2008

NPC Natural Product Communications Vol. 3

No. 8
1205 - 1216
Plant Secondary Metabolism: Diversity, Function and its
Evolution
Michael Wink

Institute of Pharmacy and Molecular Biotechnology, INF 364, 69120 Heidelberg, Germany

wink@uni-hd.de

Received: April 2nd, 2008; Accepted: May 23rd, 2008

A typical character of plants is the production and storage of usually complex mixtures of secondary metabolites (SM). The
main function of secondary metabolites is defense against herbivores and microbes; some SM are signal compounds to attract
pollinating and seed dispersing animals or play a role in the symbiotic relationships with plants and microbes. The distribution
of SM in the plant kingdom shows an interesting pattern. A specific SM is often confined to a particular systematic unit, but
isolated occurrences can occur in widely unrelated taxonomic groups. This review tries to explain the patchy occurrence of SM
in plants. It could be due to convergent evolution, but evidence is provided that the genes that encode the biosynthesis of SM
appear to have a much wider distribution than the actual secondary metabolite. It seems to be rather a matter of differential
gene regulation whether a pathway is active and expressed in a given taxonomic unit or not. It is speculated that the genes of
some pathways derived from an early horizontal gene transfer from bacteria, which later became mitochondria and
chloroplasts. These genes/pathways should be present in most if not all land plants. About 80% of plants live in close symbiotic
relationships with symbiotic fungi (ectomycorrhiza, endophytes). Recent evidence is presented that these fungi can either
directly produce SM, which were formerly considered as plant SM or that these fungi have transferred the corresponding
pathway gene to the host plant. The fungal contribution could also explain part of the patchy occurrence patterns of several
secondary metabolites.

Keywords: evolution, chemotaxonomy, horizontal gene transfer, molecular phylogeny.

Diversity and function of compounds. The structurally diverse SM can be

plant secondary metabolites grouped according to their biosynthetic pathways and
structures into two large groups: SM with nitrogen in
A typical character of plants is the production and
their structures and those without (Table 1). Within
storage of usually complex mixtures of secondary
both complexes we can define further units, which
metabolites (SM). The constituents of these mixtures
can be split into even finer subgroups [1,4-7, 8-10].
often belong to several classes of SM, for example
terpenoids are often accompanied by phenolics.
Whereas phytochemists were and are always
Usually we find a limited number of major SM and
enthusiastic about the high structural diversity of SM,
several minor components, which are often
biologists tried to understand why and how plants
biosynthetically related to the main constituents [1-7].
synthesize, transport and store all these different SM.
As compared to the numbers of known SM, only
For more than 60 years phytochemists have
limited numbers of SM have been studied in detail in
separated complex mixtures of SM by prep.TLC,
terms of physiology, biochemistry and ecology.
LC, and HPLC and have isolated more than
Nevertheless, it is safe to assume, that SM are not
100,000 phytochemicals. Their structures have
functionless waste products (as suggested earlier in
been determined by mass spectrometry, NMR
the 20th century) [4], but important for the plants in
spectroscopy and X-ray crystallography. It has been
an ecological context [8, 12-14].
suggested that only 20 to 30% of the known 350,000
plant species have been examined phytochemically in
Since early days when land plants evolved in
some detail; therefore, the real number of SM present
the Devonian period, they had to cope with animals,
in the plant kingdom very likely exceeds 200,000
1206 Natural Product Communications Vol. 3 (8) 2008
Table 1: Approximate numbers of known secondary metabolites
from higher plants [11]
Type of secondary metabolite Estimated numbers*
Nitrogen-containing SM
Alkaloids
Non-protein amino acids (NPAAs)
Amines
Cyanogenic glycosides
Glucosinolates
Alkylamides
Lectins, peptides, polypeptides
SM without nitrogen
Monoterpenes (C10) **
Sesquiterpenes C15)**
Diterpenes (C20)**
Triterpenes, steroids,
saponins (C30, C27)**
Tetraterpenes (C40)**
Flavonoids, tannins
Phenylpropanoids, lignin, coumarins,
lignans
Polyacetylenes, fatty acids, waxes
Polyketides
Carbohydrates, simple acids
*Approximate number of known structures; **total number of all
terpenoides exceeds 22,000 at present.
which tried to feed on plants (herbivores) and with
infectious microorganisms (bacteria, fungi). It is
obvious but still important to note that plants cannot
run away when challenged by a herbivore nor do they
have an elaborate immune system to fight off a
microbial infection. As a common defense measure,
plants and other sessile organisms evolved bioactive
natural products, which repel, deter or poison
herbivores and which can inhibit growth and
development of bacteria, fungi and even viruses.
Some of the defense compounds are constitutive,
while others can be induced under stress conditions:
several SM (so-called phytoalexins) and defense-
related proteins are synthesized de novo when a plant
is challenged by bacteria, fungi or viruses. Because
plants have to compete with other plants for light,
water and nutrients, SM often also serve as mediators
in plant-plant interactions (so-called allelopathy).
During evolution, SM were apparently optimized in
such a way that they did not only exhibit defensive
but also additional non-defense functions: Some SM
have additional physiological and ecological
functions (for example, as nitrogen storage
compounds; UV protectants) or serve as signal
compounds to attract pollinating or seed dispersing
animals and can mediate the interactions between
symbiotic bacteria and their plant hosts (for example,
Rhizobia). Because no defense system is absolute,
a number of specialists have evolved, which have
Wink
overcome the defense barriers. Some insects even use
the defense compounds of their host plants to protect
themselves against predators [8,15-19]. Herbivores
have developed a number of enzymatic means to
21,000
detoxify, inactivate or excrete toxic plant compounds
700
100
(liver enzymes, ABC transporters). Also the
60
microorganisms that are present in the rumen or
100 intestines of herbivores can play a role in inactivating
150 plant defense chemicals.
2,000
The ecological functions of SM become apparent if
2,500 plants are grown in the wild with all their enemies
5,000 and competitors around. Plant breeders have selected
2,500 varieties without toxins or which taste better as food
plants. Consequently, food crops have often lost their
5,000
500 natural resistance to herbivores and/or pathogens, so
5,000 that we have to use synthetic pesticides to protect
them. Using alkaloid-rich and alkaloid-free lupins,
2,000 we have shown that the alkaloids are important for
1,500 lupin plants to protect themselves effectively against
750 a wide diversity of herbivores [13,14,19-21].
400
Origins of plant secondary metabolism
Early in the 20th century it was argued that secondary
metabolites arise either spontaneously or with the aid
of non-specific enzymes. There is good evidence
today that biosynthetic enzymes are highly specific in
most instances and that most have been selected
towards this special task (although they often derive
from common progenitors with a function in primary
metabolism). As a consequence of specific enzymatic
synthesis, final products nearly always have a distinct
stereochemistry. Only the enzymes that are involved
in the degradation of SMs, such as glucosidases,
esterases and other hydrolases, are less substrate-
specific.
Despite the enormous diversity of secondary
metabolites (Table 1), the number of corresponding
basic biosynthetic pathways is restricted and distinct
[7,21,22]. Precursors usually derive from basic
metabolic pathways, such as glycolysis, Krebs cycle
and the shikimate pathway. For pathways leading to
cyanogenic glycosides, glucosinolates, some
alkaloids and non-protein amino acids (NPAAs),
amines, flavonoids and several terpenes, the enzymes
that catalyze individual steps have been identified. In
pathways leading to isoquinoline, indole, pyrrolidine,
pyrrolizidine and tropane alkaloids, flavonoids,
coumarins, NPAAs, mono-, sesqui- and triterpenes,
some of the genes that encode biosynthetic enzymes
have already been isolated and characterized.
Diversity, function and evolution of plant secondary metabolism
Among the unsolved problems of secondary
metabolism is the question of when, where and how
the genes evolved that encode enzymes of SM
biosynthesis, as well as those of transport, storage
and turnover.
Theoretically the following scenarios can be
considered:

Secondary metabolism could be a young
phenomenon and modern plants in the periphery
of the tree of plants have developed their
pathways independently.

Alternatively, secondary metabolism is an old
key innovation, which was developed early in the
evolution of land plants and was inherited by
present-day plants. This would mean that the
corresponding genes should be detectable in at
least most of modern plants.

Plants could have developed the genes of SM
from their own genes of primary metabolism.
Starting with duplication of a gene, the new gene
became mutated, exhibited new metabolic
functions, and was established by natural
selection.

Plants might also have inherited some of the
genes/pathways in early evolution by horizontal
gene transfer from their bacterial symbionts,
which later developed into modern mitochondria
and plastids. It is well-established that bacteria
(especially Actinomyces and Streptomyces) and
cyanobacteria produce a wide diversity of
secondary metabolites, showing similar
structures as plant SM (for example, some
anthraquinones, terpenoids and alkaloids).

About 80% of modern plants live in symbiosis
with fungi (endophytes, ectomycorrhiza); these
fungi could directly have supplied its host with
SM or might have transferred the corresponding
genes to the host’s genome (i.e. by horizontal
gene transfer).
Inference from molecular phylogenies
During the last 20 years botanists have reconstructed
a tree of life for the plant kingdom using nucleotide
sequences of plastid and nuclear marker genes
[22-24]. The resulting molecular phylogenies can be
used to map the distribution of SM in the plant
kingdom [11].
Suggesting that all extant taxa, ranging from spore to
seed producing plants, accumulate a certain class of
SM makes it very likely that all individuals possess
the genetic information for its synthesis and storage.
Natural Product Communications Vol. 3 (8) 2008 1207
Alkaloids Terpenes Flavonoids & other phenolics
Chlorophyta
Charophyta
Marchantiophyta
Anthoceratophyta
Bryophyta
Lycopodiophyta
Sphenophyta
Filicophyta
Angiospermae
Pinophyta
Gnetophyta
Ginkgophyta
Cycadophyta
Figure 1: Molecular phylogeny of land plants.
Clades that produce alkaloids are marked in red; those with terpenoids in
green and those with phenolics in black.
This in turn would suggest that the corresponding
pathway must have evolved during early evolution.
In Figure 1 a molecular phylogeny of major
lineages of land plants is shown. All lineages are
apparently able to synthesize flavonoids and other
phenolics deriving from phenylalanine/tyrosine. This
observation suggests that the genes of flavonoid
biosynthesis must be present from early land plants to
angiosperms. This is apparently true, since the
corresponding key enzymes, such as phenylalanine
ammonia lyase (PAL) and chalcone synthase (CHS)
have been detected in spore forming plants and seed
plants (Figure 2 A,B). Terpenoids, such as mono-,
sesqui-, tri-, and tetraterpenes, and steroids are also
present from mosses to higher plants, suggesting that
the pathways leading to the active isoprene and the
following combinations and cyclisation of C-5 units
were present already in the ancestors of land plants
(Figure 1).
Alkaloids have a more patchy distribution and special
alkaloidal types are usually specific for certain taxon
groups; therefore, they were often used as
chemotaxonomic markers [9,25,26]. Alkaloids, as
such, can be detected in lycopods, horsetails, and
some gymnosperms, but are predominant in
angiosperms (Figure 1). Because many alkaloids are
neurotoxins [27], which interfere with targets of
neuronal signal transduction, we may speculate that
the diversification of alkaloids in angiosperms co-
evolved with the rapid diversification of insect and
vertebrate herbivores during the Cretaceous and
Tertiary periods.
The distribution of alkaloids offers a good example
for understanding the dynamics of SM with restricted
1208 Natural Product Communications Vol. 3 (8) 2008 Wink

100 Daucus carota

73 Petroselinum crispum
Lactuca sativa
PAL 30 100 Helianthus annuus
Catharanthus roseus
O 23 Digitalis lanata
27
100 Agastache rugosa
OH
8 Nicotiana tabacum
NH 2 100 Solanum tuberosum
Lithospermum erythrorhizon
36
62 Rubus idaeus
O
55
100
Glycine max
Astragalus membranaceus
dicots
OH
Pisum sativum
13
48 Prunus avium
Citrus limon
21
37 Quercus suber
99 38 Coffea canephora
Arabidopsis thaliana
100 Isatis tinctoria
100
Beta vulgaris
46 100 Stellaria longipes
Allium cepa
Oryza sativa
monocots
100
99 Triticum aestivum
70 Zea mays
100 Saccharum officinarum
Pinus pinaster
100
Selaginella kraussiana
42
Lycopodium tristachyum
lycopods
89
98 Equisetum arvense horsetail
100 Blechnum spicant ferns
Aspergillus fumigatus
Rhodotorula mucilaginosa
fungi
100
74 Ustilago maydis
100 Amanita muscaria
100 Coprinopsis cinerea
Anabaena variabilis
bacteria
100
Nostoc punctiforme
Streptomyces maritimus
87 Photorhabdus luminescens

0.1

2B
OH
4-Coumaroyl-CoA HO OH

+
3 Malonyl-CoA CHS OH O

Solanum lycopersicum
Nicotiana tabacum
Pyrus communis
Vitis vinifera
Petroselinum hortense
Daucus carota
plants
Arabidopsis thaliana
100 Barbarea vulgaris
100 Allium cepa
Medicago sativa
62 Physcomitrella patens
Phaeosphaeria nodorum
Aspergillus oryzae
100

100
Coccidioides immitis
Magnaporthe grisea
fungi
Neurospora crassa
Cytophaga hutchinsonii

51
Arthrobacter
Rhodobacterales bacterium cyanobacteria
Prochlorococcus marinus
100 Synechococcus sp. W H 7805
Pseudomonas fluorescens
100 Streptomyces griseus
100 Stigmatella aurantiaca
93 Anaeromyxobacter dehalogenans
Deinococcus radiodurans
99 Rhodospirillum centenum

87
99 Bacillus subtilis bacteria
Bacillus licheniformis
99 Geobacillus kaustophilus
Bacillus halodurans
97 Bacillus clausii

0.1
Figure 2: Molecular phylogeny of PAL (A) and CHS (B) inferred from derived amino acid sequences of the corresponding genes (after [46]).
Diversity, function and evolution of plant secondary metabolism Natural Product Communications Vol. 3 (8) 2008 1209

occurrences. In Figure 3, a more detailed analysis of production, we can often detect the same trait in other
the distribution of two major classes of alkaloids in unrelated members of the plant kingdom. This
the angiosperms is illustrated as an example. The tree observation indicates that chemotaxonomy, which
is based on several molecular markers and the basis relies on the similarity of SM profiles, must be
for a reorganization of angiosperm phylogeny [24]. regarded with some caution. If we were to assemble
Quinolizidine alkaloids, such as lupanine and all plants with similar SM in one clade, we should
cytisine, are mainly found within the genistoid not obtain a meaningful phylogeny in most instances
legumes (order Fabales); isolated occurrences have [11,25]. Because reliable molecular phylogenies are
been reported from other plant families, which are available by now for many plant groups,
not directly related to Fabaceae [28,29]. A similar chemotaxonomic groupings need to be reanalyzed
picture can be seen for pyrrolizidine alkaloids (PAs), and reinterpreted.
such as senecionine and symphytine, which are the
main SM in Boraginaceae and in part of the Inference from genome data mining
Asteraceae (Senecioninae). PAs also occur in several
A large number of bacterial, animal and plant
unrelated monocot and dicot families.
genomes have been partly or completely sequenced,
which provides us with the chance to test our
The patchy occurrence of QA and PA within the
assumptions explained in the last paragraphs by a
angiosperms continues if we look into finer details in
data mining approach (as shown for PAL and CHS in
the Fabaceae (Figure 4). Clades with PA production
Figure 2).
(genera Crotalaria and partially Lotononis) cluster
within QA producing clades. In the genus Lotononis
Here we will concentrate on two genes of indole and
two clades are apparent [30], one producing QA the
isoquinoline alkaloid biosynthesis: Strictosidine
other PA.
synthase (STS) catalyzes the fusion of strictosidine
and tryptamine, a key step in the biosynthesis of
How can we interpret the patchy distribution of PA
monoterpene indole alkaloids [26]. These alkaloids
and QA in angiosperms and in the Fabaceae? Their
are abundant in four plant families – the
occurrence in unrelated groups could be due to
Apocynaceae, Loganiaceae, Gelsemiaceae, and
convergence; in this case we would expect an
Rubiaceae. Therefore, we would expect that STS
independent set of biosynthetic genes. In the case of
should be present in these families but not beyond.
the Genisteae s.L. the PA producing taxa are
As can be seen from Figure 5, which is based on
imbedded in clades with QA, which are ancestral.
derived amino acid sequences, related genes/proteins
This observation unambiguously implies that the PA
occur in several other plants (among them
taxa must have inherited the genes of QA
Arabidopsis thaliana), which do not produce such
biosynthesis, but that the QA pathway is down-
alkaloids. This finding indicates that plants must have
regulated in PA producing clades. The PA pathway
evolved a precursor of the STS gene much earlier
could be a new invention or caused by common PA
during their phylogeny. Interestingly, related genes or
genes or by horizontal gene transfer (see below). As a
STS precursor genes, which share many common
consequence, we could assume that the genes for QA
amino acid sites, can be detected among animals and,
and PA biosynthesis evolved at a much earlier stage
more importantly, among bacteria. This observation
in angiosperm phylogeny and that all descendants
could suggest that the progenitor of the STS gene
have a set of the corresponding genes. Whether QA
evolved in protobacteria and was imported into plants
or PA production is actually taking place would
via endosymbiotic protobacteria, which later became
then depend on a differential expression of the
mitochondria. The modern mitochondrial genomes
corresponding genes (which in general is a common
are quite small, suggesting that most of the original
theme in plants). Because the genes of QA
genes have been transferred to the nucleus of an
biosynthesis are not known we have presently no
ancestral eucaryotic or plant cell. Thus plant STS
possibility to test our assumption.
genes would be the result of an ancient horizontal
gene transfer.
The QA/PA example is also representative of many
other SM (cardiac glycosides, iridoid glycosides, As a second enzyme we have analysed the berberine
sesquiterpene lactones, diterpenoids, many alkaloids) bridge enzyme (BBE); it catalyses a specialized step
with more specific and restricted distribution in protoberberine alkaloid biosynthesis (Kutchan,
patterns: In addition to one or more centers of SM
1210 Natural Product Communications Vol. 3 (8) 2008 Wink

Figure 3: Phylogeny of angiosperms and the distribution of quinolizidine (QA) and pyrrolizidine alkaloids (PA) (after [46])
Diversity, function and evolution of plant secondary metabolism Natural Product Communications Vol. 3 (8) 2008 1211

NJ
rbcL Argyrolobium lanceolatum 1
Argyrolobium lunaris
Argyrolobium lotoides
Goodia lotifolia
Argyrolobium polyphyllum
Argyrolobium marginatum
Polhillia obsoleta 36
Polhillia brevicalyx 30
Polhillia canensis 31
Calicotome villosa I
Cytisus hirsutus LG19 H
N
Chamaecytisus purpureus
Cytisus villosus
Cytisus scoparius
Laburnum anagyroides
Genista florida
Genista januensis
Genista tinctoria
Petteria ramentacea
Teline monspessulana AS67
Teline osyroides

N
Ulex europaeus
Ulex parviflorus
Retama monosperma

H
Retama raetam
Retama sphaerocarpa
Genista germanica
Genista radiata
Spartium junceum I
Lupinus albus I
Lupinus angustifolius I
Lupinus arboreus

QAs
Lupinus arcticus
Lupinus polyphyllus
Lupinus atlanticus
Echinospartum horridum C38
Polhillia pallens 10
Dichilus pilosus 24
Dichilus strictus 26
Dichilus reflexus 25
Dichilus lebeckioides 40
Dichilus gracilis 6
Melolobium candicans 18a
Melolobium adenodes 12a
Aspalathus pendula 39
Aspalathus linearis 42
Aspalathus nivea 12
Aspalathus cephalotes
Lebeckia inflata 13
Lebeckia pauciflora 28
Lebeckia bowieana 80
Lebeckia bowieana 127
Lebeckia pluckenetiana 29
Wiborgia tetraptera 46
Wiborgia obcordata 47
Wiborgia fusca 11

Lebeckia
Lebeckia carnosa 7
Lebeckia meyeriana 33
Lebeckia meyeriana 118
Rafnia perfoliata 20a
Rafnia amplexicaulis W40
Lebeckia meyeriana 11a
Lebeckia wrighttii 81
Lebeckia wrighttii 59
Lebeckia wrighttii 128
Lebeckia wrighttii 129
Lebeckia simsiana 98
Lebeckia ambigna 99
Lebeckia leipoldtia 58
Lebeckia leipoldtia 130
Lotononis curvicarpa 14
Lotononis calycina 9
Lotononis foliosa 67
Lotononis eriantha 109
Lotononis adpresse 93
Lotononis decumbens 73

Lotononis
Lotononis platicarpa 53
Lotononis platicarpa 96
Lotononis bentamiana 8
Lotononis digitata 77
Lotononis magnifica 112
Lotononis magnifica 92
Lotononis mollis 52
Lotononis mollis 63
Lotononis plicata 66
Lotononis polycephala 114
Lotononis digitata 65
Lotononis globulosa 69
Lotononis densa ssp leu 15
Lotononis exstipulata 75
Lotononis filiformis 68
Lotononis elongata 108
Lotononis pulchella 51
Lotononis carnosa 104
Lotononis pulchella 115
Lotononis meyeri 123

OH
Lotononis stricta 117
Lotononis meyeri 113
Lotononis eriocarpa 105
Lotononis lotononoides 111
Lotononis prostrata 18
Lotononis involucrata 70
Lotononis
Lotononis leptoloba 71
Lotononis maximiliani 91
Lotononis maximiliani 126
Lotononis pungens 124
O
Lotononis brevicaulis 94
Lotononis sericophylla 19
Lotononis divaricata 121
Lotononis sericophilla 122

O
Lotononis sericophilla 125
Lotononis umbellata 55
Lotononis falcata 16

O
Lotononis laxa 106
Lotononis macrosepala 76

O
Lotononis laxa 72
Lotononis crumania 54
Lotononis sabulosa 95
Lotononis parviflora 17
Lotononis fruticoides 107
Lotononis lenticula 110
Lotononis sparsiflora 116
Lebeckia cytisioides 27
Lebeckia cytisioides 34a
Lebeckia sericea 102
Lebeckia multiflora 34
Lebeckia multiflora 119

N
Lebeckia macrocantha 103
Lebeckia pungens 101
Lebeckia melilotoides 100
Pearsonia aristata 20
Pearsonia aristata 17a
Pearsonia grandifolia 21
Crotolaria distans 48
Crotalaria cunninghami AS26
Crotalaria laburnifolia 131

Crotalaria
PAs
Crotalaria capensis
Crotalaria pumila F27
Crotalariea eremeae AS52
Crotalaria pallida
Bolusia amboensis 23
Bolusia amboensis 22 Bolusia
Bolusia capensis 35
Lotononis hirsuta 16a
Lotononis hirsuta 64
Lotononis hirsuta 97
Lot. hirsuta
Bolusanthus speciosus 26a
Virgilia oroboides 22a
Virgilia divaricata 8a
Virgilia divaricata
Podalyria calyptrata 31a
Liparia splendens
Cyclopia genistoides
Calpurnia sericea W19
Sophora davidii
Sophora jaubertii
Sophora flavescens
Anagyris foetida
Piptanthus nepalensis
Baptisia leucantha F5
Thermopsis lupinoides F79
Baptisia tinctoria
Thermopsis rhombifolia LG42
Maackia amurensis
Hovea elliptica
Lotus corniculatus 84
Lotus corniculatus 87
Chordospartium stevensonii F20
Swainsonia lessertifora AS65
Vicia ervilia AS11
Erythrina lysistimon 15a
Erythrina caffra LG
Erythrina falcata F35
Psoralea aphylla 28a
Vigna unguiculata
Daviesia brevifolia AS28
Daviesia ulicifolia AS61
Goodia lotifolia Belair AS47
Platylobium obtusangelicum AS62
Ormosia formosa 78
Ormosia hengchuniana 79
Ormosia formosana M20
Castanospermum australe
Sophora secundiflora I
Dichrostachis cinerea 7a
Acagia saligna 37a
Schotia afra 4a
Polygala chamaebuxus I
Polygala amara

0.001 substitutions/site

Figure 4: A molecular phylogeny of part of the Fabaceae, based on nucleotide sequences of the rbcL gene (after [30,46]) and distribution of QA and PA
producing taxa.
1212 Natural Product Communications Vol. 3 (8) 2008 Wink

NH2
NH
H O
N
H O N O H3CO H3CO
Glc H Glc
O O O O N N
HO HO
STS
O O OH OH

OCH3 BBE
OCH3

STS 100
Arabidopsis thaliana NP 563818
Arabidopsis thaliana AAF22901
76 Oryza sativa AAR87254
100 Helianthus annuus
93 Lycopersicon esculentum
Brassica napus plants BBE Lactuca sativa
Medicago truncatula
Zea mays Vigna unguiculata
100 100
100 Oryza sativa NP 001049635 54 Cannabis sativa
87 Triticum aestivum Daucus carota
Medicago truncatula Arabidopsis thaliana
97
Ophiorrhiza pumila Glycine max
96 Cynodon dactylon
100 Rauvolfia serpentina
100 Phleum pratense plants
100 Rauvolfia mannii 100 Lolium perenne
100 100
Drosophila melanogaster AAC47118 100 Triticum aestivum
100 Drosophila pseudoobscura Hordeum vulgare
Anopheles gambiae Secale cereale
Xenopus tropicalis Oryza sativa
98
99 Berberis stolonifera
Danio rerio
86
100
Gallus gallus animals 86
100
Thalictrum flavum
Eschscholzia californica
Rattus norvegicus 100 Papaver somniferum
100
Bos taurus 100 Bacillus thuringiensis
100 Canis familiaris 86 Bacillus cereus
87
Homo sapiens BSCv
0.1
76 Clostridium beijerincki bacteria
96 Streptomyces avermitilis
100 Pan troglodytes
Ustilago maydis
100 Mesorhizobium loti 100 Aspergillus fumigatus
Bradyrhizobium japonicum
fungi
81 Aspergillus oryzae
100 Pseudomonas mendocina Phaeosphaeria nodorum
91
100
bacteria
94 Pseudomonas aeruginosa 100 Neurospora crassa
Marinobacter aquaeolei 90 Magnaporthe grisea

81 Solibacter usitatus
Alcanivorax borkumensis
51
84 Deinococcus geothermalis
0.1
Figure 5: Phylogeny of strictosidine synthase (STS) inferred from
derived amino acid sequences (after [46]).
Taxa that produce monoterpene alkaloids are marked by an arrow and
printed in bold.
1995). These alkaloids mainly occur in
Ranunculaceae, Papaveraceae, Menispermaceae,
Berberidaceae, Annonaceae and several other plant
families. As can be seen from Figure 6, BBE genes
and proteins are, however, widely present in higher
plants, independent of whether a taxon actually
makes these alkaloids or not, indicating common
ancestry. Interestingly, BBE or similar proteins,
which share a number of common conserved sites
could also be found in fungi and bacteria. This
finding implies again that the progenitor of the BBE
gene evolved in protobacteria and was imported into
plants and fungi via the endosymbiosis, which had
led to mitochondria. Thus plant BBE would also be
the result of an ancient horizontal gene transfer,
similar to the situation in STS.
These few examples support the assumption of a
common distribution of SM genes in plants based on
ancestral horizontal gene transfer. We do not know
Figure 6: Phylogeny of berberine bridge enzyme (BBE) inferred from
derived amino acid sequences (after [46]). Taxa that produce
protoberberine alkaloids are marked by an arrow and printed in bold.
however, whether the genes are functional in species
that do not express the corresponding pathway. These
genes could be either silent copies or inactive
pseudogenes; this possibility could be tested by
expressing the genes in a heterologous system and
then to evaluate their biological activity.
A number of SM (for example, many terpenoids,
quinolizidine alkaloids, the piperidine alkaloid
coniine) are produced completely or partly in
chloroplasts and/or mitochondria [21,31-33]. The
corresponding genes are mostly nuclear. It is
tempting to speculate that these localisations are
indirect indicators of a former bacterial origin of the
corresponding pathways. The PAL, CHS, STS and
BBE examples indicate that plants have probably
obtained the corresponding gene copies through early
horizontal gene transfer from either protobacteria or
cyanobacteria that gave rise to mitochondria and
chloroplasts, respectively. It is likely that plants have
also contributed some of their own genes to further
modify the base structure of alkaloids, terpenoids and
flavonoids that would produce the variation observed
within most classes of SM.
Diversity, function and evolution of plant secondary metabolism
Contribution from endophytes and
ectomycorrhizal fungi
Mycorrhizal associations have apparently existed
since the advent of vascular plants about 400 million
years ago. It has been estimated that 80% of plants
live in permanent contact with fungi [34-36]. Many
of them are ecto- and endomycorrhiza that live in
close contact with roots. The fungi substantially
enlarge the root surface area and thus increase the
possibility to catch more water and nutrients. The
role and benefits of endophytes is less obvious.
Some of these fungi apparently are also able to
produce SM and contribute to the chemical defense
of their host plants. An interesting example of a
symbiotic fungus-plant interaction can be seen in
ergot alkaloids. These alkaloids are mainly
produced by saprophytic fungi: Claviceps purpurea,
C. microcephala, C. paspali (Clavicipitaceae) and
more than 40 further members of this genus live as
symbionts on grasses (tribes Festucaceae, Hordeae,
Avenae, Agrosteae). Rye is especially affected
among cereals. Claviceps is not a parasite, but
obviously a symbiotic organism. It takes nutrients
from its host but provides chemical defense against
herbivores as compensation (the ergot alkaloids are
powerful neurotoxins). Field experiments have
shown that such a fungal infection is an ecological
advantage for grasses in the wild. Related symbiotic
fungi such as Epichloe, Balansia and Myriogenspora
(Clavicipitaceae) infect grasses and also produce
alkaloids. Ergot alkaloids (such as agroclavine,
chanoclavine, ergine, ergosine, and ergometrine) are
also common SM of some genera of the
Convolvulaceae (including Argyreia, Ipomoea,
Turbina corymbosa, Stictocardia tiliafolia). It has
been shown recently, that the ergot alkaloid
formation in the Convolvulaceae is due to an
endophytic clavicipateceous fungus that lives
together with certain species in this plant family [37].
In this case, the isolated occurrence of ergot alkaloids
in Convolvulaceae is due to a symbiotic relationship.
Certain simple pyrrolizidine alkaloids, such as loline,
were detected in the grass (Lolium pratense, ex.
Festuca pratensis) and in a root hemiparasitic plant
(Rhinanthus serotinus). The alkaloids derived from a
symbiotic endophytic fungus (Neotyphodium
uncinatum; Clavicipitaceae), which lives on the grass
[38]. Similar to the situation in Claviceps, the fungus
provides defense compounds, which help the grass
and its hemiparasite to ward off herbivores.
Natural Product Communications Vol. 3 (8) 2008 1213
The legume genera Astragalus and Oxytropis are
famous for the production of toxic indolizidine
alkaloids, such as swainsonine, causing locoweed
poisoning in livestock. These alkaloids are apparently
produced by an endophyte (Embellisia spp.;
Pleosporaceae) [39].
The maytansinoid ansa antibiotics are produced by
Actinosynnema pretiosum (Actinomycetes), but also
in a number of angiosperms, including Celastraceae,
Rhamnaceae, and Euphorbiaceae. It has been
suggested that the occurrence of these SM depends
on infection by this actinomycete [40].
Naphthodianthrones, such as hypericin are well-
known constituents of St. John’s wort (Hypericum
perforatum; Clusiaceae). Kusari et al. [41] have
isolated an endophytic fungus from H. perforatum
that produces hypericin in culture. Also emodin, a
precursor of hypericin is produced by both endophyte
and the plant. Fungi are known producers of
anthraquinones; in lichens, which are symbioses
between an alga and a fungus, anthraquinones
represent common defense metabolites, being
probably produced by the fungal partner. Since
anthraquinone producing plants are isolated over the
plant kingdom (Asphodelaceae, Fabaceae,
Rhamnaceae, Polygonaceae) it would be interesting
to search for a former or extant endophyte
association.
Camptothecin (structurally a quinoline alkaloid, but
derived from the tryptamine/secologanin pathway)
occurs in unrelated families such as Camptotheca
acuminata (Cornaceae), Nothapodytes foetida,
Pyrenacantha klaineana, Merrilliodendron
megacrapum (Icacinaceae), Ophiorrhiza pumila,
O. mungos (Rubiaceae), Ervatamia heyneana
(Apocynaceae) and Mostuea brunonis
(Gelsemiaceae). It has been shown recently, that
camptothecin can be produced by endophytes from
Nothapodytes foetida [42]. It is tempting to speculate
that the patchy distribution was originally caused by
endophytes, which have either infected the respective
plants or which have transferred their genes.
The anticancer drug taxol is produced by endophytes
from Taxus brevifolia [43]. An isolated occurrence of
the taxan alkaloids has been reported from Corylus
avellana (Betulaceae), which is not associated with
an endophyte (although the pathway genes might
haven imported by an endophyte at an earlier stage)
[44]. Podophyllotoxin is produced by endophytes of
1214 Natural Product Communications Vol. 3 (8) 2008
EARLY EVOLUTION protobacteria
Symbiotic fungi
DNA
-endophytes
-ectomycorrhiza
SM
Figure 7: Hypothetical scheme for the evolution of secondary metabolism in plants.
Podophyllum peltatum [45]. Although the
endophytes, which were isolated from these plants,
are capable of the biosynthesis of hypericin, taxol,
camptothecin and podophyllotoxin in vitro, it is less
likely that they alone perform the productions in the
plant. It has been speculated that a horizontal gene
transfer has taken place at some stage, thus importing
the respective pathways from the fungi into the host
plant [41]. It is a challenging question to determine
the degree and contribution of endophytic and
ectophytic SM pathways to the SM profiles of plants.
If it were to be a more common phenomenon than
usually assumed, it would offer an additional
explanation for the patchy distribution of certain SM
in the plant kingdom.
It has been speculated that horizontal gene transfer
could also be taking place when viruses or insects
invade plants: Viruses could integrate some SM
genes of a certain host plant into their genomes; if
they are transferred to a different host plant species,
they might transfer the corresponding genes. Insects
often harbour endosymbiotic bacteria, which may
carry genes of SM biosynthesis. Such bacteria could
be transferred between species by sucking insects.
Conclusions and Outlook: One of the main
questions discussed in this communication concerns
the origin and evolution of plant secondary
metabolism. We have started with the observation
that some SM (such as phenolics and terpenoids) are
Wink
cyanobacteria
mt cp
DNA DNA
Genes
Duplication
Mutation
genome Selection
Specialization
mRNA
protein
SM SM
SM SM
produced by nearly all vascular plants, whereas
others, especially those with, for example, alkaloids,
cardiac glycosides, and anthraquinones, show a more
restricted but usually patchy distribution.
The patchy distribution could be due to convergence,
which is certainly the case in several instances
(Figure 7). It could also be due to a wider distribution
of SM pathway genes in the plant kingdom, which
are silent or inactivated in most places, but become
activated under certain conditions or in particular
clades. Evidence is provided from the distribution of
a few key genes/proteins that SM pathways might
have been introduced into plants from SM producing
bacteria via an early horizontal gene transfer; it is
established that protobacteria became mitochondria
and cyanobacteria plastids. Another external source
for plant SM could be ectomycorrhizal fungi and
endophytes, which either directly produce a particular
SM or indirectly by transferring the pathway genes
from fungi to plants. Because fungal infections do not
necessarily follow plant phylogeny, they could cause
(at least partly) the patchy distribution seen in some
SM groups.
It is likely that horizontal gene transfer only
introduced a limited number of pathway genes and
that the host plants developed and contributed their
own set of genes/enzymes, leading to the various
structural variations seen in nature as a sort of
biotransformation reaction.
Diversity, function and evolution of plant secondary metabolism Natural Product Communications Vol. 3 (8) 2008 1215

Acknowledgements - I would like to thank various
coworkers and collaborators who have contributed
to the results in the past (E. Käss, M. Kaufmann,
B. Gemeinholzer, G. Mohamed, T. Morazova,
F. Botschen, C. Gosmann, H. Schäfer, F. Sporer,
H. Sauer-Gürth, H. Staudter, B.-E. van Wyk,
J. Boatwright). Ted Cole kindly allowed me to use
an Indesign file of an APG II tree. Peter Waterman
provided helpful comments on the manuscript.

References
[1] Luckner M. (1990) Secondary Metabolism in Microorganisms, Plants and Animals. Springer, Heidelberg.
[2] Conn EE. (1981) Secondary Plant Products, in The Biochemistry of Plants. Vol. 7, Academic Press, New York.
[3] Bell EA, Charlwood BV (1980) Secondary Plant Products. Springer, Heidelberg.
[4] Mothes K, Schütte HR, Luckner M. (1985) Biochemistry of Alkaloids. Verlag Chemie, Weinheim.
[5] Dey PM, Harborne JB. (1997) Plant Biochemistry. Academic Press, San Diego.
[6] Seigler DS. (1998) Plant Secondary Metabolism. Kluwer Academic Boston, Dordrecht, London, 1-759
[7] Dewick PM. (2002) Medicinal Natural Products. A Biosynthetic Approach. Wiley, New York, 1-507
[8] Harborne JB (1993) Introduction to Ecological Biochemistry. 4th Ed.. Academic Press, London.
[9] Harborne JB, Baxter H. (1993) Phytochemical Dictionary. A Handbook of Bioactive Compounds from Plants. Taylor & Francis,
London; Harborne JB, Turner BL. (1984) Plant Chemosystematics. Academic Press, London.
[10] DNP (1996) Dictionary of Natural Products on CD-Rom version 5:1, Chapman & Hall, London.
[11] Wink M. (2003) Evolution of secondary metabolites from an ecological and molecular phylogenetic perspective. Phytochemistry
64, 3-19.
[12] Stahl E. (1888) Pflanzen und Schnecken. Jenaische Zeitschrift für Naturwissenschaften, 22, 557-684.
[13] Wink M. (1988) Plant breeding: Importance of plant secondary metabolites for protection against pathogens and herbivores.
Theoretical and Applied Genetics, 75, 225-233.
[14] Wink M. (1999) Function of plant secondary metabolites and their exploitation in biotechnology. Annual Plant Reviews Vol. 3,
Sheffield Academic Press, Sheffield.
[15] Levin DA. (1976) The chemical defences of plants to pathogens and herbivores. Annual Review of Ecology and Systematics, 7,
121-159.
[16] Swain T. (1977) Secondary compounds as protective agents. Annual Review of Plant Physiology, 28, 479-501.
[17] Rosenthal GA, Berenbaum MR. (1991) Herbivores: Their Interactions with Secondary Plant Metabolites. Vol. 1, The chemical
participants; Academic press, San Diego; Rosenthal GA, Berenbaum MR (1992): Herbivores: Their interactions with secondary
plant metabolites. Vol. 2. Ecological and evolutionary processes. Academic press, San Diego.
[18] Bernays EA, Chapman RF. (1994) Host-Plant Selection by Phytophagous Insects. Chapman and Hall, London.
[19] Wink M. (1992) The role of quinolizidine alkaloids in plant insect interactions. In Insect- Plant Interactions. Bernays EA. (Ed.),
Vol. IV, 133-169, CRC-Press, Boca Raton.
[20] Wink M. (1985) Chemische verteidigung der lupinen: zur biologischen bedeutung der chinolizidinalkaloide. Plant Systematics and
Evolution, 150, 65-81.
[21] Wink M. (1999) Biochemistry of Plant Secondary Metabolism. Annual Plant Reviews, Vol. 2, Sheffield Academic Press, Sheffield.
[22] Soltis P, Soltis DE, Doyle JJ. (1992) Molecular Systematics of Plants. Chapman & Hall, London.
[23] Doyle JJ. (1993) DNA, phylogeny, and the flowering of plant systematics. Bio Science, 43, 380-389.
[24] APG II (2003) An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG
II. Botanical Journal of the Linnean Society, 141, 399-436.
[25] Wink M, Waterman P. (1999) Chemotaxonomy in relation to molecular phylogeny of plants. In Biochemistry of Plant Secondary
Metabolism. Wink M. (Ed.), Sheffield Academic Press, Annual Plant Reviews, Vol. 2, pp 300-341.
[26] Roberts MF, Wink M. (1998) Alkaloids-Biochemistry, Ecological Functions and Medical Applications. Plenum, New York.;
Kutchan TM (1995) Alkaloid biosynthesis: the basis for metabolic engineering of medicinal plants. The Plant Cell, 7, 1959–1970.
[27] Wink M. (2000) Interference of alkaloids with neuroreceptors and ion channels. In Bioactive Natural Products. Vol. 11, Atta-ur-
Rahman (Ed.), Elsevier. pp. 3-129.
[28] Wink M, Witte L. (1983) Evidence for a wide spread occurrence of the genes of quinolizidine alkaloid biosynthesis. Induction of
alkaloid accumulation in cell suspension cultures of alkaloid-"free" species. FEBS letters, 159, 196-200.
[29] Wink M, Mohamed GIA. (2003) Evolution of chemical defence traits in the Leguminosae: Mapping of distribution patterns of
secondary metabolites on a molecular phylogeny inferred from nucleotide sequences of the rbcL gene. Biochemical Systematics
and Ecology, 31, 897-917.
1216 Natural Product Communications Vol. 3 (8) 2008 Wink

[30] Boatwright JS, le Roux MM, Wink M, Morozova T, van Wyk B-E. (2008) Systematics of the tribe Crotalarieae (Fabaceae).
American Journal of Botany (submitted)
[31] Wink C, Hartmann T (1981) Properties and subcellular localisation of L-alanine: aldehyde aminotransferase. Concept of an
ubiquitous plant enzyme involved in secondary metabolism. Zeitschrift für Naturforschung, 36c, 625-632.
[32] Wink M, Hartmann T. (1982) Localization of the enzymes of quinolizidine alkaloid biosynthesis in leaf chloroplast of Lupinus
polyphyllus. Plant Physiology, 70, 74-77.
[33] Roberts MF. (1981) Enzymic synthesis of coniceine in Conium maculatum chloroplasts and mitochondria. Plant Cell Reports, 1,
10-13.
[34] Remy W, Taylor TN, Hass H, Kerp H (1994) Four hundred-million-year-old vesicular arbuscular mycorrhizae. Proceedings
National Academy of Sciences, USA, 91, 11841-11843.
[35] Simon L, Bousquet J, Levesque C, Lalonde M. (1993) Origin and diversification of endomycorrhizal fungi and coinidence with
vascular land plants. Nature, 263, 67-69.
[36] Sawers RJH, Gutjahr C, Paszkowski U. (2008) Cereal mycorrhiza: an ancient symbiosis in modern agriculture. Trends Plant
Science, 13, 93-97.
[37] Ahimsa-Müller MA, Markert A, Hellwig S, Knoop V, Steiner U, Drewke C, Leistner E. (2007) Clavicipitaceous fungi associated
with ergoline alkaloid-containing Convolvulaceae. Journal of Natural Products, 70, 1955-1960.
[38] Lehtonen P, Helander M, Wink M, Sporer F, Saikkonen K. (2005) Transfer of endophyte origin defensive alkaloids from a grass to
hemiparasitic plant. Ecology letters, 8, 1256-1263.
[39] Ralphs MH, Creamer R, Baucom D, Gardner DR, Welsh SL, Graham JD, Hart C, Cook D, Stegelmeier BL. (2008) Relationship
between the endophyte Embellisia spp. and the toxic alkalod swainsonine in major locoweed species (Astragalus and Oxytropis).
Journal of Chemical Ecology, 34, 32-38.
[40] Cassady JM, Chan KK, Floss H, Leistner E. (2004) Recent developments in the maytansinoid antitumor agents. Chemistry,
Pharmacy Bulletin, 52, 1-26.
[41] Kusari S, Lamshöft M, Zühlke S, Spiteller M. (2008) An endophytic fungus from Hypericum perforatum that produces hypericin.
Journal of Natural Products, 71, 159-162.
[42] Puri SC, Verma V, Amna T, Qazi GN, Spiteller M. (2005) An endophytic fungus from Nothapodytes foetida that produces
camptothecin. Journal of Natural Products, 68, 1717-1719.
[43] Stierle A, Strobel G, Stierle D. (1993) Taxol and taxane production by Taxomyces andreanae, an endophytic fungus of Pacific yew
Science, 260, 214-216.
[44] Ottagggio L, Bestoso F, Amirotti A, Balbi A, Damonte G, Mazzei M, Sancandi M, Miele M. (2008) Taxanes from shells and leaves
of Corylus avellana. Journal of Natural Products, 71, 58-60.
[45] Eyberger AL, Dondapati R, Porter JRJ. (2006) Endophyte fungal isolates from Podophyllum peltatum produce podophyllotoxin.
Journal of Natural Products, 69, 1121-1124.
[46] Wink M, Botschen F, Gosmann C, Schäfer H, Waterman PG. (2008) Evolution and origin of plant secondary metabolism-
Implications for chemotaxonomy. In Wink M. Biochemistry of plant secondary metabolism. Blackwell-Wiley-VCH (in press).