Precursor-Directed Biosynthesis of Phenylbenzoisoquinolindione Alkaloids and the Discovery of a Phenylphenalenone-Based Plant Defense Mechanism

Article abstract of DOI:10.1021/acs.jnatprod.7b00885

Phenylbenzoisochromenone glucosides (oxa-phenylphenalenone glucosides) occurring in some phenylphenalenone-producing plants of the Haemodoraceae undergo conversion to phenylbenzoisoquinolindiones (aza-phenylphenalenones) in extracts of Xiphidium caeruleum

Full text of DOI:10.1021/acs.jnatprod.7b00885

Article
pubs.acs.org/jnp
Cite This: J. Nat. Prod. XXXX, XXX, XXX−XXX
Precursor-Directed Biosynthesis of Phenylbenzoisoquinolindione
Alkaloids and the Discovery of a Phenylphenalenone-Based Plant
Defense Mechanism
Yu Chen,†,‡ Christian Paetz,† and Bernd Schneider*,†
†
̈
Max Planck Institut fur Chemische Okologie, Hans Knoll Straße 8, 07745 Jena, Germany
̈
̈
‡Jiangsu Key Laboratory for the Research and Utilization of Plant Resources, Institute of Botany, Jiangsu Province and Chinese
Academy of Sciences, Qianhu Houcun 1, 210014 Nanjing, China
S
* Supporting Information
ABSTRACT: Phenylbenzoisochromenone glucosides (oxa-phenylphenale-
none glucosides) occurring in some phenylphenalenone-producing plants of
the Haemodoraceae undergo conversion to phenylbenzoisoquinolindiones
(aza-phenylphenalenones) in extracts of Xiphidium caeruleum. Precursor-
directed biosynthetic experiments were used to generate a series of new
phenylbenzoisoquinolindiones from native phenylbenzoisochromenone gluco-
sides and external amines, amino acids, and peptides. Intermediates of the
conversion were isolated, incubated with cell-free extracts, and exposed to
reactions under oxidative or inert conditions, respectively, to elucidate the
entire pathway from phenylbenzoisochromenones to phenylbenzoisoquino-
lindiones. An intermediate in this pathway, a reactive hydroxylactone/aldehyde,
readily binds not only to amines in vitro but may also bind to the N-terminus
of biogenic peptides and proteins of herbivores and pathogens in vivo. The
deactivation of biogenic amino compounds by N-terminal modification is
discussed as the key reaction of a novel phenylphenalenone-based plant defense mechanism. According to these data, the
ecological function of phenylphenalenone-type compounds in the Haemodoraceae, subfamily Haemodoroideae, has been
substantiated.
henylphenalenones and their oxa- and aza-derivatives
(Figure 1) belong to a group of polycyclic phenyl-
may function as phytoalexins. Alternatively, the accumulation of
phenylphenalenone derivatives reported for the roots7 and the
stamens6a suggests that these compounds could be constitutive
and therefore may be considered as phytoanticipins rather than
phytoalexins. The hypothesis that these are phytoanticipins
seems valid, especially for plants of the Haemodoroideae,
because this subfamily is rich in glucosidic phenylphenalenones
and phenylbenzoisochromenones (PBICs, oxa-phenylphenale-
nones, C18O skeleton).7,8 Other constitutive glycosidic plant
defense compounds, such as the glucosinolates, cyanogenic
glycosides, benzoxazinoid glucosides, and avenacosides,9 are
stored in the vacuole. During an attack by herbivores or
pathogens, vacuoles are mechanically damaged, and defensive
glucosides are exposed to hydrolytic enzymes present in the
cytosol or neighboring cells. Owing to their hydrophilicity,
phenylphenalenone- and PBIC glucosides are also stored in the
vacuole and, after cell disruption, may undergo hydrolytic
cleavage. After the hydrolysis of glucosides, phenylphenalenone
aglucones could, as photodynamic agents or as Michael
acceptors, act to deter pathogens or predators. The photo-
dynamic activity of phenylphenalenone-producing Lachnanthes
P
Figure 1. Classification of phenylphenalenones (C19), their oxa-
derivatives (phenylbenzoisochromenones, PBICs, C18O), and aza-
derivatives (phenylbenzoisoquinolindiones, PBIQs, C18N).
propanoid-derived natural products1 that occur mainly in the
monocotyledonous plant families Haemodoraceae2 and Musa-
ceae.3 In banana (Musa species), the role of phenyl-
phenalenones as inducible plant defense metabolites (phytoa-
lexins) has been well documented.4
Owing to the relatively high levels of phenylphenalenones
that occur in some Haemodoraceae species, plants and root
cultures of this plant family have served as model systems for
biosynthetic studies5 and for locating phenylphenalenone-type
compounds in the plant tissue.6 Just as in the family Musaceae,
phenylphenalenones in some species of the Haemodoraceae
Received: November 15, 2017
© XXXX American Chemical Society and
American Society of Pharmacognosy
A
DOI: 10.1021/acs.jnatprod.7b00885
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Figure 2. Conversion of intravacuolar phenylphenalenone-derived phenylbenzoisochromenone (PBIC) glucosides (1a−c) from X. caeruleum to
phenylbenzoisoquinolindiones (PBIQs) by a series of enzymatic and spontaneous reactions. The occurrence of the ß-glucosidase and the esterase in
the same cell is still hypothetical; alternatively, they may be located in adjacent cells. 1-Carboxy-5,6-dihydroxy-4-phenylnaphthalen-8-aldehyde is in
equilibrium with its cyclic isomer, the hydroxylactone (4a). The aldehyde represents a highly reactive defense compound that binds to primary
amines such as amino acids and proteins.
tinctoria plants in pigs was first described by Darwin.10 The
phenalenones are reported to act as photosensitizers with the
formation of singlet oxygen as the active agent against
pathogens.11 The ability of phenalenone-type compounds to
function as Michael acceptors has been discussed as another
possible mode of action.12
which could be derived from a native PBIC. To test this
possibility, excised flowers of X. caeruleum, which are rich in
PBIC glucosides (see Figure S3 and Scheme S1), were
suspended in aqueous solutions of (3-13C)-L-phenylalanine
and (1-13C)-L-leucine, respectively. After incubation at ambient
temperature for 12 h, the 13C-labeled PBIQs 2-[(1″S)-1″-
carboxy-2″-phenylethyl]-5-hydroxy-7-phenyl-2H-benzo[de]-
isoquinoline-1,6-dione (5) and 2-[(1″S)-1″-carboxy-3″-methyl-
n-butyl]-5-hydroxy-7-phenyl-2H-benzo[de]isoquinoline-1,6-
dione (6) were isolated from the flower extract (Figure S1).
Unlike the 13C NMR spectra of nonlabeled compounds, the
spectra of labeled 5 and 6 displayed enhanced signals at δ 35.5
and 172.5, respectively, corresponding to 13C-enriched C-2″ of
5 and 13C-enriched C-5″ of 6 (Figure S2). The isotopic
enrichments of labeled carbon atoms were 91.8 and 90.9% for 5
and 6, respectively. This finding indicated that the nitrogen-
containing side chain of PBIQs originated from amino acids.
To examine whether PBIQ alkaloids are formed from a
spontaneous reaction or an enzyme-catalyzed reaction, we
incubated various amino acids with acetone extracts of
macerated aerial plant material instead of a suspension of
fresh flowers. Using this precursor-directed biosynthetic
approach, 10 new PBIQs (7−16) (Figure 3) and nine known
PBIQs (5, 6, 17, 27−32) (Table S1) were obtained.6a,8 For MS
and NMR data, see Tables S2−S5. Two of the new
compounds, 7 and 8, are enantiomers of compounds 5 and 6
generated by incubation of the plant material with D-
phenylalanine and D-leucine instead of the natural L-amino
acids. Likewise, the new 5-O-methyl derivatives 9 and 10 were
obtained from D-phenylalanine and D-leucine. In the case of L-
lysine, the two primary amino groups both reacted with the
aldehyde form of compound 4a to give bis-PBIQ 15. The
formation of the latter compound and of (2-hydroxyethyl)-
lachnanthopyridone (17) shows that the reaction is not
restricted to the amino group of α-amino acids but is generally
accessible for primary amines. Incubation with L-cysteine
resulted in the formation of another unusual compound, 16,
The occurrence of phenylphenalenone-related compounds
such as PBICs and phenylbenzoisoquinolindiones (PBIQs, aza-
phenylphenalenones, C18N skeleton) in Xiphidium caeruleum2b
and the ease with which hydrolytic and oxidative conversion of
such compounds13 occurs prompted us to investigate whether
the bioactivation of phenylphenalenone and PBIC glucosides
results in chemically reactive derivatives. Reactive derivatives
and biosynthetic intermediates may act as defense compounds
against herbivores and pathogens. On the basis of the recent
observation of the co-occurrence of phenylphenalenone
glucosides, PBIC glucosides, and PBIQs in X. caeruleum,8
especially in the stamens,6a and on the basis of the results of
further phytochemical and metabolic studies on phenyl-
phenalenone derivatives in X. caeruleum,6c herein we propose
a novel defense mechanism (Figure 2) that may apply not only
to the studied species and the Haemodoroideae subfamily but
also to other phenylphenalenone-producing plants.
To substantiate the hypothetical pathway shown in Figure 2
and demonstrate precursor-product relationships, we con-
ducted phytochemical studies and isotopic labeling experi-
ments. Isolated compounds 1a−4b were incubated with cell-
free extracts to observe metabolic reactions, and spontaneous
conversions were tracked under oxidative and inert conditions.
PBIQ alkaloids (C18N scaffold) were first reported from L.
tinctoria14 and later found in X. caeruleum6a,8 and in its sister
species, Wachendorfia thyrsiflora.7 Until now, their biosynthesis
had not been investigated. On the basis of the chemical
structures of PBIQs,6a it is reasonable to hypothesize that
nitrogen-containing side chains are derived from plant
proteinogenic amino acids. If the hypothesis shown in Figure
2 is correct, the amino acid should react with an aldehyde,
B
DOI: 10.1021/acs.jnatprod.7b00885
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Figure 3. Structures of PBIQs 5−17, phenylphenalenone glucosides 18 and 19, and PBICs 20 and 21.
which has a thiazolo[3,2-b]isoquinoline backbone. The use of
acetone as a solvent and the preparation of D-amino acid
derivatives demonstrated that PBIQs are spontaneously formed
by a nonenzymatic reaction. Starting from primary amines and
excised flowers or extracts of aerial plant material afforded these
alkaloids in one preparative step. It is noteworthy that extracts
from green plant parts resulted in a mixture of 5-hydroxy- and
5-methoxy-PBIQs (Table S1), whereas flower material gave
only the 5-hydroxy-PBIQs, suggesting the plant parts contained
tissue-specific PBICs as precursors (see below). Because X.
caeruleum can be hydroponically grown and vegetatively
propagated, plant material is accessible for the precursor-
directed biosynthesis of PBIQs on a preparative scale.
To explore the biosynthetic precursors of PBIQs, HPLC-
ESIMS analyses and further phytochemical investigation of
different lyophilized tissues of X. caeruleum were performed.
The results showed that phenylphenalenones mainly occur in
the form of the five glucosides (1a - 1c, 18, 19) (Figure S3,
Supporting Information). PBIC glucosides 1a and 1b were the
two major constituents in flowers, while 1a - 1c were the three
major glucosides in green leaves. The C19-phenylphenalenone
glucosides (18, 19) (Figure 3) are only present in the roots.
However, no PBIQ was found in the acetone extracts of
lyophilized plant material, which supports the finding that
PBIQs are artifacts generated during extraction and isolation.
Incubation experiments using cell-free extracts demonstrated
that 1a and 1b were converted to their aglycones (2a, 2b)
(paragraph S4) by hydrolytic enzymes from the green leaves of
X. caeruleum when oxidation was simultaneously suppressed
using (NH4)2SO3 (Scheme 1). It should also be noted that the
purification of 1b required the compound to be collected in an
(NH4)2SO3 solution to prevent oxidation. The hydrolysis of
glucosides could be blocked by castanospermine, a glucosidase
inhibitor, which suggested that the hydrolysis of the glucosidic
bond, as well as of the malonyl ester bond, was one of the first
steps operating during the metabolic conversion of malonyl-
glucoside 1a. In the absence of (NH4)2SO3, both 1a and 1b
were transformed into the same oxidized product, lachnantho-
pyrone (3a) (Scheme S3.1−3). The transformation from 1a to
3a was shown to be time- and temperature-dependent (Scheme
S4). The oxidation of 2a in air gave 3a (Scheme S6), which
indicated that this conversion happens spontaneously and
without the catalytic activity of an oxidase. Aglycone 2b, though
too unstable to be isolated, was clearly detected by HPLC-
HRESIMS. Thus, 2b underwent spontaneous decarboxylation
followed by oxidation to generate 3a. Our investigation
C
DOI: 10.1021/acs.jnatprod.7b00885
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Scheme 1. Conversions of Glucosides 1a and 1b to 3a by a Cell-Free Extract from X. caeruleum
Scheme 2. pH-Dependent Conversion of Compound 3a to PBIQs 5 and 16
unambiguously establishes the early steps of the metabolic
pathway 1a/1b → 2a/2b → 3a (Scheme 1).
through a series of pH-dependent nonenzymatic reactions. This
pathway was verified by the detection of 5, which was generated
by directly incubating L-phenylalanine with 1a and 1b in cell-
free extracts at pH 8.0 (Scheme S5). Compound 16, formed
from 1-carboxy-5,6-dihydroxy-4-phenylnaphthalen-8-aldehyde
with L-cysteine, is a special case because the reaction did not
stop after formation of the isoquinolindione but continued in
annelation of a thiazole ring involving both a 1,6-Michael
addition reaction and an oxidation step.
Likewise, methoxy-PBIQs arose from the methoxy malonyl-
glucoside (1c) via a parallel pathway comprising its aglycone
(2c) (Scheme S2.3), 5-methoxylachnanthopyrone (Scheme
S3.4) (3b), and 3,6-dihydroxy-5-methoxy-7-phenyl-3H-benzo-
[de]isochromen-1-one (4b) as the intermediates. This “5-OMe”
pathway was also supported by the organ-specific occurrence of
5-OMe PBICs in aerial green plant parts (Scheme S1).
Further experiments were performed to identify the pathway
from 3a to PBIQ 5 (Scheme 2). As demonstrated for the
conversions in buffer solutions of different acidities (pH 4.0−
7.0), 3a was readily hydrated to 4a at pH 4.0−5.0 (Scheme
S7A−C), which was confirmed by a parallel experiment in
weakly acidic solution (Scheme S8). When the pH of the buffer
was increased to pH 5.0−6.0, 4a was oxidized to 22 (Scheme
S7D). In the presence of an amine (e.g., L-phenylalanine) in
the above mixtures, PBIQ 5 was obtained from 4a at pH 5.0
(Scheme S9A−C), and its yield increased with the basicity of
the buffer (Scheme S9D and E). 1-Carboxy-5,6-dihydroxy-4-
phenylnaphthalen-8-aldehyde was postulated to be the reactive
intermediate during the formation of 5. These results clearly
indicated that, after the hydrolysis of the flower glucosides 1a
and 1b, PBIQs (Table S1) are formed via intermediate 3a
D
DOI: 10.1021/acs.jnatprod.7b00885
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Scheme 3. PBIC 4a Coupling with Various Peptides
The formation of PBIQs from PBICs prompted the question
of whether PBICs will react with proteins or peptides to mimic
the interaction with biomolecules of pathogens or herbivores.
Therefore, a standard peptide mixture of Tyr-Gly, Val-Tyr-Val,
Met-Phe-Gly-Gly-Tyr (methionine enkephalin), and Leu-Phe-
Gly-Gly-Tyr (leucine enkephalin), all of which have a free
terminal amino group, was added to 4a as a PBIC prototype in
different buffer solutions (pH from 4 to 8) at 30 °C for 2 h
(one-pot reactions). HPLC-HRESIMS analysis was used to
identify the adducts 23−26 (Scheme 3). The results revealed
that each peptide was bound via the free amino group to a
PBIQ moiety originating from 4a. The yields of adducts 23−26
increased as the basicity of the reaction mixture changed from
pH 5.0 to 7.0 (Scheme S10A−E). However, at pH 8.0, 23 and
24 became the major products (Scheme S10 F). From these
data, it is evident that plant-derived PBICs are involved in the
N-terminal modification of peptides and proteins. The reaction
mechanism may be effective in a chemical environment, as it
exists in parts of the insect gut that have appropriate pH values.
Thus, this reaction is thought to be part of the phenyl-
phenalenone-based defense against florivores, herbivores, and
plant cell-disrupting microorganisms. This mild conjugation
reaction also represents a new preparative approach to
chemically modify proteins in a site-selective way.
be inactive metabolic end products, PBIC glucosides provided
evidence via bioassays of their antimicrobial activity.
ASSOCIATED CONTENT
* Supporting Information
■
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.jnat-
prod.7b00885.
Experimental procedures, plant materials, extraction and
isolation of PBICs from X. caeruleum and PBIQs from
precursor-directed biosynthetic experiments, metabolite
analysis by LC-ESIMS and NMR, NMR spectra, and
bioassay data (PDF)
AUTHOR INFORMATION
Corresponding Author
*E-mail: schneider@ice.mpg.de.
■
ORCID
Bernd Schneider: 0000-0003-4788-1565
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank the greenhouse team of the Max Planck
Institute for Chemical Ecology (Jena, Germany) for growing
and photographing plants and Emily Wheeler for editorial
assistance. Y.C. acknowledges a research fellowship (No.
50015316) awarded from the Sino-German (CSC-DAAD)
Postdoc Scholarship Program. C. Weigelt and Dr. K. Voigt
(HKI Jena) are acknowledged for performing antimicrobial
bioassays.
The naturally occurring PBICs, their transformed aglycones,
and the PBIQs derived from precursor-directed biosynthetic
experiments were subjected to antimicrobial tests (Table S8).
Compared to ciprofloxacin and amphotericin B, all tested
PBICs and their aglycones showed moderate activity against E.
coli and S. salmonicolor. In addition, 3a, 4a, 20, and 21 (Figure
3) showed activity against multiresistant Staphylococcus aureus
strains. Tested PIBQs showed no bioactivity.
The present study reveals a novel phenylphenalenone-based
plant defense mechanism operating in X. caeruleum and other
species of the Haemodoraceae, subfamily Haemodoroideae.
The hydroxylactone 4a and its acyclic aldehyde form, which is
oxidatively generated from plant PBICs, have been identified as
the reactive structures that bind to primary amines, amino
acids, and peptides. Hence, biogenic peptides and proteins of
herbivores and pathogens may be deactivated by N-terminal
modification. The reaction was also used for precursor-directed
biosynthetic experiments to prepare PBIQs, being substituted
at the ring nitrogen with amino acid-, peptide-, and protein-
derived side chains. In contrast to PBIQs, which were shown to
REFERENCES
■
(1) Cooke, R.; Edwards, J. Prog. Chem. Org. Nat. Prod. 1981, 40,
153−190.
(2) (a) Cooke, R.; Segal, W. Aust. J. Chem. 1955, 8, 107−113.
(b) Opitz, S.; Holscher, D.; Oldham, N. J.; Bartram, S.; Schneider, B. J.
̈
Nat. Prod. 2002, 65, 1122−1130.
(3) (a) Luis, J. G.; Echeverri, F.; Quinones, W.; Brito, I.; Lopez, M.;
̃
Torres, F.; Cardona, G.; Aguiar, Z.; Pelaez, C.; Rojas, M. J. Org. Chem.
1993, 58, 4306−4308. (b) Otal
́
varo, F.; Nanclares, J.; Vasquez, L. E.;
Quinones, W.; Echeverri, F.; Arango, R.; Schneider, B. J. Nat. Prod.
̃
2007, 70, 887−890.
(4) (a) Flors, C.; Nonell, S. Acc. Chem. Res. 2006, 39, 293−300.
́
(b) Hidalgo, W.; Chandran, J. N.; Menezes, R. C.; Otalvaro, F.;
E
DOI: 10.1021/acs.jnatprod.7b00885
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Schneider, B. Plant, Cell Environ. 2016, 39, 492−513. (c) Holscher, D.;
̈
Dhakshinamoorthy, S.; Alexandrov, T.; Becker, M.; Bretschneider, T.;
Buerkert, A.; Crecelius, A. C.; De Waele, D.; Elsen, A.; Heckel, D. G.;
Heklau, H.; Hertweck, C.; Kai, M.; Knop, K.; Krafft, C.; Maddula, R.
K.; Matthaus, C.; Popp, J.; Schneider, B.; Schubert, U. S.; Sikora, R. A.;
Svato, A.; Swennen, R. L. Proc. Natl. Acad. Sci. U. S. A. 2014, 111,
105−110. (d) Kamo, T.; Kato, N.; Hirai, N.; Tsuda, M.; Fujioka, D.;
Ohigashi, H. Biosci., Biotechnol., Biochem. 1998, 62, 95−101.
(5) (a) Opitz, S.; Schneider, B. Phytochemistry 2003, 62, 307−312.
̌
(b) Brand, S.; Holscher, D.; Schierhorn, A.; Svatos, A.; Schroder, J.;
̈
̈
Schneider, B. Planta 2006, 224, 413−428. (c) Munde, T.; Brand, S.;
Hidalgo, W.; Maddula, R. K.; Svatos, A.; Schneider, B. Phytochemistry
2013, 91, 165−176.
̌
(6) (a) Chen, Y.; Paetz, C.; Menezes, R. C.; Schneider, B.
Phytochemistry 2016, 128, 95−101. (b) Opitz, S.; Schneider, B.
Phytochemistry 2002, 61, 819−825. (c) Opitz, S.; Schnitzler, J. P.;
Hause, B.; Schneider, B. Planta 2003, 216, 881−889. (d) Holscher, D.;
̈
Schneider, B. Planta 2007, 225, 763−770.
(7) Fang, J.-J.; Kai, M.; Schneider, B. Phytochemistry 2012, 81, 144−
152.
(8) (a) Fang, J.-J.; Holscher, D.; Schneider, B. Phytochemistry 2012,
̈
82, 143−148. (b) Schneider, B. Phytochem. Lett. 2017, 21, 104−108.
(9) Morant, V.; Jørgensen, K.; Jørgensen, C.; Paquette, S. M.;
San
1795−1813.
́ ́
chez-Perez, R.; Møller, B. L.; Bak, S. Phytochemistry 2008, 69,
(10) Darwin, C. R. In On The Origin of Species by Means of Natural
Selection or the Preservation of Favoured Races in the Struggle for Life.
Watts, p. 1; Appleton and Co., New York, 1859.
(11) Lazzaro, A.; Corominas, M.; Marti, C.; Flors, C.; Izquierdo, L.
R.; Grillo, T. A.; Luis, J. G.; Nonell, S. Photochem. Photobiol. Sci. 2004,
3, 706−710.
(12) Hidalgo, W.; Duque, L.; Saez, J.; Arango, R.; Gil, J.; Rojano, B.;
́
Schneider, B.; Otalvaro, F. J. Agric. Food Chem. 2009, 57, 7417−7421.
(13) Chen, Y.; Paetz, C.; Menezes, R. C.; Schneider, B.
Phytochemistry 2017, 133, 15−25.
(14) (a) Bazan, A. C.; Edwards, J. Phytochemistry 1976, 15, 1413−
1415. (b) Edwards, J.; Weiss, U. Tetrahedron Lett. 1972, 13, 1631−
1634.
F
DOI: 10.1021/acs.jnatprod.7b00885
J. Nat. Prod. XXXX, XXX, XXX−XXX