Induction of a Complete Secondary-Product Pathway in a Cultured Lichen Fungus

EXPERIMENTALMYCOLOGY 16,52-63(1992)

CHICITA F. CULBERSON AND DANIELE ARMALEO

Department of Botany, Duke University, Durham, North Carolina 27706 Accepted for publication October 21, 1991

CULBERSON, C. F., AND ARMALEO, D. 1992. Induction of a complete secondary-product pathway in a cultured lichen fungus. Experimental Mycology 16, 52-63. Experimental studies on the secondary metabolism characteristic of lichens have been impeded by the slow growth of the fungi and by the inconsistent results of many attempts to induce the pathways in the fungi isolated from their photosynthetic partners. In the present study, a lichen-specific secondary pathway was consistently induced in a lichen fungus (Cladonia grayi) grown in the absence of the alga. The depside (4-0-demethylsphaerophorin) and two depsidones (grayanic and 4-0-demethylgrayanic acids) found in the natural lichen began to accumulate a few days after the transfer of lightly fragmented mycelia from liquid to solid medium. Induction was enhanced on drier substrates and was correlated with the proliferation of aerial hyphae, where the major product (grayanic acid) accumulated in extracellular patches visible by fluorescence microscopy. The time course was analyzed by quantitative high-performance liquid chromatography of extracts from small cultures grown on nylon filters. Induction was rapid in view of the slow growth of the fungus, and secondary productivity was comparable to that of some nonlichen fungi. These results cot&m that the alga is not needed for catalysis in lichen depside and depsidone biosynthesis and suggest instead that the characteristic secondary metabolism of natural lichen is linked to their aerial habit of growth. 0 1992 Academic Press, Inc. INDEX DESCRIPTORS: Secondary metabolism; polyketides; lichen fungi; Cladonia grayi; grayanic acid; depsides; depsidones; aerial hyphae; photobiont; hyphal differentiation.

Lichens produce hundreds of unique ex- symbiotic associations between fungi (my- tracellular secondary products of which cobionts) and green algae or cyanobacteria many belong to classes of compounds (photobionts). The partners can be cultured rarely found in other organisms (Elix et al., separately, but the natural equilibrium be- 1984). Their chemical structures and occur- tween the combined symbionts is not easily rences in thousands of species have been maintained under laboratory conditions. extensively studied using natural material For some, however, symbiotic associations (C. F. Culberson et al., 1977; C. F. Culber- have been successfully established or main- son and Elix, 1990). Some lichen products tained in vitro (reviews: Ahmadjian, 1980; have biological activity of interest in medi- Ahmadjian and Jacobs, 1985; Bubrick, cine and agriculture and many are used in 1988). phylogenetic reconstructions for lichen While the controls and genetics of sec- fungi (W. L. Culberson, 1986; Lawrey, ondary metabolism in nonlichen fungi are 1984; Sankawa et al., 1982; Umezawa et extensively studied (reviews: Drew and De- al., 1983). Despite extensive knowledge of main, 1977; Garraway and Evans, 1984; the structures and distributions of the com- Martin and Liras, 1989), the literature of- pounds, the regulation and biological signif- fers no clear clues to critical factors affect- icance of secondary metabolism remains ing analogous pathways in isolated lichen unclear, in part because of complications fungi. The many efforts to detect secondary arising from the lichen’s dual nature. Li- products in cultures of lichen fungi have chens are stable, structurally organized, yielded inconsistent results. A few isolates

52 0147-5975192 $3.00 Copyright 0 1992 by Academic Press, Inc. AU rights of reproduction in any form reserved. INDUCTION OF A LICHEN SECONDARY PATHWAY 53 have spontaneously produced these com- grayi Merr. ex Sandst. (North Carolinas pounds, some even copiously (Leuckert et Mitchell Co., Roan Mountain, W, L. Cua- al., 1990), but more often the cultured fun- berson 19971, DUKE). The culture mediu gus either lacks secondary products or syn- was a modified Lilly and Barnett (LB)” me- thesizes compounds not found in the natu- dium with proline (2 g/liter) instead of as- ral lichen (Ahmadjian and Culberson, un- paragine and half (10 g/liter) the usual con- published results; Fox and Huneck, 1969; centration of glucose (Ahmadj’ Hamada and Ueno, 1987). In contrast to the Solid media containing 1.5% largely negative results for the fungus cul- (Difco) or agarose (Bio-Rad) were delivere tured alone, many isolates consistently bio- @-ml aliquots in 5-cm plastic Petri d synthesize characteristic secondary prod- by pipet to increase uniformity. E ucts when they are artificially lichenized or moisture was removed by leavi even simply grown in the presence of their ered plates in the reverse-flow sterile hood normal photobiont (e.g., C. F. Culberson for 24 h (“moist” plates) or for approxi- and Ahmadjian, 1980; Ahmadjian and Ja- mately 3 days (“semi-day” pla cobs, 1985; Kon et al., 1990; Yamamoto et drying and during the course of al., 1985). This frequent correlation of li- iment, plates were inverted and chenization or coculture with the produc- Parafilm. Mycelia for the induction experi- tion of lichen-specific secondary com- ments came from stock cultures, typi pounds has reinforced the idea that both in 50-ml glass tubes containing 10 ml o partners play essential roles in secondary dium and IO g of acid-washed beads (Thom- metabolism (C. F. Culberson and Ahmad- as Scientific, 0.5 mm diam.). The liquicl jian, 1980; Hamada, 1988). However, the medium was changed every 2-3 onth * definition of these roles has remained elu- and mycelia were fragmented by rapid vor- sive mostly because the slow growth of li- texing (5-10 s) once a month (Arm chen fungi has impeded experimental studies. 199 1). In the preparation of HPLC sa By combining two technologies, a culture all solvents and reagents were HPLC- method on microfilters over solid media grade. Plastics were used sparingly (Mulvany, 1969; Oliver et al., 1989) and were prewashed, nylon filters wi.th me quantitative high-performance liquid chro- no1 and microfuge tubes with acetone, Sub- matography, the present study shows that stitution of electrophoresis-grade agarose secondary metabolism can be induced and for agar in the solid medium irnpr~v~~l followed in submilligram cultures of an iso- HPLC analysis of trace amounts of second- lated lichen fungus. The fungus was Clu- ary compounds by reducing ba~k~~~~ doniu gruyi, from a common lichen produc- peaks. Blank filters from LB Petri plate ing characteristic lichen-specific polyketide served as controls for residual mediu derivatives. The experiments define critical derived artifacts ~ factors affecting the induction of this typi- cal secondary pathway in culture, the hy- Transfer of Mycelia to Solid ~ed~~~ phal sites and patterns of secondary- product accumulation, and the biosynthetic The nylon filters (Fisher, Magna Nymph capacity of the mycobiont. 66, 20 pm pore-size) to receive liquid- grown mycelia were washed five times in MATERIALS AND METHODS

Strains, Media, and Reagents 1 Abbreviations used: LB, Lilly and Barnett medium; GR, grayanic acid; DMGR, 4-O- The fungus used in this work was a sin- demethylgrayanic acid; DMSPH, 4-O-demethyl- gle-spore isolate (30-197#8) of Cludonia sphaerophorin. 54 CULBERSON AND ARMALEO methanol and dried, and the weights (Cahn times), each extraction for ca. 2 min at 25 microbalance) were penciled along the room temperature. The extracted filters margins. The filters were autoclaved and all and attached fungal mats, which remained subsequent manipulations were performed intact, were used for the determination of under sterile conditions in a reverse-flow biomass (see below). The six extracts were hood. Immediately before transfer onto fil- pooled and evaporated under a stream of ters, a liquid culture (typically 2-4 weeks nitrogen, at reduced pressure, with gentle after the last medium change) was diluted heat (ca. 40°C). The residue was taken up in with fresh medium to l-2 mg mycelium (dry acetone (washing five times with 0.1 ml) weight)/ml. Equal-volume samples (0.5 or and transferred quantitatively to a 1.5-ml 1 .O ml) were withdrawn from the stirred hy- microfuge tube. After evaporating the sol- phal suspension with an automatic pipet fit- vent, the solid on the wall of the microfuge ted with tips cut off to enlarge their opening tube was concentrated to the bottom with to about 3 mm. Each sample was delivered ca. 100 ~1 of acetone, vortexing and spin- onto a separate nylon filter placed under ning the solution, and evaporating the sol- suction on a fritted glass surface padded vent. For early time-point samples (through with filter paper (Whatman No. 1). To con- 16 days), the use of acetone to transfer the fine the mycelia to uniform areas on the residue from the original methanol/water nylon filters, a tube of appropriate size was extract to the microfuge tube gave the centered on the filter until all of the liquid cleaner sample needed to quantify the small had passed through. Most experiments amounts of secondary products. At the used filters 13 mm in diameter, with 0.1-0.5 longer time points, however, an acetone- mg mycelia seeded through inverted 1.5-ml insoluble residue began to accumulate, polypropylene microfuge tubes with lids trapping a fraction of the secondary prod- and bottoms cut off. Larger filters (25 mm ucts. To quantify the trapped fraction, this in diameter) used with larger amounts of residue was dissolved in 90% methanol. Al- mycelia (l-3 mg) were seeded through 17 X ternatively, to bypass the acetone-insoluble loo-mm polypropylene test tubes. The rim precipitate at the longer time points, the of the tube pressed against the padded filter residue from the initial 90% methanol ex- provided an efficient seal so that each filter tract was transferred to a microfuge tube received a well-defined circular mat of and concentrated to the bottom using only mycelia covering all but 2-3 mm at the mar- 90% methanol. In either case, the final res- gin. Each inoculated filter was placed im- idues were quantitatively resuspended in mediately onto solid medium (typically one measured volumes of methanol immedi- per plate) and incubated at room tempera- ately before HPLC. ture (25’C). Identification of Secondary Products Filter Harvest, Extraction, and HPLC Sample Preparation The isocratic HPLC method for the initial identification of compounds and the Filters were removed from agar media at first time-course experiment used a Beck- each time point and stored immediately at man (Gold Star) Ultrasphere ODS 5-pm - 100°C for analysis after the last time (4.6 x 250 mm) column, a methanol-water- point. Upon removal from the freezer, fil- acetic acid solvent (82: 18: 1.6, v/v/v) flow- ters from the same time point were gener- ing at 1.0 ml/min, and detection at 270 nm. ally extracted together in a lo-ml beaker For compounds in extracts of filter cul- with OS- to l-ml portions of 90% methanol- tures, capacity factors (k’) and retention in- water (four times) and methanol (two dex values relative to methyl ketone con- INDUCTION OF A LICHEN SECONDARY PATHWAY 55 trols (Z) were compared to an extensive da- containing dodecanophenone (0 g/ml; tabase for lichen products, including the Aldrich Chemical Co.) as an inte stan- secondary products of natural C. grayi and dard, to give on-scale peaks with 5-p,l injec- its related chemotypes in the C. chlo- tions in all analyses so that the limits of rophaea complex (C. F. Culberson et al., solubility of the secondary pr~d~ct§ in 1985; C!. F. Culberson and Elix, 1990). Fi- methanol were never approached. Peak nal identifications were based on direct broadening, observed in chromatograms of TLC and HPLC comparisons with authen- longer time-point samples, could be virtu- tic samples (Table 1): natural GR and 4-0- ally eliminated by adding 1% trifl~or~acetic demethylgrayanic (DMGR) acids from C. acid to the injection solvent. Quan grayi and C. anitae (W. L. Culberson et by HPLC-peak area used stand al., 1982), respectively, and synthetic 4-0- tions of pure GR and DMSP demethylsphaerophorin (DMSPH) (Elix sponse factors relating HPLC-peak area to and Wardlaw, 1987). Thin-layer chromatog- nanomoles of compound are not rn~as~r~ raphy (TLC) was on Merck Silica Gel F-254 ably different for depsidones like plates, in three solvent systems (Solvents DMGR that differ only by 4~~~rn~tbylati~ A, B, and C) by methods standardized for (unpublished result), but those for GR an lichen products (C. F. Culberson and Elix, DMSPH (570,000 and 858,000 area units/ 1990). DMSPH, tentatively identified (by nmol, respectively) reflected the st HPLC) as a trace component of an acetone absorbance of depsides compared to extract of freshly collected natural C. grayi dones at the detection wavelengt (350 mg) (North Carolina: Orange Co., near Chapel Hill, Armako, 1991, DUKE), was Biomass Determination confirmed as follows: After precipitation of Mycelia on each filter increased in den- most of the GR from the crude extract, pre- sity and height during the experiments, but parative TLC (Solvent C) enrichment of hyphae did not extend horizontally keypad trace components comigrating with authen- the original boundary of the se patch tic DMSPH, and elution from the spot fur- and only to a limited extent trated ther separated by 2D-TLC (Solvent C, then through the filter into the agar. The biomass B), the purity and identity of the iso- and secondary-product measurements in- lated DMSPH were confirmed by HPLC. volved only the material within and above the filter. After the metha~ol/water extrac- IX Quantification of tion of secondary s, the filter cul- Secondary Products tures were dried o at room ternary- ature in a vacuum desiccator over 5 and A gradient HPLC method, providing additional confirmations of identifications and TABLE 1 Representative Chromatographic Data improved quantifications of the secondary products, used a Beckman C8 (4.6 x 250 TLC (Rfx 100) in solvents mm; 5-km) column with Solvent I (70% wa- NPLG ter/methanol + 1% ortho-phosphoric acid) Compound” Ab c k’l1 and Solvent II (methanol). The 20-min lin- DMGR 21 49 20 2.24/10@4 ear gradient from 35 to 10% Solvent I effi- DMSPH 3.5 59 32 4.47/1168 ciently separated all of the secondary prod- GR 37 63 50 6.21/1259 ucts (Fig. 1). Between runs, a 2-min gradient to 2% Solvent I (held for 5 min) cleaned a Chemical structures are in Fig. 2. b Norstictic acid and atrrmorin, standard TLC ccm- the column. Flow was 1.0 ml/min through- trols for lichen products, ran at 36 and 82 in Sobent A, out, and detection was at 270 nm. The sam- at 30 and 18 in Solvent B, and at 29 and 87 in Sd- ple was dissolved in sufficient methanol, vent C. 56 CULBERSON AND ARMALEO weighed. Subtracting the original weight of were then inverted onto a microscope slide the empty filter gave the biomass of the cul- with the specimen adhering to the under- ture after extraction. Since the methanol! side of the coverglass. The tape fastened water-soluble but acetone-insoluble residue the coverglass to the slide, allowing obser- (possibly carbohydrate) described above vations under oil, and formed a chamber constituted a sizable fraction (ll-13%) of between the two glass surfaces to accom- the biomass after Day 16, this weight (mea- modate the often clumpy specimen. A Leitz sured on a dried aliquot of each methanol- DiaIux 20 fluorescence microscope with a water extract) was added to the values de- camera attachment was used to photograph termined from the weights on the filters. (Kodak TriXpan 400 black-and-white film) The accuracy of the biomass estimate im- samples under long-wave uv light. proves as the number or the size of filters per time point increases. Using additional RESULTS filters for each time point also averages slight variations in conditions (e.g., dry- Identification of Secondary Products ness) between different plates. Further, the extra filters allowed removal of any that showed contamination during the long Three characteristic, biogenetically re- course of the experiments. lated lichen products were identified in induced cultures of the isolated C. grayi mycobiont, the depside DMSPH and two dep- Fluorescence Microscopy sidones DMGR and GR (Figs. 1 and 2). The Diluted samples from suspensions of natural lichen accumulates high amounts of fragmented mycelia were spread directly GR and small amounts of DMGR (C. F. onto LB plates and incubated at room tem- Culberson et al., 1985). The previously un- perature. (Fluorescence microscopy on fil- reported trace of DMSPH was demon- ter cultures was impeded by the high back- strated by preparative TLC enrichment of ground generated by the filter itself). Ob- the minor compounds followed by two- servations at less than 250~ enlargement dimensional TLC and HPLC confirmation. were made without a coverglass on colo- Time-course experiments described below nies on small agar sections removed asep- showed the order and speed of induction of tically from the plates and placed onto these secondary products. slides immediately before microscopy. Due to uv-radiation damage, the same colony could not be followed over time. When DAY 22 CR higher resolution was needed, a single colony was excised from the plate and freed of most agar under a dissecting microscope. It was then transferred onto a coverglass held under the dissecting microscope by two DMSPH strips of double-stick tape placed over the right and left edges of the glass. The colony was fragmented with forceps, and individ- 0 5 10 i5 20 is ual fragments were carefully dragged Retention Time (min) across the coverglass so as to stretch out FIG. 1. Sample HPLC (gradient method) of the ex- aerial hyphae . Manipulations were prompt tract from a 22-day culture of the C. gruyi fungus grown on a nylon filter. The chromatogram represents after removal of the colony from the plate, 1/4Oth of the total extract. For the identification and since drying out impeded the sliding of hy- quantitation of secondary products, see Materials and phae. The coverglass and attached tape Methods. STD is the internal standard. INDUCTION OF A LICHEN SECONDARY PATHWAY 7

CH3 CH3 OR Acetyl Pol malonyl + HO COOH + HO coo Pat 6 way

OH &HI 5

FIG. 2. A possibie biosynthetic pathway to GR (adapted from Ehx et nl., 1987). The two reactions summarized within the brackets are (1) an acyl migration and (2) a Smiles rearrangement. For further explanation, see Discussion.

Induction of Secondary Metabolism and reached a plateau. DMGR, always the Effect of Water Loss in small amounts, kept increasing (F Since induction followed tbe Mycelia of C. gruyi grown in liquid LB from liquid to solid medium, we tested the medium for weeks or months produced no effect of different degrees of hydration. F& significant amounts of DMSPH, DMGR, ter cultures were placed on regular and GR, Upon transfer from liquid to solid (“moist”) agar plates and on others tha LB medium, the cultures made secondary had been predried so that the thickness o products that were detectable by 1 week the medium had been reduced by about ha and increased rapidly by 3 weeks (Figs. 3A (“semi-dry”). During the first and 3B). In this and similar induction ex- (Figs. 3B and 3C), secondary-pro periments, the three identified components synthesis was induced much more r of the pathway showed distinct patterns of in the semi-dry cultures. To confa change. DMSPH accumulated most rapidly the observed effect was due to de in the earliest phases, initially often sur- water content and not to the higher conce passing GR (Fig. 3B). Later, however, tration of medium components in while GR continued to accumulate rapidly dry plates, the time course of early ~~d~c~ over a long period of time, DMSPH tion was analyzed in a separate ex~e~me~t

30 1.2 5 .z 25 1.0 A 20 0.8

3 15 0.6

a 10 0.4 a8 5 0.2

0 0.0 0 10 20 30 40 50 60 0 5 10 I5 20 Time (days) FIG. 3. Time course of secondary product synthesis in filter-cultures on regular (“moist”) and on partially dried (“semi-dry”) LB agar medium. Symbols represent the amounts of DMSPH (0), DMGR (U), and GR (1) per filter-culture. For each time-point, the combined extracts from three to five cultures on 13-mm falters were analyzed by HPLC (isocratic method). (A) Sixty days on “moist” L agar. (B) The first 20 days on “moist” LB agar. (C) Twenty days on “semi-dry” LB agar. The scale for nmol product/filter is the same in panels B and C. 58 CULBERSON AND ARMALEO

comparing filter cultures grown on identical ated-growth phase between Days 8 and 16 LB media containing either 1.5 or 3% aga- was correlated with the proliferation of rose. The concentrations of all components aerial hyphae emerging from an aggluti- were the same except for the higher agarose nated hyphal base resting on the filter. The and the consequently lower free water in growth curve measured for this mycobiont the 3% medium. By Day 20 the 3% cultures on the solid substrate resembles those of contained, per milligram biomass, 2.5 x the other lichen mycobionts measured in liquid amount of grayanic acid found in the 1.5% media (Ahmadjian, 1961; Lawrey, 1984). controls. Clearly, the level of secondary- Many nonlichen fungi display similar product induction in this isolate was in- growth curves but with vastly shortened versely correlated with the amount of avail- time scales. Most nonlichen secondary me- able water. tabolism is studied in liquid culture; mycophenolic acid production by Penicillium Relation of Biomass to the Timing and brevicompactum, however, has been ana- Rate of Biosynthesis lyzed in sufficient detail on solid media (Bartman et al., 1981) to provide an appro- Seeding mycelia on individually pre- priate comparison for our results. The ac- weighed filters allowed measurement of celerated-growth phase between Days 8 biomass for the samples analyzed chemi- and 16 in our fungus spanned Days 2 to 3 in cally. Even with the limited availability of the nonlichen fungus. The lichen fungus re- the slow-growing mycobiont, the small size sembled P. brevicompactum (and many of the samples needed for each microcul- other nonlichen fungi in liquid culture) by ture permitted replicates for each time showing a significant increase in secondary point. Growth, after an initial lag, acceler- products during the growth spurt, the rate ated and then slowed (Fig. 4). The acceler- of accumulation per unit of biomass peak- ing just after growth slowed, and then de- clining. Also the maximum secondary pro- Time (days) ductivity (0.7 nmol product/day/mg bio- 0 5 10 15 20 25 30 mass) of our slow-growing lichen fungus was comparable to that of the faster- growing P. brevicompactum.

Hyphal Sites and Pattern of Secondary-Product Accumulation

uv-Irradiated GR and DMSPH crystals emit a blue fluorescence that allows patterns of localization of these compounds on Time Intervals (days) hyphae to be followed microscopically dur- FIG. 4. Daily rate of secondary-product synthesis ing induction (Fig. 5). Mycelial fragments by the lichen fungus per milligram of biomass (bar were seeded at low density directly on agar graph) superimposed on the growth curve (line graph). The biomass at each time-point is the average of two or agarose plates, and colonies were ob- 2.5mm filter-cultures. The secondary products per fil- served periodically by fluorescence micros- ter-culture (averaged from two filters) were quantified copy. The earliest signs of accumulation by HPLC (gradient method). The time intervals used were small patches of blue fluorescence on for calculation of daily productivity are indicated be- the surface of aerial hyphae about 250 pm low the lower abscissa. Bars represent the nanomoles product per filter made during interval/number of days or more from the tip (Fig. 5A). During in- in interval/average biomass per filter in interval. Bar duction, the patches increased in size and markings: q DMSPH, q DMGR, q GR. intensity from the apical region (mostly de- INDUCTION OF A LICHEN SECONDARY PATHWAY 59

FIG. 5. Fluorescence microscopy of the cultured lichen fungus. Bars represent 100 km. (A) Aerial hypha showing new deposits of secondary product 260-350 p,rn from the tip. (B) Two weeks into the induction the apical region of an active aerial hypha shows no deposits (bracket). In the basal region, the more distal and smaller deposits appear banded (arrow), while larger ones predominate closer to the base emerging from the body of the colony (out of the plane of focus). Also in the surrounding hyphae, most deposits are near the base (out of the plane of focus). (C) Top overview of the same colony as in (B), showing the largest deposits at the sites of emergence of aerial hyphae. (D) After 4 weeks, the aerial mycelium is almost completely covered with secondary product. In (A) the hypha was stretched under a coverglass, and no coverglass was used in B-D (see Materials and Methods). void of patches) to the base of individual patches could not be seen at the lower aerial hyphae (Fig. 5B), a polarity consis- edges of the colony, where hyphae h tent with a model proposed by Moss (1984). etrated the medium. Often the sites of accumulation displayed When acetone or met characteristic banding patterns (Fig. 5B). through a colony under However, not all aerial hyphae were af- blue fluorescent patche fected during induction, many lacking quickly, leaving a uniform patches even when growing near active hy- surface autofluorescence ty phae. Many aerial hyphae both branched walls, as seen after removal of secondary and coalesced in bundles as growth pro- products or in uninduced mycelia. From gressed, and fluorescent patches became their fluorescence, kinetics of ac~~rn~~a- particularly conspicuous where such aerial tion, and solubihty in acetone or me bundles emerged from the central aggluti- we believe that the blue patches represent nated mass of the colony (Fig. SC). Even- extracellular accretions of secondary prod- tually, the entire aerial mycelium became ucts, mostly grayanic acid. covered with fluorescent accretions, oblit- erating most traces of discontinuity and of Effects of Other Variables yphal tip-to-base gradations (Fig. 5D). In general, the density of the accretions The results reported here for ~~~t~res on tended to be the highest in the aerial regions nylon filters over an agar medium also ap- around the top of the dome-shaped colony, ply to mycelia growing directly on agar. RX and decrease toward the agar surface. Blue experiments, the nylon-filter metho 60 CULBERSON AND ARMALEO more versatile and sensitive. In one exper- the biosynthesis of specific polyketide- iment using Durapore (Millipore Corp.) in- derived secondary metabolites was re- stead of nylon fdters, growth and produc- ported for the nonlichen fungi Penicillum tion of secondary compounds although ini- brevicompactum with mycophenolic acid tiated were not sustained. This is probably (Bartman et al., 1981) and P. pat&urn with due to the fact that hyphae cannot easily 6-methylsalicylic acid (Peace et al., 1981). penetrate the Durapore matrix (Oliver et The overlap of maximum productivity and al., 1989), whereas they can pass through the end of the aerial-growth spurt conforms the nylon mesh. The mycelia for all of our to the view that links secondary metabo- experiments were fragmented (Ahmadjian, lism to conditions of reduced growth 1987; Armaleo, 1991) when transferred to (Bu’Lock, 1975; Moss, 1984). However, solid media. However, differences in the the localization of compounds on the aerial degree and timing of the preplating frag- mycelium suggests that on solid media not mentation probably contributed to varia- only growth limitation but also specific bio- tions in the length of the initial lag phase chemical consequences of dryness and air observed in different experiments. Work in trigger secondary metabolism as a compo- progress using a variety of lichen myco- nent of aerial differentiation. bionts grown on media other than LB indi- In the few studies reporting characteris- cates that the choice of medium, as well as tic lichen depsides and depsidones from the genetic makeup of the fungus, can have isolated lichen mycobionts and two nonli- quantitative and qualitative effects on in- then fungi, the cultures were almost invari- duction. Given the proper match of medium ably grown on solid media (Komiya and and genotype, however, the results out- Shibata, 1969; Umezawa et al., 1983; Ejiri lined here apply to a diversity of lichen my- and Shibata, 1975; Yamamoto et al., 1976; cobionts . Hamada and Ueno, 1987; Leuckert et al., 1990). The secondary products characteris-

DISCUSSION tic of lichen fungi seem more stringently linked to aerial growth than are the com- Induction of Secondary Metabolism and pounds commonly associated with nonli- Aerial Growth on Solid Substrates then fungi and so extensively studied in liquid media. The parallels between our lichen We have found that a complete second- mycobiont and the two penicillia, however, ary-product pathway is consistently in- suggest an aspect of secondary metabolism duced in an isolated lichen mycobiont when shared by lichen and nonlichen fungi that is hyphal suspensions are transferred from a best studied on solid media. Conditions liquid to a solid medium. The three poly- promoting aerial growth are especially rel- ketide-derived compounds identified in the evant for lichen fungi where adaptations for cultures are all found in the natural lichen. life above a substrate, exposed for years to In the isolated mycobiont, accumulation light and air and without protection from begins after a lag of several days and is cor- water loss, are more likely to be expressed related with the growth of an aerial than in liquid culture. Indeed, the critical mycelium, where the end-product of the importance of drying for establishing and pathway (grayanic acid) is deposited. In- maintaining the lichen symbiosis has been duction is enhanced on progressively drier amply verified in artificial lichenizations. substrates, suggesting that decreased water is a critical activating factor (review: Jen- Patterns of Accumulation and Biogenesis nings, 1984), probably in conjunction with of Grayanic Acid exposure to air. The same correlation be- Grayanic acid is progressively and un- tween differentiation of aerial hyphae and evenly deposited in an amorphous matrix INDUCTION OF A LICHEN SECONDARY PATHWAY I on the surface of aerial hyphae. The discon- are activated first. With age, tinuities in the deposition are best observed quent steps gain efficiency and during the first 3 weeks of induction. Some most all the DMSPH to depsido hyphae appear to be inactive whereas oth- tually the proportions should mate ers accumulate secondary product in de- in the natural lichen, which contai fined but irregular areas, as if the enzymes traces of DMSPH and large amo responsible for grayanic acid synthesis and/ or translocation were fixed in localized Factors Affecting Froductivi~ patches, perhaps within the cell wall. The and Detection fluorescent accretions do not reflect any obvious underlying cell structure, but in- Besides growth on solid substrates, o crease in size and intensity toward the hy- factors contribute to the production an tection of secondary products in cuitu phal base, perhaps a consequence of the lichen fungi. As the mycobionts gen larger build-up of the pathway’s machinery do not conidiate in culture, suspe~siQ~s of or of decreased NADPH (Moss, 1984) in lightly fragmented mycelia are sown on fil- the older vs. the younger regions of the hy- ter or agar surfaces to provide large nu pha. Eventually, as the metabolic pathway bers of synchronous growth initials amph- accelerates and aerial growth slows, the en- fying aerial development and secondary tire aerial mycelium becomes covered with productivity. While fragmentation does not secondary product. The localization of induce the characteristic componnd§ i grayanic acid suggests that the aerial hy- submerged culture, mycobionts transfer-r-e phae in the culture are equivalent to the to solid media without fragmentation re- medullary hyphae inside the natural lichen main agglutinated and make only small thallus where most of the characteristic amounts of secondary products. C~~pl~~~ secondary products accumulate. the manipulation of small amounts of myce- The joint occurrence of the depside ha on filters with current microana~ytica~ DMSPH with DMGR and GR in the myco- methods (TLC, HPLC, and ~~oresce~c biont cultures and in the natural lichen is microscop:y) provides the versatility an consistent with two theories that see dep- sensitivity needed for experiments with the sides giving rise to depsidones either by ox- slow-growing lichen fungi. Most workers idative coupling (Erdtman and Wachtmeis- have used various microchemical analyses ter, 1957) or by a route illustrated in Fig. 2 for the biosynthesis of GR (Elix et al., on extracts of milligram (or less) of the fungus, and only few have 1987). According to the latter proposal, to extract and purify characteristrc h&en acetyl-polymalonyl-derived phenolic acids products (IKomiya and Shibata, 1969; Ejiri (here orsellinic acid and sphaerophorol car- and Shibata, 1975; Hamada and Ueno, boxylic acid) are esterified to form a dep- 1987). The earlier methods (e. side (DMSPH). 5’-Hydroxlation followed chromatography and microcry by acyl migration and a Smiles rearrange- were not sufficiently discriminating or sen- ment yields a diphenyl ether, which cy- sitive for such small cultures. The compo- clizes to a depsidone (DMGR) by intramo- sition of the solid medium and the lecular esterification. O-Methylation of variability of spores from natural DMGR would give the depsidone GR char- also affect the production of set acteristic of the natural lichen. During compounds in cultured myco induction, the proportion of depside (DMSPH) to depsidones (DMGR and GR) The Roles of the Symbionts in is initially high and decreases with time, a Secondary metabolism pattern consistent with the above propos- als. The pathway steps leading to DMSPH The ability of our culture to c~rn~~e~e t 62 CULBERSON AND ARMALEO pathway to grayanic acid indicates that the tween natural and synthetic thalli of Usnea strigosa. biosynthetic capacity for this depsidone lies Lichenologist II: 149-165. ARMALEO, D. 1991. Experimental microbiology of li- entirely within the mycobiont. Work in chens: Mycelial fragmentation, a novel growth progress in our laboratory shows that other chamber, and the origins of thallus differentiation. solid-media cultures of related and unre- Symbiosis 11: 163-178. lated mycobionts produce a variety of dep- BARTMAN, C. D., DOERFLER, D. L., BIRD, B. A., sides and depsidones. These results indi- REMALEY, A. T., PEACE, J. N., AND CAMPBELL, cate that the alga is not required for catal- I. M. 1981. Mycophenolic acid production by Peni- cillium brevicompactum on solid media. Appl. En- ysis at any step, and confirm the reports vir. Microbial. 41: 72%736. discussed earlier in which occasional pro- BUBRICK, P. 1988. Methods for cultivating lichens and ductivity was described in isolated myco- isolated bionts. In Handbook of Lichenology (M. bionts but could not be tied to any specific Galun, Ed.), Vol. 3, pp. 127-138. Academic Press, conditions of culture. Regulatory effects of London. BU’LOCK, J. D. 1975. Secondary metabolism in fungi the alga on secondary metabolism might and its relationship to growth and development. In still be found, and may perhaps contribute The Filamentous Fungi (J. E. Smith and D. R. to the synthesis of compounds seen in arti- Berry, Eds.), Vol. 1, pp. 33-58. Edward Arnold, ficially lichenized or cocultured symbionts. London. However, the use in those experiments of CULBERSON, C. F., AND AHMADJIAN, V. 1980. Artiti- cial reestablishment of lichens. II. Secondary prod- dispersed inocula deposited onto solid sub- ucts of resynthesized Cladonia cristatella and Le- strates (e.g., mica, soil, or agar) seems a canora chrysoleuca. Mycologia 72: 90-109. more likely explanation of the observed CULBERSON, C. F., CULBERSON, W. L., AND productivity. JOHNSON, A. 1977. Second Supplement to “Chem- ical and Botanical Guide to Lichen Products.” American Bryological and Lichenological Society. ACKNOWLEDGMENTS St. Louis, MO. CULBERSON, C. F., CULBERSON, W. L., AND We thank John A. Elix for a sample of 4-0- JOHNSON, A. 1985. Orcinol-type depsides and dep- demethylsphaerophorin, Stephen Johnston and Bruce sidones in the lichens of the Cladonia chlorophaea R. Nicklas for the use of fluorescence microscopes, group (Ascomycotina, Cladoniaceae). Bryologist 88: Anita Johnson and Gerda Vergara for technical assis- 380-387. tance, and William Louis Culberson for helpful discus- CULBERSON, C. F., AND ELIX, J. A. 1990. Lichen sub- sions. This study was supported in part by Grants stances. In Methods in Plant Biochemistry (J. B. BSR-8806659 from the National Science Foundation Harborne, Ed.), Vol. 1, pp. 509-535. Academic and ARIG- from the North Carolina Biotech- Press, London. nology Center. CULBERSON, W. L. 1986. Chemistry and sibling spe- ciation in the lichen-forming fungi: Ecological and REFERENCES biological considerations. Bryologist 89: 123-13 1. CULBERSON, W. L., CULBERSON, C. F., AND AHMADJIAN, V. 1961. Studies on lichenized fungi. JOHNSON, A. 1982. A new endemic red-fruited Cla- Bryologist 64: 168-179. donia from the North Carolina coast. Mycologia 74: AHMADJIAN, V. 1967. The Lichen Symbiosis. Blais- 662-667. dell, Waltham, MA. DREW, S. W., AND DEMAIN, A. L. 1977. Effect of pri- AHMADJIAN, W. 1980. Separation and artificial syn- mary metabolites on secondary metabolism. In An- thesis of lichens. In Cellular Interactions in Symbi- nual Review of Microbiology @I. P. Starr, J. L. In- osis and Parasitism (C. B. Cook, P. W. Pappas, and graham, and A. Balows, Eds.), Vol. 31, pp. 343-356. E. D. Rudolph, Eds.), Ohio State University Bio- Annual Reviews Inc., Palo Alto, CA. sciences Colloquia, No. 5, pp. 3-29. EJIRI, H., AND SHIBATA, S. 1975. Squamatic acid from AHMADJIAN, V. 1987. Laboratory culture of lichens the mycobiont of Cladonia crispata (Ach.) Flot. and lichen symbionts. In Proceedings of Symposium Phytochemistry 14: 2505. on Tissue Culture of Lichen and Bryophyte, pp. ELIX, J. A., JENIE, U. A., AND PARKER, J. L. 1987. A l-10. Nippon Paint Co., 19-17 Ikedanaka-machi, novel synthesis of the lichen depsidones divaronic Neyagawa, 572, Japan. acid and stenosporonic acid, and the biosynthetic AHMADJIAN, V., AND JACOBS, J. B. 1985. Artiticial implications. Aust. J. Chem. 40: 1451-1466. re-establishment of lichens. IV. Comparison be- ELIX, J. A., AND WARDLAW, J. H. 1987. Synthesis of INDUCTION OF A LICHEN SECONDARY PATHWAY 3

lichenpara-sphaerophorol depsides. Aust. J. Chem. expression of genes involved in the biosynthesis of 40: 425-429. antibiotics and other secondary metabolites. In An- ELIX, J. A., WHITTON, A. A., ANI) SARGENT, M. V. nual Review of Microbiology (L. N. Ornston,A. 1984. Recent progress in the chemistry of lichen Balows, and E. P. Greenberg, Eds.), Vol. 43, pp substances. Progr. Chem. Org. Nat. Prod. 45: 103- 173-206. Annual Reviews, Inc., Palo Alto, CA. 234. Moss, M. 0. 1984. The mycelial habit and secondary ERI)TMAN, H., AND WACHTMEISTER, C. A. 1957. Phe- metabolite production. In The Ecology and Physioi- noldehydrogenation as a biosynthetic reaction. In ogy of the Fungal Mycelium. Symposium of the &it- Festschrift Arthur Stall, pp. 144-165. Sandox Ltd., ish Mycological Society. (D. I-I. Jennings and Basel. A. 14. M. Rayner, Eds.), pp. 127-142. Cambridge Fox, C. H., AND HUNECK, S. 1969. The formation of Univ. Press, Cambridge. roccellic acid, eugenitol, eugenetin, and rupicolon MULVANY, J. G. 1969. Membrane filter ~~c~ni~ues in by the mycobiont Lecanora rupicola. Phytochemis- microbiology. In Methods in Mjcrobio~ogy (J. R. try 8: 1301-1304. Norris and D. W. Ribbons, Eds.), Vol. 1, pp. 205- GARRAWAY, M. O., AND EVANS, R. C. 1984. Fungal 253. Academic Press, London. Nutrition and Physiology. Wiley, New York. OLIVER, E., ‘CRITTENDEN, P. D., BECKETT, A., AND HAMADA, N. 1988. Depside from isolated lichen my- BROWN, D. H. 1989. Growth of ~c~en-fo~ingf~~~~ cobiont. II. Lichenologist 20: 294-295. on membrane filters. Lichenologist 21: 387-392. HAMADA, N., AND UENO, T. 1987. Depside from an PEACE, J. N., BARTMAN, C. D., DOERFLER, D. L., isolated lichen mycobiont. Agric. Biol. Chem. 51: AND CAMPBELL, I. M. 1981. 6-Methyls~i~y~~c acid 1705-1706. production in solid cultures of Penicillium pat&m JENNINGS, D. H. 1984. Water flow through mycelia. occurs only when an aerial mycelium is present. In The Ecology and Physiology of the Fungal Appl. Environ. Microbial. 41: 1407-1412. Mycelium. Symposium of the British Mycological SANKAWA, U., SH~~IJYA, M., EBIZUKA, Y., Noou- Society. (D. H. Jennings and A. D. M. Rayner, Eds.), pp. 143-164. Cambridge Univ. Press, Cam- CHI, H., KTNOSHITA, T., AND IITAKA, Y. 1982. Dep- side as potent inhibitor of prostaglandin biosynthe- bridge ~ sis: A new active site model for fatty acid cycloox- KOMIYA, T., AND SHIBATA, S. 1969. Formation of lichen substances by mycobionts of lichens. Isolation ygenase. Prostaglandins 24: 21-34. of (+)-usnic acid and salazinic acid from myco- UMEZAWA, K., MURAMATSU, S., ISHIZUKA, bionts of Ramalina spp. Chem. Pharm. Bull. 11: SAWA, T., TAKEUC~I, T., AND MATSUSNIMA, 1305-1306. 1983. Inhibition of histidine decarboxylase and tu- KON, Y., KASHIWADANI, H., AND KUROKAWA, S. mour promoter-induced arachidonic acid relea 1990. Induction of lichen thalli of Usnea confusa lecanoric acid analogues. Biochem. ~iQ~~ys” Asah. ssp. kitamiensis (Asah.) Asah. in vitro. J. Ja- Commun. 110: 733-739. pan. Bot. 65: 24-33. YAMAMOTO, Y., MITZUGUCHI, R., AND YAMADA, Y. LAWREY, J. D. 1984. Biology of Lichenized Fungi. 1985. Tissue culture of Usnea rubescens and Rama- Praeger, New York. ha yasudae and production of usnic acid in their LEUCKERT, C., AHMADJIAN, V., CULBERSON, C. F., cultures. Agric. Biol. Chem. 49: 3347-3348. AND JOHNSON, A. 1990. Xanthones and depsidones YAMAMOTO, Y., NISHIMURA, K., AND KI~YA~A~ N. of the lichen Lecanora dispersa in nature and of its 1976. Studies on the metabolic products of Aspergil- mycobiont in culture. Mycologia 82: 370-378. fus terreus. I. Metabolites of the strain IF0 Ghena. MAarfN, .I. F., AND LIRAS, P. 1989. Organization and Pharm. Bull. 24: 1853-1859.