Semi-automated separation of the epimeric dehydropyrrolizidine alkaloids lycopsamine and intermedine: Preparation of their N-oxides and NMR comparison with diastereoisomeric rinderine and echinatine

Article abstract of DOI:10.1002/pca.2511

Introduction The diversity of structure and, particularly, stereochemical variation of the dehydropyrrolizidine alkaloids can present challenges for analysis and the isolation of pure compounds for the preparation of analytical standards and for toxicology studies. Objective To investigate methods for the separation of gram-scale quantities of the epimeric dehydropyrrolizidine alkaloids lycopsamine and intermedine and to compare their NMR spectroscopic data with those of their heliotridine-based analogues echinatine and rinderine. Methods Lycopsamine and intermedine were extracted, predominantly as their N-oxides and along with their acetylated derivatives, from commercial samples of comfrey (Symphytum officinale) root. Alkaloid enrichment involved liquid-liquid partitioning of the crude methanol extract between dilute aqueous acid and n-butanol, reduction of N-oxides and subsequent continuous liquid-liquid extraction of free base alkaloids into CHCl₃. The alkaloid-rich fraction was further subjected to semi-automated flash chromatography using boronated soda glass beads or boronated quartz sand. Results Boronated soda glass beads (or quartz sand) chromatography adapted to a Biotage Isolera Flash Chromatography System enabled large-scale separation (at least up to 1-2 g quantities) of lycopsamine and intermedine. The structures were confirmed using one- and two-dimensional ¹H- and ¹³C-NMR spectroscopy. Examination of the NMR data for lycopsamine, intermedine and their heliotridine-based analogues echinatine and rinderine allowed for some amendments of literature data and provided useful comparisons for determining relative configurations in monoester dehydropyrrolizidine alkaloids. A similar NMR comparison of lycopsamine and intermedine with their N-oxides showed the effects of N-oxidation on some key chemical shifts. A levorotatory shift in specific rotation from +3.29° to -1.5° was observed for lycopsamine when dissolved in ethanol or methanol respectively. Conclusion A semi-automated flash chromatographic process using boronated soda glass beads was standardised and confirmed as a useful, larger scale preparative approach for separating the epimers lycopsamine and intermedine. The useful NMR correlations to stereochemical arrangements within this specific class of dehydropyrrolizidine alkaloid cannot be confidently extrapolated to other similar dehydropyrrolizidine alkaloids.

Full text of DOI:10.1002/pca.2511

Research Article
Received: 30 December 2013,
Revised: 29 January 2014,
Accepted: 2 February 2014
Published online in Wiley Online Library
(wileyonlinelibrary.com) DOI 10.1002/pca.2511
Semi-automated Separation of the Epimeric
Dehydropyrrolizidine Alkaloids Lycopsamine
and Intermedine: Preparation of their N-oxides
and NMR Comparison with Diastereoisomeric
Rinderine and Echinatine
a
a
b
a
Steven M. Colegate, * Dale R. Gardner, Joseph M. Betz and Kip E. Panter
ABSTRACT:
Introduction – The diversity of structure and, particularly, stereochemical variation of the dehydropyrrolizidine alkaloids can
present challenges for analysis and the isolation of pure compounds for the preparation of analytical standards and for
toxicology studies.
Objective – To investigate methods for the separation of gram-scale quantities of the epimeric dehydropyrrolizidine alkaloids
lycopsamine and intermedine and to compare their NMR spectroscopic data with those of their heliotridine-based analogues
echinatine and rinderine.
Methods – Lycopsamine and intermedine were extracted, predominantly as their N-oxides and along with their acetylated
derivatives, from commercial samples of comfrey (Symphytum officinale) root. Alkaloid enrichment involved liquid–liquid
partitioning of the crude methanol extract between dilute aqueous acid and n-butanol, reduction of N-oxides and subsequent
3
continuous liquid–liquid extraction of free base alkaloids into CHCl . The alkaloid-rich fraction was further subjected to semi-
automated flash chromatography using boronated soda glass beads or boronated quartz sand.
Results – Boronated soda glass beads (or quartz sand) chromatography adapted to a Biotage Isolera Flash Chromatography System
enabled large-scale separation (at least up to 1–2 g quantities) of lycopsamine and intermedine. The structures were confirmed
1
13
using one- and two-dimensional H- and C-NMR spectroscopy. Examination of the NMR data for lycopsamine, intermedine and
their heliotridine-based analogues echinatine and rinderine allowed for some amendments of literature data and provided useful
comparisons for determining relative configurations in monoester dehydropyrrolizidine alkaloids. A similar NMR comparison of
lycopsamine and intermedine with their N-oxides showed the effects of N-oxidation on some key chemical shifts. A levorotatory
shift in specific rotation from +3.29° to À1.5° was observed for lycopsamine when dissolved in ethanol or methanol respectively.
Conclusion – A semi-automated flash chromatographic process using boronated soda glass beads was standardised and
confirmed as a useful, larger scale preparative approach for separating the epimers lycopsamine and intermedine. The useful
NMR correlations to stereochemical arrangements within this specific class of dehydropyrrolizidine alkaloid cannot be
confidently extrapolated to other similar dehydropyrrolizidine alkaloids. Published 2014. This article is a U.S. Government
work and is in the public domain in the USA.
Supporting information can be found in the online version of this article.
Keywords: Boronated soda glass chromatography; HPLC–ESI/MS; NMR; optical rotation; echinatine; dehydropyrrolizidine alkaloids;
intermedine; lycopsamine; rinderine; comfrey; Symphytum officinale
The C9 monoester DHPAs lycopsamine (1) and intermedine
2) are epimeric about C13 (Fig. 1) (Mackay et al., 1983) and are
components in many DHPA-producing plants (Bull et al., 1968).
Introduction
In vivo oxidative metabolites of the plant-derived esters of
dehydropyrrolizidine alkaloids (DHPAs) and their N-oxides are
potentially acutely and chronically toxic to livestock and humans
(
(
Edgar et al., 2011; Molyneux et al., 2011). Consequently, to enable
*
Correspondence to: Steven M. Colegate, Poisonous Plant Research Laboratory,
Agriculture Research Service, US Department of Agriculture, 1150 East 1400
North, Logan, Utah 84341, USA.
monitoring of animal feed, the human food supply, dietary
supplements and herbal medicines for the presence of the DHPAs,
and to define the actual risk to health presented by individual
DHPAs, pure alkaloids are required for analytical standards and,
on a larger scale, for toxicological studies. However, the sometimes
subtle diversities of structure and stereochemistry (relative and
absolute) of the several hundreds of DHPAs known can present a
challenge for analysis and for the isolation of pure alkaloids.
E-mail: steven.colegate@ars.usda.gov
a
Poisonous Plant Research Laboratory, Agriculture Research Service, US
Department of Agriculture, 1150 East 1400 North, Logan, Utah, 84341, USA
b
Office of Dietary Supplements, National Institutes of Health, 6100 Executive
Blvd., Room 3B01, Bethesda, Maryland, 20892, USA
Phytochem. Anal. 2014
Published 2014. This article is a U.S. Government work and is in the public domain in the USA.

S. M. Colegate et al.
Lycopsamine (1) and intermedine (2) in extracts of Amsinckia
intermedia have been previously separated from each other by
taking advantage of their different affinities for the boronate
anion (Frahn et al., 1980). Two methods were described: (i) an
affinity-type column where 1 and 2 were selectively partitioned
between aqueous borax-coated powdered soda glass and the
CHCl elution solvent; and (ii) an ion exchange process in which
3
the anionic nature of the respective boronate complexes
differentially hindered retention on a cation exchange resin.
Only one of these approaches, that is, the cation exchange,
seems to subsequently have been utilised successfully, albeit
only once and with mannitol as the lycopsamine/intermedine
displacement ligand instead of glucose (Roitman, 1983). In the
present study, both methods were re-examined for their ability
to separate lycopsamine, intermedine and related alkaloids
derived from Symphytum officinale roots with the intent of
standardising an approach with modern availability of reagents,
substrates and equipment.
Experimental
Chemicals and reagents
Dry, powdered comfrey (S. officinale) root (Frontier Natural Products
Co-op) was purchased from Take Herb (Alhambra, CA, USA). The
following were reagent ACS/USP/NF grade (Pharmaco Products,
Brookfield, CT, USA): methanol, for extractions; petroleum ether (b.p.
Figure 1. Structures of dehydropyrrolizidine alkaloids.
3
3
5–60°C) for column packing; and CHCl , for flash column chromatog-
Each can be biosynthesised relatively exclusively of the other,
such as the lycopsamine-N-oxide chemotype of Amsinckia
intermedia (Colegate et al., 2013) and intermedine-N-oxide from
Cerinthe glabra (Luber et al., 2012), but more often they co-occur
at various relative levels as the free bases and/or their N-oxides
raphy. All other reagents, that is, zinc powder, n-butanol, ammonia
solution and concentrated sulphuric acid were of analytical grade
and generally sourced. Acetonitrile was HPLC-certified solvent
(
Honeywell Burdick and Jackson, Muskegon, MI, USA) and water was
Milli-Q-purified (18.2 MΩ/cm) (Millipore, Bedford, MA, USA). Formic
acid was ‘For Analysis’ grade (> 99%; Acros Organics/Thermo Fisher
Scientific, Bridgewater, NJ, USA). Ammoniated methanol was prepared
by bubbling dry ammonia gas through cooled (ice bath) methanol. The
resultant saturated ammoniated methanol solution was diluted 1:9
with methanol for use in strong cation exchange, solid phase
extraction (SCX SPE). The HPLC columns, with their guard columns,
bulk SCX packing material (Sepra SCX, 50 μm, 65A) and SCX SPE
cartridges (Strata SCX 55 μm, 70 A; 5 g/20 mL Giga tube), were sourced
from Phenomenex (Torrence, CA, USA). The indigo-carmine-based
redox resin was prepared and used as previously described (Colegate
et al., 2005; Williams et al., 2011). Echinatine (3), rinderine (4), heliotrine
(Kelley and Seiber 1992; Williams et al., 2011).
In some cases a fortuitous combination of solubility properties
and a lack of other DHPA co-metabolites has allowed for a
relatively simple extraction, purification and precipitation
process to result in large quantities of relatively pure DHPA. Such
was the case, for example, in the isolation of the macrocyclic
diester DHPAs riddelliine (Adams et al., 1942) and senecionine
(
Culvenor, 1962). However, when the DHPAs are less tractable,
methods of separation include thin-layer chromatography
TLC), column chromatography, centrifugal planar chromatogra-
(
phy, semi-preparative and preparative scale high-pressure liquid
chromatography (HPLC), rotation locular counter current
chromatography (RLCCC), droplet counter current chromatogra-
phy (DCCC) and high-speed counter current chromatography
(
5) (Fig. 1) and lasiocarpine (all > 99% pure based on HPLC–ESI(+)/MS
and NMR analysis) were sourced from the stocks of extracted and
purified pyrrolizidine alkaloids kept by the USDA/ARS Poisonous
Plant Research Laboratory. One- and two-dimensional (COSY, HSQC)
1
13
H- (300 MHz) and C- (75 MHz) NMR data were acquired using a JEOL
Eclipse NMR spectrometer using solutions in deuterochloroform
Sigma-Aldrich, St Louis, MO, USA) and the residual proton in the CHCl
(HSCCC). Because of the time, effort and higher levels of losses
through degradation or adsorption-related effects usually
involved with necessary repetitive application of the solid
stationary phase approaches, they have mostly been used to iso-
late relatively small quantities of pure alkaloid for NMR-based
structural studies (e.g. Roeder et al., 1991; Colegate et al.,
(
3
as the chemical shift reference. Optical activity was measured using an
Autopol IV Automatic Polarimeter (Rudolph Research Analytical,
Hackettstown, NJ, USA).
2
012). High-speed CCC, a hydrodynamic equilibration system,
has been applied to up to 800 mg of sample for the larger scale
separation of DHPAs from Amsinckia tessellata, Symphytum spp.,
Trichodesma incanum and Senecio douglasii var. longilobus (Cooper
et al., 1996). However, while the resolution reported was sufficient
to separate the open chain diesters (symphytine and related
isomers) from the monoesters (lycopsamine, intermedine and their
acetylated derivatives) extracted from Symphytum spp., it was not
sufficient to separate the individual diesters or monoesters (such
as the epimeric lycopsamine and intermedine) from each other.
HPLC–ESI(+)/MS
Analytical samples were prepared by adding an aliquot (10 μL) to a
lasiocarpine-spiked solution of 0.1% formic acid in water and methanol
(1:1, v/v) (100 μL). The lasiocarpine was included as an internal standard to
normalise all responses. An aliquot (5 μL) of an analytical sample was subse-
quently injected (Agilent 1260 Infinity HPLC System, Agilent Technologies,
CA, USA) onto a Synergi Hydro RP column (150 × 2 mm, 4 μ) fitted with a
guard column of similar adsorbent (AC C18, 4 mm diameter × 2 mm)
(Security Guard Cartridge system). Sample components were eluted from
wileyonlinelibrary.com/journal/pca
Published 2014. This article is a U.S. Government work and is in
the public domain in the USA.
Phytochem. Anal. 2014

Isolation of Lycopsamine and Intermedine
the column using a gradient flow (400 μL/min) of 0.1% formic acid in water
Isolera 1 flash chromatography system (Biotage, Charlotte, NC, USA)
and washed with CHCl . At this stage the column effluent bypassed
(mobile phase A) and acetonitrile (mobile phase B). Thus, mobile phase B
3
(
3%) was held for 2 min before linearly increasing to 70% by 10 min. After
the fraction-collection module of the system to avoid unnecessary
contamination of the UV detector cell and collection lines. Visual
inspection of the column effluent showed when any more excess water
was flushed from the column and when the petroleum ether was
holding at 70% for another 5 min the column was re-equilibrated to 3% mo-
bile phase B over 2 min and held for a further 7 min before the next injection.
The column effluent was directed into a Velos Pro LTQ mass
spectrometer (Thermo Scientific, USA) equipped with
a
heated
3
replaced completely by CHCl .
electrospray ionisation (HESI) source. Mass spectra were acquired using
alternating scans in the positive ion mode over the 15 min chromato-
graphic run. The first full scan (m/z 200 – 800) was followed by a data
dependent, collision-induced dissociation (CID) scan using a generic
CID energy of 32%, Activation Q of 0.25 and an Activation Time of 10.0
ms. The ionisation spray voltage was set at 3.45 kV, the source heater
temperature at 305°C and the sheath gas flow was 40 units with an
auxiliary flow of 5 units. The capillary temperature was set at 275°C.
During typical use the packed and conditioned boronated soda glass
bead column was disconnected from the Isolera system and the column
end cap removed. The headspace CHCl
face of the packed glass bead column and a solution of reduced extract
of S. officinale (ca. 4 g) in CHCl (15 mL) was added slowly to the head of
3
was drained down to the sur-
3
the column. The slow addition ensured effective exposure of the
alkaloids to the aqueous borax film coating the glass beads such that
the orange–red band occupied the first 1.5–3 cm of the column by the
end of the addition. Chloroform (ca. 20 mL) was added to the head of
the column, the top end cap was refitted and the column was connected
Extraction and isolation of crude alkaloid mixtures from
to the Isolera system. The column was washed with CHCl at 2 mL/min.
3
Symphytum officinale
The first 100 mL was collected separately followed by up to 140 × 22
mL fractions using the Isolera 1 fraction collector. The flow rate was then
increased to 10 mL/min and ca. 10 fractions of 900 mL were collected
separately. The separation was monitored using HPLC–ESI(+)/MS.
Decreased separation efficiency after repeated use indicated a need for
cleaning and regeneration of the column. To regenerate the column, the
glass beads were removed and efficiently cleaned by washing with metha-
nol. Once thoroughly dried of all methanol, the clean beads were re-coated
with aqueous borax and repacked into the column as described.
Dry comfrey (Symphytum officinale) root powder (ca. 5 kg) was Soxhlet-
extracted with methanol (15 L) for 4 × 20 h and 1 × 60 h. After each 20
h extraction, the extract-loaded solvent was exchanged for fresh solvent
to decrease any degradation of DHPAs in the boiling methanol. The
progress of the extraction was monitored using HPLC–ESI(+)/MS.
Each extract was separately evaporated to near dryness, affording a
red oil–solid mixture that was partitioned between n-butanol (500 mL)
and 0.05 M sulphuric acid (250 mL). The pale orange aqueous acid layer
was separated from the deep red n-butanol layer. The n-butanol layer
was back-washed with 0.05 M sulphuric acid (2 × 250 mL). The
partitioning and washes were monitored using HPLC–ESI(+)/MS. The
Preparation of N-oxides
combined aqueous acid fraction was made more acidic (to ca. 0.5
M
A sample of lycopsamine (ca. 450 mg, ca. 1.5 mmol) and a sample of
intermedine (ca. 700 mg, 2.3 mmol) were each separately dissolved in
ethanol (10 mL). An aliquot of 30% hydrogen peroxide (5 mL, ca. 44
mmol) was added to the filtered ethanolic solutions and left to stand
at room temperature. The reaction solutions became slightly cloudy with
a fine white precipitate or suspension. There was a slight pressure build-
up that was released every hour or so for the first 4–5 h and then every
24 h. The progress of the reactions was monitored using HPLC–ESI(+)/MS
and MS/MS.
sulphuric acid) and stirred with zinc powder. The reduction was
monitored as usual with HPLC–MS and took about 24 h to be complete.
The filtered reduction product was made basic with ammonium hydrox-
ide, until the copious precipitate of zinc hydroxide just redissolved, and
was then continuously extracted with CHCl
displacement liquid–liquid extractor unit (Sigma-Aldrich). The extracting
solvent was replaced after 20 h. The resultant, pale yellow CHCl extract
3
using a 2 L downward
3
solution was dried over anhydrous sodium sulphate and evaporated to
dryness to afford a red–orange gum (ca. 16 g combined yield from the
After about a week, the slightly cloudy solutions were evaporated un-
der reduced pressure to remove most of the ethanol. The aqueous resi-
due from each reaction was diluted with 0.05 M sulphuric acid (ca. 10 mL)
and each applied to a separate, conditioned SCX SPE cartridge. The
loaded columns were treated as usual (Colegate et al., 2012) by washing
with water and methanol before eluting the components using 10%
saturated ammoniated methanol. All column manipulations were
monitored using HPLC–ESI(+)/MS.
The first basic methanol fraction (ca. 20 mL) from each column was
immediately evaporated to dryness to afford lycopsamine-N-oxide (402
mg, 85% yield) and intermedine-N-oxide (622 mg, 84% yield), each as a
white glassy foam. The foam structure was readily broken using a spatula
to yield white solids for each alkaloid N-oxide.
repeat extractions) that was redissolved in CHCl
°C until used.
3
(60 mL) and stored at
4
Semi-automated flash column chromatography of alkaloidal
fractions
Soda glass beads (0.09–0.135 mm, ca. 1 kg) (Thomas Scientific,
Swedesboro, NJ, USA) in a screw-cap media storage bottle (2 L) were
treated with 5% w/v aqueous borax (300 mL) by roller-mixing for 20 h
to ensure effective coating of the beads. The moistened beads were
added in small aliquots to an ACE Large Chromatography column
(600 × 50 mm; Supelco/Sigma-Aldrich, Bellefonte, PA, USA) with a
bottom-end fitting that included a flow valve (adapter, filter, ptfe, #50
ace-thred to 1/4in-28unf, flow valve, 100 micron polyethylene filter, fetfe
o-ring; ACE Glass, Vineland, NJ, USA) and an appropriate top-end fitting
Results and discussion
(
adapter, filter, ptfe, #50 ace-thred to 1/4in-28unf, 100 micron polyethyl-
Extraction and concentration of comfrey root DHPAs
ene filter, fetfe o-ring; ACE Glass) and containing petroleum ether. After
each addition, the clumped aliquot of coated beads was carefully
compressed using a plunger made by screwing a rubber stopper (slightly
smaller diameter than the column) to a long wooden handle (60 cm).
This packing process was continued until all the beads were used,
producing the final packed column (550 × 45 mm). The column was
washed with petroleum ether until no more water was visible in the
column effluent, at which stage the petroleum ether was drained to
the surface of the column and the headspace then topped up with
In contrast to an earlier study of S. officinale that reported the
isolation of the open chain diester DHPAs symlandine, symphytine
and echimidine (Kim et al., 2001), the present HPLC–ESI(+)/MS anal-
ysis of the crude methanol extract of S. officinale root powder
showed the major presence of the N-oxides and free base forms
of the monoester DHPAs lycopsamine, intermedine and their acet-
ylated derivatives. There was a minor presence of the symlandine-
like isomers and their N-oxides (Fig. 2, Table 1). The MS/MS data
(Table 1) for the acetylated derivatives that, under these
3
CHCl . After fitting the column top-end cap, the column was connected
to a flow- (composition and rate) and fraction-collection-programmable
Phytochem. Anal. 2014
Published 2014. This article is a U.S. Government work and is in
wileyonlinelibrary.com/journal/pca
the public domain in the USA.

S. M. Colegate et al.
100
4
90
80
70
60
50
40
30
20
10
0
A
6
3
*
2
5
1
8
2
90
80
70
60
50
40
30
20
10
0
B
1
5
*
3
4
7
1
2
3
4
5
6
7
8
9
10
11
12
13
Figure 2. The HPLC–ESI(+)/MS base ion (m/z 200–800) chromatograms for (A) the crude methanol extract and (B) the zinc/acid-reduced extract of dried,
powdered roots of Symphytum officinale. Peaks labelled with ‘*’ are the internal standard, lasiocarpine. Peak identities are: 1, intermedine; 2, lycopsamine;
3, intermedine-N-oxide; 4, lycopsamine-N-oxide; 5, acetylated lycopsamine/intermedine; 6, acetylated lycopsamine/intermedine-N-oxides; 7, symlandine-
like isomers; 8, symlandine-like N-oxide isomers. The MS and MS/MS data are shown in Table 1.
Table 1. MS and MS/MS data for dehydropyrrolizidine alkaloids and N-oxides in a methanol extract of dried comfrey (Symphytum
officinale) root
+
Alkaloid
MH m/z
MS/MS m/z (% abundance)
Symlandine isomers
382
382 (1.6), 338 (1.6), 300 (1), 238 (2), 220 (15.9),
138 (1), 121 (1.3), 120 (100), 118 (2)
Symlandine isomers N-oxides
398 (100)/795(20)
398 (100), 380 (7), 354 (32.5), 308 (11.4),
2
2
98 (2.8), 292 (1.9), 282 (1.5), 254 (68.6),
37 (3), 220 (7.8), 218 (1.4), 154 (1.8), 137
(
2.6), 136 (1.9), 120 (1.7)
198 (2), 180 (76.6), 162 (6.3), 124 (1),
20 (100), 118 (1.4), 103 (1.7)
358 (62.4), 340 (9.6), 315 (3.4),
Acetylated lycopsamine/intermedine
342
1
Acetylated lycopsamine/intermedine N-oxides
358(40)/715(100)
3
2
14 (31.3), 298 (2.5),
96 (1.8), 268 (14.5), 252 (4), 242 (2.2),
214 (100), 197 (4), 180 (6.5), 178 (5.7),
154 (1.9), 137 (4.7), 136 (2.9), 120 ( 2.9)
Lycopsamine/intermedine
300
300 (2.7), 256 (2.8), 210 (1.3), 156 (6.8),
38 (100), 120 (14.8), 94 (27.9)
316 (55.3), 298 (4.7), 272 (22.8), 254 (1.8), 226
1
Lycopsamine/intermedine N-oxides
316 (30)/ 631 (100)
(
(
18.3), 210 (3), 172 (100), 156 (1.7), 155 (6.9), 154
4.8), 138 (15), 137 (4.6), 136 (7.3), 111 (1.6), 94 (3.4)
chromatographic conditions, effectively co-eluted, indicated that
the acetylation was at the C7 hydroxyl (rather than the C13 hy-
droxyl) by virtue of the strong ion at m/z 180 (Colegate et al., 2005).
The major problem associated with the Soxhlet methanol
extraction of bulk, powdered root material from S. officinale is
the presence of large amounts of white solid and deep red
pigmented co-extractives. Some of this solid was able to be
removed by filtration as the bulk methanol extract was concen-
trated in vacuo. Further concentration of the deep red filtrate
afforded a red oily residue that was subsequently extracted with
dilute aqueous acid. The resultant red–brown aqueous suspen-
sion–mixture was quite intractable with respect to further
filtration, and therefore, after trialling several solvents, it was
partitioned with water-saturated n-butanol. The deep-red
pigments, along with some of the DHPAs, partitioned into the
n-butanol. Monitoring using HPLC-ESI/MS showed that a subse-
quent wash of the n-butanol fraction with 0.05 M sulphuric acid
recovered the DHPAs. Two approaches were investigated for
the recovery of DHPAs from the combined aqueous acid
fractions: (i) strong cation exchange (SCX) capture of DHPAs
wileyonlinelibrary.com/journal/pca
Published 2014. This article is a U.S. Government work and is in
the public domain in the USA.
Phytochem. Anal. 2014

Isolation of Lycopsamine and Intermedine
and N-oxides followed by indigo-carmine-based redox resin-medi-
ated reduction of the N-oxides; (ii) zinc/acid reduction of the N-ox-
ides followed by liquid–liquid partitioning of the free base DHPAs.
The SCX-based capture of DHPAs and their N-oxides directly
from the aqueous acid fraction was a scaled-up version of similar
analytical applications (Betteridge et al., 2005) and required
pumping of large volumes of aqueous acid solution through
large columns (70 × 160 mm) of the silica-based, bulk SCX resin.
Issues associated with monitoring column overloading compli-
cated this approach with uncertainty. However, once the
columns were at capacity, they were washed and the DHPAs
and N-oxides eluted as usual with ammoniated methanol.
Evaporation of the solvent and reconstitution of the residue in
methanol was followed by treatment with indigo-carmine-based
redox resin to effect reduction of the N-oxides. However, and
despite high-efficiency recovery of DHPAs with this method,
the slow rate of application combined with the limited capacity
of the column and the consequent need for regular monitoring
of the column effluent to prevent overloading, rendered the
whole process quite slow for a large-scale preparation.
The simplest approach, requiring less ‘hands-on’ and less time,
was indeed the conventional liquid–liquid partitioning of the
aqueous acid-solubles after treatment with zinc dust to reduce
the N-oxides. Implementation of a continuous liquid–liquid
extractor instead of the repeated extractions using a separating
funnel expedited this approach. An overnight extraction of the
3
basified solution of reduction products with CHCl using the
liquid–liquid extractor was sufficient to recover most if not all
the DHPAs without any significant degradation of extracted
was possible using the more usual Biotage column cartridges.
Maintaining a narrow band of applied DHPA extract and
subsequent slow elution flow rates (2–5 mL/min) was essential
for efficient, differential partitioning of the various DHPAs along
the column length. Consequently the chromatographic runs
often exceeded 24 h duration.
The first eluate collection of 100 mL, every fifth fraction
collector tube, and each of the 900 mL final wash fractions were
monitored using HPLC–ESI/MS (Fig. 3). The bulk of the colour
eluted first, overlapping with elution of the symlandine-like
isomers and the acetylated derivatives of 1 and 2. Then,
intermedine (2) eluted over a relatively tight band compared
with the subsequent elution of lycopsamine (1). Fractions were
pooled based upon their constituents. Thus, evaporation of
fraction collector tubes 5–28 afforded a mixture of acetylated
derivatives, the symlandine isomers and intermedine as a red
oil (0.8 g). Fractions 29–75 yielded intermedine as a pale yellow
oil (0.87 g). Fractions 76–129 yielded a mixture of intermedine
and lycopsamine as a yellow oil (0.6 g) and fractions 130–140
plus the first six of the 900 mL fractions afforded lycopsamine
as a white foamy glass (1.4 g) (Fig. 3). The very slow rate of
elution of lycopsamine with respect to intermedine presumably
reflects a stronger boronate complex of the former.
Those fractions containing both 1 and 2 were variously
combined and re-chromatographed. The column is remarkably
stable and reusable. In one instance a column was prepared,
used once then allowed to stand for 4 months before it was then
reused another eight times over 5 months. The column does
retain some colour and, with repeated use, the efficiency of
separation of intermedine from lycopsamine reduces, indicating
that the column requires cleaning and regeneration.
3
DHPAs in the boiling CHCl for this time. The HPLC–ESI(+)/MS
analysis showed that the reduced comfrey root extract
+
comprised mainly intermedine and lycopsamine (MH m/z 300)
It was found that boronated quartz sand, treated in the same
way as the soda glass beads, performed just as well as the
boronated soda glass beads and so provides an alternative
matrix. However, most of the data and observations relating to
stability, cleaning, regeneration and reuse of the column
reported herein are derived from the soda glass columns.
+
along with their acetylated derivatives (MH m/z 342) and minor
amounts of the open chain diester symlandine-like isomers
+
(
MH m/z 382) (Fig. 1).
Semi-automated purification of lycopsamine (1) and
intermedine (2)
NMR comparison of lycopsamine stereoisomers 1–4
After a careful consideration of many options for separating
lycopsamine (1) and intermedine (2) from co-extracted DHPAs
and from each other, it was decided to reinvestigate the use of
boronated complexes. Despite the lack of literature reports
describing its use following the initial description by Frahn
et al. (1980), their differential partitioning approach using
crushed soda glass, rather than their ion exchange approach,
was considered the most straightforward and more likely to
be amenable to semi-automated, larger scale (ca. ≥ 1 g)
purifications using a standardised treatment of a standardised
matrix (soda glass beads or quartz sand).
Commercial flash chromatography systems, such as the
Biotage Isolera 1 used in this study, offer the ability to pump
various solvents in controlled isocratic or gradient elution modes
through packed chromatographic columns. Eluant fractions can
then be automatically collected based on the most effective
parameters. As the DHPAs lack a strong chromophore that
would allow for sensitive detection in real time, equivolume
fractions were collected and then analysed post-run using
HPLC–ESI(+)/MS, usually in overnight, automated analytical
sequences.
1
13
Examination of the one- ( H and C) and two-dimensional
(COSY and/or HSQC) NMR spectroscopy data for lycopsamine
(1), intermedine (2), echinatine (3) and rinderine (4) enabled
unambiguous confirmation of structures and assignment of
resonance signals (Tables 2 and 3) (see online Supporting
information for copies of the NMR spectra). Although most of
the assignments were consistent with previous literature reports
(Jones et al., 1982; Zalkow et al., 1985; Asibal et al., 1989;
Wiedenfeld and Roeder, 1991), the current two-dimensional
NMR experiments enabled an unambiguous reassignment for
the methyl carbons 14, 16 and 17 of 1–4 relative to published
data (Jones et al., 1982; Asibal et al., 1989; Wiedenfeld and
Roeder, 1991).
The structures and absolute configurations of 1–4, in particu-
lar those of the esterifying viridifloric and trachelanthic acids,
have been confirmed using X-ray diffraction crystallography or
by synthesis (Mackay et al., 1983; Zalkow et al., 1985; Asibal
1
13
et al., 1989). A comparison of the H- (Fig. 4) and C- (Fig. 5)
NMR spectra of the four diastereoisomers highlights some
potentially diagnostic chemical shift differences within the two
pairs of C13 epimers (i.e. 1/2 and 3/4) and the two pairs of C7
epimers (i.e. 1/3 and 2/4). The former relates to the difference
The use of the large glass column, adapted to the Biotage
Isolera 1 system, allowed for a higher scale of separation than
Phytochem. Anal. 2014
Published 2014. This article is a U.S. Government work and is in
the public domain in the USA.
wileyonlinelibrary.com/journal/pca

S. M. Colegate et al.
Figure 3. The HPLC–ESI(+)/MS ion chromatograms (left column) of pooled fractions from semi-automated separation of dehydropyrrolizidine
alkaloids on boronated soda glass beads. In the ion chromatograms, the peak marked with an ‘*’ is the internal standard, lasiocarpine. Also shown
are the MS/MS data (right column) for the major components (see also Table 1).
1
Table 2. H-NMR spectroscopy data for the diastereoisomeric monoester dehydropyrrolizidine alkaloids lycopsamine (1),
a
intermedine (2), echinatine (3) and rinderine (4)
b
H
2
Lycopsamine
Intermedine
Echinatine
Rinderine
5.92, s
5.93, s
5.65, s
5.67, s
3
3
5
5
6
6
7
8
9
9
1
1
1
1
1
d
3.96, sm
3.33, m
3.44, dm
2.77, dm
ca. 2.0, sm
ca. 2.0, sm
4.32, bs
3.92, bd, J = 15.5
3.26, m
3.41, dm
2.72, dm
1.96, sm
1.96, sm
4.25, bm
4.16, bm
4.83, d, J = 12.7
4.75, d, J = 12.7
4.08, q, J = 6.4
1.18, d, J = 6.4
2.02, sept, J = 6.9
0.92, d, J = 6.9
0.91, d, J = 6.9
3.91, bdm
3.29, bdm
3.86, sm
u
d
u
d
u
3.29, bdm
ca. 3.2, m
2.58, dm
1.94, sm
1.80, m
3.21, dt, J = 10.6, J = 6.3
d
t
2.59, bm
1.91, m
1.80, m
4.11, td, J
d
= 3.7, J
t
= 5.8
4.10, sm
3.87, sm
4.17, bs
ca. 3.96, sm
d
u
3
4
5
6
7
4.86, d, J = 13.6
4.76, d, J = 14.2
ca. 4.0, sm
1.25, d, J = ca. 9
2.16, sept, J = 6.8
0.93, d, J = 6.8
0.88, d, J = 6.8
5.01, d, J = 13.2
4.78, d, J = 13.5
ca. 3.95, sm
1.29, d, J = 6.6
2.16, sept, J = 6.8
0.90, d, J = 6.8
0.85, d, J = 6.7
4.91, d, J = 13.0
4.84, d, J = 13.0
4.06, sm
1.18, d, J = 6.4
2.02,sm
0.92, d, J = 6.9
bd = broad doublet; bdm = broad doublet of multiplets; bm = broad multiplet; bs = broad singlet; d = doublet; dm = doublet of
multiplets; dt = doublet of triplets; m = multiplet; q = quartet; s = singlet; sept = septet; sm = superimposed multiplet; td = triplet
of doublets.
a
All assignments were supported by two-dimensional COSY and/or HSQC experiments and the spectra are included online as
Supporting information.
b
‘u’ and ‘d’ represent the upfield and downfield protons of a geminal pair.
wileyonlinelibrary.com/journal/pca
Published 2014. This article is a U.S. Government work and is in
the public domain in the USA.
Phytochem. Anal. 2014

Isolation of Lycopsamine and Intermedine
1
3
Table 3. C-NMR spectroscopy data for the diastereoiso-
in the esterifying acid, whereas the latter is the difference
between the heliotridine and retronecine-based DHPA
monoesters. 1
meric monoester dehydropyrrolizidine alkaloids lycopsamine
1), intermedine (2), echinatine (3) and rinderine (4)
a
(
From the H-NMR data, an R configuration about C13 is
indicated by a slight upfield chemical shift of the C14 methyl
protons of 2 (δ 1.18) and 4 (δ 1.18) relative to those with an S
configuration, that is, 1 (δ 1.24) and 3 (δ 1.29). Additionally, the
C17 methyl protons on the iso-propyl group are shifted slightly
downfield in 2 (δ 0.91) and 4 (δ 0.92) relative to those with an
S configuration, that is, 1 (δ 0.88) and 3 (δ 0.85) such that the
signal is superimposed (i.e. 4) or barely resolved (i.e. 2) from their
geminal C16 methyl protons. It has been reported that a
difference in chemical shift (Δδ) of the C9 protons of 0–0.2
ppm indicates a C12 S configuration, whereas a Δδ of ≥ 0.25
ppm indicates a C12 R configuration (Wiedenfeld and Roeder,
C
Lycopsamine Intermedine Echinatine Rinderine
1
2
3
5
6
7
8
9
132.74
130.08
62.85
53.94
36.3
71.11
78.52
62.65
174.54
83.71
71.06
17.31
32.46
16.04
17.80
132.69
131.02
62.88
53.95
36.32
70.87
78.78
62.56
175.41
83.16
69.3
136.23
125.88
62.12
54.32
33.74
74.46
79.81
62.0
174.09
84.18
71.67
17.34
32.33
15.82
17.93
136.13
126.73
61.86
54.2
33.9
74.87
79.88
62.45
175.37
83.49
69.33
17.25
33.08
17.19
17.07
1
1
1
1
1
1
1
1
2
3
4
5
6
7
1
991). This general rule is supported by the H9 Δδ observed
for 1, 2 and 4 but, with a Δδ of 0.23 ppm, is ambiguous for 3
Table 2). Although the Δδ of the C9 protons between the C13
17.33
33.09
17.21
17.02
(
S and R epimers is very pronounced for the heliotridine-based
epimeric pair 3 and 4, it is not so for the retronecine-based
C13 epimers 1 and 2 (Fig. 4) and thus should be considered
cautiously and only in concert with other NMR stereochemical
indicators.
b
b
a
All assignments were supported by two-dimensional COSY
and/or HSQC experiments and the spectra are included
online as Supporting information.
A clear resolution of the C14, C16 and C17 methyl carbons in
13
b
the C-NMR spectrum was observed with the C13 S epimers 1
and 3 compared with the C13 R epimers 2 and 4 (Fig. 5). This
improved resolution is due to a significant downfield shift of
Assignments based on comparison with intermedine but
can be interchanged.
1
4
16, 17
heliotridine C13 S
A
2
7
8
15
6d,u
9
d
9u
5u
1
4
1
6, 17
heliotridine C13 R
B
C
2
9d, 9u
7 8
5
u
15 6d,u
1
4
1
6, 17
retronecine C13 S
2
2
9d, 9u
6d,u
7
8
5u
15
1
4
1
6, 17
retronecine C13 R
D
E
9
d, 9u
6d,u
7
8
5u
15
heliotridine C13 R
6
.0 5.8 5.6 5.4 5.2 5.0 4.8 4.6 4.4 4.2 4.0 3.8 3.6 3.4 3.2 3.0 2.8 2.6 2.4 2.2 2.0 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4
f1 (ppm)
1
Figure 4. The H-NMR spectra for: (A) echinatine (3), (B) rinderine (4), (C) lycopsamine (1), (D) intermedine (2) and (E) heliotrine (5). For assignments
see Table 2, but some key assignments described in the text are annotated on the spectra. Also annotated for each spectrum are the necine base and
C13 absolute stereochemistry for ease of comparison.
Phytochem. Anal. 2014
Published 2014. This article is a U.S. Government work and is in
the public domain in the USA.
wileyonlinelibrary.com/journal/pca

S. M. Colegate et al.
A
heliotridine C13 S
heliotridine C13 R
1
2
8
8
7
17, 14, 16
1
3
6
B
C
1
14, 16, 17
2
7
13
6
17, 14, 16
13, 7
8
6
retronecine C13 S
1
1
2
14, 16, 17
retronecine C13 R
heliotridine C13 R
D
E
8
7
13
6
2
180
170
160
150
140
130
120
110
100
90
80
70
60
50
40
30
20
10
f1 (ppm)
1
3
Figure 5. The C-NMR spectra for: (A) echinatine (3), (B) rinderine (4), (C) lycopsamine (1), (D) intermedine (2) and (E) heliotrine (5). For assignments
see Table 3, but some key assignments described in the text are annotated on the spectra. Also annotated for each spectrum are the necine base and
C13 absolute stereochemistry for ease of comparison.
C17 and a significant upfield shift of C16, while the chemical
(δ ca. 132.72, 78.65, 71, 130.5 and 36.3 respectively) (Fig. 5,
Table 3), may also be diagnostic for the heliotridine-based
alkaloids including, as shown in this study, heliotrine.
shift of C14 remained fairly similar for all of the diastereoisomers
13
(
Table 3). Additionally, the C-NMR chemical shift of C13 is
diagnostically downfield for the C13 S-configured lycopsamine
1) and echinatine (3), that is, δ 71.06 and 71.67 respectively,
(
NMR comparison of the N-oxides of lycopsamine and
intermedine
compared with the C13 R-configured intermedine (2) and
rinderine (4), that is, δ 69.3 and 69.33 respectively.
The applicability of these observations is limited to the
esterified viridifloric and trachelanthic acid epimers, because
Despite the increased water solubility imparted to a DHPA by N-
oxidation, the N-oxides of both lycopsamine and intermedine
1
13
extrapolation to, for example, the H- (Fig. 4) and C- (Fig 5)
NMR spectra of heliotrine (5) (Fig. 1) would result in incorrect
conclusions. Heliotrine has been shown to be an ester of
heliotric acid (methylated trachelanthic acid), that is, with a
C12 S/C13 R configuration (Kochetkov et al., 1969). However,
isolated from other considerations, the large Δδ (ca. 0.44 ppm)
of the C9 protons would indicate an incorrect C12 R configura-
tion. Furthermore, the downfield shift of C13 (δ ca. 78) and the
clear resolution of the C16 and C17 methyl protons might
indicate a C13 S configuration rather than the established R
configuration.
Potentially diagnostic for the configuration of the necine base
are the chemical shifts of H2, H7, H8 and H6u that in the
heliotridine-based 3 and 4 are all significantly upfield of their
retronecine counterparts in 1 and 2 (Fig. 4, Table 2). Further-
more, significant downfield shifts of C1(δ ca. 136), C8 (δ ca.
were remarkably soluble in CHCl . This meant that the more
3
usual approach to synthesising these N-oxides, that is, using a
solution of ethanol and CHCl from which the newly formed N-
3
oxide precipitates (Molyneux et al., 1991), was not applicable.
Therefore, the reaction solutions were processed using SCX
SPE from the aqueous acid-soluble fraction of the reaction
1
13
products. The H- and C-NMR data for lycopsamine-N-oxide
(Table 4), acquired using a solution in CDCl , are similar to
published data acquired using D O (Roeder and Bourauel,
1992) and 10% D O in CDCl (Nishida et al., 1991). Apart from
differences in some chemical shifts (up to about 2.7 ppm in
3
2
2
3
13
1
the C data and about 0.24 ppm in the H data), presumably
related to solvent effects, the major apparent discrepancies
involve the current observation of a singlet and a broad
multiplet for the geminal protons on C3 and C5, respectively,
compared with two doublets when recorded in D
2
O and a
7
1
9.85) and C7 (δ ca. 74–75), and upfield shifts of C2 (δ ca.
26.3) and C6 (δ ca. 33.8), relative to their retronecine analogues
doublet (C3 protons) and broad multiplet (C5 protons) when
recorded in 10% D
2
O in CDCl
3
. In addition, when acquired in
wileyonlinelibrary.com/journal/pca
Published 2014. This article is a U.S. Government work and is in
the public domain in the USA.
Phytochem. Anal. 2014

Isolation of Lycopsamine and Intermedine
1
13
Table 4. H- and C-NMR spectroscopy data for the N-oxides of the epimeric monoester dehydropyrrolizidine
a
alkaloids lycopsamine (1) and intermedine (2)
b
Carbon
Lycopsamine-N-oxide
Intermedine-N-oxide
1
13
1
13
H
C
H
C
1
2
NA
5.79
4.43 (s, 2H)
132.89
122.30
78.21
NA
5.80 (s)
4.42 (s, 2H)
132.95
122.25
78.14
3
3
5
5
6
6
7
8
9
9
1
1
1
1
1
1
1
d
u
d
u
d
u
3.77 (bm, 2H)
69.51
34.72
3.72 (m, 2H)
69.38
34.79
2.58 (bm)
2.01 (bm)
4.70 (sbs)
4.77 (sbs)
4.86 (d, J9d.9u = ca. 12)
4.76 (sd, J9u.9d = ca. 12)
NA
NA
3.86 (q, J13,14 = 6.6)
1.24 (d, J14,13 = 6.5)
2.14 (sept, J15,16/17 = 6-7)
0.89 (d, J16,15 = 6.7)
0.84 (d, J17,15 = 6.7)
2.55 (m)
1.98 (bd, J6u,6d = 14)
4.66 (bs)
4.76 (bs)
4.84 (s, 2H)
69.89
96.43
61.50
69.62
96.0
61.4
d
u
1
2
3
4
5
6
7
174.21
84.70
72.87
17.31
32.33
15.65
18.18
NA
NA
4.11 (q, J13,14 = 6.3)
1.17 (d, J14,13 = 6.4)
1.85 (sept, J15,16/17 = 6-7)
0.97 (d, J16,15 = 7)
174.93
84.22
70.08
16.80
33.92
17.25
17.75
0.91 (d, J17,15 = 6.8)
NA = not applicable; bd = broad doublet; bm = broad multiplet; bs = broad singlet; d = doublet; m = multiplet; q =
quartet; s = singlet; sbs = superimposed broad singlet; sd = superimposed doublet; sept = septet.
a
All assignments were supported by two-dimensional COSY and/or HSQC experiments and the spectra are in-
cluded online as Supporting information.
b
‘u’ and ‘d’ represent the upfield and downfield protons of a geminal pair.
22
1
0% D
2
O in CDCl
3
the C7 and C8 protons are reported as a single
D
(Frahn et al., 1980). Luber et al. (2012) report [α] = +12.0°
peak. However, despite these differences, the current assign-
ments based on the two-dimensional homonuclear (COSY) and
heteronuclear (HSQC) correlation experiments were in accord
with previous data and confirmed the structure.
There seems to be only two reports of NMR data for
intermedine-N-oxide, one of a synthetic intermedine-N-oxide
(c = 0.01, methanol) for intermedine-N-oxide, but there do not
seem to be values reported for lycopsamine-N-oxide. Further-
more, Souza et al. (2005) isolated lycopsamine from Heliotropium
20
indicum and reported [α] = +0.12° (c = 0.05, chloroform).
D
2
2
In the present study the [α] for lycopsamine, intermedine,
D
lycopsamine-N-oxide and intermedine-N-oxide were: À1.4°
(c = 0.98, methanol), +3.7° (c = 0.81, methanol), +10.6° (c = 1.04,
methanol), +7.3° (c = 0.98, methanol) respectively. Obviously,
despite the MS and NMR evidence for lycopsamine herein, the
levorotatory value for the lycopsamine solution was a concern
because all reported values have been dextrorotatory, even the
very small value determined for the chloroform solution of
lycopsamine (Souza et al., 2005). Therefore, to investigate poten-
tial levorotatory contamination of the lycopsamine sample,
specific rotations were determined for another two samples of
lycopsamine isolated using the boronated soda glass bead
column and for lycopsamine isolated from a lycopsamine
chemotype of Amsinckia intermedia (Colegate et al., 2013) that
had not been exposed to the borax solutions. In each case the
(Nishimura et al., 1987) and one of a natural compound isolated
from Cerinthe glabra (Luber et al., 2012). Using two-dimensional
heteronuclear NMR experiments, including HMQC, Luber et al.
(2012) were able to correct some of the earlier assignments by
Nishimura et al. (1987), however, they could not unequivocally
assign C7 and C13. The current NMR data (Table 4) were similar
to those of Luber et al. (2012) except that the C9 protons
1
resonated as a single peak (evident in both the H and HSQC
spectra) at 4.84 ppm in contrast to the doublet and a multiplet
at 4.84 and 4.97 ppm, respectively, observed by Luber et al.
(2012). However, the use of the HSQC experiment for determin-
ing direct C–H coupling allowed for a higher resolving power in
13
the C direction than the HMQC experiment used by Luber et al.
2012) (Mandal and Majumdar, 2004) and, thereby, confirmed
2
2
(
[α]
D
was À1.56° (c = 2.5, methanol), À1.5° (c = 1.4, methanol)
their tentative assignment of C3 and C13.
and À1.8° (c = 1.6, methanol), respectively, thereby confirming
the levorotatory optical rotation of a methanolic lycopsamine
solution.
Optical rotation observations
Subsequently, to better compare with literature reports, the
lycopsamine and intermedine were dissolved in ethanol.
Somewhat surprisingly, despite the recognition that the magni-
tude, and sometimes the sign, of the optical rotation is affected
by both solvent and temperature (Gottarelli et al., 1991), there
Reported measurements of specific rotation ([α]) for ethanolic
solutions of lycopsamine vary from +1.2° (Zalkow et al., 1985)
to +5.7° (Frahn et al., 1980). For intermedine, also in ethanolic so-
lution, they vary from +4.8° (Culvenor and Smith, 1966) to +9.8°
Phytochem. Anal. 2014
Published 2014. This article is a U.S. Government work and is in
the public domain in the USA.
wileyonlinelibrary.com/journal/pca

S. M. Colegate et al.
was a significant dextrorotatory effect on the optical rotations
measured and, consequently, the specific rotations calculated
when changing the solvent from methanol to ethanol. Thus for
Jones AJ, Culvenor CCJ, Smith LW. 1982. Pyrrolizidine alkaloids – a
carbon-13 N.M.R. study. Aust J Chem 35: 1173–1184.
Kelley RB, Seiber JN. 1992. Pyrrolizidine alkaloid chemosystematics in
Amsinckia. Phytochemistry 31: 2369–2387.
Kim N-C, Oberlies NH, Brine DR, Handy RW, Wani MC, Wall ME. 2001.
Isolation of symlandine from the roots of common comfrey
2
2
lycopsamine [α] = +3.75° (c = 2.4, ethanol) and for intermedine
D
2
2
[
α] = +6.27° (c = 0.75, ethanol). The change in sign of the optical
D
(Symphytum officinale) using countercurrent chromatography. J Nat
rotation for lycopsamine was specifically confirmed by evaporat-
ing the ethanolic solution to dryness under a stream of nitrogen,
and redissolving the residue in methanol for measurement of its
optical rotation.
Prod 64: 251–253.
Kochetkov NK, Likhosherstov AM, Kulakov VN. 1969. The total synthesis
of some pyrrolizidine alkaloids and their absolute configuration.
Tetrahedron 25: 2313–2323.
Luber M, Musa A, Kadry HA, Bracher F. 2012. Isolation of the pyrrolizidine
alkaloid intermedine-N-oxide from Cerinthe glabra and ab initio
Acknowledgements
1
3
calculation of its C NMR shifts. Z Naturforschung 67B: 411–416.
The authors thank the anonymous reviewers for their comments,
suggestions and constructive critique.
Mackay MF, Sadek M, Culvenor CCJ. 1983. Lycopsamine and intermedine,
15 5
C H25NO : Diastereoisomeric pyrrolizidine alkaloids. Acta Cryst C39:
7
85–788.
Mandal PK, Majumdar A. 2004. A comprehensive discussion of HSQC and
HMQC pulse sequences. Concepts Magn Resonan A 20A: 1–23.
Molyneux RJ, Johnson AE, Olsen JD, Baker DC. 1991. Toxicity of
pyrrolizidine alkaloids from Riddell groundsel (Senecio riddellii) to
cattle. Amer J Vet Res 52: 146–151.
Molyneux RJ, Gardner DR, Colegate SM, Edgar JA. 2011. Pyrrolizidine
alkaloid toxicity in livestock: a paradigm for human poisoning? J Food
Additiv Contamin 28: 293–307.
Nishida R, Kim C-S, Fukami H, Irie R. 1991. Ideamine N-oxides:
Pyrrolizidine alkaloids sequestered by the Danaine butterfly, Idea
leuconoe. Agric Biol Chem 55: 1787–1792.
Nishimura Y, Kondo S, Takeuchi T, Umezawa H. 1987. Total syntheses of
indicine N-oxide, intermedine N-oxide, and their enantiomers using
carbohydrate as a chiral educt. Bull Chem Soc Jpn 60: 4107–4113.
Roeder E, Bourauel T. 1992. Pyrrolizidine alkaloids from Neatostema
apulum. Phytochemistry 31: 3613–3615.
References
Adams R, Hamlin Jr KE, Jalinek CF, Phillips RF. 1942. The structure of
riddelliine, the alkaloid in Senecio riddellii. I. J Am Chem Soc 64:
2760–2763.
Asibal CF, Glinski JA, Gelbaum LT, Zalkow LH. 1989. Pyrrolizidine alkaloids
from Cynoglossum creticum. Synthesis of the pyrrolizidine alkaloids
echinatine, rinderine, and analogues. J Nat Prod 52: 109–118.
Betteridge K, Cao Y, Colegate SM. 2005. An improved method for extrac-
tion and LCMS analysis of pyrrolizidine alkaloids and their N-oxides in
honey: Application to Echium vulgare honeys. J Agric Food Chem 53:
1894–1902.
Bull LB, Culvenor CCJ, Dick AT. 1968. In The Pyrrolizidine Alkaloids: Their
Chemistry, Pathogenicity and other Biological Properties. North-Hol-
land: Amsterdam.
Colegate SM, Edgar JA, Knill AM, Lee ST. 2005. Solid phase extraction and
LC–MS profiling of pyrrolizidine alkaloids and their N-oxides: a case
study of Echium plantagineum. Phytochem Anal 16: 108–119.
Colegate SM, Gardner DR, Joy RJ, Betz JM, Panter KE. 2012.
Dehydropyrrolizidine alkaloids, including monoesters with an
unusual esterifying acid, from cultivated Crotalaria juncea (Sunn
Hemp cv.‘Tropic Sun’). J Agric Food Chem 60: 3541–3550.
Colegate SM, Gardner DR, Davis TZ, Welsh SL, Betz JM, Panter KE. 2013.
Identification of a lycopsamine-N-oxide chemotype of Amsinckia
intermedia. Biochem Syst Ecol 48: 132–135.
Cooper RA, Bowers RJ, Beckham CJ, Huxtable RJ. 1996. Preparative
separation of pyrrolizidine alkaloids by high-speed counter-current
chromatography. J Chromat A 732: 43–50.
Culvenor CCJ. 1962. Senecio magnificus F. Muell., a source of senecionine.
Aust J Chem 15: 158–159.
Culvenor CCJ, Smith LW. 1966. The alkaloids of Amsinckia species: A.
intermedia Fisch. & Mey., A.hispida (Ruiz. & Pav.) Johnst. and A.
lycopsoides Lehm. Aust J Chem 19: 1955–1964.
Edgar JA, Colegate SM, Boppré M, Molyneux RJ. 2011. Pyrrolizidine alka-
loids in food: a spectrum of potential health consequences. J Food
Additiv Contamin 28: 308–324.
Roeder E, Breitmaier E, Birecka H, Frohlich MW, Badzies-Crombach A.
1
991. Pyrrolizidine alkaloids of Heliotropium spathulatum. Phytochem-
istry 30: 1703–1706.
Roitman JN. 1983. The pyrrolizidine alkaloids of Amsinckia menziesii. Aust
J Chem 36: 769–778.
Souza JSN, Machado LL, Pessoa ODL, Braz-Filho R, Overk CR, Yao P,
Cordell GA, Lemos TLG. 2005. Pyrrolizidine alkaloids from
Heliotropium indicum. J Braz Chem Soc 16: 1410–1414.
Wiedenfeld H, Roeder E. 1991. Pyrrolizidine alkaloids from Ageratum
conyzoides. Planta Med 57: 578–579.
Williams MT, Warnock BJ, Betz JM, Beck JJ, Gardner DR, Lee ST, Molyneux
RJ, Colegate SM. 2011. Detection of high levels of pyrrolizidine-N-ox-
ides in the endangered plant Cryptantha crassipes (Terlingua Creek
Cat’s-eye) using HPLC–ESI-MS. Phytochem Anal 22: 532–540.
Zalkow LH, Glinski JA, Gelbaum LT, Fleischmann TJ, McGowan LS, Gordon
MM. 1985. Synthesis of pyrrolizidine alkaloids indicine, intermedine,
lycopsamine, and analogues and their N-oxides. Potential antitumor
agents. J Med Chem 28: 687–694.
Frahn JL, Culvenor CCJ, Mills JA. 1980. Preparative separation of the
pyrrolizidine alkaloids, intermedine and lycopsamine, as their borate
complexes. J Chromatogr 195: 379–383.
Gottarelli G, Osipov MA, Spadat GP. 1991. A study of solvent effect on the
optical rotation of chiral biaryls. J Phys Chem 95: 3879–3884.
Supporting information
Supporting information can be found in the online version of
this article.
wileyonlinelibrary.com/journal/pca
Published 2014. This article is a U.S. Government work and is in
the public domain in the USA.
Phytochem. Anal. 2014