Comprehensive understanding of acetohydroxyacid synthase inhibition by different herbicide families

Significance

Acetohydroxyacid synthase (AHAS), also known as acetolactate synthase, is the target for more than 50 commercial herbicides that are used globally to protect essential rice, corn, wheat, and cotton crops. Two newly developed chemical classes of AHAS inhibitors are the pyrimidinyl-benzoates and sulfonylamino-cabonyl-triazolinones. These are the active components of more than 12 successfully marketed herbicide products. Here we have determined the crystal structures of two members of each of these families in complex with plant AHAS. In addition, we have established a precise explanation of the inhibition kinetics for all of the AHAS-inhibiting herbicide families. These data will be an important resource for the design of herbicides with a reduced propensity for developing weed resistance.

Abstract

Five commercial herbicide families inhibit acetohydroxyacid synthase (AHAS, E.C. 2.2.1.6), which is the first enzyme in the branched-chain amino acid biosynthesis pathway. The popularity of these herbicides is due to their low application rates, high crop vs. weed selectivity, and low toxicity in animals. Here, we have determined the crystal structures of Arabidopsis thaliana AHAS in complex with two members of the pyrimidinyl-benzoate (PYB) and two members of the sulfonylamino-carbonyl-triazolinone (SCT) herbicide families, revealing the structural basis for their inhibitory activity. Bispyribac, a member of the PYBs, possesses three aromatic rings and these adopt a twisted “S”-shaped conformation when bound to A. thaliana AHAS (AtAHAS) with the pyrimidinyl group inserted deepest into the herbicide binding site. The SCTs bind such that the triazolinone ring is inserted deepest into the herbicide binding site. Both compound classes fill the channel that leads to the active site, thus preventing substrate binding. The crystal structures and mass spectrometry also show that when these herbicides bind, thiamine diphosphate (ThDP) is modified. When the PYBs bind, the thiazolium ring is cleaved, but when the SCTs bind, ThDP is modified to thiamine 2-thiazolone diphosphate. Kinetic studies show that these compounds not only trigger reversible accumulative inhibition of AHAS, but also can induce inhibition linked with ThDP degradation. Here, we describe the features that contribute to the extraordinarily powerful herbicidal activity exhibited by four classes of AHAS inhibitors.

The sulfonylurea (SU), imidazolinone (IMI), triazolopyrimidine (TP), pyrimidinyl-benzoate (PYB), and sulfonylamino-cabonyl-triazolinone (SCT) (Fig. 1) herbicides constitute one of the essential pillars in modern agricultural practices. These herbicides capture a major share of the crop-protection market, which is valued at ∼$US30 billion per annum. The outstanding success of these herbicides is due to a number of factors, including their ability to selectively kill weeds in preference to crops, their low toxicity to animals, and extremely potent herbicidal activity, allowing very low application rates (1).

Fig. 1.

Examples of the chemical structures of the five different classes of AHAS herbicides. PYBs (BS and PB), SCTs (PC and TCM), SUs (CE), TPs (penoxsulam), and IMIs (IQ).

These five families of herbicides exert their activity by inhibiting acetohydroxyacid synthase (AHAS, E.C. 2.2.1.6), also known as acetolactate synthase, which is the first enzyme in the branched-chain amino acid (i.e., valine, leucine, and isoleucine) biosynthesis pathway. The reaction performed by AHAS involves the decarboxylation of a molecule of pyruvate, yielding an enzyme-bound hydroxyethyl (HE) intermediate that then reacts with either a second molecule of pyruvate or 2-ketobutyrate to produce (S)-2-acetolactate or (S)-2-aceto-2-hydroxybutyrate, respectively (2). This activity is dependent on the presence of three cofactors, thiamine diphosphate (ThDP), magnesium ion (Mg²⁺), and flavin adenine dinucleotide (FAD). Of the more than 50 AHAS inhibitors approved for commercial use as herbicides none of these resemble the substrates or products of the reaction.

The inhibition constants of the commercially available SU and IMI herbicides for AHASs from several different sources have been determined previously, with K_i values in the low nanomolar and micromolar ranges, respectively (3–5). For Arabidopsis thaliana AHAS (AtAHAS), chlorimuron ethyl (CE) (Fig. 1) is the most potent SU with a K_iapp of 10.8 nM, whereas the IMI imazaquin (IQ) (Fig. 1) has a K_iapp of only 3.0 μM, some 300-fold higher (4). IC₅₀ values for the other two families of AHAS herbicides have also been reported. For the PYBs, values range from 0.7 nM to 11.0 nM for rice and barley AHAS (6). For the SCTs values range from 9.3 nM to 24.8 nM for barley AHAS (7). One of the difficulties in determining K_i values has been due to the fact that inhibition of AHAS by these commercial herbicides has a complex mechanism that does not follow the classical behavior for mixed, uncompetitive, or competitive inhibition. Indeed, until recently, it has been described as a slow tight-binding mechanism (8). A point of intrigue that concerns the IMIs is that they are only modest inhibitors of AHAS but are equally effective herbicidal compounds compared with the other four families of AHAS inhibitors [e.g., application rates are 10–100 g/ha (1)]. This suggests that factors other than affinity of binding must contribute to the biological activity of these compounds.

It has been ascertained that inhibition of fungal AHAS by these herbicides fits a model of reversible accumulative inhibition (9). Therefore, two general aspects need to be considered when studying inhibition of AHAS: (i) the affinity of the inhibitor for the enzyme (K_i) and (ii) the effect of accumulative inhibition.

The crystal structures of the catalytic subunit of AtAHAS in complex with the SUs and IMIs showed that the positions occupied by these two families of inhibitors partially overlap with each other and that they bind using significantly different mechanisms of induced fit (10). In addition, the structural data when the SUs bind revealed degradation of ThDP. Whether this degradation is due to radiation damage or to inhibitor–enzyme interactions remained unresolved. Curiously, the ThDP remains completely intact in the crystal structure of AtAHAS in complex with the IMI, IQ (10). An intact ThDP is also observed in the uninhibited structure of the catalytic subunit of Saccharomyces cerevisiae AHAS (ScAHAS) (11). These data suggested that the modification of ThDP is herbicide induced but the type of modification depends on the chemistry of the different families of AHAS inhibitors.

AHAS inhibitors have been used for crop protection since the late 1980s. The extensive use of these herbicides has resulted in the emergence of 54 weed-resistant species that show at least one single-point mutation occurring to the target enzyme (12), with many of these mutations conferring cross-resistance to several families of AHAS-inhibiting herbicides (3, 4, 13). Thus, a better understanding of the interactions between these inhibitors and their target enzyme will help to advance our armory to control resistance.

In the present study we have determined the crystal structures of the catalytic subunit of AtAHAS in complex with two PYBs, pyrithiobac (PB) and bispyribac (BS), and two SCTs, propoxycarbazone (PC) and thiencarbazone methyl (TCM) (Fig. 1). The chemical structures of these herbicides differ significantly from the SUs and IMIs, and thus these crystallographic data represent a significant advance in our understanding of how all of the families of AHAS-inhibiting herbicides function. These structures further show that the PYBs and SCTs modify the thiazole ring of ThDP in different ways. This result was unexpected and appears to be one of the factors that explains the remarkable efficacy of these herbicides when in action. In addition, we have applied an assay method to determine the K_i values of these herbicides in the absence of accumulative inhibition, a value that more precisely reflects the affinity of herbicide binding.

Results and Discussion

Crystal Structures of AtAHAS in Complex with Four Commercial Herbicides.

The crystal structures of AtAHAS in complex with PB, BS, PC, and TCM have been determined at resolution values in the range of 2.73–2.87 Å (Table 1). The former two compounds are members of the PYB family and the latter two are members of the SCT family. In all of the structures there is a single monomer in the asymmetric unit (Fig. 2A) with the AtAHAS tetramer generated by crystallographic symmetry. The structures have an overall fold similar to that reported for uninhibited AtAHAS (PDB code 5K6Q) and the complexes with the SUs (e.g., CE; PDB code 1YBH) or the IMIs (e.g., IQ: PDB code 1Z8N) (10), with rmsd values in the range of 0.16–0.27 Å after superimposition of all Cα atoms. In all four complexes the difference electron density for the inhibitor unequivocally allowed the positioning of all of the atoms into the final models (Fig. 2B). Likewise, the complete FAD, including the flavin ring, could be fitted unambiguously to difference electron density maps and in all cases the flavin ring is in a bent conformation, suggesting it is in a reduced state.

Table 1.

Data collection and refinement statistics for AtAHAS in complex with commercial herbicides

Data and statistics	PB	BS	PC	TCM
Crystal parameters
Unit cell, Å, a = b, c	179.09, 184.97	179.72, 184.78	180.02, 186.16	180.36, 186.47
Space group	P6422	P6422	P6422	P6422
Diffraction data*
Temperature, K	100	100	100	100
Resolution range, Å	40.30–2.87	42.04–2.87	48.55–2.76	48.63–2.73
Observations [I > 0σ(I)]	865,462 (42,989)	788,399 (37,243)	986,206 (86,605)	1,009,017 (87,188)
Unique reflections [I > 0σ(I)]	40,366 (1,972)	40,616 (1,981)	45,984 (4,308)	47,622 (4,415)
Completeness, %	99.9 (100.0)	99.9 (100.0)	99.7 (96.9)	99.6 (96.1)
Rmerge†	0.087 (0.769)	0.183 (0.897)	0.131 (0.741)	0.195 (0.790)
Rpim	0.027 (0.247)	0.052 (0.371)	0.136 (0.169)	0.061 (0.250)
〈I〉/〈σ(I)〉	41.0 (2.7)	37.5 (3.9)	16.2 (2.7)	10.2 (2.1)
Refinement statistics
Resolution limits, Å	40.30–2.87	42.04–2.87	45.01–2.76	44.67–2.73
Rfactor	0.1465	0.1731	0.1438	0.1446
Rfree	0.1786	0.2112	0.1790	0.1781
rmsd‡ bond lengths, Å	0.002	0.002	0.002	0.002
rmsd bond angles, °	0.800	0.712	0.647	0.683
Ramachandran plot, %
Favored	98.6	98.1	98.1	97.9
Outliers	0	0	0	0

^*Values in parentheses are for the outer-resolution shells: 2.92–2.87 Å for PB, 2.92–2.87 for BS, 2.86–2.76 for PC, and 2.83–2.73 for TCM.

^†R_merge = Σ∣I –〈I〉|/Σ〈I〉, where I is the intensity of an individual measurement of each reflection, and 〈I〉 is the mean intensity of that reflection.

^‡rmsd, root-mean-square deviation.

Fig. 2.

The AtAHAS–herbicide complex. (A) The overall fold of a single subunit of AtAHAS in complex with BS. The α (86–280), β (281–451) and γ (463–639) domains are colored light blue, purple, and blue, respectively. The C-terminal tail (640–668) is depicted in red. FAD (yellow), ThDP (orange), and BS (light green) are shown as stick models, whereas the Mg²⁺ ion is represented as a pink sphere. (B) The fold of the inhibitors with the difference electron density superimposed. Carbon is green; nitrogen, blue; oxygen, red; and sulfur, yellow.

In all four structures of the enzyme–inhibitor complexes the electron density for the pyrimidine ring and the diphosphate moiety of ThDP and the magnesium ion that holds ThDP to the enzyme are well resolved. However, the density for the thiazolium ring differs when the structures for the two separate families are compared. For PB and BS there is no density at the C2 carbon, indicating that the thiazolium ring has been cleaved, forming thiamine aminoethenethiol diphosphate (ThAthDP) (Fig. S1A). This was subsequently confirmed by mass spectrometry analysis of a sample obtained from a solution of enzyme incubated with PB in buffer and from crystals of the AtAHAS–PB complex (Fig. S2A). The mass spectrometry data showed a peak that corresponds to a fragment that would have resulted from cleavage of ThAthDP. Thus, it appears that when the complete ThAthDP is released from the enzyme and the sample is subjected to the conditions of the mass spectrometry experiment, the amide bond is broken (scheme in Fig. S2A). By comparison a control sample of the uninhibited enzyme did not show the presence of this fragment. A similar modification to ThDP has been previously reported in the crystal structures of Zymomonas mobilis pyruvate decarboxylase (14) (PDB code 1ZPD) and Streptomyces clavuligerus carboxethylarginine synthase (15) (PDB code 2IHU). In the structures where the SCTs are bound, the electron density for the thiazolium ring is intact and furthermore has density extending beyond the C2 atom, which we assigned to be an oxygen atom. Thus, we modeled the structure as thiamine thiazolone diphosphate (ThThDP) (Fig. S1B). Mass spectrometry analysis of a solution of enzyme incubated with PC in buffer also showed a mass for the ThThDP derivative and, as expected from the electron density maps, no peak could be detected for the ThAthDP. The same set of peaks was also obtained from dissolved crystals of the AtAHAS–PC complex (Fig. S2B). For an additional control experiment we also performed mass spectrometry analysis of the cofactor isolated from uninhibited AtAHAS and this showed only the presence of intact ThDP (Fig. S2C) and neither ThAthDP nor ThThDP was present, confirming that the herbicides alter the structure of this cofactor. ThThDP has been observed in several other ThDP-dependent enzymes, including those for Pseudomonas putida benzoylformate decarboxylase (PDB code 1YNO), human branched-chain alpha-ketoacid dehydrogenase (16) (PDB code 2BFC), Oxalobacter formigenes oxalyl CoA decarboxylase (17) (PDB code 2C31), and Escherichia coli pyruvate dehydrogenase complex E1 component (18) (PDB code 1RP7).

Fig. S1.

ThDP and Mg²⁺ in AtAHAS in complex with herbicides of the PB and SCT families. (A and B) 2F_o – F_c (blue) and F_o – F_c (green or red) electron density maps contoured at 2.0 σ and 3.5 σ, respectively, for ThDP in (A) PYB and (B) PC complexes. Similar electron density maps were observed in the BS and TCM complexes.

Fig. S2.

High-resolution mass spectrometry analysis of ThDP. (A) A fragment of ThAthDP isolated from a solution of AtAHAS incubated with PB. Molecular formula: C₆H₁₁N₄ [M + H]⁺ calculated 139.0978 m/z, found 139.0985 m/z (4.5 ppm error). (B) ThThDP isolated from a solution of AtAHAS incubated with PC. Molecular formula: C₁₂H₁₉N₄O₈P₂S [M + H]⁺ calculated 441.0393 m/z, found 441.0378 m/z (2.6 ppm error). m/z peaks were identical when crystals of AtAHAS in complex with PB or PC were dissolved in water and subjected to LC-MS. (C) Control sample, ThDP isolated from AtAHAS incubated in buffer without herbicide. Molecular formula: C₁₂H₁₉N₄O₇P₂S [M]⁺ calculated 425.0444 m/z, found 425.0461 m/z (4.0 ppm error).

Modification of ThDP has also previously been reported in AHAS structures in complex with the disubstituted SUs (10, 19, 20), but these modifications differ from those observed in these four structures. In some of the complexes the degradation of ThDP results in the loss of several atoms from the thiazolium and/or the pyrimidine rings. On the other hand, the structures of AtAHAS in complex with the monosubstituted SUs, monosulfuron or monosulfuron ester have the ThDP-HE intermediate bound (21). However, the binding of IQ has no effect on the structure of ThDP (10). It is noteworthy here that IQ binds to the enzyme at the entrance of the active site ∼2 Å farther away from the C2 of ThDP than the SUs (10) and the four compounds whose structures have been determined here. These structural data strongly indicate that the binding of each herbicide results in its own chemical modification to ThDP, or in the case of IQ it remains unmodified.

An alternative explanation for the degradation/modification of ThDP could have been due to the accumulation of free radicals as a result of exposure to X-rays during data collection. To rule out this possibility, two datasets for uninhibited AtAHAS, the AtAHAS–BS complex, and the AtAHAS–PC complex were collected. One of these was taken instantaneously upon positioning the crystals in the X-ray beam and a second one after 24 min of exposure to synchrotron radiation. A comparison of the electron density maps showed that synchrotron radiation does not affect the structure of ThDP in any of the three samples tested (Fig. S3). The possibility that modification of ThDP could have occurred due to interactions formed between the herbicide and the cofactor in crystallo was also tested. This was done by diffusion of CE or BS into preformed crystals of uninhibited AtAHAS. In both cases the electron density maps showed that ThDP was intact with no additional density at the C2 carbon. This result suggests the chemical modification of this cofactor occurs before cocrystallization, that is, when the protein and the inhibitors are interacting and exchanging in solution, which is also evidenced by the mass spectrometry results.

Fig. S3.

ThDP and Mg²⁺ in AtAHAS. (A–F) 2F_o – F_c (blue) electron density map contoured at 2.0 σ and F_o – F_c (green and red) electron density maps contoured at ±3.5 σ, for ThDP and Mg²⁺ in crystals of AtAHAS in complex with BS (A and B), with PC (C and D), and for the free enzyme (E and F). A, C, and E are the initial electron density maps, and B, D, and F are electron density maps after 24 min of exposure to X-ray radiation. These data show that X-ray radiation does not modify the structure of ThDP. The resolution of the electron density maps for A, C, and E is 2.88 Å, 2.76 Å, and 2.92 Å, respectively. The resolution for the electron density maps in B, D, and F is 3.0 Å.

Binding mode of the PYBs.

The scaffold of the PYBs consists of a benzoate and a pyrimidine ring linked by a sulfur or an oxygen atom (Fig. 1). In PB, two methoxy groups are attached to the pyrimidine ring and a chlorine atom is attached to the benzoate. The linker between the two rings is a single sulfur atom. There are six commercial herbicides that use this scaffold; four of these are proherbicides where the charge on the benzoate is masked by a hydrophobic group. The weeds that these compounds target have an esterase in the cell membrane that converts the proherbicide into the active parent compound (1, 22). Thus, this crystal structure represents the binding mode for a range of commercial products, not just a single compound. BS has a similar scaffold to PB, but differs in that it possesses an oxygen atom in the linker between the rings and has an additional dimethoxy-pyrimidine group instead of a chlorine atom. BS and PB make interactions with 18 and 11 residues of AtAHAS, respectively (Fig. 3A). The common dimethoxy-pyrimidine moiety is inserted deepest into the herbicide-binding site, with the pyrimidine ring forming a displaced π-stacking interaction with W574. This interaction is important for anchoring this class of herbicide to the enzyme. In total, there are >10 van der Waals contacts between this residue and the pyrimidine ring (in both complexes). One of the methoxy oxygen atoms of the dimethoxy-pyrimidine ring forms a hydrogen bond with the side chain of R377 and is located 3.7 Å from the C7 carbon atom of FAD (Fig. S4 A and B). The second methoxy group is within hydrophobic contact of G121, M124, M570, and W574 and is oriented toward the cleaved thiazolium ring of ThDP (∼5.5 Å). In the enzyme–PB complex a water molecule is 2.8 Å from the methoxy oxygen. This water is not observed when BS is bound and may be a consequence of the slightly different angles the dimethoxy-pyrimidine moieties adopt in the two structures. The linkers between the two rings, a sulfur atom for PB and an oxygen atom for BS, form contacts with G121 (hydrophobic) and K256 (polar). The bond angles for the C-S-C of PB and the C-O-C of BS are 108.6° and 113.3°, respectively, consistent with sp³ bond hybridization. The benzoate rings of BS and PB occupy similar positions in the two complexes and form T-shaped π-stacking interactions with F206 and hydrophobic contacts with V196, G201, and A202. The carboxylate group forms salt bridges to the side chains of R377 and K256 and a hydrogen bond with the side chain of S653. In PB the additional substituent, a chlorine atom, forms contacts to P197, M200, and R377. In BS, the chlorine is replaced by a second dimethoxy-pyrimidine, the oxygen linker to this substituent is stabilized by R377 and M200, and the pyrimidine ring forms hydrophobic contacts with P197 and M200. One of the methoxy groups forms contacts with R199, D376, and G654 and the other makes contacts with K256 and Q260 (Figs. 3A and 4). In both complexes, an ordered water molecule stabilizes the cleaved thiazolium ring of ThDP, which then forms part of a hydrogen-bonding network leading back to interactions with the two herbicides (Fig. 5A).

Fig. 3.

Stereoview of the PYB and SCT binding sites in AtAHAS. (A) BS and PB. (B) PC and TCM. The herbicides are shown in ball and stick models, whereas key residues for herbicide binding are depicted as stick models. The ’ indicates that these residues are from the neighboring subunit. The carbon atoms for the enzyme–inhibitor complexes are green (BS), light blue (PB), gray (PC), and gold (TCM). The color scheme for the other atoms is as in Fig. 2.

Fig. S4.

Stereoview of the AtAHAS herbicide-binding site. (A) BS. (B) PB. (C) PC. (D) TCM. The herbicides, FAD, ThDP, and the nearby amino acids are shown as stick models. The ‘ indicates that these residues are from the neighboring subunit. The carbon atoms are green for the herbicides, gold for FAD and ThDP, and light blue for the amino acids, and chlorine is orange. The color scheme for other atoms is as in Fig. 2. Dashed lines represent hydrogen bonds.

Fig. 4.

Connolly surface and herbicide blocking the substrate access channel in AtAHAS. BS is represented in a ball and stick model. The color scheme for the herbicide is as in Fig. 2.

Fig. 5.

Stereoview of the water molecules stabilizing the modified ThDP in the AtAHAS–herbicide complexes. Three ordered water molecules stabilize the modified thiazolium ring via a hydrogen-bonding network extended to the nearby residues and the herbicide. (A) ThAthDP in the PB complex. (B) ThThDP in the TCM complex. The ThDP analogs, the herbicides, and water molecules in each panel are superposed onto the difference electron density maps contoured at 3.5 σ. The carbon atoms for ThDP, FAD, herbicides, and the nearby residues are green, yellow, cyan, and light blue, respectively. Water molecules are represented as orange spheres. The color scheme for the other atoms is as in Fig. 2.

Binding mode of the SCTs.

The scaffold of PC and TCM consists of a triazolinone and a sulfonylamino-carbonyl linker. TCM has a disubstituted thienyl attached to the linker. Alternatively, PC has a monosubstituted aromatic ring attached in the same location (Fig. 1). PC and TCM make interactions with 15 and 10 residues of AtAHAS, respectively (Fig. S4 C and D). Many are in common with residues that bind to the PYBs. The triazolinone moiety in both compounds inserts deepest into the herbicide-binding site (Fig. 3B), forming π-stacking interactions with W574 and G121. A hydrogen bond also forms between the exposed N2 nitrogen and R377 (Fig. S4 C and D) and the propyl group forms hydrophobic contacts with M351, H352, and the C6 (3.59 Å) and C7 (3.62 Å) atoms of the flavin ring. Because this moiety is elongated compared with the other compounds in this series, it uniquely fills the crevice between the FAD and H352. In the TCM complex the carbon atom of the methoxy group is 4.1 Å from the C7 atom of the flavin ring. The oxygen of the ether linkage forms hydrogen bonds with two water molecules, and a third water molecule is also nearby where it hydrogen bonds to the amide of G121. These three waters stabilize the carbonyl oxygen atom attached to ThThDP, filling the space between the cofactor and the herbicide (Fig. 5B). In both structures, the methyl substituent forms a hydrophobic contact with G121, and the carbonyl substituent hydrogen bonds to the side chain of K256.

The sulfonylamino-carbonyl linker is important for inhibitor binding. The carbonyl oxygen forms two polar contacts to R377 and S653. The amino group in this linker has a pk_a of 2.1 and 3.0 for PC and TCM, respectively. Thus, under physiological conditions and in the crystallization buffer is charged, allowing a salt bridge to form with the side chain of K256. One of the oxygen atoms of the sulfonyl group (in both compounds) forms a hydrogen bond with the side chain of S653 and a water molecule. The other sulfonyl oxygen forms a hydrogen bond with a water molecule and makes hydrophobic interactions with K256 and P197. The aromatic ring in PC located at the entrance of the herbicide-binding site is held in place by hydrophobic contacts to V196, M200, and S653. Likewise the thienyl ring forms limited contacts with the enzyme (i.e., to the side chain of V196 and the backbone of F206). The methyl ester substituent is common to both compounds and forms hydrophobic contacts with the side chain of F206, K256, G121, and Q207. Somewhat surprisingly, none of the oxygen atoms in this group forms hydrogen bonds with the enzyme or cofactors.

Comparisons of the different modes of herbicide binding.

The structural data for the four compounds presented here emphasizes the difficulty in trying to predict the mode of binding of the different herbicide classes to AHAS. Key to the binding of the PYBs and SCTs is the modifications to the ThDP cofactor that occur and the unique interactions that are formed as a result. The effects these have on binding could not have been predicted in the absence of these crystal structures. The degree of flexibility that the PYBs and SCTs and the residues in the herbicide-binding site possess adds another level of complexity in predicting binding modes. The side chains of R199, M200, K256, R377, and W574 are the common residues where the most significant conformational changes are observed upon PYBs and SCTs binding to AtAHAS (Fig. S5). Superposition of the crystal structures of AtAHAS in complex with herbicides from the PYB, SCT, SU, and IMI families (e.g., BS, PC, and CE, PDB code 1YBH; and IQ, PDB code 1Z8N, respectively) (Fig. 6) shows that the herbicides from these four inhibitor classes fit in the herbicide-binding site by significantly different mechanisms. Indeed, only partial overlap is observed among the four families. This mainly occurs at the deepest part of the herbicide pocket, where the heterocyclic rings of the PYBs (pyrimidine), SUs (pyrimidine or triazine), and SCTs (triazolinone) are inserted. These rings adopt similar positions due to the common π-stacking interactions formed with W574. However, the shorter linker in the PYBs compared with the SUs and SCTs means that in this family of compounds the pyrimidine ring has a different orientation with respect to the indole moiety of W574. Compared with the heterocyclic rings of CE or PC, which are parallel to W574, the pyrimidine ring of BS is rotated by ∼20°. The positions of the heterocyclic ring substituents also vary by 1.05–2.90 Å (Fig. 6, Left). This results in the formation of a different set of contacts with FAD, M124, M351, H352, M570, and V571. Key to the binding of PYBs, SCTs, and SUs is the hydrogen bonding formed between R377 and the oxygen atoms of the carboxylate (PYBs) or the carbonyl groups (SCTs and SUs) whose positions in the binding site overlap (Fig. 6). The disubstituted aromatic rings of SCTs and SUs located at the entrance of the access channel occupy similar positions. However, the number of contacts formed between the SU herbicides and the enzyme is greater than those observed in the SCT complexes. Surprisingly, the locations of the benzoate ring of the PYBs and the second pyrimidine ring of BS do not overlap with any of the aromatic rings from the SCT or SU families. These two ring systems do, however, partially overlap with the positions of the dihydroimidazolone and the quinolone rings of IQ, respectively. It is worth noting the IMIs are the only family of AHAS herbicides that bind at the entrance of the substrate access channel, with a larger portion of the inhibitor molecule protruding toward the protein surface. This is a unique feature for this class of compounds. The very specific role that the ordered water molecules play, especially in forming bridging interactions between ThThDP and the herbicides, is a finding not previously anticipated. BS has the special feature of having three aromatic rings in its structure. Together these adopt a twisted “S” shape (Fig. 2B), a conformation not readily predictable by modeling. PB and BS each have a single atom in the linker between two of the aromatic rings. In the light of the structures of the SUs, having an extended sulfonylurea group as the bridge, it was difficult to envisage how such a compact linker could provide enough contacts to anchor the inhibitor to the enzyme. However, both members of the PYB family competently achieve this result.

Fig. S5.

Stereoview of the conformational changes in the AtAHAS herbicide-binding site. Five residues of the herbicide-binding site adopt a different position depending on whether (A) BS or (B) PC is bound compared with the uninhibited enzyme (PDB code 5K6Q). Amino acid residues are represented as stick models and BS and PC are represented as ball and stick models. The carbon atoms for the uninhibited enzyme and the enzyme in complex with herbicide are colored pink and cyan, respectively. The color scheme for the other atoms is as in Fig. 2. The ’ indicates that these residues are from the neighboring subunit. Dashed lines represent hydrogen bonds.

Fig. 6.

Stereoview of the binding mode of the four different families of herbicides in complex with AtAHAS. The herbicides and amino acids of the herbicide-binding site are colored light blue for BS, a member of the PYB family; cyan for PC, a SCT; gold for chlorimuron ethyl (CE), a SU with PDB code 1YBH; and pink for IQ, an IMI with PDB code 1Z8N.

Inhibition of AHAS by the Different Chemical Classes of Herbicide.

A challenge in the determination of the inhibition constants for these herbicides that inhibit AHAS is to account for the accumulative inhibition that occurs during the assay. Previously, we have shown that partial O₂ depletion decreases the amount of accumulative inhibition (9), and that condition was used here for measurement of K_i values. For illustration, the inhibition of AtAHAS by PB under partial anaerobic conditions (Fig. 7A) was significantly weaker than the control carried out under aerobic conditions (Fig. 7B). Another factor for consideration is the presence of an initial lag phase (23). To circumvent this effect, inhibitors were added to the assay mixture 12 min after the enzyme, a point where the maximum enzyme activity was attained (Fig. 7). Using this partially anaerobic method, the K_i values of six commercial herbicides representing four herbicide families were determined (Table 2).

Fig. 7.

Time course for the inhibition (K_i) of AtAHAS by PB. (A and B) Monitoring the AtAHAS (15 nM) catalyzed reaction under (A) partial anaerobic and (B) aerobic conditions, in the absence (●) or presence (▼) of PB (100 nM), using a colorimetric assay (45). The dashed lines represent the linear section of the curve after the injection of PB (arrow). In this region, the enzymatic rate of the uninhibited enzyme is linear. The rates obtained under partial anaerobic conditions are used for the calculation of K_i (Materials and Methods). The bars represent experimental errors from triplicate measurements.

Table 2.

Inhibition constants and kinetic rate constants for accumulative inhibition of the commercial herbicides for AHAS

Inhibitor	Ki, nM	kiapp, min−1	k3, min−1
BS	40.9 ± 6.1	0.32 ± 0.03	0.022
CE	74.7 ± 6.1	1.22 ± 0.07	0.012
PB	179 ± 11.6	0.81 ± 0.02	0.037
PC	434.8 ± 33.8	0.90 ± 0.02	0.037
TCM	670.8 ± 64.0	0.43 ± 0.04	0.015
IQ	18.5 ± 2 × 103	0.87 ± 0.03	1.3 × 10−12

SEM is shown. BS, bispyribac; CE, chlorimuron ethyl; IQ, imazaquin; PB, pyrithiobac; PC, propoxycarbazone; TCM, thiencarbazone methyl.

The results show that there is a 5- to 10-fold increase in the K_i values for all of the compounds compared with the values obtained under aerobic conditions and where the lag phase was not considered. For instance, the K_i value for CE measured here is 74.7 ± 6.0 nM, ∼7-fold higher than the value reported previously by Chang and Duggleby (4) (K_i = 10.8 nM). Across the four families of commercial herbicides there is a wide variation in the K_is with BS having the lowest value of 41.9 nM and IQ the highest with a K_i of 18.5 μM. The wide disparity in K_i values across all of these commercial herbicides indicates that there are other reasons than the straight binding event contributing to their herbicidal activity. Among these are the absorption properties of each herbicidal molecule and their ability to translocate to meristematic tissues (24, 25), their ion-trapping capacity in the plant cell (26), and the rate of herbicide metabolism via detoxification (27). However, here we explain aspects of the inhibition of AHAS that play a significant role in the extraordinary herbicidal activity of these compounds. These are (i) accumulative inhibition and (ii) modification of the cofactor ThDP.

Accumulative Inhibition.

Accumulative inhibition is observed when each of the herbicides is assayed against AtAHAS (Fig. 8). The values of the rate constants k_iapp and k₃ (Table 2) were obtained by fitting the inhibition curves to Eq. 1 (9) (Fig. S6). It is notable that the relatively low affinity of IQ, TCM, and PC for the enzyme (Table 2) had to be taken into account in fitting to Eq. 1 (Materials and Methods).

Fig. 8.

Progress curves for the inhibition of 1 μM AtAHAS by six different herbicides at an inhibitor concentration of 100 nM. BS, bispyribac; CE, chlorimuron ethyl; IQ, imazaquin; PB, pyrithiobac; PC, propoxycarbazone; TCM, thiencarbazone methyl.

Fig. S6.

Progress curves for the inhibition of AtAHAS by six different herbicides. BS, bispyribac; CE, chlorimuron ethyl; IQ, imazaquin; PB, pyrithiobac; PC, propoxycarbazone; TCM, thiencarbazone methyl. The inhibition of 1 μM AtAHAS was assayed in the presence of 100 nM herbicide, except for CE (50 nM). The experimental data were fitted using Eq. 1 (9) (dashed lines), giving the rates k_iapp and k₃ for each herbicide (Table 2). The effective ratio of free enzyme/enzyme–inhibitor complex (F value) for PC, TCM, and IQ has been calculated according to the inhibition constant formula (Eq. 2), taking into account the lower affinity of these herbicides for the enzyme (Materials and Methods).

Based on these data, several general conclusions could be made: (i) the k_iapp and k₃ rates are variable and there is no obvious correlation between k_iapp and k₃ values, in agreement with these two rates representing independent events. k_iapp represents the apparent rate of inactivation in presence of the inhibitor, whereas k₃ represents the enzyme recovery rate that occurs in absence of inhibitor (9). (ii) The k₃ value of IQ is so small that it is effectively zero. This result is significant because it infers that the inhibition by IQ is virtually irreversible. A largely unexplained feature of the herbicidal activity of the IMIs, including IQ, is that they have significantly higher K_i values (>1,000-fold) than those of the other AHAS-inhibiting herbicide classes, yet they are equally potent as herbicides and require comparable application rates in the field. The high K_i value of IQ for the enzyme suggests that under physiologic conditions, the free enzyme/enzyme–inhibitor complex ratio is low in the plant cell compared with other herbicides that have a higher affinity (i.e., lower K_i value). However, the very low k₃ value implies that if given enough time, IQ should reach an inhibition efficiency that is similar to those of the more tight-binding herbicides. To illustrate, we have compared the percentage of enzyme inhibition in two conditions based on predictions using Eq. 1 with the parameters determined in this study (k_iapp and k₃, Table 2).

In the first scenario, where the free enzyme/enzyme–inhibitor complex ratio (F parameter in Eq. 1) is given an arbitrary value of 10 (to reflect a situation where herbicides have a high affinity for the enzyme) and an incubation of 2 h, more than 50% of the enzyme molecules are predicted to be inhibited for CE, PB, and BS (Fig. 9, shaded bars). For IQ, PC, and TCM the free enzyme/enzyme–inhibitor ratio has to be adjusted to 196, 14.7, and 17.2, respectively, to take into account that these compounds are weaker inhibitors compared with the SUs and the PYBs (Materials and Methods). Accordingly they show a low efficiency in this situation. In the second scenario, a free enzyme/enzyme–inhibitor complex ratio value of 1,000 (F = 1,000) takes in account the low affinity of IQ. In such conditions, only IQ is shown to be efficient with an inhibition of more than 80% of the enzyme if incubation is allowed for 2 d (Fig. 9, solid bars).

Fig. 9.

Prediction of the percentage of enzyme molecules inhibited by the AHAS herbicides. Concentrations for AtAHAS and inhibitors are as in Fig. 8. Shaded/solid bars represent the expected percentage after 2 h/2 d when the ratio of free enzyme to enzyme–inhibitor complex is 10:1/1,000:1. *Effective value of F has been calculated according to the inhibition constant formula (Eq. 2) (Materials and Methods).

The rate of metabolism of IMIs is markedly slow. For instance, the half-life values of IQ in velvetleaf and cocklebur are 12 d and 30 d, respectively (28). This renders the possibility of a long period of activity for imazaquin at the site of action. Accordingly, weeds do show a delayed response to the application of IMIs compared with other families of herbicides (29, 30), which correlates with our predictions derived from Eq. 1 (Fig. 9). These findings provide a rational explanation for this “unexpected” potency of IQ and show that “weak” binders may be very efficient, due in part to accumulative inhibition.

Inhibition Through ThDP Modification.

The crystal structures of the AtAHAS–herbicide complexes show that ThDP can be modified, raising the question of whether such a chemical reaction occurs in solution and during catalysis in the presence of herbicides. The modification of ThDP (where the pyrimidine and/or the thiazolium rings are cleaved or the thiazolium ring is oxidized to the thiazolone form) was assessed in the presence and absence of herbicide by fluorescence spectroscopy (31), where the concentrations of ThDP and AtAHAS are identical. A small change in the fluorescence of ThDP was observed in absence of any herbicide (17% modification after 60 min) in accordance with that observed for E. coli AHAS isoenzyme II (31). The results show that compared with the control, CE, PB, and PC trigger a significant modification of ThDP (50% change in fluorescence after 30 min) (Fig. 10A). However, the kinetics show a biphasic event, where the second phase matches the rate of the control, starting from when pyruvate is depleted in the reaction mixture (Fig. 10B). In correlation with this fact is that ThDP remains intact in the presence of herbicide (e.g., CE) when pyruvate has been omitted from the reaction mixture (Fig. 10A). These results suggest that the production of an enzyme-bound intermediate is required for the modification of ThDP by herbicides. Accordingly, our data show that under turnover conditions PB, CE, and PC produce a 5.6-, 6.7-, and 9.5-fold increase, respectively, in the modification of ThDP relative to the control experiment in the presence of pyruvate. In contrast, the degradation of ThDP induced by imazaquin is significantly lower (1.2-fold increase) and correlates with the structural data that show this herbicide has no effect on the enzyme cofactor. These data, the crystallographic studies, and mass spectrometry together suggest that the modification of ThDP forms an integral part of the mechanism of AtAHAS inhibition by these herbicides.

Fig. 10.

ThDP degradation triggered by herbicide inhibition. (A) A total of 80 μM AtAHAS was incubated with 80 μM of ThDP at 30 °C, in the (●) absence or presence of 16 μM of herbicide [(○) CE, (□) IQ, (■) PB, and (△) PC] in buffer containing 100 mM pyruvate. (▲) A total of 80 μM AtAHAS incubated with 80 μM ThDP, 16 μM CE, and without pyruvate was used as a control. The decay of ThDP was monitored by fluorescence spectroscopy (31). (B) The enzyme activity was monitored at 525 nm, using a colorimetric assay (45). The bars represent experimental errors from duplicate measurements. (C) Effect of ThDP degradation on the time-dependent inhibition of AtAHAS by CE. Curve a: 10 μM AtAHAS incubated with a suboptimal concentration of ThDP (1 μM) and 0.05 μM CE. Curve b: 1 μM AtAHAS incubated with a saturating concentration of ThDP (1 mM) and 0.05 μM CE. AtAHAS activity was monitored at 333 nm. The more pronounced time-dependent inhibition of curve a is attributed to the degradation of ThDP.

The inhibition of AtAHAS by CE increases in intensity when the ratio of ThDP vs. enzyme decreases from 1,000 to 0.1 in the assay solution (Fig. 10C), suggesting that reversible accumulative inhibition and inhibition by the way of ThDP degradation affect the enzyme activity simultaneously. This implies that, within the cell, AHAS inhibitors could cause a deficiency in active ThDP molecules. A crucial factor that determines the bioactivity of herbicides from the five different families AHAS inhibitors is their ability to translocate to actively dividing meristematic cells where AHAS is primarily expressed (2). In these cells, ThDP plays crucial roles in the process of cell division and production of energy and thus an impaired biosynthesis (32, 33) or transport (34) of the cofactor causes an immediate reduction of plant growth, accompanied by poor shoot and root development, chlorosis, and necrosis. Furthermore, the biosynthesis of ThDP in growing tissues (shoots and roots) is low, and it has to be to transported from mature tissues to satisfy the demand for this cofactor (34). In this condition, the degradation of ThDP by herbicide can play a very important role in inhibiting the plant growth and, accordingly, the symptoms mentioned above correlate with those observed when AHAS herbicides are in action (2).

Conclusions

The crystal structures of AtAHAS in complex with the PYBs and SCTs have revealed unexpected modes of binding that could not have been predicted by docking or modeling. In general, the PYBs and SCTs occupy a similar location to that where the SUs bind. The crystal structures explain why the linker between the benzoate and pyrimidine ring in the PYBs can be as short as a single atom and indeed addition of any extra atoms is not practical as there is insufficient space for this to occur. The presence of ThAthDP and ThThDP was initially detected in the electron density maps for the crystals of PYB and SCT complexes with AtAHAS, respectively. Mass spectrometry and kinetics confirmed that modification of ThDP is an essential part of the mechanism of inhibition of AHAS by the commercial herbicide families.

The kinetics of inhibition of AtAHAS by herbicides from four different chemical classes of AHAS inhibitors were also characterized and these were shown to be enhanced by reversible accumulative inhibition. Our findings provide a rational explanation for the extraordinary efficacy of the AHAS inhibitors, including those with weak binding to the target enzyme (e.g., IQ), and provide unique insights into the mode of action of AHAS inhibitors via an herbicide-induced degradation/modification of ThDP with likely implications in the plant cell metabolism.

Materials and Methods

Protein Expression and Purification.

Expression and purification of the catalytic subunit of AtAHAS were carried out as reported previously, but with some differences (35). Here, a single colony was grown overnight at 37 °C in 100 mL of Luria–Bertani (LB) media supplemented with kanamycin (50 μg/mL). A total of 40 mL of this culture was then transferred to 1 L of Terrific Broth (TB) supplemented with kanamycin (50 μg/mL) and incubated at 37 °C with shaking [150 oscillations per minute (opm)]. When the OD₆₀₀ reached 2.0, expression of AtAHAS was induced with 0.5 mM IPTG, and cells were further grown for 18 h at 15 °C with shaking (150 opm). Cells were collected by centrifugation at 5,100 × g for 20 min, and the cell paste was snap frozen with liquid nitrogen and stored at −70 °C. All subsequent operations were performed at 4 °C, protecting the enzyme from light as much as possible. The cell paste was thawed and resuspended in 50 mM Tris, pH 7.9, 10 mM MgCl₂, 1 mM ThDP, 10 μM FAD, and 20 mM imidazole (binding buffer). For lysis, 10 mg lysozyme per gram of cell, 50 μL DNase (50 μg/μL), and a capsule of protease inhibitor mixture (Complete, EDTA-free from Roche) were added. The cell suspension was incubated on ice for 15 min and then sonicated in a Branson Sonifier 250 for 12 × 10 s at constant duty cycle, with rest intervals of 10 s. The cell debris was removed by centrifugation at 39,100 × g for 60 min. The supernatant was loaded onto a Ni-NTA resin column (Qiagen) previously equilibrated with binding buffer, using an ÄKTAprime plus FPLC system (GE). After the supernatant was loaded, the column was washed with binding buffer. The enzyme was then eluted with 50 mM Tris, pH 7.9, 10 mM MgCl₂, 1 mM ThDP, 10 μM FAD, and 300 mM imidazole (elution buffer). The flow-through was collected, and the fractions containing the protein were pooled and 50 mM sodium pyruvate, 250 mM NaCl, and 1 mM DTT were subsequently added to the samples. The protein was incubated on ice for 10 min and then loaded onto an S-200 HR gel filtration column (Pharmacia) previously equilibrated with 50 mM KPO₄ buffer, pH 7.2, 10 mM MgCl₂, 10 μM FAD, and 1 mM DTT (GF buffer). The protein was eluted with 400 mL GF buffer and the flow-through was collected in 5-mL fractions. The fractions containing the protein were pooled and concentrated using an Amicon concentrator device. The purified enzyme was stored in aliquots of 30 μL at −70 °C. All chemicals used were of analytical grade and were purchased from Sigma-Aldrich, unless otherwise stated. Protein concentration was measured with Direct Detect system (Millipore).

Protein Crystallization and Structure Determination.

Crystallization trials were performed using the hanging-drop vapor-diffusion method as described previously (35). The protein solution was prepared before starting the experiment by incubating freshly thawed protein (40 mg/mL) with 5 mM MgCl₂, 1 mM ThDP, 1 mM FAD, 5 mM DTT, and when appropriate 1 mM inhibitor. Crystals of AtAHAS in complex with PC or TCM were obtained by mixing well solution and protein solution in a 1:2 ratio. The well solution contained 1 M Na/K tartrate, 0.1 M 2-(N-cyclohexylamino)ethanesulfonic acid (CHES), and 0.19 M (NH₄)₂SO₄. The pH of the solution ranged from 9.4 to 9.8. Crystals of the AtAHAS in complex with PB or BS were obtained by mixing equal volumes of protein solution and well solution containing 20% wt/vol PEG 3350 and 0.2 M sodium citrate tribasic dihydrate (PB) or 0.2 M potassium citrate tribasic dihydrate (BS). The pH of these solutions was adjusted to 8.3. The crystallization experiments were kept at 290 K and protected from light until the crystals reached their maximum size. For cryocooling the crystals were transferred to a drop containing well solution and 35% vol/vol ethylene glycol, cofactors, and inhibitor. The crystals were mounted in cryoloops and frozen in liquid nitrogen.

The datasets were collected with an ADSC QUANTUM 210r detector, using the MX1 beam line at the Australian Synchrotron, operated at 13 keV. The datasets were indexed, integrated, and scaled using XDS (36) and Aimless (37) or HKL2000 (38). The crystal structures were solved by molecular replacement with PHASER (39), using AtAHAS in complex with metsulfuron-methyl as the starting model (PDB code 1YHY). Model building was carried out using Coot 0.8.1 (40) and refinement was with Phenix 1.9-1692 (41). The electron density maps in the two AtAHAS–SCT complexes and in the AtAHAS–PB complex also showed a molecule from the crystallization buffers, being the CHES and citrate, respectively. These molecules are remote from the active and herbicide-binding site and are not likely to influence the enzyme structure or the binding of the herbicides. C340 was modified to its sulfonic acid derivative as in previous structures of AtAHAS (10, 21). Figs. 2B and 3–6 and Figs. S1 and S3–S5 were generated with CCP4mg (42) and Fig. 2A by PyMOL (43).

Mass Spectrometry Assays.

The mass spectrometry analyses for ThDP were performed in the presence and absence of inhibitor. Crystals of AtAHAS in complex with PB or PC were washed and dissolved in deionized (DI) water at room temperature and the protein was precipitated by mixing with an equal volume of methanol containing 5% (vol/vol) acetic acid on ice. The precipitate was removed by centrifugation at 11,000 × g for 20 min at 4 °C, using a 3-kDa Nanosep device (PALL). In parallel 700 μM AtAHAS was incubated with 200 mM potassium phosphate (pH 7.2), 100 mM sodium pyruvate, 10 mM MgCl₂, 6 mM ThDP, 1 mM FAD, and 70 μM inhibitor (PB or PC) for 5 d at 18 °C. DI water was added instead of inhibitor for the control sample. The samples were then dialyzed (10-kDa dialysis tubing) exhaustively against DI water at 4 °C. The protein was precipitated by incubation overnight with an equal volume of methanol containing 5% (vol/vol) acetic acid. The precipitate was removed by centrifugation at 11,000 × g for 25 min at 4 °C, using a 3-kDa Nanosep device. All samples were then freeze dried and resuspended in 20 μL of DI water. Samples were first injected into a low-resolution Bruker HCT ion trap ESI mass spectrometer (5 μL/min) in positive ion mode and measured across the mass range of 300–900 m/z with a spray voltage of 4,000 V. Exact masses for ThDP and its analogs were determined by high-resolution analysis, using a Bruker micrOTOF-Q mass spectrometer. Samples were separated using reversed-phase chromatography on a Dionex Ultimate 3000 RSLC nanosystem. Samples were injected onto a Zorbax Eclipse XDB-C18 column (2.1 × 50 mm, 3.5 μm), using 100% buffer A, where buffer A consisted of 0.1% formic acid in MilliQ water, and buffer B was 100% acetonitrile. The liquid chromatography (LC) program included 100% A (0–3 min), 0–98% B (3.01–7 min), held at 98% B for 2 min, followed by reequilibration of the column in buffer A. The eluted sample was directly analyzed on a Bruker micrOTof mass spectrometer, using an ESI interface. Source parameters included capillary, 3,500 V; dry gas, 5 L/min; dry temp, 200 °C; end plate offset, 500 V; and collision energy, 2 V. The MS scan was performed across 50–1,500 m/z. Using a valve switcher, Agilent ESI tune mix was infused in the LCMS run for 30 s before acquisition of data eluted from the C18 column to allow internal calibration of LCMS data, post-MS acquisition. Calibration and determination of accurate molecular formulas was performed using DataAnalysis v4.2.

AHAS Assay.

For all assays, otherwise mentioned, the standard reaction buffer contained 200 mM potassium phosphate (pH 7.2), 100 mM sodium pyruvate, 10 mM MgCl₂, 1 mM ThDP, and 10 μM FAD and the assays were conducted at 30 °C.

For accumulative inhibition assays, 1 μM AtAHAS was assayed using the continuous method, measuring the disappearance of pyruvate (43). The inhibitors were added 12 min after the enzyme so that inhibition could be measured when maximum activity of the enzyme was achieved. The accumulative inhibition induced by the herbicides was observed for a 40-min period. All experiments were carried out in triplicate and the data were fitted using Eq. 1 (9) for the calculation of the rates (k_iapp and k₃) involved in accumulative inhibition:

$$\left[\text{P}\right]=V_{max}\left(F\left(\frac{k_3}{k_{iapp}+F{k}_3}\right)t+\left(\frac{k_{iapp}}{k_{iapp}+F{k}_3}\right)\frac{1}{k_{iapp}/F+k_3}\left(1-exp\left[-\left(\frac{k_{iapp}}{F}+k_3\right)t\right]\right)\right).$$ [1]

For the inhibitors IQ, TCM, and PC, their relatively high K_i value was taken into account in the estimation of the value F (free enzyme/enzyme–inhibitor complex ratio). The effective concentration of enzyme–inhibitor was calculated using the inhibition constant formula

$$\left(K_i=\left[\mathrm{E}\right]\left[\mathrm{l}\right]/\left[\mathrm{EI}\right]\right),$$ [2]

the known initial concentrations of enzyme and inhibitor, and the K_i value of Table 2.

The effect of O₂ in the AHAS inhibition assay was determined by the colorimetric method (45), using standard reaction buffer purged with air or N₂ for 60 min. A total of 2 μL of 22.5 μM AtAHAS was added to 3 mL purged buffer (which incorporated ∼0.06% of O₂ present in the air-purged solution) to initiate the reaction. After 12 min of reaction at 37 °C, 2 μL of 0.15 M PB purged with air or N₂ was added to the assays. To monitor the enzyme activity 100-μL samples were taken from the reaction mixture every 3 min over a 45-min period and the reaction was immediately stopped by the addition of 10% H₂SO₄ (vol/vol) (to reach a final concentration of 1%). In this step, acetolactate is decarboxylated and transformed to acetoin after incubation at 60 °C for 15 min. The acetoin formed was quantified by measuring the absorbance at 525 nm after incubation with creatine (0.17%, wt/vol) and α-naphthol (1.7%, wt/vol) at 60 °C for 15 min and later at room temperature for 15 min to develop maximum color. The extinction coefficient used for the colored complex is ε_M = 22,700 M⁻¹⋅cm⁻¹ (8). The experimental data observed for the reaction in the absence of O₂ remained linear for ∼15 min after adding the inhibitor to the reaction, longer than that observed for the experiment carried out under aerobic conditions. This offered a suitable window for the measurement of initial rates.

The measurement of the K_i values was carried out at a range of 10–12 different inhibitor concentrations, including no inhibitor, in conditions identical to the one described for the partial anaerobic reaction mentioned above. The K_i values were obtained by nonlinear regression, using Eq. 3 [weak to medium binding (4)], taking into account the first 6 min after adding the inhibitor (from t = 12 min to t = 18 min):

$$v_i=v_o/{\left(1+\left[I\right]/K\right)}_i.$$ [3]

Anoxic conditions produced higher enzyme activity in comparison with the assay carried out in the presence of O₂. This increase in enzyme activity was expected. Tse and Schloss (44) showed that O₂ competes with pyruvate for binding to HeThDP in the second step of the reaction, producing peracetate-ThDP instead of acetolactate-ThDP. Therefore, the observed increase of enzyme activity is attributed to the displacement of O₂ by N₂, inducing an effective increase of acetolactate production and a decrease in oxygenase activity. This further demonstrates the role N₂ plays in the minimization of factors that lead to oxidative inactivation of the enzyme and constitutes the basis for the establishment of a method to determine the K_i of AHAS inhibitors in the absence of accumulative inhibition.

ThDP Assay.

ThDP was assayed by the thiochrome method as reported previously (31), but with some modifications. AtAHAS (80 μM) was incubated in the absence and presence of herbicide (16 μM) in buffer containing 100 mM potassium phosphate (pH 7.2), 10 mM MgCl₂, 80 μM ThDP, and 80 μM FAD. The reaction was conducted at 30 °C and monitored for 60 min after adding pyruvate and herbicide or just herbicide alone. Samples of 10 μL were collected at intervals of 10 min and mixed with 600 μL 50% ethanol (vol/vol) before adding 100 μL 0.04% K₃Fe(CN)₆ (wt/vol) in 15% NaOH (wt/vol). The reaction was completed by the addition of 2 μL 30% H₂O₂ (vol/vol) after 2 min of incubation at room temperature. Fluorescence of the formed thiochrome was measured immediately with a CytoFluor series 4000 fluorescence multiwell plate reader (PerSeptive Biosystems), using excitation at 360 nm and emission at 460 nm, with bandwidths set at 20 nm. The residual cofactor was expressed as the percentage of that present initially, which was determined by reference to standard solutions of 0 μM, 10 μM, 20 μM, 40 μM, 60 μM, 80 μM, 100 μM, and 200 μM ThDP assayed using the same procedure. The enzyme activity was monitored in parallel at intervals of 5 min, using a colorimetric method (45) as described above, with the only difference being that the reaction mixture was diluted before the addition of 10% H₂SO₄. This was accounted for by transferring samples of 5 μL (reaction mixture) to 20 μL of 50 mM potassium phosphate (pH 7.2) and then 2 μL of this solution to 98 μL of 50 mM potassium phosphate (pH 7.2), which was finally used for the colorimetric assay. All experiments were carried out in duplicate.