www.newphytologist.org

787

Research

Blackwell Publishing, Ltd.

Resistance cost of a cytochrome P450 herbicide metabolism mechanism but not an ACCase target site

mutation in a multiple resistant

Lolium rigidum

population

M. M. Vila-Aiub, P. Neve and S. B. Powles

Western Australian Herbicide Resistance Initiative (WAHRI), School of Plant Biology, Faculty of Natural & Agricultural Sciences, The University of Western

Australia (UWA), 35 Stirling Highway, Crawley, WA 6009, Australia

Summary

• Costs of resistance are predicted to reduce plant productivity in herbicide-resistantweeds.•

Lolium rigidum

herbicide-susceptible individuals (S), individuals possessingcytochrome P450-based herbicide metabolism (P450) and multiple resistant individualspossessing a resistant ACCase and enhanced cytochrome P450 metabolism (ACCase/P450) were grown in the absence of mutual plant interaction to estimate plantgrowth traits.• Both P450 and ACCase/P450 resistant phenotypes produced less above-groundbiomass than the S phenotype during the vegetative stage. Reduced biomassproduction in the resistant phenotypes corresponded to a reduced relative growthrate and a lower net assimilation rate and rate of carbon fixation. There were no signifi-cant differences between the two resistant phenotypes, suggesting that costsof resistance are associated with P450 metabolism-based resistance. There were nodifferences in reproductive output among the three phenotypes, indicating thatthe cost of P450 resistance during vegetative growth is compensated during theproduction of reproductive structures.• The P450-based herbicide metabolism is shown to be associated with physiologicalresistance costs, which may be manipulated by agronomic management to reducethe evolution of herbicide resistance.

Key words:

ACCase target site mutation, dry matter acquisition/ allocation,evolution of resistance, fitness cost, growth analysis,

Lolium rigidum

, P450 herbicidemetabolism.

New Phytologist

(2005)

167

: 787–796

© New Phytologist

(2005)

doi

: 10.1111/j.1469-8137.2005.01465.x

Author for correspondence:

Martin Vila-Aiub Tel:

+

61 8 64887980Fax:

+

61 8 64887834 Email: mvilaaiu@plants.uwa.edu.au

Received:

18 January 2005

Accepted:

4 April 2005

Introduction

The evolution of herbicide resistance in weeds is a worldwidephenomenon in modern cropping systems (reviewed inLeBaron & Gressel, 1982; Powles & Holtum, 1994; Powles& Shaner, 2001). This evolutionary process involves theselection and enrichment of adaptive mutant alleles thatendow resistance. If a resistance allele or allelles compromisesplant function, then the mutant genotype may suffer a relativedisadvantage or reduction in one or more components of

plant fitness (termed physiological or direct cost of resistance)compared with the wild type in the absence of herbicideselection (Fisher, 1928; Strauss

et al

., 2002). A number ofexplanations are considered to account for the existence ofphysiological herbicide resistance costs (Strauss

et al

., 2002):(1) a mutation associated with target site resistance may insome way compromise normal functioning of the target siteenzyme, resulting in pleiotropic changes in cellular metabolism(Gronwald, 1994; Purrington & Bergelson, 1999; Berticat

et al

., 2002); (2) where the resistance mechanism is nontarget

New Phytologist

(2005)

167

: 787–796

www.newphytologist.org

©

New Phytologist

(2005)

Research788

site based then the resistance may be associated with diversionof resources to increased or novel production of detoxificationenzymes (Coley

et al

., 1985; Herms & Mattson, 1992); or (3)resistance loci may be tightly linked and cosegregate with otherloci affecting plant growth attributes (Bergelson & Purrington,1996; Purrington & Bergelson, 1997).

The existence, magnitude and nature of any resistance costdepends on the mechanism(s) and genetic basis of resistance(Coustau

et al

., 2000; Roux

et al

., 2004). For target site-basedherbicide resistance a physiological cost is well documentedfor triazine herbicide resistance. In this case, a specific pointmutation in the

psbA

gene that encodes for a Ser (264) Glyamino acid substitution in the D1 protein confers triazineresistance in many weed species and this mutation incurs aresistance cost by reducing photosynthetsis (see explanation1) and hence compromising vegetative and reproductive plantgrowth (reviewed by Gronwald, 1994). However, for othertarget site-based mechanisms endowing herbicide resistancethe results on the expression of resistance costs are less clear.For example, there is a diversity of target site gene mutationsendowing resistance to the acetyl CoA carboxylase (ACCase)and acetolactate synthase (ALS) inhibiting herbicides, eachhaving the potential to endow different resistance costs,meaning that studies on costs of herbicide resistance mustbe assessed for each specific mutation.

Resistance costs that could be associated with nontarget sitebased herbicide resistance have been little studied. Evolvedherbicide resistant populations of

Lolium rigidum

and

Aloper-cus myosuroides

can have a nontarget site based resistancemechanism of enhanced rates of herbicide metabolism owingto elevated activity of cytochrome P450 enzymes (Hall

et al

.,1995; Preston

et al

., 1996). It is not known whether there isa physiological resistance cost associated with this nontargetsite-based P450-catalysed herbicide metabolism mechanism.Increased or novel production of detoxifying enzymes (non-target site resistance) may incur resistance costs if there is anincreased allocation of resources to the production of theseenzymes (explanation 2) (Herms & Mattson, 1992). In thiscase, the expression and magnitude of these costs will dependon whether enzyme production is constitutive or inducedby herbicide application (Karban

et al

., 1997; Strauss

et al

.,2002).

Individual

L. rigidum

plants can display multiple resistanceto ACCase and ALS inhibiting herbicides owing to bothtarget site- and nontarget site-based resistance mechanisms(Tardif

et al

., 1993; Tardif & Powles, 1994; Preston

et al

., 1996;Preston & Mallory-Smith, 2001). This phenomenon of mul-tiple resistance provides an excellent opportunity to compareresistance costs associated with different herbicide resistancemechanisms in individuals within a single resistant population(similar genetic background). In a well characterized multipleresistant

L. rigidum

population (SLR31), the majority of indi-viduals are resistant to aryloxyphenoxypropionate (APP) andsulfonylurea (SU) herbicides because of enhanced herbicide

metabolism mediated by cytochrome P450 monooxygenaseenzymes (Christopher

et al

., 1991; Christopher

et al

., 1994;Tardif & Powles, 1994; Preston & Powles, 1998). In addition,15% of individuals possess a resistant ACCase enzyme (target-site mechanism) (Tardif & Powles, 1994) conferred by an Ileto Leu substitution at amino acid 1781 of the ACCase gene(X. Q. Zhang & S. B. Powles, unpublished). This point muta-tion endows resistance to APP and cyclohexanedione (CHD)herbicides (Tardif & Powles, 1994). Individuals with theresistant ACCase also have enhanced P450-mediated herbicidemetabolism and are therefore multiple resistant individuals(Vila-Aiub

et al

., 2005). A small proportion (6%) of thepopulation is susceptible to APP, CHD, and SU herbicides. Aplant cloning technique (Vila-Aiub

et al

., 2005) has permittedthe identification and isolation of these three distinctphenotypes from within the SLR31 population, herein termedherbicide susceptible individuals (S), enhanced P450metabolism-based resistant individuals (P450) and multipleherbicide resistant individuals (ACCase/P450). Researchpresented here aims to determine direct resistance costs byassessing plant performance (i.e. resource acquisition andallocation patterns at the vegetative and reproductivestage) of these herbicide resistance phenotypes (S, P450 andACCase/P450).

Materials and Methods

Plant material

A multiple herbicide-resistant

Lolium rigidum

Gaudin popula-tion (SLR31), originating from Bordertown, South Australia(140

°

46

′

E, 36

°

18

′

S) was used in all experiments. The P450and ACCase/P450 phenotypes comprising SLR31

L. rigidum

population exhibit differential responses to the ACCaseinhibiting APP and CHD herbicides. The insensitive ACCaseconfers resistance to certain CHD herbicides (e.g. sethoxydim),while those individuals in the population possessing onlyenhanced herbicide metabolism (P450) are susceptible to theseCHD herbicides, but resistant to diclofop-methyl (Tardif& Powles, 1994). Based on this difference, susceptible (S), andP450 and ACCase/P450-based resistant individuals could beidentified within the SLR31 population (Vila-Aiub

et al

.,2005). A large number of individuals were asexually cloned,and the parent plants and their respective clones were numberedto ensure subsequent identification. All individuals weresprayed with diclofop-methyl and dead individuals wereclassified as the herbicide susceptible (S) phenotype within theSLR31 population. Individuals resistant to diclofop-methylwere sprayed with sethoxydim. Diclofop-methyl resistant plantsthat were susceptible to sethoxydim were classified as the P450phenotype. The remaining fraction of the SLR31 populationwhich survived exposure to diclofop-methyl and then tosethoxydim, was classified as the ACCase/P450 phenotype(Vila-Aiub

et al

., 2005). Seeds of each phenotype from plants

©

New Phytologist

(2005)

www.newphytologist.org

New Phytologist

(2005)

167

: 787–796

Research 789

grown under identical conditions were harvested in December2001 and stored at constant 20

°

C (Vila-Aiub

et al

., 2005)before use in experiments.

Growth analysis and biomass allocation during the vegetative phase

In 2002, seeds of the individual S, P450 and ACCase/P450phenotypes were germinated on 0.7% (w : v) agar (12 h lightat 25

°

C, 12 h dark at 15

°

C). After 4 d, individual seedlings(2 cm high) were transplanted into pots (9 cm diameter,13 cm high) filled with an equal weight of washed white sandand supplied with the following nutrient solution: 90 mg kg

−

1

KH

2

PO

4

, 140 mg kg

−

1

K

2

SO

4

, 150 mg kg

−

1

CaCl

2

, 40 mg kg

−

1

MgSO

4

, 10 mg kg

−

1

MnSO

4

, 9 mg kg

−

1

ZnSO

4

, 2 mg kg

−

1

CuSO

4

, 0.7 mg kg

−

1

H

3

BO

3

, 0.4 mg kg

−

1

CoSO

4

, 0.2 mg kg

−

1

Na

2

MoO

4

. The nutrient solution was supplied at half strengthat the time of seeding and 1 wk after seeding. Nitrogen (asNH

4

NO

3

) was applied (50 mg kg

−

1

) at weekly intervals overthe course of the experiment. The field capacity of the sandsubstrate (12%) was measured before commencement ofexperiments and pots were watered to field capacity on a dailybasis. Plants were grown in a controlled environment facility(phytotron) under natural solar radiation at 20

°

C/15

°

C (day/night) and arranged in a completely randomized design. Potswere regularly rearranged to randomize any environmentaldifferences within the phytotron. There were 18–24 replicatesper treatment (three phenotypes

×

two harvest times (29 dand 57 d after seeding)). The experiment was repeated using25–28 replicates per treatment in 2003.

For both experiments, shoots and roots were harvested29 d and 57 d after sowing (DAS) (shoots were divided intoleaf material and stems including the leaf sheath). Leaf area perplant was determined with a digital leaf area meter (LI-3100;LiCor, Lincoln, Nebraska, USA). Above-ground material androots were oven dried at 80

°

C for 72 h, and dry biomassrecorded. Relative growth rate (RGR) and its components (net

assimilation rate, NAR; specific leaf area, SLA; leaf weightfraction, LWR) were calculated for each treatment combina-tion (phenotype

×

harvest). A software program developed byHunt

et al

. (2002) was used to calculate growth parameters,which are derived according to classical growth analysis:

Eqn 1

(RGR)

=

(NAR) (SLA) (LWR)

(

t

is time;

W

is total dry weight per plant;

L

A

is total leaf areaper plant;

W

L

is total leaf dry weight per plant). Plant growthtraits and derived variables (Hunt, 1982; Poorter & Nagel,2000) estimated after each harvest, are described in Table 1.At the final harvest (57 DAS), the number of vegetative tillersper plant was recorded.

Statistical analysis

The unbiased formula proposed by Hoffmann and Poorter(2002) was used to calculate RGR. The variance (

σ

2

or

V

)associated with RGR was estimated with Causton and Venus’sformula (1981):

Eqn 2

(ln

W

2

is the mean of the ln-transformed plant weight at time2; ln

W

1

is the mean of the ln-transformed plant weight attime 1). The degrees of freedom associated with RGR and itscomponents is

n

−

2, where

n

is the total number of plantsused in two harvests.

One-way analyses of variance (

) were performed tocompare RGR and its components for herbicide susceptibleand resistant

L. rigidum

phenotypes. Tukey’s honestly signifi-cant difference (HSD) test was used to compare mean values(

α

=

5%).

Attribute (dimension) Symbol or formula

Number of vegetative tillers TVRoot biomass (g) WRStem biomass (g) WSTLeaf biomass (g) WLLeaf area (cm2) LATotal biomass (g) W = WS + WR + WLRelative growth rate (RGR) (mg mg−1 d−1) (ln W2 − ln W1)/(t2 − t1)Specific leaf area (SLA) (cm2 mg−1) LA/WLNet assimilation rate (NAR) (mg cm−2 d−1) ((W2 − W1)/(T2 − T1))((ln2W − ln1W)/(2LA − 1LA))Leaf weight ratio (LWR) (mg mg−1) WL/WRoot weight ratio (RWR) (%) WR/WStem weight ratio (SWR) (%) WST/W

Table 1 Measured and estimated vegetative plant growth variables for the S, P450 and ACCase/P450 phenotypes of Lolium rigidum

dW

Wdt

dW

L dt

L

W

W

WA

A

L

L = × ×

V RGRV W V W

t t( )

(ln ) (ln )

( )=

+−

2 1

2 12

New Phytologist (2005) 167: 787–796 www.newphytologist.org © New Phytologist (2005)

Research790

Multivariate one-way analysis of variance () wasconducted to test for the effect of phenotype on correlateddependent growth traits measured on single plants (numberof vegetative tillers, leaf area, total biomass, and root, leaf andstem weight ratios) harvested 57 DAS (2002 and 2003 exper-iments). If multivariate tests were significant at P < 0.05, one-way s were performed individually on each growthtrait. To comply with the assumptions of normal distributionand homoscedasticity, data for number of tillers, leaf area andtotal biomass were log-transformed (y = log10 x) , and root,leaf and stem weight ratios were angular-transformed (y = arc-sin √x) (Sokal & Rohlf, 1969) before and .Where appropriate, means were separated using Tukey’s HSDtest (α = 5%).

Photosynthesis and gas exchange

Leaf CO2 assimilation rates (µmol m−2 s−1) were determinedusing a LiCor 6400 gas exchange apparatus. Photosyntheticresponses were measured at photon flux densities between0 and 1500 µmol m−2 s−1 (6400-02B red-blue LED lightsource). Seven plants per phenotype were randomly selectedbefore the final harvest of the 2003 experiment for thesemeasurements. Photosynthetic gas exchange was measuredon one fully expanded leaf at three 60-s intervals. The gasanalyser was calibrated to provide constant conditions of20°C, 360 µmol CO2 mol−1 (6400-01 CO2 injector system)and 500 µmol s−1 air flow rate. The response of CO2 assimilationrate to photon flux was described using a rectangularhyperbola model (Long & Hällgren, 1993):

A(µmol m−2s−1) = Eqn 3

(A is the leaf CO2 exchange rate at x light intensity; Asat is thelight-saturated rate of CO2 uptake (i.e. maximum photo-synthesis); Kq is the light intensity of photosyntheticallyactive radiation (PAR) at which photosynthesis is half of thelight saturated maximum (Asat)). The nonlinear least squaresregression curves were generated with version6.0 (SPSS Science, Chicago, IL, USA) after transforming data(y = √x + 6). Comparison of model parameters between thephenotypes was performed by and means wereseparated by Tukey’s HSD test (α = 5%) ( version 3.0; GraphPad Software, Inc., San Diego, CA, USA).

Biomass allocation during the reproductive phase

Glasshouse experiment Seeds of the S, P450 and ACCase/P450 phenotypes were germinated on 0.7% agar, as previouslydescribed, and after 3 d seedlings (2-cm high) transplantedinto pots (17-cm diameter, 14-cm high) containing washedwhite sand (after 3 d). Mineral nutrients and N were suppliedas described for the 2002 vegetative growth experiment.Plants were grown in controlled glasshouse conditions (mean

temperatures were between 17 and 24°C) with natural solarradiation. Plants (n = 24–30) of each phenotype were arrangedin a completely randomized design. Daily watering to fieldcapacity was as described for the previous experiment.

Before flowering, cross-pollination between phenotypes wasprevented by grouping plants of the same phenotype withinpollen-proof enclosures. At maturity, seed heads were harvestedfrom individual plants and the number of reproductive tillerswas counted. Total above-ground biomass was determinedafter drying plants at 80°C for 72 h. Seed heads (spikes) werethen threshed and total seed mass was measured. The numberof seeds produced per plant (Sn) was estimated as:

Eqn 4

(TSw is the total seed weight produced per plant; Sw is themean weight of 50 seeds per plant; n = 3). Individual seedweight was determined from the average weight of 50 seeds.Harvest index (HI) (%) per plant was calculated as the ratioof seed mass to total above-ground biomass (seed mass +vegetative biomass).

Field experiment Individual seedlings (2 cm high) of thethree phenotypes (S, P450, ACCase/P450) were transplantedinto large pots (25 cm diameter, 23 cm high) containing astandard potting mix (50% peat moss and 50% river sand).Plants of each phenotype (20 replicates) were grown outdoorsduring the normal growing season for L. rigidum ( July–November) in a completely randomized design. Plants werefertilized with a slow-release fertilizer (Macrocote: Blue Plus;Langley Chemicals, Welshpool, WA, Australia) at a rate of 12 gper pot (w : w nutrient concentrations: N 16.3% (NH2 8.4%,NH4 6.4%, NO3 1.5%); phosphorus (P) 4%; potassium (K)10%; sulphur (S) 5%; magnesium (Mg) 0.63%; iron (Fe)0.20%; copper (Cu) 0.03%; zinc (Zn) 0.03% and manganese(Mn) 0.08%) and with three applications of liquid fertilizer (N19% (NH2 15%, NH4 1.9%, NO3 2.1%); P 8%; K 16%, Mg1.2%; S 3.8%; Fe 0.4; Mn 0.2 g l−1; Zn 0.2 g l−1; Cu 0.1 g l−1;B 0.1 g l−1 and Mo 0.01 g l−1). Plants were watered regularly.Pollen proof enclosures were erected before flowering, asdescribed previously. Plants were harvested as described forthe glasshouse experiment.

Statistical analysis

The effect of phenotype on reproductive traits was assessed bymultivariate one-way analysis of variance (). To complywith assumptions of normal distribution and homoscedasticity,harvest index values were angular transformed (y = arcsin √x)and values for other experimental variables were log-transformed(y = log10 x or y = log10 x + 1) (Sokal & Rohlf, 1969). Meanswere compared using Tukey’s HSD test where appropriate(α = 5%).

A x

K xsat

q +

STS

Snw

w

=× 50

© New Phytologist (2005) www.newphytologist.org New Phytologist (2005) 167: 787–796

Research 791

Results

Growth analysis, biomass allocation and photosynthesis during the vegetative phase

A classical growth analysis was performed on plants of the S,P450 and ACCase/P450 phenotypes grown under nutrientand temperature controlled conditions. The experiment wasconducted in 2002 and repeated in 2003. A showeda significant (P < 0.001) effect of phenotype on dependentgrowth variables estimated 57 DAS for both experiments,accounting for 58% (1 – Wilk’s lambda) and 46% of the totalvariance in 2002 and 2003, respectively.

Herbicide susceptible (S) vs metabolic (P450) resistant phenotype

In the 2002 experiment, the resistant P450 phenotypeproduced significantly less total biomass (at 57 DAS) thanplants of the S phenotype (Table 2). Leaf area and number ofvegetative tillers produced per plant did not differ between thetwo phenotypes (Table 2). The P450-based resistant plantshad a significantly lower mean RGR than S plants during the29–57 DAS interval (Table 3). This differential response wascorrelated with a lower NAR in the P450 phenotype (Table 3).

Differences in resource allocation were observed: P450 resistantplants allocated 5% less resources to roots (RWR) (Fig. 1a)and 5% more to leaves (Fig. 1b). There were no differencesin stem weight ratio and SLA between the two phenotypes(Fig. 1c).

In the 2003 experiment, the resistant P450 phenotype (57DAS) again produced significantly less total dry biomass thanthe S phenotype, and once again there were no differences inleaf area and number of vegetative tillers between phenotypes(Table 2). In common with the 2002 experiment, the P450phenotype exhibited a significantly lower RGR than the Sphenotype (Table 3) which was correlated with a lower NAR(Table 3). Photosynthetic rates (leaf CO2 assimilation rates) atsaturating light intensities were consistently lower (P < 0.01)for the P450 phenotype (Fig. 2), as indicated by the value ofparameter Asat (Table 4). Light intensities at which photosyn-thesis rates were 50% of maximum (parameter Kq) were notsignificantly different between the two phenotypes (Table 4,Fig. 2). Individuals of the S and P450 phenotypes allocatedproportionally the same biomass to roots, leaves and stems(Fig. 1d–f, Table 3). The results from the 2002 and 2003experiments confirm that the P450 phenotype producesless vegetative biomass than the S phenotype and that thisdifference is the result of a lower RGR, NAR and CO2 assim-ilation rate.

Table 2 Mean total biomass, leaf area and number of vegetative tillers at 57 d after sowing (DAS) for the S, P450 and ACCase/P450 phenotypes of Lolium rigidum growing under temperature- and nitrogen-controlled conditions (2002 and 2003)

Growth traits

2002 Experiment (n = 18–24) 2003 Experiment (n = 25–28)

S P450 ACCase/P450 P S P450 ACCase/P450 P

Total biomass (mg) 9044a (546) 6792b (479) 6859b (326) < 0.001 9426a (430) 8113b (321) 8252b (358) 0.02Leaf area (cm2) 1019 (36) 974 (57) 1013 (61) 0.59 946 (43) 993 (38) 908 (49) 0.32Tillers 111 (5.5) 111 (5.8) 115 (5.6) 0.83 134 (8.8) 126 (6.4) 114 (4.7) 0.28

P-values from one-way ANOVA on individual traits. Different superscript letters indicate significant differences between mean values within rows for each experiment according to Tukey’s HSD (α = 0.05). Values in parentheses denote SE of the mean.

Table 3 Mean estimates of relative growth rate (RGR) and its components for the S, P450 and ACCase/P450 phenotypes of Lolium rigidum growing under temperature and nitrogen controlled conditions (2002 and 2003)

Growth rates and related components

2002 Experiment (n = 18–24) 2003 Experiment (n = 25–28)

S P450 ACCase/P450 S P450 ACCase/P450

29−57RGR (mg mg−1 d−1) 0.146a 0.130b 0.138ab 0.150a 0.140b 0.144ab

NAR (mg cm−2 d−1) 1.319a 0.981b 1.021b 1.580a 1.282b 1.401b

SLA (cm2 mg−1) 0.270a 0.289a 0.273a 0.221a 0.245a 0.224a

LWR (mg mg−1) 0.443a 0.479b 0.504b 0.464a 0.477a 0.474a

NAR, Net assimilation rate; SLA, specific leaf area; LWR leaf weight ratio. Parameters were estimated between the 29 and 57 d after sowing (DAS) harvest interval. Different superscript letters indicate significant differences within rows for each experiment according to Tukey’s HSD (α = 0.05).

New Phytologist (2005) 167: 787–796 www.newphytologist.org © New Phytologist (2005)

Research792

Herbicide susceptible (S) vs multiple resistant (ACCase/P450) phenotype

At 57 DAS (2002 experiment), ACCase/P450 plantsaccumulated significantly less total biomass than S plants(Table 2). No differences in the number of vegetative tillers orleaf area were observed between the phenotypes (Table 2). Alower RGR was recorded for the ACCase/P450 phenotype,although this difference was not statistically significant(Table 3). As observed for the P450 phenotype, ACCase/P450 individuals showed significantly lower NAR than the Sindividuals (Table 3). The ACCase/P450 phenotype exhibiteda higher LWR (Fig. 1b) and partitioned proportionally lessresources to roots and stems than the S phenotype (Fig. 1a,c).

In the 2003 experiment, as observed in 2002, the ACCase/P450 phenotype accumulated less total dry matter than the Sphenotype, and leaf area and number of vegetative tillers pro-duced were not significantly different (Table 2). The ACCase/P450 phenotype had a significantly lower NAR (Table 3) anda marginally reduced CO2 assimilation rate in relation tothe S phenotype (P = 0.05) (Table 4, Fig. 2). Accordingly, theACCase/P450 phenotype exhibited a reduced RGR, though thedifference between phenotypes was not statistically significant(Table 3). Individuals of the resistant phenotype allocatedproportionately less resources to stems (Fig. 1f ).

Metabolic (P450) vs multiple resistant (ACCase/P450) phenotype

In both experiments, there were no significant differences intotal dry matter acquisition or allocation to different organsbetween the two resistant phenotypes (Table 2, Fig. 1). Growthanalysis revealed that individuals of the P450 and ACCase/P450 phenotypes had similar RGR, which correlated well withsimilar net assimilation (Table 3) and photosynthetic rates(Table 4, Fig. 2).

Biomass allocation during the reproductive stage

Two independent experiments were conducted to assessreproductive growth traits for plants of the S, P450 andACCase/P450 phenotypes grown under glasshouse (2002)and field conditions (2003). For the glasshouse experiment, analysis identified a significant (P = 0.0015) overalleffect of phenotype on reproductive growth attributes. However,when analysed individually by , the only significantlydifferent trait was the number of reproductive tillers, with theS phenotype producing significantly more tillers than eitherresistant phenotype (Table 5). There were no significantdifferences between phenotypes in total seed mass, the numberof seeds produced or individual seed weight (Table 5).

Fig. 1 Biomass allocation (%) to roots (root–weight ratio, RWR) (a and d), leaves (leaf–weight ratio, LWR) (b and e) and stems (stem–weight ratio, SWR) (c and f) for S, P450 and ACCase/P450 phenotypes of Lolium rigidum estimated 57 d after sowing (DAS) (2002 and 2003). Vertical bars denote SE of the mean. Different letters indicate significant differences between mean values according to Tukey’s HSD (α = 0.05).

© New Phytologist (2005) www.newphytologist.org New Phytologist (2005) 167: 787–796

Research 793

In the field experiment, did not reveal any signif-icant (P = 0.60) effect of phenotype on reproductive traits ofindividuals of the S, P450 and ACCase/P450 phenotypes.Reproductive output for all phenotypes was higher thanobserved in the glasshouse experiment (except for number ofreproductive tillers) but was not significantly different betweenphenotypes (Table 5).

Discussion

This study conducted with a single multiple-resistant L. rigidumpopulation (SLR31) has identified a consistent reduction invegetative biomass (resource acquisition) in the herbicide-resistant P450 and ACCase/P450 phenotypes compared withthe susceptible phenotype (S > P450 = ACCase/P450) after57 d growth (Table 2). This indicates a physiological or directresistance cost associated with P450-endowed herbicideresistance expressed at the vegetative growth stage. This is the

first report that a resistance mechanism of enhanced herbicidemetabolism results in a physiological cost.

These differences in biomass production during the vege-tative stage, did not result in differences in reproductive output(seed mass and seed number) (S = P450 = ACCase/P450)(Table 5). This lack of correspondence between vegetative andreproductive performance has also been noted in some previ-ous studies of herbicide-resistant and -susceptible biotypes(Stowe & Holt, 1988; Reboud & Till-Bottraud, 1991; Purbaet al., 1996). Reboud and Till-Bottraud (1991) have sug-gested that the advantage of herbicide susceptible over resist-ant individuals may be most apparent during early vegetativegrowth and that compensation of resistance costs over timemay occur in the absence of stressful conditions, leading tosimilar reproductive outputs. Also, the expression of pleio-tropic costs associated with herbicide resistance mechanismsmay vary between plant organs (stems, leaves or flowers) andgrowth stages (Purrington & Bergelson, 1999), making it nec-essary to evaluate the basis of these costs (which may be relatedto P450 expression levels) in vegetative and reproductive plantstages.

Variations in RGR necessarily correspond to linear varia-tions in its physiological (NAR) and morphological (SLA and

Fig. 2 Leaf CO2 assimilation rate plotted against photosynthetic photon flux density (PPFD) for the S (open circles), P450 (open squares) and ACCase/P450 (closed squares) phenotypes of Lolium rigidum. Photosynthetic rates (n = 7) were measured for single fully expanded leaves in an atmosphere enriched with 360 mmol CO2 mol−1 at constant 20°C and 500 µmol s−1 air flow rate. Vertical bars denote SE of the mean.

Table 5 Mean values for reproductive traits of the S, P450 and ACCase/P450 phenotypes of Lolium rigidum under glasshouse and field conditions

Reproductive growth traits

Glasshouse experiment Field experiment

S P450 ACCase/P450 P S P450 ACCase/P450 P

Number of reproductive tillers 169a (8.8) 142a (8.3) 116b (4.7) < 0.001 133 (7.4) 120 (8.9) 117 (7.2) 0.25Total seed mass (g) 9.3 (0.7) 10.3 (0.6) 9.3 (0.8) 0.40 20.3 (1.2) 18.6 (1.1) 20.9 (1.2) 0.44Harvest index (%) 16.6 (1.1) 18.6 (0.9) 17.0 (1.1) 0.35 29.7 (1.5) 28.3 (0.9) 30.4 (1.1) 0.52Seed number 5310 (292) 5643 (351) 5213 (456) 0.38 9847 (835) 8488(516) 9600 (689) 0.28Individual seed weight (mg) 1.79 (0.07) 1.86 (0.07) 1.83 (0.06) 0.83 2.35 (0.4) 2.23 (0.1) 2.07 (0.1) 0.82

P-values from one-way ANOVA on individual traits. Different superscript letters indicate significant differences between phenotypes within rows according to the Kruskal–Wallis test (α = 0.05). Values in parentheses denote SE of the mean.

Table 4 Parameter estimates and regression coefficients for the hyperbolic model (y = Asatx/(Kq + x)) describing leaf CO2 assimilation rate (rate of photosynthesis) in response to light intensity for the S, P450 and ACCase/P450 phenotypes of Lolium rigidum

Phenotype

Parameter

Asat Kq R2

S 36.1a (33.5–38.8) 81.3a (64–99) 0.79P450 30.8b (29.3–32.4) 71.1a (60–82) 0.88ACCase/P450 32.6b (30.1–35.1) 79.9a (62–98) 0.78

Data were square root transformed and parameter estimates and their 95% confidence limits (values in parentheses) were back-transformed. Different superscript letters denote significant differences between the phenotypes by Tukey’s HSD test (P ≤ 0.05).

New Phytologist (2005) 167: 787–796 www.newphytologist.org © New Phytologist (2005)

Research794

LWR) components. According to previous studies, inherentvariations in RGR in terrestrial plants may be alternativelyexplained by variations in either NAR (Garnier, 1992) orSLA (Poorter & Remkes, 1990; Hunt & Cornelissen, 1997),although other studies have found no clear explanatory corre-lation between RGR and any of its components (Meziane &Shipley, 1999). In this study, results from two experiments,have shown that significant differences in the RGR of S andP450 phenotypes are well correlated with differences in netassimilation rate (NAR) (Table 3).

In order to limit differences in RGR which resulted fromthe lower NAR and photosynthetic rate of resistant pheno-types, compensatory changes in LWR were observed. Bothresistant phenotypes allocated more resources to leaves (LWR)than the S phenotype, in order to maximize total carbonfixation. This morphological adjustment was not sufficientto compensate for their intrinsically lower NAR, which stillresulted in lower total dry matter production (Table 2).

Herbicide resistance costs

The magnitude (%) of resistance costs (1 − (resistant biomass/susceptible biomass) × 100) associated with enhanced P450herbicide metabolism at the vegetative stage were 25% and13% for the 2002 and 2003 experiments, respectively. Ofcourse, it is not valid to widely extrapolate from this resultas it remains to be established whether the physiologicalcost accompanying the nontarget site P450-based enhancedherbicide metabolism resistance mechanism observed in thisSLR31 L rigidum population is associated with resistance inother L. rigidum populations and other resistant species. Withinsecticide-resistant insect species, P450-enhanced metabolismhas been shown to incur significant resistance costs (Boivinet al., 2003).

Our study suggests that the mutation Ile (1781) Leu inthe ACCase gene incurred no measurable physiological cost, asresponses of multiple resistant individuals (ACCase/P450)were not significantly different from those of individualspossessing only enhanced P450 metabolism during vegetativeand reproductive growth. This Ile (1781) Leu mutation hasnow been observed in several species (Zhang & Devine,2000; Brown et al., 2002; Christoffers et al., 2002; Délye et al.,2002a, b) and studies on resistance costs should be conductedwith these resistant biotypes to establish cost of this mutationin other resistant populations/species. This result with theACCase gene Ile (1781) Leu mutation of no resistance costcannot of course be extrapolated to other mutations of theACCase gene that endow herbicide resistance. Resistance coststudies need to be conducted with the other known resistanceendowing ACCase gene mutations (Délye et al., 2003, 2005).It is acknowledged that in the present study a direct compar-ison of the fitness of ACCase mutants without the presence ofaltered P450 metabolism was not possible and is required tounequivocally assert the absence of a resistance cost for this

ACCase mutation. Two studies found no direct costs ofresistance to ACCase inhibiting herbicides (APP-CHD) inDigitaria sanguinalis (Wiederholt & Stoltenberg, 1996a) andSetaria faberi (Wiederholt & Stoltenberg, 1996b). Althoughis probable that these resistant biotypes had a target site muta-tion, the lack of knowledge of the molecular basis of resistanceand the comparison of resistant and susceptible individualswith dissimilar genetic backgrounds precludes a clear compar-ison with the results of our study.

Despite this apparent lack of a resistance cost associatedwith the Ile (1781) Leu mutation in this L. rigidum popula-tion, other studies have demonstrated significant differencesin germination requirements and seedling emergence charac-teristics for the ACCase phenotype (Vila-Aiub et al., 2005).

Ecological significance of herbicide resistance costs

When they occur, resistance costs associated with allelesendowing herbicide resistance may potentially limit thefrequency and alter the evolutionary dynamics of these allelesin weed populations. The extent and consequences of thesecosts will depend on the intensity of competitive interactionsthat will ultimately modify plant fitness. These competitiveinteractions will interplay with environmental conditions andlife history traits to modify the intensity of plant competitionand ecological fitness relations between resistant and susceptiblecomponents of weed populations (Whittaker et al., 1973;Harper, 1977; Purba et al., 1996).

The extent and magnitude of any physiological cost associ-ated with herbicide resistance mechanisms is an importantparameter influencing the preselection frequency of resistancein populations and the subsequent dynamics of resistanceduring and after herbicide selection. Knowledge of resistancecosts is invaluable in predictive models of herbicide resistancedynamics under various management scenarios (Gressel &Segel, 1990; Maxwell et al., 1990; Diggle & Neve, 2001; Neveet al., 2003). The results of this study suggest that attempts topredict the rate of evolution of herbicide resistance shouldconsider that the potential expression of resistance costs mayvary depending on the mechanism of resistance and the plantlife stage involved. Based on predictions from asymmetriccompetition (Shipley & Keddy, 1994), in which large indi-viduals obtain more resources than small plants and size vari-ability becomes magnified over time (i.e. large plants suppresssmall plants) (Weiner, 1986), individuals possessing anenhanced P450 metabolism might potentially be at adisadvantage when competing with herbicide-susceptibleindividuals. Agronomic practices such as the inclusion ofcompetitive crops and pasture phases that exploit the reducedgrowth of individuals with P450-enhanced herbicide metab-olism may help to select against herbicide-resistant individualsand moderate rates of resistance evolution in the field.

In conclusion, the data presented in this study provide thefirst evidence for a resistance cost associated with a P450-based

© New Phytologist (2005) www.newphytologist.org New Phytologist (2005) 167: 787–796

Research 795

nontarget site herbicide detoxification mechanism and theabsence of a resistance cost for a target site ACCase genemutation (Ile (1781) Leu). This information is needed for usein predictive modeling and in designing resistance manage-ment strategies.

Acknowledgements

WAHRI is funded by the Australian Grains Research andDevelopment Corporation (GRDC). An International Post-graduate Research Scholarship supported M. M. V. A. duringthis research at The University of Western Australia. Theauthors thank Dr K. Steadman and Dr M. Walsh for criticalreading of the manuscript.

References

Bergelson J, Purrington CB. 1996. Surveying patterns in the cost of resistance in plants. American Naturalist 148: 536–558.

Berticat C, Boquien G, Raymond M, Chevillon C. 2002. Insecticide resistance genes induce a mating competition cost in Culex pipiens mosquitoes. Genetical Research 79: 41–47.

Boivin T, Bouvier JC, Chadoeuf J, Beslay D, Sauphanor B. 2003. Constraints on adaptive mutations in the codling moth Cydia pomonella (L.): measuring fitness trade-offs and natural selection. Heredity 90: 107–113.

Brown AC, Moss SR, Wilson ZA, Field LM. 2002. An isoleucine to leucine substitution in the ACCase of Alopecurus myosuroides (black-grass) is associated with resistance to the herbicide sethoxydim. Pesticide Biochemistry and Physiology 72: 160–168.

Causton DR, Venus JC. 1981. The biometry of plant growth. London, UK: Edward Arnold.

Christopher JT, Powles SB, Lijegren DR, Holtum JAM. 1991. Cross-resistance to herbicides in annual ryegrass (Lolium rigidum) II. Chlorsulfuron resistance involves a wheat-like detoxification system. Plant Physiology 95: 1036–1043.

Christopher JT, Preston C, Powles SB. 1994. Malathion antagonizes metabolism-based chlorsulfuron resistance in Lolium rigidum. Pesticide Biochemistry and Physiology 49: 172–182.

Christoffers MJ, Berg ML, Messersmith CG. 2002. An isoleucine to leucine mutation in acetyl-CoA carboxylase confers herbicide resistance in wild oat. Genome 45: 1049–1056.

Coley PD, Bryant JP, Chapin FS III. 1985. Resource availability and plant antiherbivore defense. Science 230: 895–899.

Coustau C, Chevillon C, ffrench-Constant R. 2000. Resistance to xenobiotics and parasites: can we count the cost? Trends in Ecology and Evolution 15: 378–383.

Délye C, Matéjicek A, Gasquez J. 2002a. PCR-based detection of resistance to acetyl-CoA carboxylase-inhibiting herbicides in black-grass (Alopecurus myosuroides Huds) and ryegrass (Lolium rigidum Gaud). Pest Management Science 58: 474–478.

Délye C, Wang T, Darmency H. 2002b. An isoleucine-leucine substitution in chloroplastic acetyl-CoA carboxylase from green foxtail (Setaria viridis L. Beauv.) is responsible for resistance to the cyclohexanedione herbicide sethoxydim. Planta 214: 421–427.

Délye C, Zhang XQ, Chalopin C, Michel S, Powles SB. 2003. An isoleucine residue within the carboxyl-transferase domain of multidomain acetyl-coenzyme A carboxylase is a major determinant of sensitivity to aryloxyphenoxypropionate but not to cyclohexanedione inhibitors. Plant Physiology 132: 1716–1723.

Délye C, Zhang XQ, Chalopin C, Michel S, Matéjicek A, Powles SB. 2005.

Molecular bases for sensitivity to acetyl-coenzyme A carboxylase inhibitors in black-grass (Alopecurus myosuroides Huds.). Plant Physiology 137: 794–806.

Diggle AJ, Neve P. 2001. The population dynamics and genetics of herbicide resistance. A modeling approach. In: Powles SB, Shaner DL, eds. Herbicide Resistance and world grains. Boca Raton, Florida, USA: CRC Press, 61–99.

Fisher RA. 1928. The possible modification of the response of the wild type to recurrent mutations. American Naturalist 62: 115–126.

Garnier E. 1992. Growth analysis of congeneric annual and perennial grass species. Journal of Ecology 80: 665–675.

Gressel J, Segel LA. 1990. Modelling the effectiveness of herbicide rotations and mixtures as strategies to delay or preclude resistance. Weed Technology 4: 186–198.

Gronwald JW. 1994. Resistance to photosystem II inhibiting herbicides. In: Powles SB, Holtum JAM, eds. Herbicide resistance in plants. Biology and biochemistry. Boca Raton, FL, USA: Lewis Publishers, 27–60.

Hall LM, Moss SR, Powles SB. 1995. Mechanism of resistance to chlorotoluron in two biotypes of the grass weed Alopecurus myosuroides. Pesticide Biochemistry and Physiology 53: 180–192.

Harper JL. 1977. Population biology of plants. London, UK: Academic Press.Herms DA, Mattson WJ. 1992. The dilemma of plants: to grow or defend.

Quarterly Review of Biology 67: 283–335.Hoffmann WA, Poorter H. 2002. Avoiding bias in calculations of relative

growth rate. Annals of Botany 80: 37–42.Hunt R. 1982. Plant growth curves. The functional approach to plant growth

analysis. London, UK: Edward Arnold.Hunt R, Cornelissen JHC. 1997. Components of relative growth rate and

their interrelations in 59 temperate plant species. New Phytologist 135: 395–417.

Hunt R, Causton DR, Shipley B, Askew AP. 2002. A modern tool for classical plant growth analysis. Annals of Botany 90: 485–488.

Karban R, Agrawal AA, Mangel M. 1997. The benefits of induced defenses against herbivores. Ecology 78: 1351–1355.

LeBaron H, Gressel J. 1982. Herbicide resistance in plants. New York, USA: Wiley.

Long SP, Hällgren JE. 1993. Measurement of CO2 assimilation by plants in the field and the laboratory. In: Hall DO, Scurlock JMO, Bolhar-Nordenkampf HR, Leegood RC, Long SP, eds. Photosynthesis and production in a changing environment: a field and laboratory manual. London, UK: Chapman & Hall, 129–167.

Maxwell BD, Roush ML, Radosevich SR. 1990. Predicting the evolution and dynamics of herbicide resistance in weed populations. Weed Technology 4: 2–13.

Meziane D, Shipley B. 1999. Interacting components of interspecific relative growth rate: constancy and change under differing conditions of light and nutrient supply. Functional Ecology 13: 611–622.

Neve P, Diggle AJ, Smith FP, Powles SB. 2003. Simulating evolution of glyphosate resistance in Lolium rigidum I: population biology of a rare resistance trait. Weed Research 43: 404–417.

Poorter H, Nagel O. 2000. The role of biomass allocation in the growth response of plants to different levels of light, CO2, nutrients and water: a quantitative review. Australian Journal of Plant Physiology 27: 595–607.

Poorter H, Remkes C. 1990. Leaf area ratio and net assimilation rate of 24 wild species differing in relative growth rate. Oecologia 83: 553–559.

Powles SB, Holtum JAM. 1994. Herbicide resistance in plants. Biology and biochemistry. Boca Raton, FL, USA: Lewis Publishers.

Powles SB, Shaner DL. 2001. Herbicide Resistance and World Grains. Boca Raton, FL, USA: CRC Press.

Preston C, Mallory-Smith CA. 2001. Biochemical mechanisms, inheritance, and molecular genetics of herbicide resistance in weeds. In: Powles SB, Shaner DL, eds. Herbicide Resistance and World Grains. Boca Raton, FL, USA: CRC Press, 23–60.

New Phytologist (2005) 167: 787–796 www.newphytologist.org © New Phytologist (2005)

Research796

Preston C, Powles SB. 1998. Amitrole inhibits diclofop metabolism and synergises diclofop-methyl in a diclofop-methyl-resistant biotype of Lolium rigidum. Pesticide Biochemistry and Physiology 62: 179–189.

Preston C, Tardif FJ, Christopher JT, Powles SB. 1996. Multiple resistance to dissimilar herbicide chemistries in a biotype of Lolium rigidum due to enhanced activity of several herbicide degrading enzymes. Pesticide Biochemistry and Physiology 54: 123–134.

Purba E, Preston C, Powles SB. 1996. Growth and competitiveness of paraquat-resistant and -susceptible biotypes of Hordeum leporinum. Weed Research 36: 311–317.

Purrington CB, Bergelson J. 1997. Fitness consequences of genetically engineered herbicide and antibiotic resistance in Arabidopsis thaliana. Genetics 145: 807–814.

Purrington CB, Bergelson J. 1999. Exploring the physiological basis of costs of herbicide resistance in Arabidopsis thaliana. American Naturalist 154: S82–S91.

Reboud X, Till-Bottraud I. 1991. The cost of herbicide resistance measured by a competition experiment. Theoretical and Applied Genetics 82: 690–696.

Roux F, Gasquez J, Reboud X. 2004. The dominance of the herbicide resistance cost in several Arabidopsis thaliana mutant lines. Genetics 166: 449–460.

Shipley B, Keddy PA. 1994. Evaluating the evidence for competitive hierarchies in plant communities. Oikos 69: 340–345.

Sokal RR, Rohlf FJ. 1969. Biometry. The principles and practice of statistics in biological research. San Francisco, CA, USA: Freeman & Co.

Stowe E, Holt JS. 1988. Comparison of triazine resistant and susceptible biotypes of Senecio vulgaris and their F1 hybrids. Plant Physiology 87: 183–189.

Strauss SY, Rudgers JA, Lau JA, Irwin RE. 2002. Direct and ecological costs of resistance to herbivory. Trends in Ecology and Evolution 17: 278–285.

Tardif FJ, Holtum JAM, Powles SB. 1993. Occurrence of a herbicide-resistant acetyl-coenzyme A carboxylase mutant in annual ryegrass (Lolium rigidum) selected by sethoxydim. Planta 190: 176–181.

Tardif FJ, Powles SB. 1994. Herbicide multiple-resistance in a Lolium rigidum biotype is endowed by multiple mechanisms: isolation of a subset with resistant acetyl-CoA carboxylase. Physiologia Plantarum 91: 488–494.

Vila-Aiub MM, Neve P, Steadman KJ, Powles SB. 2005. Ecological fitness of a multiple herbicide resistant Lolium Rigidum population: dynamics of seed germination and seedling emergence of resistant and susceptible phenotypes. Journal of Applied Ecology 42: 288–298.

Weiner J. 1986. How competition for light and nutrients affects size variability in Ipomoea tricolor populations. Ecology 67: 1425–1427.

Whittaker RH, Levin SA, Root RB. 1973. Niche, habitat and ecotope. American Naturalist 107: 321–338.

Wiederholt RJ, Stoltenberg DE. 1996a. Similar fitness large crabgrass (Digitaria sanguinalis) accessions resistant or susceptible to acetyl-coenzyme A carboxylase inhibitors. Weed Technology 10: 42–49.

Wiederholt RJ, Stoltenberg DE. 1996b. Absence of differential fitness between giant foxtail (Setaria faberi) accessions resistant and susceptible to acetyl-coenzyme A carboxylase inhibitors. Weed Science 44: 18–24.

Zhang XQ, Devine MD. 2000. A possible point mutation of plastidic ACCase gene conferring resistance to sethoxydim in green foxtail (Setaria viridis) (abstract no. 81). Weed Science Society of America 40: 33.

About New Phytologist

• New Phytologist is owned by a non-profit-making charitable trust dedicated to the promotion of plant science, facilitating projectsfrom symposia to open access for our Tansley reviews. Complete information is available at www.newphytologist.org.

• Regular papers, Letters, Research reviews, Rapid reports and both Modelling/Theory and Methods papers are encouraged.We are committed to rapid processing, from online submission through to publication ‘as-ready’ via OnlineEarly – the 2004 averagesubmission to decision time was just 30 days. Online-only colour is free, and essential print colour costs will be met if necessary.We also provide 25 offprints as well as a PDF for each article.

• For online summaries and ToC alerts, go to the website and click on ‘Journal online’. You can take out a personal subscription tothe journal for a fraction of the institutional price. Rates start at £109 in Europe/$202 in the USA & Canada for the online edition(click on ‘Subscribe’ at the website).

• If you have any questions, do get in touch with Central Office (newphytol@lancaster.ac.uk; tel +44 1524 594691) or, for a localcontact in North America, the US Office (newphytol@ornl.gov; tel +1 865 576 5261).