Breakdown of Chlorophyll in Higher Plants—Phyllobilins as Abundant, Yet Hardly Visible Signs of Ripening, Senescence, and Cell Death

Abstract Fall colors have always been fascinating and are still a remarkably puzzling phenomenon associated with the breakdown of chlorophyll (Chl) in leaves. As discovered in recent years, nongreen bilin‐type Chl catabolites are generated, which are known as the phyllobilins. Collaborative chemical‐biological efforts have led to the elucidation of the key Chl‐breakdown processes in senescent leaves and in ripening fruit. Colorless and largely photoinactive phyllobilins are rapidly produced from Chl, apparently primarily as part of a detoxification program. However, fluorescent Chl catabolites accumulate in some senescent leaves and in peels of ripe bananas and induce a striking blue glow. The structural features, chemical properties, and abundance of the phyllobilins in the biosphere suggest biological roles, which still remain to be elucidated.


Introduction
Ther ejuvenating appearance of chlorophyll (Chl) in spring and the seemingly pompous disappearance of the green plant pigments in the autumnal foliage of deciduous trees and bushes are very colorful natural phenomena. The seasonal breakdown of Chl, in particular, has always been enchanting,and also most puzzling. Indeed, Chl metabolism is probably the most visual sign of life on Earth, even observable from outer space. [1] It has been estimated that more than 1000 million tons of the green plant pigment are biosynthesized and degraded every year on Earth. [2] Once,i tw as believed that colored products would result from Chl breakdown, similar to the bilins from the breakdown of heme, [3] or to photo-oxygenolysis products of Chl. [4] Thus,a ll the early searches for Chl catabolites concentrated on the detection of such hypothetical colored leftovers of green Chl. [5] As we now understand better, all of these endeavors were futile, [1a, 6] and Chl seemed to disappear without leaving at race. [2] Colored Chl catabolites were,i ndeed, not detected in higher plants, until very recently. [7] Only in the last 25 years [8] has Chl breakdown in plants begun to reveal some of its molecular and cellular mysteries. [1a,9] An original breakthrough was achieved by the unambiguous identification and structure elucidation of colorless Chl catabolites from vascular plants, [1a, 8, 10] thereby paving the way for fundamental insights into Chl breakdown. [9a,c, 11] In distantly related studies,K ishi, Shimomura, and co-workers identified Chl-derived tetrapyrroles as luminescent compounds from marine photoorganisms. [12] However,a sb ecame apparent in the early 1990s,t hese linear tetrapyrroles [12] differed basically from the Chl catabolites from higher plants. [8a,10] 1.1. ABreakthrough:I dentification of aFirst Colorless Chl Catabolite from Higher Plants By comparison of the pigment patterns of natural, wildtype senescent leaves of the grass Festuca pratensis and of barley (Hordeum vulgare)w ith the ones of corresponding "stay-green" mutants (that do not degreen), Matile,T homas et al. were able to identify several colorless compounds that accumulated in senescent wild-type leaves,but were absent in the mutants. [5,13] Some of these colorless compounds were suspected to be products of Chl breakdown. Several of them decomposed readily into rust-colored compounds and were called "rusty pigments". [14] "Rusty pigment 14" (1)w as am ajor colorless fraction in extracts of senescent primary leaves of barley,and suspected Fall colors have always been fascinating and are still aremarkably puzzling phenomenon associated with the breakdown of chlorophyll (Chl) in leaves.Asdiscovered in recent years,nongreen bilin-type Chl catabolites are generated, whicha re known as the phyllobilins. Collaborative chemical-biological efforts have led to the elucidation of the key Chl-breakdown processes in senescent leaves and in ripening fruit. Colorless and largely photoinactive phyllobilins are rapidly produced from Chl, apparently primarily as part of adetoxification program. However,f luorescent Chl catabolites accumulate in some senescent leaves and in peels of ripe bananas and induce astriking blue glow.T he structural features,c hemical properties,a nd abundance of the phyllobilins in the biosphere suggest biological roles,whichs till remain to be elucidated. to be aChl catabolite. [14,15] 14 C-Labeled d-aminolevulinic acid, the biosynthetic precursor of the natural porphyrinoids,w as incorporated into the "rusty pigment 14" fraction, thereby providing further support for its presumed role as aC hl catabolite. [15] Indeed, 1 could be identified as aC hl-derived linear tetrapyrrole [8a] by ac ombination of mass spectrometry [16] as well as UV/Vis,CD, and NMR spectroscopy. [17] It was characterized as an optically active metal-free,c olorless,a nd nonfluorescent linear tetrapyrrole with unconjugated pyrrole units.T he linear tetrapyrrole 1 featured aC hl-diagnostic cyclopentanone unit, annealed to the a-a nd b-positions of ap yrrole ring, and carrying am ethoxycarbonyl group.T he colorless and nonfluorescent bilin-type tetrapyrrole 1 was, thus,i dentified as the first nongreen Chl catabolite from higher plants. [1a, 8a, 9a] Thep olar linear tetrapyrrole 1 from barley (H. vulgare) was classified as anonfluorescent Chl catabolite (NCC), and provisionally named Hv-NCC-1. [9a, 18] Its detailed structure analysis established 1 as a3 2 ,18 1 ,18 2 -trihydroxy-16,19-dihydro-1-formyl-19-oxophyllobilane (Figures 1a nd 2), [8,19] named on the basis of as emisystematic,s tructure-based nomenclature,a ccording to which linear tetrapyrrolic Chl catabolites are phyllobilin derivatives. [9c, 10] The" phyllobilin" terminus refers to the basic bilin-type structures of such Chl catabolites,a nd to chlorophyll as their origin. Then amegiving compound is the phyllobilane (I,F igure 2). In analogy to the nomenclature of linear tetrapyrroles and bile pigments, [3] NCCs with three saturated meso positions are,hence, classified as 16,19-dihydro-1-formyl-19-oxophyllobilanes, such as,f or example,t he 16,19-dihydro-1-formyl-19-oxo-16epi-phyllobilane (epi-5), identified in senescent leaves of the Katsura tree (Cercidiphyllum japonicum;s ee Table 1i n Section 4.1). [20] Structure elucidation of 1-formyl-19-oxophyllobilane 1 provided af irst firm central focal point for considerations on the enigmatic pathway of Chl breakdown during leaf senescence: [8a] Thechemical constitution of 1 indicated loss of the central magnesium ion and of the phytyl group during breakdown of Chl. Thes tructure of NCC 1 also implied an oxygenolytic opening of the porphyrinoid macrocycle at the northern meso position, thereby revealing aregioselectivity of Chl breakdown that was completely unexpected on the basis of model reactions with Chl derivatives. [4b] It was reminiscent, in as triking way,o ft he breakdown of heme, [3] in which the macrocycle of heme is also opened at the analogous a-meso position. However,i nc ontrast to oxidation of the methine unit (of Chl) to the formyl group that is characteristic of NCC 1,typical heme catabolites are 1,19-dioxobilins,which lack the carbon atom at the site of the cleavage of the macrocycle (which is removed as carbon monoxide by ar eaction catalyzed by the heme oxygenase). [21] As 1 carried amethyl group at its 2-position, it also appeared to be more closely related to Chl a than to Chl b. [8a,22] However,c ompared to Chl a,a dditional polar functional units were attached at the periphery of 1,thus rendering it rather soluble in water. Furthermore,the bilane-like 1 was indicated to exist in av ariety of stable conformations with respect to its three saturated meso positions.

Nonfluorescent Chl Catabolites Accumulate in Senescent Leaves
Thei nitially identified Hv-NCC-1 (1)w as obtained from leaves that degreened upon storage in the dark. Theu se of this conventional method of artificially inducing leaf senescence raised the question of the more general validity of the surprising structure of the NCC 1.Fortunately, Bn-NCC-1 (2), Bn-NCC-2 (3), and Bn-NCC-3 (4;F igure 3), related polar NCCs in naturally senescent leaves of oilseed rape (Brassica napus), were soon identified, thereby substantiating the role of NCCs as colorless products of natural Chl breakdown. [18a,23] Bernhard Kräutler studied chemistry at the ETH in Zürich, where he received his PhD in 1976 working with Prof. Albert Eschenmoser.A fter postdoctorals tudies with Prof. Allen J. Bard (University of Texas, Austin) and Prof. Nicholas J. Turro (Columbia University,N ew York), he returned to the ETH to start his own research group. In 1991 he became Full Professor of Organic Chemistry at the University of Innsbruck, and since October 2015 has been Professor Emeritus. His research interestsi nclude chlorophyll breakdown, the chemical biology of vitamin B 12 ,aswell as functionalized fullerenes and porphyrinoids.   [10] the name-giving structure, depicted in two representative formulas to highlight its pseudocyclic (left) and extended conformations (right).

Chlorophyll Breakdown-a Cellular Three-Compartment Pathway
Matile et al. found evidence for the localization of NCCs in the vacuoles. [13c, 24] Thus,catabolites of Chl, which originate in the chloroplasts,would need to pass through the cytosol to gain access to the vacuoles. [9c,24] Theb reakdown of Chl was, therefore,p roposed to involve metabolic processes in the three main compartments of leaf cells,i ncluding (unidirectional) transport between them. [1a, 9a] Subsequent research indicated an original fluorescent Chl catabolite (6,anFCC or 1-formyl-19-oxophyllobilene-b) [25] as ac olorless product of Chl breakdown in the chloroplasts.F CCs (similar to 6)w ere deduced to be exported into the cytosol, where further modified FCCs (mFCCs) would be generated. [9c] Subsequently,t ypical mFCCs would be transported into the vacuoles to undergo rapid isomerization to the corresponding polar NCCs,l ose their characteristic formyl group to furnish dioxobilin-type fluorescent Chl catabolites (DFCCs, [26] the precursors of corresponding dioxobilin-type nonfluorescent Chl catabolites,D NCCs), [26,27] or become persistent hypermodified FCCs (hmFCCs) that accumulate in leaves and fruit (and give the ripe bananas their blue glow; Figure 4). [6b, 28] Thus,aq uarter of ac entury of collaborative chemical and biological research on Chl breakdown has not only revealed an unexpected structural variety of different natural Chl catabolites,but also key enzymes involved in the formation of the catabolites (see Sections 3a nd 4). [1a,8a, 9a,c, 10] 3. Common Early Part of the PaO/Phyllobilin Pathway

From Chlorophylls aand bt oPheophorbide a
Thea pparent close structural relationship of NCCs to Chl a raised the question of the whereabouts of the remains from the breakdown of Chl b.T he observed specificity of the ring cleavage reaction for pheophorbide a (Pheo a)d uring Chl breakdown underlined this problem. [29] However,isotopic labeling experiments by Folley and Engel with Hv-NCC-1 (1) showed significant incorporation of deuterium into the C2methyl group of 1,t hus supporting the origin of 1 from the degradation of both Chl a and Chl b,a nd suggesting the relevance of ar eduction of the C2-formyl group of Chl b. [30] Chl a/b interconversions were,i ndeed, revealed not only for the biosynthetic oxidative branch from chlorophyllide a to chlorophyllide b, [31] but also its surprising reductive catabolic counterpart that converts Chl b back into Chl a,a nd involves 7 1 -hydroxy-Chl a as an intermediate. [32] In this way,aso-called Chl cycle [11b,33] was shown to regulate the relative levels of Chl a and Chl b throughout the development of the plant, as well as in the initial stages of Chl breakdown in senescent plant tissues. [9c, 32] Surprisingly,t wo different pathways are available in higher plants for the degradation of Chl a to Pheo a (Figure 5). [9c] One of these,r elevant in ripening citrus fruits, involves ester hydrolysis by chlorophyllase [34] to give chlorophyllide a (Chlide a). This is followed by removal of the Mg ion to furnish Pheo a. [35] However, this sequence is not functional in some senescent leaves,i nw hich pheophytin a (Phein a), the suggested product of direct Mg-removal from Chl a,w as observed as an intermediate of Chl breakdown. [36] In A. thaliana,t he serine-type hydrolase pheophytinase was identified recently,w hich converts Phein a into Pheo a,b ut does not hydrolyse Chl a. [37] [23] depicted in representative formulas to highlight its pseudocyclic (left) and extended conformations (right). intermediates of Chl degradation. [38] Lipofuscin-like fluorescing compounds that could be identified in extracts of senescent leaves due to their blue fluorescence were suspected to be intermediate products of Chl breakdown. [39] Theb lue-fluorescing compound 6 (originally named Bn-FCC-2) was first prepared from Pheo a by the use of an enzyme-active extract of senescent leaves of oilseed rape (Brassica napus). [25] Its molecular formula (C 35 H 40 N 4 O 7 ) [40] confirmed ac lose relationship to Pheo a,w ith net addition of two Oatoms and of four Hatoms.B lue-fluorescing Bn-FCC-2 (6)was,thus,proposed to be an intermediate of Chl breakdown and suggested to represent ap roduct of Pheo a oxygenolysis. [25] Pheo a,i ndeed, appeared to be degraded in ap rocess requiring molecular oxygen. [29] Furthermore,s ince Pheo a accumulated in leaves in the absence of oxygen, but not Pheo b,Pheo a was considered the last green intermediate of Chl breakdown in senescent leaves. [29] This role of Pheo a required reduction of Chl b to Chl a as an early step of Chl breakdown. [31,33] What would be ar ational product of the oxygen-dependent degradation of Pheo a and how could it relate to blue-fluorescent 6? [25] Despite the documented failures of detecting colored remains of Chl in natural plant sources,t he unknown red bilin-type catabolite 7 (Figures 6 and 7) appeared to represent such ar ational hypothetical precursor of 6. [25] Studies by Gossauer and Engel [41] on Chl catabolites in Auxenochlorella protothecoides (earlier named Chlorella protothecoides) had, indeed, revealed the existence of similar red, Chl-derived linear tetrapyrroles in secretions of this green alga.
It became crucial to test the role of such ah ypothetical red intermediate in an early phase of Chl catabolism. Thus,t he red 12,13dihydro-1-formyl-19-oxophyllobiladiene-b,c 7 (now commonly called red Chl catabolite,RCC) was prepared from Pheo a by partial chemical synthesis. [42] In analogy to experiments by the Gossauer research group with Cd II complexes of pyropheophorbide a (13 2 -desmethoxycarbonyl-Pheo a), [43] the Cd II complex of Pheo a methyl ester (II)w as photooxidized to furnish Cd II - [4,5]-dioxo- [4,5]seco-4,5-dihydromethylpheophorbidate III in about 30 % yield ( Figure 6). Theb rownish and rather unstable oxidation product was readily reduced with NaBH 4 to give the deep-red RCC methyl ester (7-Me). [42] Enzymatic hydrolysis of 7-Me with porcine liver esterase selectively generated af irst sample of authentic RCC (7)n early quantitatively. [42]   . Photooxidation of the Cd II -pheophorbidate II and reduction of the 4,5-dioxosecophytoporphyrin III furnishes RCC (7)via its methyl ester precursor 7-Me. [42] Figure 7. Chl breakdown from Pheo a via RCC (7)t oprimary FCCs (6/epi-6)i s catalyzed by Pheo ao xygenase (PaO) and by two classes of RCC reductases (RCCR-1 and RCCR-2). [9c] When synthetic 7 was available as ar eference,t races of the very same (previously elusive) red compound were detected when Pheo a was incubated with extracts of chloroplasts of senescent Brassica napus cotyledons. [44] Indeed, RCC was eventually revealed as the product of aRieske-type monooxygenase,named Pheo a oxygenase (PaO), [45] that uses Pheo a selectively as its substrate and that is inhibited by Pheo b. [46] Therefore,R CC (7)c arries the hallmarks of the oxygenolytic ring opening by PaO,considered the key step of the PaO/phyllobilin pathway of Chl breakdown (Figure 7). [9c, 29, 46] In this enzyme-catalyzed process,t he chlorintype macroring of Pheo a is cut open between C4 and C5, and one oxygen atom is specifically incorporated into the newly produced, characteristic formyl group of (an enzyme-bound form of) RCC (7). [46] Thus,RCC (7)isthe native 1-formyl-19oxobilin-type Chl catabolite,ornative phyllobilin.

Primary Fluorescent Chl Catabolites-Epimeric Pair of Phyllobilin Ancestors
Synthetic RCC (7)w as shown in extracts of senescent B. napus leaves to be reduced to the presumed Chl catabolite 6,p rovisionally called Bn-FCC-2. [44] NMR spectroscopic studies [17] had already shown that Bn-FCC-2 (6)w as a1 2,13,16,19-tetrahydro-1-formyl-19-oxophyllobilene-b ( Figure 7). As 6 exhibited the same pattern of peripheral functional groups as Pheo a,and was not modified further by polar groups known from the NCC structures,itwas called the primary FCC (pFCC). [25] In addition to the characteristic absorption maximum of an a-formylpyrrole unit (ring A) near l = 315 nm, [8a] the UV/Vis spectra of 6 showed an ew maximum at l = 360 nm, corresponding to an ew chromophore extending over rings Ba nd C( Figure 8). [6b,25] In contrast to the nonluminescent NCCs,s olutions of FCC 6 exhibit as trong and characteristic blue fluorescence,with an emission maximum near l = 450 nm. It was this feature that led to the original phenomenological classification as an FCC. [6b, 25] By using an aerated enzyme assay based on as uspension of chromoplasts of red sweet pepper (Capsicum annum), Pheo a was transformed into another blue-fluorescing compound, originally called Ca-FCC-2, and identified as an isomer of pFCC (6). [47] Detailed NMR-spectroscopic analysis indicated ac onfigurational difference between 6 and Ca-FCC-2 at C16 (C1 in the earlier phytoporphyrin-based nomenclature [47] ). Ca-FCC-2 was,t herefore,n amed epi-pFCC (epi-6,F igure 7). [47] Reduction of RCC (7)t opFCC (6)w as shown to be accomplished by ac ofactor-free RCC reductase (RCCR). [48] Theidentification of 6 and its C16 epimer epi-6 suggested two stereospecific classes of RCCRs in higher plants.T his conclusion was verified by studies of the RCCRs from ar ange of plants and their classification as RCCR-1 and RCCR-2. [49] RCCR of B. napus is aR CCR-1 and achieves ah ighly stereo-and regioselective reduction of RCC (7)t o pFCC (6). [25,48] Thes ame (regio-and) stereoselectivity is deduced for RCCRs from some other plants,a mong them A. thaliana. [49b] In contrast, in asecond group of plants,among them spinach (Sp.oleracea) [50] and bananas (M. acuminata), [51] C16-epimeric NCCs were found as descendants of epi-pFCC (epi-6)f rom the reduction of RCC (7)b ya n enzyme of the RCCR-2 class. [49a] RCC reductase (RCCR) is aconstitutively expressed enzyme,now classified as belonging to the family of the (ferredoxin-dependent) bilin reductases. [52] Thec rystal structure of RCCR from A. thaliana was analyzed in substrate-free and RCC-loaded forms (see Refs. [53,54] for structural details).

Partial Synthesis of pFCC and epi-pFCC-A Chemical Interlude
Controlled-potential electrolytic reduction of RCC methyl ester (7-Me)i naprotic solvent system furnished FCC methyl esters 6-Me and epi-6-Me in about 25 %y ield, and with negligible stereoselectivity. [55] This (chemical) reaction represented af irst model process for the reaction catalyzed by the cofactor-free RCCRs,e xploring the ease of the one-electron reduction of 7-Me.L ikewise, pFCC (6)a nd epi-pFCC (epi-6)were similarly obtained by electrochemical reduction of synthetic RCC (7), and revealed useful regioselectivity for the addition of hydrogens at the 15-and 16positions,b ut an absence of significant stereoselectivity ( Figure 9). [20b] Hence,t his experiment provided access to both C16 epimers of the natural primary FCCs.
Thefacile electrochemical reduction of 7 to 6 and epi-6 in weakly acidic solution was in line with af erredoxin-driven reduction [54,56] catalyzed by RCCRs.I ndeed, enzymatic reduction of RCC (7)b yR CCR was deduced to follow ar elated mechanistic sequence,t hrough protonation and single electron transfer reduction steps. [54] 4. Branching of the PaO/Phyllobilin Pathway in its Later Stages Degreened leaves of barley (H. vulgare)were not only the first source of the colorless NCC 1,but also of asecond major type of nonfluorescent and colorless Chl catabolites,t he urobilinogenoidic Chl catabolites (UCCs) 8a and 8b. [57] The two epimers, 8a and 8b,f ound in the barley leaf extracts, appeared to be direct descendants of Hv-NCC-1 (1). [57] However,c onsistent with the absence of af ormyl group,t he UV spectra of these phyllobilins lacked the absorption band near l = 315 nm, which is typical of NCCs,s uch as 1.T he tetrapyrroles 8a and 8b are now classified, in as tructure-based way,a s1 ,19-dioxobilin-type NCCs (DNCCs,F igure 10). [10] More recently,t he major colorless phyllobilin in senescent leaves of Norway Maple (Acer platanoides)w as also characterized as aDNCC,and shown by its CD spectra to behave as the enantiomer (ent-8 a)ofthe barley DNCC 8a. [58] This striking finding led to the suggestion that DNCCs may not be derived from NCCs,b ut would probably originate from an earlier Chl-breakdown intermediate,thus basically indicating ad ivergent pathway of Chl breakdown (see Section 4.2). [58] 1-Formyl-19-oxobilin-type Chl catabolites only appear to be present in av ariety of senescent leaves and ripe fruit, such as spinach leaves, [50] apples,a nd pears. [59] In striking contrast, only the remarkable 1,19-dioxobilin-type Chl catabolite ent-8 a was detected in senescent leaves of Norway maple. [58] However,both lines of colorless phyllobilins were present in senescent leaves of Arabidopsis thaliana [27,60] and in broccoli florets (B. oleracea). [61] As we now know,1 ,19-dioxobilin-type Chl catabolites (DCCs), which have in the meantime been classified as type-II phyllobilins,b ranch off from the original lineage of the 1formyl-19-oxophyllobilins,o rt ype-I phyllobilins.B ranching occurs at the stage of FCCs,w here oxidative deformylation competes with pathways to other downstream type-I phyllobilins (hmFCCs,N CCs) and leads mainly to 1,19-dioxobilin-type NCCs (DNCCs). [10,26a] Twos uch branching points have,s of ar, been identified in A. thaliana ( Figure 11). [26c]
[16c] The 1 HN MR spectra of NCCs show acharacteristic singlet corresponding to the CH=Og roup at low field. [8a] The constitution of the phyllobilins and part of their relative configuration was deduced from analysis of high-field 1 H, 1 Hhomonuclear and 1 H, 13 Cheteronuclear NMR spectra. [8a, 10, 17a] All NCCs (or 16,19-dihydro-1-formyl-19-oxophyllobilanes [10] )a re flexible linear tetrapyrroles with unconjugated pyrrolic units.A sd escendants of pFCC (6)o ro f epi-pFCC (epi-6), which differ by their absolute configuration at C16 in aspeciesspecific way,N CCs also occur in two epimeric classes (see,f or example, Ref. [64] and Table 1). With one striking exception, NCCs carry am ethyl group at C2. [10] Then oted exception is At-NCC-3 (15,f rom senescent leaves of A. thaliana), in which ah ydroxymethyl group is attached at C2. [66] Furthermore,a ll NCCs, and other known natural phyllobilins from Figure 11. Branching of Chl breakdown occurs at the level of FCCs and provides pathways to downstreamtype-I and type-II phyllobilins( see Tables 1-3 for examples of R 1 ,R 2 ,a nd R 3 ). [10] [a] Compound number (see text).
[b] R 1 to R 3 refer to ageneralized NCC formula, shown in Figure 13.
The absolute configuration of NCCs at C16 is still unknown;assigned as "n" or as "epi", when the NCC is derived from pFCC (6)o rfrom epi-pFCC (epi-6), respectively.

Angewandte Chemie
Reviews mechanistic view that the isomerization was achieved by an intramolecular protonation of C10 by the propionic acid function. As econd carboxylic acid group at the C8 2 -position of the natural polar At-FCC-2 (22)f urther accelerated the FCC to NCC isomerization by afactor of about 7a tp H5(a probable consequence of alocal charge effect). [67] Therefore, an on-enzymatic isomerization of FCCs to NCCs was proposed to account for the formation of NCCs in the acidic milieu of the vacuoles,w here the free propionic acid function would be partially protonated. Thei ntramolecular protonation step was deduced to preferentially generate NCCs with an R configuration at C10. Consistent with the proposed chemical mechanism of the isomerization of natural FCCs to the corresponding NCCs,n atural NCCs basically exhibit-with few exceptions-similar CD spectra, thus supporting their common absolute configuration at the (C10) meso position between pyrrole rings Ba nd C. [20a,b] Thed educed R configuration of the NCC epi-11 at C10 was recently confirmed by detailed structural analysis,i ncluding X-ray analysis,ofayellow Chl catabolite (YCC) derived from this NCC (see Section 7). [72] Thec ritical role of the propionic acid side chain in the FCC to NCC isomerization was further demonstrated by studying this type of isomerization reaction with the related FCC methyl esters 6-Me and epi-6-Me,w hich were obtained from partial synthesis. [20b] Both epimeric pFCC esters eventually isomerized to the corresponding NCC methyl esters, but with low stereoselectivity at C10, slow reaction rate,a nd low conversion (5-Me/ent-epi-5-Me from 6 and epi-5-Me/ent-5-Me from epi-6-Me Figure 14). [20b] Thus,the mirror images of both natural type NCCs are easily accessible (as methyl esters). Clearly,a ctivation of the isomerization by the free propionic acid function is blocked by its esterification. This remarkable finding has helped to rationalize the surprising accumulation of persistent FCCs in ripening bananas (see Section 6), which are FCCs,b iologically "caged" with ap ropionate ester function. [28a] So far, natural NCCs were deduced to feature acommon R configuration at C10. However,t wo NCCs,r ecently isolated from leaves of birch trees displayed CD spectra that were essentially mirror images of the CD spectrum of 1,thus suggesting ar eversed configuration at C10. [73] This finding suggests,first of all, the relevance of alternative ways that may generate NCCs with the reversed configuration at C10. Possibly,t he FCC to NCC isomerization is achieved by an enzyme-controlled, stereochemically different process in birch leaves.T he observation of the aberrant configuration in birch NCCs emphasizes the importance of CD spectra for the characterization of new NCCs.
Va rious colorless fluorescent Chl catabolites (FCCs), or type-I phyllolumobilins,w ere detected in extracts of senescent leaves,w here they are easily traced by their (blue) fluorescence.Asdescribed above,the primary fluorescent Chl catabolites are formed in higher plants by direct enzymecatalyzed reduction of protein-bound RCC (7). Theabsolute configuration at the C16-position of the epimeric primary FCCs 6 [10,25] and epi-6 could, so far, not be specified. [21a,47] Therefore,F CCs are classified as belonging either to the normal series (for 6 and mFCCs derived from 6)ortothe epi series (epi-6 or mFCCs derived from epi-6). Thec onfiguration at C16, once introduced at the pFCC level, appears to be retained under physiological conditions in mFCCs and in their descendants (NCCs,D NCCs,e tc.). Therefore,N CCs, DNCCs,e tc. also belong either to the normal or to the epi series (Tables 1a nd 2i nt his section and Table 3i nS ection 4.2).
In contrast to the often accumulating NCCs,and with the exception of the unusually persistent hypermodified FCCs (see Section 6), [28a] typical FCCs exist only fleetingly in leaves of senescent plants,a nd their structures have only occasionally been assigned. [10] In addition, primary FCCs appear to be rapidly functionalized further, and the hydroxylation of pFCC/epi-pFCC at the C3 2 -position is an early event in Chl breakdown in higher plants,p robably taking place in the chloroplasts.

Angewandte Chemie
Reviews and turned out to represent FCC esters with b-glucopyranosyl units attached at the critical propionate at O6' (their primary OH group). Among these hmFCCs, Ma-FCC-69 featured a3 ,4-dihydroxyphenylethyl aglycon at its glucopyranosyl ester moiety,t hus representing the C16 epimer (epi-25) [76] of Sw-FCC-62 (25)from leaves of Sp.wallisii. [74] This normal/epi stereodivergence is due to the differing classes of RCC reductases:i nb anana leaves (and, likewise,i nb anana fruit; see Section 6), ar eductase of the RCCR-2 type is present, whereas in Sp.wallisii,a nR CCR-1 produces colorless Chl catabolites of the normal series. [74] FCCs accumulate in as triking abundance in senescent banana leaves,b ut NCCs were not detected, nor was there any indication of the presence of type-II phyllobilins. [76] In the two less-polar hmFCCs from banana leaves, Ma-FCC-63 (epi-27)a nd Ma-FCC-64 (epi-28), a b-o ra na-glucopyranosyl moiety,respectively,was attached through the primary 6'-OH group.The free anomeric center of epi-27 and epi-28 allowed for the mutual interconversion by spontaneous anomerization in aqueous solution. [76] Thep resence of the latter anomeric hmFCCs, epi-27 and epi-28,i ne xtracts of senescent banana leaves strengthened the suggestion that the observed hmFCCs could either be precursors,o r, possibly,r emnants or partial degradation products of still elusive further functionalized hmFCCs.E sterification of the critical propionate function of FCCs by glucopyranosyl or, alternatively,by galactopyranosyl groups provides two distinct lines among the persistent banana leaf Ma-FCCs.T he sugar units of these hmFCCs provide attachment sites for further groups.Indeed, several minor,s till less polar FCC fractions were recently analyzed structurally,a nd were revealed to be hmFCCs, formally derived from epi-27/epi-28,b ut functionalized further by unusual terpenoid aglycons with b-glycosidic linkages. [77] 4.2. The Type-II Phyllobilin Branch-Colorless 1,19-Dioxobilins Thed iscovery of 1,19-dioxobilin-type NCCs (DNCCs) raised the question of their formation and of their natural relevance in Chl breakdown. [57,58] As products of the PaO/ phyllobilin pathway,t he native phyllobilins are 1-formyl-19oxobilins (or type-I phyllobilins). [9c, 10] DNCCs,t he 1,19dioxobilin-type Chl catabolites,a re the offspring of as ubsequent step of Chl breakdown and are,thus,classified as type-II phyllobilins.I nterestingly,t he 1,19-dioxobilin-type structure of DNCCs (Figures 11 and 17) [57,58] makes them look remarkably similar to the heme-derived bilins,t he hemobilins. [3] Recent studies have revealed astriking abundance and constitutional and stereochemical variety of DNCCs that rivals that of the now better studied NCCs,a nd populating ag rowing second branch of phyllobilins,o ft he type-II phyllobilins. [10] 1,19-Dioxobilin-type nonfluorescent Chl catabolites (DNCCs) are formal products from an oxidative removal of the formyl group of NCCs. [57] Thel ack of the characteristic absorption of NCCs at l % 315 nm (Figure 9) makes DNCCs more difficult to detect by their UV absorption. Such colorless dioxobilin-type Chl catabolites (DCCs) accumulate in avariety of senescent leaves.D CCs may occur in leaves together with type-I phyllobilanes,asisthe case in senescent leaves of A. thaliana. In wild-type A. thaliana leaves,t he polar 1,19dioxophyllobilane At-DNCC-33 (30)i s, by far, the most abundant phyllobilin, with isomeric At-DNCC-45 (31 a)a nd  Figure 17. Generalized formulas of colorlessa nd nonfluorescent type-II phyllobilins: DNCCs, 4-hydroxymethyl-DNCCs and 2-hydroxymethyl-iso-DNCCs. [26] Angewandte Chemie Reviews At-DNCC-48 (31 b)being minor components. [27] The1 ,19-dioxophyllobilanes At MES -DNCC-47 (32)a nd At MES -DNCC-38 (33)w ere found in extracts of the A. thaliana MES16 mutant (besides minor NCC fractions;T able 3). [26a] Bilane 33 is aC 16 epimer of Vv-DNCC-51 (epi-33)f rom degreened grape wine leaves. [78] Thei someric DNCCs 8a, 8b,a nd ent-8 a ( Figure 10 and Table 3) were discovered in senescent leaves of barley [57] and of Norway maple. [58] Theo bservation of 1,19-dioxobilin-type NCCs (DNCCs) first raised the question of their formation and of their general metabolic relevance for natural Chl breakdown. [57,58] Based on astereochemical divergence indicated by the deduced structure of the DNCC ent-8 a from leaves of Norway maple,asplit of the PaO/phyllobilin pathway at the level of the fluorescent Chl catabolites (FCCs) was proposed, which, in consequence,w ould involve the intermediate existence of one (or of several) 1,19-dioxobilin-type fluorescent Chl catabolite (or DFCC) intermediate(s). [58] Indeed, ac ytochrome P450 enzyme (CYP89A9) was recently identified in A. thaliana that catalyzed the in vitro deformylation of epi-pFCC (epi-6), thereby furnishing four epimeric DFCCs. [27] Ap air of these DFCC epimers isomerized rapidly to ap air of DNCCs in weakly acidic solution. These in vitro experiments clarified the basic constitutional features of an FCC deformylation and of the DFCC to DNCC isomerization, proposed to be early key steps of the dioxobilin branch of Chl breakdown. [27] Clearly, the intriguing deformylation by the cytochrome CYP89A9 requires further investigation. General precedence for the removal of formyl (or acyl) groups by P450 enzymes exists. [79] However,t here appears to be none for aP 450-catalyzed oxidative deformylation at an aposition of ap yrrole unit. The inferred nucleophilic (hydro)peroxo-Fe III intermediate of the P450 cycle was suggested to induce oxidative (CÀC) bond cleavages. [79b, 80] This would thus imply an insertion of oxygen atom into the previous (C À C) bond with formation of aformate ester,r eminiscent of the Baeyer-Villiger reaction (Figure 18, bottom). Hydrolysis of this putative ester,r emoval of the currently unknown C1 fragment (possibly formic acid), and protonation at C4 could all take place without assistance by the P450 enzyme,t hereby helping to explain the lack of stereoselectivity observed in the in vitro experiment with CYP89A9. [27] Thus,t he remarkable in vitro results with CYP89A9 did not provide afirm conclusion with respect to the stereochemical outcome of the DFCC/DNCC isomerization, nor was am ajor step of the natural dioxobilin pathway clearly identified. Careful analysis of an extract of A. thaliana leaves at an early degreening stage,r evealed am inor (blue) fluorescent fraction exhibiting ac haracteristic band at l = 360 nm (from the conjugated B/C part corresponding to FCCs), but lacking the absorption at l = 320 nm of an aformylpyrrole unit ( Figure 8). As ample of the fleetingly existent natural DFCC 34 (a 3 2 -hydroxy-1,4,12,13,16,19hexahydro-1,19-dioxophyllobilene-b)w as recently isola- Table 3: Structures of natural dioxobilane-type nonfluorescent Chl catabolites (DNCCs, top section) and of iso-DNCCs (bottom section): labels R 1 ,R 2 ,a nd R 3 refer to the general constitutional formula of DNCCs and iso-DNCCs,s hown in Figure 17.
No. [a] R 1 R 2 R 3 C16 [b] Provisional names [c] Ref.  [78] [e] Hydroxymethyl group at C4 (see Figure 17 for generalized formulas of DNCCs and iso-DNCCs).  Hence,anaturally existing, functionalized DFCC was identified and, simultaneously,a n important natural branching point to the type-II phyllobilins was revealed in ah igher plant.

Chl Breakdown in Arabidopsis thaliana-A Model Case
Theg rowing systematic biological knowledge concerning Arabidopsis thaliana has also become an important resource in the field of Chl breakdown, [27] assisting the identification of an umber of enzymes in this model plant, [9c,11a] in fruitful synergy with our recent complementary work concerned with the discovery and structure elucidation of an extraordinary number of Chl catabolites. [26,27,60,67] Arange of colorless type-I phyllobilins were found in earlier analyses of extracts of senescent leaves of (wild-type) A. thaliana,including five At-NCCs (3, 4, 15-17, [60,66] see Table 1) and three At-FCCs (6, 22, 24,T able 2). [60] More recently,c olorless type-II phyllobilins were discovered in A. thaliana,a nd the At-DNCCs 30, 31 a, and 31 b were characterized (Table 3), [26b, 27] as well as the fleetingly existent DFCC 34. [26c] However,v arious further colorless,n onfluorescent phyllobilins were observed, apparently related to DNCCs,i n extracts of senescent leaves of A. thaliana (either of wild type [26b] or of the MES16 mutant [26a] ). These were,p rovisionally,classified as nonfluorescent DCCs (NDCCs) on the basis of their UV spectra, which were similar to those of the structurally characterized 1,19-dioxobilin-type NCCs (DNCCs) 30 and 31 a/31 b.H owever, as deduced from mass spectra and NMR-spectroscopic analyses,s everal of these NDCCs exhibited ap uzzling carbon-hydroxymethylation and, thus,d id not have the proper chemical constitution of DNCCs. [26a,b] A2 -hydroxymethyl-iso-DNCC (At-2HM-iso-DNCC-43, 36)a nd a4 -hydroxymethyl-DNCC (At-4HM-DNCC-41, 37)w ere discovered ( Figure 19 and Table 3) [26b] in extracts of senescent leaves of wild-type A. thaliana,while the corresponding methyl esters 2-hydroxymethyl-iso-DNCC At MES -2HM-iso-DNCC-46 (38)a nd the 4-hydroxymethyl-DNCC At MES -4HM-DNCC-44 (39)w ere found in extracts of the MES16-mutant of A. thaliana. [26a] In the leaves of this mutant, an additional minor fraction of ac olorless blue fluorescent phyllobilin was noticed, which lacked the l = 315 nm band in the UV spectra, but showed the characteristic absorption maximum of FCCs near l = 360 nm. [26a] Thus,this fluorescent compound was tentatively classified as af luorescent DCC (FDCC). It was isolated and characterized as the 2hydroxymethyl-iso-DFCC At MES -2HM-iso-DFCC (40 = 2HM-iso-pDFCC,s ee Figure 19), [26a] which differed from genuine DFCCs by the constitution of its ring A (Figure 19 >). Thes tructural properties of the FDCC 40 suggest its role as the direct precursor of the At MES -2HM-iso-DNCC-46 (38), [26a] its nonfluorescent isomer. All of these remarkable carbon-hydroxymethylated type-II phyllobilins lack an oxygen functionality at their C3 2 -position, which is typical of most natural phyllobilins. [9c, 10] Hence,t heir structures relate them to pFCC (6)a st heir (direct) precursor. Hydroxymethylation appears to be tightly associated with the oxidative deformylation of pFCC to type-II products.P ossibly,i ti sacytosolic "rescue operation" that introduces a( needed) polar functionality at ring Ao ft he catabolites. In this sense,t he hydroxymethylation (at the carbon atom) has been considered a" biosynthetic intermezzo" in the course of the type-II branch of the PaO/phyllobilin pathway. [26a,b]

Reviews
In summary,avast variety of colorless phyllobilins are produced in A. thaliana by breakdown of Chl, spearheaded by formation of pFCC (6)i nt he chloroplasts,a sw ell as, presumably,o f3 2 -hydroxy-pFCC (23). [9c, 26c] Thed escendants of pFCC (6)t hat lack an OH group at the C3 2 -position are minor components among the Chl catabolites in A. thaliana, despite their particularly diverse nature (Figures 19 and 20). Themajor phyllobilins in A. thaliana (both the wild type and MES16 mutant) are,instead, derivatives of 3 2 -hydroxy-pFCC (23). Its type-II phyllobilin descendants dominate over the type-I analogues.E vidence for export of the FCCs 6 and 23 into the cytosol and for their further independent processing to modified FCCs,D FCCs,a nd FDCCs is derived from the structures of the FCCs 22, 24,DFCC 34,and FDCC 40, as well as of their nonfluorescent descendants.T he latter are believed to arise by acid-catalyzed isomerization of their fluorescent precursors after import into the vacuoles.T he deduced in vivo deformylations of pFCC (6)a nd of 3 2 -hydroxy-pFCC (23)e stablish two natural branching points from type-I to type-II phyllobilins in leaves of A. thaliana. [26c]

Long Overlooked Chl Catabolites in Fruit and Vegetables
Thed isappearance of chlorophyll is commonly associated with the appearance of fall colors.However,ripening fruit (and also vegetables) often undergo degreening processes (Chl breakdown) that are visually similar to those observed in senescent leaves ( Figure 21). [9c,11c] Hence,t he question arose "What happens to Chl when fruit ripen and vegetables degreen?"

Colorless Chl Catabolites in Fruit and Vegetables
As arule,when apples,pears,orother fruit ripen, the associated appearance of the appetizing colors of the ripe fruit is av isual indicator of their degree of ripeness. [81] At the same time,t he Chl originally present in the unripe green fruits is broken down, presumably to produce phyllobilins.C hl breakdown in the peel of Golden Delicious apples (Malus domestica)a nd of Williams pears (Pyrus communis)w as shown to yield the nonfluorescent type-I phyllobilins epi-9 and epi-11,also named Md-NCCs and Pc-NCCs. [59] Thesame (epi-type,that is,RCCR-2 derived) NCCs were also found in senescent leaves of the corresponding apple and pear trees, thereby indicating acommon pathway in the leaves and fruit of these fruit trees. [59] Several NCCs,including Ej-NCC-2 (epi-29,T able 1), were identified (on the basis of mass-spectrometric and UV-spectroscopic data) in quince (Cydonia oblonga,M iller) [82] and in loquat fruit (Eriobotrya japonica). [83] Likewise,t he NCCs Bo-NCC-1 (3)a nd Bo-NCC-2  Figure 19 for the structures of type-II phyllobilins derived from pFCC). Figure 21. Ripening fruit (left) and degreening florets of broccoli (right) undergo Chl breakdown and accumulate colorlessp hyllobilins. [59,61] Angewandte Chemie Reviews (17), as well as Bo-DNCC-3 (30), were characterized in degreening broccoli florets (Brassica oleracea,v ar. Ital.). These three NCCs are known representatives of the normal stereochemical series of type-I and type-II phyllobilins. [61] Five NCCs were described earlier in senescent spinach leaves (Spinacia oleracea), the so-called So-NCCs (epi-1, epi-5, epi-11, epi-12, epi-13), which belong to the epi series of NCCs. [50,64] As expected, in ripe(ning) fruit and (degreening) vegetables,Chl breakdown follows the common PaO/phyllobilin pathway and furnishes colorless type-I and type-II phyllobilins,inaspecies-dependent way.Clearly,these plantderived components of our food are acommon source of Chl catabolites,w hich, hence,a re part of our daily nutrition. [6a]

Persistent Blue Luminescent Chl Catabolites in Bananas
Ther ipening of bananas is associated with the typical development of ab right yellow color,w hich, in turn, is commonly considered acritical visual indicator of the degree of ripeness of the banana fruit. Clearly,during the degreening process,t he Chl present in the peels of the unripe fruit is degraded. We were,t herefore,i ntrigued to analyze bananas for Chl catabolites.
[28a] Surprisingly,s ome of the catabolites found in the peels of freshly ripe bananas (Musa acuminata, Cavendish cultivar) were revealed to be persistent FCCs and to belong to the then unprecedented group of hypermodified FCCs (hmFCCs). These FCCs make the ripening bananas glow blue,a si sb est seen when analyzed by irradiation with black light and observation in ad ark room ( Figure 22). [28a] Several persistent hmFCCs accumulate in the peels of ripe(ning) bananas,w here the Mc-FCCs epi-23, epi-41,a nd epi-42 represent as izeable fraction of the phyllobilins.T he major hmFCCs of the peel of ripe bananas, Mc-FCC-56 (epi-41)and Mc-FCC-53 (epi-42), feature an ester at the propionic acid function that is derived from daucic acid [84] itself. [ (Table 2a nd Figure 22). [28a, 85] Thus, Mc-FCCs differ characteristically from related hmFCCs so far found in senescent leaves of bananas [51] or of the Peace Lily, [74] which are all esterified with typical natural hexopyranosyl units.
In the early phase of the ripening process of bananas,the classical FCC 3 2 -hydroxy-epi-pFCC (Mc-FCC-62, epi-23) could also be observed as amajor FCC fraction, the rational precursor of the hypermodified FCCs (hmFCCs) in the banana peel. [51] In addition, av ariety of Mc-NCCs were characterized in extracts of the banana peels.H owever, similar to the banana leaves, [76] representatives of the type-II phyllobilins were not detected. [51] These findings indicated an exclusive role of type-I phyllobilins (of the epi-type) in banana peels,a sw ell as ap athway of Chl breakdown that diverges into two branches of type-I phyllobilins at the level of FCCs (Figure 23). This stage of Chl breakdown is presumed to be located in the cytosol, where formation of specific hmFCCs through esterification of 3 2 -hydroxy-epi-pFCC (epi-23)w ith ad aucyl group competes with other modifications that furnish typical mFCCs.T he latter are presumed to be transported into the vacuoles for eventual rapid isomerization to Mc-NCCs (Table 1). [51] As is commonly observed, dark, senescence-associated spots develop naturally on the peels of very ripe bananas. [87] This deterioration can be inhibited by protection from air (or oxygen). [88] These spots arise around the stomata as asign of local senescence and eventual cell death. [85] Strangely,g lucosylated hmFCCs (especially epi-43 and epi-44)a ccumulate specifically in the senescent area around the dark spots.T his local FCC enrichment in areas encircling the growing necrotic spots can easily be observed in darkened rooms,w hen black light is used as the light source. [85] Theb lue fluorescent rings observed on the peels of overripe bananas arise in areas committed to programmed cell death, and have,t hus,b een considered to represent "blue halos of cell death". [85]

Phyllochromobilins from Oxidation of Phylloleucobilins
Early on, NCCs were called "rusty pigments". [8a, 14] Indeed, samples of these colorless products of Chl catabolism, which often accumulate as apparently final products of Chl breakdown in senescent leaves of higher plants,r eadily become rust colored. [7] Analysis of such ac olored mixture obtained by exposure of asolution of the common NCC epi-11 to sunlight in the presence of air revealed the presence of the yellow compounds 46Z and 46E,aswell as of pink-colored 47,which we classified as yellow Chl catabolites (YCCs) and pink Chl catabolites (PiCCs), respectively ( Figure 24). [6b,7,89]  Intriguingly,t he E/Z-isomeric YCCs 46Z and 46E were identified in freshly prepared extracts of yellow Cercidiphyllum japonicum leaves.T his observation supports the actual presence of such phyllochromobilins in senescent leaves and indicates their contribution to the color of senescent leaves. [7] In polar solution, YCCs show at ypical absorption band at l % 430-440 nm, [7] as do bilirubin (BR) [90] and model dipyrrinones. [3] When as olution of the YCC 46Z was exposed to daylight, 46Z isomerized (in part) to the E isomer 46E. [89] Structure analysis of YCCs 46Z and 46E by heteronuclear NMR spectroscopy indicated an unsaturated "western" meso position and, as ac onsequence,ap-conjugated chromophore extending over rings C and D. In aformal sense, 46Z and 46E are formed from NCC epi-11 by oxidative desaturation with formation of aC 15=C16 double bond. This (part of the) chromophore of YCCs features the same remarkable (local) properties as the C/D part of the chromophore of bilirubin (BR). [3,90] X-ray analysis of crystals of 46Z-Me (the methyl ester of 46Z)confirmed the structure of 46Z-Me (and, thus,o f46Z)t hat had been deduced from their spectra. [7] In addition, it revealed ah ydrogen-bonded dimer of 46Z-Me in the crystal and verified the earlier deduced R configuration of 46Z [7] and of its precursor epi-11 [20a] at C10. [72] Oxidation of the NCC epi-11 by dicyanodichlorobenzoquinone (DDQ) opened up as emipreparative pathway to the YCCs 46Z and 46E. [89] In the course of this synthetic transformation, 11 or epi-11 was stereoselectively hydroxylated to 15-OH-11 or to 15-OH-epi-11, respectively ( Figure 25). Surprisingly,selective oxidation of either endogenous NCC 11 or of added epi-11 also occurred in aqueous homogenates of green or senescent Sp.wallisii leaves in the presence of air (or molecular oxygen), with formation of 15-OH-11 or 15-OH-epi-11.Both of the oxidized NCCs 15-OH-11 or 15-OH-epi-11 are efficient precursors of the same YCC 46Z,as ar esult of selective acid-induced elimination of water from C15 and C16. [91] Homogenates of (green or senescent) Sp.wallisii leaves contain as till poorly characterized oxidative activity (likely to be enzyme-based), which provides an entry to the endogenous formation of the Figure 23. hmFCCs and NCCs are generated in the peels of ripeningbananas, thereby indicating ap athway of Chl breakdown that is split at the stage of 3 2 -OH-epi-pFCC (epi-23). [51] Figure 24. Structural formulas of yellow Chl catabolites (YCCs) and pink Chl catabolites (PiCCs), which may contribute to the color of senescent leaves of deciduous trees. [89] Figure 25. The NCCs 11 and epi-11 are oxidized by an extract from Sp. wallisii leaves to epimeric 15-OH-NCCs, which are dehydrated in weak acid to furnish the YCC 46Z. [91] Angewandte Chemie Reviews YCC 46Z from 11 or epi-11,a sw ell as from some other NCCs. [91] Thes cope of this type of "green synthesis with leaves" on the basis of their still puzzling oxidative activity remains to be explored, as does the selectivity and preparative limitations of this type of transformation. An analogous oxidation with DNCCs (as observed with NCCs 11 and epi-11 [91] ), could provide apossible pathway to the corresponding dioxobilin-type YCCs (DYCCs). YCC-type compounds were not only detected in senescent leaves of C. japonicum, [7] but also in fresh extracts of av ariety of senescent leaves,f or example,ofthe deciduous lime [92] and Egeria densa trees, [93] as well as in the peel of ripe bananas. [51]

Pink Phyllobilins-The Phylloroseobilins
Similar to bilirubin (BR), [90] YCCs are easily oxidized in the presence of air or molecular oxygen. Pink Chl catabolites, classified as PiCCs,a re obtained from the oxidation of YCCs. [89] Fors ynthetic purposes,t he oxidation of,f or example,t he YCC 46Z,c an be achieved efficiently in the presence of an excess of Zn II ions,t hereby furnishing the bright blue Zn II complex Zn-47. [94] Tr eatment of Zn-47 with acetic acid or phosphate removes the Zn ion and liberates the PiCC 47 nearly quantitatively ( Figure 26). [94a] Thep ink phyllobiladiene-b,c 47 features along-wavelength absorption band at l % 520 nm, consistent with the further extension of the conjugated p system to ring B ( Figure 27). Thec hromophore of the PiCC 47 exhibits ar emarkable correspondence to that of the heme-derived bilin phycoviolobilin. [3,56] Desaturation of the "southern" C10-position of the YCC 46Z caused the PiCC 47 to become available as ar acemate:i ts single asymmetric center (C8 2 )i sa cidified by adjacent functionalities that assist its fast racemization. [94a] Detailed NMR-spectroscopic analysis of the PiCC 47 indicated astriking E configuration of the C10=C11 double bond. An X-ray crystal-structure analysis confirmed the NMR-derived structure and revealed the oxidized phyllobiladiene-b,c 47 as apair of hydrogen-bonded enantiomers in the crystal, which p stack, thanks to their extended planar chromophore system (see Figure 26). [94] Ther ecent observations of the natural occurrence of YCCs and PiCCs points to am ore general relevance of further endogenous transformations of colorless Chl catabolites in senescent plants that go beyond the stage of the abundant type-I and type-II phylloleucobilins.S uch endogenous processes may represent important further steps of natural Chl breakdown, thereby helping to explain the eventual pronounced decrease in the amount of colorless phyllobilin-type Chl catabolites in leaves,f requently noted when analyzing leaves undergoing progressive senescence over several days to weeks.  Phyllobilins,C hl-derived linear tetrapyrroles, [10] have structures related to those of the heme-derived (hemo-)bilins. [3,90] Accordingly,p hyllobilins are expected to display diverse photo-, coordination and redox chemistry.H owever, the chemical properties of phyllobilins have barely been explored to date. [10, 94b,95]

Phyllobilins as Photoactive Tetrapyrroles
Thephotochemical properties of the breakdown products of the green photosynthetic pigment Chl have been of prime interest. Theo bserved rapid catabolic transformation of Chl into colorless phyllobilins has been rationalized primarily as the destruction of potentially phototoxic Chl. Hence,t he observed rapid formation of the nonfluorescent and colorless type-I and type-II phyllobilins (NCCs and DNCCs) is completely in line with such ad etoxification aspect of Chl breakdown. Indeed, NCCs and DNCCs,w hich typically accumulate in senescent leaves,d isplay absorptions limited to the UV region of sunlight (Figure 8), and they,f urthermore,l ack photoactivity.
Remarkably,p hotoactive breakdown intermediates do not typically accumulate during the rapid breakdown of Chl on the way to the nonfluorescent phyllobilins.Fluorescent Chl catabolites,s uch as pFCC (6)a nd 3 2 -hydoxy-pFCC (23), as well as their C16 epimers,represent an important intermediate stage in the PaO/phyllobilin pathway. [9c] TheF CC 6 absorbs very little light in the visible region ( Figure 9). As their classification suggests,F CCs are effective emitters of blue fluorescence. [28a, 95] Thus, epi-23-Me,t he semisynthetic methyl ester of 3 2 -hydoxy-epi-pFCC (epi-23)f eatures an emission with am aximum at l = 437 nm and af luorescence quantum yield of 0.21 (lifetime:1 .6 ns in ethanol). [95] The photoexcited epi-23-Me undergoes intersystem crossing into the triplet state (with aq uantum yield of 0.6), from which it generates singlet oxygen with nearly 100 %e fficiency. Thus, the FCC epi-23-Me is aremarkably potent sensitizer of singlet oxygen ( 1 O 2 ). However,F CCs exist only fleetingly in senescent leaves,e xcept when "caged" as the persistent hypermodified FCCs (hmFCCs). [28a] Thed educed, similar photochemical features of hmFCCs are noteworthy,a st hey may play ap hysiological role in senescent plant tissue and in ripening fruit (see Section 9).

Phyllobilins as Antioxidants
As is typical for bilanes and strongly reduced hydroporphinoids, [96] NCCs are easily oxidized. [89] In line with this property,t hey are also remarkable amphiphilic antioxidants, [59] as evident by their inhibitory effect in the classical autoxidation reaction of linoleic acid. [97] In such tests,t he NCC epi-11 exhibited an only five times lower capacity than bilirubin (BR), [59] an effective and physiologically important antioxidant. [97] Analogous investigations of the effect of the corresponding YCC 46Z on the autoxidation of linoleic acid indicated this YCC to even inhibit about 3-5 times more effectively than BR, not unexpected in view of the similar features of their chromophores. [98] Experiences on the antioxidant effect of other phyllobilins (such as model FCCs, DNCCs,and DY CCs) would also be of interest. This aspect of the properties of phyllobilins still remains to be studied.

Phyllobilins as Ligands in Transition-Metal Complexes
In contrast to the cyclic tetrapyrroles,which enrich nature with the important metalloporphyrinoid cofactors, [71,99] al ess typical feature of linear tetrapyrroles is their ability to bind metal ions. [3,100] However,h eme-derived bilins and phyllo-(chromo)bilins can be considered to share similar features as multidentate ligands for (transition-) metal ions. [94b,100] Four nitrogen centers are available in linear tetrapyrroles for the coordination of metal ions. [94b] However,Natoms of isolated pyrrole rings can hardly compete with polar solvent molecules for coordination at metal ions.T herefore,t he photoinactive nonfluorescent NCCs and DNCCs are judged as lacking the capacity for complexing. [94b] In contrast, the availability of nitrogen atoms of the imine and enamine types in phyllochromobilins provides centers for coordination to transitionmetal ions.Sofar,the scarcity of such phyllobilins has limited the corresponding studies to the coordination behavior of PiCC 47 and, to some extent, of YCC 46Z,a nd its methyl ester 46Z-Me. [94] PiCC 47 binds the transition-metal ions Zn II , Cd II ,N i II ,C u II ,a nd Pd II in 1:1c omplexes with high affinities and with reaction rates in the order of roughly 100 m À1 s À1 ( Figure 28). Binding of these transition-metal ions is easily observed by astrong red-shift of the absorption bands in the visible region, for example,b yDl % 100 nm to about l = 620 nm for the complex Zn-47.T ridentate binding of the Zn II ion by 47 has been deduced from NMR data in solution and ahigh affinity (in this case) with alinear 1:1stoichiometry down to 1nm solutions of Zn II . [94a] Binding at such low concentrations could be quantified by analysis of fluorescence emission. In contrast to the very weak fluorophore of 47,the monomeric and diamagnetic complex Zn-47 features astrong fluorescence with an emission maximum at l = 650 nm (see Figures 27 and 28). [94a] Coordination by such transition-metal ions restructures the E,Z-phyllobiladiene-b,c 47 to a Z,Zstructured and effectively tricoordinate ligand ( Figure 26). [94] In Zn-47,t he coordinated metal ion also constrains and rigidifies the chromophore part of the ligand, thus inhibiting deactivation pathways through light-induced E/Z isomerization, which are presumably available to the free PiCC 47. [94a] TheY CC methyl ester 46Z-Me effectively behaves as ab identate ligand for Zn II ions. 1 HNMR NOE experiments with the diamagnetic Zn complex of 46Z-Me suggested interligand distances indicative of a1 :2 arrangement in the symmetrical complex Zn(46Z-Me) 2 ( Figure 29). [94b,c] Similar to the situation with PiCC 47,the binding of the closed-shell Zn II ion to 46Z-Me in DMSO solution also shifted the longwavelength absorption maximum from l = 430 nm to 484 nm. At the same time,t he luminescence of 46Z-Me (weak emission maximum at l = 495 nm) was red-shifted in the complex Zn(46Z-Me) 2 ,w ith am aximum at l = 538 nm, and was also intensified nearly 100-fold. [94b] Clearly,t he littleexplored behavior of some of the phyllobilins,s uch as PiCC 47 and YCC 46Z,i nc oordinating transition-metal ions indicates ac onsiderable potential of these bilin-type ligands in binding metal ions at low concentrations and, thus,inalso serving as effective indicators of the presence of metal ions. [94a,b]

On the Role of Chlorophyll Breakdown in Higher Plants-Time for aNew Paradigm
Having deciphered, to some extent, how Chl is degraded in some higher plants,wemay now be in abetter position to address the challenging question of why Chl breakdown occurs. [9a,c] Considering the massive amounts of Chl degraded each year on Earth, the phenomenon of Chl breakdown cannot be understood as amere visual spectacle in nature and as apriceless tourist attraction in some areas of the world. [101] On the contrary,Chl breakdown should be considered, above all, as having beneficial consequences for the plants themselves.
[101b] Furthermore,other organisms interacting with plants may also benefit (indirectly), such as,f or example,a nimals and humans that make use of plant-based nutrition. Therefore,and in view of the sheer amount of phyllobilins produced when Chl disappears, [10] Chl breakdown deserves close scientific attention, not only from fundamental and applied research, but also from basic agrobiological and ecological points of view. [9][10][11]  In ah istorical interpretation, Chl breakdown was given arole in serving the direct recovery of the four nitrogen atoms of the Chl from leaves. [2] Taking note of the now established build-up of the phyllobilins as linear tetrapyrroles,t his view can, clearly,n ot be supported any longer. [1a, 8a,9c] However, am ajor direct consequence of Chl breakdown in senescent leaves is the destruction of the abundant and highly photoactive green plant pigment, thus,e ffectively getting rid of ac ellular constituent with as trong potential as ap hototoxic agent. [9a] In addition, removal of the Chls from their proteinaceous binding partners also renders the latter more labile for proteolysis. [102] This consequence of Chl breakdown obtained strong support from studies with "stay green" mutants (which retain their Chl), in which the relocation of proteinogenic nitrogen atom, and its recovery were found to be strongly reduced. With this knowledge,t he ambivalent role of Chl breakdown in plants cannot be overestimated: [2,5,101b] on the one hand, it eliminates the photosynthetic capacity of Chl, thus reducing am ajor supply line that drives metabolic activities;onthe other hand, by delivering important nutrient components,s uch as reduced nitrogen, it contributes to enriching protein in crops and vegetables,a nd to the development of nutritious fruit, both of which are important from an economic point of view. [11c, 103]

On Physiological Roles of Chlorophyll Catabolites in Plants
Rather than disappearing in senescent leaves "without leaving at race", [2] Chl furnishes the abundant phyllobilins. Thelatter,instead of being considered mere waste from Chldetoxification, [1a, 9a] should now be given attention with respect to their possible physiological roles in plants.S trikingly,biological functions of bilin-type Chl catabolites are so far still unknown. This may,inpart, be due to alack of reliable samples of typical Chl catabolites,a nd to difficulties in their practical application. As arule,phyllobilins are complex and rather unstable compounds.I nt his latter respect, the "persistent" hmFCCs [28a] appear to represent ar emarkable exception:i nhmFCCs,t he inherent tendencyo fF CCs to undergo isomerization to the corresponding NCCs is blocked (hmFCCs are biologically "caged" FCCs) as aconsequence of abiosynthetic effort in the (banana) plant, which is seemingly useless in the absence of af urther biological role. [85] Several phyllobilins,s uch as nonfluorescent type-I phyllobilins (NCCs), have been shown to be effective antioxidants. [59] Likewise,Y CCs are excellent antioxidants. [98] Such properties of the amphiphilic phyllobilins are of considerable interest. [59] DNCCs,the structurally related type-II analogues, are likely to show basically similar effects.A ntioxidants are frequently abundant in plants,e specially in ripe fruit, [81a] and may be crucial for prolonging the viability of cells during senescence [24] and ripening. [103] Except for the NCCs and DNCCS,m ost phyllobilins are classified as photoactive compounds,a nd may,h ence,b e considered relevant as cellular metabolites with the capacity to act as light filters.Phyllochromobilins (YCCs,PiCCs,etc.) and their metal complexes could make ap articularly important contribution in this respect. Thus,t he colors of phyllochromobilins contribute to the pigmentation of leaves [7] and fruit. [51] Thecolors of leaves [104] and fruit [105] are considered to be important signals for av ariety of animals.F CC fluorescence may,secondly,have aspecial role as optical brighteners in bananas and banana leaves [104a, 106] or act as direct fluorescence signals to animals. [107] FCCs are specifically efficient sensitizers of singlet oxygen ( 1 O 2 ): TheFCC epi-23-Me sensitized light-induced generation of 1 O 2 with aq uantum yield of about 0.6. [95] Clearly,t he photochemical capacity of FCCs in sensitizing the formation of 1 O 2 may play aphysiological role in (senescent or ripening) plant tissue. 1 O 2 can function in pathogen defense,a sw ell as acting as atransiently existing signal molecule with messages for receptors in the cell nucleus. [108] Important signals mediated by 1 O 2 are believed to originate from chloroplasts. [109] Photoexcited FCCs could behave as alternative production sites for 1 O 2 in the cytosol. In light of the important roles of the transiently existing 1 O 2 [110] in plants, [108] accumulation of the persistent hmFCCs in ripening bananas [28a, 85] and in leaves of some other evergreens [28b,74, 76] may indicate ap hysiologically relevant role of FCCs as 1 O 2 sensitizers. [95] In line with this premise,t he formation of the blue fluorescent rings around senescence-associatedd ark spots on the peel of ripe bananas,a saresult of the local accumulation of hmFCCs,m ay have af unction in detaining pathogens and in sustaining the viability of the skin tissue. [85] Phyllobilins have abroad capacity for acting as ligands for transition-metal complexes.T he pink Chl catabolite (PiCC) 47 binds some transition-metal ions down to nanomolar concentrations. [94a] Their metal-chelating ability [94b,c] could be the basis for particular physiological roles of some phyllochromobilins in plants characterized as heavy-metal accumulators. [111] Such plants concentrate Zn, Cd, Hg, and Ni ions in their vacuoles,where pigments,tentatively assigned as natural phyllochromobilins,w ere,i ndeed, first found. [13c] Thef ormation of metal complexes may,t hus,b el ikely.F urthermore, tridentate and tripyrrolic transition-metal complexes were shown to possess druglike properties. [112] Therefore,t he capacity of PiCCs for binding metal ions in at ridentate fashion may be of particular relevance from plant physiological and pharmacological points of view. [94b]

Summary and Outlook
In the past 25 years,r esearch on the identity and on the chemical properties of key Chl catabolites (or phyllobilins) has provided fundamental, structure-based answers to the question of how Chl is broken down in some higher plants. Indeed, most studies have dealt with the degradation of Chl in angiosperms, [9c] where natural Chl breakdown is inferred to follow the PaO/phyllobilin pathway.I nt his pathway,t he monooxygenase PaO plays akey role,asit cleaves the chlorin macrocycle and generates the red, native "type-I phyllobilin", the barely detectable RCC.Chl breakdown continues rapidly in the chloroplast by furnishing colorless and blue-fluorescent Angewandte Chemie Reviews linear tetrapyrroles (FCCs). In as pecies-dependent way either one of the two epimeric primary FCCs (6/epi-6)i s thereby made.A tt he FCC stage,t he PaO/phyllobilin pathway splits into (at least) two major, and several minor downstream channels ( Figure 30). This part of the catabolic pathway is mostly managed by enzymes associated with the cytosol. In some plants,o ne important second branch produces 1,19-dioxobilin-type (or type-II) phyllobilins by oxidative deformylation of FCCs.C hl breakdown appears to end with the transformation of fluorescent phyllobilins in the acidic vacuole into colorless,a nd essentially photoinactive, nonfluorescent analogues,w hich often accumulate in senescent leaves.However,insome plants,the typically short-lived FCCs are biosynthetically "caged" with complex ester groups and remain fluorescent as persistent hypermodified FCCs (hmFCCs). Theaccumulation of hmFCCs in bananas,observable by their bright blue fluorescence,i sapuzzling and striking phenomenon that may be useful as asignal for fruiteating animals.
Having learned how Chl is broken down to temporarily accumulating colorless catabolites in some higher plants,t he question arises logically:W hati st he fate of the colorless phyllobilins?A re they broken down further in controlled metabolic processes?T he recently discovered endogenous formation of phyllochromobilins may play ap articular role and hold apossible key to help answer this question. Current work in our laboratories is directed at finding evidence for this.
Chl breakdown has,s of ar, been analyzed only in at iny fraction of the vast plant kingdom. In some investigations, naturally senescent leaves of very closely related species,such as the deciduous apple and pear trees [59] as well as plum and apricot trees [68] which are all Rosaceae,w ere analyzed and showed astrongly related pattern of colorless phyllobilins.In other plants investigated, such as A. thaliana and related Brassicaceae,orinthe tropical monocots Musa acuminata and Spathiphyllum wallisii,t he PaO/phyllobilin pathway was discovered to produce other types of phyllobilins.
Our investigations have so far been driven by an interest in discovering the most relevant, basic types of phyllobilins in higher plants,r ather than exploring the plant kingdom in abiologically systematic way.Clearly,itwill be of interest to learn about Chl breakdown in (higher) plants outside the range of the angiosperms.Inthe specific case of the green alga Auxenochlorellaprotothecoides,the early steps of Chl breakdown are remarkably related to those of the PaO/phyllobilin pathway of higher plants. [11a] So far,m ost studies on bilin-type Chl catabolites have relied on their analysis in plant extracts.Clearly,besides such eye-opening in vitro studies,t he observation of Chl catabolites in vivo may provide further insight into the effective location of processes relevant for Chl breakdown in the tissue of al eaf or af ruit. Exploratory methodological mass spectrometric [113] and fluorescence spectroscopic studies [85] of specific Chl catabolites have begun to pave am ore systematic way to in vivo analyses of Chl breakdown.
In ad istantly related context, exploratory studies of Chl breakdown in herbivorous protists and insects were recently reported. Thus,i nt he gut and digestive tract of herbivorous caterpillars,P heo a was detected as the major left-over product of the ingested Chl. [114] Furthermore,i nt he aquatic ecosystem, herbivorous protists were found to degrade (and detoxify) ingested Chl to an onfluorescent cyclopheophorbide a enol. [115] Clearly,the related question is no less intriguing: What happens to Chl in the metabolism of humans and higher animals. [9d, 10] Thed iscovery of hypermodified FCCs is particularly fascinating, [28a] as is the broad occurrence of type-II phyllobilins as products of Chl breakdown. [57,58] These latter 1,19dioxobilins and the heme-derived (hemo-)bilins share substantial similarity,a st hey differ (mostly) only by the Chlderived "extra" ring Eo ft he type-II phyllobilins.I mportant chemical properties of the two lines of bilin-type natural products can be rather similar. Hence,p hyllobilins are candidates for av ariety of important biological roles,w hich are now an apparent domain of the (hemo-)bilins. [116] Hence, physiological roles of phyllobilins,f or example,a se xternal signals, [28a,117] in intra-or intercellular signaling, [108b, 118] or in plant defense,c an not be discounted. [119] In this respect, phyllochromobilins may be particularly interesting experimental targets.I ti st ruly remarkable that biological roles of the phyllobilins still wait to be discovered.
Chl breakdown is the remarkable visual sign of leaf senescence and of the ripening of some fruits.S enescence is typically associated with apoptosis,w hereas ripening is not. Senescence may be caused by endogenous developmental processes [120] or by environmental (stress) factors (such as low temperatures or the lack of light, [121] water, or certain nutrients). [122] As easonal combination of several of these factors may be the underlying cause of the apparently Figure 30. Abbreviated general outline of the PaO/phyllobilin pathway in some higher plants. [10][11] Chl is broken down in al inear sequence from pheophorbide a (Pheo a)toprimary FCCs (pFCCs). Branching out at the stage of FCCs gives several downstream lines of type-I and type-II phyllobilins. Colorless phyllobilins( NCCs and DNCCs) accumulate in senescent and ripening plant tissues. The colorless and nonfluorescent NCCs and DNCCs may be oxidized further to yellow (and then pink) phyllochromobilins.
synchronized Chl breakdown that is observed in leaves in the fall. [101b, 123] Senescence and degreening of leaves may also be induced more locally by external factors (other than the ones listed above), such as mechanical or chemical damages,orby infections with pathogens. [124] Plants may undergo specific responses,depending upon the means by which senescence is induced. [125] Ripening should be seen as ab asically different developmental process that is typically accompanied by Chl breakdown. [11c] Hence,i ti si nteresting to test the profiles of the phyllobilins formed in response to the different mechanisms that induce Chl to disappear.
Products of Chl breakdown accumulate in senescent leaves,i nd eveloping vegetables,a nd in ripening fruit. Their value as health-sustaining nutritional components for leafeating and frugivorous creatures (from insects,snails,birds,to mammals and humans), has hardly been addressed yet. [6b,59, 61, 126] It is tempting, when eating apples,t oc onsider the presence of NCCs in fruit as beneficial to our health as expressed in the old saying:" An apple aday keeps the doctor away". [59] Clearly,o nly the tip of the iceberg of the biological phenomenon of Chl breakdown in higher plants has so far been revealed. [6b,9c, 10, 11] Research on the still-elusive physiological roles of phyllobilins is mandatory,asare studies of the regulation of Chl breakdown in natural (and artificially induced) senescence and ripening,aswell as investigations of the roles and interactions of Chl catabolites during the infection of plants by pathogens.T he impact of the colorless phyllobilins (which are abundant in fruit and vegetables) on our health and wellbeing may be revealed by pharmacological studies.W ithout any doubt, the temporarily abundant remains from natural Chl breakdown leave important traces and have afundamental impact on the biosphere of our Earth. Similar to the heme-derived bilins (the hemobilins), phyllobilins will continue to fascinate,a nd the "Tale of the two Bilins" may take many more years to be told.