Copyright 2009 by the Genetics Society of AmericaDOI: 10.1534/genetics.108.095588

Genetics Education

Innovations in Teaching and Learning Genetics

Edited by Patricia J. Pukkila

Reinventing the Ames Test as a Quantitative Lab That ConnectsClassical and Molecular Genetics

Nathan Goodson-Gregg and Elizabeth A. De Stasio1

Biology Department, Lawrence University, Appleton, Wisconsin 54911

Manuscript received August 27, 2008Accepted for publication November 13, 2008

ABSTRACT

While many institutions use a version of the Ames test in the undergraduate genetics laboratory,students typically are not exposed to techniques or procedures beyond qualitative analysis of phenotypicreversion, thereby seriously limiting the scope of learning. We have extended the Ames test to includeboth quantitative analysis of reversion frequency and molecular analysis of revertant gene sequences. Bygiving students a role in designing their quantitative methods and analyses, students practice and applyquantitative skills. To help students connect classical and molecular genetic concepts and techniques, wereport here procedures for characterizing the molecular lesions that confer a revertant phenotype. Wesuggest undertaking reversion of both missense and frameshift mutants to allow a more sophisticatedmolecular genetic analysis. These modifications and additions broaden the educational content of thetraditional Ames test teaching laboratory, while simultaneously enhancing students skills in experimentaldesign, quantitative analysis, and data interpretation.

AS called for by numerous national groups (e.g.,NationalResearchCouncil2003;Handelsmanet al. 2004), biology education has moved in recent yearstoward the provision of research-rich environments inwhich undergraduate laboratories are investigative andopen ended and in which quantitative skills areemphasized. While data analysis has long been a stapleof student learning, recent research demonstrates thatstudents becomemost engagedand learnbest when theyhave a hand in the design of experiments as well as in theexecution and analysis of resulting data (Hake 1998;Merkel 2003; Handelsman et al. 2007). Inquiry-basedlabshavebeen shown to improve students research skillsin biology (Myers and Burgess 2003). Further, assuggested by BIO2010 (National Research Council2003), biology curricula should explicitly build thequantitative skills of budding biologists. Hack andKendall (2005) argue that biology curricula shouldchange because current life science students must learnto use models, to apply appropriate mathematic tools

and statistics to solve problems, and to manage andintegrate data. That these tools are best taught in thecontext of biology courses themselves has been demon-strated by Metz (2008), who has shown that undergrad-uate biology students do notmake connections betweenquantitative concepts taught in mathematics and statis-tics courses and their application tobiological problems.Metz (2008)demonstrates that inclusionof quantitativeand statistical analyses in biology laboratory courses ledto significant gains in long-term retention of suchknowledge, regardless of whether students also hadtaken courses in statistics.To address the issue of connecting quantitative

analysis and biological problem solving, we have ex-tended the open-ended Ames test for the undergradu-ate genetics lab to allow students to practice quantitativeskills during student-driven experimental design andanalysis. Students bring to the lab potential mutagens oftheir choice, and they are charged both with creatingmethods to determine the number of colony-formingunits (CFUs) per bacterial culture and with using thatfigure to determine reversion frequencies. We find thatthis is a difficult task for students, but having them

1Corresponding author:LawrenceUniversity, P.O. Box 599, Appleton,WI54911-0599. E-mail: destasie@lawrence.edu

Genetics 181: 2331 ( January 2009)

conclude what sort of serial dilutions are needed, forexample, is an important step in gaining a long-termunderstanding of the quantitative aspects of the lab andother similar analyses. This approach requires thatinstructors and students bring to the lab an attitude ofinvestigation and learning rather than a sense ofurgency to do the lab and obtain a particular result.Importantly, this lab also fulfills a second great need

within the genetics curriculum, specifically, the ability todirectly connect concepts of classical genetics with thoseof molecular genetics. Investigating both classical andmolecular genetics in the time frame of the undergrad-uate lab requires the use of fast-growing organisms witheasily selectable phenotypes. Thus, the classic Ames testusing Salmonella typhimurium in a reversion screen pro-vided the starting point for the design of this lab.Marshall (2007) has published a yeast-based versionof the Ames test that is also investigative in nature.Marshalls version does not include the level of quanti-tative analysis described here, nor does she have stu-dents perform molecular analysis of revertant genes,although her lab could be extended as we have donewith Salmonella. Wessner et al. (2000) has described aninitial qualitative spot-overlay experiment followed by asecondary dose-response experiment as a cost-effectivesubstitute for the traditional Ames test lab. This modi-fication increases the amount of quantitative and qual-itative data generated over a period of 23 weeks, but itdoes not carry the experiment further than classicalgenetics. The lab described here has students carry theirinvestigation to the molecular level, including DNApreparation, PCR, and DNA sequencing, thus connect-ing reversion analysis, an important and rather difficultclassical genetic concept, with revertant gene sequences.The lab is cost effective and produces substantial classicalandmolecular genetic data in,3 weeks (two lab sessionsplus sequence analysis).Learning outcomes at the conceptual level of molec-

ular genetics include an enhanced understanding of thefact that multiple DNA sequences can encode a func-tioning enzyme, and hence can confer the same pheno-type, thatdifferentmutagensproducedifferentmolecularlesions, and that beneficial mutations can be producedspontaneously, as well as classical concepts of mutagen-esis and reversion. Students should leave with improvedunderstanding of molecular techniques such as DNAextraction, PCR, DNA sequencing, and sequence com-parison. In addition, students quantitative skills receivepractice in the design of appropriate serial dilutions,computation andunderstanding of reversion frequency,use of descriptive statistics, and discussions of statisticalsignificance.The original Ames test (Ames 1979) is a reversion

screen using Salmonella strains with a His phenotypedue to mutations inHisD orHisG. These Salmonella areplated on histidine-deficient media. Mutagens or othercompounds are then added to the plates using one of

several application procedures and the number ofrevertant His1 colonies is enumerated. We report hereprotocols for adding a quantitative component to thislab in which both spontaneous and mutagen-inducedreversion frequencies are experimentally derived. Tra-ditional pedagogical applications of the Ames test endwith students simply comparing the numbers of re-vertant colonies produced by different substances. Thisprocess does not allow quantitative comparison betweenmutagenesis of strains carryingdifferentHismutations.In addition, when students carry out a traditional Amestest lab, they do not determine which DNA sequencechanges confer revertant phenotypes and thus cannotinfer which mutation mechanisms acted on the strainsor fully understand the concept of reversion.Recently, some research laboratories have used DNA

sequence analysis in conjunction with the traditionalAmes test (Levine et al. 1994; Abu-Shakra et al. 2000)and found strong evidence of substantial sequencevariation among revertants of a given strain (Kochet al. 1994). We report here a new protocol that allowsstudents to analyze themolecular lesions that confer theHis1 phenotype in revertants of strains carrying either amissense or frameshift His mutation in HisG or HisD,respectively. The variability in the molecular lesionsbetween these revertants will give students concrete,experimentally derived examples of the connectionbetween phenotype and genotype by showing studentsthat there are many ways that gene sequences can evolveto confer a specific phenotype. Students will also findthat treatment with known mutagens preferentiallyinduces specific types of mutations (e.g., transitionsand deletions) that correlate to the mutagenic proper-ties of these substances (Takiya et al. 2003). Dataproduced in this lab can also provide a starting pointfor discussions of mutation mechanisms, DNA repair,and molecular evolution.

MATERIALS AND METHODS

Strains and media: Salmonella strains TA98, TA100, andTA102 were obtained from BioReliance. Strains TA1535 andTA1538 were obtained from the American Type CultureCollection. Only strains TA1535 and TA1538 are used in thefinal teaching laboratory, as they are readily available andproduce consistent results easily analyzed by undergraduates.Overnight cultures were grown in 10 ml of L medium (1%tryptone, 0.5% yeast extract, 0.5% NaCl, 0.1% glucose) at 37.Nutrient-rich plates contained Lmediumwith 1.5% bacto agar;nutrient-deficientplates containedVBM(0.02%MgSO4*7H2O,0.2% citric acid, 1.0% K2HPO4, 0.35% NaNH4HPO4, 1.5%bacto agar, 2.0% dextrose). Three milliliters of overlay agar(9.7 3 1011% l-histidine, 1.1 3 103% biotin, 0.55% agar,0.45% NaCl) was used to plate Salmonella strains.

Week 1: CFU determination and Ames testone 3-hr lab

Determining CFUs in bacterial cultures: To allow studentsto quantify reversion frequencies, students determine theCFU

24 N. Goodson-Gregg and E. A. De Stasio

content of their overnight cultures. Students are told to expect23 10823 109 CFUs/ml in overnight cultures of Salmonella.Students must determine how to appropriately and accuratelydilute their cultures to reach 100 colonies/plate, taking intoaccount the expected bacterial concentration and the volumeto be plated. Students typically plate 100 ml of two dilutions,between 13 106 and 53 107, on nutrient-rich agar. Plates areincubated for 24 hr at 37 after which students count individualcolonies. Studentsmust thenuse this information todeterminehow many His bacteria were plated in their Ames test.Inducing mutagenesis via the Ames test: To produce His1

revertants, 100 ml of a TA1535 or TA1538 overnight culture ismixed with 3 ml of top agar and immediately plated on VBM.After allowing the top agar to harden, 10 ml of a potentiallymutagenic substance is pipetted onto a sterile 0.65-cm filterpaper disc. Pairs of students are encouraged to test themutagenicity of two substances they bring to the lab as wellas both control substances; solid substances such as foodproducts are placed in a student-determined portion of waterand blended. We obtained the best results when nomore thana half-volume of water was used to blend solids. The disc is thenplaced, mutagen side down, on the center of the top agaroverlay. Potentially mutagenic test substances that are notwater soluble are plated via direct application to the center ofthe plate. Positive controls include NaN3 at 0.05 mg/ml forstrain TA1535 and 4NOP at 0.5 mg/ml for strain TA1538,although both mutagens are tested with each strain. Test sub-stances, such as uncooked foods, that might contain sub-stantial amounts of other microorganisms can be autoclavedor filter sterilized, if sufficiently liquid, before application.Alternately, the potential mutagen can be plated directlywithout Salmonella on VBM to confirm the absence of thesepotential contaminants. In our experience, contaminating bac-teria are seen at very low frequency and are easily distinguishedfrom the Salmonella colonies. Plates are incubated at 37 for48 hr. Revertant colonies are counted in a manner that doesnot contaminate the plates. Students circle the revertantcolony that they would like to use for molecular analysis andplates are stored at 4. Instructors may want to provoke adiscussion concerning how one would rigorously determinethe mutation spectrum of a substance, leading students torealize that spontaneously derivedmutants will also be presentat lower frequency in any mutation-induced collection ofrevertants. In a research setting, it might therefore be best tosequence every revertant from a single plate and compare themutations observed to those seen from spontaneously gener-ated revertant sequences. To give the students more controlover the lab experiment, however, one independent revertantcolony identified by each student is picked with a sterile

inoculating loop and grown overnight in 10 ml of L mediumprior to the next lab.

Week 2: Characterizing molecular lesions conferring a His1

phenotypeone 2-hr lab

To isolate genomic DNA from liquid cultures of Salmonella,Qiagen DNeasy mini-prep spin columns were used permanufacturers instructions. We designed primers to amplify491- and 487-bp fragments that flank Hismutations of hisG46(in TA1535) and hisD3052 (in TA1538), respectively. Base-pairassignments denote position relative to the start codon of therespectiveHis coding region (Figure 1): hisG (sense)59-CGCTTTACGCATAGCT-39 (bp 16); hisG (antisense)59-AGCTTCAAGCGTCGC-39 (bp 537); hisD (sense)59-CCGTCTGAAGTACTG-39 (bp 706); and hisD (antisense)59-TCAATGGTTGATGCC-39 (bp 1223). Each thermocycler reaction contains21 ml nuclease-free water, 1 ml each of sense and antisenseprimers for the appropriate gene (at 12.5 pmol/ml), 2 mltemplate DNA, and 25 ml of clear Promega master mix (3 mmMgCl2, 400 mm dNTPs, and 50 units/ml Taq polymerase). Theoptimal annealing temperatures for hisG and hisD primersnecessitate separate thermocycling for each set. Thermocyclerprogramming is 5min at 94, 30 cycles of 2min at 94, 30 sec at38 (hisG primers)/44 (hisD primers), 2min at 72, and a finalelongation time of 20min at 72. Following PCR amplification,amplicons are purified using Qiagen QIAquick spin columns;gel electrophoresis using 1.5% agarose with TBE can beperformed to confirm the presence of desired fragments.Purified amplicons (1020 ng) and 3- to 5-pmol sense primerwere sent to the Yale University DNA facility for sequencing.Sequence data were analyzed using MacVector and comparedto both the sequences from Salmonella LT2 (wild type) andTA1535/TA1538 (NCBI). For class use, DNA sequences can becompared using NCBIs BLASTengine (http://blast.ncbi.nlm.nih.gov/Blast.cgi). Full text lab handouts may be obtainedfrom http://www.lawrence.edu/fast/DESTASIE/protocols.html.

RESULTS

Bruce Ames and colleagues produced a number ofSalmonella strains with mutations in the his operon, twoof which we recommend for use in the undergraduategenetics lab. We identified strains with different His

mutations so students could see for themselves howsecondarymutations lead to reversion of amissense vs. a

Figure 1.The hisG and hisD genes of S. typhi-murium and the location and sequence of the mu-tation sites found in TA1535 and TA1538,respectively. Amino acid (AA) sequences areshown under the base sequences. Mutations con-ferring the His phenotype are shown in bold-face type and are surrounded by bases and AAsto indicate the effect of the mutation on the pro-tein sequence. PS and PA denote binding sites forsense and antisense primers used in PCR. Baselocations are relative to the first base of the re-spective coding regions.

Genetics Education 25

frameshift mutation. After experimenting with fiveSalmonella strainsTA98, TA100, TA102, TA1535, andTA1538we identified two strains for which assaysprovided reproducible, countable numbers of sponta-neous revertant colonies from easily measured quanti-ties of overnight cultures. In three replicate experiments,both TA100 and TA102 produced .100 very largespontaneous revertants per plate; the excessive numberand size of colonies produced made accurate countingdifficult as colonies often merged. Strains TA98,TA1535, and TA1538 each produced between 10 and50 spontaneous revertants per plate. TA1535, contain-ing a missense T / C allele of hisG (hisG46), andTA1538, containing a 1 frameshift mutation in hisD(hisD3052), were selected for use, and all data that followare derived from experiments with these two strains.To accurately calculate reversion frequencies, experi-

ments were conducted to determine the average CFUcontent of overnight cultures. Overnight cultures ofTA1535 consistently grew to a slightly higher densitythan did TA1538, with an average of 4.53 108 comparedto 3.9 3 108 CFUs/ml, respectively (n 36). TA1535and TA1538 colonies looked similar after 24 hr ofincubation. Excessive incubation periods (.24 hr)produced large colonies in both strains, which mergedand made colony counting difficult.It is desirable to have students use knownmutagens as

positive controls that have the lowest possible toxicity tohumans and that have different mutation spectra. NaN3and 4NOP were identified for this purpose. NaN3 isknown to inducemissensemutations (Olsen et al. 1993)while 4NOP primarily induces deletions and insertions(Ames et al. 1975). Because TA1535 contains a missenseallele (of HisG) and TA1538 contains a frameshift allele(of HisD), each mutagen should produce a differentreversion frequency with the two Salmonella strains.

As expected, NaN3 and 4NOP did produce revertantsat very different frequencies in the two Salmonellastrains used; each treatment exceeded only the sponta-neous reversion rate in the expected strain. The spon-taneous reversion rate of TA1535 was 2.3 3 107/CFUwhile TA1538 spontaneously reverted at a rate of 1.4 3107/CFU (Figure 2). After testing mutagen concen-trations ranging from 1.0 to 0.0001 mg/ml, we usedconcentrations that yielded reversion rates significantlyhigher than the spontaneous rate, but for which thenumber of revertant colonies per plate could be enu-merated by hand. Significant increases in the reversionrate to 8.9 3 107 and 2.1 3 106 revertants/CFUwere observed in TA1535 after treatment with 0.025and 0.05 mgNaN3/ml, respectively. Similarly, treatmentof TA1538 with 0.25 and 0.5 mg 4NOP/ml producedsignificant increases in the reversion rate to 1.6 3 106

and 2.73 107 revertants/CFU, respectively. Treatmentof TA1535 with 4NOP and of TA1538 with NaN3 pro-duced reversion rates that were virtually identical to thespontaneous reversion rate for each strain. Students willnote that colonies of strain TA1535 are larger than thoseof TA1538. Students should also note that, after treat-mentwith 4NOP, strainTA1538produces large numbersof small colonies that aggregate around the center of theplate, indicating the lower diffusion rate of 4NOP. At therecommended concentrations, kill zones surroundingthe filter paper disc will not be observed with eitherstrain.It should be noted that the calculated reversion

frequencies will be more rigorously quantitative for thespontaneous reversion frequency and comparisons be-tween mutagens more difficult. Mutagen-induced fre-quencies should be considered conservative estimatesdue to possible differences in mutagen diffusion ratesthrough the top agar. In our hands, water-soluble sub-

Figure 2.Reversion frequency,measured in revertants/CFU, ofTA1535 and TA1538 after exposureto H2O (negative control) and indi-cated concentrations of NaN3 and4NOP. n 15; error bars, 1 SE.

26 N. Goodson-Gregg and E. A. De Stasio

stances appear to diffuse uniformly from the centrallyplaced discs, as indicated by uniform distributions ofrevertants across the plate. Students will, however, notethat 4NOP does not fully diffuse across the plate (it isminimally water soluble) and revertants will be clusteredwithin an 3-cm radius of the disc. This fact does notdiminish the importance of the quantitative data anal-ysis as one can still rigorously compare the response ofthe two different Salmonella strains to 4NOP and seethat reversion frequencies differ substantially. Studentsshould be led to understand that reversion frequenciesof 4NOP and any other less soluble substances should bereported as minimal reversion frequencies as CFUsplated outside the diffusion zone will not have beenexposed to mutagen, while bacteria within the diffusionzone may be exposed to a gradient of mutagen. It is agood lesson for students to realize that even quantitative

results are subject to bias of various sorts and must bebased on experiments where as many variables as pos-sible are controlled. Students should be encouraged tobrainstorm other methodologies that would ensureuniform distribution of mutagens, such as inclusion inthe top agar before plating (although the increasedtemperature here might be problematic). We note thatwe do not use other methodologies in the teaching labbecause we prefer to minimize the spread of toxic mu-tagens to glassware that is handled by undergraduates.The disc method keeps the mutagen application in adefined space with minimal spread around the lab.To determine whether students will see significant

mutagenesis after treatment with substances of theirchoice, we subjected 24 substances to a qualitative assay(Table 1) using our updated protocol. Perishable foodproducts were plated on VBMwithout Salmonella strainsto confirm that they did not contain other microorgan-isms; only bread and chewing tobacco produced bacte-rial colonies. Two application procedures were used foreach substance, and treatments that produced at least a50% increase in the reversion frequency were quantita-tively tested for potential mutagenicity (Figure 3). Fivesubstances were found to be potentially mutagenic in atleast one strain (Figure 3). Each treatment producedsignificant mutagenesis in TA1538, but only directapplication of melted Carmex lip balm induced muta-genesis in both strains. Carmex treatment increased thereversion frequency to 5.3 3 107 revertants/CFU inTA1535 and 6.4 3 107 revertants/CFU in TA1538.We started a database of variousmutations that confer

the His1 phenotype, to which students will add data.Genomic DNA was prepared from TA1535 and TA1538as described and sequenced to confirm the integrity ofour initial strains. The resulting sequences were exactly

TABLE 1

Different environmental substances qualitatively tested forpotential mutagenicity

Substances tested AC Substances tested DZ

Artificial vanilla flavoring DeodorantArtificial sweetner Diet colaBread crust Ethidium bromideBurnt bread crust Hair dyeBurnt hamburger Hand sanitizerCarmex Organic coffee (caffeinated)Celery Peanut butterCigar tobacco Rat poisonCigar ashes Stone-ground mustardChewing tobacco Super glueCoffee (caffeinated) Tea (caffeinated)Cooked hamburger Tea leaves (caffeinated)

Figure 3.Reversion frequency,measured in revertants/CFU, ofTA1535 and TA1538 after exposureto H2O (negative control) and vari-ous suspected environmental muta-gens. Application procedure was10 ml on filter paper or direct appli-cation (DA). n 15 for H2O treat-ments and n 3 for all othertreatments. Error bars, 1 SE.

Genetics Education 27

as described (NCBI accession J01804). Five spontane-ous revertants and at least five mutagen-induced rever-tants from each strain were prepared from two separateexperiments and were sequenced to determine typicalmutation tendencies. Variation in the kinds of muta-tions conferring the His1 phenotype was observed(Table 2 and Table 3).Two different reversion mutations were found in a 2-

base region of hisG in strain TA1535 (Table 2), directlyreverting the original His mutation (site 2) or chang-ing the C 1 base upstream (site 1). Reversion to the LT2(wild-type) sequence of hisG, a C/ T transition at site 2was the most common mutation (40% of sequencedhisD alleles), replacing the mutant proline with wild-type leucine. Three isolates of a C / A transversionwere also observed in TA1535 at site 2, encoding ahistidine at the same site of the enzyme. A revertantfound after treatment of TA1535 with hand sanitizerincluded a C/ T transition at site 1 in two revertantcolonies, and an identical mutation was observed in onespontaneous revertant colony. These alleles substitute aserine for the mutant proline.A great deal of variation was observed in the hisD base

sequence of TA1538 revertant alleles. The site of indelmutations (insertions and deletions) spanned a 30-baseregion encompassing the original mutation site. A sam-pling of these mutations is presented (Table 3). 4NOPwas theonly treatment that produced the samemutationtwice. The mutations conferring the His1 phenotype inTA1538 are best characterized by the net effect of theinsertions and deletions found in each allele, relative to

the wild-type (LT2) hisD sequence (Table 4). All re-vertant alleles did restore reading frame, but someadded new amino acids as well. The most commonmutation was a net deletion of 3 bases, which restoredthe correct reading frame (50% of alleles). Students willfind it interesting that none of the revertant colonies ofTA1538 contained a wild-type sequence for the hisDgene. A sequence database, including sequence of wild-type alleles, is available at http://www.lawrence.edu/fast/DESTASIE/protocols.html.We optimized the protocols for characterizing molec-

ular lesions tomaximize the success rate.We define finalsuccess as the generation of a positive sequencing result(base sequence) that could yield a His1 phenotype.Agarose gel electrophoresis was performed after eachintermediate step leading to base sequence analysis(genomic DNA preparation, PCR amplification, andPCR product purification) to ensure the presence ofdesired products. After all of these treatments, DNAfragments of the appropriate size (Figure 4) wereobserved from DNA preparations from revertant strainsthat were prepared with our optimized protocol and theappropriate primer set 100% of the time.DNA concentrations of the final purified PCR prod-

ucts were calculated by both gel electrophoresis andspectrophotometric analysis at 260 nm. Typical DNAconcentrationswere found tobebetween5and20ng/ml.To prevent lab congestion, save time, and reducematerial expenses, students do not need to determinethe DNA concentration of their final product. Instead,all DNA concentrations are assumed to be 10ng/ml afterPCR amplicon purification, and samples were preparedas per the off-site sequencing facilitys recommendation.Of 38 samples prepared in this manner and sequenced,only 1 did not produce a positive result, and a positiveresult was obtained after this sample was sequencedagain.

DISCUSSION

Wedesigned andfield-tested a 2-week investigative labfor an intermediate-level undergraduate course ingenetics using the classical Ames test as a starting point.

TABLE 2

Different molecular lesions conferring the His1 phenotype instrain TA1535 after treatment with various substances

Treatment C/ T C/ A

H2O 3 2NaN3 2 0Organic coffee 0 1Hand sanitizer 2 0

TABLE 3

Samples of insertion and deletion reversion mutations of TA1538 and the net effect of these molecular lesions

Treatment DNA sequence (base 880910) bp insertion bp deletion Net D

LT2 (wild type; none) cgcgcggacaccgcccggcaggccctgagc 0 0 0TA1538 (His mutant; none) cgcgcggacaccgcc ggcaggccctgagcg 0 1 1H2O (spontaneous revertant) cgcgcggacaccgcc ggccaggccctgagc 1 1 04NOP c ggacaccgccgcggcaggccctgagcgcc 2 5 3Organic coffee cgcgcggacaccgcc ggcagtagaacgtca 10 1 9

A 30-base sequence is shown from base 880, relative to the start codon of hisD. All bases are contiguous; gaps indicate deletionsof bases relative to LT2 sequence. Boldface letters indicate insertion. Net D, net effect of insertions and deletions in number ofbases relative to LT2.

28 N. Goodson-Gregg and E. A. De Stasio

The lab allows students to play a role in experimentaldesign, and it is open ended and discovery based.Students produce both quantitative data that can bestatistically compared and qualitative sequence data thatcan be analyzed for effect on amino acid code andreading frame. The main advantages to the enhancedlab are its emphasis on quantitative analysis and theexplicit connectionmade between reversion phenotypeand gene sequences.We have increased the efficacy, rigor, and quantitative

aspects of the lab compared to the usual implementa-tion of the classic Ames test. Students directly measure,on the basis of serial dilutions of their own design, CFUsin overnight cultures to calculate reversion frequenciesrather than simply count revertant colonies. This is aclassical genetic measurement and one that under-graduates often have difficulty grasping. The techniqueof serial dilution is often equally mysterious to students.Measuring and calculating both dilutions and usingthem to calculate frequencies aids student understand-ing of these quantitative techniques (see assessmentdiscussion below).The classic Ames test was further extended to include

a molecular genetic investigation, allowing students tosee firsthand the direct connection between a pheno-type (His and His1) and DNA sequences. Studentschoose a revertant colony for study, extract genomicDNA, use PCR to amplify large regions of the affectedgenes, use a core facility to produce DNA sequence,align and analyze the resulting sequences, and correlateto phenotype. The procedures that we designed havebeen 97% effective in producing usable molecular data(base sequences). DNA preps using Qiagens DNAeasykit, PCR done with Promegas clear master mix, and theindicated thermocycling have never failed with ourprotocols, obviating the need for an intermediate stepof gel electrophoresis. DNA sequences and chromato-grams arrive from core facilities as electronic files, andfreely available software is used to align and comparesequences. Students will still need to use their owncommon sense, however, when comparing frameshiftalleles, as the simple alignment tool does not alwayssee insertions or deletions as well as the human eye.

We suggest the use of Salmonella strains with differ-ent His mutations so students can observe for them-selves the connection between DNA sequence andphenotype, the types of lesions induced by differentmutagens, and how the secondary mutations lead toreversion of a missense vs. a frameshift mutation. Use ofboth NaN3 and 4NOP as positive controls allowsstudents to discover that mutagens can have differentmutation spectra. TA1535 is not significantly reverted by4NOP. NaN3 does not produce significant revertants ofTA1538, demonstrating that different types of muta-tions are needed for reversion in these two strains.When students receive and analyze their revertantsequences, they will note that 4NOP does indeed pro-duce specific frameshift mutations and that NaN3exposure is correlated with missense mutations.This lab lends itself very well to written analysis of the

data inmanuscript form. Tomaximize student learning,class data should be pooled to some degree so thatvariability in overnight culture concentration is obviousand the need to account for it by calculating reversionfrequency, rather than by simply comparing numbers ofrevertant colonies directly, is clear. In addition, a largerdata set will allow students to see firsthand that (1)reversion in each strain requires a different type ofrevertant mutation, (2) different mutagens and the

TABLE 4

The net effect of different molecular lesions conferring theHis1 phenotype in strain TA1538 after treatment with various

substances relative to wild-type LT2 sequence

Treatment Deletion of 3 Deletion of .10 Insertion

H2O 2 1 24NOP 2 0 0Organic coffee 0 0 1Artificial vanilla 1 1 0Hand sanitizer 1 1 0

Figure 4.Agarose gel (1.5%) of purified PCR products.Lanes 13 were amplified with the hisG primer set, and lanes46 were amplified with the hisD primer set. M, Hi-Lo DNAmarker; 13, TA1535 overnight culture, spontaneous rever-tant, and 0.05 mg/ml NaN3-induced revertant, respectively;46, TA1538 overnight culture, spontaneous revertant, and0.5 mg/ml 4NOP-induced revertant, respectively.

Genetics Education 29

spontaneous mechanisms have particular mutationspectra, and (3) multiple DNA sequences can encodea functional enzyme. We suggest that instructors begin adatabase of revertant sequences generated by theirclasses to which each years class will add data. Instruc-tors could manipulate data release to have studentscompare data sets of different sizes such that studentscan see how collecting larger data sets allows scientists tosee trends more clearly.The learning outcomes from this lab are numerous

and instructors can use data generated in this lab atmany different points in a typical genetics course. Theuse of model organisms as test subjects due to theuniversality of DNA structure and function should bemade explicit at the outset of the lab. The concept ofreversion is often difficult for students to grasp, butadding the molecular analysis to this lab makes it clearthat multiple types of mutations can create a functionalgene product from the original His genotype. Instruc-tors can then build on student knowledge from this labwhen introducing the concept of suppressor screens.The connection between phenotype, genotype, andalleles will be explicit and tangible since students willhave produced data at all three levels themselves.Instructors may then make connections between thegene sequences and protein structure, molecular evo-lution, and mechanisms of mutation as they desire.They can refer to this lab when discussing the impor-tance of DNA repair (these strains have a uvrBmutation,which eliminates the excision repair mechanism) andthe error-prone nature of DNA replication and repair.Further extension of this lab is possible. S9 rat liver

extract can be incorporated into the histidine-deficientmedia to mimic the metabolic action of the mammalianliver (with additional expense). Students could designthe PCR primers and optimize PCR or they could seetheir PCR amplicon in an agarose gel prior to or afteramplicon purification.To test the efficacy of this lab, 10 summer research

students were recruited to field-test the lab; 6 of thesestudents had taken a genetics course the previous year(and performed a traditional, qualitative Ames test) andthe rest had not. Students were given two pretests: a self-assessment of knowledge and a problem-based objectivetest in which half of the questions tested quantitativeskills (e.g., dilutions, frequency calculations) and therest covered concepts that connect classical and molec-ular genetics (e.g., what is a reading frame?). A shortprelab lecture on the development and utility of theAmes test and the concept of reversion analysis pre-ceded implementation of the lab. Students had nodifficulty following the procedures, although manyneeded help in determining how to prepare and useserial dilutions, in spite of having done lab-basedresearch full time for the previous 10 weeks. Three dayslater, students collected data and determined reversionfrequencies. To shorten the volunteer time needed,

students were provided with sequence data for fourmutant alleles and asked to compare them to wild-typesequence and determine the molecular basis of therevertant phenotype. Each pair of students was asked tosummarize their findings orally. Students were espe-cially interested in the fact that a simple base sub-stitution mutation produced a His1 reversion inTA1535, but complex indels were required to restorethe reading frame and confer the His1 phenotype inTA1538. Following the summaries, the students com-pleted the post-tests and a series of questions evaluatingthe lab directly.Pre- and post-test scores demonstrated students

marked improvement in understanding both the quan-titative and the molecular aspects of the lab (Figure 5 ).Pre- and post-tests were coded and mixed by a thirdparty prior to scoring. Pretest scores indicated that evenstudents who had taken a genetics course the previousyear had difficulty with the quantitative analysis andserial dilutions associated with this lab, skills not used inthe traditional Ames test that they had used the previousyear. In spite of the fact that one of our introductorybiology labs has students undertaking serial dilutions(with explicit directions not designed by the students),no student scored perfectly on this question on thepretest, underscoring the need to have students practicethis skill repeatedly and to design the dilutions them-selves with a particular goal in mind. Paired scores wereraised by an average of 6 points of 19 total (Figure 5).Average pre- and post-test scores were significantlydifferent (P , 0.001, paired two-sample t-test).Self-assessment of student knowledge and skills also

increased from pre- to post-lab. Students anonymouslyrated their understanding of 12 concepts or skills on a

Figure 5.Average student scores on an objective assess-ment of concepts and quantitative skills before and after com-pleting the laboratory exercise. Sample questions includeDescribe how to use serial dilution to produce 100 bacterialcolonies on a plate when the starting bacterial overnight cul-ture is expected to contain 106 bacteria per ml. Assume youplace 100 ml of your final dilution on the agar plate andWhat is meant by reading frame? Scores are out of 19 pos-sible points. n 10. Error bars 1 SD.

30 N. Goodson-Gregg and E. A. De Stasio

five-point scale (1, very poor understanding; 2, poorunderstanding; 3, neutral understanding; 4, good un-derstanding; 5, excellent understanding), includingsuch questions as I feel comfortable calculating muta-tion frequencies and I understand how to use andanalyze a DNA sequence. Students confidence in-creased after completion of the lab and its analysis;ratings increased by an average of 1.6 points on the five-point scale. The differences between average pre- andpost-lab responses for each question were all signifi-cantly different (P, 0.05, paired two-sample t-test), withthe exception of the question I understand how to useand analyze DNA sequence (P , 0.07). Studentresponses were abnormally high to this question in theprelab self-assessment (3.8), but nonetheless, an in-crease (to 4.8) was observed after the lab and analysis.Anonymous student assessment of the lab itself was

very positive. Average responses to questions such asthis lab shouldbepart of theGenetics curriculum, andOther students would benefit from the skills learned inthis lab were rated, on average, 4.4 and 4.5 on a five-point scale (1, strongly disagree; 2, somewhat disagree;3, neutral; 4, somewhat agree; 5, strongly agree). Inspace reserved for written comments about the lab,students wrote Having my own input, my mutagen, Iwas more excited to learn the results and why theyoccurred, and I like the addition of sequencing, usingmy own potential mutagen. Students also commentedon the skill-building aspects of the lab: The addition ofthe quantitative component enhanced [my] under-standing of the concept of reversion frequency and[There were] a variety of techniques from microbiol-ogy, molecular biology, and genetics. This lab reallybrings it all together.In conclusion, we developed a new multi-week in-

vestigative lab that addresses two very important issuesin a genetics curriculum. The lab adds rigorous quan-titative analysis as well as molecular genetics to theclassic Ames test. The lab has been well received bystudents and has been shown in a pilot study to improvestudent skills and knowledge on an objective test as wellas student confidence on the basis of self-assessments.

The authors thank JoAnn Stamm and Wayne Krueger for technicalassistance and the 2008 summer research students for field-testing thelab. This manuscript benefited from the comments of two anonymousreviewers. The authors received a curricular development grant fromthe Provosts office at Lawrence University.

LITERATURE CITED

Abu-Shakra, A., E. T. McQueen and M. L. Cunningham,2000 Rapid analysis of base-pair substitutions induced by muta-genic drugs through their oxygen radical or epoxide derivatives.Mutat. Res. 470: 1118.

Ames, B. N., 1979 Identifying environmental chemicals causing mu-tations and cancer. Science 204: 587593.

Ames, B. N., H. O. Kammen and E. Yamasaki, 1975 Hair dyes aremutagenic: identification of a variety of mutagenic ingredients.Proc. Natl. Acad. Sci. USA 72: 24232427.

Hack, C., and G. Kendall, 2005 Bioinformatics: current practiceand future challenges of life science education. Biochem. Mol.Biol. Educ. 33: 8285.

Hake, R. R., 1998 Interactive-engagement versus traditional meth-ods: a six-thousand-student survey of mechanics test data for in-troductory physics courses. Am. J. Phys. 66: 6474.

Handelsman, J., D. Ebert-May, R. Beichner, P. Bruns, A. Changet al., 2004 Education: scientific teaching. Science 304(5670):521522.

Handelsman, J., S. Miller and C. Pfund, 2007 Scientific Teaching:The Wisconsin Program for Scientific Teaching. Roberts & Co andW. H. Freeman, New York.

Koch, W. H., E. N. Henrikson, E. Kupchella and T. A. Cebula,1994 Salmonella typhimurium strain TA100 differentiates severalclasses of carcinogens andmutagens by base substitution specific-ity. Carcinogenesis 15: 7988.

Levine, J. G., R. M. Schaaper and D. M. DeMarini, 1994 Complexframeshift mutations mediated by plasmid pKM101: mutationmechanisms deduced from 4-aminobiphenyl-induced mutationspectra in Salmonella. Genetics 136: 731746.

Marshall, P. A., 2007 Using Saccharomyces cerevisiae to test the mu-tagenicity of household compounds: an open ended hypothesis-driven teaching lab. CBE Life Sci. Educ. 6: 307315.

Merkel, C., 2003 Undergraduate research at the research universi-ties. New Dir. Teach. Learn. 93: 3953.

Metz, A., 2008 Teaching statistics in biology: using inquiry-basedlearning to strengthen understanding of statistical analysis in bi-ology laboratory courses. CBE Life Sci. Ed. 7: 317326.

Myers, M. J., and A. B. Burgess, 2003 Inquiry-based laboratorycourse improves students ability to design experiments and in-terpret data. Adv. Physiol. Educ. 27: 2633.

National Research Council, 2003 BIO2010: Transforming Under-graduate Education for Future Research Biologists, edited by L.Stryer. National Academy Press, Washington, DC.

Olsen, O., W. Xingzhi and D. V. Wettstein, 1993 Sodium azide mu-tagenesis: preferential generation of A.T/G.C transitions in theberley Ant18 gene. Proc. Natl. Acad. Sci. USA 90: 80438047.

Takiya, T., Y. Horie, S. Futo, Y. Matsumoto, K. Kawai et al.,2003 Rapid selection of non-hotspot mutants among hisD1 re-vertants of Salmonella typhimurium TA98 in Ames test by peptidenucleic acid (PNA)-mediated PCR clamping. J. Biosci. Bioeng.96: 588590.

Wessner, D. R., P. C. Maiorano, J. Kenyon, R. Pillsbury and A. M.Campbell, 2000 Spot-overlay Ames test of potential mutagens.Assoc. Biol. Lab Educ. 22: 118.

Communicating editor: P. J. Pukkila

Genetics Education 31