|[Related articles/posters: 119 018 077 ]|
Protoberberine alkaloids display a broad spectrum of biological
activities.  The quaternary amine
derivatives, such as berberine (1) and coralyne (2),
exhibit antimicrobial and antitumor activities which are attributed
to their ability to intercalate with DNA. 
In common with other intercalation agents, this complexation
is attributed to their planar structure, a property which facilitates
intercalation and subsequent p-stacking
with base pairs. It has been established that berberine and coralyne
bind selectively for A-T rich sites. 
Recently, we explored the phthalide anion-imine reaction
for the synthesis of a series of fluorine labelled 13-hydroxy-8-oxo
protoberberines where the fluorine nucleus could be used as a
NMR probe to study intercalation. 
In order to improve the intercalation potential of these alkaloids,
which molecular modelling shows are non planar systems, we sought
ways to convert them to planar systems. In this paper, we report
our initial study to synthesise fully aromatised, planar protoberberine-8-ones
4 from readily available 13-hydroxy-8-oxo derivative 3
Results and Discussion
The starting product in our approach was the protoberberine-8-one
5, available by dehydration of the 13-hydroxy-8-oxo protoberberine
3. It is known that when treated with the strong oxidant
like lead tetraacetate, protoberberines of type 5 gave
polyoxygenated products, which serve as intermediates in the conversion
to other isoquinoline alkaloids. 
To avoid this undesired outcome, mild dehydrogenation reagents
were studied. The common reagents, like MnO2,
I2, 10% Pd/C and Hg(OAc)2,
could not effect the dehydrogenation, while DDQ gave a mess, without
evidence for formation of the desired product 4 being detected
by TLC and 1H NMR.
To achieve the dehydrogenation, we tested the old classical method using sulfur. However, under melted conditions, such as at 210 oC for 0.5 hours, the reactant 5 was recovered quantitatively. When the mixture of 5 and sulfur was heated at higher temperature with a Bunsen burner until hydrogen sulfide (rotten egg smell) was produced and foaming was finished, the desired product 4 was obtained and isolated (Scheme 2). A solution of 4 showed characteristic green fluorescence under daylight similar to that of 9-aminoacridine, a well-known DNA intercalation agent. The structure of 4 was determined by spectroscopic analysis. Typically, the 1H NMR showed low-field shifts for three aromatic singlets H1, H13 and H4 (Dd 0.25, 0.37 and 0.10 ppm, compared with reactant 5), owning to proximity between H1 and H13, H4 and H5 after a double bond is introduced between C5 and C6, or deshielding (anisotropic) effect of the fully aromatised, planar structure.
A two-step transformation involving dehydrobromination of the
preformed 5-bromo intermediate 7 was also studied. However,
treatment of 5 with NBS in the presence of a catalytic
amount of benzoyl peroxide, caused bromination to occur regiospecifically
at position-13 of the aromatic ring to produce 6, instead
of the desired benzylic position of the B-ring (C-5) (Scheme
3). The same 13-bromo product 6 was also formed from
5 under typical electrophilic reaction conditions, such
as bromine or pyridinium bromide perbromide (PBP) in pyridine.
The regioselective acetoxylation (lead tetraacetate)  or nitration (10% aqueous HNO3)  at position-13 of compound 5 type of structures have been reported in the literature. The regioselectivity is attributed to the effect of the methoxy groups in the A-ring of compound 5 promoting electron enrichment at position-13 via resonance hydrid contributors such as quininoid structure 8; participation of the ring nitrogen via iminium structure 9 occur in the same sense but are less likely.
The structure of bromo derivative 6 was determined by spectroscopic analysis. Typically, EIMS (m/e 79Br 385, relative intensity 68% and 81Br 387, 68%) and HRMS (m/e 79Br 385.0322, 100% and 81Br 387.0306, 100%) indicated the molecular formula C19H16BrNO3 (required m/e 79Br 385.0314, 51% and 81Br 387.0293, 49%). Compared with reactant 5, the 1H NMR indicated the disappearance of one aromatic singlet. It was established by a COSY experiment that the aromatic singlets at d 7.91 and 6.77 correlated with the methoxy protons (d 3.95, s, 6H) and only proton d 6.77 with methylene group at position-5 (d 2.87, t) allowing assignment of the resonance at d 7.91 to H1 and d 6.77 to H4. Similar long-range correlations were established for the dimethoxyisoquinoline structures 4-6. Two-dimensional COSY experiment was used to detect cross peaks due to couplings that are less than the observed line width such as four or five bond couplings (J = 0.1-0.5 Hz) and hence unresolvable in the 1D 1H-NMR.  Owing to the presence of bromine atom at position-13 in 6, H1 (Dd 0.62) and H12 (Dd 0.55 ppm) were strongly deshielded compared with reactant 5.
Treatment of 5 with selenium dioxide (1.15 equiv) in acetic
acid at reflux gave a mixture of products; the major product was
isolated and shown to be the unusual seleno heterocycle 10
(Scheme 4). The structure of 10 was determined
by spectroscopic and spectrometric analysis. EIMS (m/e
78Se 383, 34% and 80Se
385, 72%) and HRMS (m/e 78Se 383.0142,
42% and 80Se 385.0219, 100%) indicated the
characteristic isotope distribution of one selenium atom in the
molecular formula C19H15NO3Se
(required m/e 78Se 383.0225, 24% and 80Se
385.0217, 50%). Compared with starting material 5, the
1H NMR indicated the disappearance of two
aromatic singlets, with only one singlet at d
6.92 being observed. This singlet correlated with one of the
methoxy resonances (d 3.97, s, 3H)
and the methylene group at position-5 (d
3.25, t) in COSY. This clearly indicated it was proton H4 and
selenium insertion must have occurred at positions 1 and 13.
The mechanism for the formation of 10 is suggested involving
the electrophilic addition of selenium dioxide into C-13 and C-1
to form 12, which undergoes an unknown reductive pathways
to eliminate the oxygen on selenium.
Total assignments of 1H and 13C NMR data for the products 4, 6 and 10 were achieved by various 1D and 2D NMR experiments (such as COSY, HMQC, HMBC) and fully supported the assigned structures. The key proton-carbon long range correlations, which facilitated the total assignments and established the important connections in these molecules, are summarised in Figure.
We have synthesised the first fully aromatised, planar protoberberine-8-one
species 4. The regioselective bromination and selenation
of 5 were discovered. We are currently investigating the
role of these planar protoberberines 4 as intercalation
units and incorporating them into new bis-intercalator systems.
Melting point were measured with a Gallenkamp melting pointing
apparatus in open capillaries and were uncorrected. Mass spectra
(EI, 70 ev) were recorded on Shimadzu GC MS-QP2000A Spectrometer.
High resolution mass spectra (HRMS-EI) were recorded on a Micromass
AutoSpec spectrometer at 70 ev. NMR spectra, 1H,
13C (Proton Broad Band Decoupling), HMQC (1H-detected
Heteronuclear Multiple Quantum Coherence) and HMBC (1H-detected
Heteronuclear Multiple Bond Connectivity) experiments, were recorded
on a Bruker AMX 300 MHz Spectrometer in deuteriochloroform solution.
Proton-carbon long-range correlation experiments (HMBC) were
parameterised for JCH of 10 Hz. Chemical
shifts (d) are reported in ppm downfield
from internal tetramethylsilane (TMS). TLC was conducted using
Merck silica gel 60 F254 pre-coated aluminium sheets. Preparative
TLC (25x25 cm) or Chromatotron plates were prepared using Merck
silica gel 60 PF254. Column chromatography was operated on Merck
silica gel 60.
8-Oxoprotoberberine 5 (23 mg, 0.075 mmol) was well mixed
with sulfur (158 mg, 4.94 mmol) in a test tube. The mixture was
heated in a yellow flame of a Bunsen burner until the solid completely
melted and kept at this state for a few seconds. Heating by blue
flame (higher temperature) until foaming started then settled
down. A rotten egg smell of H2S was detected.
After cooling, the black solid was extracted with DCM and separated
by chromatotron (petrol-EtOAc 2:1 to 1:1) to give dehydrogenated
product 4 as bright yellow crystals (23 mg, 100%). mp
213-215 oC; 1H NMR:
8.61 (1H, d, J = 7.8 Hz, H6), 8.52 (1H, d, J
= 8.5 Hz, H9), 7.63-7.70 (2H, m, H12 and H11), 7.54 (1H, s, H1),
7.44 (1H, ddd, J = 8.3, 5.8, 2.4 Hz, H10), 7.25 (1H, s,
H13), 6.85 (1H, s, H4), 6.68 (1H, d, J = 7.8 Hz, H5), 4.06
(3H, s, 2-OCH3), 3.98 (3H, s, 3-OCH3);
13C NMR: 160.60
(C, 8), 151.94 (C, 3), 150.33 (C, 2), 136.69 (C, 12a), 135.79
(C, 13a), 133.07 (CH, 11), 128.85 (CH, 9), 126.40 (CH, 12), 125.84
(CH, 10), 125.32 (C, 4a), 122.57 (CH, 6), 121.24 (C, 8a or 13b),
121.18 (C, 8a or 13b), 112.73 (CH, 5), 108.48 (CH, 4), 105.60
(CH, 1), 97.46 (CH, 13), 56.88 (CH3, 2-OCH3),
56.72 (CH3, 3-OCH3).
EIMS (70 ev, m/e, relative intensity): 305 (M+,
100), 262 (31), 261 (35), 219 (11), 153 (21). HRMS: C19H15NO3
required 305.1052, found 305.1047.
Method A (with NBS):
To a solution of 8-oxoprotoberberine 5 (42 mg, 0.137 mmol)
in carbon tetrachloride (4 mL) was added benzoyl peroxide (a few
crystals) and NBS (32 mg, 0.180 mmol) in that order. The mixture
was stirred under reflux for 3.5 h, while monitored by TLC (petrol-EtOAc
1:1). After completion, the mixture was cooled and transferred
into a separation funnel. The flask was rinsed with DCM. The
combined organic solution was washed with water, dried with Na2SO4,
concentrated and separated by column chromatography (petrol-EtOAc
1:1). Concentration of the fraction (Rf = 0.58, petrol-EtOAc
1:1) gave 13-bromoprotoberberine 6 as yellow needles (53
mg, 100%). mp 138.5-140 oC. 1H NMR: 8.46
(1H, d, J = 7.9 Hz, H9), 8.11 (1H, d, J = 8.2 Hz,
H12), 7.91 (1H, s, H1), 7.73 (1H, t, J = 7.8 Hz, H11),
7.52 (1H, t, J = 7.5 Hz, H10), 6.77 (1H, s, H4), 4.72 (2H,
t, J = 5.8 Hz, H6), 3.95 (6H, s, 2xOCH3),
2.87 (2H, t, J = 5.8 Hz, H5); 13C NMR:
161.81 (C, 8), 150.86 (C, 3), 147.03 (C, 2), 136.93 (C, 12a or
13a), 136.90 (C, 12a or 13a), 133.61 (CH, 11), 133.19 (C, 4a),
128.63 (CH, 9), 127.86 (CH, 10), 127.51 (CH, 12), 125.37 (C, 8a),
122.76 (C, 13b), 115.64 (CH, 1), 110.34 (CH, 4), 100.01 (C, 13),
56.90 (CH3, OCH3), 56.64
(CH3, OCH3), 42.64 (CH2,
6), 29.86 (CH2, 5). EIMS: 387 (M+,
81Br, 68), 385 (M+, 79Br,
68), 372 (98), 370 (100), 326 (12), 292 (12), 262 (11), 248 (14),
232 (12). C19H16BrNO3:
required m/e 79Br 385.0314, 51% and 81Br
387.0293, 49%; found 79Br 385.0322, 100% and
81Br 387.0306, 100%.
Method B (with bromine):
To a solution of bromine (10 mg, 0.062 mmol, 1.05 equiv.) in dry
pyridine (3 mL) was added 8-oxoprotoberberine 5 (18 mg,
0.059 mmol). The mixture was stirred in an oil bath (80 oC) for
15 h. After cooling, the solution was poured into a mixture of
crushed ice (50 g) and 0.5 M HCl (50 mL) and extracted with dichloromethane
(2x15 mL). The combined dichloromethane extracts were washed
with brine, dried (sodium sulphate) and concentrated to dryness
to provide a yellow solid. Recrystallisation from methanol provide
6 as yellow needle (23 mg, 100%, characterised by 1H NMR).
Method C (with pyridinium bromide perbromide, PBP):
To a solution of pyridinium bromide perbromide (49 mg, 0.154 mmol,
1.05 equiv.) in dry pyridine (3.4 mL) was added 8-oxoprotoberberine
5 (45 mg, 0.147 mmol). The mixture was stirred in an oil
bath (80 oC) for 15 h. Work-up as Method B provided 6
as yellow needle (57 mg, 100%, characterised by 1H NMR).
5,6-Dihydro-2,3-dimethoxy-8-oxo-1,13-selena-8H-dibenzo[a,g]quinolizine (10) 
A mixture of 8-oxoprotoberberine 5 (36 mg, 0.117 mmol),
selenium dioxide (20 mg, 0.180 mmol) in acetic acid (4 mL) was
stirred under reflux for 22 h, while monitored by TLC (petrol-EtOAc
1:4). Acetic acid was removed by evaporation and the residue was
separated by column chromatography (petrol-EtOAc 1:4) followed
by preparative TLC (DCM-EtOAc 5:1) to give major product 10
as slightly orange crystals (14 mg, 31%). mp 238-240 oC. 1H
NMR: 8.51 (1H, dd, J = 8.7, 1.6 Hz, H9), 7.68 (1H, td,
J = 7.6, 1.2 Hz, H10), 7.44-7.49 (2H, m, H10 and H12),
6.92 (1H, s, H4), 4.45 (2H, t, J = 6.7 Hz, H6), 4.03 (3H,
s, 2-OCH3), 3.97 (3H, s, 3-OCH3),
3.25 (2H, t, J = 6.7 Hz, H5); 13C NMR:
162.48 (C, 8), 152.41 (C, 3), 145.47 (C, 2), 135.84 (C, 12a),
134.38 (C, 13a), 133.44 (CH, 11), 130.51 (C, 1), 130.01 (CH, 9),
128.56 (C, 4a), 127.05 (CH, 10), 126.36 (C, 13b), 126.14 (C, 8a),
124.85 (CH, 12), 113.35 (C, 13), 111.06 (CH, 4), 61.00 (CH3,
2-OCH3), 57.49 (CH3,
3-OCH3), 41.40 (CH2,
6), 27.27 (CH2, 5). EIMS: 385 (M+,
80Se, 72), 383 (M+, 78Se,
34), 342 (31), 340 (26), 339 (18), 338 (18), 191 (14), 44 (100).
required m/e 78Se 383.0225, 24% and 80Se
385.0217, 50%; found 78Se 383.0142, 42% and
80Se 385.0219, 100%
The authors thank Central Queensland University and the Australian
Research Council for financial support.
References and Notes
1. Reviews of biological activities: a) Simeon, S.; Rios, J. L.; Villar, A. Plant. Med. Phytother. 1989, 23, 202-250. (Chem. Abstr. 1990, 112, 195246). b) Shamma, M. The Isoquinoline Alkaloids, Chemistry and Pharmacology; Academic Press: New York and London, 1972; Vol. 25, pp 268-314. c) Shamma, M.; Moniot, J. L. Isoquinoline Alkaloids Research, 1972-1978; Plenum Press: New York and London, 1978, pp 209-259. d) Bhakuni, D. S.; Jain, S. The Alkaloids 1986, 28, 95-181.
2. Comprehensive review on anticancer activity: Suffness, M.; Cordell, G. A. "Antitumor Alkaloids" in The Alkaloids Chemistry and Pharmacology, Vol.25, Academic Press, Inc: Orlando, San Diego, New York, London, Toronto, Montreal, Sydney, Tokyo, 1985; pp188-198.
3. Warrener R. N.; Liu L.; Russell R. A. Tetrahedron 1998, in press.
4. Dorn, C. R.; Koszyk, F. J.; Lenz, G. R. J. Org. Chem. 1984, 49, 2642-2644.
5. Cushman, M.; Chinnasamy, P.; Patrick, D. A.; McKenzie, A. T.; Toma, P. H. J. Am.. Chem. Soc. 1990, 55, 5995-6000.
6. Derome, A. E. In Modern NMR Techniques for Chemistry Research; Pergamon Press: Oxford, 1989; p 194.
7. For comparison with other protoberberine derivatives, the traditional nomenclature and numbering system for protoberberine alkaloids is used for compound 10. But according to the "IUPAC Nomenclature of Fused and Bridged Fused Ring Systems." Pure & Appl. Chem. 1998, 70, 143-216, the systematic name of 10 should be: 5a-aza-1,2-dimethoxy-11-selena-6-oxo-11H-Benzo[bc]aceanthrylene based on its hydrocarbon ring skeleton.