A novel thiazolidinone herbicide is a potent inhibitor of glucose incorporation into cell wall material

Article abstract of DOI:10.1002/(SICI)1096-9063(199812)54:4<368::AID-PS845>3.0.CO;2-6

5-tert-Butyl-carbamoyloxy-3-(3-trifluoromethyl)phenyl-4-thiazolidinone (compound 1), a representative of a novel class of thiazolidinone herbicides, shows potential for the pre-emergence control of grasses and small-seeded broad-leaved weeds in crops such

Full text of DOI:10.1002/(SICI)1096-9063(199812)54:4<368::AID-PS845>3.0.CO;2-6

Pestic. Sci. 1998, 54, 368È376
A Novel Thiazolidinone Herbicide is a Potent
Inhibitor of Glucose Incorporation into Cell Wall
¤
Material
Kate R. Sharples, Tim R. Hawkes,* Glynn Mitchell, Lucretia S. Edwards,
Mike P. Langford, David W. Langton, Katherine M. Rogers, Jane K. Townson &
Youlin Wang
Zeneca Agrochemicals, JealottÏs Hill Research Station, Bracknell, Berkshire RG42 6ET, UK
(Received 18 June 1998; revised version received 19 August 1998; accepted 24 August 1998)
Abstract:
5-tert-Butyl-carbamoyloxy-3-(3-triÑuoromethyl)
phenyl-4-thi-
azolidinone (compound 1), a representative of a novel class of thiazolidinone
herbicides, shows potential for the pre-emergence control of grasses and small-
seeded broad-leaved weeds in crops such as soyabean. This compound power-
fully inhibited the growth and lateral branching of the roots of agar-grown
seedlings of susceptible species but had no apparent e†ect on the growth of
Escherichia coli, Aspergillus nidulans, Fusarium culmorum or of insects cells in
liquid culture. It inhibited the growth of plant (Daucus carota and Zea mays) cells
in liquid culture and this inhibition was not reversed by addition to the medium
of mixtures of either amino acids or of nucleosides. It did not inhibit fatty acid
biosynthesis, respiration, the biosynthesis of sterols, or the biosynthesis of protein
in the systems examined herein. By contrast, it induced potent (IC c.50 nM) and
50
rapid inhibition of the incorporation of [3H]glucose into the acid-insoluble cell-
wall fraction of roots. Thus, compound 1 exerts its herbicidal e†ect, directly or
indirectly, through inhibition of the biosynthesis of cellulose and cellulose-like
polysaccharides, in a manner similar to isoxaben and dichlobenil. Mutant lines of
Arabidopsis thaliana selected for resistance to isoxaben were cross-resistant to
compound 1. Consistent with the selective herbicidal activity observed, the com-
pound was a potent inhibitor of [3H]glucose incorporation into the poly-
saccharide of seedling roots of Zea mays and of Setaria viridis but only relatively
weakly active on Glycine max and Ipomoea hederacea. Given that compound 1
appeared to be metabolised only slowly and that [3H]glucose incorporation
experiments were conducted within a total period of less than 90 min, it seems
probable that, as in the case of the acetyl CoA carboxylase-inhibitor gramini-
cides, the observed selectivity is determined by species-dependent di†erences at
the (in this case unknown) molecular target site for the herbicide. ( 1998 Society
of Chemical Industry
Pestic. Sci., 54, 368È376 (1998)
Key words: mode of action; inhibition of glucose incorporation; cellulose bio-
synthesis; herbicide; 5-tert-butyl-carbamoyloxy-3-(3-triÑuoromethyl)phenyl-4-
thiazolidinone
* To whom correspondence should be addressed.
E-mail: tim.hawkes=aguk.zeneca.com
¤ One of a collection of papers on various aspects of agrochemicals research contributed by sta† and collaborators of Zeneca
Agrochemicals UK and Zeneca Ag Products USA. The papers were collected and collated by Dr B. C. Baldwin and Dr D.
Tapolczay.
368
( 1998 Society of Chemical Industry. Pestic. Sci. 0031È613X/98/$17.50. Printed in Great Britain

A novel thiazolidinone herbicide
369
1
INTRODUCTION
(DMSO) were prepared and diluted into the test media
described in Sections 2.2È2.8 such that the Ðnal concen-
tration of DMSO did not exceed 2 ml litre~1. Stock
solutions of other chemicals that were used as standard
inhibitors in individual tests were prepared similarly.
Control experiments contained DMSO alone.
5-tert-Butyl-carbamoyloxy-3-(3-triÑuoromethyl)phenyl-4-
thiazolidinone (Fig. 1; 1) is a representative of a novel
class of N-phenyl-lactam-carbamate herbicides recently
reported by us.1 This compound is an experimental her-
bicide which shows potential for the selective pre-
emergence control of a range of weed species in
soybean. Susceptible weeds include grasses (such as
Digitaria spp., Setaria spp., Sorghum spp., Bachiaria spp.
and Echinochloa crus-galli Beauv.) and small-seeded
broad-leaved weeds (including Amaranthus spp. and
Chenopodium spp.), though large-seeded broad-leaved
weeds (such as Ipomoea spp. and Abutilon spp.) are not
controlled. The compound shows excellent crop selec-
tivity at the rates required for control of susceptible
weed species.
2.2 Tests for herbicidal activity
The pre-emergence herbicidal activity of 1 was deter-
mined under glasshouse conditions (24/19¡C day/night
with a 14-h photoperiod, supplementary lighting pro-
vided by metal halide lamps). The seeds were sown in a
loamy sand (1É6% organic matter content, pH 7) at a
depth of 1 cm for weeds and 2 cm for crop species.
Plant species included Setaria viridis Beauv., Echino-
chloa crus-gallii Beauv., Sorghum halepense Pers., Che-
nopodium album L., Amaranthus retroÑexus L., Ipomoea
hederacea Jacq., Abutilon theophrasti (L.) Medic.,
Glycine max Merr. and Zea mays L. The seeds were
watered two hours before spraying and lightly watered
afterwards. The compound was formulated in 4%
JF4400 (21É8 g litre~1 Span 80, 78É2 g litre~1 Tween 20
made up to 1 litre with methyl cyclohexanone) and the
plants were sprayed (8001E nozzle, 400 litre ha~1 at
2 psi) at a range of application rates. The plants were
In this paper, we describe the results of a study
designed to elucidate the mode of action of 1, and the
basis of its di†erential activity towards susceptible and
non-susceptible plant species.
2
MATERIALS & METHODS
2.1 Chemical synthesis and preparation of solutions for
testing
assessed 20 days after treatment and ED values were
50
estimated for each species.
Compound 1 was also tested for activity in Ðeld trials
at ZenecaÏs Ðeld stations at Rocky Mount, Leland and
Champaign in the USA, under well-irrigated conditions.
Test species included Digitaria sanguinalis Scop., E.
crus-gallii, Setaria faberi Herrm., S. halepense, Sorghum
vulgaris Pers., A. retroÑexus and I. hederacea.
The method used for the synthesis of 1 is outlined in
Fig. 1. Stage 1 involves the cyclocondensation of 3-
triÑuoromethylaniline, thioglycolic acid and formalde-
hyde to form the thiazolidinone ring. Stage 2 comprises
an initial chlorination at the thiazolidinone 4-position,
followed by hydrolysis to the corresponding alcohol.
Stage 3 involves reaction of this alcohol with tert-butyl
isocyanate to install the carbamate side chain. Full pre-
parative details of each stage of the process have been
reported previously.1
2.3 Assay for inhibition of growth of seedlings on agar
Test compounds at various concentrations (typically
0É03È30 lM) were incorporated into MurashigeÈSkoog
Concentrated stock solutions in dimethyl sulfoxide
Fig. 1. Synthesis of 5-tert-butyl-carbamoyloxy-3-(3-triÑuoromethyl)phenyl-4-thiazolidinone (compound 1).

370
Kate R. Sharples et al.
basal medium solidiÐed with agar (Phytagel). Sterilised
wild-type seeds of S. viridis, I. hederacea, G. max and Z.
mays were sown onto the agar surface and root and
shoot lengths were measured after six days.
Seeds of the isoxaben-resistant Arabidopsis thaliana
Heynh. line Ixrl-1 were obtained from the Arabidopsis
Biological Resource Centre, Ohio State University,
USA.2 This line and wild-type A. thaliana were treated
as above and assessed visually for inhibition of growth
after 15 days.
calf serum in microtitre wells and grown for four days
at 27¡C, either in the presence or absence of 70 lM of
test compound. Then 3-(4,5-dimethylthiazol-3-yl)-2,5-
diphenyltetrazolium bromide (MTT; 1 mg ml~1) was
added to each well and incubated for 4 h. After decant-
ing the supernatant, the residue of cell layer at the
bottom was then dispersed evenly in DMSO and the
colour read at 540 nm using a Dynatech plate reader.
2.6 [14C]Acetate incorporation into sterol and fatty
acid components of Zea mays cells
2.4 Growth of plant cells in liquid culture: inhibition
and reversal
Test compound was added to maize cells growing in
liquid culture as described in Section 2.4. After 30 min,
55É5 kBq (0É5 mM) of sodium [14C]acetate was added
and the mixture stood for 10 min, after which the cells
were rapidly chilled by transfer to ice, isolated from the
medium by vacuum Ðltration onto a Ðlter paper and
washed with 10 mM sodium acetate before disruption
and extraction into acidiÐed isopropanol ] chloroform
(1 ] 2 by volume). After low-speed centriÐgation, some
of the chloroform layer was removed and evaporated
under nitrogen. The lipid residue was re-dissolved in
chloroform (1 ml) and 50 ll of this solution was loaded
onto a 0É25-mm silica TLC plate, and chromatographed
using hexane ] diethyl ether ] acetic acid (80 ] 20 ] 1,
by volume) as eluent. After drying, the TLC plate was
scanned using a FujiBAS 1500 phosphor imaging
system to produce a chromatogram of the incorporated
radiolabel. The main peaks corresponded to phospho-
lipids (R c.0É04), glycolipids (R c.0É08), sterols (R c.0É13
Black Mexican Sweetcorn (Z. mays) cells were grown in
MurashigeÈSkoog medium containing 2,4-dichloro-
phenoxyacetic acid (2,4-D; 2 mg litre~1). Daucus carota
sativa Arang. line DC3 was maintained in the same
medium but containing 2,4-D and kinetin (0É1 and
0É2 mg litre~1, respectively). Stock lines were sub-
cultured every seven days. Growth experiments were
started from a 10% inoculum into medium (10 ml) kept
gently rotary-shaken in 25-ml conical Ñasks. Controls
(zero growth) contained a powerful standard inhibitor
(e.g., a sulfonyl urea herbicide). After Ðve days the cells
were freeze-dried and weighed. Attempts to reverse inhi-
bition of growth were made by including casein hydro-
lysate (2É5 mg ml~1), 2-ketoglutarate (0É6 mg ml~1) or
mixtures of adenosine, guanosine, cytidine and uridine
(0É3 mg ml~1 each) in the medium. In control experi-
ments these additions appropriately reversed the inhibi-
tory e†ects of sulfometuron-methyl (an inhibitor of
acetolactate synthase),3 acetylmethylphosphinate (an
inhibitor of pyruvate dehydrogenase)4 and hadacidin
(an inhibitor of adenylosuccinate synthase),5 respec-
tively.
f
f
f
and 0É16), fatty acids (R c.0É44) and sterol esters (R
c.0É83).
f
f
2.7 Assay for changes in the proÐle of plant sterols
Lipid was extracted from Z. mays cells treated for three
days with 0 (control), 10 or 50 lM test compound,
exactly as described in Section 2.4. Chloroform aliquots
(50 ll) were derivatised with bis(trimethylsilyl)triÑuoro-
acetamide and loaded onto a Hewlett Packard 5890
series gas chromatograph Ðtted with a Hewlett Packard
ultra 1 (12 m ] 0É2 lm) column. The resulting chro-
matogram was analysed for sterols using a Hewlett
Packard 5971 mass selective detector. Sterols were iden-
tiÐed by comparing the mass spectra with those of stan-
dard plant sterols.6
2.5 Growth of bacteria, fungi and insect cells
Escherichia coli Castell. & Chalm. strain JM101 was
seeded as a lawn onto a plate of agar made up in
minimal medium containing glucose as sole carbon
source, and 50 ll of a solution of test-compound was
added to small wells cut into the agar. Zones of bac-
terial growth-inhibition around these wells were assess-
ed following overnight growth at 37¡C. Agar plugs of
the fungi Aspergillus nidulans (Eidam) Winter and
Fusarium culmorum Sacc. were applied to plates of
LuriaÈBertani medium solidiÐed with agar made up
with and without test compound. The diameter of
mycelial growth was measured after 60 h and six days
at 20¡C.
2.8 Assay for inhibition of DNA-directed protein
synthesis in Escherichia coli extracts
The method used was that described by Zubay,7 slightly
modiÐed and used as described by Hawkes et al.8 for
detection of inhibitors. The DNA used was pGEM-bgal
Sf21 insect cells were grown in TC100 (GIBCO)
medium supplemented with 5% heat-inactivated foetal

A novel thiazolidinone herbicide
371
TABLE 1
(lac Z) from Promega. Test reactions contained 30 ll of
“S30Ï E. coli extract, 15 ll of concentrated reagent solu-
tion and 5 ll of test compound. Herbicidal inhibitors of
plastidic protein biosynthesis, such as esters of monic
acid, which inhibit isoleucyl tRNA synthetase9 were
readily detected as potent inhibitors at \5 lM.
Herbicidal Activity of Compound 1 in the Glasshouse
Species
Glycine max
ED (g ha~1)
50
c.20 000
770
54
Zea mays
Setaria viridis
Echinochloa crus-gallii
Sorghum halepense
Chenopodium album
Amaranthus retroÑexus
Ipomoea hederacea
Abutilon theophrasti
87
113
16
2.9 Assay for incorporation of [3H]glucose into
acid-insoluble polysaccharide
19
c.3500
830
Seeds were germinated for three to Ðve days between
layers of tissue paper and sand. Samples (20 mm long)
were cut from the root tips of Z. mays. I. hederacea and
G. max seedlings and bulked into 0É3-g lots which were
placed in 15-ml glass vials containing WhiteÏs root
culture medium (10 ml) at 25¡C. Whole seedlings of S.
viridis were used and bulked up into 0É4-g lots in vials
as above. Test compounds were added at a range of
concentrations (0É001È25 lM) and incubated with gentle
rotary shaking for 20 min, then [3H]glucose (74 kBq)
was added and each tube gently rotary-shaken for a
period of 1 h. Incorporation was stopped by raising the
temperature to 100¡C for 5 min. “ZeroÏ incorporation
controls were killed by pre-heating to 100¡C for 5 min.
Plant material was washed twice in situ with water and
then homogenised for 30 s in DMSO ] water (90 ] 10
by volume) using an Ultraturrax, and de-starched
through further incubation in DMSO for 2 h at 95¡C.
Samples were then collected by low-speed centrifu-
gation, washed in water, spun again, taken up in tri-
Ñuoroacetic acid (2 M; 5 ml) and incubated at 110¡C for
2 h. Samples were cooled, collected on paper Ðlter
papers, washed, dried, combusted (Packard Oximate 80)
suppression of I. hederacea was seen in some trials at
high application rates (500È1000 g ha~1), though com-
plete control was not achieved.
3.2 Clues to the mode of action of compound 1 from
biological symptoms
Compound 1 caused severe stunting of growth and
complete inhibition of seed germination when used at
high rates. Some of the more tolerant species, such as I.
hederacea, exhibited “sub-lethalÏ e†ects which included
swelling and splitting of stems, but the most profound
e†ects were observed on roots. In agar, the roots of
treated S. viridis developed as brown, stunted, swollen
and without lateral branches or root. Root growth was
completely inhibited at a concentration of c.2 lM in
agar. Visually, the symptoms induced in roots appeared
somewhat distinct from those induced by dinitroaniline
herbicides which act through binding to tubulin.10
Some further possible modes of action were imme-
diately recognised as unlikely on the basis of the bio-
logical symptoms observed. Compound 1 did not
bleach plants (and thus was unlikely to inhibit the bio-
synthesis of carotenoids), nor did it cause rapid scorch-
ing (and thus was unlikely to act in a similar way to
inhibitors of protoporphyrinogen oxidase or redox
mediators of photosystem I). The pre-emergence activity
and weed spectrum were reminiscent neither of inhibi-
tors of photosystem II, nor of hormone (2,4-D) type her-
bicides.
and counted using
a liquid scintillation counter
(Beckman LS 6000TA). To validate the test system it
was conÐrmed that incorporation was near linear with
time over at least 2 h.
3
RESULTS AND DISCUSSION
3.1 Herbicidal activity of compound 1
Representative ED values from the glasshouse tests
50
are shown in Table 1. These results clearly demonstrate
TABLE 2
that the grasses (S. viridis, E. crus-gallii, S. halepense)
and small-seeded broad-leaved species (C. album, A.
retroÑexus) are much more sensitive to 1 than are the
large-seeded broad-leaves species (I. hederacea, A.
theophrasti). The results also show that G. max shows a
signiÐcantly greater degree of tolerance towards 1 than
does Z. mays.
In the Ðeld, 1 gave complete control of several weed
species, as shown in Table 2. No damage to soybean
was observed at application rates of 1000 g ha~1. Some
Application Rates Required for Weed Control in the Field
Species
Application rate (g ha~1)
Echinochloa crus-gallii
Sorghum halepense
Sorghum vulgaris
Digitaria sanguinalis
Setaria faberi
250È500
250
250
125È250
125È250
63È125
Amaranthus retroÑexus

372
Kate R. Sharples et al.
3.3 E†ect of compound 1 on growth of bacteria, fungi,
insect and plant cells in liquid culture
oration of [14C]acetate incorporation into the lipid
fractions. Furthermore, consistent with its lack of cyto-
toxicity versus Sf21 cells, 1 did not uncouple oxidative
phosphorylation or inhibit respiration in intact mito-
chondria extracted from cockroach Ñight muscle. Thus,
1 did not inhibit fatty acid biosynthesis, sterol biosyn-
thesis or respiration. These conclusions were further
substantiated by data from in-vitro studies (Hawkes, T.
R., unpublished), indicating that 1 did not inhibit maize
acetyl CoA carboxylase or the incorporation of
[14C]malonyl CoA into palmitate by spinach stromal
fatty acid synthase.
At a nominal concentration of 0É8 mM, 1 did not detect-
ably inhibit the growth of E. coli. Similarly at 0É3 mM, it
only slightly ( \ 20%) reduced the growth rates of A.
nidulans and of F. culmorum mycelium, but had no e†ect
on either appearance or Ðnal biomass after six days. At
0É07 mM, 1 had no e†ect on either the growth or via-
bility (as judged by the reduction of MTT dye) of Sf21
insect cells. This test is a sensitive indicator for general
respiratory uncouplers of oxidative phosphorylation
and also for other metabolic toxic e†ects such as inhibi-
tion of respiration and inhibition of dihydrofolate
reductase.11 At 1, 10 and 100 lM, compound 1 inhibited
the growth of D. carota cells by 24, 46 and 52% relative
to controls; Z. mays cells were similarly inhibited by
52% at 100 lM.
Herbicidal molecules which act by inhibiting the
major pathways of, for example, the biosynthesis of
amino acids or of purines can often be recognised on
the basis of the fact that the inhibitory e†ect which they
exert on the growth of plant cells can be speciÐcally
reversed by addition of the end-product of the inhibited
pathway. However, over a range of treatments, the
inhibitory e†ect of 1 on growth was undiminished in the
presence of casein hydrolysate, 2-ketoglutarate or
various mixtures of nucleosides. Whilst falling short of
absolute proof, these experiments strongly suggest that
it is unlikely that the herbicidal activity of 1 is due to
inhibition of the biosynthesis of amino acids, purines or
pyrimidines.
3.5 Sterol proÐle of Zea mays cells
The relative percentages of campesterol, stigmasterol
and sitosterol in control maize cells were 13, 33 and 54,
respectively. Treatment with 10 lM of a c-keto triazole
herbicide known to inhibit obtusifoliol 14C-
demethylase15 resulted in massive changes, with the
stigmasterol and sitosterol peaks being reduced to c.5%
and 22%, respectively, and the campesterol peak being
more or less abolished and replaced by a new peak,
identiÐed as 14-a-methyl-24-dihydrofecosterol (with
characteristic mass peaks at m/e 486, 471, 381, 227, 213
and 196) and accounting for c.70% of the total sterol.
Treatment for three days with 10 lM 1 resulted in no
discernible e†ect on the sterol proÐle of either maize or
carrot cells. Thus 1 did not, apparently, exert its herbi-
cidal e†ect through interference with the later steps of
plant sterol biosynthesis.
3.6 DNA-directed protein synthesis in extracts of
Escherichia coli
3.4 [14C]Acetate incorporation into fatty acids and
sterols
The machinery of protein biosynthesis within the
plastid is broadly similar to that within eubacteria.
DNA-directed protein synthesis in E. coli extracts can
serve as a proxy for protein and RNA biosynthesis
within the plastid,16 and thus, for example, inhibitors of
isoleucyl tRNA synthetase which are both antibiotics
and herbicides can be readily detected through inhibi-
tion of this process.9 However, in this case, even at
20 lM; 1 had no detectable e†ect on the rate of in-vitro
synthesis of b-galactosidase directed by plasmid DNA.
In untreated Z. mays cells the relative percentages of
incorporation of [14C]acetate were as follows: phos-
pholipids (69); glycolipids (5); sterols (8É5); fatty acids
(12) and sterol esters (6). Treatment of cells for 30 min
with either 50 lM Ñuazifop acid (a standard herbicide
which inhibits acetyl CoA carboxylase)12 or methyl
(E,E)-a-(methoxymethylene)-2-(2-phenylethenyl)phenyl-
acetate (a strobilurin-derived inhibitor of respiration)13
resulted, in both cases, in a c.80% decrease in total
incorporation and a shift to a disposition of percentage
incorporation to: phospholipids (36); glycolipids (4);
sterols (40); fatty acids (0É2); sterol esters (20). Under
similar conditions, mevinolin (a potent inhibitor of
HMGCoA reductase)14 resulted in a 25% decrease in
total incorporation and a shift to a percentage dis-
position of phospholipids (83); glycolipids (5); sterols
(0); fatty acids (5) and sterol esters (0). By contrast,
treatment with 1 at a nominal concentration of 200 lM
for more than 2 h had no discernible e†ect on either the
total level of incorporation or the disposition of incorp-
3.7 [3H]glucose incorporation into acid-insoluble
polysaccharide
A series of experiments was conducted to examine the
e†ect of 1 on the incorporation of [3H]glucose into the
acid-insoluble polysaccharide cell wall fraction of Z.
mays roots, usually assumed to be cellulose.17 The
results are summarised in Fig. 2. Compound 1 was a
potent inhibitor of glucose incorporation, similar to iso-
xaben (N-(3-[1-ethyl-1-methylpropyl]-5-isoxazoloyl)-2,

A novel thiazolidinone herbicide
373
analogues of 1 yielded much lower values (Wang, Y.,
unpublished), indicating that the value is useful and not
merely an arbitrary lower limit set by the concentration
of target site within a particular root preparation.
Further biological evidence for 1 having an action on
cell wall biosynthesis came from a closer examination of
the inhibition of growth of plant cell culture under the
light microscope. Sub-lethal treatments of 1, isoxaben
and of dichlobenil all caused a characteristic swelling of
tobacco BY2 cells (up to three-fold after Ðve days)
which was not apparent with herbicides such as tri-
Ñuralin and acetochlor which have di†erent modes of
action. Inhibition thus appeared much less when assess-
ed on the basis of the settled volume of wet cells than
on the basis of dry weight.
Fig. 2. Inhibition of cellulose biosynthesis by test compounds
(10 lM) in seedling roots of Zea mays.
3.8 An isoxaben-resistant Arabidopsis mutant is
cross-resistant of compound 1
6-dimethoxybenzamide)18 and dichlobenil (2,6-dichloro-
benzonitrile).19 Experiments using a range of standard
metabolic inhibitors and herbicides conÐrmed that inhi-
bition was not a general phenomenon but was peculiar
to 1, isoxaben and dichlobenil (Wang Y., unpublished;
see also Reference 18). The exception was the respir-
ation inhibitor, methyl (E,E)-a-(methoxymethylene)-2-
(2-phenylethenyl)phenylacetate, which also exerted a
signiÐcantÈand presumably less directÈinhibitory
e†ect. As discussed above, 1 is not an uncoupler of oxi-
dative phosphorylation, and does not inhibit respir-
ation. Thus, like isoxaben and dichlobenil, 1 acts
directly or indirectly to inhibit cellulose biosynthesis.
The results from this study into the mode of action of 1
were very similar to those reported by Heim et al.18 for
the herbicide isoxaben. Using A. thaliana plants, these
workers observed that isoxaben had no e†ect on the
incorporation of radio-labelled precursors into protein,
nucleic acids or fatty acids, but potently and speciÐcally
(and unlike a range of control metabolic inhibitors)
inhibited the incorporation of radio-labelled glucose
into the acid-insoluble fraction of the cell wall. As was
found with 1, isoxaben was more active against root
tissue than against cells in culture and, again, activity
was in the nM range with a maximum degree of inhibi-
tion that was less than complete (close to 80%).
The IC
value for inhibition was c.50 nM, with a
The mutant Ixr1-1, which is resistant to isoxaben,2 is
cross-resistant to 1 (Fig. 3). This provides some direct
evidence for 1 and isoxaben sharing the same molecular
target (which is not shared with dichlobenil).20 Isoxaben
at 0É3 lM killed wild-type Arabidopsis whilst only
reducing the growth of the resistant biotype by 20%
(Fig. 3). Thus the resistant biotype had at least a Ðve-
fold tolerance to isoxaben. At 12 lM, compound 1 killed
the wild-type Arabidopsis but reduced the growth of the
resistant biotype by only 50%. Thus the resistant
50
maximum level tending to approximately 90% at high
treatments. It is, of course, difficult to judge how such
apparent IC values might translate to real binding
constants at the (unknown) molecular target because of
the confounding e†ects of uptake and partition.
However the value clearly reports on a system which
has come to equilibrium since (i) it was not highly
dependent on time, as similar values were obtained
from experiments conducted over 1 or 2 h, and (ii) some
50
Fig. 3. The e†ect of isoxaben and compound 1 on the growth of (L) wild-type and (…) resistant Arabidopsis thaliana.

374
Kate R. Sharples et al.
CONCLUSIONS
4
Taken together, the experiments described above and
the symptoms produced by 1 on plants suggest very
strongly that its herbicidal action is due solely to inhibi-
tion of cellulose biosynthesis. However the molecular
target itself remains to be deÐned. Little is known of the
cellulose-biosynthesising machinery in higher plants.
Electron microscope evidence suggests that synthesis
occurs within an enzyme complex associated with the
plasma membrane, but the nature of this remains
obscure, since, thus far, it has proved impossible to
isolate without loss of activity.21 Recent progress in
plant genomics will perhaps expedite the identiÐcation
of its various components.22 UDP-Glucose has been
implicated as the activated form of glucose which acts
directly as substrate for catalytic polymerisation and a
membrane-associated isoform of sucrose synthase
(rather than UDP glucose pyrophosphorylase) seems
the most likely candidate to provide this.23 In principle,
1, isoxaben and dichlobenil could bind directly to any
essential part of the cellulose synthase complex or, alter-
natively, could act indirectly through interference with
regulation or some essential post-translational pro-
cessing. Whatever the mechanism, it must take e†ect
rapidly since inhibition is expressed within less than an
hour of contact with the herbicide (and probably faster,
this estimate being limited by our inability to make the
measurement more quickly). Certainly none of the her-
bicides bears any obvious structural resemblance to
UDP-Glucose and neither isoxaben nor dichlobenil
inhibits the relatively well-characterised cellulose syn-
thase from Acetobacter xylinum.24 It seems most prob-
able that all of the herbicides bind at a site other than
the catalytic site and which is in some way unique to
the plant system.
There is no evidence that the herbicides need bind at
the same site, and indeed there is evidence to the con-
trary. There are three known isoxaben-resistant
mutants of A. thaliana. Two confer a high level of resist-
ance and represent di†erent alleles at the same locus
(Ixr A)2 whilst the third confers only a low level of
resistance and maps to a locus genetically unlinked to
the other two (Ixr B).25 In all of these mutants cellulose
biosynthesis was identiÐed as the primary site of resist-
ance and resistance could not be attributed to di†er-
ences in uptake or metabolism. The mutants mapped to
two independent genes, so it is assumed that isoxaben
either acts directly on the putative cellulose synthase
complex, which consists of at least two di†erent sub-
units, or that the genes may encode proteins involved in
regulation or post-translational modiÐcations of active
site.26 These mutants were not cross-resistant to dichlo-
benil and neither were dichlobenil mutants cross-
resistant to isoxaben.20 A dichlobenil photo-affinity
analogue was demonstrated to bind to an 18 kD
protein not associated with the plasma membrane.27
Fig. 4. The inhibition of cellulose biosynthesis by compound
1 in roots of ()) Glycine max, (>) Zea mays and (…) Ipomoea
hederacea, and (]) in whole seedlings of Setaria viridis.
biotype was also signiÐcantly (at least two-fold) resist-
ant to 1.
3.9 Species-dependent di†erences in inhibition of
[3H]glucose incorporation reÑect di†erences in the
herbicidal activity of compound 1
The inhibition of glucose incorporation into the acid-
insoluble cell-wall material is clearly species-dependent,
as demonstrated in Fig. 4. In S. viridis and Z. mays,
glucose incorporation was inhibited by 80% at concen-
trations [0É1 and [1 lM, respectively, whereas sig-
niÐcantly higher treatments were required to produce
the same level of inhibition in I. hederacea and G. max.
These results closely parallel the observed herbicidal
activity of 1 against three species (see Table 1 for ED
values).
50
The rate of metabolism of 1 in plants was readily
measured as loss of parent compound after addition to
G. max cell culture. The compound was identiÐed by
HPLC using a Hi-chrom, C18, ODS2 reverse-phase
column.
A
mobile phase of acetonitrile ] water
(65 ] 35 by volume) was used and detection was by UV
at a wavelength of 210 nm. There was no signiÐcant
loss of parent compound after 24 h. Further tests using
radiolabelled close analogues of 1 demonstrated that, in
general, this class of herbicide is not metabolised in G.
max for at least three days after application (Hayhurst
I., unpublished). This slow metabolism, combined with
the relatively short period over which incorporation
assays were conducted, suggests that it is most likely
that the di†erential activity of 1 is directly determined
by species-dependent variability at an (unknown)
molecular target. Also, as (embryo-derived) maize cell
culture seemed markedly less sensitive to inhibition
than did maize roots, it seems likely that variability at
the molecular target site might also be tissue- as well as
species-dependent, and, perhaps, account for the more
severe e†ects of 1 on roots than on shoots and leaves.

A novel thiazolidinone herbicide
375
10. Appleby, A. P. & Valverde, B. E., Behavior of dinitro-
aniline herbicides in plants. W eed T echnol., 3 (1989) 198È
206.
11. Vistica, D. T., Skehan, P., Scudiero, D., Monks, A.,
Pittman, A. & Boyd, M., Tetrazolium-based assays for cel-
lular viability: a critical examination of selected param-
eters a†ecting formazan production, Cancer Res., 51
(1991) 2515È20.
12. Harwood, J. L. In Herbicides, ed. N. R. Baker and M. P.
Percival. Elsevier Science Publishers BV, 1991, pp. 209È
46.
13. Beautement, K., Clough, J. M., de Fraine, P. J. & Godfrey,
C. R. A., Fungicidal b-methoxyacrylates: from natural
products to novel synthetic agricultural fungicides. Pestic.
Sci., 31 (1991) 499È519.
14. Alberts, A. W., Chen, J., Kuron, G., Hunt, V., Hu†, J.,
Ho†man, C., Rothrock, J., Lopez, M., Joshua, H., Harris,
E., Patchett, A., Monaghan, R., Currie, S., Stapley, E.,
Albers-Schonberg, G., Hensens, O., HirshÐeld, J., Hoog-
steen, K., Leisch, J. & Springer, J., Mevinolin: A highly
potent competitive inhibitor of HMG-CoA reductase and
a cholesterol-lowering agent. Proc. Natl Acad. Sci. (USA),
77 (1980) 3957È61.
15. Schaller, H., Maillot-Vernier, P., Belliard, G. & Benve-
niste, P., Increased sterol biosynthesis in tobacco calli
resistant to a triazole herbicide which inhibits demethyla-
tion of 14a-methyl sterols. Planta, 187 (1992) 315È21.
16. Hughes, J. & Mellows, G., Interaction of pseudomonic
acid A with Escherichia coli B isoleucyl-tRNA synthetase.
Biochem. J., 191 (1980) 209È19.
The isoxaben-resistant Arabidopsis used in this study
had the IxrA mutation and was cross-tolerant to 1.
Thus, isoxaben and 1 probably share at least part of a
binding site. It is interesting to note the parallels with
quinclorac, which also appears to inhibit the biosyn-
thesis of plant cell walls and where, again, selectivity
between species is inferred to be due to di†erences at an
unknown molecular site of action. Quinclorac inhibits
the biosynthesis of both celluloses and hemicelluloses
and apparently acts at a site distinct from isoxaben and
1.28,29
Consistent with the observation that both isoxaben
and 1 never inhibited cellulose biosynthesis by more
than 80È90%, and with the observed tissue- and
species-selective e†ects, there is also molecular evidence
for there being more than one cellulose-synthesising
machinery in plants. There are Arabidopsis mutants
which are unable to synthesis cellulose only in speciÐc
tissues, indicating that there may be di†erent forms of
the enzyme in di†erent parts of the plant and that the
cellulose required for primary and secondary cell walls
may also be di†erentially regulated.30h32 Progress in
the Ðnal identiÐcation of the binding site for 1 will most
probably require
a
combination of molecular
approaches (e.g. map-based cloning of the resistant
mutants) and traditional biochemistry (e.g. identiÐca-
tion of the binding site through synthesis of a suitable
photo-affinity label).
17. Updegra†, D. M., Semimicro determination of cellulose in
biological materials. Analyt. Biochem., 32 (1969) 420È4.
18. Heim, D. R., Skomp, J. R., Tschabold, E. E. & Larrinua,
I. M., Isoxaben inhibits the synthesis of acid-insoluble cell
wall materials in Arabidopsis thaliana. Plant Physiol., 93
(1990) 695È700.
REFERENCES
19. Hogetsu, T., Shibaoka, H.
& Shimokoriyama, M.,
Involvement of cellulose synthesis in actions of gibberellin
and kinetin on cell expression. 2,6-Dichlorobenzonitrile as
a new cellulose-synthesis inhibitor. Plant Cell Physiol., 15
(1974) 389È93.
1. Zeneca patent, Preparation of [(phenyl)heterocyclyl] car-
bamates or thiocarbamates as herbicides, WO 94/13652,
1994.
2. Heim, D. R., Roberts, J. L., Pike, P. D. & Larrinua, I. M.,
Mutation of a locus of Arabidopsis thaliana confers resist-
ance to the herbicide isoxaben. Plant Physiol., 90 (1989)
146È50.
3. LaRossa, R. A. & Schloss, J. V., The sulfonylurea herbi-
cide sulfometuron-methyl is an extremely potent and
selective inhibitor of acetolactate synthase in Salmonella
typhimurium. J. Biol. Chem., 259 (1984) 8753È7.
4. Baillie, A. C., Wright, K., Wright, B. J. & Earnshaw, C. G.,
Inhibitors of pyruvate dehydrogenase as herbicides.
Pestic. Biochem. Physiol., 30 (1988) 103È12.
5. Rossomando, E. F., Maldonado, D. & Crean, E. V., E†ect
of hadacidin on growth and adenylosuccinate synthetase
activity of Dictyostelium discoideum. Antimicrob. Agents
Chemother., 14 (1978) 476È82.
6. Rahier, A. & Benveniste, P., Mass spectral identiÐcation
of phytosterols. In Analysis of Sterols and Other Bio-
logically SigniÐcant Steroids, ed. W. D. Nes and E. J.
Parish. Academic Press, San Diego, 1989, pp. 223È50.
7. Zubay, G., In-vitro synthesis of protein in microbial
systems. Ann. Rev. Genet., 7 (1973) 267È87.
8. Hawkes, T. R., Merrick, M. & Dixon, R., Interaction of
puriÐed NtrC protein with nitrogen regulated promoters
from Klebsiella pneumoniae. Mol. Gen. Genet., 201 (1985)
492È8.
20. Heim, D. R., Bjelk, L. A. & Larrinua, I. M., Isolation and
characterisation of a dichlobenil-resistant mutant of Ara-
bidopsis thaliana. W eed. Sci. Soc. Am., 31 (1991) 76.
21. Delmer, D. P., Cellulose biosynthesis. Annu. Rev. Plant
Physiol., 38 (1987) 259È90.
22. Cutler, S. & Somerville, C., Cellulose synthesis: Cloning in
silico. Curr. Biol., 7 (1997) R108ÈR111.
23. Delmer, D. P. & Amor, Y., Cellulose biosynthesis. Plant
Cell, 7 (1995) 987È1000.
24. Tanaka, Y., Sugoh, M., Ji, W., Iwabuchi, J., Yoshida, H. &
Omura, S., Screening method for cellulose biosynthesis
inhibitors with herbicidal activity. J. Antibiot., 48 (1995)
720È5.
25. Heim, D. R., Roberts, J. L., Pike, P. D. & Larrinua, I. M.,
A second locus, Ixr Arabidopsis thaliana, that confers
resistance to the herbicide isoxaben. Plant Physiol., 92
(1990) 858È61.
26. Heim, D. R., Skomp, J. R., Waldron, C. & Larrinua, I. M.,
Di†erential response to isoxaben of cellulose biosynthesis
by wild-type and resistant strains of Arabidopsis thaliana.
Pestic. Biochem. Physiol., 39 (1991) 93È9.
27. Delmer, D. P., Read, S. M. & Cooper, G., IdentiÐcation of
a receptor protein in cotton Ðbres for the herbicide 2,6-
dichlorobenzonitrile. Plant Physiol., 84 (1987) 415È20.
28. Koo, S. J., Neal, J. C. & DiTomaso, J. M., 3,7-Dichloro-
quinolinecarboxylic acid inhibits cell-wall biosynthesis in
maize roots. Plant Physiol., 112 (1996) 1383È9.
9. Zeneca patent, Assays for aminoacyl tRNA synthetases
and their inhibitors and their use in the identiÐcation of
herbicides, WO 95/09927, 1995.

376
Kate R. Sharples et al.
29. Koo, S. J., Neal, J. C. & DiTomaso, J. M., Mechanism of
action and selectivity of quinclorac in grass roots. Pestic.
Biochem. Physiol., 57 (1997) 44È53.
30. Baskin, T. I., Betzner, A. S., Hoggart, R., Cork, A. & Wil-
liamson, R. E., Root morphology mutants in Arabidopsis
thaliana. Australian J. Plant Physiol., 19 (1992) 427È37.
31. Arioli, T., Betzener, A., Peng, L., Baskin, T., Herth, W.,
Cork, A., Birch, R., Rolfe, B., Redmond, J. & Williamson,
R., Radial swelling mutants deÐcient in cellulose biosyn-
thesis. J. Cell Biochem., 21 (1995) 440.
32. Potikha, T. & Delmer, D. P., A mutant of Arabidopsis
thaliana displaying altered patterns of cellulose deposition.
Plant J., 7 (1995) 453È60.