Two iridoid glycosides from Chaenorrhinum minus

Phytochemisrry,Vol. 29, No. 12,PP. 3865-3868,1990 Printed in Great Britain.

0031-9422/90$3.00+0.00 0 1990Pergamon Press plc

TWO IRIDOID GLYCOSIDES

FROM CHAENORRHZNUM

MINUS

JENS BREINHOLT, JAN STEEN JENSEN, WREN ROSENDAL JENSEN and BENT JUHL NIELSEN PharmaBiotek Research Center, Department of Organic Chemistry, Build. 201, Technical University of Denmark, DK-2800 Lyngby, Denmark (Received 10 April 1990) Key Word Index-Chaenorrhinum minus; Scrophulariaceae; iridoid glucosides; chaenorrhinoside; bartsioside; antirrhinoside; prunasin.

lo-glucosyl-

Abstract-In addition to the known iridoid glucoside antirrhinoside and the cyanogenic glucoside prunasin, two new iridoid glucosides were isolated and identified from Chaenorrhinum minus. One of the new iridoid glucosides was found to be lo-glucosyl bartsioside. The other compound, named chaenorrhinoside, was shown to contain an u-pyrone ring with a /?-glucopyranosyl moiety attached in the C-6 position, structural features which have only rarely been encountered before in iridoid compounds.

INTRODUCTION In a chromatographic investigation [l] of the distribution of iridoids in Serophulariaceae, it was shown that members of the tribe Scrophularioideae-Antirrhineae almost consistently contained antirrhinoside (1). This tribe comprises among others the genera Antirrhinum (snapdragon) and Linaria (toadflax). The subject of the present work, Chaenorrhinum minus (L.) Lange [ = Linaria minor (L.) Desf.] also contained compound 1. Ac-

RO R

cording to Hegnauer [2], earlier chemical work on C. minus has shown that the cyanogenic glucoside prun-

R

asin (4) is also present in the plant. Here we report our results on the constituents of the water-soluble part of an extract of the plant. RESULTS AND

3 R=Glc 5 R=H

0 6 R=H

7 RR=

>C(CH,),

dJ

DISCUSSION

A ‘HNMR spectrum of the aqueous extract of the plant revealed the presence of several compounds. Separation was performed by reversed phase chromatography (see Experimental) to give, besides the known compounds antirrhinoside (1) and prunasin (4), the two unknown compounds 2 and 3, where the latter was an iridoid glucoside. Thus, in the ‘H NMR spectrum of compound 3 (see Experimental) a pattern similar to that of bartsioside (5) [3] was evident, except for the presence of signals corresponding to an additional sugar moiety. When comparing the spectrum of 3 with that of bartsioside, significant downfield shifts of 0.30 and 0.15 ppm were seen for the protons at C-10 in compound 3, suggesting this position for the second sugar unit. Smaller downfield shifts (ca 0.1 ppm) were evident for the protons at C-l, C7 and C-9. The signals from the two sugar moieties were overlapping in the region 6 3.1-3.9, but a signal arising from a second anomeric proton was visible as a doublet (J = 7.9 Hz) at 6 4.40. The 13C NMR spectrum of 3 (Table 1) provided more information. From the 21 peaks one aglucone and two /I-glucopyranosyl moieties, respectively, could be assigned. The second sugar moiety

' '0 *

’ ‘0 0

4

8

was linked to the C-10 position, since significant shift differences of + 8, - 7 and + 3 ppm were seen for C- 10, C8 and C-7, respectively, when comparing with the spectrum of bartsioside. This was consistent with the ‘H NMR data above. Acetylation provided an octaacetate. Thus the new compound consists of a molecule of bartsioside to which a /Gglucopyranosyl moiety is attached at the C-10 oxygen and consequently we have named it lo-glucosyl bartsioside. The second glycoside (2) had the molecular formula C, ,H,,O, in keeping with a molecular peak of m/z 342 in the mass spectrum. The 13C NMR spectrum showed the presence of 15 carbon atoms of which six could be assigned to a /?-glucopyranosyl moiety. Of the remaining nine carbon atoms only two (680.8 and 73.3) were sp3hybridized while the remaining seven (6107.5 to 162.7)

3865

J. BREINHOLT et al.

Table

I. 13C NMR data for compounds 2, Za, 3, 5, 6,6a and 7*

C

3

5

2

2a

1 3 4 5 6 7 8 9 10

95.0 139.2 109.7 32.5 38.9 132.1 138.4 48.2 68.1

95.2 139.4 109.7 33.0 38.9 129.1 141.8 48.1 60.2

162.7 153.8 107.5 158.5 80.8 73.3 145.2 124.4 112.9

151.8 104.9 155.3 77.8 71.6 140.3 123.7 114.3

103.8 74.2 76.5 70.4 77.0 61.5

100.4 71.0 72.4 68.0 72.5 61.5

1’ 2 3’ 4’ 5 6’

99.1 73.6 76.4 70.4 77.0 61.5

99.2t 73.6 76.5 70.4 77.0 61.5

6a

6 162.7 154.0 106.5 160.9 73.0” 73.5” 145.6 123.4 113.0

7

152.4 104.4 153.6 71.6” 71.7” 140.8 120.1 115.2

152.4 104.7 156.0 80.5” 80.0” 142.7 112.5 115.4

*At 62.5 MHz; 2, 3, 5 and 6 in D,O, Za, 6a and 7 in chloroform-d. tSignals from p-glucopyranosyl moiety in the C-10 position. “Exchangeable in same vertical row.

were sp’-hybridized, corresponding to three double bonds and a carbonyl group. The presence of a carbonyl group was certified by the IR spectrum which showed an intense absorption at 1720 cm-’ and was also in agreement with the UV-spectrum, where an absorption at 317 nm (log E 3.88) indicated a highly conjugated system. Because the molecular formula showed the presence in the molecule of seven double bonds or rings, the structural features mentioned above allowed the presence of two rings in the aglucone. The ‘H NMR spectrum of compound 2 in addition to the signals arising from a glucopyranosyl moiety showed the presence of two separated spin-systems. The first of these was an AB-system at 67.68 and 6.85 consisting of two vinylic protons with no further couplings. The low field shifts, combined with the large shift difference between these two protons indicated on the one hand conjugation with a carbonyl group, on the other hand the presence of an oxygen substiluent on the double bond. This, in combination with a rather small coupling constant (J = 5.1 Hz), suggested the presence of an a-pyrone ring [4,5]. The second spin-system consisted of four protons of which one was central and coupled with the remaining three. The chemical shifts as well as the coupling pattern indicated an arrangement such as that shown in Scheme 1, which together with the Dglucopyranosyl moiety and the a-pyrone ring segment accounted for the complete molecular formula. Combination of the segments of the aglucone could be performed in two ways to give the structures A or B, respectively, neither of which could be excluded by the spectral information obtained so far, although structure A due to its similarity with an iridoid seemed the most likely. A ‘H, ‘H-COSYLR spectrum of 2 gave additional information. It demonstrated the presence of the small, expected geminal coupling between the protons on the exocyclic double bond. It also disclosed the presence of a small long range coupling between H-4 and H-6. The ‘H, ‘H-NOESY spectrum on the other hand allowed discrimination between the structures A and B, as a significant NOE could be seen between H-4 and H-6, a

RO

\\

R,R

=

H,

OGlc

R

El Scheme

1

fact which only seemed compatible with structure A. A similar effect could be seen between H-l’ and H-6, and this determined the position of the sugar moiety, so the complete structure of chaenorrhinoside at this point could be depicted as 2, although without the stereochemical information. Acetylation of compound 2 as expected gave a pentaacetate (Za). Treatment of compound 2 with p-glucosidase in water caused facile hydrolysis to the aglucone (6) which in turn could be transformed to either a diacetate (6a) or to an acetonide (7), the latter by treatment with stannous chloride in dry acetone. The formation of an acetonide proved the &-configuration of the two hydroxy groups in compound 6, and thus the relative configuration of chaenorrhinoside as that depicted in formula 2. Assignment of the ‘%JNMR signals of compound 2 (Table 1) was fairly straightforward when using ‘H, 13CPCOSY and -COLOC spectra, including the signals from carbon atoms not carrying protons, namely those at 6 124.4, 145.2, 158.5 and 162.7. Thus under the conditions used, the signal at 6 124.4 could be assigned to C-9 as three-bond couplings were observed between this carbon and both H-4 (66.85) and the two C-10 protons (65.94

Iridoid glycosides from Chaenorrhinum tninus

and 5.51). Likewise, three-bond couplings were seen between H-3 (67.68) and each of the two signals at 6 162.7 (GO) and 158.5, allowing the latter signal to be assigned to C-S. Finally the signal at 6 145.2 could be assigned to C-8. When comparing the assignments with those published for isocoumarin [S] a fair similarity could be seen for some of the signals in the a-pyrone ring, but not for C3 and C-5 which showed deviations of 9 and 22 ppm, respectively. However, this was not unexpected in view of the large differences between the two systems. As indicated above, chaenorrhinoside is probably of iridoid derivation. In fact, the aglucone could in relatively few steps be formed from antirrhinoside (1). Thus, enzymatic deprotonation at C-10 with a displacement of the negative charge followed by opening of the oxirane ring, would provide the correct functionalities in the cyclopentane ring except at C-5 and C-9. Next, removal of the sugar moiety and oxidation at C-l followed by a final removal of the constituents of water could form the 5,9double bond. Although iridoid lactones are quite common, the only known example of an iridoid derived apyrone is 5,9_dehydronepetalactone (8) isolated from Nepeta cataria [6]. EXPERIMENTAL Microanalyses were performed by LEO Microanalytical Laboratory, Ballerup, Denmark. Mps: uncorr. The plant material was grown in Tilstrup, Denmark at the experimental station of The Botanical Garden of Copenhagen. A voucher specimen (IOK-21/89, verified by Dr Alfred Hansen) has been deposited in the Herbarium of the Botanical Museum, Copenhagen. Frozen plant material (320 g, collected in August 1989) was blended with EtOH (2 x 750 ml) and the filtrate taken to dryness. The residue was partitioned in H,O-Et,0 and the aq. fraction passed through alumina (200 g) followed by elution with H,O (SOOml). After concn the residue was dissolved in MeOH (ca 20 ml) and clarified by passage through active carbon to give the almost colourless crude extract (6.0 g). The ‘H NMR spectrum showed antirrhinoside (1) to be the main glucoside present but signals from other constituents were visible. Chromatography in two portions on a Merck Lobar column (RP-18; size C), using H,O-MeOH mixtures (10: 1 to 1: 1) as eluents, gave after elution of the polar compounds, 4 fractions. Fraction 1 consisted of antirrhinoside (1; 1.1 g), fraction 2 was crude chaenorrhinoside (2; 300 mg), fraction 3 contained lo-glucosyl bartsioside (3; 130 mg) while fraction 4 was mainly prunasin (4; 1.0 g) as seen by comparison of the ‘H NMR spectrum with that of a known [7] specimen. Chaenorrhinoside (2). Rechromatography of fr. 2 from above on the same column (H,O-MeOH; 6: 1) gave pure 2 (190 mg). Crystd from H,O: mp 218-219” (dec); [a];’ + 133” (H,O, c 0.6); ‘H NMR (250 MHz, D,O): 67.68 (d, J=5.1 Hz, H-3), 6.85 (d, J =5.1 Hz,H-4), 5.94(d,J=1.9Hz,H-10),5.51 (d,J=l.l Hz,Hlo), 5.10 (d, J = 5.9 Hz, H-6), 4.84 (ddd, J = 5.9,1.9 and 1.1 Hz, H7), 4.70 (d, J = 7.9 Hz, H-l’), 3.3-4.0 (pattern cooresponding to HT-H-6 of /l-glucopyranosyl moiety); 13C NMR data in Table 1; EIMS (probe) 70 eV, m/z (rel. int.): 342 [M]’ (1.5), 180 [M -C,H,,O,+H]+ or [M-C,H,OB+H]+ (Sl), 163 (25), 151 (70), 73 (100). 2:::” nm (log E):245 (3.78), 317 (3.88); 7::; cm-‘: 1530 and 1610 (C=C), 1720 (conj. ester), 3260 and 3440 (OH). (Found: C, 51.6; H, 5.4. C 15H 180 9 requires: C, 51.3; H, 5.5). Two-dimensional NMR spectra of 3 were recorded in D,O on a 500 MHz spectrometer [S]. ‘H, ‘H-COSYLR: Relaxation delay D, and t, both 0.2 set; spectral range 2326 HZ size in F,: 1 K, in F,: 512. ‘H, 13C-COSY (F,-decoupling); Relaxation

3867

delay D,: 0.55 t,: 3.5 msec, t,: 1.75 msec; size in F,: 1 K, F,: 256; spectral range (Hz): 8197 (F2), 1250 (F,). ‘H, lJC-COLOC: Relaxation delay D,: 1.5 set, t,: 0.025 set, t,: 0.013 s; size in F,: 1 K, in F, 128; spectral range (Hz): 8196 (F2) and 1250 (F,). ‘H, ‘H-NOESY: (Phase sensitive (TPPI), presaturation of solvent) Relaxation delay D,: 2 set, t,: 1 set, delta-t,: 5%; size in F, and F,: 1 K; spectral range (Hz): 2500. Chaenorrhinoside pentaacetate (34. Acetylation of 3 with A@-pyridine (1: 1) overnight provided the tryst. prod. mp (EtOH) 168”; [a];’ +72” (CHCl,; c 1.0); ‘H NMR (250 MHz, CDCl,): 6 7.48 (d, J = 5.4 Hz, H-3) 6.61 (d, J = 5.4 Hz, H-4), 6.18 (br d, J = 1.5, H-lo), 5.69 (d&like, J = 5.3 and 1 Hz, H-7), 5.48 (br d,J=l Hz,H-10),5.05(d,.J=5.5 Hz,H-6),4.68(d,J=6.9 Hz,H1’); ‘“CNMR data in Table 1. (Found: C, 54.00, H, 5.1. C2sH,,0,1 requires; C, 54.4; H, 5.1). Chaenorrhinoside aglucone (6). Compound 6 was prepared by hydrolysis of 3 (250mg) in H,O (20ml) with p-glucosidase (Sigma; 25 mg) for 18 hr when TLC (silica gel; CHCl,-MeOH, 4: 1) showed the reaction to be complete. Extraction with EtOAc (5 x 25 ml) gave after drying and evaporation 6 (130 mg). Cryst. (EtOH) gave the pure compound, mp 145-146”; [a]$’ +155 (MeOH; c 0.6); ‘H NMR (250 MHz, D,O): 67.68 (d, .I = 5.2 Hz, H-3),6.67(d,J=5.2 Hz, H-4), 5.90(d,J=1.8 Hz, H-10),5.47(d,J = 1.3 Hz, H-10),4.9O(d, J=6.0 Hz, H-6), 4.68 (d&like, 5=6.0 Hz and 1.5 Hz, H-7). (Found: C, 59.9; H, 4.8. C,H,O, requires: C, 60.0; H, 4.5). Chaenorrhinoside aglucone diacetate (6a). Compound 6a was prepared as above by acetylation of 6. Crystd from MeOH, mp 133-135”; [a]:: + 121” (CHCI,; ~0.7). ‘H NMR (250 MHz, CDCl,): 6 7.45 (d, J=5.2 Hz, H-3), 6.35 (d, J=5.2 Hz, H-4), 6.19 (d,.I= 1.9 Hz, H-lo), 5.97(d,.l=6.2 Hz, H-6), 5.80(dt-like, J=6.2 and 1.5 Hz, H-7), 5.49 (br d, J= 1.2 Hz, H-lo), 2.06 (s, 6H, 2 x Me); WNMR data in Table 1. (Found: C, 58.9; H, 4.5. CIaH,,O, requires: C, 59.1; H, 4.6). Chaenowhinoside aglucone acetonide (7). The aglucone 6 (100 mg) was dissolved in dry Me&O (50 ml) and dry CuSO, (650 mg) added. The mixture was refluxed with stirring for 2.5 hr, but TLC indicated that almost no reaction had taken place. After addition of SnCI, (50 mg) complete reaction took place in 0.5 hr. The mixture was filtered and taken to dryness. The residue redissolved in CH,CI, and washed with satd NaHCO, soln, dried and evapd to give crude 7 (117 mg; 96%). Cryst. from H,O-MeOH (10: 1); mp 85-87”; [a];’ + 121” (CHCl,; ~0.7); ‘H NMR (250 MHz; CDCI,): 67.53 (d, J = 5.1 Hz, H-3), 6.47 (d, J =5.1 H&H-4),6.26(d,J=ca0.7Hz,H-10),5.63(brs,H-10),5.33 (d, J = 5.9 Hz, H-6), 5.12 (d&like, J = 6 Hz and ca 0.7 Hz, H-7), 1.35 and 1.45 (s’s, 2 x Me), 13C NMR data in Table 1. (Found: C, 65.3; H, 5.6. C,,H,,O, requires: C, 65.4; H, 5.5). lO-Glucosyl bartsioside (3). Rechromatography of fraction 3 from above gave the pure glucoside as an amorphous glass; [a] hz -64” (MeOH; ~0.7); ‘H NMR (250 MHz; D,O): 66.17 (dd, J =6.2 and 1.5 Hz, H-3), 5.84 (m. H-7), 5.39 (d, 3=3.8 Hz, H-l), 4.86 (dd, J=6.2 and 3.0Hz, H-4), 4.70(d, J=7.9 Hz, H-l’), 4.48 and 4.24 (br A&system, J = 12.5 Hz, lo-CH,), 4.40 (d, J = 7.9 Hz, H-1”),2.98(m, H-9), 2.90(m, H-5),2.6O(brdd,J= 16and 7 Hz, H6), 2.05 (br d, .I= 16 Hz, H-6); 13C NMR data in Table 1. (Found: C, 49.0; H, 6.8. C21H320,7, H,O requires: C, 49.4; H, 6.3). lo-Glucosyl bartsioside octaacetate (3a). Compound 3a was obtained as a syrup by acetylation of 3. Due to the small amount available, the compound was only characterized by spectroCDCI,): 66.17 (dd, 5=6.1 and scopy: ‘HNMR(250MHz, 1.2 Hz, H-3), 5.74 (m, H-7), 4.90 (d, J = 8.0 Hz, H-l’), 4.76 (dd-like, 5=6.1 and 2.7 Hz, H-4),4.58(d,J=S.OHz, H-1”),4.51 and4.12 (br AB-system, J = 12.5Hz, lo-CH,), 2.84 (m. 2H, H-5 and H-9), 2.1c2.00 (8 x OAc); FAB/SIMS (Cs-gun; 18.5 kV) showed a peak at 829.3 [M + H] + (MEaIC, = 828.3).