Arabidopsis thaliana mutants altered in homologous recombination

Abstract

Homologous recombination contributes both to the generation of allelic diversity and to the preservation of genetic information. In plants, a lack of suitable experimental material has prevented studies of the regulatory and enzymatic aspects of recombination in somatic and meiotic cells. We have isolated nine Arabidopsis thaliana mutants hypersensitive to x-ray irradiation (xrs) and examined their recombination properties. For the three xrs loci described here, single recessive mutations were found to confer simultaneous hypersensitivities to the DNA-damaging chemicals mitomycin C (MMC^s) and/or methyl methanesulfonate (MMS^s) and alterations in homologous recombination. Mutant xrs9 (Xray^s, MMS^s) is reduced in both somatic and meiotic recombination and resembles yeast mutants of the rad52 epistatic group. xrs11 (Xray^s, MMC^s) is deficient in the x-ray-mediated stimulation of homologous recombination in somatic cells in a manner suggesting a specific signaling defect. xrs4 (Xray^s, MMS^s, MMC^s) has a significant deficiency in somatic recombination, but this is accompanied by meiotic hyper-recombination. A corresponding phenotype has not been reported in other systems and thus this indicates a novel, plant-specific regulatory circuit linking mitotic and meiotic recombination.

Many similarities in DNA repair and recombination have been noted between plants, yeast, and higher eukaryotes. Meiotic recombination events in Arabidopsis thaliana (1) and maize (2) are unequally distributed along chromosomes, resulting in silent or hot spots of recombination, as in mammalian cells or yeast (3). In mammalian cells and yeast, high recombination frequency is confined to transcriptionally active regions, and this also has been concluded for maize (4). As in other organisms (5), the frequency of meiotic recombination in plants is influenced by environmental factors (6), sequence diversity in interspecific crosses (7), and the direction of crossing (8). In somatic cells, rates of homologous recombination can be stimulated by DNA-damaging treatments (9–11). This also has been observed in bacteria, yeast, and mammalian cells (12). However, a closer comparison reveals that plant cells are highly resistant to DNA-damaging treatments and are significantly more prone to the induction of increased recombination levels (9–11) than animal cells (13). This feature may be relevant for the late differentiation of germ cells in plants and the potential transmission of somatic recombination events to the germ-line (14, 15). Interestingly, genomic changes are influenced by environmental stresses (16, 17), and specific induction of intrachromosomal homologous recombination by heat and salt stress has been documented (10, 11). Thus, regulation of genome stability during plant development is especially important. To help decipher enzymatic and regulatory aspects of DNA metabolism in plants, we isolated A. thaliana mutants hypersensitive to x-ray irradiation (xrs for x-ray sensitive), which also showed elevated sensitivity to DNA-damaging chemicals (18). Such a phenotype is indicative of a deficiency in DNA repair, most probably in recombinational repair. Here we examined homologous recombination properties of xrs mutants in somatic cells and during meiosis.

MATERIALS AND METHODS

Extrachromosomal Recombination Assay.

The mutants xrs4, xrs9, and xrs11 were isolated from ethyl methanesulfonate-mutagenized A. thaliana cv. Landsberg erecta; BC1 plants were used for this study (18).

Leaves from plants grown aseptically for 4 weeks [in 12 hr light of 60 μE/m² per sec (Biolux lamps, Osram, Munich) at 21°C and 12 hr dark at 15°C, 80% relative humidity] were placed, adaxial side down on agar-solidified (0.8%) germination medium (18). A surface of 10 cm² densely covered by leaves was used for particle bombardment. DNA-coated gold particles were delivered with a Bio-Rad particle gun following the manufacturer’s instructions, using a rupture disk of 1,100 psi (16, 19). The 5′ end (pN1) and 3′ end (pC1) deletion derivatives of the uidA gene were used as recombination substrates (17). The recombination substrates were delivered as single-stranded circular molecules of opposite polarity or as double-stranded molecules linearized at the end of the homologous overlap (pN1 cut with SalI and pC1 with BstBI). Equimolar amounts of both recombination substrates were precipitated on gold particles as described in the manufacturer’s instructions (Biolistic, Bio-Rad). Single-stranded or double-stranded DNA molecules carrying a functional uidA gene were used as controls for transformation efficiency. To decrease experimental variability, projectiles prepared with the same DNA precipitation were used to compare the different genotypes. After transformation, the plates were incubated for 48 hr [in 16 hr light of 60 μE/m² per sec (Osram Natura de Luxe lamps) and 8 hr dark at a constant temperature of 25°C, 80% relative relative humidity]. Leaves subsequently were stained for β-glucuronidase activity, and the number of blue marks scored (20). Transformation frequencies (TFs) and relative recombination frequencies (RRFs) were compared, using Student’s test with the help of the SAS statistical software (at P ≤ 0.05) (21).

Extrachromosomal Recombination Assay Under X-Ray Stimulation.

Leaves placed on the germination medium were irradiated with 20 Gray and incubated for 24 hr, 16 hr light of 60 μE/m² per sec (Osram Natura de Luxe lamps). Irradiation was repeated directly before transformation with ssDNA substrates. Control, not irradiated, leaves were incubated under the same conditions. TFs and RRFs were compared, using the Student test (21) (at P ≤ 0.05).

Meiotic Recombination Assay.

The following marker lines (provided by the Arabidopsis Seed Stock Center, Nottingham, United Kingdom) containing linked mutations (ab/ab) were used for meiotic recombination assay: NW 4 containing ch1–1 gl2–1/ch1–1 gl2–1 on chromosome 1 resulting in green-yellow leaves without trichomes, NW 5 containing hy1–1 as-1/hy1–1 as-1 on chromosome 2 resulting in elongated hypocotyl, yellow asymmetric and lobed leaves, NW9 containing ttg-1 yi-1/ttg-1 yi-1 on chromosome 5 resulting in no trichomes on stem and leaves, yellowish flower buds and yellowish sharper leaves. Plants were grown in soil, illuminated for 16 hr at 60 μE/m² per sec using Osram Natura de Luxe lamps. The temperature during the day period was 22°C ± 0.5 and during the night 16°C ± 0.5. The relative humidity was adjusted to 80%.

RESULTS

Homologous Recombination Properties of the Mutants in Somatic Cells.

Recombination efficiency of xrs mutants was compared using an assay for extrachromosomal homologous recombination (ECR). A pair of deletion derivatives of the uidA gene (coding for β-glucuronidase) sharing a 1-kb overlapping sequence was used as substrate (20) and delivered into freshly excised leaves by particle bombardment (19). Homologous recombination between two plasmid molecules within the common sequence of the uidA gene restored a functional gene. Leaves were stained for β-glucuronidase activity, and the number of stained cells or small sectors was scored (= RF). Equimolar amounts of the intact uidA gene were used for monitoring the efficiency of transformation and transient expression in control samples (TF). The TFs of the wild type and the three mutant lines were not significantly different (at P ≤ 0.05, Tables 1 and 2), which suggests similar efficiencies of delivery and expression of the intact gene in all lines. Corresponding TFs for each genotype were used for calculation of the RRFs (RRF = RF/TF). Mean RRFs in the wild type were 136.46% and 1.73% with single-stranded ssDNA and double-stranded dsDNA substrates, respectively (Tables 1 and 2) and are in agreement with previous reports (20, 22). The recombination properties of xrs11 with dsDNA and ssDNA did not differ from the wild type (at P ≤ 0.05, Tables 1 and 2). In contrast, the RRF in mutant xrs9 was approximately 20-fold lower with dsDNA and about 5-fold lower with ssDNA than in the wild type. In mutant xrs4, the reduction of RRF was less severe than in xrs9 but significant (at P ≤ 0.05) with ssDNA substrates where RRF was ca. 2-fold lower than the wild type (Table 1).

Response of Recombination to X-Ray Stimulation in Somatic Cells.

Somatic recombination can be stimulated by genomic and/or environmental stress (10, 11). X-ray-induced recombination in the wild type and mutants was compared by applying x-ray doses (20–60 Gy) as single or multiple treatments before transformation with the recombination substrates. A double irradiation with 20 Gy separated by a 24-hr incubation and transformation 1 hr after the last irradiation was most suitable. TFs in unirradiated samples did not differ from TFs in x-ray-treated leaves (at P ≤ 0.05 Table 3). The low RRFs characteristic of xrs4 and xrs9 (Tables 1 and 2) also were found after x-ray irradiation (data not shown). However, x-ray treatment resulted in about 2-fold increase in RRF in the wild-type tissue (Table 3). Mutant xrs11, which under noninductive conditions has an RRF similar to the wild type (Tables 1, 2, and 3), showed ca. 2-fold reduction in RRF after x-ray treatment, when compared with wild type (Table 3). Because the radiation dose used for induction did not impair TF, the unusual reaction of xrs11 to x-ray-mediated stimulation of recombination was not due to an overall reduction of fitness but rather to an alteration in this specific response.

Meiotic Recombination Properties of xrs4 and xrs9.

Defects in recombination in somatic cells often are accompanied by alteration of the meiotic recombination frequency (12, 23). To examine this relationship in xrs4 and xrs9, these methyl methanesulfonate (MMS) hypersensitive lines were crossed to marker lines carrying a pair of linked recessive mutations (ab/ab), resulting in visible phenotypes (see Materials and Methods) (24). Individual plants of the segregating F₂ population with a wild-type phenotype for the marker mutations (ab/AB or AB/AB) (18) were chosen for further analysis. The F₃ progenies were tested for hypersensitivity to MMS to characterize their XRS alleles and selected visually for the ab/AB genotype. As an additional control to the XRS/XRS, xrs/XRS, and xrs/xrs genotypes segregated in F₂, the wild-type L. erecta was used as a donor of the XRS9 allele in crosses with the marker lines. Meiotic recombination between marker alleles should generate recombinants with phenotypes of a single marker (aB/aB, aB/ab, Ab/Ab, or Ab/ab), and the proportion of recombinant phenotypes in the F₃ reflects the frequency of meiotic recombination at a given chromosome region. From the scores of individuals in the four phenotypic groups, meiotic recombination frequencies were calculated using the REC F2 program (25). Recombination tests for xrs9 and xrs4 were performed with markers on two different chromosomes (Table 4). Because both mutations map to chromosome 4 (18), markers on chromosomes 1, 2, and 5 were used for the experiment. Apart from a low certation (differential chance of fertilization due to pollen tube competition) at the gl locus (0.57) in xrs9 and the hy locus (0.44) in xrs4, the genetic segregation of marker alleles was not significantly distorted. The meiotic recombination frequencies of the two mutants for the genotype XRS/xrs did not differ (at P ≤ 0.05) from the values obtained for the wild-type XRS/XRS and XRS/XRS genotypes recovered from F₂-segregating individuals. In contrast, the meiotic recombination efficiency was approximately 5-fold lower for xrs9/xrs9 with both marker lines (chromosomes 1 and 5) (Table 4). In the xrs4/xrs4 genetic background, the meiotic recombination efficiency for markers on chromosome 2 was significantly increased (approximately 2.5-fold) in two independent families (Table 4, nos. 1 and 2). The recombination frequency between markers on chromosome 5 in the wild type was 35% and also was increased in xrs4/xrs4 to 51% level breaking marker linkage (Table 4).

DISCUSSION

It has been shown that in plants homologous recombination can be stimulated by genomic stress caused by x-rays, mitomycin C, and MMS treatments (9–11). This indicates that homologous recombination is involved, at least in part, in the repair of the DNA damage caused by these agents. Thus, mutant plants, which are simultaneously hypersensitive to x-rays and DNA-damaging chemicals, may be defective in recombinational repair. Our results show that homologous recombination properties in somatic cells of three chosen xrs mutants and during meiosis for two of them are significantly different from the wild-type strain.

ECR assay was used for determination of homologous recombination frequencies in somatic cells. ECR was able to reveal recombination deficiencies of rad1, rad10, and rad52 yeast mutants (26) but failed to reflect lack of mismatch repair hMLH protein, Ku p80 or ADP- ribosyl transferase in mammalian cells (27). This suggests that some of the cellular defects may be perceived by ECR and others not. In plants, efficiency of ECR correlates with the length of provided homology (20, 22) and occurs with high fidelity (28). Use of single-stranded substrates of different polarities suggested that single-strand annealing is the main ECR recombination pathway (22, 29); however, involvement in double-strand break repair also was postulated (29). All of these observations point toward specificity of ECR in plant cells, which may help to reveal possible recombination differences between DNA damage-hypersensitive mutant strains and the wild type. The recombination properties of xrs11 did not differ from the wild type (Tables 1 and 2). In contrast, the RRF in mutant xrs9 was about 20-fold lower with dsDNA and approximately 5-fold lower with ssDNA than in the wild type. This drastic reduction of RRF with dsDNA substrates may reflect a deficiency in processing dsDNA ends. In mutant xrs4, the reduction of RRF was less severe than in xrs9 but significant with ssDNA substrates where the RRF was around 2-fold lower than the wild type (Table 1). The deficiencies of xrs4 and xrs9 in ssDNA recombination resemble the phenotypes of the radiation-sensitive yeast strains rad51, 52, and 54, which seem to be affected in recombination steps after the exonucleolytic processing of double-strand break ends (30).

Homologous recombination efficiency was increased in x-ray-treated leaves of wild type, indicating that the induced pathway of recombinational repair also uses extrachromosomal molecules as substrates and that an ECR assay may reveal alterations of this pathway in mutants. Mutant xrs11 showed unaltered levels of homologous recombination under standard assay conditions but was unable to respond to stimulation of recombination by x-ray irradiation. Because such an increase in somatic recombination is a reaction to genomic and environmental stresses (10, 11), xrs11 may be affected in the transduction of signals linking DNA damage to proper cellular responses (31) rather than in the recombination process itself. It resembles yeast DUN kinase mutation, which renders cells sensitive to hydroxyurea, UV, and MMS, and abolishes specific activation of DNA damage-inducible genes (32).

Both xrs9 and xrs4 mutants affect meiotic recombination independent of chromosomal position and in a allele-specific manner. Mutant xrs9 is decreasing meiotic recombination efficiency. This again resembles the phenotypes of the yeast mutants rad51, rad52, rad54 and rad57 (12, 30) and Schizosaccharomyces pombe group three mutants altered in meiotic recombination (33). In addition, mutants of the rad52 group sporulate at a low level, and spores have reduced viability (12). The fertility of xrs4 and xrs9 also is reduced (to 10% and 50%, respectively) (18). This lower fertility is unlikely to be a consequence of aberrant gametogenesis, because genetic transmission of markers was not significantly altered in crosses in different directions. Environmental factors were shown to influence rates of meiotic recombination, which often affect particular genomic regions (6). For both mutants examined we have determined levels of meiotic recombination at two unlinked genomic locations. Results presented in Table 4 for both mutants at both genomic positions seem to be rather consistent. This suggests that genotype, rather than environment, determined the altered recombination frequencies.

Mutant xrs4 has a rather unexpected phenotype combining hypersensitivity to x-ray radiation, MMS, and mitomycin C with a reduction in homologous ECR efficiency and enhanced meiotic recombination. The only other plant mutant locus known to increase the level of meiotic recombination is Rm1 in petunia (34). In contrast to xrs4, Rm1 is dominant and thought to alter the spatial distribution of meiotic crossing-overs; its effect on somatic recombination is unknown. Mutations leading to increased meiotic recombination have been found in yeast, e.g., pms1–3, cdc5, sir2 (12), and srs2 (23); however, in contrast to xrs4, their somatic recombination levels are unaltered or increased. Thus xrs4 has a novel phenotype, which may result from a mutation in a gene regulating the mitotic and meiotic recombination pathways. Yeast mutations with phenotypes of DNA-damage hypersensitivity have been isolated recently that map to genes of signaling kinases. For example, a mutation of the HRR25 kinase gene results in increased sensitivity to x-ray radiation and MMS, reduced sporulation and spore viability, as well as decreased mitotic recombination (35). A mutation in the regulatory protein kinase C (PKC1) also influences rates of mitotic recombination (36). The effects of these mutations on meiotic recombination are not known. It is possible that XRS4 belongs to this class of regulatory genes, but with an unexpected converse influence on somatic and meiotic recombination.

Table 1.

Homologous recombination efficiency in wild type, xrs11, xrs9, and xrs4, using ssDNA substrates delivered by particle gun

Genotype	Transformation, no. of stained cells/field	Recombination, no. of stained cells/field			RRF, %			Genotype	Statistical evaluation
		Exp. 1	Exp. 2	Exp. 3	Exp. 1	Exp. 2	Exp. 3		Mean RRF, %	SE	RRF in % of wt
wt	1,345	2,169	1,020	—	161.26	75.83	—	wt	136.46	23.54	100.00
xrs11	3,627	3,441	819	1,096	94.87	22.58	30.21	xrs11	80.48	20.53	58.97
xrs9	3,665	741	379	755	20.21	10.34	20.60	xrs9	29.66*	5.99	21.73*
xrs4	2,873	913	556	1,208	31.77	19.35	42.04	xrs4	74.89*	24.46	54.88*
wt	2,914	—	3,200	6,593	—	109.81	226.25
xrs11	2,331	3,014	2,925	—	129.30	125.48	—
xrs9	2,993	1,454	1,454	2,477	48.58	48.58	82.75
xrs4	1,348	1,753	2,039	—	130.04	151.26	—
wt	1,360	1,485			109.19
xrs9	2,420	1,424			58.84

Exp., experiment; wt, wild type.

^*Significantly different from wild-type’s value at P ≤ 0.05 (Student’s t test).

Table 2.

Homologous recombination efficiency in wild type, xrs11, xrs9, and xrs4, using dsDNA substrates delivered by particle gun

Genotype	Transformation, no. of stained cells/field		Recombination, no. of stained cells/field			RRF, %			Genotype	Statistical evaluation
	Exp. 1	Exp. 2	Exp. 1	Exp. 2	Exp. 3	Exp. 1	Exp. 2	Exp. 3		Mean RRF, %	SE	RRF in % of wt
wt	2,531	2,932	44	30	69	1.61	1.09	2.50	wt	1.73	0.33	100.00
xrs11	2,371	1,265	16	83	16	0.88	4.50	0.88	xrs11	2.08	0.98	120.23
xrs9	1,431	1,512	0	3	1	0.00	0.20	0.06	xrs9	0.08*	0.04	4.62*
xrs4	1,521	2,095	18	16	23	0.99	0.88	1.27	xrs4	1.04	0.09	60.11

Exp., experiment; wt, wild type.

^*Significantly different from wild-type’s value at P ≤ 0.05 (Student’s t test).

Table 3.

Response of extrachromosomal recombination to x-ray stimulation in the wild type and the mutant xrs11

Genotype	Transformation, no. of stained cells/field		Recombination, no. of stained cells/field		RRF, %		Fold induction by x-rays	Genotype/treatment	Statistical evaluation
	− x-rays	+ x-rays	− x-rays	+ x-rays	− x-rays	+ x-rays			Mean RRF, %	SE	RRF in % of wt
wt	745	903	773	1,961	103.75	217.16	2.09	wt	75.64	15.79	100.00
xrs11	1,325	1,462	1,906	1,374	143.84	93.98	0.65	xrs11	94.50	20.43	124.93
wt	418	573	139	395	33.25	68.93	2.07	wt + x-rays	148.01*	31.73	197.32*
xrs11	875	856	399	196	45.60	22.89	0.50	xrs11 + x-rays	46.45*	16.12	61.40*
wt	279	317	265	555	94.98	175.07	1.84
xrs11	403	697	342	221	84.86	31.70	0.37

wt, wild type.

^*Significantly different from wild type’s value (− x-rays) at P ≤ 0.05 (Student’s t test).

Table 4.

Meiotic recombination properties of the mutants xrs4, xrs9 and the wild type

Genotypes tested	Phenotypic frequencies				Meiotic recombination frequency %	χc2
	P1	R1	R2	P2
XRS9/XRS9; ch1-1 gl2-1/CH1-1 GL2-1	246	51	43	34	32.19	—
†XRS9/XRS9; ch1-1 gl2-1/CH1-1 GL2-1	266	78	72	104	28.33	1.5
†xrs9/XRS9; ch1-1 gl2-1/CH1-1 GL2-1	506	133	70	101	29.37	1.18
†xrs9/xrs9; ch1-1 gl2-1/CH1-1 GL2-1	297	12	8	61	5.99	57.55*
XRS9/XRS9; ttg-1 yi-1/TTG-1 YI-1	154	33	41	27	35.12	—
†xrs9/xrs9; ttg-1 yi-1/TTG-1 YI-1	42	2	2	13	6.9	12.68*
XRS4/XRS4; hy1-1 as-1/HY1-1 AS-1	208	17	15	47	12.49	—
†xrs4/XRS4; hy1-1 as-1/HY1-1 AS-1	258	27	15	62	12.81	0.03
†1xrs4/xrs4; hy1-1 as-1/HY1-1 AS-1	134	17	25	18	27.80	11.05*
†2xrs4/xrs4; hy1-1 as-1/HY1-1 AS-1	130	16	18	5	31.00	6.90*
XRS4/XRS4; ttg-1 yi-1/TTG-1 YI-1	154	33	41	27	35.12	—
†xrs4/xrs4; ttg-1 yi-1/TTG-1 YI-1	231	57	69	15	50.99	1.65

P1 and P2, parental phenotypes; R1 and R2, recombinant phenotypes.

^*Significantly different at P ≤ 0.05.

^†Genotypes recovered from F₂ segregating individuals. Meiotic recombination of the population wild type for XRS allele was used as reference for contingency tests (χ_c²).