The surprising pairing of 2-aminoimidazo[1,2-a][1,3,5]triazin-4-one, a component of an expanded DNA alphabet

Article abstract of DOI:10.1107/S2053229618016923

Synthetic biologists demonstrate their command over natural biology by reproducing the behaviors of natural living systems on synthetic biomolecular platforms. For nucleic acids, this is being done stepwise, first by adding replicable nucleotides to DNA, and then removing its standard nucleotides. This challenge has been met in vitro with `six-letter' DNA and RNA, where the Watson–Crick pairing `concept' is recruited to increase the number of independently replicable nucleotides from four to six. The two nucleobases most successfully added so far are Z and P, which present a donor–donor–acceptor and an acceptor–acceptor–donor pattern, respectively. This pair of nucleobases are part of an `artificially expanded genetic information system' (AEGIS). The Z nucleobase has been already crystallized, characterized, and published in this journal [Matsuura et al. (2016). Acta Cryst. C72, 952–959]. More recently, variants of Taq polymerase have been crystallized with the pair P:Z trapped in the active site. Here we report the crystal structure of the nucleobase 2-aminoimidazo[1,2-a][1,3,5]triazin-4-one (trivially named P) as the monohydrate, C₅H₅N₅O·H₂O. The nucleobase P was crystallized from water and characterized by X-ray diffraction. Interestingly, the crystal structure shows two tautomers of P packed in a Watson–Crick fashion that cocrystallized in a 1:1 ratio.

Full text of DOI:10.1107/S2053229618016923

research papers
The surprising pairing of 2-aminoimidazo[1,2-a]-
[1,3,5]triazin-4-one, a component of an expanded
DNA alphabet
ISSN 2053-2296
Roberto Laos,a Christos Lampropoulosb and Steven A. Bennera,c
*
aFoundation for Applied Molecular Evolution (FfAME), 13709 Progress Blvd, Box 7, Alachua, FL 32615, USA,
bDepartment of Chemistry, University of North Florida, 1 UNF Drive, Jacksonville, FL 32224, USA, and cFirebird
Biomolecular Sciences LLC, 13709 Progress Blvd, Box 7, Alachua, FL 32615, USA. *Correspondence e-mail:
manuscripts@ffame.org
Received 7 September 2018
Accepted 28 November 2018
Edited by D. S. Yufit, University of Durham,
England
Synthetic biologists demonstrate their command over natural biology by
reproducing the behaviors of natural living systems on synthetic biomolecular
platforms. For nucleic acids, this is being done stepwise, first by adding replicable
nucleotides to DNA, and then removing its standard nucleotides. This challenge
has been met in vitro with ‘six-letter’ DNA and RNA, where the Watson–Crick
pairing ‘concept’ is recruited to increase the number of independently replicable
nucleotides from four to six. The two nucleobases most successfully added so far
are Z and P, which present a donor–donor–acceptor and an acceptor–acceptor–
donor pattern, respectively. This pair of nucleobases are part of an ‘artificially
expanded genetic information system’ (AEGIS). The Z nucleobase has been
already crystallized, characterized, and published in this journal [Matsuura et al.
(2016). Acta Cryst. C72, 952–959]. More recently, variants of Taq polymerase
have been crystallized with the pair P:Z trapped in the active site. Here we
report the crystal structure of the nucleobase 2-aminoimidazo[1,2-a][1,3,5]-
triazin-4-one (trivially named P) as the monohydrate, C5H5N5OꢀH2O. The
nucleobase P was crystallized from water and characterized by X-ray diffraction.
Interestingly, the crystal structure shows two tautomers of P packed in a
Watson–Crick fashion that cocrystallized in a 1:1 ratio.
Keywords: artificially expanded genetic infor-
mation system (AEGIS); heterocycle; crystal
structure; synthetic biology; imidazotriazinone;
expanded DNA alphabet.
CCDC reference: 1881877
Supporting information: this article has
supporting information at journals.iucr.org/c
1. Introduction
Synthetic biologists have rearranged hydrogen-bond donor
and acceptor moieties in the bases of DNA nucleotides to add
four new base pairs to the two found in standard DNA (i.e.
G:C and A:T) (Benner et al., 2016). These additional pairs fit
the Watson–Crick geometry but, because their hydrogen-
bonding patterns are unique, can be replicated independently
of standard pairs as part of an ‘artificially expanded genetic
information system’ (AEGIS) (Fig. 1).
One of these AEGIS pairs is formed between the nucleo-
sides 2-amino-8-(10-ꢀ-d-20-deoxyribofuranosyl)imidazo[1,2-a]-
[1,3,5]triazin-4(8H)-one (trivially named P) and 6-amino-3-
(20-deoxy-d-ribofuranosyl)-5-nitro-1H-pyridin-2-one (trivially
named Z) (Fig. 1). The Z:P pair is special among AEGIS pairs
in that both of its constituent nucleobases place electron
density into the minor groove (green lobes in Fig. 2), density
that may be a recognition feature for polymerases (Hen-
drickson et al., 2004). This may account for the relative ease
with which many DNA polymerases, RNA polymerases, and
reverse transcriptases accept dZTP and dPTP to synthesize
Z:P pairs (Yang et al., 2010; Laos et al., 2013; Leal et al., 2015).
Accordingly, the Z:P pair has been a focus of our efforts to
develop a comprehensive molecular biology for an expanded
genetic system. Those efforts have yielded improved DNA
# 2019 International Union of Crystallography
22 https://doi.org/10.1107/S2053229618016923
Acta Cryst. (2019). C75, 22–28

research papers
Figure 1
Artificially expanded genetic information system (AEGIS), where R is deoxyribose. The P nucleobase used in this study is at the bottom right.
polymerases able to replicate GACTZP DNA (Laos et al.,
2013), methods to sequence GACTZP DNA (Yang et al.,
2011), and procedures to allow GACTZP to evolve in the
laboratory to create functional molecules (Zhang et al., 2015,
2016; Sefah et al., 2014; Biondi et al., 2016). Recently, the
crystal structure of an evolved polymerase has been elucidated
while incorporating Z opposite P (Singh et al., 2018).
Crystallography of double-stranded DNA molecules that
contain Z:P pairs shows that the pairs form the expected
planar structure joined by three hydrogen bonds, even in
duplexes with six consecutive Z:P pairs (Georgiadis et al.,
2015). However, during polymerase chain reaction (PCR) with
many polymerases, Z:P pairs are slowly replaced by C:G pairs
(Fig. 2). This loss implies a mismatching between either
deprotonated Z and G (Zꢁ:G), or between C and protonated
P (C:P+), or (with a proton shift) between deprotonated G and
Z (Z:Gꢁ), and between protonated C and P (C+:P). All of
these pairs retain both size and hydrogen-bonding comple-
mentarity.
These observations drive our interest in the detailed
structure of the individual Z and P heterocycles, in part since
detailed analysis of the structural properties of the hetero-
cycles themselves might help in the design of polymerases that
exclude these charged mismatches. A structure of the
nucleobase Z, which unexpectedly cocrystallized as a mixture
of its protonated and deprotonated forms in a Z-only crystal
(Matsuura et al., 2016), has already served this role. We
considered it might be possible that P also crystallizes as a
mixture of P and P+ in a P-only crystal. Here, we report the
crystal structure of P crystallized by itself, showing that it
Figure 2
(Top left) The Z:P pair from an AEGIS. Hydrogen-bond-donor groups are in blue and hydrogen-bond-acceptor groups are in red. The green lobes
indicate electron density presented to the minor groove. The charged Zꢁ:G and C:P+ mismatches are proposed to be sources of error in polymerase-
catalyzed AEGIS DNA replication. (Bottom right) The standard Watson–Crick C:G pair, for reference.
ꢂ
Acta Cryst. (2019). C75, 22–28
Laos et al.
A component of an expanded DNA alphabet 23

research papers
Table 1
Experimental details.
solved its crystal packing problem not by protonation, but
rather by tautomerization.
Crystal data
Chemical formula
Mr
C5H5N5OꢀH2O
169.16
Monoclinic, C2/c
Crystal system, space group
Temperature (K)
2. Experimental
2.1. Synthesis of P
100
16.357 (13), 3.712 (3), 22.723 (19)
101.72 (3)
1350.9 (19)
8
Mo Kꢂ
0.13
0.40 ꢄ 0.10 ꢄ 0.10
˚
a, b, c (A)
ꢀ (ꢃ)
3
˚
The synthesis of 2-aminoimidazo[1,2-a][1,3,5]triazin-4(3H)-
one from 2,4,6-trichloro-1,3,5-triazine (‘cyanuric chloride’)
was first reported by Kim et al. (1978) (Fig. 3).
V (A )
Z
Radiation type
ꢃ (mmꢁ1
Crystal size (mm)
)
Cyanuric chloride (1) (150 g, 0.8134 mol; Caution: care
should be taken to avoid skin contact) was dissolved in
acetone (1140 ml) in a round-bottomed flask (5 l) equipped
with a mechanical stirrer and submerged in an ice/salt bath,
and a slurry of ice and water (1200 ml) was added to the
mixture, forming immediately a white precipitate. A solution
of ammonium hydroxide (324 ml, 5 M in water, 1.62 mol) was
then added dropwise over a period of 15 min using a dropping
funnel. The temperature of the reaction was monitored and
kept below 2 ꢃC by adding additional ice to the reaction
mixture. After all of the NH4OH had been added, the mixture
was incubated on an ice/salt bath for a further 15 min. The ice
bath was then removed and the reaction was allowed to
proceed without cooling for 30 min. Thin-layer chromato-
graphy (TLC; chloroform/ethyl acetate 9:1 v/v) was used to
follow the progress of the reaction; the product had an RF
value of 0.3. Product 2 (Fig. 3) was recovered by filtration,
washed with water, and dried under high vacuum (yield
95.28 g, 71%). 1H NMR (DMSO-d6): ꢁ 8.5 (s, 2H), assigned to
the NH2 protons, whose frequency depended on the dryness of
the sample.
Compound 2 (171 g, 1.0365 mol) was dissolved in aqueous
NaOH (84.5 g, 1.3620 mol, 4 l). As soon as the solution was
clear, it was acidified with acetic acid until precipitation of a
white solid ceased. The product was collected by centrifuga-
tion and dried in a lyophilizer to give 3 (yield 48.61 g, 32%). 1H
NMR (DMSO-d6): ꢁ 8.8 (s, 2H), 11.45 (s, 1H), 11.7 (s, 1H).
Compound 3 (1.0 g, 6.82 mmol) was dissolved in aqueous
NaOH (0.275 g in 20 ml water) to yield a clear solution.
Aminoacetaldehyde dimethyl acetal (0.93 ml, 8.58 mmol) was
added. The mixture was placed on an oil bath (120–130 ꢃC);
foaming lasted for 3 h. The mixture was cooled. The resulting
grey powder was recovered by filtration, washed with cold
water, and dried under high vacuum to give 4 (yield 1.3 g,
88%). 1H NMR (D2O): ꢁ 2.79 (s, 2H), 2.81 (s, 6H), 3.99 (t, J =
5.4 Hz, 1H).
Data collection
Diffractometer
Absorption correction
Bruker D8 VENTURE
Multi-scan (SADABS; Bruker,
2013)
No. of measured, independent and
observed [I > 2ꢄ(I)] reflections
Average redundancy
Rint
33557, 1397, 1086
24.021
0.089
0.626
ꢁ1
˚
(sin ꢅ/ꢆ)max (A
)
Refinement
R[F2 > 2ꢄ(F2)], wR(F2), S
No. of reflections
No. of parameters
0.060, 0.166, 1.16
1397
112
H-atom parameters constrained
H-atom treatment
˚
ꢁ3
Áꢇmax, Áꢇmin (e A
)
0.27, ꢁ0.56
Computer programs: APEX2 (Bruker, 2014), SAINT (Bruker, 2013), SHELXT2014
¨
(Sheldrick, 2015a), SHELXL2017 (Sheldrick, 2015b), ShelXle (Hubschle et al., 2011), XP
in SHELXTL-Plus (Sheldrick, 2008) and SHELXL2014 (Sheldrick, 2015b).
heated for 1.5 h. This mixture was then added to ice/water to
give a white precipitate. The mixture was neutralized with
aqueous NaOH (10 M, ca 103 ml). The precipitate was
recovered by filtration, washed with cold water, and dried to
give 5 (yield 8.33 g, 90%). 1H NMR (D2O): ꢁ 7.1 (d, J = 2.4 Hz,
1H), 7.2 (d, J = 2.4 Hz, 2H).
2.2. Crystallization
Synthetic P was recrystallized from pure water by seeding
with small crystallites of P obtained by neutralizing a basic
solution of P in water with acetic acid vapor. Crystals recov-
ered at the surface of the solution were then used as seeds to
grow larger crystals of P from a neutral water solution
(1 mg mlꢁ1). After about a month at 4 ꢃC, long needle-like
colourless crystals were recovered.
2.3. Refinement
Compound 4 (13.2 g, 61.3 mmol) was placed in a round-
bottomed flask and dissolved in cold concentrated sulfuric
acid (31 ml). The flask was placed in an oil bath at 105 ꢃC and
Crystal data, data collection and structure refinement
details are summarized in Table 1. The H atoms were obtained
from a difference Fourier map and refined either with
Figure 3
Synthesis of of imidazo[1,2-a][1,3,5]triazin-4(3H)-one, trivially called the P heterocycle.
ꢂ
24 Laos et al.
A component of an expanded DNA alphabet
Acta Cryst. (2019). C75, 22–28

research papers
Table 2
Hydrogen-bond geometry (A, ).
constraints to maintain equivalent bonds and isotropic
displacement parameters or using a riding model. The water
ꢃ
˚
˚
O—H distances were fixed at 0.98 A, and Uiso(H) values were
constrained to 1.5Ueq(O). H atoms attached to C and N atoms
˚
were placed in calculated positions (C—H = 0.95 A and
D—Hꢀ ꢀ ꢀA
D—H
Hꢀ ꢀ ꢀA
Dꢀ ꢀ ꢀA
D—Hꢀ ꢀ ꢀA
O2—H2Yꢀ ꢀ ꢀO2i
O2—H2Xꢀ ꢀ ꢀN1
O20—H2Wꢀ ꢀ ꢀO20
0.98
0.98
0.98
0.98
0.88
0.88
0.88
0.88
0.88
0.95
0.95
1.62
1.84
1.30
1.34
1.98
2.16
2.12
2.02
1.92
2.39
2.50
2.582 (6)
2.788 (5)
2.164 (15)
2.164 (15)
2.853 (5)
3.026 (5)
2.993 (12)
2.897 (3)
2.792 (5)
3.315 (5)
3.396 (4)
166
162
143
137
174
168
170
172
173
163
158
i
˚
N—H = 0.88 A) and refined as riding on their parent atoms,
ii
O20—H2Zꢀ ꢀ ꢀO20
N3—H3ꢀ ꢀ ꢀN3iii
N2—H2Aꢀ ꢀ ꢀO2i
N2—H2Aꢀ ꢀ ꢀO20
with Uiso(H) values constrained to 1.2Ueq(C,N). The aniso-
tropic displacement parameters for the O atoms of the two
partially occupied water sites were constrained to be equal.
N2—H2Bꢀ ꢀ ꢀO1iii
N8—H8ꢀ ꢀ ꢀN8iv
C7—H7ꢀ ꢀ ꢀO2v
C6—H6ꢀ ꢀ ꢀO1vi
3. Results and discussion
3.1. Description of structure
3
2
1
1
2
3
2
1
2
1
2
3
3
Symmetry codes: (i) ꢁx þ ; y þ ; ꢁz þ ; (ii) ꢁx þ ; y ꢁ ; ꢁz þ ; (iii) ꢁx þ ; ꢁy þ ,
2
2
2
1
1
2
ꢁz þ 1; (iv) ꢁx þ 1; y; ꢁz þ ; (v) ꢁx þ 1; y ꢁ 1; ꢁz þ ; (vi) ꢁx þ 1; ꢁy; ꢁz þ 1.
2
IUPAC systematic numbering has been used to better
describe the structure, which also matches the crystallographic
numbering. As seen in Figs. 4 and 5, P crystallized as a pair
joined by three hydrogen bonds (Table 2) originating from
donor atoms O1, N3, and N2. The acceptor atoms on the
symmetry-related P molecule are N2iii, N3iii, and O1iii,
respectively [symmetry code: (iii) ꢁx + 32, ꢁy + 23, ꢁz + 1]. The
Hirshfield surface plot (Hirshfeld, 1977) was generated using
CrystalExplorer17 (Turner et al., 2017). It showed short
contacts between these paired sites (Fig. 5). The fingerprint
plot for the areas of interest shows that 25.9% of all of the
intermolecular contacts arise from Oꢀ ꢀ ꢀH interactions, and
44.1% from Nꢀ ꢀ ꢀH interactions. The distance between the
donor atoms for two of these assigned hydrogen bonds are
dard valence rules in light of these distances suggest that in
this pair, atom O1 must serve as a hydrogen-bond acceptor
from the N—H donor carried by N2iii, while O1iii must serve as
a hydrogen-bond acceptor from the N—H donor carried of
N2.
However, the geometry, with valence rules, allows only a
single hydrogen to mediate the contact between N atoms 3 and
30. Thus, hydrogen 3–30 was refined with 50% occupancy and
placed in an idealized position on N3 using a riding model.
This refinement model implies that two different tautomers of
P (Fig. 6) exist in the one crystal, quite possibly in rapid
equilibrium even within the crystal. One tautomer carries its
N3 without a proton, where it serves as a hydrogen-bond
acceptor, and a proton on the nitrogen in the 30 position. In the
other tautomer, N3 carries a proton and is a hydrogen-bond
donor to nitrogen 30, and serves as a hydrogen-bond acceptor
where a proton is absent. In addition to permitting N atoms 3
and 30 to be joined by a single hydrogen, these two tautomers
permit a hydrogen bond from the nitrogen in position 8 to
nitrogen 80 of an adjacent pair to create a hydrogen-bonding
iii
iii
˚
˚
N2 ꢀ ꢀ ꢀO1 = 2.897 (3) A and N3ꢀ ꢀ ꢀN3 = 2.853 (5) A. Stan-
Figure 4
Labeled ball-and-stick representation (generated using DIAMOND;
Brandenburg et al., 1999) of the crystal structure of P, showing the
hydrogen bonding (fragmented orange bonds) between the two
tautomers paired in a Watson–Crick hydrogen-bonding fashion. The
crystallographic numbering is given as the element identifiers, which also
matches the IUPAC systematical numbering of the heterocyclic atoms.
Atom H3 was refined as being bonded to N3 with 50% occupancy, which
allows one of the N atoms to accept a hydrogen bond from the paired
N—H unit. Maintaining valence rules requires that the heterocycle with
the N—H group also lacks a hydrogen on N8, while the heterocycle
accepting the hydrogen bond at N3iii must have a hydrogen on N8iii,
suggesting tautomerization between the two heterocycles. The H atoms
that move between tautomeric inner conversion are called the 3–30
hydrogen (the hydrogen that is moving dynamically between N atoms 3
and 30) and the 8–80 hydrogen (the hydrogen that is moving dynamically
between N atoms 8 and 80). Main features of this figure: (a) Watson–Crick
fashion pairing between tautomers (N2ꢀ ꢀ ꢀO1iii, N3ꢀ ꢀ ꢀN3iii, and
O1ꢀ ꢀ ꢀN2iii); (b) hydrogen bonding shown for N atoms at the 8 and 80
positions (N8iiiꢀ ꢀ ꢀN8iv); (c) hydrogen bonding mediated by a water
molecule (N1iiiꢀ ꢀ ꢀO2). Color code: O red, N blue, C dark grey, and H light
grey. [Symmetry codes: (iii) ꢁx + 32, ꢁy + 23, ꢁz + 1; (iv) ꢁx + 1, y, ꢁz + 21.]
Figure 5
(a) Hirshfield surface of P (dnorm), showing the short contacts at the sites
of hydrogen bonding (McKinnon et al., 2004; Spackman & Byrom, 1997).
(b) The Oꢀ ꢀ ꢀH filtered fingerprint plot for the Hirshfield surface of only
the donor atoms, corresponding to 25.9% of all the intermolecular
contacts being due to Oꢀ ꢀ ꢀH interactions. (c) The Nꢀ ꢀ ꢀH filtered
fingerprint plot for the Hirshfield surface of only the donor atoms,
corresponding to 44.1% of all the intermolecular contacts due to Nꢀ ꢀ ꢀH
interactions.
ꢂ
Acta Cryst. (2019). C75, 22–28
Laos et al.
A component of an expanded DNA alphabet 25

research papers
network (Fig. 4). The presence of tautomerism was also
corroborated experimentally, with the disorder being clearly
evident in the electron-density map.
In addition, ꢈ-stacking interactions are present between
neighboring heterocycle pairs. Here, the distance between the
˚
mean heterocycle planes is 3.228 A. One disordered water
molecule is also found in the asymmetric unit. The disordered
water was refined in two positions, with O2 having 75%
occupancy and O20 having 25% occupancy. One of the two
water positions, namely the H2X—O2—H2Y water, is
hydrogen bonded to N1 with one of its H atoms, namely H2X;
˚
the H2Xꢀ ꢀ ꢀN1 distance is a standard 1.840 A. Fig. 7 shows the
labeled asymmetric unit of the compound using ellipsoids at
the 50% probability level.
Figure 7
Displacement ellipsoid drawing at the 50% probability level of the
labeled asymmetric unit of P. Atoms H3 and H8 each have 50%
occupancy to satisfy tautomerism (see text for details). Atom O2 and its
H atoms (H2X and H2Y) were refined with 75% occupancy, while O20
and its H atoms (H2W and H2Z) were refined with 25% occupancy. Color
code: O red, N blue, C dark grey, and H light grey.
3.2. Discussion
As nucleotides are added to DNA, we must revisit a range
of topics that were first visited with natural DNA over half a
century ago, but which are relevant anew now that DNA is
gaining new nucleotide ‘letters’. These include the placement
of protons on various heteroatoms in the heterocyclic rings
(tautomerism), the stacking of heterocycles on top of each
other (found in the double helix, as well as in the crystal
structures of the nucleobases), and the context dependence of
these nonstandard nucleotides in double helices. This makes it
interesting to analyze the structures of these heterocyclic
molecules.
substituents in P are analogously placed in guanine (Fig. 8), at
positions 2 and 6, given the different purine-ring numbering
nomenclature. Likewise, the planes between the stacked
˚
˚
heterocycles are similarly spaced, i.e. 3.228 A for P and 3.30 A
for guanine, as reported by Thewalt et al. (1971), and more
˚
recently reported as 3.263 A (Guille & Clegg, 2006). The
spacing between sheets is characteristic of nucleobases, for
The structure of P has some similarities and differences
when compared to the structure of its nearest natural analog,
guanine. First, and most broadly, the 2- and 4-position
˚
instance, the reported interplanar spacing for thymine is 3.4 A
(Ozeki et al., 1969). However, a single tautomeric form is
adopted by this guanine structure. This is different from the
structure of P reported here. Unexpectedly, the heterocycle
separated from the ribose in the P nucleotide self-pairs to
form a purine–purine ‘fat’ pair. These are known in natural
DNA duplexes (Geyer et al., 2003) and do not excessively
destabilize the duplex thermodynamics. However, they
require hydrogen-bonding complementarity of the Watson–
Crick ‘edges’ to join the two purine-ring systems. This is
possible, for example, in the ‘fat’ pair between xanthine and
diaminopurine. Such hydrogen-bonding complementarity is
not possible in the self-pairing of P, at least not if P is drawn in
its conventional tautomeric form, as in Fig. 1.
To get hydrogen-bonding complementarity in this ‘fat’ pair
between two P heterocycles, P must adopt two tautomeric
forms. The second tautomer of P is, of course, not possible in
the P nucleoside (Seela et al., 2002; Jiang et al., 2010). This is
because ribose is attached to N8, which in turn requires (under
standard rules of valence) N3 to lack a proton. Therefore, the
P:Ptaut self-pair is not possible in duplex DNA or RNA.
Interestingly, this is the first case where this has been seen for a
purine analog, but is possible with other AEGIS purines. For
example, isoguanine also forms a ‘fat’ pair via a tautomer, but
isoguanosine does not seem to have been studied crystal-
lographically as an isolated species. However, isolated crystal
structures have been reported for some of the pyrimidine
analogs found in certain AEGIS implementations. Thus, one
single-component crystal of isocytosine also cocrystallizes as a
Figure 6
Tautomerism observed in imidazo[1,2-a][1,3,5]triazin-4(3H)-one (P).
IUPAC numbering has been used for the heterocycles. (Top) P in its
conventional tautomeric form can tautomerize to move a hydrogen from
atom N8 to N3, changing the pattern of hydrogen-bond donors/acceptors
from ‘acceptor–acceptor–donor’ (AAD) to ‘acceptor–donor–donor’
(ADD). (Bottom) The pair of tautomeric forms are hydrogen-bonding
complementary in the ‘reverse’ purine–purine ‘fat’ pair formed by
interactions between the two along their Watson–Crick edges.
ꢂ
26 Laos et al.
A component of an expanded DNA alphabet
Acta Cryst. (2019). C75, 22–28

research papers
Figure 8
(a) Comparison of the structures of P, guanine, and isoguanine, with their respective ring numbering. (b) Three cases of AEGIS nucleobases (P,
isoguanine, and isocytosine) that cannot pair with themselves (left) but can pair with their respective tautomers (right).
mixture of tautomers. This arrangement also allows it to solve
a crystal packing problem by creating a network of hydrogen-
bonding interactions across a plane of a crystal (McConnell et
al., 1964; Helene & Douzou, 1964; Sharma & McConnell,
1965). The surprising existence of both tautomers in this
structure suggest that these two need not be greatly different
in energy. This discovery illustrates the value of letting the
biomolecules ‘speak to us’ about their intrinsic preferences,
using small molecule crystallography to understand what they
are ‘saying’.
of the US government, nor NASA, nor the NSF, nor the
TWCF. CL thanks the Department of Chemistry at the
University of North Florida, the Research Corporation for
Science Advancement, the Dreyfus Foundation, and the
National Science Foundation.
Funding information
Funding for this research was provided by: Templeton World
Charity Foundation (grant No. 0092/AB57); John Templeton
Foundation (grant No. 54466); National Science Foundation
(grant No. CHE-1507816; grant No. DMR-1429428; grant No.
DMR-1626332); National Aeronautics and Space Adminis-
tration (award No. NNX15AF46G).
Acknowledgements
The Benner laboratory wishes to acknowledge support from
the Templeton World Charity Foundation, Inc. and the John
Templeton Foundation, the National Science Foundation and
NASA. The views and conclusions contained in this document
are those of the authors and should not be interpreted as
representing the official policies, either expressly or implied,
References
Benner, S. A., Karalkar, N. B., Hoshika, S., Laos, R., Shaw, R. W.,
Matsuura, M., Fajardo, D. & Moussatche, P. (2016). Cold Spring
Harb. Perspect. Biol. 8, 26.
ꢂ
Acta Cryst. (2019). C75, 22–28
Laos et al.
A component of an expanded DNA alphabet 27

research papers
Biondi, E., Lane, J. D., Das, D., Dasgupta, S., Piccirilli, J. A., Hoshika,
S., Bradley, K. M., Krantz, B. A. & Benner, S. A. (2016). Nucleic
Acids Res. 44, 9565–9577.
Brandenburg, K., Putz, H. & Berndt, M. (1999). DIAMOND. Crystal
Impact GbR, Bonn, Germany.
Ozeki, K., Sakabe, N. & Tanaka, J. (1969). Acta Cryst. B25, 1038–1045.
Seela, F., Rosemeyer, H., Melenewski, A., Heithoff, E.-M., Eickmeier,
H. & Reuter, H. (2002). Acta Cryst. C58, o142–o144.
Sefah, K., Yang, Z. Y., Bradley, K. M., Hoshika, S., Jimenez, E.,
Zhang, L. Q., Zhu, G. Z., Shanker, S., Yu, F. H., Turek, D., Tan,
W. H. & Benner, S. A. (2014). Proc. Natl Acad. Sci. USA, 111, 1449–
1454.
Bruker (2013). SAINT and SADABS. Bruker AXS Inc., Madison,
Wisconsin, USA.
Bruker (2014). APEX3. Bruker AXS Inc., Madison, Wisconsin, USA.
Georgiadis, M. M., Singh, I., Kellett, W. F., Hoshika, S., Benner, S. A.
& Richards, N. G. J. (2015). J. Am. Chem. Soc. 137, 6947–6955.
Geyer, R. C., Battersby, T. R. & Benner, S. A. (2003). Structure, 11,
1485–1498.
Guille, K. & Clegg, W. (2006). Acta Cryst. C62, o515–o517.
Helene, C. & Douzou, P. (1964). C. R. Hebd. Seances Acad. Sci. 259,
4387.
Sharma, B. D. & McConnell, J. F. (1965). Acta Cryst. 19, 797–806.
Sheldrick, G. M. (2008). Acta Cryst. A64, 112–122.
Sheldrick, G. M. (2015a). Acta Cryst. A71, 3–8.
Sheldrick, G. M. (2015b). Acta Cryst. C71, 3–8.
Singh, I., Laos, R., Hoshika, S., Benner, S. A. & Georgiadis, M. M.
(2018). Nucleic Acids Res. 46, 7977–7988.
Spackman, M. A. & Byrom, P. G. (1997). Chem. Phys. Lett. 267, 215–
220.
Hendrickson, C. L., Devine, K. G. & Benner, S. A. (2004). Nucleic
Acids Res. 32, 2241–2250.
Hirshfeld, F. L. (1977). Theor. Chim. Acta, 44, 129–138.
Thewalt, U., Bugg, C. E. & Marsh, R. E. (1971). Acta Cryst. B27,
2358–2363.
Turner, M. J., McKinnon, J. J., Wolff, S. K., Grimwood, D. J.,
Spackman, P. R., Jayatilaka, D. & Spackman, M. A. (2017).
CrystalExplorer17. University of Western Australia.
Yang, Z., Chen, F., Alvarado, J. B. & Benner, S. A. (2011). J. Am.
Chem. Soc. 133, 15105–15112.
¨
Hubschle, C. B., Sheldrick, G. M. & Dittrich, B. (2011). J. Appl. Cryst.
44, 1281–1284.
Jiang, D., Budow, S., Glac¸on, V., Eickmeier, H., Reuter, H., He, Y. &
Seela, F. (2010). Acta Cryst. C66, o194–o197.
Kim, S. H., Bartholomew, D. G., Allen, L. B., Robins, R. K., Revankar,
G. R. & Dea, P. (1978). J. Med. Chem. 21, 883–889.
Laos, R., Shaw, R., Leal, N. A., Gaucher, E. & Benner, S. (2013).
Biochemistry, 52, 5288–5294.
Leal, N. A., Kim, H. J., Hoshika, S., Kim, M. J., Carrigan, M. A. &
Benner, S. A. (2015). ACS Synth. Biol. 4, 407–413.
Matsuura, M. F., Kim, H.-J., Takahashi, D., Abboud, K. A. & Benner,
S. A. (2016). Acta Cryst. C72, 952–959.
Yang, Z., Chen, F., Chamberlin, S. G. & Benner, S. A. (2010). Angew.
Chem. Int. Ed. 49, 177–180.
Zhang, L., Yang, Z., Sefah, K., Bradley, K. M., Hoshika, S., Kim, M.-J.,
Kim, H.-J., Zhu, G., Jimenez, E., Cansiz, S., Teng, I. T.,
Champanhac, C., McLendon, C., Liu, C., Zhang, W., Gerloff,
D. L., Huang, Z., Tan, W. & Benner, S. A. (2015). J. Am. Chem. Soc.
137, 6734–6737.
Zhang, L. Q., Yang, Z. Y., Trinh, T. L., Teng, I. T., Wang, S., Bradley,
K. M., Hoshika, S., Wu, Q. F., Cansiz, S., Rowold, D. J., McLendon,
C., Kim, M. S., Wu, Y., Cui, C., Liu, Y., Hou, W. J., Stewart, K., Wan,
S., Liu, C., Benner, S. A. & Tan, W. H. (2016). Angew. Chem. Int.
Ed. 55, 12372–12375.
McConnell, J. F., Sharma, B. D. & Marsh, R. E. (1964). Nature, 203,
399–400.
McKinnon, J. J., Spackman, M. A. & Mitchell, A. S. (2004). Acta
Cryst. B60, 627–668.
ꢂ
28 Laos et al.
A component of an expanded DNA alphabet
Acta Cryst. (2019). C75, 22–28

supporting information
supporting information
Acta Cryst. (2019). C75, 22-28 [https://doi.org/10.1107/S2053229618016923]
The surprising pairing of 2-aminoimidazo[1,2-a][1,3,5]triazin-4-one, a
component of an expanded DNA alphabet
Roberto Laos, Christos Lampropoulos and Steven A. Benner
Computing details
Data collection: APEX2 (Bruker, 2014); cell refinement: SAINT (Bruker, 2013); program(s) used to solve structure:
SHELXT2014 (Sheldrick, 2015a); program(s) used to refine structure: SHELXL2017 (Sheldrick, 2015b) and ShelXle
(Hübschle et al., 2011); molecular graphics: XP in SHELXTL-Plus (Sheldrick, 2008); software used to prepare material
for publication: SHELXL2014 (Sheldrick, 2015b).
2-Aminoimidazo[1,2-a][1,3,5]triazin-4(3H)-one monohydrate
Crystal data
C5H5N5O·H2O
Mr = 169.16
F(000) = 704
Dx = 1.663 Mg m−3
Monoclinic, C2/c
a = 16.357 (13) Å
b = 3.712 (3) Å
c = 22.723 (19) Å
β = 101.72 (3)°
V = 1350.9 (19) Å3
Z = 8
Mo Kα radiation, λ = 0.71073 Å
Cell parameters from 9591 reflections
θ = 2.6–26.3°
µ = 0.13 mm−1
T = 100 K
Very thin needle, colourless
0.40 × 0.10 × 0.10 mm
Data collection
Bruker D8 VENTURE
diffractometer
Multilayer mirror monochromator
φ and ω scans
1397 independent reflections
1086 reflections with I > 2σ(I)
Rint = 0.089
θ
max = 26.4°, θmin = 2.5°
Absorption correction: multi-scan
(SADABS; Bruker, 2013)
h = −20→20
k = −4→4
l = −28→28
33557 measured reflections
Refinement
Refinement on F2
Least-squares matrix: full
R[F2 > 2σ(F2)] = 0.060
wR(F2) = 0.166
S = 1.16
1397 reflections
112 parameters
0 restraints
Hydrogen site location: mixed
H-atom parameters constrained
w = 1/[σ2(Fo2) + (0.0742P)2 + 4.2384P]
where P = (Fo2 + 2Fc2)/3
(Δ/σ)max < 0.001
Δρmax = 0.27 e Å−3
Δρmin = −0.56 e Å−3
Acta Cryst. (2019). C75, 22-28
sup-1

supporting information
Special details
Geometry. All esds (except the esd in the dihedral angle between two l.s. planes) are estimated using the full covariance
matrix. The cell esds are taken into account individually in the estimation of esds in distances, angles and torsion angles;
correlations between esds in cell parameters are only used when they are defined by crystal symmetry. An approximate
(isotropic) treatment of cell esds is used for estimating esds involving l.s. planes.
Refinement. The structure was solved and refined using the Bruker SHELXTL software package (Bruker, 2014; Hübschle
et al., 2011; Sheldrick, 2015b), in the space group C2/c, with Z = 8 for the formula unit, C5H7N5O2. The final anisotropic
full-matrix least-squares refinement on F2 with 116 variables converged at R1 = 7.14%, for the observed data and wR2 =
19.41% for all data. The goodness-of-fit was 1.089. The largest peak in the final difference electron density was 0.269 e
Å-3 and the largest hole was -0.562 e Å-3 with an r.m.s. deviation of 0.086 e Å-3. On the basis of the final model, the
calculated density was 1.663 Mg m-3 and F(000) = 704 e.
All H atoms bound to C atoms were positioned geometrically ( C—H = 0.93/1.00 Å) and allowed to ride with Uiso(H)=
1.2/1.5Ueq(C).
Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Å2)
x
y
z
Uiso*/Ueq
Occ. (<1)
O2
0.6958 (3)
0.730670
0.667110
0.7846 (9)
0.774086
0.752094
0.61373 (11)
0.64639 (13)
0.77095 (14)
0.780006
0.6401 (14)
0.854966
0.607746
0.953 (4)
1.213329
0.765008
0.3616 (5)
0.5857 (6)
0.8537 (6)
0.902557
0.918339
0.4099 (7)
0.2871 (6)
0.304919
0.1238 (8)
0.013066
0.1454 (8)
0.056609
0.6152 (6)
0.692018
0.4336 (7)
0.3267 (6)
0.6820 (7)
0.23204 (16)
0.242044
0.265584
0.2572 (6)
0.255700
0.232740
0.51107 (8)
0.34214 (9)
0.38555 (10)
0.349521
0.0541 (13)
0.081*
0.081*
0.0541 (13)
0.081*
0.081*
0.0169 (5)
0.0136 (5)
0.0156 (6)
0.019*
0.75
0.75
0.75
0.25
0.25
0.25
H2Y
H2X
O2′
H2W
H2Z
O1
N1
N2
H2A
H2B
C9
N8
H8
C7
H7
C6
H6
N3
H3
0.807740
0.417693
0.019*
0.57886 (16)
0.51426 (13)
0.506724
0.46028 (17)
0.408713
0.49153 (16)
0.467115
0.69230 (13)
0.730825
0.35240 (11)
0.31270 (10)
0.273345
0.34504 (12)
0.327525
0.40446 (12)
0.436123
0.44912 (9)
0.479352
0.0123 (6)
0.0152 (5)
0.018*
0.0164 (6)
0.020*
0.0156 (6)
0.019*
0.0140 (5)
0.017*
0.5
0.5
C4
N5
C2
0.62514 (16)
0.56777 (13)
0.70163 (16)
0.46086 (12)
0.40962 (9)
0.39144 (11)
0.0132 (6)
0.0127 (5)
0.0124 (6)
Atomic displacement parameters (Å2)
U11
U22
U33
U12
U13
U23
O2
O2′
O1
N1
N2
0.050 (2)
0.050 (2)
0.0200 (10)
0.0149 (11)
0.0163 (12)
0.087 (4)
0.087 (4)
0.0225 (11)
0.0170 (12)
0.0230 (13)
0.0280 (18)
0.019 (2)
0.019 (2)
−0.0018 (8)
0.0019 (9)
−0.0035 (10)
0.0143 (17)
0.0143 (17)
0.0047 (7)
0.0007 (9)
0.0025 (9)
0.0103 (19)
0.0103 (19)
0.0009 (8)
0.0000 (9)
0.0007 (9)
0.0280 (18)
0.0088 (9)
0.0082 (11)
0.0076 (11)
Acta Cryst. (2019). C75, 22-28
sup-2

supporting information
C9
N8
C7
C6
N3
C4
N5
C2
0.0143 (13)
0.0129 (11)
0.0143 (13)
0.0148 (13)
0.0136 (11)
0.0146 (13)
0.0129 (11)
0.0144 (13)
0.0118 (13)
0.0185 (12)
0.0166 (14)
0.0163 (14)
0.0183 (12)
0.0137 (13)
0.0153 (12)
0.0126 (13)
0.0107 (13)
0.0135 (11)
0.0178 (14)
0.0170 (14)
0.0096 (11)
0.0108 (13)
0.0096 (11)
0.0104 (13)
0.0045 (10)
0.0021 (9)
0.0019 (11)
0.0008 (11)
0.0010 (9)
0.0046 (11)
0.0024 (9)
0.0046 (11)
0.0021 (10)
0.0006 (9)
0.0020 (11)
0.0063 (11)
0.0014 (9)
0.0015 (10)
0.0016 (9)
0.0029 (10)
−0.0002 (10)
−0.0022 (9)
−0.0023 (11)
−0.0001 (11)
0.0005 (9)
−0.0019 (10)
0.0009 (9)
0.0012 (10)
Geometric parameters (Å, º)
O2—H2Y
O2—H2X
O2′—H2W
O2′—H2Z
O1—C4
N1—C2
N1—C9
N2—C2
0.9800
0.9802
0.9800
0.9800
1.223 (3)
1.337 (3)
1.343 (4)
1.331 (4)
0.8800
C9—N5
1.384 (3)
1.397 (4)
0.8800
1.345 (4)
0.9500
1.401 (4)
0.9500
1.361 (4)
1.373 (3)
0.8800
N8—C7
N8—H8
C7—C6
C7—H7
C6—N5
C6—H6
N3—C4
N3—C2
N3—H3
C4—N5
N2—H2A
N2—H2B
C9—N8
0.8800
1.323 (4)
1.396 (3)
H2Y—O2—H2X
H2W—O2′—H2Z
C2—N1—C9
C2—N2—H2A
C2—N2—H2B
H2A—N2—H2B
N8—C9—N1
N8—C9—N5
N1—C9—N5
C9—N8—C7
C9—N8—H8
C7—N8—H8
C6—C7—N8
C6—C7—H7
N8—C7—H7
105.5
127.7
115.1 (2)
120.0
120.0
C7—C6—N5
C7—C6—H6
N5—C6—H6
C4—N3—C2
C4—N3—H3
C2—N3—H3
O1—C4—N3
O1—C4—N5
N3—C4—N5
C9—N5—C4
C9—N5—C6
C4—N5—C6
N2—C2—N1
N2—C2—N3
N1—C2—N3
105.2 (2)
127.4
127.4
121.8 (2)
119.1
119.1
125.0 (2)
120.8 (2)
114.2 (2)
121.7 (2)
108.2 (2)
130.0 (2)
119.2 (2)
116.4 (2)
124.4 (2)
120.0
128.3 (2)
108.9 (2)
122.7 (2)
107.1 (2)
126.5
126.5
110.5 (2)
124.7
124.7
C2—N1—C9—N8
C2—N1—C9—N5
N1—C9—N8—C7
N5—C9—N8—C7
C9—N8—C7—C6
N8—C7—C6—N5
C2—N3—C4—O1
C2—N3—C4—N5
N8—C9—N5—C4
179.2 (3)
−1.4 (4)
179.9 (3)
0.4 (3)
−0.1 (3)
−0.2 (3)
179.0 (2)
−0.4 (4)
N1—C9—N5—C6
O1—C4—N5—C9
N3—C4—N5—C9
O1—C4—N5—C6
N3—C4—N5—C6
C7—C6—N5—C9
C7—C6—N5—C4
C9—N1—C2—N2
C9—N1—C2—N3
179.9 (2)
179.0 (2)
−1.5 (4)
2.3 (4)
−178.2 (2)
0.5 (3)
177.5 (3)
179.2 (2)
−0.6 (4)
−177.9 (2)
Acta Cryst. (2019). C75, 22-28
sup-3

supporting information
N1—C9—N5—C4
N8—C9—N5—C6
2.5 (4)
−0.6 (3)
C4—N3—C2—N2
C4—N3—C2—N1
−178.3 (2)
1.6 (4)
Hydrogen-bond geometry (Å, º)
D—H···A
D—H
H···A
D···A
D—H···A
O2—H2Y···O2i
O2—H2X···N1
O2′—H2W···O2′i
O2′—H2Z···O2′ii
N3—H3···N3iii
N2—H2A···O2i
N2—H2A···O2′
N2—H2B···O1iii
N8—H8···N8iv
C7—H7···O2v
C6—H6···O1vi
0.98
0.98
0.98
0.98
0.88
0.88
0.88
0.88
0.88
0.95
0.95
1.62
1.84
1.30
1.34
1.98
2.16
2.12
2.02
1.92
2.39
2.50
2.582 (6)
2.788 (5)
2.164 (15)
2.164 (15)
2.853 (5)
3.026 (5)
2.993 (12)
2.897 (3)
2.792 (5)
3.315 (5)
3.396 (4)
166
162
143
137
174
168
170
172
173
163
158
Symmetry codes: (i) −x+3/2, y+1/2, −z+1/2; (ii) −x+3/2, y−1/2, −z+1/2; (iii) −x+3/2, −y+3/2, −z+1; (iv) −x+1, y, −z+1/2; (v) −x+1, y−1, −z+1/2; (vi) −x+1,
−y, −z+1.
Acta Cryst. (2019). C75, 22-28
sup-4