J. Mol. Biol. (1974) 83, 369-378
Heterogeneity of RNA Polymerase in Escherichia c&i II. Polyadenylate .Polyuridylate
Synthesis by Holoenzyme II
YOICH~O IWAKURA, RYUJI PUKUDA AND AILIRA ISEUXAMA Department of Biochemistry Institute for Virus Research Kyoto University Kyoto 606, Japan (Received 20 Awgnst 1973, and in revised form 15 November 1973) Of the two RNA polymerases of Escherichia coli, only holoenzyme II, consisting of the core enzyme and the 0’ subunit, was found to catalyze unprimed polyadenylate*polyuridylate synthesis from ATP and UTP, whereas holoenzyme I was incapable of synthesizing any polyribonucleotides in the absence of DNA. The isolated a’ subunit alone was inactive in this reaction but converted the core enzyme derived from holoenzyme I into holoenzyme II or poly(A) .poly(U) polymerase. Unlike holoenzyme I, holoenzyme II was also active in polyadenylate synthesis even on double-stranded DNA. The 0’ subunit therefore appears to have additional functions besides the stimulation of transcription at certain DNA templates.
1. Introduction Besides the DNA-dependent synthesis of RNA or transcription of genetic messages, RNA polymerase (ribonucleosidetriphosphate : RNA nucleotidyltransferase (DNAdependent) (EC 188.8.131.52)) is known to be capable of catalysing polyadenylate-polyuridylate synthesis in the absence of DNA (Mehrotra & Khorana, 1965; Gometos et al., 1964). The reaction requires ATl? and UTP as substrates, manganese as a sole divalent metal cofactor and a larger quantity of enzyme than that needed for the DNA-dependent synthesis of RNA. This system shows unique kinetics; a burst of poly(A)*poly(U) synthesis is observed after a long lag period (Schafer & Cramer, 1970). This reaction has been considered to be a side reaction carried out by RNA polymerase in unusual reaction mixtures, and to be useful for analysis of the complex process involving several consecutive steps in the enzymic synthesis of RNA. However, it has been left unexplained why the ratio of DNA-dependent RNA synthesis to unprimed poly(A) *poly(U) synthesis varies with different enzyme preparations (Smith et al., 1967; Schafer & Cramer, 1970). Recently, we found two forms of RNA polymerase holoenzyme in Escherichia coli, each containing different species of the (T factor (Fukuda et al., 1974). Of the two holoenzymes, only holoenzyme II, consisting of the core enzyme and the newly discovered (T’subunit, was found to catalyse poly(A) *poly(U) synthesis in the absence of DNA. This is the first indication of functional difference between the two holoenzymes. 25
2. Materials and Methods Materials and Methods have been described in the preceding paper (Fukuda et al., 1974). Except where noted, poly(A) *poly(U) synthesis was assayed in the standard reaction mixture which contained in 0.25 ml: 30 pm01 Tris*HCl buffer (pH 7.8 at 37°C) ; 1 pmol manganese sulphate; 1.25 pmol 2-mercaptoethanol; 100 nmol each of UTP and [3H]ATP. The reaction was carried out at 37% and the incorporation of AMP into an acid-insoluble form was then determined as in the standard RNA polymerase assay (Fukuda et al., 1974).
3. Results (a) IsoEatioB
of holoenzyme II by phosphocelldose column chromatography
During purification of DNA-dependent RNA polyrnerase, polynucleotide phosphorylase (RNA: orthophosphate nucleotidyltransferase (EC 184.108.40.206)) was known HoloenzymeU ;: xi i:
:: :: ::
Fraction no. c, -0
Sodium dodecyl sulphate gel electrophoresis
1. Reohromatography of holoenzyme II on a phosphocellulose column. An RNA polymerase preparation obtained as desoribed in the preceding paper (Fukuda et al., 1974) was fractionated by chromatography on a phosphocellulose column, and holoenzyme II fractions, which were eluted slightly before the peak of core enzyme and contained the o’ subunit, were combined and dialysed against 0.05 an-Tris*HCI buffer (pH 7.8 at 4°C) containing 0.1 mMEDTA, O-1 mnn-dithiothreitol and 5% glycerol. The dialysed enzyme (9 mg protein) was applied on a column (0.9 cm x 10 cm) of phosphocellulose previously equilibrated with the same buffer and eluted with 160 ml of a linear gradient (0 M to 0.6 M) of KCl. Fractions of 3 ml were collected every 45 min. RNA polymerase activity (--+----) of 0.005-ml portions was measured accord(0.25 ml) ing to the standard reaction procedures (Fukuda et al., 1974): the reaction mixture contained 30 pm01 Tris.HCl buffer (pH 7.8 at 37’C); 1.25 -01 magnesium acetate; 0.5 pmol manganese sulphate; 1.25 pmol 2-mercaptoethanol; 40 nmol each of the 3 unlabelled ribonucleoside triphosphates; 40 nmol [“HIATP (4300 cts/min/nmol) and 1.7 pg poly[d(A-T)]copolymer. The reaction was carried out at 37°C for 10 min. Poly(A).poly(U) synthesis (--x--x--) was measured for O.O&ml portions in the reaction mixture, which contained in 0.25 ml: 30 pmol Tris.HCl buffer (pH 7.8 at 37°C); 0.5 pmol manganese sulphate; l-25 pmol 2-mercaptoethanol; 100 nmol each of UTP and [3H]ATP (4300 ots/min/nmol). The reaction was carried out at 37°C for 2 h. --O-O--, Absorbance at 280 nm. FIG.
to accompany closely (McConnell & Bonner, 1972) and to disturb analysis of unprimed RNA synthesis by the polymerase. However, the contaminant could be removed by means of phosphocellulose column chromatography. It did not bind to the column and was recovered in the flow-through fraction, whereas the G subunit and the core polymerase were eluted at about 0.1 M and 0.3 M-KC& respectively and, in addition, there could be observed a peak of weak polymerase activity (less than 10% of whole activity) just before the core enzyme. When analysed on sodium dodecyl sulphate/polyacrylamide gel electrophoresis, the subunit a’, as well as the core enzyme subunit, was found in this minor peak; moreover, the activity of unprimed poly(A) *poly(U) synthesis was detected only in this peak and not in the core enzyme peak or in the undissociated holoenzyme eluted between the (3 subunit and the enzyme containing the u’ subunit. The enzyme fractions containing 0’ subunit were combined and rechromatographed for further purification. As can be seen in Figure 1, the enzyme containing the CT’subunit was almost completely separated from the core enzyme and poly(A) *poly(U) synthesis was found only in this peak and not in the core enzyme. In contrast, the activity for poly[d(A-T)] copolymer-dependent poly(A-U) copolymer synthesis was detected in either of the two peaks. Subunit analysis revealed that the first peak represents RNA polymerase holoenzyme II. In agreement with this result, the enzyme was found to be active in native phage T7 DNA-dependent RNA synthesis, like holoenzymes I and II isolated by DNA cellulose column (Fukuda et al., 1974). The core enzyme was virtually inactive on this DNA. Thus it became clear that chromatography on phosphocellulose is also useful for isolation of holoenzyme II.
Time (mini (a)
Enzyme (units) (b)
Pm. 2. Poly(A) +poly(U) synthesis by holoenzyme II. Holoenzyme II was purified as shown in Fig. 1 and the enzyme activity as determined in the standard reaotion mixture for RNA synthesis with phage T7 DNA as tempkte was 6020 units/mg protein, whereas that of the original enzyme preparation was 8100 units/mg protein. One unit of enzyme activity is defined as that amount of protein which catalyzes the incorporation of 1 nmol of labelled ATP into RNA in 60 min at 37°C under standard reaction conditions. (a) Various amounts of the enzyme (units) (-O-O---) were incubated at 37°C for the indicated times in the standard poly(A).poly(U) reaction mixture. [3H]ATP (10,600 ots/min/nmol) was used as a labelled substrate. As a reference, poly(A).poly(U) synthesizing activity of the unfractioned original RNA polymerase (--e--e--) was also measured. (b) To show the linear relationship between the activity of poly(A).poly(U) synthesis and the enzyme protein used, the data as obtained were replotted.
of holoenzyme II a.spoZy(A) -poly( U) polymerase
To determine the activity of holoenzyme II in unprimed poly(A) *poly(U) synthesis, increasing amounts of the enzyme were incubated in the standard reaction mixture for poly(A) +poly(U) synthesis, which contained [3H]ATP and UTP as substrates and manganese as divalent metal. In agreement with previous observations (Smith et al., 1967; Schafer & Cramer, 1970), the reaction starts after a long lag and slows down as substrates are used up (Fig. 2(a)). After 120 minutes incubation with 35 units polymerase, more than 60 nmol of [3H]ATP (60% of the input ATP) were incorporated into the poly(A)*poly(U) product. It can be calculated from the data shown in Figure 2(b), that one unit of holoenzyme II, that allows incorporation of one nmol of [3H]ATP into RNA in 60 minutes with excess phage T7 DNA, catalysed incorporation of more than 2 nmol [3H]ATP into in 60 minutes reaction. Thus holoenzyme II appears to permit the poly(A).poly(U) synthesis of approximately equimolar amounts of poly(A) *poly(U) and RNA, under the conditions used, 15+ (a) PolyfALpolyW
” s I
(b) RNA .--------
0 _____ _a----------
P’ subunit (ml) I I
Q subunit (ml)
Fra. 3. Stimuletion of poly(A)*poly(U) synthesis by the d subunit. The core enzyme and the CIsubunit were isolated by phosphocellulose column chromatography es shown in Fig. 1, whilst the U’ subunit was purified by glycerol gradient centrifugation following DEAR-Sephadex A50 column ohrometography in 6 M-urea, es reported in the preceding p&per (Fukuda et al., 1974). (a) To 5.0 pg core polymerese, indicated amounts of the purified (r (--e--e--; 26 pg/ml) or u’ subunit (--O-O--; 75 pg/ml) were added and the activity of unprimed poly(A)*poly(U) synthesis was determined in 0.50 ml standard reaction mixture. The reaction w&s carried out at 37°C for 120 min. (b) Stimulation of RNA synthesis by the D and U’ subunits wes also measured in the stsnderd reaction mixture for RNA synthesis with 1-O pg core enzyme and 7 .ag intact phsge T7 DNA.
On the other hand, the original preparation of RNA polymerase before passing through the phosphocellulose column, exhibited very low activity of poly(A) *poly(U) synthesis; 15 units of enzyme gave no detectable product, and 154 units polymerase were needed to make 05 nmol of poly(A) -poly(U) in 60 minutes, which could be synthesized by less than seven units of purified holoenzyme II. The ratio (less than 5%) of the two enzyme preparations required for synthesis of an equimolar amount of poly(A) *poly(U) is in good agreement with the content of holoenzyme II in the original enzyme preparation, supporting the previous interpretation that only holoenzyyme II catalyses the synthesis of poly(A)*poly(U). Antisera against holoenzyme, subunits ,l3and ,!l’, effectively blocked poly(A) *poly(U) synthesis by holoenzyme II, suggesting that not only the a’ subunit but also the core unit in holoenzyme II is actually engaged in this reaction. In support of this contention, rifampicin and streptolydigin, which are known to inhibit RNA synthesis by binding to the /? subunit of RNA polymerase (Heil & Zillig, 1970; Iwakura et al., 1973), inhibited unprimed poly(A) *poly(U) synthesis. Since anti-o subunit serum did not inhibit the activity of unprimed poly(A) -poly(U) synthesis by RNA polymerase holoenzyme II, it appears clear that the u subunit is not involved in this reaction. In order to prove the hypothesis that the (r’ subunit renders the core enzyme active in unprimed synthesis of poly(A) *poly(U), enzyme reconstitution experiments were done. The (T’ subunit was isolated from holoenzyme II as described in the preceding paper (Fukuda et al., 1974), while the core enzyme and the u subunit were prepared by passing holoenzyme I through a phosphocellulose column. Neither u nor 0’ subunits could catalyze poly(A) .poly(U) or poly(A) synthesis by themselves, but both subunits were active in stimulating the core enzyme to transcribe T7 DNA to the same extent (Fig. 3(b)). In contrast, the core enzyme derived from holoenzyme I, which was absolutely inactive in poly(A) *poly(U) synthesis, became active only when the (T’ subunit was provided (Fig. 3(a)). Thus, it is clear that poly(A) *poly(U) synthesis is catalyzed only by RNA polymerase holoenzyme II. (c) Characteristics of poZy(A) -poly( U) synthesis by holoenzyme II Contrary to DNA-dependent RNA synthesis by holoenzymes I and II, the presence of manganese was found to be essential for synthesis of poly(A) *poly(U) by purified holoenzyme II. The maximal activity was achieved at 4 mM-manganese, whereas magnesium was totally inactive (Fig. 4). The absolute requirement of manganese was reported for poly(A) *poly(U) synthesis by RNA polymerase preparations (Schafer & Cramer, 1970; Krakow et al., 1968) suggesting that the entity involved might have been holoenzyme II. The remarkable effect of ionic strength was also observed on poly(A) *poly(U) synthesis by holoenzyme II. In DNA-dependent RNA synthesis (So et al., 1967), it is known that the addition of salt stimulates the reaction several-fold and the reaction proceeds for many hours. However, as shown in Figure 5, increasing concentrations of KC1 were inhibitory rather than stimulatory on poly(A) *poly(U) synthesis, and no detectable synthesis was observed at concentrations of KCl higher than OS15 M.
The reaction requires both ATP and UTP as substrates, and no detectable formation of homopolymer could be observed when ATP, UTP, GTP or CTP was added as a sole substrate (Fig. 6(a)). In contrast, addition of GTP and/or CTP
Metal ion concn tmM)
Fm. 4. Influence of metal ion concentration. Holoenzyme I was reconstituted from core enzyme and an excess amount of (I subunit, whereas holoenzyme II was isolated as described in Fig. 1. Activities of 1.25 pg holoenzyme I and 2.90 pg holoenzyme II were measured in standard assay conditions for the synthesis of RNA (37”C, 15 was varied as min) or poly(A).poly(U) (37”C, 120 min), except that the metal ion concentration indicated. [3H]ATP (4300 cts/min/nmol) was used as a labelled substrate. -a---a--. Mga+ alone added to the assay mixture; --O--O--, Mne+ alone added to the assay mixture.
FIG. 5. Influence of salt concentration. Holoenzymes I and II were prepared as in Fig. 4 and activities of poly(A)-poly(U) synthesis and T7 DNA-direoted RNA synthesis were measured in standard assay mixtures, except that the indicated concentrations of KC1 were added. The reaction W&S carried out at 37% for 10 min (RNA synthesis) or 120 min (poly(A)=poly(U) synthesis). --O-O-, Holoenzyme I (10 units)/ holoenzyme II (6 units)/RNA synthesis; --O--O--, holoenzyme RNA synthesis ; --e-e--, II (4 units)/poly(A) *poly(U) synthesis.
FIG. 6. Influence of substrate combination on poly(A) .poly(U) synthesis. Isolation of holoenzyme II and determination of enzyme activity were as described in Fig. 1. (a) Homopolymer synthesis was carried out at 37°C for the indioated periods. The reaction mixture (0.25 ml) contained 30 q101 Tris-HCI buffer (pH 7.8 at 37’C); 1.25 pmol 2-mercaptoethanol; 0.5 pmol manganese sulphate (2 mm); 4 units of holoenzyme II; and 100 nmol of one of the following labelled substrates: [“HIATP, 4300 cts/min/nmol; [l*C]GTP, 11,000 cts/min/nmol; [3H]UTP, 3500 cts/min/nmol; [l*C]CTP, 5000 cts/min/nmol. Poly(A) *poly(U) synthesis was assayed in the same reaction mixture, except that 100 nmol each of [3H]ATP and m&belled UTP were added as substrates. (b) E3H]ATP incorporation at the indicated incubation times at 37°C was determined with 4 units of holoenzyme II in the reaction mixture as in (a), except that Mn2+ concentration was increased to 4 mu and unlabelled substrates were added as indicated. No incorporation was detected when [l*C]GTP or [14C]CTP was added as a labelled substrate instead of [sH]ATP.
inhibited poly(A) *poly(U) synthesis (Fig. 6(b)) and the product of such reduced incorporation was found to be poly(A) . poly(U) ; no polynucleotide containing GMP or CMP was detected in the absence of DNA. (d) DNA-dependent poZy(A) synthesis by holoewyme II The above results demonstrated that in the absence of DNA only poly(A) *poly(U) can be synthesized by the action of holoenzyme II. On the other hand, it has been reported that RNA polymerase of E. coli catalyses poly(A) synthesis in the presence of DNA, presumably by slipping of transcribed poly(A) chains on thymidylate residues in template DNA (Chamberlin & Berg, 1964; Stevens, 1964). As expected from this mechanism, not only holoenzyme II, but also holoenzyme I and core enzyme were also found to be capable of synthesizing poly(A) in the presence of single-stranded DNA (Fig. 7(a)). The reaction starts without any lag periods and continues linearly fox at least t,wo hours. Single-stranded DNA of either circular (phage fd DNA) or linear (denatured T7 DNA) forms is effective and a more favourable template for both enzymes. Since poly[d(A-T)] copolymer failed to prime poly(A) synthesis by either holoenzyme,
iii4 6 ‘i *Cl 3-
.---- ZENY ------.
Time (mid FIG. 7. DNA-dependent poly(A)-. synthesis. -_-. _ Holoenzymes I and II were prepared as described in Fig. 4. UNA from phage fd w&8 kindly supplied by Dr M. Takanami. (a) Poly(A) synthesis was measured with 5.0 1-18of holoenzyme I (--O--O-) or 2.9 pg of holoenzyme II ([email protected]
) in a reaction mixture which oontained in O-25 ml: 30 ~01 Tris*HCl (pH 7.8 at 37°C); 1.25 pm01 2-mercaptoethanol; 0.25 pm01 manganese sulphate (1 mM); 100 nmol of [3H]ATP (10,600 cts/min/nmol); and 7 pg fd DNA. Core enzyme was as active as holoenzyme I. (b) Poly(A) synthesis was assayed as in (a), except that the template was replaced with 7 pg
T7 DNA. Core enzyme was virtually
inactive on T7 DNA.
(c) Enzyme activities of the same amounts of holoeneymes I and II as used for the experiments described above were determined in the standard reaction mixture for RNA synthesis with 7 /.cg T7 DNA. Unprimed poly(A)*poly(U) synthesis was measured with twice as much of the enzyme.
residues appears to be necessary, as suggested by others (Stevens, 1964). In accord with this observation, addition of substrates other than ATP inhibited the reaction, presumably by preventing slippage of product poly(A). Although holoenzymes I and II exhibited similar activities in RNA synthesis directed by double or single-stranded DNA, and in single-stranded DNA-dependent poly(A) synthesis, a marked difference between the two enzymes was observed when double-stranded DNA was used as template for poly(A) synthesis. Holoenzyme I could hardly synthesize poly(A) on double-stranded DNA (Fig. 7(b)). a sequence of thymidylate
4. Discussion The present report together with the preceding paper (Fukuda et al., 1974) presented evidence that E. coli contains two species of RNA polymerase holoenzyme. The newly discovered holoenzyme II has the following unique properties: (i) the enzyme consists of the standard core subunits (a, 8, ,E’) and an additional polypeptide, designated u’, with a molecular weight of 56,000; (ii) both the native and reconstituted holoenzyme II exhibited transcription of native double-stranded DNA just as standard holoenzyme I, indicating that the a’ subunit has a function similar to the a subunit; (iii) however, only holoenzyme II can catalyse poly(A) -poly(U) synthesis in the absence of DNA; and in addition (iv) holoenzyme II synthesizes poly(A) even on double-stranded DNA. Several possible mechanisms have been proposed for unprimed poly(A) . poly(U) synthesis by RNA polymerase preparations. Polynucleotide phosphorylase catalyzes poly(A-U) random copolymer synthesis from ADP and UDP contaminants in ATP and UTP substrates, and the product may multiply by the action of RNA polymerase. However, holoenzyme II was shown to be free from polynucleotide phosphorylase; no polynucleotide other than poly(A) -poly(U) is synthesized and only manganese is active as cofactor, whereas magnesium is active in the polynucleotide phosphorylase reaction. Moreover, inorganic phosphate, a potent inhibitor of the polynucleotide phosphorylase reaction, gave no effect on poly(A) *poly(U) synthesis. Ohasa & Tsugita (1972) reported that the a subunit of RNA polymerase exhibited poly(A) synthesizing activity. Then, product poly(A) may direct poly(U) synthesis by RNA polymerase followed by regeneration of poly(A). However, no detectable poly(A) synthesis has so far been detected with our purified holoenzymes I and II under the reported reaction conditions (Ohasa & Tsugita, 1972). Thus, in poly(A) poly(U) synthesis by holoenzyme II, poly(A) and poly(U) chains might be synthesized simultaneously depending on the complementary strands. Schafer et al. (1970) proposed that contaminated DNA fragments serve as primer for poly(A) -poly(U) synthesis by RNA polymerase. The ratio of 1.77 for optical density at 280 nm versus 260 nm of holoenzyme II appears to rule out contaminations of nucleic acids; moreover, the CT’subunit, isolated by DEAE-Sephadex column chromatography of polymerase dissociated with 6 M-urea and free from nucleic acids, was still capable of converting core enzyme to poly(A).poly(U) polymerase. Thus, it can be concluded that holoenzyme II itself catalyses poly(A) *poly(U) synthesis in the absence of DNA and without the help of polynucleotide phosphorylase or poly(A) polymerase. Variable activities of poly(A)+poly(U) synthesis of RNA polymerase preparations may therefore reflect variable contents of holoenzyme II in those preparations. Krakow et al. (1970) proposed that during initiation of RNA synthesis the cr subunit binds in the region of product site in the catalytic unit of DNA-bound polymerase and fixes the substrate to be incorporated into the 5’ end of the product. In accordance with this proposal, the affinity of the polymerase to substrates increases once the enzyme binds to template DNA (Ishihama $ Hurwitz, 1969). From this viewpoint, it seems reasonable to suppose that the 0’ subunit is highly reactive with substrates even when the polymerase is not bound to DNA, and thus allows the enzyme to initiate poly(A) .poly(U) synthesis without DNA.
Since only holoenzyme II synthesizes poly(A) even on double-stranded DNA, it appears that the u’ subunit is also more active in the proposed function (Ishihama et al., 1971; Hinkle & Chamberlin, 1971) of enforcing polymerase to activate or unwind local regions of DNA. Such unique functions so far observed suggest that in cells holoenzyme II plays a role different from holoenzyme I. Further attempts are being made in our laboratory to investigate if different sequences in DNA are recognized by the two holoenzymes. The authors are grateful to Dr K. Ito and Mr S. Naito for advice and support received throughout this work, and to Dr T. Yura for help in preparation of the manuscript. Part of this work was supported by grants from the Ministry of Education of Japan and the Matsunaga Science Foundation. REFERENCES Chamberlm, M. & Berg, P. (1964). J. Mol. Biol. 8, 708-726. Fukuda, R., Iwakura, Y. & Ishihama, A. (1974). J. Mol. Biol. 83, 353-367. Gomatos, P. J., Krug, R. M. & Tamm, I. (1964). J. Mol. Biol. 9, 193-207. Heil, A. & Zillig, W. (1970). PEBS Letters, 11, 165-167. Hinkle, D. C. & Chamberlin, M. (1971). Cold Spring Harbor Symp. Quant. Biol. 35, 65-72. Ishihama, A. & Hurwitz, J. (1969). J. Biol. Chem. 244, 6680-6689. Ishihama, A., Murakami, S., Fukuda, R., Matsukage, A. & Kameyama, T. (1971). Mol. Gen. Genet. 111, 66-77. Iwakura, Y., Ishihama, A. & Yura, T. (1973). Mol. Gen. Genet. 121, 181-196. Krakow, J. S. & von der Helm, K. (1970). Cold Spritig Harbor Symp. Quant. Biol. 35,73-83. Krakow, J. S., Daley, K. & Fronk, E. (1968). Biochem. Biophys. Res. Commun. 32, 98-100. McConnell, D. J. & Banner, J. (1972). Biochemistry, 11, 4329-4336. Mehrotra, B. B. & Khorana, H. G. (1965). J. BioZ. Chem. 240, 1750-1753. Ohasa, S. & Tsugita, A. (1972). Nature New Biol. 240, 35-38. Schafer, K. P. & Cramer, F. (1970). Eur. J. Biochem. 15, 103-110. Smith, D. A., Ratliff, R. L., Williams, D. L. & Martinez, A. M. (1967). J. Biol. Chem. 242, 590495.
So, A. G., Davie, E. W., Epstein, R. & Tissi&es, A. (1967). Proc. Nat. Acad. Sk., 58, 1739-1746. Stevens, A. (1964). J. Biol. Ohem. 239, 204-209.