Tricin, A Flavonoid Monomer in Monocot Lignification

Tricin was recently discovered in lignin preparations from wheat straw, and subsequently in all monocot samples examined. To provide proof that tricin is involved in lignification and establish the mechanism by which it incorporates into the lignin polymer, the 4'– O – β -coupling products of tricin with the monolignols ( p- coumaryl, coniferyl, and sinapyl alcohols) were synthesized along with the trimer that would result from its 4'– O – β -coupling with sinapyl alcohol and then coniferyl alcohol. Tricin was also found to cross-couple with monolignols to form tricin-(4'– O – β )-linked dimers in biomimetic oxidations using peroxidase/H 2 O 2 or silver (I) oxide. NMR characterization of GPC-fractionated acetylated maize ( Zea mays ) lignin revealed that the tricin moieties are found in even the highest molecular weight fractions, ether-linked to lignin units, demonstrating that tricin is indeed incorporated into the lignin polymer. These findings suggest that tricin is fully compatible with lignification reactions, is an authentic lignin monomer and, because it can only start a lignin chain, functions as a nucleation site for lignification in monocots. This initiation role helps resolve a long-standing dilemma that monocot lignin chains do not appear to be initiated by monolignol homo-dehydrodimerization as they are in dicots that have similar syringyl-guaiacyl compositions. The term flavonolignin is recommended for the racemic oligomers and polymers of monolignols that start from tricin (or incorporate other flavonoids) in the cell wall, in analogy with the existing term flavonolignan that is used for the low molecular mass compounds composed of flavonoid and lignan moieties.


INTRODUCTION
Lignin, a complex phenylpropanoid polymer in the plant cell wall, is predominantly deposited in the cell walls of secondary-thickened cells (Vanholme et al., 2010). It is synthesized via oxidative radical coupling reactions from three prototypical monolignols: p-coumaryl, coniferyl, and sinapyl alcohols, differentiated by their degree of methoxylation ortho to the phenolic hydroxyl group. Considered within the context of the entire polymer, the main structural features of lignin can therefore be defined in terms of its p-hydroxyphenyl (H), guaiacyl (G), and syringyl (S) units, respectively derived from these three monolignols (Ralph, 2010). Several novel monomers, all deriving from the monolignol biosynthetic pathway, have been found to incorporate into lignin in wild-type and transgenic plants.
For example, monolignol acetate, p-hydroxybenzoate, and p-coumarate ester conjugates have all been shown to incorporate into lignin polymers and are the source of naturally acylated lignins (Ralph, 2004;Lu and Ralph, 2008); lignins derived solely from caffeyl alcohol were found in the seed coats of both monocot and dicot plants (Chen et al., 2012;Chen et al., 2013); lignins derived solely from 5hydroxyconiferyl alcohol were found in a cactus seed coat (Chen et al., 2012); a Medicago truncatula transgenic deficient in cinnamyl alcohol dehydrogenase (CAD) exhibited a lignin that was overwhelmingly derived from hydroxycinnamaldehydes (instead of their usual hydroxycinnamyl alcohol analogs) (Zhao et al., 2013); and iso-sinapyl alcohol was implicated as a monomer in caffeic acid O-methyltransferase (COMT) down-regulated switchgrass (Tschaplinski et al., 2012). These findings imply that plants are quite flexible in being able to use a variety of monomers during lignification to form the heterogeneous lignin polymer. Most recently, and as addressed more fully here, the flavonoid tricin has been implicated as a monomer in monocot lignins (del Rio et al., 2012).
Tricin is the first monomer from outside the monolignol biosynthetic pathway to be implicated in lignification.
Tricin (5,7-dihydroxy-2-(4-hydroxy-3,5-dimethoxyphenyl)-4H-chromen-4-one), a member of the In 2012 we reported the first evidence that tricin was incorporated into lignin, as implicated by two previously unassigned correlation peaks at δ C /δ H 94.1/6.56 and 98.8/6.20 in an HSQC NMR spectrum from the whole cell wall and an isolated milled wood lignin of (unacetylated) wheat straw (del Rio et al., 2012). The same evidence has now been found in the HSQC spectra of wheat straw lignin isolated via different methods (Yelle et al., 2013;Zeng et al., 2013). Additional studies have verified the presence of tricin in lignin fractions from a variety of monocots including bamboo (You et al., 2013), coconut coir (Rencoret et al., 2013), maize, and others examined in our laboratories. The implication that tricin is the first phenolic from outside the monolignol biosynthetic pathway found to be integrated into the polymer has prompted further study with the aim of identifying and mechanistically delineating the role of tricin in lignin and its biosynthetic incorporation pathway.
Tricin, unlike the monolignols that derive from the shikimate biosynthetic pathway (Sarkanen and Ludwig, 1971), is derived from a combination of the shikimate and acetate/malonate-derived polyketide pathways (Winkel-Shirley, 2001), as shown in Figure S1. After p-coumaroyl-CoA is synthesized from p-coumaric acid by 4-coumarate:CoA ligase (4CL), it branches from the monolignol biosynthetic route to be transformed via chalcone synthase (CHS) and chalcone isomerase (CHI) into naringenin, the central precursor of most flavonoids. Naringenin is subsequently converted into apigenin by flavone synthase (FNS). Further hydroxylation at C-3' and C-5' followed by Omethylation furnishes tricin (Koes et al., 1994;Winkel-Shirley, 2001). The incorporation of tricin into lignin therefore suggests that an additional biosynthetic pathway, namely the polyketide pathway, may be associated with cell wall lignification in monocots.
The revelation that tricin is incorporated into the lignin polymer was precipitated by closer study of signals found within the NMR spectra of various monocot samples. Before this discovery, tricin had not been noted in any lignin fractions and, although it is reasonable to anticipate compatibility based on its chemical structure, there is no direct and reliable evidence to date showing that tricin is able to react with monolignols through radical coupling; the efficiency and selectivity of the coupling reactions between tricin and various monolignols were also, therefore, unknown. Synthetic model compounds that would facilitate the elucidation of the role of tricin within plant cell walls were desirable as aids to be used in a mechanistic study of "flavonolignin" generation. [We coin the term flavonolignin to describe the racemic oligomers and polymers of monolignols that start from tricin (or other flavonoids) in the cell wall, in analogy with the existing term flavonolignan that is used for the low molecular mass compounds composed of flavonoid and lignan moieties that are presumably made in the cytoplasm (Begum et al., 2010;Niculaes et al., 2014;Dima et al., 2015)].
The overall objective of this study is to demonstrate that tricin incorporates into the lignin polymer of monocots, with maize/corn stover as the representative experimental material. To this end, we have synthesized tricin and various model compounds in which tricin is conjugated to monolignols in the manner expected for the lignification process. Next, we verified whether these synthetic compounds could be made from their assumed precursors under biomimetic radical conditions anticipated for lignification. Subsequently, NMR data generated from these synthetic and biomimetic coupling products were compared with NMR data from native maize stover lignin, including high molecular weight fractions. We conclude that tricin is a monomer in monocot lignification and, because little syringaresinol is found in maize lignin, tricin is functioning as a nucleation site that initiates the lignin polymer chain.

Synthesis of tricin-monolignol cross-coupled oligomers
The synthetic scheme shown in Figure 1 outlines the syntheses of oligomers 14a-c and 19 containing a tricin (T) unit linked to arylglyceryl derivatives. 2,4,6-Trihydroxyacetophenone 1 and 4-hydroxy-3,5-dimethoxybenzaldehyde 2 were selected as the starting materials, and the phenolic hydroxyl groups were appropriately protected as the corresponding MOM (methoxymethyl) ether and benzyl ether, respectively (St. Denis et al., 2010). The flavone base structure was synthesized via chalcone 5, which was made via a Claisen-Schmidt condensation between di-protected triol 3 and benzylated syringaldehyde 4 in KOH/methanol. The initial attempt to synthesize protected tricin proceeded through cyclization of chalcone 5 to provide the flavanone by first refluxing in ethanol with sodium acetate, followed by several attempts at oxidative dehydrogenation using a variety of reagents; 2,3dichloro-5,6-dicyano-1,4-benzoquinone (DDQ) (Shanker et al., 1983), manganese (III) acetate (Singh et al., 2005), and copper (II) acetate were all tried without success. Successful cyclization of chalcone 5 used a molar equivalent of I 2 in pyridine (Lin et al., 2007). Acetylation of the crude products (not shown in Figure 1) was performed to facilitate the isolation and purification of the target compound by chromatography. The desired flavonoid product 6 was obtained in ~20% yield from compound 5, O-β)-G/S/H (the G, S, or H moiety in the dimer depending on whether the formal monolignol involved is coniferyl, sinapyl, or p-coumaryl alcohol, respectively)] followed traditional β -ether dimer synthetic methods (Kratzl et al., 1959;Ralph et al., 1992), starting with the standard S N 2 reaction between brominated 4-acetoxyacetophenone derivatives 10a-c and mono-protected tricin 7, followed by addition of formaldehyde and final deprotection (removal of the MOM and acetyl groups). After reduction of the carbonyl group at C α in 13a-c by NaBH 4 in ethanol, each compound 14a-c contained two chiral centers, and was therefore a mixture of two diastereomers [anti (or erythro) and syn (or threo)], as was readily seen in the NMR spectra. The anti-diastereomer was the major product according to the Felkin-Anh model (Paddon-Row et al., 1982;Lodge and Heathcock, 1987), with an anti/syn ratio of about 80:20, as determined by relative integration of the β -or γ -protons. As noted previously for such β -ethers, the syn-isomer has the highest field (lowest δ ) γ -proton (Ralph and Wilkins, 1985;Ralph and Helm, 1991;Ralph, 1993). The anti and syn structure assignments were further confirmed by the magnitude of coupling constant 3 J α -H,β-H : the doublet with the larger coupling constant of 7.0 Hz indicated the anti-isomer, whereas the doublet with the smaller one (4.8 Hz) corresponded to the syn-isomer (Ralph, 1993;Bouaziz et al., 2002). In the obtained HSQC spectra, correlation signals belonging to tricin unit C3/H3, C6/H6, and C8/H8 were well dispersed from the lignin signals. The type of monolignol linked to tricin at the 4'-OH did not affect the location of these three C-H correlation peaks; i.e., they were invariant for 14a-c and 19. All of these correlations, however, shifted in predictable ways upon acetylation. The chemical shifts (before and after acetylation) in various NMR solvents are listed in Table 1. Trimer 19, T-(4'-O-β)-S-(4-O-β)-G, was synthesized from 12b via a similar pathway ( Figure 1). There are a total of 8 possible diastereomers (2 4 = 16 optical isomers) of 19 because this trimer possesses four chiral carbons; LC-MS separated only four components. High-resolution 1 H NMR of the isomer mixture also readily distinguished four isomers designated as anti-syn (anti structure in the internal polymer chains serve to connect them, thus increasing the polymer size and, as previously thought but recently questioned (Ralph et al., 2008;Crestini et al., 2011), causing the polymer to branch. The peroxidases/H 2 O 2 system, a two-step one-electron transfer system, is therefore commonly used as a biomimetic system for the preparation of dimeric lignin model compounds or dehydrogenation polymers. Oxidative radical coupling using Ag 2 O as the one-electron oxidant is another convenient approach to the synthesis of lignin model compounds (Zanarotti, 1985;Quideau and Ralph, 1994a).
Herein we applied both methods to determine whether tricin is capable of reacting with monolignols under radical coupling conditions and to elucidate the nature of the resulting products. Peroxidasecatalyzed reactions were carried out using horseradish peroxidase in acetone/aqueous buffer (v/v, 2:3), which contained a larger proportion of acetone than is typically used in this procedure.
Furthermore, due to the poor solubility of tricin in this solvent system, a larger overall volume of solvent (500 mL) was needed to completely dissolve ~25 mg tricin along with the monolignols. A reaction time of two hours, longer than the time used in a prior study (Zhang et al., 2009), was then used to ensure complete reaction at this lower concentration of reactants. Acetone was used as the solvent for the Ag 2 O-catalyzed reaction because it has been shown to give the highest yield of β -O-4-structures in the dimerization of coniferyl alcohol (Quideau and Ralph, 1994b). NMR data acquired from synthetic model compounds 14a-c were used for qualitative analysis of the products of tricinmonolignol cross-coupling. Additionally, the synthesized model compounds were purified by semipreparative HPLC and used to produce standard curves for measuring the yields of T-(4'-O-β)-G/S/H products resulting from the cross-coupling reactions.
When tricin and a monolignol were oxidized by peroxidase/H 2 O 2 , various coupling products were formed, as evidenced by NMR spectra of the total crude products. Generally, most of the tricin remained unreacted, but no monolignol remained in the mixture after reaction, as indicated by the strong signals of C3/H3, C6/H6, and C8/H8 from tricin, and the disappearance of C7/H7 and C8/H8 correlations at δ C /δ H ~130.0/6.50 and ~128.0/6.21 from the side-chain double bonds of the monolignols. Reaction between coniferyl alcohol and tricin generated a low yield of a cross-coupled product, compound 14a, as demonstrated by the correlation peaks at coumaryl alcohol under peroxidase/H 2 O 2 coupling conditions occurred in relatively high yield, with only a small amount of tricin remaining, as was confirmed by comparing the major dimeric product with the synthesized authentic compound. However, compound 14b, T-(4'-O-β)-S, was barely detected in the HSQC spectrum from the tricin-sinapyl alcohol coupling mixture. The major product in this reaction was, as in the reaction with sinapyl alcohol alone, syringaresinol, the β -βhomocoupled dimer.
Silver(I) oxide, Ag 2 O, was also used as an oxidant to invoke cross-coupling between tricin and monolignols. If monolignols were added in a single shot, no T-(4'-O-β)-monolignol cross-coupling adducts were detected by NMR. In contrast, compounds 14a-c were successfully obtained by slow addition of each respective monolignol into the tricin solution containing Ag 2 O; this sequence is viable because tricin is inert and does not undergo homo-dimerization under these conditions.
Logically, the reactivity of tricin in cross-coupling reactions with a monolignol is lower than that of simple dimerization of the monolignol. Therefore, as seen in the results described earlier, in the presence of equivalent quantities of tricin and a given monolignol, the monolignol will predominantly undergo homocoupling. However, if the concentration of monolignols is limited, the occurrence of cross-coupling is enhanced. identical m/z values, were obtained from the anti-diastereomers of compounds 14a-c. The yields of cross-coupling products were determined by first combining the peak areas of the peaks from both isomers, the value of which was then used for quantification based on the standard curve. Radical coupling between tricin and monolignols effected by peroxidase/H 2 O 2 produced compounds 14a-c in 12.6, 1.2 and 51.4% yields; when Ag 2 O was used as the oxidant, the yields were 6.9, 15.3, and 10.2%.
Previous studies have revealed that the oxidation rate of sinapyl alcohol itself by horseradish peroxidase is quite low (Takahama et al., 1996). Transfer of the radical from p-coumarate to this alcohol may aid in the formation of syringyl-rich lignin in some plant species (Grabber, 2005;Hatfield et al., 2008;Ralph, 2010). Without the use of radical transfer agents (or a direct oxidant like Ag 2 O), a much lower yield for 14b than for 14a and 14c would therefore be expected from radical coupling using peroxidase as the oxidant.   Figure   3) containing a flavonoid moiety (C 3 ) connected to a monolignol-derived unit (C β ) that has the appearance of deriving from 3-β-cross-coupling was isolated from Avena sativa and Hydnocarpus wightiana (Parthasarathy et al., 1979;Wenzig et al., 2005). This type of structure could not be identified among the radical coupling products examined in this study, either by NMR or LC-MS analysis. Furthermore, a reasonable reaction mechanism for the formation of compound 20a following the coupling of tricin and coniferyl alcohol is not obvious; an earlier intermediate in the biosynthetic pathway may react to form this adduct. No evidence has yet been found to suggest that linkages between monolignols and tricin occur at the 5-O-, 7-O-, 3-C-, 6-C-, or 8-C-positions under biomimetic lignification conditions. Instead, in this study, we prove the incorporation of tricin into lignin via 4'-O-β-coupling with monolignols, as might be anticipated. To determine whether tricin in the lignin polymer has linkages to carbohydrates will require further investigation.

Evidence for the incorporation of tricin into the lignin polymer
The objective of this study was to experimentally support the initial claims made regarding the existence of tricin in the milled wood lignin from wheat straw (del Rio et al., 2012) synthetic and authenticated model compounds (before and after acetylation) were acquired and used for comparison with tricin-containing moieties in maize lignin. As noted above, results from biomimetic radical coupling reactions between tricin and monolignols suggest that tricin is compatible with lignification. Therefore, lignin preparations extracted by acetic acid pretreatment of maize stover (Pan andSano, 1999, 2005)  lignin. To further elucidate whether tricin is incorporated into high molecular mass lignin chains, rather than simply being bonded to monolignols to form dimers or short-chain oligomers, the acetylated maize stover lignin was fractionated via GPC. Eight fractions were collected, with the first two fractions containing high molecular weight components (M w =5670, M n =1580 for the first fraction, M w =2440, M n =970 for the second fraction) accounting for 73% of the sample. Based on NMR characterization, the first four fractions with large to medium molecular weight components all contained covalently bonded tricin. The HSQC spectrum of the highest molecular weight fraction is shown in Figure 4A.

Tricin initiates lignin chains
To date, the accumulated evidence has indicated that tricin is only incorporated into the polymer tricin 4'-O-5-coupled units can not arise from the coupling of a tricin (radical) with a lignin oligomer (radical), nor that units such as compound 20 will not be found in the lignin. Theoretically, 3-coupled products (like β -coupled products from normal monolignols) are possible but they have not been evidenced here in either the biomimetic coupling reactions or in the polymers of the natural samples we examined, notwithstanding the report of compound 20, Fig. 3, in the literature (Parthasarathy et al., 1979;Wenzig et al., 2005). All of these theoretical products would have very different NMR characteristics from those noted here. We therefore deduce that tricin predominantly incorporates into lignin via 4'-O-β-coupling, which by necessity localizes each tricin unit at one terminus of its lignin chain, and that terminus must be at the starting end of that chain. Tricin therefore acts as a nucleation site for lignin chain growth in monocots, a role that has previously been proposed for ferulate on arabinoxylans (Ralph et al., 1995).

Resolution of a monocot lignin dilemma
The observation that tricin may be the initiator of many of the lignin chains in the polymer helps to explain an old dilemma arising from many maize lignin spectra: i.e., that, despite its being an S-G lignin with an S:G ratio similar to those found in many dicots and hardwoods, there is little or no evidence for (syringa)resinol structures in maize lignin. That maize lignin has essentially no resinol structures, the correlation positions for which are indicated by the magenta-colored dashed ellipses, can be seen in Figure 4. This means that the polymer chain is not (significantly) started via monolignol dimerization per se (as it is in hardwoods/dicots and softwoods). The data here provide the two-part explanation for the first time. In large measure, the lack of resinols starting polymerization is because the chain is initiated/nucleated by tricin in monocots. Such nucleation behavior has previously been attributed to ferulates on arabinoxylans and this may also be present here -it is hard to observe or quantify (Ralph et al., 1995). However, careful examination of the NMR spectra reveals that monomer dimerization is in fact occurring, but the monomers in this case are acylated monolignols. Maize lignin is γ -acylated (by p-coumarate or acetate) (Ralph, 2010). Such acylated monomers cannot cyclize after β -β-coupling to give resinol structures (Lu and Ralph, 2008;Ralph, 2010). In fact, the β -β-product arising from the coupling of two acylated monolignols, the tetrahydrofuran C' in Figure 4, is readily seen in the side-chain region of the HSQC spectra, Figure 4. At some point it will be intriguing to understand how and why the dimerization reactions are dominated by the acylated monolignols rather than the parent monolignols themselves, but this is not the primary concern here. The observation that syringaresinol is found in single-shot coupling reactions with tricin and sinapyl alcohol, but is less prominent when sinapyl alcohol is slowly added to the tricin solution, further supports the theory that lignification is an endwise process. confidently be asserted that tricin is not only able to couple with monolignols and participate in lignification, but that it regularly does so in monocots, where tricin is found covalently bound into the very lignin polymer itself. In addition, as the tricin that is observed in lignin NMR spectra can only arise from the participation of tricin in the initial coupling reactions with a monolignol, it must be placed at the beginning of a polymer chain, thus acting as an "initiator" of sorts. The prevalence of tricin in these lignins strongly suggests that tricin units have a role in nucleating the growth of the lignin polymer in monocots. We have also resolved the dilemma of the almost complete absence of syringaresinol units (the dimer that typically starts a lignin chain in pure S-G lignins) in maize and other monocot lignins; the chains are either started by tricin, or by dimerization of acylated monolignols that give rise to novel β -β-linked dimers that are readily seen in the spectra of the maize lignin here.

CONCLUSIONS
Taken together, these findings provide reliable and substantive evidence that tricin is incorporated into maize (and other monocot) lignins via a free radical coupling mechanism and is covalently bound into the lignin polymer, i.e., that tricin should be regarded, in a general sense, as an authentic lignin monomer in monocots. This study therefore not only supports the striking observation that monocots routinely incorporate a flavonoid (derived from an entirely different biosynthetic pathway) into their lignins, but it also serves to highlight the remarkable ways in which lignification in monocots differs from the process in other plant classes.

General
All chemicals and solvents used in this study were purchased from commercial sources and used without further purification. Horseradish peroxidase (type II, 180 pyrogallol units/mg) was provided by Sigma (Ronkonkkoma, NY).
Maize (Zea mays) stover lignin was obtained from acetic acid pretreated (Pan andSano, 1999, 2005)  Flash chromatography was performed with Biotage snap silica cartridges on an Isolera One instrument (Biotage, Charlottesville, VA) using a hexane/ethyl acetate gradient as the eluent.
Preparative thin-layer chromatography plates (1 or 2 mm thickness, normal-phase) were purchased from Analtech (Newark, DE) and were run using hexane/ethyl acetate or methanol/dichloromethane as the eluent. NMR spectra were recorded on a Bruker Biospin (Billerica, MA) AVANCE 500 MHz or 700 MHz spectrometer fitted with a cryogenically cooled 5 mm TCI (500 MHz) or TXI (700 MHz) gradient probe with inverse geometry (proton coil closest to the sample). Bruker's Topspin 3.1 (Mac) software was used to process spectra. The central solvent peaks were used as internal references A total of 8 fractions were collected, out to 180 min.

Synthesis of oligomers
Procedure for acetylation: The starting material was dissolved in pyridine/acetic anhydride (v/v, 2:1) and stirred for 2 h at room temperature. The solution was transferred to a separatory funnel and extracted with ethyl acetate (EtOAc) and washed several times with acidic water to eliminate most of the pyridine. The organic phase was washed with saturated ammonium chloride solution (NH 4 Cl), dried over anhydrous magnesium sulfate (MgSO 4 ), filtered, and evaporated under reduced pressure to give the acetylated products. The yield ranged from 92-96%.
Peroxidase and Ag 2 O were used as oxidative reagents for the radical coupling reactions in this study. The detailed procedure is illustrated using coniferyl alcohol. Tricin (15.7 mg,47.5 μ mol) and coniferyl alcohol (8.6 mg, 47.5 μ mol) were dissolved in 200 mL acetone/phosphate buffer (5.00 pH, 20 mM, v/v, 2:3). Hydrogen peroxide-urea complex (4.9 mg, 52.1 μ mol) and peroxidase (0.5 mg) were added. The reaction solution was stirred at room temperature for 2 h. After acidification to pH 3.0 with 1 M HCl, the reaction mixture was placed in a hood to allow the acetone to evaporate until the radical coupling compounds precipitated, EtOAc (3 × 100 mL) was added to extract the products.
After separation, drying over anhydrous MgSO 4 , filtration, and evaporation, a mixture of coupling products was obtained (86.4% overall yield). The radical coupling reactions of tricin/sinapyl alcohol (87% overall yield) and tricin/p-coumaryl alcohol (90% overall yield) were carried out via the same method. The oxidation reaction using silver (I) oxide (Ag 2 O) was carried out according to the method used in a previous study (Quideau and Ralph, 1994a) with only slight modifications. Tricin (15.7 mg,47.5 μ mol) was dissolved in acetone (5 mL), and Ag 2 O (16.5 mg, 71.3 μ mol) was added. Coniferyl alcohol (8.6 mg, 47.5 μ mol) dissolved in acetone (30 mL) was added dropwise via an addition funnel over 5 h, and the resulting mixture allowed to stir overnight. The reaction was quenched with 1 M HCl (2 mL). The inorganics were filtered off, and the filtrate was collected for evaporation to obtain the crude product mixture (99.0% overall yield). The same method was applied to the tricin/sinapyl alcohol (99.5% overall yield) and tricin/p-coumaryl alcohol (92.5% overall yield) adducts. Yields of the desired tricin-monolignol coupling products 14a-c [T-(4'-O-β)-G/S/H] were 12.6, 1.2, and 51.4% under peroxidase-catalyzed oxidation conditions and 6.9, 15.3, and 10.2% under Ag 2 O oxidation.

SUPPORTING INFORMATION
Additional Supporting Information may be found in the online version of this article.            19