Metabolism of N-alkyldiamines and N-alkylnortropinones by transformed root cultures of Nicotiana and Brugmansia

Article abstract of DOI:10.1016/S0031-9422(99)00292-7

A range of analogues of N-methylputrescine and tropinone were fed to transformed root cultures of Nicotiana rustica and/or a Brugmansia candida x aurea hybrid. These cultures were made by the transformation of the relevant plant species with Agrobacterium rhizogenes. A number of the metabolites, notably those showing a relatively modest alteration in the N-alkyl substituent, were metabolized in vivo to form homologues of the normal alkaloids biosynthesized by these roots. These products were identified by GC/MS and comparison with some synthetic reference materials. Analogues with major alterations in the size of the N-alkyl substituent were not metabolized at all. In the N. rustica cultures, the analogues fed at 1 mM significantly affected the profile of normal alkaloids, with up to a 4-fold diminution in nicotine being found in the presence of N-n-propylputrescine. The ratio between alkaloids of the pyrrolidine series and the piperideine series was also affected. In contrast, the presence of the analogues in the B. candida x aurea hybrid culture at 1 mM did not inhibit or substantially interfere with the accumulation of the normal spectrum of alkaloids. The potential for using these cultures to make complex novel products from simple precursors is discussed.

Full text of DOI:10.1016/S0031-9422(99)00292-7

Phytochemistry 52 (1999) 855±869
Metabolism of N-alkyldiamines and N-alkylnortropinones by
transformed root cultures of Nicotiana and Brugmansiap
Henry D. Boswella, Birgit Dragerc, 1, John Eaglesd, Carol McClintocka, Adrian Parrd,
Andreas Portste€enb, David J. Robinsa, Richard J. Robinsd, e, Nicholas J. Waltond,*,
Chi Wongf
aChemistry Department, The University, Glasgow G12 8QQ, UK
bInstitut fur Pharmazeutische Biologie und Phytochimie, Westfalische Wilhelms-Universitat, Hittorfstrasse 56, D-48149 Munster, Germany
È
È
È
È
cInstitut fur Pharmazeutische Biologie der Martin-Luther-Universitat Halle-Wittenberg, Hoher Weg 8, D-06120 Halle, Germany
È
È
dInstitute of Food Research, Norwich Laboratory, Norwich Research Park, Colney, Norwich NR4 7UA, UK
e
Â
Á
Laboratoire d'Analyse Isotopique et Electrochimique de Metabolismes, CNRS ESA 6006, 2 rue de la Houssiniere, F-44322 Nantes cedex 03, France
fDepartment of Chemistry, University of Durham, Science Laboratories, South Road, Durham DH1 31E, UK
Received 8 February 1999; received in revised form 12 May 1999; accepted 12 May 1999
Abstract
A range of analogues of N-methylputrescine and tropinone were fed to transformed root cultures of Nicotiana rustica and/or
a Brugmansia candida  aurea hybrid. These cultures were made by the transformation of the relevant plant species with
Agrobacterium rhizogenes. A number of the metabolites, notably those showing a relatively modest alteration in the N-alkyl
substituent, were metabolized in vivo to form homologues of the normal alkaloids biosynthesized by these roots. These products
were identi®ed by GC/MS and comparison with some synthetic reference materials. Analogues with major alterations in the size
of the N-alkyl substituent were not metabolized at all. In the N. rustica cultures, the analogues fed at 1 mM signi®cantly a€ected
the pro®le of normal alkaloids, with up to a 4-fold diminution in nicotine being found in the presence of N-n-propylputrescine.
The ratio between alkaloids of the pyrrolidine series and the piperideine series was also a€ected. In contrast, the presence of the
analogues in the B. candida  aurea hybrid culture at 1 mM did not inhibit or substantially interfere with the accumulation of
the normal spectrum of alkaloids. The potential for using these cultures to make complex novel products from simple precursors
is discussed. # 1999 Elsevier Science Ltd. All rights reserved.
Keywords: Alkaloid (analogue); Biotransformation; Brugmansia candida  aurea; Solanaceae; Hybrid; Nicotiana rustica; Transformed root culture
1. Introduction
(Fig. 1) from L-arginine or L-ornithine, via putrescine
1 and N-methylputrescine 2, to N-methylpyrrolinium
(Leete, 1983). In Nicotiana, nicotine 23 is formed by
the direct decarboxylative condensation of N-methyl-
pyrrolinium with nicotinic acid (Leete, 1983).
Anabasine 21, a minor alkaloid in many nicotine-pro-
ducing species, is formed from L-lysine via cadaverine
6 (Leete, 1983). This diamine is directly oxidised,
apparently by the same diamine oxidase as N-methyl-
The pyrrolidine alkaloid, (S )-nicotine 23, and the
tropane alkaloids, (S )-hyoscyamine 18 and (S )-scopo-
lamine 19, are produced by members of the Solanaceae
and share in part a common biosynthetic pathway
p Part 1 in the series `Biosynthesis of novel alkaloids'.
* Corresponding author. Fax: +44-1603-507723.
putrescine,
N-methylputrescine
oxidase
(MPO)
E-mail address: nicholas.walton@bbsrc.ac.uk (N.J. Walton).
1 Previous address: Institut fur Pharmazeutische Biologie und
Phytochimie, Westfalische Wilhelms-Universitat, Hittorfstrasse 56,
D-48149, Munster, Germany.
(Walton & McLauchlan, 1990) and the product, D1-
piperideine 20, condensed with nicotinic acid (Fig. 1).
In the biosynthesis of the tropanes (Fig. 1), metab-
0031-9422/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.
PII: S0031-9422(99)00292-7

856
H.D. Boswell et al. / Phytochemistry 52 (1999) 855±869
olism is now understood to proceed via a formal con-
densation of N-methylpyrrolinium with activated acet-
oacetate (Robins, Abraham, Parr, Eagles & Walton,
1997), followed by conversion to tropinone 9, re-
duction to the alcohol, tropine, and esteri®cation with
a phenyllactoyl residue to produce littorine 17, the tro-
pane precursor of hyoscyamine 18 (Robins & Walton,
1993; Robins et al., 1995). Many species that accumu-
late tropane alkaloids also accumulate the pyrrolidine
alkaloids, hygrine 25 and cuscohygrine 26, also derived
from N-methylpyrrolinium (Fig. 1).
tropinone 9 remain to be determined, although recent
studies have helped to rede®ne the probable intermedi-
ates, notably 4-(1-methyl-2-pyrrolidinyl)-3-oxobutano-
ate 24 (Robins et al., 1997). Recently, the S-
adenosylmethionine-stimulated skeletal rearrangement
of nonnatural (S )-littorine by a cell-free extract of
Datura stramonium has been reported (Ollagnier,
Kervio & Retey, 1998). Whether the same activity will
be observed with the natural isomer, (R )-littorine,
remains to be con®rmed.
It has been known for some years that the pyrroli-
dine alkaloid pathway in Nicotiana will accept and
metabolize nonnatural derivatives of putrescine or nic-
otinic acid (Leete, 1983). Methylated pyrrolinium ions
were incorporated into nicotine analogues by N. gluti-
nosa plants (Rueppel & Rapoport, 1971) and wick-
feeding 5-¯uoronicotinic acid to plants of N. glauca
led to the formation of 5-¯uoronicotine and 5-¯uoroa-
Much of the enzymology of these pathways remains
poorly characterized (Robins, 1998). The mechanism
of the condensation reaction to produce nicotine 23
remains to be elucidated and only a very low activity
in vitro has been reported (Friesen & Leete, 1990).
The enzymological characterization of the steps of the
tropane pathway between N-methylpyrrolinium and

H.D. Boswell et al. / Phytochemistry 52 (1999) 855±869
857
Fig. 1. Schematic pathway for the biosynthesis of various alkaloids in Nicotiana and Brugmansia. Note that (i) arrows do not necessarily imply a
single step and (ii) the exact form in which acetoacetate and phenyllactate are supplied is not known.
nabasine (Leete, 1979; Leete, Bodem & Manuel, 1971).
Similarly, feeding N-methyl-D1-piperideine to N. glauca
caused the accumulation of N-methylanabasine (Leete
& Chedekel 1972), a known natural alkaloid that nor-
mally occurs in only trace amounts (Demole & Berthet
1972). The more distal precursor, cadaverine 6, is
actively biotransformed to anabasine 21 when supplied
to transformed root cultures of Nicotiana spp. (Walton
& Belshaw, 1988; Walton, Robins & Rhodes, 1988),
while its isomer, N-ethylputrescine 3, is readily con-
verted to the nonnatural (S )-N-ethylnornicotine
(Boswell, Watson, Walton & Robins, 1993).
Such studies are of value on two counts. First, they
can demonstrate the degree of plasticity of alkaloid
biosynthetic pathways, illustrating in the absence of
detailed enzymological characterization the extent to
which analogues of the normal substrates might be uti-
lised in successive reaction steps. These data can be
valuable in establishing the substrate speci®city and
giving some mechanistic information about as yet uni-
denti®ed enzymes. Secondly, they enable the potential
for biotransformations to be assessed, using cultures
or extracted enzymes as a means to generate biologi-
cally-active novel compounds from chemically-syn-
thesized nonnatural substrates.
Root cultures of Nicotiana rustica Ð established by
genetic transformation with Agrobacterium rhizogenes
LBA9402 Ð accumulate nicotine 23 as the major al-
kaloid, with minor amounts of nornicotine, anatabine
22 and anabasine 21 (Hamill, Parr, Robins
Rhodes, 1986). Transformed root cultures of
&
a
Brugmansia (ex. Datura) candida  aurea hybrid (line
DB5) accumulate a large range of tropane and pyrro-
lidine alkaloids, the pro®le being dominated by the
aromatic esters of tropine, littorine 17, hyoscyamine
18 and scopolamine 19 (Robins, Parr, Payne, Walton
& Rhodes, 1990). In this paper we report the ability
of these cultures to metabolize nonnatural analogues
of intermediates to form novel alkaloids. As a group
of precursors distal to the ®nal products, lower N-
alkylputrescines and N-alkylcadaverines were syn-
thesized and supplied to growing cultures as potential
substrates. Cultures of the B. candida  aurea hybrid
were also supplied with tropinone analogues in order
to evaluate the relative eciency of supplying distal
or proximal precursors of the desired products. In a
further paper (Boswell et al., 1999), we describe the
in vitro activity of relevant enzymes of the tropane
alkaloid pathway extracted from the B. candi-
da  aurea hybrid culture with these substrates.

858
H.D. Boswell et al. / Phytochemistry 52 (1999) 855±869

H.D. Boswell et al. / Phytochemistry 52 (1999) 855±869
859
Fig. 2. Gas chromatograms of alkaloidal extracts from cultures of a Brugmansia candida  aurea hybrid supplied with (a) no alkylated diamine;
(b) N-ethylputrescine; (c) N-(2-¯uoroethyl)putrescine; (d) N-n-propylputrescine; (e) cadaverine; (f) N-methylcadaverine. Gas chromatography con-
ditions were as described in the Experimental. The following abbreviations are used: Y: hygrine; N: oxohygrine; C: cuscohygrine; R: tropinone;
T: tropine; A: acetyltropine; G: tigloyltropine; L: littorine; H: hyoscyamine; S: scopolamine; these symbols in lower case (y, n, u,. . .) denote ana-
logues of these alkaloids. TIC (a.u.)=total ion current, (arbitrary units).
2. Results
exhibited a mass ion and fragmentation pattern con-
sistent with the mass increment of the N-alkyl sub-
strate relative to N-methylputrescine 2, the precursor
of nicotine 23, or to cadaverine 6, the precursor of
anabasine 21 (Fig. 1). These can be assigned to hom-
ologous series of alkaloids. The extent of conversion
(Table 2) declined with increasing chain length of the
N-alkyl substituent. Thus, N-n-propylputrescine 4 was
actively converted to a novel alkaloid (N-n-propylnor-
2.1. N-alkyldiamine metabolism by Nicotiana rustica
root cultures
N. rustica root cultures supplied with simple N-
alkylputrescines and N-alkylcadaverines at 1 mM gave
rise to new alkaloidal products. These were character-
ized by GC and GC/MS (Table 1). Each of these

860
H.D. Boswell et al. / Phytochemistry 52 (1999) 855±869
Fig. 2 (continued)
nicotine) which accounted for about 33% of the total
alkaloid produced. In contrast, N-n-butylputrescine 5
was relatively poorly metabolized and its alkaloidal
product accounted for only 8% of the total alkaloid.
It should be noted, however, that much the same yield
of product was obtained, the total alkaloid content
being markedly decreased in the presence of N-n-pro-
pylputrescine. No metabolism was detected when N-
benzylputrescine, N-phenylethylputrescine or N-cyclo-
hexylputrescine were supplied.
verine 7 and N-ethylcadaverine 8 were substantially
metabolized (Table 2); conversion of N-n-propylcada-
verine was feeble and there was no detectable product
from N-iso-propylcadaverine, N-n-butylcadaverine or
N-benzylcadaverine.
With N-n-propylputrescine 4 and N-methylcadaver-
ine 7, where the novel alkaloid accounted for 20% or
more of the total alkaloid, nicotine accounted for a
reduced proportion of the total Ð 20% in the case of
N-n-propylputrescine 4 and 33% in the case of N-
In the N-alkylcadaverine series, only N-methylcada-
methylcadaverine 7. N-n-Propylputrescine
4
also

H.D. Boswell et al. / Phytochemistry 52 (1999) 855±869
861
Fig. 2 (continued)
increased the proportion of anatabine 22, though this
2.2. N-alkyldiamine metabolism by Brugmansia
candida  aurea root cultures
was not observed when N-methylcadaverine 7 was sup-
plied. In contrast, N-methylcadaverine-feeding was as-
sociated with a large rise in the proportion of
anabasine 21. Those compounds found not to be
metabolised to novel alkaloids (N-benzylputrescine, N-
phenylethylputrescine, N-cyclohexylputrescine) did not
in¯uence the proportions of the major alkaloids
present.
Corresponding experiments were also performed in
which N-alkyldiamines were fed at 1 mM to B. candi-
da  aurea root cultures and the degree of metab-
olism determined by GC/MS. In this species, these
compounds also represent analogues of the initial
substrates of the pathway but, in contrast, are more

862
H.D. Boswell et al. / Phytochemistry 52 (1999) 855±869
Table 3
GC/MS Fragmentation characteristics for the series of alkaloids identi®ed from transformed root cultures of Brugmansia fed with
N-ethylputrescine
Novel alkaloid
Relative Rta M+ ion
(% abundance) (% abundance)
Characteristic fragments
(A) Pyrrolidine alkaloid
N-Ethylnorhygrine
N-Ethylnoroxohygrine
1.287
1.259
155(1)
171(1)
140(4), 126(2), 112(2), 98(83), 84(8), 82(17), 70(22), 55(15), 43(100)
156(4), 114(3), 110(1), 98(100), 84(5), 82(5), 41(16)
N-Ethyl-N'-methylnorcuscohygrine 1.045
238(<1)
252(<1)
223(2), 140(5), 110(4), 98(50), 84(100), 82(64), 68(33), 55(34), 42(95)
98(100), 84(5), 82(45), 68(43), 55(31), 43(71), 41(88)
N,N'-Diethylnorcuscohygrine
1.085
(B) Tropane alkaloids
N-Ethylnortropinone
N-Ethylnortropine
1.140
1.123
1.096
1.036
1.011
1.009
1.005
153(14)
155(21)
197(11)
237(2)
303(3)
303(4)
317(8)
138(9), 124(11), 110(20), 96(95), 95(48), 82(21), 69(89), 56(80), 42(72), 41(100)
138(37), 126(21), 110(54), 97(100), 96(94), 82(40), 69(76), 56(85), 41(79)
182(4), 154(11), 138(100), 108(42), 96(44), 82(17), 67(43), 56(42), 43(95)
222(4), 194(6), 154(12), 138(76), 110(100), 98(52), 80(62), 68(72), 55(95), 41(91)
154(8), 138(100), 110(10), 108(25), 97(28), 96(31), 67(28), 56(16)
3-Acetyl-N-ethylnortropine
3-Tigloyl-N-ethylnortropine
N-Ethylnorlittorine
N-Ethylnorhyoscyamine
N-Ethylnorscopolamine
154(10), 138(100), 110(9), 108(22), 97(28), 96(26), 67(21), 56(17)
302(6), 168(42), 152(100), 150(31), 122(51), 108(76), 103(39), 56(78)
a Relative to the normal alkaloid (at Rt=1.00). Hygrine=7.77 min; oxohygrine=8.72 min; cuscohygrine=27.90 min; tropine=13.13 min; tro-
pinone=13.47 min; 3-acetyltropine=17.62 min; 3-tigloyltropine=28.27 min; littorine=39.29 min; hyoscyamine=39.79 min; scopolamine=42.47
min.
distal from the ®nal products (Fig. 1). GC traces of a
control extract and extracts of roots supplied with
various N-alkyldiamines are presented in Fig. 2. The
spectra of extracts from fed roots contain, to varying
extents, additional compounds that are absent from
the control. These were shown by GC/MS to have
the expected mass increment relative to the normal
alkaloids and to give corresponding fragmentation
patterns (Table 3). The precursors were found to be
extensively metabolized into alkaloids of both the
monocyclic/bismonocyclic pyrrolidine series and the
bicyclic [3.2.1]tropane alkaloid types (Table 4).
Within the former category, peaks corresponding to
the higher N-alkyl-series for hygrine, oxohygrine (an
uncharacterized oxidation product of hygrine) and
the N-methyl,N'-alkyl- and N-alkyl,N'-alkyl- cuscohy-
grine analogues are present. These alkaloids constitute
the major metabolic products in these feeds and may
readily be distinguished on the basis of the ratio of
the mass fragments at 84 and 98 m.u. With the
exception of N-ethylputrescine 3, the yield of ana-
logues of the tropane alkaloids was poor, the yield
broadly diminishing as the N-alkyl constituent
increased in mass. The majority of the products that
accumulated were of the endo/a con®guration, i.e.
analogues of tropine derivatives. No products could
be detected when N-cyclohexylputrescine or N-pheny-
lethylputrescine were fed.
verine 7 proved to be a good substrate, producing
both the cuscohygrine-like bispiperidines and a number
of esters of reduced pseudopelletierine. Apart from
some slight oxidation of N-ethylcadaverine 8 to an
uncharacterized homologue of oxohygrine, however,
higher homologues were not metabolized and no pro-
ducts were found with N-benzylcadaverine, N-butylca-
daverine, N-n-propylcadaverine or N-iso-propyl-
cadaverine.
In contrast to the N. rustica cultures, the feeding of
these putative precursors at 1 mM had a minimal
e€ect on the pro®le and yield of normal alkaloids. As
shown in Table 5, where the pro®le is compared to the
control grown in parallel, the yields of both tropane
and nontropane alkaloids is comparable and is overall
still strongly dominated by littorine 17, hyoscyamine
18 and scopolamine 19.
2.3. N-alkylnortropinone metabolism by Brugmansia
candida  aurea root cultures
In a further series of feeds, N-alkyl analogues of tro-
pinone 9 or its higher homologue, pseudopelletierine
14, were fed to examine the di€erence between provid-
ing distal and proximal precursors to the same root
cultures.
As can be seen (Table 6), all three of N-ethyl- 10,
N-(2-¯uoroethyl)- 11 and N-iso-propylnortropinone 12
were e€ectively metabolized to a range of aliphatic
and aromatic esters. No analogues of the pyrrolidine-
type alkaloids were observed (data not shown). The
yield was in all cases superior to that found with the
When the higher homologous series was fed, both
cadaverine 6 and N-methylcadaverine 7 were found to
be metabolized. Although cadaverine 6 was only con-
verted to pelletierine (the higher homologue of
hygrine) and norpseudopelletierine 13, N-methylcada-

H.D. Boswell et al. / Phytochemistry 52 (1999) 855±869
863

864
H.D. Boswell et al. / Phytochemistry 52 (1999) 855±869

H.D. Boswell et al. / Phytochemistry 52 (1999) 855±869
865
Table 6
Production of alkaloid analogues by transformed root cultures of Brugmansia fed with N-alkylated nortropinones
Analogue alkaloid yield (% of total alkaloid)a
Tropinone additive
(1 mM)a
N-ethylnortropinone
(N = 2)
N-(2-¯uoroethyl)-
nortropinone
N-iso-propyl-
nortropinone
N-ethylnor-
pseudopelletierine
Analogue of
Tropine
20.4210.0
8.027.4
0.5520.55
11.823.2
0.020.0
0.4520.05
0.3520.15
0.1520.15
13.326.8
1.921.4
21.6
0.0
4.1
6.2
0.0
0.6
0.0
0.0
6.2
0.0
0.0
19.8
0.7
0.0
0.9
0.0
0.2
0.1
0.0
9.7
0.0
0.0
10.4
15.4
0.9
2.9
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Tropinone (substrate)
Pseudotropine
3-Acetyltropine
3-Acetylpseudotropine
3-Tigloyltropine
3-Tigloylpseudotropine
Phenylacetyltropine
Littorine
Hyoscyamine
Scopolamine
4.924.0
Analogue as % total alkaloid
Total normal alkaloid (mg)b
Total novel alkaloid (mg)bc
62.021.8
4.821.0
7.221.6
48.5
15.5
14.6
32.6
12.7
6.1
48.2
11.6
2.6
a The values represent the average of N separate experiments, each representing the mean of between 5 and 10 ¯asks, pooled and extracted
together. Values given are the average2range.
b Some minor alkaloids are not included.
c Corrected for the presence of unmetabolised substrate.
equivalent N-alkyldiamine (Table 4) but, in the same
manner, showed a diminution with the mass of the
N-alkyl substituent. Metabolism of the N-ethyl ana-
logue 10 was particularly strong, the N-ethyl products
exceeding 60% of the total alkaloid. While the N-(2-
¯uoroethyl)- 11 and N-iso-propyl- 12 nortropinone
analogues were very eciently reduced Ð leaving
negligible quantities of the substrate in the extract Ð
the products of reduction were less e€ectively esteri-
®ed. Again, the reduction strongly favoured the for-
mation of the endo/a-tropan-3-ol. The majority of the
esteri®ed alkaloid accumulated with all three sub-
strates was the phenyllactoyl ester and only with N-
ethylnortropinone 10 was there any measurable re-
arrangement to the tropoyl ester (hyoscyamine plus
scopolamine analogues).
With the higher homologues of pseudopelletierine
[3.3.1], reduction and acetylation of the N-ethyl ana-
logue 15 was observed but no aromatic products could
be detected. No metabolism of N-(2-¯uoroethyl)norp-
seudopelletierine or N-iso-propylnorpseudopelletierine
16 occurred.
The presence of the N-alkylnortropinones qualitat-
ively had no obvious e€ect on the accumulation of
the normal alkaloids (Table 7). Quantitatively, it is
dicult to assess accurately their impact, in part due
to overlapping peaks in some of the GC spectra.
Overall, there appears to be no signi®cant diminution
of normal alkaloid accumulation, with the possible
exception of N-ethylnortropinone-treated cultures.
This aspect of the analogue metabolism merits further
attention.
3. Discussion
These studies clearly show that a considerable
degree of plasticity exists in the substrate speci®city
of many of the enzymes in the alkaloid biosynthetic
pathways of N. rustica and B. candida  aurea. In
this, they extend considerably previous studies in
which nonnatural precursors were fed (Boswell et al.,
1993; Demole & Berthet, 1972; Leete, 1983; Leete,
1979; Leete & Chedekel, 1972; Leete et al., 1971;
Rueppel & Rapoport, 1971; Walton & Belshaw, 1988;
Walton et al., 1988). Nevertheless, it was found that
only with a precursor analogue closely similar to the
natural intermediate could the yield of analogue-de-
rived alkaloids in the culture exceed that of the natu-
ral alkaloids. With more structurally-divergent
analogues, only small amounts (if any) of novel alka-
loids were produced. E€ectively, therefore, the poten-
tial to produce novel alkaloids was limited to
relatively minor structural variations of the normal
alkaloid repertoire. With the B. candida  aurea cul-
tures, it was found that the yield of analogues of
hyoscyamine and scopolamine was greatest when pre-

866
H.D. Boswell et al. / Phytochemistry 52 (1999) 855±869
Table 7
E€ect of N-alkylated tropinone analogues on the production of normal alkaloids by transformed root cultures of Brugmansia
Alkaloid yield
(% of total normal alkaloid)a
Tropinone additive
(1 mM)
N-ethylnortropinone
(N = 2)
N-(2-¯uoroethyl)nortropinone N-iso-propylnortropinone N-ethylnorpseudopelletierine
control
etNT-fed
control FetNT-fed
control iprNT-fed
control etNP-fed
Alkaloid produced
Tropine
Tropinone
22.321.7 18.0216.4 14.4
10.8
1.4
20.6 11.2
24.0
1.4
43.3
0.0
0.920.5
0.0520.05 0.220.2
0.320.3
1.2
0.1
0.4
< 0.1
0.5
0.2
Pseudotropine
3-Acetyltropine
3-Tigloyltropine
Phenylacetyltropine
Littorine
0.1
0.0
5.9
0.0
5.4
5.320.6
3.220.4
2.121.2
12.727.8
1.020.5
15.521.3
5.923.5 10.3
15.4
2.7
4.6 14.9
5.423.0
0.120.1
3.2
0.3
3.6
3.3
5.7
0.1
2.7
0.8
1.9
0.7
0.8
9.7
3.821.8 12.9
0.620.6 0.3
9.027.2 10.8
20.5 15.8
0.5 3.3
4.9
1.5
4.1
1.0
Aposcopolamine
Hyoscyamine
Scopolamine
0.7
12.6
10.4
14.3 14.0
16.4 12.4
16.8
26.6
16.9
19.2
21.525.1 13.127.7 13.9
Total normal alkaloid (mg)b 7.823.5
Total novel alkaloid (mg)bc
4.821.0 12.0
7.221.6
15.5
14.6
22.5 12.7
±
4.3
±
11.6
2.6
±
±
6.1
a The values represent the average of N separate experiments, each representing the mean of between 5 and 10 ¯asks, pooled and extracted
together. Values given are the average2range.
b Some minor alkaloids are not included.
c Corrected for the presence of unmetabolised substrate.
cursor analogues close to these tropane alkaloids
were supplied; with more distant precursor analogues
(N-alkyldiamines), there was substantial metabolism
via a competing pathway leading to pyrrolidine alka-
loids.
the biotransformation product is the (S )-isomer or
whether, like naturally occuring anabasine 21 from
Nicotiana species (Leete, 1983) and anabasine extracted
from transformed root cultures of N. rustica and N.
tabacum fed with cadaverine 6 (Watson, Brown,
Colquhoun, Walton & Robins, 1990), it is a racemic
mixture.
An ecient conversion of N-ethylputrescine by N.
rustica transformed root cultures to the anticipated
higher homologue of nicotine, N'-ethyl-(S )-nornico-
tine has already been reported (Boswell et al. 1993).
This product accounted for about 40% of the total
alkaloid present in the roots. The results presented
here suggest that N-n-propylputrescine is metabolised
similarly and to a comparable extent, although the
small quantity of this and other precursors available
for feeding did not permit characterization of the al-
kaloidal product by NMR or by CD. N-iso-propyl-
nornicotine is known as a trace natural metabolite in
Burley tobacco (Leete, 1983). Three N-alkylcadaver-
ines (N-methyl-, N-ethyl- and N-n-propyl-) were
examined for bioconversion; each gave rise to a novel
alkaloid, but only in the case of N-methylcadaverine
7 did this account for a substantial (ca. 20%) pro-
portion of total alkaloid content. This new alkaloid
exhibited a mass spectrum consistent with N'-methy-
lanabasine, a known natural product of N. tabacum
where it occurs as the (S )-isomer (Demole & Berthet,
1972). It would be valuable to establish whether, in
common with N'-methylanabasine isolated from N.
glauca and N. tabacum fed N-methyl-D1-piperideine,
The inhibitory e€ects of cadaverine 6 on the for-
mation of nicotine 23 in transformed root cultures of
Nicotiana spp. have been discussed elsewhere (Walton
& Belshaw, 1988; Walton & McLauchlan, 1990;
Walton et al., 1988; Walton et al., 1994) and were
similarly observed here with N-methylcadaverine 7.
Whether or not this is a direct or indirect e€ect
needs to be tested with labelled substrates. If there
were a direct e€ect, since the N-methylcadaverine was
free of cadaverine, the stimulation of anabasine 21
formation by N-methylcadaverine 7 must result from
the demethylation either of the compound itself or of
N-methylpiperideine: N-methylanabasine has already
been ruled out as an intermediate in anabasine for-
mation by wick-feeding experiments with N. glauca
plants (Leete & Chedekel, 1972). Feeding of N-n-pro-
pylputrescine, though not of the other compounds
tested, caused an increase in the proportion of anata-
bine 22. Although anatabine arises by the conden-
sation of two molecules of nicotinic acid (Leete,
1983), increased anatabine formation both in response
to cadaverine-feeding of transformed root cultures

H.D. Boswell et al. / Phytochemistry 52 (1999) 855±869
867
(Walton & Belshaw, 1988; Walton et al., 1988) and
to the expression of a bacterial lysine decarboxylase
gene in N. glauca (Fecker et al., 1992) has been
reported previously. While the e€ect of excess cada-
verine could be due to an aberrant route to anatabine
were metabolized to norpseudopelletierine 13 and
pseudopelletierine 14 respectively, showing that the
enzyme(s) responsible for the ring-closure of 4-(1-
methyl-2-pyrrolidinyl)-3-oxobutanoate 24 can appar-
ently catalyse this reaction with either a 5- or 6-mem-
bered ring.
22 from cadaverine (Mothes, Schutte, Simon,
&
Weygand, 1959), for example via anabasine or the
condensation of two piperideine units, this e€ect
could equally well be indirect. Certainly, it is hard to
explain the in¯uence of N-n-propylputrescine as other
than indirect. Since these compounds are both sub-
strates for MPO (Boswell et al., 1999; Walton and
McLauchlan, 1990) the diminution could perhaps
result from inhibition of the formation of N-methyl-
pyrrolinium, with a consequent stimulation of the
condensation of two molecules of decarboxylated nic-
otinic acid to give rise to anatabine. This interpret-
ation ®nds support in the observed overall diminution
of alkaloid yield in cultures fed N-n-propylputrescine
(Table 2).
In the B. candida  aurea cultures, all four N-alkyl-
diamines tested and cadaverine 6 were metabolized
beyond the N-methylpyrrolinium level to analogues of
hygrine and cuscohygrine (Table 4). No such com-
pounds arose in the cultures fed N-alkylnortropinones
(10, 11, 12), con®rming that the pyrrolidine alkaloids
diverge from the tropanes before ring closure.
Cadaverine 6 has previously been shown to be an
e€ective substrate for MPO of N. tabacum (Walton &
In view of the established requirement for an N-
alkyl substituent on the tropane ring for tropinone
reductase (TR)
I and TRII activity (Portste€en,
Drager, & Nahrstedt, 1994), it is not surprising that
reduction only occurs with the N-alkylated sub-
strates. The reduction products, in particular the
analogues of tropine (3a-ol), accumulated much
more e€ectively from the N-alkylnortropinone feeds
than from N-alkyldiamines. Notably, also, the more
proximal substrates gave a wider range of esteri®ed
tropines. Although with all of N-ethyl- 10, N-(2-
¯uoroethyl)- 11 and N-iso-propyl- 12 nortropinones
only one peak is found in the GC/MS chromato-
gram with the appropriate mass for the acetyl ester,
N-ethyl- and N-iso-propyl- nortropinones show two
peaks for the tigloyl ester, indicating products of
both the endo/3a and exo/3b series. By analogy with
the natural (N-methyl-) products, the earlier peak is
assigned to the 3a-con®guration. Similarly, in view
of the 3a:3b ratio normally found in the root cul-
tures (Boswell et al., 1999; Drager et al. 1992;
Robins et al., 1990) (Tables 5 and 7), the single
acetyl ester is assigned to the 3a-con®guration. It
should be noted, however that this interpretation is
in contrast to the in vitro enzyme activities (Boswell
et al., 1999). Similarly, the accumulation of the (3a-
ol) reduction product of N-ethylnorpseudopelletierine
15 in vivo is surprising in view of the apparent
absence in vitro of any activity of TRI with this
substrate (Boswell et al., 1999).
McLauchlan,
1990)
and
Hyoscyamus
niger
(Hashimoto, Mitani, & Yamada, 1990) and both cada-
verine 6 and N-methylcadaverine 7 are good substrates
for the MPO from B. candida  aurea cultures
(Boswell et al., 1999). These data indirectly con®rm
the wide substrate range for this enzyme.
Conversion of the N-alkyldiamines into the tropane
pathway metabolites, however, was limited. N-
Ethylputrescine 3 was metabolized through the entire
length of the tropane biosynthetic pathway, leading
to an appreciable production of analogues of tropine
esters including littorine and, to a lesser extent, hyos-
cyamine and scopolamine. N-(2-Fluoroethyl)putrescine
3a and N-iso-propylputrescine 4a were also metab-
olized as far as the littorine analogue, but only the
former was rearranged to an hyoscyamine analogue
and neither gave any scopolamine analogue. A similar
pattern occurred with the N-alkylnortropinones with
the notable di€erence that much higher levels of the
reduction products Ð analogues of tropine and pseu-
dotropine Ð accumulated. Almost no tropane alka-
loids arose from N-n-propylputrescine 4, although
this was an e€ective substrate for analogues of the
pyrrolidine alkaloid, cuscohygrine. Thus, it appears
that the linear C3 unit may interfere with acetoacetyla-
tion or ring closure to form the tropane skeleton. In
contrast, both cadaverine 6 and N-methylcadaverine 7
Phenyllactoyl ester formation (littorine analogue)
occurred fairly readily with the N-alkylnortropanes
but the complete absence of a littorine analogue
from N-methylcadaverine 7 or N-ethylnorpseudopelle-
tierine 15 suggests that the phenyllactoylation reac-
tion is restricted to the [3.2.1]tropane ring-containing
structures, the [3.3.1]system failing to act as a sub-
strate. Thus, it appears that for phenyllactoylation
the structural requirements for the N-alkyl group are
less stringent than for the ring structure. Similarly,
rearrangement to hyoscyamine occurred with N-
ethyl- and N-(2-¯uoroethyl)- norlittorines (Tables 4
and 6), but not with the N-n-propyl- or N-iso-pro-
pyl- norlittorines. Only with N-ethylputrescine 3 was
a N-ethylnorscopolamine analogue detected. This is
in accord with the report that N-acetylnoratropine
but not N-butanoylnoratropine are accepted as sub-
strate by puri®ed hyoscyamine 6b-hydroxylase
(Hashimoto & Yamada, 1987). The present ®ndings
imply that this enzyme may similarly accept a N-

868
H.D. Boswell et al. / Phytochemistry 52 (1999) 855±869
ethyl substituent, though not a N-(2-¯uoroethyl) or
N-propyl group.
4.3. Alkaloid extraction
In conclusion, it would appear that the most
stringent structural speci®city in tropane alkaloid
formation is shown by the enzymes catalysing the
phenyllactoylation of tropine to littorine 17 and
the subsequent rearrangement of littorine 17 to
hyoscyamine 18. These restrictions inevitably limit
the biotransformation capacity of the cultures and
their ability to accumulate large amounts of novel
alkaloids. However, limited alterations to structure,
such as the replacement of the N-methyl of tropi-
Crude alkaloid extracts were prepared as described
previously (Drager et al., 1992).
4.4. GC and GC/MS
Alkaloids were analysed by GC and GC/MS as
described previously (Drager et al., 1992).
Acknowledgements
none
exploited.
9
with an N-ethyl can be e€ectively
We have many colleagues to thank for their help
with this work: at IFR, Louise Tatton for GC/MS,
Ian Colquhoun for NMR and Abbi Peerless for grow-
ing cultures; at Nantes, Marc Berry for GC/MS. HDB
gratefully acknowledges receipt of a SERC CASE stu-
dentship. RJR and BD gratefully acknowledge the
®nancial assistance of the British Council and the
DAAD through the scienti®c exchange programme
No. ARC215.
4. Experimental
4.1. Transformed root cultures
Root cultures of Nicotiana rustica and of the
Brugmansia candida  aurea hybrid, transformed with
Agrobacterium rhizogenes, were grown as described
elsewhere (Hamill et al., 1986; Robins et al., 1990). In
general, 0.2 g fresh mass of roots was subcultured into
50 ml of B50 medium and grown in the absence of
antibiotics.
References
Alonso Garrido, D. O., Buldain, G., & Frydman, B. (1984). J. Org.
Chem., 49, 2021.
4.2. Synthetic substrates
Boswell, H. D., Drager, B., Portste€en, A., Robins, D. J., Robins,
R. J., Walton, N. J. (1999). Phytochemistry (in press).
Boswell, H. D., Watson, A. B., Walton, N. J., & Robins, D. J.
(1993). Phytochemistry, 34, 153.
Analogues of putrescine and cadaverine were pre-
pared following published procedures (Alonso
Garrido, Buldain & Frydman, 1984). N-Alkyl ana-
logues of tropinone were prepared by a modi®ed
Robinson synthesis (Robinson, 1917), involving the
acid-catalysed condensation (pH 4.8; 258C; 48 h) of
butan-1,4-dialdehyde, 3-ketoglutaric acid and the
appropriate alkylamine. The unstable tropan-2,4-
dicarboxylate product was decarboxylated by treat-
ment with HCl (6 M; 708C; 1 h) and the product
worked up by solvent extraction and crystallisation
as either the HCl or the oxalate salt. Higher hom-
ologues were similarly prepared but with pentan-1,5-
dialdehyde. Purity was assessed by GC/MS (Drager
et al., 1992) and by NMR spectroscopy. Standards
for the identi®cation of products by GC and GC/
MS were prepared by the chemical reduction with
sodium borohydride of N-alkylnortropinones and N-
alkylnorpseudopelletierines to a mixture of the 3a-
and 3b-isomers. This mixture was in turn used to
synthesize the phenyllactoyl-esters by heating
(1408C; 48 h) a fused mix of the N-alkylnortropine
with (R )-phenyllactic acid under an HCl atmos-
phere.
Demole, E., & Berthet, D. (1972). Helv. Chim. Acta, 55, 1898.
Drager, B., Portste€en, A., Schaal, A., McCabe, P., Peerless, A. C.
J., & Robins, R. J. (1992). Planta, 188, 581.
Fecker, L. F., Hillebrandt, S., Rugenhagen, C., Herminghaus, S.,
Landsmann, J., & Berlin, J. (1992). Biotechnol. Lett., 14, 1035.
Friesen, J. B., & Leete, E. (1990). Tetrahedron Lett., 31, 6295.
Hamill, J. D., Parr, A. J., Robins, R. J., & Rhodes, M. J. C. (1986).
Plant Cell Rep., 5, 111.
Hashimoto, T., & Yamada, Y. (1987). Eur. J. Biochem., 164, 277.
Hashimoto, T., Mitani, A., & Yamada, Y. (1990). Plant Physiol., 93,
216.
Leete, E. (1979). J. Org. Chem., 44, 165.
Leete, E. (1983) In S. W. Pelletier, Alkaloids: chemical and biological
perspectives, 7, (p. 85). John Wiley & Sons Inc.
Leete, E., & Chedekel, M. R. (1972). Phytochemistry , 11, 2751.
Leete, E., Bodem, G. B., & Manuel, M. F. (1971). Phytochemistry,
10, 2687.
Mothes, K., Schutte, H. R., Simon, H., & Weygand, F. Z. (1959).
Naturforsch. Teil B, 14, 49.
Ollagnier, S., Kervio, K., & Retey (1998). J. FEBS Lett., 437, 309.
Portste€en, A., Drager, B., & Nahrstedt, A. (1994). Phytochemistry,
37, 391.
Robins, R. J. (1998). In M. F. Roberts, & M. Wink, Alkaloids: bio-
chemistry, ecology and medicinal applications (p. 199). Plenum
Press.
Robins, R.J., Walton, N.J. (1993). In G. A. Cordell, The alkaloids,
44, (p. 115). Academic Press.

H.D. Boswell et al. / Phytochemistry 52 (1999) 855±869
869
Robins, R. J., Abraham, T. W., Parr, A. J., Eagles, J., & Walton, N.
J. (1997). J. Am. Chem. Soc., 119, 10929.
Walton, N. J., & Belshaw, N. J. (1988). Plant Cell Rep., 7, 115.
Walton, N. J., & McLauchlan, W. R. (1990). Phytochemistry, 29,
1455.
Robins, R. J., Chesters, C. J. E., O'Hagan, D., Parr, A. J., Walton,
N. J., & Woolley, J. G. (1995). J. Chem Soc. Perkin Trans., 1,
481.
Walton, N. J., Peerless, A. C. J., Robins, R. J., Rhodes, M. J. C.,
Boswell, H. D., & Robins, D. J. (1994). Planta, 193, 9.
Walton, N. J., Robins, R. J., & Rhodes, M. J. C. (1988). Plant Sci.,
54, 125.
Robins, R. J., Parr, A. J., Payne, J., Walton, N. J., & Rhodes, M. J.
C. (1990). Planta, 181, 414.
Robinson, R. (1917). J. Chem. Soc., 111, 762.
Rueppel, M. L., & Rapoport, H. (1971). J. Am. Chem. Soc., 93,
7021.
Watson, A. B., Brown, A. M., Colquhoun, I. J., Walton, N. J.,
& Robins, D. J. (1990). J. Chem. Soc. Perkin Trans., 1,
2607.