Multiple Mismatches Research Articles

Neoral use in the cardiac transplant recipient

CyA is the core immunosuppressant of choice for the majority of transplant patients. The introduction of Neoral, a new microemulsion formulation of CyA. and more recently a range of adjunctive immunosuppressants have further enhanced the efficacy and tolerability of CyA-based immunosuppression. In the first year following transplantation the major causes of morbidity and death are graft failure, acute rejection, and systemic infection. Patients with deteriorated pulmonary circulation before transplantation are at increased risk of early postoperative death. Risk factors for early acute rejection include female donor sex, young donor age, and multiple HLA-DR mismatches. The principal cause of death in the long term is graft vasculopathy which accounted for 40% of all deaths. Risk factors that have been hypothesized to play a role in the pathogenesis of graft vasculopathy include hyperlipidemia, recipient age and gender, donor age, the number of HLA AB and DR mismatches, and CMV infection. Strategies proposed to reduce the risk of graft vasculopathy include aggressive use of lipid-lowering agents, avoidance of low CyA doses, and the use of adjunctive rapamycin or RAD therapy. Rejection surveillance therefore relies on routine serial endomyocardial biopsy. Recent research suggests that a more accurate assessment of the state of the graft can be obtained by considering the results across a number of biopsy samples obtained from different parts of the heart, rather than basing clinical judgment on the worst single result obtained. New molecular markers such as granzyme A mRNA are likely to improve the power of histology to diagnose and predict rejection. Neoral pharmacokinetics give greater bioavailability and less intrapatient variability than Sandimmune. In the keynote OLN 351 study comparing Neoral with Sandimmune in de novo heart transplant recipients, fewer Neoral patients needed antilymphocyte therapy to treat rejection, fewer female patients had rejection episodes in the Neoral group, the tolerability of the two formulations was equivalent, and there was a lower incidence of infections in the Neoral group. The clinical impact of Neoral in comparison with Sandimmune in de novo heart transplant patients has been investigated in a number of additional trials, including long-term studies, which have confirmed that Neoral is associated with: Lower CyA doses than Sandimmune. Equal or greater antirejection efficacy than Sandimmune. Comparable tolerability to Sandimmune. During the administration of intravenous CyA as an induction therapy in the days immediately following transplantation, there is evidence to suggest that a 6-hour infusion given twice daily, which mimics the pharmacokinetic profile of oral dosing, may be clinically more effective than a continuous 24-hour infusion. Milligram-for-milligram dose conversion from Sandimmune to Neoral is feasible. Following conversion, a reduction in the CyA dose may be required in the majority of patients to maintain target levels. In pediatric patients, the rate of elimination of CyA is greater and bioavailability increases with increasing age. Younger patients (less than 8 years of age) may be managed more effectively with a 3-times-daily, rather than a twice-daily dosing schedule. A number of studies have compared the clinical effects of Sandimmune and Neoral in maintenance therapy for cardiac transplant patients. As with de novo patients, these studies have found the new formulation of CyA to be associated with lower rates of acute rejection, lower therapeutic doses, and comparable tolerability. Milligram-to-milligram conversion from the old to the new CyA formulation is generally well tolerated, although in a minority of patients there is a significant increase in CyA levels. These may be associated with a transient increase in side effects which resolve on dose reduction. There is a dose-sparing effect with Neoral. Routine monitoring of both CyA and serum creatinine levels are adv

In vitro reactivity of allospecific cytotoxic T lymphocytes does not explain the taboo phenomenon

Matching for human leucocyte antigens (HLA) is important for graft survival in kidney transplantation. Nevertheless, most patients receive a kidney graft with multiple HLA mismatches. Some of these mismatches seem to be more harmful than others. By studying the effect of single HLA mismatches in the context of the patients' own HLA, we have previously identified donor/recipient combinations with a significantly higher incidence of early graft failure, the so-called taboo combinations. In the present study we investigated whether a higher cytotoxic T lymphocyte (CTL) response towards taboo mismatches may be involved in this phenomenon. CTL reactivity was determined both in taboo and control combinations by in vitro CTL precursor assays, using peripheral blood mononuclear cells and proximal tubular epithelial cells as target cells. Inhibition studies with CD8-antibody as well as Cyclosporin A were performed to identify high avidity and primed CTLs. Furthermore, in committed CTLp assays indirect recognition of the taboo mismatch was tested using synthetic peptides. The CTL precursor frequencies in taboo combinations were always lower than the CTL precursor frequencies in control combinations. No difference in avidity and activation status of the CTLs could be detected when taboo combinations were compared with the controls. In the committed CTLp assays no reactivity towards any of the synthetic peptides was observed. The significantly poorer graft survival of taboo combinations cannot be explained by a higher number of donor-specific CTLs. Furthermore, the avidity or activation status of these CTLs does not provide a clue to the taboo phenomenon.

Evidence for epitope spreading and active suppression in skin graft tolerance after donor-specific transfusion.

To clarify the controversial results in the literature regarding the role of donor-specific transfusion (DST) on allograft survival, we have examined the influence of the following on DST-induced allograft survival in a 2C transgenic mouse model: varying the time between DST and transplantation; the role of MHC disparities between donor and recipient; whether tolerance induced by DST spreads to skin allografts expressing other alloantigens; and whether cyclosporine (CsA) treatment could further modulate skin allograft tolerance after DST. The studies were performed in both 2C anti-Ld (MHC class I) transgenic and normal (nontransgenic) mice. Our data demonstrate that a single infusion of Ld-mismatched lymphocytes 7 days before transplantation leads to permanent acceptance of donor-specific skin allografts in both transgenic (58/58) and nontransgenic (8/8) mice in the absence of any other nonspecific immunosuppressive treatment. Pretransplantation DST from donors mismatched for more than one MHC antigen (Ag) has no beneficial effect on subsequent donor skin allograft survival. However, Ld plus multiple minor histocompatibility (mH) Ag-mismatched DST induced permanent acceptance of donor-specific skin allografts. Tolerance induced by one-locus Ld-mismatched DST spreads to skin allografts expressing either two-locus Ld or one-locus Ld plus multiple mH Ags. Administration of CsA after DST diminished skin allograft survival, rather than enhancing it, suggesting that tolerance in this model system is established by an active immunological process sensitive to CsA. (1) Pretransplantation infusion of Ld-mismatched lymphocytes in the presence or absence of multiple mH mismatches induces permanent survival of donor-specific skin allografts. (2) CsA abrogates DST-induced transplantation tolerance.

Hybridization of glass-tethered oligonucleotide probes to target strands preannealed with labeled auxiliary oligonucleotides.

In this article we introduce a strategy of preannealing labeled auxiliary oligonucleotides to single-stranded target DNA, prior to hybridization of the DNA target to oligonucleotide arrays (genosensors) formed on glass slides for the purpose of mutation analysis. Human genomic DNA samples from normal individuals and cystic fibrosis (CF) patients (including homozygous delta F508 and heterozygous delta F508/wild type (wt) in the region examined) were used. A PCR fragment of length 138 bp (wt) or 135 bp (mutant) was produced from exon 10 in the cystic fibrosis transmembrane conductance regulator (CFTR) gene, using a new pair of polymerase chain reaction (PCR) primers. This fragment contains four of the most frequent mutation sites causing the disease (Q493X, delta I507, delta F508, and V520F). Each of these mutations was tested using a pair of nonamer (9-mer) probes covalently attached to glass slides, representing the normal (wt) and the mutant alleles. Single-stranded target DNA was isolated from the PCR fragment using one PCR primer labeled with biotin and a streptavidin minicolumn to capture the biotin-labeled strand. Prior to hybridization to the 9-mer array on a glass slide, the unlabeled target strand was preannealed with one, three, or four auxiliary oligonucleotides, at least one being labeled with 32P. As observed previously in several laboratories, the discrimination between normal (wt) and mutant alleles at each site using oligonucleotide array hybridization ranged from very good to poor, depending on the number and location of mismatches between probe and target. Terminal mismatches along the probe were difficult to discriminate, internal mismatches were more easily discriminated, and multiple mismatches were very well discriminated. An exceptionally intense hybridization signal was obtained with a 9-mer probe that hybridized contiguously (in tandem) with one auxiliary oligonucleotide preannealed to the target DNA. The increased stability is apparently caused by strong base stacking interactions between the "capture probe" and the auxiliary oligonucleotide. The presence of the delta F508 mutation was detected with this system, including discrimination between homozygous and heterozygous conditions. Base mismatch discrimination using the arrayed 9-mer probes was improved by increasing the temperature of hybridization from 15 to 25 degrees C. Auxiliary oligonucleotides, preannealed to the single-stranded template, may serve several purposes to enable a more robust genosensor-based DNA sequence analysis: 1. A convenient means of introducing label into the target DNA molecule. 2. Disruption of interfering short-range secondary structure in the region of analysis. 3. Covering up of redundant binding sites in the target strand (i.e., where a given probe has more than one complement within the target). 4. Tandem hybridization with the capture probe (providing contiguous stacking) as a means for achieving efficient mismatch discrimination at the terminal position of the capture probe (adjacent to the auxiliary oligonucleotide). By use of multiple auxiliary oligonucleotides, all of the above benefits can be derived simultaneously.

DR Non-B1 Mismatches Influence Allogeneic MLR-Induced TH1- or TH2-Like Cytokine Responses in Rhesus Monkeys

Human and nonhuman primates have multiple DR B1 and non-B1 alleles. However, the role of mismatched DR non-B1 alleles in primary alloimmune responses is not well understood. Macaques, which share close DNA homologies with human MHC genes and have a high number of β-chain genes in the DR subregion, are preeminent preclinical models for immunologic studies of transplant tolerance and immunosuppression. In this study, we examined the effect of allogeneic MHC Class II DRB mismatches in Th1- and Th2-like cytokine responses elicited in one-way MLR cultures in rhesus macaques. An ELISPOT method was used to estimate cytokine secretion at the single cell level. Molecular typing for DRB1 and DR non-B1 alleles was performed by a moderate-high resolution PCR-SSP method using a panel of 55 primer pairs covering 74 DRB alleles and clusters. Of 35 unrelated combinations, 66% had multiple (≥2) allelic MM at DRB1 and DR non-B1 with no significant correlation between numbers of DRB1 and DR non-B1 mismatches. Pairs with 1 or 0 MM were assigned to a mono/null MM group to obtain sufficient numbers for statistical analysis. The pairs differing by multiple vs. mono/null DRB1 MM showed no significant difference in cytokine prevalence ( P = 0.69). In contrast, high IFN-γ/IL4 SFC ratios were noted in pairs with multiple vs. mono/null DR non-B1 MM ( p = 0.0009). IFN-γ/IL-10 spot forming cell (SFC) ratios were consistent with IFN-γ/IL-4 SFC ratios ( r = 0.98). Multiple DR non-B1 mismatches showed a trend towards higher MLR proliferative responses, although the stimulation index did not reflect the dominant cytokine response. These observations suggest a bias towards Th1-like cytokine production under allostimulation with multiple DR non-B1 gene products. Further study of the primary structure of DR non-B1 determinants may be helpful in understanding the fine molecular mechanisms governing the regulation of cytokine profiles during allostimulation in primates.

Specific tolerance without demonstrable chimerism: an alternative paradigm

5 Tolerance to kidney allotransplants (KTx) without chronic immunosuppressive therapy is achievable in several preclinical primate models, many of which show a requirement for donor bone marrow (DBM) induced chimerism. When initiated on KTx day 0, anti-CD3 immunotoxin (IT) induces stable tolerance without DBM if the proinflammatory INF-γ and TNF-α responses are prevented by peritransplant (peri-tx) treatment with NF-κB inhibitors 15-Deoxyspergualin (DSG) and methylprednisolone (MP). To determined if chimeric cells might be seeded from the KTx, we examined recipients given peri-tx IT, DSG and MP for evidence of chimerism and acquired specific unresponsiveness. Methods: 16 rhesus monkey received a KTx with immunosuppression as follows: for the 1st 3 days, IT at 150 ug/ kg total dose, MP tapered from 7 mg/kg/day, and DSG at 2.5 mg/kg/day with the DSG continued up to 14 days. All were followed for ≥ 3 months. 8 received only the KTx and 8 received additional donor CD34+ cells. All had multiple mismatches for donor MHC Class II DRB alleles and shared 1 donor DRB-1 allele. Chimerism was performed by PCR-SSP for mismatched DRB alleles. Frequency (f) changes in alloreactive PBL CTLp were assayed in limiting dilution cultures containing rhIL-2 and donor (D) or indifferent 3rd party (I) stimulator cells, and f was estimated byχ2 minimization. Recipients over 250 days were challenged with D& I skin allografts. Results: Overall, 56% of recipients had stable, long term KTx function (mean Cr =0.9 ± 0.3) and no rejection at a median 245 days with the longest survivor in good health at 513 days. Graft loss to rejection was 12.5% and 25% with and without D-CD34+ cell infusion, respectively. A mortality of 37.5% was from mixed unrelated and iatrogeneic causes without rejection, lymphomas or definable infections. At a 0.002% sensitivity level for chimerism detection, 15/16 recipients each tested on multiple (3-12) occasions in >100 assays exhibited no chimerism in blood or skin at any time; 1/16 had PBL chimerism in the 1st week. 1 had chimerism only in lymph nodes, liver & spleen, and 1 showed only thymic chimerism. 75% had a 0.6-0.9 log10 specific reduction in anti-D CTLpf without significant change in the f of anti-I CTLp. Challenge D skin grafts were accepted for 4-5 weeks while I grafts rejected in 6-9 days, Conclusions: Despite a lack of demonstrable chimerism, even after donor CD34+ cell infusion, the majority of recipients developed evidence of acquired specific tolerance. This protocol promotes development of a robust tolerance mechanism that appears to be independent of peripheral chimerism.

Rapid Detection of Genetic Mutations Using the Chemiluminescent Hybridization Protection Assay (HPA): Overview and Comparison with Other Methods

The detection of genetic mutations is of paramount importance for the study, diagnosis, and treatment of human genetic disease. Methods of detection generally fall into one of two categories: those to scan for unknown mutations and those to detect known mutations. This review focuses on methods for the detection of known mutations. The hybridization protection assay (HPA) is described in detail. The HPA method utilizes short oligonucleotide probes covalently labeled with a highly chemiluminescent acridinium ester (AE). The assay format is completely homogeneous, requiring no physical separation steps, and can rapidly and sensitively detect all single-base mismatches as well as multiple mismatches, insertions, deletions, and genetic translocations. When very low copy number targets are assayed, HPA is coupled with transcription-mediated amplification (TMA), an isothermal method that amplifies DNA or RNA targets. Other methods that are described for the detection of known mutations include hybridization with sequence-specific oligonucleotides, hybridization to oligonucleotide arrays, allele-specific amplification, ligase-mediated detection, primer extension, and restriction fragment analysis. The advantages and limitations of each of these methods are discussed. Methods to scan for unknown mutations are briefly described.

RNase YI* and RNA structure studies.

The enzymology of RNase YI*, a recently discovered endoribonuclease from yeast, was studied and compared to other endonucleases for detection of single-strand regions and base pair mismatches in RNA. Its value for RNA structure analyses was assessed with Escherichia coli 5S rRNA as a model substrate. The generally accepted structure of the 5S rRNA is based on thermodynamic energy considerations as well as structures conserved in regions of the molecule during evolution. S1 and mung bean nucleases gave similar results with very marked preference only for the longest single-stranded region in the model. RNase YI* was much more discriminating for detecting unpaired nucleotides as well as short single-strand regions and predicted the generally accepted 5S rRNA structure. Preliminary experiments also indicated that RNase YI* was more sensitive than RNase I for detecting single or multiple base pair mismatches in an RNA-DNA hybrid.

The presence of tRNA pseudogenes in mammalia and plants and their absence in yeast may account for different specificities of pre-tRNA processing enzymes

Six of 13 cloned members of the human tRNA Val gene family code for tRNA Val pseudogenes, of which all but one are transcribed efficiently in HeLa cell extracts. Due to single or multiple mismatches in stem regions, the corresponding pre-tRNAs are resistant against the action of human 5′- and 3′-processing enzymes and are thus prevented from being converted to mature tRNAs. Surprisingly, all of them are accurately and efficiently processed to mature-sized tRNA in yeast nuclear extract. This is in agreement with corresponding studies of plant pre-tRNAs which are not processed in wheat germ extract but are rapidly processed in yeast extract. These observations imply that the yeast pre-tRNA 5′-and 3′-maturases do not monitor the three-dimensional structure of their substrates as stringently as mammalian and plant enzymes, possibly because tRNA pseudogenes do not occur in yeast.

A hybrid approach to address normalization

Accuracy is critical when multiple databases are merged into a single system, because an error in a single record could lead to multiple mismatches. Address normalization is fairly common in database merging. We have developed a system to accurately and efficiently normalize mailing addresses. However, our system differs from other neural network architectures. Its key ingredients are an address dictionary and a scoring system. The scoring system is based on analog neural network systems, but the address dictionary follows a digital approach. The two key processes in our system are learning and address normalization. Learning is further split into dictionary creation updating and system parameters training. >

Misincorporation of nucleotides by calf thymus DNA primase and elongation of primers containing multiple noncognate nucleotides by DNA polymerase alpha.

Misincorporation of nucleotides by calf thymus DNA primase was examined using synthetic DNA templates of defined sequence. Primase seldom misincorporated NTPs during initiation of a new primer (i.e. polymerization of two NTPs to generate the dinucleotide). Following dinucleotide formation, however, primase readily misincorporated NTPs. Although the rate of misincorporation varied according to both the identity of the mismatch and the template sequence, primase is by far the least accurate nucleotide-polymerizing enzyme known. In some cases primase discriminated against incorrect NTPs by less than a factor of 100. After primase incorporated a noncognate nucleotide into the primer, the next correct NTP was readily added. Remarkably, primase could also polymerize consecutive noncognate nucleotides and generate primers containing multiple mismatches. Generation of a correctly base-paired primer-template negatively regulated further primer synthesis; however, generation of a primer-template containing multiple mismatches did not. After primase synthesized a primer containing multiple mismatches, the primer was transferred to the polymerase alpha active site via an intramolecular mechanism. Importantly, polymerase alpha readily elongated this primer if dNTPs were present. These data are discussed with respect to the question of why primase is required for DNA replication.

13] Site-directed mutagenesis of single-stranded and double-stranded DNA by phosphorothioate approach

Publisher Summary Oligonucleotide-directed mutagenesis allows the introduction of almost any precisely defined mutation into a cloned, sequenced gene. The mutation may comprise single or multiple mismatches or it may involve the insertion or deletion of a large number of bases. The chapter describes a number of methods for the efficient production of site-directed mutations. The gapped duplex, uracil-containing template, and coupled primer approaches have all been used to improve the basic method described in detail. However, these methods are limited to protocols using single-stranded vectors, involve the transfection of heteroduplex DNA, and very often require plaque purification steps. The original phosphorothioate-based mutagenesis method has undergone a number of improvements and has several advantages over the other methods addressed in the chapter. Finally, the phosphorothioate approach is not limited to the use of single-stranded or phagemid DNA but has been extended to double-stranded DNA vectors.

Septic Pulmonary Emboli from Pulmonic Valvular Endocarditis Demonstrated by Serial Ventilation-Perfusion Lung Imaging

The evolution of multiple ventilation-perfusion mismatches on serial lung scintigraphy in a fully anticoagulated, bacteremic patient with small pulmonary infiltrates and repeatedly negative Doppler ultrasonography of the lower extremities suggested pulmonary emboli from right-sided valvular endocarditis. The demonstration of pulmonic valvular vegetations on follow-up transesophageal echocardiography supported this diagnosis. In the proper clinical setting, septic pulmonary emboli should be included in the differential diagnosis of high-probability ventilation-perfusion lung imaging.

In vitro mutagenesis in the lacI gene of Escherichia coli: Fate of 3′-terminal mispairs versus internal base mispairs in a transfection assay

The fate of G · T mismatches and frameshifts, present at the 3′-terminus of primer-template or internally, has been studied with a combined transfection and electrophoretic assay following in vitro polymerization by DNA polymerase I (Klenow enzyme) of Escherichia coli. Several synthetic oligodeoxynucleotide primers were synthesized and annealed to uracil-containing single-stranded DNA of M13 phage bearing the lacI gene, to produce 1–3 consecutive G · T mismatches in the middle of the duplex region or at the 3′-OH end of the primer. Additional mismatched primer-templates were prepared, in which the primer had a deleted nucleotide, an extra nucleotide or both G · T mismatch and an extra nucleotide. The extension or degradation of these primers during in vitro DNA synthesis in the presence of all 4 dNTPs (‘complete’ reaction) or in the absence of dATP (‘-A’ reaction) was monitored by gel electrophoresis. Duplex DNA products were used in a transfection assay and the nucleotide changes in i − mutant progeny were determined by sequence analysis. The results suggest that whereas a single 3′-terminal G · T mismatch is relatively stable in chain elongation by Klenow enzyme, multiple terminal G · T mismatches are degraded by the 3′-exonuclease activity of this polymerase prior to primer extension. This editing activity is increased with the number of 3′-terminal mispairs. Single, double and triple T → C base substitutions were efficiently recovered when the mismatches occurred internally. Also, single-base eliminations or additions were readily recovered when the mutagenic primers contained an internal base deletion or addition, respectively. When products of the ‘-A’ misincorporation reaction (catalyzed by Klenow enzyme) were assayed by transfection, base substitutions (exclusively T → C), but no frameshifts, were recovered. The results indicate that the absence of multiple tandem base substitutions among i − mutants recovered following primer elongation under mutagenic ‘minus’ conditions was due to the efficient action of the 3′-exonuclease activity of the Klenow enzyme on multiple terminal mismatches during in vitro polymerization, rather than to in vivo events (lack of expression or occurrence of mismatch repair) in the M13- lacI transfection assay.

The induction of skin graft tolerance in major histocompatibility complex‐mismatched or primed recipients: primed T cells can be tolerized in the periphery with anti‐CD4 and anti‐CD8 antibodies

Mice given short courses of anti-CD4 and anti-CD8 monoclonal antibodies became tolerant of allogeneic skin grafted at the same time. Tolerance could be obtained without T cell depletion across multiple minor antigen mismatches, both in naive and primed animals, demonstrating that peripheral T cells could be tolerized, even if they had been previously activated. Where donor and recipient were incompatible across the whole major histocompatibility complex, specific tolerance could be achieved by using a combination of depleting followed by non-depleting antibodies, where each alone was unsuccessful. Although mice clearly tolerated their original skin grafts, we observed in some strain combinations that a second fresh, but genotypically identical graft, was slowly rejected. Such mice also possessed T cells which could proliferate to donor-type stimulator cells in vitro. Whatever the mechanisms, we have demonstrated that operational transplantation tolerance can be achieved with simple, non-toxic antibody therapy. The introduction of comparable tolerance-inducing regimens in clinical organ transplantation could obviate the need for long-term immunosuppression and its unfortunate side effects.

Mismatch repair-induced meiotic recombination requires the pms1 gene product.

The presence of multiple heterologies in a 9-kilobase (kb) interval results in a decrease in meiotic crossovers from 26.0% to 10.1%. There is also an increase from 3.5% to 11.1% in gene conversions and ectopic recombinations between the flanking homologous MAT loci. The hypothesis that mismatch repair of heteroduplex DNA containing several heterologies would lead to a second round of recombination has now been tested by examining the effect of a mutation that reduces mismatch correction. The repair-defective pms1-1 allele restores the pattern of recombination to nearly that seen in congenic diploids without the heterologies. Mismatch repair-induced recombination causes a significant increase in MAT conversions and ectopic recombination events with as few as two heterozygosities separated by 0.3-0.7 kb, but not when the mismatches are separated by greater than 1 kb. The frequency of these events depends on both the number and position of the heterozygosities relative to the flanking homologous MAT loci used to detect the events. The creation of recombinogenic lesions by mismatch repair in yeast could be analogous to the creation of recombinogenic lesions in dam- Escherichia coli. We suggest that the repair of heteroduplex DNA containing multiple mismatches may produce chromosomal rearrangements and gamete inviability when naturally polymorphic chromosomes undergo meiotic recombination.

Effect of base pair mismatches on recombination via the RecBCD pathway

The effect of base pair mismatches on recombination via the RecBCD pathway was studied in mutS and wild-type Escherichia coli, using substrates that contain single or multiple mismatches. Recombination between homologous DNA inserts in lambda phage and pBR322-derived plasmids forms phage-plasmid cointegrates that result from an odd numbers of crossovers. In the mutS host, when the sequence homology of a pair of 405 bp substrates decreased from 100% to 89%, the recombinant frequency decreased by about 9-fold, while in the wild-type host the decrease was about 240-fold. These results suggest that multiple mismatches can reduce recombinant frequencies by impeding the mechanism of recombination itself, and by provoking mismatch repair. Single mismatches in 31 bp substrates caused reductions in recombinant frequencies of 2- or 12-fold, depending on the location of the mismatch. However, unlike the reduction by multiple mismatches, the reduction of the recombinant frequencies by single mismatches was the same in both mutS and wild-type hosts. Thus a single mismatch is sufficient to impede recombination, and mismatch repair seems unable to act on single mismatches in very short homologies during recombination.

The contrasting structures of mismatched DNA sequences containing looped-out bases (bulges) and multiple mismatches (bubbles).

We have studied the structure and reactivities of two kinds of mismatched DNA sequences--unopposed bases, or bulges, and multiple mismatched pairs of bases. These were generated in a constant sequence environment, in relatively long DNA fragments, using a technique based on heteroduplex formation between sequences cloned into single-stranded M13 phage. The mismatched sequences were studied from two points of view, viz 1. The mobility of the fragments on gel electrophoresis in polyacrylamide was studied in order to examine possible bending of the DNA due to the presence of the mismatch defect. Such bending would constitute a global effect on the conformation of the molecule. 2. Sequences in and around the mismatches were studied using enzyme and chemical probes of DNA structure. This would reveal more local structural effects of the mismatched sequences. We observed that the structures of the bulges and the multiple mismatches appear to be fundamentally different. The bulged sequences exhibited a large gel retardation, consistent with a significant bending of the DNA at the bulge, and whose magnitude depends on the number of mismatched bases. The larger bulges were sensitive to cleavage by single-strand specific nucleases, and modified by diethyl pyrocarbonate (adenines) or osmium tetroxide (thymines) in a non-uniform way, suggesting that the bulges have a precise structure that leads to exposure of some, but not all, of the bases. In contrast the multiple mismatches ('bubbles') cause very much less bending of the DNA fragment in which they occur, and uniform patterns of chemical reactivity along the length of the mismatched sequences, suggesting a less well defined, and possibly flexible, structure. The precise structure of the bulges suggests that such features may be especially significant for recognition by proteins.