The revised structure of the cyclic octapeptide surugamide A

Article abstract of DOI:10.1248/cpb.c19-00002

Surugamides are a group of non-ribosomal peptides isolated from marine-derived Streptomyces. Surugamide A (1) and its closely related derivatives, surugamides B–E (2–5), are D-amino acid containing cyclic octapeptides with cathepsin B inhibitory activity.

Full text of DOI:10.1248/cpb.c19-00002

476
Chem. Pharm. Bull. 67, 476–480 (2019)
Vol. 67, No. 5
Regular Article
Highlighted Paper selected by Editor-in-Chief
The Revised Structure of the Cyclic Octapeptide Surugamide A
,a
Kenichi Matsuda,a Takefumi Kuranaga,* Ayae Sano,a Akihiro Ninomiya,b Kentaro Takada,b and
,a
Toshiyuki Wakimoto*
a Faculty of Pharmaceutical Sciences, Hokkaido University; Sapporo 060–0812, Japan: and b Graduate School of
Agricultural and Life Sciences, The University of Tokyo; Tokyo 113–8657, Japan.
Received January 4, 2019; accepted January 25, 2019
Surugamides are a group of non-ribosomal peptides isolated from marine-derived Streptomyces. Suru-
gamide A (1) and its closely related derivatives, surugamides B–E (2–5), are D-amino acid containing cyclic
octapeptides with cathepsin B inhibitory activity. The D-isoleucine (Ile), the nonproteinogenic amino acid
residue embedded in 1, is less common in natural peptides because a rare Cβ-epimerization is required for
its biosynthesis. Taking advantage of the synthetic route of 2 previously established by our group, we synthe-
sized the cyclic octapeptide 1 containing D-Ile by solid phase peptide synthesis. The structure of 1 actually
contains D-allo-Ile in place of D-Ile, which was corroborated by chemical syntheses and chromatographic
comparisons.
Key words marine natural product; structural revision; non-ribosomal peptide; D-amino acid
assembly lines lack the canonical thioesterase (TEs) domains
Introduction
Surugamides, two groups of structurally unrelated non-ribo- that generally catalyze the offloading and macrocyclization of
somal peptides (NRP), are produced by marine actinomycetes, the grown peptides.7) Instead, SurE, a penicillin-binding pro-
Streptomyces sp. JAMM992 and several other Streptomyces tein that is physically discrete from the mega-synthetase, acts
strains.1–4) Surugamide A (Fig. 1, 1), a cyclic octapeptide in trans with both assembly lines to release the cyclooctapep-
with cathepsin B inhibitory activity was identified in 2013, tides (1–5) and the cyclodecapeptide (6), respectively, high-
along with the structurally related derivatives surugamides lighting their unique biosynthetic systems4) (Fig. 2).
B–E1) (Fig. 1, 2–5). Its biosynthetic gene cluster consists of
Another feature of the octapeptidic surugamides is the D-
four NRPS genes, surA–D5) (Fig. 2). The two NRPS genes, isoleucine (Ile) residue present at the position 4 in 1 and 3–5
surA/surD, which are responsible for the synthesis of the cy- (Fig. 1, 2 is a derivative with D-Val at the position 4). D-Ile is
clic octapeptides are separated by two additional NRPS genes, an enantiomer of L-Ile with the epimerization at the two ste-
surB/surC, responsible for the synthesis of cyclosurugamide reogenic centers at the Cα-, and Cβ-positions. The epimeriza-
F (Fig. 1, 6), a structurally unrelated cyclic decapeptide, and tion at the Cα-position commonly observed in NRP is general-
its linear derivative, surugamide F5,6) (Fig. 1, 7). Both peptide ly proceeds via two ways: the direct activation of the D-amino
Fig. 1. Structures of Surugamides
1a (proposed structure): X=Me, Y=H, 1b (revised structure): X=H, Y=Me “p” in structure of surugamide A stands for “position.”
*To whom correspondence should be addressed. e-mail: kuranaga@pharm.hokudai.ac.jp; wakimoto@pharm.hokudai.ac.jp
© 2019 The Pharmaceutical Society of Japan

Vol. 67, No. 5 (2019)
Chem. Pharm. Bull.
477
Fig. 2. Biosynthesis of Surugamides
The NRPS genes surA/surD for syntheses of 1–5, surB/surC for syntheses of 6, 7, and the trans-acting thioesterase SurE are colored in yellow, gray, and orange, respec-
tively. The colored balls represent functional domains: blue, condensation domain; red, adenylation domain; gray, peptidyl-carrier protein; green, epimerization domain;
orange, trans-acting thioesterase SurE. (Color figure can be accessed in the online version.)
Fig. 3. LC-MS Comparison of the Natural Surugamide A (1), and the Synthetic 1a and 1b
Extracted ion chromatographs (EIC) for m/z 912.7 are depicted.
acid by the D-amino acid-specific adenylation (A) domain, or we first synthesized D-Ile-containing 1 by taking advantage of
the equilibration of the Cα configuration of the PCP-tethered the established route for the total synthesis of 2.4) This led us
biosynthetic intermediate catalyzed by the epimerization (E) to realize that 1 actually does not possess D-Ile; therefore, the
domain.8) In the case of surugamide biosynthesis, whereas reported structures of the cyclic octapeptide surugamides 1,
the epimerization of the Cα-position of the Ile residue at the 3–5 need to be corrected.
position 4 (Ile-4) should be facilitated by the E domain in
To confirm the structure of 1, we conducted the solid phase
module 4 of SurA, the biosynthetic mechanism for the epimer- peptide synthesis of 1 from 9-fluorenylmethyloxycarbonyl
ization at its Cβ-position is obscure. The epimerization at the (Fmoc)-D-Leu loaded on 2-chlorotrityl resin (8). The Fmoc
Cβ-position is relatively rare in nature and, to the best of our group was deprotected by a treatment with piperidine to af-
knowledge, only one type of biosynthetic machinery has been ford 9, and then three rounds of N,N′-diisopropylcarbodiimide
identified for L-allo-Ile biosynthesis in the desotamides9) and (DIC)/Oxyma-mediated amide coupling and Nα-deprotection
marformycin10) pathways. In the biosynthesis of L-allo-Ile, with piperidine yielded the resin-bound tetrapeptide 10. The
pyridoxal 5′-phosphate (PLP)-dependent aminotransferases amine 11 was conjugated to Fmoc-D-Ile-OH to afford 12a, and
(DsaD and MfnO in the biosyntheses of desotamide and mar- three further rounds of DIC/Oxyma-mediated amide coupling
formycin, respectively) and the cognate isomerases (DsaE and were performed to generate the resin-bound octapeptide (13a).
MfnH in the biosyntheses of desotamide and marformycin, Successive treatment of 13a with (CF3)2CHOH cleaved the
respectively) cooperate to convert L-Ile to L-allo-Ile.11) How- peptide from the resin to give 14a, which was subsequently
ever, a pair of homologous enzymes is not encoded in the cyclized by using PyBOP and HOAt to produce the cyclic
biosynthetic gene cluster of the surugamides. In this study, peptide 15a. Finally, 15a was treated with trifluoroacetic acid

478
Chem. Pharm. Bull.
Vol. 67, No. 5 (2019)
Fig. 4. Marfey’s Analysis of the Natural Surugamide A (1), the Synthetic 1a, 1b, and Standard Amino Acids
EICs for m/z 384.15 corresponding to derivatized DL-Leu, DL-Ile, and DL-allo-Ile are depicted.
The gray ball represents 2-chlorotrityl resin.
Chart 1. Synthetic Scheme of Surugamide A
(TFA)–i-Pr3SiH–H2O (95:2.5:2.5) to remove the tert-butoxy-
a D-Ile residue, since the derivatized D-Ile and D-Leu were
carbonyl (Boc) group, which yielded the originally reported coeluted at almost the same retention time (Fig. 4). However,
structure of surugamide A (1a). However, the HPLC analysis the data clearly showed the presence of a D-allo-Ile residue in
revealed that the synthetic 1a was not identical to the natural the natural 1 (Fig. 4). According to the domain architecture of
surugamide A (Fig. 3). This result prompted us to compare NRPS for surugamide A biosynthesis, the Cα-epimerization of
the synthetic and natural 1 by the C3 Marfey’s method.12–14) Ile occurs only at the position 4, suggesting that 1 possesses
The result did not exclude the possibility of the presence of D-allo-Ile residue instead of D-Ile in the position 4. To confirm

Vol. 67, No. 5 (2019)
Chem. Pharm. Bull.
479
this, 1 with the substitution of D-Ile to D-allo-Ile (1b) was syn- was stirred for 20min, and then filtered. This procedure was
thesized in the same manner as 1a, as shown in Chart 1. As repeated twice. The filtrate was azeotropically dried with
expected, the HPLC analysis revealed that 1b and the natural toluene (×3) to afford a crude mixture of peptide 14a, which
1 are identical (Fig. 3), which was further corroborated by the was used in the next reaction without further purification. To a
C3 Marfey’s methods14) (Fig. 4).
solution of peptide 14a in CH2Cl2–DMF (=9:1) (25mL) were
In the previous study,1) the configuration of Ile was added 2,4,6-collidine (26µL, 0.196mmol), HOAt (13.5mg,
identified by a subtle chromatographic difference between 0.100mmol), and PyBOP (52.4mg, 0.100mmol), and then
the 2,3,4,6-tetra-O-acetyl-β-D-glucopyranosyl isothiocyanate the reaction mixture was stirred for overnight. The solvent
(GITC) derivatives, resulting in the misassignment. In this was removed and the residues was dissolved in EtOAc and
sturdy, our synthetic efforts confirmed that 1 possesses D-allo- saturated aqueous NH4Cl. The resulting mixture was extracted
Ile at position 4, instead of D-Ile. Judging from the common- with EtOAc (×3), washed with brine, dried over MgSO4, and
ality of the biosynthetic pathways, the D-Ile residue present concentrated to give the crude 15a. To the residue was added
in other derivatives, such as 3–5, should also be corrected to a mixture of TFA–H2O–iPr3SiH (=95:2.5:2.5) (1.0mL), and
D-allo-Ile.
the mixture was stirred for 1min for the removal of the Boc
group. The reaction mixture was diluted with Et2O (24mL),
and centrifuged at 3500×g for 5min at 4°C, and then Et2O
Experimental
General Remarks 1H- and 13C-NMR spectra were record- layer was removed by decantation. This procedure was
1
ed on a JEOL ECA 500 (500MHz for H-NMR) spectrometer. repeated twice. The crude 1a was purified by reversed-phase
Chemical shifts are denoted in δ (ppm) relative to residual sol- HPLC to afford 1a (21.1mg, 46% for 18 steps) as a white
vent peaks as internal standards (dimethyl sulfoxide (DMSO)- solid.
d6, δΗ 2.50, δC 39.5). Electrospray ionization (ESI)-MS
1a: [α]D19 −2.98 (c=0.1 MeOH); 1H-NMR (500MHz,
spectra were recorded on a Thermo Scientific Exactive mass DMSO-d6): see Fig. S1; 13C-NMR (500MHz, DMSO-d6): see
spectrometer. Optical rotations were recorded on a JASCO Fig. S2; high resolution (HR)-MS (ESI) Calcd for C48H82N9O+8
P-1030 polarimeter. LC-MS analyses were performed with a [M+H]+ 912.6281. Found 912.6275.
SHIMADZU HPLC system equipped with a LC-20AD intel-
1b: Compound 1b (18.7mg) was synthesized from 13b in
ligent pump coupled with an amaZon SL-NPC spectrometer the same manner as 1a in 41% yield for 18 steps. [α]D21 −3.85
1
(Bruker Daltonics). All reagents were used as supplied unless (c=0.1 MeOH); H-NMR (500MHz, DMSO-d6): see Fig. S3;
otherwise stated. Column chromatography was performed 13C-NMR (500MHz, DMSO-d6): see Fig. S4; HR-MS (ESI)
using 40–50µm Silica Gel 60N (Kanto Chemical Co., Inc.).
Procedure for Solid-Phase Peptide Synthesis (SPPS)
Calcd for C48H82N9O+8 [M+H]+ 912.6281. Found 912.6278.
Analytical Conditions For the comparison of the natu-
Step 1: The Fmoc group of the solid supported peptide was ral surugamide A (1) and the synthetic compounds 1a, and
removed by using a 20% piperidine–N,N-dimethylformamide 1b, the samples were analyzed by an LC-MS system with a
(DMF) solution (10min, room temperature).
Step 2: The resin in the reaction vessel was washed with spectrometer. The samples were separated by chromatogra-
DMF (×3) and CH2Cl2 (×3).
phy on COSMOSIL 5C18-AR-II (4.6 i.d. ×150mm) (Nacalai
SHIMADZU HPLC system coupled with an amaZon SL-NPC
Step 3: To the solution of carboxylic acid (4eq) were added Tesque) using H2O–MeCN (55:45)+0.05% TFA as the mo-
DIC (4eq, 0.50M in NMP) and Oxyma (4eq, 0.50M in DMF). bile phase (flow rate: 0.8mL/min).
After 2–3min of pre-activation, the mixture was injected
into the reaction vessel. The resulting mixture was stirred for portion of 1a or 1b was hydrolyzed in 6M HCl at 110°C
30min at 37°C. overnight. The solution was dried under a N2 stream and
Step 4: The resin in the reaction vessel was washed with dissolved in H2O. To the solution were added 100µL of 1%
DMF (×3) and CH2Cl2 (×3).
FDAA (2,4-dinitro-5-fluorophenyl-L-alaninamide) in acetone
Amino acids were condensed onto the solid support by re- and 20µL of 1M NaHCO3. The reaction mixture was stirred
peating Steps 1–4. at 50°C for 30min and quenched with 10µL of 2M HCl. The
Total Hydrolysis and Derivatization with FDAA A
Peptides 13a and 13b: The 2-chlorotrityl resin (156.0mg, FDAA-derivatized amino acids were analyzed as described
0.25mmol) in a Libra tube was swollen with CH2Cl2, and below. Samples were injected into an Agilent Zorbax SB-C3
then the excess solvent was removed by filtration. To the (4.6 i.d. ×150mm) column and separated using H2O with a
resin was added a solution of Fmoc-D-Leu-OH (176.7mg, 5% isocratic modifier of 1% formic acid in MeCN as mobile
0.50mmol) and i-Pr2NEt (262µL, 1.50mmol) in CH2Cl2 phase A, and MeOH with a 5% isocratic modifier of 1% for-
(0.5mL), and the mixture was stirred for 30min. The reaction mic acid in MeCN as mobile phase B. The column was eluted
mixture was filtered, washed with DMF (×3), CH2Cl2 (×3), in a 0.8mL/min, linear gradient mode from 30% to 60% over
and methanol. The Fmoc-D-Leu-2-chlorotrityl resin (50.8mg, 50min for mobile phase B.14)
0.05mmol) in the Libra tube was swelled in CH2Cl2 for 1h,
and then subjected to 7 cycles [Fmoc-D-Phe-OH, Fmoc-
Acknowledgments This work was partly supported by
L-Ile-OH, Fmoc-L-Lys-OH, Fmoc-D-Ile-OH (for 1a) or Fmoc- the Takeda Science Foundation, the Asahi Glass Foundation,
D-allo-Ile-OH (for 1b), Fmoc-L-Ile-OH, Fmoc-D-Ala-OH, the SUNBOR GRANT, the NOASTEC Foundation, the Aki-
Fmoc-L-Ile-OH] of the SPPS protocol (steps 1–4) to afford the yama Life Science Foundation, the Japan Agency for Medical
mixture of the resin-bound peptides 13a or 13b.
Research and Development (AMED Grant Number 18061402)
Surugamide (proposed Structure: 1a and Revised and Grants-in-Aid from the Ministry of Education, Cul-
A
Structure: 1b): To peptides 13a was added CH2Cl2– ture, Sports, Science and Technology (MEXT), Japan (JSPS
(CF3)2CHOH (=70:30) (0.5mL), and the reaction mixture KAKENHI Grant Numbers JP16703511, and JP18056499).

480
Chem. Pharm. Bull.
ChemBioChem, 17, 1709–1712 (2016).
Vol. 67, No. 5 (2019)
Conflict of Interest The authors declare no conflict of
interest.
6) Matsuda K., Kobayashi M., Kuranaga T., Takada K., Ikeda H.,
Matsunaga S., Wakimoto T., Org. Biomol. Chem., (2019), in press.
7) Trauger J. W., Kohli R. M., Mootz H. D., Marahiel M. A., Walsh C.
T., Nature (London), 407, 215–218 (2000).
Supplementary Materials The online version of this ar-
ticle contains supplementary materials.
8) Marahiel M. A., Stachelhaus T., Mootz H. D., Chem. Rev., 97,
2651–2673 (1997).
References
9) Li Q., Song Y., Qin X., Zhang X., Sun A., Ju J., J. Nat. Prod., 78,
944–948 (2015).
10) Zhou X., Huang H., Li J., Song Y., Jiang R., Liu J., Zhang S., Hua
Y., Ju J., Tetrahedron, 70, 7795–7801 (2014).
11) Li Q., Qin X., Liu J., Gui C., Wang B., Li J., Ju J., J. Am. Chem.
Soc., 138, 408–415 (2016).
1) Takada K., Ninomiya A., Naruse M., Sun Y., Miyazaki M., Nogi Y.,
Okada S., Matsunaga S., J. Org. Chem., 78, 6746–6750 (2013).
2) Mohimani H., Gurevich A., Mikheenko A., Garg N., Nothias L.-F.,
Ninomiya A., Takada K., Dorrestein P. C., Pevzner P. A., Nat.
Chem. Biol., 13, 30–37 (2017).
3) Xu F., Nazari B., Moon K., Bushin L. B., Seyedsayamdost M. R., J.
Am. Chem. Soc., 139, 9203–9212 (2017).
4) Kuranaga T., Matsuda K., Sano A., Kobayashi M., Ninomiya A.,
Takada K., Matsunaga S., Wakimoto T., Angew. Chem. Int. Ed., 57,
9447–9451 (2018).
12) Marfey P., Carlsberg Res. Commun., 49, 591–596 (1984).
13) Bhushan R., Brückner H., J. Chromatogr.
Biomed. Life Sci., 879, 3148–3161 (2011).
B Analyt. Technol.
14) Vijayasarathy S., Prasad P., Fremlin L. J., Ratnayake R., Salim A.
A., Khalil Z., Capon R. J., J. Nat. Prod., 79, 421–427 (2016).
5) Ninomiya A., Katsuyama Y., Kuranaga T., Miyazaki M., Nogi Y.,
Okada S., Wakimoto T., Ohnishi Y., Matsunaga S., Takada K.,