Chlorophyll Catabolites in Fall Leaves of the Wych Elm Tree Present a Novel Glycosylation Motif

Abstract Fall leaves of the common wych elm tree (Ulmus glabra) were studied with respect to chlorophyll catabolites. Over a dozen colorless, non‐fluorescent chlorophyll catabolites (NCCs) and several yellow chlorophyll catabolites (YCCs) were identified tentatively. Three NCC fractions were isolated and their structures were characterized by spectroscopic means. Two of these, Ug‐NCC‐27 and Ug‐NCC‐43, carried a glucopyranosyl appendage. Ug‐NCC‐53, the least polar of these NCCs, was identified as the formal product of an intramolecular esterification of the propionate and primary glucopyranosyl hydroxyl groups of Ug‐NCC‐43. Thus, the glucopyranose moiety and three of the pyrrole units of Ug‐NCC‐53 span a 20‐membered ring, installing a bicyclo[17.3.1]glycoside moiety. This structural motif is unprecedented in heterocyclic natural products, according to a thorough literature search. The remarkable, three‐dimensional bicyclo[17.3.1]glycoside architecture reduces the flexibility of the linear tetrapyrrole. This feature of Ug‐NCC‐53 is intriguing, considering the diverse biological effects of known bicyclo[n.3.1]glycosidic natural products.


Quantification of Chl in green and senescent Elm tree leaves by UV/Vis Spectroscopy.
Chl a and b in green leaves. A total area of 9 cm 2 was cut out of a green Ulmus glabra leaf. The leaf was frozen in liquid N 2 , pulverized in a mortar, and extracted with MeOH. The slurry was filtered through a sintered glass filter, and the residue was ground in a mortar and extracted with MeOH. The procedure was repeated until the residue was colorless. The extracts were combined and diluted with MeOH to 100.00 ml in a volumetric flask. After filtration using a Sartorius filter the extracts were analyzed by UV/Vis spectrometry. In green Ulmus glabra leaves 51.1  3.7 µg cm -2 (56.9  4.2 nmol cm -2 ) of chlorophyll a and b were found (n=4), the data analysis was based on ref. [5] .
Chl a and b in senescent leaves: A total area of 141 cm 2 (second and third measurement: 81 cm 2 ) was cut out of several senescent Ulmus glabra leaves. The extraction and the UV/Vis analysis Scherl-etal-Elmtree-SuppInfo 13/04/2016 S6 were performed as described above. Yellow Ulmus glabra leaves were found to contain 0.5  0.3 µg cm -2 (0.5  0.1 nmol cm -2 ) of residual chlorophyll a and b (n=3).

Quantification of nonfluorescent Chl-catabolites (NCCs) in senescent leaves by anal. HPLC.
NCCs in senescent (yellow) leaves: The quantification of the Ug-NCCs was accomplished by anal. HPLC. A sample of Cj-NCC-1 [6] from Cercidiphyllum japonicum was used to prepare a standard soln. (calculated with ε 320 = 17,000 as described). Senescent Ulmus glabra leaves were found to contain 16 Collection, isolation and structure elucidation of NCCs.

Identification of Ug-NCC-43 (3) and Tc-NCC-2 (from Lime tree) using HPLC.A yellow Wych
Elm tree leaf and (in a second experiment) a yellow Lime tree leaf were ground in a mortar with       Figure S15). In order to include macrocyclic rings of the desired size in the queries, the repeating group/atom feature was used in both SciFinder and Reaxys, with A = any atom (except for H-atoms) and any bond type (dotted lines).
Any substitution as well as annulation of further rings were permitted anywhere in the query structure. . Substructure queries (screen dumps) used for the different macrocycle-bridged pyranoses in CAS SciFinder (left) and Elsevier Reaxys (right). As a first restriction to these macrocyclic lactones, a carbonyl group was added at the higher-numbered attachment point of the pyranose ring. Further restrictions to compounds possessing at least one N atom were achieved by the "Analyze by: Elements" feature in SciFinder, and with "Filter by Substructure" in Reayxs using a single N-atom as query. The final structural restriction was done by using a ring with 4 C and 1 N-atom and bonds of any type in both database systems. These searches gave the results shown in Table S3.  Table S3) -26 were found in both database systems. One of the five compounds retrieved exclusively in SciFinder was a false hit, a two-component system with the pyrrole ring in one, and the bridged pyranose in the other component (this compound was consequently not retrieved in Reaxys, and could not have been retrieved anyway as the corresponding publication is not present there). The other four compounds exclusively in SciFinder were missed by Reayxs because they had not been indexed from the two respective publications present also in Reaxys.
Regarding the six compounds retrieved in Reaxys but not in SciFinder, two had not been indexed by CAS (Chemical Abstracts Service) for a publication otherwise covered; the other four compounds did only appear as products of reactions which had been indexed for the two respective publications in Reaxys, but not by CAS. -A total of 14 references were retrieved for these compounds (8 for the 1,5-bridged macrocycles), 10 common to both sources, and two each exclusively in SciFinder and Reayxs, respectively. Of the two references retrieved exclusively in SciFinder but missed in Reaxys, one was missing in the latter database, for the other reference present, the relevant compound had not been indexed in Reaxys. The two references retrieved exclusively in Reaxys were both present in SciFinder, but had the respective relevant compounds not indexed, and were thus not retrieved. These differences in our search results, as many other similar examples, exemplify the necessity to search BOTH large compound databases.
The only "natural product" retrieved (cf. Table S3) among macrocyclic lactones containing at least one nitrogen was found to be actually only a synthetic derivative of the nitrogen-free natural product Aleurinin [13]. Such a restriction to natural products had to be done quite differently in the two database systems involved: in Reayxs, this information is placed in a specific data field "Isolation from Natural Product", and "Filter by: Natural Product" allows a simple and precise restriction to such compounds. In the literature database CAPLUS, natural products are in principle indexed with the substance role [14] "natural product occurrence"; this precise role, however, is not directly searchable in SciFinder (in contrast to another user interface to the same CAS databases, STN Messenger [15], not available to us). One can search only for the more general substance role "Occurrence" to determine the numbers given in Table S3; as this broader role is also used for occurrence in the environment (e.g., pollutants), or in analytical samples, such results may give too high a number.