Molecular Mechanisms of Herbicide Resistance

Molecular Mechanisms of Herbicide Resistance

Christophe Delye, Arnaud Duhoux, Fanny Pernin, Chance W. Riggins, and Patrick J. Tranel*

Key words: Allele-specific PCR, dCAPS, DNA sequencing, gene expression, mutation, PCR, qPCR.

Resistance to herbicides occurs in weeds as theresult of evolutionary adaptation (Jasieniuk et al.1996). Basically, two types of mechanisms areinvolved in resistance (Beckie and Tardif 2012;Delye 2013). Target-site resistance (TSR) is causedby changes in the tridimensional structure of theherbicide target protein that decrease herbicidebinding, or by increased activity (e.g., due toincreased expression or increased intrinsic activity)of the target protein. Nontarget-site resistance(NTSR) is endowed by any mechanism notbelonging to TSR, e.g., reduction in herbicideuptake or translocation in the plant, or enhancedherbicide detoxification (reviewed in Delye 2013;Yuan et al. 2007).

Mutations endowing herbicide resistance can beclassified into two types. The first type is structuralchanges in a DNA sequence encoding a protein, i.e.,structural mutations. Structural mutations endow-ing herbicide resistance are expected to cause astructural modification in the tridimensional struc-ture of a protein that will lead to a decrease in theefficacy of an herbicide. For example, mutationsconferring an amino acid substitution at theherbicide-binding site of a target protein candecrease the affinity of the herbicide for the targetprotein (TSR). Alternatively, mutations at the activesite of a metabolic enzyme or a transporter proteincan improve the activity of these proteins inherbicide degradation or compartmentation awayfrom its site of action, respectively (NTSR). In thecase of structural changes in DNA sequence, seekingthe cause for resistance means identifying and beingable to detect the relevant structural mutations inthe DNA of resistant plants.

The second type of mutations associated withherbicide resistance results in a difference in the

expression of one or several genes in resistant plantscompared to sensitive plants, i.e., regulatory muta-tions (Delye 2013; Yuan et al. 2007). Thesemutations are changes in a DNA sequence that cancause an increase in the expression of the herbicidetarget protein that compensates for the herbicideinhibitory action (TSR), or a variation in theexpression of herbicide-metabolizing enzyme(s) orof transporter proteins that will lead to an increase inherbicide degradation or compartmentation awayfrom its site of action, respectively (NTSR).Identifying regulatory mutation(s) responsible forchanges in gene expression is not straightforward,because these mutations can be of diverse nature.Examples include whole-gene amplification (e.g.,Gaines et al. 2010), structural changes in thepromoter sequence of the gene encoding the proteinshowing a variation in expression, or even structuralchanges in the promoter sequence or in the codingsequence of a gene encoding a protein that regulatesthe expression of the protein showing a variation inexpression (Delye 2013). Epigenetic processes (e.g.,DNA methylation) can also be involved in theregulation of gene expression (Delye 2013). Thus, inthe case of regulatory mutations or of epigeneticregulation, seeking and detecting herbicide resistanceis most easily achieved by identifying and being ableto detect significant differences in expression of genesbetween resistant and sensitive plants.

A variety of approaches are available for con-firming and evaluating herbicide resistance in weeds(Burgos et al. 2013). The aim of this paper is toprovide the inexperienced researcher with informa-tion to investigate herbicide resistance at the DNAlevel. Protocols and guidelines are provided forinvestigating both structural changes in DNAsequence and changes in gene expression.

DNA and RNA Basics

Most DNA-based assays for herbicide resistancerely on the polymerase chain reaction (PCR) toselectively amplify a DNA sequence of interest fromthe milieu of DNA that is not of interest. Thus, we

DOI: 10.1614/WS-D-13-00096.1* First, second, and third authors: Researcher, PhD Student,

and Research Technical Assistant, INRA, UMR1347 Agroeco-logie, F-21000 Dijon, France; fourth and fifth authors:Postdoctoral Research Assistant and Professor, Department ofCrop Sciences, University of Illinois, Urbana, IL 61801.Corresponding author’s E-mail: [email protected]

Weed Science 2015 Special Issue:91–115

Delye et al.: Methods for HR mechanisms N 91

begin with a discussion of extracting DNA for usein PCR, performing the PCR, and analyzing thePCR products.

Most standard ‘‘genomic’’ DNA extraction proce-dures yield DNA from the nuclear, chloroplastic, andmitochondrial genomes, and thus are suitable for awide range of downstream molecular analyses,including PCR. Although DNA can be extractedfrom all types of plant material (e.g., leaves, roots,stems, seeds, preserved tissue), young, newly emergedleaves are often best for extraction because they areeasier to grind up than mature tissue and typicallycontain lower amounts of polysaccharides, polyphe-nolics, and other secondary metabolites that caninterfere with DNA isolation. Attention should begiven to selecting healthy, clean tissue (e.g., withoutvisible signs of infection, fungi, insects, soil) and properhandling to prevent plant-to-plant contamination. Inthe absence of fresh tissue, high-quality DNA can alsobe extracted from preserved material. For example, ifimmediate processing of fresh material is not possible(such as under field conditions), then the plant tissuecan be kept on ice for several hours or preserved forfuture workup by drying or freezing. Leaves should bedried relatively quickly and stored in a moisture-freeenvironment to prevent rotting. Leaf tissue can bepressed between paper towels and air-dried for severaldays at room temperature or placed in a zip-lock bagwith silica gel to absorb moisture. Alternatively, leaftissue can be frozen by first rinsing and blotting dry,and then storing at 220 C or indefinitely at 280 C.Frozen tissue should not be allowed to thaw beforeprocessing because this increases the risk of DNAdegradation by endogenous nucleases.

Following are two common and relatively simpleprotocols that we use routinely to obtain DNAsuitable for PCR and other molecular assays. Theseprotocols are only two of numerous DNA extractionmethods that have been published in the literature orare available online, and there is also a variety ofcommercial DNA extraction kits available, such asthe DNeasy Plant kits (Qiagen, Valencia, CA) andE.Z.N.A. Plant DNA kits (Omega Bio-Tek, Nor-cross, GA). These and other commercial kits typicallyyield DNA of higher purity, but one must factor inthe cost of the kits with the time, expertise, scale, andgoals of the project.

Easy Way: Brutus DNA Extraction. Brutus DNAextraction basically consists of grinding a plantfragment in a salty buffer, then boiling it to releaseDNA from plant tissues (Delye et al. 2002). Thequality and quantity of the DNA solution issufficient to provide DNA templates suitable for arange of PCR-based techniques, ranging from basicPCRs to sophisticated techniques such as TaqMan(e.g., Delye et al. 2010). This easy DNA extractionprotocol is fast and costs almost nothing. It can beperformed from different plant tissues (leaf, root,green stem), either fresh or dried (Delye et al.2011); however, Brutus extraction from ripe seeds isoften not suitable for PCR.

Materials needed: Extraction buffer (to make500 ml, mix 50 ml 1 M Tris-HCl, pH 5 9.5, 10 ml0.5 M EDTA pH 5 8, and 37.275 g KCl in 200 mldeionized water, bring to 500 ml final volume withdeionized water and sterilize in an autoclave); ameans to grind the tissue, e.g., a bead mill ordisposable plastic pestles; and a tube-heating deviceat 95 C (e.g., a water bath).

General steps are as follows:

1. Collect 10 to 20 mm2 section of plant tissue in amicrocentrifuge tube.

2. Add 100 ml extraction buffer per sample.3. Grind tissue. If using a bead mill, use two rounds

of 90 s at 30 shakes s21.4. Incubate tubes at 95 C for 5 min.5. Place tubes in ice for 10 min.6. Centrifuge tubes (at least 3,000 to 4,000 3 g for

1 min, so that tissue fragments lie at the bottomof the tube and are not pipetted to reaction mixesin downstream experiments).

7. Store DNA samples at 220 C, or use immedi-ately.

Tip: Use molecular biology-grade reagents for allof the following protocols.

Tip: Any type of grinding device can be used,provided it efficiently shreds plant tissues. Forprocessing a small number of samples, tissue canbe ground with pestles designed for use in 1.5 mlmicrocentrifuge tubes (e.g., Fisher Scientific 12-141-364; Pittsburgh, PA). For a large number ofsamples, tissue can be ground in 1.5 ml micro-centrifuge tubes with a 3-mm glass bead or in0.5 ml or 0.2 ml microcentrifuge tubes with a 2-mm glass bead, using a bead mill (e.g., RetschMM400; Hann, Germany).

92 N Weed Science 63, Special Issue 2015

CTAB DNA Extraction. One of the most popularplant DNA extraction buffers is CTAB (cetyltri-methyl ammonium bromide; also called hexadecyl-trimethyl ammonium bromide). The most com-monly used CTAB extraction protocol is that ofDoyle and Doyle (1990), and it has been successfullyemployed for a wide variety of plant species.Numerous modifications of this basic method areavailable. The following protocol works for severalspecies, and would be a good option to try if theBrutus method is not successful. The CTABprocedure includes chloroform extraction andDNA precipitation steps, which result in cleanerDNA than is obtained with the Brutus method.

Material needed: CTAB extraction buffer (tomake 100 ml, mix 10 ml 1M Tris-HCl pH 8, 28 ml5M NaCl, 4 ml 0.5M EDTA, 2 g CTAB, and 40 mldeionized water; heat to dissolve; bring to 100 mlfinal volume with deionized water; and filtersterilize with e.g., Nalgene Rapid-Flow TissueCulture Filter Unit, 0.22 mm; (Rochester, NY);pestles for grinding samples; microcentrifuge capa-ble of spinning 1.5 ml polypropylene tubes at13,500 3 g at room temperature; water bath.

The extraction steps are as follows:

1. Collect 10 to 20 mm2 tissue in a 1.5 ml tubeand grind with pestle.

2. Add 600 ml CTAB buffer to ground tissue withpestles still in tubes. Several samples can beprocessed at once. Additional grinding withpestles might be necessary to fully pulverizetissue before proceeding to the next step.

3. Incubate homogenate at 65 C for 20 min to 1 hin water bath. Occasionally mix by inversion toavoid aggregation of homogenate.

4. Add 400 ml chloroform and mix by inversion.

5. Centrifuge 5 min at $ 13,500 3 g in amicrocentrifuge at room temperature.

6. Transfer upper aqueous layer (approximately500 ml) into a new 1.5 ml tube and add an equalvolume of isopropanol. Mix by inversion.

7. Centrifuge 10 min at $ 13,500 3 g at roomtemperature.

8. Discard supernatant and add 250 ml of 80%ethanol. Centrifuge for 2 min as before.

9. Discard supernatant and add 250 ml of 95%ethanol. Centrifuge for 2 min as before.

10. Discard supernatant and dry pellet completely.

11. Resuspend dry DNA pellet in 100 ml water orTE buffer. Store at 220 C.

Tip: Brutus DNA samples are very stable, evenunder repeated freeze/thaw cycles. Furthermore,tubes containing the mixture of boiled buffer andtissues can be reused when the DNA solution hasrun out: simply add 50 to 100 ml fresh extractionbuffer to the ground plant tissue remaining in thetube, and repeat the extraction procedure.

Tip: The protocol provided uses a pestle forgrinding samples, but could be modified to use abead mill for higher sample throughput.

Tip: Chloroform is often used as a solvent inDNA extraction protocols; however, dichloro-methane is less toxic and can be substituted forchloroform (Chaves et al. 1995).

Tip: The nucleic acid pellet is usually visible atthis stage and attention must be given not to losethe pellet when decanting.

Tip: Residual ethanol should be removed beforedissolving the DNA pellet in sterile distilled wateror TE buffer because it can inhibit downstreamenzymatic reactions. The pellet can be air-driedfor 30 min, or a Savant SpeedVac (ThermoScientific, Waltham, MA) with medium heat canalso be used to dry the pellet. Do not dry thepellets too long in the SpeedVac (5 min is usuallylong enough) or they will be difficult toresuspend.

Tip: DNA is more stable in TE buffer (10 mMTris-HCl pH 8, 1 mM EDTA) than in water, soTE buffer is recommended for long-term storage.However, EDTA can interfere with PCR, so weoften store DNA in a 0.13 EDTA TE buffer(10 mM Tris-HCl pH 8, 0.1 mM EDTA).

Delye et al.: Methods for HR mechanisms N 93

RNA Extraction and cDNA Synthesis. DNA canalso be synthesized from messenger RNA (mRNA)using a reverse-transcriptase enzyme. This enzymesynthesizes DNA complementary to RNA (cDNA)from the 39 end of a primer (a small, single-strandDNA molecule) hybridized on the RNA strand,using the RNA strand as a template. cDNA is ofparticular interest when working on genes withcomplex intron–exon structure, because, likemRNAs, cDNAs do not contain introns. For‘‘simple’’ genes such as the gene encoding acet-olactate synthase (ALS) (ca. 2,000 base pairs [bp]with no introns in most plants), genomic DNA andcDNA are identical (within transcribed regions).For genes such as the gene encoding chloroplasticacetyl CoA carboxylase (ACCase) in grasses (familyPoaceae), the cDNA sequence is ca. 7,000 bp long.The corresponding genomic DNA is ca. 10,000 to12,000 bp long, and contains 32 introns (Delye2005; Huang et al. 2002).

Traditional protocols for RNA extraction andpurification and cDNA synthesis are time consum-ing, and often involve either dangerous reagents,and/or complex purification procedures. However,RNA extraction and cDNA synthesis are now mosteasily done using combinations of commercial kits,which can be used without particular training. Anexample of a combination of widely distributed kitsyielding good quality cDNA is RNeasy Plant MiniKit (Qiagen) plus RNAse Free DNAse Set for RNAextraction and genomic DNA removal (Qiagen),followed by 5PRIME Masterscript Kit (FisherScientific) for reverse-transcription reactions.

RNA must be treated with care. RNA is a single-stranded molecule and, due to its chemistry, and tothe presence of RNAses everywhere, including in

your plant samples and on your skin, RNA is morefragile than double-stranded DNA. It is essentialthat RNAses are absent or neutralized whenhandling RNA (efficient RNAse-neutralizing solu-tions are commercially available). The extractionsteps are as follows:

1. Proceed to the extraction of RNA from planttissue samples immediately after collection.Because collection of plant tissue can inducewound stress response or RNA degradation dueto RNAse release, it is best to freeze samples inliquid nitrogen immediately after collection. Ifplant tissue sample are not to be processedimmediately, store at 280 C.

2. Use RNAse-free plasticware and reagents.3. Perform reverse-transcription reactions as soon as

possible after RNA extraction.4. Check RNA quantity and quality using a UV

spectrophotometer. RNA quality is assessed usingthe A260/A280 and A260/A230 absorptionratios. RNA samples with A260/A280 andA260/A230 ratios between 1.8 and 2.2 aregenerally considered suitable.

5. Do not store RNA samples more than a few daysat 220 C. Longer-term storage must be at280 C. When storing RNA samples for severalmonths, check RNA quantity and quality againprior to any experiment to estimate RNAdegradation.

6. Do not let RNA pellets dry. Rapidly dissolve inRNAse-free water or buffer.

7. Do not subject RNA samples to frequent freeze-and-thaw cycles. Let RNA samples gently thawon ice. Work on ice as much as possible.

After RNA is extracted, cDNA synthesis usingreverse-transcription mixes can be conducted withseveral types of primers. If no specific gene is

Tip: RNA can be extracted from all living tissues.Young, actively growing tissues typically willprovide the highest yields of RNA. However,when extracting RNA (unlike when extractingDNA), one must select tissue in which the geneof interest is likely to be expressed (e.g.,aboveground plant part for foliar herbicides,roots for soil-applied herbicides).

Tip: Never let a frozen sample thaw, because thiscauses RNA degradation.

Tip: This is only one example. Other kits can becombined to produce cDNA.

Tip: cDNA is much more stable than RNA and,depending on the downstream applications,might be a better option for long-term storage.RNA degradation can occur within a few weekseven if stored at 280 C.

94 N Weed Science 63, Special Issue 2015

targeted, or the sequence of gene(s) of interest is notknown in the species considered, then randomhexamer primers (i.e., a mixture of DNA moleculesconsisting of six randomly chosen nucleotides) orpoly-dT primers (i.e., a primer including a poly-Tsequence that is complementary to the poly-A tail ofmRNAs), or both, can be used for reversetranscription. Random hexamer primers are expect-ed to generate short cDNA fragments by hybridiz-ing randomly on their complementary sequences onmRNAs. Poly-dT primers are expected to generatecDNA fragments corresponding to the 39 end ofmRNAs. If a specific, known gene or region of agene is targeted, then gene-specific primers can beused for the reverse transcription.

PCR. The polymerase chain reaction (PCR) canmassively replicate a given DNA region (amplicon)from small or minute amounts of DNA. PCRenables easy and rapid gene sequencing. PCR is alsothe tool of choice for DNA-based mutationdiagnosis. A PCR reaction mix typically consistsof a buffer containing template DNA or cDNA,primers (see below), a thermostable DNA polymer-ase (e.g., Taq polymerase), and the four DNAnucleotides (dNTPs). PCRs are run as a successionof cycles, each with three steps, which are carriedout in a thermocycler. At each cycle, the quantity ofamplicon present in the reaction mix is theoreticallydoubled, until the dNTP stock is exhausted. Thethree steps in each cycle are:

1. Denaturation: the reaction mix is heated at 93 to95 C to dissociate double helix DNA into singlestranded DNA.

2. Annealing: the temperature is lowered to allowprimer hybridization to their target DNAsequences.

3. Extension: the temperature is increased to thereaction temperature optimal for DNA polymer-ase activity (68 to 72 C). The DNA polymerasesynthesizes a new DNA strand by adding dNTPsto form a sequence complementary to that of theDNA template strand, starting from the 39 end ofthe primers.

Primer Design. A primer is a short (10 to 30 bp)single strand of nucleic acid (oligonucleotide) thatserves as a starting point for DNA synthesis. A pairof primers is used in PCR to amplify the DNAregion flanked by the primers. During the PCR, theprimers hybridize with their target sequence on

DNA. Subsequent DNA synthesis occurs from the39 end of a primer using the DNA strand the primerhybridized with as a template.

For a pair of primers, one primer (‘‘forward’’primer) has the same sequence as a short region onthe coding strand of template DNA, and the other(‘‘reverse’’ primer) has the same sequence as a shortregion on the complementary strand of templateDNA. Forward primers hybridize to their comple-mentary strand, which is the noncoding strand oftemplate DNA, whereas reverse primers hybridize tothe coding strand of template DNA. The degree ofspecificity of a primer is directly linked to howclosely its sequence matches its target sequence. Thelast five bases at the primer 39 end are particularlycrucial for specificity.

If the sequence of the targeted gene is known inthe species of interest, primer design can be directlyperformed using software such as Primer3 (http://www.bioinformatics.nl/cgi-bin/primer3plus/primer3plus.cgi).Software programs such as these design and rankprimers based on their stability and specificity, takinginto account the PCR conditions and input data, such asexpected primer size range, expected amplicon size, andpossibility for the primers to hybridize together oroutside of the targeted region.

If the sequence of the targeted gene is not knownin the species of interest, then primer design can bedone based on the alignment of sequences of thehomologous gene from closely related species (ifavailable) as follows:

1. Obtain nucleotide sequences for homologues ofthe gene of interest that are available in theGenBank/EMBL database (http://www.ncbi.nlm.nih.gov/gquery/gquery.fcgi).

2. Align these sequences. Several tools are availablefor this purpose, including online tools (e.g.,http://multalin.toulouse.inra.fr/multalin/).

3. Search for DNA regions containing fully or veryhighly conserved sequences among species.

4. Design primers targeting these conserved regions.

Tip: The longer the primers, the more stable thehybridization to its target sequence. Optimalprimer length for basic PCR is generallyconsidered to be 20 to 25 nucleotides. The pHof the PCR mix and the temperature used forprimer hybridization in PCRs can influence thespecificity of primers.

Delye et al.: Methods for HR mechanisms N 95

If few DNA regions are highly conserved amongthe sequences aligned, there are several optionsregarding the variable nucleotides: (1) use thenucleotide most often present at a given position;(2) use inosine at variable positions (inosine formshydrogen bonds with all four natural DNA bases)(Ohtsuka et al. 1985); or (3) use a mixture ofoligonucleotides that vary at one or a small numberof nucleotides (often called degenerate primers).

Once the primers are designed, they can becustom synthesized for you by any one of severalcompanies (e.g., Integrated DNA Technologies,Coralville, IA; Life Technologies, Carlsbad, CA;Eurofins MWG Operon, Huntsville, AL).

Optimizing PCR. Several parameters can be adjustedto optimize PCR. Mostly, PCR is optimized byadjusting the composition of the reaction mix(including primer and DNA concentration) andthe program entered in the thermocycler. In thispaper, we do not detail PCR optimization becausethis has been done elsewhere (e.g., Roux 1995,2009). Our aim is to provide guidelines on how toset up a robust PCR assay.

PCR mix. There is a broad variety of commercialand in-lab mixes available for PCR. Optimizing aPCR mix can be a tedious issue. It is generally atime-saving approach to use a PCR mix that hasproved effective for a purpose similar to yours, andto optimize PCR cycling programs. The PCR mixprovided in Table 1 has proved robust and efficientfor PCR using DNA from different extractionprocedures and a variety of thermostable polymer-ases. (Delye et al. 2002)

PCR programs. In many instances, short durationscan be successfully used for denaturation, annealingand extension steps. This saves time and can sparethermocyclers. A typical PCR program is as follows:

1. Initial denaturation step (optional): 5 min, 95 C.2. Three-step cycles consisting of: denaturation (5 s

at 95 C), annealing (10 to 30 s at annealingtemperature), and extension (15 to 120 s at 72 C);repeated 30 to 40 times.

3. Final extension step (optional): 5 to 10 min at72 C.

The initial denaturation step can be used toensure complete DNA denaturation for difficultDNA targets (i.e., high guanine and cytosine [GC]contents) or for crude DNA extracts (e.g., Brutusextracts). Also, this step is generally mandatory ifyou are using an antibody-bound DNA polymerase(e.g., a ‘‘Hot Start’’ DNA polymerase; availablefrom numerous sources). For the following (cycled)

Tip: The nucleotides at the 39 end of the primersare crucial for primer specificity and should bepositioned on nucleotide positions very highlyconserved in the alignment. The last 39 nucleo-tide of the primer should not be positioned onthe third nucleotide of a codon, because thisnucleotide can vary among individuals in aspecies.

Tip: If DNA sequence alignments do not enableeasy identification of conserved regions, tryaligning protein sequences. Protein sequencesare often more conserved than DNA sequences;alignment of protein sequences allows you tofocus in on areas of the gene that are mostconserved.

Tip: Generally, ordering oligonucleotides usingthe default or least-expensive options in terms ofquantity and purity is sufficient for most PCRapplications.

Tip: Custom oligonucleotides are now veryinexpensive (if ordering just a pair of primers,the shipping cost might be more expensive thanthe oligonucleotides themselves). For this reason,and because primers predicted to work sometimesdo not, it often is advantageous to design andorder a few different primers and then determinewhich combinations work best.

Tip: PCR efficacy is dependent on the primerconcentration. It is often a good idea to test arange of primer concentrations when optimizinga PCR reaction.

Tip: Commercial kits exist that are designed tooptimize PCR efficacy and/or specificity.

96 N Weed Science 63, Special Issue 2015

denaturation steps, 5 s is generally sufficient. Shortdenaturation steps also preserve the activity of thepolymerase. For annealing, 10 to 30 s is generallysufficient. Try several annealing temperatures tooptimize the PCR. Generally, testing the annealingtemperature computed for the primers and temper-atures 5 C above and below this temperature will besufficient. Alternatively, test 55, 60, and 65 C.

For the extension step, a general rule is to use1 min extension time per 1,000 bp of the targetedamplicon. For amplicons longer than 2,000 to2,500 bases, dedicated types of thermostableDNA polymerases or mixtures of polymerasesdedicated to ‘‘long range’’ PCRs are commerciallyavailable.

Amount of DNA. The PCR method is extremelysensitive, requiring only a few DNA molecules to bepresent in the reaction mix to yield successfulamplification.

Table 1. Example mix for polymerase chain reaction (PCR), and stock solutions for preparing the mix.a,b

103 buffer (for 10 ml) 1.66 ml (NH4)2SO4 1 M [1.66 mM]6.7 ml Tris-HCl 1 M, pH 5 8.8 [670 mM]200 ml MgCl2 1 M [20 mM]Bring to 10 ml final volume with distilled waterSterilize with autoclaveAdd 70 ml b-mercaptoethanol 14.3 M [100 mM]Store at 220 C

23 stock solution (for 1 ml) 200 ml 103 buffer100 ml Brij 58 at 1% wt/vol (Sigma-Aldrich, P5884)c [0.1%]14 ml dNTP mixture (dATP, dCTP, dTTP, and dGTP each at

25 mM) [350 mM each dNTP]d

20 ml bovine serum albumin 20 mg ml21 [400 mg ml21]Bring to 1 ml final volume with distilled waterStore at 220 C

PCR mix (20 ml) 10 ml 23 stock solution5 ml primer solution 0.4 to 1.6 mM [0.1 to 0.4 mM]e

5 ml DNA0.2 to 0.5 units polymerase

a Use molecular biology-grade reagents.b Final concentration in the solution is given between brackets.c Source: Sigma-Aldrich, St. Louis, MO.d Abbreviations: dNTP, deoxynucleoside triphosphate; dATP, deoxyadenosine triphosphate; dCTP, deoxycytidine triphosphate,

dTTP, deoxythymidine triphosphate; dGTP, deoxyguanosine triphosate.e Range of primer concentrations that most often works.

Tip: Annealing temperature is an importantdeterminant for PCR success. Decreasing theannealing temperature might enhance PCRefficiency but increases the risk for nonspecificPCR amplification. Conversely, increasing theannealing temperature can decrease the PCRefficiency, but decreases the risk for nonspecificPCR amplification. The trick is to select atemperature that is a compromise betweenspecificity and efficacy of PCR amplification.

Tip: Annealing temperatures cannot exceed thetemperature used for the extension step (generally72 C). If your primers have an annealingtemperature close to the extension temperature,try two-step cycles: 5 s denaturation, directlyfollowed by extension.

Tip: Many papers describing PCR methods use afinal extension step at the end of the cyclingprogram, typically consisting of 10 min at 72 C.In many cases, this is not necessary.

Tip: Most thermocyclers provide the option ofholding the samples at 4 C after the PCRprogram. This is not necessary, because PCRmixes can safely remain at room temperature forseveral hours. Not using this option can also sparethermocyclers.

Delye et al.: Methods for HR mechanisms N 97

Designing a Proper PCR Experiment. The PCRmethod is extremely efficient in amplifying DNA.Thus extreme care must be taken to reduce thelikelihood for sample-to-sample contamination. Atypical PCR experiment should include: (1) theuntested samples to be amplified; (2) positivecontrols, i.e., samples for which successful amplifi-cation had been obtained consistently (to check forthe efficacy of PCR); and (3) negative controls, i.e.,samples without DNA (no amplification should beobtained for these samples, which check for DNAcontamination of the PCR reagents).

Gel Electrophoresis of PCR Products. Horizontalagarose gel electrophoresis is a simple method forthe separation of PCR products or amplicons basedon their size and charge. An electric field is applied tothe gel and induces the negatively charged DNAmolecules to migrate through an agarose matrixtowards the anode. Short, lightweight moleculesmigrate faster through the agarose matrix of the gel,and are therefore separated from longer, heaviermolecules. At the end of the electrophoresis, DNAmolecules are revealed using one of various DNAstaining reagents. The following materials are needed:

N Electrophoresis buffer: Tris Borate EDTA(TBE) buffer 0.53 (obtained by dilution of a103 stock solution with deionized water. Tomake 1 L 103 stock solution, dilute in 500 mldeionized water: 108 g Tris base, 55 g boricacid, and 20 ml 1 M EDTA; bring to 1 L final

volume with deionized water and sterilize in anautoclave).

N Loading buffer. To make 100 ml, mix 50 ml1 M EDTA pH 5 8, 1.5 g ficoll, 24 g urea,0.05 g bromophenol blue, and 0.05 g xylen–cyanol, then bring to 100 ml final volume withautoclaved deionized water.

N Agarose, molecular biology grade.N Microwave oven.N Horizontal electrophoresis system with power

source.N UV transilluminator and camera.N Ethidium bromide (can be purchased as a 10 mg

ml21 solution) or other DNA staining agent.

General steps for gel electrophoresis are as follows:

1. Weigh the agarose. See Table 2 to determine theamount of agarose. Place it in an Erlenmeyerflask with half the volume of 0.53 TBE bufferrequired for the gel. Do not close the flask.

2. Melt the agarose gel in the buffer by microwavingup to ebullition, then gently agitating the flask(careful, it is hot!) until obtaining a homogenous,translucent solution (no solid agarose grains mustbe seen).

3. Add the rest of the volume of 0.53 TBE bufferrequired for the gel. Homogenize by gentlyagitating.

4. Pour the gel mixture on its support with a comb(to form the loading wells). Wait for completesolidification.

5. Place the gel in the electrophoresis device. Coverwith 0.53 TBE buffer.

6. Add 1ml loading buffer per 20 ml PCR product toeach sample. Centrifuge briefly.

Tip: In case of DNA contamination in your PCRassay, it is often most effective to use new batchesof all reagents (PCR mixes, polymerase, primersand distilled water) than to try to identify thesource of contamination. Also, we recommendusing pipette tips with aerosol barriers to reducethe potential for sample contamination.

Tip: Ethidium bromide is considered a mutagenand must be handled with care. Dedicated, clearlyidentified areas must be set up for manipulationsinvolving ethidium bromide. There are alterna-tives to ethidium bromide that are advertised asbeing safer (e.g., SYBRgreen-based dyes availablefrom several sources), but ethidium bromide isstill most broadly used.

Tip: Brutus DNA solutions are not quantified. Testa range of 5-fold or 10-fold dilutions to identify thebest working dilution in your lab. Generally, 25-fold to 100-fold dilutions are adequate. If you useda similar amount of plant tissues for extraction forall your samples, the dilution test only needs to beperformed on a few samples. Then, use the samedilution for all samples.

Tip: After the PCR is complete, the samples areready for electrophoretic analysis. Samples can beanalysed immediately after PCR, or stored at 4 Cfor later analysis (220 C for long-term storage).

98 N Weed Science 63, Special Issue 2015

7. Load 1 to 15 ml of each sample on the gel. Loadinga commercially available DNA ladder on one gellane is generally useful for amplicon size estimation.

8. Run electrophoresis at 100V for 20 min to45 min, depending on the size of the gel and thedegree of separation expected (the longer themigration, the better the separation).

Amplicons are generally visualized with a UVtransilluminator after dipping the gel into anethidium bromide solution (0.5 mg ml21). Alterna-tively, ethidium bromide can be added to the gel(0.5 mg ml21) just prior to casting it. Ethidiumbromide renders DNA fluorescent under UV lightby intercalating between the DNA bases. OtherDNA-staining reagents can also be used. Use acamera to save a photograph of the gel.

DNA Sequencing

The easiest way to obtain sequence data for yourgene of interest is to directly use the PCR amplicon

as a template for Sanger sequencing (Sanger et al.1977). Sanger sequencing consists of two steps: (1)sequencing reaction using a DNA fragment as atemplate, and special nucleotides that terminate andlabel the amplified molecules; and (2) visualizationof the labelled molecules by capillary electrophore-sis. This second step requires expensive equipment,and so is carried out by, for example, a sequencingcompany or a centralized campus facility. The firststep can be contracted along with the second step,or can be done in-lab using for example, theApplied Biosystems BigDye Terminator CycleSequencing Kit (Life Technologies).

It is generally advisable to first purify the PCRamplicon that will be used as the template prior tothe sequencing reaction. There is a range of kitsavailable to purify PCR amplicons of different sizes.Alternatively, sequencing companies propose am-plicon purification as part of their sequencingservice. If amplicons other than the one of expectedsize are observed during gel electrophoresis (even ifminor in abundance), then the amplicon of interestshould be isolated by gel purification, because thesequence of all amplicons in the PCR mix willsuperimpose to generate the consensus sequenceobserved. Sequencing different amplicons in thesame Sanger reaction generally yields illegiblesequences. To ‘‘gel-purify’’ the amplicon:

Tip: To avoid transferring amplicons from onePCR reaction to another, the loading buffer canbe deposited at the top of the tube. It will blendwith the PCR mix during the centrifugation step.If this is done carefully, there is no need tochange pipette tips between samples. Alternately,there are commercial PCR mixes incorporating aloading buffer, which allows for sample loadingon electrophoresis gels after PCR completionwithout further workup.

Table 2. Concentration of agarose to be used for gelelectrophoresis according to the expected size in base pairs (bp)of your amplicon of interest.

Amplicon size

Agarose concentration(g agarose 100 ml21 0.53 TBE

[tris borate EDTA] buffer)

Above 1,000 bp 0.8600 to 1,000 bp 1200 to 600 bp 2, 200 bp 3 to 3.5

Tip: Be sure to check the negative controls, whichshould not show any trace of an amplicon. If theydo, all the PCR experiment results are invalidat-ed. Do not be ashamed: PCR is a very sensitivetechnique, and contamination can happen evenwith experienced molecular biologists.

Tip: Classical thermostable DNA polymerases donot have a proofreading activity. Thus, erroneousnucleotides are incorporated in some of the DNAmolecules generated in the PCR mixes (Tindalland Kunkel 1988). It is important to realize thatthe amplicon used for sequencing is actually apopulation of individual DNA molecules. Whenan amplicon is directly sequenced, the sequenceobtained is a consensus of the sequences of theDNA molecules generated by PCR and present inthe PCR mix. If misincorporation occurs duringthe first PCR cycles, erroneous DNA moleculeswill represent a substantial part of the DNAmolecules in the amplicon. Thus, erroneousnucleotides will show in the consensus sequence.One option to lower the risk of sequencing errors,is to perform sequencing from at least three tofive different PCR mixes (i.e., perform three tofive identical PCR reactions in separate tubes,pool them, and sequence the pool). Alternatively,use a thermostable polymerase with proofreadingactivity.

Delye et al.: Methods for HR mechanisms N 99

1. Pour an agarose gel with wide combs and leave atleast one empty lane between samples.

2. Load the entire PCR contents into the wells andrun the gel for 30 to 45 min (allow enough timeto fully separate the bands).

3. Stain the gel with ethidium bromide.4. Place the gel on a UV transilluminator (wear eye

and skin protection to prevent direct exposure toUV light), turn on the UV lamp, and carefullyexcise the band with a razor or bistoury blade.The amplicon can then be purified with acommercial gel isolation kit (e.g., QIAquickGel Extraction Kit, Qiagen) and can subsequent-ly serve as the template for sequencing.

For the sequencing reaction, the same PCRprimers used to generate the amplicon can be used.Only one primer is used in each sequencingreaction. It is generally recommended to performtwo sequencing reactions, one with each primer.Sequencing from the forward primer will yield thesequence of the coding DNA strand, and sequenc-ing from the reverse primer will yield the sequenceof the complementary DNA strand. Thus, theamplicon will be sequenced on both strands. Theforward and reverse sequences are then compared todetect and eliminate sequencing errors. The forwardsequence of the amplicon can be obtained from thereverse sequence by reverse complementing. Reversecomplementing and alignment of the sequences ofboth strands can be performed using sequencingsoftware (see below). Currently, Sanger sequencingplatforms yield highly accurate sequence data of upto about 1,000 nucleotides in length (Metzker2005). To sequence larger amplicons, additionalsequencing reactions are performed using primersspaced 500 to 800 nucleotides apart (the partialsequences must overlap to allow assembly into thefull sequence of the amplicon).

There are many programs (freeware and non-freeware) available for analysis of sequence data, soit is really up to the user’s personal preference as towhich one to use. A simple and free general-purposeprogram that can be routinely used to manipulate,edit, align, and assemble sequences is BioEdit (http://www.mbio.ncsu.edu/bioedit/bioedit.html). An exten-sive compilation of software programs for a variety ofevolutionary and molecular studies can be found atJoe Felsenstein’s website (http://evolution.genetics.washington.edu/phylip/software.html).

The sequences of the DNA molecules in the PCRproduct can differ due to biological reasons (i.e.,nucleotide polymorphism). For example, polyploid

or heterozygous plants can contain different genecopies, and/or different alleles of a gene, whichexhibit different nucleotide sequences. Sangersequencing yields a single sequence resulting fromthe superimposition of the sequences of thedifferent genes or alleles. This sequence thusdisplays multiple nucleotides at the variablenucleotide positions. If there is an interest in theprecise sequence of individual genes or alleles, it isnecessary to separate individual DNA moleculesfrom the amplicon mixture by cloning, and thenindividually sequence a few randomly selectedones. In this case, each sequencing reaction willyield a sequence derived from a single DNAmolecule.

Cloning of PCR products involves ligating DNAfragments into plasmid vectors carrying an antibi-otic resistance gene, and then transferring theserecombinant vectors into bacteria cells (generallyEscherichia coli cells). Electroporation (Dower et al.1988) is most effective for this purpose, butchemical or heat shock-based transfer proceduresthat do not require an electroporation device arecurrently implemented in commercial kits. Trans-formed cells are selectively identified from non-transformed cells on a solid agar medium contain-ing the appropriate antibiotic. Individual colonies of

Tip: Cloning involves a substantial amount ofwork. Before embarking into it, be sure it is reallynecessary for your purpose.

Tip: Any individual DNA molecule obtained in aPCR may contain one or more errors introducedby PCR. Thus, when sequencing cloned prod-ucts, several clones should be individuallysequenced. Sequencing three to five clones pergene or allele is usually sufficient, but keep inmind that you cannot identify which moleculederives from which gene or allele prior tosequencing.

Tip: A proofreading polymerase can be used forPCR when downstream applications call forcloning, to reduce sequencing errors.

100 N Weed Science 63, Special Issue 2015

transformed cells are then grown in liquid medium,whereupon the recombinant plasmid vector isamplified. The plasmid is then purified from thecells and used for DNA sequencing or otherdownstream applications. Several kits are availablefor cloning PCR products, including the TOPO TACloning Kit (Life Technologies), the pGEM-T EasyVector System (Promega, Madison, WI), or thepDrive system (Qiagen).

The approaches discussed above generally areused to obtain sequence information from one or afew genes. For large-scale gene sequencing projects,or to sequence one or a few amplicons in very largenumbers of individual plants, using one of severalnext-generation sequencing (NGS) techniques isadvisable. Current NGS technologies includeRoche/454 GSFLX pyrosequencing (454 LifeSciences, Branford, CT), Illumina (San Diego,CA), and SOLiD (Life Technologies) (Thudi etal. 2012). NGS technology can generate enormousamounts of sequence data, but up-front costs ofsample preparation, sequencing, and data processingare high. NGS technologies are thus of interest ifsequencing a few genes or amplicons in a massivenumber of plants or of weed populations isrequired. Another use of NGS techniques is to gainaccess to the sequence of genes of interest in a weedspecies where no or few genomic data are available(Vigueira et al. 2013). This can be achieved bysequencing and assembling the transcriptome or adraft of the genome of the species of interest (e.g.,Riggins et al. 2010). Sequencing and assembly canbe carried out by a sequencing company orcentralized campus facilities.

The recent rise of the NGS technologies has alsoopened the possibility to study weed adaptive traitswith a complex genetic determinism involving bothstructural and regulatory mutations (Vigueira et al.2013), such as NTSR (Delye 2013; Yuan et al.2007). An approach of choice for this purpose isquantitative transcriptome sequencing, also knownas RNA sequencing or RNAseq. It implies RNAsamples are converted to cDNA. Millions of shortsequences are then generated using NGS. Currently,the Illumina NGS technology is most adequate forquantitative transcriptome sequencing, because ofthe massive number of short reads generated thatenable to maximize transcriptome coverage even forrare transcripts (Ness et al. 2011). Illumina shortreads thus simultaneously contain transcriptomesequence information and a measure of geneexpression (Ozsolak and Milos 2010). Yet, Illuminareads are short (currently 100 to 200 nucleotides),

which renders de novo transcriptome assemblytricky (Ness et al. 2011). The short sequence readscan be mapped against a reference genome sequenceor assembled de novo to produce transcriptomeresources, including the structure and abundance ofeach transcript (see Martin and Wang 2011 andWard et al. 2012 for reviews). Guidelines fortranscriptome-based analysis of NTSR in weedshave been proposed (Delye 2013). Yet, thisapproach is still in its infancy (Delye et al. 2013;Vigueira et al. 2013), and only a few attempts havebeen made to date (e.g., Gaines et al. 2013; Gardinet al. 2013).

Detecting Mutations

Amplification of a gene of interest, followed byDNA sequencing, can be used to detect DNApolymorphisms that can confer herbicide resistance.However, when a particular DNA polymorphism(mutation) is known to cause resistance, theresearcher might want to screen plants only forthe presence of this mutation (mutation genotyp-ing). For this purpose, methods are available thatare simpler, faster, and cheaper then DNAsequencing.

Easy, Robust Method: dCAPS, CAPS. The mostrobust, in-lab PCR-based techniques that can beused to detect DNA mutations are based on thedifferential cleaving by restriction enzymes betweenamplicons carrying and not carrying the targetedmutation(s). Restriction enzymes recognise a spe-cific DNA sequence (recognition site) and hydrolyze(‘‘cut’’) the DNA molecule at, or close to, theirrecognition sites.

The basic approach involves first performing aPCR to amplify a gene region of interest, incubating(digesting) the amplicon with the appropriaterestriction enzyme, and analyzing the PCR productsby gel electrophoresis. Observing two small molec-ular-weight DNA fragments in place of one largermolecular-weight fragment indicates that the frag-ment has been digested.

The Cleaved Amplified Polymorphic Sequence(CAPS) technique uses a restriction site that isnaturally present in one type of amplicon (mutantor wild-type) and absent in the other due to theresistance-endowing mutation. Thus, CAPS canonly be implemented if the specific mutation soughteither creates or abolishes a restriction site. Becausethere are hundreds of commercially availablerestriction enzymes with unique recognition sites,

Delye et al.: Methods for HR mechanisms N 101

the CAPS approach might work in some cases.However, the mutation of interest most often willnot create a restriction site polymorphism; there-fore, the derived Cleaved Amplified PolymorphicSequence (dCAPS) technique has been developed.dCAPS uses PCR to create a restriction enzymerecognition site in a sequence where none exists.The dCAPS primer is located close to the mutationof interest, and contains one or more mismatchednucleotides so as to create a restriction enzymerecognition site encompassing the mutation ofinterest (Neff et al. 1998). The (d)CAPS techniquehas been successfully used to detect mutationsconferring herbicide resistance, including in theACCase and ALS genes (e.g., Delye and Boucan-saud 2008; Delye et al. 2011; Kaundun andWindass 2006).

Obviously, to use the dCAPS method, thenucleotide sequences of wild-type and mutantalleles should be known. Ideally, development andtesting of the assay also will require DNA samplesfrom both wild-type and mutant plants. Heterozy-gous plants should be available to verify thereliability of the assay.

Primer Design for CAPS. Primers must flank themutation (and thus the restriction site) and yield anamplicon that can be easily discriminated from theresulting digestion products by gel electrophoresis ifthe enzyme recognition site is present. For example,you can design primers to amplify a 300 bpfragment with the restriction site near the middle.

Primer Design for dCAPS. Here, primer pairs consistof one ‘‘dCAPS’’ primer that creates the restrictionsite in the amplicon, and of one ‘‘conventional’’PCR primer. dCAPS primer sequences targeting thecodon of interest can be generated using the freesoftware dCAPS Finder 2.0 (http://helix.wustl.edu/dcaps/dcaps.html) (Neff et al. 2002). This softwareprovides a range of dCAPS primer and restrictionenzyme combinations to be used in dCAPS assays.dCAPS primers are then selected on the basis of: (1)restriction enzyme availability (and cost), and (2)absence of mismatch at the last 39 nucleotideposition in the primer sequence, which introduces arisk for primer inefficiency.

The following rules should be followed whendesigning a dCAPS assay:

1. Design assays so that only amplicons containingwild-type codons are cut (provides a control forrestriction enzyme activity).

2. Use long dCAPS primers (ca. 40 nucleotides) toenable easy discrimination of undigested anddigested (about 40 bp removed) amplicons bystandard agarose gel electrophoresis. For thisreason, design the ‘‘conventional’’ primer so thatthe amplicon generated is 150 to 350 nucleotideslong.

3. Ensure that all nucleotides in the restrictionenzyme recognition site are exclusively located inthe dCAPS primer sequence and in the part ofthe targeted codon where any variation wouldcause amino acid substitution, so as to detectonly nucleotide polymorphisms endowing resis-tance (to avoid false positive detection).

Tip: If heterozygous plants are not available, onecan mix equal amounts of DNA from wild-typeand mutant plants to obtain a ‘‘mock’’ heterozy-gote.

Tip: Keep in mind that the same restriction sitemight be present elsewhere in the gene. If it ispresent more than once in your amplicon, it canconfound interpretation of the digestion produc-tions.

Tip: Design the primers so that they amplify arelatively small fragment (, 400 bp).

Tip: If the target gene has introns and genomicDNA is used as a PCR template (this is generallymore convenient than using cDNA), design theprimers to hybridize within a single exon or besure to use the genomic DNA sequence whendesigning primers.

Tip: Setting dCAPS Finder ‘‘mismatch’’ optionto 1 might not yield satisfactory primers. Theoption can be increased to 3 if dCAPS primerscontaining no more than one mismatch in thelast three to five 39 nucleotides are selected.

102 N Weed Science 63, Special Issue 2015

4. When several restriction enzymes can be used,select one with no other recognition site in theamplicon. Several freeware programs are availableto identify restriction sites in DNA sequence (e.g.,Webcutter, http://rna.lundberg.gu.se/cutter2/).Alternatively, position the non-dCAPS primer toexclude any other enzyme recognition site in theamplicon.

When the primers are designed, PCR optimiza-tion is performed as described above.

Restriction Digests. The composition of PCR mixesis usually compatible with reaction mixes forrestriction enzymes. Thus, no purification of thePCR product is necessary prior to digestion byrestriction enzymes. Digestions are performed in10 ml volumes using the following general steps:

1. Put 1 to 5 ml PCR mix in tubes. The amount ofPCR mix used for digestion depends on theintensity of the amplicons observed.

2. Add 5 units of restriction enzyme.3. Add 0.5 ml of 103 enzyme reaction buffer

(supplied with the enzyme).4. Fit the total volume to 10 ml with distilled water.5. Incubate digestion mixes 3 h at the temperature

optimal for the activity of the restriction enzymeselected.

6. Visualize digestion patterns by electrophoresis(see above).

7. Interpret the results. See Figure 1 for an exampleof dCAPS patterns.

Another Simple Method: Allele-Specific PCR.Allele-specific PCR utilizes a pair of primers inwhich one primer (allele-specific primer) selectivelybinds to only one allele (Sommer et al. 1992). Allelespecificity of the PCR is due to the presence of anucleotide exactly matching one of the alleles butnot the other at the 39 end of the allele-specificprimer. At a specific annealing temperature, a 39mismatch does not prime in a PCR (Sommer et al.1992). The presence or absence of one given

Tip: When dCAPS primers are designed followingthese rules, it is not necessary that mutant plants beavailable. For genes where resistance-endowingmutations are well known (e.g., ACCase, ALS,EPSPS, and psbA; Burgos et al. 2013), one candesign assays to detect these mutations beforeresistance has evolved in a weed species.

Tip: dCAPS assays are based on the absence ofdigestion of amplicons carrying the mutation(s)targeted. An excess of amplicon must be avoided,because it can create false positive detection due topartial digestion of excessively abundant amplicons.

Tip: Run samples of undigested PCR product onthe gel as a negative control for digestion (allowsone to more easily discriminate digested andundigested fragments).

Figure 1. Band patterns that can be observed after agarose gelelectrophoresis of the digestions of PCR mixes using a (d)CAPSassay digesting wild-type amplicons. Interpretation is as follows:(A) expected patterns for a theoretical dCAPS assay. Lane 1,undigested amplicon: homozygous mutant; Lane 2, fully digestedamplicon: homozygous wild-type; Lane 3, mixture of digested(two bottom fragments) and undigested (top fragment)amplicons with similar intensity on the gel: heterozygousmutant. Note that, although not depicted here, a faint trace ofundigested amplicon can be observed in homozygous wild-typesamples if amplicon material is very abundant (restrictionenzyme can be saturated by its substrate). (B) example of dCAPSpatterns obtained with a dCAPS assay targeting ALS codon 197in Lolium spp. Lanes h, heterozygous mutants; lanes W,homozygous wild-types; lanes M, homozygous mutants. LaneH, water control (no DNA added to the PCR mix).

Tip: Duration of digestion can be cut down to15 min using ‘‘fast digestion’’ mixes, if availablefor the enzyme used.

Delye et al.: Methods for HR mechanisms N 103

nucleotide at the targeted position will thus result inthe presence or absence of an amplicon, respectively.

Standard allele-specific PCR using two primersresults in the detection of one allele by yielding asingle amplicon when this allele is present, and noamplicon when it is absent (e.g., Wagner et al.2002). The lack of an amplicon can also occur dueto PCR failure (e.g., due to poor-quality DNAtemplate) and, therefore, standard allele-specificPCR is prone to false negatives.

An improvement of the technique is bidirectionalallele-specific PCR that uses four primers. Bidirec-tional allele-specific PCR enables concurrent detec-tion of the presence or absence of two distinct allelesin a single PCR (e.g., Delye et al. 2002; Kaundunet al. 2011). Two primers are classical forward andreverse primers flanking the mutation site. The othertwo are allele-specific primers. In a given assay, oneforward allele-specific primer targets one allele, andone reverse allele-specific primer targets the otherallele. Using bidirectional allele-specific PCR, threeamplicons sizes are obtained: one is specific for oneallele (e.g., wild-type), the second is specific for theother allele (e.g., mutant), and the third one isamplified in all samples (positive internal control). Asfor (d)CAPS markers, sequences of the alleles ofinterest must be known for primer design. Also, to setup the assay, it is mandatory to have biologicalsamples with homozygous wild-type and mutantgenotypes, and heterozygous mutants.

The following rules should be followed whendesigning a bidirectional allele-specific PCR assay:

1. One allele-specific primer is designed on thecoding DNA strand, the second, on the noncod-ing strand.

2. The sizes of the amplicons specific to each alleleshould easily be separated on agarose gels.

3. Allele-specific primers to be used in the sameassay should have melting temperature as close aspossible.

4. The last nucleotide in the sequence of each allele-specific PCR primers must be the nucleotidespecific to the allele targeted.

5. Test a range of annealing temperatures until onetemperature is found that allows specific ampli-fication.

Interpretation of bidirectional allele-specific PCRpatterns is illustrated in Figure 2.

Bidirectional allele-specific PCR is sensitive tothe annealing temperature used and to the compo-sition of the PCR mix (especially to the pH).Because no digestion step is involved, bidirectionalallele-specific PCR is faster than (d)CAPS, but canrequire a fair amount of troubleshooting to developa robust assay. Also, because the diagnostic stepoccurs during the PCR, variations among templatesamples (DNA quantity, quality, and purity) can

Tip: The specificity of an allele-specific primercan be improved by creating an additionalmismatch at the penultimate 39 nucleotideposition in the primer. Kwok et al. (1994) andPettersson et al. (2003) provide guidelines fordesigning allele-specific primers.

Figure 2. Bidirectional allele-specific PCR patterns that can beobtained after agarose gel electrophoresis for an assay designed togenerate a smaller allele-specific amplicon from the mutant allele.(A) expected patterns for a theoretical bidirectional allele-specificPCR assay. Up to three amplicons are generated: ‘‘a,’’ positiveinternal control (should always be observed); ‘‘b,’’ wild-typeallele; ‘‘c,’’ mutant allele. Interpretation is as follows: Lane 1,occurrence of the mutant allele; Lane 2: occurrence of the wild-type allele; Lane 3, heterozygous mutant. If only two differentnucleotides are known to be present at the targeted position,lanes 1 and 2 can be interpreted as homozygous mutant andwild-type, respectively. If nucleotide(s) different from the twonucleotides targeted by the assay is (are) present on all genecopies at the targeted position, none of the allele-specific primerswill hybridize, and pattern 4 will be obtained. In this case,sequencing fragment ‘‘a’’ will allow identification of the newallele. (B) example of patterns obtained with a bidirectionalallele-specific PCR assay targeting the first nucleotide at ACCasecodon 1781 in Alopecurus myosuroides. There are four possiblenucleotide variants at this position: A (if the codon encodes anisoleucine residue; wild-type allele), C or T (if the codon encodesa leucine residue; mutant alleles), and G (if the codon encodes avaline residue; mutant allele). The allele-specific assay is designedto detect the A and T variants. Lanes T, occurrence of the Tallele; lanes A, occurrence of the A allele; lane AT, heterozygousmutant; lane O, plant not carrying A or T alleles (homozygousmutant for the C allele). Lane H, water control (no DNA addedto the PCR mix).

104 N Weed Science 63, Special Issue 2015

decrease the robustness of the assay. In contrast,the diagnostic step of the (d)CAPS assay is therestriction enzyme digest, which is a much lessfinicky reaction than PCR. Bidirectional allele-specific PCR allows for detection of only two allelesper assay, but you know precisely which allele(s) is(are) detected. The (d)CAPS assay can be used todetect several mutations at the same nucleotide oradjacent nucleotides (on the basis of the digestion ofthe wild-type amplicons vs. absence of digestion ofamplicons carrying any mutation disrupting therestriction enzyme recognition site). In cases wereseveral mutations are possible at the targetedposition, as observed for instance in ALS or ACCase(Beckie and Tardif 2012; Tranel and Wright 2002),(d)CAPS does not enable identification of thespecific mutation disrupting the restriction enzymerecognition site. Both allele-specific PCR and(d)CAPS assays (as well as other assays not discussedhere) can and have been used to detect mutationsconferring herbicide resistance. Choosing betweendCAPS and bidirectional allele-specific PCR there-fore depends on the planned study. If severalmutations are present at the nucleotide positions ofinterest, or if the robustness of a bidirectional allele-specific PCR becomes a concern, then (d)CAPS isthe assay of choice.

‘‘High-Tech’’ Methods. Other PCR-based tech-niques can be used to detect DNA mutations. Theyrequire expensive equipment and reagents, and arebest suited for very high-throughput analyses(hundreds or thousands of samples). Three exam-ples of high-tech methods are TaqMan, MultiplexSNaPshot (both from Life Technologies) andScorpions (Sigma-Aldrich, St. Louis, MO).

TaqMan uses a pair of PCR primers flanking themutation site, and a short DNA probe carrying afluorophore (reporter dye) at the 59 end and aquencher at the 39 end. During the PCR reaction,the PCR primers and the TaqMan probe annealsimultaneously on the DNA sequence. During thePCR elongation step, exonuclease activity of thethermostable DNA polymerase degrades the probe,which releases the reporter dye from the quencher,thereby producing fluorescence. The TaqMantechnology can combine two probes with twodifferent fluorochromes, making it possible todetect two different alleles. TaqMan assays havebeen developed to genotype ALS (e.g., Warwick etal. 2008) or ACCase (e.g., Delye et al. 2010).

Scorpions uses a probe directly carrying afluorophore that has a specific target sequence for

PCR. This system has the advantage over systemssuch as TaqMan in that no separate probe carrying afluorophore is required to bind to the amplifiedtarget, making detection both faster and moreefficient. Assays combining Scorpions, allele-specificPCR, and quantitative PCR technology have beendeveloped to quantify mutant ACCase alleles inbulk samplings of Lolium spp. (Kaundun et al.2006).

Multiplex SNaPshot can detect mutations at upto 10 sites in a single assay. It is a combination ofPCR and single-base sequencing. The DNA regionscarrying a mutation are amplified by PCR. DNAoligonucleotide probes anneal to a target sequenceimmediately adjacent to the variable nucleotidepositions. A DNA polymerase then extends theprobe by incorporating one dye-labelled nucleotidecorresponding to the nucleotide present at thevariable position. Thus, the size of the DNAfragment obtained for each mutation site targetedis the size of the probe plus one fluorescent base, forwhich the color depends on the nucleotideincorporated. The mixture of probes with incorpo-rated dye-labelled nucleotides is then separated on asequencer, and the nucleotide present at eachposition investigated is detected based on the sizeof the corresponding fragment and the wavelengthof the incorporated dye-labelled nucleotide. SNaP-shot assays have been developed to genotypeACCase (Alarcon-Reverte et al. 2013).

Detecting Changes in Gene Expression

Quantitative PCR (qPCR) is a versatile an-alytical tool in which a fluorescent dye is used tomonitor the amount of PCR product produced inreal time during each PCR cycle. There are twobasic categories of detection dye chemistries:nonspecific dyes that intercalate with any double-stranded DNA fragment, and target-specific dyesthat utilize fluorescent probes and/or primers(Nolan et al. 2006). Nonspecific dyes, such asSybr Green and EvaGreen, are typically less costlyand simpler to use than target-specific dyes, butthey do have one important drawback: any form ofdouble-stranded DNA, including nontemplatefragments and primer-dimers, will be detected bynonspecific dyes. However, these unwanted prod-ucts can readily be detected using a melting curveanalysis and eliminated by assay optimization and/or redesigning the primers.

Herbicide resistance caused by an increase in theamount of a protein can be due to a change in the

Delye et al.: Methods for HR mechanisms N 105

regulation of the corresponding gene or to anincrease in the number of genomic copies of thisgene. Both origins of resistance can be detected bytechniques based on qPCR. Although qPCR ispredominately used for gene expression analyses withcDNA as the template (discussed in the next section),it is also compatible with genomic DNA. Thus, wefirst focus on the use of qPCR for detecting geneamplification, which was associated with glyphosateresistance initially in Palmer amaranth (Amaranthuspalmeri S. Wats.) (Gaines et al. 2010).

Testing for Gene Amplification Using qPCR.The basic qPCR experiment for testing geneamplification involves calculating the relativequantity of a target gene based on an endogenouscontrol or reference gene. Ideally, the reference geneshould be single-copy in all individuals. Examplereference genes that have been used in studies ofglyphosate-resistant weeds include ALS (Gaines etal. 2010) and carbamoylphosphate synthetase(Tranel et al. 2011) for Amaranthus spp., andcinnamoyl-CoA reductase for Italian ryegrass[Lolium perenne L.ssp. multiflorum (Lam.) Husnot](Salas et al. 2012).

Primer Design. The same basic principles of primerdesign discussed for regular PCR also apply toqPCR, with only slight amendments. Both forwardand reverse qPCR primers should be designed toanneal at 60 to 64 C and within 2 C of each other.The amplicon should also be 70 to 300 bp inlength, which is optimum for the shortened cyclingconditions of a typical qPCR experiment. Anotherconsideration when designing qPCR primers for agene amplification analysis is the region of the genein which the primers are anchored. Ideally, theprimers should target a region of the gene showinglittle nucleotide polymorphism to maximize assaysensitivity and reliability. Exon–intron junctions,introns, and 59 untranslated regions tend to be morevariable across genes and species, so these regionsshould be avoided for a qPCR assays. In any case,

several different primer sets should be evaluatedexperimentally before large-scale screening. Inte-grated DNA Technologies (www.idtdna.com) offersa free online qPCR primer design tool along with adownloadable qPCR Application Guide with addi-tional background information and useful tips onprimer design and troubleshooting.

A qPCR experiment for gene amplificationshould include positive and negative controls(samples with and without gene amplification)and no-template controls. Three technical replicates(i.e., replicates of the same biological sample)should be run and the resultant quantification cycle(Cq) values averaged. Plate-to-plate variation can beassessed by including interplate control samples(i.e., including one or more common samples oneach plate). Individuals from multiple populationscould also be included to evaluate background genelevels in the species.

Absolute vs. Relative (Comparative Cq Method)Quantitation. Before beginning the qPCR experi-ment, one must decide how the data are to beanalyzed. The results of your assay can be analyzedeither by absolute or relative quantitation methods.Absolute quantitation involves determining quanti-ties of unknown samples from a standard curve.This method can be useful if one wants to know theexact copy number of a particular gene of interest ina particular individual. Although this methodresults in a high level of accuracy, a standard curveof known quantities must be performed for eachexperimental run, which requires more reagents andtakes up space on the plate. Alternatively, relativequantitation using the Comparative Cq Method(Pfaffl 2001) does not require standard curves foreach plate as long as the PCR efficiencies of thetarget and control genes are essentially equivalent.To use this method, a validation experiment (Livakand Schmittgen 2001) is first performed byestablishing standard curves for the target andendogenous control genes from the same biologicalsample. The standard curves are done using five tosix serial dilutions of genomic DNA (gDNA) foreach gene with three replicates for each dilutionpoint. Resultant Cq values are averaged for eachinput amount and DCqs calculated (Cq target 2

Tip: The Cq value refers to the PCR cycle duringwhich fluorescence increases above some thresh-old level. It also is referred to as a Ct (thresholdcycle) value.

Tip: Several candidate reference genes should beevaluated experimentally—rather than just pick-ing one from the literature—to confirm suitabil-ity in a particular plant species. General guide-lines for selecting a reference gene are provided ina subsequent section.

106 N Weed Science 63, Special Issue 2015

Cq control). The log input of the gDNA is thenplotted against the DCqs. To pass the test, theabsolute value of the slope must be , 0.1.

qPCR Protocol for Detection of Gene Amplification.This protocol was originally designed to detectEPSPS gene amplification in Amaranthus spp. withthe 7900-HT detection system (Life Technologies),which uses 384-well plates. It can easily be adaptedfor 96-well plates and other detectors, and for othergenes/species.

1. Dilute all genomic DNA samples to 10 ng ml21.Include positive and negative control samples(resistant and sensitive plants) and no-templatecontrols (NTC) for each gene.

2. Prepare a dilution series to evaluate primerefficiency and determine the dynamic range ofthe assay. A 5-fold serial dilution, such as 13,0.23, 0.043, 0.0083, 0.00163, is sufficientand should be prepared for the target and controlgenes using one sample template. The properrange must be experimentally determined foreach assay.

3. Prepare a master mix for each gene. Account forall samples and replicates in calculating themaster mix plus some extra (10%) for pipettingvariation. Our calculations are based on a 10 mlfinal volume (9 ml mix + 1 ml template) for a 384-well plate. For example, a master mix for onegene for 10 samples would be calculated as shownin Table 3.

4. Pipette 9 ml mix in each well, then add 1 ml DNAindividually.

5. Gently tap the plate on the benchtop to settle themix in the bottom of the wells and overlay withABI Prism optical adhesive cover (#4311971).Spin the plate in a centrifuge for 2 min at roomtemperature before loading into the 7900 HT.

Data Analysis. Following the run, the data are storedin a Sequence Detection System (SDS) file (formatfor the 7900-HT). First, open the file and verify thatthe baseline and threshold were generated correctly.These values will likely be set automatically by thesoftware (default settings) and no adjustments mightbe necessary. Otherwise, the threshold can be adjustedmanually and should be within the geometric phase ofthe PCR amplification curve and above the baseline.Second, visually inspect each well and remove orexclude those with poor amplification. Replicates ofthe samples should have similar Cq values, otherwiseremove outliers before analysis. The dissociationcurves (for SYBR Green assays only) will help identifywells that should be removed, including empty wellsand wells with nonspecific amplification products.Primer–dimer peaks are common in the no-templatecontrols, but should disappear in the standard curveand unknown samples. If you ran a standard curve,make sure the slope is around 23.3 (indicating thatyour assay is 100% efficient) and the correlationcoefficient (R2) is . 0.98. Variation in the data can be

Table 3. Preparation of master mix for each gene in aquantitative polymerase chain reaction (qPCR) assay. Volumesare based on a 10-ml reaction using 1 ml of template DNA.

ReagentAmount forone reaction

Amount for threetriplicates each of

10 samplesa

---------------------------ml --------------------------

EvaGreen 5.0 165.0forward primer (0.2 to 0.4 mM) 0.3 9.9reverse primer (0.2 to 0.4 mM) 0.3 9.9water 3.4 112.2Total 9.0 297.0

a Includes 10% extra to allow for pipetting error.

Tip: The starting DNA concentration is notcritical, as long as the Cq values of the samplesare obtained within the dilution series range. Forexample, most of our DNA yields from CTABaverage 200 to 400 ng ml21, so our dilution seriesstarts with this as the upper range (basicallyundiluted stock) and is serially diluted five times(200 ng ml21 (13), 40 ng, 8 ng, 1.6 ng, 0.32 ng)to produce a standard curve.

Tip: PCR efficiency is satisfactory if the slopes of thedilution curves are close to 23.3 (+/2 10%). Slopeshigher or lower than this range indicate the needfor assay optimization. Look for an R2 value. 98%. Percentage efficiency (E) is calculated byE 5 [1021/slope2 1]*100 (Radstrom et al. 2003).

Tip: Thaw all samples and reagents on ice.EvaGreen is light-sensitive, so handle according-ly. Master mix reagents should be mixedthoroughly but not vortexed. Accurate pipettingis crucial with such small volumes (an electronicpipette is helpful, but not necessary).

Delye et al.: Methods for HR mechanisms N 107

caused by a variety of factors, but the more commonsources are imprecise pipetting, incomplete mixing ofreagents, poor quality of DNA/RNA template,improper threshold or baseline setting, or improperhandling of the qPCR plate. Low precision also can becaused by poorly designed primers or inherentbiological properties of the genes under investigation.After quality checking the results, the data can then beexported to one of a variety of qPCR programsavailable online or to Microsoft Excel for analysis.

Assuming good standard curves and a validationexperiment that demonstrated equal amplificationefficiencies of the target and reference genes, theComparative Cq Method can then be used to testfor gene amplification. The following steps caneasily be done in Excel:

1. Calculate the mean and standard deviation valuesof the replicates for both genes of each biologicalsample. Sample Cq values should fall within thelimits of the standard curve and standarddeviations should be # 0.3 in order to distin-guish between a one- to twofold difference incopy number (Bubner and Baldwin 2004).

2. Calculate the DCq value for each sample bysubtracting the mean Cq of the reference genefrom the mean Cq of the target gene.

3. Calculate the standard deviation of the DCq value.This is calculated from the standard deviations (s)of the target and reference gene values by thefollowing formula: s 5 (s1

2 + s22)1/2.

4. Calculate the DDCq value by subtracting theDCq of the reference sample (plant without geneamplification) from the DCq of the unknownsamples (plants in which gene amplification issuspected).

5. Calculate the fold-difference and range byincorporating the DCq standard deviations usingthe formula: 22DDCq, with DDCq 6 the standarddeviation of the DCq value.

Measuring Gene Expression Using Reverse-Tran-scription-qPCR (RT-qPCR). RT-qPCR is cur-rently the technique of choice for gene expressionquantification. However, a range of factors can

influence the reliability of direct gene expressionmeasurement (e.g., experimental error due tovariations in RNA or cDNA concentration amongsamples, operator error, and variation in the efficacyof the reverse-transcription reaction). Thus, expres-sion data is generally analyzed as relative expressionratios using a normalization strategy: raw geneexpression data is normalized using expression dataof one (or preferably several) reference gene(s).

Reference Genes. A reference gene must have aconstitutive and constantly stable expression in allexperimental conditions studied. For herbicide-resistance studies, reference gene sets should consistof genes for which expression has been proven stablebefore and after herbicide application, and amongresistant and sensitive plants. To date, reference geneswith a stable expression under herbicide action haveonly been validated for two grass weeds, blackgrass(Alopecurus myosuroides Huds.) (Petit et al. 2012) andLolium spp. (Duhoux and Delye 2013).

Three validated reference genes are generallyconsidered to be adequate for accurate normalization(Vandesompele et al. 2002). To obtain threevalidated reference genes, it is advisable to test thestability of at least six candidate reference genes(Bustin et al. 2009, 2010). Candidate genes shouldbe involved in different metabolic pathways, so thatthey are under independent regulations of expres-sion. This reduces the risk for a similar effect ofherbicide application on all candidate referencegenes. A list of potentially useful candidate genes(that is, genes that have frequently been used asvalidated reference genes in the literature) is givenin Table 4. Another possibility to identify geneswith a stable expression that can be used ascandidate reference genes is to use data fromtranscriptome-wide studies, when available (e.g.,Czechowski et al. 2005: the genes identified inArabidopsis thaliana (L.) Heynh.in this work arecandidate reference genes of interest for studiesaddressing species in the family Brassicaceae).

Tip: The fold-difference for the referenceindividual and any other individual lacking geneamplification should be approximately equal to 1,assuming the reference gene is single copy andstable.

Tip: There are no ‘‘universal’’ reference genesidentified to date. Reference genes suitable forone herbicide and one weed species might not besuitable for a different herbicide/species pair. It isa prerequisite to any study aiming at comparinggene expression among samples that the stabilityof the reference genes used has been adequatelyvalidated in the experimental system studied.

108 N Weed Science 63, Special Issue 2015

Experimental Design for the Validation ofReference Genes. The stability of reference geneexpression must be checked across all experimentalconditions studied. cDNA samples used to testcandidate reference genes must therefore cover thefull range of modalities intended in the study: (1)plant material (e.g., resistant and sensitive pheno-types, different geographical origins); (2) herbicideapplication modalities (e.g., herbicide dose, timebefore and after herbicide application); and (3)plant tissue (select the tissues where herbicides areapplied where herbicide target is most abundant oractive; this generally means the youngest, mostactively growing plant parts, such as leaf andmeristems for foliar herbicides, and root tips forsoil-applied herbicides).

Obtaining Expression Data for Candidate ReferenceGenes. This process is as follows:

1. Choose a set of candidate reference genes.

2. Design primer pairs for each candidatereference gene (see section ‘‘primer design’’above). The expected amplicon should be 70to 300 bp long for an optimal efficiency ofPCR.

3. Test the primer pairs in classical PCR followedby agarose gel electrophoresis. This step can beconducted on genomic DNA or cDNA. Ifconducted on genomic DNA, consider thepossibility for introns to be present in thetargeted region. A single, clear amplicon mustbe obtained.

4. Remove the candidate reference genes notpassing this step from your list. Alternatively,design new primers for the reference gene andgo back one step.

Tip: It is advisable not to use ribosomal genes asreference genes, because of their high expressionlevel compared to most other genes.

Table 4. Examples of candidate reference genes for use in reverse-transcription quantitative polymerase chain reaction (RT-qPCR).

Metabolic pathway (Gene ontology) Candidate reference gene

Microtubule-based process Beta-tubulina

Oxidation-reduction process Glyceraldehyde-3-phosphate dehydrogenasea,b

Translational initiation Cap binding proteinb

Protein catabolic process Ubiquitina.b

Translational elongation Eukaryotic elongation factor 1 alphaTranslational elongation Eukaryotic elongation factor 4 alphaTransmembrane transport Sucrose proton symporterCarbon fixation Ribulose 1-5-bisphosphateMicrotubule-based process ActinProtein folding CyclophilinProtein catabolic process Ubiquitin conjugating enzyme E2DNA binding Histone 3Vacuolar fusion SAND protein

a Reference genes validated in the grass weed Alopecurus myosuroides with herbicides inhibiting ACCase (Petit et al. 2012).b Reference genes validated in the grass weed Lolium sp. with herbicides inhibiting ALS (Duhoux and Delye 2013).

Tip: A crucial point here is to collect plantmaterial rapidly and to neutralize plant metabo-lism without delay to avoid modifications of thegene expression patterns induced by response towounding. The most widespread and effectivemethod for this purpose is snap freezing of plantmaterial in liquid nitrogen, followed by storage at280 C. Frozen plant material must not beallowed to thaw, and must be processed asquickly as possible to avoid RNA degradation(see RNA extraction above).

Tip: At least six candidate genes should be testedfor stability of expression using RT-qPCR. It isthus wise to start the procedure with more thansix potential candidate genes, because this meansthat at least six candidate genes must pass steps 2to 13 below.

Tip: Designing primers flanking intron-contain-ing regions can be useful to check genomic DNAcontamination in cDNA samples.

Delye et al.: Methods for HR mechanisms N 109

5. Check the sequence of the amplicon.6. Remove the candidate reference genes not

passing this step from your list.7. Extract RNA from your samples. Check RNA

quality and perform cDNA synthesis from asimilar amount of starting RNA for all samples.

8. Check efficiency and specificity of the primerpairs in qPCR using dilution series as describedin the previous section.

9. Remove the candidate reference genes that donot show a single-peak melting curve (seeFigure 3).

10. Remove the candidate reference genes withPCR efficiency values outside of the 90% to110% range from your list.

11. Chose one working dilution for all your cDNAsamples.

12. Perform qPCR for every remaining candidategene on the diluted cDNA samples. Include thedilution series in each qPCR run as a control forthe efficiency of the reaction.

13. Extract Cq values for each cDNA sample andfor each gene after positioning the threshold.

Selecting a Set of Validated Reference Genes forExpression Data Analysis. Three software packagesare commonly used to estimate gene stability andidentify the more stable genes. They are available at:http://www.gene-quantification.de/. A brief descrip-tion is provided below for the three softwarepackages:

1. BestKeeper (Pfaffl et al. 2004):Input:

N qPCR efficiency values for each candidatereference gene.

N Cq values for each gene and each sample.

Output:

N Samples that should be removed from theanalysis due to experimental errors for goodstability analysis.

N Variation in Cq values and its standarddeviation (SD) for each gene. Genes withSD , 1.00 and P value for the Pearsoncorrelation coefficient below 0.01 are consid-ered stable. Most stable genes are genes withlowest standard deviation.

Tip: It is crucial to use several biological replicatesto adequately test the stability of reference genes,taking into account biological variation (e.g.,different individual plants used in a sameexperimental modality). Technical replicatesshould also be used for each biological replicateto account for experimental variation (e.g.,independent reverse-transcription reactions).

Tip: If the efficiency value is not satisfactory for agiven gene, designing new primers is generallyeasier, faster, and cheaper than trying to optimizeqPCR.

Tip: If the dilution series used enabled satisfactoryassessment of qPCR efficiency, the midpointdilution of the series can generally be used as theworking dilution for all cDNA samples. Thisrequires that a similar amount of RNA is used forall samples when performing the reverse-tran-scription reaction.

Tip: The threshold for fluorescence detection canbe set automatically or manually. When settingthe threshold manually, use the logarithmicamplification plot to position the threshold sothat it is above the background fluorescence,below the linear region, and at the beginning ofthe region of exponential amplification.

Tip: The three software packages are each basedon a specific algorithm. They provide comple-mentary results, and should be used together toidentify the most stable genes to be used asreference genes.

Tip: it is possible to analyse the stability of thegenes within a subgroup of sample by onlyentering the Cq values of samples in thissubgroup.

110 N Weed Science 63, Special Issue 2015

1. NormFinder (Andersen et al. 2004):

Input:

N Cq data transformed using the 22DDCq method(Livak and Schmittgen 2001). This method usesa calibrator defined as the lowest Cq valueobtained for each gene (i.e., the Cq value fromthe sample with the highest expression level forthe gene considered). Input data for the sample iand the gene j will thus be: Iij 5 22(Cqij 2 Cqminj).

Output:

N Stability value (SV) for each gene. The moststable genes are the genes with the lowest SVs.

1. geNorm (Vandesompele et al. 2002):

Input:

N qPCR efficiency for each candidate referencegene.

N Cq data transformed using the 22DDCq method.

Output:

N Expression stability value (M) for each gene.Genes with M , 1.5 are considered stable.

N Ranking of the genes from the two most stabledown to the least stable.

N Optimal number of reference genes for ade-quate normalization.

Figure 3. Primer specificity test for RT-qPCR. (A) melting curves generated for one primer pair that does not yield a single-peakmelting curve because of primer dimer formation or nonspecific amplification. This primer pair is thus not suitable for RT-qPCR. (B)melting curves generated for one primer pair that yield a single-peak melting curve. This primer pair can be proceeded to the next step(step # 10) of the candidate reference gene validation process.

Tip: Stability should be assessed without consid-ering subgroups in a first step (overall stability).NormFinder allows assigning subgroups to eachsample in the sampling (e.g., sensitive or resistant,treated or untreated). In a second step, stabilityshould be analysed within subgroups. Subgroupsshould include a minimum of eight samples.

Tip: The 0.15 cutoff threshold for the pairwisevariation between consecutive normalizationfactors that is proposed to identify the numberof reference genes most adequate for normaliza-tion is not to be considered as an absolutethreshold value. If pairwise variation values arealways above 0.15, using the number of reference

2. 3.

Delye et al.: Methods for HR mechanisms N 111

How Many Reference Genes? Never use a singlereference gene for normalization. On the otherhand, keep in mind that it will be necessary tomeasure the expression of all reference genes in allsamples analysed to normalize the expression of thetarget genes investigated. Thus, the more referencegenes used, the more labor-intensive the experi-ments. In most cases, using the three most stablegenes allows adequate normalization of geneexpression (provided all genes have been foundstable using the three software packages) (Vande-sompele et al. 2002).

Quantification and Comparison of Relative ExpressionLevels of Genes of Interest. Primer design for genesof interest. Follow steps 2 to 10 in section ObtainingExpression Data for Candidate Reference Genes usingthe sequence for your gene(s) of interest.

Comparison of the expression data generated for

genes of interest among samples is performed on thebasis of the respective relative expression levels.Relative expression levels are computed based on thePCR efficiency of the target gene and of thereference genes, and on the Cp deviation of thetarget and the reference genes in the analysed

sample compared to a reference sample (Pfaffl2001). Relative expression level quantificationsoftware implement this approach and allow theuser to compare the expression of a target geneamong different samples. A ‘‘reference sample’’ orgroup of samples must be defined beforehand.

A useful software for this purpose is REST-mcs(Relative Expression Software Tool—multiple con-dition solver, Pfaffl et al. 2002) that is available athttp://www.gene-quantification.de/.

Input:

N qPCR efficiency for each target gene andreference genes.

N Cq values for each sample and each gene(reference and target genes).

Output:

N Relative expression ratio in the ‘‘sample group’’using the ‘‘reference group’’ as a baseline.

N Significance of the difference observed amongsamples (P value , 0.05).

Another option for data analysis is the qpcRlibrary developed for the R software (http://www.dr-spiess.de/qpcR.html)

genes yielding the lowest normalization factorvalue is also appropriate.

Tip: It is possible to assess the stability of thegenes within a subgroup of samples by onlyentering the transformed Cq values of samples inthe subgroup.

Tip: Remember it is preferable to use genesinvolved in different metabolic pathways.

Tip: qPCR efficiency values can be largely aboveor below 100% when working with genes ofinterest showing important variation in theirexpression among samples. This is acceptable,provided that the associated correlation coeffi-cient that indicates the linearity of the expressionis . 98%. Caution: this tip does not apply toreference genes.

Tip: In experiments involving response toherbicides, obvious reference samples would becDNA from untreated, herbicide-sensitive plants.

Tip: In some samples, no expression of a targetgene can be detected, and thus no Cq value canbe generated, but expression of the referencegenes is as expected (e.g., in untreated sensitiveplants when studying a gene potentially involvedin resistance). In such cases, a possibility is to usemore concentrated dilutions of these samples inqPCR so as to be able to generate a Cq value bystarting with more copies of the target gene. Itmight be that, even so, expression of the targetgene still is not detected. This indicates extremelylow, or even absent expression of the target genein the samples in question. For the purpose ofgene expression comparison, a possibility is toarbitrarily attribute a Cq value equal to themaximum number of cycles in the qPCR for thetarget gene in the samples in question.

112 N Weed Science 63, Special Issue 2015

Dealing with Polyploid Species

The approaches discussed above apply straight-forwardly to diploid species, i.e., species with asingle genome that contains single-copy genes ofinterest. Yet, there are quite a few major weedspecies that are polyploid, i.e., they contain severalgenomes [e.g., Avena spp., Chenopodium spp.,Echinochloa spp., ricefield bulrush [Schoenoplectusmucronatus (L.) Palla]]. The ploidy level can varyamong species within a given genus, or even amongplants in a given species. For instance, commonlambsquarters (Chenopodium album L.) plants canbe diploid (one genome), tetraploid (two genomes),or hexaploid (three genomes) (Bhargava et al.2006). Studying polyploid species means facingtwo types of additional challenges.

First, because polyploid plants have severalgenomes, they are expected to contain several copiesof a gene of interest. The genomic organisation ofsuch species is not simple (Scarabel et al. 2010), andthe number of copies of a given gene is notnecessarily the same in the different genomes. Thus,a species with X number of genomes might containmore than X number of copies of a given gene; forexample, a recent study conducted on the tetraploidspecies rice barnyardgrass [Echinochloa phyllopogon(Stapf) Koso-Pol.] identified two ALS copies, asexpected, but also four chloroplastic ACCase copies(Iwakami et al. 2012). As every copy of a geneencoding a protein involved in herbicide resistancemight carry mutations endowing herbicide resis-tance (e.g., Yu et al. 2013), it is important to obtainand sequence all copies of this gene present in agiven species. Specific techniques such as Southernhybridization (Southern 1975 [ not reviewed in thispaper]) can be used to assess the number of copiesof the gene of interest in a polyploid species.

The second additional challenge resides in thepossibility that some of the gene copies presentacross the different genomes of polyploid species arepseudogenes or silenced genes (e.g., Huang et al.2002). Thus, it is important to check that all genesanalysed are readily expressed. It is thereforeessential to know the ploidy level of a species ofinterest beforehand, and to be aware that studyingpolyploid weeds presents additional challenges.

Mistakes to Avoid after a Genetic Change

is Identified

It must be kept in mind that having identified agenetic change (either a structural mutation or gene

expression differences reflecting occurrence ofregulatory mutation[s]) between resistant andsensitive plants does not necessarily mean that yourwork is done. Ultimately, you will need evidencethat the identified genetic change is actuallyresponsible for the observed resistance phenotype.Such evidence can be obtained using genetic andbiochemical approaches. For example, one couldclone genes that are identical except for theimplicated mutation, produce the encoded enzymesusing an E. coli expression system, and thencompare herbicide sensitivities of the enzymes usingan in vitro biochemical assay (Dayan et al. 2014).Alternatively, one could compare herbicide sensi-tivities of whole organisms (E. coli, yeast, or plant)bearing transgenes that differ only by the implicatedmutation. If multiple mutations are found within acandidate resistance gene, systematic experimenta-tion is required to determine which specificmutation, or combination of mutations, confersresistance. If an identified mutation is identical toone that has been previously demonstrated to conferresistance in a different species, ‘‘guilt by homolo-gy’’ provides reasonable—but not complete—certainty of the mutation’s involvement in resis-tance. Nevertheless, this ‘‘guilt by homology’’ saysnothing about the potential involvement of anadditional resistance mechanism within your par-ticular population. Genetic analysis (Mallory-Smithet al. 2014) can be used to determine if multipleresistance mechanisms are present within a popu-lation, and to test for cosegregation of the candidategenetic difference. Cosegregation analysis also canbe a useful approach to obtain evidence that achange in gene expression confers resistance.

As a final tip with regard to making inferencesbased on identified genetic changes, one needs tokeep in mind that weed populations typically existas collections of genetically diverse individuals andthat resistance can be endowed by a range of genes(reviewed in Delye et al. 2013). Identification of aparticular genetic change conferring resistance inone or a few individuals of a population certainly isno guarantee that other genetic changes conferringresistance are not also present within the populationfrom which the plants were obtained, or even withinthe plants in which the particular genetic changewas identified.

Literature Cited

Alarcon-Reverte R, Hanley S, Kaundun SS, Karp A, Moss SR(2013) A SNaPshot assay for the rapid and simple detection of

Delye et al.: Methods for HR mechanisms N 113

known point mutations conferring resistance to ACCase-inhibiting herbicides in Lolium spp. Weed Res 53:12–20

Andersen L, Jensen JL, Ørntoft TF (2004) Normalization ofreal-time quantitative reverse transcription-PCR data: amodel-based variance estimation approach to identify genessuited for normalization applied to bladder and colon cancerdata sets. Cancer Res 64:5245–5250

Beckie HJ, Tardif FJ (2012) Herbicide cross-resistance in weeds.Crop Prot 35:15–28

Bhargava A, Shukla S, Ohri D (2006) Karyotypic studies onsome cultivated and wild species of Chenopodium (Chenopo-diaceae). Genet Resour Crop Evol 53:1309–1320

Bubner B, Baldwin IT (2004) Use of real-time PCR fordetermining copy number and zygosity in transgenic plants.Plant Cell Rep 23:263–271

Burgos NR, Tranel PJ, Streibig JC, Davis VM, Shaner D,Norsworthy JK, Ritz C (2013) Review: confirmation ofresistance to herbicides and evaluation of resistance levels.Weed Sci 61:4–20

Bustin SA, Beaulieu JF, Huggett J, Jaggi R, Kibenge FSB, OlsvikPA, Penning LC, Toeggel S (2010) MIQE precis: Practicalimplementation of minimum standard guidelines for fluores-cence-based quantitative real-time PCR experiments. BMCMol Biol 11:74

Bustin SA, Benes V, Garson JA, Hellemans J, Huggett J, KubistaM, Mueller R, Nolan T, Pfaffl MW, Shipley GL, Vande-sompele J, Wittwer CT (2009) The MIQE guidelines:minimum information for publication of quantitative realtimePCR experiments. Clin Chem 55:611–622

Chaves AL, Vergara CE, Mayer JE (1995) Dichloromethane asan economic alternative to chloroform in the extraction ofDNA from plant tissues. Plant Mol Biol Rep 13:18–25

Czechowski T, Stitt M, Altmann T, Udvardi MK, Scheible WR(2005) Genome-wide identification and testing of superiorreference genes for transcript normalization in Arabidopsisthaliana. Plant Physiol 139:5–17

Dayan FE, Owens DK, Corniani N, Silva FML, Watson SB,Howell J, Shaner DL (2015) Biochemical markers and enzymeassays for herbicide mode of action and resistance studies.Weed Sci 63:23–63

Delye C (2005) Weed resistance to acetyl coenzyme Acarboxylase inhibitors: an update. Weed Sci 53:728–746

Delye C (2013) Unravelling the genetic bases of non-target-site-based resistance (NTSR) to herbicides: a major challenge forweed science in the forthcoming decade. Pest Manag Sci69:176–187

Delye C, Boucansaud K (2008) A molecular assay for theproactive detection of target site-based resistance to herbicidesinhibiting acetolactate synthase in Alopecurus myosuroides.Weed Res 48:97–101

Delye C, Jasieniuk M, Le Corre V (2013) Deciphering the evolutionof herbicide resistance in weeds. Trends Genet 29:649–658

Delye C, Matejicek A, Gasquez J (2002) PCR-based detection ofresistance to acetyl-CoA carboxylase-inhibiting herbicides inblack-grass (Alopecurus myosuroides Huds.) and ryegrass(Lolium rigidum Gaud.). Pest Manag Sci 58:474–478

Delye C, Michel S, Berard A, Chauvel B, Brunel D, GuilleminJP, Dessaint F, Le Corre V (2010) Geographical variationin resistance to acetyl-coenzyme A carboxylase-inhibitingherbicides across the range of the arable weed Alopecurusmyosuroides (blackgrass). New Phytol 186:1005–1017

Delye C, Pernin F, Michel S (2011) ‘Universal’ PCR assaysdetecting mutations in acetyl-coenzyme A carboxylase or

acetolactate-synthase that endow herbicide resistance in grassweeds. Weed Res 51:353–362

Dower WJ, Miller JF, Ragsdale CW (1988) High efficiencytransformation of E. coli by high voltage electroporation.Nucleic Acids Res 16:6127–6145

Doyle JJ, Doyle JL (1990) A rapid total DNA preparationprocedure for fresh plant tissue. Focus 12:13–15

Duhoux A, Delye C (2013) Reference genes to study herbicidestress response in Lolium sp.: up-regulation of P450 genes inplants resistant to acetolactate-synthase inhibitors. PLOS ONE8:10.1371/journal.pone.0063576

Gaines T, Lorentz L, Maiwald F, Ott MC, Perret P, Yu Q, HanH, Busi R, Strek H, Beffa R, Powles S (2013) Identifyingdifferential expression in non-target-site herbicide resistancegenes in Lolium rigidum using high-throughput sequencing.Page 69 in Global Herbicide Resistance Challenge. Perth,Australia: Australian Herbicide Resistance Initiative [Abstract]

Gaines TA, Zhang W, Wang D, Bukun B, Chisholm ST, ShanerDL, Nissen SJ, Patzoldt WL, Tranel PJ, Culpepper AS, GreyTL, Webster TM, Vencill WK, Sammons RD, Jiang J, PrestonC, Leach JE, Westra P (2010) Gene amplification confersglyphosate resistance in Amaranthus palmeri. Proc Natl AcadSci U S A 107:1029–1034

Gardin J, Beffa R, Gouzy J, Carrere S, Delye C (2013) Atranscriptomics-based approach enables the first identificationof candidate genes involved in non-target-site-based resistanceto herbicides in a grass weed (Alopecurus myosuroides). Page 68in Global Herbicide Resistance Challenge. Perth, Australia:Australian Herbicide Resistance Initiative [Abstract]

Huang S, Sirikhachornkit A, Faris JD, Su X, Gill BS, HaselkornR, Gornicki P (2002) Phylogenetic analysis of the acetyl-CoAcarboxylase and 3-phosphoglycerate kinase loci in wheat andother grasses. Plant Mol Biol 48:805–820

Iwakami S, Uchino A, Watanabe H, Yamasuea Y, Inamura T(2012) Isolation and expression of genes for acetolactatesynthase and acetyl-CoA carboxylase in Echinochloa phyllopo-gon, a polyploid weed species. Pest Manag Sci 68:1098–1106

Jasieniuk M, Brule-Babel AL, Morrison IN (1996) Theevolution and genetics of herbicide resistance in weeds. WeedSci 44:176–193

Kaundun SS, Cleere SM, Stanger CP, Burbidge JM, Windass JD(2006) Real-time quantitative PCR assays for quantification ofL1781 ACCase inhibitor resistance allele in leaf and seed poolsof Lolium populations. Pest Manag Sci 62:1082–1091

Kaundun SS, Dale RP, Zelaya IA, Dinelli G, Marotti I, McIndoeE, Cairns A (2011) A novel P106L mutation in EPSPS and anunknown mechanism(s) act additively to confer resistance toglyphosate in a South African Lolium rigidum population.J Agric Food Chem 59:3227–3233

Kaundun SS, Windass JD (2006) Derived cleaved amplifiedpolymorphic sequence, a simple method to detect a key pointmutation conferring acetyl CoA carboxylase inhibitor herbi-cide resistance in grass weeds. Weed Res 46:34–39

Kwok S, Chang SY, Sninsky JJ, Wang A (1994) A guide to thedesign and use of mismatched and degenerate primers. PCRMethods Appl 3:S39–S47

Livak KJ, Schmittgen TD (2001) Analysis of relative geneexpression data using real-time quantitative PCR and the2-DDCt Method. Methods 25:402–408

Mallory-Smith C, Hall L, Burgos N (2015) Experimentalmethods to study gene flow. Weed Sci 63:12–22

Martin JA, Wang Z (2011) Next-generation transcriptomeassembly. Nat Rev Genet 12:671–682

114 N Weed Science 63, Special Issue 2015

Metzker ML (2005) Emerging technologies in DNA sequencing.Genome Res 15:1767–1776

Neff MM, Neff JD, Chory J, Pepper AE (1998) dCAPS, asimple technique for the genetic analysis of single nucleotidepolymorphisms: experimental applications in Arabidopsisthaliana genetics. Plant J 14:387–392

Neff MM, Turk E, Kalishman M (2002) Web-based primerdesign for single nucleotide polymorphism analysis. TrendsGenet 18:613–615

Ness RW, Siol M, Barrett SCH (2011) De novo sequenceassembly and characterization of the floral transcriptome incross- and self-fertilizing plants. BMC Genomics 12:298

Nolan T, Hands RE, Bustin SA (2006) Quantification of mRNAusing real-time RT-PCR. Nat Protoc 1:1559–1582

Ohtsuka E, Matsuki S, Ikehara M, Takahashi Y, Matsubara K(1985) An alternative approach to deoxyoligonucleotides ashybridation probes by insertion of deoxyinosine at ambiguouscodon positions. J Biol Chem 260:2605–2608

Ozsolak F, Milos PM (2010) RNA sequencing: advances,challenges and opportunities. Nat Rev Genet 12:87–98

Petit C, Pernin F, Heydel JM, Delye C (2012) Validation of a setof reference genes to study response to herbicide stress ingrasses. BMC Res Notes 5:18

Pettersson M, Bylund M, Alderborn A (2003) Molecularhaplotype determination using allele-specific PCR andpyrosequencing technology. Genomics 82:390–396

Pfaffl MW (2001) A new mathematical model relativequantification in real-time RT-PCR. Nucleic Acids Res 29:e45

Pfaffl MW, Horgan GW, Dempfle L (2002) Relative expressionsoftware tool (RESTE) for group-wise comparison andstatistical analysis of relative expression results in real-timePCR. Nucleic Acids Res 30:e36

Pfaffl MW, Tichopad A, Prgomet C, Neuvians TP (2004)Determination of stable housekeeping genes, differentiallyregulated target genes and sample integrity: BestKeeper-Excel-based tool using pair-wise correlations. Biotechnol Lett26:509–515

Radstrom P, Lofstrom C, Lovenklev M (2003) Strategies forovercoming PCR inhibition. Pages 149–161 in DieffenbachCW, Dveksler GS, eds. PCR Primer: A Laboratory Manual.Cold Spring Harbor: Cold Spring Harbor Laboratory Press

Riggins CW, Peng Y, Stewart CN, Jr, Tranel PJ (2010)Characterization of de novo transcriptome for waterhemp(Amaranthus tuberculatus) using GS-FLX 454 pyrosequencingand its application for studies of herbicide target-site genes.Pest Manag Sci 66:1042–1052

Roux KH (1995) Optimization and troubleshooting in PCR.Genome Res 4:S185–S194

Roux KH (2009) Optimization and troubleshooting in PCR.Cold Spring Harbor Protoc 1–6, doi:10.1101/pdb.ip66. http://cshprotocols.cshlp.org/content/2009/4/pdb.ip66. Accessed June18, 2013

Salas RA, Dayan FE, Pan Z, Watson SB, Dickson JW, Scott RC,BurgosNR(2012)EPSPSgeneamplificationinglyphosate-resistantItalian ryegrass (Lolium perenne ssp. multiflorum) fromArkansas. Pest Manag Sci 68:1223–1230

Sanger F, Nicklen S, Coulson AR (1977) DNA sequencing withchain-terminating inhibitors. Proc Natl Acad Sci U S A74:5463–5467

Scarabel L, Locascio A, Furini A, Sattin M, Varotto S (2010)Characterisation of ALS genes in the polyploid speciesSchoenoplectus mucronatus and implications for resistancemanagement. Pest Manag Sci 66:337–344

Sommer SS, Groszbar AR, Bottema CDK (1992) PCRamplification of specific alleles (PASA) is a general methodfor rapidly detecting known single base-pair changes. Bio-techniques 12:82–87

Southern EM (1975) Detection of specific sequences amongDNA fragments separated by gel electrophoresis. J Mol Biol98:503–517

Thudi M, Li Y, Jackson SA, May GD, Varshney RK (2012)Current state-of-art of sequencing technologies for plantgenomics research. Brief Funct Genomics 2:3–11

Tindall KR, Kunkel TA (1988) Fidelity of DNA synthesis bythe Thermus aquaticus DNA polymerase. Biochemistry 27:6008–6013

Tranel PJ, Riggins CW, Bell MS, Hager AG (2011) Herbicideresistances in Amaranthus tuberculatus: a call for new options.J Agric Food Chem 59:5808–5812

Tranel PJ, Wright TR (2002) Resistance of weeds to ALS-inhibiting herbicides: what have we learned. Weed Sci 50:700–712

Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N,De Paepe A, Speleman F (2002) Accurate normalization ofreal-time quantitative RT-PCR data by geometric averaging ofmultiple internal control genes. Genome Biol 3:10.1186/gb-2002-3-7-research0034

Vigueira CC, Olsen KM, Caicedo AL (2013) The red queen inthe corn: agricultural weeds as models of rapid adaptiveevolution. Heredity 110:303–311

Wagner J, Haas HU, Hurle K (2002) Identification of ALSinhibitor-resistant Amaranthus biotypes using polymerasechain reaction amplification of specific alleles. Weed Res42:280–286

Ward JA, Ponnola L, Weber CA (2012) Strategies fortranscriptome analysis in nonmodel plants. Am J Bot99:267–276

Warwick SI, Xu R, Sauder C, Beckie HJ (2008) Acetolactatesynthase target-site mutations and single nucleotide polymor-phism genotyping in ALS-resistant kochia (Kochia scoparia).Weed Sci 56:797–806

Yu Q, Ahmad-Hamdani MS, Han H, Christoffers MJ, PowlesSB (2013) Herbicide resistance-endowing ACCase genemutations in hexaploid wild oat (Avena fatua): insights intoresistance evolution in a hexaploid species. Heredity110:220–231

Yuan JS, Tranel PJ, Stewart CN, Jr (2007) Non-target-siteherbicide resistance: a family business. Trends Plant Sci12:6–13

Received June 18, 2013, and approved November 19, 2013.

Delye et al.: Methods for HR mechanisms N 115