Acylphloroglucinols and flavonoid aglycones produced by external glands on the leaves of two dryopteris ferns and currania robertiana

\

Pergamon PII] S9920Ð8311"86#90992Ð9

Phytochemistry\ Vol[ 37\ No[ 5\ pp[ 820Ð828\ 0887 Þ 0887 Elsevier Science Ltd[ All rights reserved Printed in Great Britain 9920Ð8311:87 ,08[99¦9[99

ACYLPHLOROGLUCINOLS AND FLAVONOID AGLYCONES PRODUCED BY EXTERNAL GLANDS ON THE LEAVES OF TWO DRYOPTERIS FERNS AND CURRANIA ROBERTIANA ECKHARD WOLLENWEBER\ JAN F[ STEVENS\% MONIKA IVANIC$ and MAX L[ DEINZER% Institut fur Botanik der Technischen Universitat\ Schnittspahnstrasse 2\ D!53176 Darmstadt\ Germany^ % Dept[ of Chemistry^ $ Dept[ of Biochemistry and Biophysics\ ALS Building 0996\ Oregon State University\ Corvallis\ OR 86220\ U[S[A[ "Received 11 September 0886#

Key Word Index*Dryopteris villarii ^ D[ ar`uta^ Currania robertiana^ Pteridaceae^ ferns\ leaf exudates^ acylphloroglucinols^ ultrastructure of glandular trichomes[

Abstract*The sparse resinous leaf exudates of Dryopteris villarii and D[ ar`uta contain acylphloroglucinols that have previously only been found in the rhizomes of Dryopteris species[ The phloroglucinol derivatives were characterized as representatives of the para!aspidin\ desaspidin and albaspidin series by 1!D NMR spectroscopy and atmospheric pressure chemical ionization mass spectrometry "APCI!MS!MS#[ Mass frag! mentation proved to be of diagnostic value to identify the acyl side chains and also allowed distinction between desaspidins and isomeric albaspidin analogues[ The frond exudate of Currania robertiana was devoid of acylphloroglucinols\ but was found to contain ~avonoid aglycones[ The acylphloroglucinols and ~avonoid aglycones are presumably produced by external glands on the leaves\ the presence and ultrastructure of which was examined by scanning electron microscopy[ Þ 0887 Elsevier Science Ltd[ All rights reserved

exudates produced by external glands is reported here for the _rst time[

INTRODUCTION

Extensive studies on the chemical composition of frond exudates from Gymnogrammoid ferns "genera Cheilanthes\ Notholaena\ Pityro`ramma and some other Pteridaceae# have revealed the presence of a rich array of ~avonoid aglycones ð0Ð2Ł\ diterpenoids and triterpenoids ð3Ð5Ł[ All fern species previously exam! ined for exudate ~avonoids or other phenolics ð6Ł originate from both Americas\ from Asia\ and from Australia "Platyzoma ð7Ł#[ This is the _rst study of the foliar exudates of two European Alpine ferns\ Currania robertiana "Ho}m[# Wherry ðSyn[] Gym! nocarpium robertianum "Ho}m[# Newm[^ Dryopteris robertiana "Ho}m[# C[ Chr[Ł and Dryopteris villarii "Bell[# Woynar[ Preliminary TLC analysis of the frond exudates of D[ villarii and D[ ar`uta "Kaulf[# Watt[ from Cal! ifornia indicated the presence of non!~avonoid phe! nolics which were subsequently isolated and identi_ed as acylphloroglucinols by spectroscopic techniques[ Although acylphloroglucinols are well!known as internal gland products from rhizomes of many ferns\ in particular Dryopteris ð8\ 09Ł\ their presence in leaf

 Author to whom correspondence should be addressed[

RESULTS AND DISCUSSION

Identi_cation of leaf exudate constituents In the following\ Penttila and Sundman|s system for assigning names to acylphloroglucinols ð00Ð02Ł and numbering of atoms as in Ref[ ð03Ł is used[ The frond exudate of D[ ar`uta yielded para!aspidin AA "0!AA# and four acylphloroglucinols of the desaspidin series "1!AP\ 1!PP\ 1!AB\ and 1!PB# by column chro! matography on polyamide and semi!preparative HPLC on RP!07[ In addition\ six acylphloroglucinols were isolated from the frond exudate of D[ villarii which were identi_ed as the albaspidin homologues 2! AP\ 2!PP\ 2!AB\ 2!PB\ 2!BB and aspidinol "3#[ Struc! ture assignment was achieved by a combination of atmospheric pressure chemical ionization mass spec! trometry "APCI!MS# and 1!D NMR spectroscopy[ By co!TLC with markers\ the frond exudate of C[ robertiana was found to contain the ~avones\ apig! enin\ apigenin!6!methylether\ apigenin!3?!methy! lether\ apigenin!6\3?!dimethylether and the ~avonols\ kaempferol\ kaempferol!2!methylether\ kaempferol! 3?!methylether\ kaempferol!2\3?!dimethylether\

820

821

ECKHARD WOLLENWEBER et al[

kaempferol!2\6\3?!trimethylether and herbacetin! 2\7\3?!trimethylether[ The latter ~avonol is a rather rare compound which\ to our knowledge\ was pre! viously found only _ve times from natural sources\ including a fern Pityro`ramma trian`ularis var[ tri! an`ularis "now Penta`ramma# ð04Ł[ The foliar exu! dates of D[ villarii and D[ ar`uta were found to be devoid of ~avonoids[ Compound 1!AB "Mr 307\ C11H15O7# was obtained as the major constituent of the frond exudate of D[ ar`uta[ Its 0H NMR spectrum showed characteristic signals for an acylphloroglunicol[ Four major low! _eld resonances attributable to "hydrogen!bonded# hydroxyls "dH 8[87\ 00[20\ 05[28 and 07[39#\ accompanied by some four minor signals due to keto! enol tautomerism[ The methylene bridge linking the two ring systems appeared as a broad singlet at dH2[42 which interacted with C!0 and C!0? of both rings in the HMBC spectrum[ The H!4? singlet at dH5[94 was assigned to an aromatic proton\ indicating that one ring system was aromatic in nature[ The other ring was identi_ed as a _licinic acid unit with a `em!dime! thyl group resonating at dH0[44[ This signal coupled with the 02C resonance at dC13[7 in the HMQC spec! trum of 1!AB[ The substitution pattern of the aro! matic "phloroglucinol# ring was determined to be 2?! acyl\ 4?!H\ 5?!OMe[ The aromatic proton signal showed interactions with all ring carbons except C!1?

bearing a free hydroxyl with dH 05[28\ which excludes the possibility of 1!AB being an iso!desaspidin hom! ologue[ Furthermore\ a methoxy group at C!5? fol! lowed from the observation that the free hydroxyl ortho to C!4? interacted with C!2?\ C!4? and with the carbon to which the hydroxyl is attached\ C!3?[ The 0 H NMR spectrum further demonstrated the presence of an acetyl and a butyryl side chain[ Table 0 lists all 0 H and 02C shifts for 1!AB assigned on the basis of 0 HÐ02C couplings observed in the HMQC and HMBC spectra as well as by comparison with NMR data reported for desaspidin BB ð03Ł[ The acetyl function was located on the _licinic acid unit by mass spec! trometry "see below#[ Compound 1!AB was thus ident! i_ed as desaspidin AB[ The accompanying acylphloroglunicols from D[ ar`uta were identi_ed by 0H NMR and mass spec! trometry only[ Compounds 1!AA\ 1!AP and 1!PB showed 0H NMR spectra similar to that of 1!AB\ except for the acyl side chain signals\ and were there! fore identi_ed as desaspidin homologues[ Compound 0!AA "Mr 289\ C19H11O7# lacked the aromatic H!4? signal but showed a C!methyl at dH 1[03 instead[ The 0 H NMR spectrum of 0!AA further showed two acetyl methyls[ The structure of 0!AA was assigned as para! aspidin AA and con_rmed by mass spectrometry[ The Mr of compound 2!BB "Mr 359\ C14H21O7# sug! gested that 2!BB was identical with or closely related

Leaf surface acylphloroglucinols from Dryopteris

822

Table 0[ 0H and 02C NMR data for phloroglucides 1!AB\ 2!PB and 2!BB Atom No[ 0 1 2 3 4 5 6 7 8 09 00 01 02 0? 1? 2? 3? 4? 5? 7? 8? 09? 00? 01? 02? 3!OH 5!OH 1?!OH 3?!OH 5?!OH 5?!OMe

1!AB dH

2[42 br s 1[62 s

0[44 s 0[44 s

5[94 s

1[88 m 0[69 m 9[88 t "6[3#$

1!AB dC 097[54 076[52 000[28 088[93 33[37 061[98 05[70 192[60 18[43

13[66 13[66 093[68 052[30 095[48 053[19 82[01 051[26 195[51 34[80 07[26 03[14

07[39 s 8[87 s 05[28 s 00[20 s 2[74 s

2!PB dH

2!PB dC

2[20 s 2[05 m 0[06 t "6[1#$ 0[440 s 0[37 s

2[12 m 0[58 m 0[90 t "6[2#$ 0[434 s 0[36 s 07[53 s 01[24 s 07[46 s 01[24 s

009[76 076[65 097[10 088[17 33[41 062[30 07[07 195[41 32[02 07[12 14[40 13[27 009[74 076[79 097[03 087[56 33[39 062[30 196[33 24[97 7[48 03[06 14[37 13[27

2!BB dH

2[20 s 2[04 m 0[58 m 0[90 t "6[3#¦ 0[43 0[37 s

2[04 m 0[58 m 0[90 t "6[3#$ 0[43 s 0[37 s 07[53 s 01[24 s

2!BB dC 009[77 076[66 097[10 088[17 33[41 062[31 07[12 195[42 32[02 07[12 03[06 14[49 13[27 009[77 076[66 097[10 088[17 33[41 062[31 195[42 32[02 07[12 03[06 14[49 13[27

07[53 s 01[24 s

44[71

 Assignments interchangeable[ $ "J in Hz#[

to the known fern phloroglucinol\ aspidin BB[ However\ this possibility was immediately ruled out because of the absence of O!methyl resonances in the 0 H NMR spectrum[ In the 02C NMR spectrum\ there were only 01 carbon signals indicating a symmetrical molecule consisting of two butyryl_licinic acids linked through a methylene bridge[ Unambiguous assign! ment of the 0H and 02C resonances followed from 0HÐ 02 C HMQC and HMBC spectroscopy "Table 0#[ The structure of 2!BB being identical with albaspidin BB was con_rmed by comparison of our NMR data with those reported for a number of acylphloroglucinols ð03Ł[ In the HMBC spectrum\ the `em!dimethyl groups did not interact with C1O carbon of the _l! icinic acid rings[ This indicates that the preferred solu! tion con_rmation is the tautomer with the `em!dime! thyl para to the keto function[ From the HMQC spec! trum it became clear why only 01 signals were observed and not the expected 02 signals] it appeared that carbons 6\ 09 and 09? all resonated at dc 07[12\ since there were two cross peaks for this signal\ one with H!6 at dH 2[20 and the other with the multiplet at

dH 0[58 assigned to H!09:09? by HMBC correlations "Table 0#[ Compound 2!PB "Mr 335\ C13H29O7# was identi_ed as albaspidin PB in a similar fashion[ Due to the near!symmetry of the molecule\ double signals were observed for the ring carbons of both _licinic acid units[ The carbons resonating at the same frequencies as the ring carbons in albaspidin BB were identi_ed as those belonging to the butyryl_licinic acid unit[ This piece of information could not be inferred from HBMC correlations because there were no inter! actions between nuclei across the acyl C1Os[ Assign! ment of the butyryl and propionyl signals readily fol! lowed from the HMBC spectrum[ The acyl carbon at dC 195[4 showed cross peaks with two methylene pro! ton signals "dH 2[05 and 0[58# and the terminal methyl protons at dH 0[90\ which were therefore attributed to the butyryl function[ Likewise\ within the propionyl side chain\ acyl C!7? interacted with the methylene protons at dH 2[12 "H!8?# and the methyl protons at dH 0[06 "H!09?#[ Full assignment of signals is given in Table 0[

823

ECKHARD WOLLENWEBER et al[

Fig[ 0[ HPLC!APCI!MS analysis of crude phloroglucinol mixtures of D[ ar`uta "upper panels# and D[ villarii "lower panels#[ Total ion current "TIC# chromatograms were recorded in the positive ion mode from m:z 059 to m:z 459 "left#[ The acylphloroglucinols yielded predominantly pseudo!molecular ðMHŁ¦ ions "middle# which were collisionally activated to give X and Y cleavage ions in the MS!MS mode "CAD spectra\ right#[

The accompanying minor acylphloroglucinols\ 2! AP\ 2!AB\ and 2!PP\ were obtained from the frond exudate of D[ villarii in very small amounts[ Their structures were tentatively assigned as the albaspidin homologues AP\ PP and AB by mass spectrometry[ Compound 3 "Mr 113\ C01H05O3# was identi_ed as aspidinol B by 0H\ 02C\ 0HÐ02C HMQC and HMBC spectroscopy[ The butyryl function was located para to the methoxy substituent\ because the aromatic pro! ton "H!4?\ at dH 4[85# correlated with all ring carbons but C!1?[ This indicates that the aromatic proton is para to C!1? and meta to C!0? carrying the butyryl side chain[ Mass spectrometry The composition of the crude phloroglucinol mix! tures of D[ ar`uta and D[ villarii was examined by HPLC!APCI!MS in the single and tandem MS modes "Fig[ 0#[ The acylphloroglucinols yielded intense pro! tonated molecular ðMHŁ¦ ions without signi_cant fragmentation in the single MS mode[ Fragmentation of the ðMHŁ¦ ions was induced by collisions with Ar gas molecules in the tandem!MS mode[ Basically\ two fragmentation pathways were discerned\ "0# cleavage of the methylene bridge on either side of C!6\ and "1# loss of the acyl side chain ð05Ł[ The _rst route yielded the most intense fragment ions which proved to be of great diagnostic value in the identi_cation of both acyl residues "Table 1\ Fig[ 1#[ In the albaspidins\ the charge is preferentially retained on the _licinic acid fragments\ XH¦ and YH¦\ after cleavage of the meth! ylene bridge[ By contrast\ para!aspidin!AA and the desaspidins yielded predominantly Y?H¦ and YH¦ ions as inferred from homologues whose XH¦ and

YH¦ fragments di}er in mass[ Hence it seems reason! able to assume that the contribution of the YH¦ ions to the ion current is signi_cantly larger than the con! tribution of XH¦ ions for all desaspidins[ This indi! cates that the charge is retained on the phloroglucinol unit rather than on the _licinic acid unit "cf ion inten! sities in Table 1#[ Clearly\ charge retention is favoured on the least acidic oxygen substituents\ that is\ on the phloroglucinol hydroxyl because the _licinic acid oxygen substituents participate in tautomeric proton! exchange processes[ As can be seen from Table 1\ di}erences in the ion intensities of the Y?H¦ ions allow distinction between desaspidins and their isomeric albaspidins[ Consider\ for instance\ the two isomeric acylphloroglucinols\ 1!AB and 2!AP\ whose X cleav! age products are of equal mass "and Y products of equal mass#[ They di}er signi_cantly in the intensities of the Y?H¦ ions] 75) observed for 1!AB and only 01) for 2!AP "Fig[ 0\ Table 1#[ In the second pathway\ fragment ions are formed by loss of an acyl neutral\ O1C1CHR\ from the ðMHŁ¦ ion which are of additional diagnostic value in the albaspidins[ Loss of the acyl substituent from both XH¦ and YH¦ yields protonated _licinic acid with m:z 044[ Although the desaspidins contain an acyl_licinic acid unit\ the m:z 044 ion was only observed in spectra of the albaspidins "ðX−O1C! 1CHR\ usually × 09) of the base peak#\ and thus provides another diagnostic ion to distinguish between desaspidins and albaspidins[ Additional frag! ment ions may arise as a result of loss of water from ðMHŁ¦\ XH¦ and YH¦ ions[ Acylphloroglucinols have previously been analysed by other {soft| ionization techniques\ i[e[ CI!MS\ FI! MS\ FD!MS and FAB!mass "cf ð06Ł and refs cited

Leaf surface acylphloroglucinols from Dryopteris

824

Table 1[ APCI!CAD mass spectral data for phloroglucides 0Ð2 ðm:z "rel[ int[# $Ł Compound Para!aspidin 0!AA Desaspidins 1!AA 1!AP 1!AB 1!PB Albaspidins 2!AA 2!AP 2!AB 2!PP 2!PB 2!BB

Mr

XH¦

Y?H¦

X?H¦

YH¦

393

086 "099#

198 "87#

198 "87#

086 "099#

289 393 307 321

086 "01# 086 "099# 086 "05# 100 "099#

084 "099# 198 "77# 112 "75# 112 "77#

198 "47# 198 "77# 198 "21# 112 "77#

072 "82# 086 "099# 100 "099# 100 "099#

393 307 321 321 335 359

086 "099# 086 "099# 086 "099# 100 "099# 100 "88# 114 "099#

198 "57# 112 "01# 126 "12# 112 "25# 126 "124# 126 "45#

198 "57# 198 "8# 198 "14# 112 "25# 112 "14# 126 "45#

086 "099# 100 "80# 114 "86# 100 "099# 114 "099# 114 "099#

 Atmospheric pressure chemical ionization!collision activated decomposition[ Compare Fig[ 1 for structures of fragment ions[ Minor fragments are listed in the Experimental section[ $ Principal ions are printed in bold type face[ See also text for further explanation[

Fig[ 1[ APCI!CAD mass fragmentation of desaspidins "top# and albaspidins "bottom#[ For intensities of X and Y cleavage ions\ see Table 1[

therein#[ By FAB mass spectrometry\ Widen et al[ ð06Ł were able to obtain ðMŁ¦s of albaspidins up to the hexamer[ Dimeric albaspidins yielded similar spectra

with APCI!MS in our study[ Without further exam! ples of APCI!MS analysis of olgomeric phloro! glucinols it is di.cult to estimate the upper limit of

825

ECKHARD WOLLENWEBER et al[

the molecular range up to which molecular infor! mation can be obtained by APCI!MS[ Clearly\ APCI has the advantage over FAB mass spectrometry in that the lower mass region is not obscured by matrix ions and that it is well!suited for coupling with ana! lytical HPLC[ Although vaporization of the HPLC e/uent involves heat\ artifacts due to Rottlerone rearrangement\ often seen in EI mass spectra of asym! metric oligomeric acylphloroglucinols ð07Ł\ were not observed in our APCI mass spectra of asymmetric albaspidins[ For instance\ APCI of 2!PB ðMH¦ at m:z 336Ł did not give rise to expectable Rottlerone artifacts with m:z 322 and m:z 350[ Hypothetically\ Rottlerone rearrangement of an albaspidin X0CH10Y may give rise to two symmetric albaspidins X0CH10X and Y0CH10Y by recombination of a neutral fragment X with X?H¦ or Y with Y?H¦ in the ion source "cf Fig[ 1#[

Localization and distribution of acylphloro`lucinols in Dryopteris Light microscopic examinations of the three fern species revealed the presence of glandular trichomes on their leaf surfaces[ The ultrastructure of the leaf glands of D[ villarii and C[ robertiana was investigated by scanning electron microscopy of fresh material "Fig[ 2#[ The glandular trichomes of both species were similar[ They appear to be unicellular trichomes with round to orbicular heads "Fig[ 2"c##\ thus cor! responding with the glandular trichomes depicted earlier for D[ sparsa ð08Ł or reported for Arachnoides species ð19Ł[ According to these authors\ the exudate is deposited on the distal end of these trichomes\ pos! sibly between cell wall and cuticle ð10Ł[ In view of the earlier discussed correlation between ~avonoid agly! cones and the presence of secretory structures ð11Ł\ the leaf surface phloroglucinols of both Dryopteris ferns are presumably produced by their glandular trich! omes[ This hypothesis also parallels earlier obser! vations that Dryopteris phloroglucinols are excreted along with other resinous materials by internal glands present in the rhizomes and petiolar bases ð12Ł[ Draw! ings as well as electron microscopic pictures showing the internal glandular hairs of D[ assimilis Walker "  D[ expansa "Presl# Fraser!Jenkins and Jermy# were published in 0868 ð10Ł[ In that species\ glandular hairs of 59Ð89 mm length were observed in large inter! cellulars in the rhizomes[ It was demonstrated that the phloroglucinols were biosynthesized in the glandular cytoplast and subsequently accumulated in a sub! cuticular excretion layer[ Some years earlier\ a draw! ing of the external secreting hairs of D[ fra`rans "L[# Schott was published ð13Ł[ These glandular hairs occur on the epidermis of the rhizomes and petiolar bases[ They were said to contain much secreted material between the cuticle and the outer layer of the cell[ Internal secreting hairs were not round in this species[ Similar drawings exist for D[ campyloptera "Kuntze#

Clarkson\ D[ expansa\ and D[ intermedia "Muehl# A[ Gray ð14Ł[ Comparison of the leaf surface phloroglucinols identi_ed in this study with rhizome phloroglucinols reported for D[ villarii and D[ ar`uta ð15Ł indicates that the leaves di}er from the rhizomes with regard to phloroglucinol composition[ Widen et al[ ð15Ł detected ~avaspidic acids "BB and AB#\ para!aspidin BB\ and _lixic acids "BBB and ABB# in the rhizomes of both D[ villarii and D[ ar`uta[ In D[ villarii rhizomes\ these acylphloroglucinols were found along with small amounts of desaspidin BB and albaspidin BB[ The qualitative compositions of D[ villarii and D[ ar`uta rhizomes from di}erent sources proved to be similar ð15Ł[ At this point it is unknown to what extent the qualitative di}erences are associated with organ to organ variability[ Comparative analyses of leaves and rhizomes from individual plants will shed more light on this question[ On the other hand\ the absence of acylphloroglucinols on the leaf surface of C[ robertiana is in agreement with an earlier report of absence of acylphloroglucinols in the rhizomes of this fern ð8Ł[ At this point we can only speculate about the eco! physiological function of leaf surface phloroglucinols of Dryopteris[ It is conceivable that the distribution of leaf surface acylphloroglucinols reaches beyond these two species of Dryopteris[ As with externally accumu! lated ~avonoid aglycones\ the aspect of localization of phloroglucinols in the plant certainly deserves attention when biological e}ects are of interest[ For example\ literature reports on the neurotoxic e}ects of acylphloroglucinols in cattle "drowsiness and blind! ness# eating the rhizomes and leaves of male fern "D[ _lix!mas# are exclusively focused on the rhizomes ð16Ł[

EXPERIMENTAL

Plant material Dryopteris villarii was collected by R[ Mues "Saar! brucken# in August 0884 and by E[ Wollenweber "Darmstadt# in October 0885 at the Schynige Platte in the Berner Oberland\ Switzerland "alt[ ca 0859 m#[ Dryopteris ar`uta was collected by J[ N[ Roitman on 03 March\ 0883\ in oak woodland near the El Carise Oaks Campground on hwy 63\ 12 miles east of Inter! state 4 "San Juan Capistrano#"elev[ ca 799 m#\ Riv! erside Co[\ CA[ Currania robertiana was collected by R[ Mues in July 0884 near Ilanz in Graubunden\ Swit! zerland and by E[ Wollenweber in the Botanischer Garten der TH Darmstadt[ Fronds were carefully clipped in the _eld and air!dried in paper bags[ Vou! chers are kept at the Universities of Saarbrucken and Darmstadt[ NMR 0

H "599 MHz# and 02C "049[8 MHz#^ CDCl2 at room temp[ Albaspidins 1!PB and 1!BB were run at 172 K

Leaf surface acylphloroglucinols from Dryopteris

826

Fig[ 2[ SEM views of glandular trichomes on the leaf surface of "a# D[ villarii "epidermal walls shrank on drying#^ "b# C[ robertiana "lower leaf surface\ survey#^ "c# C[ robertiana "detail#[ For magni_cation see scale bar on the photographs[

to reduce keto!enol tautomerism[ TMS and CHCl2 resonances "dH 6[15 ppm\ dC 62[12 ppm# were used as int[ chemical shift references[ 0HÐ02C HMQC and

HMBC experiments were performed using standard pulse sequences[ Spectral widths of 19 and 119 ppm were used in the H and C dimensions\ respectively[

827

ECKHARD WOLLENWEBER et al[

Mass spectrometry APCI!MS] PE Sciex API III plus triple quadrupole instrument[ Samples were introduced by loop injec! tion or by HPLC via a heated nebulizer interface kept at 399> which heats the column e/uent to ca 019>[ Ionization of the analyte vapour mixt[ was initiated by a corona discharge needle at ca 5 kV and a discharge current of ca 2 mA[ The ori_ce plate voltage was set at 44Ð54 V in the positive ion mode[ MS!MS experiments were performed with ArÐN1 "8]0# as collision gas at a thickness of ca 0[8×0903 atoms−1[ The collision energy was 04 V[ Other operating conditions were standard[ Isolation and chromato`raphic procedures Field!collected air!dried materials as well as fresh leaves from the Botanical Garden were brie~y rinsed with Me1CO to dissolve the exudates[ After evapn of the solvent\ the residues of D[ villarii and D[ ar`uta were redissolved in a small amount of boiling MeOH\ cooled to −07> and centrifuged to eliminate the major part of the lipophilic constituents[ The supernatant was chromatographed on polyamide SC!5\ eluted with toluene and increasing amounts of MeCOEt and MeOH[ In C[ robertiana the residue was directly sub! jected to CC on acetylated polyamide[ Fractions were monitored by TLC on polyamide "DC!00\ Macherey! Nagel# with solvents A ðpetrol "bp099Ð039>#ÐtolueneÐ MeCOEtÐMeOH 01]5]0]0Ł\ B ðtolueneÐpetrol "bp099Ð039>#Ðmethyl ethyl ketoneÐMeOH 01]5]1]0Ł and C "tolueneÐMeCOEtÐMeOH 01]4]2# and on sil! ica with solvents D "tolueneÐMeCOEt 8]0# and E "tolueneÐdioxaneÐHOAc 07]4]0#[ Chromatograms were viewed under UV "255 nm# before and after spraying with Natursto}reagenz A[ Identi_cation of ~avonoids was achieved by co!TLC with authentic markers available in E[W|s lab[ HPLC Analyt[ RP!HPLC] 4 mm LiChrospher RP!07 column "149×3 mm# at 0 ml min−0 with a linear solvent gradient from 79) to 84) MeCN in 0) aq[ HCOOH over 04 min followed by a further 04 min at 84) MeCN[ Semi!prep RP!HPLC] 09mm Econosil RP!07 column "149×09 mm# at 4 ml min −0[ D[ villarii phloroglucinols were separated isocratically with 89) MeCN in 0) aq[ HCOOH[ For semi!prep isolation of D[ ar`uta phloroglucinols\ a linear solvent gradient from 79) to 84) MeCN in 0) aq[ HCOOH over 04 min was used\ followed by a further 04 min at 84) MeCN[ The UV trace was recorded at 179 nm[ Peak frs were collected manually and evapd on a rotavapor and by lyophilization[ Milligram or sub!mg amounts were obtained as white to cream!coloured amorphous powders[ Para!aspidin!AA "0!AA#[ 0H NMR] d 0[44 "5H\ s\ × C"Me#1#\ 1[03 "2H\ s\ arom[ Me#\ 1[60 and 1[63

"each 2H\ s\ 1×acetyl Me#\ 2[46 "1H\ br\ s\ H!6#\ 2[63 "2H\ s\ OMe#\ 09[99\ 00[49\ 04[47 and 07[30 "s\ 3×OH#^ APCI!MS\ m:z "rel[ int[#] 394 ðMHŁ¦"099#[ APCI!CAD!MS] Table 1 and m:z 276 ðMH−H1OŁ¦"09#\ 068 ðYHÐH1OŁ¦ "11#[ Desaspidin "1!AA#[ 0H NMR] d 0[43 "5H\ s\ × C"Me#1#\ 1[52 and 1[62 "each 2H\ s\ 1×acetyl Me#\ 2[43 "1H\ br s\ H!6#\ 2[74 "2H\ s\ OMe#\ 5[94 "0H\ s\ H!4?#\ 8[81\ 00[24\ 05[12 and 07[30 "s\ 3×OH#^ APCI!MS\ m:z "rel[ int[#] 280 ðMHŁ¦"099#[ APCI! CAD!MS] Table 1 and m:z 262 "MHÐH1OŁ "5#\ 054 ðYHÐH1OŁ¦ "11#[ Desaspidin AP "1AP#[ 0H NMR] d 0[06 "2H\ t\ J  6[1 Hz\ H!09?#\ 0[45 "5H\ s\ ×C"Me#1#\ 1[62 "2H\ s\ acetyl Me#\ 2[94 "1H\ m\ H!8 ?#\ 2[42 "1H\ br s\ H!6#\ 2[74 "2H\ s\ OMe#\ 5[95 "0H\ s\ H!4?#\ 8[86\ 00[20\ 05[23 and 07[30 "s\ 3×OH#[ APCI!MS\ m:z "rel[ int[#] 394 ðMHŁ¦ "099#^ APCI!CAD!MS] Table 1 and m:z 276 ðMHÐH1OŁ¦ "3#\ 068 ðYHÐH1OŁ¦ "10#[ Deaspidin AB "1!AB#[ 0H and 02C NMR] Table 0^ APCI!MS\ m:z "rel[ int[#] 308 ðMHŁ¦ "099#^ APCI! CAD!MS] Table 1 and m:z 390 ðMHÐH1OŁ "4#\ 082 ðYHÐH1OŁ¦ "13#[ Desaspidin PB "1!PB#[ 0H NMR] d 9[88 "2H\ t\ J  6[3 Hz\ H!00?#\ 0[06 "2H\ t\ J  6[1 Hz\ H!09#\ 0[43 "5H\ s\ ×C"Me#1#\ 0[69 "1H\ m\ H!09?#\ 1[82Ð2[96 "3H\ m\ H!8 and H!8?#\ 2[42 "1H\ br s\ H!6#\ 2[74 "2H\ s\ OMe#\ 5[94 "0H\ s\ H!4?#\ 8[83\ 00[22\ 05[27 and 07[36 "s\ 3×OH#^ APCI!MS\ m:z "rel[ int[#] 322 ðMHŁ¦ "099#[ APCI!CAD!MS] Table 1 and m:z 304 ðMHÐH1OŁ¦ "3#\ 252 ðMH−O1C1CH0C1H4Ł¦ "1#\ 082 ðYHÐH1OŁ¦ "07#[ Albaspidin AA "2!AA#[ APCI!MS\ m:z "rel[ int[#^ 394 ðMHŁ¦ "099#^ APCI!CAD!MS] Table 1 and m:z 252 ðMH−O1C1CH1Ł¦ "5#\ 068 ðXHÐH1OŁ¦ "32#\ 044 ð_licinic acid¦HŁ¦ "16#[ This product was not isolated but was kindly provided by Prof[ C[ J[ Widen for comparison purposes[ Albaspidin AP "2!AP#[ APCI!MS\ m:z "rel[ int[#] 308 ðMHŁ¦ "099#[ APCI!CAD!MS] Table 1 and m:z 266 ðMHÐO1C1CH1Ł¦ "1#\ 252 ðMHÐ O1C1CH0CH2Ł¦ "5#\ 068 ðXHÐH1OŁ¦ "3#\ 082 ðYHÐH1OŁ¦ "00#\ 044 ð_licinic acid¦HŁ¦ "3#[ Albaspidin AB "2!AB#[ APCI!MS\ m:z\ "rel[ int[#] 322 ðMHŁ¦ "099#[ APCI!CAD!MS] Table 1 and m:z 280 ðMHÐO1C1CH1Ł¦ "3#\ 252 ðMHÐ O1C1CH1C1H4Ł¦ "11#\ 196 ðYHÐH1OŁ¦ "34#\ 068 ðXHÐH1OŁ¦ "03#\ 044 ð_licinic acid ¦ HŁ¦ "03#[ Albaspidin PP "2!PP#[ APCI!MS\ mz "rel[ int[#] 322 ðMHŁ¦ "099#[ APCI!CAD!MS] Table 1 and m:z 266 ðMHÐO1C1CH0CH2Ł¦ "05#\ 082 ðXHÐH1OŁ¦ "31#\ 044 ð_licinic acid ¦ HŁ¦ "03#[ Albaspidin PB "2!PB#[ 0H and 02C NMR] Table 0^ APCI!MS\ m:z "rel[ int#] 336 ðMH#¦ "099#[ APCI! CAD!MS] Table 1 and m:z 280 ðMHÐ O1C1CH0CH2Ł¦ "05#\ 266 ðMHÐ O1C1CH0C1H4Ł¦ "17#\ 196 ðYHÐH1OŁ¦ "15#\ 082 ðXHÐH1OŁ¦ "15# ð_licinic acid ¦ HŁ¦ "03#[ Albaspidin PB "2!PB#[ 0H and 02C NMR] Table 0^ APCI!MS\ m:z "rel[ int#] 350 ðMHŁ¦ "099#[ APCI!

Leaf surface acylphloroglucinols from Dryopteris

CAD!MS] Table 1 and m:z 280 ðMHÐ O1C1CH0C1H4Ł¦ "52#\ 196 ðXHÐH1OŁ¦ "44#\ 044 ð_licinic acid ¦ HŁ¦ "08#[ 2!BB was identical with authentic albaspidin BB by HPLC and APCI!MS! MS[ Aspidinol B "3#[ "1?\5?!Dihydroxy!3?!methoxy!2?! methyl!l?!butyrophenone#[ 0H NMR] d 0[99 "2H\ t\ J  6[1 Hz\ H!3#\ 0[62 "1H\ m\ H!2#\ 1[91 "2H\ s\ Me! 2?#\ 2[97 "1H\ t\ J  6[1 Hz\ H!1#\ 2[70 "2H\ s\ OMe! 3?#\ 4[85 "0H\ s\ H!4?#\ 8[58 and 09[13 "1? and 5?!OH#^ 02 C NMR] d 6[1 "Me!2?#\ 03[0 "C!3#\ 07[4 "C!2#\ 35[2 "C!1#\ 44[8 "OMe!3?#\ 093[2 "C!2?#\ 81[0 "C!4?#\ 049[4 "C!0?#\ 059[2 "C!1?#\ 050[3 "C!5?# 052[5 "C!3?#\ 193[5 "C!0#[ APCI!MS\ m:z "rel[ int[#\ 114 ðMHŁ¦ "099#^ APCI!CAD!MS\ m:z] 114 ðMHŁ¦ "099#\ 196 ðMHÐ H1OŁ¦ "68#\ 072 "34#\ 068 ð196−COŁ¦ "69#\ 040 ð196− 1COŁ¦ "69#\ 002 "25#[ Compound 3 was identical with authentic aspidinol B by HPLC and APCI!MS!MS[ Acknowled`ements*E[W[ wishes to thank Prof[ Dr R Mues "Saarbrucken# who drew his interest to these fern species and supplied some of the material studied here[ Thanks are also due to Prof[ Dr J[ Schneller "Zurich# who supplied two plants of Currania rober! tiana\ to Dr J[ N[ Roitman "Albany\ CA# who col! lected Dryopteris ar`uta\ to Mrs Marion Dorr "Darm! stadt# for technical assistance^ to Mrs H[ Wilhelmi and Prof[ Dr W[ Barthlott "both at the Botanisches Institut der Universitat Bonn# for SEM photographs and permission to publish them\ and to Prof[ Dr V[ L[ Hsu "OSU\ Corvallis# for access to NMR facilities[ We further wish to thank Prof[ Dr C[!J[ Widen Ý yras "Turku\ "Helsinki\ Finland# and Prof[ Dr P[ A Finland# for a gift of marker compounds\ for reprints and for valuable discussions[

REFERENCES

0[ Wollenweber\ E[\ Biochemical Systematics and Ecolo`y\ 0878\ 06\ 030[ 1[ Wollenweber\ E[ and Roitman\ J[ N[\ Zeitschrift fur Naturforschun`\ 0880\ 35c\ 214[ 2[ Roitman\ J[ N[\ Wong\ R[ Y[ and Wollenweber\ E[\ Phytochemistry\ 0882\ 23\ 186[ 3[ Ruedi\ P[\ Wollenweber\ E[\ Marx\ D[ and Scheele\ C[\ Zeitschrift fur Naturforschun`\ 0878\ 33c\ 890[ 4[ Arriaga!Giner\ J[ F[\ Rumbero\ A[ and Wol! lenweber\ E[\ Zeitschruft fur Naturforschun`\ 0881\ 36c\ 497[

828

5[ Wollenweber\ E[\ in] Flavonoids and Bio~a! vonoids\ Proc[ Int[ Bio~[ Symp[\ Vienna\ 0884\ ed[ S[ Antus\ M[ Gabor and K[ Vetschera[ Akademiai Kiado\ Budapest\ 0885\ pp[ 100Ð119[ 6[ Wollenweber\ E[ and Arriaga!Giner\ F[ J[\ Phyto! chemistry\ 0880\ 29\ 1296[ 7[ Wollenweber\ E[\ Scheele\ C[ and Tryon\ A[ F[\ American Fern Journal\ 0876\ 66\ 17[ 8[ Widen\ C[!J[\ Sarvela\ J[ and Britton\ D[ M[\ Annales Botanici Fennici\ 0872\ 19\ 396[ 09[ Widen\ C[!J[\ Fraser!Jenkins\ C[\ Reichstein\ T[\ Gibby\ M[ and Sarvela\ J[ Annales Botanici Fennici\ 0885\ 22\ 58[ 00[ Penttila\ A[ and Sundman\ J[\ Journal of Phar! macy and Pharmacolo`y\ 0869\ 11\ 282[ 01[ Euw\ J[ v[\ Lounasmaa\ M[\ Reichstein\ T[ and Widen\ C[ J[\ Studia Geobotanica\ 0879\ 0\ 164[ 02[ Murakami\ T[ and Tanaka\ N[ in Pro`ress in the Chemistry of Or`anic Natural Products\ 0877\ 43\ 0Ð209[ Springer\ Verlag\ Vienna[ 03[ A Ý yras\ P[\ Lotojonen\ S[ and Widen\ C[!J[\ Or`anic Ma`netic Resonance\ 0870\ 05\ 198[ 04[ Wollenweber\ E[\ Dietz\ V[ H[\ Schilling\ G[\ Favre!Bonvin\ J[ and Smith\ D[ M[\ Phyto! chemistry\ 0874\ 13\ 854[ 05[ Lounasmaa\ M[\ Karjalainen\ A[\ Widen\ C[ J[ and Huhtikangas\ A[\ Acta Chemica Scan! dinavica\ 0861\ 15\ 77[ 06[ Widen\ C[!J[\ Pyysalo\ H[ and Reichstein\ T[\ Hel! vetica Chimica Acta\ 0883\ 66\ 0874[ 07[ Lounasmaa\ M[\ Widen\ C[ J[ and Reichstein\ T[\ Helvetica Chimica Acta\ 0860\ 33\ 1749[ 08[ Widen\ C[ J[\ Sarvela\ J[ and Iwatsuki\ K[\ Bot! anical Ma`azine Tokyo\ 0865\ 78\ 166[ 19[ Widen\ C[ J[\ Mitsuka\ S[ and Iwatsuki\ K[\ Bot! anical Ma`azine Tokyo\ 0870\ 83\ 016[ 10[ Huurre\ A[\ Huhtikangas\ A[ and Widen\ C[ J[\ Planta Medica\ 0868\ 24\ 151[ 11[ Wollenweber\ E[ and Dietz\ V[ H[\ Phyto! chemistry\ 0870\ 19\ 758[ 12[ Huhtikangas\ A[\ Huurre\ A[ and Partanen\ A[\ Planta Medica\ 0879\ 27\ 51[ 13[ Widen\ C[!J[ and Britton\ D[ M[ Canadian Journal of Botany\ 0860\ 38\ 878[ 14[ Widen\ C[!J[ and Britton\ D[ M[\ Canadian Jour! nal of Botany\ 0858\ 36\ 0226[ 15[ Widen\ C[!J[\ Widen\ K[\ Vida\ G[ and Reichstein\ T[\ Botanica Helvetica\ 0880\ 090\ 66[ 16[ Clarke\ M[ L[\ Harvey\ D[ G[ and Humphreys\ D[ J[\ Veterinary Toxicolo`y\ 1nd Ed[\ Bailliere Tindall\ London\ 0870\ pp[ 086Ð087[