Characterization of Pathogenic Human MSH2 Missense Mutations Using Yeast as a Model System: A Laboratory Course in Molecular Biology

INTRODUCTION

Seminal reports from committees examining undergraduate education in the research sciences strongly support an inquiry-based approach to learning (National Research Council Committee on Undergraduate Biology Education to Prepare Research Scientists for the 21st Century, 2003; The Boyer Commission on Educating Undergraduates in the Research University, 1998). The course outlined in this paper is designed to use an inquiry-based approach to teach students a variety of current research laboratory methods in the context of a novel research project. Students gain practical skills and theoretical understanding while experiencing the progression and complexity of a real research project.

Future research biologists are predicted to need skills from many disciplines, including mathematics, physics, chemistry, engineering, and computer science (National Research Council Committee on Undergraduate Biology Education to Prepare Research Scientists for the 21st Century, 2003). The training of these multi-faceted future research scientists requires an integrated curriculum with courses that are accessible and interesting to a variety of students. The Advisory Committee to the National Science Foundation Directorate for Education and Human Resources (1996) and the authors of a report for Project Kaleidescope (Rothman and Narum, 1999) encourage endeavors to educate a variety of science students. The subject for the project outlined in this paper was selected to capture the interest of a broad spectrum of enrolled students. The topic is centered on a highly conserved, fundamental biological process with a direct impact on a significant human disease.

Background for the Research Project

In the United States in the year 2003, the American Cancer Society estimates that colorectal cancer (CRC) will be diagnosed in 147,500 individuals and will result in 57,100 deaths. Approximately 2–7% of these CRC cases are a consequence of the most commonly inherited form of the disease, hereditary non-polyposis colorectal cancer (HNPCC) (reviewed recently in Boland, 2000; Muller and Fishel, 2002; Peltomaki, 2001). As far back as 1913, HNPCC was recognized as having a genetic basis (Warthin, 1913). Eighty years later, an exciting convergence of basic science and medical research proved that HNPCC results from defects in DNA mismatch repair (MMR) genes (Aaltonen et al., 1993; Fishel et al., 1993; Ionov et al., 1993; Leach et al., 1993; Parsons et al., 1993; Peltomaki et al., 1993; Strand et al., 1993; Thibodeau et al., 1993).

DNA MMR is a highly conserved process (reviewed recently in Harfe and Jinks-Robertson, 2000; Marti et al., 2002). Early work in prokaryotes revealed that certain genes when inactivated led to a pronounced increase in the rate of spontaneous mutations (Cox et al., 1972; Nevers and Spatz, 1975; Siegel and Bryson, 1967). Identification of several of these genes including mutH, mutL, mutS, and uvdR/mutU established the importance of prokaryotic methyl-directed DNA mismatch repair in maintaining genomic integrity (Bauer et al., 1981; Glickman, 1979; Glickman and Radman, 1980; Jones and Wagner, 1981; Lu et al., 1983; Pukkila et al., 1983; Rydberg, 1978; Shanabruch et al., 1983).

In nature, DNA mismatches form primarily as a consequence of replication errors and chemical modification (reviewed recently by Aquilina and Bignami, 2001; Kunkel and Bebenek, 2000). Without an intact mismatch repair system, DNA begins to accumulate mutations, some of which will be deleterious to the cell or organism. In addition to the increased rate of single base-pair mutations, researchers observed that repetitive DNA elements (for example mono, di and tri-nucleotide repeats) are highly unstable in mismatch repair defective strains (Levinson and Gutman, 1987; Strand et al., 1993). Interestingly, these findings contributed to the establishment of the link between HNPCC and mismatch repair, given that a hallmark of HNPCC tumors is highly unstable repetitive DNA elements (Aaltonen et al., 1993; Ionov et al., 1993; Parsons et al., 1993; Thibodeau et al., 1993).

Closely related eukaryotic homologs of MutS (MutS Homolog or MSH genes) (Reenan and Kolodner, 1992) and MutL (MutLHomolog or MLH and some Post-Meiotic Segregation or PMS genes) (Horii et al., 1994; Lindblom et al., 1993; Nicolaides et al., 1994; Papadopoulos et al., 1994) are found in diverse organisms ranging from yeast to humans. While MutS functions as a homodimer to bind mispaired DNA in prokaryotes, the similar function is accomplished in eukaryotes by two distinct heterodimers known as MutS α, composed of Msh2 and Msh6, and MutSβ, composed of Msh2 and Msh3 (Figure 1). MutSα binds primarily to single-base pair mismatches and small insertion/deletion loops. MutS β binds to larger insertion/deletion loops. A heterodimer composed of Mlh1 and Pms1 (PMS2 in humans) functions as the primary MutL equivalent in DNA mismatch repair in eukaryotes. The MutH endonuclease confers stand specificity for repair in prokayotes, where the newly synthesized unmethylated strand is cleaved for repair. A strand specificity factor in eukaryotes is postulated to be Pol30/PCNA, although the precise mechanism of stand discrimination is not clear. Downstream events including strand displacement, excision and new synthesis are currently being investigated and include helicases (Mechanic et al., 2000; Yamaguchi et al., 1998), exonucleases (Burdett et al., 2001; Genschel et al., 2002), polymerases, and ligases (Tomkinson and Mackey, 1998).

After the establishment of the connection between HNPCC and DNA MMR, many germline mutations have been mapped in afflicted families. Most mutations reside in one two genes, MSH2 and MLH1. Individuals inheriting a defective mismatch repair locus have an 80–90% chance of manifesting cancer before the age of 50 (Lynch and Lynch, 2000). The high penetrance and early age of onset of the disease underscores the need for aggressive identification and screening of susceptible individuals.

The focus of this research project is MSH2. Since the discovery of the role hMSH2 in HNPCC many alterations in the gene have been found in afflicted families.¹ Included among these changes is a collection of nearly 30 missense mutations. With missense mutations one must always consider that a single amino acid change does not necessarily result in a dysfunctional protein. Certain changes are functionally silent. In some cases, there are not sufficient family members to do a careful pedigree analysis to ensure that the mapped mutation is indeed pathogenic. It is part of the objective of the course to determine whether the mapped missense mutations found at hMSH2 loci result in a dysfunctional protein (pathogenic polymorphisms) or if the missense changes are likely to be benign polymorphisms in the human population. If missense mutations are to be used in genetic counseling, it is imperative to determine which changes result in a defective protein. We also use the mutations that have been found in clinical isolates to define important structure-function relationships in the Msh2 protein.

1Compiled by the Human Gene Mutation Database, Copyright © University of Wales College of Medicine 2003)

MATERIALS AND METHODS

Basic Molecular and Microbial Techniques

The mutagenesis procedures are provided in detail elsewhere (Muta-Gene Phagemid In Vitro Mutagenesis, Version 2, Bio-Rad Laboratories, Hercules, CA) (Ausubel et al., 1994). The mutagenesis components not commercially available include the mutagenic oligonucleotides (Princeton Syn/Seq Facility) and template DNA (see Figure 2). Transformation of CaCl₂ and electro-competent bacterial cells and subsequent microbial manipulations are conducted according to Ausubel et al. (1994). Small-scale preparations of plasmid DNA are accomplished by the alkaline lysis method (Ausubel et al., 1994). Large-scale preparation of plasmid DNA is achieved by using anion exchange chromatography (Midi-prep procedure, Qiagen Inc., Valencia, CA). Polymerase chain reactions (PCR) are conducted using standard conditions (Ausubel et al., 1994). Plasmid DNA and PCR fragments are digested using the conditions specified by the manufacturer (New England Biolabs, Inc., Beverly, MA). Agarose gel reagents and electrophoresis parameters are detailed elsewhere (Ausubel et al., 1994). Missense mutations are sequenced at the nucleotide level using dideoxy nucleotide sequencing Sequenase Version 2 chain-termination sequencing reagents, [α-³⁵S]dATP (Amersham Biosciences, Piscataway, NJ), MSH2 specific primers (Princeton Syn/Seq facility) and template DNA.

Preparation of Yeast Strains for the Mismatch Repair Functional Assays

Mutagenized plasmid, pMSH2 and pRS413 are used to transform AGY75 (Table 1), a reporter strain constructed prior to the course using the lithium acetate method (Burke et al., 2000). The selective medium is lacking histidine (–HIS) to select for transformants harboring the mutagenized plasmid, pMSH2, or pRS413. The medium also is missing tryptophan (–TRP) to select for the pSH44 dinucleotide reporter construct (described in Figure 4 and 5A) as well as leucine (–LEU) and threonine (–THR) to induce expression of the GT_16.5–URA3 fusion transcript from the LEU2 inducible promoter. Finally, the medium lacks uracil (–URA) to ensure that no instabilities in the dinucleotide repeat had been incorporated before the functional assays are initiated. These selective plates are described as– HIS –TRP –LEU –THR –URA. All yeast manipulations, including media preparation, are conducted as described previously (Burke et al., 2000).

Qualitative Mismatch Repair Assays

Three colonies from each of the transformations are transferred using sterile wooden sticks to a defined position approximately 20 mm in diameter on a fresh –HIS –TRP –LEU –THR –URA plate. After 2 d of growth at 30°C, yeast strains are replica-plated onto drug-containing media to assess the mismatch repair phenotype (described in detail below) and to –HIS –TRP medium as a control to ensure that the cells were properly delivered to all of the plates.

The drug selection for dinucleotide instability is synthetic medium lacking histidine, tryptophan, leucine, threonine and supplemented with 5-fluororotic acid monohydrate, 5-FOA (Toronto Research Chemicals Inc., North York, ON, Canada). The medium is designated –HIS –TRP –LEU –THR + 5-FOA to reflect the nutritional deficiencies and the presence of the drug 5-FOA. A second type of medium to score primarily for single base pair mismatch repair defects is supplemented with 60 μg/ml canavanine, CAN (Sigma-Alderich, Milwaukee, WI). The medium is also lacking histidine, tryptophan, arginine and is designated –HIS –TRP –ARG + CAN. Images of the plates are taken with a FOTO/Analyst Fotodyne Camera and FOTO/Analyst PC Image software (Fotodyne, Hartland, WI) and saved to a network drive so that the students are able to access their images from remote computers.

Quantitative Assays for Mismatch Repair

The same strains employed in the qualitative test above are used to inoculate liquid synthetic medium lacking histidine and tryptophan (–HIS– TRP). Cultures are grown overnight at 30°C with aeration. The approximate cell concentration is determined (Burke et al., 2000) and students either concentrate the culture or perform serial dilutions so that the appropriate number of cells is delivered to each plate.

Based on the qualitative assay, students are able to determine whether the msh2 allele has a mismatch repair phenotype that is similar to wild type (MMR+) or similar to msh2 null (MMR–). Based on this information, the cells with the msh2 missense allele are either processed like wild type or like msh2 null cells. The cultures must be treated differently because if the mutation rate is higher (as is the case for the msh2 null allele strain), fewer cells are plated on the drug-containing plates to allow for an accurate colony count. To select for Can^r mutant ∼6 × 10⁷ cells of the MMR + cells are plated on –HIS –TRP –ARG + CAN, by concentrating the cultures in a centrifuge for 5 min at 3000 rpm to allow for efficient plating volumes. Approximately 10⁶ cells for each of the MMR–cells are plated on –HIS –TRP –ARG + CAN. To select for FOA^r mutant∼1.4 × 10⁶ MMR + and∼ 10⁴ of the MMR–cells are plated onto –TRP– HIS –LEU –THR + FOA plates. To obtain a viable cell count for each culture ∼ 100 cells for each of the samples are plated onto– HIS –TRP plates. The plates are placed at 30°C for 2–5 d.

Steady-State Expression Levels of the Msh2p Variants by Immunoblotting

Proteins extracts (Burke et al., 2000) are resolved using SDS polyacrylamide gel electrophoresis (SDS-PAGE) (Ausubel et al., 1994). The resolving gel is a 7% acrylamide gel (37.5:1 acrylamide to bis-acrylamide ratio). Samples and markers are resolved in duplicate halves on the gel. One half is subjected to for Coomassie Brilliant Blue R (Sigma–Alderich) staining and the duplicate is processed for immunoblotting (Ausubel et al., 1994). Immunodetection of Msh2p expressing the hemagglutinin epitope (HA) is conducted according to the ECL method (Amersham Biosciences). The primary antibody used is mouse 12CA5 monoclonal antibody specific for the HA epitope (Princeton Monoclonal Facility). The blot is incubated at room temperature in the presence of the primary antibody for 1 hour. After the specified washes, the secondary antibody, α-mouse-horse radish peroxidase (HRP)-conjugated secondary antibody (Amersham Biosciences) is applied for 45 min. Both antibodies are used at a 1:2500. After the final washes, the ECL detection reagents are applied and the blot is processed as recommended by the manufacturer.

Localization of the Msh2p Variants Using Indirect Immunofluorescence

The procedure for indirect immunofluorescence of HA epitope tagged Msh2 (Msh2::HA) is based on previous protocols (Pringle et al., 1989). Cultures of exponentially growing yeast strains (the same strains used in the functional assays) are fixed in ∼4% formaldehyde for 30 min at 30°C, washed and digested for 1 hour with Zymolyase 100T (Seikagaku Corporation, Japan) as recommended. The primary antibody for detection of Msh2::HA is 12CA5 diluted 1:400. In addition, control wells, are prepared with a mixture of 12CA5 and with rat monoclonal α-tubulin (YL1/2, Serotec, Raleigh, NC) each diluted 1:400. The primary antibody incubation is for 30 min at room temperature. Wells are washed and the secondary antibody mixture is applied for 45 min at room temperature in a moist chamber. The secondary antibody mixture is a combination of goatα -mouse IgG Alexa Fluor 488 diluted 1:200 and donkey α-rat IgG Alexa Fluor 594 diluted 1:400 (both from Molecular Probes, Eugene, OR). Cells are washed extensively, stained with freshly diluted 4',6'-diamidino-2-phenylindole, DAPI (1 μg/ml in PBS) for 5 min at room temperature, washed twice and prepared for microscopy as recommended (Pringle et al., 1989). Cells are examined with a Nikon Eclipse E600 (Nikon, Melville, NY) containing a DIC H Plan Fluor 100X oil immersion objective lens (NA = 1.3). Fluorescence was visualized using Nikon filter sets: UV-2A (for DAPI stained nuclei and mitochondria), B-2A (for Msh2 localization), and G-2A (for tubulin localization). Images are recorded using a Nikon DXM1200 digital camera and ACT-1 software, version 2 (Nikon) and stored as electronic files onto network drives.

Recombinant DNA Cloning

The entire sample containing digested pMSH2-allele construct and 1 μg of digested pGBD-MSH2 are resolved by agarose gel electrophoresis. The appropriate bands are excised and extracted using the Bio101 Gene Clean Spin Extraction Kit (Qbiogene, Carlsbad, CA). The yield of extracted insert and vector DNA is determined by comparing a portion of the sample to a concentration curve of digested pGBD-MSH2. The pmol/μl of 5' termini for each fragment are determined and ligation reactions are prepared such that∼ 3- to 5-fold molar excess of 5' termini of insert DNA compared to vector is present in the sample. A control tube is prepared with pGBD-MSH2 vector fragment alone with no insert fragment. In addition, ligase controls include a singly cut vector with and without T4 DNA ligase. Ligation reaction reactions are conducted according to the specifications of the enzyme manufacturer (New England Biolabs). Products of the ligation reaction and the control samples are used to transform bacterial cells by electroporation (Ausubel et al., 1994). A control tube in which no DNA has been added is also prepared to control for sterility. A final uncut vector control is prepared to compare with the singly cut vector with and without ligase. Screening for recombinant molecules employs basic molecular techniques described above, including the alkaline lysis extraction procedure, PCR, restriction endonuclease digestions, and analytical agarose gel electrophoresis. The control DNA for these diagnostic tests is pGBD-MSH2. In addition, a second diagnostic restriction endonuclease digestion is conducted to verify that the digestion sites used in the procedure are still encoded in the recombinant plasmids.

Yeast 2-hybrid Assays

pGBD-C2, pGBD-MSH2, and the recombinant plasmid containing the fusion between GBD and the msh2 allele (pGBD-MSH2-allele) are introduced into PJ69-4A, a MATa yeast 2-hybrid reporter strain (James et al., 1996) using the lithium acetate transformation method (Burke et al., 2000). Cells are plated onto medium lacking tryptophan (–TRP) to select for the transformants. Before the course, the GAD fusion constructs (Figure 9A) were introduced into PJ69-4α, a MATα yeast 2-hybrid reporter strain (James et al., 1996) and distributed. The students set up crosses to form diploid yeast strains on YEPD agar plates. The plates were incubated for 24 h at 30°C and replicaprinted to selective plates lacking leucine and tryptophan (–LEU– TRP) or lacking leucine, tryptophan, and histidine (–LEU– TRP –HIS). The plates are incubated for 2–5 d at 30°C and are photographed as described above for the qualitative mismatch repair assays.

RESULTS

Overview of the Basic Project

Students work in pairs to study a MSH2 missense mutation that is listed as pathogenic in the human databases (see Table 1 for a listing of alleles generated). The missense mutation is introduced in the cognate yeast MSH2 coding sequence so that the MSH2 gene harboring the mutation is distinguishable from the wild type by restriction endonuclease digestion. Students definitively prove that the mutation has been engineered by nucleotide sequence analysis of the mutated region. Ultimately, the mutant alleles are analyzed using in vivo DNA mismatch repair assays. The remainder of the course is focused on discerning in what way are the altered Msh2 proteins malfunctioning in mismatch repair. Students examine steady-state protein levels by immunoblotting to determine if the altered amino acid might affect the stability of the protein. Additionally, they determine whether the Msh2p variants are properly localized to the nucleus using indirect immunofluorescence. Finally, the students test whether the altered proteins have lost their ability to interact with other subunits of the mismatch repair complex using the yeast 2-hybrid assay.

Mutagenesis Strategy

Using site-directed mutagenesis the students create human missense mutations designated as pathogenic in the cognate yeast MSH2 gene. To introduce the changes the students employ a mutagenesis technique known as the dut ung method (Kunkel, 1985). The technique uses single-stranded uracil-containing template DNA, a mutagenic oligonucleotide to prime DNA synthesis, deoxynucleotides and T7 DNA polymerase to synthesize the mutagenized DNA strand, and T4 DNA ligase to create a covalently closed double stranded DNA plasmid. After the in vitro synthesis, the synthesized plasmids are introduced into a bacterial strain that will preferentially repair the parental uracil-containing strand, resulting in a high yield of transformants containing the mutagenized plasmid. Most of the components for the mutagenesis are commercially available. The unique reagents include the template DNA and the mutagenic oligonucleotides. The MSH2 containing template DNA (pMSH2) was constructed prior to the course and is described in detail in Figure 2 and the mutagenic oligonucleotides are designed as detailed in the following section.

Designing Mutagenic Oligonucleotides

The students begin the project by designing mutagenic oligonucleotides. The procedure relies on Accelrys SeqWeb Version 2, a Web-based sequence analysis program, to aid in the design. Each pair is assigned a missense allele found in the human databases. The databases consulted include the International Collaborative Group on Hereditary Non-Polyposis Colorectal Cancer (ICG-HNPCC), and the Human Gene Mutation Database (in association with Celera). The cognate amino acid is found in yeast gene by comparing the human and yeast sequences. A yeast codon preference table ( ftp://genome-ftp.stanford.edu/pub/codon/ysc.g63.cod) is consulted when planning the mutagenesis to ensure that the mutated protein will not have decreased expression as a consequence of rare codon usage at the altered amino acid site. In addition, the students engineer a mutation that will either create or destroy a restriction endonuclease site within the coding region of the gene. Oligonucleotides are designed to be 23 nucleotides in length with 10 nucleotides flanking the codon to be altered. Finally, the proper strand must be specified so that the mutagenic oligonucleotide will anneal to the single-stranded template DNA using the information given in Figure 2. Before ending the computer session, the students determine the restriction endonuclease digestion pattern that would be predicted for the unmutagenized and mutagenized plasmids.

Site-Directed Mutagenesis

Over several laboratory sessions, the mutagenic oligonucleotide is phosphorylated and used to prime synthesis of the complementary strand of the uracil-containing DNA template in an in vitro mutagenesis reaction. Each student pair prepares three samples. One contains all of the listed components, a second lacks the mutagenic oligonucleotide (no primer control), and a third lacks the T4 DNA ligase (the no ligase control).

One tenth of the reaction products and controls are resolved by agarose gel electrophoresis in the presence of ethidium bromide to determine whether the expected double-stranded covalently closed products are generated before introducing the molecules into bacteria.The students are provided with double-stranded and single-stranded pMSH2 controls for comparison purposes (see Figure 3A for an example of a student's results).

The synthesized molecules from the mutagenesis are used to transform bacterial cells by means of the calcium chloride method. The bacterial strain produces a functional uracil N-glycosylase that allows for selection against the parental, unmutagenized strands. Students transform the bacteria with the complete reaction mixture, and with two controls. One control lacks input DNA (the negative, or sterility control, which should give no transformants), and the second contains double-stranded pMSH2 DNA (a positive control, which should give a high yield of transformants).

Confirming the Mutation by Restriction Endonuclease Digestion of Plasmid DNA and PCR Amplified Fragments

Six transformants are chosen to recover the plasmid DNA for restriction endonuclease screening to confirm the presence of the mutation. Digestions are performed on the whole plasmid and on polymerase chain reaction (PCR) amplified fragments of the mutagenized regions. Many of the relevant restriction endonucleases cleave the pMSH2 plasmid into multiple fragments (some as many as 73 fragments), making the diagnosis difficult. Thus, analysis of the amplified region of interest proves to be a highly effective way for the students to identify the mutagenized plasmids.

Students select two primers from a listing of available primers to amplify the mutagenized region. The goal is to amplify a fragment that upon restriction endonuclease digestion will give a banding pattern distinct from that of unmutagenized DNA. The students are directed to choose primers keeping in mind considerations such as the annealing positions and the orientation of the 3'-OH end of the primers with respect to the altered codon.

The purified plasmid DNA and the unmutagenized plasmid DNA control (pMSH2) as well as the PCR products are digested with diagnostic restriction endonucleases. Students are required to find the correct buffer and temperature conditions for each diagnostic digestion and prepare an agarose gel of a suitable percentage, depending on the desired resolving power. Students identify the mutagenized plasmids by analyzing the digestion patterns and comparing them to the predicted patterns and to the unmutagenized DNA control lanes (see Figure 3B and 3C for examples of student gels). Typically, greater than 50% of the samples analyzed encode the desired mutation.

Dideoxy Nucleotide Sequence Analysis of the Mutagenized Region

As a final confirmation that the mutagenesis was precisely as engineered, the students perform dideoxy nucleotide sequencing reactions of the relevant region using highly purified plasmid DNA. Students must select an appropriate primer from the listing of MSH2 primers stocks to prime the synthesis reaction. The students take into consideration the orientation of the mutation with respect to the position of the primer's 3'-OH end. The primer is chosen so that it will anneal ∼100 nucleotides away from the altered codon to allow for optimal resolution of the sequence in the mutagenized region (see Figure 3D for a student result).

Qualitative and Quantitative DNA Mismatch Repair Assays

After confirmation of the mutagenesis at the nucleotide level, the mutagenized plasmid (containing the msh2 allele of interest), the pMSH2 unmutagenized plasmid (containing wild-type MSH2), and the pRS413 vector control (no MSH2) are used to transform the msh2Δ yeast reporter strain AGY75 (Table 2, Figure 4, and Figure 5A). As a control for sterile technique, the students prepare a transformation sample with no plasmid DNA.

Using the created yeast strains, qualitative and quantitative mismatch repair assays are performed. Two types of drug-containing medium are used to assay DNA mismatch repair. The drugs select for yeast cells that have incorporated mutations into their DNA at an elevated rate. In these assays, an increased number of survivors on the drug-containing medium reflects a pronounced defect in DNA mismatch repair. The selection for dinucleotide instability (typically seen as a consequence of MutSβ loss of function) is synthetic medium supplemented with 5-fluororotic acid monohydrate, 5-FOA, whichprovides a potent selection for yeast cells that are ura3 defective (Boeke et al., 1987). In the assay, the loss of Ura3p is a consequence of a frameshift in the reporter construct (Figure 4) dinucleotide repeat region that is fused in frame with URA3 (Henderson and Petes, 1992). Dinucleotide tracts expand or contract at an elevated rate as a consequence of replication errors (reviewed in Kunkel and Bebenek, 2000). If not repaired by the MMR machinery, the reading frame for the URA3 gene will be disrupted and the cell will no longer produce a functional Ura3p gene product. Mismatch repair defective cells display an elevated rate of resistance to 5-FOA (FOA^r) using this reporter, which is indicative of dinucleotide instability (Henderson and Petes, 1992). The second type of medium to score primarily for single base pair mismatch repair defects (typically seen as a consequence of MutSα loss of function) is supplemented with canavanine, a toxic analog of arginine (Rosenthal, 1977). Both canavanine and arginine are taken into the cells via the arginine permease (Grenson et al., 1966) encoded at the CAN1 locus (Broach et al., 1979). If the CAN1 gene is defective, then the yeast become resistant to canavanine, Can^r (Whelan et al., 1979). Mismatch repair defective strains show an elevated rate of canavanine resistance (Marsischky et al., 1996).

Students examine the results of growth on the drug plates and non-selective plates and compare the msh2-allele strain to the wild-type strain and the msh2 null strain. The wild type strain typically shows very little growth on the drug containing plates compared to the msh2 null strain, which characteristically shows pronounced growth. By comparison with the controls, the students make a qualitative judgment as to whether the msh2 allele results in defects in DNA mismatch repair (see Figure 5B for an example of student results).

Using the same strains and drug selection media, the students perform quantitative assays to determine the rate of mutation in the yeast strains described above. In order to calculate the rate, the students determine the number of Can^r or FOA^r mutants in the culture and the total number of viable cells. From this information the students determine the rate as an expression of the number of mutations (derived from the mutant count) per cell division (derived from the number of viable cells plated). The median number of colonies per plate from each drug-containing medium (either FOA or CAN) is used to calculate the mean number of mutations that must have occurred (Lea and Coulson, 1949). For each strain, the number of colonies on the nonselective plates is averaged and the appropriate dilution factor is applied to give the number of viable cells. For microbial growth, it is an accepted approximation that the number of cells in a population is equivalent to the number of cell divisions. Having gathered the required data and after making the appropriate conversions, the rate of mutation is expressed in terms of the number of mutations per cell division.

Each pair determines the rates for the MSH2 wild-type, the msh2 missense allele and msh2 null control strain. Class data are pooled to gather more accurate mutation rates for each strain and to illustrate the importance of multiple trials to obtain meaningful data. Four pairs typically analyze each allele so that a total of 12 trials for each msh2 allele strains and 75–90 trials for each of the control strains are performed yearly. Students are asked to compare the calculated mutation rates with those derived from the combined data.

By the end of the first half of the semester, the students have engineered a mutation and established whether the change effects DNA mismatch repair. In a typical year, at least one of the human-derived allele has a wild-type phenotype in yeast. The students are encouraged to carefully scrutinize the original publications that designated the wild type-like alleles as pathogenic and to include their findings in the first laboratory report.

Determination of the Steady-State Levels of Msh2 and Msh2 Variant Proteins

To begin to understand why certain missense mutations in MSH2 result in mismatch repair defects, the steady-state expression levels of the Msh2 and Msh2-variant proteins are determined. This assay determines whether the amino acid change results in dramatically altered stability of the Msh2 protein. The same three strains used in the mismatch repair assays are used in this analysis. The immunodetection of wild type Msh2p and the variant Msh2p are performed using antibodies that bind the hemagglutinin epitope (HA) encoded in the MSH2 gene (see Figure 2). The no epitope control sample is protein extracts from the msh2 null strain harboring the pRS413 vector. Student results from an immunoblot and Coomassie staining can be seen in Figure 6. Each year, many students observe a decrease in overall expression of the variant Msh2 proteins. Portions of the lectures and informal discussions are devoted to understanding the possible models for decreased expression. Students are encouraged to devise a strategy for distinguishing between the possibilities.

Localization of Msh2 and Msh2 Variant Protein Using Indirect Immunofluorescence

Students also test the hypothesis that the defect in mismatch repair phenotype conferred by some msh2 alleles might be attributable to failure of the variant Msh2 protein to localize to the nucleus. Indirect immunofluorescence is employed to localize the wild-type and mutated forms of the Msh2 protein. The same three yeast strains used in the previous analyses are used for the localization studies. The strains include the wildtype (expressing Msh2p containing the HA epitope), the msh2 missense mutant strain (expressing the variant Msh2p containing the HA epitope), and the msh2 null strain (that does not express Msh2 or the HA epitope).

Two important negative controls for the experiment include a sample of wild-type, antigen expressing cells with no primary antibody added and a sample with msh2 null cells that do not express the antigen (no epitope control), but are treated similarly to the experimental cells. These controls emphasize the importance of including a no epitope control if possible, since the background level of fluorescence observed in the no epitope control is typically much higher than the no primary antibody control. The positive control for the experiment consists of cells expressing the wild-type Msh2 protein. In addition, a double labeling of Msh2 and tubulin (see Materials and Methods) is performed as an internal control to prove that the cells were adequately prepared for immunofluorescence. All cells should show tubulin staining since tubulin is essential for viability. In addition, all cells are treated with the DNA specific dye, DAPI, to indicate nuclear position.

Students observe the overall cellular morphology with differential interference contrast (DIC) optics and visualize the Msh2 or tubulin localization and the cellular DNA using a microscope equipped with epi-fluorescence capabilities. After examining and photographing the control and experimental samples, the students are able to determine whether the variant Msh2p co-localizes with the nuclear DNA. Students are introduced to the practical and theoretical aspects of fluorescence microscopy while obtaining a tangible result. Results from a student immunofluorescence experiment may be seen in Figure 7. This laboratory session has a high rate of success in that the controls work well. One caveat is that the variant proteins that display decreased overall expression are sometimes difficult to detect. For these exceptions, the staff provides prepared slides where the allele has been over expressed, so that the localization may be observed. Many alleles are properly expressed and the variant protein is localized to the nucleus. The final experiment detailed below often proves helpful in understanding the defect for certain alleles.

Formation of a Recombinant DNA Molecule to Test MMR Subunit Interaction

The purpose of the final experiment is to determine whether the changed amino acids in the variant Msh2 proteins effect interactions with other members of the MMR complex. The students employ the yeast 2-hybrid assay to monitor the subunit interaction (see James et al., 1996 for a full explanation of the technique). Prior to the course, the basic components of the 2-hybrid MMR interaction were constructed (Table 3 and Figure 9A). Students design and create a recombinant DNA molecule that allows for expression of the variant Msh2 protein in the yeast 2-hybrid assay.

The SeqWeb program described above is employed to devise an efficient strategy for constructing the recombinant DNA molecule. A construct prepared before the class, pGBD-MSH2, serves as the “vector backbone” (Figure 9A) and a fragment of DNA containing the msh2 missense mutation represents the“ insert.” Students select two unique restriction enzymes that cut within the MSH2 coding region on the pGBD-MSH2 construct. Upon double digestion pGBD-MSH2 should be cleaved at sites within the MSH2 coding sequence that flank the mutation and generate non-ompatible, cohesive-ends. Using noncompatible cohesive ends improves the efficiency of obtaining the desired construct and also ensures that the reading frame is in tact upon ligation. Each enzyme cleavage site must be unique in pGBD-MSH2 to preserve the vector backbone, but does not need to be in the pMSH2-derivative construct carrying the mutation. The students generate a restriction map with unique cut sites in pGBD-MSH2. When selecting the enzymes, consideration is given to buffer compatibilities of the selected enzymes and how well the fragments can be resolved by agarose gel electrophoresis. Provided the recombinant DNA cloning strategy is reasonable, the students proceed with their own design.

The students digest the DNA, resolve the fragments with agarose gel electrophoresis and purify the desired bands from agarose slices. The students determine the yield of insert and vector DNA and ligation reactions and controls are prepared. Controls include the pGBD-MSH2 vector fragment alone with no insert fragment, and a singly cut vector with and without T4 DNA ligase to measure for the efficiency of the ligation. The ligated molecules are introduced into bacterial cells using electroporation. An uncut plasmid DNA control is included to measure the efficiency of the electroporation and to serve as a comparison for how efficiently the ligase worked to generate covalently closed plasmids. If the ligation is 100% efficient, then the cut plasmid with ligase should give the same number of transformants as the uncut control; whereas the cut control with no ligase should result in considerably fewer transformants. Standard molecular procedures screen for the recombinant molecules. For an example of student results see Figure 8.

The recombinant plasmid DNA as well as pGBD-C2 vector DNA and pGBD-MSH2 are used to transform a MATa yeast 2-hybrid strain (Table 1, Figure 9A). In addition, a no-plasmid DNA sterility control is also prepared. Prior to the course, the GAD fusion constructs (Figure 9A) were introduced into a MATα yeast 2-hybrid reporter strain. Thus, a panel of haploid yeast strains is available, each carrying a construct that will express a relevant hybrid protein. The MATa strains harbor the GBD fusions or GBD-alone control, while the MATα strains contain the GAD-partner fusions.

The students set up crosses to form diploid yeast strains. Ultimately, each diploid expresses both of the hybrid proteins necessary to examine the MMR subunit interactions (Figure 9A). Control diploid strains are formed to illustrate no interaction (cells expressing GBD alone in combination with each of the GAD-Partner fusions) and to illustrate a positive interaction (GBD-Msh2 in combination with each of the GAD-Partner fusions). In the yeast reporter strain, the HIS3 gene is fused downstream of the GAL1 promoter, thus the 2-hybrid interaction may be assayed using medium lacking histidine (–HIS). Two other reporter genes, ADE2 and lacZ are encoded in the 2-hybrid reporter strain, but the expression of these genes is not sufficient to accurately assess the interactions of the MMR components. Growth on –HIS indicates that the hybrid proteins are interacting. Thus, the determination of whether the mutated amino acids interfere with the ability of Msh2p to bind its functional partners can be assayed using a simple growth assay. Results from a typical 2-hybrid experiment can be seen in Figure 9B. Many of the missense mutations appear to prevent mismatch repair complex formation. Students are encouraged to examine the crystal structure of the closely related MutS protein (Lamers et al., 2000; Obmolova et al., 2000) to identify the amino acid of interest and to develop a hypothesis for why subunit formation might be effected.

Instruction in Writing in the Scientific Style

Students enrolled in the course are required to execute all of the experiments detailed above and to write two 25-page papers in the format of a scientific publication. This method of student evaluation was chosen since most assessments of scientific achievement are dependent upon written documents.

Although the students are writing up an ongoing project, each laboratory report has a different emphasis. The first report has an important implication for human health, thus the background material should include the relevant medical literature. The first report contains the following sections: Designing a mutagenic oligonucleotide; Creating a msh2 missense mutation by site-directed mutagenesis; Confirmation of the missense mutation by restriction enzyme analysis and dideoxy sequencing; and Mutation frequency assays of msh2 missense alleles. In the first report, the students have the satisfaction of discussing a concrete result with a medically significant conclusion. The second report investigates how the variant Msh2 proteins are defective and thus the background should emphasize the molecular mechanism of DNA mismatch repair. The second report specifically includes: Immunoblot analysis of altered Msh2 proteins; Cloning msh2 alleles into the yeast 2-hybrid system; Interactions with members of the mismatch repair machinery; and Msh2p localization studies. The second report has many results and allows for integration of several pieces of data to form a coherent model for a particular MSH2 allele.

As with most research articles, laboratory reports include the following sections: Title Page, Abstract, Introduction, Materials and Methods, Results (including data presentation), Discussion, and References. High standards are imposed in terms of scholarship, accuracy, scientific writing style, figure quality, citation professionalism, proposals for future work, and critical analysis. A page limit is imposed to encourage the students to write succinctly and to reflect reality since most scientific documents, including manuscripts and grant proposals, have page constraints.

Guidance for writing the reports is provided through several avenues. Seven pages of the manual are devoted to providing advice and suggestions for writing the laboratory reports. In addition to typical office hours, review sessions are held and a discussion board is posted on the course web site to answer specific technical questions asked electronically.

When evaluating the written work we strive to provide sufficient feedback on the document and to maintain grading equity. A rubric has been devised to aid teaching assistants in evaluating the reports (supplement supplied). The rubric has been useful for supplying students with concrete feedback as to their level of achievement in a variety of areas and for providing a more quantitative form of evaluation to aid in establishing equity across the grading groups. The rubric designed for this course allows for distinguishing among very accomplished Princeton molecular biology students. Educators might need to shift the emphasis depending on the typical student. Overall, the rubric has been well received by the students and the teaching assistants.

The quantitative assessment also allows for an analysis of specific areas of improvement in student writing. For the two years during which the rubric listed in the supplement was used, students showed an overall improvement of 5.4 ± 0.4 percent between the first and second report. The areas of improvement are in almost all assessment areas (see Figure 10); however, the students showed reproducible improvement in sections that are unique to scientific documents. The data is consistent with the observation that the students enter the class with basic writing and analytical skills and therefore do not show pronounced improvement in these areas. In addition, the data suggests that the feedback given after the first report aids the students specifically in the areas of writing in the scientific style.

Course Assessment

Princeton University's Office of the Registrar provides an assessment service for all courses. The assessment includes numerical scores in response to specific categories and written evaluations. All evaluations are anonymous and conducted under the supervision of a student volunteer rather than the course director. Based on the numerical and written evaluations, the course described in this paper has been well received by the students (see Tables 4 and 5). The breakdown of the numerical score for 2003 is shown in Table 4. Similar marks were given for the 5 years the course has been taught using the format described in this paper (not shown). The scores from a year prior to when the project was implemented (1997) and the University average are included for comparison purposes (Table 4). There was a significant improvement in the approval rating in all areas surveyed in the year that the course described in the paper was first introduced. Of particular interest is the score regarding the student's assessment of the overall quality of the laboratories the year that the project was implemented (Table 5 compare 1997 scores to those following).

Many of the written evaluations indicated the student interest in the project. Below are several examples of student comments regarding the project.

A survey of Molecular Biology seniors conducted in the spring of 2002, provides an indication of the broader utility of the course. In response to a senior thesis survey question “Which classes helped you the most in preparing you for your independent work?” 85% of the 26 respondents cited MOL350, the course described in this paper. In that survey, specific comments were solicited about the course described in this paper, since it is part of the Princeton molecular biology core curriculum. In response to the following question: “Now that it has been one or more years since you have taken the core requirements, what are your views on these courses?” Student testimony about MOL350 (also known as core lab) in response to the above question was as follows: 4 had no comment, 1 was neutral (“alright”), 1 was negative (“too formulaic”), and 25 were positive. 41% of the positive comments specifically mentioned the value of the course with respect to conducting research laboratory work and writing a molecular biology thesis.

In addition, the course confers a technical benefit beyond completing a degree at Princeton University. A recent example is from Jeffrey Kim, Class of 2003, who started working after graduation as a research technician at Rockefeller University.

DISCUSSION

Summary

This paper outlines an advanced laboratory course in cell and molecular biology. The students enrolled in the course gain experience with current laboratory techniques and scientific writing while conducting original research. The topic was selected to appeal to a wide range of students and the format of the project was designed to provide a practical and relevant training for future work in a molecular biology research laboratory setting.

Project Format: Advantages of Using Model Organisms

We use the facile model organism Saccharomyces cerevisiae, to study eukaryotic DNA mismatch repair. Because mismatch repair is a highly conserved process, we are able to ask medically relevant questions using a safe and noncontroversial model organism. Other practical considerations include the vast research resources available that have been generated by the yeast scientific community over the years. Furthermore, yeast research is inexpensive when compared to many model organisms or tissue culture systems. Yeast cells are easily manipulated, rapidly growing, efficiently stored for long periods of time, genetically tractable, readily transformed, and non-hazardous. Finally, most molecular biology laboratories employ either bacteria or yeast to manipulate cloned DNA, therefore becoming proficient in microbial work is a useful skill. Employing bacteria for the molecular manipulations and yeast for the mismatch repair functional assays, we are able to cover many topic areas and use a variety of techniques in the time allocated for the course.

Project Format: Techniques

The format of the project ensures exposure to commonly used research laboratory methods, including site-directed mutagenesis, the polymerase chain reaction, restriction enzyme analysis, agarose gel electrophoresis, DNA sequencing, immunoblotting, microbial manipulations, fluorescence microscopy, recombinant DNA technology, and in vivo reporter assays. The procedures for the project are provided in a loose-leaf manual. Protocols are derived from a variety of sources and the heterogeneity of styles has been preserved with the objective of introducing the students to an assortment of protocol formats. Original procedures have been modified slightly to eliminate the most confusing ambiguities.

Project Format: Time Frame

Each student works in the laboratory twice a week (Mon and Wed or Tue and Thu) for 3 h each session for a total of 12 wk (see the supplied Timeline). Thus, over the semester the students spend ∼72 h in the laboratory. In addition, they are often required to come in during“ off-hours” to perform minor tasks to keep the project progressing.

To ensure execution of all of procedures, it is necessary to have the intermediates as well as the final constructs and strains available when a particular aspect of the experiment fails. The “back-ups” allow the students to experience the entire project and ensure that a substantial amount of data is available to describe and analyze in their laboratory reports.

If an educator is considering implementing this course, but class time is limited, an option is to implement only the first half of the course. Biophysics majors currently exercise this option at Princeton. Many of the most basic molecular techniques are covered in the first half of the course. If resources are limited, the dideoxy nucleotide sequencing laboratory could be eliminated. In addition, many teaching laboratories are not equipped with fluorescent microscopes and thus this module could be eliminated. Prior to purchasing the microscope, gene dosage experiments were conducted rather than the immunofluorescence.

Supervision and Encouragement of Independence

The course employs a technician, a lecturer, a postdoctoral instructor, and four graduate student teaching assistants to prepare reagents and to teach and supervise 50–75 undergraduate students, most of whom are juniors in the Department of Molecular Biology. Each laboratory session has ∼30 students, or 15 pairs, divided into two laboratory rooms. Each room has one teaching assistant and either the lecturer or the postdoctoral instructor present. Thus, the ratio of instructional staff to students is ∼2:15.

The teaching staff provides instruction and support, but whenever possible encourages self-sufficiency. For example, students prepare reagents and maintain stock solutions. In addition, they routinely calculate the appropriate dilutions from a stock to the final solution. Students are required to plan as well as execute their experiments. For example, they design a mutagenesis, calculate diagnostic restriction digestion patterns, choose appropriate primers for PCR and sequencing reactions, and devise a recombinant DNA cloning strategy.

Student Motivation and Learning

An ongoing challenge to educators is to understand what motivates their students to learn and retain the material being presented. Many recent reports on how students learn have supported the approach outlined in this paper (for example see The National Research Council's, Bio2010, Transforming Undergraduate Education for Future Research Biologists, 2003).

The course has been designed to expose student to relevant and useful techniques and to let the students experience the excitement of original research in an interesting and applicable field. The students are able to apply what they have learned to their senior thesis independent research projects conducted in the various molecular biology research laboratories at Princeton. In addition, many of the Princeton molecular biology students go on to use the technical skills in their future careers, some 40% of which will earn a Ph.D.

The topic for this course was chosen to capture the interest of a wide array of enrolled students, including physics, engineering, chemistry, and biology majors. The Advisory Committee to the National Science Foundation Directorate for Education and Human Resources (1996) and the authors of a report for Project Kaleidescope (Rothman and Narum, 1999) support this effort to appeal to a variety of science students. The diverse students enrolled in this course express a high degree of enthusiasm for the topic. Because DNA mismatch repair is a fundamental biological process with an understandable relevance to human health, the subject matter engages those interested in basic science, medicine, or both.

The students are also excited about participating in a novel research project and express in written evaluations that they find it preferable to performing a set of experiments where the answer is known. The inquiry-based approach to learning described in this paper is supported by the National Research Council's recent recommendations for undergraduate education for future research biologists (National Research Council Committee on Undergraduate Biology Education to Prepare Research Scientists for the 21st Century, 2003). Other influential educational reports recommend making research-based learning for undergraduates the standard (The Boyer Commission on Educating Undergraduates in the Research University, 1998). In summary, in an effort to motivate the students to learn and retain the material, a great deal of energy is directed toward making this course exciting, real, and relevant for future scientists.

FOOTNOTES

ACKNOWLEDGMENTS