Identification and Characterization of Enzymes Catalyzing Pyrazolopyrimidine Formation in the Biosynthesis of Formycin A

Article abstract of DOI:10.1021/acs.orglett.7b00355

Genome scanning of Streptomyces kaniharaensis, the producer of formycin A, reveals two sets of purA, purB, purC, and purH genes. The Pur enzymes catalyze pyrimidine assembly of purine nucleobases. To test whether enzymes encoded by the second set of pur genes catalyze analogous transformations in formycin biosynthesis, formycin B 5′-phosphate was synthesized and shown to be converted by ForA and ForB to formycin A 5′-phosphate. These results support that For enzymes are responsible for formycin formation.

Full text of DOI:10.1021/acs.orglett.7b00355

Letter
pubs.acs.org/OrgLett
Identification and Characterization of Enzymes Catalyzing
Pyrazolopyrimidine Formation in the Biosynthesis of Formycin A
Yeonjin Ko,†,§ Shao-An Wang,†,§ Yasushi Ogasawara,‡,# Mark W. Ruszczycky,‡ and Hung-wen Liu*,†,‡
†Department of Chemistry, University of Texas at Austin, Austin, Texas 78712, United States
‡Division of Chemical Biology and Medicinal Chemistry, College of Pharmacy, University of Texas at Austin, Austin, Texas 78712,
United States
S
* Supporting Information
ABSTRACT: Genome scanning of Streptomyces kaniharaensis, the producer of formycin A, reveals two sets of purA, purB, purC,
and purH genes. The Pur enzymes catalyze pyrimidine assembly of purine nucleobases. To test whether enzymes encoded by the
second set of pur genes catalyze analogous transformations in formycin biosynthesis, formycin B 5′-phosphate was synthesized
and shown to be converted by ForA and ForB to formycin A 5′-phosphate. These results support that For enzymes are
responsible for formycin formation.
ormycin A (i.e., 8-aza-9-deazaadenosine (1)) is an
1
Fantibiotic isolated from Nocardia interforma and Strepto-
myces kaniharaensis SF-557.2 It has been reported to
demonstrate antiproliferative properties with respect to Ehrlich
carcinoma, mouse leukemia L-1210, Yoshida sarcoma, and
HeLa cells.3 It also has antiviral activity against influenza virus
A14 and human immunodeficiency virus type 1.5 Being a C-
nucleoside analogue of adenosine (2), formycin A has been
shown to be a potent inhibitor of various enzymes that accept
adenosine as a substrate such as adenosine kinase from
Mycobacterium tuberculosis, which is involved in the purine
salvage pathway.6 Likewise, formycin A and its analogues are
also known to inhibit bacterial purine nucleoside phosphor-
Figure 1. Biosynthetic origins of formycin A and adenosine.
ylase, which has pharmaceutical significance.7
While the biological activities of formycin A and its
derivatives have been well documented, how they are
biosynthesized in nature has remained obscure. Early in vitro
studies revealed that formycin B 5′-phosphate (12b) is a likely
precursor to formycin A (1).8 Feeding experiments in N.
interforma showed that the ribosyl moiety of formycin A is
derived directly from ribose likely via an intermediary produced
from phosphoribosyl pyrophosphate (PRPP, 3).9 However, the
N3 and N8 nitrogen centers of formycin A originate from the
ε-amino group of L-lysine (5).10 Importantly, L-glutamate (4)
was found to be asymmetrically incorporated into formycin A
such that the C6, C5, C4, and C9 carbons of formycin A are
derived from the C1, C2, C3, and C4 carbons of glutamate (see
Figure 1).10b,11a Similar incorporation patterns were also
observed in the formation of other C-nucleosides, such as
pyrazomycin, showdomycin, and minimycin,11 making L-
glutamate a common precursor of this class of nucleoside
antibiotics. These incorporation patterns differ from those
observed for adenosine (2), where the C4, C5, and N7 are from
glycine (6), N3 and N9 are from glutamine (7), and C6 is from
bicarbonate, while C2 and C8 are from N10-formyl-tetrahy-
drofolate (N10-formyl-THF, see Figure 1).12 Thus, despite the
marked structural resemblance between formycin A (1) and
adenosine (2), they are expected to be biosynthesized via
distinct pathways. Although early isotope tracer experiments
have shed some light on the possible precursors of formycin A,
the link between these precursors and the final product remains
a mystery.
To investigate the biosynthesis of formycin A, the genome of
S. kaniharaensis was sequenced to generate 3011 contiguous
Received: February 3, 2017
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.7b00355
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
sequences (contigs) with an average size of 2641 bp. Although
the relatively short size of the contigs hampered gene assembly,
preliminary genomic analysis revealed the presence of two sets
of purA, purB, purC, and purH genes. The pur genes are known
to encode enzymes catalyzing the transformation of carbox-
yaminoimidazole ribonucleotide (CAIR) to adenosine 5′-
monophosphate (8a → 14a) during the biosynthesis of purine
nucleobases (see Scheme 1).13 Given the high sequence
forA (purA homologue) and forB (purB homologue) are
proposed to be responsible for its conversion to formycin A 5′-
phosphate (12b → 13b → 14b).
To test this hypothesis, formycin B 5′-phosphate (12b), the
putative substrate for ForA, was synthesized according to
reported procedures with some modifications as shown in
Scheme 2 and described in the Supporting Information.14−20
Scheme 2. Chemical Synthesis of Formycin B 5′-Phosphate
Scheme 1. Later Steps of the Established Biosynthetic
Pathway for Adenosine (2) and the Proposed Analogous
Pathway for Formycin A (1) Formation
(12b)
Meanwhile, the forA gene (GenBank accession no. KY629733)
was heterologously overexpressed and isolated as an N-terminal
His6-tagged protein in Escherichia coli. When 1 mM 12b was
incubated with 10 μM ForA in the presence of 2.5 mM L-
aspartate and 0.125 mM guanosine triphosphate (GTP) in 25
mM potassium phosphate buffer (pH 8) containing 16 mM
magnesium acetate, 1 mM phosphoenolpyruvate (PEP), and 2
units pyruvate kinase at room temperature, the consumption of
12b could be observed by anion exchange HPLC (see Figure
2). Pyruvate kinase and PEP are used to regenerate GTP from
GDP in this assay. While no new peak was discernible by
HPLC likely due to prolonged retention on the anion exchange
HPLC column, analysis by reversed phase HPLC did reveal a
new peak eluting near the void volume. High-resolution LC−
MS analysis of the reaction mixture demonstrated a signal at
462.0658 m/z (negative ion mode), which is consistent with
the mass calculated for the [M − H]− ion of 13b (462.0668
Da). Steady state kinetic parameters for ForA were determined
by coupling pyruvate formation to oxidation of NADH in the
presence of lactate dehydrogenase, which can be monitored by
UV absorbance. At 2.5 mM L-aspartate and 0.125 mM GTP,
the kcat for saturating 12b is 18 2 min−1, and the KM for 12b
is 74 25 μM. These are comparable to the reported kinetic
constants of E. coli PurA (kcat of 82 7 min−1 and KM for 12a
of 21 2 μM at 5 mM L-aspartate and 0.2 mM GTP).21
The forB gene (GenBank accession no. KY629734) was also
heterologously expressed in E. coli and purified as an N-His6-
tagged protein in order to further investigate the reactions
catalyzed by the for gene cluster. When both ForA and ForB
were incubated with 12b, a new peak appeared with a retention
similarity (ca. 90%) among the conserved adenosine bio-
synthetic genes (i.e., the pur genes) present in all Streptomyces
species, the genuine pur genes responsible for purine nucleotide
biosynthesis in S. kaniharaensis could be readily identified. In
contrast, the second set of pur-like genes other than the PurB
homologue exhibit only ca. 50% sequence similarity to the
genuine pur genes found in S. kaniharaensis (see Table S1).
This distinction suggests that the second set of pur genes,
which will be referred to as the for genes, are more likely
involved in a secondary rather than a primary metabolic
pathway.
Even though formycin A (1) contains a pyrazolopyrimidine
nucleobase instead of the imidazolopyrimidine base observed in
adenosine (2), the presence of a seemingly second copy of pur
genes in S. kaniharaensis strongly hinted that construction of
the nucleobases in both compounds may follow similar
biosynthetic routes. Namely, the for genes in S. kaniharaensis
may be involved in formycin A biosynthesis. In particular, it was
hypothesized that ForA, -B, -C, and -H are responsible for
catalyzing the transformation of 8b to formycin A (1) as shown
in Scheme 1. Formycin B 5′-phosphate (12b) represents a late
intermediate in this putative pathway, and the gene products of
B
DOI: 10.1021/acs.orglett.7b00355
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
logously overexpressed and purified as a C-terminal His6-tagged
protein in E. coli. Despite the sequence similarity of these
enzymes (41% identity, 58% similarity) and the structural
resemblance of the ForA substrate 12b to the PurA substrate
inosine 5′-monophosphate (12a, IMP), no turnover was
observed upon incubation of 12b with PurA (Figure S2).
This result implies that S. kaniharaensis cannot produce
formycin A using the Pur enzymes alone, and therefore, a
second biosynthetic gene cluster is indeed necessary.
In contrast to PurA, ForA can accept both 12b and 12a as
substrates though 12b is preferred. The relative specificity of
ForA for the two substrates can be assessed quantitatively from
the ratio of the corresponding specificity constants (i.e., kcat/
KM) determined from a competition experiment.22 This was
performed by adding ForA to a mixture of approximately 0.15
mM 12b and a 3−30-fold excess of 12a in the presence of 5
mM L-aspartate, 0.5 mM GTP, 0.2 mM 5-aminoimidazole-4-
carboxamide 1-β-D-ribofuranoside (AICAR, internal standard),
1 mM PEP, 1 mM MgCl2, and 1 unit pyruvate kinase in 20 mM
HEPES buffer at pH 7.7. As shown in Figure 3, the HPLC
Figure 2. In vitro assay of ForA and ForB. (A) Reaction mixture
containing formycin B 5′-phosphate (12b) without enzymes; (B) 12b
+ ForA; (C) 12b + ForA + ForB. Reaction mixtures for traces (A) to
(C) were analyzed after a 1.5 h incubation by HPLC using a Dionex
CarboPac PA1 anion exchange column in a two-solvent system of
H2O versus 1.5 M ammonium acetate in H2O (pH 7.0). (D) Isolated
ForA/ForB product; (E) isolated ForA/ForB product + CIP; (F)
coinjection of isolated ForA/ForB/CIP product and formycin A (1)
standard. HPLC traces (D) to (F) were obtained using a Microsorb-
MV 100-5 C18 column in a two-solvent system of 1% ammonium
acetate in H2O (pH 5.1) versus acetonitrile (see Supporting
Information and text for details).
Figure 3. Competition assay of ForA with 12a and 12b. (A) Sample
HPLC traces at three time points from a single trial showing relative
consumption of 12b versus 12a. HPLC traces utilized a Microsorb-
MV 100-5 C18 column with a two-solvent system of 1% ammonium
acetate in H2O (pH 5.1) versus acetonitrile. (B) Plots of log([12a]/bi)
versus log([12b]/a) for the five separate trials where bi and a are
scaling factors relating the concentrations to the standardized HPLC
peak integrations and shift the intercepts to facilitate visualization (see
text and Supporting Information for details).
time of 14 min by anion exchange HPLC (see Figure 2, trace
C). High-resolution ESI−MS analysis (negative ion mode) of
the isolated peak demonstrated a signal at 346.0556 m/z, which
matches the predicted monoisotopic mass of the [M − H]− ion
of formycin A 5′-phosphate (14b) (346.0558 Da). Treatment
of the ForA/ForB reaction product with calf intestinal alkaline
phosphatase (CIP) resulted in a new species observed by
reversed phase HPLC (Figure 2, trace E) that coeluted with a
formycin A standard (Figure 2, trace F). Isolation and MS
analysis of this reaction product revealed a signal at 266.0893
m/z (negative ion mode) consistent with the [M − H]− ion of
formycin A (266.0895 Da). These observations demonstrate
that ForA and ForB catalyze the introduction of the C6 amino
group to generate formycin A 5′-phosphate (14b) from
formycin B 5′-phosphate (12b) with L-aspartate serving as
the nitrogen donor. These activities are analogous to those of
PurA and PurB during the biosynthesis of adenosine (see
Scheme 1).
peaks for 12a and 12b from reaction aliquots at some time t
could be separately integrated and standardized based on their
extinction coefficients at the detector wavelength (278 nm) and
the AICAR peak integration to yield the observables y(t) and
x(t), respectively. These observables are related according to
log y(t) = β2 log x(t) + β1
(1)
where β2 is the reciprocal of the desired specificity constant
ratio and β1 is a nuisance parameter. Approximately seven time
points were taken per experiment, and the entire competition
experiment was repeated 5 times. The parameter β2 was then
estimated from the five data sets using a linear regression model
developed from eq 1 indicating that ForA is approximately 8-
fold (95% conf. int.: 6.9−9.2) more specific for 12b than 12a
(see Supporting Information for details). This supports the
hypothesis that the for gene cluster is more than a redundant
copy of the pur cluster and is instead specific for formycin A
biosynthesis.
Complementary to the genomic analyses, these observations
provide experimental evidence that ForA and ForB represent
components of the formycin A biosynthetic pathway. In order
to establish that the for and pur gene clusters encode enzymes
for two distinct biosynthetic pathways, we sought to determine
the substrate specificity of the ForA and PurA enzymes. The
purA gene (GenBank accession no. KY629735) was hetero-
C
DOI: 10.1021/acs.orglett.7b00355
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
(3) (a) Ishizuka, M.; Takeuchi, T.; Nitta, K.; Koyama, G.; Hori, M.;
Umezawa, H. J. Antibiotics. Ser. A 1964, 17, 124−126. (b) Ishizuka, M.;
Sawa, T.; Hori, S.; Takayama, H.; Takeuchi, T.; Umezawa, H. J.
Antibiot. 1968, 21, 5−12.
In conclusion, the discovery of two sets of homologous genes
(i.e., pur and for) in the genome of S. kaniharaensis led to the
hypothesis that the pyrazolopyrimidine of formycin A is
biosynthesized in a manner analogous to the formation of the
imidazolopyrimidine moiety in adenosine. This hypothesis was
supported by in vitro demonstration of the catalytic activities of
two enzymes (ForA and ForB) encoded by the for genes.
Whereas enzymes of the pur gene cluster are specific for
pyrimidine nucleoside biosynthesis, the for-encoded enzymes
are selective for pyrazolopyrimidine nucleoside biosynthesis.
This implies that the two sets of enzymes are indeed
responsible for the production of different metabolites with
the for genes encoding enzymes for the biosynthesis of
formycin A. Furthermore, ForA together with ForB catalyze
the conversion of formycin B 5′-phosphate (12b) to formycin
A 5′-phosphate (14b) analogous to the reactions catalyzed by
PurA and PurB during the biosynthesis of adenosine. This
study shows how sequence information garnered by genome
scanning can be directly translated into the discovery of an
uncharted biosynthetic pathway. Further investigation of
formycin biosynthesis is being pursued with an emphasis on
identifying the full gene cluster and reconstructing the
biosynthetic pathway in vitro. Detailed knowledge of the
biosynthesis of C-nucleoside antibiotics is essential for the
repurposing and modification of these pathways to produce
new antibiotics with improved biomedical properties.
(4) Takeuchi, T.; Iwanaga, J.; Aoyagi, T.; Umezawa, H. J. Antibiotics.
Ser. A 1966, 19, 286−287.
(5) Dapp, M. J.; Bonnac, L.; Patterson, S. E.; Mansky, L. M. J. Virol.
2014, 88, 354−363.
(6) Long, M. C.; Parker, W. B. Biochem. Pharmacol. 2006, 71, 1671−
1682.
(7) (a) Bzowska, A.; Kulikowska, E.; Shugar, D. Biochim. Biophys.
Acta, Protein Struct. Mol. Enzymol. 1992, 1120, 239−247. (b) Kierdas-
zuk, B.; Modrak-Wojcik, A.; Wierzchowski, J.; Shugar, D. Biochim.
́
Biophys. Acta, Protein Struct. Mol. Enzymol. 2000, 1476, 109−128.
(8) Sawa, T.; Fukagawa, Y.; Homma, I.; Wakashiro, T.; Takeuchi, T.;
Hori, M.; Komai, T. J. Antibiot. 1968, 21, 334−339.
(9) Kunimoto, T.; Sawa, T.; Wakashiro, T.; Hori, M.; Umezawa, H. J.
Antibiot. 1971, 24, 253−258.
(10) (a) Ochi, K.; Iwamoto, S.; Hayase, E.; Yashima, S.; Okami, Y. J.
Antibiot. 1974, 27, 909−916. (b) Ochi, K.; Yashima, S.; Eguchi, Y.;
Matsushita, K. J. Biol. Chem. 1979, 254, 8819−8824.
(11) (a) Buchanan, J. G.; Hamblin, M. R.; Sood, G. R.; Wightman, R.
H. J. Chem. Soc., Chem. Commun. 1980, 917−918. (b) Elstner, E. F.;
Suhadolnik, R. J. Biochemistry 1972, 11, 2578−2584. (c) Isono, K.;
Suhadolnik, R. J. J. Antibiot. 1977, 30, 272−273.
(12) Buchanan, J. M.; Hartman, S. C. Adv. Enzymol. Relat. Areas Mol.
Biol. 2006, 21, 199−261.
(13) (a) Zalkin, H.; Dixon, J. E. Prog. Nucleic Acid Res. Mol. Biol.
1992, 42, 259−287. (b) Zhang, Y.; Morar, M.; Ealick, S. E. Cell. Mol.
Life Sci. 2008, 65, 3699−3724.
ASSOCIATED CONTENT
* Supporting Information
(14) (a) Xu, G.; Moeller, K. D. Org. Lett. 2010, 12, 2590−2593.
(b) Zeng, J.; Vedachalam, S.; Xiang, S.; Liu, X.-w. Org. Lett. 2011, 13,
42−45.
■
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.7b00355.
(15) Le Coq, A.; Gorgues, A. Org. Synth. 1979, 59, 10.
(16) Buchanan, J. G.; Edgar, A. R.; Hutchison, R. J.; Stobie, A.;
Wightman, R. H. J. Chem. Soc., Perkin Trans. 1 1980, 2567−2571.
(17) Harusawa, S.; Matsuda, C.; Araki, L.; Kurihara, T. Synthesis
2006, 2006, 793−798.
Experimental details including synthesis of compounds,
genomic analysis, gene cloning and expression, protein
isolation, and analytical methodologies (PDF)
(18) Zhou, J.; Yang, M.; Akdag, A.; Schneller, S. W. Tetrahedron
2006, 62, 7009−7013.
(19) Buchanan, J. G.; Stobie, A.; Wightman, R. H. J. Chem. Soc.,
Perkin Trans. 1 1981, 2374−2378.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: h.w.liu@mail.utexas.edu.
ORCID
(20) Buchanan, J. G.; Stobie, A.; Wightman, R. H. Can. J. Chem.
1980, 58, 2624−2627.
(21) Kang, C.; Sun, N.; Poland, B. W.; Gorrell, A.; Honzatko, R. B.;
Fromm, H. J. J. Biol. Chem. 1997, 272, 11881−11885.
(22) Cornish-Bowden, A. Fundamentals of Enzyme Kinetics, 4th ed.;
Wiley-Blackwell: Weinheim, Germany, 2012.
Hung-wen Liu: 0000-0001-8953-4794
Present Address
#Laboratory of Applied Biochemistry, Graduate School of
Engineering, Hokkaido University, Sapporo, Hokkaido 060-
8628, Japan.
Author Contributions
§Y.K. and S.-A.W. contributed equally to this work.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported in part by grants from the National
Institutes of Health (GM035906 and GM040541) and the
Welch Foundation (F-1511).
REFERENCES
■
(1) Hori, M.; Ito, E.; Takita, T.; Koyama, G.; Takeuchi, T.; Umezawa,
H. J. Antibiotics. Ser. A 1964, 17, 96−99.
(2) Suhadolnik, R. J. Nucleoside Antibiotics; Wiley-Interscience: New
York, 1970.
D
DOI: 10.1021/acs.orglett.7b00355
Org. Lett. XXXX, XXX, XXX−XXX