Phenotypic reversions at the W/Kit locus mediated by mitotic recombination in mice.

The mouse W locus encodes Kit, the receptor tyrosine kinase for stem cell factor (SCF). Kit is required for several developmental processes, including the proliferation and survival of melanoblasts. Because of the nearly complete failure of Wrio/+ melanoblasts to colonize the skin, the costs of Wrio/+ mice are characterized by a majority of white hairs interspersed among pigmented hairs, giving a roan effect. However, 3.6% of Wrio/+ mice exhibit phenotypic reversions, i.e., spots of wild-type color on their coats with an otherwise mutant phenotype. Melanocyte cell lines were derived from each of six independent reversion spots on the skin of (C57BL/6 x DBA/2)F1 Wrio/+ mice. All six melanocyte cell lines exhibited the general characteristics common to normal, nonimmortal mouse melanocytes. Of these, three revertant cell lines had lost the dominant-negative Wrio allele following mitotic recombination between the centromere and the W locus. One of the cell lines remained Wrio/+ but showed (i) stimulation in response to SCF and (ii) increased Kit expression, suggesting that the Wrio mutation can be rescued by increased endogenous expression of the c-kit proto-oncogene. Finally, two cell lines showed no detectable genetic change at the W/Kit locus and failed to respond to SCF stimulation in vitro. These results demonstrate that mitotic recombination can create large patches of wild-type hair on the coats of Wrio/+ mutant mice. This shows that mitotic recombination occurs spontaneously in normal healthy tissue in vivo. Moreover, these experiments confirm that other mechanisms, not associated with loss of heterozygosity, may account for the coat color reversion phenotype.     

Duplications That Suppress and Deletions That Restore Expression from a Chalcone Synthase Multigene Family

Seed coat color in soybean is determined by four alleles of the classically defined / (inhibitor) locus that controls the presence or absence as well as the spatial distribution of anthocyanin pigments in the seed coat. By analyzing spontaneous mutations of the / locus, we demonstrated that the / locus is a region of chalcone synthase (CHS) gene duplications. Paradoxically, deletions of CHS gene sequences allow higher levels of CHS mRNAs and restore pigmentation to the seed coat. The unusual nature of the / locus suggests that its dominant alleles may represent naturally occurring examples of homology-dependent gene silencing and that the spontaneous deletions erase the gene-silencing phenomena. Specifically, mutations from the dominant ii allele (yellow seed coats with pigmented hila) to the recessive i allele (fully pigmented) can be associated with the absence of a 2.3-kb Hindlll fragment that carries CHS4, a member of the multigene CHS family. Seven independent mutations exhibit deletions in the CHS4 promoter region. The dominant / allele (yellow seed coats) exhibits an extra 12.1-kb Hindlll fragment that hybridizes with both the CHS coding region and CHS1 promoter-specific probes. Mutations of the dominant / allele to the recessive i allele (pigmented seed coats) give rise to 10.4- or 9.6-kb Hindlll CHS fragments that have lost the duplicated CHS1 promoter. Finally, gene expression analysis demonstrated that heterozygous plants (I/i) with yellow seed coats have reduced mRNA levels, indicating that the 12.1-kb Hindlll CHS fragment associated with the dominant / allele inhibits pigmentation in a trans-dominant manner. Moreover, CHS gene-specific expression in seed coats shows that multiple CHS genes are expressed in seed coats.     

Revertant somatic mosaicism by mitotic recombination in dyskeratosis congenita

Revertant mosaicism is an infrequently observed phenomenon caused by spontaneous correction of a pathogenic allele. We have observed such reversions caused by mitotic recombination of mutant TERC (telomerase RNA component) alleles in six patients from four families affected by dyskeratosis congenita (DC). DC is a multisystem disorder characterized by mucocutaneous abnormalities, dystrophic nails, bone-marrow failure, lung fibrosis, liver cirrhosis, and cancer. We identified a 4 nt deletion in TERC in a family with an autosomal-dominant form of DC. In two affected brothers without bone-marrow failure, sequence analysis revealed pronounced overrepresentation of the wild-type allele in blood cells, whereas no such skewing was observed in the other tissues tested. These observations suggest that this mosaic pattern might have resulted from somatic reversion of the mutated allele to the normal allele in blood-forming cells. SNP-microarray analysis on blood DNA from the two brothers indeed showed independent events of acquired segmental isodisomy of chromosome 3q, including TERC, indicating that the reversions must have resulted from mitotic recombination events. Subsequently, after developing a highly sensitive method of detecting mosaic homozygosity, we have found four additional cases with a mosaic-reversion pattern in blood cells; these four cases are part of a cohort of 17 individuals with germline TERC mutations. This shows that revertant mosaicism is a recurrent event in DC. This finding has important implications for improving diagnostic testing and understanding the variable phenotype of DC.     

A transgenic,visual screenable marker for soybean seeds

Most soybean cultivars produce buff colored seeds due to a seed coat specific siRNA mechanism. This phenomenon is specifically limited to the seed coat and produces a strong visual effect, thus, a strategy to evade the silencing was used to produce a maternal transgenic marker for soybeans. Expression of a rice chalcone synthase transgene with little DNA sequence homology to the soybean siRNAs resulted in dark colored seed coats. This phenotype is the result of anthocyanin pigment production and does not appear to affect other tissues. This novel approach for producing an easily scored transgenic marker for soybean will facilitate high-throughput screening and analysis of transgenic soybean.     

Mutant gene expression in mouse aggregation chimeras. 8. The effect of the white gene on coat pigmentation

Aggregation of mouse embryos produced 11 chimaeras Miwh/+C/C----+/+c/c and 8 chimaeras +/+C/C----+/+c/c (control). Chimaerism was detected by mosaicism of coat retinal pigment epithelium and by electrophoretic pattern of glucose phosphate isomerase. All chimaeras showed a common pattern of pigmented and unpigmented hair regions that alternated as stripes of different length and width and extended from spine in lateral-ventral direction. However, white coat color predominated in Miwh/+C/C----+/+c/c chimaeras due to a higher proportion of unpigmented zones as well as to weakening of hair color in pigmented areas. Besides, distal regions of limbs were always unpigmented in Miwh/+C/C----+/+c/c chimaeras and completely or partially pigmented in +/+C/C----+/+c/c chimaeras. Pigmented hair regions are often located on the ventral trunk surface where the Miwh/+ heterozygotes usually had an unpigmented spot. The examination of hairs, taken from the same regions of gray coloration, revealed the presence of pigmented, unpigmented and mosaic hairs. The proportion of unpigmented hairs was much higher in Miwh/+C/C----+/+c/c chimaeras than in +/+C/C----+/+c/c chimaeras. The data obtained indicate that a single Miwh gene dose reduced proliferative activity of melanoblasts which resulted in weakening of coat pigmentation.     

Effect of the W mutation, for white belly spot, on testicular germ cell differentiation in mice.

The effect of the mutation for white belly spot controlled by the dominant gene W on spermatogenesis in mice was examined by experimental cryptorchidism and its surgical reversal. The course of spermatogenesis from spermatogonia to spermatid was normal in intact testes of W/+ mice. In cryptorchid testes, there was no difference in the number and activity of Type A spermatogonia between the testes of W/+ and +/+ mice, in mitotic and labelling indices. Although surgical reversal of the cryptorchid testis resulted in regenerative differentiation of germ cells in both genotypes, the recovery of cell differentiation in the W/+ testis was slower than in the +/+ testis. There were fewer germ cells, such as intermediate-Type B spermatogonia or more advanced ones, in W/+ testes. On Day 17 after surgical reversal, cell associations in W/+ testes were abnormal and the numbers of intermediate-Type B spermatogonia, spermatocytes and spermatids were approximately 70, 50 and 15%, respectively, of those in +/+ testes. These results indicate that the W gene affects spermatogenic cell differentiation in adult mice.     

Investigation of the potential for mitotic recombination in the mouse

A variation of the mouse spot test is described that is designed to distinguish between spots of altered coat colour that arise by reciprocal mitotic recombination and those caused by somatic mutation or non-disjunction. Mouse fetuses that were heterozygous for two, linked coat colour genes were irradiated (1.5 Gy X-rays) in utero at 10.25-10.50 days post coitum (p.c.) or left untreated. Subsequently, the coats were classified for the presence of spots of altered colour. The irradiated embryos were heterozygous for the linked genes pink-eyed dilution (p) and albino (c) and were produced by both the repulsion and coupling crosses. Half of the reciprocal recombination events between the centromere and the proximal marker (p), in heterozygotes with p and c in repulsion, should produce twin spots. No such twin spots would be expected from a similar event in the coupling heterozygotes. The coats of 238 irradiated and 208 untreated repulsion heterozygotes plus 107 irradiated and 314 untreated coupling heterozygotes were classified for spots. One irradiated, repulsion heterozygote had a diffuse twin spot that was only recognisable by microscopic examination of the hairs. We conclude that if the treatment described induces mitotic recombination in the mouse, it does so with low efficiency.     

Disrupted retinal development in the embryonic belly spot and tail mutant mouse

The Belly spot and tail (Bst) semidominant mutation, mapped to mouse Chromosome 16, leads to developmental defects of the eye, skeleton, and coat pigmentation. In the eye, the mutant phenotype is characterized by the presence of retinal colobomas, a paucity of retinal ganglion cells, and axon misrouting. The severity of defects in the Bst/+ retina is variable among individuals and is often asymmetric. In order to determine the role of the Bst locus during retinal morphogenesis, we searched for the earliest observable defects in the developing eye. We examined the retinas of Bst/+ and +/+ littermates from embryonic day 9.5 (E9.5) through E13.5 and measured retinal size, cell density, cell death, mitotic index, and cell birth index. We have found that development of the Bst/+ retina is notably dilatory by as early as E10.5. The affected retinas are smaller than their wildtype counterparts, and optic fissure fusion is delayed. In the mutant, there is a marked lag in the exit of retinal cells from the mitotic cycle, even though there are no observable differences in the rate of cellular proliferation or cell death between the two groups. We hypothesize that Bst regulates retinal cell differentiation and that variability of structural defects in the mutant, such as those affecting optic fissure fusion, is a reflection of the extent of developmental delay brought about by the Bst mutation.     

Effects of multiple yeast rad3 mutant alleles on UV sensitivity, mutability, and mitotic recombination.

A yeast strain was constructed that had a disruption of the chromosomal RAD3 gene and carried a series of centromeric plasmids with defined mutations in this gene. Using this isogenic collection, we examined sensitivity to UV radiation, spontaneous and UV radiation-induced mutagenesis, and mitotic recombination. Several alleles resulted in a marked increase in UV sensitivity. Most of these alleles were found to carry mutations located in consensus motifs for DNA helicases. Other alleles caused a modest or no increase in UV sensitivity and carried mutations in regions of the Rad3 polypeptide that are apparently not conserved. This correlation suggests that the DNA helicase activity of Rad3 protein is required for nucleotide excision repair of DNA. Some rad3 alleles conferred a marked increase in the frequency of spontaneous mutagenesis, including nonsuppressor reversion of the lys2-1 ochre mutation. These alleles also showed a good correlation with conserved DNA helicase domains, suggesting that the Rad3 DNA helicase also plays a role in the fidelity of DNA synthesis or postreplicative mismatch correction. Several rad3 mutator alleles also resulted in increased levels of mitotic recombination. Increased spontaneous mutagenesis and mitotic recombination are characteristic features of the Rem- phenotype. However, in contrast to the prototypic Rem- phenotype, the rad3 mutator alleles identified in this study did not confer inviability in the presence of mutations in the RAD50 or RAD52 gene required for strand break repair of DNA.     

Radiation and haematopoiesis in Harwell steel mice

Haematological information on steel (Sl) mice is limited largely to Sl/Sld mice of Bar Harbor stock (WC.B6 F1). Therefore, two Harwell alleles, SlgbH and Slcon, were investigated. In the steady state both heterozygotes were modestly anaemic, homozygous Slcon and compound Slcon/SlgbH more so. On perturbation by X-irradiation Slcon/SlgbH showed a decrease in median lethal dose (MLD)--6.5 Gy, Slcon/+ and Slcon/Slcon slightly less change (7.5 Gy) compared with +/+, 8 Gy. In recovery from sublethal doses single heterozygotes, double heterozygotes with Wv, and compounds showed no delay in restoration of the count of red blood corpuscles (RBC) such as that seen in typical W mice (e.g. Wv/+, W/Wv). Effects on Slcon/Slcon and Slcon/SlgbH differ from those reported for Sl/Sld in that they show normal growth of spleen colonies when used as lethally irradiated recipients of bone marrow, they support growth of implanted bone marrow to form radiation chimaeras. When Harwell steel mice are donors of bone marrow to lethally irradiated +/+ mice the chimaeras ultimately are not anaemic; when lethally irradiated Harwell steel mice are recipients of +/+ marrow they remain macrocytically anaemic. One deduces that, for normal development and production of normal RBC in the steady state, the erythron requires intrinsic factors determined by wild type alleles at the W locus and extrinsic factors determined by wild type alleles at the Sl locus. Mutant alleles at either locus may determine macrocytosis. Two mutant alleles at either locus are still more deleterious, often lethal. Whereas mutant W alleles may also influence the pluripotent haematopoietic stem cell (HSC) leading to reduced MLD on X-irradiation, a similarly reduced MLD for Sl mutants may represent an increased need for and consumption of products of the haematopoietic stem cells rather than truly increased radiosensitivity, since the Do for spleen colony-forming units is the same for Slcon/SlgbH as +/+ mice.     

Chalcone synthase mRNA and activity are reduced in yellow soybean seed coats with dominant I alleles.

The seed of all wild Glycine accessions have black or brown pigments because of the homozygous recessive i allele in combination with alleles at the R and T loci. In contrast, nearly all commercial soybean (Glycine max) varieties are yellow due to the presence of a dominant allele of the I locus (either I or i) that inhibits pigmentation in the seed coats. Spontaneous mutations to the recessive i allele occur in these varieties and result in pigmented seed coats. We have isolated a clone for a soybean dihydroflavonol reductase (DFR) gene using polymerase chain reaction. We examined expression of DFR and two other genes of the flavonoid pathway during soybean seed coat development in a series of near-isogenic isolines that vary in pigmentation as specified by combinations of alleles of the I, R, and T loci. The expression of phenylalanine ammonia-lyase and DFR mRNAs was similar in all of the gene combinations at each stage of seed coat development. In contrast, chalcone synthase (CHS) mRNA was barely detectable at all stages of development in seed coats that carry the dominant I allele that results in yellow seed coats. CHS activity in yellow seed coats (I) was also 7- to 10-fold less than in the pigmented seed coats that have the homozygous recessive i allele. It appears that the dominant I allele results in reduction of CHS mRNA, leading to reduction of CHS activity as the basis for inhibition of anthocyanin and proanthocyanin synthesis in soybean seed coats. A further connection between CHS and the I locus is indicated by the occurrence of multiple restriction site polymorphisms in genomic DNA blots of the CHS gene family in near-isogenic lines containing alleles of the I locus.     

Pigmented Soybean (Glycine max) Seed Coats Accumulate Proanthocyanidins during Development

The dominant I gene inhibits accumulation of anthocyanin pigments in the epidermal layer of soybean (Glycine max) seed coats. Seed-coat color is also influenced by the R locus and by the pubescence color alleles (T, tawny; t, gray). Protein and RNA from cultivars with black (i,R,T) and brown (i,r,T) seed coats are difficult to extract. To determine the nature of the interfering plant products, we examined seed-coat extracts from Clark isogenic lines for flavonoids, anthocyanins, and possible proanthocyanidins by thin-layer chromatography. We show that yellow seed-coat varieties (I) do not accumulate anthocyanins (anthocyanidin glycosides) or proanthocyanidins (polymeric anthocyanidins). Mature, black (i,R,T) and imperfect-black (i,R,t) seed coats contained anthocyanins, whereas mature, brown (i,r,T) and buff (i,r,t) seed coats did not contain anthocyanins. In contrast, all colored (i) genotypes tested positive for the presence of proanthocyanidins by butanol/ HCl and 0.5% vanillin assays. Immature, black (i,R,T) and brown (i,r,T) seed coats contained significant amounts of procyanidin, a 3[prime],4[prime]-hydroxylated proanthocyanidin. Immature, black (i,R,T) or brown (i,r,T) seed-coat extracts also tested positive for the ability to precipitate proteins in a radial diffusion assay and to bind RNA in vitro. Imperfect-black (i,R,t) or buff (i,r,t) seed coats contained lesser amounts of propelargonidin, a 4[prime]-hydroxylated proanthocyanidin. Seed-coat extracts from these genotypes did not have the ability to precipitate protein or bind to RNA. In summary, the dominant I gene controls inhibition of not only anthocyanins but also proanthocyanidins in soybean seed coats. In homozygous recessive i genotypes, the T-t gene pair determines the types of proanthocyanidins present, which is consistent with the hypothesis that the T locus encodes a microsomal 3[prime]-flavonoid hydroxylase.     

Unstable elements at the locusLuteus-1 inPetunias hybrida: Genetic analysis of their mobility

At theLul (Luteus-1) locus from Petunia, semi-dominant mutations occur, characterized by partial chlorophyll deficiencies. They can show phenotypic instability with green spot reversions. Two types of unstable mutations are described. Thelu-ml type presents a low rate of somatic reversions that are regulated by non-linked various activating factors. In contrast thelu-m3 instability shows high reversion rates regulated by a linked specific activating factor (La3), not found otherwise among the genome. Under the effect of therecombination modulator (Rm1), lu-m3 andLa3 were separated inducing a stabilization oflu-m3. This allele,lu-m3rs, however, seems to be able to respond again to a transactivation by reintroduction ofLa3 through crossing. The high transposition potential of the associationlu-m3rs-La3 was implemented to test its capacity to reinster. This experiment, conducted on a large scale, has allowed us to detect and study two unstable mutations concerning the flower pigmentation. Those two genes that control the anthocyanin production and the flavonol synthesis, respectively, are strictly linke toLul. Consequently it can be assumed that those two unstable mutations are due to reinsertion of thelu-m3rs transposons through proximal transpositions.     

Molecular Characterization of the pun Allele of the Mouse Pink-Eyed Dilution Locus

The mouse pink eyed dilution locus, p, located on chromosome 7, mediates coat and eye color. The human correlate of this gene may underlie some forms of tyrosinase-positive oculocutaneous albinism. Mutations at the p locus result in a reduction in pigmentation of the eyes and coat. Although most mutant p alleles (including all spontaneous mutations) affect only pigmentation, several mutant alleles (all radiation induced) are also associated with a variety of other phenotypes. We have focused our attention on the p^un mutant allele, a spontaneous mutation, exhibiting one of the highest reversion frequencies reported for a mammalian mutation. Using a new technique, genome scanning, we have cloned fragments of genomic DNA from the p locus that are associated with a DNA duplication in p^un DNA. These fragments can now be used to locate the p gene-encoding sequences and aid in the molecular characterization of complex mutant p alleles.     

A small tandem duplication is responsible for the unstable white-ivory mutation in Drosophila

The white-ivory (wi) mutation, an unstable allele of the white locus in Drosophila, reverts to wild-type at frequencies of 5 X 10(-5) in homozygous females, and 5 X 10(-6) in males and deletion heterozygous females. We show by molecular cloning and Southern blot analysis of DNA from wi flies that a 2.9 kilobase tandem duplication within the white locus is responsible for the mutation. Phenotypic reversion appears, in most cases, to be due to an exact excision of the extra copy of the sequence. Two derivative alleles of wi, one phenotypically wild-type, the other a partial revertant, carry insertions of moderately repetitive DNA from outside the locus, in addition to suffering deletions of some white locus DNA. Earlier genetic data preclude unequal crossing-over between homologs as an explanation for the precise reversions. Rather, an intrachromosomal meiotic event seems to be responsible. Our results suggest that intrachromosomal recombination may be responsible in other systems for a larger number of rearrangements than has been suspected, and that interallelic recombination frequencies in Drosophila do not always correlate in a simple way with DNA length or extent of homology.     

Mouse reporter strain for noninvasive bioluminescent imaging of cells that have undergone Cre-mediated recombination

Conditional alleles containing LoxP recombination sites, in conjunction with Cre recombinase delivered by a variety of means, allows for spatial and temporal control of gene expression in mouse models. Here we describe a mouse strain in which a luciferase (Luc) cDNA, preceded by a LoxP-stop-LoxP (L-S-L) cassette, was introduced into the ubiquitously expressed ROSA26 locus. Mouse embryo fibroblasts derived from this strain expressed luciferase after Cre-mediated recombination in vitro. ROSA26 L-S-L-Luc/+ mice expressed luciferase in a diffuse or liver-restricted pattern, as determined by noninvasive, bioluminescent imaging, when crossed to transgenic mice in which Cre was under the control of a zygotically expressed (EIIA-Cre), or a liver-restricted (albumin-Cre), promoter, respectively. Organ-specific luciferase expression was also seen after intraparenchymal administration of an adenovirus encoding Cre. The ROSA26 L-S-L-Luc/+ strain should be useful for characterizing Cre mouse strains and for following the fate of cells that have undergone Cre-mediated recombination in vivo.     

Enhanced susceptibility of mouse embryos heterozygous for oligosyndactyly (Os/+) to mitomycin C-induced skeletal abnormalities

The mouse mutation, oligosyndactyly (Os), results in syndactyly, muscle anomalies, and deficiency of nephrons in heterozygous animals and early embryonic lethality in homozygotes. Since the homozygous lethality results from mitotic arrest with intact spindles at the time of implantation, we have hypothesized that the heterozygous manifestations may result from impairment of cell proliferation in regions with high proliferative rates. To test this hypothesis, Os/+ and +/+ mouse embryos at 6.5 days of gestation were exposed to mitomycin C (MMC), an agent that causes a high degree of embryonic cell death which is "compensated" for by a period of rapid cell proliferation. 17.9% of MMC-treated +/+ fetuses had fused vertebrae, a significant increase over untreated fetuses, and this frequency was further increased to 33.6% in MMC-treated Os/+ fetuses. Saline treated Os/+ and +/+ fetuses showed the same low rate (0-3%) of vertebral fusion. These results indicate that Os/+ embryos have an increased sensitivity to the vertebral fusion-inducing effect of MMC at 6.5 days of gestation, a finding compatible with the hypothesis that rapid cell proliferation may be impaired in Os/+ embryos and fetuses.     

L-Asparagine auxotrophs of Saccharomyces cerevisiae: genetic and phenotypic characterization.

L-Asparagine auxotrophy in Saccharomyces cerevisiae is the result of mutation in each of two unlinked cistrons, ASN1 and ASN2. Mutation in only one of these cistrons yields growth indistinguishable from that of wild-type cells under a variety of nutritional stresses. Relatively high concentrations of L-asparagine are required to permit maximal growth of the auxotrophs, and the amino acid requirement cannot be satisfied by a variety of other amino acids that serve as nitrogen sources for cell growth. Although reversion of the mutations can occur, haploid populations of cells containing only low frequencies of prototrophs can be maintained easily. In diploid cells heteroallelic for certain combinations of alleles of the two genes, mitotic recombination gives rise to prototrophic cells that accumulate to high frequency in populations of the cells.     

The we gene is a modifier of the wal gene in mice

Interaction of gene wellhaarig (we) with genes waved alopecia (wal) and hairless (hr) was studied in mice. The mutant gene we is responsible for the development of a specific waved coat in homozygotes. Homozygous mice carrying mutant gene wal also have a wavy coat, though a partial alopecia develops with time in these animals. In homozygotes for the hr gene, hair loss is observed beginning from the age of ten days. A series of crosses we/we and wal/wal yielded animals with we/+wal/wal and we/we wal/wal genotypes. In mice we/+wal/wal carrying gene we at a single dose, alopecia is accelerated significantly as compared to the single-dose homozygotes +/+wal/wal. In we/we wal/wal mice, alopecia starts earlier than in we/+wal/wal mice; by the age of one month, the double homozygotes are almost hairless except for small body areas covered with a sparse coat. In addition, curliness of the first-generation hair in mice we/we wal/wal is much more expressed than in +/+wal/wal and we/we+/+ mice. The obtained evidence suggests that the we gene is a modifier of the wal gene because the former enhances the effects of the wal gene, which is confirmed by the earlier onset of alopecia and progression of the latter in mice having the we/+wal/wal genotype and especially in we/we wal/wal animals. The we/we hr/+ mice do not differ in coat from we/we+/+ mice; in both cases, the coat is wavy. The coat of double homozygotes we/we hr/hr, is similar to that of we/we+/+ mice until ten days of age, when the signs of alopecia appear. By the age of 21 days, mice we/we hr/hr have lost their coat completely like mice +/+ hr/hr. Hence, the we gene is a modifier of the wal gene though it does not interact with hr gene during the coat formation.     

Opposing Roles of the Holliday Junction Processing Systems of Escherichia Coli in Recombination-Dependent Adaptive Mutation

Aspects of the molecular mechanism of ``adaptive' mutation are emerging from one experimental system: reversion of an Escherichia coli lac frameshift mutation carried on a conjugative plasmid. Homologous recombination is required and the mutations resemble polymerase errors. Reports implicating a role for conjugal transfer proteins suggested that the mutation mechanism is ordinary replication error occurring during transfer synthesis, followed by conjugation-like recombination, to capture the replicated fragment into an intact replicon. Whereas conjugational recombination uses either of two systems of Holliday junction resolution, we find that the adaptive lac reversions are inhibited by one resolution system and promoted by the other. Moreover, temporary absence of both resolution systems promotes mutation. These results imply that recombination intermediates themselves promote the mutations.