Pyrrole aminoimidazole alkaloid metabiosynthesis with marine sponges Agelas conifera and Stylissa caribica

Article abstract of DOI:10.1002/anie.201108119

Game-SET-match: Pyrrole aminoimidazole alkaloids (PAIs) are metabiosynthesized from chlorinated analogues of oroidin by cell-free enzyme preparations from PAI-producing sponges. Evidence and implications for the biosynthesis of PAIs include putative single-electron transfers (SETs) that promote Ci-C bond-forming reactions of precursors. Copyright

Full text of DOI:10.1002/anie.201108119

Angewandte
Chemie
DOI: 10.1002/anie.201108119
Electron Transfer
Pyrrole Aminoimidazole Alkaloid Metabiosynthesis with Marine
Sponges Agelas conifera and Stylissa caribica**
E. Paige Stout, Yong-Gang Wang, Daniel Romo, and Tadeusz F. Molinski*
Pyrrole aminoimidazole alkaloids (PAIs, Figure 1),[1] the
characteristic class of compounds found in marine sponges
of the genre Agelas, Stylissa, Phakellia, Axinella, and Hyme-
niacedon, have drawn contemporary interest because of their
complex structures and, in some cases, potent biological
activities.[2]
properties, binding to MreB protein in E. coli,[4c] and
inhibition of cell motility in a variety of cancer cell lines.[4d]
It is widely accepted that the biosynthesis of complex PAIs
begins with three simple building-block PAIs, hymenidin
(3a),[5] oroidin (3b),[6] and clathrodin (3c),[7] which are
produced naturally in certain sponges. Detailed field experi-
ments support the contention that oroidin (3b) and related
molecules, longamide and dispacamide, most likely serve as
the primary chemical defenses of Stylissa and Agelas
sponges.[8] Sceptrin (1a) is formally derived by [2+2] cyclo-
addition of 3a, and benzosceptrins A–C (4a–c)[9] are cyclized
analogues of 1. The potent antitumor agents, agelastatins A
(2a) and B (2b) can be envisioned as a product of the enzyme-
mediated cyclization/oxidation of 3a,b. The total synthesis of
several PAIs, including 3b,[10] 1a,[11] and 2,[12] have been
reported, thus culminating in the elegant synthesis of palauꢀ-
amine (5) by Baran and co-workers.[13]
Various schemes have been advanced to explain the
biosynthesis of 1a,[3] 5,[14] and related PAIs.[15,16] Most
proposals rationalize carbon–carbon bond formation with
mechanisms based on the electrophile/nucleophile duality
inherent in enamine–imine tautomerism of conjugated 2-
aminopyrroles, or concerted electrocyclic reactions that even
claim to explain the order of appearance of intermediates.[11]
While accounting for the general assembly of PAIs, to date, all
two-electron arrow-pushing schemes fail to address a funda-
mental discrepancy: no biosynthetic experimental evidence
has been advanced to support any of the proposed schemes.
Herein we report the first experimental evidence for the
biosynthesis of PAIs by sponges with enzyme-catalyzed
conversion of the oroidin analogue 3d (dichloroclathrodin)
into chlorinated versions of known benzosceptrins[13] and
nagelamides. Further, we provide the first evidence of
conversion of 3b into two known PAIs, benzosceptrin C
(4c) and nagelamide H (7a), which is strong evidence that
oroidin is the precusor to these natural products. These
successful “metabiosynthesis”[17] experiments employ cell-
free enzyme extracts obtained from the marine sponges
Stylissa caribica, Agelas conifera, and A. sceptrum, and carry
Figure 1. Representative PAIs from sponges Agelas spp. and Stylissa,
and the unnatural chlorinated analogues 3d,e, and 4d.
Sceptrin (1a),[3] an intriguing dimeric optically active C2-
symmetric cyclobutane, exhibits a broad range of biological
activities,[4] including antimicrobial[4a] and antimuscarinic[4b]
[*] Dr. E. P. Stout, Prof. Dr. T. F. Molinski
Department of Chemistry and Biochemistry
and Skaggs School of Pharmacy and Pharmaceutical Science
University of California, San Diego
9500 Gilman Drive, La Jolla, CA 92093 (USA)
E-mail: tmolinski@ucsd.edu
Dr. Y. G. Wang, Prof. Dr. D. Romo
À
out C C bond formation of PAIs through an oxidative
mechanism likely involving single-electron transfer (SET)
and radical couplings.
Department of Chemistry, Texas A&M University
P.O. Box 30012, College Station, TX 77842 (USA)
[**] Funding for this work from NIH (GM052964) and a Ruth L.
Kirschstein National Research Service Award (NIH/NCI T32
CA009523) is gratefully acknowledged. We thank J. J. LaClair and M.
Burkart for help with fluorescence spectra, B. Morinaka and Y. Su
(UCSD) for NMR and HRMS measurements, respectively. We are
grateful to J. Pawlik (UNC Wilmington), and the captain and crew of
the RV Cape Hatteras for logistical support during collecting
expeditions and in-field assays.
Many reported PAI structures share the same carbon/
nitrogen framework, but differ only in the number of Br
atoms in the pyrrole ring (e.g., 3a,[3] 3b,c,[18] and 2a,b[5]), thus
suggesting that the late-stage enzymatic steps in the biosyn-
thesis of PAIs for 3a–c may be indifferent to halogenation in
the pyrrole ring. We exploited this feature to clearly differ-
entiate biosynthesis of PAIs under cell-free conditions using
a non-natural chlorinated precursor, dichloroclathrodin (3d),
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201108119.
Angew. Chem. Int. Ed. 2012, 51, 1 – 6
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
which was prepared from the known 4,5-dichloropyrrol-2-yl
trichloromethyl ketone[19] using a strategy we previously
developed for 15N-oroidin.[10c]
Live sponges, Stylissa caribica, Agelas conifera, and A.
sceptrum were freshly collected by hand using scuba (À80 ft,
Bahamas) and immediately converted into cell-free prepara-
tions (sc-, ac-, and as-CFP respectively) which contained
active enzymes. An aqueous-organic suspension of sc-CFP
(0.2 mgmLÀ1 protein, Bradford assay) in 1:9 CH3CN/0.1m Na-
phosphate buffer (pH 7–9) rapidly consumed 3d (ca. 90%
within 30 min; Table 1) and generated several new chlorin-
Table 1: Enzyme-catalyzed reactions of PAIs with cell-free preparations
from A. conifera (ac), A. sceptrum (as), and S. caribica (sc).[a]
Entry Substrate[a] Enzyme Conversion [%][b] R[c]
pH,
conditions
Scheme 1. Structures of nagelamide H (7a) and the chlorinated prod-
ucts 6, 7d, and 8 obtained from cell-free metabiosynthesis of 3d and
base-promoted elimination of 8 to 4d.
4d 7d
8
[mmsÀ1
]
1
2
3
4
5
6
7
8
3d
3d
3d
3d
3d
3d
3d
3d
3d
3d
3d
3d
3d
3d
3d
1a
3b
3e
sc
sc
sc
ac
as
sc
sc
sc
ac
sc
ac
sc
ac
sc
sc
sc
as
sc
0
0
0
0
–
–
–
–
–
–
–
–
–
5
6
7
7
7
8
52
19
21
37
30
0
21
13
10
60
55
13
15
2
10.3
10.2
10.2
10.9
10.9
Compound 8, C22H20Cl4N10O5S (HR-APCI-TOFMS
m/z 675.0015, [MÀH]À), showed a new red-shifted chromo-
phore in its UV/vis spectrum (MeOH, lmax = 366 nm, e =
4000), in addition to the expected band from dichloropyr-
rol-2-carboxamide (lmax = 274 nm, e = 13600), and strong
fluorescence (lex = 366 nm; lem = 450). The 1H NMR spec-
trum of 8 (600 MHz, 1.7 mm microcryoprobe, [D6]DMSO)
showed two relatively unchanged pyrrole NH signals (d =
12.60, s; 12.62, s), thus suggesting two intact 2,3-dichloropyr-
role-2-carboxamide groups. The vinyl protons from the 1,2-
disubstituted alkene of 3d were absent and replaced by four
contiguously coupled CH signals (COSY) assigned to
a 1,2,3,4-tetrasubstituted cyclobutane ring reminiscent of 1a
but lacking symmetry. Interpretation of one-bond sp3 hetero-
9
50 10.3
19 5.22
7[d]
9
5
13
7[d]
10
11
12
13
14
15
16
17
18
–
–
–
–
–
–
–
–
–
6.88
0.88
3.00
6.56
–
2.44
–
11.0
–
7, CO[e]
7, CO[e]
7, FÀ[f]
7, FÀ[f]
7, heat[g]
8, Ar[h]
7
0.4 0.1
7
1.2 1.7
1
0
3
0
0
5
0
72[i] 60[i]
7
7
0
0
[a] For entries 1—15: [3d]=17 mm in 1:11 CH3CN/0.1m sodium phos-
phate buffer, 0.2 mgmLÀ1 total protein (Bradford assay) of cell-free
preparation (CFP) from either S.caribica (sc), A. conifera (ac), or A. scep-
trum (as). For further details and the conditions for entries 16–18, see the
Supporting Information. [b] Determined by LCMS (UV). [c] R is an
estimated initial rate of consumption of substrate, assuming substrate-
saturated conditions. [d] NaHSO3 aq. added during workup (final
[HSO3À]=0.1m). [e] Run at 1 atm; 15 min pre-saturation. [f] KF (aq,
10 mm), 15 min pre-incubation. [g] CFP was boiled (1008C) for 30 min
prior to addition of substrate. [h] Run at 1 atm; anaerobic conditions
were achieved by purging substrate solution and CFP with Ar for 15 min
prior to addition. [i] Observed products were tetrabrominated natural
products 4c and 7a, respectively.
1
nuclear coupling constants, JCH, which reflect ring strain,[20]
confirmed the four-membered ring [coupled HSQC of 8: 1JCH
1
136 Hz (C9), 134 (C9’), 148 (C10), 138 (C10’); see 1a: JCH
135 Hz (C9/9’), 136 (C10/10’)]. Interpretation of COSY and
HMBC data allowed connection of the partial structures (see
the Supporting Information), including key HMBC correla-
tions from both H9’ (d = 1.78) and H10’ (d = 3.35) to C11’ (d =
86.4) and from H10 (d = 3.58) to C12 (d = 116.0). NOESY
data were used to assign the relative configuration of the
cyclobutane ring. NOEs observed between H9 (d = 2.54),
H10, and H10’ supported a cis ring fusion, while correlations
between both H8 protons (d = 2.78, 3.00) and H9’ and
between both H8’ protons (d = 3.17, 3.25) and H9 supported
ated compounds including two with molecular masses about a trans orientation for the dichloropyrrol-2-carboxamide side
twice that of 3d, as well as an unstable compound 6. The two chains, all of which is consistent with known cyclobutane-
major products (Figure 1 and Scheme 1; 4d, [MÀH]À containing PAI natural products.[3]
m/z 593.13; 7d [MÀH]À m/z 718.21) displayed pseudomolec-
By comparing the formulas of 6 and 8, it is apparent that
ular ions with isotopic compositions corresponding to Cl4. the latter is the bisulfite addition product of the former
Improved conversion of 3d was realized by quenching the captured from a reactive intermediate in the reaction pathway
reaction mixture with bisulfite (100 mm, NaHSO3 aq.), (Table 1, entries 3 versus 8). Treatment of 8 with base (aq.
conditions that produced a different major product 8 along NaOH, aq. 238C) resulted in elimination of H2SO3 to give the
with 7d. Scaled-up enzyme-catalyzed preparative reactions postulated intermediate A (Scheme 1), with loss of both the
with 3d provided sufficient quantities of products (ca. 1.0– red-shifted chromophore and its associated fluorescence, and
1.5 mg) to allow complete structure elucidation of 7d and 8 by transformation into 4d, the Cl4-derivative of benzosceptrin A
NMR spectroscopy and MS.
(4a; Figure 1).[9] The 1H NMR spectrum of 4d is almost
2
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1 – 6
These are not the final page numbers!

Angewandte
Chemie
identical to that of the natural product benzosceptrin B (4b).
We observed that sceptrin (1a), itself, is not a viable substrate
for the enzyme-catalyzed oxidation; exposure of 1a to sc-CFP
produced no 4b or 7a and returned only starting material,
which suggests that sceptrin is not an intermediate in the
biosynthesis of benzosceptrins, nor is it isomerized under
these reaction conditions into ageliferin.[21]
of five co-occurring sponges that do not produce PAIs failed
to generate 4d/7d or oxidize 3d at all (see Table S1 in the
Supporting Information).
We detected appreciable concentrations of 3b in both A.
conifera and S. caribica, but not in A. sceptrum (see Figure S1
in the Supporting Information). Neither S. caribica nor A.
conifera from the Caribbean appear to contain natural
benzosceptrins (4a–c). These natural products were described
only from sponges collected in the Mediterranean Sea and
Pacific Ocean. Yet, remarkably, the chlorinated analogues of
the latter natural products are rapidly formed when 3d is
exposed to cell-free preparations of the sponges A. conifera
and S. caribica from the Caribbean Sea, thus revealing the
permissive nature of the sponge oxido reductases. Cell-free
preparations of both sponges unleash competent enzymes
which catalyze oxidative couplings and give rise to the same
products, 4d and 7d, found in sponge of the same genus from
different geographic locations. Disruption of the living sponge
tissue and cellular structure, apparently liberates enzymes
that retain their oxidative capacity to carry out oxidative
biosynthesis of PAIs, albeit without the level of control that
presumably operates within intact sponge cells.
The second, more hydrophobic product 7d has the
formula C24H24Cl4N11O5S (HR-APCI-TOFMS m/z 718.0439,
[MÀH]À), which corresponds to four-electron oxidative
dimerization of 3d and addition of the elements of taurine,
NH2(CH2)2SO3H.[22] Analysis of the NMR data (800 MHz,
1
[D6]DMSO; H NMR, COSY, HSQC, HMBC) revealed the
structure of 7d as the chlorinated analogue of 7a.[23]
Finally, natural 3b was also converted into known PAI
natural products under similar reaction conditions. Cell-free
preparations of A. sceptrum, which contains 1a but lacks 3b,
4c, and 7a at detectable concentrations (LCMS), rapidly
oxidized 3b to give the known natural products 4c and 7a
(Table 1, entry 17 and see the Supporting Information). This
is the first experimental evidence of formation of complex
PAIs from the widely presumed precursor oroidin.[24]
As dimerization of 3a into 4b is a net four-electron
oxidation (formal loss of 2H2), we propose that all conver-
sions of 3 into 1,2, and 5–8 involve discrete enzyme-catalyzed
reactions, which possibly involve one or more oxido reduc-
tases[25] as supported by the following lines of evidence.
Optimal conversion was found at pH ꢀ 7 (Table 1, entries 1–
7). Boiled sc-CFP (1008C) resulted in loss of activity and no
conversion of 3d (Table 1, entry 14). Similarly, activity was
greatly attenuated when incubation of 3d was carried out
under an atmosphere of CO (1 atm, Table 1, entries 10 and
11) or in the presence of aqueous fluoride (10 mm, KF aq.,
Table 1, entries 12 and 13), conditions known to inhibit
cytochrome oxidases. Addition of H2O2 (50 mm) to the
enzyme incubation made no change, however incubation of
3d with sc-CFP under anaerobic conditions also led to
significantly reduced yields of 4d and 7d (Table 1, entry 15).
Collectively, these data suggest that O2 is the terminal
oxidant. Ultracentrifugation of the CFP (10 kDa molecular
weight cut-off) resulted in retention of all activity (conversion
of 3d into 4d/7d) in the fraction containing species of
molecular weight greater than 10 kDa (see Figure S3 in the
Supporting Information); results that support enzyme-medi-
ated conversion of 3d.
We hypothesize benzosceptrins, nagelamides, and 6–8,
arise from sequential SETs which result in the four-electron
oxidation of 3a–c. By association, it is also likely, that
biosynthesis of many other PAIs are mediated by orches-
trated enzymatic SETs (Scheme 2) which initiate formation of
resonance-stabilized radical cation intermediates and subse-
quent reactions.[26] We propose that the biosynthesis of PAIs
proceeds through intermediates that partition between at
least two discrete enzyme-mediated pathways (Scheme 2),
thus leading to oxidized products 4d, 6, and 8 or redox neutral
products (e.g., 1a) with respect to the starting materials. In
the case of 1a, SET from the p-rich conjugated vinyl-2-
aminoimidazole of 3a to the metal center of the oxido
reductase generates the radical cation A which adds to
The observed conversions of 3d appear to be paired to
oxido reductases associated with Stylissa and Agelas sponges.
Replacement of sponge-CFPs with horseradish peroxidase
(Sigma) gave neither 4d nor 7d and only produced complex
mixtures of unidentifiable products. Furthermore, 9,10-dihy-
dro-dichloroclathrodin (3e; Table 1, entry 18), prepared in
a similar manner to 3d, does not undergo oxidation under the
reaction conditions, which is consistent with requirements for
a p-electron-rich donor substrate and stabilization of radical
cation intermediates through delocalization.
The enzyme-catalyzed dimerization of 3d is specifically
linked to PAI-producing sponges. Cell-free preparations of
several PAI-producing Agelas sponges all converted 3d into
4d/7d, while incubation of 3d with similarly prepared extracts
Scheme 2. Proposed mechanisms for enzyme-catalyzed SET of 3d and
cyclization to either 1a (path a) or 6/8 (path b).
Angew. Chem. Int. Ed. 2012, 51, 1 – 6
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
+
À
a neutral molecule of 3a with loss of H to form the first C C
bond of the pre-cyclobutane radical intermediate B followed
by cyclization to C (Scheme 2, path a). Subsequent SET
reduction by an electron carrier provides 1a with a net-zero
electron transfer. Thus, 1a is formed through a cryptic
oxidative mechanism. Compounds 4d, 6, and 8 arise from
path b, possibly through the cationic intermediate D, and
In conclusion, we have demonstrated cell-free enzymatic
conversion of oroidin (3b) and the oroidin-like precursor 3d
into the natural the products benzosceptrin C (4c) and
nagelamide H (7a), and their tetrachloro analogues 4d and
7d, respectively. The results implicate oxidoreductase-like
activity in the biosynthesis of PAIs in marine sponges through
single-electron transfer, and we suggest an SET pathway
which may explain the origin of sceptrin (1a) and related
oroidin-derived dimeric PAIs.
À
nucleophilic capture (H2O or HSO3 ) with a net four-electron
oxidation. Elimination of HX gives 4d after tautomerism
(Scheme 1).[27] The nagelamide H analogue 7d may emerge
from additional oxidation at the imidazole ring of B
(hydroxylation/tautomerism, not shown) and nucleophilic
capture of an incipient 2-iminoimidazolidinone by taurine.
Even numbers of SET oxidation steps could, conceivably,
give rise to the even-electron cationic species, including
acyliminium ions that have been implicated in the biosynthe-
sis of agelastatins,[12b] phakellins, palaua’mine, and other
PAIs.[15] High-energy radical cation intermediates would
better explain the formation of the ring-strained PAI alka-
loids, for example, the cyclobutanes 1, 4, 6, and 8 and the
trans-bicyclo[3.3.0]octane 5, because the enthalpic cost is paid
by reduction of the terminal oxidant, O2. An important
feature of the SET hypothesis is its compatibility with many
Received: November 18, 2011
Revised: February 3, 2012
Published online: && &&, &&&&
Keywords: alkaloids · biosynthesis · electron transfer ·
.
metabiosynthesis · natural products
[1] For clarity, each PAI is drawn as the free base and a discrete
tautomer, although most are isolated as salts, and counterions
and tautomer equilibria may vary with structure.
[2] A. Aiello, E. Fattorusso, M. Menna, O. Taglialatela-Scafati in
Modern Alkaloids. Structure, Isolation, Synthesis and Biology
(Eds.: E. Fattorusso, O. Taglialatela-Scafati), Wiley-VCH,
Darmstadt, 2008. The possibility that PAIs are the products of
bacterial consortia has not been excluded.
À
late-stage two-electron C C bond-forming reactions cur-
rently favored in PAI biosynthetic proposals. An additional
attractive feature of odd-electron pathways from 3 is acces-
sibility to reactive intermediates (radical) having sufficiently
high free energy to bypass the kinetic barriers that disfavor
ionic or concerted mechanisms (e.g., thermal [2+2] cyclo-
addition or cycloreversion[21]) without the need to invoke
unlikely excited states or violations of orbital symmetry
rules.[26]
[3] R. P. Walker, D. J. Faulkner, D. Van Engen, J. Clardy, J. Am.
Chem. Soc. 1981, 103, 6772 – 6773.
[4] a) V. S. Bernan, D. M. Roll, C. M. Ireland, M. Greenstein, W. M.
Maiese, D. A. Steinberg, J. Antimicrob. Chemother. 1993, 32,
539 – 550; b) R. Rosa, W. Silva, G. E. de Motta, A. D. Rodrꢁguez,
J. J. Morales, M. Ortiz, Experientia 1992, 48, 885 – 887; c) A. D.
Rodrꢁguez, M. J. Lear, J. J. La Clair, J. Am. Chem. Soc. 2008, 130,
7256 – 7258; d) A. Cipres, D. P. OꢀMalley, K. Li, D. Finlay, P. S.
Baran, K. Vuori, ACS Chem. Biol. 2010, 5, 195 – 202.
The evidence we present (Table 1, entries 10–13) strongly
[5] J. Kobayashi, Y. Ohizumi, H. Nakamura, Y. Hirata, Experientia
1986, 42, 1176 – 1177.
À
argues for a metalloenzyme-catalyzed C C bond formation in
PAI biosynthesis. The modular nature of PAI biosynthesis is
now succinctly explained by single-electron processes, not
unlike the modular biosynthesis of lignans which also
proceeds through radical phenolic couplings catalyzed by
metalloenzymes. Fe-porphyrin oxidoreductases are known to
carry out lignan biosynthesis,[28] for example conversion of
coniferyl alcohol into pineol, with dioxygen (O2) as the
terminal oxidant. We favor the latter in PAI biosynthesis,
although we cannot rule out the possibility that other
metalloenzymes, for example laccases, are involved.
Several details of the enzyme-mediated formation of PAIs
remain to be elucidated. For example, the mechanism by
which optical activity is induced. Compounds 4d, 7d, and 8
appear to have lower optical activity than the natural
products,[29] however, as in lignan biosynthesis, asymmetry
may be induced by the action of dirigent proteins[30] which are
coupled with oxidases to control stereochemistry, a property
that may be disrupted in cell-free preparations. The sub-
cellular location, structural ordering, and compartmentaliza-
tion of the biosynthetic enzymes, and additional oxidative
steps required to generate higher order PAIs including
palauꢀamine and “tetrameric” members—and purification of
these oxidoreductase(s)—and other aspects of SET oxidation
in the biosynthesis of PAIs are the subjects of ongoing
investigations in our laboratories.
[6] a) S. Forenza, L. Minale, R. Riccio, E. Fattorusso, J. Chem. Soc.
Chem. Commun. 1971, 18, 1129 – 1130; b) E. E. Garcia, L. E.
Benjamin, R. I. Fryer, J. Chem. Soc. Chem. Commun. 1973, 78 –
79.
[7] J. J. Morales, A. D. Rodriguez, J. Nat. Prod. 1991, 54, 629 – 631.
[8] T. Lindel, J. Hoffmann, M. Hochgꢂrtel, J. R. Pawlik, J. Chem.
Ecol. 2000, 26, 1477 – 1496.
[9] J. Appenzeller, S. Tilvi, M. T. Martin, J. F. Gallard, H. Elbitar,
E. T. H. Dau, C. Debitus, D. Laurent, C. Moriou, A. Al-
Mourabit, Org. Lett. 2009, 11, 4874 – 4877.
[10] a) T. Lindel, M. Hochgꢂrtel, J. Org. Chem. 2000, 65, 2806 – 2809;
b) M. R. Bhandari, R. Sivappa, C. J. Lovely, Org. Lett. 2009, 11,
1535 – 1538; c) Y. Wang, B. I. Morinaka, J. C. P. Reyes, J. H.
Wolff, T. F. Molinski, D. Romo, J. Nat. Prod. 2010, 73, 428 – 434.
[11] a) P. S. Baran, D. OꢀMalley, A. Zografos, Angew. Chem. 2004,
116, 2728 – 2731; Angew. Chem. Int. Ed. 2004, 43, 2674 – 2677;
b) P. S. Baran, L. Zografos, D. OꢀMalley, J. Am. Chem. Soc. 2004,
126, 3726 – 3727; c) V. B. Birman, X. T. Jiang, Org. Lett. 2004, 6,
2369 – 2371.
[12] a) D. Stien, G. T. Anderson, C. E. Chase, Y. H. Koh, S. M.
Weinreb, J. Am. Chem. Soc. 1999, 121, 9574 – 9579; b) M.
Movassaghi, D. S. Siegel, S. Han, Chem. Sci. 2010, 1, 561 – 566;
c) For a comprehensive review of other agelastatin A syntheses,
see S. M. Weinreb, Nat. Prod. Rep. 2007, 24, 931 – 948.
[13] a) I. B. Seiple, S. Su, I. S. Young, C. A. Lewis, J. Yamaguchi, P. S.
Baran, Angew. Chem. 2010, 122, 1113 – 1116; Angew. Chem. Int.
Ed. 2010, 49, 1095 – 1098; Angew. Chem. 2010, 122, 1113 – 1116;
b) I. B. Seiple, S. Su, A. Nakamura, J. Yamaguchi, L. Jørgensen,
4
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1 – 6
These are not the final page numbers!

Angewandte
Chemie
R. A. Rodriguez, D. P. OꢀMalley, T. Gaich, M. Kçch, P. S. Baran,
J. Am. Chem. Soc. 2011, 133, 14710 – 14726.
[23] T. Endo, M. Tsuda, T. Okada, S. Mitsuhashi, H. Shima, K.
Kikuchi, Y. Mikami, J. Fromont, J. Kobayashi, J. Nat. Prod. 2004,
67, 1262 – 1267.
[14] a) R. B. Kinnel, H. P. Gehrken, R. Swali, G. Skoropowski, P. J.
Scheuer, J. Org. Chem. 1998, 63, 3281 – 3286; b) K. S. Feldman,
A. Y. Nuriye, Org. Lett. 2010, 12, 4532 – 4535; c) Z. Zhiqiang Ma,
J. Lu, X. Wang, C. Chen, Chem. Commun. 2011, 47, 427 – 429.
[15] a) H. Hoffmann, T. Lindel, Synthesis 2003, 12, 1753 – 1783; b) A.
Al Mourabit, P. Potier, Eur. J. Org. Chem. 2001, 237 – 243; c) M.
Kçck, A. Grube, I. B. Seiple, P. S. Baran, Angew. Chem. 2007,
119, 6706 – 6714; Angew. Chem. Int. Ed. 2007, 46, 6586 – 6594;
d) A. Al-Mourabit, M. Zancanella, S. Tilvi, D. Romo, Nat. Prod.
Rep. 2011, 28, 1229 – 1260.
[16] For a bioinspired strategy to phakellin, see: a) L. H. Foley, G.
Bꢂechi, J. Am. Chem. Soc. 1982, 104, 1776 – 1777. For bioinspired
approaches to palauꢀamine and related PAIs, see: b) A. S. Dilley,
D. Romo, Org. Lett. 2001, 3, 1535 – 1538; c) S. G. Koenig, S. M.
Miller, K. A. Leonard, R. S. Lowe, B. C. Chen, D. J. Austin, Org.
Lett. 2003, 5, 2203 – 2206; d) Q. Li, P. Hurley, H. Ding, A. G.
Roberts, R. Akella, P. G. Harran, J. Org. Chem. 2009, 74, 5909 –
5919.
[24] Redox-neutral conversion of 3b into dibromosceptrin (1c) was
not detected in this experiment, the result is possibly due to
dysregulation of electron donor/acceptors in disrupted cell-free
preparations. High conversions of 3b/d into benzosceptrins (4c/
d) and nagelamides H (7a/d) were also achieved with CFPs from
other PAI-producing Caribbean sponges (see Table S1 in the
Supporting Information).
À
[25] Peroxidase-promoted radical pairings generate C C bonds
without incorporation of oxygen: a) T. M. Kutchan, W. M.
Chou, Plant J. 1998, 15, 289 – 300; b) K. A. Nielsen, B. L.
Lindberg-Møller in Cytochrome P450 (Ed.: P. R. Ortiz de Mon-
tellano), Kluwer, Dordrecht, 2005, pp. 553 – 574.
[26] There is no a priori reason to assume the involvement of radical
cation intermediates over radical anions except for abundant
À
precedence of peroxidases in C C coupling reactions, relatively
low oxidation potential of vinyl-2-aminoimidazoles, and the
experimental evidence presented herein. An efficient photo-
chemical catalytic redox-neutral synthesis of cyclobutanes by
radical anion coupling has been demonstrated. J. Du, T. P. Yoon,
J. Am. Chem. Soc. 2009, 131, 14604 – 14605.
[17] The term metabiosynthesis is coined here to define cell-free,
enzymatic reactions conducted on non-natural substrates to give
natural product analogues.
[18] a) M. Kçck, M. Z. Assmann, Z. Naturforsch. C 2002, 57, 157 –
160; b) P. A. Keifer, R. E. Schwartz, M. E. S. Koker, R. G.
Hughes, Jr., D. Rittschof, K. L. Rinehart, J. Org. Chem. 1991,
56, 2965 – 2975.
[19] P. Bꢃlanger, Tetrahedron Lett. 1979, 20, 2505 – 2508.
[20] D. S. Dalisay, B. I. Morinaka, C. K. Skepper, T. F. Molinski, J.
Am. Chem. Soc. 2009, 131, 7552 – 7553.
[21] Baran and co-workers propose that 1a is biosynthetically
transformed into the isomeric ageliferin by thermal cyclobutane
ring opening of 1a into a diradical which re-closes to the
cyclohexene, see: B. H. Northrop, D. P. OꢀMalley, A. L. Zogra-
fos, P. S. Baran, K. N. Houk, Angew. Chem. 2006, 118, 4232 –
4236; Angew. Chem. Int. Ed. 2006, 45, 4126 – 4130. Al-Mourabit
and co-workers favor an alternative even-electron pathway from
3b to 4b. See Ref. [9].
[27] For example, photochemical [2+2] cycloaddition of 3b, which
requires an S1 excited state and intersystem crossing to T1, is
unlikely in Stylissa caribica, a sponge that biosynthesizes 1b but
lives in subdued light conditions, at depths no less than À25 m. In
any case, failure of photochemical dimerization of 3b into 1a
under laboratory conditions is likely due to other factors.
[28] W. Bao, D. M. OꢀMalley, R. R. Sederoff, Science 1993, 260, 672 –
674. We thank an anonymous reviewer for this analogy.
[29] Enantio-heterogeneity of PAIs is not unexpected. For example,
both optical antipodes of the natural product, (+)-dibromopha-
kellin, have been reported. See: a) G. Sharma, B. Magdoff-
Fairchild, J. Org. Chem. 1977, 42, 4118 – 4124; b) G. de Nanteuil,
A. Ahond, J. Guilhem, C. Poupat, E. Tran Huu Dau, P. Potier, M.
Pusset, J. Pusset, P. Laboute, Tetrahedron 1985, 41, 6019 – 6033.
[30] L. B. Davin, H.-B. Wang, A. L. Crowell, D. L. Bedgar, D. M.
Martin, S. Sarkanen, N. G. Lewis, Science 1997, 275, 362 – 366.
[22] Taurine is an osmolyte commonly found in sponges in high
concentrations and, evidently, entrained in our CFP prepara-
tions.
Angew. Chem. Int. Ed. 2012, 51, 1 – 6
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!

.
Angewandte
Communications
Communications
Electron Transfer
Game-SET-Match: Pyrrole aminoimida-
zole alkaloids (PAIs) are metabiosynthe-
sized from chlorinated analogues of
oroidin by cell-free enzyme preparations
from PAI-producing sponges. Evidence
and implications for the biosynthesis of
PAIs include putative single-electron
E. P. Stout, Y. G. Wang, D. Romo,
T. F. Molinski*
&&&& — &&&&
Pyrrole Aminoimidazole Alkaloid
Metabiosynthesis with Marine Sponges
Agelas conifera and Stylissa caribica
À
transfers (SETs) that promote C C bond-
forming reactions of precursors.
6
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1 – 6
These are not the final page numbers!