Cytoplasmic Influence on the Expression of Nuclear Genes Affecting Life Span in Drosophila M Elanogaster

Heredity 66 (1991) 259—264 Received 26 June 1990 Genetical Society of Great Britain

Cytoplasmic influence on the expression of nuclear genes affecting life span in Drosophila m elanogaster

ISAMUYONEMURA, TOMIO MOTOYAMA,* HAYATO HASEKURA & BARRY BOETTCHERt Departmentof Legal Medicine, Shinshu University School of Medicine, Matsumoto, 390 Japan, *Department of Physiology, College of Environmental Health, Azabu University, Sagami-hara, 229 Japan and tDepartment of Biological Sciences, The University of Newcastle, New South Wales, 2308 Australia

Inearlier studies we have found that the difference between short and long life spans of two inbred strains of Drosophila melanogaster is controlled by nuclear major genes. The present study has revealed a cytoplasmic factor that influences the expression of the nuclear longevity genes. The factor shows a typical maternal inheritance and is considered to be an extranuclear gene, such as mitochondrial DNA (chondriome). This paper marks the discovery of two basic forms of inheritance, nuclear and extra-nuclear, in the genetics of life span of D. melanogaster. These findings suggest that further studies, including genetic engineering, on longevity and aging might enable direct manipulation of these characters.

Keywords:Drosophilamelanogasrer, extranuclear genes, gene expression, life-span, nuclear major genes.

Introduction The present work reports the discovery of a new genetic influence on life span of D.m. involving extra- Fromthe results obtained in a number of studies, it has nuclear genes as well as nuclear genes, Jm A and Jm X. been concluded that life span is under polygenic Thus, two fundamental modes of inheritance, nuclear control. Genetic contribution to variation of life span is and extranuclear, are shown to influence life span. not large; environmental conditions such as temperature or food are major determinants. Heritability of life span has been calculated as 0.13 3—0.442 in Drosophila Materialsand methods (Maynard Smith, 1959; Tantawy & Rakha, 1964), 0.2—0.3 in nematodes (Johnson & Wood, 1982), and Twoshort-lived purebred strains of D. m. were used as 0.21—0.36 in mice (Storer, 1966). the parental stocks, P1 and P2. These were mated Goodrick (1975), however, obtained high herit- according to the combinations listed in Table 1 to ability (0.47—0.787) of life span in mice, and his calcu- produce reciprocal F1 hybrids, F2 and all backcross lations indicated that the inheritance of life span in generations. Flies were bred at 27°C on food contain- mice was controlled by a single gene. The present ing 8 per cent dry yeast and 5 per cent sugar. The dead authors have undertaken mating experiments with two were counted every day and were removed from the purebred strains of Drosophila melanogaster (D. m.) culture vials. The P1 strain was established from a wild and have demonstrated that major genes at two loci, type stock of D.m. through 13 or more generations of autosomal and X chromosomal, determined their life sib-mating; their genes -are presumed to be homo- spans. These longevity genes were analysed for their zygous at 95 per cent or more loci. This strain is identi- mode of inheritance and were named Jm (Yonemura cal with the P1 generation (the short-lived parent) used et a!., 1989, 1990). Luckinbill et al. (1988) also in previous work (Yonemura et al., 1989, 1990). The succeeded in selection of life span in D. m., and demon- P2 strain in the present work is a purebred line estab- strated that the difference in life span in their stocks lished by the same procedure from Oregon R. Details was mainly determined by a few genes on chromosome of culture and handling methods were the same as in an 1 (X chromosome) and chromosome 3 (autosome). earlier report (Yonemura et al., 1989). 259 260 I. YONEMURA ETAL.

These results show that the inheritance of life span in Results this experiment was controlled by few major genes. Themeans and standard deviations of the parental On the other hand, the scaling test of Mather (1949) stocks, the F1 and F2 generations, and the reciprocal requires the following calculations, taking the means of backcross generations are given in Table 2. These data B1, B2 and F2 generations as ,B2and F2 and those of — — ignore the obvious differences between the sexes or P1, P2 and F1 as P1, P2 and F1: A =2B1 P1 Fl, between different crosses in the same generations. B=2B2—P2--F1, C=4F2—2F1—P1—P2. However, the procedure is necessary to enable the If A =B=C=0, genes in the model are considered following calculations. to have only additive and dominance effects and to If we take the phenotypic variance in the F2 as Vi,, show no epistatic effect; epistatic effect is an interac- the genotypic variance as VG, and the environmental tion between gene loci. Calculations based on the data variance as VE, then V =V+ VE. VE is given by in Table 2 result in A=9.4±0.814, B3.4±0.804 (V+ Vp2+2VF)/4 using variances of P1, P2 and F1. and C 12.8 2.12, which are significantly larger than From Table 2 it can be calculated that V,= 289.00 and 0 (P< 0.05). This shows interaction variance is large. VE= 153.05. Heritability in the broad sense, h2b, is This is inconsistent with the above conclusion that VG given by V0/V. Then h2b can be calculated as solely consists of VA. Consequently, the quantitative h2b=0.4704. VG is divided into additive variance genetic analysis is invalid. (VA), dominance variance (VD) and interaction variance The distribution of life spans of the different genera- (V1): VG= VA+ V0+ V1. VA is given by 2VF2— VB, tions of D.m. studied here are shown in Fig. 1 (F, F1 which is calculated from Table 2 as VA= 305.75. This and F2 generations) and Fig. 2 (backcross generations). figure is larger than VG =135.95calculated before, The progenies showing the shortest mean life spans indicating that VD and V1 take no part in VG. In such a were P2 and generations originating from P2 females: genetic model the number of loci of the contributing F1_1, F2_1, B1_1, B2_1, B2_3, B2_4. The exception to genes can be calculated as N =R2/8VA= 0.03 (where R this general observation was B1_3. It is pointed out is the difference of mean life span between P1 and P2). that, although P2 females iii the parental generation initiated the B13 generation, the females in the back- Table 1 Notation of generations in reciprocal crosses cross (P1) contained no contribution (either nuclear or cytoplasmic) from P2 females. Mating Progeny In the progenies where the females used for breed- ing did not have a contribution from F2 females, F1 -2' P1dxp2Q 1I F22,B12, B13, B14, B22, long-lived flies were P19xP2d F12 observed. Such long-lived flies were not observed F1_1dxF1_19 F2_1 among the parental strains, including P1 which, simi- F12dF129 F22 larly, did not have a contribution from P2 females. P,dxF119 B11 P1dxF1_29 B12 P19xF1_1d B13 Discussionsand conclusions P19xF1_2d B14 P2dxF119 B21 Thestriking observation in this study is the marked P2dxF129 B22 difference in the life spans of progenies from reciprocal '2Qx1'11d B23 crosses. For example, F11 and F12; F21 and F22 P29xF12d B2_4 (Fig. 1). A difference between the mean life spans of male F1 flies from reciprocal crosses might be Table 2 Sample number in each generation and the mean accounted for by differences in genes located on the X life span. Values are means s.d. (days) and Y chromosomes of the parental stocks. However, such an explanation could not account for the differ- Generation Number Life span ence in the mean life spans of female F1 flies observed here (F1 and F1 -2'Fig.1). However, an extranuclear P1(d+9) 641 36.5±5.6 influence appears to be involved. All progenies where P2(d+9) 640 28.1±3.1 the cytoplasm of the zygote originated from the P2 F1(total) 1297 48.1 16.9 F2(total) 1303 43.4± 17.0 parental strain, F2, F1 —1' F2 —1' B1 —1' B2 —1' B2 , showeda unimodal, short, life span. On the other B1(total) 2539 47.0± 15.8 B24' B2(total) 2582 39.8 16.3 hand, progenies where the cytoplasm of the zygote B(total) 5121 43.4± 16.5 originated from the P1 parental strain, but the nuclear genetic material contained a contribution from the P2 CYTOPLASMIC INFLUENCE ON LIFE SPAN IN DROSOPHILA 261

A IA IX IY parental strain, F1_2 F2_2, B1_2, B1_3, B1_4, B2_2, AIAIXIXIEl showed a wide range of life spans, including long life spans not observed in the parental stocks. Consequently, the results indicate that P2 possesses a cytoplasmic factor which has a life-shortening effect which can over-ride the nuclear longevity genes, and which follows maternal inheritance accurately. This is the reason why the variance in the F1 (VF =VF + VF2) increased, and why the above quantitative genetic analysis is invalid. A2A2X2X2E2 In generations where P2 females made no contribution to the cytoplasm of the zygote, longevity graphs appeared nearly identical with those in the previous reports and their genotypic values could be likewise assigned (Yonemura etal.,1989, 1990). The nuclear genes which influenced life span to lengths far greater than those of the parental stocks, are concluded as being from the P2 parents, in which their effects were countered by the life-shortening cytoplasmic factor. In order to account for the life spans of the flies observed 20 a, A2X2YlE2 here, we have to add the cytoplasmic factor, or extra- 15 AIA2XIX2E2 nuclear gene, to the Jm longevity gene systems 0 pr323 observed and studied in our previous reports (Yone- 010 - mura et a!., 1989, 1990). The cytoplasmic factor, or extranuclear gene, is 5 denoted as El for P1 strain and E2 for P2. In the 0 I previous reports (Yonemura et al., 1989, 1990), the I0-F: nuclear longevity genes for P1 were assigned as Jm Al AIA2XIY2El7k AIA2XIX2-EI (autosomal) and Jm Xl and Jm Yl (sex chromosomal). In the present study, nuclear genes for P2 were denoted 51_?0330 \J Jm A2, Jm X2 and Jm Y2 (although identity of these 0 i I 'i I genes with those of the P2 strain in the previous reports 20- F2_ is not certain). From the present experimental results, 15 extranuclear genes are to be added to the genotypes as (n:327 below: o P1 males: A1A1X1Y1 El, P1 females: AJA1XJXJ El, H h\- I I I males:A2A2X2Y2E2, 0-F2_2 2 (3n35I P2 females: A2A2X2X2E2. 5 Genotypes of other generations may be simply 0 10 20 30 40 50 60 70 80 90 100 assigned. In Figs 1 and 2, the genotypes for corre- Life span (days) sponding peaks are indicated (except for F2, to avoid congestion in the figure). Fig.1 Life spans of stocks of Drosophila melanogaster, All generations possessing E2 were short-lived, where day of emergence is given as day 0. irrespective of their other Jm genotypes. The shortest Distributions of mortality rates in parent generations mode of the mortality peak in males was 27 days in the (P1,P2) and their reciprocal crosses (F1,F2). The solid line is F2, and the longest was 32 days in the F1 -1;infemales, for males and the dotted line is for females. Genotypes corresponding to each peak are shown in the figure. Jm Al and Jm the peak commonly appeared a little later. In the A2 are autosomal longevity genes; Jm Xl and Jm X2 are generations with El, the observed segregation of life X-chromosomal genes; Jm Yl and Jm Y2 are genes assumed span was according to the other Jm genotypes; geno- on the Y chromosome; Jm El and Jm E2 are extranuclear typic values could be estimated from the data in Figs 1 genes, influencing life span. and 2. Taking the mode of the mortality peak as the 262 I. YONEMURA ETAL.

AIA2XlY2E2 A2A2XIY2.E2 AlA 2X2Y2 E2 AIAIXIYIE2 A2A2X2Y2E2 AIAIX2YIE2 5 A]A2XlYIE2 AlA2X2YlE2 l0 AIAIXlXIE2 A LA IX IX2E2 5 Al A2 XIX IE2 AIA2XIX2E2 0

B AlAIXlX2EIAIAIX2YI•EI /. AIA2X1YIE1 B22 _—T AlA 2X IX 2E I 101311AIAIXIXlEI .::::- AIA2XtXlEI AIA2X2X2EI n330 A2A2XlX2EI 51 EEII fl AlA2XlY2EI A2A2XIY2EI A2A2X2X2EI

El,_1\frVI 1 - I I A I. J,., I I I 0 AIAIXIX2EI >' ALA2X2YI•E2 C A2A2X2YIE2 0 AIA2X2X2. E2 A2A2X2X2E2

I I 25 - B_4 524 AIA2X2Y2E2 20- n32I n32I A2A2X2Y2E2 nr350 5 r——AA2XIY2.EI AIA2XIXI.El AIA2X lX2E2 10 - AIAIXIX1•EI A2A2X IX2E2

I I I I 0901000 10 20 30 40 50 60 70 80 90 100 Life span (days) Fig.2 Life spans of stocks of Drosophilamelanogaster,where day of emergence is given as day 0. Distributions of mortality rates in backcross generations. The solid line is for males and the dotted line is for females. Genotypes corresponding to each peak are shown. Jm Al and Jm A2 are autosomal longevity genes; Jm Xl and Jm X2 are X-chromosomal genes; Jm Yl and Jm Y2 are genes assumed on the Y chromosome; Jm El and Jm E2areextranuclear genes, influencing life span.

A]A2XJX1 = genotypic value, genotypes possessing El showed the 66days (B1_1: 65, B1 _.66). (5) following values. AIA2X2YI =AJA2X2Y2 =AIA2X1X2 =AJA2X2X2 = = = AIA1X1Y1AJA1X1Y2=37days(P1:35,B1_2:39, A2A2X2Y2A2A2X1X2A2A2X2X2=72 days B1_3: 36, B1_4: 36). (1) (F1_2: 69, B1_2: 72, B2_2: 74, B.3: 68). (6) From these equations, genotypic values (in days) AIA1X1X1 =42 days (F1: 39, B1_3: 43, B1_4: 43). (2) were calculated as follows: A1A1X1X2=A1A1X2Y1 = 50days (B1_2: 48, Yl=Y2=0, B1_3: 48, B1_2: 53). (3) AIA1=32, A1A2XIYI =A1A2XJY2=A2A2X1Y2= 59 days (F12: 57, B1: 61, B13: 57, B1_: 60, AJA2=52, B2_2: 58, 60). (4) A2A2=52, CYTOPLASMIC INFLUENCE ON LIFE SPAN IN DROSOPHILA 263

Xi =5 (toAlA])and 7 (to AJA2 and A2A2), D.m. has been noted to diminish with aging; Massie & Williams (1987) noticed an association between this X2= 18 (toAlAl)and 20 (to A1A2 and A2A2). quantitative variation and senility. Yl and Y2 show no effect of longevity genes. A2 is The present experiment has shown that the E2gene dominant to Al. A1AI presents 32 days of basic life has a strong suppressive effect on nuclear genes, since span, while AIA2 and A2A2 give 52days.Xi adds 5 A2A2X2Y2 individuals, which have a life span of 73 days of supplemental life to A1AI and 7 days to A1A2 days with El, the longest life span of D.rn. at 27°C, live and A2A2; this effect shows a dosage effect since only 30 days with E2, the shortest life span of D.m. homozygous females live longer than hemizygous This difference of life span is attributed to a qualitative males. X2 has a higher additive effect than Xl, adding difference between El and E2 in influencing life span, 18 days to AJAI and 20 days to AJA2 and A2A2. In rather than a quantitative difference of mitochondrial long-lived generations where segregation of life span is DNA, as suggested by Massie & Williams (1987). It is observed, the short-lived trait originates from P1 and possible that a large difference of life span among the long-lived trait from P2. These data clearly species is based on differences between extranuclear demonstrate that the short-lived P2 possesses nuclear longevity genes rather than between nuclear genes. longevity genes for a long life span, but an extranuclear In D.m., developmental speed in the preimaginal factor suppresses their expression. Table 3 shows all stage has been found to have inter-relations with the genotypic values under the control of extranuclear quantitative traits such as adult life span (Wattiaux, factor El, calculated from the above data. 1962; Burcombe & Hollingsworth, 1970; Lints & The life span mode of P1 males, the flies with the Lints, 1971). We have found that nuclear longevity shortest life span in the presence of El is 35 days, and genes control developmental speed in the preimaginal the longest mode under the control of E2 is 32 days. stage, from larval stage to emerging (Yoriemura et al., This indicates that the life-shortening effect of E2 1991). It will be important to determine whether the influences not only long-lived genotypes but also short- extranuclear genes disclosed by the present report also lived genotypes. That is to say, all Jm nuclear longevity influence developmental speed. It seems easier to genes are influenced by E2. investigate the molecular mechanisms whereby long- The present study has demonstrated a typical evity genes influence developmental speed rather than maternal inheritance of short life span, indicating the to demonstrate how they determine life span. Further, presence of cytoplasmic factor whichinfluences the now that an extranuclear, and presumably mito- expression of Jm longevity genes. It is inappropriate to chondrial, factor has been shown to have a large assume that this factor is a product of the mother, such influence on life span, it appears feasible to undertake as a hormone or other metabolic substance, trans- physical manipulations, including studies involving mitted through cytoplasm of the ovum. If this was the recombinant DNA technology, with a view to altering case, it would be expected to produce aneffect only in life spans of stocks on D. m. the F11 generation (P2 female X P1 male), but it could hardly leave an effect on all individuals in the F2 -1' the Acknowledgment next generation from the F1 -1• Consequently, this factor can be assessed as a substance reproduced by Thisstudy was partly supported by the Smoking the F11 and transmitted through the F11 females to Research Foundation (Japan). the F2 -1' that is to say, a self-reproducing extranuclear gene, most probably part of themitochondrial DNA (chondriome). We tentatively name the factor in P1 as References Jm El and in P2 as Jm E2.MitochondrialDNA in BURCOMBE,J. V. AND HOLLINGSWORTh, M. i. 1970. The relationship between developmental temperature and longevity in Drosophila. Gerontologia, 16, 172—181. Table 3 Life span given by each nuclear longevity genotype 000DRICK, C. L 1975. Life-span and the inheritance of long- (genotypic value in days) under the influence of Jm El, an evity of inbred mice. J. Gerontol., 30,257—263. extranuclear gene JOHNSON, T. E. AND wOOD, W. B. 1982. Genetic analysisof life- span in Caenorhabditis elegans. Proc.Nati. Acad. Sci. X1X2 X1Y1 USA,79,6603—6607. — X1Y2 X1X1 X2X2 LINTS, F. A. AND LINTS, c. v. 1971. Relationship between growth and ageing in Drosophila melanogaster. Nature (New 50 A1A1 32 37 42 Biol.), 229, 86—88. A1A2 52 59 66 72 LUCKINBILL, L. 5., GRAVES, J. L., REED, A. H. AND KOESTAWANG, S. A2A2 52 59 66 72 1988. Localizing genes that defer senescence in Droso- 264 LYONEMURA ETAL.

phila melanogaster. Heredity, 60, 367—374. to speed of development in Drosophila melanogasrer. MASSIE, H. R. AND WILLIAMS, T. R. 1987. Mitochondrial DNA Nature, 194, 706—707. and life span changes in normal and dewiriged Drosophila YONEMURA, I., ABE, M., ISHIDATE, R., ISHIYAMA, T., MOTOYAMA, T., at different temperatures. Exp. Gerontol., 22, 139—153. HASEKURA, H. AND BOETFCIIER, B 1990. Influence of tempera- MAThER,K. 1949.Biometrical Genetics, Methuen, London. ture on the inheritance of adult life span in Drosophila MAYNARD SMITH, .1959.Sex-linked inheritance of longevity in melanogaster. Hereditas, 112, 117—127. Drosophila subobscura. J. Genet., 56, 227—23 5. YONEMURA, I., MOTOYAMA, T. AND HASEKURA, H. 1989. Mode of STORER, J. B. 1966. Longevity and gross pathology at death in inheritance of major genes controlling life span differences 22 inbred mouse strains. J. Gerontol., 21, 404—409, between two inbred strains of Drosophila melanogaster. TANTAWY, A. 0. AND RAKHA, F. A. 1964. Studies on natural Hereditas, 111,207—214. populations of Drosophila, IV Genetic variances of and YONEMURA, I., MOTOYAMA, T., HASEKIJRA, H. AND B0ETrcHER, B. correlations between four characters in D. melanogaster 1991. Relationship between genotypes of longevitygenes and D. simulans. Genetics, 50, 1349—1355. and developmental speed in Drosophila melanogaster. WArFIAUX, j.M.1962. Variation of bristle number in relation Heredity, 66, 143—149.