Single-site mutations in the carboxyltransferase domain of ... · 1/4/2007  · Single-site mutations in the carboxyltransferase domain of plastid acetyl-CoA carboxylase confer resistance

Single-site mutations in the carboxyltransferasedomain of plastid acetyl-CoA carboxylase conferresistance to grass-specific herbicidesWenjie Liu*, Dion K. Harrison*†, Dominika Chalupska‡, Piotr Gornicki‡, Chris C. O’Donnell§, Steve W. Adkins§,Robert Haselkorn†‡, and Richard R. Williams*

*Agricultural Molecular Biotechnology Laboratory, School of Agronomy and Horticulture, University of Queensland, Gatton 4343, Queensland, Australia;§Tropical and Subtropical Weeds Research Unit, School of Land and Food Sciences, University of Queensland, Brisbane 4072, Queensland, Australia;and ‡Department of Molecular Genetics and Cell Biology, University of Chicago, Chicago, IL 60637

Contributed by Robert Haselkorn, January 4, 2007 (sent for review December 20, 2006)

Grass weed populations resistant to aryloxyphenoxypropionate(APP) and cyclohexanedione herbicides that inhibit acetyl-CoAcarboxylase (ACCase; EC 6.4.1.2) represent a major problem forsustainable agriculture. We investigated the molecular basis ofresistance to ACCase-inhibiting herbicides for nine wild oat (Avenasterilis ssp. ludoviciana Durieu) populations from the northerngrain-growing region of Australia. Five amino acid substitutions inplastid ACCase were correlated with herbicide resistance: Ile-1,781-Leu, Trp-1,999-Cys, Trp-2,027-Cys, Ile-2,041-Asn, and Asp-2,078-Gly(numbered according to the Alopecurus myosuroides plastid AC-Case). An allele-specific PCR test was designed to determine theprevalence of these five mutations in wild oat populations sus-pected of harboring ACCase-related resistance with the result that,in most but not all cases, plant resistance was correlated with one(and only one) of the five mutations. We then showed, using ayeast gene-replacement system, that these single-site mutationsalso confer herbicide resistance to wheat plastid ACCase: Ile-1,781-Leu and Asp-2,078-Gly confer resistance to APPs and cyclohex-anediones, Trp-2,027-Cys and Ile-2,041-Asn confer resistance toAPPs, and Trp-1,999-Cys confers resistance only to fenoxaprop.These mutations are very likely to confer resistance to any grassweed species under selection imposed by the extensive agriculturaluse of the herbicides.

aryloxyphenoxypropionate � Avena � cyclohexanedione

Aryloxyphenoxypropionates (APPs) and the cyclohexanedio-nes (CHDs) are often used to control grass weeds selec-tively. These herbicides target the fatty acid biosynthetic pathwayof grasses by inhibiting the plastid form of the enzyme acetyl-CoA carboxylase (ACCase; EC 6.4.1.2). However, many weedpopulations have become resistant to APPs and CHDs, includingsome of the major grass weeds, such as wild oat (Avena fatua L.and Avena sterilis ssp. ludoviciana Durieu), rigid ryegrass (Lo-lium rigidum Gaudin), black grass (Alopecurus myosuroidesHudson), and green foxtail (Setaria viridis L. Beauv) (1). Here wedescribe the distribution of mutations conferring resistance tothese herbicides in several wild oat populations in the northernwheat-growing areas of Australia.

In both eukaryotes and prokaryotes, ACCase is a biotinylatedenzyme that catalyzes the first committed step of de novo fatty acidbiosynthesis by carboxylation of acetyl-CoA to malonyl-Co in atwo-step reaction: carboxylation of the biotin group of the enzyme,followed by transfer of the carboxyl group from carboxybiotin toacetyl-CoA by the carboxyltransferase (CT) activity. In plants,ACCase activity is found in both plastids where primary fatty acidbiosynthesis occurs and the cytosol where synthesis of very long-chain fatty acids and flavonoids occurs. Selectivity of APP andCHD herbicides is due to the different types of plastid ACCasefound in plants. The multidomain type found in the cytosol of allplants and the multisubunit type found in plastids of dicots areinsensitive to APPs and CHDs. In contrast, the plastid ACCase in

grasses is herbicide-sensitive. Expression of the latter is high in themeristematic region of young plants (2), reflecting the demand formalonyl-CoA in dividing and fast-growing cells and consistent withthe high efficacy of postemergence application of the herbicides.APP and CHD herbicides interact with the CT domain of ACCase(3). The APP-binding site has been inferred from the 3D structureof the CT domain of yeast ACCase complexed with haloxyfop (4).

Five amino acid substitutions in the CT domain have beenimplicated in resistance to APP and/or CHD herbicides: anIle-1,781-Leu substitution in A. myosuroides (5–7) as well ashomologous substitutions in L. rigidum (8, 9), S. viridis (10), A.fatua (11), and Lolium multiflorum (12); an Ile-2,041-Asn sub-stitution in A. myosuroides as well as homologous substitution inL. rigidum (13); and Trp-2,027-Cys, Gly-2,096-Ala, and Asp-2,078-Gly substitutions in A. myosuroides (14). Ile-1,781-Leu andAsp-2,078-Gly mutations are correlated with resistance to APPsand CHDs, whereas Trp-2,027-Cys, Ile-2,041-Asn, and Gly-2,096-Ala are correlated with resistance to APPs but not CHDs.Knowledge of the molecular basis of resistance to ACCase-inhibitors caused by mutations in the enzyme is based primarilyon characterization of the diploid weed species A. myosuroidesand L. rigidum. It was also shown that a single amino acid changein wheat (Triticum aestivum L.) plastid ACCase, correspondingto the Ile-1,781-Leu substitution, makes the enzyme resistant toAPPs and CHDs (8). Other mechanisms of resistance to APPsand CHDs have been proposed, for example, rapid herbicidedetoxification (reviewed in ref. 1). In this study, we identifyACCase mutations in nine populations of hexaploid wild oat A.sterilis ssp. ludoviciana, a major weed in winter crops in NorthAmerica and the northern grain-growing region of Australia. Weshow that each of these single-site mutations affects herbicidesensitivity of the susceptible ACCase. A simple PCR-based testmakes it possible to determine herbicide sensitivity of A. sterilisweeds rapidly.

ResultsFour amino acid substitutions in the CT domain of ACCase fromherbicide-resistant A. sterilis ssp. ludoviciana plants were iden-

Author contributions: W.L. and D.K.H. contributed equally to this work; W.L., D.K.H., D.C.,P.G., S.W.A., and R.R.W. designed research; W.L., D.K.H., D.C., C.C.O., and S.W.A. performedresearch; W.L. contributed new reagents/analytic tools; W.L., D.K.H., D.C., P.G., S.W.A., R.H.,and R.R.W. analyzed data; and W.L., D.K.H., P.G., S.W.A., and R.H. wrote the paper.

The authors declare no conflict of interest.

Freely available online through the PNAS open access option.

Abbreviations: APP, aryloxyphenoxypropionate; CHD, cyclohexanedione; ACCase, acetyl-CoA carboxylase; CT, carboxyltransferase.

†To whom correspondence may be addressed. E-mail: [email protected] [email protected].

This article contains supporting information online at www.pnas.org/cgi/content/full/0611572104/DC1.

© 2007 by The National Academy of Sciences of the USA

www.pnas.org�cgi�doi�10.1073�pnas.0611572104 PNAS � February 27, 2007 � vol. 104 � no. 9 � 3627–3632

PLA

NT

BIO

LOG

YSU

STA

INA

BILI

TYSC

IEN

CE

Dow

nloa

ded

by g

uest

on

Apr

il 7,

202

1

http://www.pnas.org/cgi/content/full/0611572104/DC1http://www.pnas.org/cgi/content/full/0611572104/DC1

tified by sequencing genomic DNA and cDNA: Trp-1,999-Cys(TGG to TGT) in the Shk population; Trp-2,027-Cys (TGG toTGT) in the Nx99 population; Ile-2,041-Asn (ATT to AAT) inthe UQT population; and Asp-2,078-Gly (GAT to GGT) in theUQM population (Fig. 1 and Table 1). We confirmed that theseamino acid changes are sufficient to alter wild oat ACCasesensitivity to herbicides by using yeast gene-replacement strainscontaining wheat ACCase (see below). These changes accountfor the resistant phenotype of plants in the A. sterilis populations.Our findings are consistent with other studies described in theIntroduction, except for the Trp-1,999-Cys substitution, whichhas not been implicated previously in resistance of any species.Genomic DNA and cDNA sequences were consistent, indicatingthat all of the mutant alleles were transcribed, a necessary stepfor the expression of the resistant phenotype. Partial sequencecomparisons with the three homoeologous sequences of thewild-type (susceptible) A. fatua revealed that Trp-1,999-Cys,Ile-2,041-Asn, and Asp-2,078-Gly are located each in the three

ACC homoeologous genes of the hexaploid A. sterilis genome.Trp-2,027-Cys could not be assigned to any particular homoe-olog because of the lack of polymorphism in the sequencedfragment.

Allele-specific PCR tests developed for each of the fivemutations implicated in herbicide resistance (Fig. 2), includingthe Ile-1,781-Leu substitution previously identified in other grassspecies, were used to screen all of the individual plants of theNx99, UQT, UQM, and Shk populations, which survived her-bicide treatment (Table 1). As expected, all 28 plants from theNx99 population that survived the APP treatment contained theCys-2,027 allele; all 30 UQT plants surviving the APP herbicidetreatment, as well as two plants surviving the CHD herbicidetreatment, contained the Asn-2,041 allele; and all 29 UQMplants surviving the APP herbicide treatment, as well as 26 plantssurviving the CHD herbicide treatment, contained the Gly-2,078allele. The control-susceptible biotype did not contain thesealleles. These correlations confirmed our conclusion from theinitial sequencing experiment.

Fig. 1. Amino acid sequence comparisons of the herbicide target site in the CT domain of the plant plastid (pla) and cytosolic (cyt) multidomain ACCase.Mutations associated with resistance are shown with residue numbering following the full-length A. myosuroides plastid ACCase (GenBank accession no.AJ310767) highlighted by dark-gray vertical strips with amino acids found at the corresponding position in yeast shown in the bottom row (underlined withnumbering following the yeast ACCase sequence). Amino acid residues implicated in APP binding (5) are highlighted by light-gray vertical strips. Dots indicateidentical residues. The composite sequences of the plastid and cytosolic ACCase were derived from sequences available from GenBank in December 2006.

Table 1. Herbicide sensitivity and ACCase mutations in individual resistant plants of nine populations of A. sterilis ssp. ludovicianasuspected of carrying herbicide resistance

Population

Number of plants

APP* CHD

Fenoxaprop Clodinafop Haloxyfop Sethoxydim Tralkoxydim

R S R S R S R R R S Mutation in resistant plants

Shk 28 2 W-1,999-Cb,c

11 D-2,078-Gc

1 I-2,041-Nc

8 none detectedNx99 28 22 W-2,027-Cb,c

UQT 30 0 2 28 I-2,041-Nb,c

UQM 29 1 26 4 D-2,078-Gb,c

McdNl 20 0 14 6 Leu-1,781c

McdLed 20 0 14 6 D-2,078-Gc

Crooble 20 0 4 16 D-2,078-Gc

Nx03 20 0 10 10 W-2,027-Cc

Ew 1 14 1 10 W-2,027-Cc

1 5 none detected

R, resistant plants; S, susceptible plants.*Applied in form of esters.†Detected by DNA sequencing.‡Detected by allele-specific PCR.

3628 � www.pnas.org�cgi�doi�10.1073�pnas.0611572104 Liu et al.

Dow

nloa

ded

by g

uest

on

Apr

il 7,

202

1

In the Shk population, among the 48 individuals surviving thefenoxaprop treatments, 28 contained the Cys-1,999 allele, 11contained the Gly-2,078 allele, and one contained the Asn-2,041allele (Table 1). There were eight individuals in the Shk popu-lation that survived fenoxaprop treatment but carried none ofthe five mutations.

The allele-specific PCR test was then used to screen fiveadditional resistant populations (Table 1). For populationMcdNl, all 20 clodinafop-resistant and all 14 tralkoxydim-resistant plants contained the Leu-1,781 allele. For populationMcdLed, all 20 haloxyfop- and all 14 tralkoxydim-resistant plantscontained the Gly-2,078 allele. For population Crooble, all 20fenoxaprop- and all four sethoxydim-resistant plants containedthe Gly-2,078 allele. For population Nx03, all 20 clodinafop-plants and all 10 tralkoxydim-resistant plants contained theCys-2,027 allele. For population Ew, only one of the twoclodinafop- and one of the six tralkoxydim-resistant plantscontained the Cys-2,027 allele. None of the mutations assessedin this study was detected in the remaining six ACCase-inhibitor-resistant EW plants.

Two hundred seventy-nine plants from nine populations eachcarried only one of the five ACCase mutations associated withresistance. Plants from population Shk carrying mutant allelesCys-1,999 and Gly-2,078 and plants from population Nx99carrying the Cys-2,027 allele survived fenoxaprop treatment atdoses 2–16 times the recommended rate of application. On thecontrary, the Asn-2,041 allele was never found in plants thatsurvived the treatment at higher doses, suggesting a lower levelof resistance for this allele.

Four yeast gene-replacement strains depending for growth ona chimeric wheat ACCase, each carrying a single mutationassociated with herbicide resistance, were tested. The chimericACCase consisted of the N-terminal half of the wheat cytosolicACCase and C-terminal half of the wheat plastid ACCase,including the entire herbicide-binding domain shown in Fig. 1.We showed previously that such chimeric ACCase complementsthe yeast ACC1 null mutation, and that growth inhibition of theresulting yeast gene-replacement strain reflects sensitivity of thewheat plastid ACCase to inhibitors (3, 8). We further showedthat a single amino acid change corresponding to the Ile-1,781-Leu mutation makes the yeast strain resistant to CHDs (sethoxy-dim) and APPs (haloxyfop) (8). Wild-type yeast is resistant toboth APPs and CHDs (3, 15). Yeast with chimeric ACCasescarrying Trp-2,027-Cys, Ile-2,041-Asn, and Asp-2,078-Gly mu-tations grow as well as the strain with the wild-type residues. Astrain with the Trp-1,999-Cys mutation grows significantlyslower (2-fold longer doubling time). Chimeric ACCase with theGly-2,096-Ala mutation did not complement the yeast ACC1 nullmutation, presumably due to lack of enzymatic activity sufficientto sustain yeast growth.

The Trp-1,999-Cys mutation renders the yeast gene-replacement strain resistant to fenoxaprop only but has no

effect on sensitivity to haloxyfop and sethoxydim (Fig. 3) andclodinafop (data not shown). This mutation was found in someplants of the Shk population resistant to fenoxaprop but not inpopulations resistant to other APPs or CHDs (Table 1). TheTrp-2,027-Cys, Ile-2,041-Asn, and Asp-2,078-Gly mutationsrender the strains resistant to fenoxaprop, haloxyfop, andclodinafop (Fig. 3 and data not shown). The effect of the

Fig. 2. Allele-specific PCR tests for ACCase mutations in herbicide-resistantpopulations of A. sterilis ssp. ludoviciana. L, 1-kb DNA marker; c, no DNAcontrol; S, template DNA from a susceptible plant; R, template DNA from aresistant plant from each of the populations.

Fig. 3. Response of yeast gene-replacement strains that depends for growthon chimeric wheat ACCase carrying single-site mutations to fenoxaprop-P-ethyl, haloxyfop, and sethoxydim. The control (sensitive) strain (w-t) withwild-type chimeric wheat ACCase was described previously (4).

Liu et al. PNAS � February 27, 2007 � vol. 104 � no. 9 � 3629

PLA

NT

BIO

LOG

YSU

STA

INA

BILI

TYSC

IEN

CE

Dow

nloa

ded

by g

uest

on

Apr

il 7,

202

1

Trp-2,027-Cys mutation on sensitivity to haloxyfop and clodi-nafop is not as strong as that of the other two mutations andnot as strong as its effect on sensitivity to fenoxaprop. TheAsp-2,078-Gly mutation causes complete resistance to se-thoxydim, but only partial sethoxydim resistance was observedfor Trp-2,027-Cys and Ile-2,041-Asn (Fig. 3). In the latter case,the partial resistance was observed only in some experiments,suggesting that the effect of these mutations is rather smallleading to the variable result, depending, for example, on theyeast growth conditions. We have observed such variability inother experiments using the yeast gene-replacement system.Most plants of the UQM, McdLed, and Crooble populationsare resistant to both APPs and CHDs, and they carry theAsp-2,078-Gly mutation, whereas only a smaller number ofplants from populations UQT and Nx03, which are consistentlyresistant to APPs and carry the Ile-2,041-Asn and Trp-2,027-Cys mutation, respectively, are resistant to CHDs. Thus, theresults of the yeast experiments are fully consistent with theresults of the whole plant phenotypic assays.

Haloxyfop-methyl ester and clodinafop-propargyl ester werecompared with haloxyfop and clodinafop in the yeast system(data not shown). We found no significant difference betweeninhibitory properties of APP free acids and their esters, sug-gesting that yeast can hydrolyze the esters efficiently to providethe APPs in the active form as free acids.

DiscussionWild oats (A. fatua and A. sterilis) occur throughout all ofAustralia’s cereal-growing regions. A. sterilis ssp. ludoviciana isthe predominant weed in the northern grain region of Australia,particularly prevalent in cropping systems with winter cereals,primarily wheat, and winter pulse rotation. It has been success-fully controlled by postemergence application of APPs andCHDs. The excellent efficacy of these herbicides, which do notaffect broadleaf crops (insensitive ACCase), as well as somecereals, such as wheat (rapid herbicide detoxification), on manygrass weed species encouraged their widespread and repeateduse. Thirty-five resistant species have been reported in 26countries (the first case in 1982) with increasing numbers of sitesand areas of infestation, and nine resistant species in Australia,including the first case of resistance in wild oat in WesternAustralia in 1985. Twelve randomly collected A. sterilis ssp.ludoviciana samples from in-crop sites and fallow paddocks in thenorthern grain region of Australia in the winter of 2002 weretested in this study for herbicide resistance. All were sensitive,suggesting that the herbicide resistance is not yet widespread.However, plants from nine populations from the same regionsupplied by farmers as suspected to be resistant to ACCaseherbicides, based on poor weed control in the field, all showeda high level of resistance in subsequent pot tests, either to APPsor CHDs or both (Table 1). The herbicide-use history for thepopulations is incomplete, but the available data indicate at least2 years of herbicide application at all nine sites. For example,population Nx03 was subjected to four APP treatments [fenoxa-prop (1996), clodinafop (1997), haloxyfop (1998), and haloxyfop(2001)], followed by two CHD treatments [sethoxydim (2002)and clethodim (2003)].

Using a combination of DNA sequencing and allele-specificPCR, we identified five amino acid substitutions in the plastidACCase gene of nine A. sterilis ssp. ludoviciana populationsresistant to ACCase-inhibiting herbicides (Table 1 and Fig. 1).All of these mutations cluster in the CT domain, whichcontains the binding site of APP and CHD herbicides (3, 4, 8).These mutated amino acid residues are not found at corre-sponding ACCase positions in any susceptible grass species.Furthermore, four of these mutations, Ile-1,781-Leu, Trp-2,027-Cys, Ile-2,041-Asn, and Asp-2,078-Gly, have been re-ported in other ACCase-inhibitor resistant grass weed species

(5–14), mostly in A. myosuroides. The Trp-1,999-Cys mutationhas not been previously described in any herbicide-susceptiblegrass. We have also identified the Gly-2,078 allele in resistantL. rigidum (unpublished data), which further supports involve-ment of this substitution in herbicide resistance in differentplants. The presence of any one of these mutations in plastidACCase is correlated with plant resistance to the herbicides,suggesting strongly that the lost sensitivity of the enzyme to theinhibitors is responsible for the resistant phenotype of theplant.

The 400-aa domain shown in Fig. 1 contains all of theresidues shown to be present in the APP-binding pocket as wellas residues of the CoA-binding site (4, 16). In the 3D structureof the CT domain from yeast ACCase, the two sites are veryclose to each other but do not overlap. The six herbicide-resistant mutations are located in this domain: I-1,781-L (yeastL-1,705), W-1,999-C (yeast W-1,924), I-2,041- (yeast V-1,967),D-2,078-G (yeast D-2,004) in the immediate vicinity of boundinhibitor, and W-2,027-C (yeast W-1,953) and G-2,096-A(yeast A-2,022), a short distance away (Fig. 4). The multido-main plastid ACCase gene in grasses originated by duplicationof the cytosolic gene early during evolution of the grass/monocot lineage (17), a relationship ref lected in their highsequence similarity (Fig. 1). All but one of the residuesinvolved in APP binding and resistance are conserved in bothACCase isozymes in grasses as well as all cytosolic ACCasesfrom dicot plants, the next closest group of relatives of thegrass ACCases (not shown). I-1,781-L is the only amino aciddifference consistent with the resistance of all plant cytosolicACCases to APPs and CHDs. However, pea and corn cytosolicACCase are somewhat sensitive to APPs (18). Furthermore,the Toxoplasma gondii apicoplast ACCase, which is sensitive toAPPs, although less than the plastid ACCase from grasses (8,19, 20), has all of the critical residues identified in Fig. 1, exceptfor L-1,781 and S-2,073, but changing the Leu to Ile in thiscontext reduces sensitivity to APPs.

Consistent with previous studies on other grassy weeds,including L. rigidum, A. myosuroides, S. viridis, and A. fatua(5–12), the Leu-1,781 allele was found in the McdLed popu-lation resistant to both APP and CHD herbicides. Alsoconsistent with the study on A. myosuroides by Delye et al. (14),the Gly-2,078 allele was detected in UQM, McdLed, andCrooble populations resistant to both APPs and CHDs (Table1). The Asn-2,041 allele was found in UQT populationsresistant to APP but sensitive to CHD herbicides, which againconcurs with previous studies on A. myosuroides and L. rigidum(13). However, two plants from this population carrying theAsn-2,041 allele survived sethoxydim treatment, which isunexpected. These plants might carry the Gly-2,096-Ala sub-stitution, which was not assayed in this study but was found to

Fig. 4. Position of amino acid residues corresponding to the six herbicide-resistance mutations (green) in the 3D structure of the CT domain of yeastACCase in complex with CoA (blue) and haloxyfop (red) (5, 17). Amino acidsshown in yellow were determined to contribute to the haloxyfop-binding site(5). This illustration was prepared by using PyMol (DeLano Scientific, South SanFrancisco, CA) and coordinates from Protein Data Bank ID codes 1UYS and1OD2. The numbering follows the yeast ACCase sequence.

3630 � www.pnas.org�cgi�doi�10.1073�pnas.0611572104 Liu et al.

Dow

nloa

ded

by g

uest

on

Apr

il 7,

202

1

confer resistance to APPs in A. myosuroides (14) or a previ-ously undescribed resistance allele in ACCase. The Gly-2,096-Ala substitution, which was not found in any of the A. sterilisssp. ludoviciana populations investigated in the initial se-quence-based analysis, could explain the eight Shk and one Ewplants resistant to APPs (and not containing any of the otherknown resistance alleles), but it does not explain the five EWand two UQT plants resistant to CHDs. Other resistancemechanisms such as enhanced metabolism could also be atplay, as previously proposed (reviewed in ref. 1). Ten plantsfrom the Nx03 population, which were resistant to the CHD,tralkoxydim, contained the Cys-2,027 allele, which was re-ported to confer resistance to APPs only (14). However, thisparticular CHD has not been studied previously in relation tothe Cys-2,027 allele.

Sequence comparisons of the Cys-1,999, Cys-2,027, Asn-2,041,and Gly-2,078 alleles with the three ACC1 homoelogs of A. fatuademonstrated that any of the three homoelogs in hexaploid A.sterilis ssp. ludoviciana can harbor the herbicide-resistance mu-tations. Therefore, the chance of hexaploid wild oats accumu-lating a mutated allele is greater than that of diploid grass weeds.Crossing plants containing the different resistance alleles toproduce progeny containing resistance mutations on each of thehomoeologous chromosomes could prove useful to investigationof the contribution of gene dosage to differing resistance levelsin polyploid weed species. Interestingly, one population (Shk) inthis study contained at least three different resistance mutations,but no individual plant contained more than one mutation. It ispossible that a fitness penalty may be associated with multiplemutations.

Our experiments using yeast gene-replacement strains con-firm the results from the phenotypic and sequence analysis:single amino acid changes Ile-1,781-Leu and Asp-2,078-Glyconfer strong resistance to both APPs and CHDs; Trp-2,027-Cys and Ile-2,041-Asn confer strong resistance to APPs (andmild resistance to CHDs), and Trp-1,999-Cys confers strongresistance to only one of the tested APPs, fenoxaprop. Theseamino acid substitutions are correlated with resistance inplants such as A. sterilis, A. myosuroides, and L. rigidum but alsowork in the structural context of the wheat plastid ACCase,suggesting that such single-amino acid mutant alleles can easilybe selected in other grassy weeds subjected to herbicidetreatment.

The slow growth of the gene-replacement strain with theTrp-1,999-Cys ACCase mutant suggests this mutation has asignificant effect on the enzymatic activity. Such a decreasedactivity could impose a significant fitness penalty on a plantcarrying the mutant allele affecting its persistence in wildpopulations once the selective pressure is removed. The effect onfitness of hexaploid plants, such as wild oats, could be less severe.Decreased enzymatic activity was suggested for ACCases withthe Trp-2,027-Cys and Asp-2,078-Gly mutations but not theIle-2,041-Asn and G-2,096-A mutations (13, 14). In this context,lack of complementation by ACCase with the G-2,096-A muta-tion must be due to a specific problem in the structural contextof wheat plastid ACCase. Wild-type yeast ACCase has an Alaresidue at the corresponding position, and therefore this residueby itself is not deleterious.

Using the allele-specific PCR tests developed in this study, fiveresistant populations were quickly confirmed to contain muta-tions in ACCase, in addition to the four populations confirmedby more extensive DNA sequencing. Molecular-based diagnosticmethods have the advantage of quickly detecting resistant plantscompared with conducting pot assays, which are time- andspace-consuming and labor-intensive. Reliable, fast, simple de-tection of herbicide resistance is necessary for farmers to adopttimely alternative herbicide application strategies to preventfurther spread of herbicide-resistant weeds. However, only the

phenotypic assays can identify individuals with resistance mech-anisms that have not yet been characterized at the molecularlevel.

Materials and MethodsPlant Material and Herbicide Treatment. Herbicide resistance ofnine A. sterilis ssp. ludoviciana populations (Table 1) from thenorthern grain region of Australia (Southern Queensland andNorthern New South Wales) was confirmed by pot assay byusing the same herbicide(s) that failed in the field. Seedlingsat the three- to four-leaf stage (five per pot) were sprayed withherbicides in a wind tunnel using a traversing boom sprayoperated at a height of 70 cm above the plants, with f lat fannozzles and a spray pressure of 0.3 MPa. Herbicides wereapplied at the manufacturer’s recommended field applicationrate: 0.3 liter�ha�1 fenoxaprop-P-ethyl [110 g active ingredient(a.i.) liter�1]; 0.125 liter�ha�1 clodinafop-propargyl (240 g ofa.i. liter�1); 0.3 liter�ha�1 haloxyfop (130 g of a.i. liter�1); 1.0liter�ha�1 sethoxydim (186 g of a.i. liter�1), and 500 g�ha�1tralkoxydim (400 g of a.i. kg�1). APP treatments includednonionic surfactant Agral (0.25% vol/vol), and CHD treat-ments included 1.0 liter�ha�1 adjuvant oil D-C-TRATE, (763g�liter�1). Shk and Nx99 populations were also treated withdoses 2, 4, 8, and 16 times the recommended rate. Plants weremonitored daily and assessed visually at 30 days after treat-ment to be resistant (alive) or susceptible (dead). Twelverandom A. sterilis ssp. ludoviciana populations from the sameregion of Australia tested under the same regime were herbi-cide-sensitive.

Sequencing of the Plastid ACCase CT Domain and Allele-Specific PCR.Total DNA and RNA were extracted from leaf tissue from asingle plant. RNA extracted using TRIzol method (Life Tech-nologies, Gaithersburg, MD) and treated with DNase (Pro-mega, Madison, WI) was used as a template to synthesizesingle-stranded cDNA with a polydT primer according to themanufacturer’s instructions (Qiagen, Hilden, Germany). Uni-versal PCR primers [supporting information (SI) Table 2] foramplification of all homoeologous DNA fragments encodingpart of the CT domain of the plastid ACCase of A. sterilis ssp.ludoviciana were designed based on DNA sequences of A. fatua(GenBank accession nos. AF231334, AF231335, AF231336,AF231337, AF464875, and AF464876) and L. rigidum (Gen-Bank accession nos. AF359513, AF359514, AF359515, andAF359516). Allele-specific primers used for PCR detection ofmutants are shown in SI Table 2. The PCR primers were alsoused for sequencing (see SI Materials and Methods). Theallele-specific PCR test was used to screen all of the resistantA. sterilis ssp. ludoviciana plants identified in this study for thepresence or absence of the respective mutations. Amino acidnumbering for A. sterilis follows the full-length A. myosuroidesplastid ACCase (GenBank accession no. AJ310767).

Herbicides. Sethoxydim (Bayer Cropscience Australia, EastHawthorn, Victoria, Australia; Crescent Chemicals, Haup-pauge, NY; or Sigma-Aldrich, Castle Hill NSW, Australia),tralkoxydim (Crop Care Australasia, Murrarie, Queensland,Australia), fenoxaprop-P-ethyl (Bayer Cropscience Australiaor Sigma-Aldrich), clodinafop-propargyl plus clodinafop-cloquintocet-methyl (Syngenta Crop Protection, North RydeNSW, Australia), haloxyfop-R-methyl ester (Nufarm Austra-lia, Laverton, Victoria, Australia), haloxyfop (Crescent Chem-icals, or Sigma-Aldrich), clodinafop (Syngenta, Research Tri-angle Park, NC), agral (Syngenta Crop Protection, North RydeNSW, Australia), D-C-TRATE (Caltex Australia Petroleum,Sydney NSW, Australia).

Liu et al. PNAS � February 27, 2007 � vol. 104 � no. 9 � 3631

PLA

NT

BIO

LOG

YSU

STA

INA

BILI

TYSC

IEN

CE

Dow

nloa

ded

by g

uest

on

Apr

il 7,

202

1

http://www.pnas.org/cgi/content/full/0611572104/DC1http://www.pnas.org/cgi/content/full/0611572104/DC1http://www.pnas.org/cgi/content/full/0611572104/DC1

Yeast Gene-Replacement Strains. Single amino acid substitutions inthe CT domain in wheat plastid ACCase were introduced byreplacing a 1.6-kb BlpI-ApaI fragment in construct C50P50 invector pRS423 (4) with the corresponding mutated fragmentprepared by two-step PCR procedure. Structures of the newconstructs was verified by DNA sequencing. Saccharomycescerevisiae strain W303D-ACC1�Leu-2 (ACC1/acc1::LEU2) usedfor complementation was provided by S. Kohlwein (TechnicalUniversity, Graz, Austria). Yeast transformation, sporulation,tetrad analysis, marker- and galactose-dependence tests, yeast

growth, and inhibition measurements were carried out as de-scribed (3, 15). Yeast growth experiments were carried out byusing either 3-ml cultures in tubes or 0.2-ml cultures in 96-wellculture plates.

We thank H. Y. Lee (University of Chicago, Chicago, IL) for technicalassistance. We also thank the Centre for Pesticide Application andSafety, University of Queensland, Gatton, for providing assistance withherbicide application and access to the wind tunnel facility. This researchwas supported by the Australian Grains Research and DevelopmentCorporation and the University of Queensland.

1. Delye C (2005) Weed Sci 53:728–746.2. Podkowinski J, Jelenska J, Sirikhachornkit A, Zuther E, Haselkorn R, Gor-

nicki P (2003) Plant Physiol 131:763–772.3. Nikolskaya T, Zagnitko O, Tevzadze G, Haselkorn R, Gornicki P (1999) Proc

Natl Acad Sci USA 96:14647–14651.4. Zhang H, Tweel B, Tong L (2004) Proc Natl Acad Sci USA 101:5910–5915.5. Brown AC, Moss SR, Wilson ZA, Field LM (2002) Pest Biochem Physiol

72:160–168.6. Delye C, Calmes E, Matejicek A (2002) Theor Appl Genet 104:1114–1120.7. Delye C, Matejicek A, Gasquez J (2002) Pest Manag Sci 58:474–478.8. Zagnitko O, Jelenska J, Tevzadze G, Haselkorn R, Gornicki P (2001) Proc Natl

Acad Sci USA 98:6617–6622.9. Zhang XQ, Powles SB (2006) Planta 223:550–557.

10. Delye C, Wang TY, Darmency H (2002) Planta 214:421–427.11. Christoffers MJ, Berg ML, Messersmith CG (2002) Genome 45:1049–1056.12. White GM, Moss SR, Karp A (2005) Weed Res 45:440–448.

13. Delye C, Zhang XQ, Chalopin C, Michel S, Powles SB (2003) Plant Physiol132:1716–1723.

14. Delye C, Zhang XQ, Michel S, Matejicek A, Powles SB (2005) Plant Physiol137:794–806.

15. Joachimiak M, Tevzadze G, Podkowinski J, Haselkorn R, Gornicki P (1997)Proc Natl Acad Sci USA 94:9990–9995.

16. Tong L, Harwood HJ Jr (2006) J Cell Biochem 99:1476–1488.17. Huang SX, Sirikhachornkit A, Faris JD, Su XJ, Gill BS, Haselkorn R, Gornicki

P (2002) Plant Mol Biol 48:805–820.18. Herbert D, Price LJ, Alban C, Dehaye L, Job D, Cole DJ, Pallett KE, Harwood

JL (1996) Biochem J 318:997–1006.19. Jelenska J, Sirikhachornkit A, Haselkorn R, Gornicki P (2002) J Biol Chem

277:23208–23215.20. Zuther E, Johnson JJ, Haselkorn R, McLeod R, Gornicki P (1999) Proc Natl

Acad Sci USA 96:13387–13392.

3632 � www.pnas.org�cgi�doi�10.1073�pnas.0611572104 Liu et al.

Dow

nloa

ded

by g

uest

on

Apr

il 7,

202

1