Mutations induced by 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) in the lacZ and cII genes of Muta™ Mouse

4-(Methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) found in chewing tobacco, snuff, cigarettes, and cigars is a tobacco-specific nitrosamine and classified as a possible human carcinogen (Class 2B) by the International Agency for Research on Cancer (IARC). NNK given intraperitoneally was seen to induce lung and liver adenomas.To evaluate the genotoxicity of NNK in vivo, NNK was intraperitoneally administered to Muta™ Mouse at two concentrations (125 and 250 mg/kg, once a week for 4 weeks) followed by the measurement of mutant frequencies in the lacZ and cII genes from lung and liver in the same mice. Characterization of the types of the mutation was determined by sequencing the cII genes from mutant plaques. The mutant frequencies in both target genes from both organs dose-dependently increased up to 10 times compared to those of the control group. For the types of mutations, the ratio of the G:C to A:T mutation in the total number of mutants was less than the ratio of A:T to T:A and A:T to C:G transversion, contrary to a previous report [Cancer Res, 49 (1989) 5305]. The A:T to T:A transversion was the most highly induced mutation both in the lung and liver cII genes. The increasing rate of mutant frequencies in lung and liver over the vehicle control was 55 and 56 times, respectively, while the increasing rate of G:C to A:T transition was only 1.9 and 2.8 times, respectively.These observations show that NNK predominantly induces DNA adducts leading to A:T to T:A and/or A:T to C:G mutations in the transgene.     

DNA sequence analysis of spontaneous and γ-radiation (anoxic)-induced lacI mutations in Escherichia coliumuC122::Tn5: differential requirement for umuC at G · C vs. A · T sites and for the production of transversions vs. transitions

Escherihica coliumC122::Tn5 cells were γ-radiated (¹³⁷Cs, 750 Gy, under N₂), and lac-constitutive mutants were produced at 36% of the wild-type level (the umC strain was not deficient in spontaneous mutagenesis, and the mutational spectrum determined by sequencing 263 spontaneous lacI^d mutations was very similar to that for the wild-type strain). The specific nature of the umC strain's partial radiation was determined by sequencing 325 radiation-induced lacI^d mutations. The yields of radiation-induced mutation classes in the umC strain (as a percentage of the wild-type yield) were: 80% for A · T → G · C transitions, 70% for multi-base additions, 60% for single-base deletions, 53% for A · T → C · G transversions, 36% for G · C → A · T transitions, 25% for multi-base deletions, 21% for A · T → T · A transversions, 11% for G · C → C · G transversions, 9% for G · C → T · A transversions and 0% for multiple mutations. Based on these deficiencies and other factors, it is concluded that the umC strain is near-normal for A · T → G · C transitions, single-base deletions and possibly A · T → C · G transversions; is generally deficient for mutagenesis at G · C sites fro transversions, and is grossly deficient in multiple mutations. Damage at G · C sites seems more difficult for translesion DNA synthesis to bypass than damage at A · T sites, and especially when trying to produced a transversion. The yield of G · C → A · T transitions in the umC strain *36% of the wild-type level) argues that a basic sites are involved in no more than 64% of γ-radiation-induced base substitutions in the wild-type strain. Altogether, these data suggest that the UmuC and UmuD′ proteins facilitate, rather than being absolutely required for, translesion DNA synthesis; with the degree of facilitation being dependent both on the nature of the noncoding DNA damage, i.e., at G · C vs A · T sites, and on the nature of the misincorporated base, i.e., whether it induces transversions or transitions.     

The comet assay with MCL-5 cells as an indicator of genotoxic treatment with chemicals and cigarette smoke condensates

The metabolically competent human lymphoblastoid cell line MCL-5 was treated with a panel of mutagens to assess the induction of DNA damage. Treatment effects were observed by monitoring cell proliferation and by single-cell gel electrophoresis (SCGE). The direct-acting mutagens benzo[a]pyrene-7,8-diol 9,10-epoxide (BPDE) and 1-methyl-3-nitro-1-nitrosoguanidine (MNNG), as well as pro-mutagens requiring metabolic activation, i.e. benzo[a]pyrene (BaP), 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP), 4-N-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK), and cigarette-smoke condensate (CSC), were assayed by SCGE. Assay schemes were adapted for the MCL-5 cell line and for low levels of strand break induction, by inclusion of the DNA synthesis inhibitors cytosine arabinoside and hydyroxyurea, and by extending the electrophoresis time. For all mutagens tested, dose-dependent increases of median and average tail moment values among 50 nucleoids per slide were observed. The determining factors for selecting the treatment doses for mutation-induction experiments were the solubility of BaP and PhIP in the exposure medium, and the cytotoxicity exhibited by BPDE, MNNG and CSC. Induction of DNA strand breaks was obtained at mutagen concentrations permitting sufficient cell proliferation, except in the case of MNNG.     

A strand bias occurs in point mutations associated with variant surface glycoprotein gene conversion in Trypanosoma rhodesiense.

We previously described a bloodstream Trypansoma rhodesiense clone, MVAT5-Rx2, whose isolation was based on its cross-reactivity with a monoclonal antibody (MAb) directed against a metacyclic variant surface glycoprotein (VSG). When the duplicated, expressed VSG gene in MVAT5-Rx2 was compared with its donor (basic copy) gene, 11 nucleotide differences were found in the respective 1.5-kb coding regions (Y. Lu, T. Hall, L. S. Gay, and J. E. Donelson, Cell 72:397-406, 1993). Here we describe a characterization of two additional bloodstream trypanosome clones, MVAT5-Rx1 and MVAT5-Rx3, whose VSGs are expressed from duplicated copies of the same donor VSG gene. The three trypanosome clones each react with the MVAT5-specific MAb, but they have different cross-reactivities with a panel of other MAbs, suggesting that their surface epitopes are similar but nonidentical. Each of the three gene duplication events occurs at a different 5' crossover site within a 76-bp repeat and is associated with a different set of point mutations. The 35, 11, and 28 point mutations in the duplicated VSG coding regions of Rx1, Rx2, and Rx3, respectively, exhibit a strand bias. In the sense strand, of the 74 total mutations generated in the three duplications, 54% are A-to-G or G-to-A (A:G) transitions and 7% are C:T transitions, while 26% are C:A transversions and 13% are C:G transversions. No T:G or T:A transversions occurred. Possible models for the generation of these point mutations are discussed.     

Investigation of mutant frequency at the HPRT locus and changes in microsatellite sequences in healthy young adults

In an attempt to understand the inter-individual variation that occurs in in vivo mutant frequency at the HPRT locus, we have examined the effect of polymorphisms in genes for metabolic enzymes on the mutation rate. In the same population of human volunteers, the background variant frequency in a number of microsatellite sequences was studied to determine individual variation in the capacity to repair mismatches in these sequences. The HPRT mutant frequency of T-cells isolated from a group of 49 healthy, non-smoking adults varied from 0.25 to 9.64×10⁻⁶. The frequency of polymorphisms in CYP1A1, GSTM1 and NAT2 among these individuals was similar to those published, and when subjected to univariate analysis these polymorphisms showed no influence on the HPRT mutant frequency. However, there was a significant interaction between the GSTM1 null genotype and the slow acetylator status in NAT2 (P<0.05) which was associated with higher mutant frequency. Analysis of 30 microsatellite sequences in 20 HPRT proficient clones per individual showed only six alterations in total, giving an overall mutation rate per allele of 0.01%, whilst three alterations were found in five HPRT deficient clones per individual examined for changes in 10 microsatellites, giving an overall mutation rate per allele of 0.3%. Thus, the alterations detected are probably due to background mutations and not to differences in mismatch repair capacity.     

Impact of maternal lifestyle factors on newborn HPRT mutant frequencies and molecular spectrum — Initial results from the Prenatal Exposures and Preeclampsia Prevention (PEPP) Study

Epidemiological studies have demonstrated associations between maternal tobacco smoke exposure and consumption of alcohol during pregnancy and increased risk of pediatric malignancies, particularly infant leukemias. Molecular evidence also suggests that somatic mutational events occurring during fetal hematopoiesis in utero can contribute to this process. As part of an ongoing multi-endpoint biomarker study of 2000 mothers and newborns, the HPRT T-lymphocyte cloning assay was used to determine mutant frequencies (M_f) in umbilical cord blood samples from an initial group of 60 neonates born to a sociodemographically diverse cohort of mothers characterized with respect to age, ethnicity, socioeconomic status, and cigarette smoke and alcohol exposure. Non-zero M_f (N=47) ranged from 0.19 to 5.62×10⁻⁶, median 0.70×10⁻⁶, mean±SD 0.98±0.95×10⁻⁶. No significant difference in M_f was observed between female and male newborns. Multivariable Poisson regression analysis revealed that increased HPRT M_f were significantly associated with maternal consumption of alcohol at the beginning [Relative Rate (RR)=1.84, 95% CI=0.99–3.40, P=0.052) and during pregnancy (RR=2.99, 95% CI=1.14–7.84, P=0.026). No independent effect of self-reported active maternal cigarette smoking, either at the beginning or throughout pregnancy, nor maternal passive exposure to cigarette smoke was observed. Although based on limited initial data, this is the first report of a positive association between maternal alcohol consumption during pregnancy and HPRT M_f in human newborns. In addition, the spectrum of mutations at the HPRT locus was determined in 33 mutant clones derived from 19 newborns of mothers with no self-reported exposure to tobacco smoke and 14 newborns of mothers exposed passively or actively to cigarette smoke. In the unexposed group, alterations leading to specific exon 2–3 deletions, presumably as a result of illegitimate V(D)J recombinase activity, were found in five of the 19 mutants (26.3%); in the passively exposed group, two exon 2–3 deletions were present among the seven mutants (28.6%); and in the actively exposed group, six of the seven mutants (85.7%) were exon 2–3 deletions. Although no overall increase in HPRT M_f was observed and the number of mutant clones examined was small, these initial results point to an increase in V(D)J recombinase-associated HPRT gene exon 2–3 deletions in cord blood T-lymphocytes in newborns of actively smoking mothers relative to unexposed mothers (P=0.011). Together, these results add to growing molecular evidence that in utero exposures to genotoxicants result in detectable transplacental mutagenic effects in human newborns.     

A novel mutation screening system for Ehlers-Danlos Syndrome, vascular type by high-resolution melting curve analysis in combination with small amplicon genotyping using genomic DNA

Ehlers-Danlos syndrome, vascular type (vEDS) (MIM #130050) is an autosomal dominant disorder caused by type III procollagen gene (COL3A1) mutations. Most COL3A1 mutations are detected by using total RNA from patient-derived fibroblasts, which requires an invasive skin biopsy. High-resolution melting curve analysis (hrMCA) has recently been developed as a post-PCR mutation scanning method which enables simple, rapid, cost-effective, and highly sensitive mutation screening of large genes. We established a hrMCA method to screen for COL3A1 mutations using genomic DNA. PCR primers pairs for COL3A1 (52 amplicons) were designed to cover all coding regions of the 52 exons, including the splicing sites. We used 15 DNA samples (8 validation samples and 7 samples of clinically suspected vEDS patients) in this study. The eight known COL3A1 mutations in validation samples were all successfully detected by the hrMCA. In addition, we identified five novel COL3A1 mutations, including one deletion (c.2187delA) and one nonsense mutation (c.2992C>T) that could not be determined by the conventional total RNA method. Furthermore, we established a small amplicon genotyping (SAG) method for detecting three high frequency coding-region SNPs (rs1800255:G>A, rs1801184:T>C, and rs2271683:A>G) in COL3A1 to differentiate mutations before sequencing. The use of hrMCA in combination with SAG from genomic DNA enables rapid detection of COL3A1 mutations with high efficiency and specificity. A better understanding of the genotype–phenotype correlation in COL3A1 using this method will lead to improve in diagnosis and treatment.     

Chloroplast DNA base substitutions: an experimental assessment

An experimental assessment was carried out to determine directly the frequency and types of spontaneous base substitutions that occur in chloroplast DNA. A target site within the chloroplast 16S rRNA gene of the green alga Chlamydomonas reinhardtii was chosen for the assay. Mutations at this site were known to confer spectinomycin resistance and simultaneously result in the loss of an AatII cleavage site. In the experiments reported here, base substitutions at any individual base occurred at a frequency in the range of 0.9–11 per 10⁹ viable cells plated. Four new mutations that confer resistance to spectinomycin were identified at the target site in the Chlamydomonas chloroplast 16S rRNA gene. When the relative rates of transition and transversion mutations were quantified, a bias toward transversions was observed. The prominence of A/T C/G transversions in the observed mutation spectrum suggests that oxidative damage may be the major cause of base substitution mutations within the chloroplast.     

Mutational specificity of thymine deprivation-induced mutation in the lacI gene of Escherichia coli

LacI⁻ mutants obtained following 2 and 6 h of thymine deprivation were cloned and sequenced. The mutational spectra recovered were dissimilar. After 2 h of starvation the majority of mutations were base substitutions, largely G: C→C: G transversions. Frameshift mutations but not deletions were observed. In contrast, following 6 h of starvation, with the exception of the G: C→C: G transversion, all possible base substitutions were recovered. Moreover, several deletions but no frameshift events were observed. The differences in the mutational spectra recovered after two periods of thymine deprivation highlight the role of altered nucleotide pools and the potential influence of DNA replication mechanisms.     

Ribavirin-Resistant Mutants of Human Enterovirus 71 Express a High Replication Fidelity Phenotype during Growth in Cell Culture

It has been shown in animal models that ribavirin-resistant poliovirus with a G64S mutation in its 3D polymerase has high replication fidelity coupled with attenuated virulence. Here, we describe the effects of mutagenesis in the human enterovirus 71 (HEV71) 3D polymerase on ribavirin resistance and replication fidelity. Seven substitutions were introduced at amino acid position 3D-G64 of a HEV71 full-length infectious cDNA clone (26M). Viable clone-derived virus populations were rescued from the G64N, G64R, and G64T mutant cDNA clones. The clone-derived G64R and G64T mutant virus populations were resistant to growth inhibition in the presence of 1,600 μM ribavirin, whereas the growth of parental 26M and the G64N mutant viruses were inhibited in the presence of 800 μM ribavirin. Nucleotide sequencing of the 2C and 3D coding regions revealed that the rate of random mutagenesis after 13 passages in the presence of 400 μM ribavirin was nearly 10 times higher in the 26M genome than in the mutant G64R virus genome. Furthermore, random mutations acquired in the 2C coding regions of 26M and G64N conferred resistance to growth inhibition in the presence of 0.5 mM guanidine, whereas the G64R and G64T mutant virus populations remained susceptible to growth inhibition by 0.5 mM guanidine. Interestingly, a S264L mutation identified in the 3D coding region of 26M after ribavirin selection was also associated with both ribavirin-resistant and high replication fidelity phenotypes. These findings are consistent with the hypothesis that the 3D-G64R, 3D-G64T, and 3D-S264L mutations confer resistance upon HEV71 to the antiviral mutagen ribavirin, coupled with a high replication fidelity phenotype during growth in cell culture.     

Mutagenic specificity of N-acetoxy-3-aminobenzanthrone, a major metabolically activated form of 3-nitrobenzanthrone, in shuttle vector plasmids propagated in human cells

3-Nitrobenzanthrone (3-NBA) is a potent environmental mutagen and a potential human carcinogen present in diesel exhaust and airborne particulates. N-acetoxy-3-aminobenzanthrone (N-Aco-ABA) has been shown to be a major reactive metabolite of 3-NBA, which mainly produces adducts with guanine and adenine in cellular DNA. Here we analyzed mutations induced by N-Aco-ABA using supF shuttle vector plasmids to elucidate the mutagenic specificity of 3-NBA in human cells. Base sequence analysis of more than 100 plasmids with supF mutations induced in wildtype and DNA repair-deficient XP cells revealed that the major mutation was base substitutions of which the majority (42 and 38%, respectively) were G:C to T:A transversions. The next major mutation was G:C to A:T and A:T to G:C base substitutions in wildtype and XP cells, respectively. The DNA polymerase stop assay using N-Aco-ABA-treated plasmids as a template showed that most stop signals, i.e., adducted sites, appeared at G:C sites. These results suggest that N-Aco-ABA binds preferably to guanine rather than adenine, and adducted adenine is repaired more efficiently by the nucleotide excision repair. Error-prone DNA polymerases could insert adenine at sites opposite to N-Aco-ABA-adducted guanine, which leads to G:C to T:A transversion. These findings could be very important to evaluate the human lung cancer risk of environmental 3-NBA.     

Analysis of oxidative DNA damage and HPRT mutations in humans after hyperbaric oxygen treatment

DNA damage induced by reactive oxygen species (ROS) seems to play an important role in the induction of mutations and cancer. We have recently shown that hyperbaric oxygen (HBO) treatment of volunteers (i.e., exposure to 100% oxygen at a pressure of 2.5 ATA) induces DNA damage detected in leukocytes with the comet assay. Using formamidopyrimidine-DNA glycosylase (FPG protein) we provided indirect evidence for the induction of oxidative DNA base damage. We now comparatively evaluated FPG-sensitive sites with the comet assay and 7,8-dihydro-8-oxo-deoxyguanosine (8-OHdG) with HPLC analysis after a single HBO. As 8-OHguanine (8-OHgua) is one of the major DNA modifications induced by ROS and a pre-mutagenic lesion, we looked for HBO-induced mutations at the HPRT locus with the T cell cloning test. We also determined the genotypes for glutathione transferases (GST) and tested a possible influence of the GSTM1 and GSTT1 genotypes on the sensitivity of subjects against HBO-induced genotoxicity. Our results indicate that despite a clear induction of FPG-sensitive sites no increased levels of 8-OHdG and no induction of HPRT mutations was detected in lymphocytes after HBO. Furthermore, the DNA effects in the comet assay and the mutant frequencies in the HPRT test seem to be unrelated to the GST genotypes of the test subjects.     

Total gene deletions and mutant frequency of the HPRT gene as indicators of radiation exposure in Chernobyl liquidators

This study was conducted to determine the utility of deletion spectrum and mutant frequency (MF) of the hypoxanthine phosphoribosyl transferase gene (HPRT) as indicators of radiation exposure in Russian Liquidators who served in 1986 or 1987 in the clean up effort following the nuclear power plant accident at Chernobyl. HPRT MF was determined using the cloning assay for 117 Russian Controls and 122 Liquidators whose blood samples were obtained between 1991 and 1998. Only subjects from whom mutants were obtained for deletion analysis are included. Multiplex PCR analysis was performed on cell extracts of 1080 thioguanine resistant clones from Controls and 944 clones from Liquidators. Although the deletion spectra of Liquidators and Controls were similar overall, the Liquidator deletion spectrum was heterogeneous over time. Most notable, the proportion of total gene deletions was higher in 1991–1992 Liquidators than in Russian Controls (χ²=10.5, p=0.001) and in 1993–1994 Liquidators (χ²=8.3, p=0.004), and was marginally elevated relative to 1995–1996 Liquidators (χ²=3.3, p=0.07). This type of mutation has been highly associated with radiation exposure. Total gene deletions were not increased after 1992. Band shift mutations were also increased in the 1991–1992 Liquidators but were associated with increased MF of both Liquidators and Controls (p=0.009), not with increased MF in 1991–1992 Liquidators (p=0.7), and hence are not believed to be associated with radiation exposure. Regression analysis demonstrated that relative to Russian Controls HPRT MF was elevated in Liquidators overall when adjusted for age and smoking status (37%, p=0.0001), and also was elevated in Liquidators sampled in 1991–1992 (72%, p=0.0076), 1993–1994 (22%, p=0.037), and 1995–1996 (62%, p=0.0001). In summary, HPRT MF was found to be the more sensitive and persistent indicator of radiation exposure, but the specificity of total gene deletions led to detection of probable heterogeneity of radiation exposure within the exposed population.     

Identification of two new nucleotide mutations (HPRTUtrecht and HPRTMadrid) in exon 3 of the human hypoxanthine-guanine phosphoribosyltransferase (HPRT) gene

Mutations in the X-linked hypoxanthine-guanine phosphoribosyl transferase gene (HPRT) result in deficiencies of HPRT enzyme activity, which may cause either a severe form of gout or Lesch-Nyhan syndrome depending on the residual enzyme activity. Mutations leading to these diseases are heterogeneous and include DNA base substitutions, DNA deletions, DNA base insertions and errors in RNA splicing. Identification of mutations has been performed at the RNA and DNA level. Sequencing genomic DNA of the HPRT gene offers the possibility of direct diagnostic analysis independent on the expression of the mature HPRT mRNA. We describe a Dutch and a Spanish family, in which the Lesch-Nyhan syndrome and a severe partial HPRT-deficient phenotype, respectively, were diagnosed. Direct sequencing of the exons coding for the HPRT gene was performed in both families. Two new exon 3 mutations have been identified. At position 16676, the normally present G was substituted by an A in the Dutch kindred (HPRT_Utrecht), and led to an arginine for glycine change at residue 70. At position 16680, the G was substituted by a T in the Spanish family (HPRT_Madrid); this substitutes a valine for glycine at residue 71. These new mutations are located within one of the clusters of hotspots in exon 3 of the HPRT gene in which HPRT_Yale and HPRT_{New Haven} have previously been identified.     

Nucleotide Correspondence between Protein-Coding Sequences of <Emphasis Type="Italic">Helicobacter pylori</Emphasis> Strains 26695 and J99

Comparison of ORFs between H. pylori strains 26695 and J99 showed that transitions (more than 3%) prevail over transversions (less than 1%). The predominance of transitions was explained by the high rates of cytosine replacement by thymine in the coding (3.5–5.3%) and noncoding (2.9–3.9%) DNA strands. The proportion of transversion-type correspondences (A → C, A → T, C → A, C → G, G → C, G → T, T → A, and T → G) did not exceed 0.84%. The highest proportion (28.3%) was observed for correspondences between C and T in ACGT-ATGT, the target site of active methyltransferase of H. pylori J99 (M.Hpy99XI). It was assumed that C → T mutations due to cytosine methylation-deamination are prevalent in H. pylori.     

Polymorphism of prion protein gene in Arctic fox (<Emphasis Type="Italic">Vulpes lagopus</Emphasis>)

Prion diseases are fatal neurodegenerative disorders of humans and certain other mammals. Prion protein gene (Prnp) is associated with susceptibility and species barrier to prion diseases. No natural and experimental prion diseases have been documented to date in Arctic fox. In the present study, coding region of Prnp from 135 Arctic foxes were cloned and screened for polymorphisms. Our results indicated that the Arctic fox Prnp open reading frame (ORF) contains 771 nucleotides encoding 257 amino acids. Four single nucleotide polymorphisms (SNPs) (G312C, A337G, C541T, and A723G) were identified. SNPs G312C and A723G produced silent mutations, but SNPs A337G and C541T resulted in a M–V change at codon 113 and R–C at codon 181, respectively. The Arctic fox Prnp amino acid sequence was similar to that of the dog (XM 542906). In short, this study provides preliminary information about genotypes of Prnp in Arctic fox.     

Determination of tobacco-specific N-nitrosamines in urine of smokers and non-smokers

Tobacco-specific N-nitrosamines (TSNA) include 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK), N′-nitrosonornicotine (NNN), N′-nitrosoanabasine (NAB) and N′-nitrosoanatabine (NAT) and are found in tobacco and tobacco smoke. TSNA are of interest for biomonitoring of tobacco-smoke exposure as they are associated with carcinogenesis. Both NNK and NNN are classified by IARC as Group 1 carcinogens. Samples of 24?h urine collections (n?=?108) were analysed from smokers and non-smokers, using a newly developed and validated LC-MS/MS method for determining total 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol (NNAL, the major metabolite of NNK), and total NNN, NAB and NAT. TSNA levels in smokers’ urine were significantly higher than in non-smokers. In smokers, urinary excretion of total TSNA correlated significantly (r?>?0.5) with markers of smoking dose, such as daily cigarette consumption, salivary cotinine and urinary nicotine equivalents and increased with the ISO tar yield of cigarettes smoked. The correlation between urinary total NNN and the smoking dose was weaker (r?=?0.4–0.5). In conclusion, this new method is suitable for assessing tobacco use-related exposure to NNK, NNN, NAB and NAT.     

Inosine induced mutations

Summary Two complementary 24 base single stranded oligonucleotides containing randomly located inosine residues were synthesized in vitro. Once annealed, the two oligonucleotides were cloned into derivatives of ColE1 and transformed into Escherichia coli. Sequence analysis of 157 clones yielded 305 mutations. The pattern of the mutations revealed the following: (1) The frequency of inosine induced mutations was significantly less than predicted from its content in the oligonucleotides; (2) Inosine incorporation resulted almost exclusively in base changes to guanine; (3) The mutation distribution is biased towards A/T to G/C substitutions; (4) There were reproducible position biases; and (5) There was a reproducible strand bias which was independent of the cassette orientation with respect to the plasmid origin of replication.