Insights into lignin primary structure and deconstruction from Arabidopsis thaliana COMT (caffeic acid O-methyl transferase) mutant Atomt1

View Article Online / Journal Homepage / Table of Contents for this issue

PAPER

www.rsc.org/obc | Organic & Biomolecular Chemistry

Insights into lignin primary structure and deconstruction from Arabidopsis

thaliana COMT (caffeic acid O-methyl transferase) mutant Atomt1†

Syed G. A. Moinuddin,‡ Michae¨l Jourdes,‡ Dhrubojyoti D. Laskar, Chanyoung Ki, Claudia L. Cardenas,

Kye-Won Kim, Dianzhong Zhang, Laurence B. Davin and Norman G. Lewis*

Received 31st March 2010, Accepted 19th May 2010

First published as an Advance Article on the web 22nd July 2010

DOI: 10.1039/c004817h

The Arabidopsis mutant Atomt1 lignin differs from native lignin in wild type plants, in terms of sinapyl

(S) alcohol-derived substructures in ﬁber cell walls being substituted by 5-hydroxyconiferyl alcohol

(5OHG)-derived moieties. During programmed lignin assembly, these engender formation of

benzodioxane substructures due to intramolecular cyclization of their quinone methides that are

transiently formed following 8-O-4¢ radical-radical coupling. Thioacidolytic cleavage of the 8-O-4¢

inter-unit linkages in the Atomt1 mutant, relative to the wild type, indicated that cleavable sinapyl (S)

and coniferyl (G) alcohol-derived monomeric moieties were stoichiometrically reduced by a circa 2 : 1

ratio. Additionally, lignin degradative analysis resulted in release of a 5OHG–5OHG–G trimer from the

Atomt1 mutant, which then underwent further cleavage. Signiﬁcantly, the trimeric moiety released

provides new insight into lignin primary structure: during polymer assembly, the ﬁrst 5OHG moiety is

linked via a C8–O–X inter-unit linkage, whereas subsequent addition of monomers apparently involves

sequential addition of 5OHG and G moieties to the growing chain in a 2 : 1 overall stoichiometry. This

quantiﬁcation data thus provides further insight into how inter-unit linkage frequencies in native lignins

are apparently conserved (or near conserved) during assembly in both instances, as well as providing

additional impetus to resolve how the overall question of lignin macromolecular assembly is controlled

in terms of both type of monomer addition and primary sequence.

in tobacco clearly established that it was instead involved in

the second methylation step leading to sinapyl alcohol (3) and

Introduction

S-moieties of lignin.⁹This was conclusively demonstrated by

massive reduction of S-lignin deposition in these transformants,

as shown by thioacidolysis release of monomers 19 and 21 (S/G

ratios of ~1 for wild type (WT) and 0.07–0.09 in most COMT

I down-regulated plants), with these resulting from presumed 8–

O–4¢ interunit linkage cleavage. There were also no signiﬁcant

differences in estimated Klason lignin contents for WT (~18.4%

cell wall residue, CWR) and the COMT down-regulated (~18.3%

CWR) lines. Thus, COMT was responsible for the second methyla-

tion step and was, therefore, effectively mono-functional in terms

of substrate utilized in vivo.

The lignin pathway in angiosperms results in formation of the three

main monomeric lignin precursors, p-coumaryl (1), coniferyl (2)

and sinapyl (3) alcohols, which lead to the H (p-hydroxyphenyl), G

(guaiacyl), and S (syringyl) components of lignins in their vascular

cell walls, respectively.^1,2The S-component is, however, mainly, if

not exclusively, deposited in ﬁber cell walls in plant species such

as Arabidopsis.³It was initially considered that for formation of

the S and G precursors, both methylation steps were catalyzed by

a single enzyme, the so-called caffeic acid O-methyl transferase

(COMT).^4,5This was due to its ability to methylate both caffeic

(5) and 5-hydroxyferulic (6) acids in vitro to afford ferulic (12) and

sinapic (13) acids, respectively (Fig. 1). Accordingly, COMT was

considered to be substrate versatile, and responsible for producing

both G and S lignins derived from coniferyl (2) and sinapyl (3)

alcohols in angiosperms.^4,5

In subsequent pioneering work from the Legrand/Fritig lab-

oratory, various putative COMT isoforms (I, II and III) were

puriﬁed from tobacco (Nicotiana tabacum)⁶and their correspond-

ing cDNAs cloned.^7,8In contrast to long held assumptions of

COMT being substrate versatile, down-regulation of COMT I

Institute of Biological Chemistry, Washington State University, Pullman,

WA 99164-6340, USA. E-mail: lewisn@wsu.edu.; Fax: +1 509 335 8206;

Tel: +1 509 335 2682

† This work was previously reported in oral presentations at the annual

retreat of the BioEnergy Science Center (June 2009, Chattanooga, TN) and

at the Phytochemical Society of North America (PSNA) annual meeting

(August 2009, Towson, MD).

‡ Both contributed equally to the work, with S. G. A. Moinuddin

completing all the synthetic chemistry.

Yet just before this seminal study, reports from the Chiang¹⁰

and Dixon¹¹laboratories had concluded that heterologous

3928 | Org. Biomol. Chem., 2010, 8, 3928–3946

This journal is

The Royal Society of Chemistry 2010

©

View Article Online

Fig. 1 In vitro substrate versatile COMT-catalyzed conversions.

down-regulation of COMT in tobacco using aspen (Populus

tremuloides) and alfalfa (Medicago sativa) COMTs, respectively,

had no signiﬁcant effect on the G/S ratios of the lignins.

Furthermore, down-regulation of the alfalfa COMT in tobacco¹¹

gave transformants whose overall putative lignin contents were re-

portedly reduced by approximately 2/3 relative to WT, whereas the

transformants resulting from down-regulation of aspen COMT in

tobacco¹⁰had no apparent changes in lignin contents. However,

determinations in the ﬁrst of these studies were based on the

unreliable thioglycolic acid lignin methods (discussed in Anterola

and Lewis),¹²as well as in the use of an alkaline nitrobenzene

protocol which can release p-hydroxybenzaldehydes (e.g., 22, 23

and 25) and p-hydroxybenzoic acids (e.g., 26, 27 and 29) from

various phenylpropanoids, including those from lignins. Both

studies had thus incorrectly supported the earlier view that COMT

was substrate versatile in vivo.

Earlier studies, undertaken nearly eight decades ago, had also

described various brown-midrib (bm1–bm4) mutants of maize (Zea

mays) with putative colored lignins in cell walls of leaf midribs

and sclerenchyma cells surrounding the vascular bundles of the

stems.^21,22Of these, the so-called bm₃mutant had a putative

lower (“acid detergent”) lignin content as estimated by the

potassium permanganate method,²³as well as lower stalk strength

as suggested by 17–26% less crushing strength and 8–14% less

stalk-section weight than WT plants.²⁴Later, it was reported that

when all then known monolignol pathway enzymes were assayed

in the bm₃mutant in vitro, apparently only the activity of the

COMT purportedly catalyzing methylation of caffeic acid (5) was

strongly reduced.²⁵Additionally, a later study of the bm₃mutant

line by Lapierre et al.,²⁶employing thioacidolytic treatment of

the corresponding stem tissue, resulted in a reportedly reduced

level of releasable S-monomers (by a factor of ca. 4.4) relative to

WT, together with an unidentiﬁed, putative, 5-hydroxyguaiacyl

(5OHG) derived monomeric component (representing ~7% of

total monomeric units released). For the WT line, the recovery

of monomeric units 18, 19 and 21 relative to estimated Klason

lignin content was quite low (~20.2%), with an H : G : S ratio =

1 : 9.3 : 16, whereas the yield of estimated monomers 18, 19 and 21

was ~16.1% in the mutant line, with a putative H : G : S : 5OHG

ratio = 1 : 16 : 6.2 : 1.7. The data also suggested a small reduction

in estimated lignin contents (Klason lignin: WT = 15.2%, bm₃=

11.3%), as well as an overall reduction in S-lignin deposition.

However, given the variability in the data, the study did not

conclusively ascertain that the effect of this mutation was only

on S-lignin deposition.

The overall purpose of this investigation was thus to begin to

gain urgently needed insight into lignin primary structure using the

Arabidopsis COMT1 (At5g54160) knockout mutant line Atomt1.

Accordingly, an Atomt1 lignin-derived isolate was subjected

to non-exhaustive thioacidolytic cleavage in order to begin to

precisely identify the substructure(s) resulting from coupling of

the putative 5OHG units in the lignin primary structure, such

as benzodioxanes (Fig. 2), and thus the effect on overall lignin

polymer assembly (lignin amounts/monomeric compositions).

This treatment resulted in release of cleavable benzodioxane

substructures, a 5OHG–5OHG–G trimer and a 5OHG–G dimer

from the Atomt1 lignin isolate, with the latter presumed generated

through cleavage of the corresponding trimer. Signiﬁcantly, the

The pioneering work by Atanassova et al.,⁹establishing COMT

as involved in the second methylation step in vivo leading to the

S-component of its lignin, was later conﬁrmed using poplar,^13–15

aspen,¹⁶and alfalfa.¹⁷Accordingly, confusion over COMT being

bi-functional/substrate versatile in vivo for a period encompassing

nearly two decades had been unambiguously resolved. Inter-

estingly, other in vitro studies using heterologously expressed

COMTs from tobacco,¹⁸aspen,¹⁹and alfalfa²⁰in Escherichia

coli later showed they were also able to methylate caffeic (5)/5-

hydroxyferulic (6) acids, caffeoyl (7)/5-hydroxyferuloyl (8) CoAs,

caffeyl (9)/5-hydroxyconiferyl (10) aldehydes or caffeyl (11)/5-

hydroxyconiferyl (4) alcohols. This once again illustrates dif-

ferences that can result between in vitro assays and the much

more metabolically relevant biochemical processes in vivo due to

substrate/enzyme compartmentalization in the latter.

This journal is

The Royal Society of Chemistry 2010

Org. Biomol. Chem., 2010, 8, 3928–3946 | 3929

©

View Article Online

Fig. 2 Potential 8–O–4¢ (30) and benzodioxane (31–34) lignin subunits in Arabidopsis via (a) H, G and S monolignol coupling and (b) H, G and S

monolignol coupling as above, but with 5OHG for formation of substituted benzodioxanes (31–34).

G unit placement in the trimer/dimer isolated was either that of a

terminal unit in this segment of the lignin biopolymer, or perhaps

more likely as a 5OHG–5OHG–G repeating unit.

This study thus now clariﬁes further how the programmed

ligniﬁcation response is predictably modulated when COMT

is either down-regulated or “knocked-out”, while apparently

nearly conserving overall 8–O–4¢ inter-unit linkage frequency, as

well as that of other inter-unit linkages. This limited substrate

degeneracy/polymerization process also helps provide further

insight into why, from an evolutionary perspective, this mutation

was unfavorable throughout the plant kingdom. Additionally, of

the 17 putative COMT multigene members in Arabidopsis, 12

were heterologously expressed in Escherichia coli with only one

being able to O-methylate caffeic (5)/5-hydroxyferulic (6) acid,

caffeyl (9)/5-hydroxyconiferyl (10) aldehyde and caffeyl (11)/5-

hydroxyconiferyl (4) alcohol in vitro (Fig. 1). Taken together, these

ﬁndings place further constraints on ligniﬁcation in vivo in general,

and further underscore the need to develop new methods to fully

establish lignin primary structure(s).

Fig. 3 Plots of growth/development parameters between Arabidopsis WT

(black squares) and Atomt1-knockout (grey circles) lines until senescence;

(a) stem lengths and (b) basal stem widths. [All measurements are an

average of 20 plants per sampling point].

Results and discussion

Growth characteristics and quantiﬁcation of thioacidolytically

released lignin-derived monomeric entities

noted between both lines (WT and Atomt1) in terms of overall

gross growth/development indices of the stems from rosette leaf

Prior to investigating the nature of the lignin formed in the cell-wall

stem tissues of A. thaliana WT (ecotype Wassilewskija) and the

corresponding Atomt1 knock-out lines, a comparison of various

gross growth characteristics of their developing stems at different

time intervals was carried out. Sampling of plant stem tissues

was conducted from germination until maturation/senescence,

with particular attention being directed to the vascular appa-

ratus. The plots of overall stem lengths and basal stem widths

during growth and development until senescence are depicted

in Fig. 3. There were no readily visible phenotypical differences

development stage (3 weeks) until maturation/senescence (10

weeks).

It was next considered instructive to plot the amounts of

cleavable lignin 8–O–4¢ interunit linkages, in terms of release

of their G and S thioacidolysis-derived monomeric moieties 19

and 21, relative to the putative lignin contents of both lines

as a function of growth/development.^1,2,27For the WT line,

the ꢀ, ꢀ and ꢀ values in Fig. 4a depict the monomeric S,

G, and G+S amounts released (mmol g^-1putative AcBr lignin

content) at weekly intervals. By contrast, for the Atomt1 line, the

3930 | Org. Biomol. Chem., 2010, 8, 3928–3946

This journal is

The Royal Society of Chemistry 2010

©

View Article Online

These data could thus be hypothetically interpreted as follows:

(i) 5-hydroxyconiferyl alcohol (4) moieties presumed present in the

Atomt1 mutant were, at least in part, covalently linked to adjoining

G residues in the benzodioxane substructure(s) of the cell-wall

lignin, in the ratio of 2 : 1 (Fig. 4a), and (ii) since overall levels of

ligniﬁcation had not, in fact, been signiﬁcantly altered in terms of

amounts in either line, the overall assembly process had apparently

only been modulated due to limited substrate degeneracy during

programmed (template) polymerization.^28–30

Note also for both lines (Fig. 4b), the plot of monomeric

moieties released relative to putative AcBr lignin contents again

established that these data points do not intersect at the origin

of the X and Y axes, as previously noted.^1–3,27,31In other work,

we have demonstrated that some of the putative ‘lignin’ at the

early stages of plant growth and development contain other non-

lignin moieties which interfere with these estimations (manuscript

in preparation).

NMR spectroscopic analyses of lignin-derived-isolates of WT and

Atomt1 lines

In order to examine the possible formation of benzodioxane

substructure VI (Fig. 5c) during programmed ligniﬁcation in the

Atomt1 mutant line, the lignin-enriched isolates from the WT and

Atomt1 lines were next obtained following published procedures.²⁷

Each was then individually subjected to ¹³C, 2D heteronuclear

multiple quantum coherence (HMQC) and 2D heteronuclear

multiple bond coherence (HMBC) NMR spectroscopic analyses.

In this regard, analyses of the oxygenated carbon resonances

in the aliphatic region (50–90 ppm) conﬁrmed that the lignin-

isolates of the WT line (Fig. 5a) contained the ﬁve expected

lignin substructures I–V, namely: 8–O–4¢ aryl ether (substructure

I), resinol-like (substructure II), phenylcoumaran (substructure

III), dibenzodioxocin (substructure IV) and cinnamyl alcohol end

groups (substructure V) (Fig. 5c).

Fig. 4 (a) Plots of G and S thioacidolysis monomers (as mmol g^-1AcBr

lignin) released from both Arabidopsis WT and Atomt1 mutant lines, at

^{different stages of growth and development. WT:}ꢀ, S-derived monomer

^(21);ꢀ, G-derived monomer (19); ꢀ, G+S derived monomers (19)/(21).

Atomt1: ᭢, S-derived monomer (21); ᭜, G-derived monomer (19); ᭹

hypothetical inclusion of S : G monomers in 2 : 1 ratio. (b) Correlations

of estimates of amounts of total monomeric G+S derivatives releasable by

thioacidolysis as a function of estimated AcBr lignin contents; WT = dark

symbols, Atomt1, open circles.

The substructures in the WT lignin-enriched isolates (Fig. 5a)

were also individually conﬁrmed by analysis of their characteristic

one-bond d_C/d_Hcorrelations: 60.3/3.19 and 60.3/3.56 (C-9/9-

H), 84.1/4.20 (C-8/8-H) and 71.7/4.68 (C-7/7-H) ppm for

substructure I; 71.8/3.78 and 71.8/4.11 (C-9/9-H, C-9¢/9¢-H),

54.3/3.01 (C-8/8-H, C-8¢/8¢-H) and 85.7/4.59 (C-7/7-H, C-

7¢/7¢-H) ppm for substructure II; 63.4/3.69 and 63.4/3.45 (C-

9/9-H), 53.8/3.38 (C-8/8-H) and 87.6/5.42 (C-7/7-H) ppm, for

substructure III; 62.3/4.08 and 62.3/3.95 (C-9/9-H), 86.6/4.08

(C-8/8-H) and 84.2/4.69 (C-7/7-H) ppm for substructure IV, as

well as 63.1/4.13 (C-9/9-H) ppm for cinnamyl alcohol end groups

(V). Other inter-unit linkages, such as 8–1¢, diphenyl and/or

diphenyl ether bonds, were not further characterized individually

as earlier^27,32due to overlapping resonances, although they are

also present. However, there was no detection of any cross-peaks

consistent with substructure VI (down to baseline) under the

conditions employed for analysis of the WT lignin-isolate.

The carbon–proton correlations in the 2D-HMQC spectra of

the lignin-enriched isolates from the Atomt1 mutant line also

established the presence of the ﬁve expected lignin substructures I–

V (Fig. 5b). These were individually detected by their characteristic

cross-peak d_C/d_Hcorrelations as follows: 59.9/3.12 and 59.9/3.47

(C-9/9-H), 84.2/4.19 (C-8/8-H) and 71.2/4.65 (C-7/7-H) ppm

for 8–O–4¢ aryl ether substructure I; 71.5/3.65 and 71.5/4.06

S-component (᭢) was essentially undetectable as anticipated, with

the G-amounts released (᭜) also being ca. 20% lower than that of

^{WT G levels (}ꢀ) at maturity. The estimated AcBr lignin contents

of stems of both WT and Atomt1 lines from the earliest stages of

stem development/elongation (circa 3 weeks) until maturity were

also plotted relative to releasable monomeric G and S moieties

(Fig. 4b). In the WT line, after circa 6% “AcBr lignin” deposition,

there is a linear increase in amounts of G and S monomers

(19 and 21) released until maturation (highest lignin content).

For the Atomt1 line, there is also a linear increase in releasable

G monomers (19) from early stages of stem development until

maturation, with this again increasing proportionally relative

to putative lignin level. It was useful at this stage of enquiry

to attempt to correlate, for each time point, the differences in

monomer release between the WT and Atomt1 lines. Provisionally,

this corresponded to that of the reductions in both S amount

released, as well as the reduction in G values between the WT

(G) and Atomt1 (G) lines. For each time point, the reduction in

monomer release corresponded to a S : G reduction in a 2 : 1 ratio.

When hypothetically factoring in this stoichiometric correction

(᭹, Fig. 4a) for each stage of growth/development, the corre-

sponding plot became essentially superimposable over that of WT

(G+S).

This journal is

The Royal Society of Chemistry 2010

Org. Biomol. Chem., 2010, 8, 3928–3946 | 3931

©

View Article Online

Fig. 5 Partial NMR spectra of lignin-enriched isolates from Arabidopsis WT and Atomt1 lines. 2D HMQC spectra of oxygenated aliphatic region of

lignin-enriched isolate from (a) WT and (b) Atomt1. (c) Lignin substructures I–VI assignable by NMR spectroscopic analyses.

(C-9/9-H, C-9¢/9¢-H), 53.4/2.84 (C-8/8-H, C-8¢/8¢-H) and

85.4.6/4.54 (C-7/7-H, C-7¢/7¢-H) ppm for resinol-like substruc-

ture II; 62.8/3.68 and 62.8/3.39 (C-9/9-H), 52.9/3.32 (C-8/8-H)

and 87.5/5.41 (C-7/7-H) ppm for phenylcoumaran substructure

III; 61.7/4.03 and 61.7/3.92 (C-9/9-H) and 84.1/4.67 (C-7/7-H)

ppm for dibenzodioxocin substructure IV, and 62.6/4.08 (C-9/9-

H) ppm for cinnamyl alcohol end groups (V). In contrast to the

WT sample, analysis of the 2D HMQC NMR spectrum of the

Atomt1 mutant lignin-enriched isolate (Fig. 5b) established the

presence of an additional substructure that was characterized by

two correlations (d_C/d_H) at 76.0/4.81 (C-7/7-H) and 78.2/4.0 (C-

8/8-H) consistent with benzodioxane substructure VI (Fig. 5c).

In addition, the observed long-range HMBC correlation from

the 5OHG aromatic proton-2 at d_H~6.80 ppm to the aromatic

ring carbon-4 (d_C~133.9 ppm) further extended the correlation

from d_C~133.9 ppm across to the ether-linked proton-8 (d_H~4.1

ppm) (data not shown). This established the presence of a 5OHG

aromatic ring in the benzodioxane subunit linked via an 8–O–

4¢/7–O–5¢ bonding pattern to the aliphatic side-chains of both G,

as well as 5OHG–G, units in benzodioxane substructure VI. This,

in turn, suggested that the 5OHG moieties were both homo- and

hetero-(G) linked in the lignin of the Atomt1 line.

3932 | Org. Biomol. Chem., 2010, 8, 3928–3946

This journal is

The Royal Society of Chemistry 2010

©

View Article Online

Table 1 Provisional relative frequencies (%) of inter-unit linkages by 2D

HMQC NMR analyses in lignin isolates of WT (Arabidopsis and alfalfa)

versus the Atomt1 mutant and alfalfa COMT down-regulated lines

be further veriﬁed through other means such as by quantitative

degradative analyses.

Furthermore, a re-evaluation of the previously estimated levels

of comparable inter-unit linkages in alfalfa (Medicago sativa L.),

as reported in an earlier study of COMT down-regulation,³³

resulted in our interpretation that the overall 8–O–4¢ inter-unit

linkage frequency was also conserved in both WT and COMT

alfalfa lines. We deduced, from the previously reported NMR

spectroscopic analysis, that the WT line had ~81% of overall

inter-unit linkages, whereas the COMT-line was considered to

have ~38% of benzodioxane (VI) and ~44% of 8–O–4¢ (I) inter-

unit linkages.¹Once again, the combined relative amounts totaled

~82%, indicating that estimated levels of 8–O–4¢ inter unit linkages

were conserved in both alfalfa lines. Yet, by contrast, this same

alfalfa lignin inter-unit-linkage frequency data and the presence

of the benzodioxane subunits was considered as further evidence

for random assembly, via a “combinational or combinatorial

biochemistry” process.^34,35These same researchers are committed

though to a random assembly process, with lignins having up

to 10⁶⁶possible isomers via monolignol-derived polymerization,

as well as putatively being able to utilize a wide range of non-

monolignol phenolics in cell-wall assembly, etc.³⁵Additionally, it

was also ventured that no two lignin structures are the same within

all plant species in the plant kingdom,³⁵but with no proof for such

an assertion.

By contrast, our own consideration of how ligniﬁcation occurs

is diametrically opposite to that above.^1,2,36Firstly, the frequency

of interunit linkages for both the Arabidopsis and alfalfa WT

and COMT knockout/mutant appears to be quite similar for all

linkage types within each plant line, albeit with some small changes

(e.g. in resinol (II) levels from 7 to 5% in Arabidopsis and 6 to 3%

in alfalfa). Whether such estimated differences are signiﬁcant, or

whether the programmed polymerization template is perturbed

because of the presence of benzodioxane substructures, cannot

be distinguished at this time. This high level of interunit linkage

conservation has occurred even though there are substantial dif-

ferences in the levels of benzodioxane-linked subunits in both cases

(i.e. 18 versus 38% subunits in Arabidopsis and alfalfa, respectively)

which could, in part, have also differentially perturbed the pro-

posed template-assisted lignin assembly process.^1,2,28–30Yet, while

the S : G (cleavable monomer) ratio is quite different in both lines

at plant maturity, (i.e. 0.28 and 0.64 for Arabidopsis and alfalfa,

respectively), these ratios when divided by the benzodioxane level

(as a fraction of 1) give near identical ratios of circa 1.6 and

1.7. Secondly, differences noted in S : G ratios between both plant

lines presumably also reﬂect differences in overall numbers and

thickness of S-rich ﬁber cell types in both species. Taken together,

there is thus apparently a remarkably unplastic assembly process

during ligniﬁcation in both Arabidopsis and alfalfa, albeit with

substantial differences in numbers of speciﬁc cell types that result

in the distinct S : G ratios noted. We consider this indicative of

a tightly controlled/conserved assembly process in distinct cell

types, as we have already reported in other studies at the cell wall

speciﬁc level.³

Arabidopsis

Alfalfa ^a

Inter-unit linkage

WT Atomt1 WT COMT down-regulated

8–O–4¢ (I)

79

7

9

2

60

5

9

4

3

81

6

8

1

3

44

3

8

4

Resinol (II)

Phenylcoumaran (III)

Dibenzodioxocin (IV)

CA end groups (V)

Benzodioxane (VI)

0

18

0

38

^aRe-calculated from Marita et al.³³

Conservation of 8–O–4¢ inter-unit lignin frequencies in WT and

COMT knocked-out/down-regulated lines (Arabidopsis versus

alfalfa).

Currently, neither thioacidolysis nor nitrobenzene oxidation

degradative analyses of the lignins in either WT or Atomt1

lines can account for the bulk of the putative lignins in these

plant samples, since recoveries of monomers relative to overall

putative lignin contents are circa 20–40%.¹Moreover, accurate

quantiﬁcation of benzodioxane substructure (VI) in the lignin

of the Atomt1 line is not yet currently achievable using either

degradation method (unpublished results). Accordingly, attempts

to provisionally estimate the relative overall frequencies of lignin

substructures (i.e., I, II, III, IV, V and VI) were next carried

out using NMR spectroscopic analyses. These were estimated

by measuring the volume integrals in the 2D-HMQC spectra

of the lignin-enriched isolates of both WT and Atomt1 mutant

lines, as had been previously reported for alfalfa lignin-derived

preparations.³³In this way, the NMR spectroscopic analysis of

the HMQC cone volumes of C-7/7-H correlations at 71.7/4.68

and 76.2/4.87 ppm, that are characteristic for the 8–O–4¢ (I)

and benzodioxane (VI) substructures, respectively, suggested that

substructure VI can contribute to ≤25–26% of the total detectable

8–O–4¢ interunit linkages present in the lignin-enriched isolate of

the Atomt1 line. Estimated inter-unit linkage (substructures I, II,

III, IV, V and VI) frequencies in the lignin isolates of WT and

Atomt1 mutant lines are summarized in Table 1.

Based on this approach, the WT Arabidopsis line provisionally

had ~79% of 8–O–4¢ (I), ~7% of resinol-like (II), ~9% of phenyl-

coumaran (III), ~2% of dibenzodioxocin (IV), ~2% of cinnamyl

alcohol end groups (V) and no detectable benzodioxane (VI)

inter-unit linkages, whereas the Atomt1 line accounted for ~60%

of 8–O–4¢ (I), ~5% of resinol-like (II), ~9% of phenylcoumaran

(III), ~4% of dibenzodioxocin (IV), ~3% of cinnamyl alcohol end

groups (V) and ~18% of benzodioxane (VI) inter-unit linkages,

respectively. The benzodioxane substructure (VI), however, is a

chemical variation of the 8–O–4¢ (I) inter-unit linkage. Thus, taken

together, the total 8–O–4¢ linkage levels in the Atomt1 line are

estimated as ~78%, i.e. the sum of ~60% 8–O–4¢ (I) and ~18%

benzodioxane (VI) inter-unit linkages, respectively. This, in turn,

suggests that the 8–O–4¢ linkage frequency in the lignin isolates

of both the Atomt1 mutant and the WT line was conserved. In

addition, the relative ratios and amounts of the other inter-unit

linkages detected in this way are provisionally also reasonably well

conserved using this approach. However, these data now need to

Molecular weight distributions of lignin-isolates

The molecular weight distributions of the lignin-enriched isolates

from Atomt1 and WT lines were estimated by gel permeation

This journal is

The Royal Society of Chemistry 2010

Org. Biomol. Chem., 2010, 8, 3928–3946 | 3933

©

View Article Online

The crude thioacidolysis reaction mixture was then fraction-

ated using preparative silica gel thin-layer chromatography and

reversed-phase HPLC, with the corresponding fractions individu-

ally subjected to 1D and 2D HMQC/HMBC NMR spectroscopic

analyses to determine which of those contained benzodioxane

subunits. Fractions containing the traces of cleaved trimeric 36

and the corresponding dimeric 35 benzodioxane were subjected

to mass spectroscopic analyses using LC-APCI-MS. This gave

masses in the positive ion mode at m/z 672.9 (M + H–HSEt)⁺,

611.1 (M + H–2HSEt)⁺(Fig. 7a) and m/z 478.9 (M + H–HSEt)⁺,

417.1 (M + H–2HSEt)⁺(Fig. 7b) indicative of the presence of

trimer 36 and dimer 35, respectively. This was further conﬁrmed

by analysis of three-bond long range HMBC correlations of

the isolated fractions. Speciﬁcally, the trimer structure 5OHG–

5OHG–G (36) was deduced from two distinct sets of three

bond correlations, correlating the benzodioxane carbon-7 at d_C

~76.2 ppm to the G aromatic ring proton-2 (d_H~7.18 ppm) and 6

(d_H~6.98 ppm), as well as the 5OHG aromatic ring proton-2 (d_H

~6.80 ppm) and 6 (d_H~6.69 ppm), respectively. The structure of

the 5OHG–G dimer (35) was also deduced from the correlation

of proton-7 (d_H~4.87 ppm) of the benzodioxane functionality

to carbon-2 (d_C~111.9 ppm) and carbon-6 (d_C~121.5 ppm) of

the guaiacyl ring (G) unit, as well as from extended individual

cross-peaks that correlated with aromatic protons-2 (d_H~7.12

ppm) and 6 (d_H~6.98 ppm) and aromatic carbon-4 (d_C~147.9

ppm), respectively. That G and 5OHG in 36, as well as G in 35,

moieties were individually linked to the 7 and/or 7¢ position of the

benzodioxane ring was conﬁrmed further from the strong cross-

peak correlation of protons-2 and 6 for the respective aromatic

ring (G and/or 5OHG) to that of their corresponding carbons-2,

4 and 6 in the respective HMBC spectra. Interestingly, both 1D

and 2D NMR spectroscopic analyses of 36 and 35 established that

the hydroxyl groups at the 9- and 9¢-positions in trimer 36 and the

9-position in dimer 35 were not thioethylated under thioacidolytic

reaction conditions. Noother related products were detectedunder

the conditions employed.

Fig. 6 Molecular weight distribution proﬁles of the lignin-enriched

isolates from Atomt1 (grey) and WT (black) lines.

chromatography using a pre-calibrated Sephadex G-100 column³²

eluted with 0.1 M NaOH, in order to minimize lignin association

effects.³⁷The molecular weight distribution elution proﬁles of the

enriched lignin isolates from Atomt1 and WT lines were relatively

broad, but very similar, with both centered at ~4500 Da (Fig. 6).

These values are also in good agreement with previously reported

Mw values of lignin-enriched isolates from A. thaliana WT.^27,32

The polydispersity index values (Mw/Mn) were also similar, being

calculated as ~2.2 and ~2.5 for WT and Atomt1 lignin-enriched

isolates, respectively. However, due to the preparation procedure

for the lignin-enriched isolates, the polydispersity index does not

reﬂect the in situ molecular weight distributions of the actual

lignin polymers but presumably better reﬂects only the “random”

cleavage occurring during mechanical ball milling treatment

(~4 days).

Lignin deconstruction and identiﬁcation of isolable

benzodioxane-linked subunits

Isolation of Atomt1 lignin-derived trimer (36) and dimer (35)

thioacidolysis derivatives. From the initial thioacidolysis analyses

of intact Atomt1 stem tissues described above, it was provi-

sionally considered that the benzodioxane substructure (VI) was

more resistant to cleavage than the 8–O–4¢ inter-unit linkage,

i.e. given that monomeric 5OHG moieties were not detected

to any signiﬁcant level in the mutant following thioacidolysis.

Therefore, to investigate this question, thioacidolysis was next

carried out under conventional conditions³⁸using the previously

discussed Atomt1 lignin-derived-isolate harboring benzodioxane-

linked substructures. As anticipated, NMR spectroscopic analysis

of the resulting crude thioacidolysis reaction mixture veriﬁed

presence of benzodioxane substructures [4.81 (7-H) and 76.0 (C-

7)/4.0 (8-H) and 78.2 (C-8)], i.e. they had not been (fully) cleaved

under these conditions.

Conﬁrmation of lignin subunit structures by total synthesis. To

further unambiguously conﬁrm the structures of the putative

thioacidolytically cleaved lignin-derived trimer 36 and dimer 35,

a synthetic strategy was next developed, the details of which

are summarized in the Experimental section. In this regard,

various coupling products were synthesized by modiﬁcation of

an Ag₂CO₃oxidation³⁹in circa 58–62% yield from coniferyl

(2) and sinapyl (3) alcohols, methyl 5-hydroxyferulate (37), and

its tert-butyldimethylsilyl ether derivative (38a/b), and O-tert-

butyldimethylsilyl coniferyl alcohols (41a/b) (Scheme 1). This

afforded the corresponding dimers 5OHG–G (32),³⁹5OHG–S

(34),^40–43and 5OHG–5OHG (43)^40,43with the 5OHG–5OHG–G

trimer (46) being obtained from coupling of methyl ester of the

5OHG–5OHG dimer (44) and coniferyl alcohol (2) followed by

DIBAL reduction. Initially, the Ag₂CO₃oxidation of methyl 5-

hydroxyferulate (37) and coniferyl alcohol (2) afforded a mixture

of the methyl esters of the threo/erythro-5OHG–G benzodioxane

dimer 39 in a 9 : 1 ratio, which was later reduced with DIBAL

to form the threo/erythro-5OHG–G benzodioxane dimer 32 in a

similar diastereomeric ratio with 88% yield. The HRMS spectrum

of threo/erythro-32 gave a (M + Na) at m/z 397.1259, consistent

with the molecular formula C₂₀H₂₂O₇+ Na (calculated mass:

3934 | Org. Biomol. Chem., 2010, 8, 3928–3946

This journal is

The Royal Society of Chemistry 2010

©

View Article Online

Fig. 7 Mass spectroscopic analyses of Atomt1 lignin-derived trimer 36 and dimer 35 and their synthetic counterparts. (a) Mass spectrum of Atomt1

lignin-derived trimer 36; (b) mass spectrum of Atomt1 lignin-derived dimer 35; (c) mass spectrum of synthetic trimer 36; and (d) mass spectrum of

synthetic dimer 35, and their putative charged fragments.

397.1263), and its LC-APCI-MS analysis gave a (M - H) ion at m/z

373.0 corresponding to molecular formula C₂₀H₂₂O₇. The ¹H and

¹³C NMR spectra of threo/erythro-32 were also in close agreement

with the reported literature data for nocomtol.³⁹Further, the threo

and erythro diastereoisomers of 32 were individually separated by

chiral OJ column and their structures were conﬁrmed by 1D and

mass spectral analyses (see Instrumentation and Chromatography

Materials section). In an analogous manner, the Ag₂CO₃oxidation

of methyl 5-hydroxyferulate (37) and sinapyl alcohol (3) initially

afforded a methyl ester of the threo-5OHG–S benzodioxane dimer

40, which was later reduced with DIBAL to afford the threo-

5OHG–S benzodioxane dimer 34 in 85% yield. Speciﬁcally, the

HRMS spectrum of threo-34 gave a (M + Na) ion at m/z 427.1400,

consistent with a molecular formula of C₂₁H₂₄O₈Na (calculated

mass: 427.1369), whereas LC-APCI-MS analysis of 34 gave a

(M - H) ion at m/z 403.0 corresponding to molecular formula

directly gave the desired product 43 in its threo form, but in very

low yield (~4%). A slightly better yield of product 43 (~10%) was

obtained using horseradish peroxidase (HRP), in the presence

of H₂O₂, with the product so obtained in a 1 : 9 erythro/threo

ratio. Interestingly, the erythro/threo product from the HRP/H₂O₂

catalyzed coupling of 5-hydroxyconiferyl alcohol (4) could readily

be separated into its threo and erythro diastereoisomers by chiral

OJ column chromatography (see Instrumentation and Chro-

matography Materials section). However, the best overall synthetic

procedure involved modiﬁed Ag₂CO₃oxidative coupling of the

tert-butyldimethylsilyl ether derivatives of 5-hydroxyconiferyl al-

cohol (41a/b) with 5-hydroxyferulate methyl ester (37) to afford

initially the corresponding benzodioxane esters 42a/b which

were then converted to 43 by DIBAL reduction and TBAF de-

protection (Scheme 1). The overall yield of this conversion from

38a/b was ~90%, with the product 43 obtained in threo form.

The HRMS spectrum of threo-43 gave a (M + Na) ion at

m/z 413.1207, consistent with the molecular formula C₂₀H₂₂O₈

+ Na (calculated mass: 413.1212), with the LC-APCI-MS also

1

C₂₁H₂₄O₈. Additionally, the H and ¹³C NMR spectra of threo-

34 were in close agreement with the reported literature data for

nitidanin.^40–43

1

To obtain the 5OHG–5OHG benzodioxane dimer 43, three

different synthetic protocols were considered. First, a modiﬁed

Ag₂CO₃oxidative coupling of 5-hydroxyconiferyl alcohol (4)

giving a (M - H) ion at m/z 388.9. Comparison of the H and

¹³C NMR spectra with 5OHG–G diol 32 and 5OHG–S diol 34

also established the presence of an hydroxyl group in compound

This journal is

The Royal Society of Chemistry 2010

Org. Biomol. Chem., 2010, 8, 3928–3946 | 3935

©

View Article Online

Scheme 1 Synthetic route to dimers 32, 34, 43 and trimer 46. [*threo-(7R-8R)-43 is also formed.]

threo-43 at the 5-position. That is, the ¹H NMR spectrum

contained two sets of meta-coupled signals as doublets at d 6.7

(J 1.5, 2¢-H), 6.67 (J 1.5, 6¢-H) and 6.65 (J 1.5, 2-H)/6.59 (J

2, 6-H) ppm, along with two methoxyl groups at d 3.85 (s, 3¢-

OMe) and 3.83 (s, 3-OMe) ppm, indicative of two 5OHG units

coupled together. Furthermore, the ¹H NMR spectrum of threo-

43 displayed two trans-oleﬁnic protons as doublets at d 6.47 (J

16.0, 7¢-H), 6.27 (J 5.5 and 15.5, 8¢-H) ppm, and a doublet at d

4.20 (J 5.0, 9¢-H) ppm for an allylic alcohol group connected to

one of the two 5OHG units. The trans-conformation for threo-43

was also established from analysis of the resonances at d 4.9 (d,

J 8.0, 7-H) and 4.0 (m, 8-H) ppm corresponding to two vicinal

aliphatic oxymethines as well as two geminal methylene protons

as double doublets at d 3.78 (J 2.0 and 12.25, 9-H) and 3.52

(J 3.5 and 12.5, 9-H) ppm, respectively. The 5OHG–5OHG diol

threo-43 structure was further conﬁrmed by 2D (HMBC, HMQC)

NMR spectroscopic analyses. In this regard, the NMR spectra

for erythro-43 were also essentially identical to threo-43, except

for the vicinal aliphatic oxymethine proton resonances at d 5.2

(d, J 3.0, 7-H) and 4.46 (m, 8-H) ppm as well as the geminal

methylene protons 3.62 (br dd, J 5.0, 7.5 and 12.0, 9-H) and 3.49

(dd, J 5.0 and 12.0, 9-H) ppm, respectively. Interestingly, this is

the ﬁrst report of the chemical synthesis of 5OHG–5OHG dimer

(43), although it has been reported to occur as a natural product,

simplidin (threo conﬁguration), in extracts of Firmiana simplex.⁴³

In a similar manner, the chemical synthesis of trimer 46 was

achieved via modiﬁed Ag₂CO₃oxidative coupling of methyl

ester of 5OHG–5OHG benzodioxane dimer 44 with coniferyl

alcohol (2) followed by DIBAL reduction (Scheme 1). The

HRMS spectrum of 46 had a (M + Na) ion at m/z 591.1846,

consistent with molecular formula C₃₀H₃₂O₁₁+ Na (calculated

mass; 591.1842), with the LC-APCI-MS of 46 also giving a

(M - H) ion at m/z 567.0 corresponding to molecular formula

C₃₀H₃₂O₁₁. Interestingly, its structure closely resembles that of the

previously reported benzodioxane/benzodioxane-type sesquine-

olignans, namely isoamericanol B1-C2,⁴⁴except for the presence

of methoxyl groups at the 3, 3¢ and 3¢¢- positions in 46. Analysis of

the ¹H and ¹³C NMR spectra of 46 established that it existed as a

mixture of diastereoisomers in the ratio of ~1.3 : l (threo/erythro)

that could not readily be separated. The ¹H NMR spectrum of 46

3936 | Org. Biomol. Chem., 2010, 8, 3928–3946

This journal is

The Royal Society of Chemistry 2010

©

View Article Online

also contained three proton signals at d 7.12 (dd, J 2.0 and 4.5,

2-H), 6.97 (d, J 8.0, 6-H), 6.88 (d, J 8.0, 5-H) ppm for a G-unit and

four proton signals at d 6.78 (br s, 2¢-H), 6.71 (dd, J 2.0 and 4.5,

2¢¢-H), 6.70 (dd, J 2.0 and 4.5, 6¢¢-H) and 6.60 (d, J 2.0, 6¢-H) ppm

corresponding to two 5OHG units. Additionally, the presence of

three methoxyl groups at the 3, 3¢ and 3¢¢-positions was conﬁrmed

from the resonances at d 3.88 (s, 3-OMe), 3.87 (s, 3¢-OMe) and

4.06 (m, 8-H) ppm and 3.79 (br d, J 2.5 and 12.5, 9-H), 3.50 (br

d, J 3.5 and 12.0, 9-H) ppm, respectively.

LC-APCI-MS of synthetic 36 in the positive ion mode (Fig. 7c)

gave fragment ions very similar to the isolated 36 (Fig. 7a) at

m/z 672.9 (M + H - HSEt)⁺and 611.1 (M + H - 2HSEt)⁺

corresponding to the molecular formula C₃₆H₄₆O₁₀S₃, as well as

a (M - H) ion at m/z 733.1. The ¹H NMR spectrum of 36,

upon comparison with the 5OHG–5OHG–G triol 46, once again

demonstrated the absence of two trans-oleﬁnic protons at the

7¢¢, 8¢¢ positions and an hydroxyl group at the 9¢¢ position. As

for compound 35, these positions were substituted with three

thioethyl moieties as indicated by multiplets at d 2.63-2.33 and

1.24-1.09 ppm, respectively. Interestingly, the hydroxyl groups at

both 9 and 9¢-positions were again not thioethylated. The ¹H NMR

spectrum of 36 also contained seven aromatic resonances at d 7.12

(dd, J 1.5 and 5.5, 2-H), 6.97 (dd, J 1.5 and 8.0, 6-H) and 6.86

(dd, J 1.5 and 8.0, 5-H) ppm for the G-unit and d 6.77 (br s,

2¢-H and 6¢-H), 6.719 (dd, J 1.5 and 2.0, 2¢¢-H) and 6.717 (dd,

J 2.0, 6¢¢-H) ppm corresponding to the trithioethylated derivative

of two 5OHG units, respectively. The presence of a diastereomeric

mixture of 36 in the ratio of ~1.3 : 1 (threo/erythro) was also evident

from the resonances at d 5.0 (d, J 8.0, 7-H), 4.97 (d, J 7.5, 7¢-

H), 4.08 (m, 8-H, 8¢-H) ppm and 3.81 (m, 9-H and 9¢-H), 3.54

(m, 9-H and 9¢-H) ppm. In this way, the Atomt1 lignin-derived

thioacidolysis products 35 and 36 were fully identiﬁed.

1

3.86 (s, 3¢¢-OMe) ppm, respectively. The H spectrum included

signals attributable to trans-oleﬁnic protons at d 6.47 (d, J 16.0,

7¢¢-H), 6.27 (dt, J 5.5 and 16.0, 8¢¢-H) ppm and an hydroxy-methyl

proton at d 4.20 (t, J 5.0, 9¢-H) ppm. In addition to these proton

resonances, the ¹H NMR spectrum of 46 established the presence

of two sets of aliphatic AMXY-type signals at d 4.99 (d, J 8.0,

7-H), 4.97 (d, J 7.5, 7¢-H), 4.08 (m, 8-H and 8¢-H), 3.82 (dd, J 4.5

and 12.0, 9-H and 9¢-H) and 3.54 (dd, J 5.5 and 12.5, 9-H and

9¢-H) ppm, respectively. The structure of the 5OHG–5OHG–G

benzodioxane trimer 46 was further conﬁrmed by 2D (HMBC,

HMQC and HSQC) NMR analyses (data not shown).

Chiral HPLC analysis for 32, 34, 43 and 46. Chiral HPLC was

then used to establish conditions for baseline separation of the

(+)- and (-)-antipodes of the benzodioxane dimers threo/erythro-

5OHG–G (32), threo-5OHG–S (34), and threo-5OHG–5OHG

(43), whereas the 5OHG–5OHG–G benzodioxane trimer 46 was

only partially separable under the conditions employed (data not

shown). The chiral HPLC analysis of the three benzodioxane

dimers 32, 34 and 43 established the corresponding (+) and (-)-

antipodes were present in an ~1 : 1 ratio, respectively.

Time-course of thioacidolysis of 5OHG–G benzodioxane dimer

(32) and 5OHG–5OHG–G benzodioxane trimer (46)

It was next instructive to examine the time course of thioacidolytic

conversion of the corresponding lignin model compounds 32 and

46, and the product distribution and yield (%) at different time

points during treatment, taking into account that they contain a

7¢–8¢ unsaturated side-chain. Of particular interest was whether

the 5OHG–5OHG subunit or the 5OHG–G subunit in trimer

46 underwent preferential cleavage, as this could provide useful

insight into the nature of the chemistry of the Atomt1 lignin

biopolymer. In this regard, each model compound 32 and 46

was subjected to thioacidolysis for periods ranging from 0 to 5 h

duration, with the various products quantiﬁed as shown in Fig. 8.

Synthesis of lignin-derived thioacidolysis products of 35 and 36

Synthetic dimer 32 and trimer 46 were next subjected to

thioacidolysis under conventional conditions,^38,45to generate the

benzodioxane dimer/trimer thioethylated products 35 and 36,

respectively (see Experimental section). In both cases, their struc-

tures were unambiguously established using different NMR and

mass spectroscopic analyses, with these ultimately demonstrated

as being identical to those obtained from the Atomt1 mutant plant

line as indicated below.

Speciﬁcally, LC-APCI-MS of synthetic 35 in the positive ion

mode gave fragment ion peaks at m/z 478.9 (M + H - HSEt)⁺

and 417.1 (M + H - 2HSEt)⁺(Fig. 7d) identical to the isolated

35 (Fig. 7b) corresponding to molecular formula of C₂₆H₃₆O₆S₃,

whereas in the negative ion mode it gave a (M - H) ion at m/z

538.9. Upon comparison with the 5OHG–G benzodioxane diol 32,

its ¹H NMR spectrum indicated the absence of two trans-oleﬁnic

protons at the 7¢, 8¢ positions and an allylic hydroxyl group at the

9¢ position, with the side-chain also being substituted with three

thioethyl moieties (multiplets at d 2.64–2.33 and 1.25–1.10 ppm),

respectively. Again, the hydroxyl group at the 9-position was not

thioethylated under the reaction conditions employed, as indicated

Monothioethylation. Prior to beginning the time course under

standard thioacidolysis conditions, the 5OHG–G benzodioxane

dimer 32 (3.1 mmol) was subjected to thioacidolysis at room

temperature. Within 1 min of reaction time, the bulk of the

starting material was converted into mono-substituted derivative

47 (1.11 mmol, ~34%). The LC-APCI-MS of 47 gave a (M - H)

ion at m/z 417.0 corresponding to the molecular formula of

C₂₂H₂₆O₆S, whereas its ¹H NMR spectrum upon comparison with

the 5OHG–G benzodioxane diol (32) established the presence of

trans-oleﬁnic protons at 7¢, 8¢ positions and the absence of an

hydroxyl group at the 9¢ position which was instead substituted

with a thioethyl moiety. The ¹H NMR spectrum of 47 also

displayed ﬁve aromatic resonances at d 7.11 (d, J 2.0, 2-H),

6.96 (dd, J 2.0 and 8.0, 6-H) and 6.88 (d, J 8.0, 5-H) ppm

for the G-unit and d 6.72 (d, J 1.5, 2¢-H), 6.60 (d, J 1.5, 6¢-

H) ppm corresponding to the monothioethylated derivative of

the 5OHG unit, respectively. The presence of a benzodioxane

substructure was evident from resonances in a trans-conformation

at d 4.98 (d, J 8.0, 7-H) and 4.0 (m, 8-H) ppm, corresponding

1

earlier for the isolated dimer 35 and trimer 36. The H NMR

spectrum of 35 contained ﬁve aromatic resonances at d 7.10 (br s,

2-H), 6.96 (d, J 8.0, 6-H) and 6.88 (dd, J 1.5 and 8.5, 5-H) ppm

for the G-unit and d 6.78 (dd, J 2.0 and 7.0, 2¢-H), 6.70 (ddd,

J 1.5 and 6.5, 6¢-H) ppm corresponding to the trithioethylated

derivative of the 5OHG unit. Presence of a diastereomeric mixture

for 35 in the ratio of ~1.7 : 1 (threo/erythro) was also evident from

the resonances at d 4.98 (d, J 7.5, 7-H), 4.97 (d, J 8.0, 7-H), 4.08-

This journal is

The Royal Society of Chemistry 2010

Org. Biomol. Chem., 2010, 8, 3928–3946 | 3937

©

View Article Online

Thioacidolytic treatment of synthetic 5OHG–5OHG–G benzo-

dioxane trimer 46. After thioacidolytic treatment of trimer 46

(2.71 mmol) for one minute only, the corresponding monothioethy-

lated product 48 was formed (but not quantiﬁed) with only ~28%

of trimer 46 remaining (Fig. 8b). Within 1 h, however, trimer 36 was

mainly detected (1.58 mmol, 58.5%) together with smaller amounts

of the monomeric G-unit 19 (0.45 mmol, ~16.5%) and traces of the

5OHG–G dimer 35. The trimer 36 levels then began to decrease

over a period of 5 h (to ~0.27 mmol, ~9.9%), with the 5OHG–

G dimer 35 concomitantly increasing in amounts (0.51 mmol,

~19%). As before, the amounts of the G-monomer 19 increased

until reaching a plateau level at 3–4 h (1.04 mmol, ~38.4%).

Under the reaction conditions employed, neither the monomer

5OHG (20) nor the 5OHG–5OHG benzodioxane dimer 49 were

detected. Thus, thioacidolysis of synthetic trimer 46 resulted only

in incomplete conversion into products 19 and 35.

Although thioacidolysis of the corresponding model com-

pounds 32 and 46 currently imposes an analytical limitation in the

overall quantiﬁcation of the tri-substituted derived 5OHG unit

(20), it interestingly provides evidence for differential (selective)

cleavage of the benzodioxane functionality within the 5OHG–

5OHG–G benzodioxane trimer 46 subunit, i.e. thereby provid-

ing evidence for the release of the corresponding 5OHG–G

subunit (instead of the 5OHG–5OHG subunit). This result is

thus in agreement with the thioacidolysis of the Atomt1 lignin-

derived-isolate, which accounted for isolation of compounds 19

(monomer), 35 (dimer) and 36 (trimer), and could be considered

exclusively derived from 5OHG–5OHG–G trimer 46. Thus, it

can be provisionally concluded that a 5OHG–5OHG–G unit

corresponds to the benzodioxane substructure (Fig. 9) in the lignin

macromolecular assembly of the Atomt1 mutant line, which upon

deconstruction released quantiﬁable amounts of derived 5OHG–

G benzodioxane dimer (32) and 5OHG–5OHG–G benzodioxane

trimer (46), respectively. How this reﬂects the actual lignin

deconstruction is not completely known at this point, i.e. given

that the side-chain at C7–C8 is not substituted as is mainly the

case for lignin in situ.

Fig. 8 Thioacidolysis treatment of synthetic (a) 5OHG–G benzodioxane

dimer 32 and (b) 5OHG–5OHG–G benzodioxane trimer 46.

to two vicinal aliphatic oxymethines and two geminal methylene

protons at d 3.78 and 3.50 (m, 9-H) ppm. Presence of two trans-

oleﬁnic protons was noted as doublets at d 6.39 (d, J 15.5, 7¢-H),

6.12 (dt, J 7.5 and 16.0, 8¢-H) ppm, as well as a doublet at d 3.30

(d, J 7.5) ppm for the 9¢-H proton and a monothioethyl group

from the proton resonances at d 2.49 (q, J 7.5)/1.21 (t, J 6.5 and

7.5) ppm, respectively, for the 10¢ and 11¢-protons connected to

a 5OHG unit. The two methoxyl groups resonated at d 3.87 (s,

3-OMe) and 3.86 (s, 3¢-OMe) ppm.

Thioacidolytic treatment of synthetic 5OHG–G benzodioxane

dimer 32. After 1 min at elevated temperature and pressure

(i.e. conventional thioacidolytic conditions), circa 85% of start-

ing dimer 32 was consumed, with concomitant generation of

monosubstituted derivative 47 (1.11 mmol, ~34%), based upon

direct UPLC analysis and quantiﬁcation of the reaction mixture.

Within 1 h, the mono-substituted product 47 reached basal levels

and remained essentially constant thereafter (Fig. 8a). On the

other hand, dimer 35 increased in amount to its highest level

(1.05 mmol, ~34%) by 1 h, with this then steadily decreasing until

reaching a plateau level (0.44 mmol, ~14%) by 5 h. By contrast,

formation of the monomeric G-unit (19) increased steadily until

it also began to attain a plateau level (1.38 mmol, ~44.6%) by

5 h. The corresponding 5OHG thio-derivative 20, on the other

hand, was only detected in trace amounts (data not shown). Thus,

taken together, thioacidolysis of dimer 32 resulted in an incomplete

conversion of starting material to afford a mixture of products, and

in so doing providing some insight into the dynamics which may

also occur during lignin polymer thioacidolysis/deconstruction.

Fig. 9 Putative partial lignin primary structure in ﬁbers from WT

(S–S–G) and Atomt1 (5OHG–5OHG–G) Arabidopsis lines.

Arabidopsis COMT gene cloning, heterologous expression,

puriﬁcation and kinetic analyses

Analysis of the Arabidopsis genome indicated that 17 genes were

annotated as putative COMTs.⁴⁶Of these Arabidopsis homologues,

3938 | Org. Biomol. Chem., 2010, 8, 3928–3946

This journal is

The Royal Society of Chemistry 2010

©

View Article Online

Table 2 Kinetic parameters of AtCOMT1 with various phenyl-

propanoids and quercetin (50)

G (35) and 5OHG–5OHG–G (36) deconstruction components,

albeit where the overall 8–O–4¢ inter-unit linkage frequency in the

Atomt1 knockout line has apparently been conserved relative to

WT. Deconstruction and identiﬁcation of the released subunits

provide further support to lignin macromolecular assembly in

Arabidopsis being conserved in a highly programmed (template

guided) manner, with 5OHG units provisionally being 8–O–

4¢ linked to the primary lignin polymer via either one or two

5OHG units capped by a G-unit. It is signiﬁcant that G-moieties

apparently serve as either terminal units in the benzodioxane

substructures of the polymer or as end-groups, i.e. rather than the

previously contemplated/speculated benzodioxane substructures

linked n-fold.³³

V

_max/pkat mg^-1

Substrates

K_m/mM

protein

k

_cat/s^-1

K

_enz/M^-1s^-1

Caffeic acid (5) 24.2 3.5

14.6 1.3

30.1 4.3

0.62

1.30

26 000

40 600

5-OH Ferulic

acid (6)

Caffeyl

32.0 6.9

19.7 10.8 35.9 10.3

1.55

2.85

1.52

2.23

0.16

78 700

159 000

29 500

70 600

6800

aldehyde (9)

5-OH coniferyl 17.9 9.6

aldehyde (10)

Caffeyl alcohol 51.5 15.8 35.3 5.6

66.2 17.1

(11)

5-OH coniferyl 31.6 8.4

51.9 8.2

alcohol (4)

Quercetin (50)

23.7 19.3 3.8 2.4

AtCOMT1 has the highest similarity (78.8%) and identity (75.5%)

to COMTI from tobacco,⁷with the other homologues showing

~62–42% similarity and 54–33% identity to AtCOMT1. Twelve of

these 17 COMTs were cloned (AtCOMT1, 3, 4, 6, 8–13, 15 and

17), with the corresponding recombinant proteins expressed in

E. coli, puriﬁed to homogeneity by afﬁnity chromatography and

biochemically characterized. Only AtCOMT1 displayed COMT

activity proper in vitro: speciﬁcally, it had a very broad substrate

speciﬁcity for a wide range of potential substrates (Table 2), with

caffeyl (9) and 5-OH coniferyl (10) aldehydes being the preferred

substrates, having K_enzvalues of 78 700 and 159 000 M^-1s^-1, re-

spectively. Interestingly, AtCOMT1 was also previously described

as involved in quercetin (50) to isorhamnetin (51) formation.⁴⁷

However, the kinetic data in vitro suggest that quercetin (50)

is a much poorer substrate by at least an order of magnitude

(i.e. K_enz= 6900 M^-1s^-1for 50 versus 159 000 M^-1s^-1for 10).

The remaining other 11 putative COMTs, all obtained in soluble

recombinant protein forms, were unable to methylate any of the

potential substrates used.

Experimental

Instrumentation and chromatography materials

All solvents and chemicals used were either reagent or HPLC

grade and purchased from Sigma-Aldrich (St. Louis, MO, USA),

unless otherwise stated. Chemical reactions were carried out under

anhydrous conditions using dry freshly distilled solvents in an

argon atmosphere. Silica gel thin-layer chromatography (TLC)

employed Partisil^ꢀPK5F (Whatman, 20 ¥ 20 cm, 1 mm, 150 A)

R

˚

and AL SIL G/UV₂₅₄(Whatman, 20 ¥ 20 cm, 0.25 mm) whereas

silica gel column chromatography (CC) utilized silica gel 60

(EM Science). UV and a 10% H₂SO₄spray followed by heating

were employed for TLC plate visualization. High-performance

liquid chromatography (HPLC) analyses were performed using

an Alliance 2690 system (Waters, Milford, MA) equipped with

an UV-Vis diode array detector. Reversed-phase separations were

carried out with a SymmetryShield RP₁₈column (150 ¥ 3.9 mm,

Waters) with detection at 280 nm. Elution conditions at a ﬂow rate

of 1 cm³min^-1were: linear gradients of CH₃CN–H₂O (Optima

LC/MS, Fischer Scientiﬁc) from 5 : 95 to 40 : 60 in 5 min, then

to 80 : 20 in 40 min, to 95 : 5 in 2 min with this composition held

for an additional 2 min, to 95 : 5 in 2 min, and ﬁnally to 5 : 95 in

2 min with this being held for 5 min. Chiral HPLC separations for

determining enantiomeric compositions of diastereoisomers 32,

34, 43 and 46 employed a Chiralcel OJ column (250 ¥ 4.6 mm,

Chiral Technologies, Inc.) eluted with hexanes–EtOH (7 : 3) at a

ﬂow rate of 0.5 cm³min^-1for 32/34/43 and 0.3 cm³min^-1for

46. Eluant was monitored at 280 nm, as well as with an in-line

Advanced Laser Polarimeter detector (PDR-Chiral, Lake Park,

Florida, USA, cell volume 56 ml, length 5.17 cm). Optical rotations

were measured with a JASCO DIP-181 digital polarimeter at 20 ^◦C

(Hg-D line, 546 nm) using a quartz cell (length: 5 cm, volume

1 cm³), with values given in 10^-1deg cm²g^-1.

NMR spectra were recorded on Varian Mercury-Vx 300,

Inova 500 and Varian 600 MHz spectrometers operating at

300.1, 499.85 and 599.69 MHz (¹H), and at 75.5, 125.67 and

150.8 MHz (¹³C), respectively, with J values given in Hz. Liquid

Conclusions

Lignin primary structure in COMT mutated lines produces

benzodioxane substructures that are readily cleavable as 5OHG–

This journal is

The Royal Society of Chemistry 2010

Org. Biomol. Chem., 2010, 8, 3928–3946 | 3939

©

View Article Online

chromatography-mass spectrometry with an atmospheric pres-

sure chemical ionization interface system (LC-APCI-MS) was

performed using a Finnigan MAT (San Jose, CA, USA) ion

trap mass spectrometer equipped with a Finnigan electrospray

interface, with liquid chromatographic separations carried out

standards and by mass spectrometric analyses using a HP 5973

MS detector (EI mode, 70 eV).

Chemical syntheses

R

on an ACQUITY Ultra Performance Liquid Chromatograph^ꢀ

Coniferyl (2), sinapyl (3) and 5-hydroxyconiferyl (4) alcohols,

5-hydroxyferulic acid (6), caffeyl (9) and 5-hydroxyconiferyl (10)

aldehydes, methyl 5-hydroxyferulate (37) and methyl O-tert-

butyldimethylsilyl ferulate (38a/b). These were synthesized as

previously described in Kim et al.⁵²

(UPLC, Waters) system using a BEH C18 column (Waters, 2.1 ¥ 50

mm) eluted as follows: ﬂow rate of 0.3 cm³min^-1; linear gradients

of CH₃CN-3% (v/v) AcOH in H₂O from 10 : 90 to 45 : 55 in

1.5 min, then to 60 : 40 in 5 min, to 80 : 20 in 2 min with this

composition held for an additional 1 min, then to 100 : 0 for 1 min

and ﬁnally to 10 : 90 in 0.5 min with this being held for 4 min. High

resolution mass spectrometric analyses (HRMS) were performed

at the University of Mississippi using an Agilent HPLC 1100

series equipped with a diode array detector and mass detector

in series (Agilent Technologies, Palo Alto, CA, USA). The mass

detector was a time of ﬂight (Model G1969A) equipped with an

electrospray ionization interface controlled by Agilent software

(Agilent MassHunter Work Station, A.02.01). Data acquisition

and processing were done using the AnalystTM QS software

(Analyst, QS 1.1). All acquisitions were performed under negative

ionization mode with a capillary voltage of 3000 V. Nitrogen was

used as the nebulizer gas (30 psi) as well as the drying gas at

11 m³min^-1at a temperature of 350 ^◦C. The voltage of the PMT,

fragmentor and skimmer was set at 850, 175 and 60 V, respectively.

Full scan mass spectra were acquired from m/z 100–1300.

O-tert-Butyldimethylsilyl coniferyl alcohols (41a/b). To an ice-

cold solution of 38a/b (2 g, 5.9 mmol) in dry toluene (40 cm³)

was added 1 M diisobutylaluminium hydride (DIBAL) in toluene

(18 cm³, 17.7 mmol). After stirring the contents for 1 h, excess

DIBAL was destroyed by drop-wise addition of MeOH (10 cm³).

The reaction mixture was then extracted with EtOAc (2 ¥ 300 cm³),

washed with brine (1 ¥ 100 cm³) and dried (Na₂SO₄). Following

evaporation of the EtOAc solubles to dryness in vacuo, the residue

so obtained was subjected to silica gel CC eluted with hexanes–

EtOAc (4 : 1) to afford a mixture of O-tert-butyldimethylsilyl

coniferyl alcohols (41a/b) which was used without further puriﬁ-

cation (2.5 : 1 ratio, 1.7 g, 5.4 mmol, 90%). d_H(500 MHz; acetone-

d₆; Me₄Si) 7.45 (1 H, ArOH, 41a), 7.21 (1 H, ArOH, 41b), 6.73

(1 H, d, J 1.5, 2-H, 41a), 6.61 (1 H, dd, J 1.5 and 4.0, 2-H, 41b),

6.60 (1 H, dd, J 1.5 and 4.0, 6-H, 41b), 6.58 (1 H, d, J 2.0, 6-

H, 41a), 6.45 (1 H, d, J 16.0, 7-H, 41a/b), 6.2 (1 H, dt, J 5.5

and 16.0, 8-H, 41a/b), 4.19 (2 H, br d, J 5.0, 9-H, 41a/b), 3.85

(3 H, s, 3-OMe, 41a), 3.82 (3 H, s, 3-OMe, 41b), 1.0 (18 H, s,

2 ¥ 3Me-TBS, 41a/b), 0.2 (6 H, s, 2Me-TBS, 41a), 0.17 (6 H, s,

2Me-TBS, 41b); d_C(125 MHz; acetone-d₆; Me₄Si) 152.38, 149.52,

149.39, 144.17, 138.17, 133.12, 131.35, 130.40 (C-7, 41b), 130.33

(C-7, 41a), 129.17, 129.07, 128.44, 113.16 (C-6, 41a), 107.9 (C-2,

41b), 103.9 (C-2, 41a), 102.62 (C-6, 41b), 63.37 (C-9, 41b), 63.30

(C-9, 41a), 56.44 (C-3-OMe, 41b), 55.74 (C-3-OMe, 41a), 26.30

(C-3Me-TBS, 41a), 26.17 (C-3Me-TBS, 41b), 19.3, 19.0, -4.12,

-4.21; m/z (LC-APCI) 309.2 (M - H. C₁₆H₂₅O₄Si requires 309.2)

(t_R= 4.3 min).

Plant material and growth parameters

Seeds from the Atomt1 knockout mutant line, provided by Dr.

Lise Jouanin [Institut National de la Recherche Agronomique

(INRA), Versailles, France], were ﬁrst grown on MS medium

containing kanamycin (50 mg cm^-3). Transformants were next

transferred to soil and grown in Washington State University

greenhouses as previously described²⁷until maturation with the

seeds next harvested. These were then subjected to another round

of kanamycin screening with seedlings also subjected to GUS

screening as described.⁴⁸

threo/erythro-5OHG–G benzodioxane dimer (32). To a mix-

ture of methyl 5-hydroxyferulate (37, 0.5 g, 2.23 mmol) and

coniferyl alcohol (2, 0.4 g, 2.22 mmol) in benzene–acetone

(2 : 1, 60 cm³) was added Ag₂CO₃(1.5 g, 5.45 mmol).³⁹The

contents were then stirred at room temperature under N₂for

about 11 h. After completion of the reaction, as monitored

by TLC (hexanes–EtOAc 3 : 7), the Ag₂CO₃was removed by

ﬁltration through a Celite pad with the latter washed with

acetone (20 cm³). The resulting ﬁltrate was evaporated to dryness

in vacuo, with the residue puriﬁed further by silica gel CC

using hexanes–EtOAc (3 : 7) as eluant to afford threo/erythro-

5OHG–G benzodioxane methyl ester (~9 : 1 ratio) [(E)-

methyl 3-(3-(4-hydroxy-3-methoxyphenyl)-2-(hydroxymethyl)-8-

methoxy-2,3-dihydrobenzo[b][1,4]dioxin-6-yl)acrylate] which was

used without further puriﬁcation (39, 0.55 g, 1.36 mmol,

61%).^39,53–55d_H(500 MHz; acetone-d₆) 7.8 (1 H, br s, ArOH), 7.55

(1 H, d, J 16.0, H-7¢), 7.12 (1 H, d, J 2.0, 2-H), 6.98 (1 H, dd, J

2.0 and 8.0, 6-H), 6.96 (1 H, d, J 2.0, 2¢-H), 6.89 (1 H, d, J 8.5,

5-H), 6.88 (1 H, d, J 1.5, 6¢-H), 6.43 (1 H, d, J 16.0, 8¢-H), 5.3 (1

H, d, J 2.5, 7-H_erythro), 5.0 (1 H, d, J 8.0, 7-H_threo), 4.12 (1 H, m,

8-H), 3.90 (3 H, s, 3-OMe), 3.87 (3 H, s, 3¢-OMe), 3.81 (1 H, br d,

Both wild type [ecotype Wassilewskija (Ws)] and the selected

Atomt1 lines were grown. All lines were harvested weekly from 3

weeks until 10 weeks, with growth and development parameters

of the main (bolting) stem also measured (i.e. length and basal di-

ameters) using twenty randomly chosen plants/harvest sampling

point.

Lignin analysis: acetyl bromide, alkaline nitrobenzene oxidation

and thioacidolysis estimations

Extractive-free cell wall residues (CWR) were prepared as pre-

viously described,²⁷with lignin contents next estimated by the

acetyl bromide method^38,49using modiﬁed absorptivity values

for speciﬁc lignin type i.e. H, G or S-enriched.²⁷Monomeric

compositions were estimated using both alkaline nitrobenzene

oxidation⁵⁰and thioacidolysis^38,45,51methods, with all products

analyzed using an HP 6890 Series GC System equipped with a

RESTEK-5Sil-MS (30 m ¥ 0.25 mm ¥ 0.25 mm) column. Both

alkaline nitrobenzene oxidation and thioacidolysis products were

identiﬁed by comparison/calibration against authentic silylated

3940 | Org. Biomol. Chem., 2010, 8, 3928–3946

This journal is

The Royal Society of Chemistry 2010

©

View Article Online

J 12.25, 9-H), 3.72 (3 H, s, COOMe), 3.52 (1 H, br d, J 12.5, 9-

H); d_C(125 MHz; acetone-d₆) 167.8 (C-9¢ COOMe), 150.38 (C-3¢),

148.55 (C-3), 148.1 (C-4), 145.6 (C-5¢), 145.6 (C-7¢), 136.78 (C-4¢),

129.0 (C-1¢), 127.5 (C-1), 121.68 (C-6), 116.8 (C-8¢), 115.8 (C-5),

112.0 (C-2), 111.29 (C-6¢), 105.1 (C-2¢), 79.7 (C-8), 77.0 (C-7),

61.74 (C-9), 56.5 (C-3-OMe), 56.4 (C-3¢-OMe), 51.6 (CO-OMe);

(1 H, d, J 1.8, 2¢-H), 6.88 (1 H, d, J 1.8, 6¢-H), 6.82 (2 H, s, 2-H and

6-H), 6.43 (1 H, d, J 16.2, 8¢-H), 4.97 (1 H, d, J 7.8, 7-H_threo), 4.1 (1

H, m, 8-H), 3.91 (6 H, s, 3¢-OMe), 3.84 (3 H, s, 3-OMe, 5-OMe),

3.78 (1 H, dd, J 2.4 and 12.6, 9-H), 3.72 (3 H, s, COOMe), 3.52

(1 H, dd, J 4.2 and 12.6, 9-H); d_C(150 MHz; acetone-d₆and 3

drops CD₃OD): 167.9 (C-9¢, COOMe), 150.4 (C-3¢), 148.8 (C-3

and C-5), 145.64 (C-5¢), 145.61 (C-7¢), 137.38 (C-4), 136.8 (C-4¢),

127.9 (C-1¢), 127.6 (C-1), 116.8 (C-8¢), 111.3 (C-6¢), 106.2 (C-2

and C-6), 105.1 (C-2¢), 79.8 (C-8), 77.4 (C-7), 61.6 (C-9), 56.77

(C-3-OMe and C-5-OMe), 56.56 (C-3¢-OMe), 51.7 (CO-OMe);

m/z (LC-APCI) 401.2 (M - H. C₂₁H₂₁O₈requires 401.1) (t_R

2.2 min).

=

To a solution of threo/erythro-39 (0.50 g, 1.2 mmol) in toluene

(20 cm³) was added DIBAL (1 M in toluene, 4 cm³) at 0 ^◦C

under N₂for about 1 h, with the excess DIBAL destroyed by

adding MeOH (4 cm³) dropwise. The resulting suspension was

then extracted with excess EtOAc (2 ¥ 200 cm³), washed with dil.

HCl solution (1 ¥ 100 cm³), brine (1 ¥ 100 cm³) and dried (Na₂SO₄).

The combined EtOAc solubles were then evaporated in vacuo

with the corresponding residue puriﬁed by silica gel CC using

EtOAc–acetone (9 : 1) as eluant to afford threo/erythro-5OHG–

G benzodioxane dimer (32) in ~9 : 1 ratio (0.396 g, 1.05 mmol,

88%).³⁹The ¹H and ¹³C NMR spectra of threo/erythro-32

[(E)-4-(3-(hydroxymethyl)-7-(3-hydroxyprop-1-enyl)-5-methoxy-

2,3-dihydrobenzo[b][1,4]dioxin-2-yl)-2-methoxyphenol] were in

close agreement with the reported literature data.³⁹(32) m/z (LC-

APCI) 373.0 (M - H. C₂₀H₂₁O₇requires 373.1), (t_R= 2.05 min);

HRMS (EI) 397.1259 (M + Na. C₂₀H₂₂O₇Na requires 397.1263).

Chiral HPLC separation of ( )-threo/erythro-32 was achieved as

described above in the ‘Instrumentation and Chromatography

Materials’ section resulted in the separation of (-)-threo-32 [t_R

= 46.4 min; [a]²_D⁰-9.39 (c 0.00149 in MeOH)], (+)-threo-32 [t_R

= 57.7 min; [a]²_D⁰+8.39 (c 0.00143 in MeOH)] and ( )-erythro-

m/z (LC-APCI) 431.1 (M - H. C₂₂H₂₃O₉requires 431.1) (t_R

=

2.5 min).

1

The H and ¹³C NMR spectra of ( )-threo-5OHG-S benzo-

dioxane diol [(E)-4-(3-(hydroxymethyl)-7-(3-hydroxyprop-1-enyl)-

5-methoxy-2,3-dihydrobenzo[b][1,4]dioxin-2-yl)-2,6-dimethoxy-

phenol] (34) were in close agreement with the reported literature

data:^40–43m/z (LC-APCI) 403.0 (M - H. C₂₁H₂₃O₈requires 403.1

(t_R= 2.0 min); HRMS (EI) 427.1400 (M + Na. C₂₁H₂₄O₈Na

requires 427.1369). Chiral separation: (-)-threo 34: t_R= 77.8 min,

[a]²_D⁰-2.96 (c 0.00135 in MeOH); (+)-threo 34: t_R= 100.6 min,

[a]²_D⁰+2.48 (c 0.00161 in MeOH). The ¹H NMR and mass spectra

of (+)-and (-)-threo-34 were identical with ( )-threo-34.

5OHG–5OHG benzodioxane dimer (43). Three different

synthetic protocols were employed.

Protocol 1. To a solution of methyl 5-hydroxyferulate (37,

0.8 g, 3.57 mmol) and O-tert-butyldimethylsilyl coniferyl alcohols

(41a/b, 1.3 g, 4.19 mmol) in benzene–acetone (2 : 1) was added

Ag₂CO₃(2.7 g, 9.8 mmol) with continuous stirring at room

temperature for 15 h under a N₂atmosphere. After completion

of the reaction, the Ag₂CO₃was removed by ﬁltration through

a Celite pad with the latter washed with acetone (40 cm³). The

resulting ﬁltrate was evaporated to dryness in vacuo, with the

residue 42a/b (1.30 g, 2.4 mmol, 57%) so obtained reduced

directly with DIBAL (1 M in toluene, 11 cm³) at 0 ^◦C under

N₂for about 1 h. The resulting suspension was then extracted

with excess EtOAc (2 ¥ 400 cm³), washed with dil. HCl solu-

tion (1 ¥ 150 cm³), brine (1 ¥ 150 cm³) and dried (Na₂SO₄).

The combined EtOAc solubles were next evaporated in vacuo

with the corresponding residue so obtained latter deprotected

with tetrabutylammonium ﬂuoride (TBAF)⁵²(1 M in THF,

5 cm³) and puriﬁed by silica gel CC eluted with EtOAc–acetone

(8 : 2), to afford the ( )-threo-5OHG–5OHG benzodioxane dimer

(43, 0.77g, 1.97 mmol, 82%) [(E)-5-(3-(hydroxymethyl)-7-(3-

hydroxyprop-1-enyl)-5-methoxy-2,3-dihydrobenzo[b][1,4]dioxin-

2-yl)-3-methoxybenzene-1,2-diol]:^40,43d_H(500 MHz; acetone-d₆):

7.67 (2 H, br s, ArOH), 6.7 (1 H, d, J 1.5, 2¢-H), 6.67 (1 H, d,

J 1.5, 6¢-H), 6.65 (1 H, d, J 1.5, 2-H), 6.59 (1 H, d, J 2.0, 6-H),

6.47 (1 H, d, J 16.0, 7¢-H), 6.27 (1 H, dt, J 5.5 and 15.5, 8¢-H), 4.9

(1 H, d, J 8.0, 7-H_threo), 4.20 (2 H, d, J 5.0, 9¢-H), 4.0 (1 H, m, 8-H),

3.85 (3 H, s, 3¢-OMe), 3.83 (3 H, s, 3-OMe), 3.78 (1 H, dd, J 2.0

and 12.25, 9-H), 3.52 (1 H, dd, J 3.5 and 12.5, 9-H); d_C(125 MHz;

acetone-d₆) 150.15 (C-3¢), 149.1 (C-3), 146.3 (C-5), 145.5 (C-5¢),

135.2 (C-4), 134.1 (C-1¢), 130.6 (C-7¢), 130.1 (C-4¢), 129.4 (C-1),

128.8 (C-8¢), 109.39 (C-6), 108.7 (C-6¢), 103.86 (C-2), 103.56 (C-

2¢), 79.5 (C-8), 77.2 (C-7), 63.3 (C-9¢), 61.8 (C-9), 56.6 (C-3-OMe),

56.3 (C-3¢-OMe); m/z (LC-APCI) 388.9 (M - H. C₂₀H₂₁O₈requires

389.1) (t_R= 2.4 min); HRMS (EI) 413.1207 (M + Na. C₂₀H₂₂O₈Na

requires 413.1212). Chiral HPLC analysis (see Instrumentation

1

32 [t_R= 75.1 min]. The H NMR and mass spectra of (+)- and

(-)-threo-32 were identical with ( )-threo-32 above.

( )-erythro-32: d_H(500 MHz; acetone-d₆) 7.70 (1 H, br s, ArOH),

7.0 (1 H, d, J 2.0, 2-H), 6.88 (1 H, dd, J 2.0 and 8.0, 6-H), 6.83 (1

H, d, J 8.0, 5-H), 6.72 (1 H, d, J 2.0, 2¢-H), 6.62 (1 H, d, J 1.5, 6¢-

H), 6.48 (1 H, d, J 16.0, 7¢-H), 6.28 (1 H, dt, J 5.5 and 16.0, 8¢-H),

5.26 (1 H, d, J 2.5, 7-H_erythro), 4.48 (1 H, m, 8-H), 4.20 (2 H, dd,

J 1.0 and 5.5, 9¢-H), 3.85 (3 H, s, 3-OMe), 3.80 (3 H, s, 3¢-OMe),

3.64 and 3.48 (2 H, m, 9-H); d_C(150 MHz; CDCl_3;Shigemi NMR

tube) 149.5 (C-3¢), 147.0 (C-3), 145.9 (C-4), 144.0 (C-5¢), 131.06

(C-4¢), 131.01 (C-1¢), 129.9 (C-7¢), 127.7 (C-1), 127.6 (C-8¢), 119.2

(C-6), 114.8 (C-5), 108.6 (C-2), 108.5 (C-6¢), 103.0 (C-2¢), 78.0

(C-8), 75.57 (C-7), 63.94 (C-9¢), 59.1 (C-9), 56.3 (C-3-OMe), 56.2

(C-3¢-OMe); m/z (LC-APCI) 373.1 (M - H. C₂₀H₂₁O₇requires

373.1) (t_R= 1.95 min).

threo-5OHG–S benzodioxane dimer (34). The threo-5OHG–S

benzodioxane dimer 34 was prepared exactly as for compound

32 using the Ag₂CO₃(0.7 g, 2.52 mmol) coupling of methyl 5-

hydroxy ferulate (37, 0.25 g, 1.11 mmol) and sinapyl alcohol (3,

0.2 g, 0.95 mmol) to initially afford the methyl ester of the 5OHG–

S benzodioxane dimer (40, 0.26 g, 0.6 mmol, 63%). The latter

(0.20 g, 0.46 mmol) was then reduced and puriﬁed as described

above for compound 39 by addition of DIBAL (1 M in toluene,

1.5 cm³) to afford the threo-5OHG–S benzodioxane dimer (34,

0.16 g, 0.39 mmol, 85%).

threo-5OHG–S benzodioxane methyl ester [(E)-methyl 3-(3-

(4-hydroxy-3,5-dimethoxyphenyl)-2-(hydroxymethyl)-8-methoxy-

2,3-dihydrobenzo[b][1,4]dioxin-6-yl)acrylate] (40). d_H(600 MHz;

acetone-d₆and 3 drops CD₃OD): 7.55 (1 H, d, J 16.2, 7¢-H), 6.98

This journal is

The Royal Society of Chemistry 2010

Org. Biomol. Chem., 2010, 8, 3928–3946 | 3941

©

View Article Online

and Chromatography Materials section) resulted in the separation

of (-)-threo-43 [t_R= 84.2 min; [a]²_D⁰-4.7 (c 0.00234 in MeOH)] and

(+)-threo-43 [t_R= 115.1 min; [a]²_D⁰+4.63 (c 0.00216 in MeOH)].

residue (~ 0.9 g) was puriﬁed by silica gel CC using hexanes–ethyl

acetate (2 : 8) as eluant to afford threo/erythro-5OHG–5OHG

benzodioxane methyl ester in 9.5 : 0.5 ratio (44, 0.69 g, 1.65 mmol,

88%). Further, to a solution of ester 44 (0.68 g, 1.62 mmol) in

benzene–acetone (2 : 1) was added coniferyl alcohol (2, 0.3 g, 1.66

mmol) and Ag₂CO₃(1.14 g, 4.15 mmol). After completion of

the reaction, the reaction mixture was puriﬁed by silica gel CC

using hexanes–EtOAc (1 : 9) as eluant to afford the methyl ester of

5OHG–5OHG–G benzodioxane trimer (45, 0.58 g, 0.97 mmol,

58%). A portion of the latter ester 45 (0.3 g, 0.5 mmol) was

next reduced with DIBAL (2 cm³) to afford 5OHG–5OHG–G

benzodioxane trimer 46 (0.22 g, 0.40 mmol, 80%).

1

The H NMR and mass spectra of (+)-threo-43 and (-)-threo-43

were identical with ( )-threo-43.

Protocol 2. To a solution of 5-hydroxyconiferyl alcohol (4,

53 mg, 0.27 mmol) in benzene and acetone (2 : 1, 15 cm³) was added

Ag₂CO₃(90 mg, 0.32 mmol), with the suspension then stirred

at room temperature under N₂for 3 h. The contents were next

ﬁltered over a Celite pad, washed with acetone (20 cm³), and dried

in vacuo. The residue so obtained was subjected to preparative

silica gel TLC (eluant, CHCl₃–MeOH, 9 : 1) to afford ( )-threo-

5OHG–5OHG (43) (4.5 mg, 0.011 mmol, 4.1%): The ¹H and

¹³C NMR spectra of ( )-threo-43 were identical to that described

in Protocol 1; m/z (LC-APCI) 388.9 (M - H. C₂₀H₂₁O₈requires

threo/erythro-5OHG–5OHG benzodioxane methyl ester [(E)-

methyl

3-(3-(3,4-dihydroxy-5-methoxyphenyl)-2-(hydroxyme-

thyl)-8-methoxy-2,3-dihydrobenzo[b][1,4]dioxin-6-yl-)-acrylate]

(44): d_H(500 MHz; acetone-d₆; Me₄Si) 7.68 (2 H, br s, ArOH),

7.56 (1 H, d, J 16.0, 7¢-H), 6.98 (1 H, d, J 1.5, 2¢-H), 6.89 (1 H,

d, J 2.0, 6¢-H), 6.7 (1 H, d, J 2.0, 2-H), 6.67 (1 H, d, J 2.0, 6-H),

6.44 (1 H, d, J 16.0, 8¢-H), 5.24 (1 H, d, J 2.5, 7-H_erythro), 4.94 (1

H, d, J 8.0, 7-H_threo), 4.09 (1 H, m, 8-H), 3.91 (3 H, s, 3¢-OMe),

3.84 (3 H, s, 3-OMe), 3.83 (1 H, m, 9-H), 3.73 (3 H, s, COOMe),

3.54 (1 H, br dd, J 4.5 and 12.5, 9-H); d_C(125 MHz; acetone-d₆;

Me₄Si) 167.8 (C-9¢, COOMe), 150.4 (C-3¢), 149.1 (C-3), 146.37

(C-5), 145.6 (C-5¢), 145.61 (C-7¢), 136.8 (C-4), 135.3 (C-4¢), 128.47

(C-1¢), 127.5 (C-1), 116.8 (C-8¢), 111.3 (C-6), 109.43 (C-6¢), 105.15

(C-2), 103.9 (C-2¢), 79.8 (C-8), 77.21 (C-7), 61.73 (C-9), 56.63

(C-3-OMe), 56.60 (C-3¢-OMe), 51.6 (CO-OMe); m/z (LC-APCI)

416.9 (M - H. C₂₁H₂₁O₉requires 417.1) (t_R= 2.2 min).

389.1) (t_R= 2.4 min). Chiral HPLC analysis; (-)-threo-43: t_R

=

84.0 and (+)-threo-43: t_R= 114.5 min.

Protocol 3: enzymatic synthesis. To 5-hydroxyconiferyl alcohol

(4, 55 mg, 0.28 mmol) was added acetone (20 cm³) and 0.2 M citric

acid/_◦phosphate buffer (pH 6.0, 5 cm³) with the contents cooled

to 0 C.⁵⁶Two drops of 30% H₂O₂and 7.3 mg of horseradish

peroxidase type (II) (1100 units mg^-1) in H₂O (3 cm³) were then

added every 15 min, with the mixture stirred at 0 ^◦C for 2 h. The

progress of the reaction was monitored by reversed-phase HPLC

(see Instrumentation and Chromatography Materials section).

The reaction mixture was then warmed to room temperature,

extracted with EtOAc (2 ¥ 50 cm³), washed with brine (1 ¥ 25 cm³),

dried (Na₂SO₄) and evaporated to dryness in vacuo. The residue

so obtained was subjected to preparative silica gel TLC eluted

with CHCl₃–MeOH (9 : 1), to afford 5OHG–5OHG benzodioxane

diol: ( )-threo-43 (9.8 mg, 0.024 mmol, 8.6%) and ( )-erythro-43

(1.2 mg, 0.003 mmol, 1.1%), respectively.

5OHG–5OHG–G benzodioxane methyl ester [(E)-methyl

3-(3¢-(4-hydroxy-3-methoxyphenyl)-2¢,3-bis(hydroxymethyl)-5,8¢-

dimethoxy-2,2¢,3,3¢-tetrahydro-2,6¢-bibenzo[b][1,4]dioxin-7-yl)-

acrylate] (45): d_H(500 MHz; acetone-d₆; Me₄Si) 7.77 (1 H, br s,

ArOH), 7.56 (1 H, d, J 16.0, 7¢¢-H), 7.13 (1 H, dd, J 1.5 and 4.5,

2-H), 6.98 (1 H, dd, J 1.5 and 8.25, 6-H), 6.89 (1 H, dd, J 3.5

and 8.0, 5-H), 6.80 (2 H, br d, J 1.0, 2¢-H and 6¢-H), 6.73 (1 H,

d, J 2.0, 2¢¢-H), 6.72 (1 H, d, J 2.0, 6¢¢-H), 6.43 (1 H, dd, J 1.5

and 16.0, 8¢¢-H), 5.0 (2 H, d, J 8.0, 7-H and 7¢-H), 4.13 (2 H, m,

8-H and 8¢-H), 3.91 (3 H, s, 3-OMe), 3.88 (3 H, s, 3¢-OMe), 3.87

(3 H, s, 3¢¢-OMe), 3.82 (2 H, dd, J 3.5 and 12.5, 9-H and 9¢-H),

3.72 (3 H, s, COOMe), 3.54 (2 H, dd, J 3.0 and 12.5, 9-H and

9¢-H); d_C(125 MHz; acetone-d₆; Me₄Si) 167.7 (C-9¢¢, COOMe),

150.3, 150.10 (d), 148.47 (d), 148.0, 145.48 (d), 145.40 (d), 136.6,

134.82 (d), 129.54 (d), 129.18, 127.5, 121.59 (d), 116.8, 115.74 (d),

111.90 (d), 111.2, 109.9 (d), 105.10 (d), 105.0, 79.53 (d), 79.47 (d),

77.09 (d), 76.85 (d), 61.76, 61.63, 56.45 (br s), 56.34 (d), 51.59

(COOMe); m/z (LC-APCI) 595.1 (M - H. C₃₁H₃₁O₁₂requires

595.2) (t_R= 2.5 min).

( )-threo-43: the ¹H and ¹³C NMR spectra were identical to that

described in Protocol 1; m/z (LC-APCI) 388.9 (M - H. C₂₀H₂₁O₈

requires 389.1) (t_R= 2.4 min). Chiral HPLC analysis; (-)-threo-43:

t_R= 84.6 and (+)-threo-43: 115.1 min.

( )-erythro-43: d_H(500 MHz; acetone-d₆) 7.61 (2 H, br s, ArOH),

6.71 (1 H, d, J 1.5, 2¢-H), 6.62 (1 H, d, J 1.5, 6¢-H), 6.62 (1 H, d,

J 2.0, 2-H), 6.59 (1 H, d, J 1.5, 6-H), 6.49 (1 H, d, J 16.0, 7¢-H),

6.28 (1 H, dt, J 5.5 and 16.0, 8¢-H), 5.2 (1 H, d, J 3.0, 7-H), 4.46

(1 H, m, 8-H), 4.2 (2 H, d, J 4.5, 9¢-H), 3.85 (3 H, s, 3¢-OMe), 3.79

(3 H, s, 3-OMe), 3.62 (1 H, br dd, J 5.0, 7.5 and 12.0, 9-H), 3.49

(1 H, dd, J 5.0 and 12.0, 9-H); d_C(125 MHz; acetone-d₆; Me₄Si)

150.4 (C-3¢), 148.86 (C-3), 146.32 (C-5), 144.91 (C-5¢), 134.6 (C-4),

132.84 (C-1¢), 130.9 (C-7¢), 130.0 (C-4¢), 129.5 (C-1), 128.4 (C-8¢),

108.66 (C-6), 108.1 (C-6¢), 103.85 (C-2), 103.1 (C-2¢), 78.48 (C-8),

76.56 (C-7), 63.26 (C-9¢), 59.88 (C-9), 56.47 (C-3-OMe), 56.41 (C-

3¢-OMe); m/z (LC-APCI) 388.9 (M - H. C₂₀H₂₁O₈requires 389.1)

(t_R= 2.4 min). Chiral HPLC analysis; ( )-erythro-43: t_R= 83.6

and 90.6 min, respectively.

5OHG-5OHG-G benzodioxane triol [(E)-(3¢-(4-hydroxy-3-

methoxyphenyl)-7-(3-hydroxyprop-1-enyl)-5,8¢-dimethoxy-2,2¢,3

,3¢-tetrahydro-2,6¢ - bibenzo[b][1,4]dioxine - 2¢,3diyl - )dimethanol]

(46):⁴⁴d_H(500 MHz; acetone-d₆): 7.77 (1 H, br s, ArOH), 7.12 (1 H,

dd, J 2.0 and 4.5, 2-H), 6.97 (1 H, d, J 8.0, 6-H), 6.88 (1 H, d, J 8.0,

5-H), 6.78 (1 H, br s, 2¢-H), 6.71 (1 H, dd, J 2.0 and 4.5, 2¢¢-H), 6.70

(1 H, dd, J 2.0 and 4.5, 6¢¢-H), 6.60 (1 H, d, J 2.0, 6¢-H), 6.47 (1H,

d, J 16.0, 7¢¢-H), 6.27 (1 H, dt, J 5.5 and 16.0, 8¢¢-H), 4.99 (1 H, d, J

8.0, 7-H), 4.97 (1 H, d, J 7.5, 7¢-H), 4.20 (2 H, t, J 5.0, 9¢-H), 4.08 (2

H, m, 8-H and 8¢-H), 3.88 (3 H, s, 3-OMe), 3.87 (3 H, s, 3¢-OMe),

3.86 (3 H, s, 3¢¢-OMe), 3.82 (2 H, dd, J 4.5 and 12.0, 9-H and

9¢-H), 3.54 (2 H, dd, J 5.5 and 12.5, 9-H and 9¢-H); d_C(125 MHz;

5OHG–5OHG–G benzodioxane trimer (46). To a solution of

42a/b (1 g, 1.87 mmol) in dry THF (30 cm³) was added TBAF

(1 M in THF, 4 cm³),⁵²with the contents allowed to stir under N₂

in an ice-bath for 30 min. After completion of the reaction, the

reaction mixture was quenched with saturated aqueous NH₄Cl

(100 cm³), with the whole extracted with EtOAc (3 ¥ 100 cm³).

The combined organic fractions were washed with brine (1 ¥

100 cm³), dried (Na₂SO₄) and evaporated to dryness in vacuo. The

3942 | Org. Biomol. Chem., 2010, 8, 3928–3946

This journal is

The Royal Society of Chemistry 2010

©

View Article Online

acetone-d₆): 150.07, 148.48, 148.47, 148.0, 145.43, 145.40, 145.26,

134.73, 134.71, 134.0, 130.6, 130.0, 129.91, 129.88, 129.4, 129.2,

121.60 (d), 115.75, 115.73, 111.91, 111.88, 109.87, 108.6 (br d),

105.07, 104.96, 103.5, 79.47, 79.45, 79.28, 79.24, 77.11, 77.06,

76.89, 76.81, 63.27, 61.77 (d), 56.44, 56.34 (d), 56.28; m/z (LC-

APCI) 567.0 (M - H. C₃₀H₃₁O₁₁requires 567.2) (t_R= 2.0 min);

HRMS (EI) 591.1846 (M + Na. C₃₀H₃₂O₁₁Na requires 591.1842).

Chiral HPLC analysis; t_R= 162.6, 192.0 (br) min and 254.2, 292.85

(br) min respectively; the enantiomers were not baseline separable.

dioxine-2¢,3-diyl)dimethanol] (36). Thioacidolysis of the 5OHG–

5OHG–G benzodioxane trimer (46) (50 mg, 0.088 mmol) was

carried out exactly as described above for 32, with the crude

thioacidolysis reaction products puriﬁed by silica gel TLC (eluant,

EtOAc–acetone, 8 : 2) and further puriﬁed by reversed-phase

HPLC to afford the threo/erythro-mixtures of G triethanethiol

(19, 9.8 mg, 0.028 mmol, ~32%), 5OHG–5OHG–G benzodioxane

triethanethiol (36, 3.3 mg, 0.0045 mmol, ~5%) and 5OHG–G

benzodioxane triethanethiol (35, 6 mg, 0.011 mmol, ~13%),

respectively.

threo and erythro 5OHG–G benzodioxane triethanethiol.

threo- and erythro-36: d_H(500 MHz; acetone-d₆) 7.75 (1 H, br s,

ArOH), 7.12 (1 H, dd, J 1.5 and 5.5, 2-H), 6.97 (1 H, dd, J 1.5

and 8.0, 6-H), 6.86 (1 H, dd, J 1.5 and 8.0, 5-H), 6.77 (2 H, br s,

2¢-H and 6¢-H), 6.719 (1 H, dd, J 1.5 and 2.0, 2¢¢-H), 6.717 (1 H,

dd, J 2.0, 6¢¢-H), 5.0 (1 H, d, J 8.0, 7-H), 4.97 (1 H, d, J 7.5,

7¢-H), 4.39 - 4.34 (1 H, m, 7¢¢-H), 4.08 (2 H, m, 8-H, 8¢-H), 3.877

(3 H, br s, OMe), 3.866 (3 H, br s, OMe), 3.834 (3 H, br s, OMe),

3.81 (2 H, m, 9-H and 9¢-H), 3.54 (2 H, m, 9-H and 9¢-H), 3.22

- 3.04 (1 H, m, 8¢¢-H), 3.02 - 2.70 (2 H, m, 9¢¢-H), 2.63 - 2.33

(6 H, m, 3 ¥ –SCH₂Me), 1.24 - 1.09 (9 H, m, 3 ¥ –SCH₂Me);

d_C(125 MHz; acetone-d₆) 150.15 (t), 149.51, 149.47, 148.55 (d),

148.07, 145.52 (d), 145.47 (d), 144.77 (d), 144.73 (d), 134.79 (d),

134.30 (d), 133.69 (t), 132.57, 132.50 (d), 129.97, 129.92, 129.27

(d), 121.66 (d), 115.80 (d), 111.98, 111.93 (d), 111.40 (br d), 110.89,

110.85 (d), 109.95 (d), 106.73, 106.26, 105.17, 105.13 (d), 105.03

(d), 79.54 (d), 79.31 (d), 79.26 (d), 77.19, 77.12, 76.96 (d), 76.89

(d), 61.84 (C-9 and C-9¢), 56.52, 56.45 (d), 56.41 (d), 53.86 (d),

53.78, 53.70, 53.18 (d), 52.58 (d), 37.53 (d), 37.43, 27.40, 27.35,

27.22 (d), 27.00 (dd), 26.42, 26.30 (d), 15.47, 15.41, 15.25, 15.10

(d), 14.99; m/z (LC-APCI) 733.1 (M - H. C₃₆H₄₅O₁₀S₃requires

733.2) (t_R= 5.6 min); 672.9 (M + H - HSEt)⁺, 611.1 (M + H - 2

HSEt)⁺.

[4-(3-(Hydroxymethyl)-5-methoxy-7-(1,2,3-tris(ethylthio)propyl)-

2,3-dihydrobenzo[b][1,4]dioxin-2-yl)-2-methoxyphenol]

(35).

Thioacidolysis of compound 32 was performed using the

procedure described earlier^38,45with the following modiﬁcations.

5OHG–G benzodioxane dimer 32 (5 ¥ 5 mg each, 0.013 mmol

each) was dissolved in dry dioxane (1 cm³) of which an aliquot

(100 mm³) was added to a 20 cm³glass tube ﬁtted with a Teﬂon-

lined screwcap containing the thioacidolysis reagent solution

[15 cm³, 0.2 M BF₃·Et₂O in dioxane–ethanethiol (8.75 : 1, v/v)].

◦

The tubes were then heated at 100 C for 4 h with concomitant

magnetic stirring of the solution. After completion of the

reaction, the reaction mixture vessel was cooled using ice–H₂O,

with the reaction mixture then poured into CH₂Cl₂(40 cm³);

the reaction tubes were next rinsed with H₂O (2 ¥ 10 cm³). The

pH of the combined aqueous solubles was then adjusted to

3–3.2 with 0.4 M NaHCO₃with the CH₂Cl₂solubles separated

and the aqueous phase re-extracted with CH₂Cl₂(2 ¥ 25 cm³).

The combined organic solubles were next dried (Na₂SO₄) and

concentrated in vacuo, with the residue puriﬁed by a combination

of preparative silica gel TLC eluted with hexanes–EtOAc (7 : 3,

1 : 1 and 3 : 7, respectively) and further puriﬁed by reversed-phase

HPLC (see instrumentation and chromatography materials

section), to afford threo/erythro G triethanethiol in ꢁ1.2 : 1 ratio

(19, 9.1 mg, 0.026 mmol, ~40%) and threo/erythro 5OHG-G

benzodioxane triethanethiol in ꢁ1.7 : 1 ratio (at C7–C8) (35,

4.1 mg, 0.0076 mmol, ~12%), respectively.

Time course of thioacidolytic cleavage of 5OHG–G benzodioxane

dimer (34) and 5OHG–5OHG–G benzodioxane trimer (46)

The 5OHG–G benzodioxane dimer (32) (3.1 mmol) was initially

subjected to standard quantitative thioacidolytic conditions [(+)-

sesamin as internal standard, C₂₀H₁₈O₆, ca 75 mg],^45,51while

quenching the reaction at hourly intervals from 0 to 5 h. With

the resulting thioacidolysis reaction mixture dissolved in THF

(10 cm³), aliquots (1 mm³) were then subjected to reversed-phase

UPLC analyses (BEH C18 column, Waters, 2.1 ¥ 150 mm) with

elution conditions at a ﬂow rate of 0.2 cm³min^-1: linear gradients

of MeOH-3% (v/v) AcOH in H₂O from 40 : 60 to 70 : 30 in 3 min,

then to 80 : 20 in 4 min, to 100 : 0 in 2.25 min with this composition

held for an additional 0.25 min, and ﬁnally to 40 : 60 in 0.5 min

this being held for 5 min). The 5OHG–5OHG–G benzodioxane

trimer (46) (2.7 mmol) was then subjected to standard quantitative

thioacidolytic conditions exactly as described above for 5OHG–

G benzodioxane dimer (32). The various thioacidolysis products

from 32 and 46 were identiﬁed by comparison to retention times

of authentic standards and by mass spectrometric analyses. The

chemical characterization of intermediate 5OHG–G benzodiox-

ane monoethanethiol 47, as obtained from 5OHG–G dimer (32)

at the early phase thioacidolysis reaction, was established by 1D

¹H and ¹³C NMR spectroscopy and mass spectral analysis. For

quantitative analyses, authentic standards G (19), 5OHG (20),

S (21), 5OHG–G (35), and 5OHG–5OHG–G (36) were used to

threo- and erythro-35: d_H(500 MHz; acetone-d₆; Me₄Si) 7.75 (1

H, br s, ArOH), 7.10 (1 H, br s, 2-H), 6.96 (1 H, d, J 8.0, 6-H),

6.88 (1 H, dd, J 1.5 and 8.5, 5-H), 6.78 (1 H, dd, J 2.0 and 7.0,

2¢-H), 6.70 (1 H, ddd, J 1.5 and 6.5, 6¢-H), 4.98 (1 H, d, J 7.5,

7-H), 4.97 (1 H, d, J 8.0, 7-H), 4.40-4.34 (1 H, m, 7¢-H), 4.08-4.06

(1 H, m, 8-H), 3.87 (3 H, s, 3-OMe), 3.83 (3 H, br s, 3¢-OMe), 3.79

(1 H, br d, J 2.5 and 12.5, 9-H), 3.50 (1 H, br d, J 3.5 and 12.0,

9-H), 3.24–3.05 (1 H, m, 8¢-H), 3.03-2.71 (2 H, m, 9¢-H), 2.64-

2.33 (6 H, m, 3 ¥ -SCH₂Me), 1.25-1.10 (9 H, m, 3 ¥ -SCH₂Me);

d_C(125 MHz; acetone-d₆; Me₄Si) 149.68, 149.62, 149.44, 149.40,

148.43 (d), 147.94 (d), 145.0 (d), 144.84 (d), 134.15, 133.71 (d),

133.60, 133.55, 132.4, 132.3, 129.3, 121.55 (d), 115.7, 111.90 (t),

111.38, 111.33, 110.85, 110.80, 110.82 (d), 106.6, 106.1, 79.39,

77.05 (d), 77.03 (d), 61.8, 56.38 (d), 56.33, 53.80 (d), 53.73, 53.6,

53.1, 52.5, 37.49, 37.44, 37.34 (d), 27.3, 27.27, 27.16, 27.12, 26.96,

26.91 (d), 26.34, 26.22 (d), 15.40 (d), 15.33, 15.17 (d), 15.05, 15.02

(d), 14.92 (d); m/z (LC-APCI) 538.9 (M - H. C₂₆H₃₅O₆S₃requires

539.2); 417.1 (M + H - 2HSEt)⁺, 478.9 (M + H - HSEt)⁺(t_R

=

5.7 min).

threo and erythro 5OHG–5OHG-G benzodioxane triethane-

thiol. [(3¢-(4-Hydroxy-3-methoxyphenyl)-5,8¢-dimethoxy-7-(1,2,

3-tris(ethylthio)propyl)-2,2¢,3,3¢-tetrahydro-2,6¢-bibenzo[b][1,4]-

This journal is

The Royal Society of Chemistry 2010

Org. Biomol. Chem., 2010, 8, 3928–3946 | 3943

©

View Article Online

calculate the corresponding response factors. All samples were

analyzed in duplicate.

and Fr-3 (17.7 mg) contained benzodioxane substructures [4.81

(7-H) and 76.0 (C-7)/4.0 (8-H) and 78.2 (C-8)]. These were

individually subjected to reversed-phase HPLC separation (see

Instrumentation and Chromatography Materials section) with

fractions corresponding to dimer 35 (from Fr-2) and trimer 36

(from Fr-3) individually pooled, extracted with EtOAc (40 cm³),

5OHG–G benzodioxane monoethanethiol [(E)-4-(7-(3-(ethyl-

thio)prop-1-enyl)-3-(hydroxymethyl)-5-methoxy-2,3-dihydro-

benzo[b][1,4]dioxin-2-yl)-2-methoxyphenol] (47): d_H(500 MHz;

acetone-d₆; Me₄Si) 7.77 (1 H, ArOH), 7.11 (1 H, d, J 2.0, 2-H),

6.96 (1 H, dd, J 2.0 and 8.0, 6-H), 6.88 (1 H, d, J 8.0, 5-H), 6.72 (1

H, d, J 1.5, 2¢-H), 6.60 (1 H, d, J 1.5, 6¢-H), 6.39 (1 H, d, J 15.5,

7¢-H), 6.12 (1 H, dt, J 7.5 and 16.0, 8¢-H), 4.98 (1 H, d, J 8.0, 7-H),

4.0 (1 H, m, 8-H), 3.87 (3 H, s, 3-OMe), 3.86 (3 H, s, 3¢-OMe), 3.78

and 3.50 (2 H, m, 9-H), 3.30 (2 H, d, J 7.5, 9¢-H), 2.49 (2 H, q,

J 7.5, 10¢-H), 1.21 (3 H, t, J 6.5 and 7.5, 11¢-H); d_C(125 MHz;

acetone-d₆; Me₄Si) 150.12 (C-3¢), 148.45 (C-3), 148.44 (C-4),

145.44 (C-5¢), 134.22 (C-4¢), 132.41 (C-7¢), 130.2 (C-1¢), 129.34

(C-1), 125.38 (C-8¢), 121.55 (C-6), 115.7 (C-5), 111.9 (C-2), 108.67

(C-6¢), 103.4 (C-2¢), 79.47 (C-8), 77.04 (C-7), 61.75 (C-9), 56.32 (C-

3-OMe), 56.31 (C-3¢-OMe), 34.19 (C-9¢), 25.0 (C-10¢), 14.97 (C-

11¢); m/z (LC-APCI) 417.0 (M - H. C₂₂H₂₆O₆S requires 417.1) (t_R=

3.7 min).

1

evaporated to dryness in vacuo and freeze-dried. The 1D H and

¹³C NMR spectra of 35 and 36 contained identical chemical shift

resonance values to synthetic 35 and 36, respectively.

5OHG–G benzodioxane triethanethiol (35, 0.8 mg); m/z (LC-

APCI) 478.9 (M + H - HSEt)⁺and 417.1 (M + H - 2HSEt)⁺(t_R=

5.7 min).

5OHG–5OHG–G benzodioxane triethanethiol (36, 3.5 mg);

m/z (LC-APCI) 672.9 (M + H - HSEt)⁺and 611.1 (M + H -

2HSEt)⁺(t_R= 5.6 min).

Arabidopsis COMT gene cloning, heterologous expression,

puriﬁcation and kinetic analyses

Twelve of 17 putative COMTs (AtCOMT1, 3, 4, 6, 8–13, 15

and 17)⁴⁶were isolated from Arabidopsis (ecotype: Columbia)

as follows: After isolation of total RNA using RNeasy Mini kit

(Qiagen), single-strand cDNA was generated with reverse tran-

scriptase II (Invitrogen) following the manufacturer’s instructions.

The twelve putative COMTs were then individually ampliﬁed with

PCR; their products were cloned using the TOPO TA cloning

kit (Invitrogen) and their sequences were conﬁrmed by individual

sequencing.

Isolation and molecular weight distributions of lignin-derived

isolates from WT and Atomt1 lines

Whole stem sections from Atomt1 mutant and WT lines were

subjected to a modiﬁed Bjo¨rkmann⁵⁷lignin isolation procedure

as described by Bernard-Vailhe´ et al.⁵⁸The extractive-free CWRs

obtained from Atomt1 mutant (6.95 g) and WT (12.15 g) lines,

were individually ground by ball-milling for 96 h at 4 ^◦C and

then extracted twice using a mixture of dioxane–water (96 : 4,

v/v, 30 cm³g^-1) for 24 h. The crude lignins thus obtained for

Atomt1 mutant (362 mg) and WT (710 mg) were then individually

subjected to various precipitation conditions as described by

Jourdes et al.²⁷to furnish the corresponding lignin-enriched

fractions: Atomt1 mutant (157 mg, 11.6%); l_max(dioxane)/nm 281

(e/g^-1cm^-115.7 0.11). WT (282 mg, 10.8%); l_max(dioxane)/nm

281 (e/g^-1cm^-116.6 0.16). Molecular weight distributions of

both lignin-enriched isolates were estimated by gel permeation

chromatography on Sephadex G-100 column (100 cm ¥ 2.5 cm)

eluted with 0.1 M NaOH at 40 cm³h^-1and precalibrated with

sodium polystyrene sulfonates (Mw’s: ~1430; 5180; 29 000; 79 000)

and coniferyl alcohol (2, 180) as previously described.³²

For heterologous protein expression, each gene was re-ampliﬁed

with directional nucleotide containing forward primer. The prod-

ucts were then directionally cloned into a pET101/D-TOPO

expression vector. After sequence conﬁrmation, each of the

vectors was individually transformed into E. coli expression cell

[BL21 Star (DE3)]. E. coli cells were cultured in Luria-Bertani

(LB) broth supplemented with carbenicillin (0.1 mg cm^-3) at

37 ^◦C. When cell growth reached an OD₆₀₀value between 0.5 and

0.9, 0.5 mM isopropyl thio-b-D-galactoside (IPTG) was added

for induction. Cells were individually harvested by centrifugation

at 3,000 ¥ g for 20 min and stored at -80 ^◦C until needed.

Pellets were thawed, suspended in a 1 : 1 mixture of Tris-HCl

(20 mM; pH 7.5, buffer A) and BugBuster^TMProtein Extraction

reagent. Cellular debris was removed by centrifugation (15 000 ¥

g, 30 min), with the resulting pellet resuspended in Buffer A

(20 mM Tris-HCl, pH 7.9; 500 mM NaCl; 5 mM imidazole). Each

suspension was ﬁltered (0.45 mm) and applied to a POROS 20 MC

column (100 mm ¥ 4.6 mm) pre-equilibrated with Buffer A) at

a ﬂow rate of 10 cm³min^-1and eluted as previously described.⁵⁹

Aliquots (12 mm³) of every other fraction were analyzed by SDS-

polyacrylamide gel electrophoresis.⁶⁰Fractions containing His-

NMR spectroscopic analyses of lignin-enriched isolates from

Atomt1 mutant and WT lines

Each lignin isolate (~30 mg) of both WT and Atomt1 lines was

individually dissolved in DMSO-d₆(0.5 cm³), with the ¹³C NMR,

two-dimensional phase-sensitive gradient-selected, HMQC and

HMBC spectra individually recorded using similar acquisition

parameters as described earlier.³²

R

tagged AtCOMTs were pooled and concentrated using Centricon^ꢀ

Plus-80 concentrators, diluted with 20 mM Tris-HCl (pH 7.5) and

reconcentrated. Protein concentrations were determined using the

Bradford method,⁶¹with BSA as standard.

Thioacidolysis of lignin-enriched isolates harboring trimer (36) and

dimer (35) from Atomt1 mutant

Thioacidolytic cleavage of the Atomt1 lignin-enriched isolate

(90 mg) was carried out as described for 32 and 46. The crude

reaction mixture so obtained (120 mg) was ﬁrst subjected to

silica gel TLC eluted with hexanes–EtOAc (7 : 3), with isolated

bands individually subjected to NMR and mass spectroscopic

analyses. NMR (1D and 2D) spectra of two fractions Fr-2 (3.5 mg)

Enzyme assays for candidate AtCOMTs were performed as fol-

lows: Each assay (300 mm³) contained Tris-HCl buffer (100 mM,

pH 7.5), puriﬁed candidate AtCOMT (5 mg), substrates [caf-

feic (5)/5-hydroxyferulic (6) acid, caffeyl (9)/5-hydroxyconiferyl

(10) aldehyde and caffeyl (11)/5-hydroxyconiferyl (4) alcohol,

0.33 mM] and S-adenosylmethionine (0.33 mM). Assays were

3944 | Org. Biomol. Chem., 2010, 8, 3928–3946

This journal is

The Royal Society of Chemistry 2010

©

View Article Online

initiated by addition of each candidate AtCOMT, incubated for

10 min and stopped by acidiﬁcation (glacial AcOH, 10 mm³).

Controls included both boiled enzyme extracts and omission of

substrates. An aliquot (80 ml) of each assay mixture was then

subjected to HPLC analysis.

Note added in proof: A paper (Lu et al., Plant Physiology,

doi/10.1104/pp.110.154278) has just appeared on an incompletely

COMT-suppressed poplar line and analysis of the cell-wall

isolates.

17 D. Guo, F. Chen, K. Inoue, J. W. Blount and R. A. Dixon, Plant Cell,

2001, 13, 73–88.

18 S. Maury, P. Geoffroy and M. Legrand, Plant Physiol., 1999, 121, 215–

223.

19 L. Li, J. L. Popko, T. Umezawa and V. L. Chiang, J. Biol. Chem., 2000,

275, 6537–6545.

20 K. Parvathi, F. Chen, D. Guo, J. W. Blount and R. A. Dixon, Plant J.,

2001, 25, 193–202.

21 R. A. Emerson, G. W. Beadle and A. C. Fraser, Cornell University

Agricultural Experimental Station Memoir, 1935, 180, 1–83.

22 L. R. Jorgenson, J. Am. Soc. Agronomy, 1931, 23, 549–557.

23 L. D. Muller, R. F. Barnes, L. F. Bauman and V. F. Colenbrander, Crop

Sci., 1971, 11, 413–415.

Abbreviations

24 M. S. Zuber, T. R. Colbert and L. F. Bauman, Z. Pﬂanzen., 1977, 79,

310–314.

COMT, caffeic acid O-methyltransferase; CWR, cell-wall residue;

G, guaiacyl; S, syringyl; 5OHG, 5-hydroxyguaiacyl; WT, wild type;

HMQC, heteronuclear multiple quantum coherence; HMBC,

heteronuclear multiple bond coherence

25 C. Grand, P. Parmentier, A. Boudet and A. M. Boudet, Physiol. Veg.,

1985, 23, 905–911.

26 C. Lapierre, M.-T. Tollier and B. Monties, C.R. Acad. Sci., Ser. III,

1988, 307, 723–728.

27 M. Jourdes, C. L. Cardenas, D. D. Laskar, S. G. A. Moinuddin, L. B.

Davin and N. G. Lewis, Phytochemistry, 2007, 68, 1932–1956.

28 Y.-r. Chen and S. Sarkanen, Phytochem. Rev., 2003, 2, 235–255.

29 S.-Y. Guan, J. Mlyna´r and S. Sarkanen, Phytochemistry, 1997, 45, 911–

918.

Acknowledgements

This work was supported by the Chemical Sciences, Geosciences

and Biosciences Division, Ofﬁce of Basic Energy Sciences, Ofﬁce

of Science, U.S. Department of Energy, by the BioEnergy Science

Center, the U.S. Department of Energy Bioenergy Research Center

supported by the Ofﬁce of Biological and Environmental Research

in the DOE Ofﬁce of Science, as well as the G. Thomas and

Anita Hargrove Center for Plant Genomic Research. Thanks are

extended to Prof. Ferreira and Dr Bharathi Avula for carrying out

the HRMS analyses and to Dr Lise Jouanin for providing seeds of

the Atomt1 knockout.

30 S. Sarkanen, in Lignin and Lignan Biosynthesis, ed. N. G. Lewis and

S. Sarkanen, ACS Symposium Series, Washington, DC, USA, 1998,

pp. 194-208.

31 A. M. Patten, C. L. Cardenas, F. C. Cochrane, D. D. Laskar, D. L.

Bedgar, L. B. Davin and N. G. Lewis, Phytochemistry, 2005, 66, 2092–

2107.

32 D. D. Laskar, M. Jourdes, A. M. Patten, G. L. Helms, L. B. Davin and

N. G. Lewis, Plant J., 2006, 48, 674–686.

33 J. M. Marita, J. Ralph, R. D. Hatﬁeld, D. Guo, F. Chen and R. A.

Dixon, Phytochemistry, 2003, 62, 53–65.

34 J. Ralph, C. Lapierre, J. M. Marita, H. Kim, F. Lu, R. D. Hatﬁeld, S.

Ralph, C. Chapple, R. Franke, M. R. Hemm, J. Van Doorsselaere, R. R.

Sederoff, D. M. O’Malley, J. T. Scott, J. J. MacKay, N. Yahiaoui, A.-M.

Boudet, M. Pean, G. Pilate, L. Jouanin and W. Boerjan, Phytochemistry,

2001, 57, 993–1003.

Notes and references

35 J. Ralph, K. Lundquist, G. Brunow, F. Lu, H. Kim, P. F. Schatz, J. M.

Marita, R. D. Hatﬁeld, S. A. Ralph, J. H. Christensen and W. Boerjan,

Phytochem. Rev., 2004, 3, 29–60.

1 L. B. Davin, M. Jourdes, A. M. Patten, K.-W. Kim, D. G. Vassa˜o and

N. G. Lewis, Nat. Prod. Rep., 2008, 25, 1015–1090.

36 L. B. Davin and N. G. Lewis, Curr. Opin. Biotechnol., 2005, 16, 407–

2 L. B. Davin, A. M. Patten, M. Jourdes and N. G. Lewis, in Biomass

Recalcitrance. Deconstructing the Plant Cell Wall for Bioenergy, ed.

M. E. Himmel, Blackwell Publishing, Oxford, UK, 2008, pp. 213-305.

3 A. M. Patten, M. Jourdes, C. L. Cardenas, D. D. Laskar, Y. Nakazawa,

B.-Y. Chung, V. R. Franceschi, L. B. Davin and N. G. Lewis, Mol.

BioSyst., 2010, 6, 499–515.

415.

37 S. Dutta, T. M. Garver Jr and S. Sarkanen, ACS Symp. Ser., 1989, 397,

155–176.

38 K. Blee, J. W. Choi, A. P. O’Connell, S. C. Jupe, W. Schuch, N. G. Lewis

and G. P. Bolwell, Phytochemistry, 2001, 57, 1159–1166.

39 K. Morreel, J. Ralph, F. Lu, G. Goeminne, R. Busson, P. Herdewijn,

J. L. Goeman, J. Van der Eycken, W. Boerjan and E. Messens, Plant

Physiol., 2004, 136, 4023–4036.

4 M. Shimada, Wood Research, 1972, 53, 19–65.

5 M. Shimada, H. Fushiki and T. Higuchi, Mokuzai Gakkaishi, 1973, 19,

13–21.

40 T. H. Kim, H. Ito, K. Hayashi, T. Hasegawa, T. Machiguchi and T.

Yoshida, Chem. Pharm. Bull., 2005, 53, 641–644.

41 A. Kuboki, T. Yamamoto, M. Taira, T. Arishige and S. Ohira,

Tetrahedron Lett., 2007, 48, 771–774.

6 C. Hermann, M. Legrand, P. Geoffroy and B. Fritig, Arch. Biochem.

Biophys., 1987, 253, 367–376.

7 E. Jaeck, F. Martz, V. Stiefel, B. Fritig and M. Legrand, Mol. Plant-

Microbe Interact., 1996, 9, 681–688.

42 C. Ma, H. J. Zhang, G. T. Tan, N. V. Hung, N. M. Cuong, D. D. Soejarto

and H. H. S. Fong, J. Nat. Prod., 2006, 69, 346–350.

43 Y. K. Son, M. H. Lee and Y. N. Han, Arch. Pharmacal Res., 2005, 28,

34–38.

8 L. Pellegrini, P. Geoffroy, B. Fritig and M. Legrand, Plant Physiol.,

1993, 103, 509–517.

9 R. Atanassova, N. Favet, F. Martz, B. Chabbert, M.-T. Tollier, B.

Monties, B. Fritig and M. Legrand, Plant J., 1995, 8, 465–477.

10 U. N. Dwivedi, W. H. Campbell, J. Yu, R. S. S. Datla, R. C. Bugos,

V. L. Chiang and G. K. Podila, Plant Mol. Biol., 1994, 26, 61–71.

11 W. Ni, N. L. Paiva and R. A. Dixon, Transgenic Res., 1994, 3, 120–126.

12 A. M. Anterola and N. G. Lewis, Phytochemistry, 2002, 61, 221–294.

13 J. Van Doorsselaere, M. Baucher, E. Chognot, B. Chabbert, M.-T.

Tollier, M. Petit-Conil, J.-C. Leple´, G. Pilate, D. Cornu, B. Monties,

M. Van Montagu, D. Inze´, W. Boerjan and L. Jouanin, Plant J., 1995,

8, 855–864.

44 R. Waibel, G. Benirschke, M. Benirschke and H. Achenbach, Phyto-

chemistry, 2003, 62, 805–811.

45 C. Lapierre, B. Monties and C. Rolando, Holzforschung, 1986, 40, 113–

118.

46 M. A. Costa, R. E. Collins, A. M. Anterola, F. C. Cochrane, L. B.

Davin and N. G. Lewis, Phytochemistry, 2003, 64, 1097–1112.

47 I. Muzac, J. Wang, D. Anzellotti, H. Zhang and R. K. Ibrahim, Arch.

Biochem. Biophys., 2000, 375, 385–388.

48 T. Goujon, R. Sibout, B. Pollet, B. Maba, L. Nussaume, N. Bechtold,

F. Lu, J. Ralph, I. Mila, Y. Barrie`re, C. Lapierre and L. Jouanin, Plant

Mol. Biol., 2003, 51, 973–989.

14 C. Lapierre, B. Pollet, M. Petit-Conil, G. Toval, J. Romero, G. Pilate,

J.-C. Leple´, W. Boerjan, V. Ferret, V. de Nada¨ı and L. Jouanin, Plant

Physiol., 1999, 119, 153–163.

49 K. Iiyama and A. F. A. Wallis, J. Sci. Food Agric., 1990, 51, 145–161.

50 K. Iiyama and T. B. T. Lam, J. Sci. Food Agric., 1990, 51, 481–491.

51 C. Rolando, B. Monties and C. Lapierre, in Methods in Lignin

Chemistry, ed. S. Y. Lin and C. W. Dence, Springer-Verlag, Berlin,

Germany, 1992, pp. 334-349.

15 L. Jouanin, T. Goujon, V. de Nada¨ı, M.-T. Martin, I. Mila, C. Vallet,

B. Pollet, A. Yoshinaga, B. Chabbert, M. Petit-Conil and C. Lapierre,

Plant Physiol., 2000, 123, 1363–1373.

16 C.-J. Tsai, J. L. Popko, M. R. Mielke, W.-J. Hu, G. K. Podila and V. L.

Chiang, Plant Physiol., 1998, 117, 101–112.

This journal is

The Royal Society of Chemistry 2010

Org. Biomol. Chem., 2010, 8, 3928–3946 | 3945

©

View Article Online

52 S.-J. Kim, M.-R. Kim, D. L. Bedgar, S. G. A. Moinuddin, C. L.

Cardenas, L. B. Davin, C. Kang and N. G. Lewis, Proc. Natl. Acad.

Sci. U. S. A., 2004, 101, 1455–1460.

53 W. Gu, X. Jing, X. Pan, A. S. C. Chan and T.-K. Yang, Tetrahedron

Lett., 2000, 41, 6079–6082.

54 X. She, S. Qi, W. Gu and X. Pan, J. Chem. Res. (S), 1998, 436–437.

55 X. She, W. Gu, T. Wu and X. Pan, J. Chem. Res. (S), 1999, 100–

101.

56 N. R. Guz and F. R. Stermitz, J. Nat. Prod., 2000, 63, 1140–1145.

57 A. Bjo¨rkman, Nature, 1954, 174, 1057–1058.

58 M.-A. Bernard-Vailhe´, A. Cornu, D. Robert, M.-P. Maillot and J.-M.

Besle, J. Agric. Food Chem., 1996, 44, 1164–1169 .

59 F. C. Cochrane, L. B. Davin and N. G. Lewis, Phytochemistry, 2004,

65, 1557–1564.

60 U. K. Laemmli, Nature, 1970, 227, 680–685.

61 M. M. Bradford, Anal. Biochem., 1976, 72, 248–254.

3946 | Org. Biomol. Chem., 2010, 8, 3928–3946

This journal is

The Royal Society of Chemistry 2010

©

Article Doi

Insights into lignin primary structure and deconstruction from Arabidopsis thaliana COMT (caffeic acid O-methyl transferase) mutant Atomt1

DOI: 10.1039/c004817h

Source and publish data:

Authors:

Article abstract of DOI:10.1039/c004817h

Full text of DOI:10.1039/c004817h

Products guided by the article

R&D Labs maybe for 97141-24-7

Relevant to this article

Hot Product