- Email: [email protected]
Expression of the mitochondrial genome in HeLa cells
J. Mol. Biol. (1971) 55, 271-276
LETTERS TO THE EDITOR
Expression of the Mitochondrial IV.? Titration
of Mitochondrial
Genome in HeLa Cells
Genes for 16 s, 12 s and 4 s RNA
RNA-DNA hybridization experiments utilizing separated strands of HeLa mitochondrial DNA and purified 16 s, 12 s and 4 s mitochondrial RNA species from HeLa cells have indicated the existence of one 16 s gene and one 12 s gene per mitochondrial DNA molecule, which are located in the heavy mitochondrial DNA strand, and of approximately eleven genes for 4 s RNA: of these about eight appear to be located in the heavy strand and about three in the light strand.
We have shown previously that mitochondrial RNA is synthesized in HeLa cells, as in rat liver cells (Borst & Aaij, 1969), almost exclusively on the heavy strand of mitochondrial DNA, and, moreover, that the whole or almost whole length of this strand is transcribed (Aloni & Attardi, 1971). RNA-DNA hybridization experiments (Attardi & Attardi, 1969; Dawid, 1969; Nass & Buck, 1969,1970), kinetics and inhibition studies (Vesco & Penman, 1969; Zylber & Penman, 1969 ; Dubin & Montenecourt, 1970 ; Attardi C%Attardi, 1971) have revealed that the bulk RNA synthesized on mitochondrial DNA in HeLa and other animal cells is represented by discrete species with sedimentation coefficients of 4 S, 12 s and 16 S. In addit’ion, there are heterogeneous RNA components spread in the 4 to 50 s region and minor discrete fast sedimenting components (Aloni & Attardi, 1971; Attardi et al., 1970). In this letter we present the results of RNA-DNA hybridization experiments which strongly suggest that there is one gene for 12 s RNA and one gene for 16 s RNA per mitochondrial DNA molecule, which are located in the heavy mitochondrial DNA strand, and about 11 genes for mitochondria-associated 4 s RNA, 8 of which are located in the heavy strand and 3 in the light mitochondrial DNA strand. The methods for preparation of the mitochondrial fraction, extraction and analysis of RNA and RNA-DNA hybridization have been described in detail in previous reports (Aloni & Attardi, 1971; Attardi & Attardi, 1969,1971). Figure 1 shows the sedimentation pattern of RNA extracted from the EDTAtreated mitochondrial fraction of HeLa cells exposed for four hours to [S3H]uridine in the presence of 0.04 pglml. actinomycin D (to inhibit the synthesis of ribosomal RNA (Perry, 1964; Dubin, 1967; Penman, Vesco & Penman, 1968)). One can see that the 28 s and 18 s RNA are not labeled, while the 16 s, 12 s and 4 s peaks stand out clearly. The ratio of label in the 16 s and 12 s RNA in this experiment was about 1.5. On the basis of the molecular weight estimates derived from the electrophoretic mobilities of the two RNA components (0.7 x lo6 and 0.4 x 106) (Attardi & Attardi, 1971), of the relative sedimentation velocities of the two species after denaturation with formaldehyde, which are consistent with a molecular weight ratio of about 1.7 (Attardi & Attardi, 1970), and of electron microscope length measurements which indicat’e a molecular weight of about 3.5 x lo5 for the 12 s RNA and 55 x lo5 for the 16 s RNS (Robberson, Davidson, Aloni & Attardi, manuscript in preparation), it is reasonable t Paper III in this series is Attardi 8: Ojala (1971). 271
Y. ALONI
272
AND
G. ATTARDI
3.010 2.0106 1 I.O(1 0.8 0 0.6 0 04 04 I 02 0 06 0 1
04 0:
02 0 I
. o4 : 0.8 Ot
04
4
Cd) w c 4 I i
:
I
I .
-I
I t
-0 -0
:
t.-#4r, 0
; 4
IO
20
30
40 Frocticn
50
’ 60
70
no
FIU. 1. Purification of the 16 s, 12 s and 4 s RNA components from the mitoohondrial fraction of HeLa cells labeled for 4 hr with [5-3H]uridine in the presence of actinomyoin D. RNA was extracted from the EDTA-treated crude mitochondrial fraction isolated by differential centrifugation from the homogenate of 8 x lo* HeLa cells labeled with [6-3H]uridine for 4 hr in the presence of 0.04 pg/ml. actinomycin D. The RNA was run through a 15 to 30% (w/w) sucrose gradient in sodium dodecyl sulfate buffer (0.57, sodium dodecyl sulfate, 0.1 aa-NaGl, 0.01 aa-Tris buffer pH 7.0, 0.001 M-EDTA) in the SW27 Spinco rotor (1.69 cm x 10.16 cm buckets)
LETTERS
TO
THE
EDITOR
273
to assume that the two RNA species are synthesized in equimolar amounts. The recent demonstration that these species represent the structural components of specific 60 s mitochondrial ribosomes is in agreement with this assumption (Swanson & Dawid, 1970; Attardi & Ojala, 1971). After rerunning in sucrose gradient, the 16 s RNA (Fig. l(b)) and the 12 s RNA (Fig. l(c)) appear as fairly sharp peaks of coinciding 3H radioactivity and o.~.se~. The 4 s RNA was separated from other slow sedimenting components by acrylamide gel electrophoresis (Fig. l(d)). In order to eliminate any DNA contaminant, the three discrete RNA species were subjected to extensive DNase treatment, sodium dodecyl sulfate-phenol extraction and Sephadex GlOO chromatography, as previously described (Aloni & Attardi, 1971). For the determination of the specific activity, a portion of the Sephadex eluate corresponding to the peak of radioactivity was plated directly on a nitrocellulose membrane of the same type used for the RNA-DNA hybridization experiments and counted in the scintillation counter. The concentration of RNA solutions was determined from o.D.,,,, measurements by using a value of 214 for E$? at 260 nm in 0.1 M-NaCl (Hall & Doty, 1959). The final specific activity in cts/min/pg was 20,000 for 12 s RNA, 18,000 for 16s RNA and 9,000 for 4s RNA. While the small difference in specific activity between 12 s and 16 s RNA may be due to a slight contamination of the 16 s component by 18 s ribosomal RNA, the lower specific activity of the 4 s RNA is probably the reflection of its metabolic stability as contrasted with a certain instability of the 16 s and 12 s RNA (Zylber, Vesco & Penman, 1969; Attardi 6 Attardi, 1971). For the determination of the DNA saturation level in the hybridization experiments described below, a specific activity value of 20,000 cts/min/pg was used for both the 16 s and 12 s components. Furthermore, it was assumed that the approximate 20% of the mitochondria-associated 4 s RNA which pertains to ribosomes of contaminating endoplasmic reticulum not removed by EDTA (Zylber & Penman, 1969; Attardi & Attardi, 1971), has the same specific activity as the mitochondrial components; in fact, direct determinations showed that the 4 s RNA released from the endoplasmic reticulum by the EDTA treatment has a similar specific activity (corresponding to more than 80% of that of mitochondria-associated 4 s RNA). In the present experiments, separated strands of [14C]thymidine-labeled mitochondrial DNA, banded in an alkaline CsCl density gradient as detailed previously (Aloni t Attardi, 1971), were utilized. The specific activity of the heavy and light mitochondrial DNA strands was determined by measuring their optical densities at 260 nm (by using the equivalence 48 ,ag/ml. = 1.0 O.D.,,, unit, derived from a value of E(p), molar extinction coefficient with respect to phosphorus, of 6,500, as reported for native calf thymus DNA at 22°C in 0.01 iv-Nacl (Shack, 1958), and a correction factor for for 29 hr at 26,000 rev./min at 20°C. (a) The portions of the gradient correspondingto the 16 s and 12 s components(indicated by arrows) were pooled, collected by ethanol precipitation and centrifugation and run through two oycles (c) or three cycles (b) of sucrose gradient centrifugation under the same conditions described above. In the inserts the sedimentation profiles of the final preparations run in the presence of l*C-labeled 18 8 ribosomal RNA marker are shown. (d) The components sdimenting in the 2 to 7 s region of the gradient in the pattern shown in (a) were collected by ethanol precipitation and centrifugation and run through a polyacrylamide gel (10% acrylamide) at 6 mu for 3.6 hr, as described before (Attardi t Attardi, 1971). The portions of the patterns shown in (b), (c) and (d) indicated by arrows were collected by ethanol precipitation and centrifugation, subjeoted to DNase digestion, sodium dodecyl sulfatephenol extraction and Sephadex chromatography, and used for RNA-DNA hybridization experiments (Figs 2 and 3). -O--O-, O.D.; -a-@--, sH radioaotivity.
274
Y.
ALONI
AND
G. ATTARDI
hyperchromic effect of 40% (Nass, 1969)), and by counting in the scintillation counter portions directly plated on nitrocellulose membranes. The specific activity in cts/min/ pg was 1,400 and 1,100 for the heavy and light mitochondrial DNA strand, respectively. This difference in specific activity reflects the difference in thymine content of the two strands (Aloni & Attardi, 1971). As shown in Figure 2, the saturation curves of the heavy DNA strand by 12 s RNA and 16 s RNA and by a mixture of equal weight amounts of these components rise fairly steeply and level off at an input RNA to DNA ratio of 4 to 6. The slight upward slope of the fmal portion of the saturation curves presumably reflects the presence of heterogeneous RNA contaminants. The actual saturation level was estimated by extrapolating the near-to-horizontal portion of the saturation curves to the axis of the ordinates (Fig. 2). The 12 s and 16 s RNA appear to saturate the heavy strand of
FIa. 2. Hybridization of the heavy and light strands of “W-labeled HeLa mitoohondrial DNA with increasing amounts of purified 3H-labeled 16 s and 12 s components. Samples of 0.025 pg of the heavy strand or 0.2 pg of the light strand of 14C-labeled mitochondrial DNA were immobilized on nitrocellulose membranes and incubated with increasing amounts of 3H-labeled 16 s and 12 s RNA components, purified as explained in the legend of Fig. 1, in 2 ml. of 4 x SSC (SSC is 0.15 M-NaCl, 0.015 M-sodium citrate) containing 0.01 M-Tris buffer, pH 7.8, for 24 hr at 68°C. The nitrocellulose membranes were washed and treated with pancreatic ribonuclease (Gillespie & Spiegelman, 1965). The data are corrected for non-specific background, as determined with SV40 DNA (13 to 20% of the hybrid values at saturation for the heavy strand, about 50% for the light strand). Other experiments involving only a few points and using 0.2 pg of mitochondrial DNA heavy strand gave hybridization values for 16 s and 12 s RNA which fell very close to the saturation curves shown here (the background in these experiments was about 3% of the values obtained with the heavy strand).
mitochondrial DNA at levels of about 9% and 13%, respectively. These levels correspond fairly closely to those expected if there were one gene for each of these RN/4 species per mitochondrial DNA molecule, assuming a molecular weight of about 3.5 x lo5 for 12 s RNA and about 5.5 x lo5 for 16 s RNA (as estimated by electron microscopy, see above), and assuming that HeLa mitochondrial DNA in homogeneous, as suggested for chick and guinea pig mitochondrial DNA by reassociation kinetics studies (Borst, 1969). The DNA saturation level of 18% obtained with combined 12 s and 16 s RNA, though not corresponding to that expected if the hybridization with these two species were strictly additive, does indicate that the 16 s RXA contains sequences different from those of 12 s RNA. This result, together with the kinetic evidence of lack of precursor-to-product relationship between the two species (Attardi $ Attardi, 1971), indicates that the 16 s and 12 s cistrons are distinct. Further experiments are needed to answer the question whether the lack of additivity is the result of some sequences being in common between the two RNA species or, as appears more
LETTERS
TO
THE
EDITOR
275
likely, of the presence in the 12 s preparation of degradation products of 16 s RNA. The saturation level, somewhat higher than expected, obtained with 12 s RNA is in agreement with the last mentioned possibility. It should be mentioned that mitochondrial DNA from XenoplLs Levis has been found to be saturated by whole high molecular weight mitochondrial RK,4 at a level su ggesting the presence of one gene per mitochondrial DNA molecule for each of the two major RNA species (Dawid, 1969). The levels of saturation of the light strand of mitochondrial DNA by the 12 s and 16 s RNA components were about 0.6 and l*Oo/o, respectively (Fig. 2). These levels are much lower than those corresponding to a single gene for either of these RNA species, and, therefore, they are presumably due to a small amount of contaminating heavy strand in the light st,rand preparation. In fact, this level of hybridization with the light strand was observed previously with total mitochondrial RNA and could be reduced considerably by using light strand purified through two cycles of alkaline CsCl density gradient centrifugation (Aloni & Attardi, 1971). As shown in Figure 3, mitochondria-associated 4 s RNA gave, with the heavy mitochondrial DNA strand, a saturation level of about 4%, corresponding to eight genes for molecules of 25,000 average molecular weight per mitochondrial DNA molecule, and, with the light mitochondrial DNA strand, a saturation level of about lG5o/o, which would correspond to three genes for molecules of the same molecular weight. The much lower level of saturation observed here with 4 s RNA, as compared to t$ose obtained with 16 s or 12 s RNA or with total mitochondrial RNA (Aloni & Attardi, 1971), and the difference in the shape of t’he saturation curves tend to exclude t,he presence in the 4 s RNA preparation of appreciable amounts of degradation products of t,he high molecular weight mitochondrial RNA. It is interesting to notice that the relative rates of labeling of the 12 s and 4 s RNA components had suggested that about ten molecules of 4 s RNA are synthesized per 12 s molecule (Attardi & Attardi, 1971); this would indeed imply the existence of about ten cistrons for 4 s RNA for each 12 s cistron if mitochondrial DNA is transcribed in a co-ordinate fashion. IP .N z4 1
I ,‘ 11 .----
H -e--
.-.
,” 3-:f’ 2 0 2t” $
‘1
,‘ 3,
1sl /*-•-•-----• I
0
/ IO
L
/
20
/
30
/
40
hFUt RNA/DNA
FIG. 3. Hybridization of the heavy and light strands of r4C-labeled HeLa mitochondrial DNA with increasing amounts of purified sH-labeled mitochondria-associated 4 s RNA. Portions of 0.2 pg of the heavy strand or of the light strand of ‘W-labeled mitochondrial DNA were immobilized on nitrocellulose membranes and incubated with increasing amounts of 3Hlabeled 4 s RNA, purified as described in the legend of Fig. 1, under the conditions detailed in Fig. 2. The nitrocellulose membranes were washed and treated with pancreatic ribonuclease. The data are corrected for non-specific background (about 15% of the hybrid values at saturation for the heavy strand, about 30% for the light strand). The hybridization values for the light strand have also been corrected for the small presumptive contribution of contaminating heavy strands (about 4%), as estimated from the hybridization data for 16 s and 12 s RNA with the light strand (Fig. 2).
Y. ALONI
276
AND
G. ATTARDI
RNA-DNA hybridization experiments utilizing mitochondrial DNA from Xenopus Levis oocytes and mitochondria-associated 4 s RNA have given results suggesting the presence of about 12 tRNA genes per mitochondrial DNA molecule (Dawid, 1969). Furthermore, Nass & Buck (1970) have recently reported that two mitochondrialspecific tRNA species hybridize with the heavy strand of rat liver mitochondrial DNA and two different ones with the light strand. The results presented here are in agreement with these findings. In view of the evidence previously reported (Aloni & Attardi, 1971) that almost all transcription of mitochondrial DNA in HeLa cells occurs from the heavy strand, probably in the form of a continuous long RNA chain, the very short stretch of the light strand which is apparently transcribed may correspond to the initiation or termination region of the transcription of the circular mitochondrial DNA duplex. These investigations were supported by a grant from the National Institutes of Health (GM-11726) and were undertaken during the tenure of a Research Training Fellowship awarded by the International Agency for Research on Cancer to one of us (Y. A., on leave from the Weizmann Institute of Science, Rehovot, Israel). Actinomycin D was a gift of Merck, Sharpe & Dohme. The excellent help of Mrs LaVerne Wenzel and Mrs Benneta Keeley is gratefully acknowledged. Division of Biology California Institute of Technology
YOSEF ALONI GIUSEPPE ATTARDI
Pasadena, Calif. 91109, U.S.A. Received 2 September
1970
REFERENCES Aloni, Y. & Attardi, G. (1971). J. Mol. BioZ. 55, 251. Attardi, B. & Attardi, G. (1969). Nature, 224, 1079. Attardi, B. & Attardi, G. (1971). J. Mol. BioZ. 55, 231. Attardi, G., Aloni, Y., Attardi, B., Ojela, D., Pica-Mattocia, L., Robberson, D. L. & Storrie, B. (1970). Cold Spr. Harb. Synarp. Qua&. Biol. 35, in the press. Attardi, G. & Ojala, D. (1971). Nature, in the press. Borst, P. (1969). In International Symposium on Autonomy and Biogenesis of Mitochondria and Chloroplasta, Canberra. Amsterdam: North-Holland Publishing Co. in the press. Borst, P. & Aaij, C. (1989). Biochem. Biophys. Ree. Comm. 34, 205. Dawid, G. B. (1969). In Symp. Sot. Exptl Biol. 24, 227. Dubin, D. T. (1967). Biochem. Biophyys. Res. Comm. 29, 655. Dubin, D. T. & Montenecourt, B. S. (1970). J. Mol. BioZ. 48, 279. Gillespie, D. & Spiegelman, S. (1965). J. Mol. BioZ. 12, 829. Hall, B. D. & Doty, P. (1959). J. Mol. BioZ. 1, 111. Ness, M. M. K. (1969). J. Mol. Biol. 42, 521. Nass, M. M. K. & Buck, C. A. (1969). Proc. Nat. Acud. Sci., Wash. 62, 506. Nass, M. M. K. & Buck, C. A. (1970). J. Mol. Biol. 54, 187. Penman, S., Vesco, C. & Penman, M. (1968). J. Mol. BioZ. 34, 49. Perry, R. P. (1964). Nat. Cancer Inst. Monogrqh, 14, 73. Shack, J. (1958). J. Biol. Chem. 233, 677. Swanson, R. F. & Dawid, I. B. (1970). Proc. Nat. Acud. Sci., Wash. 66, 117. Vesco, C. & Penman, S. (1969). Proc. Nat. Acud. Sci., W&L 62, 218. Zylber, E., Vesco, C. & Penman, S. (1969). J, Mol. BioZ. 44, 195. Zylber, E. & Penman, S. (1969). J. Mol. BioZ. 46, 201.
LETTERS TO THE EDITOR
Expression of the Mitochondrial IV.? Titration
of Mitochondrial
Genome in HeLa Cells
Genes for 16 s, 12 s and 4 s RNA
RNA-DNA hybridization experiments utilizing separated strands of HeLa mitochondrial DNA and purified 16 s, 12 s and 4 s mitochondrial RNA species from HeLa cells have indicated the existence of one 16 s gene and one 12 s gene per mitochondrial DNA molecule, which are located in the heavy mitochondrial DNA strand, and of approximately eleven genes for 4 s RNA: of these about eight appear to be located in the heavy strand and about three in the light strand.
We have shown previously that mitochondrial RNA is synthesized in HeLa cells, as in rat liver cells (Borst & Aaij, 1969), almost exclusively on the heavy strand of mitochondrial DNA, and, moreover, that the whole or almost whole length of this strand is transcribed (Aloni & Attardi, 1971). RNA-DNA hybridization experiments (Attardi & Attardi, 1969; Dawid, 1969; Nass & Buck, 1969,1970), kinetics and inhibition studies (Vesco & Penman, 1969; Zylber & Penman, 1969 ; Dubin & Montenecourt, 1970 ; Attardi C%Attardi, 1971) have revealed that the bulk RNA synthesized on mitochondrial DNA in HeLa and other animal cells is represented by discrete species with sedimentation coefficients of 4 S, 12 s and 16 S. In addit’ion, there are heterogeneous RNA components spread in the 4 to 50 s region and minor discrete fast sedimenting components (Aloni & Attardi, 1971; Attardi et al., 1970). In this letter we present the results of RNA-DNA hybridization experiments which strongly suggest that there is one gene for 12 s RNA and one gene for 16 s RNA per mitochondrial DNA molecule, which are located in the heavy mitochondrial DNA strand, and about 11 genes for mitochondria-associated 4 s RNA, 8 of which are located in the heavy strand and 3 in the light mitochondrial DNA strand. The methods for preparation of the mitochondrial fraction, extraction and analysis of RNA and RNA-DNA hybridization have been described in detail in previous reports (Aloni & Attardi, 1971; Attardi & Attardi, 1969,1971). Figure 1 shows the sedimentation pattern of RNA extracted from the EDTAtreated mitochondrial fraction of HeLa cells exposed for four hours to [S3H]uridine in the presence of 0.04 pglml. actinomycin D (to inhibit the synthesis of ribosomal RNA (Perry, 1964; Dubin, 1967; Penman, Vesco & Penman, 1968)). One can see that the 28 s and 18 s RNA are not labeled, while the 16 s, 12 s and 4 s peaks stand out clearly. The ratio of label in the 16 s and 12 s RNA in this experiment was about 1.5. On the basis of the molecular weight estimates derived from the electrophoretic mobilities of the two RNA components (0.7 x lo6 and 0.4 x 106) (Attardi & Attardi, 1971), of the relative sedimentation velocities of the two species after denaturation with formaldehyde, which are consistent with a molecular weight ratio of about 1.7 (Attardi & Attardi, 1970), and of electron microscope length measurements which indicat’e a molecular weight of about 3.5 x lo5 for the 12 s RNA and 55 x lo5 for the 16 s RNS (Robberson, Davidson, Aloni & Attardi, manuscript in preparation), it is reasonable t Paper III in this series is Attardi 8: Ojala (1971). 271
Y. ALONI
272
AND
G. ATTARDI
3.010 2.0106 1 I.O(1 0.8 0 0.6 0 04 04 I 02 0 06 0 1
04 0:
02 0 I
. o4 : 0.8 Ot
04
4
Cd) w c 4 I i
:
I
I .
-I
I t
-0 -0
:
t.-#4r, 0
; 4
IO
20
30
40 Frocticn
50
’ 60
70
no
FIU. 1. Purification of the 16 s, 12 s and 4 s RNA components from the mitoohondrial fraction of HeLa cells labeled for 4 hr with [5-3H]uridine in the presence of actinomyoin D. RNA was extracted from the EDTA-treated crude mitochondrial fraction isolated by differential centrifugation from the homogenate of 8 x lo* HeLa cells labeled with [6-3H]uridine for 4 hr in the presence of 0.04 pg/ml. actinomycin D. The RNA was run through a 15 to 30% (w/w) sucrose gradient in sodium dodecyl sulfate buffer (0.57, sodium dodecyl sulfate, 0.1 aa-NaGl, 0.01 aa-Tris buffer pH 7.0, 0.001 M-EDTA) in the SW27 Spinco rotor (1.69 cm x 10.16 cm buckets)
LETTERS
TO
THE
EDITOR
273
to assume that the two RNA species are synthesized in equimolar amounts. The recent demonstration that these species represent the structural components of specific 60 s mitochondrial ribosomes is in agreement with this assumption (Swanson & Dawid, 1970; Attardi & Ojala, 1971). After rerunning in sucrose gradient, the 16 s RNA (Fig. l(b)) and the 12 s RNA (Fig. l(c)) appear as fairly sharp peaks of coinciding 3H radioactivity and o.~.se~. The 4 s RNA was separated from other slow sedimenting components by acrylamide gel electrophoresis (Fig. l(d)). In order to eliminate any DNA contaminant, the three discrete RNA species were subjected to extensive DNase treatment, sodium dodecyl sulfate-phenol extraction and Sephadex GlOO chromatography, as previously described (Aloni & Attardi, 1971). For the determination of the specific activity, a portion of the Sephadex eluate corresponding to the peak of radioactivity was plated directly on a nitrocellulose membrane of the same type used for the RNA-DNA hybridization experiments and counted in the scintillation counter. The concentration of RNA solutions was determined from o.D.,,,, measurements by using a value of 214 for E$? at 260 nm in 0.1 M-NaCl (Hall & Doty, 1959). The final specific activity in cts/min/pg was 20,000 for 12 s RNA, 18,000 for 16s RNA and 9,000 for 4s RNA. While the small difference in specific activity between 12 s and 16 s RNA may be due to a slight contamination of the 16 s component by 18 s ribosomal RNA, the lower specific activity of the 4 s RNA is probably the reflection of its metabolic stability as contrasted with a certain instability of the 16 s and 12 s RNA (Zylber, Vesco & Penman, 1969; Attardi 6 Attardi, 1971). For the determination of the DNA saturation level in the hybridization experiments described below, a specific activity value of 20,000 cts/min/pg was used for both the 16 s and 12 s components. Furthermore, it was assumed that the approximate 20% of the mitochondria-associated 4 s RNA which pertains to ribosomes of contaminating endoplasmic reticulum not removed by EDTA (Zylber & Penman, 1969; Attardi & Attardi, 1971), has the same specific activity as the mitochondrial components; in fact, direct determinations showed that the 4 s RNA released from the endoplasmic reticulum by the EDTA treatment has a similar specific activity (corresponding to more than 80% of that of mitochondria-associated 4 s RNA). In the present experiments, separated strands of [14C]thymidine-labeled mitochondrial DNA, banded in an alkaline CsCl density gradient as detailed previously (Aloni t Attardi, 1971), were utilized. The specific activity of the heavy and light mitochondrial DNA strands was determined by measuring their optical densities at 260 nm (by using the equivalence 48 ,ag/ml. = 1.0 O.D.,,, unit, derived from a value of E(p), molar extinction coefficient with respect to phosphorus, of 6,500, as reported for native calf thymus DNA at 22°C in 0.01 iv-Nacl (Shack, 1958), and a correction factor for for 29 hr at 26,000 rev./min at 20°C. (a) The portions of the gradient correspondingto the 16 s and 12 s components(indicated by arrows) were pooled, collected by ethanol precipitation and centrifugation and run through two oycles (c) or three cycles (b) of sucrose gradient centrifugation under the same conditions described above. In the inserts the sedimentation profiles of the final preparations run in the presence of l*C-labeled 18 8 ribosomal RNA marker are shown. (d) The components sdimenting in the 2 to 7 s region of the gradient in the pattern shown in (a) were collected by ethanol precipitation and centrifugation and run through a polyacrylamide gel (10% acrylamide) at 6 mu for 3.6 hr, as described before (Attardi t Attardi, 1971). The portions of the patterns shown in (b), (c) and (d) indicated by arrows were collected by ethanol precipitation and centrifugation, subjeoted to DNase digestion, sodium dodecyl sulfatephenol extraction and Sephadex chromatography, and used for RNA-DNA hybridization experiments (Figs 2 and 3). -O--O-, O.D.; -a-@--, sH radioaotivity.
274
Y.
ALONI
AND
G. ATTARDI
hyperchromic effect of 40% (Nass, 1969)), and by counting in the scintillation counter portions directly plated on nitrocellulose membranes. The specific activity in cts/min/ pg was 1,400 and 1,100 for the heavy and light mitochondrial DNA strand, respectively. This difference in specific activity reflects the difference in thymine content of the two strands (Aloni & Attardi, 1971). As shown in Figure 2, the saturation curves of the heavy DNA strand by 12 s RNA and 16 s RNA and by a mixture of equal weight amounts of these components rise fairly steeply and level off at an input RNA to DNA ratio of 4 to 6. The slight upward slope of the fmal portion of the saturation curves presumably reflects the presence of heterogeneous RNA contaminants. The actual saturation level was estimated by extrapolating the near-to-horizontal portion of the saturation curves to the axis of the ordinates (Fig. 2). The 12 s and 16 s RNA appear to saturate the heavy strand of
FIa. 2. Hybridization of the heavy and light strands of “W-labeled HeLa mitoohondrial DNA with increasing amounts of purified 3H-labeled 16 s and 12 s components. Samples of 0.025 pg of the heavy strand or 0.2 pg of the light strand of 14C-labeled mitochondrial DNA were immobilized on nitrocellulose membranes and incubated with increasing amounts of 3H-labeled 16 s and 12 s RNA components, purified as explained in the legend of Fig. 1, in 2 ml. of 4 x SSC (SSC is 0.15 M-NaCl, 0.015 M-sodium citrate) containing 0.01 M-Tris buffer, pH 7.8, for 24 hr at 68°C. The nitrocellulose membranes were washed and treated with pancreatic ribonuclease (Gillespie & Spiegelman, 1965). The data are corrected for non-specific background, as determined with SV40 DNA (13 to 20% of the hybrid values at saturation for the heavy strand, about 50% for the light strand). Other experiments involving only a few points and using 0.2 pg of mitochondrial DNA heavy strand gave hybridization values for 16 s and 12 s RNA which fell very close to the saturation curves shown here (the background in these experiments was about 3% of the values obtained with the heavy strand).
mitochondrial DNA at levels of about 9% and 13%, respectively. These levels correspond fairly closely to those expected if there were one gene for each of these RN/4 species per mitochondrial DNA molecule, assuming a molecular weight of about 3.5 x lo5 for 12 s RNA and about 5.5 x lo5 for 16 s RNA (as estimated by electron microscopy, see above), and assuming that HeLa mitochondrial DNA in homogeneous, as suggested for chick and guinea pig mitochondrial DNA by reassociation kinetics studies (Borst, 1969). The DNA saturation level of 18% obtained with combined 12 s and 16 s RNA, though not corresponding to that expected if the hybridization with these two species were strictly additive, does indicate that the 16 s RXA contains sequences different from those of 12 s RNA. This result, together with the kinetic evidence of lack of precursor-to-product relationship between the two species (Attardi $ Attardi, 1971), indicates that the 16 s and 12 s cistrons are distinct. Further experiments are needed to answer the question whether the lack of additivity is the result of some sequences being in common between the two RNA species or, as appears more
LETTERS
TO
THE
EDITOR
275
likely, of the presence in the 12 s preparation of degradation products of 16 s RNA. The saturation level, somewhat higher than expected, obtained with 12 s RNA is in agreement with the last mentioned possibility. It should be mentioned that mitochondrial DNA from XenoplLs Levis has been found to be saturated by whole high molecular weight mitochondrial RK,4 at a level su ggesting the presence of one gene per mitochondrial DNA molecule for each of the two major RNA species (Dawid, 1969). The levels of saturation of the light strand of mitochondrial DNA by the 12 s and 16 s RNA components were about 0.6 and l*Oo/o, respectively (Fig. 2). These levels are much lower than those corresponding to a single gene for either of these RNA species, and, therefore, they are presumably due to a small amount of contaminating heavy strand in the light st,rand preparation. In fact, this level of hybridization with the light strand was observed previously with total mitochondrial RNA and could be reduced considerably by using light strand purified through two cycles of alkaline CsCl density gradient centrifugation (Aloni & Attardi, 1971). As shown in Figure 3, mitochondria-associated 4 s RNA gave, with the heavy mitochondrial DNA strand, a saturation level of about 4%, corresponding to eight genes for molecules of 25,000 average molecular weight per mitochondrial DNA molecule, and, with the light mitochondrial DNA strand, a saturation level of about lG5o/o, which would correspond to three genes for molecules of the same molecular weight. The much lower level of saturation observed here with 4 s RNA, as compared to t$ose obtained with 16 s or 12 s RNA or with total mitochondrial RNA (Aloni & Attardi, 1971), and the difference in the shape of t’he saturation curves tend to exclude t,he presence in the 4 s RNA preparation of appreciable amounts of degradation products of t,he high molecular weight mitochondrial RNA. It is interesting to notice that the relative rates of labeling of the 12 s and 4 s RNA components had suggested that about ten molecules of 4 s RNA are synthesized per 12 s molecule (Attardi & Attardi, 1971); this would indeed imply the existence of about ten cistrons for 4 s RNA for each 12 s cistron if mitochondrial DNA is transcribed in a co-ordinate fashion. IP .N z4 1
I ,‘ 11 .----
H -e--
.-.
,” 3-:f’ 2 0 2t” $
‘1
,‘ 3,
1sl /*-•-•-----• I
0
/ IO
L
/
20
/
30
/
40
hFUt RNA/DNA
FIG. 3. Hybridization of the heavy and light strands of r4C-labeled HeLa mitochondrial DNA with increasing amounts of purified sH-labeled mitochondria-associated 4 s RNA. Portions of 0.2 pg of the heavy strand or of the light strand of ‘W-labeled mitochondrial DNA were immobilized on nitrocellulose membranes and incubated with increasing amounts of 3Hlabeled 4 s RNA, purified as described in the legend of Fig. 1, under the conditions detailed in Fig. 2. The nitrocellulose membranes were washed and treated with pancreatic ribonuclease. The data are corrected for non-specific background (about 15% of the hybrid values at saturation for the heavy strand, about 30% for the light strand). The hybridization values for the light strand have also been corrected for the small presumptive contribution of contaminating heavy strands (about 4%), as estimated from the hybridization data for 16 s and 12 s RNA with the light strand (Fig. 2).
Y. ALONI
276
AND
G. ATTARDI
RNA-DNA hybridization experiments utilizing mitochondrial DNA from Xenopus Levis oocytes and mitochondria-associated 4 s RNA have given results suggesting the presence of about 12 tRNA genes per mitochondrial DNA molecule (Dawid, 1969). Furthermore, Nass & Buck (1970) have recently reported that two mitochondrialspecific tRNA species hybridize with the heavy strand of rat liver mitochondrial DNA and two different ones with the light strand. The results presented here are in agreement with these findings. In view of the evidence previously reported (Aloni & Attardi, 1971) that almost all transcription of mitochondrial DNA in HeLa cells occurs from the heavy strand, probably in the form of a continuous long RNA chain, the very short stretch of the light strand which is apparently transcribed may correspond to the initiation or termination region of the transcription of the circular mitochondrial DNA duplex. These investigations were supported by a grant from the National Institutes of Health (GM-11726) and were undertaken during the tenure of a Research Training Fellowship awarded by the International Agency for Research on Cancer to one of us (Y. A., on leave from the Weizmann Institute of Science, Rehovot, Israel). Actinomycin D was a gift of Merck, Sharpe & Dohme. The excellent help of Mrs LaVerne Wenzel and Mrs Benneta Keeley is gratefully acknowledged. Division of Biology California Institute of Technology
YOSEF ALONI GIUSEPPE ATTARDI
Pasadena, Calif. 91109, U.S.A. Received 2 September
1970
REFERENCES Aloni, Y. & Attardi, G. (1971). J. Mol. BioZ. 55, 251. Attardi, B. & Attardi, G. (1969). Nature, 224, 1079. Attardi, B. & Attardi, G. (1971). J. Mol. BioZ. 55, 231. Attardi, G., Aloni, Y., Attardi, B., Ojela, D., Pica-Mattocia, L., Robberson, D. L. & Storrie, B. (1970). Cold Spr. Harb. Synarp. Qua&. Biol. 35, in the press. Attardi, G. & Ojala, D. (1971). Nature, in the press. Borst, P. (1969). In International Symposium on Autonomy and Biogenesis of Mitochondria and Chloroplasta, Canberra. Amsterdam: North-Holland Publishing Co. in the press. Borst, P. & Aaij, C. (1989). Biochem. Biophys. Ree. Comm. 34, 205. Dawid, G. B. (1969). In Symp. Sot. Exptl Biol. 24, 227. Dubin, D. T. (1967). Biochem. Biophyys. Res. Comm. 29, 655. Dubin, D. T. & Montenecourt, B. S. (1970). J. Mol. BioZ. 48, 279. Gillespie, D. & Spiegelman, S. (1965). J. Mol. BioZ. 12, 829. Hall, B. D. & Doty, P. (1959). J. Mol. BioZ. 1, 111. Ness, M. M. K. (1969). J. Mol. Biol. 42, 521. Nass, M. M. K. & Buck, C. A. (1969). Proc. Nat. Acud. Sci., Wash. 62, 506. Nass, M. M. K. & Buck, C. A. (1970). J. Mol. Biol. 54, 187. Penman, S., Vesco, C. & Penman, M. (1968). J. Mol. BioZ. 34, 49. Perry, R. P. (1964). Nat. Cancer Inst. Monogrqh, 14, 73. Shack, J. (1958). J. Biol. Chem. 233, 677. Swanson, R. F. & Dawid, I. B. (1970). Proc. Nat. Acud. Sci., Wash. 66, 117. Vesco, C. & Penman, S. (1969). Proc. Nat. Acud. Sci., W&L 62, 218. Zylber, E., Vesco, C. & Penman, S. (1969). J, Mol. BioZ. 44, 195. Zylber, E. & Penman, S. (1969). J. Mol. BioZ. 46, 201.