A new para-tridepside, named trivaric acid, composed of three units of divaric acid is the major product of a new chemotype in the Ramalina americana complex. It was detected in a single thallus, and in spite of wide surveys of R. americana and related species, no additional material of this chemistry has been found. The structure of trivaric acid is based on TLC and HPLC correlations, microhydrolysis, and micromethanolysis. Some lichens of uniform or similar morphology show a high level of chemical differentiation, and the chemotypes may have distinctive geographic distributions and different ecologies where they are sympatric. The best understood chemotype complex of this sort is the Ramalina siliquosa group of Europe, which has been the subject of numerous studies. In North America, however, the R. americana complex exhibits an even greater range of chemical variation, actually exceeding that of many entire lichen genera (Culberson et al. 1990; Dey 1978; Hale 1978). As part of a recent molecular phylogenetic study of the R. americana chemotype complex (LaGreca 1997), chemical analyses of 628 herbarium specimens and new collections from throughout the range of the group found several new chemotypes, bringing the total to 17, including one with a major product not matching any known lichen substance. Acid hydrolysis suggested that the new substance is a tridepside composed of three divaric acid units. Although acid hydrolysis is the method most widely used to assist the microchemical identification of new lichen products, many depside linkages are also readily cleaved by alcohols. Indeed, some methyl and ethyl esters reported as natural lichen products are known to be artifacts from prolonged extraction with chloroform stabilized with methanol or ethyl ether containing small amounts of ethanol or from the use of these alcohols for recrystallizations. For example, Yoshimura and Kurokawa (1991) showed that methanol solutions of extracts of five lichens containing lecanoric and gyrophoric acids gave additional HPLC peaks identified as orsellinic acid and methyl orsellinate. Some of the orsellinic acid occurred naturally in these species, but all of the methyl orsellinate came from methanolysis. In the present study, we examine the methanolysis of lecanoric and gyrophoric acids and use the results to propose a chemical structure for the new lichen product from the rare new R. americana chemotype. The single thallus (Fig. 1) of this new chemotype is too small to isolate its major secondary product and verify the chemical structure spectroscopically, but it is sufficient to provide a fragment for future chromatographic comparison if the structure proposed here is synthesized. MATERIALS AND METHODS The specimen studied (U.S.A. LOUISIANA. St. Martin Parish, St. Martinville, on fence pickets, Langlois, 26 June 1895 (us)) was originally placed near R. complanata (Sw.) Ach., but it lacks the distinctive papillae of that species and is within the range of morphological variation currently defining R. americana Hale. Thallus fragments were extracted with toluene at room temperature (3 X, totaling 20 min.) and then with warm acetone (3 X, totaling 20 min.). Residues were analyzed by TLC with three standardized solvent systems (A, B, and C) (Culberson et al. 1981; Culberson & Johnson 1982) and by HPLC. HPLC used a Beckman ODS column (4.6 X 250 mm; 5jm) and one of two gradients (G80 and G60) formed by mixing methanol:water:o-phosphoric acid (30:70:1) (Solvent 1) with methanol. The 40min. gradient of G80 was from 80% to 16% Solvent 1, holding at 16% for 15 min. The 40-min. gradient of G60 was from 60% to 12% Solvent 1, holding at 12% for 15 min. For both gradients, the column was flushed at 2% Solvent 1 and re-equilibrated at the initial conditions for 10 min. The flow rate was 1.0 ml/min., and detection was at 270 nm. Microhydrolyses used concentrated H2SO4, and, unless stated otherwise, all identifications were based on both TLC and HPLC comparisons with authentic samples. Micromethanolysis used the general method described below. Before investigating the scarce prodSCurrent address: Farlow Herbarium, Harvard University, 20 Divinity Avenue, Cambridge, MA 02138, U.S.A. 0007-2745/99/595-601$0.85/0 This content downloaded from 157.55.39.17 on Fri, 02 Sep 2016 05:12:52 UTC All use subject to http://about.jstor.org/terms 596 THE BRYOLOGIST [VOL. 102 ii' 1IIIIIIII1) 111111111 1111111 1 FIGURE 1. The entire thallus (119 mg) of the single specimen of Ramalina americana producing trivaric acid. Scale in millimeters. uct from the Ramalina, the mode of cleavage was studied with the known compounds lecanoric and gyrophoric acids, their methanolysis products, and a lichen extract containing a mixture of lecanoric and gyrophoric acids. General method for micromethanolysis.-An acetone solution of each compound or extract was prepared to give a major on-scale peak with an injection volume of 2-8 1il. A measured volume was transferred to an microfuge tube and evaporated to dryness. All evaporations were with moderate heat (400C) under reduced pressure in a stream of N2. These conditions were terminated as soon as the solvent disappeared to minimize decarboxylation and subsequent loss of the more volatile decarboxylated phenolic units. Nevertheless, some loss of orcinol and divarinol was expected. Test compounds and extracts were redissolved in methanol and allowed to stand in the capped tube at 440C. To avoid error due to slow evaporation of solvent, after each reaction interval methanol was evaporated and residues were redissolved in freshly measured volumes, adjusted downward to account for previous injections and to maintain directly comparable chromatograms with equal injection volumes. HPLC used a 40-min. gradient (G80 or G60) optimized to give a spread of peaks with retention times (Rt) of 4-40 min. for the reactants and expected products. Tests of the methanolysis of gyrophoric acid, lecanoric acid, orsellinic acid, orcinol, methyl lecanorate, and methyl orsellinate were semiquantative, using weighed samples. After the final time interval (4-10 days), while most products and starting materials were still well represented, samples were also analyzed by TLC. TLC and HPLC data are in Tables 1-2 and chemical structures in Fig. 2-7. RESULTS AND DISCUSSION The single small thallus (Fig. 1) of the new chemotype of R. americana was collected in St. Marti ville, Louisiana, a century ago by the French priest-botanist Auguste Barthdlemy Langlois (1832-1900). One of us made an extensive field search in the environs of that locality, paying particular attention to the unusual substrate for the origi al collection (wooden fences), but no new material could be found. The medulla reacted C+, and TLC showed one major product with typical orcinol-type depside coloration with 10% H2SO4 and heat (30 min. at 110oC). The pattern of R, classes (A3B5C4) in the three standard solvent systems was typical of a depside or tridepside lacking Omethylation, but the compound did not correspond to any known lichen products available for TLC and HPLC comparisons. Acid hydrolysis yielded one major product that was identified as divaric acid by TLC and HPLC comparisons with an authentic sample from the hydrolysis of divaricatic acid. The new compound was not identical to nordivaricatic acid, the depside composed of two divaric acid units, but behaved by TLC and HPLC as TABLE 1. Representative TLC data. Values following the virgule (/) are for the internal controls, norstictic acid and atranorin.