(deletion analysis of mitochondrial genome/molecular hybridization). A. W. LINNANE ..... ized for locus retention and genome size is measured in an analogous ...
Proc. NatI. Acad. Sci. USA Vol. 73, No. 6, pp. 2082-2085, June 1976 Genetics
Biogenesis of mitochondria: Molecular mapping of the mitochondrial genome of yeast* (deletion analysis of mitochondrial genome/molecular hybridization)
A. W. LINNANE, H. B. LUKINS, P. L. MOLLOY, P. NAGLEY, J. RYTKA, K. S. SRIPRAKASH, AND M. K. TREMBATH Department of Biochemistry, Monash University, Clayton, Victoria, 3168, Australia
Communicated by R. N. Robertson, March 29, 1976
We have developed a new procedure for the ABSTRACT detailed molecular mapping of any allele of the yeast (Saccharomyces cerevisiae) mitochondrial genome. The procedure employs a collection of different genetically characterized petite strains whose genomes have been physically defined by molecular hybridization. The map position of an allele is within the DNA segment common to all defined petites that can be shown by marker rescue to retain the locus. The same collection of petites can be used to locate the positions of mitochondrial rRNA and tRNA cistrons and DNA fragments produced by restriction endonucleases.
ferring resistance to different antibiotics. In such crosses, each individual zygote when formed contains both resistance and sensitivity alleles. During a small number of subsequent cell divisions the mitochondrial genomes segregate to yield a mixed population of diploids, each of which now carries only one allele. Polarity of transmission referred to the finding that amongst the progeny of a population of zygotes there was an unequal proportion of antibiotic-resistant and sensitive diploids. The extent of the observed polarities was locus dependent, and on the basis of relative transmission frequencies, marked loci could be arranged in a linear order as in Fig. 1, for example, The identification of mitochondrial genes and the determiSubsequent developments have shown the capl-eryl-olil. nation of their relative arrangement on the genome and the have limited application for a number of reasons. to technique to central are behavior their controlling mechanisms genetic Polarity of transmission is a feature only of a certain restricted the study of mitochondrial biogenesis. We have developed a class of cross (termed a polar cross) in which the two parent new and rapid mapping method which provides a physical map strains carry different alleles at the mitochondrial locus desigof the mitochondrial DNA (mtDNA) rather than one based nated wv (1). Furthermore, only loci which are in close physical purely on the genetic behavior of marked loci. Hence, many proximity to w in the polar region of the genome, such as capl conmore the of the difficulties and uncertainties inherent in and eryl, show differential polarity of transmission; loci in the ventional mitochondrial mapping techniques of gene transmajor or nonpolar part of the genome, such as olil, miki, and mission and recombination analysis can be overcome. par, are inseparable by this method (2). The procedure is dependent upon the establishment of a library of different Saccharomyces cerevdsiae petite strains.whose Recombination analysis genomes have been physically characterized by molecular Nonpolar crosses. In crosses in which both parents carry the hybridization. Each petite strain lacks a part, small or large, of allele at the w locus (known as nonpolar crosses) there has same the mitochondrial genome, so that one or more of the genes that been an equivalence observed in the frequency of all generally determine respiratory enzymes or are involved in their synthesis recombinant types (1, 2) such that it has not been possible to deas those such mitochondrial the of. absent. genes, are Any from single recombinants. Thus, ordering double differentiate termining reaction to various antibiotics, may or may not be analysis has not usually been posclassical three-point loci by retain of which petites lost in particular petites. From a study loci in the polar region of the those to even This sible. applies a particular mutation and which DNA segments these petites and as such eryl. capI genome are known to have in common, the map position of the mutation Recent evidence suggests that the principles of classical may be derived. All mitochondrial mutations, the genes deby recombination analysis may apply to nonpolar mapping termining mitochondrial rRNA and tRNA, and specific when all the genes under consideration are linked and crosses mtDNA fragments, such as restriction endonuclease fragments, the nonpolar regions of the genome. Thus within located are may be mapped using these petites. were able to confirm by recombination al. et (3) Trembath This paper briefly outlines the nature and limitations of order the olil-mikl-Oji as established by physical gene analysis the for previous mapping methods and develops the rationale means. new approach. The nonpolar cross has been important for the testing of alof independently derived mutations which have the same lelism TECHNIQUES FOR MAPPING antibiotic-resistance phenotype (4). When two such antibiMITOCHONDRIAL MARKERS otic-resistant mutants are crossed to each other the recombinant Analysis of the polarity of transmission class possessing the resistance alleles from both parents is indistinguishable from either parental class. Thus, recombination, The first attempts to map the mitochondrial genome were based which is evidence of nonallelism, can only be detected by the upon the analysis of diploid progeny arising from crosses begeneration of antibiotic-sensitive diploids which carry the of strains a set strain and tween a single antibiotic-sensitive sensitivity alleles of both parents. Since recombinant classes are isogenic except for a number of mitochondrial mutations conknown to occur with approximately equal frequency in nonpolar crosses, failure to detect sensitive diploids cannot be atDNA. mitochondrial mtDNA, Abbreviation: * This is paper 45 in the series. "Biogenesis of Mitochondria 44" is ref. tributed to problems of polarity as may occur in a polar cross. Although gene order cannot generally be determined in 3. 2082
Genetics: Linnane et al. par oli I mik I .....mm a a
eryl ery2 spi2
Proc. Natl. Acad. Sci. USA 73 (1976) capi spi3spi4
L IkL Er
FIG. 1. A map of mitochondrial antibiotic resistance markers based on the frequencies of transmission and recombination in polar crosses. The markers are defined by resistance to the antibiotics spiramycin (spi), chloramphenicol (cap), erythromycin (ery), paromomycin (par), oligomycin (oli), and mikamycin (mik). Each marker is phenotypically distinct with respect to antibiotic crossresistance. The derivation of the markers is documented by Trembath et al. (3). par
nonpolar crosses, linkage can be demonstrated. Linkage, which results in a recombination frequency of less than 20-25% (5) has been shown between different chloramphenicol loci (6, 7), different spiramycin loci (4), and different oligomycin loci (3, 8). Further, the chloramphenicol loci are linked to the erythromycin loci as first shown by Coen et al. (1), and the oligomycin loci are linked to the mikamycin locus (3, 9). The paromomycin locus is unlinked to either of these groups (5, 6, 10). Fig. 2 presents the genetic map which is most consistent with the data obtained from nonpolar crosses (and from other genetic analyses discussed below) and which illustrates the three independent linkage groups. Polar Crosses. As described above, polar crosses result in an ordered transmission of genes. In addition, recombination is nonreciprocal, i.e., there is a polarity of recombinant types. Once again this polarity is only a feature of a small part of the genome in the region of the chloramphenicol and erythromycin loci. For crosses involving capI or eryl loci, polarity of both recombination and transmission is observed with respect to other loci, such as olil, miki, and par. Conventional three-point analysis of these two loci in relation to a third locus clearly identifies a double recombinant class, though positive interference is generally observed (see refs. 2 and 5). Polarities of recombination between other loci not in the polar region are very low and double recombinant classes cannot be readily distinguished (5, 9). The map derived from recombination analysis in polar crosses is essentially similar to that based on polar transmission frequencies (see Fig. 1), although at present the significance of the actual recombination frequencies observed is not understood in molecular terms. Coretention and codeletion of markers in petites The loss or retention of a number of antibiotic resistance mutations in petite mutants may occur separately or in combination with each other. This finding has permitted us to develop a mapping technique based on the use of petites as random deletion mutants (11). Thus two closely linked genes will be either coretained or codeleted when a population of petites is isolated. The analysis assumes that spontaneous petites retain continuous segments of the genome, and there is now both genetic and physical evidence to suggest that the great majority of spontaneous petite mutants arise as the result of a single deletion event (11-13), implying that continuous stretches of the genome are retained. The use of coretention frequency analysis has partially overcome the difficulty of mapping by transmission or recombination. On the basis of earlier transmission or recombination data it was not possible to-position the olil, OHI, miki, or par loci relative to each other or to other marked loci. From a large number of spontaneous petite mutants it was determined that the separation of the olil locus from the mikl locus was less common than separation of either locus from the eryl and cap) loci and, where separation was seen, it was the olil rather than
FIG. 2. The genetic map of mitochondrial markers that is most consistent with data from recombination analysis in nonpolar crosses and from analysis of the frequencies of coretention and codeletion of markers in petites. The markers are as defined in Fig. 1, with addition of oligomycin-resistance markers tso, 01, 011, OII, OLGi, and OLG2, whose origins are detailed in Trembath et al. (3). co is a mitochondrial polarity locus (1).
the miki locus which was retained or lost with the eryl and capl loci (11). The interpretation of these findings is that mikl and old are closer together than either is to eryl and that the order is eryl-olil-mikl. Similarly, whenever par and miki were either retained or deleted together, the OHI locus was always retained or deleted, respectively; however, when par and mikl were separated by petite deletion the OII locus was more frequently coretained or codeleted with mik) locus than with par (3). The implied order is thus mikl-01i-par, with Oil considerably closer to miki than to par. Fig. 2 shows the genetic map most consistent with our data on gene retention in petites and is derived from studies involving the cap), ery), ohil, OLG2, mik), OII, par, and w loci (3, 11). Other loci are included on the basis of recombination analyses in nonpolar crosses. The only apparent discrepancy between this map and that based on transmission or recombination is in the extent of separation of eryl and cap). Very tight linkage between these two loci has been demonstrated in petites (see ref. 11), whereas recombination frequencies indicate a relatively large separation. Because of the proximity of the polarity locus co, the high recombination observed between cap) and eryl may be misleading in terms of physical separation, and the close linkage indicated by the petite analysis may be more representative of the true distance between these loci.
Physical mapping by molecular hybridization As a result of the petite deletion analysis, a number of stable petite strains, known to carry particular marked loci in various combinations, are available for physical analysis by molecular
hybridization (13). Hybridization of Petite mtDNA with Grande mtDNA. A small quantity of sonicated grande mtDNA of average singlestranded molecular weight 130,000, which is labeled to a high specific activity and is in solution, is hybridized exhaustively
to an excess of filter-bound DNA from petite cells. The fraction of label hybridized measures the fraction of the grande genome sequences present in the mtDNA of the various petites. Assuming the petites contain a continuous segment of mtDNA, this measurement represents the length of grande genome represented in the petite genome and imposes a maximum physical separation between the two outside markers retained in any petite. For example, a petite denoted U4 retains the loci cap), ery), old, and mik) and contains 36% of grande mtDNA sequences, indicating that the maximum possible distance be-
Genetics: Linnane et al.
Proc. Natl. Acad. Sci. USA 73 (1976)
tween the distal markers capi and miki is 36 units, where the grande mtDNA is arbitrarily defined as 100 units (13). Having measured the fractional length of the grande mtDNA present in a number of petites, it is possible to arrange these segments with respect to one another and simultaneously to determine the distance between marked loci by the following
method. Hybridization of mtDNA of a Reference Petite with Other Petite mtDNAs. The principle of mapping by this method is that the extent to which the mtDNA sequences of two petites overlap provides a direct estimate of the maximum distance encompassing any genetic loci retained in common and by difference helps to position any unique loci. The sequence homology or overlap between the mtDNA of a reference petite (U4) and that of a series of petite strains previously characterized for locus retention and genome size is measured in an analogous manner to the grande to petite hybridization, using labeled U4 reference mtDNA in solution. Consideration of the extent of homology of petite sequences with the reference petite, together with the loci retained, allows a physical positioning of the petite DNA sequences, within the limits of the experimental technique (13). Using these procedures with 13 petites in addition to U4 we have been able to assign physical map positions to the loci capi, eryl, olil, miki, and par as represented in Fig. 3 (13). Consider, for example, a strain denoted Y1.S which has been shown to retain 58 units of grande mtDNA sequences, of which 18 units are in common with U4. In its length Y1.3 retains the loci olil, mikl, and par, but has lost capi and eryl. Those retained in common with U4 are olil and miki and thus it can be concluded that olil and miki reside in one half (18 units of a total 36) of the U4 genome and by exclusion capi and eryl lie in the other half. Furthermore, because U4 does not retain par this locus must lie within 40 units of the olil-mikl end of U4. A second petite Y9 retains only the miki locus and has 25 units of grande mtDNA, of which 14 are common to U4. Thus olil, which lies within 18 units of one end of U4, cannot lie within these 14 units and thus must be within the remaining 4 units. The sequences not in common with U4 must project in the direction of par, allowing a slightly more precise positioning of par between 11 and 40 units from the mikl end of U4. In this manner the physical map in Fig. 3 has been derived. Other laboratories have made more limited estimates of the distances between their loci for chloramphenicol resistance and erythromycin resistance. Casey et al. (14) estimate a maximum separation of 12.5 units, whilst Faye et al. (15) infer a separation of between 2.5 and 10 units, both findings being consistent with our determination. No data on the location of other loci are available. The existence of petites in which certain regions of the genome are differentially amplified or in which "new" DNA sequences are present is recognized. The procedure used to overcome these problems involves labeling grande mtDNA and hybridizing it to a large excess of filter-bound petite mtDNA. The hybridized labeled sequences which are then melted off the filter are termed "petite equivalent" mtDNA and contain all the sequences the petite has in common with the grande in the proportion that they exist in the grande mtDNA and exclude all "new" sequences. Labeled "petite equivalent" mtDNA prepared from the hybridization of the reference petite mtDNA with labeled grande mtDNA may then be used for subsequent hybridization to the DNA of other petites. Hybridization of Petite mtDNA with Defined Mitochondrial RNA Species. Thegenes for the two mitochondrial rRNAs and the many mitochondrial tRNAs can be located by
FIG. 3. Physical map of Saccharomyces cerevisiae mitochondrial genome. The bars represent the region within which each mutation or rRNA cistron has been mapped. The markers are as described in Figs. 1 and 2. In addition, CS denotes a mutation conferring cold sensitivity for growth on ethanol and CO denotes a mutation responsible for a specific deficiency in cytochrome oxidase; the notations at this time are phenotypic descriptions and assignment of gene names will be made only after more extensive biochemical investigation of the lesions.
direct hybridization to the mtDNA of genetically and/or physically defined petites. Recent studies by Sriprakash et al. (16) show that the gene for 15S rRNA maps within 13 DNA units in a region that includes the par locus and that the gene for 21S rRNA maps within 15 DNA units in the region between capi and olil that includes the eryl locus (see Fig. 3). The location of the 21S rRNA is in agreement with previous reports (15, 17). The data impose a minimum separation between the genes of 27 units, a distance comparable to the one-third genome separation determined in Saccharomyces carlsbergensis on the basis of rRNA hybridization to restriction endonuclease fragments of mtDNA
Based on the results of hybridization studies of tRNAs with the mtDNA of petites characterized for the loss or retention of loci for chloramphenicol, erythromycin, and oligomycin resistance, it was concluded that (a) the histidine tRNA gene is located very close to the chloramphenicol locus; (b) the leucine and glutamic acid tRNA genes, with respect to the erythromycin locus, lie distal to the chloramphenicol locus and to the histidine tRNA gene; and (c) the genes for valine, aspartic acid, proline, alanine, tyrosine, phenylalanine, and isoleucine tRNAs do not lie in the chloramphenicol-erythromycin region of the genome (14). The use of genetically and physically defined petites which encompass regions of the genome other than the chloramphenicol-erythromycin region will no doubt facilitate the location of tRNA genes. Similarly, hybridization of mtDNA endonuclease restriction fragments to a number of genetically and physically characterized petites will rapidly locate these fragments with respect to the known genetic markers. Analysis of locus retention patterns in physically defined petites The possession of a diverse series of stable spontaneously derived petites each known to contain a physically and genetically defined region of the mitochondrial genome facilitates the location of any mutation not carried by the parental grande strain from which the petites were derived. A grande strain carrying the mutation in question is crossed to each of the series of characterized petites and the diploid progeny are analyzed for the segregation of the mutant and wild-type alleles. The detection of the wild-type allele amongst the progeny indicates the presence of the locus on the mtDNA of the petite parent. The physical position of the locus can thus be restricted to the
Genetics: Linnane et al. region of the genome in common to the petites retaining the
wild-type allele. This technique was first employed to map the oligomycinresistance loci OLG2 and Oil physically (Fig. 3) (3). Nonpolar crosses had indicated that the OLG2 locus was closely linked to olil with a recombination frequency of less than 2%, but the order of the two loci with respect to other mitochondrial loci was not clearly established. The grande strain carrying the OLG2-R allele was crossed to the test series of petites. In some crosses all the diploid progeny were oligomycin resistant and thus all carried the OLG2-R allele. The petite parent in these crosses was therefore assumed not to carry the OLG2-S allele. Other crosses yielded a mixture of oligomycin-resistant and sensitive cells, indicating the transmission of the OLG2-S allele from the petite parent. Analysis of the other markers retained or lost from the petites retaining the OLG2 locus revealed that the OLG2 locus is between mikI and o)il rather than eryl and olil. The determined physical characteristics of the petite genomes involved supported this arrangement of markers and confirmed, in physical terms, the close proximity of the olil locus to the OLG2 locus. The application of this method is not restricted to the mapping of antibiotic-resistance loci, but can be used for any locus, provided that diploids of both mutant and wild-type classes can be generated by crosses of the mutant with at least some petites. Hence, we have recently defined the genome region in which two new mutations reside. First, a mitochondrially determined mutant which is normal at 280 but unable to grow on non-fermentable substrates at low temperature (18°), when crossed to certain of the series of physically and genetically defined petite strains, gives rise to a population of diploids which consists of both cold-resistant and cold-sensitive cells. Analysis of the mtDNA of these petites shows they have common DNA segments between the arbitrary map positions 89 and 2 units (Fig. 3). The cold-sensitive mutant is as yet biochemically uncharacterized. The second mutation probably belongs to a class recently described by Tzagoloff et al. (19). The strain is a presumptive cytochrome oxidase mutant, as it contains the other cytochromes b, c, and cl as well as retaining the capacity for mitochondrial protein synthesis. The diploid progeny of crosses with some, but not all, of the test series of defined petites include a proportion of respiratory competent cells, so that petites able to generate such recombinant cells must carry the wild-type allele of the mutant locus. The mtDNA sequences unique to these petites are confined to the region 76 to 84 units, between the cap) and par loci (Fig. 3). Details of these two studies will
Proc. Natl. Acad. Sci. USA 73 (1976)
be reported elsewhere. We are unable to position these mutations with greater precision, as our library of defined petites at present does not include sufficient strains whose genomes are known to end in this region. The procedures outlined herein are of general applicability for the mapping of any mitochondrial mutation. Once a collection of physically defined petites is assembled in a laboratory new mutations with recognizable phenotypes can be physically mapped within days. The way is now open for a complete map of the mitochondrial genome to be rapidly developed. 1. Coen, D., Deutsch, J., Netter, P., Petrochilo, E. & Slonimski, P. P. (1970) Symp. Soc. Exp. Biol. 24, 109-127. 2. Linnane, A. W., Howell, N. & Lukins, H. B. (1974) in The Biogenesis of Mitochondria. Transcriptional, Translational and Genetic Aspects, eds. Kroon, A. M. & Saccone, C. (Academic Press, New York), pp. 193-213. 3. Trembath, M. K., Molloy, P. L., Sriprakash, K. S., Cutting, G. J., Linnane, A. W. & Lukins, H. B. (1976) Mol. Gen. Genet., in press. 4. Trembath, M. K., Bunn, C. L., Lukins, H. B. & Linnane, A. W. (1973) Mol. Gen. Genet. 121, 35-48. 5. Wolf, K., Dujon, B. & Slonimski, P. P. (1973) Mol. Gen. Genet. 125,53-90. 6. Kleese, R. A., Grotbeck, R. C. & Snyder, J. R. (1972) J. Bacteriol. 112, 1023-1025. 7. Molloy, P. L., Howell, N., Plummer, D. T., Linnane, A. W. & Lukins, H. B. (1973) Biochem. Biophys. Res. Commun. 52,9-14. 8. Lancashire, W. E. & Griffiths, D. E. (1975) Eur. J. Biochem. 51, 403-413. 9. Howell, N., Molloy, P. L., Linnane, A. W. & Lukins, H. B. (1974) Mol. Gen. Genet. 128,43-54. 10. Kutzleb, R., Schweyen, R. J. & Kaudewitz, F. (1973) Mol. Gen. Genet. 125, 91-98. 11. Molloy, P. L., Linnane, A. W. & Lukins, H. B. (1975) J. Bacteriol. 122,7-18. 12. Clark-Walker, G. D. & Miklos, G. L. G. (1975) Proc. Natl. Acad. Sci. USA 72,372-375. 13. Sriprakash, K. S., Molloy, P. L., Nagley, P., Lukins, H. B. & Linnane, A. W. (1976) J. Mol. Biol., in press. 14. Casey, J. W., Hsu, H.-J., Getz, G. S., Rabinowitz, M. & Fukuhara, H. (1974) J. Mol. Biol. 88, 734-749. 15. Faye, G., Kujawa,. C. & Fukuhara, H. (1974) J. Mol. Biol. 88, 185-203. 16. Sriprakash, K. S., Choo, K. B., Nagley, P. & Linnane, A. W. (1976) Biochem. Biophys. Res. Commun., 69,85-91. 17. Nagley, P., Molloy, P. L., Lukins, H. B. & Linnane, A. W. (1974) Biochem. Biophys. Res. Commun. 57,232-239. 18. Sanders, J. P. M., Heyting, C. & Borst, P. (1975) Biochem. Biophys. Res. Commun. 65,699-707. 19. Tzagoloff, A., Akai, A. & Needleman, R. B. (1975) Proc. Natl. Acad. Sci. USA 72,2054-2057.