from 1a (instead of 1b),13 but such a product is not
observed in the absence of periodate. Moreover, periodate has not
to our knowledge been reported to induce nonspecific oxidation or
depurination of DNA, and indeed we do not observe any such
nonspecific reactivity.§ One hypothesis for the role of periodate in
the initial formation of 1a from 10FN10 is that the deoxyribozyme
uses periodate as a cofactor to oxidize its 59-terminal G1
nucleobase,15 and the oxidized nucleoside then spontaneously
depurinates. An alternative hypothesis is that periodate oxidizes
some other site on the deoxyribozyme, forming a nucleophile that
catalyzes depurination in analogy to certain DNA repair
enzymes.16,17 Because the mass spectrometry data are inconsistent
with alteration of nucleotides G2 through G16 of the deoxyr-
ibozyme, any such oxidation must occur after position 16. Both of
these hypotheses are reasonable, and further mechanistic investiga-
tions are required for their evaluation.
3 N. Carmi and R. R. Breaker, Bioorg. Med. Chem., 2001, 9, 2589–2600.
4 N. Carmi, L. A. Shultz and R. R. Breaker, Chem. Biol., 1996, 3,
1039–1046.
5 S. Tsukiji, S. B. Pattnaik and H. Suga, J. Am. Chem. Soc., 2004, 126,
5044–5045.
6 S. Tsukiji, S. B. Pattnaik and H. Suga, Nat. Struct. Biol., 2003, 10,
713–717.
7 C. Wilson and J. W. Szostak, Chem. Biol., 1998, 5, 609–617.
8 T. J. McCarthy, M. A. Plog, S. A. Floy, J. A. Jansen, J. K. Soukup and
G. A. Soukup, Chem. Biol., 2005, 12, 1221–1226.
9 J. C. Cochrane, S. V. Lipchock and S. A. Strobel, Chem. Biol., 2007, 14,
97–105.
10 R. Fiammengo and A. Ja¨schke, Curr. Opin. Biotechnol., 2005, 16,
614–621.
11 N. K. Navani and Y. Li, Curr. Opin. Chem. Biol., 2006, 10, 272–281.
12 T. M. Shaikh, L. Emmanuvel and A. Sudalai, J. Org. Chem., 2006, 71,
5043–5046.
13 J. Lhomme, J. F. Constant and M. Demeunynck, Biopolymers, 1999, 52,
65–83.
14 L. Xue and M. M. Greenberg, Angew. Chem., Int. Ed., 2007, 46,
561–564.
15 S. Bjelland and E. Seeberg, Mutat. Res., 2003, 531, 37–80.
16 A. K. McCullough, M. L. Dodson and R. S. Lloyd, Annu. Rev.
Biochem., 1999, 68, 255–285.
17 O. D. Scha¨rer, Angew. Chem., Int. Ed., 2003, 42, 2946–2974.
18 T. L. Sheppard, P. Ordoukhanian and G. F. Joyce, Proc. Natl. Acad.
Sci. U. S. A., 2000, 97, 7802–7807.
19 O. Amosova, R. Coulter and J. R. Fresco, Proc. Natl. Acad. Sci. U. S. A.,
2006, 103, 4392–4397.
20 A. Flynn-Charlebois, Y. Wang, T. K. Prior, I. Rashid, K. A. Hoadley,
R. L. Coppins, A. C. Wolf and S. K. Silverman, J. Am. Chem. Soc.,
2003, 125, 2444–2454.
In summary, we have identified the 10FN10 deoxyribozyme
that mediates site-selective depurination of its 59-terminal
This journal is ß The Royal Society of Chemistry 2007
Chem. Commun., 2007, 2255–2257 | 2257