Synthesis and Characterization of Novel Biologically Active Platinum(I1) and Palladium(I1) Complexes of Some P-Carboline Alkaloids Talal A. K. Al-Allaf, M&dad T. Ayoub, and Luay J. Rashan Chemistry Department, College of Science.--WR. College of Education, University of Mosul, Mosul, Iraq
TAKA,
MTA.
Bhlogy
Department,
ABSTRACT The preparation of novel biologically active platinum(H) and palladium(I1) complexes of some P-carboline alkaloids (harmaline, harmalol, hannine, and hat-mane) is described. These complexes, characterized on the basis of their CHN elemental analysis, infrared, Raman, and ‘H and 13C nuclear magnetic resonance spectral data, were shown to have the empirical formula [M(alka10id)C12], M = Pt. Pd. The antitumor and antiviral activities of some of these complexes have been demonstrated.
INTRODUCTION A large body of platinum complexes and, to a lesser extent, palladium complexes have been prepared and their antitumor activity examined since the discovery of Rosenberg et al. Cl], that cisplatin compound (I) has considerable antitumor activity (the first generation). Much work was done thereafter with the aim of decreasing the toxicity of platinum compounds and increasing their activity, as well as their solubility. Hence, the second generation of analogs to cisplatin was discovered, that called carboplatin, or paraplatin (II) [2]. Other cisplatin analog complexes having amino sugars as ligands have been synthesized and tested for antitumor activity 131. Address reprint requests to: Dr. T. A. University of Mosul, Mosul, Iraq.
K. AI-Allaf,
Chemistry
Department,
College
of Science,
Journal of Inorganic Biochemistry 38 47-56 (1990)
0
1990 Elseviet’ Science Publishing
Co., Inc., 655 Avenue of the Americas,
47 NY, NY 10010
0162-0134/90!$3.50
48
T. A. K. AI-All&
et al.
OH
(1)
Cisplatin
(II)
Carboplatin (Paraplatin)
@II)
Iproplrltin
The wide range physiologic activity of @-carboline alkaloids has stimulated a wealth of different approaches to their synthesis [4,5]. Some of these compounds, harmaline (4,9-dihydro-7-mzthoxy-l-methyl-3H-pyrido[3,4-b]indole), harmalol (4,9_dihydro-lmethyl-3H-pyrido[3,4-blindole-7-ol), haxmine (7-methoxy- 1-methyl-9H-pyrido[3,4b]indolc), and harmane (1 -methyl-9H-pyrido[3,4-blindole) (IV-VII), are extremely effective as antituberculosis 161, analgesic 171, and antimicrobial [8] agents. Attempts have been made to find other chemical substances having at least as much, if not more, similar activity. especially recently; this has led to the finding that harmine (VI) has some antiviral activity [9]. As a continuation of our previous studies [lo] on the synthesis and characterization of platinum(IIj complexes of substituted 3-aryld,S-dimethyl-2-pyrazolines and as a part of an investigation of the biologic activity of these complexes, we required the synthesis of platinum(H) and palladium@) complexes of some fi-carboline alkaloids of the type 3H- (and 9H-) pyrido-[3,4-blindole (IV-VII). In our previous article (1 l] on this subject, we reported the first synthesis and characterization of platinum@) and palladium(I1) complexes of some hydrochloride derivatives of P-carboline alkaloids from the reaction of the alkaloid hydrochlorides with &MC4 (M = Pt, Pd) in water. Since the complexes we reported, i.e., trans-[M(alkaloid - HClh Clz] are rather insoluble in water and in most common organic solvents and since they decompose in DMSO, we could not manage to study their biologic activity as pure complexes. Because of this and in order to overcome these difficulties, we decided to prepare the platinum(I1) and palladium(II) complexes of the HCI-free alkaloids and to study their antitumor and antiviral activities. EXPERIMENTAL Natural abundance, proton-decoupled spectra were recorded at 25°C on a 22.63 MHz. A pulse width of 6 ps and scans (depending on the nature of the points, for a spectral width of 4,807
J?iJ--q-q Harma’o1 FIGURE
R
FT 13C nuclear magnetic resonance (NMR) Bruker-WH 90 DS spectrometer, operating at a 3-s delay were used, and approximately 3,000 compound) were accumulated with 16,000 data Hz.
1.
(IV) R = OCH3, Hannaline (V) R = OH,
Pt@) & Pd(II) COMPLEXES
H
FIGURE “3
OF SOME
2.
B-CARHOLINE
ALKALOlDS
(VI) R = OCHx, Harmine
49
(VII) R = H,
Harmane
Proton NMR spectra were recorded on the same spectrometer, using the deuterium signal of the solvent as a field lock signal. Infrared spectra were recorded on an SP 2000 spectrophotometer at a range of 200 using Nujol mull or RI disks. to 4,000 cm”, Raman spectra were recorded on solids using a Laser Raman spectrophctometer. The samples were done at Dortmund University, West Germany. Analyses of the complexes were carried out on a CHN Analyzer, type 1106 (Carlo Erba). Starting Materials The compounds &PtCb, PdCl2, harmaline, harmalol, harmine, and harmane were obtained commercially from Fluka Company. The complexes [M(DMS0)2Cl,] (M = Pt, Pd) were prepared according to the method described in the literature [12]. Preparation of Complexes All solvents used were dried and degassed under dry nitrogen.
before
use, and reactions
were carried
out
Cis-[pt (harmaline) Cr,)/ (VIII+‘. The complex cis-[Pt(DMSO)zCIZ] (0.2 g, 0.474 mmol) was suspended in chloroform (15 mL), and ham&line (0.11 g, OS 13 mmol) was added gradually at ambient temperature. The turbid solution became clear yellow, and after about 30 min, the solution was filtered through Celite. The filtrate was takeq to dryness, and yellow oily material was obtained. This was stirred with light petroleum spirts 30” to 40°C (30 mL) overnight. The mother liquor was decanted from the fine yellow crystalline product, and the solid was dried under vacuum. The product can be recrystallized from ethanol/n-hexane. The yield is almost quantitative. A similar method was used to prepare the analogous palladium complex, i.e., cis[Pd(harmaline)C12] (VIIIb). Cis-fPt (Harmalcd) ClJ
(VIZZc). The complex
cis-[Pt(DMSO)zClz ] (0.422 g, 1 .O mmol) was suspended in warm ethanol (20 mL), and a warm solution of harmalol (0.22 g, 1.1 mmol) in ethanol (10 mL) was added to the suspension under nitrogen. The solution became clear yellow. It was allowed to stand for about 30 min and then filtered through Celite to remove any unreacted starting materials. Slow evaporation of the solvent gave yellow crystals from the product. These were washed with small portions of ethanol and then with hexane and dried in vacua for several hours. This method is not satisfactory for the preparation of the analogous palladium (VIIId), due to some decomposition. However, complex, i.e., cis-[Pd(harmalol)Clz] on replacing ethanol with acetone, the yield is excellent. The procedure is as follows: A mixture of trans-[Pd(DMSO)zCl;?) (1 g, 3 mmol) and harmalol (061 g, 3 mmol)
50
T. A. K. AI-Aliaf et al.
was suspended in acetone (70 mL), and the mixture was refluxed for approximately 4 hours. The solution became clear yellow-orange. It was filtered through @elite \vhi!e it was hot. The solution was taken to dryness. The solid thus formed was dried in vacua for several hours. [Pd(harmine) Cr,/, (IX&). This can be prepared by a similar method to that for VIIIa (above), apart from the stage when the solution becomes clear, when an immediate filtration through Celite should be carried out. This method is not satisfactory for the preparation of IXa, i.e., [Pt(harmine)Clz]z, because the formation of the product is so fast that one cannot filter the solution through Celite to remove the unreacted starting material. Therefore, the following method is superior for both IXa and IXb: The complex cis-[Pt(DMSO)?Clz] (0.21, 0.5 mmol) was suspended in hot ethanol (20 mL), and harmine (0.12 g, 0.57 mmol) was dissolved in warm ethanol (10 mL) and added at once to the suspension. The solution became clear only for a few seconds, and immediately a pale yellow precipitate started to form. The reaction mixture was stirred for about 6 hours while it was warm and then filtered. The solid thus formed was washed with small portions of warm ethanol (to remove any unreacted harmine), then with n-hexane, and dried in vacua. The product is pure enough for further purposes. fikf(harmane) Cl2jt (M = Pt [ZXc] and Pd fZXd]). This was prepared by a method similar to the second method for preparing complex IXa or IXb (above): by stirring a slight excess of harmane (as tetrahydrate) with [M(DMSO)nCla J (M = Pt or Pd) in warm ethanol or acetone for about 6 hours.
RESULTS
AND DISCUSSION
Our previous studies 1131 on harmaplatin ([Pt(harmaline)Cl~]) have shown that this complex has some antitumor activity. Therefore, in connection with these studies, we thought it might be of interest to extend this work to the synthesis of the palladium complex of harmaline as well as the platinum and palladium complexes of compounds with structures related to &carboline compounds, i.e., harmalol, hat-mine, and harmane. The syntheses of VIIIa-d and IXa-d have been accomplished by treatment of the complex [M(DMS0)2Clz J in chloroform, ethanol, or acetone with the corresponding derivative of &carboline. The yields were almost quantitative. The structural assignments of these complexes were achieved on the basis of their elemental analyses and infrared (IR), Raman, and NMR (‘H and 13C) spectra. However, the results obtained for platinum complexes of these alkaloids are similar to those for analogous palladium complexes. Therefore, we are going to ccncentrate our discussions mai&y on the platinum complexes. The NMR data for the ligands (harmaline, harmalol, harmine, and harmane) were tabulated for comparison. It should be noted that none of the complexes is soluble in water, indicating that none of them is ionic. On the other hand, type VIII complexes are much more soluble in the common solvents, i.e., chloroform and ethanol, compared to those of type IX. Structural
Assignments
Models suggest that the lone pair of electrons in the case of harmine and harmane of N2 and NH are almost in the same plane, so that coordination of the metal (Pt or Pd) with these sites is blocked by steric hindrance of the carbon-l atom of the pyridine
Pt(II) & Pd(II) COMPLEXES
OF SOME @-CARBOLINE
ALKALOIDS
51
_--
A
Cl
FIGURE 3. Cl
(VIII) (a) M = Ft; R = 0CH3 (b) M = Pd; R = 0CH3 (c) M = Pt; R = OH (d) M = Pd; R = OH
ring, which would be in the way of it, so that a complex of type (IX) has been suggested, i.e., a dimer. In the case of harmaline and harmalol a complex of type VIII might be present, i.e., as a monomer. This type of ligand has the potential to form five-membered ring chelates with the metal ion and does not undergo dimerization, as type IX does. This is clear from the space-filling model, which shows that the N-2 ring has a twisted boat form, which in turn makes for close proximity of the NH and N-2 sites, leading to a rapid complexation with the metal. Because of their almost identical e1ementat-y composition, the monomer and the dimer structures are not distinguishable by analyses (CHN analyses and iH and 13C NMR). Furthermore, x-ray single crystal diffraction was unsuccessful in establishing the structures of these complexes due to the difficulty of obtaining a single crystal that is amenable to x-ray crystallography. Infrared and Raman spectroscopy might be helpful to give some idea about the structures. We now describe in detail the characterization of these complexes (VIII and IXa-d). ‘W NMR measurements. The ‘H NMR spectra were recorded at room temperature by dissolving the sample in DMSO-d,, and the results obtained are listed in Table 1. Inspection of Table 1 shows a clear change in the ‘H chemical shifts of both the ligands and their corresponding platinum and palladium complexes. For example, the 6CH3 and 6NH values for harmaline are 2.08 and 11.14 ppm, respectively, whereas they are 2.52 and 11.71 ppm for cis-[Pt(harmaline)Clz] and 2.56 and 11.60 ppm for cis[Pd(harmaline)Clz J. In other words, both the 6CH3 and 6NH values have been shifted downfield by approximately 0.5 and 0.6 ppm, respectively, on coordinating the ligand (harmaline) to platinum and palladium metals. Similarly, the 6CH3 and 6NH values for harmalol, harmine, and hat-mane are also different from those obtained for their
FIGURE 4.
(IX) (a) M = Pt; R = OCH3 (b) M = Pd; R = OCHl (c) M = Pt; R = H (d) M = Pd; R = H
2.08 2.52 2.56 2.10 2.50 2.40 2.70 2.54 2.53 2.11 2.04 2.06
Iv
3.80 3.82 3.86 (ll.OO)b ( 9.80) ( 9.60) 3.87 3.91 3.90 -
Cl 156.7 164.4 163.5 155.1 164.2 -c 141.9 143.5 143.1 142.2 144.5 _c
NH 11.14 11.71 11.60 ll.OOb 11.50 11.30 11.40 12.00 12.10 11.55 12.13 12.24
_
19.1 19.4 19.4 19.1 19.9 _ 111.8 113.0 113.1 112.6 112.3 _
C4
_ _ _.
47.5 54.9 52.5 47.4 54.9 _ 141.2 142.7 142.9 140.4 14i.6 _
C3
o Downfieldfrom internalTMS, using DhlSO-&as solvent. b Overlappingbetween NH and OH signals, c Cl throughClOa, CHJ, and CH30 not recorded.
IXd
IXC
VIIId VI IXa IXb VII
VIiIa VIIIb V vluc
CHJ
Compwnds
CHqO (GH)
‘H NMR Chemical Shifts
119.4 118.0 117.7 115.1 117.8 _ 114.8 113.8 113.6 121.2 122.5 _
C5 110.2 112.1 111.6 110.8 112.4 _ 109.0 1.0,9 110.7 126.9 127,X -
C6 157.1 158.9 158.5 157.3 157.2 _ 160.1 161.5 161.3 119.2 120.0 _
C7 94.6 94.0 94.1 96.8 96.3 94.6 94.4 94.4 112.0 112.3 _
C8
137.5 139.6 138.8 138.2 140.0 139.7 139.7 139.5 137.6 139.6 _
C9
13CNMR Chemical Shifts
120.3 118S 118.5 118.6 118.3 135.3 135.3 135.1 127.7 129.7 _
C9a
128.4 128.0 127.9 128.0 127.6 128.4 128.4 128.4 134.6 135.2 -
C!O
125.3 121.7 121.3 120.2 121.6 _ 123.5 123.5 123.4 121.6 120.4 -
ClOa
55.0 55.3 55.2 55.4 55.4 55.4
-
21.9 24.1 24.0 21.2 24.0 21.2 21.2 21.9
20.5 22.3 -
CH3 CH:O
-.
TABLE 1. ‘H and 13CNMR Chemical ShiftsQ;G(ppm)for the Ligands (IV-VII) and the Corresponding Platinum and PalladiumComplexes (VIII and IXa-d)
Pt(II) & Pd(II) COMPLEXES
H
CH3
OF SOME P-CARBOLINE
ALKALOIDS
53
FIGURE 5.
corresponding platinum and palladium complexes. This, of course, reveals the presence of an inductive effect due to the coordination of both nitrogen atoms on the ligand. This is confirmed by the value of 60CH3, which remsins almost constant on going from the 1ig;inds to their platinum and palladium complexes (Table 1); i.e., no significant change is observed due to the fact that OCH3 group is far away from the coordination sphere. 13C NMR measurements. To provide another check on the effects caused by the coordination of the ligand with platinum and palladium metals, i3C NMR spectra were recorded in DMSO+ fGi Seth the ligands and the complexes, and the data obtained are listed in Table 1. In order to specify the chemical shifts of carbons in these complexes, it is important to examine the general trends in behavior exhibited by the carbons in the ligands. The chemical assignments of the carbons of these ligands were achieved on the basis of the chemical shift rules [14] and mainly by comparison with structurally related &carboline moieties [ 1%171. The 13C NMR spectra of these ligands showed signals corresponding to each of the carbon atoms present. The most interesting feature in the spectra of the complexes is the chemical shifts of carbons 1, 3, and 9, as well as the methyl group of the fl-carboline moiety (a species adjacent to the coordination sites). The chemical shift values measured for the carbons of the ligand are different from those measured for the corresponding platinum and palladium complexes due to the effects of the coordination of both nitrogens on the ligands with the metal, while the chemical shift value of the carbon atom of the OCH3 group showed no significant change when measured for the ligand and for its platinum and palladium complexes (Table I). Roman and IR measurements. Raman spectra were recorded using solid samples and IR spectra were recorded using XI disks or Nujol mull. The figures obtained are listed in Table 2. Type VIII complexes are expected to have the c&-configuration, i.e., a chelate system through both nitrogen atoms available for coordination in the alkaloid moieties. Infixred spectral data showed two characteristic bands for vPt-Cl as well :astwo bands for vPt-N, Compare this with ~~z+[P~(NH,)~C~~], i.e., cisplatin, which shows vPt-Cl = 280 and 310 cm- 1 and one value for vPt-N (510 cm-r) [ 18). In our case, the complex cis-[Pt(harmaline)C12] give IR data as follows: vPt-Cl = 330 and 348 cm-’ and vPt-N = 258 and 373 cm-r due to both nitrogen atoms, the pyridinic 1191and the indolic ones. However, the Raman spectra for the same complex gave one vPt-Cl at 342 cm- * with a shoulder and two bands at 150 and 245 cm- *, probably due to @t-N. For IX complexes, assuming a planar structure of the complex with Cl atoms coplanar with the whole molecule, it follows that the type of symmetry may be deduced from the activity of vPt-Cl or vPt-N in the Raman and IR spectra. Hence, the symmetry likely for the complex type IX may be considered to be C&,. The calculated
deep yellow
pale yellow
off-white
IXb
Ixc
IXd
32.0 (32.64) 40.00 (40.10) 31.98 (32.14) 41.15 (40.1 I) (2.79;
(2.51) 3.01 (3.08) 2.16 (2.23) 2.96
2.42 5.64 (5.86) 7.14 (7.20) 6.07 (6.25) 8.04 (7.80) -,
?295 m
3317 s
3262 s
3310 s, b 3195 s
3290 s
3250 m
3290 m
UN-H
318, 372 m
340, 385 m
262, 366 m
246, 380 m 265, 376 m
262, 367 m
267, 355 m
258, 373 m
VP&N .-
325 m
360m
323 s
336, 358 m 348 s
330, 346 m
320. 3h3 m
330, 348 m
VPt-Cl -~-
Characteristic IR data”(cm- ‘) ~_------
---
145,248m
150, 245 m
vPt-N
.--
332 s
342 s, sh
#t-Cl
Raman (cm- I)
0 IR samples for VIIIa, b. c and iXa, b were recorded using KI disks and the rest were recorded using Nujol mull. m = medium; s = small; sh = shoulder, and b = broad bnnds. b Not recorded.
>300
206-210
210-224
180-190 242-246
orange-red pale yellow
(6.00) -b
(2.60) -b
V:!Id IXa
5.8
(7.16)
7.13
5.63 (5.83)
N
2.70
222-228
yellow
VUIC
30.33 (30.90) -b
3.55
(3.583
40.09
(39.90)
160- 168
yellow-orange
VIIIb
3.07 (2.92)
H -~I
32.42 (32.50)
yellow
VIIIa
C -
Analyses: found (cab.)(%) -
186-190
Color
Complex
Meltine point (dec.) (“C)
TABLE 2. The Properties of Platinum and Palladium Complexes (VIII and IX a-d) --___-
modes are as follows: Mode CZh
Activity
Ag R(p)
BU
N
R
P
IR
C
IR
2
1
1
1
0
where N = number of normal modes; R = number of Raman active modes; P = polarized Raman active modes; IR = infrared active modes; and C = number of coincidence,
The data given from the Raman and IR spectra of complex IXa, [Pt(harmine)Cl& , show that the complex seems to have one Raman active and one IR active mode (Table 2), i.e., one Raman and one IR for vPt-Cl(332 and 348 cm- *, respectively) as well as one Raman and one IR for vPt-N for each of the pyridinic and the indolic nitrogens (145,248 cm” and 265,376 cm-l, respectively). Therefore, it is possible to suggest, as mentioned above, that this type of complex has Czh symmetry. However, the other indication about the structure is the stretching frequency of the N-H bond, which is obscured in the spectra of the ligands, p
TAKA,
MTA.
Bhlogy
Department,
ABSTRACT The preparation of novel biologically active platinum(H) and palladium(I1) complexes of some P-carboline alkaloids (harmaline, harmalol, hannine, and hat-mane) is described. These complexes, characterized on the basis of their CHN elemental analysis, infrared, Raman, and ‘H and 13C nuclear magnetic resonance spectral data, were shown to have the empirical formula [M(alka10id)C12], M = Pt. Pd. The antitumor and antiviral activities of some of these complexes have been demonstrated.
INTRODUCTION A large body of platinum complexes and, to a lesser extent, palladium complexes have been prepared and their antitumor activity examined since the discovery of Rosenberg et al. Cl], that cisplatin compound (I) has considerable antitumor activity (the first generation). Much work was done thereafter with the aim of decreasing the toxicity of platinum compounds and increasing their activity, as well as their solubility. Hence, the second generation of analogs to cisplatin was discovered, that called carboplatin, or paraplatin (II) [2]. Other cisplatin analog complexes having amino sugars as ligands have been synthesized and tested for antitumor activity 131. Address reprint requests to: Dr. T. A. University of Mosul, Mosul, Iraq.
K. AI-Allaf,
Chemistry
Department,
College
of Science,
Journal of Inorganic Biochemistry 38 47-56 (1990)
0
1990 Elseviet’ Science Publishing
Co., Inc., 655 Avenue of the Americas,
47 NY, NY 10010
0162-0134/90!$3.50
48
T. A. K. AI-All&
et al.
OH
(1)
Cisplatin
(II)
Carboplatin (Paraplatin)
@II)
Iproplrltin
The wide range physiologic activity of @-carboline alkaloids has stimulated a wealth of different approaches to their synthesis [4,5]. Some of these compounds, harmaline (4,9-dihydro-7-mzthoxy-l-methyl-3H-pyrido[3,4-b]indole), harmalol (4,9_dihydro-lmethyl-3H-pyrido[3,4-blindole-7-ol), haxmine (7-methoxy- 1-methyl-9H-pyrido[3,4b]indolc), and harmane (1 -methyl-9H-pyrido[3,4-blindole) (IV-VII), are extremely effective as antituberculosis 161, analgesic 171, and antimicrobial [8] agents. Attempts have been made to find other chemical substances having at least as much, if not more, similar activity. especially recently; this has led to the finding that harmine (VI) has some antiviral activity [9]. As a continuation of our previous studies [lo] on the synthesis and characterization of platinum(IIj complexes of substituted 3-aryld,S-dimethyl-2-pyrazolines and as a part of an investigation of the biologic activity of these complexes, we required the synthesis of platinum(H) and palladium@) complexes of some fi-carboline alkaloids of the type 3H- (and 9H-) pyrido-[3,4-blindole (IV-VII). In our previous article (1 l] on this subject, we reported the first synthesis and characterization of platinum@) and palladium(I1) complexes of some hydrochloride derivatives of P-carboline alkaloids from the reaction of the alkaloid hydrochlorides with &MC4 (M = Pt, Pd) in water. Since the complexes we reported, i.e., trans-[M(alkaloid - HClh Clz] are rather insoluble in water and in most common organic solvents and since they decompose in DMSO, we could not manage to study their biologic activity as pure complexes. Because of this and in order to overcome these difficulties, we decided to prepare the platinum(I1) and palladium(II) complexes of the HCI-free alkaloids and to study their antitumor and antiviral activities. EXPERIMENTAL Natural abundance, proton-decoupled spectra were recorded at 25°C on a 22.63 MHz. A pulse width of 6 ps and scans (depending on the nature of the points, for a spectral width of 4,807
J?iJ--q-q Harma’o1 FIGURE
R
FT 13C nuclear magnetic resonance (NMR) Bruker-WH 90 DS spectrometer, operating at a 3-s delay were used, and approximately 3,000 compound) were accumulated with 16,000 data Hz.
1.
(IV) R = OCH3, Hannaline (V) R = OH,
Pt@) & Pd(II) COMPLEXES
H
FIGURE “3
OF SOME
2.
B-CARHOLINE
ALKALOlDS
(VI) R = OCHx, Harmine
49
(VII) R = H,
Harmane
Proton NMR spectra were recorded on the same spectrometer, using the deuterium signal of the solvent as a field lock signal. Infrared spectra were recorded on an SP 2000 spectrophotometer at a range of 200 using Nujol mull or RI disks. to 4,000 cm”, Raman spectra were recorded on solids using a Laser Raman spectrophctometer. The samples were done at Dortmund University, West Germany. Analyses of the complexes were carried out on a CHN Analyzer, type 1106 (Carlo Erba). Starting Materials The compounds &PtCb, PdCl2, harmaline, harmalol, harmine, and harmane were obtained commercially from Fluka Company. The complexes [M(DMS0)2Cl,] (M = Pt, Pd) were prepared according to the method described in the literature [12]. Preparation of Complexes All solvents used were dried and degassed under dry nitrogen.
before
use, and reactions
were carried
out
Cis-[pt (harmaline) Cr,)/ (VIII+‘. The complex cis-[Pt(DMSO)zCIZ] (0.2 g, 0.474 mmol) was suspended in chloroform (15 mL), and ham&line (0.11 g, OS 13 mmol) was added gradually at ambient temperature. The turbid solution became clear yellow, and after about 30 min, the solution was filtered through Celite. The filtrate was takeq to dryness, and yellow oily material was obtained. This was stirred with light petroleum spirts 30” to 40°C (30 mL) overnight. The mother liquor was decanted from the fine yellow crystalline product, and the solid was dried under vacuum. The product can be recrystallized from ethanol/n-hexane. The yield is almost quantitative. A similar method was used to prepare the analogous palladium complex, i.e., cis[Pd(harmaline)C12] (VIIIb). Cis-fPt (Harmalcd) ClJ
(VIZZc). The complex
cis-[Pt(DMSO)zClz ] (0.422 g, 1 .O mmol) was suspended in warm ethanol (20 mL), and a warm solution of harmalol (0.22 g, 1.1 mmol) in ethanol (10 mL) was added to the suspension under nitrogen. The solution became clear yellow. It was allowed to stand for about 30 min and then filtered through Celite to remove any unreacted starting materials. Slow evaporation of the solvent gave yellow crystals from the product. These were washed with small portions of ethanol and then with hexane and dried in vacua for several hours. This method is not satisfactory for the preparation of the analogous palladium (VIIId), due to some decomposition. However, complex, i.e., cis-[Pd(harmalol)Clz] on replacing ethanol with acetone, the yield is excellent. The procedure is as follows: A mixture of trans-[Pd(DMSO)zCl;?) (1 g, 3 mmol) and harmalol (061 g, 3 mmol)
50
T. A. K. AI-Aliaf et al.
was suspended in acetone (70 mL), and the mixture was refluxed for approximately 4 hours. The solution became clear yellow-orange. It was filtered through @elite \vhi!e it was hot. The solution was taken to dryness. The solid thus formed was dried in vacua for several hours. [Pd(harmine) Cr,/, (IX&). This can be prepared by a similar method to that for VIIIa (above), apart from the stage when the solution becomes clear, when an immediate filtration through Celite should be carried out. This method is not satisfactory for the preparation of IXa, i.e., [Pt(harmine)Clz]z, because the formation of the product is so fast that one cannot filter the solution through Celite to remove the unreacted starting material. Therefore, the following method is superior for both IXa and IXb: The complex cis-[Pt(DMSO)?Clz] (0.21, 0.5 mmol) was suspended in hot ethanol (20 mL), and harmine (0.12 g, 0.57 mmol) was dissolved in warm ethanol (10 mL) and added at once to the suspension. The solution became clear only for a few seconds, and immediately a pale yellow precipitate started to form. The reaction mixture was stirred for about 6 hours while it was warm and then filtered. The solid thus formed was washed with small portions of warm ethanol (to remove any unreacted harmine), then with n-hexane, and dried in vacua. The product is pure enough for further purposes. fikf(harmane) Cl2jt (M = Pt [ZXc] and Pd fZXd]). This was prepared by a method similar to the second method for preparing complex IXa or IXb (above): by stirring a slight excess of harmane (as tetrahydrate) with [M(DMSO)nCla J (M = Pt or Pd) in warm ethanol or acetone for about 6 hours.
RESULTS
AND DISCUSSION
Our previous studies 1131 on harmaplatin ([Pt(harmaline)Cl~]) have shown that this complex has some antitumor activity. Therefore, in connection with these studies, we thought it might be of interest to extend this work to the synthesis of the palladium complex of harmaline as well as the platinum and palladium complexes of compounds with structures related to &carboline compounds, i.e., harmalol, hat-mine, and harmane. The syntheses of VIIIa-d and IXa-d have been accomplished by treatment of the complex [M(DMS0)2Clz J in chloroform, ethanol, or acetone with the corresponding derivative of &carboline. The yields were almost quantitative. The structural assignments of these complexes were achieved on the basis of their elemental analyses and infrared (IR), Raman, and NMR (‘H and 13C) spectra. However, the results obtained for platinum complexes of these alkaloids are similar to those for analogous palladium complexes. Therefore, we are going to ccncentrate our discussions mai&y on the platinum complexes. The NMR data for the ligands (harmaline, harmalol, harmine, and harmane) were tabulated for comparison. It should be noted that none of the complexes is soluble in water, indicating that none of them is ionic. On the other hand, type VIII complexes are much more soluble in the common solvents, i.e., chloroform and ethanol, compared to those of type IX. Structural
Assignments
Models suggest that the lone pair of electrons in the case of harmine and harmane of N2 and NH are almost in the same plane, so that coordination of the metal (Pt or Pd) with these sites is blocked by steric hindrance of the carbon-l atom of the pyridine
Pt(II) & Pd(II) COMPLEXES
OF SOME @-CARBOLINE
ALKALOIDS
51
_--
A
Cl
FIGURE 3. Cl
(VIII) (a) M = Ft; R = 0CH3 (b) M = Pd; R = 0CH3 (c) M = Pt; R = OH (d) M = Pd; R = OH
ring, which would be in the way of it, so that a complex of type (IX) has been suggested, i.e., a dimer. In the case of harmaline and harmalol a complex of type VIII might be present, i.e., as a monomer. This type of ligand has the potential to form five-membered ring chelates with the metal ion and does not undergo dimerization, as type IX does. This is clear from the space-filling model, which shows that the N-2 ring has a twisted boat form, which in turn makes for close proximity of the NH and N-2 sites, leading to a rapid complexation with the metal. Because of their almost identical e1ementat-y composition, the monomer and the dimer structures are not distinguishable by analyses (CHN analyses and iH and 13C NMR). Furthermore, x-ray single crystal diffraction was unsuccessful in establishing the structures of these complexes due to the difficulty of obtaining a single crystal that is amenable to x-ray crystallography. Infrared and Raman spectroscopy might be helpful to give some idea about the structures. We now describe in detail the characterization of these complexes (VIII and IXa-d). ‘W NMR measurements. The ‘H NMR spectra were recorded at room temperature by dissolving the sample in DMSO-d,, and the results obtained are listed in Table 1. Inspection of Table 1 shows a clear change in the ‘H chemical shifts of both the ligands and their corresponding platinum and palladium complexes. For example, the 6CH3 and 6NH values for harmaline are 2.08 and 11.14 ppm, respectively, whereas they are 2.52 and 11.71 ppm for cis-[Pt(harmaline)Clz] and 2.56 and 11.60 ppm for cis[Pd(harmaline)Clz J. In other words, both the 6CH3 and 6NH values have been shifted downfield by approximately 0.5 and 0.6 ppm, respectively, on coordinating the ligand (harmaline) to platinum and palladium metals. Similarly, the 6CH3 and 6NH values for harmalol, harmine, and hat-mane are also different from those obtained for their
FIGURE 4.
(IX) (a) M = Pt; R = OCH3 (b) M = Pd; R = OCHl (c) M = Pt; R = H (d) M = Pd; R = H
2.08 2.52 2.56 2.10 2.50 2.40 2.70 2.54 2.53 2.11 2.04 2.06
Iv
3.80 3.82 3.86 (ll.OO)b ( 9.80) ( 9.60) 3.87 3.91 3.90 -
Cl 156.7 164.4 163.5 155.1 164.2 -c 141.9 143.5 143.1 142.2 144.5 _c
NH 11.14 11.71 11.60 ll.OOb 11.50 11.30 11.40 12.00 12.10 11.55 12.13 12.24
_
19.1 19.4 19.4 19.1 19.9 _ 111.8 113.0 113.1 112.6 112.3 _
C4
_ _ _.
47.5 54.9 52.5 47.4 54.9 _ 141.2 142.7 142.9 140.4 14i.6 _
C3
o Downfieldfrom internalTMS, using DhlSO-&as solvent. b Overlappingbetween NH and OH signals, c Cl throughClOa, CHJ, and CH30 not recorded.
IXd
IXC
VIIId VI IXa IXb VII
VIiIa VIIIb V vluc
CHJ
Compwnds
CHqO (GH)
‘H NMR Chemical Shifts
119.4 118.0 117.7 115.1 117.8 _ 114.8 113.8 113.6 121.2 122.5 _
C5 110.2 112.1 111.6 110.8 112.4 _ 109.0 1.0,9 110.7 126.9 127,X -
C6 157.1 158.9 158.5 157.3 157.2 _ 160.1 161.5 161.3 119.2 120.0 _
C7 94.6 94.0 94.1 96.8 96.3 94.6 94.4 94.4 112.0 112.3 _
C8
137.5 139.6 138.8 138.2 140.0 139.7 139.7 139.5 137.6 139.6 _
C9
13CNMR Chemical Shifts
120.3 118S 118.5 118.6 118.3 135.3 135.3 135.1 127.7 129.7 _
C9a
128.4 128.0 127.9 128.0 127.6 128.4 128.4 128.4 134.6 135.2 -
C!O
125.3 121.7 121.3 120.2 121.6 _ 123.5 123.5 123.4 121.6 120.4 -
ClOa
55.0 55.3 55.2 55.4 55.4 55.4
-
21.9 24.1 24.0 21.2 24.0 21.2 21.2 21.9
20.5 22.3 -
CH3 CH:O
-.
TABLE 1. ‘H and 13CNMR Chemical ShiftsQ;G(ppm)for the Ligands (IV-VII) and the Corresponding Platinum and PalladiumComplexes (VIII and IXa-d)
Pt(II) & Pd(II) COMPLEXES
H
CH3
OF SOME P-CARBOLINE
ALKALOIDS
53
FIGURE 5.
corresponding platinum and palladium complexes. This, of course, reveals the presence of an inductive effect due to the coordination of both nitrogen atoms on the ligand. This is confirmed by the value of 60CH3, which remsins almost constant on going from the 1ig;inds to their platinum and palladium complexes (Table 1); i.e., no significant change is observed due to the fact that OCH3 group is far away from the coordination sphere. 13C NMR measurements. To provide another check on the effects caused by the coordination of the ligand with platinum and palladium metals, i3C NMR spectra were recorded in DMSO+ fGi Seth the ligands and the complexes, and the data obtained are listed in Table 1. In order to specify the chemical shifts of carbons in these complexes, it is important to examine the general trends in behavior exhibited by the carbons in the ligands. The chemical assignments of the carbons of these ligands were achieved on the basis of the chemical shift rules [14] and mainly by comparison with structurally related &carboline moieties [ 1%171. The 13C NMR spectra of these ligands showed signals corresponding to each of the carbon atoms present. The most interesting feature in the spectra of the complexes is the chemical shifts of carbons 1, 3, and 9, as well as the methyl group of the fl-carboline moiety (a species adjacent to the coordination sites). The chemical shift values measured for the carbons of the ligand are different from those measured for the corresponding platinum and palladium complexes due to the effects of the coordination of both nitrogens on the ligands with the metal, while the chemical shift value of the carbon atom of the OCH3 group showed no significant change when measured for the ligand and for its platinum and palladium complexes (Table I). Roman and IR measurements. Raman spectra were recorded using solid samples and IR spectra were recorded using XI disks or Nujol mull. The figures obtained are listed in Table 2. Type VIII complexes are expected to have the c&-configuration, i.e., a chelate system through both nitrogen atoms available for coordination in the alkaloid moieties. Infixred spectral data showed two characteristic bands for vPt-Cl as well :astwo bands for vPt-N, Compare this with ~~z+[P~(NH,)~C~~], i.e., cisplatin, which shows vPt-Cl = 280 and 310 cm- 1 and one value for vPt-N (510 cm-r) [ 18). In our case, the complex cis-[Pt(harmaline)C12] give IR data as follows: vPt-Cl = 330 and 348 cm-’ and vPt-N = 258 and 373 cm-r due to both nitrogen atoms, the pyridinic 1191and the indolic ones. However, the Raman spectra for the same complex gave one vPt-Cl at 342 cm- * with a shoulder and two bands at 150 and 245 cm- *, probably due to @t-N. For IX complexes, assuming a planar structure of the complex with Cl atoms coplanar with the whole molecule, it follows that the type of symmetry may be deduced from the activity of vPt-Cl or vPt-N in the Raman and IR spectra. Hence, the symmetry likely for the complex type IX may be considered to be C&,. The calculated
deep yellow
pale yellow
off-white
IXb
Ixc
IXd
32.0 (32.64) 40.00 (40.10) 31.98 (32.14) 41.15 (40.1 I) (2.79;
(2.51) 3.01 (3.08) 2.16 (2.23) 2.96
2.42 5.64 (5.86) 7.14 (7.20) 6.07 (6.25) 8.04 (7.80) -,
?295 m
3317 s
3262 s
3310 s, b 3195 s
3290 s
3250 m
3290 m
UN-H
318, 372 m
340, 385 m
262, 366 m
246, 380 m 265, 376 m
262, 367 m
267, 355 m
258, 373 m
VP&N .-
325 m
360m
323 s
336, 358 m 348 s
330, 346 m
320. 3h3 m
330, 348 m
VPt-Cl -~-
Characteristic IR data”(cm- ‘) ~_------
---
145,248m
150, 245 m
vPt-N
.--
332 s
342 s, sh
#t-Cl
Raman (cm- I)
0 IR samples for VIIIa, b. c and iXa, b were recorded using KI disks and the rest were recorded using Nujol mull. m = medium; s = small; sh = shoulder, and b = broad bnnds. b Not recorded.
>300
206-210
210-224
180-190 242-246
orange-red pale yellow
(6.00) -b
(2.60) -b
V:!Id IXa
5.8
(7.16)
7.13
5.63 (5.83)
N
2.70
222-228
yellow
VUIC
30.33 (30.90) -b
3.55
(3.583
40.09
(39.90)
160- 168
yellow-orange
VIIIb
3.07 (2.92)
H -~I
32.42 (32.50)
yellow
VIIIa
C -
Analyses: found (cab.)(%) -
186-190
Color
Complex
Meltine point (dec.) (“C)
TABLE 2. The Properties of Platinum and Palladium Complexes (VIII and IX a-d) --___-
modes are as follows: Mode CZh
Activity
Ag R(p)
BU
N
R
P
IR
C
IR
2
1
1
1
0
where N = number of normal modes; R = number of Raman active modes; P = polarized Raman active modes; IR = infrared active modes; and C = number of coincidence,
The data given from the Raman and IR spectra of complex IXa, [Pt(harmine)Cl& , show that the complex seems to have one Raman active and one IR active mode (Table 2), i.e., one Raman and one IR for vPt-Cl(332 and 348 cm- *, respectively) as well as one Raman and one IR for vPt-N for each of the pyridinic and the indolic nitrogens (145,248 cm” and 265,376 cm-l, respectively). Therefore, it is possible to suggest, as mentioned above, that this type of complex has Czh symmetry. However, the other indication about the structure is the stretching frequency of the N-H bond, which is obscured in the spectra of the ligands, p
