A cell-free assay for the functional analysis of variants of the mismatch repair protein MLH1.
ABSTRACT The hereditary colon and endometrium cancer predisposition Lynch Syndrome (also called HNPCC) is caused by a germ-line mutation in one of the DNA mismatch repair (MMR) genes. A significant fraction of the gene alterations detected in suspected Lynch Syndrome patients is comprised of amino acid substitutions. The relevance for cancer risk of these variants is difficult to assess, as currently no time- and cost-effective, validated, and widely applicable functional assays for the measurement of MMR activity are available. Here we describe a rapid, cell-free, and easily quantifiable MMR activity assay for the diagnostic assessment of variants of the MLH1 MMR protein. This assay allows the parallel generation and functional analysis of a series of variants of the MLH1 protein in vitro using readily available, or preprepared, reagents. Using this assay we have tested 26 MLH1 variants and of these, 15 had lost activity. These results are in concordance with those obtained from first-generation assays and with in silico and pathology data. After its multifocal technical and clinical validation this assay could have great impact for the diagnosis and counseling of carriers of an MLH1 variant and their relatives.
- SourceAvailable from: umich.edu[show abstract] [hide abstract]
ABSTRACT: Lynch syndrome, also known as hereditary nonpolyposis colon cancer (HNPCC), is the most common known genetic syndrome for colorectal cancer (CRC). MLH1/MSH2 mutations underlie approximately 90% of Lynch syndrome families. A total of 24% of these mutations are missense. Interpreting missense variation is extremely challenging. We have therefore developed multivariate analysis of protein polymorphisms-mismatch repair (MAPP-MMR), a bioinformatic algorithm that effectively classifies MLH1/MSH2 deleterious and neutral missense variants. We compiled a large database (n>300) of MLH1/MSH2 missense variants with associated clinical and molecular characteristics. We divided this database into nonoverlapping training and validation sets and tested MAPP-MMR. MAPP-MMR significantly outperformed other missense variant classification algorithms (sensitivity, 94%; specificity, 96%; positive predictive value [PPV] 98%; negative predictive value [NPV], 89%), such as SIFT and PolyPhen. MAPP-MMR is an effective bioinformatic tool for missense variant interpretation that accurately distinguishes MLH1/MSH2 deleterious variants from neutral variants.Human Mutation 07/2008; 29(6):852-60. · 5.21 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: Germline mutations in the DNA mismatch repair (MMR) gene MLH1 are associated with a large percentage of hereditary non-polyposis colorectal cancers. There are approximately 250 known human mutations in MLH1. Of these, one-third are missense variants that are often difficult to characterize with regards to pathogenicity. We analysed 28 alleles of baker's yeast MLH1 that correspond to non-truncating human mutant alleles listed in online HNPCC databases, 13 of which had not been previously studied in functional assays. Using the highly sensitive lys2::InsE-A(14) reversion rate assay, we determined the MMR proficiency conferred by each allele in the S288c strain of Saccharomyces cerevisiae. Seven alleles conferred a null phenotype for MMR and eight others showed significant MMR defects, suggesting that all 15 are likely to be pathogenic in humans. In addition, we observed a strong correlation between these results, limited results from previous functional assays and clinical data. To test whether the potential pathogenicity of certain alleles depends on the genetic background of the host, we examined the mutation rates conferred by the mlh1 alleles in a second yeast strain, SK1, which is approximately 0.7% divergent from S288c. Many alleles displayed a difference in MMR efficiency between strain backgrounds with decreasing differences as the severity of the MMR defect increased. These findings suggest that genetic background can play an important role in determining the pathogenicity of MMR alleles and may explain cases of atypical colorectal cancer inheritance.Human Molecular Genetics 03/2007; 16(4):445-52. · 7.69 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: Nuclear extracts derived from HeLa and Drosophila melanogaster KC cell lines have been found to correct single base-base mispairs within open circular DNA heteroduplexes containing a strand-specific, site-specific incision located 808 base pairs from the mismatch. Correction in both extract systems is strand specific, being highly biased to the incised DNA strand. Different mispairs within a homologous set of heteroduplexes were processed with different efficiencies (G.T greater than G.G approximately equal to A.C greater than C.C), and correction was accompanied by mismatch-dependent DNA synthesis localized to the region spanning the mispair and the strand break, thus demonstrating that mismatch recognition is associated with the repair reaction. Correction of each of these heteroduplexes was abolished by aphidicolin but was relatively insensitive to the presence of high concentrations of ddTTP, indicating probable involvement of alpha and/or delta class DNA polymerase(s). These findings suggest that higher eukaryotic cells possess a general, strand-specific mismatch repair system analogous to the Escherichia coli mutHLS and the Streptococcus pneumoniae hexAB pathways, systems that contribute in a major way to the genetic stability of these bacterial species.Proceedings of the National Academy of Sciences 09/1990; 87(15):5837-41. · 9.74 Impact Factor
A Cell-Free Assay for the Functional Analysis of Variants
of the Mismatch Repair Protein MLH1
Mark Drost,1Jose ´ B.M. Zonneveld,1Linda van Dijk,1Hans Morreau,2Carli M. Tops,3Hans F.A. Vasen,4,5Juul T. Wijnen,3
and Niels de Wind1?
1Departments of Toxicogenetics, Leiden University Medical Center, Leiden, The Netherlands;2Departments of Pathology, Leiden University
Medical Center, Leiden, The Netherlands;3Departments of Human & Clinical Genetics, Leiden University Medical Center, Leiden, The
Netherlands;4Departments of Gastroenterology, Leiden University Medical Center, Leiden, The Netherlands;5The Netherlands Foundation for
the Detection of Hereditary Tumours, Leiden, The Netherlands
Communicated by Rolf Sijmons
Received 16 September 2009; accepted revised manuscript 23 November 2009.
Published online 17 December 2009 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/humu.21180
ABSTRACT: The hereditary colon and endometrium
cancer predisposition Lynch Syndrome (also called
HNPCC) is caused by a germ-line mutation in one of
the DNA mismatch repair (MMR) genes. A significant
fraction of the gene alterations detected in suspected
Lynch Syndrome patients is comprised of amino acid
substitutions. The relevance for cancer risk of these
variants is difficult to assess, as currently no time- and
cost-effective, validated, and widely applicable functional
assays for the measurement of MMR activity are
available. Here we describe a rapid, cell-free, and easily
quantifiable MMR activity assay for the diagnostic
assessment of variants of the MLH1 MMR protein. This
assay allows the parallel generation and functional
analysis of a series of variants of the MLH1 protein in
vitro using readily available, or preprepared, reagents.
Using this assay we have tested 26 MLH1 variants and of
these, 15 had lost activity. These results are in
concordance with those obtained from first-generation
assays and with in silico and pathology data. After its
multifocal technical and clinical validation this assay
could have great impact for the diagnosis and counseling
of carriers of an MLH1 variant and their relatives.
Hum Mutat 31:247–253, 2010. & 2009 Wiley-Liss, Inc.
KEY WORDS: Lynch syndrome; DNA mismatch repair;
MMR; variants of uncertain significance; functional
Lynch Syndrome (or Hereditary Nonpolyposis Colorectal
Cancer, HNPCC; MIM]s 120435, 609310) is an autosomal
dominant disorder that predominantly predisposes to colon and
endometrial cancer [Stoffel et al., 2009; Vasen et al., 2007]. Lynch
Syndrome is caused by a germ-line mutation in one of the DNA
mismatch repair (MMR) genes. In a Lynch syndrome patient
stochastic loss of the wild-type allele leads to MMR deficiency,
which results in a spontaneous mutator phenotype that drives
rapid carcinogenesis [Li, 2008].
A definitive diagnosis of Lynch syndrome can only be made
by finding a causative germ-line mutation in an MMR gene
[Stoffel et al., 2009; Vasen et al., 2007]. Additionally, the
identification of a causative mutation allows the presymptomatic
screening of affected relatives [Vasen et al., 2007]. Of all mutations
found in MMR genes in suspected Lynch Syndrome patients,
?15% in MSH2 (MIM] 609309), ?30% in MLH1 (MIM]
120436), and ?40% in MSH6 (MIM] 600678) give rise to single
amino acid alterations [Lagerstedt Robinson et al., 2007;
Peltoma ¨ki and Vasen, 2004] of which the pathological significance
is often unclear (so-called Variants of Uncertain Significance, or
VUS). Lack of classification of these variants precludes diagnosis
for carriers and their relatives. Therefore, when a suspected
Lynch syndrome patient carries a VUS, all first-degree relatives
currently enroll lifelong periodic screening, irrespective of
their mutation status [Castells et al., 2009]. Screening poses a
physical and psychological burden on these families, but also a
burden on the preventive health care apparatus. To enable correct
diagnosis and to avoid unnecessary screening it is of great
relevanceto develop a widely
standardized approach for the diagnostic assessment of VUS in
The MMR pathway corrects misincorporations arising during
DNA replication, and thereby prevents the accumulation of
spontaneous mutations [Hsieh and Yamane, 2008; Li, 2008].
MMR is initiated by the recognition of a misincorporation by the
heterodimeric MutSa protein, consisting of the MSH2/MSH6
subunits. This triggers the recruitment of the MLH1/PMS2
heterodimeric protein (called MutLa) and of Exonuclease 1 that
removes the misincorporation-containing DNA strand. Removal
is directed toward the misincorporation-containing daughter
strand by the presence of a strand discontinuity (‘‘nick’’), 50or
30of the mismatch. The excised DNA strand is subsequently
Functional MMR activity of a VUS is considered an important
diagnostic criterion [Couch et al., 2008; Goldgar et al., 2008;
Plon et al., 2008]. In vitro MMR activity assays have been
developed [Li and Modrich, 1995], and in such a complementa-
tion assay the inability of a variant to restore MMR activity to a
MMR-deficient cell extract almost certainly assigns pathogenicity
to the VUS [Couch et al., 2008; Ou et al., 2007]. However, the
requirement for molecular cloning, cell culture, and the use
of bacteria precludes their independent validation and wide
& 2009 WILEY-LISS, INC.
Additional Supporting Information may be found in the online version of this article.
?Correspondence to: Niels de Wind, Leiden University Medical Center,
Toxicogenetics, P.O. Box 9600, Leiden, 2300 RC, The Netherlands.
applicability, complicating diagnosis and screening of suspected
patients and their relatives.
Here, we present the analysis of in vitro MMR activity of 26
MLH1 VUS using a completely cell-free MMR complementation
assay. The results from this assay are consistent with those from
other functional assays and with in vitro predictions of protein
function. We infer that this assay is amenable to independent
validation and applicable in molecular diagnostic laboratories,
and therefore could greatly contribute to the diagnostic classifica-
tion of MLH1 VUS.
Materials and Methods
The plasmid pUC19CPD has been used for the generation of
MMR substrates [Wang and Hays, 2006], and was modified by
introducing additional recognition sites for nicking endonu-
cleases. The resulting plasmid was named pJH and, after the
introduction of a mismatch (see below), allows to measure MMR
activity at each DNA strand, directed by the presence of a nick
either at the 50or the 30side of the mismatch (Fig. 1A and Supp.
Fig. S1). The substrate for MMR was prepared from pJH,
essentially as described [Wang and Hays, 2006], with modifica-
tions. Briefly, a 26 basepair (bp) single-stranded DNA stretch
between two Nt. BstNBI sites was removed by heating and
annealing to an excess of complementary oligonucleotide. The
resulting gapped pJH was purified on Sephacryl S-400 HR colums
(GE Healthcare, Piscataway, NJ, USA). Phosphorylated oligonu-
cleotide G/T-Fam, containing a mismatching nucleotide and an
internal 6-FAM fluorescent label, was annealed into the gapped
plasmid and ligated. After purification the resulting mismatched
substrate was nicked at the bottom strand, 138bp 30to the
mismatch with Nb.BsmI. Finally, the substrate, named pJHGT30
lnFAM (Fig. 1B, top panel), was purified, aliquoted, and frozen
until use. All enzymes were purchased from New England Biolabs
(Boston, MA, USA).
Preparation of Nuclear Extracts
HCT116 MutLa-deficient colon cancer cells [Jiricny, 2003] were
grown at 371C in RPMI-1640 (Gibco, Paisley, UK) supplemented
with 10% FCS (PAA, Pasching, Austria), penicillin, streptomycin,
and Pyruvate (Gibco). Cells were harvested in log-phase and
nuclear extracts were prepared as described [Holmes et al., 1990].
These extracts can be prepared in large quantities, distributed, and
stored for prolonged periods.
Generation of MLH1 Mutants by PCR and In Vitro
Expression of MMR VUS Proteins
The human MLH1 (GenBank NM_000249.3) and PMS2
(GenBank U14658.1) cDNAs were cloned into the pCITE4a
vector (Novagen, Beeston, UK). The MLH1 plasmid was used to
generate MLH1 mutants, containing vector-derived T7 promoter-
and CITE sequences, required for efficient expression in vitro in a
two-step PCR procedure (Fig. 2A–C). In the first round of PCR,
overlapping mutant 50and 30MLH1 fragments were generated
separately. In a second round of amplification, the 50and 30
overlapping mutant MLH1 fragments were pooled and amplified
to generate the full-length mutant MLH1 gene. For all PCR
reactions Pfx Platinum polymerase (Invitrogen, Carlsbad, CA) was
used. All oligonucleotide primer sequences can be found in Supp.
Table S1. Nucleotide numbering reflects cDNA numbering with 1
1 corresponding to the A of the translation initiation codon in the
GenBank reference sequence. Amino acid numbering is based on
the cDNA with 11 corresponding to the translation initiation
MLH1-mutant PCR fragments were directly used for in vitro
protein expression in the TnT Quick Coupled Transcription/
Translation Kit (Promega, Madison, WI, USA) in the presence of
PCR Enhancers, according to the manufacturer’s instructions. To
quantify expression levels, proteins were labeled with
Methionine in parallel reactions. Alternatively, protein expression
was investigated by Western blotting using rabbit a-MLH1 (Santa
Cruz Biotechnology, Santa Cruz, CA, USA, 1:1,000) or mouse a-
PMS2 (BD Pharmagen, 1:1,000). Proteins were visualized by
enhanced chemiluminescence detection. Signals were quantified
using OptiQuant software.
Detailed protocols for all procedures are available upon request.
In Vitro MMR Complementation Assays
After estimation of quantities of expressed VUS MLH1 proteins
by Western blotting, the proteins were diluted to similar
concentrations in expression mix. This was followed by dimeriza-
tion with in vitro produced wild-type PMS2 protein in expression
mix (1:1vol/vol), for 30min at room temperature.
Assay reactions were performed in a total volume of 25mL
containing 75mg of HCT116 nuclear extract. The extract was
complemented with 12ml of in vitro produced MutLa and 100ng
of substrate pJHGT30lnFAM. Reaction conditions for MMR were
as described [Wang and Hays, 2006]. After the MMR reaction and
purification using the MinElute kit (Qiagen, Germantown, MD,
USA), the substrate was digested with HinDIII and BsrBI (both
Fermentas, Burlington, Ontario, Canada). One-fifth (2ml) of
digested substrate was mixed with 8ml Hi-Di Formamide
containing 0.2ml GeneScan-500 ROX size standards (Applied
Biosystems, Bedford, MA, USA) and fragment analysis was
performed on a 3730 DNA analyzer (Applied Biosystems) at the
Leiden Genome Technology Center. Data was analyzed using
PeakScan Software (Applied Biosystems). Repair levels are
calculated by dividing the height of the MMR-specific peak by
the total fluorescent signal.
Generation of Cloned Mutants
Most mutations were additionally introduced into MLH1,
cloned in pCITE4a, by site-directed mutagenesis (Quick Change
Site-Directed Mutagenesis, Stratagene, LaJolla, CA, USA). Primer
sets (Biolegio, Nijmegen, The Netherlands) were designed (Supp.
Table S2) and PCR was performed using Pfx Platinum polymerase
(Invitrogen). PCR products were ligated and used to transform
Escherichia coli. Resulting plasmids were verified by direct
sequencing of the entire mutant MLH1 insert.
In Silico Analysis
All MLH1 VUS tested in this work were subjected to in silico
analyses [Tavtigian et al., 2008] using the the Sorting Intolerant
From Tolerant (SIFT) [Ng and Henikoff, 2003] and the MAPP-
MMR [Chao et al., 2008] algorithms.
HUMAN MUTATION, Vol. 31, No. 3, 247–253, 2010
Construction of Fluorescent Heteroduplex Substrate for
For the current study we constructed a T/G mismatch-contain-
ing substrate, called pJHGT30lnFAM (Fig. 1B, top panel). MMR is
directed toward the G-containing (bottom) strand by a nick at that
strand, 30from the mismatch. Repair of the T/G mismatch to T/A
by MMR creates a recognition site for HinDIII, within a 174-bp
BsrBI restriction fragment. This results in the generation of a 75-bp
fluorescent diagnostic fragment (Fig. 1B and C, bottom panels).
A 174-bp BsrBI fragment represents the unrepaired substrate.
A Functional Assay Using In Vitro-Produced MutLa
Wild-type MLH1 and PMS2 were cloned into pCITE4A. The
MLH1 clone was used as a template to amplify by PCR the open
reading frame of the gene, together with sequences required for in
cleavage with Nt.BstNBI and heating. To generate the substrate pJHGT3’lnFAM, the latter fragment is replaced with a mismatch- and fluorescent
label (star)-containing fragment. Then, the bottom strand is nicked at Nb.BsmI to provide strand specificity to the MMR reaction. See Supp. Figure S1
for more details. B: Flow-scheme of the MMR assay. Top panel: MMR substrate, derived from pJH. To the substrate in vitro expressed MutLa-variant,
protein and MutLa-deficient HCT116 cell extract are added in the presence of cofactors. After incubation, substrate is purified and cleaved with
BsrBI and HinDIII (bottom panel). The latter enzyme only cleaves when the mismatch is repaired, resulting in a diagnostic fluorescent fragment of
75bp. Numbers: sizes of restriction sites. C: Examples of automated fragment analysis after a MMR assay. Left panel: fragment analysis after testing
an MMR-proficient variant. Right panel: example of a fragment analysis chromatogram after testing of a mock MutLa expression. D: Left panel:
Western blot showing PMS2 and MLH1 proteins produced in vitro from PCR-amplified genes, as well as a mock expression. Right panel:35S-
Methionine-labeled PMS2 and MLH1, expressed in vitro from cloned PMS2, and from PCR-amplified MLH1 genes. E: Quantification of relative
MLH1:MSH6 ratios in VH10 nuclear extracts and in a MMR assay, complemented with in vitro synthesized MutLa.
Outline of the cell free MMR complementation assay. A: Plasmid pJH. Bold line: single-stranded DNA fragment that can be removed by
HUMAN MUTATION, Vol. 31, No. 3, 247–253, 2010
vitro transcription/translation. MLH1 and PMS2 were expressed from
the PCR fragment and from the expression vector, respectively, as
judged by W blotting (Fig. 1D, left panel). Importantly, the
reticulocyte lysate was devoid of detectable rabbit MLH1 or PMS2
that might give false-positive results. Both proteins were synthesized
in approximately equimolar amounts as judged by35S-Methionine
levels in both proteins, using a parallel expression reaction (Fig. 1D,
right panel). Next, MLH1 and PMS2 were used to complement an
MLH1-deficient HCT116 nuclear extract to test for the repair of
substrate pJHGT3’lnFAM (Fig. 1B). Using this assay, nearly half of the
substrate was repaired (Fig. 1C, left panel). This result is similar to the
repair efficiency in other assays that use nonfluorescent substrates and
in vivo produced MutLa [Plotz et al., 2006; Raevaara et al., 2005].
Importantly, mock-expression reactions and reactions containing only
MLH1 or PMS2 did not result in significant repair (Fig. 1C, right
panel, see also Fig. 3). Moreover, the MMR protein stochiometry in
our MMR assay mimics the in vivo stochiometry, as evidenced by
comparing MLH1:MSH6 ratios between an MMR complementation
assay and wild-type cells (Fig. 1E). Taken together, these results
indicate that our cell-free assay is suited for testing MLH1 activity.
Production of MLH1 Mutants in a Two-Step PCR Procedure
To create MLH1-mutant genes suited for transcription/translation in
vitro without the need of prior cloning we developed a two-step PCR-
based method to produce 26 MLH1 mutants together with regulatory
sequences (Table 1, Fig. 2A–D, Supp. Table S1, and Methods and
Materials). This procedure was successful for all mutants, although final
DNA yields varied somewhat (Fig. 2E). Sequencing of the MLH1-
mutant fragments showed no visible contamination with wild-type
input MLH1 (Fig. 2F and Supp. Fig. S2). To investigate the PCR fidelity
more extensively, a PCR product of one of the mutants was cloned and
10 independent clones were fully sequenced. This revealed no
mutations, additional to the desired one (not shown). We conclude
that the PCR-based approach to generate mutant MLH1 genes for
expression in vitro is efficient and free of detectable artifacts.
MMR Activities of In Vitro Produced MLH1 VUS
The panel of 26 MLH1 mutants included three putative
polymorphisms and four presumed pathogenic mutants (for
‘‘half’’ fragments, respectively. C: The 50and 30mutant MLH1 fragments are joined in a second PCR step, resulting in mutant MLH1, preceded by
regulatory sequences. D: Functional domains of MLH1 [Raevaara et al., 2005; Wanat et al., 2006], with all 26 constructed and tested VUS. Amino
acid numbering is based on the cDNA with 11 corresponding to the translation initiation codon. E: Representative agarose gel picture of the PCR-
based generation of 26 MLH1-mutant PCR fragments. Top panel: 50mutant ‘‘half’’ fragments, middle panel: 30mutant ‘‘half’’ fragments, bottom
panel: full-length mutant MLH1 fragments. Left: size markers (kbp). F: Example of sequencing results from three randomly chosen PCR-generated
full-length MLH1-mutant fragments. The desired nucleotide alterations are encircled. See Supp. Figure S2 for a color version. G: Representative
Western blot of all 26 in vitro produced MLH1-variant proteins, expressed directly from mutagenic PCR-generated templates.
In vitro generation of MLH1-variant genes and proteins. A, B: PCR design and primer positions for generating 50and 30MLH1-mutant
HUMAN MUTATION, Vol. 31, No. 3, 247–253, 2010
references, see Table 1). All mutants are described in database of the
International Society for Gastrointestinal Hereditary Tumours
[Peltoma ¨ki and Vasen, 2004] (http://chromium.liacs.nl/LOVD2/
colon_cancer) and most of them also in the MMR Gene Unclassified
Variants Database [Ou et al., 2008] (http://www.mmrmissense.net).
An overview of pathology data is shown in Table 1. Mutant MLH1
proteins (black bars) and of plasmid clone-based MLH1 variant proteins (gray bars). Pathogenic, putative pathogenic Lynch Syndrome VUS;
polymorphisms, putative MLH1 polymorphisms. Results are mean7SEM of at least three completely independent (from PCR to MMR assay)
experiments.?Significantly higher repair activity compared to the four putatively pathogenic variants (two-tailed t-test, po0.05). The repair activity of
wild-type MLH1 was set at 100%. Representative examples of fragment analysis diagrams for each variant can be found in Supp. Figure S3.
MMR activity as measured in the in vitro MMR complementation assay. MMR activity assays of PCR fragment-based MLH1-variant
In Silico Data, Pathology Data, and References for All MLH1 VUS Tested in This Work
In vitro MMR
In silico analysisReferences to functional studiese
MMR assaysOther functional assays
1 2, 3, 4
2, 3, 6, 71, 5
0.061, 5, 89
0 1, 10, 11122, 3
0.361, 5, 11, 12 2, 3, 6
1, 10129, 13
1, 54, 6, 9f, 133, 9f
1, 5, 82, 3, 4
aNucleotide numbering reflects cDNA numbering with 11 corresponding to the A of the translation initiation codon in the GenBank reference sequence, amino acid
numbering is based on the cDNA with 11 corresponding to the translation initiation codon.
bMLH1: GenBank NG_007109.1. MSI5microsatellite instability; H5high; L5low; NA5not available; IHC5immunohistochemistry.
cSIFT scores as calculated on the SIFT webpage (http://sift.jcvi.org/). Scores o0.06 are considered pathogenic/not tolerated.
dMAPP-MMR [Chao et al., 2008] scores were calculated from http://mappmmr.blueankh.com/. Values 44.55 are considered pathogenic.
eReferences: 1 [Takahashi et al., 2007], 2 [Shimodaira et al., 1998], 3 [Kondo et al., 2003], 4 [Guerrette et al., 1999], 5 [Raevaara et al., 2005], 6 [Blasi et al., 2006], 7 [Avdievich
et al., 2008], 8 [Nystro ¨m-Lahti et al., 2002], 9 [Wanat et al., 2007], 10 [Plotz et al., 2006], 11 [Spina et al., 2008], 12 [Trojan et al., 2002], 13 [Perera and Bapat, 2008]. For
references employing more than in vitro MMR assays alone, only the result for the MMR assay has been included in this table.
fThe K618A mutant does not fit either category in reference 9 [Wanat et al., 2007].
HUMAN MUTATION, Vol. 31, No. 3, 247–253, 2010
PCR fragments (Fig. 2E) were used to produce proteins invitro (Fig.
2G). No relation between quantities of the input PCR fragments and
the amount of expressed protein was observed, and incidental
differences in expression levels were not reproducible between
experiments (not shown). Next, approximately equal amounts of
each protein were tested for MMR activity. The putative
polymorphisms all repaired more than 84% of wild type, whereas
the presumed pathogenic mutants all repaired less than 20% (Fig. 3,
black bars and Supp. Fig. S3). This difference allows to selectively
identify pathogenic MLH1 mutants. In the other 19 MLH1 mutants
we found a large variation in MMR activity (Fig. 3). As a tentative
cutoff for MMR deficiency we assumed a repair level that was not
significantly higher (Student’s two-tailed t-test, po0.05) than the
average repair efficiency of the presumed pathogenic mutants. In
support, all repair deficient VUS were predicted to be deleterious
using SIFT analysis and most also by MAPP-MMR, an algorithm
focused at MMR gene variants (Table 1).
Repair Efficiencies of Cloned Mutants Mirror Repair
Efficiencies of PCR-Based Mutants
To further investigate the robustness of the PCR product-based
expression system, we investigated the activity of MLH1 VUS,
expressed after their cloning. Although the activities of cloned
VUS reflected those of PCR-generated VUS (Fig. 3), repair levels
for the MMR-proficient cloned VUS were somewhat higher than
those of the PCR-based VUS, due to the presence of MMR-
inhibitory contaminants in the PCR-based expressions (not
shown). Nevertheless, the generation of MLH1 VUS in a two-
step PCR followed by their direct expression provides a rapid and
reliable alternative for cloning.
MMR VUS Classification Does Not Critically Depend on
As the expression levels between individual variant proteins was
subject to some random variation (see above) we wanted to
investigate whether the MMR assay is tolerant to variations in
MLH1 levels. To this end, wild-type MLH1, MMR-proficient VUS
E578G and MMR-deficient VUS G67R were expressed from PCR
products (Fig. 2G). Then, the reactions were diluted up to twofold
such that each contained a similar concentration of MLH1.
During further serial dilution, both the wild-type protein and
E578G retained activity that discriminates them from G67R (Fig.
4). Remarkably, at increasing dilutions the activity of E578G was
more affected than of wild-type MLH1, indicating that the
mutation may confer partial loss of function. This result shows
that the assay is tolerant to stochastic variations in in vitro
expression levels of the MLH1 variants.
Proper genetic counseling of suspected Lynch Syndrome
patients requires an accurate diagnosis based on the nature of
the germ line mutation found. Cancer risk resulting from
nontruncating, and especially missense, variants is difficult to
predict. This results in anxiety, and frequently, unnecessary
periodic screening in carriers of such a variant, but also in both
VUS-carrying and noncarrying relatives. In this work we present a
cell-free assay for the functional analysis of MLH1 VUS that
enables the rapid identification of pathogenic variants.
Repair deficiency for a VUS in our assay was, with few exceptions,
consistent with the in silico prediction of pathogenicity and with
pathology data (Table 1). Some of the same VUS were analyzed
before by others and most MLH1 VUS that were inactive in our
assay were also inactive in other assays. One notable exception is
VUS R265C that, although classified repair-deficient in three studies
[Plotz et al., 2006; Takahashi et al., 2007; Wanat et al., 2007], was
classified proficient in another [Trojan et al., 2002]. Additionally,
L550P, A589D, and P654L were active in another assay [Raevaara
et al., 2005] but inactive in yeast-based assays [Takahashi et al., 2007;
Wanat et al., 2007]. The latter VUS are located at the protein–pro-
tein interaction domain (Fig. 2G). In vitro binding studies
(Andersen et al., personal communication) support a defect for
these variants and another study has shown binding defects for other
VUS in this region of the protein [Guerrette et al., 1999].
Furthermore, the observed lack of nuclear import of PMS2 by
these VUS [Raevaara et al., 2005] supports a dimerization defect,
since heterodimerization of MLH1 and PMS2 stimulates nuclear
import of MutLa [Knudsen et al., 2009]. Additionally, all four VUS
that show discrepant results (R265C, L550P, A589D, and P654L) are
predicted to be defective by in silico analyses and were associated
with microsatellite instability, a hallmark of MMR deficiency
(Table 1). For these reasons, we conclude that a lack of activity in
our assay is in agreement with most other functional data, with
pathology data and with in silico analysis.
Because a loss of MMR is strongly predictive for Lynch Syndrome,
the positive predictive value of our assay is bound to be high.
Further validation of our assay seems warranted, using a much
larger group of accepted polymorphisms and pathogenic mutations
from well-studied patients. This could aid in determining the
relation between MMR activity and disease penetrance, enabling to
assess pathogenicity of VUS that display a low to intermediate level
of repair. However, as our assay does not address specific in vivo
MMR defects, such as a splicing, stability, or intracellular
localization, nor subtle in vitro defects, the negative predictive value
will probably be lower. As an example, E578G displayed reduction in
MMR activity only after its dilution (Fig. 4) that may reflect a subtle
dimerization defect. We infer that, in addition to a cell-free
complementation assay as described here, additional functional data
might be used for final diagnosis of VUS that display residual
activity. This may include nuclear localization assays [Raevaara et al.,
2005], splicing assays [Tournier et al., 2008], protein–protein
interaction studies [Guerrette et al., 1999], or stability studies
[Perera and Bapat, 2008]. A combination of different analytical
approaches, pathological, and family data should result in a
tion. To achieve similar protein concentrations, wild-type MLH1 and
variant G67R (Fig. 2G) were initially diluted twofold. After further dilution,
MMR activity was tested for all three proteins. Error bars: SEM.
Limited dependence of MMR activity on MLH1 concentra-
HUMAN MUTATION, Vol. 31, No. 3, 247–253, 2010
comprehensive, validated, diagnostic procedure of VUS in MLH1
and other MMR genes [Plon et al., 2008].
In conclusion, with this assay, we have developed a simple, rapid,
cost-effective, and reliable method that may fill a niche in diagnostic
labs [Heinen, 2009]. Cell extracts, substrate, and other ingredients
can be prepared, stored, and distributed in large quantities. The full
assay (recreation of the mutants by PCR, in vitro expression and in
vitro MMR assay) can be performed in 3 to 4 days, and many
mutants can be tested in parallel. The procedure can be used to
screen the large amount of mutants already detected, a number that,
due to the advent of personalized genomics, is likely to rise sharply.
Alternatively, the assay may be used for high-resolution a priori VUS
analyses as was done for p53 [Kato et al., 2003]. Currently we are
developing and testing very similar cell-free assays to test functional
activity of VUS in the MSH2 and MSH6 MMR genes.
This work was supported by grants from the Nuts/Ohra foundation (SNO-
T-03-06), the Dutch Digestive Foundation (MWO 05-16), and the
European Union (FP6-018754). We acknowledge Prof. John B. Hays and
Prof. Josef Jiricny for providing plasmids and Dr. Giancarlo Marra for
help. Drs. Anne Lu ¨tzen, Sofie Dabros Andersen and Prof. Lene Juel
Rasmussen are thanked for MLH1-mutant plasmids, sharing unpublished
data, and helpful discussions. We thank Drs. Anastasia Tsaalbi-Shtylik,
Cristina Ferra ´s, Jaap Jansen, and Prof. Robert Hofstra for critical reading.
Avdievich E, Reiss C, Scherer SJ, Zhang Y, Maier SM, Jin B, Hou H, Rosenwald A,
Riedmiller H, Kucherlapati R, Cohen PE, Edelmann W, Kneitz B. 2008. Distinct
effects of the recurrent MIh1G67R mutation on MMR functions, cancer, and
meiosis. Proc Natl Acad Sci USA 105:4247–4252.
Blasi MF, Ventura I, Aquilina G, Degan P, Bertario L, Bassi C, Radice P, Bignami M.
2006. A human cell-based assay to evaluate the effects of alterations in the
MLH1 mismatch repair gene. Cancer Research 65:9034–9044.
Castells A, Castellvı ´-Bel S, Balaguer F. 2009. Concepts in familial colorectal cancer:
where do we stand and what is the future? Gastroenterology 137:404–409.
Chao EC, Velasquez JL, Witherspoon MS, Rozek LS, Peel D, Ng P, Gruber SB,
Watson P, Rennert G, Anton-Culver H, Lynch H, Lipkin SM. 2008. Accurate
classification of MLH1/MSH2 missense variants with multivariate analysis of
protein polymorphisms-mismatch repair (MAPP-MMR). Hum Mutat 29:852–860.
Couch FJ, Rasmussen LJ, Hofstra R, Monteiro AN, Greenblatt MS, de Wind N,
Group IARCUGVW. 2008. Assessment of functional effects of unclassified
genetic variants. Hum Mutat 29:1314–1326.
Goldgar DE, Easton DF, Byrnes GB, Spurdle AB, Iversen ES, Greenblatt MS,
Group IARCUGVW. 2008. Genetic evidence and integration of various data
sources for classifying uncertain variants into a single model. Hum Mutat
Guerrette S, Acharya S, Fishel R. 1999. The interaction of the human MutL homologues
in hereditary nonpolyposis colon cancer. J Biol Chem 274:6336–6341.
Heinen CD. 2009. Genotype to phenotype: analyzing the effects of inherited
mutations in colorectal cancer families. Mutat Res. (in press).
Holmes J, Clark S, Modrich P. 1990. Strand-specific mismatch correction in nuclear
extracts of human and Drosophila melanogaster cell lines. Proc Natl Acad Sci
Hsieh P, Yamane K. 2008. DNA mismatch repair: molecular mechanism, cancer, and
ageing. Mech Ageing Dev 129:391–407.
Jiricny J. 2003. DNA repair defects in colon cancer. Curr Opin Genet Dev 13:61–69.
Kato S, Han S-Y, Otsuka K, Shibata H, Kanamaru R, Ishioka C. 2003. Understanding
the function–structure and function–mutation relationships of p53 tumor
suppressor protein by high-resolution missense mutation analysis. Proc Natl
Acad Sci USA 14:8424–8429.
Kondo E, Suzuki H, Horii A, Fukushige S. 2003. A yeast two-hybrid assay provides a
simple way to evaluate the vast majority of hMLH1 germ-line mutations. Cancer
Knudsen N Ø, Andersen SD, Lu ¨tzen A, Nielsen FC, Rasmussen LJ. 2009. Nuclear
translocation contributes to regulation of DNA excision repair activities. DNA
Lagerstedt Robinson K, Liu T, Vandrovcova J, Halvarsson B, Clendenning M,
Frebourg T, Papadopoulos N, Kinzler KW, Vogelstein B, Peltoma ¨ki P, Kolodner
RD, Nilbert M, Lindblom A. 2007. Lynch syndrome (hereditary nonpolyposis
colorectal cancer) diagnostics. J Natl Cancer Inst 99:291–299.
Li G. 2008. Mechanisms and functions of DNA mismatch repair. Cell Res 18:85–98.
Li GM, Modrich P. 1995. Restoration of mismatch repair to nuclear extracts of H6
colorectal tumor cells by a heterodimer of human MutL homologs. Proc Natl
Acad Sci USA 92:1950–1954.
Ng PC, Henikoff S. 2003. SIFT: predicting amino acid changes that affect protein
function. Nucleic Acids Res 31:3812–3814.
Nystro ¨m-Lahti M, Perrera C, Ra ¨schle M, Panyushkina-Seiler E, Marra G, Curci A,
Quaresima B, Costanzo F, D’Urso M, Venuta S, Jiricny J. 2002. Functional
analysis of MLH1 mutations linked to hereditary nonpolyposis colon cancer.
Genes Chromsomes Cancer 33:160–167.
Ou J, Niessen RC, Lu ¨tzen A, Sijmons RH, Kleibeuker JH, de Wind N, Rasmussen LJ,
Hofstra RM. 2007. Functional analysis helps to clarify the clinical importance of
unclassified variants in DNA mismatch repair genes. Hum Mutat 28:1047–1054.
Ou J, Niessen RC, Vonk J, Westers H, Hofstra RM, Sijmons RH. 2008. A database to
support the interpretation of human mismatch repair gene variants. Hum Mutat
Peltoma ¨ki P, Vasen H. 2004. Mutations associated with HNPCC predisposition—
update of ICG-HNPCC/INSiGHT mutation database. Dis Markers 20:
Perera S, Bapat B. 2008. The MLH1 variants p. Arg265Cys and p. Lys618Ala affect
protein stability while p. Leu749Gln affects Heterodimer Formation. Hum
Mutat 994: Online.
Plon SE, Eccles DM, Easton D, Foulkes WD, Genuardi M, Greenblatt MS, Hogervorst
FB, Hoogerbrugge N, Spurdle AB, Tavtigian SV, Group IARCUGVW. 2008.
Sequence variant classification and reporting: recommendations for improving
the interpretation of cancer susceptibility genetic test results. Hum Mutat
Plotz G, Welsch C, Giron-Monzon L, Friedhoff P, Albrecht M, Piiper A, Biondi RM,
Lengauer T, Zeuzem S, Raedle J. 2006. Mutations in the MutSalpha interaction
interface of MLH1 can abolish DNA mismatch repair. Nucleic Acids Res
Raevaara T, Korhonen M, Lohi H, Hampel H, Lynch E, Lonnqvist K, Holinskifeder E,
Sutter C, Mckinnon W, Duraisamy S. 2005. Functional significance and clinical
phenotype of nontruncating mismatch repair variants of MLH1. Gastroenter-
Shimodaira H, Filosi N, Shibata H, Suzuki T, Radice P, Kanamaru R, Friend SH,
Kolodner RD, Ishioka C. 1998. Functional analysis of human MLH1 mutations
in Saccharomyces cerevisiae. Nat Genet 19:384–389.
Spina Welsch C, Stallmach A, Zeuzem S, Schmidt C. 2008. Evaluation of the MLH1
I219Valteration in DNA mismatch repair activity and ulcerative colitis. Inflamm
Bowel Dis 14:605–611.
Stoffel E, Mukherjee B, Raymond VM, Tayob N, Kastrinos F, Sparr J, Wang F,
Bandipalliam P, Syngal S, Gruber SB. 2009. Calculation of risk of colorectal and
endometrial cancer among patients with Lynch Syndrome. Gastroenterology
Takahashi M, Shimodaira H, Andreutti-Zaugg C, Iggo R, Kolodner RD, Ishioka C.
2007. Functional analysis of human MLH1 variants using yeast and in vitro
mismatch repair assays. Cancer Res 67:4595–4604.
Tavtigian SV, Greenblatt MS, Lesueur F, Byrnes GB, Group IARCUGVW. 2008. In
silico analysis of missense substitutions using sequence-alignment based
methods. Hum Mutat 29:1327–1336.
Tournier I, Vezain M, Martins A, Charbonnier F, Baert-Desurmont S, Olschwang S,
Wang Q, Buisine MP, Soret J, Tazi J, Fre ´bourg T, Tosi M. 2008. A large fraction
of unclassified variants of the mismatch repair genes MLH1 and MSH2 is
associated with splicing defects. Hum Mutat 29:1412–1424.
Trojan J, Zeuzem S, Randolph A, Hemmerle C, Brieger A, Raedle J, Plotz G,
Jiricny J, Marra G. 2002. Functional analysis of hMLH1 variants and HNPPC-
related mutations using a human expression system. Gastroenterology
Vasen HFA, Mo ¨slein G, Alonso A, Bernstein I. 2007. Guidelines for the clinical
management of Lynch syndrome (hereditary non-polyposis cancer). Br Med J
Wanat J, Singh N, Alani E. 2006. The effect of genetic background on the function of
Saccharomyces cerevisiae mlh1 alleles that correspond to HNPCC missense
mutations. Hum Mol Genet 16:445–452.
Wanat JJ, Singh N, Alani E. 2007. The effect of genetic background on the function of
Saccharomyces cerevisiae mlh1 alleles that correspond to HNPCC missense
mutations. Hum Mol Genet 16:445–452.
Wang H, Hays JB. 2006. Construction of MMR plasmid substrates and analysis of
MMR error correction and excision. Methods Mol Biol 314:345–353.
HUMAN MUTATION, Vol. 31, No. 3, 247–253, 2010